SULTR2;1 Adjusts the Bolting Timing by Transporting Sulfate from Rosette Leaves to the Primary Stem

Abstract

Sulfur (S) is an essential macronutrient for plant growth and metabolism. SULTR2;1 is a low-affinity sulfate transporter facilitating the long-distance transport of sulfate in Arabidopsis. The physiological function of SULTR2;1 in the plant life cycle still needs to be determined. Therefore, we analyzed the sulfate transport, S-containing metabolite accumulation and plant growth using Arabidopsis SULTR2;1 disruption lines, sultr2;1–1 and sultr2;1–2, from seedling to mature growth stages to clarify the metabolic and physiological roles of SULTR2;1. We observed that sulfate distribution to the stems was affected in sultr2;1 mutants, resulting in decreased levels of sulfate, cysteine, glutathione (GSH) and total S in the stems, flowers and siliques; however, the GSH levels increased in the rosette leaves. This suggested the essential role of SULTR2;1 in sulfate transport from rosette leaves to the primary stem. In addition, sultr2;1 mutants unexpectedly bolted earlier than the wild-type without affecting the plant biomass. Correlation between GSH levels in rosette leaves and the bolting timing suggested that the rosette leaf GSH levels or limited sulfate transport to the early stem can trigger bolting. Overall, this study demonstrated the critical roles of SULTR2;1 in maintaining the S metabolite levels in the aerial part and transitioning from the vegetative to the reproductive growth phase.

Introduction

Sulfur (S) is an essential macronutrient for plant growth and metabolism (Long et al. 2015). Plants take up S from the soil as sulfate through the function of several sulfate transporters (SULTRs) (Takahashi et al. 2011, Long et al. 2015). Sulfate absorbed by the cells is converted to 5ʹ-adenylyl sulfate (APS) and reduced to sulfite by APS reductase. Then, sulfate is reduced to sulfide by sulfite reductase, and the resulting sulfide reacts with O-acetyl-l-serine to form cysteine (Long et al. 2015). Cysteine is the primary product of sulfate assimilation and is used for the synthesis of various organic S-containing compounds, such as glutathione (GSH), methionine and glucosinolates (Saito 2004, Hell and Wirtz 2011, Takahashi et al. 2011, Long et al. 2015, Maruyama-Nakashita 2017).

Twelve SULTRs in Arabidopsis thaliana are classified into four groups according to their protein sequences (Takahashi et al. 2011), and their biochemical characteristics, tissue-specific expression and physiological functions have been studied (Takahashi et al. 2011, Takahashi 2019). Among them, group 2 low-affinity transporters, SULTR2;1 and SULTR2;2, facilitate the long-distance transport of sulfate due to their vascular tissue–specific expressions. SULTR2;1 is expressed in xylem parenchyma and pericycle cells in roots and the phloem in shoots (Takahashi et al. 1997, 2000), as well as in the funiculus, the connecting part between seeds and siliques (Awazuhara et al. 2005), whereas SULTR2;2 is predominantly expressed in shoot phloem (Takahashi et al. 2000).

Besides intensive studies on the functions of SULTRs, few reports evaluated the effects of their disruption on plant growth and development. Disruption of SULTR1;1 and SULTR1;2 resulted in severe growth defects under S-sufficient and S-deficient conditions because of the insufficient sulfate uptake (Yoshimoto et al. 2007, Barberon et al. 2008). Disruption of SULTR3 resulted in early bolting and an increase in 100 seed weight compared to the wild-type (WT) plants (Zuber et al. 2010). Additionally, the quintuple mutant of group 3 SULTRs showed noticeable growth retardation with smaller rosette leaves and shorter roots, while those were unaffected in each single SULTR3 mutant (Chen et al. 2019). These results suggest that other disruption lines of SULTRs would affect the growth and development of the plant.

Certain S-containing compounds such as sulfate, cysteine and GSH regulate sulfate uptake and distribution in plants (Hell and Wirtz 2011). Sulfate deprivation is one of the main factors affecting plant growth, quality and yield (Etienne et al. 2018). GSH is a well-known reductant in the plant cell, contributing to the detoxification of reactive oxygen species and to the heavy metal detoxification using phytochelatin, which is synthesized from GSH (Freeman et al. 2004, Foyer and Noctor 2009, Noctor et al. 2011, 2012, Cheng et al. 2015). In addition, GSH participates as a plant growth regulator in various developmental stages of plants, such as embryo development, root initiation, pollen germination and flower development (Vernoux et al. 2000, Cairns et al. 2006, Schnaubelt et al. 2013, Zechmann et al. 2011, Hatano-Iwasaki and Ogawa 2012, García-Quirós et al. 2019).

