G. pentaphyllum was obtained from the Guangxi Botanical Garden of Medicinal Plants (Nanning, Guangxi, China). The seeds were germinated on moist perlite, after which seedlings of 3~5 leaf stage were selected and cultivated in hydroponic boxes with the Japanese Yamzaki formula solution (pH 6.0) (Li et al., 2023b). All plants were cultivated at a constant temperature of 25 ± 2°C, with a photoperiod of 14 h/10 h (day/night) and light intensity of 100 μmol m^-2 s^-1). After 7 days of culture, the seedlings were subjected to a hydroponic medium containing different concentrations of CdCl₂ (0, 25, 50, 100, 150 and 200 μM) for 7 days, and seedlings grown in Cd-free hydroponic medium were used as controls (Li et al., 2022b). The G. pentaphyllum plants were then collected and assessed for height and root length. Fresh leaf samples were acquired for the assessment of physiological indexes as well as transcriptomics and metabolomics analysis. All samples were ground in liquid nitrogen by using a JXFSTPRP-64L grinding instrument (Shanghai, China) and stored at -80°C.

The enzyme activities of superoxide dismutase (SOD), peroxidase (POD), catalase (CAT), L-ascorbate peroxidase (APX), as well as the contents of malondialdehyde (MDA) and proline were assessed and analyzed utilizing the methods as previously described (Pan et al., 2023). Dithizone (DTZ) staining was employed to evaluate Cd localization at the cellular levels as previously described (He et al., 2013). Periodic acid-Schiff (PAS) staining was employed using the Periodic Acid Schiff (PAS) Stain Kit (Beijing Solarbio Technology) by following the manufacturer’s instructions. A Zeiss Axioscope 5 microscopy (Zeiss, Germany) was utilized to capture the images. Each experiment utilized three replicates of each treatment regimen.

Fresh leaf samples of Cd0 (CK), Cd25 (LC) and Cd100 (HC) were used in transcriptomics profiling analysis. The total RNA of nine samples from the three conditions was extracted using a Trizol reagent kit (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. The purity, concentration and integrity of the RNA samples were assessed. The cDNA libraries were prepared using the NEBNext Ultra RNA Library Prep Kit for Illumina (NEB #7530, New England Biolabs, Ipswich, MA, USA). Finally, the cDNA library was sequenced using the Illumina Novaseq 6000 platform in the paired-end mode by Gene Denovo Biotechnology Co., Ltd (Guangzhou, China).

Following the sequencing process, the raw sequence data underwent filtration in order to eliminate low-quality reads (reads with a mass value below 10 and constituting more than 50% of the total bases) by FASTP v0.18.0. The short reads alignment tool Bowtie2 v2.2.8 was used for mapping reads to ribosome RNA (rRNA) database, and the rRNA mapped reads were removed. Then, the clean reads were mapped to the G. pentaphyllum reference genome (PRJNA720501, https://www.ncbi.nlm.nih.gov/search/all/?term=PRJNA720501) with HISAT2 v2.2.4. The fragments per kilobase of transcript per million mapped reads (FPKM) for all transcripts were quantified using Trapnell’s method (Trapnell et al., 2010). The differential expressed genes (DEGs) were calculated by using the NOISeq method, employing a |log2 (fold change) |>2 and a false discovery rate (FDR) <0.05 (Tarazona et al., 2011). Gene ontology (GO) enrichment analysis was conducted with the GO database (http://www.geneontology.org/), while KEGG pathway enrichment analysis was conducted using the KEGG database (http://www.kegg.jp/).

