Pseudomonads associated to salt-stressed plants facilitate stress adaption of soybean through enhanced lignin biosynthesis

Abstract

Root-associated microbiota play a critical role in plant tolerance to salt stress. However, the conservation of beneficial interactions across diverse crops and soils and the underlying mechanisms remain unclear. Here, we show that pseudomonads were consistently enriched in salt-stressed plant roots across multiple soil types and most crop species. Comparative genomics revealed that these pseudomonads harbored unique genomic signatures associated with high salinity tolerance, such as Na⁺ transporters. Pseudomonad isolates from salt-stressed plants robustly colonized soybean roots and significantly improved salt tolerance under both greenhouse and field conditions. Pseudomonads-dependent plant salt stress tolerance was mediated through plant lignin biosynthesis stimulation rather than the canonical mechanism of Na⁺ homeostasis. Overexpression of the key plant lignin biosynthesis genes, including GmCAD, GmCOMT, and Gm4CL, significantly enhanced soybean growth under salt stress. Furthermore, mutant plants deficient in lignin biosynthesis no longer showed pseudomonads-induced salt tolerance. Collectively, our findings reveal a previously unrecognized microbial-mediated pathway that enhances plant resilience to salt stress.

Pseudomonads enhance soybean salt tolerance via plant lignin biosynthesis rather than the canonical mechanism of Na⁺ homeostasis.

INTRODUCTION

Soil salinization is a major global threat to soil fertility (1), biodiversity (2), and crop yield (3). Through climate change, the extent and persistence of salinization are predicted to exceed those observed in recent decades (4), posing a serious threat to global food security. Numerous studies on salt-tolerant plants have uncovered a range of adaptive strategies to cope with salt stress, such as ion homeostasis, accumulation of compatible solutes, and hormonal regulation (5, 6). Soil microbiota are key drivers of plant productivity and ecosystem function, and they also play a critical role in enhancing plant survival under saline conditions (7, 8). Therefore, understanding the interactions between soil microorganisms and plants is essential for improving crop performance in saline-affected soils and for supporting sustainable agriculture.

The“cry for help”hypothesis has recently been proposed to describe a strategy whereby plants under biotic or abiotic stresses actively secret specific compounds to recruit beneficial microorganisms that enhance stress adaptation (9, 10). Increasing evidence suggests that conserved microbial recruitment patterns may occur across diverse plant species in response to specific stresses. For example, Pseudomonas are frequently enriched in rhizospheres to resist pathogen attacks (11, 12). Under nitrogen deprivation, Massilia have been reported to be recruited across diverse plant species to improve nitrogen acquisition and host performance (13–15). Similarly, under drought stress, Streptomyces are consistently enriched across diverse plants and soil types (16–18). In a recent study, Zheng et al. (19) reported a significant enrichment of Pseudomonas within the salt-stressed roots of wild soybean (Glycine soja) and that such Pseudomonas isolates enhanced plant resilience to salt stress. On the basis of these observations, we hypothesize that a conserved plant-microbe interaction mechanism may exist for salt stress tolerance, likely mediated by a common group of beneficial bacteria.

The microbial-driven mechanisms for enhancing plant salt tolerance are proposed to include regulation of ion homeostasis, osmoprotectant synthesis, and reactive oxygen species (ROS) scavenging (20–22). Rhizospheric pseudomonads are one of the most widely recognized microorganisms to benefit plants (23), and growing evidence shows that they enhance plant resistance to both abiotic and biotic stresses by producing plant growth regulators, such as indole-3-acetic acid (IAA), siderophores, and antimicrobial compounds, as well as by inducing systemic resistance in plants (24–26). Recent studies have reported that some pseudomonads alleviate plant salt stress by moderating antioxidant enzyme activities, proline accumulation, and the Na⁺/K⁺ ratio (27, 28). Despite widespread claims regarding the physiological efficacy of pseudomonads inoculation (29), the molecular mechanisms underlying its role in enhancing plant salt tolerance remains limited.

This study hypothesized that salt stress enhances root-associated Pseudomonas populations across diverse plant species and soil types. To test this hypothesis, we used genome-resolved metagenomics to investigate the rhizosphere microbiome shifts associated with salt stress of wild soybean. We included multiple soil types spanning a latitudinal gradient across China, classified as alkaline or acidic, as well as diverse plant species, including the key crops maize (Zea mays) and sorghum (Sorghum bicolor), to determine whether the observed microbial response is species specific or broadly conserved. Functional differences between salt-enriched and salt-depleted microbial taxa were identified by comparative genomics, to uncover potential genetic adaptations underlying their stress-associated proliferation. Moreover, we conducted a series of genetic, transcriptomic, and inoculation experiments to elucidate the mechanisms by which Pseudomonas enhance plant salt tolerance.

RESULTS

Identification of salt-enriched taxa from metagenome-assembled genome analysis of wild soybean rhizosphere soils

To determine how salt stress shapes the rhizosphere microbiome, we revisited the metagenomic data from rhizosphere soils of salt-stressed and control wild soybean generated by Zheng et al. (19). A total of ∼264-Gb high-quality data were retrieved from 16 rhizosphere soils of wild soybean treated with salt stress (100, 200, and 300 mM NaCl) or without (0 mM NaCl; control). We extended the previous analysis, which focused on functional genes from the nonredundant gene set (19), by recovering metagenome-assembled genomes (MAGs) to better resolve the salt stress–induced shift in taxonomy at the genomic level. Binning of assembled contigs from these metagenomic data resulted in 164 nonredundant MAGs (completeness ≥ 50% and contamination ≤ 10%), including 157 bacterial MAGs and 7 archaeal MAGs (table S1). The most dominant phyla were Pseudomonadota (n = 48), Actinomycetota (n = 32), and Acidobacteriota (n = 20) (Fig. 1A). Consistent with the 16S ribosomal RNA (rRNA) gene barcoding analysis (19), the MAG dataset also demonstrated a strong enrichment of Pseudomonas and a large decrease in the relative abundance of Acidovorax in the rhizosphere community under all levels of salt stress, as compared to the control (Fig. 1B).

Fig. 1. Relative abundance and function of all MAGs obtained in this study.

(A) The maximum likelihood tree (IQ-TREE, based on concatenation of 400 ubiquitously conserved proteins) of the 164 MAGs was constructed using PhyloPhlAn 3.0. The tree scale indicates the number of substitutions per site. Tips of the tree are colored according to phylum-level taxonomy. The relative abundances of MAGs in different treatments (control, 100, 200, and 300 mM NaCl) are shown in the heatmap, with four biological replicates for each treatment. Bar plots represent the fold enrichment under salt stress. The outer filled stars indicate MAGs exhibiting a significant change (two-tailed Student’s t test, P < 0.05) between the control and at least one of the three salt treatment groups, whereas unfilled stars denote nonsignificant differences. (B) The top five enriched or depleted genera in salt-treated rhizosphere soils of wild soybean based on 16S rRNA gene barcoding (19) and MAG datasets. Red and blue fonts indicate the consistent enriched or depleted taxon between 16S rRNA gene barcoding and MAG datasets, respectively. (C and D) The principal coordinate analysis (PCoA) of COG (C) and KEGG (D) functions for the enriched and depleted MAGs. Statistical analysis was performed using ANOSIM. (E) Number of genes assigned to COG category N (cell motility) in salt-enriched MAGs.