Previous studies demonstrated the effects of SULTR2;1 disruption on S metabolism and sulfate distribution in Arabidopsis. SULTR2;1 disruption affected the sulfate translocation from root to shoot and sulfate translocation from old to young leaves (Liang et al. 2010, Kawashima et al. 2011). In SULTR2;1 antisense lines, sulfate, cysteine and GSH concentration decreased in seeds when grown under S-sufficient conditions, indicating the SULTR2;1 role in sulfate transport to seeds (Awazuhara et al. 2005). However, the physiological function of SULTR2;1 during seedling growth to seed maturation is to be analyzed. Therefore, we examined the metabolic phenotypes of the SULTR2;1 disruption lines, sultr2;1–1 and sultr2;1–2, at the whole plant level and observed their growth phenotypes from seedling to mature growth stages to clarify the physiological role of SULTR2;1. We observed that S-containing metabolite levels and sulfate distribution to the above-ground parts were affected in sultr2;1. In addition, sultr2;1 bolted earlier than the WT without the plant biomass being affected. This suggested the vital role of SULTR2;1 in maintaining the S metabolite levels in the plant aerial part and regulation of bolting timing.

Results

Sulfate distribution to the aerial part was affected in mature sultr2;1 mutants

To analyze the physiological function of SULTR2;1, we first isolated a T-DNA insertion line (Supplementary Fig. S1A). The T-DNA mutant line was selected using specific primers in genomic PCR and RT-PCR analysis (Supplementary Table S1). The T-DNA insertion position was detected at the lower part of the third exon. After confirming the homozygosity with genomic PCR and RT-PCR, we named the T-DNA insertion mutant sultr2;1–1.

We analyzed the sulfate uptake and distribution in sultr2;1–1 and the WT, Col-0, at different growth stages before and after bolting as well as in 6-week-old mature plants using [³⁵S] sulfate as a tracer to demonstrate the effects of SULTR2;1 disruption on sulfate distribution in plants (Fig. 1). By measuring the [³⁵S] levels on the imaging plates, we quantified sulfate uptake and the distribution to aerial parts. The sulfate uptake and distribution were similar between sultr2;1–1 and the WT before and after bolting stages; however, sultr2;1–1 sulfate uptake and distribution in the mature plants were almost half of those in the WT (Fig. 1). In more detail, the [³⁵S] distribution in roots and rosette leaves of mature sultr2;1–1 mutant was similar to that in the mature WT plants, and it decreased in primary and lateral stems, flowers and siliques compared to the WT (Supplementary Fig. S2).

Fig. 1

Sulfate uptake and distribution in sultr2;1–1 and the WT. Sulfate uptake and distribution in 3-week-old plants before bolting (top, n = 6) and 4-week-old plants after bolting (middle, n = 5) and 6-week-old mature plants (bottom, n = 6). The plants were grown on MGRL agar medium at 22°C under an 18-h light and 6-h dark cycle before and after bolting. For mature plant analysis, the seedlings were grown on MGRL agar media for 2 weeks and transferred to the MGRL hydroponic system for 4 weeks. Plants were treated with [³⁵S] SO₄²⁻ in MGRL liquid media for 1 h (before bolting), 3 h (after bolting) and 6 h (mature plants) before being transferred to the imaging plates for 3 d. The left panel shows the plants scanned with a FLA7000 laser scanner. The plant parts were collected separately, weighed and extracted with 500 µl of 10 mM HCl. The [³⁵S] levels were determined using a liquid scintillation counter after mixing the extract with 2 ml of liquid scintillation counting cocktail. The values and error bars represent averages + standard error (SE). The asterisks indicate significant differences (Student’s t-test; *0.05 ≤ P < 0.1, **0.01 ≤ P <0.05) between sultr2;1–1 and the WT.

We observed the GFP fluorescence in P_SULTR2;1-GFP plants harboring a 2535-bp upstream region and a 1077-bp downstream region of SULTR2;1 fused 5ʹ-upstream and 3ʹ-downstream of sGFP, respectively, to confirm the tissue-specific expression of SULTR2;1 in mature plants (Maruyama-Nakashita et al. 2015). P_SULTR2;1-GFP and WT seeds were sown on the vermiculite, and the GFP fluorescence was examined at different growth stages, i.e. before bolting, after bolting and mature plants (Fig. 2, Supplementary Fig. S3). GFP fluorescence was detected in rosette leaf veins and early primary stems in plants before bolting. After bolting, the fluorescence was more intense in the leaf veins, especially in the main veins directly connected to the primary stem base, and was detected in the upper part of the primary stems, flowers and cauline leaves. In the mature plants, lateral stems and siliques showed fluorescence in addition to those parts detected in the plants after bolting, with the primary stem base and lateral stem bases showing the most intense fluorescence.

Fig. 2

GFP fluorescence in P_SULTR2;1-GFP plants. Three-week-old (before bolting, top), 4-week-old (after bolting, middle) and 6-week-old (mature, bottom) plants were grown on the vermiculite and supplied with half-strength MGRL solution twice a week at 23°C and 24-h light at 33 µmol m⁻²s⁻¹ light intensity. Plant images (left) were taken with an iPhone, and the plants were scanned using an Amersham Typhoon scanner for visualizing GFP and autofluorescence (right). Scale bar: 1 cm.

