MicroRNA414c affects salt tolerance of cotton by regulating reactive oxygen species metabolism under salinity stress

ABSTRACT

Salinity stress is a major abiotic stress affecting the productivity and fiber quality of cotton. Although reactive oxygen species (ROS) play critical roles in plant stress responses, their complex molecular regulatory mechanism under salinity stress is largely unknown in cotton, especially microRNA (miRNA)-mediated regulation of superoxide dismutase gene expression. Here, we report that a cotton iron superoxide dismutase gene GhFSD1 and the cotton miRNA ghr-miR414c work together in response to salinity stress. The miRNA ghr-miR414c targets the coding sequence region of GhFSD1, inhibiting expression of transcripts of this antioxidase gene, which represents the first line of defense against stress-induced ROS. Expression of GhFSD1 was induced by salinity stress. Under salinity stress, ghr-miR414c showed expression patterns opposite to those of GhFSD1. Ectopic expression of GhFSD1 in Arabidopsis conferred salinity stress tolerance by improving primary root growth and biomass, whereas Arabidopsis constitutively expressing ghr-miR414c showed hypersensitivity to salinity stress. Silencing GhFSD1 in cotton caused an excessive hypersensitive phenotype to salinity stress, whereas overexpressing miR414c decreased the expression of GhFSD1 and increased sensitivity to salinity stress, yielding a phenotype similar to that of GhFSD1-silenced cotton. Taken together, our results demonstrated that ghr-miR414c was involved in regulation of plant response to salinity stress by targeting GhFSD1 transcripts. This study provides a new strategy and method for plant breeding in order to improve plant salinity tolerance.

Introduction

Cotton (Gossypium spp.) is a widely cultivated polyploid economic crop which provides fiber, seed oil, and protein meal. Although cotton is a relatively salt-tolerant crop, exposure of cotton to high salinity can lead to the generation of reactive oxygen species (ROS), which directly leads to a considerable negative impact on cotton growth and development and lint yield [1]. High levels of ROS can affect many cellular functions by damaging nucleic acids, oxidizing proteins, and causing lipid peroxidation [2]. To survive the damaging effect of salinity-induced ROS accumulation, plants have evolved multifaceted strategies to minimize these adverse effects, such as regulating the effects of ROS by enzymatic and non-enzymatic antioxidant defense systems [3]. As the first antioxidant enzyme (‘antioxidase’) involved in cellular defense against oxidative stress, superoxide dismutases (SODs) can reduce the oxidative damage caused by ROS through conversion of superoxide anion radicals (O₂^·-) into oxygen and hydrogen peroxide (H₂O₂) [4].

In plants, SODs have been classified into three groups based on the metal cofactor used: copper-zinc SOD (Cu/Zn-SOD, CSD), manganese SOD (Mn-SOD, MSD) and iron SOD (Fe-SOD, FSD) [5]. Overexpression of a cytosolic Cu/Zn-SOD gene from rice (Oryza sativa L.) in transgenic tobacco (Nicotiana tabacum L.) plants enhanced tolerance to salt stress and water deficit [6], while overexpression of a cytosolic Cu/Zn-SOD gene from maize (Zea mays L.) in chloroplasts of transgenic Chinese cabbage plants resulted in increased tolerance against salt stress [7]. A cytosolic CSD-overproducing transgenic plum showed improved tolerance to salt stress [8] while overexpression of an Mn-SOD gene from Tamarix androssowii Litw. (TaMSD) in transgenic poplar (Populus davidiana Dode. × P. bolleana Lauche.) also led to increased tolerance against salt stress [9]. This growing evidence has suggested that SODs respond to salt stress and that SOD-overexpressing plants have an enhanced tolerance of salt stress. The most extensively investigated studies of the regulation of SOD gene expression involve the mitogen-activated protein kinase (MAPK) signaling cascade [10] and microRNAs (miRNAs) [11].

MicroRNAs are a class of small (18–24 nt), endogenous, non-coding RNA molecules that negatively regulate gene expression at multiple levels by targeting mRNAs for degradation and/or by repressing translation [12,13]. The role of miRNAs as transcriptional regulators of SOD expression in plant salt stress responses has received increasing attention [14]. For examples, in Arabidopsis thaliana (L.) Heynh., miR398 targets CSD1 and CSD2 genes, which suggests a direct connection between the miRNA pathway and SOD regulation [15]. When Arabidopsis plants were exposed to salt stress, miR398 expression was down-regulated at the transcriptional level, and this down-regulation was important for post-transcriptional CSD1 and CSD2 mRNA accumulation, and consequently salt tolerance [16]. Similar results were also reported by Lu et al. (2010) [17] and Jia et al. (2009) [18], who found that the miR398-mediated regulation of CSD in response to salt stress was conserved in rice and poplars, respectively. Although SOD and miRNAs play important roles in plant salt stress response, little is currently known about their respective functions and the interaction between them in cotton.

Previous studies reported that SOD-mediated control of ROS metabolism also plays crucial roles in plant growth and development, as well as stress responses in many plants [8,15,19–21]. Iron superoxide dismutase (Fe-SOD) was thought to be the most ancient SOD of the metalloenzymes and is an important component of the primary enzymatic antioxidant system controlling ROS accumulation in plants; they are most abundantly localized inside plant chloroplasts, to which they are indigenous [22]. In our previous study, we found that the expression of an upland cotton (Gossypium hirsutum L.) iron superoxide dismutase 1 (GhFSD1) gene was regulated by salinity stress, and that the coding sequence (CDS) of GhFSD1 harbored a potential ghr-miR414c targeting site [23]. Previous sequencing data showed that ghr-miR414c belongs to a member of the large miR414 miRNA family, which contains seven members in cotton, from ghr-miR414a to ghr-miR414g, the mature sequences of which are identical [24]. Previous studies had also indicated that ghr-miR414b and ghr-miR414e might be associated with fiber development in cotton by targeting cellulose synthesis-related genes [24] and showed that miR414 were significantly down-regulated by lead stress in cotton [25]. However, whether ghr-miR414c plays a role in salt tolerance of cotton is still unclear.

