Sultr4;1 mutant seeds of Arabidopsis have an enhanced sulphate content and modified proteome suggesting metabolic adaptations to altered sulphate compartmentalization

In recent decades, sulphur deficiency has become an increasing problem in crops of many countries, notably in Western Europe, leading to sulphur deficiency symptoms and resulting in decreased crop yields and quality parameters [1–3]. In this context, sulphur acquisition and metabolism in plants has become a major concern for research and crop improvement. Sulphur is an essential macronutrient required for plant growth. It is mainly acquired by the plant roots as inorganic sulphate and then distributed within the tissues [4]. After entering the cells, sulphate is reduced to sulphide with any excess sulphate being stored in the vacuole [5–7]. Sulphate assimilation leads to the synthesis of many compounds, including sulphur amino acids (cysteine and methionine), secondary products (glucosinolates and flavonoids) and numerous essential metabolites derived from sulphur-containing amino acids, such as cofactors (e.g. glutathione, S-adenosylmethionine, coenzyme A) and vitamins (e.g. biotin and thiamine). Therefore, sulphur assimilation is an essential part of plant metabolism and plays a crucial role in many plant processes. For example, as part of glutathione, sulphur has a role in the maintenance of the cellular redox balance and mitigation of oxidative stress in response to environmental changes. Furthermore, as constituent of sulphur amino acids, sulphur is tightly connected to protein synthesis and metabolism and is of great importance in terms of nutritional value. Indeed, animals and humans are unable to synthesize methionine and are dependent on dietary sources of this amino acid. Seeds of many agronomically important crops, including grain legumes, while rich in proteins, are deficient in the sulphur amino acids. Tabe and Droux [8] demonstrated that sulphate is the dominant form of sulphur found in the phloem supplying pods during lupin seed development and that the seed is able to reduce and assimilate sulphate in sufficient quantities for its needs. Therefore, identifying the mechanisms involved in sulphate acquisition and distribution in the developing seed is relevant to seed quality improvement.

The movement of sulphate around the plant and between cell compartments is facilitated by specific sulphate transporters (SULTR), which are encoded by a large gene family, consisting of 14 members in Arabidopsis and rice. Phylogenetic analyses indicate this gene family can be divided into four closely related groups (SULTR1 to 4) each with 12 membrane-spanning domains and a STAS (Sulphate Transporter and Anti-sigma Antagonist) domain at the carboxy-terminus [9], and a fifth more distinct group, SULTR5, lacking the STAS domain [4]. Notably, the Arabidopsis Sultr5;2 gene has recently been demonstrated to encode a high-affinity root molybdate transporter [10], which raises the question of the role of group 5 genes in sulphate transport. Considering differences in their kinetics and expression patterns, the different groups have been proposed to represent functional subtypes [11, 12]. Group 1 and 2 sulphate transporters, which are localised at the plasma membrane, have been the subject of several studies and are the best characterized groups. Members of group 1 represent high-affinity transporters that facilitate uptake of sulphate by the root (SULTR1;1 and SULTR1;2) or translocation of sulphate from source-to-sink organs (SULTR1;3) [13–17]. Group 2 is composed of low-affinity sulphate transporters whose gene products may rather play a role in vascular tissues, faciliting the translocation of sulphate around the plant [[4, 14] and [18]]. Group 3 is composed of low affinity transporters localized at the plasma membrane. The data available in the litterature indicate a differential expression of these genes in plant tissues, which is not stimulated by sulphur deficiency [19], and a role for SULTR3;5 in the root-to-shoot transport of sulphate in cooperation with SULTR2;1 in Arabidopsis [20].

Unlike groups 1 to 3, group 4 sulphate transporters have been localized to the vacuolar membrane: a SULTR4;1-GFP fusion protein was specifically accumulated in the vacuoles of roots and hypocotyls from young seedlings [21]. The Sultr4;1 gene was shown to be expressed in roots under sulphur-sufficient and deficient conditions, where it may play a role in the efflux of sulphate from the vacuolar lumen into the cytoplasm and influence the vacuolar storage capacity for sulphate [21]. In contrast, Sultr4;2 gene expression was shown to be highly inducible by sulphur limitation in the same tissue. The sultr4;1/sultr4;2 double knock-out mutants contained higher amounts of sulphate than did wild-type plants. Comparison of single and sultr4;1/sultr4;2 double knock-out mutants suggested that Sultr4;1 plays a major role and Sultr4;2 has a supplementary function [21]. Although sulphate transport has been extensively studied in roots, to date there has been no investigation of the roles of individual sulphate transporters within seeds. The vacuole may certainly play a role in the storage and unloading of sulphate within the developing seed, in which case SULTR4 transporters would be key players.