These results indicated that SULTR2;1 contributed to sulfate distribution among the aerial parts of mature plants, especially the distribution from rosette leaves to the primary stem.

SULTR2;1 disruption reduced S-containing metabolite levels in aerial parts except for the GSH levels in the rosette leaves

We isolated another SULTR2;1 disruption mutant using a CRISPR-Cas9 system to increase the credibility of the data (Supplementary Fig. S1B). The gRNA sequence in the upper region of the third exon was selected, cloned into the binary plasmid (Ueta et al. 2017) and transformed into the WT plants using Agrobacterium. The genome-edited line was selected from the transgenic plants by sequencing the corresponding genomic DNA region after PCR amplification with specific primers (Supplementary Table S1). The segregated homozygous mutant line was named sultr2;1–2 (Supplementary Fig. S1B).

As sultr2;1–1 had decreased sulfate distribution in aerial parts, we expected decreased S-containing metabolite levels in both mutants. To investigate the S metabolic changes, we collected rosette leaves at 28 d after seed sowing (DAS), primary and lateral stems, flowers and siliques from both sultr2;1 mutants and the WT at 42 DAS and analyzed sulfate, cysteine, GSH and total S concentration (Fig. 3). As expected, the sultr2;1 mutants had decreased metabolite levels in most parts of the plants compared to those in the WT, with the highest decrease in the stems. Interestingly, the total GSH contents in rosette leaves were increased to almost 4-fold in the mutants compared to that of the WT (Fig. 3).

Fig. 3

Sulfate, cysteine, GSH and total sulfur (S) contents in sultr2;1 and the WT. Plants were grown on the vermiculite and supplied with half-strength MGRL solution twice a week at 23°C and 24-h light at 33 µmol m⁻²s⁻¹ light intensity. Sulfate, cysteine, GSH and total S levels (top to bottom) in the rosette leaves (28 DAS), primary and lateral stems, flowers and siliques (42 DAS) (left to right) were analyzed. The values and error bars represent averages + SE (n = 5). The asterisks indicate significant differences detected by one-way ANOVA followed by Dunnett’s test between each sultr2;1 mutant and the WT (**0.01 ≤ P < 0.05, *** P < 0.01).

These results indicated that SULTR2;1 contributed to sulfate distribution among the aerial part of mature plants, especially distribution from rosette leaves to stems, and increased the GSH accumulation in the rosette leaves.

The sultr2;1 mutants bolted earlier than the WT without reducing the plant biomass

We observed the growth phenotypes of sultr2;1–1 and sultr2;1–2 throughout the life cycle to investigate the function of SULTR2;1 in plant growth and development. The seeds of sultr2;1–1, sultr2;1–2 and the WT were sown on vermiculites and supplied with half-strength mineral nutrient (MGRL) solution twice a week. In the seedling stages, sultr2;1 mutant growth was comparable to that of the WT (Supplementary Fig. S4). Interestingly, sultr2;1 mutants started bolting 23 DAS, 2 d earlier than the WT (Fig. 4A). Due to the early bolting, the plant height of sultr2;1 mutants was higher than that of the WT on 35, 40 and 45 DAS and was comparable after 50 DAS (Fig. 4B, C). As nutrient deficiency sometimes causes an early flowering phenotype with a reduced number and size of rosette leaves, we analyzed the number of rosette leaves before bolting (21 DAS) and the day of first bolting in the mutants and the WT, 23 and 25 DAS, respectively (Fig. 4D, E). They were similar among these DAS and between sultr2;1 mutants and the WT. In addition, sultr2;1 mutants also demonstrated a similar number of lateral stems and siliques and a similar dry weight (DW) of the stems, seeds and 100 seeds to the WT (Supplementary Fig. S5). These results indicated that the sultr2;1 mutants maintained the rosette leaf and total plant biomass despite the early bolting phenotype.

Fig. 4

Bolting stimulation in sultr2;1 mutants without loss of plant biomass. (A) The rate of plants having a primary stem more than 0.5 cm (n = 35). (B) Typical observation of bolted plants 35 DAS. Scale bar: 1 cm. (C) Plant height after bolting (n = 25). (D) Number of rosette leaves (RL) (n = 40) at 21, 23 and 25 DAS. (E) FW of RL (n = 5) before (22 DAS) and after (27 DAS) bolting. Plants were grown on the vermiculite and supplied with half-strength MGRL solution twice a week at 23°C and 24-h light at 33 µmol m⁻²s⁻¹ intensity. The values and error bars represent averages + SE. The asterisks indicate significant differences detected by one-way ANOVA followed by Dunnett’s test between each sultr2;1 and the WT (**0.01 ≤ P < 0.05, ***P < 0.01).