To determine whether salt-enriched MAGs exhibited distinct functional profiles compared to salt-depleted MAGs, we identified 35 of 164 MAGs with significant abundance changes between control and salt treatments, including 21 salt-enriched and 14 salt-depleted MAGs (Fig. 1 and fig. S1). Functional annotation using the Clusters of Orthologous Groups (COG) database revealed no significant differences between the enriched and depleted groups [analysis of similarities (ANOSIM): R = 0.089, P = 0.053] (Fig. 1C and table S2). Using Kyoto Encyclopedia of Genes and Genomes (KEGG) annotations, we further assessed the plant growth–promoting (PGP) potential of MAGs and similarly found no statistically significant differences (ANOSIM: R = 0.046, P = 0.172; Fig. 1D and table S3). These results suggested that salt-enriched MAGs lacked common functional signatures, at least within the tested databases. Note that, previously, Zheng et al. (19) observed a marked enrichment of genes associated with cell motility (COG category N) in salt-treated samples. This enrichment could largely be attributed to the high number of motility-related genes in Pseudomonas MAG (e.g., bin149; Fig. 1E). Together, these findings provided genomic-level support for the previously observed enrichment of Pseudomonas in salt-stressed root systems (19). In contrast, Acidovorax was identified as the most strongly depleted taxon under salt stress.

Salt-mediated Pseudomonas enrichment is conserved across diverse soils

To examine whether plant enrichment of Pseudomonas under salt stress was conserved across different soils, we collected experimental soils from 10 different fields in China spanning a latitudinal range of 12° and different pH gradients (Fig. 2A and tables S4 and S5). Wild soybean seedlings were grown in the above soils for 10 days and then treated with 300 mM NaCl or sterile water (control) for 2 weeks in a green greenhouse (fig. S2). 16S rRNA gene barcoding analysis showed that salt stress significantly decreased the α-diversity of root microbiota in alkaline soils, while a similar but nonsignificant trend was observed in acidic soils (Fig. 2B). For β-diversity, microbial communities significantly differed between wild soybean grown in alkaline and acidic soils. Moreover, salt stress exerted a stronger influence on the microbial communities associated to plants grown in acidic soils compared to alkaline soils (Fig. 2C).

Fig. 2. The root-associated bacterial communities of wild soybean growing in alkaline and acidic soils.

(A) The sampling sites of soil used for the greenhouse experiment in this study. Word in red and blue colors indicate the sampling sites of alkaline soil and acidic soil, respectively. Sampling sites: NM, Neimeng; DZ, Dezhou; HD, Handan; XC, Xuchang; ZC, Zoucheng; TA, Taian; AH, Anhui; XX, Xiangxi; NP, Nanping; GY, Guiyang. (B) Bacterial community α-diversity (Shannon index) between control and salt stress for alkaline soil and acidic soil. Statistical analyses were performed by two-tailed Student’s t test. Tops and bottoms of boxes represent 25th and 75th percentiles, respectively. Horizontal bars within boxes denote medians, and the upper and lower whiskers represent the range of nonoutlier data values. (C) PCoA of root-associated bacterial communities (OTU level) between control and salt stress for alkaline soil and acidic soil. Statistical analysis was performed using ANOSIM. (D) The relative abundance and fold change (FC) of Pseudomonas in the roots of wild soybean growing in alkaline and acidic soils. (E) The linear discriminant analysis (LDA) scores to identify the salt-enriched taxa at the genus level, determined by the LDA of effect size (LEfSe) analysis. Only taxa with LDA scores of >3.5 in alkaline soil and >3.3 in acidic soil (to visualize Pseudomonas) are shown. ANPR, Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium; BCP, Burkholderia-Caballeronia-Paraburkholderia. (F) Manhattan plot displaying the taxonomic information of OTUs salt enriched or depleted in the root of wild soybean growing in alkaline soil and acidic soil. The threshold of significant changed OTU is P < 0.05 and |log₂(fold change)| > 1. Abs, absolute value. Salt stress in this experiment was applied with 300 mM NaCl. Statistical analyses were performed by two-tailed Student’s t test.

To determine how salt stress affected bacterial recruitment by wild soybean across different soils, we first analyzed the compositional profiles at the order level. Burkholderiales and Rhizobiales were dominant across both control and salt-treated samples, regardless of alkaline or acidic soils (fig. S3). Salt stress significantly increased the relative abundance of Pseudomonadales by 9.6-fold in the wild soybean grown in alkaline soils (Student’s t test, P = 9.5 × 10⁻⁷; from 1.4% of control to 13.6% of salt-treated samples), whereas no significant change was observed in acidic soils (Student’s t test, P = 0.71). Of the top 10 abundant genera, Pseudomonas exhibited the most pronounced positive response to salt stress (fig. S4), with fold increases ranging from 3.9 to 96.9 in alkaline soils (Fig. 2D). Note that a salt-induced enrichment of Pseudomonas was also observed in acidic soils (Fig. 2E), albeit with modest increases ranging from 0.7- to 4.0-fold. Further analysis of salt-enriched operational taxonomic units (OTUs) in alkaline soils revealed Pseudomonas OTU12019 as the most significantly enriched taxon in salt-treated wild soybean roots [Student’s t test, P = 2.1 × 10⁻⁶; log₂(fold change) = 5.2], with a weaker but still significant enrichment in acidic soils [Student’s t test, P = 0.01; log₂(fold change) = 1.7] (Fig. 2F). Consistent with the MAG-based observations (Fig. 1), there was also a significant depletion of Acidovorax in the salt-treated roots, both in alkaline and acidic soils (Fig. 2E). Combined, these data implied that Pseudomonas had a competitive advantage over other root-associated bacteria under salt stress conditions, which was conserved across distinct soils but more pronounced in alkaline soils.

Salt-mediated Pseudomonas enrichment is conserved across plant hosts

To determine whether salt stress–induced Pseudomonas enrichment is conserved across different plants, six common crops (table S5), including maize (Z. mays), sorghum (S. bicolor), rice (Oryza sativa), wheat (Triticum aestivum), rapeseed (Brassica napus), and tomato (Solanum lycopersicum), were planted and studied in alkaline soils. Ten-day-old plants were subjected to 200 mM NaCl or sterile water (control), and root samples were collected after 2 weeks. For all plant species, salt stress significantly decreased bacterial α-diversity (Shannon index; fig. S5) and induced shifts in community compositions (Fig. 3A and fig. S6). Consistent with observations in wild soybean, the abundance of Acidovorax markedly decreased in the rhizosphere of all tested salt-stressed plant species (Fig. 3B). Similarly, Pseudomonas were the most strongly enriched taxa in the salt-stressed maize, sorghum, rapeseed, and tomato roots with their relative abundance increasing from 1.42 to 21.56% in control to 22.84 to 41.12% in salt-stressed samples (Fig. 3C). However, Pseudomonas were not enriched in the salt-stressed roots of rice and wheat (Fig. 3B). These data implied that salt stress–induced enrichment of Pseudomonas was conserved across many diverse cropped plants but that this was not a universal phenomenon.

Fig. 3. The root-associated bacterial communities across different plant species.