To further determine their biological functions in upland cotton, we cloned the primary transcript of ghr-miR414c (pri-ghr-miR414c) and the full-length open-reading frame (ORF) of GhFSD1 from upland cotton in this study. The expression profiles of ghr-miR414c and GhFSD1 under salinity stress were monitored, and their targeting relationship was verified. Transgenic Arabidopsis constitutively expressing GhFSD1 exhibited improved salt tolerance in Arabidopsis, while expression of ghr-miR414c in Arabidopsis resulted in plants showing increased sensitivity to salinity stress. We performed knockdowns of GhFSD1 in cotton by virus-induced gene-silencing (VIGS) that resulted in increased sensitivity to salinity stress. Thus, the results from the current study suggested that ghr-miR414c functions in the cotton salinity stress-response process by mediating GhFSD1 expression. Our present findings provide important new information regarding the relationship between SOD and miRNA in the salinity stress-response in cotton.

Results

Molecular cloning and characterization of GhFSD1 and ghr-miR414c in cotton

One 1075-bp cDNA sequence, GhFSD1, was cloned from the leaf of G. hirsutum cv. SF06 and submitted to GenBank (Accession No. MK305851). The resulting full-length GhFSD1 cDNA consisted of a 64-bp 5ʹ-UTR (untranslated region), 930-bp ORF, and an 81-bp 3ʹ-UTR (Data S1). The ORF of the GhFSD1 gene encoded a 309-aa protein with a predicted molecular weight of 35.645 kDa and a predicted isoelectric point of 4.84. The GC content of the ORF was 42%. Multiple sequence alignments showed that the GhFSD1 protein shared 81% identity with TcFSD1 and 70% identity with AtFSD2 (Figure 1(a)). The deduced amino acid sequence of GhFSD1 contained a Sod_Fe_N domain (Pfam: PF00081) from amino acids 48 to 133 and a Sod_Fe_C domain (Pfam: PF02777) from amino acids 140 to 258 (Figure 1(b) and Data S1). To identify the clade of this GhFSD, a phylogenetic tree was constructed using the NJ method of MEGA 6.0.6, based on FSD sequences from some model plants. Plant FSD proteins form two groups [26,27]. We found that GhFSD1 belonged to the group-I FSDs and was supported by very strong bootstrap values with a Theobroma cacao L. FSD (TcFSD1) (Figure 1(c)).

Figure 1.

Identification and characterization of GhFSD1 and ghr-miR414c. (a) Multiple amino acid sequence alignments of GhFSD1, TcFSD1, and AtFSD2. The conserved secondary structure information was indicated at the top of aligned sequences: α, alpha-helix; β, beta-sheet. (b) The conserved domains of GhFSD1 were annotated using the Pfam database. (c) Phylogenetic analysis. The numbers above the branches indicate bootstrap values (> 60%) from 1000 replicates with the neighbor-joining (NJ) method using the Jones-Taylor-Thornton (JTT) substitution model. Groups I and II of plant FSDs are indicated in red and blue, respectively. (d) The genomic location of ghr-miR414c. The red line indicated the position of ghr-miR414c on the cotton genome. The arrows indicate transcription directions. The rectangles filled with green, blue and yellow represent the genomic location of the gene locus, coding sequence and potential alternative transcripts from RNA-seq, respectively. The potential alternative transcripts were identified from CottonFGD (https://cottonfgd.org/). (e) Stem-loop structure of the ghr-miR414c precursor. The hairpin structures of ghr-miR414c precursor was detected by the mfold Web Server (http://unafold.rna.albany.edu/?q=mfold). The sequence colored with green indicated the mature sequence of miR414c.

Meanwhile, we cloned the primary transcript of ghr-miR414c (pri-ghr-miR414c) from the leaf of G. hirsutum cv. SF06 and submitted it to GenBank (Accession No. MK305852). The pri-ghr-miR414c was composed of 346 nt, which contained a 94-nt precursor sequence (ghr-MIR414c) and a 21-nt mature sequence (ghr-miR414c) (Data S2). The ghr-miR414c was located in the 5ʹ-UTR of one protein-coding gene (Gene ID: Gh_D09G1263) on the upland cotton D09 chromosome (Figure 1(d)). The BLAST searching in the genome of two diploid cotton species (Gossypium arboreum L. and Gossypium raimondii Ulbr.), with pri-ghr-miR414c as a query sequence, was used to analyze the origin of ghr-miR414c. We found the homologous sequence in only G. raimondii (Figure S1), but not in G. arboreum. The result indicated that the ghr-miR414c descended from G. raimondii, which is regarded as the donor of D-sub-genome to upland cotton [28]. In addition, the secondary structure analysis by the mfold Web Server predicted a well-developed hairpin structure in the ghr-MIR414c sequence (Figure 1(e) and S2).

Expression analysis of GhFSD1 and ghr-miR414c

We analyzed the expression patterns of GhFSD1 and ghr-miR414c in cotton. The transcriptional pattern of GhFSD1 was determined using RNA-seq data. As shown in Figure 2(a), GhFSD1 was expressed in all tissues investigated in this study, particularly in the torus. In addition, we found that the expression of GhFSD1 was induced by various abiotic stresses, including salt, polyethylene glycol, and hot or cold treatments. To determine how the expression of GhFSD1 and ghr-miR414c was affected by salinity stress, we conducted a time-course experiment with cotton grown hydroponically supplemented or not with salt treatment, and the expression levels of GhFSD1 and ghr-miR414c were analyzed in cotton using qPCR. The results of qPCR showed that the expression of ghr-miR414c peaked at 12 h, then decreased after 24 h following salt treatment to below the pre-stress expression level, a response which was opposite to the expression pattern of GhFSD1 in response to salt stress (Figure 2(b)). These results indicated that GhFSD1 was transcriptionally and/or post-transcriptionally regulated by ghr-miR414c in vivo in response to salt stress. In addition, we also detected the expression level of ghr-miR414c by qPCR in cotton root, cotyledon, hypocotyl, stem, leaf, and flower under normal growth conditions (Figure 2(c)). The results showed that the expression of ghr-miR414c was highest in the hypocotyl.

Figure 2.

