Integrated morphological, physiological, and transcriptomic profiling unveils a root-centric defense network mediating aluminum tolerance in Trifolium repens L

Abstract

Aluminium (Al) toxicity in acid soils severely limits forage productivity, and dissecting Al-tolerance mechanisms is crucial for securing forage supply in acid-soil regions. White clover (Trifolium repens L.), an excellent acid-soil-adapted forage with pronounced Al tolerance, serves as an ideal model for studying legume Al-tolerance mechanisms. We exposed white clover seedlings to gradient Al^³⁺ concentrations (0, 2, 4, 6, 10 mmol·L^⁻¹) to characterize the core physiological and molecular responses underlying its Al adaptation. Al^³⁺ stress inhibited growth in a concentration-dependent manner, suppressing both root and shoot development. Roots adopted a prioritised defence strategy, with enhanced antioxidant enzyme activity and soluble sugar accumulation mitigating oxidative damage. Transcriptome analysis revealed coordinated regulation of key pathways: flavonoid biosynthesis showed a core inhibition–branch-specific activation pattern, photosystem-related genes were upregulated to reinforce photosynthetic function, and hormone signalling networks were extensively rewired with divergent responses among auxin, gibberellin, cytokinin, ABA and JA pathways. White clover copes with Al^³⁺ stress via an integrated mechanism featuring root-prioritised defence, photosynthetic maintenance and hormone network remodelling. These findings provide new insights into legume Al tolerance and a framework for breeding Al-tolerant forages. Future studies will quantify Al content in shoots and roots, and perform functional validation of upregulated hormone-related genes to clarify their roles in the Al-tolerance regulatory network.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12870-026-08323-1.

Introduction

Aluminum (Al) is the third most abundant element in the Earth’s crust, accounting for approximately 7% of its total mass, second only to oxygen and silicon [1]. Under neutral to weakly acidic soil conditions, Al primarily exists in inert forms such as aluminosilicates and oxides, posing little threat to plant growth. However, when soil pH falls below 5.0, these insoluble forms are solubilized into phytotoxic ionic species, mainly Al³⁺, which severely inhibits root elongation, nutrient absorption, and overall plant productivity [2, 3]. Acidic soils account for an estimated 30–40% of the world’s arable land and are particularly prevalent in tropical and subtropical regions, where they represent a significant constraint on agricultural output [4, 5].The root apex is the primary site for Al³⁺ perception and injury [6]. Upon contact, Al³⁺ compromises cell wall integrity, enhances Fe-mediated lipid peroxidation and protein oxidation, disrupts intracellular signal transduction, and directly binds to nucleic acids, thereby inhibiting cell division [7]. These disruptions lead to physiological and biochemical imbalances, resulting in impaired root growth and development [8]. For instance, Al³⁺ toxicity has been shown to suppress root cell growth in maize [9], rice [10], and tea [11]. Moreover, Al³⁺ stress triggers excessive accumulation of reactive oxygen species (ROS), resulting in oxidative damage to membranes and macromolecules [12, 13].

To mitigate Al toxicity, plants have evolved two major tolerance strategies: external exclusion and internal detoxification [14]. The external exclusion mechanism primarily involves root exudation of organic acids such as malate, citrate, and oxalate, which chelate Al³⁺ ions in the rhizosphere to form non-toxic complexes, thereby preventing their entry into root apex cells. For example, wheat secretes organic acids under acidic conditions, effectively reducing Al³⁺ accumulation in roots [15]. The internal detoxification mechanism, on the other hand, entails the sequestration of Al³⁺ into vacuoles or its complexation with organic ligands within the cytosol [16]. In Medicago sativa, overexpression of MsPG1 reduces Al binding to cell walls, improving Al tolerance [17]. Similarly, the induction of transporter genes such as AhFRDL1 in peanut [18] and MsMATE in alfalfa [19] enhances organic acid secretion for Al detoxification. Furthermore, transcription factors such as ASR5 and STOP1 play key roles in the regulation of Al-responsive genes in rice [20] and Arabidopsis [21].

White clover (Trifolium repens L.) is a perennial legume of global agricultural significance, widely used as a forage crop, ground cover, and soil-restoring species due to its high biomass production, strong nitrogen-fixing capacity, and tolerance to diverse environmental conditions [22–26] It demonstrates notable adaptability to acidic soils (pH 4.5–6.0) and can accumulate heavy metals such as Cd²⁺, Pb²⁺, Zn²⁺, and Cr₂O₇²⁻ in its roots and shoots with minimal physiological damage [27–29]. In southern China, a region among the world’s largest areas of acidic soils, Al toxicity represents a major limitation to forage productivity and ecological restoration efforts. However, the molecular and physiological bases of Al tolerance in T. repens remain insufficiently understood. Most existing studies have been limited to either morphological or physiological responses, with only recent work integrating physiology and transcriptomics under specific conditions. Furthermore, systematic analyses across Al concentration gradients that simultaneously assess morphology, physiology, and transcriptomics are still lacking.

Based on these gaps, we hypothesized that T. repens exhibits distinct adaptive responses to varying intensities of Al³⁺ stress, coordinated through root-based defense and systemic metabolic regulation. To test this hypothesis, we exposed white clover seedlings to five Al³⁺ concentrations under controlled conditions and combined morphological, physiological, and transcriptomic profiling to reveal the underlying tolerance mechanisms. Our objectives were to: (1) evaluate the concentration-dependent effects of Al stress on growth and oxidative metabolism; (2) identify differentially expressed genes and enriched pathways involved in Al tolerance; and (3) construct an integrative model describing the root-prioritized defense, photosynthetic maintenance, and hormone signaling coordination under Al stress. This study provides a comprehensive understanding of Al tolerance mechanisms in T. repens and offers theoretical support for improving Al-tolerant legume germplasm and sustainable grassland management in acidic soil ecosystems.