Correlation between the GSH levels in rosette leaves and bolting timing

Due to the increased GSH in rosette leaves of sultr2;1 mutants, we hypothesized that GSH accumulation in rosette leaves correlated with bolting. To test this hypothesis, we analyzed the total GSH levels in rosette leaves of sultr2;1 mutants and the WT from 21 to 27 DAS (Fig. 5A). The sultr2;1 mutants accumulated approximately 200 pmol mg DW⁻¹ GSH in the rosette leaves from 21 to 23 DAS, which was almost two-fold that in the WT. After 23 DAS, the first bolting day for sultr2;1 mutants, GSH levels in sultr2;1 rosette leaves increased to about 450 pmol mg DW⁻¹ on 24 DAS and gradually decreased then (Fig. 5A). In contrast, GSH levels in WT rosette leaves continuously increased from approximately 100 pmol mg DW⁻¹ at 21 DAS to approximately 300 pmol mg DW⁻¹ at 26 DAS, one day after the first bolting. The GSH levels between the rosette leaves in sultr2;1 and those in the WT were almost equivalent, approximately 200 pmol mg DW⁻¹, upon the start of bolting at 23 and 25 DAS, respectively.

Fig. 5

GSH contents and the related gene expression levels in rosette leaves of sultr2;1 mutants and the WT. (A) GSH levels in the rosette leaves in 21–27 DAS plants were analyzed using an HPLC fluorescence system. The gray and white arrows indicate the first bolting dates in sultr2;1 mutants and the WT, respectively. (B) The transcript levels of SULTR2;1, OPT6, GSH1 and GSH2 genes on 21, 23, 25 and 27 DAS were determined by qRT-PCR. Plants were grown on the vermiculite and supplied with half-strength MGRL solution twice a week at 23°C and 24-h light at 33 µmol m⁻² s⁻¹ intensity. The values and error bars represent averages + SE (n = 3). The asterisks indicate significant differences detected by one-way ANOVA followed by Dunnett’s test between each sultr2;1 mutant and the WT (*0.05 ≤ P < 0.1, **0.01 ≤ P < 0.05, ***P < 0.01).

To clarify the reason for the increased GSH levels in the mutants’ rosette leaves and the relation among SULTR2;1 disruption, GSH accumulation and early bolting phenotype, we analyzed the transcript levels of SULTR2;1, OPT6, GSH1 and GSH2 at 21, 23, 25 and 27 DAS (Fig. 5B). SULTR2;1 transcript was stable among these DAS in the WT. Meanwhile, the transcript levels of GSH1 and GSH2 in the sultr2;1 mutants were higher than those in the WT at 21, 23 and 25 DAS. Their levels were decreased along with DAS to similar levels to those in the WT at 27 DAS, which supported the increased GSH levels in the mutants’ rosette leaves. The OPT6 expressions were also higher in sultr2;1 mutants than in the WT at 21, 23 and 25 DAS. The expressions peaked at 23 DAS, the first bolting day, in the mutants, but those were stable in the WT.

Since the reduction rate of GSH regulates plant development (Vernoux et al. 2000, Cairns et al. 2006, Noctor et al. 2011, 2012, Zechmann et al. 2011, Hatano-Iwasaki and Ogawa 2012, Schnaubelt et al. 2013, García-Quirós et al. 2019), we analyzed the reduced and oxidized GSH levels in the same samples used in Fig. 5 (Supplementary Fig. S6). Unexpectedly, the reduced GSH rate (%) was similar between sultr2;1 mutants and the WT. These results indicated that the total or reduced GSH level, not the rate of GSH reduction, in the rosette leaves positively correlated with the bolting timing.

It is widely accepted that bolting and flowering timing is regulated by a combination of environmental and genetic factors (Xu et al. 1997, Putterill et al. 2004, Sanchez et al. 2012, Bluemel et al. 2015, Grossniklaus 2015, Soyars et al. 2016, Cao et al. 2021). Environmental signals such as light and temperature are integrated by SUPPRESSOR OF OVEREXPRESSION OF CONSTANS 1 (SOC1) and FLOWERING LOCUS T (FT) in Arabidopsis (Soyars et al. 2016, Cao et al. 2021). Several transcription factors, including CONSTANS (CO), promote FT expression, and the FT protein is transported to the shoot apex, forms a complex with FLOWERING D (FD) and stimulates flowering through the transcriptional induction of APETARA 1 (AP1) and LEAFY (LFY) (Abe et al. 2005, Cao et al. 2021). SOC1 expression is stimulated by both the FT-FD complex and CO, and SOC1 protein stimulates the expression of AP1 and LFY (Lee and Lee 2010). Suppression of SOC1 and FT is induced by FLOWERING LOCUS C (FLC), the SHORT VEGETATIVE PHASE complex, as well as TEMPRANILLO1 (TEM1) and TEM2 (He et al. 2020). We analyzed their transcript levels in the rosette leaves to confirm the involvement of these flowering-related genes in the early bolting phenotype of sultr2;1 mutants (Supplementary Fig. S7). The transcript levels of all tested genes, FT, SOC1, LFY, CO, FLC, TEM1 and TEM2, were similar between sultr2;1 mutants and the WT.