The fresh leaf samples of Cd0 (CK), Cd25 (LC), and Cd100 (HC) were incubated overnight at 4 °C, using 1.0 mL 70% aqueous methanol. The extracts were absorbed and filtered through an SPE cartridge and microporous membrane (0.22 μm pore size) before LC-MS analysis. The metabolomic analysis was conducted utilizing an ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS) system. The chromatographic separation was achieved using a Waters C18 column, employing mobile phase A (0.04% acetic acid in water) and mobile phase B (0.05% acetic acid in acetonitrile) and operating at 40°C. Solvent gradient changes were linear for all the steps and these were: 95:5 Phase A/Phase B at 0 min, 5:95 Phase A/Phase B at 11.0 min, 5:95 Phase A/Phase B at 12.0 min, 95:5 Phase A/Phase B at 12.1 min and 95:5 Phase A/Phase B at 15.0 min. The flow rate was 0.4 mL/min, and the injection volume was 2 μL. An electrospray ionization (ESI) mode was employed to acquire the high-resolution mass spectra (HRMS), while operating in the positive ion mode. Further data processing was performed using Analyst 1.6.1 software. Metabolite information was identified by conducting searches in both internal and public databases (MassBank, KNApSAcK, HMDB, MoTo DB, and METLIN). PCA and OPLS-DA were utilized to identify significantly different metabolite levels (p-value < 0.05). Differentially altered metabolites (DAMs) were designated as a log2 fold change (FC) ≥ 2 and FC ≤ 0.5, along with variable importance in projection (VIP) scores > 1. The identified metabolites underwent metabolic pathways analysis using the KEGG database and MetaboAnalyst 4.0 (http://www.metaboanalyst.ca).

Correlation network analysis was utilized to generate and analyze the association characteristics between metabolites and genes. The correlation characteristics between genes and metabolites were obtained using the KEGG pathway shared with genes and metabolites. Bidirectional orthogonal projections to latent structures (O2PLS) were employed to analyze both gene expression and metabolite abundance, and the best models derived from this analysis were derived from integration analysis. Gene-metabolite pairs were ranked by absolute correlation coefficients, which were calculated for gene expression and metabolite abundance. The resulting networks were visualized using the Cytoscape version 3.7.2 software package.

Fourteen DEGs were selected for qRT-PCR analysis and actin was used as an internal reference gene. The primers used were based on the transcript sequences using Primer 5.0 (Table S6). The qRT-PCR was performed using the LightCycler 96 instrument (Roche, Basel, Switzerland). The 2^–ΔΔCt method was employed to calculate the relative expression levels of the genes (Livak and Schmittgen, 2001). When the transcriptomic and qRT-PCR data were combined, the candidate genes appeared to show similar expression patterns (Figure S7).

Statistical analyses were conducted by using SPSS v 26.0 software. All values were expressed as the means ± standard deviations. All the data were tested using one-way ANOVA with Duncan’s multiple-range test. The p < 0.05 was considered to be statistically significant. The graphs were drawn using GraphPad Prism 8.

With increasing the Cd concentrations, the average root lengths of the Cd treatment groups were significantly reduced by 10.33%~40.96% compared to that of the control group, except for the Cd5 group (Figure S1A). In addition, with the increase of Cd concentrations, the height of plants also exhibited a notable reduction (Figure S1B). The growth of seedlings displayed visible poisoning symptoms and withered in the Cd200 (200 μM) group. The seedlings treated with Cd25 and Cd100 showed the minimum reduction and significant inhibitory in root length and height, respectively (Figure S1). These also exhibited different growth behavior (Figure 1A), and therefore these two concentrations were chosen for further analysis.

The activities of POD and APX, as well as the concentrations of MDA and proline content, showed a gradual trend with increased Cd stress (Figure 1B). In contrast, there was a notable and consistent decline in SOD activity when compared to the Cd0 group. In addition, the activity CAT was reduced by 24.53% in the Cd 50 group and increased by 22.64% compared with that of the Cd0 group (Figure 1B). Cd-dithizone precipitates were observed in the roots, stems and leaves (Figure 1C). Obvious Cd-dithizone precipitates were also observed in the root tips. Cross-section analysis of the stems showed Cd was mainly located in parenchymatous cells of the cortex. In the leaves, Cd-dithizone precipitates were observed mainly in parenchymatous cells as well as the phloems and xylems of the main veins. However, in the Cd100 group, bright pink polysaccharide staining was seen in the stems and leaves when they were compared with the Cd0 group (Figure 1D). All these results indicated that the G. pentaphyllum seedlings could efficiently activate the oxidative enzymes, proline content, Cd transport, and polysaccharides when response to Cd stress.

