Genome-Wide Identification and Expansion Patterns of SULTR Gene Family in Gramineae Crops and Their Expression Profiles under Abiotic Stress in Oryza sativa

Sulfate transporters (SULTRs), also known as H+/SO42− symporters, play a key role in sulfate transport, plant growth and stress responses. However, the evolutionary relationships and functional differentiation of SULTRs in Gramineae crops are rarely reported. Here, 111 SULTRs were retrieved from the genomes of 10 Gramineae species, including Brachypodium disachyon, Hordeum vulgare, Setaria italica, Sorghum bicolor, Zea mays, Oryza barthii, Oryza rufipogon, Oryza glabbermia and Oryza sativa (Oryza sativa ssp. indica and Oryza sativa ssp. japonica). The SULTRs were clustered into five clades based on a phylogenetic analysis. Syntheny analysis indicates that whole-genome duplication/segmental duplication and tandem duplication events were essential in the SULTRs family expansion. We further found that different clades and orthologous groups of SULTRs were under a strong purifying selective force. Expression analysis showed that rice SULTRs with high-affinity transporters are associated with the functions of sulfate uptake and transport during rice seedling development. Furthermore, using Oryza sativa ssp. indica as a model species, we found that OsiSULTR10 was significantly upregulated under salt stress, while OsiSULTR3 and OsiSULTR12 showed remarkable upregulation under high temperature, low-selenium and drought stresses. OsiSULTR3 and OsiSULTR9 were upregulated under both low-selenium and high-selenium stresses. This study illustrates the expression and evolutionary patterns of the SULTRs family in Gramineae species, which will facilitate further studies of SULTR in other Gramineae species.


Introduction
Sulfur is the least abundant macronutrient and plays an essential role in plant growth and response to various stresses [1,2]. In higher plants, sulfur is predominantly acquired from the soil in the form of anionic sulfate (SO 4 2− ) mediated by plasma membrane-localized H + /SO 4 2− co-transport process [3][4][5]. Various transporters, including the plasma membrane sulfate transporter, ATP-dependent sulfate and thiosulfate transporter, have evolved in plants for transportation of the sulfate [6]. These transporters are responsible for the absorption and transport of the anionic sulfate, which is associated with plant yield, stress responses and environmental adaption [7,8].
Sulfate transporters (SULTRs), which contain 12 trans-membrane domains and one sulfate transport anti-sigma domain (STAS) at the C-terminal region, are known as H + /SO 4 2− symporters and are located on the plasma membrane [9]. SULTRs have been reported to be involved in sulfate transport and are well studied in Arabidopsis thaliana [10]. The reported SULTRs are classified into four groups according to their affinity [11]. The high-affinity transporters AtSULTR1;1 and AtSULTR1;2 co-localized in roots, function in soil sulfate uptake [12]. Another high-affinity AtSULTR1;3 is localized in the phloem and participates in transport of sulfate. Low-affinity transporters (AtSULTR2;1 and AtSULTR2;1) mediate long-distance transport of sulfate in the vascular tissue. To date, the roles of the third group of SULTRs in Arabidopsis remain unclear. The last group of transporters (AtSULTR4;1 and AtSULTR4;2), localized in the tonoplasts, function in the efflux of sulfate from the vacuolar space into the cytoplasm [13]. In addition to the functional analysis of the SULTRs in Arabidopsis, several studies of SULTRs have been conducted in major crops, such as maize, rice, wheat, Brassica oleracea and soybean [1,11,13,14]. A total of 14 SULTRs have been identified in the Oryza sativa ssp. indica genome with varying expression patterns [15]. According to the microarray analysis, OsSULTR1;1, OsSULTR1;2, OsSULTR2;1, and Os-SULTR5;2 showed high expression in the roots of seedlings [16]. Further, the sulfur status of the plant has a significant impact on the expression of sulfate transporters under biotic and abiotic stress [11,17]. The mutation of OsSULTR3;3 contributes to altering the accumulation of phosphate and sulfate in rice seedlings [18]. Additionally, a recent study reported that SULTRs in maize function in adaptation to sulfur deficiency conditions [11].
Although SULTRs have been characterized in maize, rice and Arabidopsis, in-depth studies are lacking in other Gramineae plants, which limits further understanding of the expansion pattern of the SULTRs in Gramineae. Gramineae species represent major food crops with high economic and industrial value. High-quality genome sequences are available for these species, and their phylogeny has been well-studied [19,20]. This makes the Gramineae an ideal model system for studying the evolutionary dynamics of the SULTR family.
In this study, the SULTRs families from 10 Gramineae species, including Brachypodium distachyon, Hordeum vulgare, Oryza barthii, Oryza glabbermia, Oryza rufipogon, Oryza sativa ssp. japonica, Oryza sativa ssp. indica, Sorghum bicolor, Setaria italica and Zea mays were identified. To investigate the gene expansion and evolution patterns of SULTRs, we conducted several analyses related to phylogeny, chromosomal distribution, gene structure, protein motif, orthogroups, duplication events, selective forces, expression profiles and comparative genetics. Additionally, we analyzed the expression patterns of SULTRs under various stress conditions, including salt, heat, drought, low-selenium and high-selenium stress. Our results will facilitate further investigations of SULTRs in Gramineae.

