The GmPRL1b‐GmST2‐GmAOC3/4 Module Confers Salt Tolerance and Botrytis cinerea Resistance by Inducing Jasmonic Acid Biosynthesis in Soybean

ABSTRACT

Abiotic and biotic stress significantly limit crop yields. However, most stress‐tolerance genes identified to date provide resistance to either biotic or abiotic stress and inhibit normal plant growth, limiting their application in breeding. We identified the soybean ( Glycine max ) NAC transcription factor gene GmST2, which is induced by salt and Botrytis cinerea stresses. Through comprehensive analyses of transgenic GmST2‐overexpressing lines, CRISPR‐Cas9 mutants, and heterologous expression systems in Arabidopsis thaliana and Nicotiana tabacum , we demonstrate that GmST2 confers dual resistance to salinity stress and B. cinerea while simultaneously enhancing plant growth under non‐stressed conditions. RNA‐seq revealed the enrichment of genes in the jasmonic acid (JA) pathway in GmST2‐overexpressing soybean. GmST2 directly binds to the promoters of GmAOC3 and GmAOC4, encoding key enzymes involved in JA biosynthesis, to promote their transcription, thereby enhancing JA accumulation and salt tolerance. Furthermore, we identified GmPRL1b, a WD40‐repeat protein, as a nuclear interactor that promotes GmST2 protein accumulation and functions upstream of GmST2 in regulating JA pathway genes. The elucidated GmPRL1b‐GmST2‐GmAOC3/4 regulatory module coordinates JA‐mediated cross‐protection against concurrent abiotic and biotic stresses. Evolutionary analysis indicates that GmST2 has undergone selection during soybean domestication, with the identified elite haplotype exhibiting enhanced salt tolerance. Our findings provide a molecular framework for developing stress‐resilient soybean cultivars, effectively addressing the trade‐off between stress resistance and plant productivity.

1. Introduction

Abiotic stress and biotic stress are major factors contributing to global reductions in crop yields. Soybean ( Glycine max ), a major crop for oil and protein production, experiences yield losses of > 40% due to salt stress and > 15% due to grey mould disease caused by Botrytis cinerea (Papiernik et al. 2005). Therefore, identifying genes that enhance tolerance to multiple types of stress concurrently and elucidating their underlying mechanisms could greatly facilitate crop breeding.

NAC (NAM, ATAF1,2, CUC2) transcription factors constitute one of the largest transcription factor families in plants. Most studies of NAC transcription factors have focused on their functional and mechanistic roles in tolerance to a single type of stress, with less attention paid to whether an individual NAC transcription factor can simultaneously promote or inhibit tolerance to both biotic and abiotic stress. For example, overexpressing OsNAC6 in rice ( Oryza sativa ) enhanced plant tolerance to rice blast, dehydration, and high‐salinity stress, although it negatively affected plant growth and yield (Nakashima et al. 2007). The membrane‐bound NAC transcription factor NTL6 positively regulates cold‐stress tolerance and disease resistance (Seo et al. 2010). To date, there have been no reports of genetic modifications of a single NAC transcription factor that can simultaneously enhance crop resistance to both abiotic and biotic stresses without adversely affecting crop growth and development.

The phytohormone jasmonic acid (JA), ubiquitous in land plants, serves as a central regulator of plant responses to diverse biotic and abiotic stresses (Rehman et al. 2023; Wang et al. 2021). Both exogenous JA application and genetic manipulation of JA biosynthesis/signalling pathways can enhance plant stress resistance. Key JA biosynthetic genes like TaAOC1 in wheat (Zhao et al. 2014), LOX3 in Arabidopsis (Ding et al. 2016), TomLoxD in tomato (Yan et al. 2013) and OsAOS1/2 in rice (Zeng et al. 2021) confer tolerance to salt stress or herbivory, while signalling components such as TaJAZ1 enhance powdery mildew resistance in wheat (Jing et al. 2019). However, JA may also exacerbate stress effects, as demonstrated by MeJA‐induced growth inhibition under salinity (Chen et al. 2017; Gao et al. 2021), highlighting the need for precise regulation when engineering JA pathways for crop improvement.

NAC transcription factors play pivotal roles in modulating JA‐mediated defence responses through regulating JA biosynthesis. In Arabidopsis, ANAC019/ANAC055 increase JA levels by enhancing the expression of the JA synthase gene LOX2, thereby activating the JA signal‐mediated defence responses (Bu et al. 2008). The rice NAC transcription factor OsNAP enhances JA biosynthesis and accumulation by increasing the expression of OsLOX2 and OsAOC2, thereby positively regulating drought tolerance (Zhou et al. 2013). Heterologous overexpression of the VaNAC26 in wild grape ( Vitis amurensis ) enhanced drought tolerance in Arabidopsis by regulating the JA biosynthesis and signalling genes (Fang et al. 2016). Most recently, IbNAC087 was shown to directly activate IbLOX and IbAOS promoters to enhance JA biosynthesis, conferring salt and drought tolerance in sweet potato (Li et al. 2024). These findings suggest that promoting JA biosynthesis by NAC transcription factors represents an important approach for regulating plant tolerance to both biotic and abiotic stress.

We previously reported the mechanism by which overexpressing the soybean NAC transcription factor gene GmSIN1 simultaneously enhanced plant growth and salt tolerance (Li et al. 2019). Here, we identified another abiotic and biotic stress‐induced soybean NAC transcription factor gene, GmST2, which positively regulates plant growth, salt tolerance, and resistance to B. cinerea . We identified a molecular module consisting of GmST2, its target genes, and an interacting protein that cooperatively regulates salt tolerance and resistance to B. cinerea , providing promising targets for soybean breeding.

2. Results

2.1. GmST2 Positively Regulates Plant Growth, Salt Tolerance, and Resistance to B. cinerea

To investigate the roles of soybean NAC transcription factors in plant growth and salt tolerance, we cloned the NAC transcription factor gene (Glyma.06G157400) which is highly induced by salt stresses (Song et al. 2012). Soybean plants generated from hairy roots overexpressing Glyma.06G157400 had significantly higher seedling weight and improved salt tolerance compared to plants generated from hairy roots transformed with the empty vector, while their RNAi plants were significantly shorter than vector control plants (Figure S1). We therefore named this gene GmST2 (SALT TOLERANT 2).

To validate its function, we generated three transgenic soybean lines overexpressing GmST2 (Figure S2), which exhibited lower photosystem II efficiency (ΦPSII) reduction and higher fresh weight compared to the wild type (WT) under 200 mM NaCl treatment for 5 days in laboratory conditions (Figure 1A–C, Figure S3). We also measured physiological indicators of salt tolerance, including proline content, malondialdehyde content, SOD, CAT, and POD activities, and Na⁺/K⁺ ratio in GmST2‐OE and WT plants with or without NaCl treatment. After NaCl treatment, GmST2‐OE plants exhibited more proline accumulation, lower malondialdehyde content, higher SOD, POD, and CAT activities, and a lower Na⁺/K⁺ ratio compared to the WT (Figure S4A).

FIGURE 1.

