Calmodulin antagonist affects peroxisomal functionality by disrupting both peroxisomal Ca²⁺ and protein import

Ca²⁺ is a cellular second messenger whose homeostasis is highly regulated in cells through different components including Ca²⁺ pumps, channels, cation exchangers and Ca²⁺-binding proteins, since this cation regulates many cellular activities (Choi et al., 2012, 2014; Charpentier and Oldroyd, 2013; Cheval et al., 2013; Steinhorst and Kudla, 2013; Simeunovic et al., 2016). Moreover, crosstalk among Ca²⁺ and other signaling molecules, including H₂O₂ and nitric oxide (NO), which affects gene expression and Ca²⁺-dependent protein kinase, has also been described (Rentel and Knight, 2004; González et al., 2012; Han et al., 2014; Niu and Liao, 2016). Ca²⁺ signals can remain local or be propagated throughout an entire cell within milliseconds to seconds (Monshausen, 2012). Ca²⁺ has been detected in almost all cellular compartments including the cytosol, chloroplasts, mitochondria and nucleus. But, there can be a 10,000-fold difference between cytoplasmic and non-cytoplasmic Ca²⁺ concentrations, which allows the formation of Ca²⁺ signals by rapid alterations of cytoplasmic Ca²⁺ levels via membrane-localized Ca²⁺-permeable channels (Stael et al., 2012). However, information regarding Ca²⁺ in peroxisomes is scarce; to our knowledge, only one study has claimed that it is present in plants (Costa et al., 2010). Nevertheless, its potential function in peroxisomes has been analysed, as Ca²⁺ is necessary for key antioxidant peroxisomal enzymes, such as catalase (Yang and Poovaiah, 2002; Schmidt et al., 2006; Costa et al., 2013), and for peroxisomal NO generation (Corpas et al., 2009; Corpas and Barroso, 2014b).

Peroxisomes are single-membrane-bound subcellular compartments present in almost all types of eukaryotic cells. They contain catalase and H₂O₂-producing flavin oxidases as indispensable enzymatic elements (Reumann et al., 2007; Palma et al., 2009; Corpas, 2015). These organelles are characterized by metabolic flexibility, as their enzymatic content can vary according to the organism (i.e. animal, plant or yeast), cell or tissue type (i.e. liver, root or leaf), development stage and external environmental conditions (Mullen et al., 2001; del Río et al., 2002; Hayashi and Nishimura, 2006; Pracharoenwattana and Smith, 2008; Hu et al., 2012).

In plant cells, peroxisomes house a large number of antioxidative enzymes, such as catalase, superoxide dismutase, components of the ascorbate–glutathione cycle and several NADP dehydrogenases, which are involved in different functions (del Río et al., 2002; Fernández-Fernández and Corpas, 2016; Hölscher et al., 2016; Leterrier et al., 2016; Corpas et al., 2017). On the other hand, accumulating data have shown that plant peroxisomes have the capacity to generate NO through a L-arginine-dependent nitric oxide synthase (NOS) activity, which strictly depends on NADPH and requires calmodulin (CaM) and Ca²⁺ (Barroso et al., 1999; Corpas et al., 2004). Furthermore, peroxisomal NO, together with other related reactive nitrogen species (RNS) such as peroxynitrite, has been shown to participate in the response to abiotic stresses such as salinity (Corpas et al., 2009) and heavy metals such as cadmium or lead (Corpas and Barroso, 2014a, 2016). In addition, it has been demonstrated that the protein responsible for generating NO appears to be imported by a type 2 peroxisomal-targeting signal (PTS2) in a process that depends on the cytosolic receptor PEX7, and CaM and Ca²⁺ (Corpas and Barroso, 2014b).