Discussion

Sulfate taken up by the roots is transported to the shoot via the xylem and then distributed to various tissues and sink organs (Takahashi et al. 2011, Long et al. 2015). SULTR2;1 is responsible for the long-distance, i.e. root-to-shoot sulfate transport through the retrieval of apoplastic sulfate to xylem parenchyma and pericycle cells (Takahashi et al. 2011), sulfate transport to seeds by being expressed at the funicles (Awazuhara et al. 2005) and sulfate transport from old to young leaves via the phloem (Liang et al. 2010, Kawashima et al. 2011). This study demonstrated that sultr2;1 mutants reduced sulfate distribution to the parts above the rosette leaves, indicating the role in sulfate transport from rosette leaves to the primary stem (Fig. 1, Supplementary Fig. S2). The SULTR2;1 expression in rosette leaf veins and the base of the primary stem also supports this conclusion (Fig. 2, Supplementary Fig. S3). Considering the SULTR2;1 expression in xylem parenchyma and phloem cells in leaves and its contribution in sulfate transport from old to young leaves, SULTR2;1 should transport sulfate from rosette leaves to the early primary stem through phloem (Takahashi et al. 2000, Liang et al. 2010). As a result, sulfate, cysteine, GSH and total S levels were decreased in stems, flowers and siliques (Fig. 3). GFP fluorescence derived from P_SULTR2;1-GFP plants was intense in the base of lateral stems and the cauline leaves, suggesting that SULTR2;1 controls the sulfate transport preferred to the lateral stems rather than the upper part of the primary stem (Sugita et al. 2016).

After being absorbed by the cells, part of the sulfate is transported into the vacuole as tentative storage or into the chloroplast/plastid for reduction and assimilation into cysteine (Takahashi et al. 2011, Long et al. 2015). GSH is synthesized from cysteine by γ-glutamylcysteine (γ-EC) synthase (γ-ECS, GSH1 in Arabidopsis) and GSH synthase (GSH2, Noctor et al. 2011, 2012, Lu 2013). Due to the localization of GSH1 in plastids and that of GSH2 in plastids and the cytosol, the main site of GSH synthesis is considered as plastids. Since S is stored and transported as GSH, the GSH levels are influenced by the S levels in plants (Noctor et al. 2011, 2012, Maruyama-Nakashita 2017). GSH is distributed to all subcellular compartments in the plant cell with the help of several γ-EC, GSH and oxidized GSH transporters such as CRT-LIKE TRANSPORTER (Zechmann and Müller 2010, Noctor et al. 2011, Heyneke et al. 2013, Zechmann 2014). In the rosette leaves of sultr2;1 mutants, only GSH levels increased, while sulfate and cysteine were decreased (Fig. 3). Although there is no direct evidence to explain the mechanism of the increased GSH, the insufficient sulfate transport to the primary stem may cause the oxidative stress in the rosette leaves and induce GSH synthesis (Noctor et al. 2011, 2012, Cheng et al. 2015), which is supported by the higher transcript levels of GSH1 and GSH2 in the mutants’ rosette leaves (Fig. 5B). Spatiotemporal increase of sulfate and cysteine in plastids undetectable with the analysis using whole tissues as in Fig. 3 could be another explanation for the inconsistency between the levels of sulfate, cysteine and GSH when the metabolic flow is shifted to the GSH synthesis as well as the increased storage of GSH in the rosette leaf vacuole. These possibilities need experimental validation.

After the transition from the vegetative to reproductive growth stage, the physiological status of rosette leaves shifts to export solutes to the reproductive tissues, which reduces the vacuolar function for sulfate storage and increases the GSH synthesis for the transport of organic S through the phloem (Hawkesford and de Kok 2006, Tegeder and Rentsch 2010, Breeze et al. 2011). The sum of sulfate, cysteine and GSH contents was 52 ± 1, 40 ± 1 and 28 ± 2 nmol mg DW⁻¹, while total S contents were 1,082 ± 81, 885 ± 30 and 842 ± 25 nmol mg DW⁻¹ in rosette leaves of the WT, sultr2;1–1 and sultr2;1–2, respectively, implying that the remaining S levels were 1,030, 845 and 816 nmol mg DW⁻¹, respectively (Supplementary Table S3). The remaining S, being attributed to mainly protein and glucosinolates in Arabidopsis, was lower in sultr2;1 mutants than in the WT, indicating that the increase of GSH is an exception among S-containing compounds in the rosette leaves. In addition, the effects of SULTR2;1 disruption on partitioning the S-containing compounds among the plant tissues could vary according to the developmental stages. In SULTR2;1 antisense lines, GSH levels in the rosette leaves and seeds of 6-week-old plants grown under S-sufficient conditions and 16-h light/8-h dark cycles decreased to 70–80% and 60–70% of the WT, respectively (Awazuhara et al. 2005). In this study, we observed increased GSH levels at earlier stages, 21–28 DAS (Figs. 3, 5A), the transcript levels of GSH synthetic enzymes, GSH1 and GSH2, in sultr2;1 mutants went down along with the DAS and became similar levels to that in the WT (Fig. 5B) and GSH levels in the rosette leaves could decrease to the maturation.