Gene Structure and Conserved Motif Analysis
The feature coordinates (exon-intron boundaries) were extracted from the GFF3 annotation files. The resulting BED format files were used as input for constructing the exonintron structure of SULTRs in Gene Structure Display Server 2.0 (http://gsds.gao-lab.org/, accessed on 20 March 2021) [26]. SULTR motifs were investigated in the MEME suite (http://meme.suite.org/, accessed on 20 March 2021) with the number of motifs parameter set to 15 [27]. The combination of phylogenetic gene structure and motif information was performed by TBtools [28].

Gene Duplication Events, Chromosomal Locations and Orthogroup Analysis
Gene duplication of the SULTR family in each species was identified using the 'du-plicate_gene_classifier' script in the Multiple Collinearity Scan Toolkit (MCScanX) with an e-value of 1 × 10 −10 [29]. Comparative genome analysis was performed by using the MCScanX toolkit. The synteny analysis of the SULTR family in the 10 genomes was also conducted. All 10 representative genome annotation files were downloaded from Ensembl Plants (https://plants.ensembl.org/index.html, accessed on 20 March 2021). The chromosomal location of all the SULTRs was determined based on the genome annotation GFF3 files (https://plants.ensembl.org/index.html, accessed on 20 March 2021).
The orthogroup was identified using OrthoFinder v2 with a cut-off e-value of 1 × 10 −3 [30]. STAG and STRID algorithms were used to rebuild the phylogenetic tree of the selected species based on the detected orthogroup. DnaSP 6.0 was used to generate the value of Tajima's D in each orthogroup [31].
All cDNA and protein sequences were aligned by the "ParaAT.pl" script in ParaAT2.0 software with muscle blast analysis [32]. The nonsynonymous substitution ratios (Ka) and synonymous substitution ratios (Ks) of the duplicated gene pairs. together with their ratio. were analyzed using KaKs_Calculator2.0 software [33]. To investigate the divergence time between each duplicated gene pair, the genetic distances were calculated and compared based on the amino acid sequences using MEGA 7.0. Moreover, the selective forces on the orthogroups of SULTRs were calculated by Tajima's D value with Dnasp 6.0 [29,34].
The  [29]. Then, the microsyntenic pairs of the SULTRs family among the 10 genomes were constructed by TBtools.

Expression Analysis of SULTR Members in O. sativa ssp. indica
A total of 175 RNA-Seq datasets of different tissues at different developmental stages were downloaded from NCBI sequence read archive (SRA) (Supplementary Table S1). For RNA-Seq analysis, rice genome sequences and annotated gene files were downloaded from MBKbase (http://www.mbkbase.org/rice, accessed on 20 March 2021). Sequencing reads were aligned to R498 genome with TopHat version 2.0.13 [35]. The fragments per kilobase of exon model per million mapped reads (FKPM) method was used to estimate the gene expression using Cufflinks version 2.2.1 [36]. The detailed SRA information is described in the Supplementary Table S1. The average FKPM from the replicates was used in this study. Based on the clustering, the log 2 values of the gene expression levels were used to create a heatmap, representing the relative expression of the 12 SULTR genes at different developmental stages.