With the aim of gaining a deeper understanding of cross-talk between Ca²⁺ and the peroxisomal NO metabolism, a pharmacological study was carried out using different NO donors, glutathione (GSH) and a CaM antagonist in Arabidopsis thaliana seedlings. The data show that Ca²⁺ is present in the peroxisomes of Arabidopsis stomata and root cells, and its presence is drastically disrupted by a CaM antagonist, which also blocks protein import into peroxisomes. Furthermore, the activity of typical peroxisomal enzymes that containing a type 1 PTS motif (PTS1), such as catalase, glycolate oxidase and hydroxypyruvate reductase, was negatively affected by the CaM antagonist, suggesting that CaM must also be necessary for plant peroxisome functionality.

Using Arabidopsis seedlings expressing CFP containing a PTS1 motif (CFP–PTS1), the potential endogenous presence of Ca²⁺ in peroxisomes was studied by using the fluorescence probe Fluo-3 AM. Although, theoretically, there is no overlap between the excitation and emission wavelengths of CFP with the specific fluorescent probe used to detect Ca²⁺ (Fluo-3 AM), potential overlap was evaluated at the experimental level. Fig. S1A,D shows in vivo confocal laser-scanning microscopy (CLSM) visualization of peroxisomes in the guard cells and root tips of transgenic Arabidopsis seedlings expressing CFP–PTS1 (excitation maximum at 458 nm and emission maximum at 475 nm). The peroxisomes appeared in the form of spherical spots in the cells. Fig. S1B,E, shows the same field observed by CLSM using an excitation wavelength of 514 nm and an emission wavelength of 610 nm but in the absence of the fluorescent probe (Fluo-3 AM) used to detect Ca²⁺ in which it is not possible to detect any fluorescence signal.

Fig. 1 shows in vivo CLSM visualization of peroxisomes, Ca²⁺ and chloroplasts in the guard and root cells of transgenic Arabidopsis seedlings expressing CFP–PTS1. Peroxisomes appeared in the form of spherical spots of a green color in guard and root cells (Fig. 1A,F, respectively). On the other hand, Fig. 1B,G show the same fields analyzed using Fluo-3 AM as fluorescence probe, which also enabled Ca²⁺ (red color) to be detected. The red fluorescence appeared in both types of spherical spots, with a pattern similar to that of CFP–PTS1, and also throughout the cells although with less intensity. Fig. 1C shows chlorophyll autofluorescence (purple color), enabling us to detect the localization of chloroplasts as spherical spots in guard cells; however, these were larger and in a different position in the cells from the spots corresponding to peroxisomes (Fig. 1A). Fig. 1D,H contains merged images of the overlap of the corresponding panels, showing a complete overlap of the punctate patterns of peroxisomes with the Ca²⁺ signals; in addition, the presence of the red color was localized throughout the cytosol, as well as being observed in other parts both inside and between the cells, for example, in the cell wall of roots cells (Fig. 1G) and colocalized with chloroplasts (see Fig. 2B,C). Fig. 1E,I shows the bright field of the guard and root cells, respectively. Apparently, peroxisomes have a high Ca²⁺ content in relation to rest of the cells, and this seems to be logical considering that the concentration of cytosolic Ca²⁺ usually is lower (around 100 nM) in comparison with the non-cytoplamisc Ca²⁺, which has a millimolar concentration (Monshausen et al., 2008; Stael et al., 2012). However, the Ca²⁺ signal has a wide distribution, such that it could be observed in images taken at a lower magnification. For example, Fig. S2 shows root cells taken at a lower magnification, illustrating the general distribution of Ca²⁺ signal using Fluo-3 AM as fluorescence probe.

Previous data have demonstrated that peroxisomal NO generation depends on Ca²⁺ (Corpas et al., 2009) and that the import of the peroxisomal NOS-like protein responsible for this NO generation also depends on Ca²⁺ and CaM (Corpas and Barroso, 2014b). With the aim of gaining a greater insight into the relationship between peroxisomal NO and Ca²⁺ metabolism, a pharmacological study was carried out using different NO donors (GSNO and DEA NONOate) and the CaM antagonist trifluoperazine (TFP). These analyses were performed in both the guard and root cells of transgenic Arabidopsis seedlings expressing CFP–PTS1 (Figs 2 and 3, respectively).

