The high-Cd cultivar (BXGZ) and low-Cd cultivar (MYQZ) of eggplant used in this study were obtained from a previous screening trial. Eggplant seeds of BXGZ and MYQZ were surface sterilized with 5% hydrogen peroxide and sown in a Petri dish for germination. Seedlings were transferred to sand and grew for one week after germination. Then eggplant seedlings with consistent growth condition were selected and planted in 500 ml cups for hydroponics with half-strength Hoagland solution. Seedlings were grown in a plant culture chamber at of 28°C and 250 μmol/m²/s light intensity (14 h per day), and the half-strength Hoagland solutions were changed every 3 days to ensure optimal growth conditions for eggplants. After 10 days of growth, seedlings of each eggplant cultivar were randomly divided into 2 groups, one was treated with Cd-free half-strength Hoagland solutions as control (CK), and the other was treated with Cd concentration of 10μM half-strength Hoagland solutions as an experimental group (Cd). The samples of eggplant seedlings were harvested after 2 weeks of culture and 3 seedlings were used as biological replicates for each treatment.

Total RNAs of eggplant root samples were isolated through EASYspin Plant RNA kit (Aidlab, China). Four cDNA libraries with 3 biological replicates each were constructed from reverse transcription of total RNA and sequenced with the Illumina HiSeq2500 (Illumina, CA, USA). The raw data have been uploaded in the SRA database of NCBI, and the BioProject accession number is PRJNA908752. To ensure high quality sequencing, the raw reads were processed to clean reads using fastp 0.18.0. Gene expression levels for each cDNA library were normalized by fragment per kilobase of transcript per million mapped reads (FPKM). Differentially expressed genes (DEGs) were identified by the criteria of false discovery rate (FDR) < 0.05 and |log₂ fold-change value| ≥1. Gene Ontology (GO) and KEGG pathway analyses of the DEGs were accomplished through Blast2GO 2.8 and BLASTX.

RNA-Seq results were verified by quantitative real-time PCR (qPCR). The cDNA templates used for qPCR were obtained through the reverse transcription of the remaining total RNA with the help of PrimeScript™ RT Master Mix kit (Takara, Japan). Five DEGs were randomly selected for qPCR experiments on CFX96 Touch instrument (Bio-Rad Laboratories, Hercules, CA) coupled with SYBR Green II PCR Master Mix kit (Takara, Japan), and the qPCR results were calculated using the _ΔΔCt method (Schmittgen and Livak, 2008).

Samples of 0.1 g of oven-dried plants were fully digested in a microwave oven (Microwave Digester XT-9900A, Shanghai Xintuo Analytical Instruments Co., Ltd., China) with HNO₃:H₂O₂ (10:3, v/v). The digested samples were fixed to 15 ml with 1% nitric acid, and then Cd concentrations were measured by flame atomic absorption spectrophotometer (Hitachi Z-2300, Japan). To ensure data accuracy and quality control of the test, plant GBW07605 (provided by the National Research Center for CRM, China) was used as the certified reference material with a recovery rate of 97.21% and a relative standard deviation of 3.2% for Cd.

The fresh root samples of BXGZ and MYQZ treated with the CK and Cd were used for this experiment. The pectin methylesterase (PME) activity, cysteine synthase, glutathione (GSH), oxidized glutathione (GSSG) and peroxidase (POD) activity were tested according to the instructions of PME, cysteine synthase, GSH, GSSG and POD assay kits from Suzhou Comin Biotechnology (China), respectively.

The transpiration rate of eggplant was measured using the portable photosynthetic apparatus HM-GH60 (Hengmei Electronic Technology Co., Ltd, China). The epidermal cells of the eggplant leaves were carefully excised and photographed by an optical microscope (Olympus BX53, Olympus Corporation, Japan). The stomatal density was analyzed via the number of stomata divided by the leaf area. All the tested leaves were picked from the same position of three different eggplant seedlings under the same light condition.

The lignin extraction of eggplant root cell wall was performed according to the method of Gao et al. (2021). Fresh samples of 2g of roots were well ground in liquid nitrogen and then thoroughly mixed with 20 ml of 75% ethanol. After standing in an ice bath for 20 min, the root samples were centrifuged at 8000g for 10 min at 4°C. The collected precipitates were sequentially washed with 14 ml of acetone, methanol/chloroform (1:1) and methanol each for 10 min, and then centrifuged at 5000g for 10 min at 4°C to refine the precipitates. The refined precipitates were air-dried at 40°C to obtain root cell walls and then stored at 4°C for further use. Thoroughly grind 1 μg of root cell walls and 150 mg of potassium bromide (KBr) with an agate mortar and then press into thin slices. The thin slices were placed into a Fourier transform infrared (FTIR) spectroscopy (Nicolet iS20, Thermo Fisher, USA) to identify the absorbance values of the functional groups of the root cell wall in the wavelength range of 400-4000 cm^-1 with a resolution of 4 cm^-1.