Plant Material and Treatments
In this study, the elite restorer line (Oryza sativa ssp. indica, 9311) of HL-CMS was used to obtain the expression profiles of the SULTR genes under heat, drought, salt, low-selenium and high-selenium treatments at different time points. All seedlings were grown in a hydroponic box filled with Yoshida solution in a growth chamber at Wuhan University with the climatic conditions set at 28 • C day/18 • C night, 14 h light/10 h dark and 65% relative humidity for 12 days. For heat treatment, 12-day-old seedlings grown in Yoshida solution were transferred to a light incubator (Bluepard, Shanghai Bluepard Instruments Co., Ltd., Shanghai, China) with the temperature set to 40 • C (14 h/day, 10 h/night). Roots were collected at 0, 3, 6, 9, 12 and 24 h after the application of heat treatments. A 20% w/v polyethylene glycol (PEG-6000) solution was applied to imitate drought stress [37]. For the salt stress, low-selenium and high-selenium treatments, the roots of each seedling were washed, followed by immediate transfer into a 200 mM NaCl solution, 2 µM, and 20 µM Na 2 SeO 4 solution, respectively [37,38]. Roots were sampled at 0, 3, 6, 9, 12 and 24 h during the light period after the application of treatments. Each treatment was conducted with 3 replicates in a completely randomized design. Three leaves at 0, 12, 24, 48 and 72 h after each treatment were randomly sampled and the fresh weight (FW) was immediately recorded. The leaf samples were then kept in a 50 mL tube filled with water for 12 h to obtain the turgid weight (TW) according to a previous report [39]. The the samples were then oven-dried at 80 • C for 48 h. The dry weight (DW) was recorded, and the relative water content was calculated according to the following expression: RWC = (FW − DW)/(TW − DW) × 100%. Student's t-test was used to determine if the mean of RWC at different times under stress was significantly different from in normal conditions.

RNA Isolation and RT-PCR Analysis of OsiSULTRs under Stress
To investigate the function of the 12 OsiSULTRs in the rice cultivar 9311, the expression patterns were analyzed in shoots and roots under various stresses using qRT-PCR. Total RNA was extracted using TRIzol reagent (Invitrogen). A total of 20 µg RNA was digested with RNA-free DNase I (New England Biolabs) and then reverse-transcribed to cDNA using random primers. The quantitative real-time PCR (qRT-PCR) was performed using 2× SYBR Green Master Mix reagent (Roche). The β-actin gene served as an internal control and the primer sequences of OsiSULTRs were listed in Supplementary Table S2. Each sample was tested for 3 technical and 3 biological replicates.  Supplementary Table S3. To understand the phylogenetic relationships of the SULTRs in the 10 Gramineae species, a phylogenetic tree was constructed and classification of the identified SULTRs was performed ( Figure 1). Clade IV and Clade V had significant gene expansion compared with Clade I and Clade II. Clade I had a larger number of SULTRs than Clade II for Oryza sativa ssp. japonica (Supplementary Table S4). Osj, Osi and Si had a higher number of SULTRs than Hv, Bd, Zm, Ob, Og and Or. We found two distinct expansion patterns during Asian cultivated rice and African cultivated rice domestication. The results showed an expansion of SULTRs during the domestication of Asian cultivated rice, while a loss of SULTRs occurred during the domestication of the African cultivated rice. To better understand the evolutionary pattern of SULTRs in Gramineae, the orthogroup (OG) clustering of the identified SULTRs was determined. A total of 111 SULTRs were classified into eight OGs, except for three unassigned SULTRs ( Figure 1, Table 1). We found that the number of genes in each OG was different, with a range from 4 to 27. Of the classified OGs, the 10 species shared seven OGs (OG2-OG8), and single-copy gene clusters were found in OG4 and OG5. OG1 exists in all rice species except for Oryza barthii. Taking all results into account, we discovered that unequal loss and expansion of most OGs occurred during the domestication process, except for the conserved OG4 and OG5. In addition, the Tajima's D values of all OGs were less than 0, indicating that the SULTRs in the different OGs were under strong negative selection (Table 1).

