Reactive oxygen species (ROS) are intriguing molecules, which are toxic to the biological structures but also play a signaling role in controlling plant growth, development and stress responses (for a recent review, see Suzuki et al., 2012; Baxter et al., 2014). They accompany the basal metabolic fluxes of aerobic organisms during the whole ontogeny. Considering the precise and multilevel control of metabolic changes it might not be surprising that ROS coming from the various cellular and extracellular sites seem to have their specific signaling targets (op den Camp et al., 2003; Avsian-Kretchmer et al., 2004; Gadjev et al., 2006; Geisler et al., 2006; Laloi et al., 2007). In plants several ROS, such as H₂O₂, O2•– and ¹O₂, are formed in chloroplasts aside the photosynthetic electron transport. They, in turn, affect the nuclear gene expression to adjust the photosynthesis to changing environment. The most stable form, which is assumed to leave this organelle and evoke the effects on the nuclear genes, is H₂O₂ (Mubarakshina et al., 2010; Borisova et al., 2012).

In an attempt to recognize the signaling effects of H₂O₂ originating from chloroplasts, Arabidopsis mutant overexpressing glycolate oxidase in chloroplasts (GO5 plants) has been developed (Fahnenstich et al., 2008). In the photorespiratory conditions GO5 mutants produce H₂O₂ in chloroplasts instead of in peroxisomes. Transcript profiling of GO5 plants (Balazadeh et al., 2012) identified the H₂O₂ regulated genes and transcription factors, whereas further work of Sewelam et al. (2014) underpinned the top 20 genes specifically up-regulated by H₂O₂ produced in chloroplasts. These studies proved that H₂O₂ can trigger different responses depending on the subcellular site of its production.

Recently, we demonstrated that thylakoids isolated from a highly stress resistant species Eutrema salsugineum (Thellungiella salsuginea) are capable of the enhanced production of H₂O₂, in comparison to Arabidopsis thaliana, already in the absence of stress (Wiciarz et al., 2015). This creates the opportunity to unravel the signaling action of chloroplast H₂O₂ generation in the natural system. E. salsugineum tolerates extreme salinity, cold, drought, ozone and over the last years this species became a plant model of stress resistance well-comparable with the Arabidopsis genome (Inan et al., 2004; Li et al., 2006; Amtmann, 2009; Hou and Bartels, 2015). So far, several studies focused on the discovery of transcriptomic footprints of high stress resistance in E. salsugineum. These studies showed that several stress-associated genes in E. salsugineum have a constitutively higher expression in comparison with A. thaliana already in the absence of stress (Inan et al., 2004; Taji et al., 2004; Gong et al., 2005). The group of up-regulated genes includes, for example, those involved in abscisic acid (ABA) biosynthesis and signaling (Taji et al., 2004; Gong et al., 2005). In contrast, after stress treatment, only a slight change in gene expression was detected in E. salsugineum in comparison with a great activation of transcription in A. thaliana plants (Taji et al., 2004; Li et al., 2006; Wong et al., 2006). Also, a comparative proteomics of A. thaliana and E. salsugineum salt responses revealed more changes in protein abundance in Arabidopsis than in Eutrema (Pang et al., 2010). Combined, these results indicate so called ‘stress preparedness’ of E. salsugineum, which supports its halophytic nature. Comparison of metabolite profiles of these two species following salt stress revealed a significant differences (Arbona et al., 2010). This suggests, that an adjustment of metabolism and activation of the already present enzymatic machinery serves as a faster and more efficient strategy to cope with stress than synthesis of new proteins.

A goal of this study was to verify whether enhanced generation of H₂O₂ in thylakoids of halophytic Eutrema in comparison to glycophytic A. thaliana is significant also in vivo. As a next objective, it was tested how this situation influences oxidative damage, antioxidant system and hormonal signaling in the control conditions and after a salinity stress.

In this work, we compared a halophytic E. salsugineum with glycophytic A. thaliana in regard to the chloroplastic H₂O₂ signature on the salt stress tolerance. As a first step, we verified whether a chloroplastic H₂O₂ signal, previously reported with use of isolated TMs (Wiciarz et al., 2015), persists in the relatively undamaged organelle and tissue.