On average, a total of 42160383 clean reads were obtained from each of the samples (Table S1). The GC content, Q20, and Q30 values of all clear reads were above 45.55%, 97.55%, and 93.16%, respectively, confirming the high reliability of the sequencing outcomes. The clean reads from all samples had a mapping rate of 85.16~86.67% when compared to the reference genome sequence, resulting in the discovery of 25,656 genes. The Pearson correlation coefficient analysis provided evidence of the biological consistency (Figure 2A). In addition, the PCA analysis demonstrated discernible differences in the expression of gene clusters between the control group (Cd0, CK) and the different Cd treatments (Cd25, LC and Cd100, HC) were distinguishable (Figure 2B). 2091 (1018 up- and 1073 down-regulated), 3985 (2010 up- and 1975 down-regulated), and 1811 (1163 up- and 648 down-regulated) DEGs were identified by comparing CK vs LC, CK vs HC, and LC vs HC, respectively (Figure 2C). In all comparison groups, a comprehensive set of 4921 DEGs were screened (Figure 2D and Table S2). Among these, 225 genes were screened that were commonly expressed across all groups (Figure 2D), manifesting that G. pentaphyllum activated the expression levels of these genes to cope with varying levels of Cd stress.

The GO enrichment analysis confirmed the impact of Cd stress on specific biological functions with the top 20 GO enriched terms (Figure S2 and Table S3). The oxidoreductase activity (GO:0016491), tetrapyrrole binding (GO:0046906) and photosystem (GO:0009521) with enriched GO terms were found mainly in CK vs LC. The protein kinase activity (GO:0004672), protein phosphorylation (GO:0006468) and photosystem (GO:0009521) with enriched GO terms were found mainly in CK vs HC. The DEGs were notably enriched in protein kinase activity (GO:0004672), protein phosphorylation (GO:0006468), and response to chitin (GO:0010200) with LC vs HC comparison. GO analysis indicated that GO terms such as oxidoreductase activity, protein kinase activity and protein phosphorylation could effectively enhance the Cd tolerance of G. pentaphyllum.

The KEGG pathway enrichment analysis confirmed the impact of Cd stress on specific biological pathways. A higher number of DEGs were significantly assigned to enrich the metabolic pathways and biosynthesis of secondary metabolites in CK vs LC, CK vs HC, and LC vs HC pairwise groups (Figure S3). In the CK vs LC comparison, several pathways, including phenylpropanoid biosynthesis, plant hormone signal transduction, photosynthesis, starch and sucrose metabolism, alpha-linolenic acid metabolism and flavonoid biosynthesis, showed significant enrichment. The significant enrichment pathways in the comparison between CK and HC included phenylpropanoid biosynthesis, photosynthesis, porphyrin metabolism, starch and sucrose metabolism, MAPK signaling pathway and biosynthesis of various alkaloids. The phenylpropanoid biosynthesis, plant-pathogen interaction, biosynthesis of various alkaloids, MAPK signaling pathway, phenylalanine metabolism and flavonoid biosynthesis were significantly enriched pathways in LC vs HC. According to the KEGG analysis, the response of G. pentaphyllum to Cd resulted in the regulation of various pathways, including the biosynthesis of other secondary metabolites, carbohydrate metabolism, amino acid metabolism, lipid metabolism, and signal transduction.

In addition, the detailed functions of the common 225 genes from the three comparison groups were investigated. The xyloglucosyl transferase activity, hydrolase activity, extracellular region, cell wall, oxidoreductase activity and phenylpropanoid metabolic process were mainly enriched in GO terms (Figure 3A). Additionally, these DEGs were categorized based on their involvement in KEGG pathways analysis. The photosynthesis, phenylalanine metabolism, plant hormone signal transduction, indole alkaloid biosynthesis, betalain biosynthesis and MAPK signaling pathway were the most enriched pathways (Figure 3B). These findings indicated that Cd stress had a substantial impact on multiple physiological processes, notably affecting amino acid metabolism, carbon metabolism, signal transduction, and secondary metabolic systems.