Fig. 2A–E corresponds to the untreated Arabidopsis seedlings in which it can be observed that Ca²⁺ clearly colocalized with peroxisomes and chloroplasts (Fig. 2D). Fig. 2F–J shows the effect of using S-nitrosoglutathione (GSNO) as an NO donor. It can be seen that this molecule significantly affects the import of CFP–PTS1, since the green signal appears to diffuse to all the guard cells, with only very few green punctuate spots being observed (Fig. 2F), and also leads to a sharp reduction in Ca²⁺ content (Fig. 2G). However, the chloroplasts do not appear to be affected (Fig. 2H). Figs 2I,J shows the merged image of the corresponding panels and the bright field, respectively. To corroborate this potential effect, another NO donor, DEA NONOate, was used. In this case, a smaller number of green spots corresponding to CFP–PTS1 content was observed (Fig. 2K) in comparison to the untreated sample (Fig. 2A); in addition, the red signal corresponding to the Ca²⁺ was diffused to the guard cells but showed an accumulation in the inner walls of the guard cells (Fig. 2L). In this regard, it must be mentioned that NO can have a wide spectrum of actions because it can directly or indirectly modulate post-translational modifications affecting the function of the target molecules (Frungillo et al., 2014; Yun et al., 2016; Begara-Morales et al., 2014, 2015, 2016; Mata-Pérez et al., 2016).

Taking in consideration that GSNO releases both NO and glutathione (GSH), the putative effect of GSH was also assessed as an internal control. We found that the pre-treatment of seedlings with 2 mM GSH prevents the import of CFP–PTS1 since the green fluorescence signal appears to diffuse through the guard cells (Fig. 2P); the red color corresponding to the Ca²⁺ signal was also diffused to the guard cells but with a strong accumulation in the inner walls of the guard cells (Fig. 2Q). On the other hand, Ca²⁺ in the chloroplasts seems to be less affected (Fig. 2R,S). Fig. 2T shows the corresponding bright field image. CaM is a Ca²⁺-binding protein involved in the signaling process. The chemical TFP, a CaM antagonist (Vandonselaar et al., 1994; Sengupta et al., 2007), was therefore used to evaluate its possible effect on both peroxisomal protein import and Ca²⁺. Fig. 2U shows that pretreatment with TFP significantly affects the import of CFP–PTS1 since the green fluorescence signal appears to diffuse to all the guard cells. On the other hand, Fig. 2V shows that it also lowers the Ca²⁺ signal (red color), with very few red punctuate signals being observed, which do not seem to correspond to peroxisomes. Ca²⁺ signal was observed in the chloroplasts (Fig. 2W). Fig. 2X,Y show the merged image of the corresponding panels and the bright field, respectively. Thus, these results collectively suggest that NO also affects both peroxisomal protein import and endogenous Ca²⁺, although the effect is less strong than that seen with GSH and TFP.

Fig. 3 illustrates an analysis of root cell similar to that carried out on guard cells. Fig. 3A,B shows peroxisomes and Ca²⁺ detected in untreated Arabidopsis roots, in which it can be observed that Ca²⁺ clearly colocalizes with peroxisomes (Fig. 3C,D). Fig. 3E–H shows the impact of GSNO. Thus, it can be observed that this molecule did not affect the import of CFP–PTS1, as peroxisomes appear in the form of green spherical spots (Fig. 3E); however, the red signal (corresponding to Ca²⁺) appears to diffuse through the cells and also appears in a smaller number of red spherical spots (Fig. 3F). Fig. 3G,H shows the merged image of the corresponding panels and the bright field, respectively. To corroborate this potential effect, another NO donor, DEA NONOate, was used. In this case, the green color was also observed to appear in green spherical spots and also with some diffuse green signals (Fig. 3I). However, the red fluorescence signal corresponding to the Ca²⁺ signal appears to be only very weakly present in spherical spots (Fig. 3J). Fig. 3K,L shows the merged image of the corresponding panels and the bright field, respectively. These effects of NO donors (GSNO and DEA NONOate) in roots are quite different to that observed in guard cells (Fig. 2F,K). This could be due to the fact that guard cells are specialized cells in the epidermis of leaves involved in stomata movement, and its regulation there is dependent on a complex interplay between NO, Ca²⁺, H₂O₂, abscisic acid (ABA), salicylic acid and Ca²⁺-dependent protein kinases (Desikan et al., 2004; Li et al., 2009; Zhang et al., 2009; Hao et al., 2010; Wang et al., 2014; Murata et al., 2015), where peroxisomes seem also to be involved (Leterrier et al., 2016). Therefore, we hypothesize that the application of exogenous NO could alter this equilibrium, with the guard cells more affected that the root cells.