Enhanced generation of H₂O₂ within the chloroplasts of control and stress-treated Eutrema was visualized by TEM, suggesting that a higher level of H₂O₂ persists in the whole chloroplasts. However, it is not uniformly distributed but spotted in the proximity of granal thylakoids, were it is produced. TEM images also revealed, that NaCl treatment caused thylakoid swelling only in A. thaliana. Thylakoid dilation is a well-documented symptom of salinity stress in plants (Niewiadomska et al., 2011; Yamane et al., 2012; Pottosin and Shabala, 2016), while structural integrity of E. salsugineum chloroplasts under mild NaCl-stress was previously shown by Chang et al. (2015). This lack of swelling might be attributed to the activity of thylakoid ion channels, and/or K⁺(Na⁺)/H⁺ antiporters which control the ion homeostasis under salinity-stress (Pottosin and Shabala, 2016).

Although we did not find any direct evidence for an export of H₂O₂ from plastids, the total H₂O₂ level in leaf extracts was also elevated in Eutrema leaves. Hence, it is likely that a chloroplast H₂O₂ generation contributes considerably to this phenomenon. Our speculation is based on a well-accepted phenomenon that illuminated chloroplast is a major site of ROS production in the mesophyll cells (Asada, 2006). Moreover, a considerable leakage of H₂O₂ from healthy chloroplasts has been shown experimentally (Mubarakshina et al., 2010; Borisova et al., 2012).

Here, performed study revealed that an increased generation of H₂O₂ in halophytic species is associated by a decreased generation of O2•–. Considering a much higher reactivity of O2•– than H₂O₂, this pattern of ROS production might shape a specific ROS signature important for so called stress-prepardeness of Eutrema. This speculation is supported by a much lower oxidative damage (MDA) in leaves of E. salsugineum than in A. thaliana. As argued by Tiwari et al. (2013), superoxide, but not H₂O₂, is required to evoke oxidative damage in photosynthetic membranes. This damage is mediated by the hydroxyl radical and is manifested by the appearance of carbon-centered radical. Experiments with the flu A. thaliana mutant and the flu overexpressing tAPX, enabled to demonstrate that H₂O₂ antagonizes the ¹O₂-mediated cell death response and growth inhibition, which might be particularly advantageous under stress (Laloi et al., 2007). A rapid conversion of a more harmful ROS (such as O2•–) into a longer-living and less toxic H₂O₂ has previously been concluded for halophytic plants on the basis of enhanced activity of superoxide dismutase (Bose et al., 2014). In halophytic E. salsugineum an operation of the protective electron cycling around PSII (Stepien and Johnson, 2009; Wiciarz et al., 2015) seems to support that view. Such a cycle could be beneficial for plants in several ways. Firstly, it may create a more safe electron sink in comparison to the production of superoxide and its toxic derivatives. Secondly, an abundance of H₂O₂ may play a signaling role.

A growing body of data indicate that H₂O₂ triggers defense responses in plant cells (for a recent review, see Suzuki et al., 2012; Baxter et al., 2014; Bose et al., 2014; Ismail et al., 2014). In this respect, ROS formation in chloroplast may also act indirectly via strengthening or amplifying retrograde redox signals (Galvez-Valdivieso and Mullineaux, 2010). One of the targets of ROS signaling route under stress is an up-regulation of antioxidants (Fryer et al., 2003; Geisler et al., 2006). However, the activity of a major H₂O₂ scavenger CAT was similar in E. salsugineum and A. thaliana in controls and in salinity-treated plants, whereas APX activity was even decreased in E. salsugineum. This is in agreement with numerous reports on salt cress. The transcript profiling analysis with two E. salsugineum ecotypes did not detect any antioxidant genes among those specifically up-regulated by salinity (Taji et al., 2004; Wong et al., 2006). Also, a proteomic data did not detect any salinity-dependent stimulation of antioxidant enzymes in this species (Pang et al., 2010). At activity level M’rah et al. (2006) previously noted that stimulation of CAT activity in E. salsugineum occurred only under a mild salinity treatment, not greater than 100 mM, whereas some stimulation was related to total peroxidase activity (assessed with guaiacol as electron donor).