Expression pattern analysis of GhFSD1 and ghr-miR414c. (a) Transcriptional profiling of GhFSD1. The log₂ FPKMs values calculated by RNA-seq data were shown as a heat map. The colors of the bar shown to the right of the heat map varied from red to green representing the relative expression levels from high to low. (b) ghr-miR414c and GhFSD1 expression patterns in cotton under salt treatments. The expression levels of ghr-miR414c and its targets were normalized to that of the cotton ubiquitin extension protein 7 (UBQ7). (c) Expression profiles of ghr-miR414c in cotton. cDNAs were synthesized from total RNA extracted from various tissues of cotton, and qPCR was used to quantify the mRNA levels. The data were normalized to expression of UBQ7 and were shown relative to the vegetative root sample. Error bars represented standard deviations of the mean values from three independent experiments. Means with a common letter are not significantly different at p < 0.05 according to Tukey’s test.

ghr-miR414c targets GhFSD1

Based on these results, we deduced that GhFSD1 played a positive role in achieving plant tolerance to salinity stress. From the expression pattern assays, we speculated that the transcriptional and/or post-transcriptional regulation of GhFSD1 may be used to alter its expression under normal conditions and in response to salinity stress. miRNAs are ubiquitous transcriptional and/or post-transcriptional regulators in plants, and previous studies had shown that cotton miRNAs were involved in the regulation of plant response to salt stress [29–31]. In our previous study, we had predicted that ghr-miR414c matched the CDS of GhFSD1 [23]. Previous reports had suggested that a functional miRNA and its target can contain up to five mismatches [32], and our analysis showed that there were two mismatches within the predicted complementary region between ghr-miR414c and GhFSD1. In addition, ghr-miR414c showed expression patterns opposite to those of GhFSD1 in response to salinity stress (Figure 2(b)). This result indicated that ghr-miR414c might regulate GhFSD1 expression.

To verify the target relationship between ghr-miR414c and GhFSD1, we performed transient co-transformation technology in tobacco (Nicotiana benthamiana). The pRI201-AN-GUS vector (harboring the GUS gene) was introduced into cells of tobacco leaves using the Agrobacterium-mediated transformation system. Control leaves infiltrated with GV3101-pRI201-AN-GUS exhibited the GUS phenotype revealed by histochemical staining, while leaves infiltrated with GV3101-pRI201-AN-GhFSD1, in which the target sequence was fused upstream of the GUS gene, showed a similar phenotype. In leaves infiltrated with GV3101-pRI201-AN-ghr-MIR414c, however, in which the GUS gene was replaced by the precursor of ghr-miR414c, the GUS phenotype was not observed. Meanwhile, GUS staining was markedly decreased in leaves co-transformed with the strain mixture GV3101-pRI201-AN-ghr-MIR414c and GV3101-pRI201-AN-GhFSD1 (Figure 3(a)). This was considered to be additional evidence that ghr-miR414c could target GhFSD1.

To confirm the results of these histochemical observations, RNA was extracted after two days of co-expression in tobacco, and GhFSD1 transcripts were analyzed by qPCR. We found that the abundance of GhFSD1 transcripts significantly declined when co-expressed with ghr-MIR414c (Figure 3(b)). In addition, we carried out RNA ligase-mediated (RLM) 5ʹ rapid amplification of cDNA ends (RACE) assay using RNA from tobacco leaves infiltrated with the strain mixture (GV3101-pRI201-AN-ghr-MIR414c and GV3101-pRI201-AN-GhFSD1) to determine the predicted cleavage site in GhFSD1. The PCR products were cloned and sequenced. The six clones showed the same site of cleavage following the 11^th nucleotide from the 5ʹ end of ghr-miR414c (Figure 3(c)). Taken together, these results show that ghr-miR414c targets GhFSD1.

Figure 3.

ghr-miR414c targets GhFSD1. (a) β-Glucuronidase (GUS) phenotype observed by histochemical staining. (b) Co-expression of the constructs containing pRI201-AN-ghr-MIR414c and pRI201-AN-GhFSD1 in tobacco leaves. Expression levels determined by qPCR were normalized to the expression levels of tobacco elongation factor 1α (NtEF-1α). Error bars represented standard deviations of the mean values from three independent experiments. Means with a common letter are not significantly different at p < 0.05 according to Tukey’s test. (c) The model of cDNA structure and mRNA cleavage site of GhFSD1 determined by 5ʹ RACE. ghr-miR414c complementary sites (black scissors) with the nucleotide positions of GhFSD1 indicated. The RNA sequence of each complementary site from 5ʹ to 3ʹ and the miRNA sequence from 3ʹ to 5ʹ are shown in the expanded regions. The black arrow indicated a cleavage site verified by RLM 5′-RACE, with the frequency of cloned PCR products shown above the alignment.

Overexpression of GhFSD1 and ghr-miR414c alters tolerance of transgenic arabidopsis plants to salinity stress

To investigate a possible role for GhFSD1 and ghr-miR414c in plant response to salinity stress, we overexpressed the full-length cDNA of GhFSD1 and pri-ghr-miR414c in Arabidopsis (Col-0 ecotype) WT. The plasmids for overexpression of GhFSD1 and ghr-miR414c were based on the pRI201-AN vector, which is under the control of the cauliflower mosaic virus 35S promoter (CaMV 35S). The pRI201-AN-GhFSD1 or pRI201-AN-miR414c plasmids were transformed into A. tumefaciens strain GV3101 by the freeze-thaw method with infiltration of GV3101 harboring pRI201-AN-GhFSD1 or pRI201-AN-miR414c into Arabidopsis being achieved by the floral dip transformation method. Three independent transgenic lines were obtained, and two of them were selected for further functional analysis (Figure S3(a)). qPCR analysis showed that the expression levels of GhFSD1 and ghr-miR414c in overexpressed Arabidopsis transgenic lines (35S::GhFSD1) were approximately 40–42-fold and 12–16-fold higher than in the WT, respectively (Figures S3(b) and 3(c)).

We then investigated the transgenic plants with respect to regulation of plant response to salinity stress. The transgenic Arabidopsis seeds were placed on half-strength MS agar medium supplemented with 0 mM, 100 mM or 150 mM NaCl, and the 35S::GhFSD1 lines displayed no abnormal phenotypes under normal (0 mM) conditions. Following exposure to 100–150 mM NaCl, the cotyledons of 35S::GhFSD1 seedlings were greener than those of the WT (Figure 4(a)). The seed germination was recorded after a 2–7-d period of NaCl treatment. Following a 3-d treatment with 100 mM NaCl, the seed germination rate of the 35S::GhFSD1 lines was 64.9–67.7%, higher than in the WT (Figure 4(b)). After 7 d salt treatment, the seed germination rate of 35S::GhFSD1 lines was about 90%, which was higher than that of WT (about 75%) (Figure 4(b)). After 2 weeks salt treatment, the root growth of 35S::GhFSD1 lines was greater than that of the WT, suggesting higher salt tolerance (Figure 4(a–d)). The seedling fresh weight of 35S::GhFSD1 lines was also examined. Compared to WT, the 35S::GhFSD1 seedlings exhibited significantly higher fresh weight (Figure 4(f)). These results suggested that GhFSD1 confers tolerance of transgenic Arabidopsis plants to salinity stress.