As performed above, GSH was used as an internal control for GSNO. We found that the pre-treatment of seedlings with 2 mM GSH avoids the import of CFP–PTS1 since the green signal appears to diffuse through the cells, especially in the cells of the root cap (Fig. 3M); the red color corresponding to the Ca²⁺ signal was also diffused through the cells in the elongation area of the primary root (Fig. 3N). Fig. 3O,P shows the merged image of the corresponding panels and the bright field, respectively. Fig. 3Q shows that pretreatment with TFP (CaM antagonist) drastically affects the import of CFP–PTS1, with the green color appearing to diffuse in the cells of the root cap. On the other hand, Fig. 3R shows a very weak red signal, indicating that the CaM antagonist very significantly affects Ca²⁺ content in root cells. Fig. 3S,T shows the merged image of the corresponding panels and the bright field, respectively. Fig. 4 shows the change in the fluorescence intensity of Ca²⁺ (red color) in guard and root cells of Arabidopsis seedlings pretreated with GSNO, DEA NONOate, GSH and TFP which corresponds to results shown in Figs 2 and 3. In this sense, it can be seen that GSH is apparently the chemical that had a stronger effect in Ca²⁺ content and distribution in guard and root cells.

Although the chemical TFP acts as a CaM antagonist (Vandonselaar et al., 1994; Sengupta et al., 2007), it might also induce non-specific effects that could not be directly related to Ca²⁺/CaM metabolism, such as inhibition of electron transport and phosphorylation in plant mitochondria (Dunn et al., 1984), inhibition of Ca²⁺ ATPase, proton excretion or membrane properties (Lichtman et al., 1982; Barr et al., 1990; Wilson, 1994; Sengupta et al., 2007). Consequently, to affirm the effect of TFP as a CaM antagonist in our experimental Arabidopsis model, another CaM antagonist was used, in particular N-(6-aminohexyl)-5-chloro-1-naphthalenesulfonamide (W7), which has been previously described to have a similar effect to TFP in animal and plant cells (Sengupta et al., 2007; Ma et al., 2008). Fig. S3 illustrates the CLSM in vivo detection of Ca²⁺ (red signal) and peroxisomes (green signal) in root tip cells of 5-day-old transgenic Arabidopsis seedlings expressing CFP–PTS1 pre-incubated with either of the CaM antagonists TFP (Fig. S3D–F) or W7 (Fig. S3G–I). In both cases and, in comparison with the untreated seedlings (Fig. S3A,B), the green signal appears to be diffuse in the cells of the root cap with a very weak red signal, indicating that both CaM antagonists caused identical effects. On the other hand, pre-incubation with the chemical W5 (Fig. S3J– L), which is an inactive form of W7, shows that Ca²⁺ clearly colocalizes with peroxisomes, similar to the localization observed in untreated seedlings (Fig. S3A–C).

