Reversible S-nitrosylation of bZIP67 by peroxiredoxin IIE activity and nitro-fatty acids regulates the plant lipid profile

Summary

Nitric oxide (NO) is a gasotransmitter required in a broad range of mechanisms controlling plant development and stress conditions. However, little is known about the specific role of this signaling molecule during lipid storage in the seeds. Here, we show that NO is accumulated in developing embryos and regulates the fatty acid profile through the stabilization of the basic/leucine zipper transcription factor bZIP67. NO and nitro-linolenic acid target and accumulate bZIP67 to induce the downstream expression of FAD3 desaturase, which is misregulated in a non-nitrosylable version of the protein. Moreover, the post-translational modification of bZIP67 is reversible by the trans-denitrosylation activity of peroxiredoxin IIE and defines a feedback mechanism for bZIP67 redox regulation. These findings provide a molecular framework to control the seed fatty acid profile caused by NO, and evidence of the in vivo functionality of nitro-fatty acids during plant developmental signaling.

Graphical abstract

Highlights

• •bZIP67 is a target of NO and NO₂-Ln that accumulates to induce expression of FAD3 desaturase
• •A non-nitrosylable version of bZIP67 is non-functional to activate FAD3 expression
• •S-nitrosylation of bZIP67 is reversed by PRXIIE activity impairing fatty acid profile
• •bZIP67 S-nitrosylation and function is enhanced in the NO₂-Ln overaccumulating aer mutant

Sánchez-Vicente et al. provide a molecular framework for the role of NO and nitro-linolenic acid on lipid accumulation during embryo development through the stabilization of the bZIP67 transcription factor. Once accumulated by S-nitrosylation, bZIP67 is responsible for the conversion of linoleic acid into linolenic acid through FAD3 desaturase transcriptional activation.

Introduction

Nitric oxide (NO) is a plant and animal gasotransmitter of great relevance to a plethora of developmental processes and under stress conditions¹ including germination promotion,² root growth and stem cell maintenance,^3,4,5,6 flowering,⁷ senescence,⁸ stomatal closure,⁹ pathogen responses,^10,11 abiotic stresses,¹² nutritional deficiencies,¹³ Fe homeostasis,¹⁴ cell death,¹⁵ nodulation, and symbiosis.^16,17 It is known that NO and phytohormones interact in a complex signaling network to regulate these processes.¹⁸ In particular, for seed development and germination, it has been described that NO is required for the correct size and number of siliques and seed yield.¹⁹ Thus, NO-deficient mutants, such as nitric oxide associated1 (noa1-2), nitrate reductases (nia1nia2), and the triple mutant nia1nia2noa1-2, present increased seed dormancy and lower germination rate, effects which highlight the vital role of NO during these developmental stages.¹⁹

NO exerts its effect mainly through post-translational modifications, which include tyrosine nitration, metal nitrosylation, and cysteine (Cys) S-nitrosylation (also known as S-nitrosation). Protein S-nitrosylation involves the covalent attachment of an NO moiety to the thiol group of specific Cys residues to form an S-nitrosothiol (SNO),²⁰ which control the cellular localization, function, and stability of many proteins. In vitro and in vivo evidence supporting NO-dependent modifications have been reported for specific regulatory proteins including PRX IIE,²¹ GSNOR,²² and PRX IIF.²³ Previous research has shown that the S-nitrosylation of NPR1¹⁰ is reversible,²⁴ a mechanism that includes more complexity and selectivity in the regulation of redox networks.

Furthermore, the transcription factors TGA1 and ABI5, containing the basic region/leucine zipper motif (bZIP), have also been described as NO targets.^2,25 bZIPs regulate many processes throughout the entire plant life cycle²⁶ and their functions are modulated by post-translational modifications,²⁷ including redox networks. In this context, we have previously reported how ABI5, a central hub for ABA seed germination repression,²⁸ is regulated by NO.² The mechanism described implies the S-nitrosylation of a key Cys residue, which promotes interaction with CULLIN4 (CUL4)-based and KEEP ON GOING (KEG) E3 ubiquitin ligases, leading to ABI5 degradation.

The Arabidopsis ABI5 homolog bZIP67 participates in the regulation of fatty acid composition during seed development and in the establishment of primary seed dormancy, the latter depending on the different ambient temperatures during seed setting and maturation. Thus, bZIP67 is known to be more abundant in seeds that mature in cooler conditions, providing a mechanism to explain how temperature regulates DOG1 expression and enhances seed dormancy.²⁹ Remarkably, bZIP67 regulates fatty acid desaturase3 (FAD3) expression in a transcriptional complex that includes LEAFY COTYLEDON1-LIKE (L1L) and NUCLEAR FACTOR-YC2 (NF-YC2).³⁰ The conversion of linoleic acid (18:2) into linolenic acid (18:3), two of the major fatty acids in seed composition, is promoted by FAD3.³¹ Indeed, a negative correlation between NO and oleic acid (18:1), the linoleic acid precursor, has been reported. Oleic acid inhibits NIA1 and NIA2 transcription, and physically binds to NOA1 leading to its degradation and, in consequence, the repression of NO synthesis.³² The generation and detection of nitro-fatty acids (NO₂-FA) in plants has also been reported, and it is known that they can exert additional functions during antioxidant and chaperone responses to abiotic stress conditions.^33,34 However, despite what is already known about NO, its direct targets with regard to the regulation of seed lipid content are still undefined. Here, based on ABI5 and bZIP67 homology, we analyzed the effect of NO and nitro-linolenic acid (NO₂-Ln) on bZIP67 S-nitrosylation and accumulation, and the reversibility of this process by the redoxin activity of PRX IIE, which ultimately influences bZIP67 accumulation and stabilization. These findings establish a molecular framework for the regulation of seed fatty acid storage accumulation by NO and the involvement of temperature in developing tissues.

Results

NO co-localizes with bZIP67 to promote its accumulation in siliques

NO modulates seed physiology, alleviating seed dormancy and promoting germination. Although the accumulation of NO in the apoplastic aleurone layer^35,36 and in Arabidopsis seed endosperm has been described to promote ABA catabolism,³⁷ less is known about NO levels and function during embryo development and seed maturation. We detected endogenous levels of NO, using the fluorescence dye 4,5-diaminofluorescein diacetate (DAF-2DA), in Arabidopsis green mature embryos with a specific localization pattern. NO-dependent fluorescence was mainly observed in the cotyledons and the radicle elongation zone (Figure 1A). To prove that this fluorescence is only prompted by endogenous NO, DAF-2DA was combined with the NO scavenger 2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide (cPTIO) or in embryos untreated with DAF-2DA where no fluorescence signal was detected under these circumstances (Figure S1A).

Figure 1.

NO and bZIP67 accumulation in green mature embryos alters FAD3 expression and seed lipid content

(A) Localization of bZIP67 in Col-0;pbZIP67:GFP-bZIP67 (top) and NO detection (bottom) using the DAF-2DA probe in Arabidopsis green mature embryos. GFP and DAF-2DA fluorescence co-localizes throughout the cotyledons. Scale bars, 100 μm.

(B) Conversion of linoleic acid into linolenic acid through the FAD3 desaturase transcriptional activation by bZIP67 (1, Mendes et al.³⁰; 2, Browse et al.³¹).

(C) Western blot analysis of bZIP67 accumulation under the NO donor GSNO (100, 250, and 500 μM) and NO scavenger cPTIO (100, 250, and 500 μM). bZIP67 protein analysis after 96 h of treatment of seeds collected from 35S:HA-bZIP67 in the bzip67 mutant background, and in bzip67 mutant as a negative control, using anti-HA antibody. Actin protein levels are shown as a loading control.

(D) Western blot analysis of bZIP67 accumulation in the siliques of the wild-type Col-0, the bzip67 mutant, and the NO mutant backgrounds cue1/nox1 and nia1nia2noa1-2 using anti-bZIP67 antibody. Actin protein levels are shown as a loading control.

(E) FAD3 transcript abundance quantified by qRT-PCR in developing seeds of Col-0 in control conditions and under NO treatments (donor 1 mM GSNO and scavenger 100 μM cPTIO) and the bzip67 mutant.

(F) FAD3 transcript abundance quantified by qRT-PCR in siliques of the wild-type Col-0 and the NO mutant backgrounds cue1/nox1 and nia1nia2noa1-2. Statistics were analyzed by one-way ANOVA. Different letters correspond to significant differences (p < 0.05).

(G) Analysis of seed lipid profile of oleic (18:1), linoleic (18:2), and linolenic (18:3) acids, in the wild-type Col-0 and in the different NO mutant backgrounds. Values are the mean of three determinations. Values for each genotype are annotated by performing a one-way ANOVA. Letters indicate significant differences (p ≤ 0.05).

To further understand the mechanism by which NO can affect this developmental stage, we analyzed bZIP67 as a master regulator during seed maturation programs.^38,39 bZIP67 contributes to storage lipid accumulation with a positive regulation of the ω-3 fatty acid desaturase FAD3.²⁹ There is evidence that bZIP67 directly binds to FAD3 promoter as a part of a transcriptional complex formed by L1L and NF-YC2.³⁰ FAD3 desaturase enzyme converts linoleic acid (18:2) into linolenic acid (18:3),³¹ two major fatty acids of the seeds (Figure 1B). Consequently, bZIP67 gene expression and protein localization were corroborated using the GFP reporter line⁴⁰ (Figures S1B and S1C). GFP-bZIP67 signal was detected in the nuclei of the embryo cotyledons (Figure 1A). Hence, bZIP67 co-localizes with NO-dependent DAF-2DA fluorescence supporting the putative involvement of this molecule in bZIP67 regulation.