(A) PCoA of root-associated bacterial communities (OTU level) between control and salt stress for all plants. Statistical analysis was performed using ANOSIM. (B) The LDA scores to identify the salt-enriched taxa at the genus (g) level, determined by the LEfSe analysis. Only taxa with the LDA score > 3.5 are shown. (C) Relative abundance of Pseudomonas in the root-associated bacterial communities of control and salt-stressed plants. Statistical analyses were performed by two-tailed Student’s t test. Values are means ± SEM (n = 5). ns, not significant. Salt stress in this experiment was applied with 200 mM NaCl, as these crops have lower salt tolerance than wild soybean.

Pseudomonas harbor versatile genetic capabilities for salt tolerance

With the contrasting responses of Pseudomonas and Acidovorax to salt stress in most plants, we hypothesized that they had contrasting salt tolerance and genetic architectures for salt stress resistance. To test this hypothesis, we reviewed published data on the salt tolerance of their type strains and found that Pseudomonas strains tolerated NaCl concentrations up to 4.9%, whereas most Acidovorax species exhibited limited growth at around 1.5% NaCl (Fig. 4A and table S6). Next, comparative genomics was conducted on 359 reference Pseudomonas genomes and 128 Acidovorax genomes publicly available in the National Center for Biotechnology Information (NCBI) database (tables S7 and S8). Compiling the geographical source of each genome showed that Pseudomonas were potentially distributed across more diverse habitats than Acidovorax (Fig. 4, B and C), although this pattern may be influenced by unequal genome number for the two genera. With a focus on genes associated with PGP traits and salt stress alleviation, functional gene annotation against the KEGG database revealed that while PGP (e.g., 1-aminocyclopropane-1-carboxylate deaminase, IAA biosynthesis, and phosphatases)–related genes showed some variability between Pseudomonas and Acidovorax, the most notable differences were observed in stress-related gene contents. Compared to Acidovorax, Pseudomonas harbored far more genes involved in antioxidation, betaine synthesis and transport, and Na⁺ transport (Fig. 4D).

Fig. 4. The genetic potential of Pseudomonas for salt tolerance.

(A) The salt tolerance of Pseudomonas (n = 78) and Acidovorax (n = 15) type strains reported in published literature (table S6). (B) The global distribution of Pseudomonas and Acidovorax species. The world map was generated using the maps and ggplot2 packages in R. (C) The maximum likelihood tree (IQ-TREE, based on concatenation of 400 ubiquitously conserved proteins) of Pseudomonas and Acidovorax genomes was constructed using PhyloPhlAn 3.0. The number of biosynthetic gene clusters (BGCs) related to salt tolerance is shown in the heatmap. Bar plots represent the total BGCs related to salt tolerance in each genome. (D) The predicted genes associated with PGP traits and salt stress alleviation in Pseudomonas and Acidovorax genomes. ACC, 1-aminocyclopropane-1-carboxylate; KO, KEGG Orthology. (E) N-acetylglutaminylglutamine amide (NAGGN) BGC in genomes of Pseudomonas strains XN05-1 and YE17. Genes highlighted in red are those involved in the synthesis of NAGGN, which were subsequently knocked out in the following experiment. (F) The growth curves of wild-type (WT) and NAGGN biosynthesis mutants (named ΔnbsABC) of Pseudomonas strains XN05-1 and YE17 across a range of NaCl concentrations (0 to 8%, w/v) over a 36-hour period. Statistical analyses were performed using repeated-measures two-way analysis of variance (ANOVA) with the Geisser-Greenhouse correction. Values are means ± SEM (n = 3).

Using antibiotics and Secondary Metabolite Analysis SHell (antiSMASH), 3873 and 616 putative biosynthetic gene clusters (BGCs) were identified in Pseudomonas and Acidovorax, respectively. On average, the number of BGCs per genome was twofold higher in Pseudomonas than in Acidovorax (fig. S7A), suggesting that Pseudomonas species have greater genetic potential for secondary metabolite biosynthesis. Moreover, these two genera displayed different BGC profiles. The predicted product classes of Pseudomonas mainly comprised nonribosomal peptides (NRPS, 32.5%), ribosomally synthesized and posttranslationally modified peptides (RiPP, 13.1%), redox cofactor (9.2%), and N-acetylglutaminylglutamine amide (NAGGN, 8.6%), while Acidovorax comprised terpenes (23.5%), RiPP (19.5%), betalactones (15.7%), and NRPS (14.4%) (fig. S7, B and C). Notably, the potential secondary metabolites related to salt tolerance, including NRPS, siderophore, NAGGN, and ectoine, were found to be far more prevalent in Pseudomonas compared to Acidovorax species (Fig. 4C). In particular, the compatible solute NAGGN was ubiquitously present across nearly all Pseudomonas species, whereas it was not predicted in any Acidovorax species. Given that some Pseudomonas and Acidovorax genome assemblies were incomplete, we reanalyzed salt tolerance–related pathways using only complete genomes to assess whether genome completeness affected our estimation of these functions. The results were consistent with those from the full genome datasets (fig. S8), indicating that the observed differences between the two genera are robust and not affected by the genome assembly level.

Since NAGGN synthesis potential was the most conspicuous predicted metabolic difference between Pseudomonas and Acidovorax, we hypothesized that it played a key role in salt tolerance in Pseudomonas. To test this, two representative Pseudomonas strains, Pseudomonas stutzeri XN05-1 and Pseudomonas frederiksbergensis YE17, previously isolated from salt-treated wild soybean roots in Zheng et al. (19), were selected for further experiments. AntiSMASH analysis revealed that both XN05-1 and YE17 strains harbored the NAGGN BGC (fig. S9). This cluster comprised three NAGGN synthesis genes encoding N-acetylglutaminylglutamine amidotransferase, N-acetylglutaminylglutamine synthetase, and peptidase (designated as nbsABC in this study; Fig. 4E and table S9). Deletion of nbsABC in strain XN05-1 severely impaired growth compared to the wild-type (WT) strain in media containing 3 to 7% NaCl (w/v), and strain YE17 exhibited markedly reduced growth under 3 to 5% NaCl (w/v) conditions (Fig. 4F). These results suggest that NAGGN biosynthesis represents one of the key strategies used by Pseudomonas to tolerate high salinity, which may contribute to their enrichment in plant roots under salt stress.

Pseudomonas alleviate soybean salt stress via a noncanonical salt tolerance mechanism

The consistent enrichment of salt-tolerant Pseudomonas in roots of many plant species under salt stress made it an ideal model system to examine their potential interaction and benefits to salt-stressed plants. Soybean (Glycine max), an economically important crop that has evolved from wild soybean, was used as the model plant. To determine whether Pseudomonas could colonize soybean roots, Pseudomonas strains XN05-1 and YE17 were inoculated onto soybean seedlings, and their root colonization was examined using scanning electron microscopy. Both Pseudomonas strains robustly adhered to the root surface (Fig. 5A). To quantify colonization dynamics under salt stress (50 mM NaCl), Pseudomonas strains were genetically tagged with green fluorescent protein (GFP), which showed no detectable effects on motility and only a minor effect on growth kinetics (fig. S10). The fluorescence intensity revealed that both strains exhibited a twofold increase in root colonization under salt stress compared to control conditions (Fig. 5, B and C). Colony-forming unit (CFU) enumeration further confirmed effective root colonization, with strains XN05-1 and YE17 reaching densities of 1.78 × 10⁶ and 1.12 × 10⁶ CFU per gram of fresh root under salt stress, respectively. Consistent with their enhanced colonization, salt-stressed soybean plants inoculated with either XN05-1 or YE17 exhibited significant growth promotion, particularly in root growth and development (fig. S11, A and B). Morphological observations further confirmed that both Pseudomonas strains significantly promoted root elongation and lateral root formation under salt stress (Fig. 5, D to F, and fig. S11C).