Additionally, to evaluate whether the CaM antagonist TFP could affect the cell viability, Arabidopsis seedlings were stained with propidium iodide (PI). This chemical is a cationic dye that does not readily cross intact membranes but that can penetrate throughout the meristem and bind to cell walls, thereby showing an outline of living cells; PI binds to nuclear structures, causing bright punctate fluorescence. Fig. S4 shows the PI staining of root cells of 5-day-old transgenic Arabidopsis seedling expressing CFP–PTS1 that were pre-incubated with TFP. It can be observed that the CaM antagonist (Fig. S4D–F) does not affect the cell viability when compared to untreated seedlings (Fig. S4A–D).

Given that the CaM antagonist TFP is the chemical with the most drastic effect on both peroxisomal protein import and Ca²⁺ content, we also performed in vitro assays with this CaM antagonist to measure the activity of representative peroxisomal enzymes, including catalase, glycolate oxidase and hydroxypyruvate reductase, using 14-day-old Arabidopsis seedlings pre-incubated with 300 µM TFP for 2 h at 25°C in darkness. Fig. 5A–C shows that these three activities were considerably affected, with a reduction of 45% for catalase, 41% for glycolate oxidase and 51% for hydroxypyruvate reductase. For the purposes of characterization, we also determined whether the CaM antagonist is capable of modifying the gene expression of these peroxisomal enzymes. The Arabidopsis genome contains three catalase genes (CAT1, CAT2 and CAT3) (Du et al., 2008), whose gene expression increased ∼2.0-, 1.4- and 1.5-fold, respectively, following pretreatment of Arabidopsis seedlings with TFP for 2 h (Fig. 5D). The CaM antagonist also causes an induction of HPR1 expression of ∼2.8-fold. Finally, glycolate oxidase is encoded by five genes, three of which, GOX1, GOX2 and GOX3, were selected for analysis in this study. While GOX1 and GOX3 expression seems to be unaffected, GOX2 expression increased 1.6-fold after pretreatment with the CaM antagonist (Fig. 5E).

Ca²⁺ is a highly important second messenger whose presence in different plant subcellular compartments, including the cytosol, chloroplasts, mitochondria, nucleus and vacuoles, has been well studied (Raychaudhury et al., 2006; Monshausen et al., 2008; Dodd et al., 2010; Kudla et al., 2010; Behera et al., 2015; Xu et al., 2015). However, information concerning Ca²⁺ in peroxisomes, particularly in plants, is scarce. In a study using human HeLa cells transiently expressing a peroxisome-targeted the GFP-based Ca²⁺ indicator D3cpv-SKL, the presence of Ca²⁺ in animal peroxisomes was shown for the first time, with the peroxisomal membrane representing a significant barrier to Ca²⁺ diffusion into peroxisomes (Drago et al., 2008). In the same year, Lasorsa et al. (2008) performed additional experiments showing that peroxisomes contain between 20- and 50-fold more Ca²⁺ than the cytosol (reaching a maximum of 100 µM). To our knowledge, only one study in the literature analyses Ca²⁺ homeostasis in plant peroxisomes; this study used the GFP-based Cameleon Ca²⁺ indicator D3cpv-KVK-SKL and demonstrated that peroxisomal Ca²⁺ affected catalase activity (Costa et al., 2010). The present study therefore aimed to gain a deeper understanding of the significance of Ca²⁺ in the peroxisomal metabolism of plants.

CaM is a protein of ∼17 kDa that is expressed in all eukaryotic cells and has the capacity to bind Ca²⁺ (Bouché et al., 2005). This protein, together with CaM-like (CML) proteins, is the primary Ca²⁺ sensor that controls the activity of various target proteins and consequently diverse cellular functions, including development processes and responses to adverse biotic and abiotic conditions (Yang and Poovaiah, 2003; Bouché et al., 2005; Reddy et al., 2011; Cheval et al., 2013; Zhao et al., 2015; Zeng et al., 2015). For instance, in the Arabidopsis genome, seven distinct CaM and 50 CMLs genes have been predicted (McCormack et al., 2005). Recently, Arabidopsis thaliana (At)CML3 (also known as AGD11) has been demonstrated to be targeted to peroxisomes (Chigri et al., 2012), where it appears to mediate the dimerization of peroxisomal processing protease AtDEG15 (Dolze et al., 2013). Moreover, there is another family of proteins which act as Ca²⁺ sensors called Ca²⁺-dependent protein kinases (CDPKs) (Boudsocq and Sheen, 2013) and in Arabidopsis the isoform AtCPK1 is targeted to peroxisomes (Dammann et al., 2003) and seems to mediate pathogen resistance (Coca and San Segundo, 2010).