Materials and methods

Plant materials and growth conditions

Seeds of Trifolium repens L. (cv. ‘Merlyn’) were provided by the Grassland Science Laboratory, College of Animal Science, Guizhou University (Guiyang, China). Prior to germination, seeds were surface-sterilized in 3% (v/v) sodium hypochlorite solution for 5 min, rinsed thoroughly with sterile distilled water, and placed on moistened filter paper in Petri dishes for germination at 25 °C conditions. After seven days, uniformly germinated seedlings were transplanted into plastic pots filled with nutrient soil and cultivated under controlled growth conditions (25 ± 2 °C, 14 h light / 10 h dark photoperiod, and 60–70% relative humidity). After 40 days of growth, uniform seedlings were transferred into sand trays containing Hoagland’s nutrient solution for acclimation over three days before Al stress treatments.

Hoagland’s nutrient solution contained KNO₃ (506 mg·L^− 1), NH₄NO₃ (80 mg·L^− 1), KH₂PO₄ (136 mg·L^− 1), MgSO₄ (241 mg·L^− 1), FeNaEDTA (36.7 mg·L^− 1), KI (0.83 mg·L^− 1), H₃BO₃(6.2 mg·L^− 1), MnSO₄·H₂O (16.9 mg·L^− 1), ZnSO₄ (8.6 mg·L^− 1), Na₂MoO₄ (0.25 mg·L^− 1), CuSO₄ (0.025 mg·L^− 1), CoCl₂ (0.025 mg·L^− 1), and Ca(NO₃)₂ (945 mg·L^− 1), which was purchased from Solarbio Science & Technology Co., Ltd.

Aluminum stress treatments

Aluminum treatments were imposed by adding analytical-grade aluminum chloride hexahydrate (AlCl₃·6H₂O) to the nutrient solution at final concentrations of 0 (control, CK), 2 mmol·L^{− 1} (B2), 4 mmol L^{− 1} (B4), 6 mmol L^{− 1} (B6), and 10 mmol L^{− 1} (B10), with the gradient designed referring to Long et al. and Chen et al. [30, 31]. The pH of all treatment solutions was adjusted to 4.0 using 0.1 mol L^{− 1} HCl or NaOH to simulate acidic soil conditions.

Each treatment consisted of three biological replicates. Nutrient solutions were renewed every two days to maintain consistent ionic balance and Al³⁺ concentration. After 21 days of treatment, ten seedlings from each group were collected for morphological measurements. Root and leaf tissues were washed with distilled water, blotted dry, and rapidly frozen in liquid nitrogen. Samples were stored at − 80 °C for subsequent physiological and transcriptomic analyses.

Morphological measurements

Morphological parameters including plant height, Root and leaf parameters were determined. Plant height was measured using a ruler (precision ± 0.1 cm).

Root and leaf morphology were analyzed by scanning samples with an Epson Perfection V800 scanner (Seiko Epson Corp., Japan), and parameters were quantified using WinRHIZO and WinFOLIA software (Regent Instruments Inc., Canada).

For biomass determination, plants were rinsed with distilled water, blotted dry, and weighed to obtain fresh weight (FW). Samples were oven-dried at 105 °C for 20 min, followed by drying at 65 °C to a constant weight to obtain dry weight (DW). The dry-to-fresh weight ratio (DW/FW) was calculated to evaluate tissue water content.

Determination of physiological and biochemical parameters

Physiological and biochemical indices were determined separately for roots and leaves. Superoxide dismutase (SOD) activity was measured using the nitroblue tetrazolium (NBT) photoreduction method. Peroxidase (POD) activity was determined by the guaiacol oxidation method. Catalase (CAT) activity was assayed using the ultraviolet absorption method. Ascorbate peroxidase (APX) activity was analyzed by the ascorbate oxidation method. Malondialdehyde (MDA) content was quantified by the thiobarbituric acid (TBA) reaction method. Proline (Pro) content was determined using the ninhydrin colorimetric method. Soluble sugar (SS) and soluble protein (SP) contents were determined using the anthrone colorimetric and Coomassie Brilliant Blue G-250 methods, respectively. Hydrogen peroxide (H₂O₂) content was determined using the titanium reagent method. All assays were performed using commercial detection kits (Solarbio Science & Technology Co., Ltd., Beijing, China) following the manufacturer’s protocols.

RNA extraction, library construction, and transcriptome sequencing

Root samples from different Al treatment groups were sent to Novogene Bioinformatics Technology Co., Ltd. (Beijing, China) for RNA extraction and sequencing. Total RNA integrity and concentration were assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies, USA).

Messenger RNA (mRNA) was enriched using Oligo(dT) magnetic beads, fragmented into short fragments, and used for first-strand cDNA synthesis with random hexamer primers and M-MuLV reverse transcriptase. Second-strand cDNA was synthesized using DNA polymerase I and dNTPs, followed by purification, end repair, A-tailing, and adaptor ligation.

Fragments of approximately 370–420 bp were selected using AMPure XP beads (Beckman Coulter, USA). After PCR amplification and purification, cDNA libraries were quantified using a Qubit 2.0 Fluorometer (Thermo Fisher Scientific, USA), and their insert sizes were verified using an Agilent 2100 Bioanalyzer. Libraries meeting quality control criteria (effective concentration > 2 nM) were sequenced on an Illumina NovaSeq 6000 platform, generating 150 bp paired-end reads.