GmST2 promotes root growth, salt tolerance, and Botrytis cinerea resistance. (A) Wei6823 (WT) and GmST2‐OE seedlings grown in wet vermiculite for 7 days and subjected to 200 mM NaCl treatment for 5 days. Bar = 5 cm. (B) GmST2 expression in GmST2 OE and WT plants, as examined by RT‐qPCR. Total RNA was extracted from the roots of V₁ seedlings. Each column represents the mean ± SD (n = 3). (C) The photosystem II efficiency (ΦPSII) reduction rate of seedlings under NaCl treatment from (A). Each column represents the mean ± SD (n = 3 independent experiments). (D) Wild type (WT, Wei6823) and GmST2‐OE transgenic soybean plants grown alongside one another in non‐saline soil in the field. Images of R₁ seedlings are shown. Bar = 10 cm. (E) The average height of plants from (D). Each column represents the mean ± SD (n = 15). (F) Representative results for six (in non‐saline field) or ten (in saline field) seedlings each for GmST2 OE1, OE2, OE3, and the WT grown in the field. The soils contained 0.1 g (non‐saline) and 0.25 g (saline) total soluble salts per 100 g dry soil. Bars = 10 cm. (G) Yield per seedling of plants grown in non‐saline and saline soil in the field. Each column represents the mean ± SD (10 < n < 25). (H) WT (W82) and gmst2 gmst2h soybean were grown in wet vermiculite for 14 days and subjected to 175 mM NaCl for 7 days. Bar = 5 cm. (I) The photosystem II efficiency (ΦPSII) reduction rate of seedlings under NaCl treatment from (H). Each column represents the mean ± SD (n = 3 independent experiments). (J) WT (W82) and gmst2 gmst2h soybean grown alongside one another in non‐saline soil in the field. Images were taken of R₁ seedlings. Bar = 20 cm. (K) The average heights of V₃ and R₃ plants from (J) were measured. Each column represents the mean ± SD (n > 10). (L) and (N) Images of leaves of WT and GmST2 OE1 transgenic soybean plants (L) or WT and gmst2 gmst2h plants (N). The leaves of V₂ seedlings were infected with B. cinerea and incubated in the dark for 3–6 days. Bars = 2 cm. (M, O) The lesion area in leaves from (L) and (N) was measured. Each column represents the mean ± SD (5 < n < 8). Significant differences between samples (labelled with different letters) were determined by one‐way ANOVA with Tukey's test (B, C, E, G), p < 0.05. Significant differences between samples (labelled with asterisks) were determined by Student's t‐test (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001) in (I, K, M, O).

In field experiments, GmST2‐OE soybeans were significantly taller than WT soybeans in both 0.3% saline and non‐saline soils, while the yield per plant was significantly higher only in saline soils (Figure 1D–G, Figure S4B).

Phylogenetic analysis of GmST2 homologous protein sequences across seven representative plant species demonstrated that GmST2, Glyma.04G208300, Glyma.05G195000, Glyma.06G114000, and ATAF1 (AT1G01720) clustered within a single clade (Figure S5), indicating their derivation from a common ancestral gene. Notably, GmST2 exhibited the highest sequence similarity with Glyma.04G208300 (designated as GmST2h), forming a putative gene duplication pair, whereas Glyma.06G114000 showed closest homology to ATAF1, suggesting their evolutionary proximity. Expression profiling under salt and drought stress conditions revealed that among the four ATAF1 orthologs in soybean, GmST2 and GmST2h displayed highly correlated expression patterns that were distinctly different from the other two paralogs. This co‐expression characteristic strongly suggests functional redundancy between GmST2 and GmST2h. To further explore their function, we generated double knockout mutants of GmST2 and GmST2h using CRISPR‐Cas9‐mediated genome editing for subsequent functional characterisation. Sequencing revealed that both GmST2 and GmST2h protein translation was terminated early in the double mutants (Figure S6). gmst2 gmst2h mutant seedlings exhibited significantly higher salt sensitivity compared to the WT under 175 mM NaCl treatment for 7 days (Figure 1H). The mutant showed a more pronounced reduction in leaf ΦPSII upon NaCl treatment (Figure 1I). The fresh weight was significantly lower in the mutant under salt stress (Figure S3B). In field experiments conducted in non‐saline soils, the height of the gmst2 gmst2h double mutant during both the vegetative and reproductive stages was significantly reduced compared to the WT (Figure 1J,K).

Given that the Arabidopsis ortholog ATAF1 of GmST2 has been demonstrated to participate in both abiotic and biotic stress responses (Wu et al. 2009), we wonder whether GmST2 functions in plant defence. We analysed resistance against B. cinerea infection using leaves of GmST2‐OE and gmst2 gmst2h plants. After 4 days of B. cinerea infection, the lesion area on the leaves of GmST2‐OE transgenic soybeans was significantly reduced compared to WT leaves (Wei6823) (Figure 1L,M), while the opposite phenotype was observed in gmst2 gmst2h plants, with significantly larger lesion areas compared to WT plants (Williams 82) (Figure 1N,O).

To investigate the effect of heterologous GmST2 expression in other plant species, we constructed 35Spro::GmST2 transgenic Arabidopsis lines. Compared to the WT, 35Spro::GmST2 transgenic Arabidopsis had significantly increased seedling root length, salt tolerance, and B. cinerea resistance in leaves (Figures S7 and S8). These results demonstrate that GmST2 positively regulates soybean growth, salt tolerance, and resistance to B. cinerea and that it functions across different plant species. GmST2 holds potential as a target gene for improving crop growth, salt tolerance, and resistance to B. cinerea simultaneously.

2.2. GmST2 is a Transcriptional Activator and is Induced by Salt Stress and B. cinerea

To investigate the molecular characteristics of GmST2, we examined its subcellular localization and transactivation activity. In transiently transfected Arabidopsis protoplasts harbouring GmST2‐green fluorescent protein (GFP) fusion protein (Figure S9A), the GFP signal was detected in the nucleus (Figure S9A). To assess GmST2's transcriptional activation activity, we fused its coding sequence with the GAL4 DNA‐binding domain in pGBKT7 and tested reporter gene activation (His prototrophy and β‐galactosidase). Yeast transformants with pGmST2‐GBKT7 grew on His‐deficient medium and showed β‐galactosidase activity, unlike empty vector controls (Figure S9B). Thus, GmST2 is a NAC transcriptional activator that localises to the nucleus.

We further explored the expression pattern of GmST2 by quantifying the relative abundance of its mRNA in various organs of soybean cv. Williams 82 (W82). GmST2 was expressed at a high level in roots, leaves, and flowers and at moderate levels in stems and seeds (Figure S9C).

Given that GmST2 enhances plant tolerance to salt stress and resistance to B. cinerea , we assessed the expression pattern of GmST2 in response to 150 mM NaCl treatment and B. cinerea , respectively. GmST2 transcripts were rapidly induced by NaCl stress or B. cinerea , reaching peak expression at 6 and 0.5 h, respectively (Figure S9D,G). Subsequently, GmST2 expression rapidly declined and returned to its baseline level at 24 h (Figure S9D,G). To investigate GmST2 protein accumulation, we conducted immunoblotting using a GmST2‐specific antibody in WT soybean cv. W82. The specificity of the GmST2 antibody was verified in tobacco transformed with p35Spro::GmST2‐Myc, the empty vector control, and soybean cv. W82 (Figure S10). GmST2 abundance continued to increase during 24 h of NaCl treatment or B. cinerea (Figure S9E,F,H,I).

We also examined its response to JA, a phytohormone that regulates both biotic and abiotic stress resistance, and H₂O₂, a common secondary signalling molecule in biotic and abiotic stress responses, in soybean roots. GmST2 expression was rapidly induced by both JA and H₂O₂ treatment (Figure S11). These findings suggest that GmST2 primarily functions in the early response to salt stress and B. cinerea stress signals.

2.3. GmST2 Positively Regulates the Expression of JA Biosynthesis Genes and Promotes JA Accumulation

To elucidate GmST2‐regulated transcriptional networks, we conducted RNA‐seq analysis of roots from GmST2‐OE and WT soybeans under normal growth conditions, given the constitutive growth enhancement observed in GmST2‐OE lines even without salt stress (Figure 1D,E). Comparative analysis revealed 1278 differentially expressed genes (DEGs; fold change > 2, q < 0.05), comprising 854 upregulated and 424 downregulated genes (Figure S12). KEGG pathway enrichment of upregulated genes revealed pronounced associations (q < 0.05) with alpha‐linolenic acid metabolism (JA biosynthesis precursor), plant hormone signal transduction, and plant‐pathogen interaction pathways. Concurrent ontology (GO) analysis demonstrated significant enrichment (q < 0.05) terms including ‘immune system process’, ‘response to biotic stimulus’, ‘defence response’, ‘hormone‐mediated signaling pathway’, ‘host programmed cell death induced by symbiont’, ‘plant‐type hypersensitive response’, ‘response to hormone stimulus and Jasmonic acid metabolic process’ (Figure 2A,B). The coordinated enrichment of JA biosynthesis pathways and defence‐related functional categories strongly suggests that GmST2 regulates of JA production and subsequent activation of defence‐related gene networks.

FIGURE 2.