Using an alternative fluorescent probe to detect Ca²⁺, the present study corroborates the presence of Ca²⁺ in the peroxisomes of two different tissues and cells – guard cells in green cotyledons and root cells – suggesting that Ca²⁺ must be necessary for the normal physiological functioning of plant peroxisomes, as has been shown for other subcellular compartments. Moreover, pharmacological techniques using different chemicals have enabled us to demonstrate that peroxisomal Ca²⁺ must be regulated by CaM and GSH and also, though less intensely, by NO. Thus, the CaM antagonist TFP blocks peroxisomal function due to its negative effect on peroxisomal protein import, as shown by CFP–PTS1 remaining in the cytosol. Additionally, the presence of Ca²⁺ in peroxisomes suggests that a CaM or CML protein is strictly necessary for the peroxisomal homeostasis of Ca²⁺, although, to our knowledge there is no experimental information about the potential effect of a CaM antagonist on the peroxisomal AtCML3 function. In any case, this is a reasonable assumption, since the function of CaM is to transport and regulate Ca²⁺ levels. Thus, these data are closely in line with previous findings that show that Ca²⁺ is required for peroxisomal NO production (Corpas et al., 2009) and that enzyme protein import responsible for its generation is blocked by CaM antagonist (Corpas and Barroso, 2014b).

Hitherto, there is clear evidence to support the cross-talk between NO and Ca²⁺/CaM in plant cells (Sang et al., 2008; Ma et al., 2012). However, in the present study, we find that NO appears to be less involved in controlling peroxisomal Ca²⁺ content; by contrast, peroxisomal Ca²⁺ is necessary and indispensable for the production of NO in peroxisomes. It has been demonstrated that the activity of peroxisomal protein L-arginine-dependent nitric oxide synthase (NOS) responsible for NO generation requires Ca²⁺ (Barroso et al., 1999; Corpas et al., 2004). More recently, it has been reported that the import of this protein into the peroxisome also depends on Ca²⁺ and CaM (Corpas et al., 2009; Corpas and Barroso, 2014b).

Reduced glutathione (GSH) constitutes an important antioxidant, and it is present in plant peroxisomes (Zechmann and Müller, 2010) where it is indispensable for the functioning of glutathione reductase (GR), a component of the ascorbate–glutathione cycle that functions in these organelles (Romero-Puertas et al., 2006; Reumann and Corpas, 2010). Under our experimental conditions, GSH was used as control during the analysis of the GSNO treatment; however, our findings were surprising as GSH, which caused a general delocalization of the Ca²⁺ signal and a marked intensification throughout the cells, drastically affected both peroxisomal protein import and the peroxisomal localization of Ca²⁺. This suggests that a change in the redox state or a possible S-glutathionylation process could be involved in the mechanism of peroxisomal protein import; the latter is regulated by a large number of cytosolic and peroxisomal proteins, known as peroxins (PEXs), which are encoded by at least 22 PEX genes in Arabidopsis (Nito et al., 2007; Brown and Baker, 2008; Hu et al., 2012).