Fig. 5. Colonization by Pseudomonas XN05-1 and YE17 and the transcriptional response of soybean roots.

(A) Scanning electron microscopy images of Pseudomonas strains alone or in the root surface. Scale bars, 2 μm. (B) Colonization patterns of strains XN05-1 and YE17 in soybean roots visualized by GFP labeling. Scale bars, 75 μm. (C) Quantitative analysis of the fluorescent intensity, n = 20 image area (15.0 × 1.0 pixel). (D to F) Roots were characterized using the WinRHIZO root analysis system (n = 5), including total root length (D), root surface area (E), and root volume (F). (G and H) The volcano plots show the differentially expressed genes (DEGs) induced by inoculation with strains XN05-1 and YE17 under salt stress. (I) Expression levels of GmNHX, GmCHX, and GmHKT genes based on normalized transcripts per million (TPM) values. GmNHX, Na⁺/H⁺ exchanger; GmCHX, cation/H⁺ exchanger; GmHKT, high-affinity K⁺ transporter. (J and K) Na⁺ and K⁺ content, n = 6 root samples. DW, dry weight. For (A) to (C), Pseudomonas colonization experiment was conducted in Hoagland solution with or without salt stress (50 mM NaCl) to obtain clean roots for observation. For (D) to (K), Pseudomonas inoculation experiment was performed in vermiculite with or without salt stress (300 mM NaCl). Exact P values were calculated using a two-tailed Student’s t test. All data are means ± SEM.

To elucidate the molecular mechanisms underlying the enhancement of salt tolerance in soybean by Pseudomonas, RNA sequencing (RNA-seq) analysis was performed on the roots of soybean inoculated and uninoculated (mock) with strains XN05-1 and YE17. Compared to the control condition, the inoculated soybean roots exhibited pronounced transcriptional reprogramming under salt stress conditions (300 mM NaCl) (Fig. 5, G and H, and fig. S12, A and B). Specifically, inoculation with strains XN05-1 and YE17 under salt stress resulted in 5972 (up: 1628, down: 4344) and 928 (up: 381, down: 547) differentially expressed genes (DEGs), respectively. In contrast, under control condition, only 177 and 149 DEGs were identified for strains XN05-1 and YE17, respectively (table S10). Given that Na⁺ toxicity will directly impair root growth, we hypothesized that the observed growth promotion by Pseudomonas involves Na⁺ homeostasis regulation. Unexpectedly, Na⁺ and K⁺ transporter genes (including members of the CHX, NHX, and HKT families) were not differentially expressed in response to either XN05-1 or YE17 inoculation (Fig. 5I and table S11). Consistent with this, ion content analysis revealed no significant differences in root Na⁺ or K⁺ contents between inoculated and mock-treated plants under either salt or control conditions (Fig. 5, J and K). Together, these data revealed that Pseudomonas strains XN05-1 and YE17 mitigated salt stress in soybean through a noncanonical salt tolerance mechanism.

Pseudomonas improve soybean salt tolerance by enhancing root lignin biosynthesis

To reveal the mechanisms underlying Pseudomonas-mediated salt tolerance, we performed KEGG enrichment analysis of the soybean root transcriptomes. A phenylpropanoid biosynthesis pathway was significantly enriched in roots inoculated with either XN05-1 or YE17 under salt stress (Fig. 6A) but not under control condition (fig. S12, C and D), indicating that Pseudomonas-mediated activation of this metabolic pathway was specific to salt stress. In the phenylpropanoid biosynthesis pathway, the gene Glyma.01G021000 was strongly induced by strains XN05-1 and YE17. This gene (named GmCAD) encodes cinnamyl alcohol dehydrogenase, an enzyme that plays a critical role in lignin synthesis. Notably, several other lignin-associated genes, including Gm4CL, GmCCR, GmCOMT, and GmPOX, were also up-regulated in the roots of soybean inoculated with Pseudomonas strains (Fig. 6B). Thus, we hypothesized that Pseudomonas may activate lignin biosynthesis in the roots to enhance soybean salt tolerance. To test this hypothesis, plant lignin content was quantified following Pseudomonas inoculation. Under salt stress, lignin levels in roots increased by 35.2 and 30.3% upon inoculation with strains XN05-1 and YE17, respectively, whereas no significant changes were observed under the control condition (Fig. 6C). Phloroglucinol staining also showed enhanced lignin deposition (red and purple coloration) in the root cell walls of inoculated plants (Fig. 6D). These results supported the notion that Pseudomonas-induced lignin biosynthesis was specifically triggered by salt stress.

Fig. 6. Pseudomonas strains XN05-1 and YE17 regulate lignin biosynthesis in soybean.

(A) KEGG enrichment analysis of DEGs in soybean root inoculated with XN05-1 and YE17 strains under salt stress (300 mM NaCl). (B) Expression level and fold change of lignin biosynthesis–related genes in soybean roots under mock inoculation and treatments with strains XN05-1 and YE17 under salt stress (300 mM NaCl). The gene heatmap was based on normalized TPM values. CAD, cinnamyl alcohol dehydrogenase; 4CL, 4-coumarate:CoA ligase; CCR, cinnamoyl-CoA reductase; COMT, caffeic acid O-methyltransferase; POX, peroxidase. (C) Lignin content in soybean roots under control and salt stress (300 mM NaCl) conditions following inoculation with XN05-1 and YE17 strains (n = 8). (D) Phloroglucinol was used for histochemical staining of root lignin, followed by sectioning and microscopic observation. Scale bars, 50 μm. (E) Phenotypes of transgenic plants under control and salt stress (50 mM NaCl) conditions in a hydroponic system. Scale bars, 1 cm. (F) Root elongation (%) of OE (overexpression) soybean lines (OE-GmCAD, OE-Gm4CL, and OE-GmCOMT) under control and salt stress (50 mM NaCl) conditions (n = 6 or 10). Fold change, transgenic plants /WT. (G) The salt injury index of WT and CR (CRISPR-Cas9 knockout) soybean lines (CR-Gmcad, CR-Gm4cl, and CR-Gmcomt) deficient in lignin biosynthesis after inoculating Pseudomonas (n = 8 to 10). (H) The soybean growth performance with or without Pseudomonas inoculation under field condition. (I to K) Root weight, nodule number, and pod number of soybean with or without Pseudomonas inoculation under field condition (n = 8 or 11). Exact P values were calculated using a two-tailed Student’s t test. All data are means ± SEM.

To further validate the role of lignin biosynthesis in soybean salt tolerance, plant lines overexpressing GmCAD, Gm4CL, and GmCOMT (OE-GmCAD, OE-Gm4CL, and OE-GmCOMT) were generated using a transgenic hairy root system (fig. S13). Compared to the WT, all overexpression lines exhibited enhanced salt tolerance, as evidenced by a marked reduction in leaf chlorosis symptoms (Fig. 6E). In addition, root elongation was significantly improved in OE-GmCAD, OE-Gm4CL, and OE-GmCOMT plants compared to WT under both control and salt stress conditions (Fig. 6F). This improvement in root elongation was more pronounced under salt stress, showing a 3.3- to 13.4-fold increase compared to the 2.5- to 3.4-fold improvement under control conditions, suggesting that overexpression of lignin biosynthesis genes conferred a salt stress–specific advantage. Supporting this conclusion, soybean plants with disrupted lignin biosynthesis genes no longer displayed Pseudomonas-induced salt tolerance (Fig. 6G), confirming lignin biosynthesis as a critical mechanism underlying this beneficial interaction.