To test the NO impact, we checked the bZIP67 transcript abundance in developing seeds after treatments with the NO donor S-nitrosoglutathione (GSNO) and the NO scavenger cPTIO, and in the battery of NO mutant backgrounds. bZIP67 transcript was highly accumulated only in the NO overproducer cue1/nox1 mutant (impaired in a plastid phosphoenolpyruvate/phosphate translocator) (Figure S1D), an effect also detected in a lesser extent after GSNO treatments (Figure S1E).

Focusing on protein level, we analyzed bZIP67 accumulation in seeds under several concentrations of the NO donor GSNO and the NO scavenger cPTIO (Figure 1C). bZIP67 accumulation was promoted under GSNO but not under cPTIO treatments at higher concentrations. To support this, we also analyzed bZIP67 accumulation in developing seeds from the NO-overproducing cue1/nox1 and the NO-deficient nia1nia2noa1-2 mutant backgrounds (Figure 1D). In addition, mutants with altered GSNO content such as hot5-2 and 35S:FLAG-GSNOR were included in our experimental setup, but no differences in bZIP67 accumulation among these two lines was detected (Figure S1F).

All these results demonstrated that the NO gasotransmitter induces bZIP67 transcription and protein accumulation.

NO increases FAD3 transcript abundance that ultimately impacts on linolenic acid accumulation in seeds

One main readout of highly accumulated bZIP67 is the increase of the expression level of the direct target FAD3. FAD3 transcript abundance after NO treatments and in the NO mutant backgrounds was analyzed in developing seeds. A 2-fold increase of the FAD3 transcript was detected in seeds after treatment with GSNO (Figure 1E) and in the cue1 samples (Figure 1F), corresponding with the NO-induced bZIP67 accumulation (Figures 1C and 1D). This response was lower in the nia1nia2noa1-2 mutant, and not found in the hot5-2 and 35S:FLAG-GSNOR lines (Figure S2A) or after cPTIO treatments.

The second readout of altering bZIP67 accumulation is the change on the fatty acid profile, and especially in the ratio between linoleic acid (18:2) and linolenic acid (18:3). Thus, Arabidopsis total seed lipid content was evaluated focusing on oleic (18:1), linoleic (18:2), and linolenic acid (18:3) cleavage from triacylglycerols in the different NO mutants (Figure 1G). A significant increment of 18:3 content was observed in cue1 seeds compared with the Col-0 that correlated with higher bZIP67 accumulation and FAD3 expression in this mutant background, as shown in Figures 1D and 1F.

Recently, the involvement of temperature during seed setting and maturation has been described to impact on bZIP67 protein abundance.²⁹ While lower temperatures—starting at 15°C—during seed maturation increased bZIP67 abundance, higher ambient temperatures—up to 25°C—decreased bZIP67 close to undetectable levels (Figure S2B). Hence, to analyze if NO could contribute to temperature-induced changes in the percentage of 18:1, 18:2, and 18:3 profiles we used nia1nia2noa1-2, cue1, and Col-0 seeds developed at 15°C, 21°C, and 25°C. Linolenic acid accumulation was always higher in cue1 than in wild-type seeds at all temperatures analyzed (Figures 1G and S2C), correlating with higher bZIP67 protein accumulation and FAD3 expression in that mutant background (Figures 1D and 1F). On the contrary, 18:3 levels in nia1nia2noa1-2 seeds showed minor changes at 21°C, but lower or higher levels than the Col-0 seeds at 15°C or 25°C, respectively (Figures 1G and S2C). The hot5-2 and 35S:FLAG-GSNOR lines showed an irregular profile in 18:3 abundance at each temperature (Figure S2D). Likewise, the GSNOR-related mutants did not accumulate the bZIP67 protein or the FAD3 transcript (Figures S1F and S2A), demonstrating that other regulatory mechanisms might be involved in 18:3 synthesis beyond the suggested canonical pathway (Figure 1B).

Also, the total quantity of fatty acids and oil bodies demonstrated that alterations on endogenous NO levels lead to a total misregulation of the fatty acid abundance and oil body distribution in the embryo (Figures S2E–S2G). This pointed out that other targets of the fatty acid biosynthetic pathway may also be regulated by NO apart from the suggested signaling model NO-bZIP67-FAD3.

Together, these results demonstrated that the NO-accumulated bZIP67 promotes FAD3 expression and the consequent 18:3 increase during seed development and maturation.

bZIP67 is S-nitrosylated in vitro and in vivo

NO can impact on protein structure, activity, and/or stability via the modification of free thiol groups (-SH) of Cys residues through S-nitrosylation. This post-translational modification consists of the attachment of an NO molecule to the sulfur atom, generating an SNO residue.²⁰ We previously showed that NO promoted ABI5 degradation through the S-nitrosylation of the Cys153 residue.² The closest homolog of ABI5 in the Arabidopsis group A of bZIPs is bZIP67. The Cys153 residue of ABI5² is conserved in bZIP67 as Cys106, and two additional Cys residues—Cys186 and Cys215—are also present in bZIP67 protein sequence (Figure 2A). By an in silico approach using the GPS-SNO 1.0 software and the dbSNO database, we analyzed the S-nitrosylation of the three Cys residues of bZIP67 in association with the flanking amino acids (Figures S3A and S3B). The results from the GPS-SNO 1.0 predicted only Cys186 as a candidate for S-nitrosylation. However, dbSNO database software projected the three Cys residues of bZIP67 as candidates for S-nitrosylation. To check these predictions, a recombinant 6xHis-bZIP67 protein (Figures S3C–S3F) was used for in vitro assays to detect S-nitrosylation using the biotin-switch method.⁴¹ Thus, the recombinant bZIP67 protein was treated with the NO donor GSNO to promote the formation of bZIP67-SNOs (Figure 2B). Since S-nitrosylation has been described as a reversible post-translational modification, a GSNO treatment coupled with the reducing agent dithiothreitol (DTT) was done to check the reversibility of bZIP67-SNO. As a result, we determined that bZIP67 was S-nitrosylated in vitro and that this post-translational modification was depleted by adding DTT (Figure 2B). To further confirm these results, the recombinant bZIP67 treated with GSNO was subjected to the biotin-switch method and analyzed by liquid chromatography coupled to electrospray tandem mass spectrometry analysis (Figure 2C). The three Cys of bZIP67 were identified as Cys106-biotin, Cys186-biotin, and Cys215-biotin with higher mass/charge ratio than standard Cys residues (Figures S4–S6).

Figure 2.

S-nitrosylation of bZIP67 in vitro and in vivo

(A) The predicted structure of the bZIP67 transcription factor. The basic domain is highlighted in gray and the bZIP domain in black. The position of the three cysteine residues is marked in the protein sequence.

(B) In vitro S-nitrosylation of bZIP67 recombinant protein by the NO donor GSNO (500 μM). This modification is reversed using DTT (20 mM). Control without ascorbate is shown (–Asc). Input protein levels are determined using anti-bZIP67 antibody.

(C) Mass spectrometric analyses for the identification of Cys106 (top), Cys186 (middle), and Cys215 (bottom) as S-nitrosylation sites. Tandem mass spectrometry spectra from the tryptic fragments QGSLSLPVPLCK, MSSSDFGYNPEFGVGLHCQNQNNYGDNR, and SVYSENRPFYSVLGESSSCMTGNGR, in which Cys106, Cys186, and Cys215 are modified by Biotin-HPDP, respectively.

(D) In vitro S-nitrosylation of bZIP67 recombinant proteins by the NO donor GSNO (500 μM). This modification is reversed using DTT (20 mM). Different versions of the protein carrying mutated Cys were used to determine the main Cys residue implicated in bZIP67 S-nitrosylation.

(E) In vivo S-nitrosylation of bZIP67 and bZIP67 w/o Cys proteins in 7-day-old seedlings. Non-treated (NT) or DTT (20 mM) and GSNO (500 μM) were added to confirm the reversibility and maintenance, respectively, of the S-nitrosylation modification. Input protein levels are determined using anti-HA antibody.

This information was used to mutagenize a bZIP67 sequence to obtain single Cys106Ser, Cys186Ser, and Cys215Ser, double Cys106Ser/Cys186Ser, and triple Cys-less Cys106Ser/Cys186Ser/Cys215Ser (bZIP67w/oC) mutants of the protein. After expression and purification of the corresponding mutant recombinant proteins of bZIP67, the S-nitrosylation ability of each protein version was studied by the biotin-switch method. It was found that single bZIP67 mutants Cys106Ser, Cys186Ser, and Cys215Ser reduced the S-nitrosylation capacity of the protein, while the double Cys mutant and the bZIP67w/oC greatly diminished the S-nitrosylation of bZIP67 (Figure 2D).

To obtain in vivo evidence of bZIP67 S-nitrosylation, transgenic lines overexpressing the wild-type HA-bZIP67 and mutant HA-bZIP67w/oC proteins were generated in the knockout bzip67 mutant background (Figures S7A and S7B). Total protein extract from 7-day-old seedlings of bzip67;35S:HA-bZIP67 and bzip67;35S:HA-bZIP67w/oC was obtained and each extract was divided into three samples, non-treated (NT), treated with GSNO, and treated with DTT (Figure 2E). Samples were subjected to the biotin-switch method, immunoprecipitated by streptavidin, and the detection of S-nitrosylated bZIP67 versions was performed using an anti-HA antibody. The HA-bZIP67 version was already detected as S-nitrosylated in the NT sample, and the signal was increased after GSNO treatment or diminished after DTT. Correspondingly, the HA-bZIP67w/oC protein lost the capacity to be S-nitrosylated, and therefore affected by the GSNO or DTT treatments.