Differential gene expression and enrichment analyses

Raw reads were filtered to remove adapters and low-quality sequences. Differentially expressed genes (DEGs) were identified using the DESeq2 package (v1.20.0). Genes with an adjusted p-value (padj) < 0.05 and |log₂(fold change)| > 1 were considered significantly differentially expressed.

Gene Ontology (GO) functional enrichment analysis was performed using the GOseq package (v1.10.0), while Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment was conducted using KOBAS (v2.0.12). Significance of pathway enrichment was evaluated by Fisher’s exact test, and enrichment factors were calculated as:

Statistical analysis

All experiments were conducted with at least three biological replicates. Data were analyzed using IBM SPSS Statistics 21 (IBM Corp., Armonk, NY, USA). Differences among treatments were determined by one-way analysis of variance (ANOVA) followed by Duncan’s multiple range test at a significance level of P < 0.05. Pearson’s correlation analysis was used to evaluate relationships between physiological and morphological parameters. Graphs were plotted using Origin 2021 (OriginLab Corp., Northampton, MA, USA).

Results

Effects of aluminum stress on the growth characteristics of white clover seedlings

Plant height, biomass, and dry-to-fresh weight ratio

With increasing Al³⁺ concentration, growth of white clover seedlings was progressively inhibited (Fig. 1d). Compared with CK, plant height was significantly reduced at ≥ 4 mmol·L^− 1 Al³⁺ (P < 0.05, Fig. 1a), and biomass decreased significantly from 4 mmol·L^− 1 upward (P < 0.05, Fig. 1c), both reaching minimum values at 10 mmol·L^− 1. In contrast, the dry-to-fresh weight ratio increased significantly at 4 and 10 mmol·L^− 1 (P < 0.05, Fig. 1b), indicating altered water status and reduced tissue moisture under aluminum stress.

Fig. 1.

Effects of aluminum stress on plant height (a), dry-to-fresh weight ratio (b), biomass (c) and phenotypic differences in shoots and roots (d) of white clover seedlings (CK: 0 mmol·L^− 1, B2: 2 mmol·L^− 1, B4: 4 mmol·L^− 1, B6: 6 mmol·L^− 1, B10: 10 mmol·L^− 1)

Leaf morphology and root system architecture

Al³⁺ stress significantly inhibited leaf development. With increasing Al³⁺ concentration, leaf length, width, perimeter, and area all decreased markedly compared with the control (CK) (P < 0.05(Figure 2a, b, c, d). Under 10 mmol·L^− 1 Al³⁺ treatment, the above parameters were reduced by 42.61%, 41.06%, 41.83%, and 67.98% respectively, relative to the control.

Fig. 2.

Effects of aluminum stress on leaf length (a), leaf width (b), leaf perimeter (c), and leaf area (d), total root length (e), root volume (f), root surface area (g), average root diameter (h), and root tips number (i) of white clover seedlings

Root system architecture was strongly affected by Al³⁺. Total root length, root volume, root surface area, and root tip number all declined steadily with increasing Al³⁺, reaching their lowest values at 10 mmol·L^− 1, where they were reduced by 66.05%, 67.47%, 66.77%, and 63.22%, respectively, compared with CK (P < 0.05, Fig. 2e, f, g, i). The total root length showed no significant difference at a concentration of 4 mmol·L^− 1 compared to that 2 mmol·L^− 1, average root diameter decreased significantly at 4 and 6 mmol·L^− 1 (P < 0.05), whereas changes at 2 and 10 mmol·L^− 1 were not significant (Fig. 2h).

Effects of aluminum stress on physiological parameters of white clover seedlings

Leaf physiological parameters

Al³⁺ stress significantly altered antioxidant activity and osmotic adjustment in leaves (Fig. 3). Activities of APX and POD, as well as soluble sugar and soluble protein contents, all decreased with increasing Al³⁺ compared with CK (Fig. 3a, c, g, h). APX activity was reduced by 45.36–68.78% across Al³⁺ treatments (P < 0.05, Fig. 3a), and POD activity by 34.67–48.57% (P < 0.05, Fig. 3c). Soluble sugar content decreased by 36.96–59.42% (Fig. 3g), and soluble protein by about 36% at several concentrations (P < 0.05, Fig. 3h). CAT activity was markedly inhibited at 2 mmol·L^− 1 and 10 mmol·L^− 1, with reductions of 87.25% and 77.91%, respectively (P < 0.05, Fig. 3d). No significant differences were observed in SOD activity and proline content between Al³⁺-treated plants and the control (Fig. 3b, f). MDA content first increased and then decreased with rising Al³⁺ concentration. At 2 and 4 mmol·L^− 1 Al³⁺, it was significantly 108.55% and 87.62% higher than CK, respectively (P < 0.05, Fig. 3e). In contrast, H₂O₂ content decreased continuously, reaching the lowest at 10 mmol·L^− 1, which was 25.06% lower than CK (P < 0.05, Fig. 3i).

Fig. 3.

Effects of aluminum stress on APX (a), SOD(b), POD (c), CAT (d), MDA (e) Proline (f), Soluble sugar (g), soluble proteins (h), and H₂O₂ (i) in the leaves of white clover seedlings

Root physiological parameters

Root responses to Al³⁺ differed markedly from those in leaves. Root APX activity and soluble sugar content increased significantly with Al³⁺ concentrations, both peaking at 10 mmol·L^− 1, and reaching 4.25 -fold and 4.11-fold of CK, respectively (P < 0.05, Fig. 4a, g). In contrast, SOD activity and soluble protein content decreased significantly with increasing Al³⁺ concentration (P < 0.05). At 2, 4, 6, and 10 mmol·L^− 1 Al³⁺, SOD activity was reduced by 38.11%, 61.14%, 58.51%, and 72.21%, respectively, while soluble protein content declined progressively by 28.47%, 46.53%, 64.06%, and 34.00% (Fig. 4b, h). POD activity decreased significantly by 54.42% and 58.74% at 2 and 4 mmol·L⁻¹Al³⁺ (P < 0.05) but showed no significant difference at higher concentrations (Fig. 4c). CAT activity decreased significantly by 77.64% and 51.55% at 4 and 10 mmol·L^− 1 Al³⁺ (P < 0.05, Fig. 4d). Proline content showed no significant difference from CK (Fig. 4f). MDA and H₂O₂ contents both decreased with rising Al³⁺ (Fig. 4e, i). At 2 and 4 mmol·L^− 1, MDA content was reduced by 52.25% and 59.23%, respectively, compared with CK (P < 0.05), suggesting alleviated lipid peroxidation in roots under aluminum stress.