GmST2 regulates JA BIosynthesis and JA is required for GmST2 to enhance salt tolerance. (A) Significantly enriched pathways of the DEGs in GmST2 OE‐1 versus Wei6823, as determined by RNA‐seq. The numbers near the columns indicate the number of DEGs with the corresponding annotations and the p‐values. (B) Significantly enriched GO terms of DEGs in GmST2 OE‐1 versus Wei6823, as determined by RNA‐seq. The numbers near the columns indicate the number of DEGs with the corresponding annotations and the q values. (C) Heatmap of the differential expression of JA biosynthesis‐related genes in GmST2 OE‐1 versus Wei6823 (WT). The numerical values for the red‐to‐white gradient bar represent log₂‐fold change compared to the control. (D, F) GmAOC3, GmAOC4, and GmPR1 transcript levels in the roots of V₁ GmST2 OE‐1 (OE) seedlings (D) and the leaves of V₃ gmst2 gmst2h seedlings and the WT (F). Data were obtained by RT‐qPCR. (E, G) JA contents of the roots of V₁ GmST2 OE‐1 wild‐type (WT) seedlings (E) or leaves of R₂ gmst2 gmst2h and wild‐type plants (G). Error bars denote the SD (n = 3). Asterisks denote significant differences between samples (Student's t‐test, *p < 0.05, **p < 0.01, ***p < 0.001). (H) Phenotype of GmST2 OE transgenic plants compared with Wei6823 under control, 200 mM NaCl, or 200 mM NaCl with 10 mM DIECA (Sodium diethyldithiocarbamate trihydrate). Seedlings were grown in wet vermiculite for 14 days and subjected to various treatments for 10 days. Bar = 5 cm. (I, J) ΦPSII reduction rate (I) and fresh weight without roots per seedling (J) measured in the seedlings shown in (H). (K) Phenotype of gmst2 gmst2h plants compared with W82 under control, 175 mM NaCl, or 175 mM NaCl with 50 μM JA. Seedlings were grown in wet vermiculite for 14 days and subjected to various treatments for 7 days. Bar = 5 cm. (L, M) ΦPSII reduction rate (L) and fresh weight without roots per seedling (M) measured in the seedlings shown in (K). NS, no significant difference. Each column represents the mean ± SD (5 < n < 8 seedlings for (J, M); n = 3 independent experiments for I and L). Significant differences between samples (labelled with asterisks) were determined by Student's t‐test (*p < 0.05; **p < 0.01) (I, L). Significant differences between samples (labelled with different letters) were determined by Brown‐Forsythe ANOVA with Tamhane's T2 post hoc test (J, M), p < 0.05.

To further investigate the impact of GmST2 on the transcription of JA biosynthesis‐related genes, we identified all gene homologous to Arabidopsis genes annotated as LOX (lipoxygenase), AOS (allene oxide synthase), AOC (allene oxide cyclase), and OPR (oxophytodienoic acid reductase) genes, encoding key enzymes involved in JA biosynthesis. We identified 3, 2, 5, and 2 homologous genes of LOX, AOS, AOC, and OPR, respectively, that were significantly upregulated in GmST2‐OE plants compared to the WT (fold change > 2 and q < 0.05) (Figure 2C).

Among JA biosynthesis‐related homologues, Glyma.18G280900 (GmAOC4) was shown to increase JA levels in soybean (Zhang et al. 2021). We therefore selected the most highly upregulated GmAOC4 and GmAOC3 (Glyma.19G044900) genes (Wu et al. 2011) to further validate the RNA‐seq results. We measured the expression levels of these genes in GmST2‐OE and gmst2 gmst2h plants. Both of them were upregulated in GmST2‐OE plants and downregulated in gmst2 gmst2h plants (Figure 2D,F). Furthermore, JA contents increased in GmST2‐OE roots (Figure 2E) and decreased in gmst2 gmst2h (Figure 2G) leaves compared to the WT. To examine whether the altered JA content affected JA signalling in these plants, we measured the transcript levels of GmPR1, a marker gene of JA signalling. GmPR1 was significantly upregulated in GmST2‐OE plants but downregulated in gmst2 gmst2h plants compared to the WT (Figure 2D,F). Collectively, these results demonstrate that GmST2 positively regulates the transcription of key genes involved in JA biosynthesis, as well as JA accumulation and JA signalling.

2.4. JA Is Required for GmST2 to Enhance Salt Tolerance

To determine whether the enhanced salt tolerance in GmST2‐OE soybean was mediated by JA accumulation, we employed sodium diethyldithiocarbamate (DIECA), a specific JA biosynthesis inhibitor, in comparative salt stress assays. Under control conditions, no phenotypic differences were observed between GmST2‐OE and wild‐type plants. However, after 10‐day exposure to 200 mM NaCl, GmST2‐OE exhibited significantly attenuated ΦPSII reduction rates (p < 0.05) and greater fresh weight retention (p < 0.05) compared to WT (Figure 2H–J). Notably, this salt tolerance phenotype was completely abolished when DIECA was co‐applied with NaCl, as evidenced by the disappearance of inter‐genotypic differences in both photosynthetic parameters and biomass accumulation (Figure 2H–J).

To further validate JA's role, we conducted complementation experiments using exogenous JA application in the gmst2 gmst2h double mutant. Strikingly, JA supplementation (50 μM) fully rescued the salt‐sensitive phenotype, normalising both the ΦPSII reduction rate and fresh weight maintenance under NaCl stress (Figure 2K–M). These results conclusively demonstrate that GmST2‐conferred salt tolerance is strictly JA‐dependent.

2.5. GmST2 Directly Regulates the Expression of GmAOCs

The core binding motifs for NAC transcription factors are “CATGT” and “CGT[G/A]” (Lindemose et al. 2014). There are four such motifs in the 1000‐bp promoter regions upstream of the translation initiation sites of GmAOC4 and GmAOC3 (Figure S13A). To investigate whether GmST2 can bind to these motifs in the GmAOC4 and GmAOC3 promoters, we conducted an electrophoretic mobility shift assay (EMSA) to detect in vitro protein‐DNA binding. Indeed, GmST2 bound to these two sites in the GmAOC4 and GmAOC3 promoters in vitro (Figure S13B). We then performed cold probe competition experiments with wild‐type and site‐mutated probes to confirm their binding, finding that GmST2 directly bound to both sites in the GmAOC4 and GmAOC3 promoters in vitro (Figure 3A,B).

FIGURE 3.

GmST2 upregulates GmAOC3 and GmAOC4 by binding to their promoters. (A) Schematic diagrams of GmST2 binding to the promoters of GmAOC3 and GmAOC4. Pink lines show the GmST2 binding site in the promoters, empty boxes indicate the untranslated regions (UTRs) of the genes, and black lines indicate ATG translation initiation sites. Roman numerals indicate potential GmST2 binding sites in the promoters. (B) EMSA examining the binding of GmST2 with the binding sites in the target gene promoters. Left to right: Lane 1, labelled probe with GST protein as a negative control; lane 2, labelled probe with GST‐GmST2 protein; lanes 3 and 4, GST‐GmST2 binding to the labelled probe in the presence of 50× or 200× unlabeled wild‐type probes (denoted by an empty wedge); lanes 5 and 6, binding to mutant probe sequences (the GmST2 binding site was mutated; lanes 5 and 6, marked by a filled wedge). The arrows indicate the protein‐probe complex. Comp, competitor probe; G‐S, GST‐GmST2 fusion protein. (C) ChIP‐qPCR assay showing that GmST2 interacts with the GmAOC3 and GmAOC4 promoters in vivo. Anti‐GmST2 antibodies were used to precipitate chromatin prepared from the roots of V₁ Williams 82 seedlings. Fold enrichment was calculated based on the relative enrichment fold of fragments with binding sites compared to fragments without binding sites. The α, and β, and γ promoter fragments are indicated in (A). Data are means ± SD (n = 4 independent experiments). Significant differences between samples (labelled with different letters) were determined by one‐way ANOVA and Tukey's test, p < 0.05. (D) Constructs used for the transient expression assay. In the constructs, 2‐kb fragments of the GmAOC3 and GmAOC4 promoters drive LUC expression. “Mutant” indicates the promoter with the GmST2 binding site(s) mutated. The arrow indicates the promoter, and the box indicates the coding sequence. (E) Transient luciferase expression assay examining GmST2‐regulated GmAOC3 and GmAOC4 expression. Data are means ± SD of seven independent biological repeats, protoplasts cotransformed with pGreenII‐0800‐LUC and p35Spro:GmST2 as a negative control. M, reporter construct containing a promoter with a mutant binding site; W, reporter construct containing the wild‐type target gene promoter driving LUC. Asterisks denote significant differences between samples (one‐way ANOVA and Tukey’s test, ***p < 0.001, ****p < 0.0001).