These findings demonstrated that NO can directly exert a modification on the Cys106, Cys186, and Cys215 residues of bZIP67 through S-nitrosylation.

bZIP67 is stabilized and functionally prompted through S-nitrosylation

The effect of NO on bZIP67 protein accumulation is shown on Figures 1 and S1. Here, we used seeds of bzip67;35S:HA-bZIP67 and bzip67;35S:HA-bZIP67w/oC lines to functionally asses the role of bZIP67 S-nitrosylation. These lines were plated on medium containing ABA, GSNO, cPTIO, or without any compound (C) during 96 h and samples were harvested to study HA-bZIP67 and HA-bZIP67w/oC accumulation. Compared with the control condition, HA-bZIP67 abundance was decreased under ABA and cPTIO treatments, but highly accumulated by GSNO (Figure 3A). Regarding the HA-bZIP67w/oC, neither cPTIO nor GSNO influenced the accumulation of the mutant version of the protein, while ABA treatment lowered its abundance (Figure 3B). This result demonstrated that abolishing bZIP67 S-nitrosylation ability in the mutant version without Cys residues, completely impairs the bZIP67 response to NO and compromises its stability.

Figure 3.

NO transcriptional regulation and stabilization of bZIP67

(A) Western blot analysis of bZIP67 accumulation under ABA (2 μM), the NO donor GSNO (500 μM), and NO scavenger cPTIO (100 μM). bZIP67 protein analysis of seeds collected after 96 h of treatment from 35S:HA-bZIP67 and (B) 35S:HA-bZIP67w/oC in bzip67 mutant background using anti-HA antibody. Actin protein levels are shown as a loading control.

(C) Western blot analysis of bZIP67 protein levels in 4-day-old seeds from 35S:HA-bZIP67 and (D) 35S:HA-bZIP67w/oC in bzip67 mutant background, in the presence of cycloheximide (1 mM) and cycloheximide (1 mM) plus GSNO (500 μM) during 0 to 4 h. Actin protein levels are shown as a loading control.

(E) FAD3 transcript abundance quantified by qRT-PCR in developing seeds of Col-0, bzip67 mutant, and 35S:HA-bZIP67w/oC in the bzip67 mutant background control and under NO donor treatment (1 mM GSNO) and the bzip67 mutant. Statistics were analyzed by one-way ANOVA. Different letters correspond to significant differences (p < 0.05).

(F) Analysis of the seed lipid profile in Col-0, bzip67 knockout mutant, 35S:HA-bZIP67, and 35S:HA-bZIP67w/oC transgenic lines in the bzip67 background. Values are the mean of three determinations. Values for each genotype are annotated by performing a one-way ANOVA. Letters indicate significant differences (p ≤ 0.05).

To go further deeper into the NO regulation of bZIP67, we blocked the synthesis of a new protein by treating the transgenic lines with the protein synthesis inhibitor cycloheximide (CHX). The CHX treatment was done either alone or in combination with GSNO and the accumulation of the wild-type and mutant bZIP67 proteins was monitored from 1 to 4 h (Figures 3C and 3D) and up to 24 h (Figures S8A and S8B). HA-bZIP67 protein accumulation was driven by NO until 12 h, while HA-bZIP67w/oC exhibited insensitivity to both treatments. This result supported that bZIP67 S-nitrosylation stabilized this transcription factor. To investigate if the stabilization of bZIP67 by NO was controlled by the proteasome, we combined the aforementioned treatments with the proteasome inhibitor MG132 and checked the protein levels. In contrast to the degradation mechanisms that undergo ABI5, bZIP67 destabilization was independent of the proteasome pathway since MG132 did not restore protein accumulation over time (Figure S8C). Considering the homology between ABI5 and bZIP67, we analyzed the levels of ABI5 in the line bzip67;35S:HA-bZIP67 and the specificity of the antibodies generated (Figures S8C and S8D). The accumulation of these two closest homologs follows an opposing trend. While the presence of ABA and cPTIO decreased bZIP67 accumulation and promoted ABI5 abundance, GSNO treatments accumulated bZIP67 protein but destabilized ABI5. Since ABI5 was described to be involved in seed size control,⁴² we also checked if bZIP67 could contribute as well to modulate these parameters by measuring seed width and length. Compared with wild-type Col-0, seed size parameters were not altered in the knockout mutant or in the bzip67;35S:HA-bZIP67. However, bzip67;35S:HA-bZIP67w/oC displayed a significant increase in seed width and length (Figures S8E and S8F).

Cys residues are essential for bZIP67 function and regulation by NO

As mentioned before, one of the main readouts of bZIP67 accumulation is the induction of FAD3 expression.³⁰ We analyzed FAD3 transcript abundance in developing seeds of Col-0, bzip67, and bzip67;35S:HA-bZIP67w/oC treated or not with GSNO (Figure 3E). As observed in the graph, GSNO increased the transcript levels of FAD3 only in the wild-type Col-0 sample as already shown in Figure 1E. The bzip67 ko and the bzip67;35S:HA-bZIP67w/oC mutants did not showed induction of FAD3 in response to GSNO. This result supported all the previous findings and restates that the non-ability of HA-bZIP67w/oC to be S-nitrosylated makes the mutant protein non-functional to transcriptionally activate FAD3 expression.

The second readout of altering bZIP67 accumulation is the change on the ratio between the 18:2 and 18:3, thanks to the enzymatic activity of FAD3. Arabidopsis total seed lipid content from Col-0, bzip67, bzip67;35S:HA-bZIP67, and bzip67;35S:HA-bZIP67w/oC was evaluated by analyzing 18:1, 18:2, and 18:3 cleavage from triacylglycerols (Figure 3F). The bzip67 ko showed higher levels of 18:2 and lower of 18:3 compared with Col-0 seeds. The transgenic line bzip67;35S:HA-bZIP67 was able to restore and even increase to the wild-type levels the abundance of 18:3 on the bzip67 ko mutant, proving the functional complementation of this line. Interestingly, the line bzip67;35S:HA-bZIP67w/oC showed the same 18:2 and 18:3 pattern than the bzip67 ko mutant. This result demonstrated that the mutant HA-bZIP67w/oC had no functional complementation and was unable to induce FAD3 expression and therefore accumulate 18:3 up to wild-type Col-0 levels. We also analyzed the total fatty acid content of these lines, but non-significant differences were observed (Figure S8G).

Since it has been recently shown that different ambient temperatures during seed setting and maturation affect bZIP67 abundance,²⁹ we tested whether HA-bZIP67 and HA-bZIP67w/oC accumulation was also compromised under these temperatures (Figure S8H). Protein extracts from leaves of bzip67;35S:HA-bZIP67 and bzip67;35S:HA-bZIP67w/oC grown at 15°C and 25°C showed that HA-bZIP67 protein abundance is enhanced at lower temperatures but decreased until total fading of the signal at higher ambient temperatures. Remarkably, an ambient temperature of 25°C did not affect HA-bZIP67w/oC abundance and a signal similar as the one observed at 15°C was detected for the Cys-less protein.

All these findings support the biological function of bZIP67 S-nitrosylation on the accumulation of bZIP67 to induce FAD3 expression that ultimately regulates 18:2 and 18:3 abundance in seeds.

bZIP67 is a component of protein complexes involved in transcriptional regulation and redox homeostasis of lipid profile in developing seeds

Understanding the dynamics of protein complex components constitutes a key point for the spatiotemporal regulation of networks. Previously, bZIP67 was described to be part of the L1L and NF-YC2 protein complexes that regulate the expression of cruciferin C (CRC), sucrose synthase 2 (SUS2), and FAD3.^30,43 To identify additional bZIP67 interactors, we performed an in vivo pull-down assay using developing siliques of pbZIP67:GFP-bZIP67 and wild-type Col-0 as control. GFP-bZIP67 was used as a bait and the proteins bound to bZIP67 were immunoprecipitated using anti-GFP beads and identified by tandem mass spectrometry.⁴⁴ From the 433 initial putative interactors of bZIP67, only 30 of them showed statistical relevance according to MaxQuant and Perseus software analysis from the three biological replicates of each genotype (Table S2). The minimum detection threshold established for this MS analysis was 17.30, and the high confidence candidates were categorized based on the GFP-bZIP67 versus Col-0 ratio. Both the GFP-bZIP67 bait and truncated GFP were recovered at the top of the resulting list as the higher abundant proteins.

Among the high confident interactors of bZIP67, we found that many were proteins involved in lipid storage regulation such as hydroxysteroid dehydrogenase 1 (HSD1), seed storage albumins (SESA1 and SESA4), and oleosins (OLEO1, OLEO2, and OLEO4). In addition, dehydrin HIRD11 and peroxiredoxin IIE (PRXIIE) proteins involved in the regulation of redox homeostasis were also identified as bZIP67 interactors. Of great relevance was the identification of PRXIIE as an interactor of bZIP67, since this peroxiredoxin is post-translationally modified by S-nitrosylation and highly involved in NO homeostasis.²¹ Three peptides from PRXIIE were identified by mass spectrometry in the pull-down assay that are represented with pink, yellow, and purple colors, in a protein model (Figure S9A). With this information, we checked again the interaction by in vivo co-immunoprecipitation assays using the prxiie;35S:PRXIIE-GFP and bzip67;35S:HA-bZIP67 lines (Figure S9B). Protein extracts from these lines were mixed and treated or not with GSNO. The HA-bZIP67 protein was selected as bait and immunoprecipitated with anti-HA beads and the presence of PRXIIE-GFP co-immunoprecipitated was detected by the anti-GFP antibody. Interestingly, the bZIP67-PRXIIE interaction was promoted by the presence of GSNO, indicating that the nature of this interaction may be related to NO homeostasis. Using embryos from developing seeds of the line prxiie;35S:PRXIIE-GFP that were also incubated with the DAPI staining, we detected the localization of PRXIIE-GFP and the nuclei by confocal fluorescence microscopy (Figure S9C). PRXIIE-GFP, assigned with green color, was detected in the chloroplasts surrounding the nuclei that were shown in blue due to DAPI staining. Previous investigations have suggested intraorganellar traffic between chloroplast and nuclei related to redox regulation through specific protuberances when these organelles appear in close proximity.⁴⁵