We observed that sultr2;1 mutants bolted earlier than the WT (Fig. 4A), suggesting that SULTR2;1 modulates the growth transition from the vegetative to reproductive phase. Although there are still many possibilities on the background of this phenotype, we observed a correlation between GSH levels in rosette leaves and the bolting timing (Fig. 5). The rosette leaf GSH levels were similar between the sultr2;1 mutants and WT upon the start of bolting. Hence, the early bolting phenotype of sultr2;1 mutants is potentially due to the reduced sulfate transport to the growing primary stem or increased GSH levels in the rosette leaves. The ratio of sulfate or S-containing metabolites between the rosette leaves and shoot meristem can trigger bolting.

In addition to the genetic backgrounds, many stresses, including nutrient stress, generally hasten the flower bud appearance and promote early flowering in Arabidopsis (Kola´r and Senˇkova 2008, Byeong-ha 2009, Xu et al. 2014, Kazan and Lyons 2016, Rankenberg et al. 2021, Zhang et al. 2023a). The early flowering phenotype often accompanies the reduced number of rosette leaves and lower yield. In contrast, the sultr2;1 mutants maintained the plant biomass such as fresh weight (FW) before bolting, rosette leaf number and FW, the final plant height and seed yield (Fig. 4D, E, Supplementary Figs. S4, S5). Based on the reduced sulfate transport to the shoot (Fig. 1) and the reduced S and S metabolite levels in the stems and reproductive tissues (Fig. 3), it is presumed that local S deficiency occurs in the early primary stem, which may eventually stimulate bolting. However, the transcript levels of SULTR2;1 do not correlate with bolting and floral transition compared to whole plant levels (Breeze et al. 2011, Del Prete et al. 2019, Hinckley and Brusslan 2020). Therefore, further study is needed to elucidate the local S status and its influences on the growth stimulation of the early primary stem.

Although we demonstrated the correlation between the rosette leaf GSH levels and bolting initiation, the bolting timing and GSH levels in the rosette leaves are not always correlated. The pad2-1, a γ-ECS (GSH1) mutant with reduced GSH levels (Parisy et al. 2006), bolted earlier than the WT (Col-0), while overexpression of γ-ECS delayed flowering compared to that in the WT (Cheng et al. 2015). The treatment of the WT with l-buthionine sulfoximine (BSO), a specific inhibitor of γ-ECS, stimulated bolting (Cheng et al. 2015). The opt6 mutants, in which long-distance transport of GSH via phloem is disrupted, delayed bolting with slightly higher GSH levels in one of two mutants’ rosette leaves compared to that of the WT (Wongkaew et al. 2018). A late flowering mutant, fca-1, accumulated higher GSH than the WT, and inhibition of GSH synthesis stimulated flowering in this mutant even though the GSH levels in the rosette leaves did not change significantly (Ogawa et al. 2001). These reports contradict our results, which might be because of the difference in light conditions as all aforementioned experiments were performed under long-day conditions, 16-h light/8-h dark cycles, while we did under 24-h light conditions. The light conditions greatly influence the bolting timing, GSH synthesis and redox status (Noctor et al. 2011, 2012, Considine and Foyer 2014, Geigenberger and Fernie 2014, Shim and Imaizumi 2014, Grossniklaus 2015, Kazan and Lyons 2016, Cejudo et al. 2019, Cao et al. 2021, Zhang et al. 2023a).

In addition, in cad2-1, another mutant allele of γ-ECS (GSH1, Cobbett et al. 1998), delayed flowering was rescued by the exogenous GSH application and extended by the treatment with BSO (Ogawa et al. 2001). Similarly, the addition of GSH-promoted bolting and vernalization-induced bolting was canceled by BSO in Eustoma grandiflorum (Yanagida et al. 2004). However, the vernalization itself did not influence the GSH levels. The application of 0.4 mM FeSO₄ to the agar medium stimulated bolting, probably through the stimulation of ferredoxin-NADP⁺-oxidoreductase activity followed by increased NADPH/NADP⁺ and ATP/AMP ratios, resulting in the increased stability of cryptochrome, a blue light receptor responsible for the extended bolting by high nitrogen (N) application (Ahmad and Cashmore 1993, Lintala et al. 2007, Yuan et al. 2016). These studies were also performed under long-day (16 h light/8 h dark) conditions. Although the GSH levels in the rosette leaves were not analyzed in these studies (Ogawa et al. 2001, Yanagida et al. 2004, Yuan et al. 2016), the application of GSH and FeSO₄ could increase the GSH levels in the rosette leaves. Although the ratio of reduced GSH was similar between sultr2;1 mutants and the WT in our study (Supplementary Fig. S6), redox status due to the NAD(P)H/NAD(P) or reduction rate of ascorbic acid might be modulated in the mutants. In addition, the ratio between S and other nutrients can be the cue for early bolting because N, phosphorus (P) and potassium (K) levels modulate the bolting timing (Zhang et al. 2023a). We need to further dissect the molecular relationships among decreases of sulfate and other S-containing metabolites in the early primary stem, increased GSH levels in the rosette leaves and early bolting phenotypes.