Expansion Patterns and Chromosome Location of SULTRs in 10 Gramineae Species
To understand the SULTRs' expansion mechanism among the 10 Gramineae species, we investigated the chromosomal distribution and duplication types of SULTRs within each species. The SULTRs of different Gramineae species showed various distributions and densities on chromosomes. According to Figure 2, a total of 25 duplicated gene pairs were identified in the 10 representative species (Table 2), including two duplicate modes: whole-genome duplication (WGD)/segmental duplication and tandem duplication. WGD/segmental duplications were only observed in rice species, including Oryza barthii, Oryza sativa ssp. indica, Orya sativa ssp. japonica and Oryza rufipogon. The duplicated gene pairs in the SULTRs accounted for 10 to 31% among the 10 Gramineae genomes. The duplicate modes and the numbers of SULTRs varied among the 10 genomes, especially in rice species. Although the total number of duplicate gene pairs was constant during the domestication from Oryza barthii to Oryza glabbermia, one WGD/segmental duplication and two tandem duplications were found in Oryza barthii and only tandem duplications were found in Oryza glabbermia. During the domestication of Asian cultivated rice, the numbers and types of duplicated gene pairs were consistent between Osi and Or, while Osj showed two more tandem duplicated gene pairs compared with Or.  (Table 1). In addition, the Tajima's D values of each clade were less than 0, suggesting that the SULTR genes in each clade were under purifying selection (Figure 3).

Gene Structure and Motif Patterns of SULTRs
The numbers of introns in the SULTRs ranged from 6 to 17, and no significant features were found between different clusters in each species (Figure 4). A total of 15 conserved motifs were identified and assigned to Motifs 1 to 15. Nine of them were found to be highly conserved. The sequences and conserved motifs are listed in Supplementary Tables S5 and S6. Genes that shared motif patterns were clustered into one group, indicating a similar function. These results suggest that the diverse motif composition among the SULTRs may have different functions in Gramineae species.

Syntenic Relationship of SULTRs among the Selected Species
The misidentification of highly conserved paralogs could result from the rapid divergence of the orthologous genes after speciation or gene loss after duplication [40]. To further elucidate the SULTR gene family size variation within the Gramineae species, the synteny of the SULTRs was investigated by comparative genome analysis of the 10 selected species using the MCScanX toolkit. We observed that the number of syntenic gene pairs among Bd, Hv, Osi, Osj, Ob, Og, Sb, Si and Zm ranged from 0 to 11 ( Figure 5). The syntenic relationship of the SULTRs was highly conserved among rice species. However, we identified only a weak syntenic relationship between Hv and Osi and between Zm and Si. Among them, Bd and Hv exhibited the lowest syntenic relationship of SULTRs, which indicateed that Bd has a distant relationship with Hv. These findings reveal closer relationships in rice species compared with other selected Gramineae species, consistent with their evolutionary distance. Figure 5. Syntenic relationships of SULTR genes among the 10 Gramineae species. The gray lines indicate the syntenic genes between two species, the blue lines show the syntenic SULTR genes between two species and the red triangles stand for the locations of the SULTR genes in each genome.

Constitutive Expression Pattern of SULTRs in Oryza sativa ssp. indica
To assess the function of the identified SULTRs in Oryza sativa ssp. indica (R498), we analyzed public rice transcriptomic data from all tissues at different stages. The results showed that the 12 OsiSULTRs exhibited distinct spatial and temporal expression patterns and can be divided into five groups ( Figure 6). The first OsiSULTR group was only expressed in grain and had almost no expression in other tissues (for example, OsiSULTR1). The second group had notable expression levels in leaves, with relatively low expression in other tissues at all stages (for example, OsiSULTR3 and OsiSULTR8). The third group (OsiSULTR6 and OsiSULTR7) was only expressed in the roots and had low expression levels in other tissues at all stages. The fourth group (OsiSULTR2) was strongly expressed in pollen and showed no expression in most tissues at all stages. The last group was expressed in all tissues at least at one stage. All OsiSULTRs were expressed in the roots except OsiSULTR1, while two OsiSULTRs (OsiSULTR1 and OsiSULTR6) had no expression in the shoots of 14-day-old seedlings. These results imply that OsiSULTRs is expressed in a tissue-specific manner in rice.