The redoxin activity of PRXIIE trans-denitrosylates bZIP67, dampening its function

To further investigate the molecular and physiological relevance of PRXIIE and bZIP67 interaction, we checked whether the expression pattern of PRXIIE and bZIP67 exhibited some similarity during seed setting and maturation. We benefited from the Arabidopsis eFP Browser that demonstrated high levels of co-expression of PRXIIE and bZIP67 during embryo development and a similar trend during seed maturation and in imbibed seeds (Figure S9D).⁴⁶ Since both genes share the same expression pattern and PRXIIE has already been involved in reactive oxygen and nitrogen species homeostasis,²¹ we investigated whether alterations on PRXIIE levels could affect the abundance of endogenous bZIP67 protein. Protein extracts from developing seeds from the prxiie knockdown (KD) mutant and the three transgenic lines 35S:PRXIIE 1, 2 and prxiie;35S:PRXIIE-GFP, were used to detect the endogenous bZIP67 and different PRXIIE versions on each genetic background (Figure 4A). We observed that lowering the levels of PRXIIE in the prxiie KD resulted in higher accumulation of bZIP67, while flooding the seeds with PRXIIE in the overexpressing lines decreased bZIP67 close to undetectable levels. The increment of bZIP67 protein was not due to an induction of the expression of bZIP67 in those lines (Figure S9E). Therefore, PRXIIE may be implicated in bZIP67 stabilization. In Figures 1 and 3 we showed that the higher accumulation of bZIP67 led to induction of FAD3 expression. Hence, we analyzed FAD3 transcript abundance in seeds of the PRXIIE mutant backgrounds and found out that the prxiie KD mutant showed a slight increased level of FAD3 transcript than the wild-type Col-0 seeds (Figure 4B). Moreover, the 18:1, 18:2, and 18:3 fatty acid profiles were analyzed in these PRXIIE lines, showing that the levels of 18:3 in the prxiie KD mutant were higher than in the Col-0 (Figure 4C). This was consistent with the increment of FAD3 expression and bZIP67 protein accumulation shown in this mutant. The total fatty acid content of the seeds in the analyzed lines showed a similar trend than the control Col-0 (Figure S9F). These results indicated that altering the content of PRXIIE has a direct impact on bZIP67 abundance and the consequent transcriptional regulation on FAD3.

Figure 4.

bZIP67 reversible S-nitrosylation by PRX IIE activity

(A) Western blot analysis of bZIP67 and PRX IIE in green siliques of PRX IIE loss-of-function (prxIIe) and gain-of-function lines (35S:PRX IIE and prx IIe;35S:PRX IIE-GFP) using anti-bZIP and anti-PRX IIE antibodies. Actin protein levels are shown as a loading control.

(B) FAD3 transcript abundance quantified by qRT-PCR in developing seeds of Col-0 as a control, bzip67 mutant and PRX IIE loss-of-function (prx IIe) and gain-of-function lines (35S:PRX IIE). Statistics were analyzed by one-way ANOVA. Different letters correspond to significant differences (p < 0.05).

(C) Analysis of the seed lipid profile in PRX IIE overexpression lines and prx IIe knockout mutant. Values are the mean of three determinations. Values for each genotype are annotated by performing a one-way ANOVA. Letters indicate significant differences (p ≤ 0.05).

(D) Trans-denitrosylation assay between bZIP67-SNO and PRX IIEΔSP. bZIP67 recombinant protein is treated previously with GSNO and then subjected to reduced PRX IIEΔSP. No signal is observed in the presence of DTT. bZIP67 and PRX IIEΔSP protein loading is detected by anti-His antibody.

(E) In vivo denitrosylation assay between bZIP67 and PRX IIE. bzip67;35S:HA-bZIP67 protein extracts are treated, or not, with GSNO and then subjected to prx IIe or 35S:PRX IIE-GFP (intact or heat inactivated) extracts. No or low signal is observed in the presence of DTT. bZIP67 and PRX IIE protein loading is detected by anti-HA and anti-GFP antibodies, respectively.

It is known that PRXIIE is involved in peroxynitrite (ONOO-) detoxification, and its enzymatic activity is inhibited by NO through S-nitrosylation.²¹ Since bZIP67 is susceptible to be modified by S-nitrosylation (Figure 2), we investigated whether the redoxin activity of PRXIIE was able to denitrosylate bZIP67 and affect its protein fate using in vitro and in vivo assays. Recombinant bZIP67, PRXIIEΔSP (without chloroplast signal peptide), and PRXIIEΔSPw/oC (a Cys-less protein without chloroplast signal peptide) were used (Figure 4D). PRXIIEΔSP version was used to mimic the in vivo effect since no complete version was detected in Arabidopsis. First, the recombinant bZIP67 samples were treated with GSNO to promote its S-nitrosylation. Second, the recombinant PRXIIEΔSP or PRXIIEΔSPw/oC were added to the GSNO-treated bZIP67 samples and the biotin-switch assay was performed. We found out that the S-nitrosylation of bZIP67 was significantly reduced after the incubation with PRXIIEΔSP, but not when incubated with PRXIIEΔSPw/oC. This demonstrated that the redoxin activity of PRXIIE was able to remove the formed SNO on bZIP67 and this was achieved through the two Cys residues of the protein. We also used the reducing agent DTT on the GSNO-treated bZIP67 samples to probe that the S-nitrosylation of bZIP67 was a reversible mechanism. To support these results, we performed in vivo denitrosylation experiments using bzip67;35S:HA-bZIP67, prxiie;35:PRXIIE-GFP, and prxiie lines (Figure 4E). Protein extracts from bzip67;35S:HA-bZIP67 were treated or not with GSNO and/or DTT to promote or remove the formation of HA-bZIP67-SNO. Then, the protein extracts from prxiie;35S:PRXIIE-GFP, prxiie, and inactive (boiled) 35S:PRXIIE-GFP were added to the respective bzip67;35S:HA-bZIP67 samples and the biotin-switch was carried out. Evidence is shown that increasing PRXIIE in the sample resulted in denitrosylation of HA-bZIP67, while inactivating the redoxin activity of PRXIIE or backgrounds lacking PRXIIE protein failed to remove the SNO from the bZIP67. All together, these results proved the trans-denitrosylation mechanisms exerted by the redoxin activity of PRXIIE to remove the NO from bZIP67 dampened its accumulation.

bZIP67 S-nitrosylation and function is enhanced in the NO₂-Ln overaccumulating aer mutant

Recent evidence has demonstrated the role of nitro-linolenic acid (NO₂-Ln) as a powerful NO donor, not only in vitro, but also as an endogenous molecule for in vivo approaches.^33,34 One enzyme involved in the regulation of NO₂-Ln in plants is the alkenal reductase (AER), which has been lately identified thanks to the isolation of the aer KD mutant that lacks this activity and overaccumulates NO₂-Ln.⁴⁷ To investigate whether NO₂-Ln may impact on bZIP67 accumulation, first we found out that AER and bZIP67 are highly co-expressed during embryo development and seed maturation (Figure S9D). Since both genes shared some developmental window, we used developing seeds from the wild-type Col-0 and aer KD mutant to investigate the accumulation of endogenous bZIP67 protein (Figure 5A). Higher levels of bZIP67 accumulation in the aer mutant compared with Col-0 were detected. This produced the first evidence that a genetic background overaccumulating NO₂-Ln also had increased abundance of the bZIP67 protein. Interestingly, we also observed high levels of bZIP67 transcript abundance in the aer background compared with wild-type Col-0, indicating some transcriptional activation (Figure S9G). As performed before in Figures 1, 3, and 4, higher levels of bZIP67 protein led to increased FAD3 expression, a result also observed in the aer mutant (Figure 5B). It has been already published that the aer mutant has increased 18:3 levels compared with wild-type plants⁴⁷, and therefore, this could be a consequence of enhancing bZIP67 levels. To explain the impact of NO₂-Ln on bZIP67 function, we investigated the possible molecular mechanism of NO₂-Ln on bZIP67 by studying the in vitro and in vivo S-nitrosylation of bZIP67. Purified recombinant bZIP67 was treated with NO₂-Ln as an NO donor and monitored using the biotin-switch method as described above (Figure 5C). The S-nitrosylation of bZIP67 by NO₂-Ln but not by linolenic acid was detected. As control, DTT treatment was used to reduce the SNO formed on bZIP67. The in vivo S-nitrosylation was studied by the biotin-switch method using developing seeds from the aer mutant and the wild-type Col-0. As mentioned before, DTT was used as a reducing agent to show the reversibility of the process. The S-nitrosylation of the endogenous bZIP67 was detected in the aer mutant, proving that the higher levels of NO₂-Ln in the developing seeds of this mutant background greatly induced this post-translational modification (Figure 5D).

Figure 5.

S-nitrosylation of bZIP67 by nitro-linolenic acid in vitro and in vivo

(A) Western blot analysis of bZIP67 accumulation in siliques of the wild-type Col-0 and the aer knockdown mutant background using anti-bZIP67 antibody. Actin protein levels are shown as a loading control.

(B) FAD3 transcript abundance quantified by qRT-PCR in developing seeds of the wild-type Col-0 and the aer mutant background. Three biological replicates were analyzed by performing a Student’s t test statistic. Asterisk indicates significant differences (p ≤ 0.05). All the assays were performed with siliques from plants grown at 21°C.

(C) In vitro S-nitrosylation of bZIP67 recombinant protein by nitro-linolenic acid (NO₂-Ln). Purified recombinant bZIP67 is treated with NO₂-Ln (500 μM) and NO₂-Ln (500 μM) and reduced with DTT (50 mM) and pretreated with the NO scavenger cPTIO (500 μM). Control treatments are done with methanol (fatty acid vehicle, lane 1) and linolenic acid (Ln, 500 μM).