Fig. 4.

Effects of aluminum stress on APX (a), SOD (b), POD (c), CAT (d), MDA (e), Proline (f), Soluble sugar (g), soluble proteins (h) and H₂O₂ (i) in the roots of white clover seedlings

Correlation analysis between morphology and physiology

Correlation analysis indicated closer associations between root morphological traits and physiological indices than between leaf traits and physiology (Fig. 5). For leaf-related traits, plant height (Ph) showed significant positive correlations with leaf area (LA), leaf width (LW), leaf length (LL), and leaf perimeter (LP). APX, POD, and soluble sugar (Ss) were also positively correlated with these leaf morphological indices. Soluble sugar was positively correlated with APX, POD, and soluble protein (SP) but negatively correlated with MDA. Significant positive correlations were also detected between APX and POD, and between CAT and SP (Fig. 5a). Correlations among leaf traits revealed that under Al³⁺ stress, antioxidant defense (APX, POD) and osmotic adjustment (soluble sugars) synergistically promoted plant growth, while the accumulation of soluble sugar auxiliarily enhanced antioxidant defense capacity and mitigated membrane lipid peroxidation. For roots, plant height was negatively correlated with root APX activity and soluble sugar, but positively correlated with SOD and SP. CAT activity positively correlated with H₂O₂, POD, and SOD. SOD activity was positively correlated with SP and negatively with Ss. Both SOD and MDA showed significant positive correlations with root volume (RV), root surface area (RS), and root tip number (RT), whereas APX activity was negatively correlated with total root length (RL) and RS (Fig. 5b). Correlations in root traits reveal that root morphogenesis (root volume, surface area, etc.) is closely linked to oxidative stress responses (SOD, MDA), while APX activity negatively regulates root elongation and surface area expansion. These patterns highlight roots as the primary organ for Al³⁺ perception and systemic tolerance responses, supporting the focus on root tissues for transcriptomic analysis.

Fig. 5.

Correlation matrices between physiological parameters and leaf morphology (a) and root architectural traits (b). Abbreviations: LA, leaf area; LW, leaf width; LL, leaf length; LP, leaf perimeter; Ph, plant height; RL, total root length; RD, average root diameter; RV, root volume; RS, root surface area; RT, number of root tips; Ss, soluble sugar; APX, ascorbate peroxidase; SOD, superoxide dismutase; POD, peroxidase; CAT, catalase; SP, Soluble protein; Pro, Proline; MDA, malondialdehyde; H₂O₂, hydrogen peroxide

Identification of differentially expressed genes (DEGs)

RNA-seq generated 328 million clean reads, with Q30 values > 91.33% and mapping rates > 66.76%, indicating high sequencing quality (Table S1). In total, 20,732 DEGs were identified across the four Al³⁺ treatment groups relative to CK (Fig. 6a), including 11,815 upregulated and 8,968 downregulated genes (Fig. 6b and c). Among them, 2,123 DEGs (402 upregulated, 1,718 downregulated) were shared by all comparisons, representing core candidate genes potentially involved in aluminum tolerance.

Fig. 6.

Venn diagrams of total (a), upregulated (b), and downregulated (c) DEGs across different Al³⁺ treatments compared with CK

Functional enrichment of DEGs

GO enrichment analysis

GO annotation of DEGs yielded 381,136 terms, including 181,313 (47.57%) in Biological Process (BP), 115,334 (30.26%) in Molecular Function (MF), and 84,489 (22.17%) in Cellular Component (CC) (Table S2). Across all four comparisons (B2vsCK, B4vsCK, B6vsCK, B10vsCK), cell wall organization or biogenesis (GO:0071554, BP), cell wall (GO:0005618, CC) and external encapsulating structure (GO:0030312, CC), and oxidoreductase activity (GO:0016491, MF) were consistently enriched, indicating that cell wall modification and redox regulation are common responses to Al³⁺ stress. Additional MF terms such as transferase activity, transferring glycosyl groups (GO:0016757), structural molecule activity (GO:0005198), and DNA-binding transcription factor activity (GO:0003700) showed treatment-specific enrichment (Fig. 7, Table S3). Carbohydrate metabolic process (GO:0005975, BP) was enriched in B2vsCK and B10vsCK, while cellular protein modification process (GO:0006464, BP) was enriched in B4vsCK and B10vsCK. Ribosome (GO:0005840, CC) appeared in B2vsCK and B6vsCK. Notably, photosynthesis (GO:0015979, BP) and thylakoid (GO:0009579, CC) were enriched only in B2vsCK, suggesting that low-level Al³⁺ stress may trigger compensatory adjustments in photosynthetic capacity, possibly reflecting chloroplast-related responses under specific stress conditions. Sulfur compound metabolic process (GO:0006790), cofactor metabolic process (GO:0051186), and cellular amino acid metabolic process (GO:0006520) in the BP were uniquely enriched in B6vsCK, indicating metabolic reprogramming at moderate stress levels (Fig. 7). B2vsCK and B6vsCK were dominated by upregulated DEGs in many terms, whereas B4vsCK and B10vsCK contained more downregulated DEGs, reflecting a shift from activation to suppression as stress intensified (Table S3).