Expression Profiling of SULTRs in Rice Roots under Stress Conditions
We further investigated the SULTRs' expression under different abiotic stresses, including salt, heat, low-selenium, high-selenium and drought stress. First, the relative water content (RWC) of seedlings under each stress was evaluated (Supplementary Figure S1). Among the stresses, the RWC values under low-selenium and normal conditions showed no significant difference, while RWC values under the remaining stresses and in the control showed significant differences. The results indicated that low-selenium stress may have no effect on the rice seedling, while other abiotic stresses have significant impact on rice (Supplementary Figure S1).
Next, the expression patterns of OsiSULTRs in roots were measured by qRT-PCR (Figure 7). OsiSULTR2 and OsiSULTR6 were significantly upregulated at almost all time points under salt stress, while the expression levels of OsiSULTR10 and OsiSULTR12 were elevated at 6 and 9 h after salt treatment, respectively. The expression levels of two SULTR genes (OsiSULTR3 and OsiSULTR10) were increased after the heat and drought treatments. Furthermore, OsiSULTR12 and OsiSULTR2 were upregulated under heat and PEG treatments, respectively. Taking together, OsiSULTR2 showed significant upregulation under salt and drought stress, combined with one time point of heat stress. OsiSULT3, OsiSULTR4, OsiSULTR8 and OsiSULTR9 were upregulated under both low-selenium and high-selenium stress. OsiSULTR10 exhibited upregulation under the low-selenium treatment and downregulation under the high-selenium treatment. OsiSULTR11 and OsiSULTR12 showed different expression patterns under the selenium treatments. LSe, HSe, PEG6000, HT, and NaCl represent the low-selenium treatment, the high-selenium treatment, polyethylene glycol treatment, high-temperature stress and salt stress. Student's t-test was used in this experiment; *: p < 0.05, **: p < 0.05, ***: p < 0.001.

Expression Profiling of SULTRs in Rice Shoots under Stress Condition
Previous studies demonstrated that OsiSULTR transcript levels in shoots were posttranscriptionally modulated in response to changes in S conditions. Hence, we evaluated the OsiSULTR expression patterns in shoots under the salt, high temperature, drought, highselenium and low-selenium treatments (Figure 8). The results showed that OsiSULTR2 was upregulated by the high-selenium treatment at all the tested time points and the expression was highest at 6 h after treatment. OsiSULTR3, OsiSULTR10 and OsiSULTR12 showed significantly higher expression levels than other OsiSULTRs under the low-selenium treatment. The expression levels of OsiSULTR10 and OsiSULTR12 were significantly upregulated under salinity stress. Under the PEG6000 treatment, the expression levels of OsiSULTR3, OsiSULTR4, OsiSULTR10, OsiSULTR11 and OsiSULTR12 were increased compared with the control. OsiSULTR3, OsiSULTR9 and OsiSULTR12 were strongly induced in shoots by the high-temperature stress compared with the control. In short, OsiSULTR3 positively responded to the low-selenium, high-selenium, high-temperature and drought stresses, while OsiSULTR12 was upregulated under all stresses except for the high-selenium treatment. Furthermore, OsiSULTR1, OsiSULTR6 and OsiSULTR8 were not expressed in the shoots of 12-day-old seedlings under the salt, high-temperature, drought, high-selenium and low selenium treatments.