(D) In vivo S-nitrosylation of bZIP67 in siliques in the wild-type Col-0 and in the aer knockdown mutant. DTT (20 mM) was added to confirm the reversibility of the S-nitrosylation modification. Input protein levels are determined using anti-bZIP67 antibody.

These results summed up the biochemical and genetic evidence for NO₂-Ln overaccumulation in the molecular and physiological function of bZIP67.

Discussion

NO is a gasotransmitter with an important regulatory role in seeds and in plant establishment. Although its function during seed development still remains elusive, NO is necessary for seed yield,¹⁹ dormancy release, and the promotion of germination and post-germinative development.^2,48,49 In this study, we detect NO accumulation in Arabidopsis embryos, which co-localizes with the basic region/leucine zipper bZIP67 transcription factor, a potential NO target during seed development. Developing embryos accumulate oil to support seedling growth and bZIP67 is thought to be a master regulator of fatty acid storage. Thus, bZIP67 activates FAD3 desaturase expression, which in turn converts linoleic acid (18:2) into linolenic acid (18:3), two main seed fatty acids³⁰ (Figure 1). Accordingly, we demonstrate that the mechanisms by which NO functions during fatty acid storage include gene expression activation (Figure S1), protein accumulation or stabilization (Figure 1), and the post-translational modification of bZIP67, mainly by S-nitrosylation (Figure 2).

The findings by Bryant et al.²⁹ and those reported here demonstrate that bZIP67 protein accumulation is also negatively affected by increasing ambient temperatures. For this reason, we extended the analysis of the fatty acid profile at different temperatures of growth (i.e., 15°C, 21°C, and 25°C). The lipid profiles obtained from different NO and GSNO homeostasis mutant backgrounds at the three temperature regimes reveal that only the NO overproducing cue1/nox1 mutant has a higher linoleic to linolenic acid ratio (step regulated by bZIP67). However, the NO-deficient nia1nia2noa1-2 mutant shows a complete deregulation of the whole lipid profile, fatty acid content, and the presence of oil bodies, while the GSNO-related mutants show small differences in the fatty acid profiles (Figures 1 and S2). These results have allowed us to conclude that both endogenous NO and temperature play a fundamental role in the regulation of oil storage.

The occurrence of bZIP67-SNO formation was analyzed in silico, in vitro, and in vivo in developing seeds and seedlings, and the results indicate that the three Cys residues (Cys106, Cys186, and Cys215) are potential targets to be S-nitrosylated (Figures 2 and S3). We corroborate that bZIP67 accumulation and stabilization by NO and low temperature depend on the three Cys residues involved in bZIP67 post-translational modification, highlighting the physiological relevance of S-nitrosylation (Figure 3). The opposite effects observed after ABA and GSNO treatments on bZIP67 stability might be considered from a developmental checkpoint and tissue specificity point of view (i.e., embryo development), while ABA typically induces the accumulation of NO and promotes global S-nitrosylation under abiotic stress conditions. In addition, bZIP67 accumulation presents an opposite pattern to that of the ABI5 seed germination repressor, which in turn is degraded by the proteasome through S-nitrosylation and accumulated by ABA in the onset of seed germination.² In this case, the differences may occur due to the presence of green tissues in the silique compared with non-green tissues in the mature seed with their corresponding NO homeostasis pathways. Proteasome inhibitor treatment does not lead to bZIP67 accumulation, suggesting a destabilization mechanism independent of the proteasome machinery (Figure S8).

Redox modifications are emerging as central hubs of protein function regulation. Thus, thioredoxins, glutaredoxins, and peroxiredoxins are involved in regulatory networks leading to the reversibility of some of these redox modifications.^50,51 Redox-related partners of bZIP67 include PRX IIE, which participates in redox regulation, the detoxification of hydrogen peroxide (H₂O₂) and peroxynitrite (ONOO⁻) and its function is inhibited by S-nitrosylation.²¹ Based on these findings, PRX IIE redoxin is a clear candidate to regulate bZIP67 transcription factor in a redox-dependent manner. Compelling evidence confirms that PRX IIE activity modulates bZIP67-SNO during seed development. The in vivo denitrosylation assay shows that PRX IIEΔSP is able to denitrosylate bZIP67-SNO and, at the same time, reduced PRX IIEΔSP is converted into PRX IIEΔSP-SNO, highlighting a trans-denitrosylating effect (Figure 4). Kneeshaw et al.²⁴ described how THIOREDOXIN-h5 functions as a protein-SNO reductase of NPR1, leading to the reverse SNO modification during plant immunity. Furthermore, although a similar expression pattern of both genes during embryo development is observed (Figure S9), we cannot disregard the possibility that other redoxins occurring in redox cascades, such as thioredoxins and glutaredoxins, may be involved in the PRX IIE-bZIP67 redox regulation.

Remarkably, NO₂-Ln, which is accumulated in the alkenal reductase mutant (aer)⁴⁷ and mainly detected in Arabidopsis seeds,³³ also modifies bZIP67 in vitro and in vivo by S-nitrosylation (Figure 5), highlighting a complex regulatory mechanism involving NO and this fatty acid. The expression of bZIP67 is induced in the presence of exogenous NO and mutant backgrounds overproducing NO and NO₂-Ln giving rise to a higher protein accumulation (Figures S1 and S9), which is in accordance with higher linolenic acid detection (Figure 1).

Our findings provide information about the molecular framework for the role of NO and NO₂-Ln on lipid accumulation during embryo development (Figure 6). First, the endogenous NO and NO₂-Ln produced in the seeds promote bZIP67 expression and protein accumulation and stabilization in a similar trend to low ambient temperatures. Higher levels of bZIP67 lead to increased expression of FAD3 desaturase, which catabolizes the conversion of linoleic acid into linolenic acid. Second, the molecular mechanism used by NO to impact on bZIP67 is through S-nitrosylation of the three Cys residues that the transcription factor has. Third, the S-nitrosylated bZIP67-SNO can be reversed by the redoxin activity of PRX IIE in a specific redox regulatory feedback. Based on our results, we hypothesize that bZIP67 is destabilized by PRX IIE activity due to this trans-denitrosylating system and that results in the control of the fatty acid profile during seed development. Finally, by using the aer mutant we found a powerful tool of endogenous NO₂-Ln to modulate in vivo targets, such as bZIP67 by S-nitrosylation. The existence of this type of redox regulatory mechanism opens new windows regarding checkpoints of NO post-translational modifications during plant development.

Figure 6.

Role of NO and NO₂-Ln in the regulation of plant lipid accumulation involving bZIP67 reversible S-nitrosylation by PRX IIE activity

bZIP67 accumulation and stabilization are promoted by NO and low temperature. The conversion of linoleic acid into linolenic acid occurs after the FAD3 desaturase transcriptional activation by bZIP67 (1, Mendes et al.³⁰; 2, Browse et al.³¹). bZIP6-SNO (either by NO or NO₂-Ln) can be reversed by the action of PRX activity, which, in turn, is converted to PRX IIE-SNO in a trans-denitrosylating system. PRX IIE inhibition by NO through S-nitrosylation leads to higher peroxynitrite (ONOO⁻) content (3, Romero-Puertas et al.²¹) able to modify by nitration the pool of fatty acids converting into nitro-fatty acids (4, Mata-Pérez et al.³³) that are accumulated in the ALKENAL REDUCTASE mutant (aer) (5, Mata-Pérez et al.⁴⁷). This regulatory mechanism is involved in embryo fatty acid storage during seed maturation.

Limitations of the study

To fine-tune FAD3 expression requires a transcriptional complex that includes bZIP67, L1L, and NF-YC2. Whether these additional factors are also regulated by NO during the conversion of linoleic acid into linolenic remains to be elucidated.