Field trials are essential for evaluating microbial inoculants under real-world conditions. A field experiment conducted in naturally saline soil (salinity of 0.42%) showed that soybean plants inoculated with Pseudomonas strains exhibited superior growth performance relative to uninoculated controls (Fig. 6H). Specifically, inoculation with Pseudomonas led to notable increases in root biomass, nodule number, and pod number compared to noninoculated plants (Fig. 6, I to K). These findings indicated that the application of Pseudomonas strains enriched from salt-stressed plants offers a promising strategy for improving crop performance in saline agriculture. Collectively, our findings establish a mechanistic link between Pseudomonas symbiosis and enhanced salt tolerance in soybean, wherein root-associated Pseudomonas activate the host’s lignin biosynthesis, leading to root architectural modifications that contribute to improved stress resilience (Fig. 7). However, further work is needed to establish exactly how Pseudomonas enhanced plant lignin gene expression, if this mechanism is conserved in the other plants for which pseudomonads were enriched by salt stress.

Fig. 7. Conceptual diagram of the conserved Pseudomonas in enhancing soybean salt tolerance.

Pseudomonas functions as a specific signal that induces the expression of lignin biosynthesis–related genes, rather than Na⁺ homeostasis, thereby promoting lignin deposition in the cell wall and improving plant salt tolerance. Created in BioRender. Zheng, Y. (2026) https://BioRender.com/5986dqo.

DISCUSSION

Although plants have evolved diverse adaptive mechanisms to cope with biotic and abiotic stresses in nature, they still rely on their microbial partners to bolster survival and provide additional protection against these stresses (30–32). In this study, genome-wide insights were provided into shifts of root microbiota composition of wild soybean under salt stress. The observed enrichment and depletion of specific taxa, exemplified by Pseudomonas and Acidovorax, respectively (Fig. 1), supported the notion that abiotic stresses substantially influenced the assembly of plant-associated microbiomes (13, 33). Extending this, we further found that the salt-induced enrichment of Pseudomonas was a highly conserved phenomenon across most tested diverse plant species and soil backgrounds (Figs. 2 and 3), indicating its selective advantage of salt adaption and beneficial roles in stress alleviation (Fig. 4). We experimentally confirmed that Pseudomonas enhanced soybean salt tolerance through a lignin-mediated mechanism, in contrast to the widely recognized mechanism of ion homeostasis (Figs. 5 and 6). This finding highlights the potential of harnessing root-associated microbiota to improve crop resilience under saline conditions.

Different plant species have unique root architectures, exudation profiles, and abiotic stress tolerance, all of which have been shown to shape symbiotic microbial communities (34, 35). Unexpectedly, most crops examined here exhibited a conserved pattern of Pseudomonas enrichment. This finding aligns with previous reports of salt-induced Pseudomonas enrichment in both salt-sensitive and salt-tolerant plants within the family Curcurbitaceae, such as Cucumis sativus, Cucurbita ficifolia, and Lagenaria siceraria (36). Together with our prior demonstration of this phenomenon in salt-stressed wild and domesticated soybean (19), these results support the conclusion that Pseudomonas enrichment is a broadly conserved response to salt stress across diverse plant species, despite notable exceptions in rice and wheat. These exceptions might be caused by either lineage-specific host selection mechanisms or differential host genetic factors that influenced microbiome recruitment. To further elucidate the interplay between salt stress and Pseudomonas enrichment in plants, future research should determine the salt stress threshold needed for Pseudomonas enrichment and assess whether this relationship follows a linear pattern or exhibits greater complexity.

Pseudomonas species exhibit substantial plant-beneficial traits, notably the production of diverse bioactive metabolites (25) and other compounds that help plants withstand biotic (37, 38) and abiotic stresses (39). These may explain its widespread occurrence in salt-stressed plants. Members of Pseudomonas have been shown to synthesize the compatible solute NAGGN in response to osmotic stress (40–42). Consistent with this, comparative genomic analysis revealed that the NAGGN BGC is widely present across nearly all Pseudomonas species (Fig. 4C). Disruption of the NAGGN biosynthetic pathway significantly impaired the growth of Pseudomonas strains under saline conditions (Fig. 4F). This osmoprotective mechanism may contribute to the observed enrichment of Pseudomonas in salt-stressed root microbiomes. Given the conserved nature of this enrichment, we propose the Pseudomonas genus as a promising model system for investigating plant-microbe interactions under saline conditions.

Microbe-mediated strategies for plant adaptation to salt stress have gained wide recognition, with core mechanisms involving the maintenance of ion homeostasis, scavenging of ROS, and activation of stress response signaling pathways (20, 21, 43). Among these, the regulation of Na⁺ homeostasis is regarded as a key mechanism by which plants adapt to salt stress conditions (6). However, our RNA-seq analysis revealed that Pseudomonas did not directly activate genes or pathways related to Na⁺ transport, such as CHX and NHX genes (Fig. 5I). We found that Pseudomonas significantly induced enrichment of the phenylpropanoid metabolic pathway, with multiple lignin biosynthesis–related genes (GmCAD, Gm4CL, and GmCOMT) exhibiting marked differential expression (Fig. 6, A and B). This led to elevated root lignin content and evident lignin deposition in the cell wall (Fig. 6, C and D). Lignin is a major component of the cell wall and plays a critical role in enhancing cell wall mechanical strength and structural stability (44). It has been demonstrated that increasing lignin content can significantly enhance plant salt tolerance and growth performance (45, 46). For instance, in tomato, overexpression of SlCOMT2 promoted both plant growth and salt tolerance (47). Notably, Pseudomonas-induced lignin biosynthesis was mainly observed under salt stress, indicating that stress condition was essential for the activation of this pathway by Pseudomonas. Our previous work has identified purines as key chemical signals mediating the enrichment of Pseudomonas under salt stress (19). These findings together suggested a coordinated interaction, in which salt stress first promoted the secretion of root-derived purines to recruit Pseudomonas. In turn, Pseudomonas functioned as a specific signal that activated the host lignin biosynthetic pathway, thereby enhancing plant stress tolerance. However, future studies are needed to elucidate how Pseudomonas activates these plant lignin biosynthetic genes.

Collectively, our study provides the most comprehensive investigation to date of how salt stress shapes root bacterial communities across diverse soil types and plant species. The discovery of the conserved enrichment of Pseudomonas in salt-stressed plant root underwrites its ecological significance as a key stress-alleviating taxon. We further establish Pseudomonas as a model system for investigating plant-microbe interactions under salt stress and evidence that specific strains enhance soybean salt tolerance through a lignin-dependent mechanism, revealing a previously unrecognized microbial strategy for modulating host salt resilience. These findings fill key knowledge gaps in understanding how environmental stress reconfigures the plant microbiome and identify lignin biosynthesis as a promising target for breeding salt-tolerant crops and screening of beneficial microbes.