Figure 4.

Overexpressing of GhFSD1 and ghr-miR414c in Arabidopsis resulted in enhanced salinity tolerance. (a) Phenotype of primary root growth of 2-week-old Arabidopsis seedlings exposed to 0, 100 or 150 mM NaCl treatment. (b, c) Seed germination of transgenic Arabidopsis over-expressing GhFSD1 (35S::GhFSD1) and miR414c (35S::miR414c) under salinity stress. Seeds were germinated on half-strength MS agar medium supplemented with 100 mM NaCl. (d, e) The elongation of thr primary root of 2-week-old Arabidopsis seedlings exposed to 0, 100 or 150 mM NaCl treatment. (f, g) Fresh weight of 2-week-old Arabidopsis seedlings exposed to 0, 100 or 150 mM NaCl treatment. Error bars represented standard deviations of the mean values from three independent experiments. Means with a common letter are not significantly different between the transgenic plants and wild-type (WT) at p < 0.05 according to Tukey’s test.

Exposure to 100–150 mM NaCl inhibited cotyledon greening in the 35S::miR414c seedlings (Figure 4(a)). After a 2- to7-d period of exposure to NaCl, the seed germination rate of the 35S::miR414c lines was lower than that of the WT (Figure 4(b)). The root growth of 35S::miR414c lines was significantly inhibited by 100–150 mM NaCl treatment (Figure 4(a–d)), with significantly lower seedling fresh weight of 35S::miR414c being observed (Figure 4(f)). These results suggested that transgenic 35S::miR414c plants were hypersensitive to salinity stress.

GhFSD1 and ghr-miR414c affect salt tolerance of cotton by scavenging excess ROS

To further investigate the biological role of GhFSD1 under salt stress, Agrobacterium-mediated VIGS was used to silence GhFSD1 in cotton. Two weeks after Agrobacterium infiltration, the bleaching phenotype of the positive controls became apparent (Figure S4), and silencing efficiency was analyzed by qPCR, with the result showing that GhFSD1 was significantly silenced (Figure 5(a)). To examine the effect of silencing GhFSD1 on the overall tolerance to salinity stress in cotton, leaf discs from silenced plants were incubated in 0 or 400 mM NaCl solutions for 4 d. Silencing of GhFSD1 showed significantly more chlorosis in the leaf discs compared to control plants infected with TRV:00 (empty VIGS construct control) (Figure 5(c)). Then, GhFSD1-silenced plants were grown in pots and irrigated with 400 mM NaCl for 2 weeks, and the silenced plants showed greater sensitivity to salinity stress, compared to control plants, a finding which was consistent with the reduced total chlorophyll concentration (Figure 5(b, d)). In addition, the Na⁺ and K⁺ concentrations in VIGS cotton leaves were also measured. The Na⁺ concentration and the Na⁺/K⁺ ratio in control and GhFSD1-silenced plants exposed to NaCl stress were increased compared to those in the 0 mM control, reflecting that the plants were suffering from severe salt stress, and significantly higher values for each variable were found in the GhFSD1-silenced plants relative to the control plants containing the empty construct (Figure 5(d)). The TRV:GhFSD1 cotton plants were also significantly shorter than the TRV:00 plants in both the presence and absence of salt stress (Figure 5(d)), indicating the possible involvement of GhFSD1 in cotton seedling growth and development in addition to the effect on salt tolerance. The increased sensitivity to salinity stress shown by the GhFSD1-silenced plants strongly suggested a role for this gene in the tolerance of cotton to salt stress, although the mechanism is still unknown.

Figure 5.

Salt sensitivity increased in GhFSD1 and ghr-miR414c VIGS cotton plants. (a) Detection of GhFSD1 and ghr-miR414c transcripts in TRV:00, TRV:GhFSD1 and TRV:miR414c cotton plants. Expression levels were determined by qPCR, and the data were normalized to the expression levels of cotton ubiquitin 7 (UBQ7) gene. Error bars represent standard deviations of the mean values from three independent experiments. Means with the same letter are not significantly different at p < 0.05 according to Tukey’s test. (b) Phenotypes of control (TRV:00) and VIGS (TRV:GhFSD1 and TRV:miR414c) cotton plants under normal 0 mM conditions (Mock) and 400 mM NaCl treatment for 3 weeks. (c) Leaf disks of control and VIGS cotton plants were incubated in water supplemented with different concentrations of NaCl (0 mM or 400 mM) for 4 d. (d) Measures of total chlorophyll concentration, plant height, and Na⁺ concentration and Na⁺/K⁺ ratio of the leaves of control and VIGS cotton plants treated with water (Mock) or 400 mM NaCl for 2 weeks. Error bars represented standard deviations of the mean values from three independent experiments. Means with a common letter are not significantly different between control and VIGS plants at p < 0.05 according to Tukey’s test.

Many studies have reported that plant virus-based vectors could be used to study miRNA functions in plants [33–35]. When the virus-based vectors carrying precursor sequences of endogenous miRNAs infects the plants, using this viral miRNA expression system, the plant displays the phenotype resulting from the overexpression of that miRNA to investigate the function of miRNAs in plants [36,37]. To further investigate the function of ghr-miR414c in cotton, we inserted the primary transcript of miR414c into the pTRV vector to produce TRV:miR414c. After Agrobacterium infiltration for two weeks, qPCR was performed to detect the expressions of ghr-miR414c and GhFSD1. As shown in Figure 5(a), the expression level of ghr-miR414c was significantly increased in TRV:miR414c cotton compared with TRV::00 cotton. As a target of ghr-miR414c, the transcript abundance of GhFSD1 was reduced in the TRV:miR414c lines.

To examine the tolerance of TRV:miR414c cotton to salinity stress, leaf discs from TRV::00 and TRV:miR414c were incubated in 0 or 400 mM NaCl solutions. At 4 d after inoculation, the TRV:miR414c leaf discs showed more severe chlorosis than did the TRV::00 leaf discs (Figure 5(c)). Under salinity stress, the phenotypes of TRV:miR414c cotton plants showed greater sensitivity to salinity stress compared to control plants (Figure 5(b)). The physiological indices of TRV:miR414c cotton, including total chlorophyll concentration, plant height, Na⁺ concentration and Na⁺/K⁺ ratio, were similar to these of GhFSD1-silenced plants, relative to those exhibited by the TRV::00 control plants under salt stress (Figure 5(d)). These results showed that ghr-miR414c-overexpressed and GhFSD1-silenced cotton exhibited similar levels of hypersensitivity to salt stress.