In conclusion, our results demonstrated that SULTR2;1 is essential for maintaining the S metabolite levels in the aerial part of mature plants. The decreased sulfate or S metabolite levels in the early primary stem or increased GSH in the rosette leaves could stimulate early bolting. The relationship between the SULTR2;1 disruption and early bolting needs further investigation.

Materials and Methods

Plant materials and growth conditions

Arabidopsis thaliana cv. Columbia (Col-0) was used as the WT plant. A T-DNA insertion mutant (SM_3_39440; Tissier et al. 1999) was provided by the Arabidopsis Biological Resource Center (ABRC), segregated using genomic PCR and RT-PCR analyses with gene-specific primers (Supplementary Table S1) and named sultr2;1–1 (Supplementary Fig. S1A). A genome-edited mutant line was generated by transforming a gRNA-inserted CRISPR-Cas9 plasmid, pEgP237-2A-GFP (Ueta et al. 2017). The gRNA sequence, 5ʹ-ACATCAAGAGAGATAGCAAT-3ʹ, was selected using the gRNA design website ‘focas’ (Osakabe et al. 2016). The transgenic lines were selected on germination agar media containing 50 mg l⁻¹ kanamycin in a growth chamber controlled at 22°C and an 18-h light and 6-h dark cycle (Valvekens et al. 1988). Genomic DNA was extracted, PCR amplified with the ST21_503+_685F and ST21_503+_1051 R primers (Supplementary Table S1) and the resultant PCR products were sequenced. The T₁ plants with mutations were transplanted to vermiculite, and the presence of the sequence in the descendant plants was confirmed. A T₃ line carrying a homozygous mutation in the SULTR2;1 genomic sequence was named sultr2;1–2 and used for subsequent experiments (Supplementary Fig. S1B).

Seeds were directly sown on vermiculite to investigate mature plant growth (Nittai, Osaka, Japan). After germination, the plant number was adjusted to 8 (phenotype observation) and 3 (GFP observation) per pot. The half-strength MGRL solution containing 750 µM sulfate was supplied to plants twice a week (Fujiwara et al. 1992), which was regarded as the S-sufficient condition. The temperature and light condition were maintained at 23°C and 24-h light at 33 µmol m⁻² s⁻¹ light intensity.

Measurement of sulfate uptake activity

Plants were grown on MGRL agar medium for 3 and 4 weeks before and after bolting, respectively. For the mature plants, 2-week-old plants grown on MGRL agar medium were transferred to MGRL hydroponic solution for 4 weeks. The roots were submerged in MGRL liquid media containing 15 µM [³⁵S] sulfate (American Radiolabeled Chemicals, Saint Louis, Missouri, USA) and incubated for 1 h (before bolting), 3 h (after bolting) or 6 h (mature plants). The samples were washed and set in transparent plastic bags before being transferred to the imaging plates for 6 d. The samples were then scanned using a fast laser scanner (Typhoon FLA7000; GE Healthcare Life Sciences, Tokyo, Japan). The roots and aerial parts (rosette leaves, primary stems, lateral stems, flowers and siliques) were collected separately and analyzed using a liquid scintillation counter (AccuFLEX LSC-8000; HITACHI ALOCA, Tokyo, Japan) as described previously (Yoshimoto et al. 2016).

Observation of GFP fluorescence

The tissue-specific expression of GFP in P_SULTR2;1-GFP plants was visualized and pictured using the Amersham Typhoon laser scanner (GE Healthcare Bio-Science AB, Uppsala, Sweden). The fluorescence of GFP was detected using a 100-µm pixel size with a 488-nm excitation laser and Cy2 525BP20 and Cy5 670BP30 emission filters.

Measurement of sulfate, cysteine and GSH contents

The plant samples were measured for FW and dried with a freeze dryer (VD-800R; TAITEC, Saitama, Japan). Then, the samples were ground into fine powder, extracted and prepared as described previously (Zhang et al. 2020).

Sulfate content was analyzed using ion chromatography (IC-2010; TOSOH, Tokyo, Japan). Briefly, 30 µl of plant extracts were injected and separated at 40°C using a TSKgel SuperIC-AZ column (4.6 mm × 15 cm; TOSOH) and a flow rate of 1.2 ml min⁻¹. The eluent contained 6.5 mM NaHCO₃ and 2.5 mM Na₂CO₃. Anion mixture standard solution 1 (Wako Pure Chemicals, Osaka, Japan) was used as a standard.