Discussion
High-quality grass family (Gramineae) genome sequences have brought opportunities to investigate the short-term evolutionary dynamics of the SULTR family. The synteny of their genomes also facilitated the comparative analysis of SULTR genes' structure and function. In our study, 111 SULTRs from 10 Gramineae species were identified. The results revealed no direct correlation between the number of SULTR genes and genome size. For instance, there were 12 SULTR genes in S. bicolor (genome size: 730 Mbp) [41], while there were 10 SULTR genes in H. vulgare (genome size: 4.79 Gbp) [42]. Zea mays, which experienced a specific WGD absent in other Gramineae crops [43], has the fewest SULTRs among the 10 Gramineae species in our study. This result indicates that Zea mays lost SULTRs after the gene duplication events. In the orthogroup clustering, Zea mays also showed fewer SULTRs than the remaining species in OG2 and OG4 (Table 2). Previous studies reported that tandem duplication and WGD/segmental duplication played a critical role in gene family expansion [44]. We found that the expansion mechanisms of the SULTRs family were different among the Gramineae species. For instance, no WGDs/segmental duplication events occurred in the studied Gramineae species apart from the rice species. SULTRs in Si had syntenic relationships with most of the SULTRs in other grass crops except Oryza. These results indicate that the expansion of the SULTR family in Si might have resulted from other unknown duplication types or large rearrangements, such as dispersed, proximal, inversions and translocations.
The number of SULTRs detected in this study is consistent with previous studies [15,16]. However, two sulfate transporters, OsSultr1;3 (LOC_os08g031410) and Os-Sultr5;1 (LOC_os08g031410), were not identified in this study. Instead, we identified one new SULTR gene (LOC_Os07g18700) in Oryza sativa ssp. japonica. Different gene structural arrangements and diverse expression patterns led to different functions of the gene subfamilies. Thus, we classified the 111 SULTR proteins into five clades. Clades I to III were classified into Group 3, whose functions have not been clearly investigated to date. Clade IV was equivalent to Group 2, corresponding to low-affinity transporters, and Clade V was classified as Group 1, involving high-affinity transporters. The gene expression profiling of the SULTRs also supported the classification of SULTRs based on Figure 6. Moreover, the SULTRs showed different gene structures between different clades, which confirmed the clade classification ( Figure 4). The Tajima's D values of all OGs in the five clades and the Ka/Ks values of the 25 duplicated gene pairs indicated that the SULTR family in the 10 Gramineae plants were under strong purifying selective forces. Additionally, with a total of eight OGs identified, OG8 (Group 3) contained 27 SULTRs, while OG1 (one from Group 3) had only four SULTRs. This suggested that different OGs might suffer from various extents of gene expansion and loss. We hypothesized that the various numbers of SULTRs identified in the OGs or clades was a result of natural or artificial selection. Multiple copies of high-affinity and low-affinity transporters were identified, while transporters with unclear functions were lost during rice domestication, according to the comparisons between Ob and Oj, Or and Osj, and Or and Osi.
Until now, increasing evidence has reported that SULTRs are critical for plant sulfate uptake and transport [45]. Our results showed that SULTRs have different expression patterns in different tissues, and some SULTRs had tissue-specific expression. We found that high-affinity SULTR genes showed increased expression in roots, which supported the idea that SULTRs participate in the uptake of sulfate from the soil and the sourcesink transport of sulfate [38]. In addition, low-affinity transporter SULTR genes had high expression levels in the stem, which suggests that the SULTRs in Group 2 are involved in regulating vascular sulfate transport. In addition, the other SULTRs did not show tissuespecific expression patterns but were expressed in most tissues during the developmental stages. Previous studies reported that SULTRs have positive responses to plant abiotic stress [11,46]. In our study, we identified 12 OsiSULTRs in shoots and roots that respond to salinity, high temperature, drought, high-selenium and low-selenium treatments. The results showed that three OsiSULTRs had no expression under five stresses in the shoots, while all OsiSULTRs were expressed in the roots. OsiSULTR1 and OsiSULTR6 were not expressed in the shoots either in constitutive expression patterns based on resequencing data or under abiotic stress detected by qRT-PCR. The expression of OsiSULTR8 has not been detected, while OsiSULTR2 was expressed in shoots, detected by qRT-PCR.
OsiSULTR10 was significantly upregulated in both shoots and roots under salinity stress. Meanwhile, OsiSULTR3 was remarkably upregulated under salt stress, high temperature and PEG6000. These results imply that OsSULTRs play different roles in low-selenium, high-temperature, drought and salt stress tolerance. Using Oryza sativa as a model species, the expression and clustering results of SULTRs in this study shed light on future functional analysis in Gramineae species.

Conclusions
In this study, 111 SULTRs were identified in Gramineae genomes, including 11, 10, 12, 11, 13, 10, 12, 11, 12 and 10 SULTRs in B. distachyon, H. vulgare, O. barthii, O. glabbermia, O. sativa ssp. japonica, O. sativa ssp. indica, O. rufipogon, S. bicolor, S. italica and Z. mays, respectively. Our results revealed the different expansion mechanisms of the SULTR family among the 10 selected Gramineae species and detected various selective pressures in the clustered clades and orthogroups. The SULTRs highly expressed in the shoots and roots may be essential for sulfur uptake and transport. The qRT-PCR results suggested that SULTR genes have important functions in response to salt, heat, drought, low-selenium and high-selenium stresses. Taken together, our results will facilitate further studies of SULTR genes' functions in other Gramineae plants.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/genes12050634/s1, Table S1: RNA-Seq datasets for build expression atlas, Table S2: List of oligonucleotide primary used for qRT-PCR, Table S3: The physiological and phylogenetic relationships of the SULTRs in Gramineae crops, Table S4: Gramineae SULTR gene numbers in different clades, Table S5: The protein sequences of rice SULTRs, Table S6: The conserved motifs of rice SULTRs, Figure S1: The relative water content of rice at different time points under different stresses.