Based on the above results and those from previous studies, we speculated that salinity stress increased the in-planta production of ROS and reduced the expression of ghr-miR414c to increase the transcript level of GhFSD1, with the elevated GhFSD1 activation reducing ROS accumulation and hence alleviating cell membrane injury in response to salinity stress. To confirm our hypothesis, we determined the SOD enzyme activities and H₂O₂ concentrations in WT cotton plants. When cotton plants were exposed to salinity stress, in addition to the induced expression of GhFSD1, the SOD enzyme activities and H₂O₂ concentrations gradually increased in the cotton plants (Figure S5A and B). Consistent with these results, we hypothesized that the ghr-miR414c-mediated regulation of GhFSD1 gene transcription contributed to the improved salinity stress response in cotton. More concretely, when cotton plants were exposed to salinity stress, the down-regulated expression of ghr-miR414c damped down the inhibition of GhFSD1, leading to an increase in GhFSD1 expression level and the elevation of SOD activity, which further resulted in increased scavenging of the excessive salinity-induced ROS accumulation in plant cells to response to salinity stress.

Discussion

It has long been known that salinity stress triggers a significant increase in the in-planta generation of ROS, which induces oxidative stress, and detrimentally causes lipid peroxidation in cellular membranes, DNA damage, and protein denaturation [38]. In plants, the ROS-scavenging system that maintains a precise balance of ROS plays important roles to prevent oxidative damage and adapt to salinity stress. The ROS scavenging in plants is mediated by an array of enzymes including SOD, ascorbate peroxidase, catalase and peroxidase [39]. As a first line antioxidant enzyme in plant defense systems against oxidative stress, SOD can reduce the oxidative damage caused by ROS through conversion of O₂^·- into oxygen and H₂O₂ [4]. To date, a substantial number of studies have revealed that SODs have been functionally characterized as regulators of plant response to salinity stress [6–9], but the roles of miRNA-mediated regulation of SOD genes have been studied only in model plants. Arabidopsis miR398 was identified to regulate the level of ROS by directing the cleavage of its two targets, Cu/Zn-SODs (cytosolic CSD1 and chloroplastic CSD2) [11,16], and similar results were also found in rice [17,18].

miRNAs have emerged as potential key regulators of transcription. Many miRNAs have recently been identified to play critical roles in plant response to salinity stress. Microarray analysis in two cotton cultivars with distinct salt sensitivity identified 17 cotton salt-responsive miRNAs belonging to eight families, 12 of them showing a genotype-specific expression model in both cultivars [31]. Solexa sequencing in radish (Raphanus sativus L.) identified 136 known miRNAs (representing 43 miRNA families) and 68 potential novel miRNAs (belonging to 51 miRNA families); of these miRNAs, 49 known and 22 novel miRNAs were differentially expressed under salinity stress [40]. Next-generation deep sequencing in cotton (G. hirsutum L.) identified 337 miRNAs, comprising 289 known miRNAs and 48 novel miRNAs, of which 155 miRNAs were expressed differentially under salinity stress [1]. Although many salt-responsive miRNAs have been identified, further functional analysis is necessary to determine their roles in salinity response and tolerance in plants. Overexpression, silencing and mutations of miRNAs and their targets have been proved highly effective for investigating the functions of stress-responsive miRNAs and their targets in plant stress responses. Gao et al. (2016) reported that overexpression of cotton miRNVL5 in transgenic Arabidopsis plants showed hypersensitivity to salt stress, indicating that miRNVL5 was a negative regulator of plant response to salinity stress [29]. Transgenic rice and Arabidopsis plants constitutively overexpressing osa-miR396c showed reduced salinity stress tolerance compared to that of WT plants [41]. Transgenic tobacco lines with tae-miR408 overexpression exhibited enhanced stress tolerance, exhibiting an improved phenotype, greater biomass, and superior photosynthesis behavior compared with the wild type under salt treatments [42].

Currently, the mechanism associated with SODs and miRNAs, which controlled cotton response to salinity stress, is still unclear. Therefore, identification of salt-responsive miRNAs and SOD genes in cotton is behind the situation in other model species, such as Arabidopsis and rice. Also, the regulatory mechanism of SOD mediated by miRNAs is still poorly understood. In this study, we characterized the expression patterns and biological functions of ghr-miR414c and its target GhFSD1, which encoded an iron superoxide dismutase, from cotton, and provided evidence of ghr-miR414c-mediated GhFSD1 expression regulation as a new regulatory mechanism in cotton response to salinity stress. Sequence analysis revealed that GhFSD1 belongs to the conserved iron superoxide dismutase domains-containing protein gene family, and the ghr-miR414c precursor could fold into a perfect secondary hairpin structure, with higher negative minimal free energies and minimal free energy index (Figure 1, Figure S2). The target relationship of ghr-miR414c and GhFSD1 was predicted and experimentally confirmed by transient transfections in tobacco and 5ʹ RLM-RACE analysis (Figure 3). However, the cleavage site may not exist in tobacco FSD1, as determined by a bioinformatics analysis. Due to the limitations of experimental conditions, constitutive transfections in cotton were very difficult. Arabidopsis was therefore used instead, and gain-of-function analyses showed that, in transgenic Arabidopsis plants, excessive GhFSD1 activation resulted in salt-tolerant phenotypes, whereas 35S::miR414c plants were hypersensitive to salinity stress (Figure 4). Our data also showed that ectopic expression of GhFSD1 in Arabidopsis promoted seed germination, seedling biomass and primary root elongation under salinity stress, whereas 35S::miR414c plants exhibited opposite trends, which were consistent with the phenotypes of transgenic Arabidopsis under salinity stress (Figure 4). Although the use of overexpression of GhFSD1 and ghr-miR414c in Arabidopsis to determine its biological function in cotton has some limitations, the results do suggest the potential function of GhFSD1 and ghr-miR414c in cotton response to salinity stress.

VIGS assays showed that silencing GhFSD1 and overexpressing ghr-miR414c decreased tolerance to salt treatment in cotton (Figure 5(b, c)). It was reported that excess Na⁺ was toxic to plants, whereas K⁺ was antagonistic to Na⁺ under salinity stress [43]. To identify the potential mechanism of VIGS plants in response to salinity stress, concentrations of Na⁺ and K⁺ were determined in both control and VIGS plants. Our results showed that there was no difference of Na⁺ and K⁺ contents in control and VIGS plants under normal conditions (Figure 5(d)). However, when exposed to salt treatment, the TRV:GhFSD1 and TRV:miR414c plants accumulated more Na⁺ and less K⁺ as compared to control plants. The imbalance of Na⁺ and K⁺ levels led to the higher Na⁺/K⁺ ratio in TRV:GhFSD1 and TRV:miR414c plants (Figure 5(d)).