In this study, we demonstrate that, the Arabidopsis Sultr4;1 gene is strongly expressed in developing seeds and that its disruption significantly increases seed sulphate content, suggesting that SULTR4;1 is involved in the efflux of sulphate from vacuoles within developing seeds. Furthermore, a proteome analysis of Sultr4;1 mutant seeds reveals metabolic modulations suggesting adaptations to altered sulphate conpartmentation, which implicates SULTR4;1-mediated sulphate transport in establishment of defence mechanisms against oxidative stress during seed development.

In Arabidopsis, the sulphate transporter group 4 is composed of two members (SULTR4;1 and SULTR4;2), localized in the vacuolar membrane and reported to be expressed in roots, where a role was proposed in the efflux of sulphate from the vacuole [21]. In the present study, we observed that during the reproductive growth phase, Sultr4;1 was preferentially expressed in developing Arabidopsis seeds at the transition between embryogenesis and seed filling and presented a relative transcript abundance much higher than Sultr4;2, which was expressed at similar levels throughout seed development (see Figure 1 and Additional file 1). These data, along with previous data reporting that SULTR4;1 plays a major role and SULTR4;2 only a small contribution in remobilising the sulphate reserves [21], prompted us to study the role of SULTR4;1 in determining seed composition. We found a remarkably high level of sulphate constrasting with an unchanged sulphur content in mature seeds of a sultr4;1 T-DNA mutant (Figure 2), which for the first time implies a function for this transporter in developing seeds. The unchanged seed sulphur content indicates that the increased sulphate level is not related to an enhanced sulphur supply to the developing seed, but rather to a reduction of sulphate utilization/assimilation within the seed. Despite the unchanged sulphur content in mature seeds, we cannot exclude that the composition of sulphur or the amount of sulphate and other sulphur metabolites transported to developing seeds could be different between the wild-type and the sultr4;1 mutant. However, as no significant difference in seed yield, leaf surface or onset of flowering were observed, the increased sulphate/sulphur ratio in sultr4;1 seeds appears not to be related to a drastic perturbation of sultr4;1 vegetative growth. The increased sulphate/sulphur ratio in sultr4;1 seeds is possibly the consequence of a reduced remobilization (i.e. assimilation) of the vacuolar sulphate pool within the developing seed. The vacuolar localisation of the transporter in seeds has not yet been proven. We therefore cannot exclude that the protein could be localised in different compartments in seeds. However, the increased sulphate/sulphur ratio in mutant mature seeds is consistent with previous data showing that, in roots of sultr4;1 mutants, sulphate is retained as a consequence of defects in sulphate efflux from the vacuole [21]. It should also be noted, that other genes encoding sulphate transporters are expressed in developing seeds [22]. In particular, the Sultr3;4 gene encoding a putative plasma-membrane sulphate transporter is strongly expressed during early seed filling at similar stages to those of Sultr4;1. Interestingly, the seed proteome profile of the sultr4;1 mutant differs from that of a sultr3;4 mutant (unpublished data), indicating that SULTR4;1 and SULTR3;4 may have distinct roles/subcellular localization during seed filling.

Overall nitrogen and protein levels were unaltered in sultr4;1 seeds, but a dissection of the sultr4;1 seed proteome revealed that ~10% of the minor protein spots detected in 2D gels varied in abundance (see Additional files 5 and 6), thus reflecting metabolic modifications (Figure 5). In particular, several proteins with roles in the oxidative stress response were up-accumulated. An oxidative stress response has frequently been observed during sulphate deprivation [32, 33], notably in sulphur-deficient seeds [34]. In the present study, most of the proteins with roles in the oxidative stress response are involved in detoxification mechanisms (Figure 5). Among them were a scavenger of free oxygen radicals (superoxide dismutase, spot 2 in Figure 4), and the glutathione S-transferase isoform 6 (spot 122 in Figure 4) that may play a protective role against oxidative damage due to hydrogen peroxide [35]. This may represent an adaptive response to oxidative stress in sultr4;1 seeds as these enzymes detoxify some of the toxic compounds, such as reactive oxygen species (ROS), produced by oxidative stress.