MATERIALS AND METHODS

Metagenomic sequencing and binning

As described by Zheng et al. (19), 10-day-old wild soybeans were treated with 100, 200, or 300 mM NaCl (60 ml per pot, applied as 20 ml/day for 3 consecutive days), with sterile water as the control. Rhizosphere soils were collected at 14 days after the salt treatment, and six plants per pot were combined to generate one composite sample. Each treatment had four biological replicates.

Genomic DNA of rhizosphere soil was extracted using the FastDNA Spin Kit for Soil (MP Biomedicals) following the manufacturer’s protocol. Libraries were prepared without amplification and sequenced on the Illumina HiSeq X-Ten platform at the Majorbio Bio-Pharm Technology. After removing adapter and reads containing low-quality bases or 10% of undefined bases, ∼264-Gb high-quality data were generated from 16 samples, as reported previously by Zheng et al. (19). This study focused on MAGs to better resolve the salt stress–induced shift in taxonomy at the genomic level. Briefly, MetaWRAP v1.3.2 (48) was used for assembly and genomic binning. Two groups of coassemblies were performed using MegaHit v1.1.3 (49) through the “assembly module” in MetaWRAP, with one including only the control samples and the other including all salt treatment samples. Assembled contigs were binned using MetaBAT2 v2.12.1 (50), MaxBin2 v2.2.6 (51), and CONCOCT v1.0.0 (52) integrated within the MetaWRAP pipeline. The retrieved MAGs were refined using the “bin_refinement” module with the options -c 50 and -x 10. All bins were subsequently dereplicated using dRep v2.2.3 (53) with the following parameters: -sa 0.99 and -nc 0.1. Last, we obtained 164 nonredundant MAGs for downstream analyses.

MAG analyses

The completeness and contamination of MAGs were evaluated by CheckM v1.0.12 (54). Taxonomic classifications of MAGs were performed using the Genome Taxonomy Database Toolkit (GTDB-Tk) v2.4.0 (55) with “gtdbtk classify_wf” against the GTDB database release 220 (56). Genes were predicted using Prodigal v2.6.3 (57) and annotated with Prokka. Additional annotations were performed by aligning genes to the EggNOG 5.0 (58) database using eggNOG-mapper v2.1.12 (59). The KEGG and COG term annotation results for each gene were extracted from the output of eggNOG-mapper. Relative abundance of the MAGs in each sample was calculated by CoverM v0.7.0 (available at https://github.com/wwood/CoverM), with the parameter --min-read-percent-identity 0.95 and --min-read-aligned-percent 0.5. A phylogenetic tree of 164 MAGs was constructed using PhyloPhlAn 3.0 (60) based on concatenated alignments of up to 400 ubiquitously conserved proteins and then visualized using Interactive Tree Of Life (iTOL) (https://itol.embl.de/) (61).

Salt-induced Pseudomonas enrichment across different soils and hosts

For experiment 1, we aimed to characterize whether salt-induced Pseudomonas enrichment is conserved across different soil types. Bulk soils at a depth of 0 to 10 cm were collected from 10 geographically distinct fields across China, spanning a latitudinal range of 12° (table S4). Soil pH was determined in a mixture with 1:5 ratio of soil:water using pH meter (Thermo Orion Star A111, Thermo Fisher Scientific, Germany). We grew wild soybean plants using these soils under controlled conditions. Briefly, wild soybean seeds were surface sterilized with 0.15% mercuric chloride for 10 min and thoroughly washed with sterile water. Thereafter, the seeds were germinated in petri dishes with sterile water in the dark at 25°C for 2 days. Seedlings were transferred to plastic pots containing 200 g of field soil and incubated in a growth chamber at 25°C under a 16/8-hour light/dark photoperiod (light intensity, 200 μmol m⁻² s⁻¹) and 65% relative humidity. Plants were regularly watered with sterile water as needed. After 10 days of growth in pots, wild soybean plants were treated with 300 mM NaCl (60 ml per pot, applied as 20 ml/day for 3 consecutive days), while control plants received 60 ml of sterile water applied in the same manner. The experiment was performed with five biological replicates (table S5).

For experiment 2, we aimed to explore whether salt-induced Pseudomonas enrichment was conserved across different plant species. The plant materials used in this study included maize (Z. mays, cv. Zhengdan 958), sorghum (S. bicolor, cv. Kangsi), rice (O. sativa, cv. Zhenghan no. 10), wheat (T. aestivum, cv. Jimai 22), rapeseed (B. napus, cv. Zhongshuang 11), and tomato (S. lycopersicum, cv. Zhongshu no. 4), all of which are important crops. All seeds were surface sterilized with 0.15% mercuric chloride for 10 min and thoroughly washed with sterile water. They were germinated in petri dishes with sterile water in the dark at 25°C for 2 to 3 days. Seedlings were transferred to plastic pots, each containing 200 g of the same soil used in our previous study (19). Plants were regularly watered with sterile water as needed. Since these crops are not salt-tolerant plants, 200 mM NaCl (60 ml per pot, applied as 20 ml/day for 3 consecutive days) was imposed on 10-day-old plants. Control plants received 60 ml of sterile water applied in the same manner. The experiment was performed with five biological replicates (table S5).

Sampling, DNA extraction, and 16S rRNA gene barcoding

Since salt-induced Pseudomonas enrichment is more pronounced in root samples than in rhizosphere soil as described by Zheng et al. (19), we focused on root samples to further evaluate whether this pattern is conserved across soils and plant species. Root samples were collected at 14 days after the salt treatment. All plants within each pot (six for wild soybean, wheat, and rice; four for maize, sorghum, rapeseed, and tomato) were combined to generate one composite sample. To collect root samples, root-adhered soil was first removed by gently washing under running water. Roots were then further washed, sonicated, and frozen at −80°C until DNA extraction. Because of failure of DNA extraction or 16S rRNA gene amplification, several treatments included less than five but more than three biological replicates. The summary of sampling information is shown in table S5.

Genomic DNA was extracted using the FastDNA Spin Kit for Soil (MP Biomedicals) according to the manufacturer’s protocol. The concentration of extracted DNA was measured with the NanoDrop spectrophotometer (ND2000, Thermo Fisher Scientific, DE, USA). Amplification of the V5-V7 regions of the 16S rRNA gene was performed using the primers 799F and 1193R (table S12). Sequencing libraries were generated using the NEXTFLEX Rapid DNA-Seq Kit (Bioo Scientific, USA) following the manufacturer’s recommendations and sequenced on the Illumina MiSeq PE300 platform at the Majorbio Bio-Pharm Technology. Quality-filtered sequences were clustered into OTUs with a 97% sequence similarity using UPARSE pipeline (62). Representative sequences for each OTU were taxonomically classified with the Ribosomal Database Project (RDP) Classifier 2.13 (63) and annotated against the SILVA database (release 138). All OTUs identified as chloroplast and mitochondria were discarded from the dataset.