In addition, it has been widely accepted that up-regulating the activities of antioxidant enzymes like SOD, which could act as the cellular signaling molecules to avoid the resulting damage under salinity stress [44]. In the present study, we determined the SOD enzyme activities and H₂O₂ concentrations in cotton plants. The results showed that salinity stress obviously increased the activities of SOD in response to the higher levels of ROS generated under salinity stress (Figure S5). Based on our data, we hypothesized that, without salinity stress, the transcript level of GhFSD1 was regulated by ghr-miR414c via post-transcriptional gene silencing to maintain the ROS levels for normal metabolism in cotton cells.. When cotton plants were exposed to salinity stress, the in-planta production of ROS was induced and the expression of ghr-miR414c was reduced. The down-regulated expression of ghr-miR414c dampened down the inhibition of GhFSD1, leading to the increased GhFSD1 expression level and elevated SOD activity. As a result, the enhanced total SOD enzyme activity contributed to scavenge the excessive salinity-induced ROS in plant cells (catalyzing the conversion of O₂^·- to H₂O₂), which alleviated cell membrane injury in response to salinity stress (Figure 6). How the downstream molecular network components respond to salinity stress, after being activated by the salt-induced ROS burst, needs to be investigated in further research.

Figure 6.

A proposed model for the role of miR414c in salt tolerance. Without salinity stress, miR414c regulated the expression of GhFSD1 to maintain the ROS levels for normal metabolism in cotton cells. When the plants were exposed to salinity stress, the repressed miR414c accumulation elevated the expression of GhFSD1. The up-regulated GhFSD1 resulted in higher total SOD enzyme activity. As a result, the enhanced total SOD enzyme activity contributed to scavenge the excessive salinity-induced ROS in plant cells (catalyzing the conversion of O₂^·- to H₂O₂), which protected against cell membrane injury, thus response to salinity stress. The size of each box was an indication of the expression level. Bold lines indicate increased pathways, while thin lines indicate decreased pathways.

Taken together, the present study indicated that both GhFSD1 and ghr-miR414c were involved in plant response to salinity stress. Ectopic expression of GhFSD1 resulted in a NaCl-tolerance phenotype in Arabidopsis, whereas a salt-hypersensitive phenotype was observed in transgenic Arabidopsis overexpressing ghr-miR414c. We have suggested that ghr-miRNA414c affects salinity tolerance of cotton by regulating ROS metabolism under salinity stress via its effect on GhFSD1 expression, with these two genes responding in opposite ways to salt stress.

Materials and methods

Plant material and stress treatments

Upland cotton cv. SF06 plants was used in this research. The seeds of upland cotton were sown in a soil mix [peat moss: perlite, 2:1 (v/v)] in plastic pots and were placed in plant growth chambers under the following conditions: 28°C/21°C day/night temperature, 16/8 h light/dark photoperiod, 3,300 lux light intensity and a relative humidity of 70%. And upland cotton plants were cultivated in a normal agronomic field from April to September under standard conditions in Tai’an, the experimental station of Shandong Agricultural University. Cotyledons, hypocotyls, roots, and stems were harvested from 10-day-old seedlings; leaves and flowers were harvested from field-grown plants for expression profiling analysis. For salt stress treatments, three-week-old seedlings were treated with NaCl or water (control) as described in our previous report [23]. The leaves of treated plantlets were harvested at 0, 3, 6, 12 and 24 h. All samples were frozen immediately in liquid nitrogen and stored at −80 °C until to analysis.

Wild-type (WT) Arabidopsis seeds, ecotype Col-0, and T₃ homozygous seeds of the GhFSD1 or ghr-miR414c transgenic Arabidopsis lines were surface sterilized and grown in half-strength Murashige–Skoog (MS) agar medium. After a 3-d stratification period at 4°C in the dark, the plates were transferred to a growth chamber at 21°C with a 16/8 h light/dark photoperiod. After one week, they were transferred to half-strength MS agar medium supplemented with 100 mM or 150 mM NaCl and cultured for two weeks. For the control conditions, seedlings were grown in half-strength MS agar medium and grown under the same culture conditions.

Seeds of tobacco (Nicotiana benthamiana Domin) were surface sterilized and germinated on half-strength MS agar medium at 25°C with a 16/8 h light/dark photoperiod. The seedlings were then transplanted at the 2- to 3-leaf stage into soil and grown under greenhouse conditions to use for co-transformation.

RNA isolation, qPCR analysis and 5ʹ RLM-RACE assay

Total RNA from 100 mg of cotton and tobacco leaves samples was isolated using RNAprep Pure Plant Kit (Polysaccharides & Polyphenolics-rich, DP441) (TIANGEN, Beijing, China). The concentrations and quality of the isolated RNA samples were determined by 1.5% agarose gel electrophoresis and a NanoDrop 2000 Spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, USA). First-strand cDNA was synthesized using HiScript® II Q RT SuperMix with gDNA wiper (Vazyme, Nanjing, China). For miRNA quantitative polymerase chain reaction (qPCR), first-strand cDNA was synthesized using a Mir-X^TM miRNA First-Strand Synthesis Kit (TaKaRa, Dalian, China). The miRNA-specific primer was designed based on the mature miRNA sequence (see Table S1). qPCR used a QuantStudio™ 6 Flex Real-Time PCR System (Applied Biosystems™, Carlsbad, CA, USA) and SYBR® Premix Ex Taq™ II (TaKaRa) in a 20 μl reaction volume with three biological replicates and three technical replicates. qPCR was performed as follows: pre-denaturation at 95°C for 30 s; 40 cycles of 95°C for 5 s and 60°C for 30 s, followed by a melt cycle from 65 to 95°C. The 2^−ΔΔCT method was used to determine the relative expression levels. Cotton ubiquitin extension protein 7 (UBQ7), tobacco elongation factor 1α (NtEF-1α) and Arabidopsis ubiquitin 5 (UBQ5) genes were used as endogenous controls.