Intriguingly, all these data suggest that a high stress resistance of Eutrema does not rely on the major enzymatic H₂O₂ scavengers of mesophyll cells. We argue, that data obtained with a partially purified (desalted) protein extracts (as used here) give a more precise information about changes in the enzymatic activities in comparison to the data typically obtained with crude extracts. The latter are strongly influenced by a different availability of low molecular compounds, such as ascorbate in case of APX, and glutathione in case of GPX. Indeed, an increased pool of reduced ascorbate and of total glutathione was found due to salinity in Eutrema, but not in Arabidopsis (Wiciarz et al., 2016). This feature might be important in the context of H₂O₂ signaling. As suggested by Bose et al. (2014), decreased activities of a major H₂O₂ scavengers allow for a persistence of ROS signal in the cells. A fine-tuning of specific antioxidants seems to be more important for stress tolerance than a bulk changes in the antioxidant status (Jiang et al., 2007; Pang et al., 2010). We are, however, aware that a relation between ROS production and scavenging is highly dynamic while a situation reported here for Eutrema relates only to certain time point. But, on the other hand, it was verified with use of several methods and plants sets to be considered as physiologically relevant.

Among the enzymatic H₂O₂ scavengers only GPX was enhanced in E. salsugineum. This points to the importance of redox active thiols. GPXs reduce not only H₂O₂ but also organic hydroperoxides using reducing power of GSH. Hence, they may contribute to the dramatically reduced MDA level. A noteworthy stimulation of antioxidant defense associated with glutathione has previously been revealed in E. salsugineum by transcript profiling of Gong et al. (2005). This was represented by GPXs (GPX2 and GPX6), glutaredoxins and glutathione-S-transferases (GSTZ1 and GSTU7). Four out of eight GPXs found in E. salsugineum have been shown to respond to salt and osmotic stresses at both transcript and protein level (Gao et al., 2014). It is worthy to note here, that GPX may serve as a specific sensor and transducer of H₂O₂ signal. A dual function of GPX associated with both scavenging of ROS and transferring the H₂O₂ signal has been pointed out repeatedly (Delaunay et al., 2002; Miao et al., 2006). Moreover, a specific activation of GPX1 promoter in tobacco by a chloroplast-generated H₂O₂ was documented by Avsian-Kretchmer et al. (2004). Whereas, study with Arabidopsis gpx3 mutant demonstrated that GPX3 participates in the ABA and H₂O₂ signaling pathway regulating the stomatal aperture and resistance to drought (Miao et al., 2006).

Considering an enhanced generation of H₂O₂ and changed O2•–/H₂O₂ ratio in chloroplasts, a major role might be attributed to the anti-oxidative protection within this organelle. Indeed, TMs of Eutrema appeared to be better protected against ROS in comparison with Arabidopsis, as indicated by a higher abundance of tAPX and FeSOD. This corresponds with enhanced level of FeSOD transcripts in the control E. salsugineum, reported by Taji et al. (2004). In regard to chloroplast stroma, we found an enhanced amount of PrxQ in E. salsugineum. These above-mentioned features of chloroplast antioxidant system are constitutive in E. salsugineum, which supports earlier conclusions on the stress preparedness of this species (Inan et al., 2004; Taji et al., 2004; Gong et al., 2005; Wong et al., 2006). They are also in line with the view that components of redox homeostasis in chloroplasts are crucial for salinity tolerance (Niewiadomska and Wiciarz, 2015; Uzilday et al., 2015). So far, recognized chloroplast redox players highly engaged in E. salsugineum are: adenosine 5′-phosphosulfate reductase involved in sulfate assimilation (Gong et al., 2005); thioredoxin CDSP32 (M’rah et al., 2007); 2-cys peroxiredoxin BAS1 (Gao et al., 2009) and plastid terminal oxidase (PTOX), an enzyme of chlororespiratory pathway (Stepien and Johnson, 2009; Wiciarz et al., 2015). In addition to these, also proteins associated with the Calvin-Benson cycle, such as a more active Rubisco (Wiciarz et al., 2015) and a more abundant glyceraldehyde 3-phosphate dehydrogenase (Pang et al., 2010; Chang et al., 2015) are likely to contribute to the improved redox balance in E. salsugineum by efficient recycling of ADP and NADP⁺.