To validate the binding of GmST2 to the GmAOCs promoters in vivo, we conducted a chromatin immunoprecipitation‐quantitative PCR (ChIP‐qPCR) experiment using the GmST2‐specific antibody in wild‐type soybean W82. qPCR primers were designed for fragments of the GmAOC4 and GmAOC3 promoters that included (β, γ) or excluded (α) the NAC core binding motifs (Figure 3A) for ChIP‐qPCR using wild‐type soybean root tissues. The enrichment levels of the β, γ fragments were significantly higher in the GmAOC4 and GmAOC3 promoters compared to the α fragments (Figure 3C), indicating that GmST2 specifically binds to the binding sites in the GmAOC4 and GmAOC3 promoters in vivo.

To explore whether GmST2 positively regulates the transcription of GmAOC4 and GmAOC3 by binding to specific sites in their promoters, we conducted a transient luciferase expression assay in Arabidopsis protoplasts. We cloned the wild‐type GmAOC4 and GmAOC3 promoters and promoters with mutated binding motifs into the reporter vector (pGreen II 0800‐LUC) and assessed the effect of GmST2 on transcription by examining changes in LUC activity (Figure 3D). Compared to promoters with mutated GmST2 binding motifs, co‐expressing GmST2‐Myc with the wild‐type GmAOCs promoters resulted in a significant increase in LUC activity (Figure 3E). Therefore, GmST2 activates the transcription of GmAOC4 and GmAOC3 by binding to specific sites in their promoters.

These results indicate that GmST2 directly targets the JA biosynthesis genes GmAOC4 and GmAOC3 to activate their expression, thereby modulating the responses of soybean to abiotic and biotic stress.

2.6. GmST2 Interacts With GmPRL1b in the Nucleus

To identify the interacting proteins of GmST2 involved in regulating soybean growth and resistance, we constructed a cDNA library from the roots of 10‐day‐old soybean seedlings and performed yeast two‐hybrid (Y2H) cDNA library screening. Glyma.15G149300, a 477‐amino acid protein with a WD40 repeat domain, was identified as the interacting protein using a Y2H assay (Figure 4A). Phylogenetic analysis across multiple species revealed that Glyma.15G149300 (named GmPRL1b) is a duplicate gene of Glyma.09G04380 (named GmPRL1a) and that both are orthologs of PLEIOTROPIC REGULATORY LOCUS 1 (PRL1, AT4G15900) in Arabidopsis, with 88.4% similarity (Figure S14). Transient expression of GmPRL1b‐GFP in tobacco leaf epidermal cells demonstrated its nuclear localization (Figure S15).

FIGURE 4.

GmPRL1b Interacts with GmST2 and promotes its accumulation. (A) GmST2 interacts with GmPRL1b in a yeast two‐hybrid assay. Yeast cells transformed with pGBKT7‐GmPRL1b and pGADT7‐GmST2 were selected on SD‐Trp/−Leu/‐His medium supplemented with 5 mM or 10 mM 3‐aminotriazol (3‐AT). pGADT7 and pGBKT7 are the empty vector controls, respectively. (B) BiFC assay of the interaction between GmST2 and GmPRL1b in planta. Vectors containing the indicated constructs were cotransformed into Nicotiana benthamiana leaves. Both YFP fluorescence images (left panel) and fluorescence images merged with bright field images (right panel) are shown. Bars = 20 μm. (C) Coimmunoprecipitation assay of GmST2 and GmPRL1b. Total proteins extracted from transiently transformed plants expressing GmPRL1b‐GFP (or GFP as a control) were immunoprecipitated (IP) by GFP antibody–conjugated agarose beads. The input and precipitated proteins were detected with an antibody recognising either GFP (α‐GFP) or GmST2 (α‐GmST2). (D) Constructs used to measure GmST2 abundance in tobacco. The constructs contain three expression cassettes: 35Spro::GmPRL1b‐GFP (or 35Spro::GFP as a vector control), 35Spro::ST2, and 35Spro::HPTII. (E) Effects of GmPRL1b expression on the abundance of GmST2 in transiently transformed tobacco leaves with the constructs shown in (A) measured by immunoblotting. α, anti. (F) Relative protein abundance (GmST2/HPTII) from seven biological replicates. (G, J) Effects of GmPRL1b overexpression (G) or RNAi (J) on the abundance of GmST2 in transiently transformed soybean hairy roots with vectors harbouring 35Spro::GmPRL1b‐GFP or 35Spro::GmPRL1b‐RNAi (35Spro::GFP as a vector control) measured by immunoblotting. (H, K) Relative protein abundance (GmST2/GmActin) from three to seven biological replicates. (I, L) GmPRL1b, GmAOC3, GmAOC4, and GmPR1 transcript levels in transiently transformed soybean hairy roots shown in (D) and (G), respectively. Data obtained by RT‐qPCR. Data are means ± SD (n = 3). Asterisks denote significant differences between samples (Student's t‐test, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001). NS, no significant difference, p > 0.05.

Bimolecular fluorescence complementation (BiFC) assays in transiently transfected tobacco leaves confirmed the interaction between GmST2 and GmPRL1b in the nucleus (Figure 4B). To further validate their interaction in soybean, we performed co‐immunoprecipitation (Co‐IP) experiments using the soybean hairy root system, demonstrating that GmST2 interacts with GmPRL1b in vivo (Figure 4C). These results indicate that GmST2 interacts with GmPRL1b in the nucleus in soybean.

2.7. GmPRL1b Enhances GmST2 Stability and JA Biosynthesis

PRL1, an Arabidopsis homologue of GmPRL1b, regulates the degradation of specific proteins (Lee et al. 2008). To investigate whether GmPRL1b affects GmST2 protein accumulation, we transiently expressed this protein in tobacco leaves using a ternary vector carrying the 35Spro::GmPRL1b‐GFP (or 35Spro::GFP as a vector control), 35Spro::GmST2, and 35Spro::HPT II expression cassettes (Figure 4D). We performed immunoblotting to detect the abundance of GmST2 and the internal control protein HPTII. The abundance of GmST2 was significantly higher in tobacco leaves expressing GmPRL1b‐GFP compared to those expressing the GFP control (Figure 4E,F).

Furthermore, we examined GmST2 protein levels in 35Spro::GmPRL1b‐GFP transgenic soybean hairy roots. GmST2 levels were significantly higher in 35Spro::GmPRL1b‐GFP roots compared to empty vector control roots (Figure 4G,H). Comparative analysis revealed significantly elevated transcript levels of GmPRL1b, GmAOC3, and GmAOC4 in 35Spro::GmPRL1b‐GFP transgenic hairy roots versus controls, whereas GmST2 transcript levels remained unchanged (Figure 4I). Concurrently, these roots exhibited substantial accumulation of JA (Figure 5E). In contrast, 35Spro::GmPRL1b‐RNAi transgenic hairy roots displayed the opposite regulatory pattern (Figure 4J–L). These findings demonstrate that GmPRL1b promotes GmST2 accumulation through post‐translational regulation rather than transcriptional control. This regulatory cascade subsequently enhances transcription of downstream target genes GmAOC4 and GmAOC3, ultimately elevating JA levels and upregulating the transcription of the JA signalling marker gene GmPR1.

FIGURE 5.