The physiological relevance of PRX IIE and bZIP67 interaction is supported by their expression patterns in similar types of tissue and during the same developmental stages (i.e., embryo development). Although PRX IIE localization is chloroplastic, through the presence of a signal peptide, the underlying mechanism by which PRX IIE is processed and can interact in the seed with bZIP67 is unknown but could be related to the in vivo subcellular localization of PRX IIE during seed maturation, where cell organelles are in close proximity and nuclei are completely surrounded by the chloroplasts, as discussed above. This fact is supported by the crosstalk between nuclei and organelles reported previously,^52,53 and could explain the feasibility of the interaction between PRX IIE and bZIP67, because both proteins may be close to each other in these subcellular organelles.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Living Colors® GFP Monoclonal Antibody	Clontech	Cat#632375, RRID:AB_2756343
Anti-bZIP67 Purified Rabbit Immunoglobulin	This manuscript	N/A
Anti-ABI5 Purified Rabbit Immunoglobulin	Albertos et al.2	N/A
Anti-PRX IIE	Romero-Puertas et al.21	N/A
Anti-Actin clone 10-B3 Purified Mouse Immunoglobulin	Sigma-Aldrich	Cat#A0480, RRID:AB_476670
ECL Rabbit HRP-linked	GE Healthcare Life Sciences	Cat#NA934, AB_772206
ECL Mouse IgG, HRP-linked	GE Healthcare Life Sciences	Cat#NA931, RRID:AB_772210
Monoclonal Anti-Biotin antibody produced in mouse clone BN-34	Sigma-Aldrich	Cat#B7653, RRID:AB_258625
Bacterial and virus strains
BL21(DE3) Competent E. coli	NEB	Cat#C2527I
One-Shot™ TOP10 E. coli	Invitrogen	Cat#K240020
Agrobacterium tumefaciens strain C58C1 (pGV2260)	Deblaere et al.54	N/A
Chemicals, peptides, and recombinant proteins
Abscisic acid, ABA	Sigma-Aldrich	Cat#A1049
2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide, Carboxy-PTIO, cPTIO	Thermofisher	Cat#C7912
4,5-Diaminofluorescein diacetate solution, DAF-2DA	Sigma-Aldrich	Cat#D225
S-Nitrosoglutathione, GSNO	Sigma-Aldrich	Cat#N4148
DL-Dithiothreitol, DTT	Sigma-Aldrich	Cat#D0632
Bleach (4–5% sodium hypochlorite)	Conejo	N/A
Triton 100X	Sigma-Aldrich	Cat#T8787
MURASHIGE & SKOOG MEDIUM + VITAMINS/MES	Duchefa Biochemie	Cat#M0255.0001
Sucrose	Sigma-Aldrich	Cat#S7903
PLANT AGAR	Duchefa Biochemie	Cat#P1001
Potassium hydroxide	Sigma-Aldrich	Cat#P1767
Recombinant protein bZIP67	This manuscript	N/A
Trizma® base	Sigma-Aldrich	Cat#T1503
Sodium chloride	Sigma-Aldrich	Cat#S7653
EGTA	Sigma-Aldrich	Cat#E3889
Immobilon-P Membran, PVDF, 0.45 μm	Millipore	Cat#IPVH00010
cOmplete Protease Inhibitor Cocktail	Roche	Cat#04693124001
Bio-Rad Protein Assay	Bio-Rad	Cat#5000001
Tween® 20	Sigma-Aldrich	Cat#P2287
Critical commercial assays
RNeasy Plant Mini Kit	QIAGEN	Cat#74904
ECL Advance Western Blotting Detection Kit	GE Healthcare Life Sciences	Cat#RPN2135
QuikChange II Site-Directed Mutagenesis Kit	Agilent	Cat#200521
Geneclean® Kit	MP Biomedicals	Cat# SKU: 111001200
Deposited data
bZIP67 mass spectrometry proteomics	This manuscript	PRIDE: PXD048872
Experimental models: Organisms
Col-0	NASC	N1093
bzip67	GABI-Kat	GABI314DO4
abi5-1	NASC	N8105
Nicotiana benthamiana	Saavedra et al.55	N/A
Oligonucleotides
Primers	This manuscript	See Table S1
Recombinant DNA
pENTR/D-TOPO	Invitrogen	Cat#K240020
pEarleyGate 201	Earley et al.56	N/A
pETM11	EMBL Protein Expression and Purification Facility	N/A
pGWB5	Nakagawa et al.57	N/A
Software and algorithms
ImageJ	Schneider et al.58	https://imagej.nih.gov/ij/
Office Excel Power Point 2016	Microsoft	N/A
GPS-SNO 1.0	Xue et al.1 (SI REFERENCES)	N/A
MaxQuant and Perseus software	Wendrich et al.44	N/A
BioRender	BioRender	https://www.biorender.com/

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Oscar Lorenzo (oslo@usal.es).

Materials availability

Plant lines generated in this study will be made available on request, but we may require a completed Materials Transfer Agreement if there is potential for commercial application.

Data and code availability

• •The mass spectrometry proteomics data of bZIP67 co-immunoprecipitation assays have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository (https://www.ebi.ac.uk/pride/) and are publicly available as of the date of publication. The dataset identifier PXD048872 is listed in the key resources table.
• •This paper does not report original code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Experimental model and study participant details

Plant materials

Arabidopsis thaliana ecotype Columbia-0 (Col-0) was the genetic background for the WT plants used in this work. The bzip67 transfer DNA insertion line from GABI-Kat (GABI314DO4) and seed stocks of abi5-1, cue1/nox1, cue1-6 and hot5-2 were obtained from the European Arabidopsis Stock Center (NASC; University of Nottingham, UK). The transfer DNA insertion line was confirmed by PCR and by germination assay in MS⁵⁹ solid medium supplemented with 5.25μM sulfadiazine (GABI-Kat). 35S:PRX IIE lines, prx IIe (SALK 064512) and nia1nia2noa1-2 mutant were a kind gift from Dra. María C. Romero-Puertas (EEZ-CSIC, Granada, Spain), Dr. Karl Josef-Dietz (University of Bielefeld-Germany)²¹ and Dr. José León (IBMCP-CSIC, Valencia, Spain).¹⁹ The ALKENAL REDUCTASE (aer) mutant and the 35S:FLAG-GSNOR transgenic line were kindly provided by Dr. Juan B. Barroso (University of Jaen, Spain).⁴⁷

Growth conditions

Arabidopsis plants were grown in a growth chamber at 15°C and 25°C or greenhouse at 21°C with a 16-h light/8-h dark photoperiod in pots containing a 1:3 vermiculite/soil mixture. For in vitro culture, Arabidopsis seeds were previously sterilized in 75% (v/v) sodium hypochlorite and 0.01% (v/v) Triton X-100 for 5min and washed 3 times in sterile water. Seeds were stratified for 3 days at 4°C under dark conditions for dormancy removal. Seeds were sowed on 10mM MES solid medium supplemented with 0.8% (w/v) agar and adjusted pH to 5.8 with KOH before autoclaving. NO scavenger (cPTIO), proteasome inhibitor (MG132) and cycloheximide were purchased from Sigma-Merck. NO donor (GSNO) was purchased from Calbiochem.

Method details

Generation of transgenic Arabidopsis plants

bZIP67 and bZIP67C106SC186SC215S were amplified using primers bZIP67-F and bZIP67-R (Table S1) and cloned into pEarleyGate 201⁵⁶ using GATEWAY technology. The constructs generated were used to transform the C58C1 (pGV2260) Agrobacterium strain.⁵⁴ The pbZIP67:GFP-bZIP67⁴⁰ construction was kindly provided by Dr. Sandra Bensmihen (CNRS-LIPM, Castanet-Tolosan, France).

35S:HA-bZIP67, 35S:HA-bZIP67C106SC186SC215S (w/o C) and pbZIP67:GFP-bZIP67 were used to generate Arabidopsis transgenic lines by the floral dip method⁶⁰ as described previously.⁵⁵ Seeds were plated on medium supplemented with the corresponding selection marker to identify T1 transgenic plants. Approximately 100 T2 transgenic seeds were again plated on selection medium, and those with 3:1 (resistant/sensitive) segregation ratio were selected, which correspond to one insertion lines. Finally, T3 homozygous seeds were selected by 100% resistance and used for further studies.

Detection of endogenous NO

Arabidopsis embryos were isolated from siliques and incubated in a 200μL solution containing 10μM of DAF-2DA (Sigma) in buffer 10mM Tris-HCl, pH 7.4 during 2h at 25°C in the dark. Embryos were then washed three times for 15 min with fresh 10mM Tris-HCl, pH 7.4 buffer. Finally, the fluorescence emitted by DAF-2DA was detected on a Leica magnifying glass by excitation at 495nm and emission at 515nm. NO depletion was also achieved by adding the scavenger cPTIO (1mM) to the solution.⁶¹

Real-time reverse-transcription PCR analysis

Total RNA for quantitative real time reverse transcription-PCR (qRT-PCR) was extracted from a pool of green silique collected from the entire Arabidopsis stem, as previously described⁶² with some variations. Samples were incubated with GSNO (1mM) and cPTIO (1mM) for 6h on liquid medium containing 10mM MES pH 5.8. Treatments were performed in six-well microplates under 80 r.p.m. and continuous light conditions. Frozen siliques were ground in liquid nitrogen using a mortar and pestle and 550μL of the extraction buffer (0.4M LiCl, 0.2M Tris, pH 8, 25mM EDTA, 1% SDS) and 550μL of chloroform was added to the tubes, which were then vortexed and incubated on ice. After centrifugation for 3 min at 15,800 g at 4°C, 1 volume of Tris-HCl saturated phenol plus 200μL of chloroform was added to the supernatant and mixed. Samples were centrifugated once more and then a 1/3 volume of 8M LiCl was added for RNA precipitation at −20°C overnight. After centrifugation for 30 min at 15,800 g at 4°C, 470μL diethyl pyrocarbonate (DEPC) treated water was added to the sample pellets and mixed with 7μL 3M NaAc, pH 5.2 and then washed with 70% ethanol and resuspended in 20μL DEPC treated water. RNA quality and concentration were tested by agarose gel electrophoresis and Spectrophotometer ND-1000 (NanoDrop). Then, 1-4μg RNA of each sample was incubated with RNase-free DNase I (Roche) for 30 min at 37°C, according to the manufacturer's instructions and heat inactivated.

Complementary DNA (cDNA) was synthesized using the SuperScript Kit (Sigma). qRT-PCRs were performed in ABI PRISM 7000 Sequence Detection System (Applied Biosystems, Framingham, MA, USA). Amplification was carried out using the “Brilliant SYBR Green qPCR MasterMix” (Stratagene), according to the manufacturer's instructions. The thermal program for SYBR Green real-time PCR was 95°C for 20s, followed by 40 cycles of 95°C for 1s and 60°C for 20s. To generate the standard curves, cDNA was diluted 10x and values were determined for each sample 4 times to ensure the slope of the standard curves and to determine the standard deviation (s.d.). The concentration of unknown samples was calculated using the ABI-Prism 7000 SDS software, which created threshold cycle values (C_t) and extrapolated relative levels of PCR product from the standard curve. The primers used are described in Table S1. ACTIN8 and EF1α were used as the control genes.

Generation of bZIP67 and PRX IIE expression vectors

The bZIP67 coding sequence was cloned into pETM-11 expression vector using EcoRI and NcoI enzymes. Klenow (Roche) was used to create 3′blunt ends, according to the manufacturer's protocol, and phosphorylated primers (bZIP67-F and bZIP67-R, Table S1) were generated to obtain fusion proteins with amino-terminal 6X His-Tag and transform Escherichia coli BL21 B834 (DE3) strain. Individual colonies were analyzed for the presence of the vector and positive clones were sequenced (Sequencing Facility, NUCLEUS, University of Salamanca).

PRX IIEΔSP and PRX IIEΔSPC121SC146S in pET28a were kindly provided by Dr. Karl Josef-Dietz (University of Bielefeld-Germany).