As part of the potential physiological functionality of Ca²⁺, the CaM antagonist TFP was observed to negatively affect the activity of the antioxidant catalase enzyme, as well as the glycolate oxidase and hydroxypyruvate reductase enzymes, which are both involved in photorespiration (Rojas et al., 2012). The decrease in their activity could be due to the blocking of the peroxisomal import of the corresponding proteins containing a PTS1; this is closely in line with the results obtained, since the CaM antagonist TFP blocks the import of CFP–PTS1. However, a previous study has demonstrated that the activity of plant catalase is stimulated by both Ca²⁺ and CaM (Yang and Poovaiah, 2002; Afiyanti and Chen, 2014). Moreover, with the aid of a recombinant AtCat3 in a CaM-binding assay, the same authors demonstrated that CaM specifically binds to AtCat3 in the presence of Ca²⁺; protein sequence analysis of Cat3 specifically enabled them to identify a CaM-binding region of 37 amino acids between Gly415 and Val451 (Yang and Poovaiah, 2002). More recently, an increase in intraperoxisomal Ca²⁺, which occurs in response to adverse conditions accompanied by oxidative stress, has been shown to stimulate catalase activity (Costa et al., 2010). However, to our knowledge, there is no information available on any potential relationship between CaM/Ca²⁺ and glycolate oxidase and hydroxypyruvate reductase activity. Nevertheless, the results show that their activity was also adversely affected, with the blocking of their import into the peroxisomes being the most plausible cause. Thus, the gene expression of these enzyme groups either increased (CAT1 to CAT3, HPR1 and GOX2) or appeared to be unaffected (GOX1 and GOX3), which could be a way of compensating for the loss of these enzyme activities.

In summary, our findings indicate that the presence of Ca²⁺ is necessary for the physiological functioning of peroxisomes in root and guard cells, since its absence negatively affects the activity of key enzymes as well as the generation of NO, as has previously been demonstrated (Corpas et al., 2009). Moreover, we suggest that Ca²⁺/CaM is also necessary for the import of peroxisomal proteins containing a PTS1. However, a drastic alteration in the redox state and/or S-glutathionylation of peroxisomal protein import of peroxins could be involved. In this latter case, further analyses would be necessary to determine which peroxins are affected by this redox regulation.

Arabidopsis thaliana transgenic seeds expressing cyan fluorescent protein (CFP) containing a type 1 peroxisomal-targeting signal (PTS1) (Nelson et al., 2007) were grown on Petri plates for 5 days or 14 days under a cycle of 16 h light (under a light intensity of 100 µE m⁻² s⁻¹) at 22°C and 8 h dark at 18°C as previously described (Corpas and Barroso, 2014a). For microscope analyses, 5-day-old seedlings were grown on vertically positioned plates. For enzymatic and gene expression assays, 14-day-old seedlings were grown on horizontally positioned plates.

Ca²⁺ was detected with 5 µM Fluo-3 AM (Biotium) prepared in 10 mM Tris-HCl (pH 7.4) (Zhang et al., 1998; Kanchiswamy et al., 2010, 2014). The Arabidopsis seedlings were incubated in darkness with 5 µM Fluo-3 AM at 25°C for 1 h. Each sample was then washed twice in the same buffer for 15 min and mounted on a microscope slide for examination with a CLSM (Leica TCS SP5 II). For Fluo-3 AM, the excitation laser wavelength was 514 nm, with an emission of 610 nm and a 40-nm band pass width (590–630 nm). For CFP, the excitation laser wavelength was 458 nm, with an emission of 475 nm and a 40-nm band pass width (455–495 nm). Chlorophyll autofluorescence was studied as follows: chlorophyll a and b, excitation of 429 and 450 nm, respectively; emission of 650 and 670 nm, respectively.

Arabidopsis seedlings (5 days old) expressing CFP–PTS1 were pre-incubated for 120 min at 25°C in darkness with different chemicals prepared in 10 mM Tris-HCl (pH 7.4) including 2 mM diethylamine NONOate (DEA NoNoate) and 2 mM S-nitrosoglutathione (GSNO), which are both NO donors (Broniowska et al., 2013; Begara-Morales et al., 2014), and 2 mM glutathione (GSH) and 300 µM trifluoperazine (TFP, a calmodulin antagonist) (Vandonselaar et al., 1994). Additionally, seedlings were also pre-incubated with 200 µM N-(6-aminohexyl)-5-chloro-1-naphthalenesulfonamide (W7), which is another CaM antagonist (Ma et al., 2008), and 200 µM N-(6-aminohexyl)-1-naphthalenesulfonamide (W5), which is the inactive W7 structural analog (Li et al., 2004). The seedlings were then incubated with 5 µM Fluo-3 AM for 1 h at 25°C in darkness to detect and visualize Ca²⁺ using CLSM. In all cases, the images obtained by CLSM from control and treated Arabidopsis seedlings were kept constant during the course of the experiment in order to produce comparable data. For each treatment, at least three independent experiments were made using, at least, four or five seedlings in each experiment.