GmPRL1b promotes salt‐stress tolerance and Botrytis cinerea resistance and functions upstream of GmST2. (A) Salt‐stress tolerance assay of seedlings from GmPRL1b‐overexpressing (GmPRL1b‐OE) soybean hairy roots. Soybean hairy roots transiently transformed with vectors harbouring 35Spro::GmPRL1b‐GFP (or 35Spro::GFP as a vector control) were treated with 125 mM NaCl for 9 days. Bars = 2 cm. (B) Average fresh weights of the aboveground parts of the seedlings shown in (A). Error bars denote the SD (n = 3 independent experiments). In each experiment, 15 seedlings were analysed per construct. (C, D) Botrytis cinerea resistance assay. N. benthamiana leaves were transiently transformed with vectors harbouring 35Spro::GmPRL1b‐GFP (or 35Spro::GFP as a vector control). The transgenic leaves were bleached with 95% ethanol and photographed (C) 2 days after inoculation with B. cinerea spores, and lesion diameters were measured (D), bar = 1 cm. Error bars denote the SD (n = 6 biological replicates from two independent experiments). Asterisks denote significant differences between samples (Student's t‐test, ***p < 0.001). VC, vector control. (E) JA contents of transiently transformed soybean hairy roots shown in (A). Error bars denote the SD (n = 3). Asterisks denote significant differences between samples (Student's t‐test, *p < 0.05, **p < 0.05, ***p < 0.001). VC, vector control. α indicates the specific antibody. (F) GmST2, GmPRL1b, GmAOC3, GmAOC4 and GmPR1 transcript levels in transiently transformed soybean hairy roots. Data obtained by RT‐qPCR. Data are means ± SD (n = 3 biological replicates). Significant differences between samples (labelled with different letters) were determined by one‐way ANOVA and Tukey's test, p < 0.05. Labels in the x‐axis indicate the construct combination used to perform soybean hairy roots transformation. EV1 and EV2, the empty vector control for 35Spro::GmST2‐RNAi and 35Spro::GmPRL1b‐GFP constructs, respectively.

2.8. GmPRL1b Positively Regulates Plant Salt Tolerance and Resistance to B. cinerea and Functions Upstream of GmST2

To elucidate the roles of GmPRL1b in salt tolerance and resistance to B. cinerea , we first examined the response pattern of GmPRL1b under different stress conditions. In the roots of soybean W82, GmPRL1b was upregulated by 6 h of NaCl treatment but downregulated by 2, 6 and 12 h of H₂O₂ treatment, as well as by 0.5 and 6 h of MeJA treatment (Figure S16A). On the other hand, the fusion protein GmPRL1b‐myc, constitutively expressed in soybean hairy roots, did not exhibit significant responses to salt and B. cinerea stress (Figure S16B,C). These results suggest that GmPRL1b transcript, but not protein levels, respond to these stress signals, and that GmPRL1b might regulate related pathways (Figure S16).

To further confirm the roles of GmPRL1b in plant tolerance to abiotic and biotic stress, we generated plants from soybean hairy roots transiently transformed with 35Spro::GmPRL1‐GFP (GmPRL1‐OE) and stably transgenic Arabidopsis lines and examined the effects of NaCl treatment on these plants. GmPRL1‐OE soybean plants had significantly higher plant weight under salt treatment compared to empty vector control plants (Figure 5A,B). There was no significant difference in primary root length between 35Spro::GmPRL1‐GFP transgenic Arabidopsis lines and wild‐type Col‐0. However, under NaCl treatment, the average primary root length of 35Spro::GmPRL1‐GFP transgenic Arabidopsis was significantly higher than that of Col‐0 (Figure S17). These results suggest that GmPRL1b positively regulates salt‐stress tolerance in plants.

Subsequently, we analysed resistance to B. cinerea in the leaves of 35Spro::GmPRL1‐GFP transgenic Arabidopsis. 35Spro::GmPRL1‐GFP Arabidopsis showed smaller lesion diameters than Col‐0, whereas the GmPRL1b ortholog mutant prl1‐1 exhibited larger lesions (Figure S18). Finally, transient expression of 35Spro::GmPRL1‐GFP in tobacco leaves resulted in significantly enhanced resistance to B. cinerea compared to the empty vector control (Figure 5C,D). These results demonstrate that, like GmST2, GmPRL1b also positively regulates salt tolerance and resistance to B. cinerea .

To elucidate the genetic interaction between GmPRL1b and GmST2 in regulating the JA pathway, we generated four hairy root co‐transformation combinations: EV1 (35Spro::GmST2‐RNAi EV) + EV2 (35Spro::GmPRL1b‐GFP EV), 35Spro::GmST2‐RNAi + EV2, 35Spro::GmPRL1b‐GFP + EV1, and 35Spro::GmPRL1b‐GFP + 35Spro::GmST2‐RNAi. qRT‐PCR validated successful overexpression and RNAi suppression (Figure 5F). Comparative analysis revealed that GmPRL1b overexpression upregulated GmAOC3/4 and GmPR1 transcripts, whereas GmST2 knockdown downregulated these genes (Figure 5F). Notably, dual‐transformed roots (combination 4) mirrored the GmST2‐RNAi phenotype, with significant suppression of both markers relative to EV controls.

Collectively, these data imply that GmPRL1b enhances salt and B. cinerea tolerance by activating the JA pathway upstream of GmST2.

2.9. Natural Variation in GmST2 Associate With Growth and Yield of Soybean in Saline Fields

Natural variation often contributes to plant adaptation to diverse environmental conditions. To investigate the effects of GmST2 allelic variation on soybean, we analysed 571 re‐sequenced soybean accessions from global sources, comprising 55 wild soybeans, 142 landraces and 374 improved cultivars (Table S5). Based on 12 SNPs (including one non‐synonymous SNP in the coding region) and one InDel identified in the genomic region of GmST2, 16 haplotypes were classified, with Hap1‐3 representing predominant haplotypes in cultivated varieties (Figure 6A, Table S6). The median‐joining haplotype network revealed close phylogenetic relationships between Hap1 and Hap3, with Hap1 potentially serving as the ancestral haplotype of Hap3 (Figure S19), while Hap2 diverged earlier. Analysis of haplotype distribution across subgroups demonstrated that Hap1 frequency significantly increased during domestication from wild soybeans to landraces, whereas Hap3 frequency markedly rose during genetic improvement from landraces to cultivars, suggesting that these two haplotypes underwent artificial selection during soybean domestication and genetic improvement, respectively (Figure 6B). XP‐CLR‐based selection signal analysis across the 100 kb flanking region of GmST2 identified a selective sweep signal in the GmST2 genomic region between wild soybeans and landraces, further supporting its selection during domestication (Figure 6C).

FIGURE 6.

GmST2 haplotype variation under artificial selection. (A) Haplotype distribution of GmST2 in a soybean natural population. Top: Schematic of GmST2 gene structure (grey: Untranslated regions [UTRs]; black: Exons; black lines: Promoter/intronic regions). Five SNPs localise to the promoter, one to the coding sequence (CDS, missense, highlighted with orange), four to introns, and one to the 3′UTR. Haplotype frequencies (Hap1‐3) are indicated at right. (B) Haplotype distribution across wild (G. soja), landrace, and cultivar soybean accessions. (C) XP‐CLR selection scores across the 100 kb flanking region of GmST2 in wild versus landrace soybean comparisons. Gene positions within this region are represented by blocks, with GmST2 highlighted in red. (D, E) Agronomic performance of Hap1, Hap2, and Hap3 (D) or HapA and HapC (E) composite soybean lines under saline field conditions. Each dot in the column represents an accession. Significant differences between samples in (D) (labelled with different letters) were determined by Kruskal–Wallis test with Dunn post hoc tests, p < 0.05. Significant differences between samples in (E) (labeled with asterisks) were determined by Student’s t‐test, *p < 0.05, ****p < 0.0001. (F) Transactivation activity assay of GmST2^HapA and GmST2^HapC activity in yeast. Yeast cells transformed with pGBKT7‐GmST2 ^HapA (or pGADT7‐GmST2 ^HapC ) and pGmPRL1b (or VC, vector control, which is pGADT7 without AD coding sequence) were selected on SD‐Trp/−Leu/‐His medium and SD/−Leu‐Trp‐His‐Ade medium supplemented with/without 4 ng/mL X‐α‐Gal. BD, pGBKT7, the empty vector controls for pGBKT7‐GmST2 ^HapA and pGADT7‐GmST2 ^HapC . The images show representative results from more than three independent yeast transformants. (G) Relative transactivation activity of yeasts from (F). The relative transcriptional activation activity was calculated as the ratio of X‐α‐Gal staining intensity between the target yeast colonies and the vector control colonies. Error bars denote the SD (n = 6 biological replicates). Significant differences between samples (labelled with different letters) were determined by Brown‐Forsythe ANOVA with Tamhane's T2 post hoc test, p < 0.05.