To further elucidate the role of Cd response for 225 genes, the ABC transporters, cell wall, phenylpropanoid biosynthesis, glutathione metabolism, photosynthesis and TFs were selected for analysis (Table 1). There were two genes related to ABC transporters that appeared to be important. ABCG23 (ABC transporter G family member 23) was up-regulated in CK vs HC and LC vs HC, and ABCG8 (ABC transporter family member 8) was down-regulated in all three comparison groups. Three down-regulated xyloglucan endotransglucosylase/hydrolase (XTH31, XTH7 and XTH9) and one down-regulated expansin (EXPA10) involved in the cell wall in the three comparison groups, and two up-regulated XTH23 and XTH2 in CK vs HC and LC vs HC, respectively. Notably, two genes associated with glutathione S-transferase (GST) and related to glutathione metabolism were up-regulated, while six chlorophyll a-b binding protein genes involved in photosynthesis were down-regulated. As many as 12 TFs were also found, underscoring their importance in mediating the Cd response and transportation processes. NAC, ERF4, MYB39, CPRF1 and HSFC1 genes were up-regulated, and bHLH70, COL16, and RL1 genes were down-regulated in three comparison groups. All these G. pentaphyllum genes were likely linked to the ability of this species to withstand the response to Cd stress.

To explore the response of G. pentaphyllum under different Cd stresses, metabolomics analysis was performed using data obtained from UPLC-MS/MS. The control (Cd0, CK) and Cd-treated groups (Cd25, LC and Cd100, HC) were distinguished by PCA and OPLS-DA score plots (Figure S4). In total, 643 metabolites were screened (Figure 4A and Table S4), of which 126 DAMs were identified in all groups (Figure 4B and Table S5). In the CK vs LC, CK vs HC, and LC vs HC comparisons, there were 84, 90 and 81 DAMs, respectively (Figure 4B). The number of DAMs was notably higher in CK vs HC than in other combinations, indicating that there was a specific impact on the stimulation of certain metabolites by a high concentration of Cd. By performing KEGG enrichment analysis (Figure S5), it was observed that pathways involved with pyruvate metabolism, citrate cycle (TCA cycle), glyoxylate and dicarboxylate metabolism were significantly enriched due to DAMs in pairwise CK vs LC. Arginine and proline metabolism, and the TCA cycle were significantly enriched in pairwise CK vs HC. Flavone and flavonol biosynthesis, valine, leucine and isoleucine biosynthesis were significantly enriched in pairwise LC vs HC. The results strongly implied that these metabolites were essential for G. pentaphyllum response to Cd stress.

The Venn diagram analysis showed that 29 DAMs were found to be affected by different Cd concentration treatments (Figure 4C and Table S4). These metabolites showed significant induction with increasing Cd intensity in CK vs LC, CK vs HC, and LC vs HC, including 10 lipids, 7 flavonoids, 3 nucleotides and derivatives, 3 organic acids, 2 amino acids and derivatives, 2 phenolic acids and 2 alkaloids (Figure 4D). Based on the log₂ fold change values, it was evident that 15 DAMs were up-regulated in CK vs LC and CK vs HC, and these included L-isoleucine, L-arginine, 5-aminoimidazole ribonucleotide, and kaempferol-3-O-robinobioside (biorobin). The down-regulated DAMs in CK vs HC and LC vs HC were protocatechuic acid-4-O-glucoside, adenine, guanosine and kaempferol-3-O-(6’’-malonyl) galactoside. Therefore, these DAMs can serve as potential candidate markers for the response of G. pentaphyllum to Cd stress and act as stimuli for distinguishing exposure to various Cd concentrations.

The histogram depicted the degree of KEGG pathway enrichment when considering both DEGs and DAMs simultaneously (Figure 5). 205 DEGs and 87 DAMs were enriched to 38 metabolic pathways in CK vs LC, 532 DEGs, and 128 DAMs were enriched to 50 metabolic pathways in CK vs HC, and 203 DEGs, and 82 DAMs were enriched to 42 metabolic pathways in LC vs HC. Interestingly, it was intriguing to observe that DEGs and DAMs in CK vs LC were simultaneously significantly enriched in the phenylpropanoid biosynthesis pathway, which was stimulated by low concentrations of Cd. The phenylpropanoid biosynthesis pathway, starch and sucrose metabolism, alpha-linolenic acid metabolism and ABC transporters were found significantly enriched for DEGs and DAMs simultaneously in CK vs HC. The results suggested that even under high Cd stress, these metabolic pathways were still active and they were stimulated simultaneously to the ABC transporter. The phenylpropanoid biosynthesis pathway, phenylalanine metabolism, beta-alanine metabolism and starch and sucrose metabolism were significantly enriched by DEGs and DAMs simultaneously in LC vs HC, and this showed that these pathways were promoted simultaneously by Cd stress. The phenylpropanoid biosynthesis pathway, starch and sucrose metabolism, alpha-linolenic acid metabolism and ABC transporter were the vital pathways for Cd response of G. pentaphyllum.