Pseudomonas and Acidovorax genomic analyses

To compare the metabolic potential of Pseudomonas and Acidovorax species, their representative genomes were collected by searching the NCBI’s Genome Browser (https://ncbi.nlm.nih.gov/datasets/genome/) using “Pseudomonas” and “Acidovorax” as keywords, respectively. Genome selection was performed using standardized filtering criteria. For Pseudomonas, only reference genomes were included, while atypical genomes, MAGs, and genomes from large multi-isolate projects were excluded. Using these criteria, 359 Pseudomonas genomes were retained from more than 45,000 available genomes (accession date: October 2024). For Acidovorax, only 12 reference genomes are available in NCBI. Therefore, we applied the same exclusion criteria (i.e., removal of atypical genomes, MAGs, and genomes from large multi-isolate projects) but did not restrict the dataset to reference genomes only, resulting in a total of 128 Acidovorax genomes (accession date: October 2024). Phylogenetic tree of the 359 Pseudomonas and 128 Acidovorax genomes were constructed using PhyloPhlAn 3.0 (60) based on concatenated alignments of up to 400 ubiquitously conserved proteins and then visualized using iTOL (61). The habitat and location for Pseudomonas and Acidovorax species were sourced from the metadata of genomes downloaded from NCBI or manually extracted from publications. Instances where information remained unclear were labeled as “unknown.”

BGCs in each genome were identified using antiSMASH v6.1.1 (64) with the following parameters: --genefinding-tool prodigal --tigrfam --cc-mibig --rre --cb-general --cb-knownclusters --cb-subclusters --asf --pfam2go --smcog-trees. Each BGC was functionally characterized on the basis of the predicted product types defined in BiG-SCAPE v1.1.5 (65).

NaCl tolerance of Pseudomonas and Acidovorax type strains

NaCl tolerance data for type strains of Pseudomonas and Acidovorax were compiled through a systematic survey of published literatures (accessed 1 February 2026), using the List of Prokaryotic names with Standing in Nomenclature (https://lpsn.dsmz.de) as the reference database. In total, NaCl tolerance information was available for 321 of 362 Pseudomonas and 14 of 22 Acidovorax type strains (table S6).

AntiSMASH prediction and construction of NAGGN synthetic gene mutants in Pseudomonas strains

On the basis of the genome sequences of Pseudomonas strains XN05-1 and YE17, the secondary metabolic gene cluster was predicted using the antiSMASH 7.1.0 online website (https://antismash.secondarymetabolites.org/). Three genes (ctg1_1282 to ctg1_1284 in strain XN05-1 and ctg1_3641 to ctg1_3643 in strain YE17, named nbsABC) are implicated in the biosynthesis of NAGGN (table S9). Mutants of nbsABC were generated via triparental conjugation using the suicide vector pK18mobsacB (19). Specifically, the upstream and downstream fragments of genomic DNA from strain XN05-1 were amplified using the primers XN05-1-NAGGN-UF/UR and XN05-1-NAGGN-DF/DR, respectively. The amplified fragments were purified and cloned into the suicide vector pK18mobsacB. The resulting recombinant plasmid was then transformed into an Amp^R derivative of XN05-1, with the aid of the helper plasmid pRK2013. Potential mutants were grown on nutrient agar (NA) medium containing 15% (w/v) sucrose to facilitate the excision of the suicide vector from the chromosome. The final mutant was confirmed using primers XN05-1-NAGGN-UF/XN05-1-NAGGN-DR and designated ΔnbsABC_{_XN05-1}. The NAGGN biosynthesis gene mutant of YE17, named ΔnbsABC_{_YE17}, was constructed following the same procedure using the corresponding primers. The primers used are listed in table S12.

Salt tolerance of WT and NAGGN biosynthesis mutants of Pseudomonas

To evaluate the salt tolerance of WT and NAGGN biosynthesis mutants, growth kinetics of each strain were monitored across a range of NaCl concentrations (0 to 8%, w/v) in nutrient broth (NB) media [peptone (10 g/liter) and beef extract (3 g/liter) pH 7.2 ± 0.2]. In a 96-well microplate, 198 μl of medium containing different NaCl concentrations was added to each well, followed by 2 μl of bacterial suspension [optical density at 600 nm (OD₆₀₀) = 0.4]. Each treatment was performed in triplicate. Microplates were incubated at 28°C with shaking at 180 rpm to ensure adequate aeration and prevent cell precipitation. The OD₆₀₀ values were measured every 4 hours over a 36-hour period to capture the full growth profile.

Root colonization analysis

To investigate the colonization of strains XN05-1 and YE17 on soybean roots under control and salt stress conditions, the strains were labeled with GFP. The GFP-expressing plasmid pBBR1MCS5-Tac-EGFP was transformed into the recipient bacterial cells via triparental conjugation, with pRK2013 used as the helper strain. GFP-labeled strains were cultured on NA plates supplemented with gentamicin (final concentration, 30 μg/ml). A single colony was then transferred into NB medium without antibiotic and cultured overnight. The culture was harvested by centrifugation and resuspended in sterile water to OD₆₀₀ = 1.0. Subsequently, 1 ml of the prepared suspension was added to 39 ml of 1/8-strength Hoagland solution (either without NaCl or supplemented with 50 mM NaCl) for cocultivation with soybean (cv. Zhonghuang 13) seedlings (66, 67). Each treatment was conducted in triplicate. After 3 days of inoculation, the roots of the soybean seedlings were rinsed under running tap water and observed using a confocal microscope (Leica Microsystems, Wetzlar, Germany). Fluorescence images were quantitatively analyzed using ImageJ software v1.54g. Briefly, each image was divided into consecutive regions (15.0 × 1.0 pixels) along the root axis to cover the entire imaged root system. The integrated density of each selected region was then measured to quantify fluorescence intensity.

Motility and growth of GFP-labeled and WT strains

Bacterial motility was determined using a semisolid agar assay (1% peptone, 0.5% NaCl, and 0.3% agar). Overnight cultures grown in NB medium at 28°C with shaking were adjusted to an OD₆₀₀ of 0.4. Then, 2 μl of each bacterial suspension was spotted onto the semisolid agar plates and incubated at 28°C for 18 hours. The diameter of the spreading zone was measured as an indicator of bacterial motility. Each treatment included six biological replicates. Growth was assessed according to the method described above for WT and NAGGN biosynthesis mutants.

Greenhouse and field experiments with Pseudomonas inoculation

Pseudomonas strains XN05-1 and YE17 were inoculated onto soybean plants with or without salt stress under greenhouse condition. Briefly, soybean seeds (cv. Zhonghuang 13) were surface-disinfected as described above. Five seeds were sown per pot containing 60 g of sterilized vermiculite, with 10 replicate pots per treatment. Pseudomonas strains XN05-1 and YE17 were cultured overnight in NB medium, followed by centrifugation at 4°C and 4000 rpm to concentrate the bacterial cells. The pellets were then resuspended in sterile water and inoculated into 10-day-old soybean seedlings with 20 ml of bacterial suspension (∼2.6 × 10⁷ CFU/ml) for each pot. Pseudomonas were applied three times at a 2-day interval. After the final inoculation, salt stress was imposed by applying 300 mM NaCl (60 ml per pot, applied as 20 ml/day for 3 consecutive days). Control plants received 60 ml of sterile water applied in the same manner. All plants were incubated in a climate chamber as described above and watered regularly with sterile water from the top. Watering was controlled to prevent drainage, and pots were placed on base trays to retain excess water and minimize bacterial loss. The growth phenotypes, root morphology, root Na⁺ and K⁺ contents, root lignin levels, and root RNA-seq were analyzed 10 days after the salt treatment. At the end of the experiment, Pseudomonas were quantified from soybean roots at a density of >4.2 × 10⁶ copies per gram of fresh root, confirming persistent root colonization. Quantification of Pseudomonas was performed as previously described by Zheng et al. (19).