Site-directed mutagenesis

The site-directed mutagenesis of bZIP67 was performed using the QuickChange II Site-Directed Mutagenesis Kit (Stratagene Corporate) with the modified PCR protocol described in Edelheit et al.⁶³ Plasmid pETM11-bZIP67 was used as template and the primers (Table S1) were designed using online tools from Stratagene and synthesized by IDTDNA, Isogen and Metabiom. Mutations were confirmed by sequencing (Sequencing Facility, NUCLEUS, University of Salamanca).

Production of recombinant proteins and polyclonal antibodies

Wild type bZIP67 recombinant protein was expressed in Escherichia coli (Biomedal). The bacterial culture was induced at different conditions and the soluble and insoluble fractions were analyzed (Figure S3). Protein extraction was carried out using a denaturing buffer containing 8M urea, followed by on-column refolding (by reducing the urea concentration) and elution was performed using an imidazole gradient (10mM to 1M) in binding buffer without urea. Figure S3 shows the captured and eluted protein at low scales from the IMAC resin (High density nickel resin, ABT. Cat. 6BCL-QHNi-500). Final protein was solubilized in 20mM Tris-HCl pH 8.0, 3.8mM DTT, and 0.2% SDS to avoid precipitation and analyzed by western blot using and anti-His antibody (Genescript) (Figure S3).

Mutated bZIP67 recombinant proteins were expressed in Escherichia coli BL21 B834 (DE3) strain and purified using the BugBuster kit (Novagen) by Ni-NTA His Bind Resin (Novagen) according to the manufacturer's protocol. The induction conditions were selected depending on the best obtained results related to the quantity and purification protein values.

Polyclonal bZIP67 antibody was produced in one rabbit as described in Albertos et al.²

Western blotting

To carry out the western blot analysis, total protein was extracted from a pool of green siliques taken from an entire stem of each of the following backgrounds: Col-0, bzip67, prx IIe, 35S:PRX IIE, 35S:PRX IIE-GFP and pbZIP67:GFP-bZIP67 transgenic lines. pbZIP67:GFP-bZIP67 was treated with the proteasome inhibitor MG132 (100μM), GSNO (1mM) and cPTIO (1mM) for 6h in liquid 10mM MES medium adjusted to pH 5.8. The treatments were performed in six-well microplates under 80 r.p.m. in continuous light conditions. For bzip67;35S:HA-bZIP67 and related transgenic lines, total protein was extracted from the seed or seedlings treated with the proteasome inhibitor MG132 (100μM), ABA (2μM), GSNO (500μM) and cPTIO (100μM) for 96h on plates containing solid 10mM MES medium adjusted to pH 5.8.

Plant tissue was ground using a mortar and pestle and homogenized in 1 volume of extraction buffer (100mM Tris-HCl, 150mM NaCl, 0.25% NP-40) containing 1mM PMSF and 1X cOmplete EDTA-free proteases inhibitors (Sigma), followed by centrifugation for 10 min at 15,800 g at 4°C. Final protein concentration was determined by the Bio-Rad Protein Assay (Bio-Rad) based on the Bradford method.⁶⁴ 70μg of total protein was loaded per well in SDS-acrylamide/bisacrylamide gel electrophoresis using Tris-glycine-SDS buffer. Proteins were electrophoretically transferred to an Inmobilon-P polyvinylidene difluoride (PVDF) membrane (Millipore) using the Trans-Blot Turbo (Bio-Rad).

Membranes were blocked in Tris buffered saline-0.1% Tween 20 containing 5% blocking agent and probed with antibodies diluted in blocking buffer with 1% blocking agent. Anti-bZIP67 purified rabbit immunoglobulin (Biomedal, 1:100), anti-ABI5 purified rabbit immunoglobulin (Biomedal, 1:10,000), anti-GFP (Clontech, 1:2,000), anti-PRX IIE (kindly provided by Dr. Karl Josef-Dietz, 1:5,000), anti-Actin (Sigma-Merck, 1: 10,000), anti-Biotin (Sigma-Merck, 1:4,000), anti-His (Qiagen, 1:2,000), anti-HA (Roche-Merck, 1:2,000), anti-Myc (Abcam, 1:10,000) and ECL-Peroxidase-labelled anti-rabbit (Amersham, 1:20,000) and anti-mouse (Amersham, 1:20,000) antibodies were used in the western blot analyses. Detection was performed using ECL Advance Western Blotting Detection Kit (Amersham) and the chemiluminescence was detected using an Intelligent Dark-Box II, LAS-1000 scanning system (Fujifilm).

S-nitrosylation assays

To test the S-nitrosylated Cys residues, we used the biotin switch method,⁴¹ in which -SNO bound to the Cys residue of the protein of interest is converted into biotinylated groups, to detect S-nitrosylated proteins with slight modifications. S-nitrosylated proteins were detected in siliques and seedlings extracts and recombinant purified proteins.

For in vitro S-nitrosylation, purified bZIP67 recombinant proteins from the WT and the mutated versions were pretreated with the reducing agent DTT (20mM, Sigma) to obtain the same redox status in all samples (-SH). Next, proteins were treated with the NO donor GSNO (200μM) in the dark at room temperature for 30min with regular mixing. To check the reversibility of the modification, proteins were treated with a reducing agent (20mM DTT, Sigma) for 1h under the same conditions after incubation with GSNO. After each step, the reagents were removed by precipitation with 3 volumes of −20°C acetone, followed by centrifugation at 2,500g for 30 min at 4°C. For -SH blocking, recombinant proteins were incubated with 20mM S-methyl methanethiosulphonate (MMTS) and 2.5% SDS at 50°C for 30min with frequent mixing. The precipitated proteins were then dissolved in 100μL of RB buffer (25mM HEPES, 1mM EDTA and 1% SDS, pH 7.7). After addition of 1mM N-[6-(biotinamido) hexyl]-3ʹ-(2ʹ-pyridyldithio) propionamide (HPDP-biotin, Pierce, Rockford, IL) and 1mM ascorbic acid (a control without ascorbate is also included), the mixture was incubated for 1h at room temperature in the dark with intermittent vortexing. The resulting final recombinant proteins were loaded onto a 10% SDS-PAGE gel and transferred to a PVDF membrane for subsequent detection of bZIP67-SNO with anti-Biotin (Sigma-Merck, 1:4,000).

For in vitro S-nitrosylation by nitro-linolenic acid (NO₂-Ln, NO donor), the nitro-fatty acid was synthesized by a nitroselenation procedure, as previously described.³³ 5μg of bZIP67 recombinant protein were pretreated with the reducing agent DTT (20mM) in the dark at room temperature for 1h and then precipitated with 2 volumes of −20°C acetone for 1h. Then, bZIP67 recombinant protein was incubated with NO₂-Ln (500μM) in the dark at room temperature for 1h. For the controls, protein was treated with methanol (fatty acid vehicle) or linolenic acid (Ln) under the same conditions. After NO₂-Ln incubation, protein was treated with DTT (50mM) in the dark at room temperature for 1h to check the reversibility of this modification. In addition, before NO₂-Ln incubation, one aliquot of bZIP67 protein was pretreated with cPTIO (500μM) in the dark at room temperature for 30min. Blocking of the non-nitrosylated free cysteine residues was carried out by incubation with 30mM MMTS and 2.5% SDS at 50°C for 20min with frequent vortexing. After precipitation with cold acetone, proteins were resuspended in HENS buffer (25mM HEPES pH 8 buffer containing 1mM DTPA, 0.1mM neocuproine, and 1% SDS). Biotinylation was achieved by adding 1mM HPDP-biotin (ThermoFisher Scientific) and 1mM ascorbic acid and by incubating the samples in the dark at room temperature for 1h with occasional mixing. Biotin-labelled proteins were separated using a non-reducing 10% SDS-PAGE gel and then transferred onto PVDF membranes (Immobilon P, Millipore, Bedford, MA, USA) using a semi-dry transfer system (Bio-Rad Laboratories). PVDF membranes were blocked with Tris-buffered saline (TBS)+1% BSA. The blot was incubated with Pierce High Sensitivity NeutrAvidin-HRP (ThermoFisher Scientific) at a dilution of 1:30,000 for 1h, and immunoreactive bands were detected using a photographic film (Hyperfilm, Amersham Pharmacia Biotech) with an enhanced chemiluminescence kit (ECL-PLUS, Amersham Pharmacia Biotech).

In vivo S-nitrosylation of bZIP67 was carried out using a pool of green siliques taken from the entire Arabidopsis stem and 7-day-old seedlings in extraction buffer (100mM Tris-HCl, 150mM NaCl, 0.25% NP-40) containing 1mM PMSF and 1X cOmplete EDTA-free proteases inhibitors (Sigma). Protein extracts (1mg) were directly assayed by the biotin switch method. Samples treated with DTT (20mM) or GSNO (500μM) were kept for 1h in the dark and were used as negative and positive controls, respectively. In vivo biotinylated proteins were purified by immunoprecipitation with an IPA (protein A/G Ultralink Resin, Pierce) anti-biotin antibody for 2h at 4°C with 15μL IPA per mg of protein preincubated with 2μL of anti-biotin antibody (Sigma-Merck). Beads were washed three times with HEN buffer (100mM HEPES, 1mM EDTA and 0.1mM neocuproine, pH 7.8) and any bound proteins were eluted with 10mM DTT in SDS-PAGE solubilization buffer.