At low levels, ROS participate in cellular signaling. In Arabidopsis seeds, abundant proteins were identified that act as scavengers of ROS generated during seed development, thereby counteracting their deleterious effects [36]. However, if ROS accumulate to harmful levels, seeds lose their ability to control ROS and cannot restart their metabolism [37], probably because ROS cause irreversible oxidative damage to lipids, proteins, and DNA. We have observed that Sultr4;1 seeds lose their viability during storage much more rapidly than wild-type seeds. Indeed, seeds collected from homozygous Sultr4;1 plants did not germinate after two years of storage at room temperature, while almost all seeds collected at the same time from wild-type plants germinated. A tetrazolium test confimed the loss of viability of the Sultr4;1 mutant seeds (see Additional file 9), which could be linked to their increased sensitivity to oxidative stress during storage, when ROS are known to be continuously produced [38]. However, the up-regulation of enzymes involved in ROS removal that we see in sultr4;1 seeds probably reflects seed developmental events and not changes during sample storage, as the seeds subjected to proteomics were frozen in liquid nitrogen immediately after harvest.

In summary, our data suggest sulphate remobilization from the vacuole to the other cell compartments is important for the seed's defence against abiotic oxidative stress during seed development and storage. This could also apply to other tissues, since in Arabidopsis roots and shoots, cadmium stress induces an up-regulation of Sultr4;1 gene expression concomitantly with an oxidative stress response [39]. Also, in Brassica juncea seedlings, Sultr4;1 gene expression was up-regulated in response to arsenic-induced stress [40, 41], which leads to the generation of ROS through the conversion of arsenate to arsenite [42].

Besides proteins related to the oxidative stress response, an up-accumulation of proteins involved in the biosynthesis of fatty acids and lipids was revealed. They correspond to enoyl- [acyl-carrier-protein] reductase (spot 59 in Figure 4), two isoforms of hydroxysteroid dehydrogenase 1 (HSD1, spots 130 and 202), and ketoacyl carrier protein synthase I (spot 235). An up-regulation of a ketoacyl carrier protein synthase was also observed in developing Arabidopsis seeds under sulphur-starved conditions [34]. It is well known that under stress conditions, toxic oxygen derivatives are produced that inactivate enzymes and damage important cellular components, such as membranes via lipid peroxidation and fatty-acid de-esterification. In this context, the up-regulation of enzymes involved in fatty acid and lipid biosynthesis may represent a mechanism to repair stress-induced membrane damage. In support of this, Li et al. [43] showed that transgenic Arabidopsis lines overexpressing HSD1 have an increased tolerance to salt stress.

Interestingly, an up-regulation in proteins of the energy and carbon metabolism was observed in sultr4;1 seeds. These are involved in glycolysis (glyceraldehyde-3-phosphate dehydrogenase, spot 187), photosynthesis or photorespiration (ribulose 1-5 biphosphate carboxylase large chain, spot 39), and anaerobic fermentation (alcohol dehydrogenase, spot 40 in Figure 4). Since an up-accumulation of enzymes involved in fatty acid and lipid biosynthesis was also observed in sultr4;1 seeds, it is possible that energy production and glycolysis were enhanced in sultr4;1 seeds to provide carbon skeletons for the synthesis of fatty acid precursors in these seeds [44]. Furthermore, members of the glycosyl hydrolases that specifically release sugar from oligosaccharides [β-glucosidase, spots 15, 38 (Figure 4), 83, 134; and β-galactosidase, spot 27 (Figure 5)] were up-accumulated in sultr4;1, probably as a way to sustain glycolysis and fatty acid biosynthesis.

This study demonstrated that the accumulation of storage proteins occurs in sultr4;1 mutant seeds (see Additional files 5). In 2D-gels, these proteins are located in the 20 to 30 kDa range (see Figure 4a and Additional files 4 and 5). Only a few storage proteins differ in abundance from wild-type seeds and these correspond to globulin 12S isoforms (Figure 5). The only protein whose level decreased was the At12S3 isoform (spot 239 in Figure 4a) and the three proteins whose level increased were At12S1 (spots 35 and 131) and At12S4 (spot 46). Interestingly, the isoforms whose level increased have a relatively poor sulphur amino acid content (2.4 and 2.8%, respectively) compared to At12S3 (4.3% of total amino acid residues) whose level decreased. A similar increase in sulphur-poor seed proteins was observed during sulphur deprivation [34, 45–47]. This indicates that, in sulphur-sufficient conditions, sulphate exported from the vacuole via SULTR4;1 participates but is not limiting for storage protein synthesis. A possibility to be considered is that SULTR4;2 could also participate to vacuolar sulphate release for storage protein synthesis. However, expression patterns of the gene and data from the literature suggest SULTR4;2 may have only a supplementary role in sulphate export from the vacuole [21].