Cd, which is classified as a non-essential heavy metal, can inhibitplant growth which is demonstrated by both physiological and biochemical indicators (Rizwan et al., 2017; Gul et al., 2021). In this study, Cd treatment showed a dwarf phenotype on G. pentaphyllum, underscoring the detrimental effects of Cd, and was consistent with previous studies (Li et al., 2022b). Prior research had demonstrated that treatment with Cd in sorghum and wheat seedlings had toxic effects and this was concentration-dependent (Jiao et al., 2023; Zhang et al., 2023). Cd treatment with low concentrations (1.6 mg/L^-1) manifested the best stimulatory growth when compared to high concentrations (6.5 mg/L^-1) in young peppermint plants (Wang et al., 2023). In this study, G. pentaphyllum root growth responds to Cd with stimulation at low doses and inhibition at high doses, as evidenced by the more pronounced growth inhibition in seedlings treated with higher Cd concentrations compared to those treated with lower concentrations (Figures 1A, S1). Previous studies have verified that the Cd accumulation of leaves is higher than that in roots of Calotropis gigantean in Cd-polluted environments (Yang et al., 2022). Our results show that Cd mainly accumulates in the root tips, as well as in different structural tissues such as the stems and leaves (Figure 1). Biomass of G. pentaphyllum in the aboveground parts is far greater than that of underground portions, and Cd is suggested to be mainly accumulated in the former.

To cope with the cytotoxicity of Cd, plants have developed multiple tolerance mechanisms. The plant’s antioxidant defense system, which includes a variety of antioxidant enzymes (POD, SOD, APX and CAT), as well as non-enzymatic antioxidants (proline, soluble sugars and proteins), plays a critical role in scavenging ROS stress generated due to Cd (Rizwan et al., 2016; Rizwan et al., 2017; Li et al., 2023c). Cellulosic polysaccharides, such as pectin, cellulose and hemicellulose, have significant functions in the binding and accumulation of Cd under Cd stress conditions (Wan and Zhang, 2012; Meyer et al., 2015). The activities of POD, CAT, APX, and proline content exhibited a consistent increase with rising concentrations of Cd (Figure 1B). Furthermore, the main components of the cell wall, which are the polysaccharides, were seen to accumulate under Cd stress (Figure 1D). This suggests that activation of antioxidant system and e stimulation of metal binding-related cell wall polysaccharides are the important physiological characteristics of G. pentaphyllum in response to Cd exposure.

Based on the transcriptomics data, 4921 DEGs were detected across the treatment groups (Figure 2 and Table S2). Analysis of DEGs found that they were associated with specific biological pathways, oxidoreductase activity, protein kinase activity and protein phosphorylation, indicating that the G. pentaphyllum seedlings could efficiently activate its defense system and induce the expression of stress-related genes to counter act the Cd toxicity it encountered (Figure S2 and Table S3). In prior studies, it was demonstrated that Cd stress exerted a substantial impact on the expression levels of DEGs which were related to amino acid, carbohydrate, and nucleotide biosynthetic pathways in sorghum and S. nigrum (Wang et al., 2022; Jiao et al., 2023). These DEGs exhibited marked induction in the biosynthesis of other secondary metabolites, carbohydrate, amino acid and lipid metabolism, as well as signal transduction pathway, which were stimulated in response of G. pentaphyllum to Cd stress (Figure S3). Earlier studies had demonstrated that the XTH gene family was implicated in the production of hemicellulose in the primary cell walls, potentially serving as an essential mechanism for alleviating Cd toxicity in tobacco and A. thaliana (Zhu et al., 2013; Han et al., 2014). Studies have also documented that overexpression of PtoEXPA12 and TaEXPA2 in tobacco plants increased the resistance to Cd accumulation (Ren et al., 2018; Zhang et al., 2018). XTHs and EXPAs were significantly expressed after Cd treatment to respond to Cd stress (Table 1). In addition, another study showed that XTHs and EXPAs might be candidate genes to regulate cell wall Cd fixation by polysaccharides (Xiao et al., 2020). A greater amount of polysaccharide accumulation was observed in the Cd treatment group (Figure 1D). Therefore, we hypothesized that XTHs and EXPAs could potentially play a role in Cd tolerance by fixation of Cd with cell wall polysaccharides. Nevertheless, further investigations are essential to verify this hypothesis. GST is a crucial enzyme involved in enzymatic detoxification, and it plays a vital role in the ROS-scavenging system by catalyzing the interaction of GST with hydrogen peroxide (Strange et al., 2001). Here we also found that two GST genes were significantly up-regulated after Cd treatment (Table 1), suggesting that high expression of GST genes in the G. pentaphyllum may be involved in Cd detoxification, thereby promoting ROS-scavenging. The LHC family gene LHCB3 was found to be specifically expressed after Pb treatment, and it was able to improve the photosynthesis efficiency of Trifolium pratense grown under Pb stress (Meng et al., 2022). Notably, the six LHC family genes were down-regulated under Cd stress, and more specifically so at high Cd concentrations (Table 1), when the leaves showed an apparent chlorosis phenotype (Figure 1A), which may regulate the photosynthesis of G. pentaphyllum.