To determine whether natural variation in GmST2 correlates with agronomic traits under saline conditions, we evaluated plant height and yield per plant of Hap1‐3 cultivars in saline fields. Two‐year field trials demonstrated that Hap1 soybeans exhibited significantly greater plant height and higher yield per plant than Hap2 (Figure 6D). Based on a non‐synonymous SNP in the GmST2 coding region, haplotypes were further classified as HapA (including Hap1 and Hap3) and HapC (including Hap2). Consistent with the initial haplotype comparison, HapA cultivars demonstrated superior salinity tolerance, with higher yield and increased plant height relative to HapC (Figure 6E).

To investigate whether the observed natural variation influences phenotypic differences through transcriptional regulation or protein function, we assessed promoter activity and transcriptional activation capacity across distinct GmST2 haplotypes. The promoters (2‐kb upstream of the translation start site) from GmST2 Hap1 and Hap2 were cloned into the pGreenII0800‐LUC vector upstream of LUC; transient expression assays in tobacco leaves revealed no significant difference in promoter activity between haplotypes, as quantified by the LUC/REN ratio (Figure S20). Subsequently, the coding sequences of GmST2 HapA and HapC were cloned into the yeast two‐hybrid vector pGBKT7 to generate DNA‐binding domain (BD) fusion proteins. Concurrently, the pGADT7 vector was modified either by removing the Activation Domain (AD) to create an empty vector control (VC) or by replacing AD with GmPRL1b. Yeast co‐transformation assays using these BD constructs and modified AD constructs demonstrated that GmST2^HapA exhibited significantly stronger basal transcriptional activation than GmST2^HapC, as measured by His, Ade, and MEL1 reporter gene activity; furthermore, co‐expression with GmPRL1b significantly enhanced the transcriptional activation by GmST2^HapA but had no significant effect on GmST2^HapC (Figure 6F,G).

These results collectively indicate that Hap1, shaped by artificial selection, confers plant height and yield advantages in saline environments. The causal basis of this agronomic improvement may lie in a non‐synonymous SNP within Hap1, which enhances GmST2's transcriptional activation capacity. This functional divergence explains the observed phenotypic superiority of Hap1‐carrying soybeans in saline fields, positioning GmST2 as a promising genetic target for salinity‐tolerant soybean breeding programmes.

3. Discussion

To enhance environmental adaptability, plants allocate their limited resources using various mechanisms, leading to trade‐offs between growth and responses to various stress factors (Wang et al. 2024; Xie et al. 2024). NAC transcription factors are crucial regulators of plant stress responses (Javed et al. 2020). Individual NAC transcription factors may participate in multiple growth‐ or stress‐related biological processes, but few are capable of simultaneously promoting growth and resistance to a wide range of stress factors (Han et al. 2023). In this study, we identified the soybean NAC transcription factor GmST2, which positively regulates growth, salt stress tolerance, and resistance to B. cinerea simultaneously, and explored its regulatory mechanism and favourable allelic variants in a saline environment.

3.1. GmST2 Promotes JA Accumulation by Directly Targeting the Key Genes Involved in JA Biosynthesis

NAC transcription factors regulate various biological processes in plants by directly targeting the promoters of key genes involved in phytohormone biosynthesis to regulate their transcription and thus modulate phytohormone biosynthesis. NAC transcription factors directly regulate the expression of key genes involved in ABA, cytokinin, gibberellic acid, salicylic acid, and ethylene biosynthesis (Dong et al. 2022; Li et al. 2019; Wang et al. 2020, 2022). Lately, the NAC transcription factor in sweet potato IbNAC087 was found to directly bind to the promoters of IbLOX and IbAOS to promote JA biosynthesis (Li et al. 2024). Besides, the direct regulation of JA biosynthesis genes to modulate JA levels has not been fully investigated.

In the current study, RNA‐seq analysis of GmST2‐OE soybean and qRT‐PCR revealed that GmST2 upregulates genes encoding four enzymes in the JA biosynthesis pathway: LOX, AOS, AOC, and OPR (Figure 2C). Analysis of GmST2‐OE transgenic soybeans and gmst2 gmst2h mutants demonstrated that GmST2 positively regulates JA accumulation (Figure 2E,G). Since GmAOC4 is known to be involved in JA biosynthesis (Zhang et al. 2021), we selected GmAOC3 and GmAOC4, which were the most highly upregulated genes in GmST2‐OE versus WT plants, as candidate genes to verify the direct regulation of JA biosynthesis genes by GmST2. Using EMSA, we confirmed the binding of GmST2 to promoter fragments of GmAOCs in vitro and identified specific binding motifs in these promoters (Figure 3A,B). In ChIP‐qPCR assays, GmST2 directly bound to promoter fragments of GmAOCs containing the GmST2‐specific binding motifs in soybean in vivo (Figure 3C). Luciferase reporter assays in Arabidopsis protoplasts showed that GmST2 positively regulates the expression of GmAOCs (Figure 3D,E) and that this positive regulation depends on the specific binding motifs (Figure 3E). These findings suggest that GmST2 directly upregulates GmAOCs, which encode enzymes involved in JA biosynthesis, thereby promoting JA biosynthesis and participating in JA‐mediated responses to abiotic and biotic stress.

3.2. GmPRL1b Facilitates the Accumulation of GmST2

NAC transcription factors are precisely regulated through a hierarchical control system encompassing transcriptional regulation, post‐transcriptional modulation as demonstrated by miR164‐mediated degradation of Arabidopsis NAC4 (Myoung‐Hoon et al. 2017), protein modification, and post‐translational control exemplified by OsPUB43‐dependent ubiquitination and subsequent proteasomal degradation of OsNAC016 (Wu et al. 2022), collectively constituting an integrated regulatory network that orchestrates plant adaptation to environmental stresses (Han et al. 2023).

In this study, we revealed the interaction between GmST2 and the WD40 protein GmPRL1b in the soybean nucleus through Y2H, BiFC, and Co‐IP experiments in transgenic hairy roots (Figure 4A–C). Subsequently, using transient expression in N. benthamiana and soybean hairy root transformation, we demonstrated that GmPRL1b promotes GmST2 accumulation through post‐translational regulation rather than transcriptional control (Figure 4D–L). Furthermore, we detected significant upregulation of the GmST2 target genes GmAOCs, JA pathway marker gene GmPR1, and increased endogenous JA contents in GmPRL1b‐OE soybean hairy roots vs. the control (Figures 4I and 5E). However, in GmPRL1‐RNAi soybean hairy roots, the situation is the opposite (Figure 4L). This indicates that GmPRL1b promotes the accumulation of GmST2 through protein–protein interactions, thereby participating in the transcriptional regulatory pathway mediated by GmST2. Is the regulation of GmST2 accumulation by GmPRL1b mediated through the 26S proteasome‐dependent degradation pathway, similar to OsNAC016 and IbNAC087 (Li et al. 2024; Wu et al. 2022), or other known protein degradation pathways? We observed that the proteasome inhibitor MG132, the cysteine proteinase inhibitor E‐64, and the autophagy inhibitor BFA all did not enhance the accumulation of GmST2, suggesting that these pathways are not involved (Figure S21). The regulatory mechanism of GmPRL1b on GmST2 accumulation remains to be further investigated.

3.3. GmST2 and GmPRL1b Cooperatively Regulate Dual Resistance to Salt Stress and B. cinerea by Promoting JA Biosynthesis

Research on the effects of a single gene in simultaneously promoting tolerance to both abiotic and biotic stress is important for improving crop stress resistance through molecular design breeding. To date, few reports describe the functions and mechanisms of a single NAC transcription factor that participates in both plant stress tolerance and disease resistance pathways. ATAF1, the Arabidopsis homologue of GmST2, functions in an antagonistic manner in stress tolerance and disease resistance pathways (Brigitte and Victor 2009; Jensen et al. 2013; Wu et al. 2009). TaNAC8 promotes tolerance to low‐temperature and high‐salt stress in wheat while also enhancing defence responses against Puccinia striiformis (Xia et al. 2010), although the underlying mechanism remains unclear. CaNAC2c enhances thermotolerance by interacting with the heat shock protein HSP70 and promotes disease resistance by interacting with CaNAC029 in pepper. However, it has an inhibitory effect on plant growth (Cai et al. 2021).