Fig. 7.

Top 20 enriched GO terms for DEGs in each Al³⁺ treatment group

KEGG pathway enrichment analysis

KEGG analysis showed that 7,930, 8,169, 12,787, and 14,075 DEGs in B2vsCK, B4vsCK, B6vsCK, and B10vsCK were assigned to 364, 358, 384, and 385 pathways, respectively, with 48, 55, 58, and 49 pathways significantly enriched in each comparison (FDR < 0.05). Eleven pathways were significantly enriched at all Al³⁺ concentrations and could be grouped into metabolic regulation, hormone signaling, and defense response. Metabolic regulation included metabolic pathways (ko01100), biosynthesis of secondary metabolites (ko01110), and phenylpropanoid biosynthesis (ko00940). Plant hormone signaling (ko04075) was the only hormone-related pathway consistently enriched. Six defense-related pathways (e.g. ko05164, ko05161, ko04624) were also significantly enriched, reflecting activation of signaling components shared with general stress and plant–pathogen responses under Al³⁺ stress (Fig. 8).

Fig. 8.

KEGG pathway enrichment analysis of DEGs between Al³⁺-treated and CK samples. The top 20 significantly enriched pathways are shown for each treatment group

Identification of candidate genes related to aluminum stress

DEGs associated with flavonoid biosynthesis

Twenty-seven DEGs related to flavonoid biosynthesis were identified, including genes encoding CHS, naringenin 3-dioxygenase, HCT, flavanone 4-reductase, caffeoyl-CoA-related enzymes, chalcone reductase, LAR, and CYP75B1 (Fig. 9; Table S5). Most were downregulated under Al³⁺ stress. Only one CHS gene (Cluster-89142.43896), two HCT genes (Cluster-89142.23766, Cluster-89142.67816), one LAR gene (Cluster-89142.8525), and two CYP75B1 genes (Cluster-90221.0, Cluster-79478.0) were upregulated, suggesting a selective modulation of flavonoid branches in response to aluminum.

Fig. 9.

Heatmap of flavonoid biosynthesis-related genes in white clover under different aluminum treatments (log₂(FPKM + 0.001)). Red denotes high expression levels, while green indicates low expression levels

DEGs associated with photosynthesis

Most Psb and Psa genes were upregulated under Al³⁺ stress. In particular, PsbS (Cluster-102254.0), Psb27 (Cluster-89142.23081), and PsbQ (Cluster-89142.2163) exhibited concentration-dependent expression patterns: PsbS was significantly upregulated at 2 mmol·L^− 1, Psb27 was significantly downregulated at 6 mmol·L^− 1, and PsbQ was significantly upregulated at 10 mmol·L^− 1 (Fig. 10). Most Pet genes were downregulated, whereas ATPF genes were consistently upregulated, especially at 10 mmol·L^− 1, suggesting an important role in maintaining ATP synthesis under severe stress.

Fig. 10.

Heatmap of photosynthesis-related genes in white clover under different aluminum treatments (log₂(FPKM + 0.001)). Red denotes high expression levels, while green indicates low expression levels

DEGs associated with plant hormone signal transduction

Seventy-six DEGs involved in plant hormone signal transduction were identified, including auxin influx carriers, auxin-responsive proteins, TIR1, ARFs, GH3, and SAUR family genes (Fig. 11; Table S5). Auxin-related genes (AUX1, TIR1, IAA) were generally upregulated under Al³⁺ stress, whereas most SAUR genes (15/20) were downregulated. signaling, four AHK genes and one AHP gene were significantly upregulated, while most ARR-A and ARR-B genes were downregulated. Gibberellin-related genes (GID1, DELLA, PIF) tended to be upregulated, indicating activation of GA signaling. By contrast, key genes in ABA and JA signaling pathways were predominantly downregulated, implying suppression of these stress-related hormone pathways. Overall, the data suggest a complex hormone network underpinning white clover adaptation to aluminum.

Fig. 11.

Heatmap of plant hormone signal transduction-related genes in white clover under different aluminum treatments (log₂(FPKM + 0.001)). Red denotes high expression levels, while green indicates low expression levels

qRT–PCR validation of RNA–Seq data

Ten DEGs representing major metabolic and signaling pathways were selected for qRT–PCR validation (Fig. 12; Table S6). LHCA3, ALDO, LHCA4, LHCA1, LHCB2, and PEL were consistently upregulated at all Al³⁺ concentrations (2–10 mmol·L^− 1), whereas GapN, HI40MT, Cah, and CHS were consistently downregulated. The expression patterns obtained by qRT–PCR were highly consistent with RNA-Seq data in both direction and relative magnitude of changes, confirming the reliability of the transcriptome dataset and supporting its use for elucidating aluminum tolerance mechanisms in white clover.

Fig. 12.

qRT–PCR validation of ten selected DEGs under different Al³⁺ treatments. Log₂ fold changes from RNA-Seq and qRT–PCR are shown for comparison

Discussion

By integrating morphological, physiological and transcriptomic evidence, this study demonstrates that white clover responds to aluminium toxicity through a root-centred, yet systemically coordinated, defence strategy. At the whole-plant level, Al³⁺ stress results in tightly coupled inhibition of root and shoot growth. At the tissue level, roots exhibit a stronger activation of antioxidant and osmotic adjustment pathways than leaves, whereas aboveground tissues appear to adopt a conservative strategy that prioritises maintenance of essential photosynthetic function over maximal growth. At the molecular level, these phenotypes are underpinned by (i) a selective reconfiguration of flavonoid biosynthesis, characterised by global repression of core structural genes with stage-specific activation of particular branches; (ii) concentration-dependent modulation of photosystem components and ATP synthase; and (iii) extensive rewiring of hormone signalling, with enhanced auxin and GA modules contrasted by a marked attenuation of ABA and JA pathways. Together, these features delineate a hierarchically organised response network in which roots act as the primary site of Al³⁺ perception and detoxification, while hormone and metabolic pathways jointly shape the whole-plant phenotype (Fig. 13).