From the metabolomics data, 126 DAMs were detected across the treatment groups (Figure 4 and Table S5). These DAMs were involved in pyruvate metabolism, TCA cycle, glyoxylate and dicarboxylate metabolism, arginine and proline metabolism, flavone and flavonol biosynthesis pathway, valine, leucine and isoleucine biosynthesis pathway (Figure S5), indicating that the G. pentaphyllum seedlings could efficiently activate these metabolites as a response mechanism to enhancing the tolerance of Cd. The TCA cycle can enhance energy supplies and elevate the Cd stress-related protein levels by increasing the intracellular carbohydrate content, which ultimately ameliorates Cd toxicity in the roots of S.nigrum (Wang et al., 2022). The TCA cycle-related metabolites, citric acid, such as α-ketoglutaric acid, L-(-)-malic acid, succinic acid and fumaric acid, were significantly accumulated after Cd treatment (Tables S3, S4 and Figure 4D), implying that they may play a vital role in maintaining energy support balance during Cd stress. Flavonoids, as secondary metabolites, are mainly used to reduce Cd poisoning through chelation and passivation in plants and these can help to confer Cd resistance (Li et al., 2015; Zhu et al., 2020; Jiao et al., 2023). Flavonoids, including luteolin-7-O-(6’’-malonyl) glucoside-5-O-rhamnoside, quercetin-3-O-glucoside (isoquercitrin), quercetin-3-O-galactoside (hyperin), kaempferol-3-O-robinobioside (biorobin) and quercetin-7-O-glucoside were found to be accumulated during Cd treatment (Figure 4D), and there must be is involved in the Cd response of G. pentaphyllum. Amino acids can effectively chelate metal ions and reduce the toxic effects of Cd on the rice plant, Noccaea caerulescens and N. praecox (Zemanová et al., 2017; Xue et al., 2022; Kocaman, 2023). The findings of this study suggest that Cd treatment resulted in significant elevations in the levels of L-isoleucine and L-arginine (Figure 4D), implying that Cd exposure could stimulate the biosynthesis of these amino acids in G. pentaphyllum, potentially enhancing stress resistance.