The RNA ligase-mediated (RLM) 5ʹ rapid-amplification of cDNA ends (RACE) assay was performed using the FirstChoice™ RLM-RACE Kit (Ambion) with 10 μg total RNA following the instructions in the user’s instruction manual. Degraded mRNAs with a 5ʹ RNA adaptor were used for reverse transcription. Subsequently, 1 μl of reverse transcription product was used for nested PCR, which was performed using a 5ʹ nested adaptor primer and a 3ʹ gene-specific nested primer (outer or inner primer). PCR amplification was performed based on the following conditions: 94°C for 3 min; 35 cycles of 94°C for 30 s, 60°C for 30 s, and 72°C for 30 s, followed by 72°C for 10 min. The PCR products obtained from 5ʹ RLM-RACE were inserted into the pEASY®-Blunt (Transgen, Bio Inc., Beijing, China), and individual clones were selected for DNA sequencing.

All gene-specific primers used in this study were gathered from the ICG database (a knowledgebase of Internal Control Genes for qPCR normalization, http://icg.big.ac.cn/index.php/Main_Page) [45], designed using Primer Premier 5.0 [46], and listed in Table S1.

Gene cloning, vector construction, genetic transformation

The cDNA sequence of the GhFSD1 gene and the primary transcript of ghr-miR414c (pri-ghr-miR414c) were isolated from a cotton cDNA library by PCR using Phanta® Max Super-Fidelity DNA Polymerase (Vazyme Bio Inc., Nanjing, China) with the primers listed in Table S1. The PCR product was ligated into the pEASY®-Blunt (Transgen, Beijing, China) to generate cloning vectors, which were transformed into bacterial strain Escherichia coli DH5α. The resultant plasmid was verified by Sanger sequencing.

For the VIGS assays, we designed specific primers containing the appropriate enzyme digestion sites (KpnI/MluI) to amplify the GhFSD1 fragments and the pri-ghr-miR414c from cloning vectors to construct VIGS vectors. The specific primers used for the construction of the VIGS vectors are listed in Table S1. The KpnI-MluI fragments of GhFSD1 or pri-ghr-miR414c were individually ligated into KpnI/MluI-digested vector pTRV2, yielding constructs TRV:GhFSD1 or TRV:miR414c, respectively.

For the stable or transient transformation assays, we designed specific PCR primers with 17-bp extensions (5ʹ) that were complementary to the ends of the linearized pRI201-AN-GUS vector (TaKaRa) digested by NdeI and/or SacI to amplify the fragments of interest. The PCR product armed with infusion connections was separately inserted into the vector pRI201-AN-GUS vector using In-Fusion® HD Cloning Kit (TaKaRa), and yielded the constructs pRI201-AN-GhFSD1-(GUS) or pRI201-AN-miR414c.

The positively ligated plasmids were verified by sequencing and individually transformed into the Agrobacterium tumefaciens strain GV3101. GV3101 carrying various vectors, was grown overnight in lysogeny broth medium with the appropriate antibiotics. A. tumefaciens cells were pelleted, resuspended, and incubated in infiltration medium (10 mM 2-[N-morpholino] ethanesulfonic acid, 10 mM MgCl₂, 200 μM acetosyringone, pH 5.6) at room temperature in the dark for 3 h.

Agrobacterium-mediated VIGS assays in cotton

For the tobacco rattle virus (TRV)-mediated silencing assays, pTRV1 vectors were mixed with pTRV2 vectors that comprised either candidate genes or empty vector pTRV2:00 in a 1:1 ratio to a final OD₆₀₀ of 1.0. Then, agroinfiltration by a 1 ml needleless syringe was used to co-infiltrate cotyledons of two-week-old cotton seedlings. The plants were covered with a clear plastic dome and the infiltrated plants left at room temperature under dim light conditions overnight. The plants were then transferred to a growth room with the temperature of 23°C, and 12 h light/12 h dark cycle. TRV:CLA1 (CLOROPLASTOS ALTERADOS 1) was utilized as the positive control to evaluate the effectiveness of VIGS and the empty vector TRV:00 as the negative control. Two weeks after infiltration, the bleaching phenotype of the positive controls became visible. The gene silencing efficiency was quantified by examining the expression level of endogenous genes using qPCR with RNA isolated from the control and silenced cotton plants. The treatment with salt stress was carried out as described previously. Each assay was performed with at least three independent replicates.

Agrobacterium-mediated transient transformation in nicotiana benthamiana

Co-transformation of tobacco leaves was used to validate the interaction between ghr-miR414c and GhFSD1. A. tumefaciens strain GV3101 carrying the different recombinant vectors (GV3101-pRI201-AN-GUS, GV3101-pRI201-AN-GhFSD1-GUS, GV3101-pRI201-AN-ghr-MIR414c and GV3101-pRI201-AN-GhFSD1-GUS + GV3101-pRI201-AN-ghr-MIR414c) was cultured as previously described. Then, the suspensions (1 ml from each treatment) was infiltrated into fully expanded leaves of six-week-old N. benthamiana plants with appropriate concentrations (OD₆₀₀ = 0.5). β-Glucuronidase (GUS) staining and GUS quantitative detection were conducted two d after infiltration. Each assay was performed with at least three independent replicates. GV3101-pRI201-AN-GUS was selected as the control. If the miRNA could cleave the target, the expression level of the GUS gene would be down-regulated.

Transformation of GhFSD1 and ghr-miR414c into arabidopsis

The cDNA sequence of the GhFSD1 gene and the primary transcript of ghr-miR414c were inserted downstream of the CaMV35S promoter in the binary vector pRI201-AN vector (Takara). The constructs were then transferred into A. tumefaciens strain GV3101 for Arabidopsis transformation by the floral dip transformation method [47]. Transformed seeds were selected on half-strength MS medium containing 50 μg/mL kanamycin (Solarbio, Beijing, China) using the fast selection protocol according to Harrison et al. [48]. Green plantlets were transferred to fresh selective half-strength MS medium and afterwards to artificial soil. Independent transgenic lines were obtained, and PCR was performed to verify the presence of 35S::GhFSD1 and 35S::miR414c in these transgenic Arabidopsis plants. T₃ generation Arabidopsis plants were used for all analyses. Overexpression levels of the GhFSD1 or ghr-miR414c sequences were quantified by RT-qPCR in 4-week-old seedlings.