Fig. 13.

Proposed model of the root-centred aluminium response network in white clover seedlings

Morphological responses of white clover seedlings to aluminium stress

Roots are the primary organs that perceive and transduce soil-borne stresses, and their architecture largely determines both the spatial extent and efficiency of nutrient and water acquisition. Root length constrains the capacity of roots to explore deeper soil layers, whereas surface area and tip number define the interface available for resource uptake [32]. In this study, increasing Al³⁺ concentrations caused a marked, dose-dependent reduction in total root length, root volume, surface area, average diameter and tip number, accompanied by significant decreases in leaf area, plant height and biomass. The strong positive relationships between Al³⁺ intensity and growth inhibition indicate that root and shoot morphogenesis in white clover are tightly coordinated under aluminium stress rather than being independently regulated.

These phenotypic changes are consistent with earlier observations that Al³⁺ preferentially targets the root apex, where it accumulates, alters cell wall properties and interferes with cell cycle progression, thereby restricting both longitudinal elongation and radial expansion [33]. Similar inhibition of root system development has been reported in crops such as rice (Oryza sativa L.) and wheat (Triticum aestivum L.) exposed to Al³⁺ stress [34, 35]. Taken together, these reports and our findings in white clover support the notion that aluminium toxicity is commonly manifested as restricted root growth, although the precise cellular and molecular mechanisms may differ among species.

The coordinated suppression of below- and aboveground growth in white clover likely reflects feedbacks between root function and shoot carbon economy. Impaired root systems reduce the capacity for nutrient and water uptake, which may in turn constrain leaf expansion and photosynthetic capacity via altered carbohydrate allocation. Conversely, diminished provision of photosynthates from leaves further exacerbates root growth inhibition, forming a self-reinforcing negative loop. In this context, the global up-regulation of photosynthesis-related genes observed in the transcriptome is particularly noteworthy, as it suggests that white clover attempts to buffer the stress-induced limitation of resource acquisition by sustaining the structural and functional integrity of the photosynthetic apparatus. However, establishing a causal link between these transcriptional changes and the observed morphological phenotypes will require direct measurements of photosynthetic performance, such as gas-exchange parameters and chlorophyll fluorescence. Photosynthetic data will be supplemented in subsequent studies to refine the mechanistic analysis.

Response of the peroxidation-scavenging and protection system to aluminium stress

Aluminium stress is widely known to trigger overproduction of reactive oxygen species (ROS), disturbing the balance between ROS generation and scavenging and leading to oxidative damage, including lipid peroxidation and protein oxidation [36]. Hydrogen peroxide (H₂O₂) is a central ROS that, when excessively accumulated, compromises membrane integrity, whereas malondialdehyde (MDA), a terminal product of lipid peroxidation, is a widely used proxy for membrane damage. The enzymatic antioxidant machinery (APX, POD, SOD and CAT), together with osmotic regulators such as proline, soluble sugars and soluble proteins, constitutes a core defence system against oxidative stress [37]. In our experiments, H₂O₂ content in both leaves and roots decreased significantly under Al³⁺ exposure and was markedly lower than in the control at 10 mmol·L^{− 1} (leaves) and 4 mmol·L^{− 1} (roots). The decrease in H₂O₂ levels under high aluminum stress is more likely due to a passive mechanism of metabolic collapse: high Al³⁺ concentrations may cause irreversible damage to cellular ROS production pathways. Leaf MDA content showed a transient increase followed by a decline, whereas root MDA remained significantly lower than that of the control across all treatments. This pattern differs from the continuously rising MDA levels reported in Ophiopogon japonicus under Al³⁺ stress [38] and indicates that white clover is able to effectively limit membrane lipid peroxidation, particularly in roots. The contrasting oxidative profiles of leaves and roots can be linked to the distinct antioxidant enzyme dynamics in these tissues. As the primary interface with Al³⁺, roots exhibited higher POD activity than leaves, and both APX activity and soluble sugar content in roots increased steadily with Al³⁺ concentration. The coordinated enhancement of these components likely enables roots to detoxify ROS more efficiently and to stabilise osmotic potential, thereby limiting the translocation of Al³⁺ to aboveground tissues. This interpretation is consistent with the notion proposed for common bean (Phaseolus vulgaris L.) that tolerant genotypes preferentially activate root antioxidant defences to confine aluminium toxicity at the root level [33]. Our data extend this concept to white clover and support the view that prioritised activation of root antioxidant and osmoprotective pathways is a key feature of its aluminium-response strategy.

Response of flavonoid biosynthesis to aluminium stress

Flavonoids are multifunctional secondary metabolites that contribute to stress adaptation by chelating metal ions, scavenging free radicals and modulating hormone transport [39]. In this study, 27 genes encoding catalytic enzymes within the flavonoid biosynthesis pathway were identified, with a predominance of downregulated transcripts under Al³⁺ stress. This global repression is consistent with the observed reduction in biomass and suggests a reallocation of metabolic resources away from general secondary metabolism. However, the transcriptional landscape is not uniformly suppressive: sustained downregulation of CHS, FH3 and DFR is accompanied by stage-specific upregulation of HCT, LAR and CYP75B1, revealing a selectively remodelled flavonoid network rather than a simple shutdown.