Root systems of BXGZ and MYQZ were scanned with Epson Perfection V700 Photo (Seiko Epson Corporation, Japan). The data of the root surface area and tip number were identified on scanned images of BXGZ and MYQZ with the help of root image analysis software Win-RHIZO Pro (Regent Instruments, Canada). Three roots of BXGZ and MYQZ under each treatment were randomly selected for replication.

Both SPSS 23.0 (IBM Inc., USA) and Excel 2016 (Professional Edition, Microsoft, USA) were used to perform statistical analysis and figure development. Two-way analysis of variance (ANOVA) and least significant difference (LSD) tests were employed to determine the significance of data. Results were considered statistically significant when p<0.05.

The results of two-way ANOVA showed that shoot and root biomasses (dry weight, DW) were significantly affected by Cd treatment and cultivar (p<0.01), but not by treatment×cultivar (p>0.05) ( Table S1, S2 , supporting information), indicating that the shoot and root biomasses eggplant were determined by both Cd treatment and cultivar. BXGZ and MYQZ were in good growth status under CK treatment, and both shoot and root biomass of BXGZ were significantly higher than those of MYQZ (p<0.05) ( Figures 1A-C ). Under Cd treatment, the shoot and root biomass of BXGZ decreased by 10.7% and 27.1%, respectively, while a 18.1% and 47.6% reduction of shoot and root biomass were observed in MYQZ, in comparison to the CK treatment ( Figures 1A-C ). Despite the significant reduction in biomass in both BXGZ and MYQZ, the reduction was more pronounced in the root system, and biomass reduction in BXGZ was significantly lower than that in MYQZ (p<0.05), indicating that Cd treatment exerts a greater impact on the root and BXGZ exhibited a stronger Cd tolerance in comparison to MYQZ.

To verify the stability of Cd accumulation of eggplant, Cd concentrations were also measured. Two-way ANOVA revealed that shoot and root Cd concentrations (DW) were significantly affected by Cd treatment and cultivar (p<0.01), as well as by treatment×cultivar (p<0.01) ( Table S3, S4 , supporting information). The Cd accumulation levels of BXGZ and MYQZ were consistent with the results of our previous soil experiments. The shoot Cd concentration of MYQZ was significantly lower than that of BXGZ (p<0.05), while the root Cd concentration of MYQZ was significantly higher than that of BXGZ (p<0.05), indicating that BXGZ and MYQZ were stable in Cd accumulation characteristics and significant differences existed between them for Cd uptake and translocation ( Figure 1D ).

To investigate the variation of molecular mechanisms between BXGZ and MYQZ in response to Cd stress, the differences in the quantity and function of DEGs between BXGZ and MYQZ under CK and Cd treatments were comparatively analyzed. A total of 1183 up-regulated and 717 down-regulated DEGs were identified in BXGZ compared to MYQZ under CK treatment (BXGZ-CK vs MYQZ-CK), and 1918 up-regulated and 1267 down-regulated DEGs were observed in BXGZ compared to MYQZ under Cd treatment (BXGZ-Cd vs MYQZ-Cd). In the comparative analysis of BXGZ-Ck vs BXGZ-Cd and MYQZ-Ck vs MYQZ-Cd, the number of both up-regulated and down-regulated DEGs was also higher in BXGZ than in MYQZ, indicating that there were significant genotypic differences between BXGZ and MYQZ, as well as BXGZ were more responsive to Cd stress ( Figure S1 ). The qPCR result of five randomly selected genes was consistent with the transcriptome profiles, indicating the credibility of the RNA-sequencing results ( Figure S2 ).

To elucidate the specific functions of the DEGs obtained in BXGZ-Cd vs MYQZ-Cd, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed. The majority of DEGs were enriched in the GO term of oxidoreductase activity (GO:0016491), nucleic acid binding transcription factor activity (GO:0001071), component of plasma membrane (GO:0005887 and 0031226), L-phenylalanine catabolic and metabolic process (GO:0006559 and GO:0006558) and sulfur compound metabolic process (GO:0006790), etc. ( Figure 2 ). For the KEGG pathways, the biosynthesis of secondary metabolites (such as phenolic and flavonoid biosynthesis), phenylpropanoid biosynthesis, glutathione metabolism, cysteine and methionine metabolism, plant hormone signal transduction and sulfur metabolism were enriched ( Figure 2 ). Most of these enriched GO and KEGG terms are associated with plant growth and stress resistance (Huang et al., 2022). Therefore, the enrichment analysis of GO and KEGG suggested that BXGZ and MYQZ have different strategies to cope with Cd stress, mainly focusing on antioxidant, sulfur metabolism, and phenylalanine synthesis, etc.