Bioinformatics analysis of GhFSD1 and ghr-miR414c

GhFSD1-homologous protein sequences were retrieved from the TAIR database (http://arabidopsis.org) and Phytozome v 12.1 (https://phytozome.jgi.doe.gov/pz/portal.html). The full-length amino acid sequences were aligned with the ClustalX program [49] with default settings. The ESPript program was used to obtain sequence similarities and secondary structure information from aligned sequences (http://espript.ibcp.fr/ESPript/ESPript/index.php) [50]. The phylogenetic tree was constructed using the neighbor-joining (NJ) method with 1,000 bootstrap replicates using the Jones-Taylor-Thornton (JTT) substitution model [51] in MEGA 6.06 with a cut-off value of 60% for the condensed tree. Physico-chemical characteristics of the GhFSD1 protein, including the molecular weight and theoretical isoelectric point, were determined using the ProtParam tool (http://www.expasy.org/tools/protparam.html). The conserved domains of the protein sequences were predicted with the Simple Modular Architecture Research Tool (SMART) (http://smart.embl-heidelberg.de/) [52]. The conserved domains of the protein sequence encoded by GhFSD1 were presented using the IBS software (http://ibs.biocuckoo.org/) [53]. Expression value (FPKMs, fragments per kilobase of transcript per million mapped reads) of GhFSD1 was obtained from the websites at CottonFGD (https://cottonfgd.org/) [54] and ccNET (http://structuralbiology.cau.edu.cn/gossypium/) [55]. Gene expression levels were calculated according to the log₂ FPKM values. The heatmaps were plotted by Cluster 3.0 software [56] and TreeView [57]. The psRNATarget (http://plantgrn.noble.org/psRNATarget/) was used to predict targets of miRNAs with default parameters. The mfold Web Server (http://unafold.rna.albany.edu/?q=mfold) was used to detect hairpin structures with pre-miRNA sequence.

Determination of ion and chlorophyll concentrations

Samples were dried at 60°C until they reached constant weight and then ground. Ion concentrations were determined in 100 mg dry weight (DW) of plant material. Ions were extracted from samples in 0.5% HNO₃, and Na⁺ and K⁺ ions were assayed by flame emission photometry (Corning Medical and Scientific, Halstead, UK).

Chlorophyll concentration was estimated using the method described by Metzner with some modifications [58]. A sample (100 mg) of fresh leaves was dipped overnight in 85% (v/v) aqueous acetone for the extraction of the chlorophyll pigments, after which the solvent was centrifuged at 4,000 rpm for 10 min and diluted with 85% aqueous acetone to a suitable concentration for spectrophotometric measurements. The chlorophyll concentrations were calculated from the absorbances at 644 and 662 nm against a blank of 85% aqueous acetone. Chlorophyll a, b and total chlorophyll were analyzed using the following equations: chlorophyll a (µg/ml) = 9.784 × OD₆₆₂ − 0.99 × OD₆₄₄; chlorophyll b (µg/ml) = 21.426 × OD₆₄₄ − 4.65 × OD₆₆₂; total chlorophyll = chlorophyll a + chlorophyll b. The pigment concentrations were presented as mg/g fresh weight.

Estimation of SOD enzyme activity

The enzymatic activity of SOD was determined in leaves using a spectrophotometric method. Extraction of SOD enzyme was performed as previously described with modification [59]. Briefly, approximately 100 mg frozen leaf sample were ground with 3 ml of 50 mM ice-cold phosphate buffer (pH 7.8) with a mortar and pestle, and the homogenates were centrifuged at 4°C at 10,000 × g for 20 min. The resulting supernatant was collected and used for determination of enzyme activity. SOD enzyme activity was assayed according to the method of Stewart and Bewley (1980) [60] based on photochemical reduction of nitro-blue tetrazolium (NBT). One unit of SOD activity was defined as the amount of enzyme required to cause 50% inhibition of the reduction rate of NBT as monitored at 560 nm.

Determination of H₂O₂ concentrations

H₂O₂ concentration was measured according to the method of Snell and Snell [61] slight modifications. Briefly, H₂O₂ was extracted by homogenizing 50 mg of fresh leaf tissue with 3 ml of phosphate buffer (50 mM, pH 6.5). To measure H₂O₂ concentration, 3 ml of extract was mixed with 1 ml of 0.1% titanium sulfate in 20% (v/v) H₂SO₄ and the mixture was centrifuged at 6000 × g for 15 min. The intensity of the yellow color of the supernatant was measured at 410 nm. H₂O₂ content was calculated from a standard H₂O₂ curve.

Histochemical staining assay

For all experiments, at least ten leaves per treatment were sampled to be stained. β-Glucuronidase (GUS) staining was carried out as described by Jefferson et al. (1987) [62], using three independent biological replicates. Briefly, detached leaves were immersed for 12 h at 37°C in a solution containing 2 mM 5-bromo-4-chloro-3-indolyl-β-d-glucuronide, 0.1 M sodium phosphate buffer (pH 7.0), 0.50 mM each of potassium ferri- and ferrocyanide, 10.0 mM EDTA (pH 7.0), and 0.10% Triton X-100, following which the leaves were incubated in 95% ethanol for 12 h and then photographed.

Experimental design and statistical analysis

For statistical analysis, all experiments were performed at least three times. Unifactorial analyses of variance (ANOVA) tests were carried out between samples using DPS (Version 7.05) [63], and, if significant, the differences between samples were compared by Tukey’s test (p < 0.05).

Accession numbers

The GhFSD1 sequence can be found at Cotton Functional Genomics Database (CottonFGD) (https://cottonfgd.org/) [52] under the Unique Name Gh_A07G0392. Sequence data for the other cotton gene GhFSD3 discussed in this paper can be found at CottonFGD under the Unique Name Gh_A13G0530. Sequence data regarding the Arabidopsis genes discussed in this paper can be found in the TAIR database (http://arabidopsis.org) under the following accession numbers: AtFSD1 (AT4G25100), AtFSD2 (AT5G51100), and AtFSD3 (AT5G23310). Sequence data for the other genes can be found in Phytozome v12.1 (https://phytozome.jgi.doe.gov/pz/portal.html) under the following accession numbers: PpFSD1 (Pp3c7_19790), PpFSD2 (Pp3c2_27440), OsFSD1 (LOC_Os06g02500), OsFSD2 (LOC_Os06g05110), TcFSD1 (Thecc1EG014798), TcFSD2 (Thecc1EG007687), BdFSD1 (Bradi1g50550), BdFSD2 (Bradi1g51140), VvFSD1 (GSVIVG01013850001) and VvFSD2 (GSVIVG01026137001).

Funding Statement

This research was mainly supported by National Key R&D Program of China (Grant Nos. 2018YFD0100303) and China Major Projects for Transgenic Breeding (Grant Nos. 2016ZX08005-004).

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data for this article can be accessed here.