Transcriptomics and metabonomics studies have demonstrated that alpha-linolenic acid metabolism plays an important role in tolerance and detoxification to Cd stresses of cotton, rice and fescue, and have identified a significant number of its metabolites and genes that are either induced or inhibited under Cd treatments (Zhu et al., 2018; Zeng et al., 2021; Li et al., 2022a). In this study, a substantial number of DEGs and DAMs were identified that were associated with alpha-linolenic acid metabolism, starch and sucrose metabolism, phenylpropanoid biosynthesis, and ABC transporters as crucial pathways in Cd stress responses in G. pentaphyllum seedlings (Figures 5–7). Moreover, alpha-linolenic acid and 12-OPDA metabolites, as well as ACX, LOX, OPR, and PLA2G genes were identified as the key metabolites and genes in this study (Figure 6A), which were related to the production of JA (Ghorbel et al., 2021). JA alleviates Cd toxicity via suppression of Cd uptake and either translocation or reduction of the accumulation of ROS in plants (Lei et al., 2020; Ahmad et al., 2021; Kaushik et al., 2022; Li et al., 2022c). These results suggest that alpha-linolenic acid metabolism was significantly affected by Cd stress, which may have contributed to the impacted key metabolite accumulation and gene expression, thereby promoting JA biosynthesis to alleviate Cd toxicity. Glucose can also alleviate Cd toxicity by enhancing Cd fixation in the cell walls of Arabidopsis roots and sequestering Cd into the vacuoles (Shi et al., 2015). Increased levels of chlorogenic acid in Kandelia obovata can reduce Cd and Zn poisoning through its hydroxyl radical scavenging capacity (Chen et al., 2020). It can be speculated that the upregulation of genes involved in the phenylpropanoid biosynthesis pathway together with the increase in chlorogenic acid metabolites is closely related to G. pentaphyllum response to Cd stress. The ABC transporter-associated TaABCC1 and OsABCC9 genes have been shown to promote Cd compartmentalization in the vacuoles which can enhance Cd tolerance in Cd-tolerant wheat and rice (Yang et al., 2021; Zhang et al., 2023). Our results provide evidence that these key metabolic pathways were significantly responsible for Cd stress, which may have contributed to the impacted key metabolite accumulation and gene expression observed in these conditions.

As crucial transcriptional regulators, WRKY, MYB, and bZIP must play a significant role in the Cd response of plants (Yuan et al., 2018; Li et al., 2023a). Overexpression of ThDRE1A, ThMYC1, and ThFEZ in T. hispida decreased the ROS content after Cd treatment by changes in the enzymatic activities of SOD, CAT, and POD (Xie et al., 2023). Overexpression of PyWRKY75 promoted the absorption and accumulation of Cd, and activated the antioxidant enzymes under Cd stress in poplar (Wu et al., 2022). PvERF15 positively regulated Cd tolerance by binging to metal response element-binding TF (PvMTF-1) (Lin et al., 2017). These findings indicated that TFs are crucial in Cd stress signal transduction, as they interact with various genes associated with Cd uptake, transport, and tolerance, functioning either as transcriptional activators or repressors. Moreover, AtMYB12 can increase the phenylpropanoid content by activating PAL, C4H, 4CL, and CHS gene expression (Xu et al., 2022). MeERF72 was found to exert a negative regulatory effect on the expression levels of MeSuS1 in cassava (Liu et al., 2018). PuERF12 and PuMYB44 were identified as candidate TFs that may regulate the expression of PuLOX2S in Nanguo pears (Zhang et al., 2022). In this study, MYB, MYB_related, ERF, GRAS, bZIP, WRKY and bHLH could effectively interact with phenylpropanoid biosynthesis, starch and sucrose metabolism and alpha-linolenic acid metabolism pathway genes (Figure 8), suggesting these TFs impact on the accumulation of these metabolites, thereby contributing to the G. pentaphyllum response to Cd stress.

A possible model of transcription- and metabolism-driven Cd response mechanisms in G. pentaphyllum is shown in Figure 9. Cd could either activate or repress the expression of TFs, including MYB, MYB_related, ERF, GRAS, bZIP, WRKY and bHLH. In addition, phenylpropanoid biosynthesis, starch, and sucrose metabolism, and alpha-linolenic acid metabolism pathway genes were activated and these increased the levels of alpha-linolenic acid, glucose, and chlorogenic acid metabolites. The up-regulated genes such as ABCCs and ABCGs upon Cd application highly activated the synthesis or transport of ABC transporter-related metabolites (L-isoleucine, L-histidine, L-arginine, L-valine, and L-leucine). Taken together, the TFs, ABC transporter genes, and metabolites (i.e. JA, glucose and phenylpropanoid) might all play important roles in G. pentaphyllum response to Cd by regulating the activities of antioxidant enzymes and the production of non-enzymatic antioxidants. These would work together to transport, fix, and sequester Cd, thereby inducing the defensive metabolite response for G. pentaphyllum to adapt better to the Cd stress. These results provide groundwork for comprehending the Cd response mechanisms in G. pentaphyllum. Nevertheless, the functions as well as the biochemical mechanisms of how these TFs, genes, and metabolites perform their tasks require further study.