A field experiment was conducted in natural saline soils (salinity of 0.42%) at Dongying, Shandong Province (37°17′28″N, 118°38′28″E), with treatments consisting of a noninoculated control and two Pseudomonas-inoculated groups (strains XN05-1 and YE17). Each treatment was replicated three times in plots measuring 5 m by 7 m, arranged in a randomized block design. Soybean seeds (cv. Zhonghuang 13) were grown with an interrow spacing of 40 cm and an intrarow spacing of 10 cm. Overnight cultures of Pseudomonas strains XN05-1 and YE17 were diluted by fourfold and then used to coat soybean seeds at a dosage of 40 ml/kg. Control seeds received equal amounts of sterile water in the same manner. The treated seeds were sown in the designated plots. At 40 days after planting, each plot was treated with 100 liters of a 50-fold diluted Pseudomonas suspension (prepared from an overnight culture) via root irrigation. Soybean plants were harvested at 70 days after planting to determine root weight, nodule number, and pod number.

Root morphology observation

Soybean roots were gently rinsed with water, spread evenly on a transparent tray, and scanned using root scanner (Epson V700, Beijing, China). Root morphological parameters, including length, surface area, and volume, were analyzed using WinRHIZO Pro v2007 (Regent Instruments, Canada).

Determination of root Na⁺ and K⁺ contents

The root samples were dried at 65°C until a constant weight was achieved, and the final dry mass was recorded. Subsequently, the samples were digested in nitric acid at 110°C for a duration of 6 hours. The concentrations of Na⁺ and K⁺ were quantified using inductively coupled plasma optical emission spectrometry (Varian Inc., USA).

Soybean root RNA-seq analysis

Root samples of soybean grown in vermiculite with or without salt stress (300 mM NaCl) were collected for RNA extraction and sequencing. Total RNA was extracted using TRIzol Reagent according to the manufacturer’s instructions. RNA-seq libraries were prepared using a polyadenylate selection strategy to enrich mRNA and sequenced on the NovaSeq 6000 platform. The raw paired end reads were trimmed and quality controlled by fastp (68) with default parameters. Quality-filtered reads were mapped exclusively to a reference soybean genome (Glycine_max_v2.1; http://plants.ensembl.org/Glycine_max/Info/Index) using HISAT2 v2.2.1 (69). Normalized read counts were calculated using the transcripts per million (TPM) method. DEGs were identified using DESeq2 v1.24.0 (70). Genes with an absolute |log₂ fold change| > 2 and a P value < 0.05 were considered significantly differentially expressed. Gene Ontology and KEGG pathway enrichment analyses of the DEGs were performed using the Goatools v0.6.5 (71) and Python SciPy v1.13.0 packages (72), respectively.

Quantification of lignin content in soybean roots

The soybean root samples were dried at 80°C until reaching a constant weight. Subsequently, the dried roots were finely ground and passed through a 40-mesh sieve. Lignin content in the soybean roots was extracted and quantified using a commercial kit purchased from Solarbio (Beijing, China). Briefly, 3 mg of dried root sample was subjected to acetylation of the phenolic hydroxyl groups in lignin. The resulting supernatant was thoroughly mixed with glacial acetic acid, followed by measurement of absorbance at 280 nm. The lignin content was then calculated using the formula provided in the manual.

Vector construction for the generation of transgenic soybean plants

To construct the GmCAD overexpression vector, the GmCAD coding sequence was first amplified from soybean root cDNA using primers GmCAD-F and GmCAD-R. Primers GmCAD-pFGC5941-F and GmCAD-pFGC5941-R were used to introduce restriction sites. The cloning vector pFGC5941 was digested with Bam HI and Sam I, and the amplified GmCAD product with restriction sites was subsequently ligated into the pFGC5941 vector (73). The recombinant vector was transformed into Escherichia coli DH5α for propagation. Plasmid DNA from positive clones was then transformed into Agrobacterium rhizogenes K599. Transformants were selected on kanamycin-containing medium (50 μg/ml) and sequenced for validation. The overexpression vectors of Gm4CL and GmCOMT were constructed following the same procedure using the corresponding primers. All primer sequences are provided in table S12.

To generate Gmcad loss-of-function mutants, CRISPR-Cas9–mediated genome editing was performed following a previous study (74). Target fragments were cloned into the pHSE401 vector via the Golden Gate assembly using Bsa I, and the resulting constructs were introduced into A. rhizogenes K599 for soybean transformation. Mutant lines of Gm4cl and Gmcomt were constructed in parallel using the same cloning and transformation workflow. Primer sequences used for cloning are provided in table S12.

Soybean hairy root transformation and phenotypic analysis

To generate composite soybean plants, A. rhizogenes strain K599 harboring overexpression vector was used to infect soybean seedlings following a previously described protocol with minor modifications (75). Briefly, seeds of the soybean cultivar Zhonghuang 13 were germinated for 4 days before inoculation with A. rhizogenes K599. Once transgenic hairy roots reached a length of 3 to 5 cm, the primary root was excised, and the plants were transferred to 1/8-strength Hoagland solution for 7 days. Initial hairy root lengths were recorded before salt treatment, after which the plants were transferred to fresh 1/8-Hoagland solution containing either 0 or 50 mM NaCl. After 2 weeks of salt stress, root lengths were measured again, and relative elongation was calculated as: (root length after salt treatment − root length before salt treatment)/root length before salt treatment. Roots were then harvested for subsequent gene expression analysis. The expression levels of GmCAD, Gm4CL, and GmCOMT were validated by quantitative reverse transcription polymerase chain reaction with specific primers (table S12).

For the Pseudomonas inoculation experiment, A. rhizogenes strain K599 carrying the CRISPR-Cas9 vector was used to infect soybean seedlings. Transgenic soybean hairy roots (20 days postinfection) were inoculated with a mixture of Pseudomonas strains XN05-1 and YE17. Coinoculation was chosen because the two isolates exhibited highly similar functional traits, including root colonization, intrinsic salt tolerance, mitigation of soybean salt stress, and induction of comparable root transcriptomic responses, and therefore was unlikely to confound validation of the lignin disruption experiment. Moreover, coinoculation may better reflect ecologically relevant conditions, as these strains naturally coexist in microbial communities. All plants were subsequently treated with 50 mM NaCl in a hydroponic system containing Hoagland solution. After 15 days of NaCl treatment, the salt injury index was calculated according to a previous study (76).

Statistical analyses

Principal coordinate analysis (PCoA) was used to ordinate the bacterial community based on the Bray-Curtis distance using the vegan v2.7.1 and ggplot2 v3.5.2 packages in R v3.5.3. Permutational Multivariate Analysis of Variance (PERMANOVA) was calculated on the basis of the Bray-Curtis distance with 999 permutations in R. Heatmap of the top 10 bacterial genera of control and salt treatment for acidic and alkaline soils was plotted using TBtools (77). Linear discriminant analysis (LDA) of effect size (LEfSe) was applied to identify the enriched bacterial taxa in control and salt treatments for all plant species. All statistical analyses were performed using two-tailed t tests, unless otherwise specified.

Supplementary Materials

The PDF file includes:

Figs. S1 to S13

Legends for tables S1 to S12

Other Supplementary Material for this manuscript includes the following:

Tables S1 to S12