Interestingly, an increased accumulation of a methionine synthase (spot 97, Figure 4) was observed in sultr4;1 seeds. It is also worth noting that under sulphate-deprived conditions, despite the almost complete depletion of cysteine, the plant is able to keep methionine levels constant [32, 48, 49]. This demonstrates the plant's capacity to regenerate methionine through alternative pathways under sulphate-deficient conditions and upon cysteine restriction. Methionine can be regenerated through methylation of homocysteine produced by transmethylation via a reaction catalyzed by S-adenosylhomocysteine hydrolase (Figure 5). Moreover, the use of the 4-carbon moiety of S-adenosylmethionine for the synthesis of polyamines and ethylene also is accompanied by recycling of the methylthio moiety and regeneration of methionine [50]. Interestingly, the protein spot identified in the present study corresponds to a cytosolic isoform involved in the regeneration of methionine from homocysteine produced in the course of the activated methyl cycle [51]. In sultr4;1 mutant seeds, the up-regulation of methionine synthesis could be a way to sustain seed protein synthesis. Moreover, maintaining the synthesis of proteins rich in methionine may be particularly important in stress conditions, since protein-endogenous methionine reacts with ROS, as a scavenger, to form sulphoxide (MetSO) without loss of activity of the corresponding protein [52, 53].

This study revealed a role for SULTR4;1 in determining sulphate content of mature Arabidopsis seeds, possibly because it controls the efflux of sulphate from the vacuoles for metabolism in other cell compartments [21]. Specific seed metabolic features were also revealed through a fine dissection of the seed proteome of a sultr4;1 mutant (Figures 4 and 5). In particular, an up-regulation of proteins involved in the oxidative stress response was revealed that suggests a function of SULTR4;1 in maintaining redox homeostasis during seed development. The maintenance of redox homeostasis is of crucial importance in plant tissues, and even more in seeds that have the ability to tolerate dehydration in the course of their development. This remarkable ability allows the orthodox seed, i.e., seed that can be stored in a state of low moisture content [54], to survive for several years in a desiccated state with metabolic activities at a standstill, until the environmental conditions become optimal for its germination. As seed desiccation enhances ROS formation, mechanisms that protect from oxidative stress are therefore crucial for maintaining cellular integrity and homeostasis [55]. Glutathione, which is a sulphur-derived product, is a major plant antioxidant. During desiccation, it is converted into the oxidized form glutathione disulphide (GSSG) and the balance between these two compounds is crucial for homeostasis and seed survival. Indeed, the GSSG versus GSH ratio can be applied to assess seed viability, since an increased ratio was typically correlated with viability loss in Pisum sativum seeds [54]. In the present study, we observed that mutant seeds, freshly collected at the dry mature stage, possessed higher GSSG levels, although the increases were not statistically significant (see Additional file 8). It is worth noting that generally, the level of GSSG is much higher in mature seeds than in other plant parts. For example, only 10% of total glutathione is oxidized in Arabidopsis seedlings [56, 57]. This indicates that the glutathione pool is relatively highly oxidized in mature Arabidopsis seeds compared to other plant organs. Despite the absence of significant changes in GSH and GSSG levels between wild-type and mutant seeds, we cannot exclude a localized disturbance of the redox balance in particular tissue or subcellular compartments of sultr4;1 mutant seeds. Indeed, as reported by Meyer et al. [58], measurements of GSSG/GSH ratios in plant tissues are prone to artefacts since they do not take into consideration the subcellular compartmentalization. Also, partial oxidation of glutathione during extraction is difficult to avoid and a slight increase in GSSG concentration, whilst barely detectable, would suffice to shift the glutathione redox potential to significantly less reducing values [59]. For these reasons, and because GSSG/GSH ratio in mature seeds is not representative of the events occurring at specific stages of seed development or during storage, future studies will be necessary to evaluate the redox status in different cell types throughout seed development and storage. Very little is known about the exact role of sulphate released from the seed vacuoles in the course of seed development in maintaining cellular homeostasis during seed development. Although further studies will be necessary, this paper is the first contribution on this topic.