A very weak response to salinity stress at the level of tested antioxidants may raise a question: whether E. salsugineum (and A. thaliana) senses the salinity stress at all? This has been clarified by analysis of stress hormones. A massive increase in stress-related hormones (ABA, SA, JA, and ethylene precursor) was detected in salinity-treated A. thaliana. These hormones are responsible for several stress effects in plants, among them stomatal closure and retardation of growth, although, their action is quite complex and may lead either to survival or to cell death. In contrast, in Eutrema the only indicator of salinity-evoked stress was a stimulation of ethylene synthesis. Although, some enhancement in ABA catabolite PA, may suggest a previous ABA induction. Earlier study of Ellouzi et al. (2014) documented that levels of ABA, JA, and ACC were increased in E. salsugineum after 2 days of salt stress, while 400 mM NaCl treatment for more than 5 days caused minimal changes in ABA and JA concentrations (Taji et al., 2004; Arbona et al., 2010). In E. salsugineum the levels of lipid-based active jasmonates (JA and JA-Ile) were low and even decreased due to salinity-treatment. This effect seems to be closely related to salt-resistance of Eutrema. As shown by Ismail et al. (2014), an interplay between ABA and JA is of particular importance for plant’s survival or death under salinity stress. A comparison of the two Vitis species differing in salt-stress tolerance revealed that in the more salt-tolerant Vitis rupestris ABA accumulated early and strongly suppressed the formation of JA-Ile, while in the salt-sensitive V. vinifera the accumulation of ABA was delayed and correlated which allow for accumulation of JA-Ile up to the high levels. A suppression of jasmonates in E. salsugineum might also be explained in accordance with the data obtained with AtGO5 mutant overproducing H₂O₂ in chloroplasts, in which a stimulated expression of several repressors of JA signaling pathway has occurred (Balazadeh et al., 2012). A very weak response to salinity at the level of ABA in E. salsugineum in contrast to A. thaliana is surprising in view of data on a higher expression of genes for ABA-biosynthesis and signaling (Taji et al., 2004; Gong et al., 2005; Wu et al., 2012) and an increased number of genes involved in ABA synthesis pathway in Eutrema genome (Wu et al., 2012). Possibly, in E. salsugineum the increased level of ABA is confined to the early stage of development or this hormone is very rapidly modulated in response to stress (indicated by slight increase in PA). A weak response of stress hormones to salinity in E. salsugineum supports the earlier conclusion of Hou and Bartels (2015), that comparing with A. thaliana the higher NaCl levels (600 mM) are required in E. salsugineum to trigger some defense reactions, such as increased transcription of two aldehyde dehydrogenase genes.

On the other hand, the importance of glucosinolates for stress-resistance of E. salsugineum is depicted by their enhanced level and a strong salinity-dependent increase. A great variation in the patterns of aliphatic, aromatic, and indole glucosinolate was documented in E. salsugineum, depending on the organ and developmental stage (Pang et al., 2012). However, a precise role of these compounds in stress tolerance is not clarified yet. An up-regulation of the synthesis of indole glucosinolates and phytoalexin camalexin was found as a specific feature of increased H₂O₂ signaling from chloroplasts in studies with AtGO5 mutant made by Balazadeh et al. (2012) and Sewelam et al. (2014). However, E. salsugineum does not produce camalexin, but wasalexins A and B and methoxybrassenin B (Pedras and Adio, 2008). Intriguingly, as another support for a constitutive stress-preparedness of Eutrema, the three phytoalexins (indolyl-3-acetonitrile, caulilexin C, and arvelexin) are found as phytoanticipins, i.e., they are constitutively produced in this species (Pedras and Adio, 2008).

In this paper we provided evidence for a specific pattern of ROS formation and accumulation in halophytic Eutrema in comparison to glycophytic Arabidopsis. In concert with that, a pronounced changes in the antioxidant system were revealed in E. salsugineum, such as a decreased activity of APX, an increased activity of GPX, an increased level of PrxQ, and the presence of thylakoid-bound forms of FeSOD and APX. All these cues allow for enhanced H₂O₂ signaling from chloroplasts of E. salsugineum already in the control conditions, at decreased level of oxidative damage in the same time. This signaling led to enhanced level of glucosinolates in Eutrema and decreased levels of stress hormones (SA and jasmonates). Salinity-stress evoked a strong up-regulation of all tested stress hormones in A. thaliana whereas in E. salsugineum only a stimulation in ethylene synthesis and ABA catabolism was noted. On this basis we hypothesize that H₂O₂ signaling is engaged in the halophytic species to modulate a hormonal responses in such a way to minimize the triggering of jasmonate pathway which leads to cell death.

MP performed analysis of antioxidants, staining and quantitation of H₂O₂, determination of glucosinolates, participated in TEM analysis and in preparation of the manuscript. MW measured MDA, participates in the analysis of antioxidants and ROS. IJ measured ROS with EPR. MK-K performed TEM analysis. PD and RV measured stress hormones, and RV participated in writing of the text. EN designed the experiment and wrote the manuscript. All authors have read and approved the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.