Plant cell integrity was evaluated with propidium iodide (PI) staining (Truernit and Haseloff, 2008). Arabidopsis seedlings were incubated for 5 min at room temperature in a 10 μg/ml PI solution (P4170 Sigma) in water. For PI, the excitation laser wavelength was 536 nm, with an emission of 617 nm and a 40-nm band pass width (597–637 nm).

Crude extracts of 14-day-old Arabidopsis seedlings were frozen in liquid N₂ and ground in a mortar with a pestle. The powder was suspended in a homogenizing medium containing 50 mM Tris-HCl (pH 7.8, ratio 1:3; w/v) containing 0.1 mM EDTA, 0.2% (v/v) Triton X-100 and 10% (v/v) glycerol. Homogenates were centrifuged at 27,000 g for 20 min, and the supernatants were used for the enzymatic assays.

Catalase activity (EC 1.11.1.6) was assayed by measuring the disappearance of H₂O₂ as described by Aebi (1984). NADH-dependent hydroxypyruvate reductase (HPR; EC 1.1.1.29) was determined according to the method used by Schwitzguébel and Siegenthaler (1984). Glycolate oxidase (GOX; EC 1.1.3.1) was assayed as previously described (Kerr and Groves, 1975), by measuring the formation of glyoxylate-phenylhydrazone.

Total RNA was extracted with Trizol according to instructions provided by Gibco BRL (Life Technologies); 5 μg of total RNA were used to produce cDNA for the reverse transcriptase (RT) reaction by adding 0.5 mM dNTPs, poly-dT₂₃, 5x RT buffer, 40 U RNase inhibitor (Invitrogen) and 200 U Reverse Transcriptase (Thermo Fisher) in a final volume of 20 μl. The reaction was carried out at 50°C for 30 min.

Semiquantitative RT-PCR amplification of actin 2 (ACT2) cDNA from Arabidopsis was chosen as control. CAT1, CAT2, CAT3, HPR1, GOX1, GOX2 and GOX3 cDNAs were amplified by a PCR as previously described (Corpas and Barroso, 2016). PCR products were then detected after electrophoresis in 2.8% (w/v) agarose gels and by staining with GelRed™. Quantification of the bands was performed using a Gel Doc system (Bio-Rad Laboratories) coupled with a high-sensitivity charge-coupled device (CCD) camera. The T/C ratio indicates the relative level of the each amplification product (T) over the ACT2 used as internal control (C) after normalization to the control samples, and it expresses the fold change with respect to the untreated control (Marone et al., 2001).

Protein concentration was determined with the aid of the Bio-Rad Protein Assay (Hercules, CA) using bovine serum albumin as a standard. For microscope analyses, around five to ten 5-day-old Arabidopsis seedlings were used for each experiment and treatment. For enzymatic activity and gene expression assays, ∼150 14-day-old Arabidopsis seedlings were used for each experiment. Data are means of at least three sets of independent experiments with three repetitions each. To estimate the statistical significance between means, the data was analyzed by the Student's test. Relative fluorescence was quantified by using ImageJ software.

Authors are grateful to Mr. Carmelo Ruiz-Torres for his excellent technical support. The authors thank Dr Juan D Alché (CSIC) for his advice on the analysis of images. Technical and human support provided by CICT of Universidad de Jaén (UJA, MINECO, Junta de Andalucía, FEDER) is also acknowledged. We also thank Michael O'Shea for proofreading the text.