In the current study, overexpressing GmST2 in soybean enhanced salt tolerance and B. cinerea resistance in leaves under both field and laboratory conditions, while the gmst2 gmst2h double mutant exhibited reduced salt tolerance and B. cinerea resistance (Figure 1, Figures S3 and S4). Transient overexpression and suppression of GmST2 in soybean hairy roots, as well as ectopic expression in Arabidopsis and N. benthamiana, yielded similar phenotypes (Figures S1, S7 and S8), indicating that GmST2 promotes salt tolerance and B. cinerea resistance across multiple species. Phenotypic restoration experiments under JA inhibitor and JA supplement indicated that GmST2 improves the salt tolerance in soybean in a JA dependent way (Figure 2H–M). Ectopic expression of GmPRL1b, encoding an interacting protein of GmST2, led to enhanced salt tolerance and B. cinerea resistance in Arabidopsis (Figures S17 and S18). Transient expression of GmPRL1b in tobacco leaves promoted B. cinerea resistance in N. benthamiana leaves and improved salt tolerance in soybean hairy roots, phenocopying the phenotypes of transgenic plants overexpressing GmST2 (Figure 5A–D). Additionally, like GmST2, overexpressing GmPRL1b promoted the accumulation of jasmonic acid (JA) in plants (Figure 5E). Furthermore, dual‐transformation assays in soybean hairy roots established GmPRL1b as an upstream regulator of GmST2, controlling the expression of JA biosynthesis genes (GmAOC3/4) and the JA pathway marker GmPR1 (Figure 5F). Previous extensive research has demonstrated the positive regulatory role of JA in plant salt tolerance and disease resistance (Wang et al. 2021). Therefore, increased JA levels may be one of the reasons for the common positive regulation of soybean salt tolerance and B. resistance by GmST2 and GmPRL1b.

During the process of coping with environmental stresses, plants exhibit a trade‐off between stress resistance and growth for efficient energy utilisation, with enhanced resistance often accompanied by growth inhibition (Liang et al. 2024; Xie et al. 2024). As one of the key hormones involved in this process, JA also causes growth inhibition (Chen et al. 2017; Gao et al. 2021; Wang et al. 2021). In this study, GmST2‐OE soybeans exhibited increased plant height compared to controls in both saline and non‐saline conditions (Figure 1D,E). Ectopic expression of GmST2 in Arabidopsis resulted in longer root lengths than controls under both mock and salt stress conditions (Figure S7). Conversely, both gmst2 gmst2h double mutants and GmST2 RNAi soybean hairy roots displayed shorter plant heights compared to controls (Figure 1J,K, Figure S1B,F), demonstrating that GmST2 positively regulates plant growth and suggesting that the JA pathway is not the sole mechanism by which GmST2 regulates soybean growth and resistance.

In summary, this study establishes a new pathway regulating plant salt tolerance and B. cinerea resistance through the GmPRL1b‐GmST2‐GmAOC3/4 module (Figure 7), providing theoretical support and genetic resources for the molecular improvement and breeding of crops with dual resistance traits.

FIGURE 7.

Model of the role of the GmPRL1b‐GmST2‐GmAOC3/4 module in regulating salt‐stress tolerance and Botrytis cinerea Resistance in Soybean. Salinity and B. cinerea induce the expression of GmST2, which promotes the transcription of the JA biosynthesis genes, such as GmAOCs (in direct way for GmAOC3/4), GmLOXs, GmAOXs and GmOPRs, thereby inducing JA accumulation. JA promotes the transcription of GmST2, and GmPRL1b promotes the accumulation of GmST2. The GmPRL1b‐GmST2‐GmAOC3/4 module functions as a positive feedback loop to mediate the accumulation of JA in response to stress and thereby improve salinity tolerance and B. cinerea resistance simultaneously. The solid and dashed arrows indicate direct and indirect regulation, respectively.

4. Materials and Methods

4.1. Salt Stress Tolerance Assay in Lab

The transgenic soybean ( Glycine max ) in cv Wei6823 background, mutants or hairy roots transiently transgenic plants in cv W82 were used for phenotypic assays in the lab, respectively. Seedlings were grown in wet vermiculite at a 16‐h photoperiod regime (light/dark temperature, 28°C/20°C) under a relative humidity of 60%. Salt stress was applied to soybean seedlings by the addition of 175 or 200 mM NaCl to the vermiculite. The fresh weight of each seedling without roots was measured. The newly expanded leaves were used for the photosystem II efficiency assay with MultispeQ v2.0 (PhotosynQ, USA).

4.2. Agronomic Trait Evaluation in Fields

For agronomic trait evaluation under field conditions, seeds were sown using a randomised complete block design with three replicates. Each plot comprised four randomly arranged rows containing wild‐type and three GmST2‐overexpression (OE) lines. Plots were spatially randomised within the experimental field. Uniform seed density per row was maintained across genotypes. Row spacing was set at 0.5 m, plot length at 2 m, and intra‐row plant spacing at 5 cm. Soil salinity, defined as the total soluble salt content per 100 g of dry soil, was quantified from surface soil samples (0–20 cm depth) collected 1 month prior to sowing. Field salinity values represent the mean of all plot measurements.

4.3. Generation of Transgenic or Gene Editing Soybean Plants

To generate GmST2‐overexpressing soybean lines, the coding sequence (CDS) of GmST2 was amplified from cDNA of soybean cultivar W82. The resulting PCR product was first cloned into the Gateway pDONR221 entry vector (Invitrogen, USA) via BP recombination. Subsequently, the CDS was transferred into the binary destination vector pB2GW7 (Plant Systems Biology, Ghent University, Belgium) through LR recombination to generate the 35Spro::GmST2 overexpression construct. Soybean transformation was performed using the Agrobacterium tumefaciens ‐mediated cotyledonary node method as previously described (Cui et al. 2013). Homozygous plants were used for subsequent phenotypic and molecular analyses. The primer sequences used are listed in Table S1.

The GmST2/GmST2h soybean mutants were generated by BioRun Co. Ltd. (Wuhan, China) using CRISPR‐Cas9 technology. Four sgRNA cassettes targeting both GmST2 and GmST2h were synthesised and ligated into the CRISPR‐Cas9 binary vector. The sgRNA sequences are provided in Table S1.

4.4. Botrytis cinerea Resistance Assay

Spores of Botrytis cinerea were inoculated onto V8 medium and incubated in the dark at 22°C for 6–10 days. The spores were then resuspended in PDA liquid medium (HB0233‐4, Hope Bio‐Technology, Qingdao, China) to achieve a concentration of 10⁶ spores per mL. Four 4 μL of the spore suspension were placed onto the centre of a leaf, and the leaves were incubated in the dark at 22°C for 1–3 days before photography. The diameters of the lesions were measured using ImageJ (1.52p) software (https://imagej.net/ij/) to assess the disease resistance of the leaves.

4.5. Hairy Root Transformation

The soybean hairy root transformation was conducted as described previously (Song et al. 2021).

4.6. cDNA Synthesis and RT‐qPCR

The cDNA synthesis, RT‐qPCR, and data analysis were conducted as described previously by (Li et al. 2016). Gene‐specific primer sequences were given in Table S2. GmTUB was used as the internal reference gene for profiling across the plant. For qPCR, the relative gene expression level was calculated using the 2^−ΔΔCT method (Livak and Schmittgen 2001).

4.7. RNA‐Seq Assay

Total RNA was extracted from roots of V₂ seedlings (Wei6823 and GmST2 OE‐1) using the RNeasy Plant Mini kit (Qiagen) according to the manufacturer's instructions. The total RNA from roots was isolated. A pool of tissues from at least three seedlings was considered a biological replicate. The mRNA sequencing libraries were constructed with barcodes using the TrueSeq RNA Sample Preparation kit (Illumina). Three biological replicates were sequenced on an Illumina HiSeq 2000 system by BGI‐Tech (Shenzhen, China), resulting in 16–18 million 49‐bp single‐end reads per sample. We used BWA (Li and Durbin 2009) to map clean reads to the soybean genome ( Glycine max Wm82.a2.v1 in Phytozome v11.0 database; https://phytozome.jgi.doe.gov/). DEG identification was based on the Noiseq method (Tarazona et al. 2011) with fold change > 2 and q < 0.05.