Total cysteine and GSH contents were analyzed as described previously (Kimura et al. 2019). Briefly, thiols were reduced with dithiothreitol (Nacalai Tesque, Kyoto, Japan) and labeled with monobromobimane (Cayman Chemical, Ann Arbor, Michigan, USA). For the reduced GSH analysis, the reduction process with dithiothreitol was omitted. The labeled samples were separated using a high-performance liquid chromatograph (JASCO, Tokyo, Japan) fitted with a TSKgel ODS-120T column (150 mm × 4.6 mm; TOSOH) and an FP-920 fluorescence detector (JASCO). The fluorescence was monitored at 478 nm under an excitation of 390 nm. The elution was performed with solvent A (12% methanol, 0.25% acetic acid) and solvent B (90% methanol, 0.25% acetic acid) with a 0–30% B gradient. Cysteine and GSH (Nacalai Tesque, Kyoto, Japan) were used as standards. The oxidized GSH (GSSG) levels were calculated by subtracting the total GSH with the reduced GSH.

Measurement of total S content

The total S content was analyzed using inductivity-coupled plasma optical emission spectrometry (ICP-OES, Agilent 5800; Agilent Technologies, Santa Clara, California, USA). The samples were prepared and analyzed as described previously (Zhang et al. 2023b).

Observation of developmental phenotypes

Plant growth on the vermiculite was observed every day after adjusting the plant number per pot. The number and the fresh weight of rosette leaves were recorded before and after bolting. Bolting time was recorded as the day after sowing when a primary stem was 0.5-cm tall (Steffen et al. 2014). Plant height was measured once every 5 d after 35 DAS. The number of lateral stems and siliques was counted on 55 DAS. Siliques were harvested when they had turned brown and before they dropped seeds. The siliques and seeds were separated using forceps, and the total seed weight, total stem DW, silique length and 100 seed weight were measured. The plant images were captured using a smartphone (iPhone SE, MHGU3J/A model; Apple, Cupertino, California, USA). Each part of the plant was weighed immediately after sampling and kept at −80°C until use.

Quantitative RT-PCR

Total RNA was isolated using Sepasol RNA I Super G (Nacalai Tesque), and reverse transcription was conducted using the PrimeScript RT Reagent Kit with gDNA Eraser (Takara, Osaka, Japan) as described previously (Zhang et al. 2020). Real-time PCR was conducted using KAPA SYBR FAST qPCR Master Mix 2× (Kapa Biosystems, USA), a qTOWER³ G touch (Analytik Jena, Jena, Germany) and gene-specific primers (Supplementary Table S2). The relative mRNA levels were calculated using the comparative threshold cycle method with EF-1 alpha (AT1G18070) as an internal standard.

Statistical analysis

The data were analyzed using the GraphPad Prism version 10.0.0 program (GraphPad Software, Boston, USA, www.graphpad.com). Statistical differences were detected using one-way ANOVA followed by Dunnett’s test (Figs. 3–5, and Supplementary Figs. S4–S7) or Student’s t-test (Fig. 1 and Supplementary Fig. S2). Significant differences detected using Dunnett’s test or Student’s t-test are shown with asterisks (*0.05 ≤ P <0.1, **0.01 ≤ P < 0.05, *** P < 0.01).

Supplementary Data

Supplementary Data are available at PCP online.

Data Availability

All data are incorporated into the article and its online supplementary material.

Funding

Japanese Society for the Promotion of Science (JSPS) KAKENHI grant numbers JP24380040, JP17H03785, JP22H02229 and JP22H05573, Takahashi Industrial and Economic Research Foundation (No. 12-003-152 (to A.M.-N.)) and JST A-STEP Program grant number JPMJTM19GH.

Acknowledgments

We gratefully acknowledge the ABRC for providing the T-DNA insertion line of SULTR2;1. ICP-OES (Agilent 5800) analysis was performed at the Center of Advanced Instrumental and Educational Supports, Faculty of Agriculture, Kyushu University, with the kind instruction by Emiko Matsunaga. The sulfate uptake assay was done at the Central Institute of Radioisotope Science and Safety Management, Kyushu University. Plant growth and seed harvesting were done at the Biotron Application Center, Kyushu University. The authors thank Japan International Cooperation Center/Japanese Grand Aid for Human Resource Development Scholarship for providing the scholarship to K.S. at Graduate School of Bioresources and Environmental Sciences, Kyushu University.

Author Contributions

K.S. and A.M.-N. designed the research, K.S., T.N., T.U. and A.M.-N. conducted the experiments, Y.O. and K.O. instructed the genome editing experiment, K.S. and A.M.-N. analyzed the data, K.S. and A.M.-N. drafted the manuscript and K.S., J.F., Y.O, K.O., and A.M.-N. revised the manuscript.

Disclosures

The authors have no conflicts of interest to declare.