Chalcone synthase (CHS) catalyses the first committed step of flavonoid biosynthesis and is widely regarded as a key control point for flavonoid accumulation and stress tolerance [40, 41]. In our dataset, 15 CHS genes were differentially expressed, representing the largest enzyme family within the flavonoid pathway. Except for Cluster-89142.43896, CHS transcripts were generally downregulated under Al³⁺ stress, which is expected to reduce the flux into downstream flavonoid branches, including the synthesis of antioxidant flavonols. A diminished pool of flavonols could weaken Al³⁺ chelation in root tissues and reduce ROS-scavenging capacity, thereby contributing to root morphological damage and plant dwarfism. Thus, the overall suppression of CHS expression likely reflects a trade-off between growth, basal metabolism and specialised defence in white clover under prolonged stress.

Anthocyanins, another major class of flavonoids, play important roles in ROS scavenging and photoprotection. The DFR gene, encoding a key enzyme in anthocyanin biosynthesis, was consistently downregulated throughout Al³⁺ exposure in white clover, in contrast to the DFR induction reported for Medicago sativa L. under salt stress [42]. This discrepancy underscores the context- and stress-specific regulation of individual flavonoid branches, implying that anthocyanin-mediated defence is not the primary strategy mobilised under aluminium toxicity in white clover.

By contrast, CYP75B1, which encodes a flavonoid 3′-hydroxylase, showed no expression at early stages but was upregulated specifically at 6 and 10 mmol·L^{− 1} Al³⁺. Although the role of CYP75B1 in aluminium responses has not been directly characterised, orthologous genes in citrus promote the accumulation of antioxidant flavonoids, and their overexpression in Arabidopsis enhances drought tolerance [43]. Based on these precedents, we propose that late-stage induction of CYP75B1 may represent a compensatory mechanism in white clover, selectively boosting the production of particular antioxidant flavonoids when stress intensifies and general flavonoid biosynthesis is compromised. This working model remains to be tested by functional analyses targeting CYP75B1 and other branch-specific enzymes in the flavonoid pathway.

Response of key photosynthesis genes to aluminium stress

Photosynthesis is central to plant energy metabolism, and its light reactions critically depend on the structural integrity and functional coordination of photosystem II (PSII) and photosystem I (PSI). The Psb gene family modulates PSII assembly, repair and electron transport, whereas Psa components are essential for PSI stability and function [44]. In our transcriptome, most Psb and Psa genes were upregulated under Al³⁺ stress, suggesting that white clover attempts to counteract damage to the photosynthetic machinery by reinforcing photosystem structure and maintaining electron transport capacity.

Within PSII, several components displayed clear concentration-dependent responses. Psb27, a factor involved in PSII assembly and repair, was generally upregulated but became transcriptionally silenced at 6 mmol·L^{− 1} Al³⁺. In contrast, PsbQ, which encodes an oxygen-evolving enhancer (OEE) subunit that stabilises the oxygen-evolving complex (OEC) and optimises the microenvironment for water splitting [45], was markedly induced at this higher Al³⁺ concentration. We interpret this as a putative compensatory relationship between Psb27 and PsbQ whereby, when Psb27-mediated assembly and repair are impaired or insufficient under severe stress, strong induction of PsbQ may enhance OEC stability and oxygen-evolving activity in the remaining functional PSII centres, thereby preventing collapse of upstream electron flow. Although this model remains speculative, it highlights a potential layer of plasticity in PSII maintenance under metal toxicity in white clover.

PsbS, a core component of non-photochemical quenching (NPQ) and thus a key regulator of photoprotection, was expressed only at 2 mmol·L^{− 1} Al³⁺. This pattern suggests that at low stress levels, white clover primarily mitigates photooxidative damage by modulating stomatal conductance and NPQ, consistent with a PsbS-mediated ROS alleviation mechanism [46]. As Al³⁺ stress intensifies, the reliance on PsbS may decrease as the plant shifts towards structural reinforcement and reorganisation of photosystems.

Despite a general downregulation of PetF, encoding ferredoxin, the overall photosynthetic electron transport chain appears to remain functional. This could be achieved through compensatory expression of other ferredoxin isoforms or through reconfiguration of electron partitioning routes, as suggested by species-specific roles of PetF observed in Chlamydomonas reinhardtii under heat stress [47]. Particularly noteworthy is the late-stage upregulation of ATPF, encoding an ATP synthase subunit. Given the central role of ATP synthase in photophosphorylation and oxidative phosphorylation [48], enhanced ATPF expression under prolonged Al³⁺ exposure likely helps secure the energy supply required for cellular repair, active transport and Al³⁺ chelation, thereby sustaining a functional “light reaction–dark reaction–energy regeneration” loop even as growth is curtailed. These findings are solely based on transcriptomic data of key photosynthesis-related genes, and further comprehensive analysis with measurements of leaf gas exchange, chlorophyll fluorescence (Fv/Fm) and other photosynthetic physiological indicators is required.

Response of plant hormone signal transduction to aluminium stress

Plant hormones orchestrate growth, development and stress adaptation through intricate signalling networks involving synergistic and antagonistic interactions, as well as extensive pathway crosstalk [49]. Auxin, through early response genes such as GH3, AUX/IAA and SAUR, regulates cell elongation and cell wall remodelling during stress adaptation [50]. We observed general upregulation of auxin-related genes but widespread downregulation of SAUR family members. SAURs are rapidly responding auxin effectors that promote cell elongation by modulating cell wall acidification and ion transport [51]. Under Al³⁺ stress, repression of elongation-promoting SAURs is therefore likely to attenuate auxin-driven growth responses. At the same time, the overall activation of the auxin pathway may enhance the expression of SAUR subtypes or other negative regulators that actively restrict elongation, resulting in dual-layer control combining functional inactivation with active repression. A similar auxin–SAUR regulatory pattern has been reported in rice under aluminium stress [52]. Functionally, this mechanism could limit root tip penetration into deeply aluminium-saturated layers and shorten internodes, thereby reducing Al³⁺ uptake while manifesting as dwarfism—a potentially adaptive but costly trade-off.