4.8. KEGG Pathway and GO Function Enrichment Analysis

Kyoto Encyclopedia of Genes and Genomes (KEGG; Kanehisa et al. (2008)), the major public pathway‐related database, was used to perform pathway enrichment analysis of DEGs. This analysis identifies significantly enriched metabolic pathways or signal transduction pathways in DEGs compared with the whole‐genome background. The calculated P‐value was subjected to Bonferroni correction, taking a corrected p‐value < 0.05 as a threshold for significance.

GO term enrichment analysis of the gene sets of interest was performed to identify enriched GO terms. The calculation of p‐value was conducted by the same method as that in KEGG pathway enrichment analysis. GO terms fulfilling this condition are defined as significantly enriched GO terms in DEGs.

4.9. Quantification of JA Content

JA contents were detected by MetWare (http://www.metware.cn/, Wuhan, China) based on the AB Sciex QTRAP 6500 LC–MS/MS platform.

4.10. Transient Assays of Gene Transcription

To study whether GmST2 can bind to specific motifs or promoters to regulate target gene expression, the pGreen II 0800‐LUC vector system (Hellens et al. 2005) was used, and the promoters were cloned into the vector to generate reporter constructs, respectively. Each reporter construct, together with either the 35Spro:GmST2 construct, was cotransformed into Arabidopsis (Col‐0) protoplasts using the methods described previously (Li et al. 2016). The signals of Firefly and Renilla LUC were assayed using the Dual‐Luciferase Reporter Assay System (Promega, USA).

4.11. Expression of Recombinant GmST2 Protein and EMSA

Full‐length GmST2 protein fused with GST was expressed in the vector pGEX‐T in the Escherichia coli strain BL21. The protein extraction and EMSA were conducted as described previously (Li et al. 2019). The oligonucleotide probe sequences are listed in Table S3.

4.12. ChIP‐qPCR Assays

The synthetic peptides of GmST2 (NDDYDLGLQLENAF) were injected into rabbits to generate the corresponding polyclonal antibodies by MW BIOTECH (HK) LIMITED. The specificities of anti‐GmST2 polyclonal antibodies were determined by immunoblot analysis using 35Spro:GmST2‐MYC or 35Spro:MYC transiently transgenic tobacco leaves (Figure S8). ChIP was conducted as described previously (Li et al. 2019). GmST2‐specific antibody and rabbit IgG protein (as mock sample) (A7016, Beyotime Biotechnology, Shanghai, China) were used. The primers used in this study are listed in Table S4.

4.13. Yeast Two‐Hybrid Assays

The full‐length GmST2 or GmPRL1b cDNA (complementary DNA) was cloned into the pGBKT7 bait vector or pGADT7 prey vector and transformed into the yeast strain AH109. Yeast cells carrying the bait vector were then transformed with the prey plasmids containing the full‐length GmST2 or GmPRL1b fragments. Transformants were selected on medium lacking histidine (His) but containing 5 mM or 10 mM 3‐AT. Primers used for plasmid construction in yeast two‐hybrid assays were listed in Table S1.

4.14. BiFC Assay

Full‐length GmST2 and GmPRL1b cDNAs were cloned into the pBiFC vectors, which contained either the N‐terminal or C‐terminal half of YFP. The resulting constructs were transformed into the Agrobacterium strain GV3101. The detailed protocol follows (Gampala et al. 2007). The primers used for plasmid construction for the BiFC assay were listed in Table S1.

4.15. Coimmunoprecipitation Assay

Agrobacteria (K599) containing the 35Spro::GmPRL1b‐GFP or 35Spro::GFP (as control) expression vector was transformed into the soybean hairy roots. About 40 h after infiltration, the transformed roots were collected for the coimmunoprecipitation assay (Gampala et al. 2007). The GmPRL1b‐GFP pulled down was detected with antibodies recognising GFP (HT801, TransGen Biotech, Beijing, China) and GmST2 detected with its specific antibody.

4.16. Statistical Methods

The distribution characteristics, the homogeneity of variances, and sample sizes of all the data were assessed, and appropriate statistical methods were used. The details for each analysis are described in figure legends. Statistical analyses were performed using GraphPad Prism version 8.0.2 (GraphPad Software, San Diego, CA, USA, www.graphpad.com).

4.17. Haplotype Analysis

To investigate natural variation at the GmST2 locus, we performed haplotype analysis on a diversity panel comprising 571 soybean accessions, including 55 wild, 142 landrace, and 374 cultivated soybeans, using geneHapR (Zhang et al. 2023).

4.18. Selection Signal Detection

To identify domestication‐associated selection signals, we conducted an XP‐CLR scan across a 100‐kb genomic region flanking GmST2 (Chen et al. 2010). Selection signatures during soybean domestication were evaluated by comparing allele frequency differences between landraces and wild soybeans ( G. soja ), with analysis parameters set as follows: ‐‐maxsnps 100, ‐‐size 40,000 and ‐‐step 2000.

4.19. Accession Numbers

Sequence data from this article can be found in the Phytozome under the following accession numbers: GmST2 (Glyma.06G157400), GmST2h (Glyma.04G249000), GmAOC3 (Glyma.19G044900), GmAOC4 (Glyma.18G280900), GmPRL1b (Glyma.15G149300), GmTUB (Glyma.08G014200), Gm60S (Glyma.13G318800), GmELF1b (Glyma.02G276600). The raw sequence data reported in this paper have been deposited in the Genome Sequence Archive (Chen et al. 2021) in the National Genomics Data Center (Members C‐N and Partners 2025), China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (CRA022563) that are publicly accessible at https://ngdc.cncb.ac.cn/gsa.

Author Contributions

S.L. and F.X. conceived the research and designed the experiments and S.L. wrote the manuscript. S.L., M.W., X.D., Y.P., Y.W., J.Y., D.J., M.S., G.B., X.L., and J.X. performed the experiments and analysed the data; S.L., F.X., and Q.L. revised the manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: GmST2 regulates the growth and salt tolerance in soybean seedlings with transgenic hairy roots.

Figure S2: Validation of GmST2 transgenic soybean.

Figure S3: The fresh weight of GmST2 OE transgenic soybeans and gmst2 gmst2h mutants under salt stress.

Figure S4: GmST2 promotes salt tolerance in OE transgenic soybean seedlings.

Figure S5: Phylogenetic tree of GmST2 and its homologues.

Figure S6: Construction of gmst2 gmst2h double mutant.

Figure S7: GmST2 promotes salt tolerance in transgenic Arabidopsis.

Figure S8: GmST2 improves B. cinerea resistance in transgenic Arabidopsis.

Figure S9: Characterisation of GmST2.

Figure S10: GmST2 antibody specificity.

Figure S11: GmST2 is induced by MeJA and H₂O₂.

Figure S12: DEG numbers in RNA‐seq assay.

Figure S13: EMSA of GmST2 and its potential binding sites in GmAOC3 and GmAOC4 promoters.

Figure S14: Phylogenetic tree of GmPRL1b and its homologues.

Figure S15: GmPRL1b localised in the nucleus.

Figure S16: The expression of GmPRL1b in response to NaCl, H₂O₂, MeJA, and B. cinerea treatment.

Figure S17: GmPRL1b promotes salt tolerance in transgenic Arabidopsis.

Figure S18: GmPRL1b improves Botrytis cinerea resistance in transgenic Arabidopsis.

Figure S19: Median‐joining network analysis of GmST2 haplotype origins.

Figure S20: Activity of GmST2 promoters of Hap1 and Hap2.

Figure S21: Effect of MG132, E‐64, and BFA on GmST2 protein stability in soybean seedlings.

Table S1: Primer sequences for vector construction.

Table S2: Primer sequences for RT‐qPCR.

Table S3: Probe sequences for EMSA.

Table S4: Primers for ChIP‐qPCR.

Table S5: Information on the soybean accessions used in this study.

Table S6: Haplotypes of GmST2.

Data Availability Statement

All data are available in the main text or the Supporting Information.