Mass spectrometry

In vitro biotinylated proteins with SDS-PAGE solubilization buffer were loaded onto a 10% SDS-PAGE gel, visualized using Brilliant Blue-G Colloidal Stain (Sigma) and the stained protein bands were excised and manually digested with trypsin or chymotrypsin in non-reducing conditions as previously described.² After digestion, the peptides were dried by speed-vacuum centrifugation and dissolved in loading solution (0.1% formic acid in water). Nano LC ESI-MS/MS (liquid chromatography coupled to electrospray tandem mass spectrometry) analysis was performed using an Eksigent 1D-nanoHPLC coupled to a 5600 TripleTOF QTOF mass spectrometer (Sciex, Framingham, MA, USA). The analytical column used was a silica-based reversed-phase column (Waters nanoACQUITY UPLC 75μm I.D. × 15cm, 1.7μm particle size). The trap column was an AcclaimPepmap, 100μm × 2cm, 5μm particle size, 100Å pore size, switched on-line with the analytical column. The loading pump delivered a solution of 0.1% trifluoroacetic acid in 98% water/2% acetonitrile (Scharlab, Spain) at 2 μL/min. The nanopump provided a flow-rate of 250 nL/min and was operated under gradient elution conditions, using 0.1% formic acid (Fluka, Switzerland) in water as mobile phase A and 0.1% formic acid in 100% acetonitrile as mobile phase B. Gradient elution was performed according to the following scheme: isocratic conditions of 96% A: 4% B for 1min, a linear increase to 50% B in 15min, a linear increase to 90% B in 30 s, isocratic conditions of 90% B for 5min and return to initial conditions in 30 s. The injection volume was 5μL. The LC system was coupled to the mass spectrometer via a nanospray source. Automatic data-dependent acquisition using dynamic exclusion allowed both the full scan (m/z 350–1250, 250msec) MS spectra followed by tandem MS CID spectra (100msec) to be obtained for the 15 most abundant ions per MS spectrum.

MS and MS/MS data were used to search against a customized target Arabidopsis thaliana protein database (31479 sequences) downloaded from UniprotKB. Database searches were done using a licensed version of Mascot v.2.5.1 and the search parameters were set as follows: methylthio-cysteine, Biotin-HPDP (cysteine), Gln->pyro-Glu (N-term Q), Glu->pyro-Glu (N-term E), deamidation (NQ) and oxidized methionine as variable modifications; peptide mass tolerance was set at 25 ppm and 0.1 Da for MS and MS/MS spectra, respectively, and 2 missed cleavages were allowed. The mascot score threshold for peptide identification was set to a value equal or higher than 20. All MS/MS spectra corresponding to biotin-HPDP modified peptides were analyzed manually.

Denitrosylation assays

For the in vitro denitrosylation assays, purified recombinant proteins bZIP67, PRX IIEΔSP and mutated PRX IIEΔSPC121SC146S were used. Purified protein of bZIP67 diluted in HEN buffer (25mM HEPES pH 7,7, 1mM EDTA, 0.1mM neocuproine) was incubated with 200μM GSNO in the dark at room temperature for 30min with regular mixing. The resulting protein-SNO was incubated for 1h together with 20mM DTT, reduced PRX IIEΔSP or PRX IIEΔSPC121SC146S. In order to detect the protein-SNO, a biotin switch method was performed as described above for the bZIP67 S-nitrosylation in vitro assay. Final recombinant proteins were loaded onto a 12% SDS-PAGE gel and transferred to a polyvinylidene difluoride membrane to subsequently detect bZIP67 and PRX IIEΔSP with anti-Biotin (Sigma-Merck, 1:4,000) and anti-His (Qiagen, 1:2,000) antibodies.

In vivo denitrosylation was carried out on 30-day-old-rosette leaves tissue in extraction buffer (100mM Tris-HCl, 150mM NaCl, 0.25% NP-40) containing 1mM PMSF and 1X cOmplete EDTA-free protease inhibitors (Sigma). Protein extracts (500 μg) were directly assayed using the biotin switch method. Samples treated with DTT (20mM) or GSNO (1mM) were kept for 1h in the dark and were used as negative and positive controls, respectively. Protein extracts from prx IIe or 35S:PRX IIE-GFP (intact or heat inactivated) were mixed with bzip67;35S:HA-bZIP67 and maintained for 1h in the dark. In vivo biotinylated proteins were purified by immunoprecipitation with an IPA (protein A/G Ultralink Resin, Pierce) anti-biotin antibody for 2h at 4°C with 15μL of IPA per mg of protein and then preincubated with 2μL of anti-biotin antibody (Sigma-Merck, 1:4,000). Beads were washed three times with HEN buffer (100mM HEPES, 1mM EDTA and 0.1mM neocuproine, pH 7.8) and any bound proteins were eluted with 10mM DTT in SDS-PAGE solubilization buffer.

Lipid profiles

For determining seed fatty acids, transmethylated lipids and fatty acids (1mg) were extracted from mature seeds according to Garcés and Mancha.⁶⁵ Fatty acids were cleaved from triacylglycerols and converted into the corresponding fatty acid methyl esters after the addition of 950μL of a methylation mixture (methanol:toluene:H₂SO₄, 34:6:1 v/v/v), 950μL heptane and heptadecanoic acid (17:0) as an internal standard, followed by incubation at 80°C for 1.5h. Once cooled, the mixture was vigorously shaken and the two phases were allowed to separate. The upper phase containing the methylated fatty acids were transferred to a vial and analyzed using an HP-6890 gas chromatograph (Palo Alto/CA) equipped with an SP2380 capillary column (30m length; 0.32mm i.d.; 0.20mm film thickness; Supelco, Bellefonte/PA). Operating conditions were as follows: H₂ as the carrier gas, flow 1 mL/min; injector at 240°C and detector at 250°C; column held for 2 min at 150°C and then programmed at 3 °C min⁻¹ to 210°C, with the injection split ratio set at 1:20. Quantification was performed by comparison of their retention times with those of known standards.

Arabidopsis embryo lipid bodies visualization

Arabidopsis embryos were dissected from green siliques, removed from the seed teguments and incubated in 1X PBS 5% formaldehyde solution for 30 min in darkness. Later, Nile Red (Thermofisher) was added (1 μg/mL) and the embryos were kept in darkness for 30 min. Then they were carefully washed at least three times with 1X PBS solution and then visualized in a Zeiss LSM510 confocal laser microscope using a 488 nm argon laser.

Co-immunoprecipitation and pull-down assays

The GFP pull-down assay was performed using pbZIP67:GFP-bZIP67 Arabidopsis transgenic lines and the WT control Col-0, where 3 replicates of 3g of silique tissue ground in liquid nitrogen, following the protocol described in Wendrich et al.,⁴⁴ were used. MS analysis was done on a Q Exactive Hybrid Quadrupole-Orbitrap mass spectrometer (Thermo Fisher). Full statistical analysis was performed using MaxQuant and Perseus software packages as described in Wendrich et al.⁴⁴

The coding sequence of PRX IIE was cloned into pGWB5 (35S:PRX IIE-GFP) vector⁶⁶ using primers PRX IIE-F and PRX IIE-R (-STOP) (Table S1) and GATEWAY technology. The construct was transiently expressed in Nicotiana benthamiana by agroinfiltration of transformed Agrobacterium tumefaciens C58C1 with pGV2260.⁵⁶ Additionally, p19 was added to avoid silencing.⁶⁷ For carrying out the haemagglutinin (HA) pull-down assays, proteins were extracted from Nicotiana benthamiana agroinfiltrated leaves, using the lysis buffer 100mM Tris-HCl, pH 7.5, 150mM NaCl, 0.1% Tween 20 or Arabidopsis 30-day-old-rosette leaves using the lysis buffer previously described⁶⁸ supplemented with 1mM PMSF and 1X cOmplete EDTA-free proteases inhibitors (Sigma). Extracts were cleared by centrifugation, and protein concentrations were determined using the Bradford assay. Then, 200μg (Nicotiana benthamiana) and 250μg (Arabidopsis 30-day-old-rosette leaves) of soluble protein were treated with or without GSNO (1mM) in darkness at room temperature during 1h. After treatment, the extracts were immunoprecipitated using anti-HA Affinity Matrix (Roche) for testing for bZIP67/PRX IIE interaction. Extracts and beads were incubated during 2h at 4°C. After incubation, the beads were washed three times in lysis buffer and the proteins were eluted from beads with 10mM DTT in SDS-PAGE solubilization buffer. The proteins were visualized using anti-HA (Roche-Merck, 1:2,000), anti-GFP (Clontech, 1:2,000) and ECL-Peroxidase-labelled anti-rabbit (Amersham, 1:20,000) antibodies.

bZIP67 subcellular localization in Arabidopsis thaliana

The subcellular localization of bZIP67 was observed using Arabidopsis transgenic lines carrying the pbZIP67:GFP-bZIP67 construction. Embryos were obtained by cutting the silique and gently squashing immature seeds in 10mM MES pH 5.8 buffer between a glass slide and a coverslip. In both cases a Leica TCS SP2 confocal microscope was used. GFP was excited using the 488nm argon laser and detection filters between 500 and 550nm. Nuclei were stained with SlowFade Diamond Antifade Mountant with DAPI (Thermofisher) and visualized immediately. Chlorophyll was differentiated by using a different channel with detection filters between 610 and 670nm.

Quantification and statistical analysis

Information about statistical analyses used and biological replicates in different experiments were stated in the corresponding Figure legends. Three biological replicates were analyzed to obtain the data, and values were expressed as mean ± SD. For the lipid profiles, the values are the mean of three determinations, and the statistical analyses were performed by a one-way ANOVA. Letters indicate significant differences (p ≤ 0.05). For the qRT-PCR, three biological replicates were analyzed by performing a one-way ANOVA, and different letters correspond to significant differences (p ≤ 0.05), except for Figures 5B and S9G, where biological replicates were analyzed by performing a T-Student statistic and asterisks indicate significant differences (p ≤ 0.05 p ≤ 0.01), expressed as ^∗ and ^∗∗, respectively. For the GFP pull-down assay full statistical analysis was performed using MaxQuant and Perseus software packages as described in Wendrich et al.⁴⁴

Published: April 11, 2024