Cytokinin signalling, mediated by AHK, AHP and type-A/B ARR transcription factors, further highlights the multilayered complexity of hormonal regulation under aluminium stress. In our dataset, four AHK receptors were consistently upregulated across Al³⁺ treatments, suggesting that white clover attempts to enhance cytokinin perception and activate the canonical “AHK→AHP→ARR” two-component pathway to maintain cell division and delay senescence. However, the early downregulation of AHP genes at 2 mmol·L^{− 1} Al³⁺ appears to interrupt this signalling cascade, limiting the activation of ARR transcription factors (ARR-A and ARR-B). As a consequence, downstream cell-division-related genes remain suppressed, and cytokinin-mediated senescence delay cannot be effectively achieved. This juxtaposition of enhanced receptor expression with impaired downstream transmission provides a mechanistic explanation for the reduced biomass and shortened lifespan of white clover under Al³⁺ stress, and is consistent with the auxin–cytokinin crosstalk implicated in Al-induced root-growth inhibition in Arabidopsis [53].

Gibberellins (GAs), key regulators of stem elongation and overall growth [54], also showed a coordinated transcriptional response. GA-related genes, including GID1, DELLA and PIF, were generally upregulated. Binding of GA to its receptor GID1 targets DELLA repressors for degradation, thereby releasing PIF transcription factors to activate cell-elongation genes. The concerted induction of these components suggests that, even under aluminium stress, white clover attempts to maintain a minimal growth programme, possibly to ensure renewal of functional tissues and to preserve basic developmental progression.

Abscisic acid (ABA) and jasmonic acid (JA) are core regulatory hormones in plant stress response, with the former primarily integrating abiotic stress-tolerance signals and the latter coordinating defence against both biotic and abiotic stresses [55, 56]. In this study, the coordinated downregulation of core signal genes in these two pathways directly reveals a key deficiency in the Al³⁺ stress defense network of white clover. In the ABA pathway, key signalling components PYL, PP2C, SnRK2 and ABF were downregulated, which is expected to impair ABA-mediated stomatal closure and stress acclimation, potentially exacerbating water loss and oxidative damage. This pattern contrasts with findings in rice, where ZFP36 positively regulates ABA-triggered antioxidant defences to alleviate Al toxicity [57]. The discrepancy may arise because ZFP36 functions as a positive modulator within otherwise an intact ABA network, whereas in white clover the concurrent downregulation of multiple core ABA components effectively disconnects the pathway, limiting its capacity to mount a protective response.

Abnormal regulation of the JA signaling pathway further amplifies this defense deficiency. General downregulation of JA-related genes (JAR1, COI, JAZ, MYC2) blocks the activation of defense genes by the JA signaling pathway, reducing plant resistance to Al³⁺ stress and secondary biotic stresses. This is highly consistent with the study by ZhongBao Yanget al. on Arabidopsis thalianathat exogenous application of JA enhances the upregulation of MYC2 expression, directly affecting Al-induced root growth inhibition, in which exogenous application of JA enhanced MYC2 expression and directly affected Al-induced root growth inhibition [58]. Among the JA-responsive genes identified here, those shared with plant–pathogen and general stress-response pathways were activated by Al³⁺ stress rather than by pathogens per se, which is broadly similar to the patterns reported by Ruipu Wang et al. in two Vitis vinifera varieties “Hö29” and “August Weiß” [59].

It is worth noting that plant hormone-related genes and genes related to various physiological indicators may participate in the regulation of plant morphological, metabolic, physiological, and biochemical changes under Al³⁺ stress through complex and coordinated interactions. It is worth noting that plant hormone-related genes and genes associated with diverse physiological indicators may jointly participate in the regulation of morphological, metabolic, physiological and biochemical changes in white clover under Al³⁺ stress through complex and tightly coordinated interactions. Subsequent studies will combine hormone-targeted assays with transgenic functional validation to thoroughly elucidate the regulatory mechanisms of hormone pathways.

Conclusions

Trifolium repens responds to aluminum stress through a root-centered multi-level response mechanism involving morphological adjustment, physiological adaptation, and molecular reconstruction. It forms a root-dominated response pattern characterized by “root-prioritized defense, coordinated photosynthetic function maintenance, and hormone network regulation in roots”, which ensures the survival of plants under stress, but inevitably at the cost of restricted growth. This study provides new insights into the root-driven aluminum tolerance mechanism of leguminous plants and lays a theoretical foundation for the breeding of aluminum-tolerant forage varieties.

Authors’ contributions

L.Z. and P.W. supervised the experiments. X.Y., L.H. and S.L. performed the experi-ments. X.Y. analyzed the data. X.Y. wrote the manuscript. All the authors readand approved the fnal manuscript.

Funding

This work was supported by Key Project of Guizhou Provincial Basic Research Program (Qiankehejichu ZD[2025]048) and Guizhou Province High-Level Innovative Talent Project (Qiankehe Platform Talent-GCC[2022]022).

Data availability

All data supporting the findings were contained in the manuscript and its supplementary files and the RNA-seq raw data. And all the RNA-seq raw data were uploaded in the SRA of NCBI (https://www.ncbi.nlm.nih.gov/sra/, the accession number is PRJNA1382770).

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.