Based on the results of DEGs, the schematic diagram of DEGs in response to Cd stress were summarized ( Figure 3 ). Two genes related to stomatal development and closure were identified, namely Ca²⁺-binding protein-1 (PCaP1) and A2-type cyclins (CYCA2-1). PCaP1 was identified to trigger stomatal closure by involving in Ca²⁺ signaling, and the overexpression of CYCA2-1 was observed to result in reduced stomatal density (Nagata et al., 2016; Qu et al., 2018). The expression level of PCaP1 in MYQZ was significantly higher than that in BXGZ, while CYCA2-1 showed the opposite ( Figure 3 ). Aquaporins has been confirmed to enhance transpiration by regulating the water uptake in roots (Sinclair et al., 2017). The expression level of aquaporins in BXGZ roots was also significantly higher than that in MYQZ (p<0.05) when subjected to Cd stress ( Figure 3 ).

The DEGs related to lateral root development such as auxin response factor (ARF19), GATA transcription factors (GATA4, 5 and 11) and auxin efflux carrier component (PIN5) were shown significantly higher levels in BXGZ than in MYQZ (p<0.05) under either CK or Cd treatment, and the expression of these genes was significantly down-regulated in both cultivars under Cd treatment ( Figure 3 ).

The expression levels of cinnamyl-alcohol dehydrogenase (CAD14) and PODs (PER3, 52, 53 and 72) related to lignin synthesis were significantly higher in MYQZ than in BXGZ (p<0.05) ( Figure 3 ). In addition, the expression levels of pectin related genes such as pectin methylesterase (PME 26, 40 and 46) were significantly higher (p<0.05) and PME inhibitor (PMEI 7 and 9) were significantly lower (p<0.05) in MYQZ than BXGZ ( Figure 3 ), respectively.

The differences in the expression of sulfur metabolism-related genes between MYQZ and BXGZ under CK treatment were not significant (p>0.05) or slightly higher in MYQZ than in BXGZ ( Figure 3 ). However, under Cd stress, the expression levels of genes concerned with sulfate transporters (SULTR1;3 and SULTR4;2), ATP sulfurylase (APS1), adenylylsulfate reductase (APR2), cysteine synthase (RCS3) and glutathione S-transferase (GST1 and 3) were significantly higher in MYQZ than in BXGZ (p<0.05), respectively, indicating a more intensive level of sulfur metabolism in MYQZ under Cd treatment.

Four types of Cd transport-related membrane proteins were identified. Among them, the Cd transporter located on the cell membrane was iron-regulated transporter 1 (IRT1), which serves to enhance the Cd uptake by roots (Barberon et al., 2014). The expression of IRT1 was relatively low in both MYQZ and BXGZ under CK treatment but was significantly elevated (p<0.05) only in MYQZ under Cd treatment. The other three types of Cd transporters were all located on the vacuole, namely ATP-binding cassette transporters (ABCC4, 9 and 17), cation exchanger (CAX5) and metal tolerance protein (MTP4 and 10), and these three transporters are mainly in charge of transporting Cd from the cytoplasm to the vacuole where it can be compartmentalized (Huang et al., 2020). In terms of overall expression levels, these three types of Cd transporters were not significantly (p>0.05) different under CK treatment, while they were significantly higher (p<0.05) in the roots of MYQZ than BXGZ under Cd treatment ( Figure 3 ).

The stomatal distribution per unit area of eggplant leaves on both the front and back sides under Cd treatment was shown in Figure 4 . The density of stomata on the front sides of both BXGZ and MYQZ leaves was significantly lower than that on the back sides, and the density of stomata of BXGZ was significantly higher than that of MYQZ (p<0.05) ( Figure 4 ). Accordingly, the differences in expression levels of PCaP1and CYCA2-1 in MYQZ and BXGZ were consistent with our microscopic observations.

The transpiration rate was significantly affected by Cd treatment and cultivar (p<0.01), but not by treatment × cultivar (p>0.05) ( Table S5 , supporting information). The transpiration rate was positively correlated with the density of stomata and the expression level of aquaporins in eggplant, which was significantly higher in BXGZ than in MYQZ under Cd stress (p<0.05) ( Figure 4 ), suggesting that BXGZ showed stronger transpiration rate under Cd stress.

BXGZ showed more developed lateral roots than MYQZ, and both the root development of BXGZ and MYQZ was affected by Cd stress ( Figure 1 ). The number of root tips was affected by Cd treatment, cultivar and treatment × cultivar (p<0.01) ( Table S6 , supporting information), while the root surface area was only affected by Cd treatment and cultivar (p<0.01) ( Table S7 , supporting information). Under Cd stress, the root tips and surface area of BXGZ were significantly higher than those of MYQZ (p<0.05), and its root tips and surface area decreased at a significantly lower level than those of MYQZ (p<0.05) ( Figure 5A, B ). Although Cd stress affected eggplant root development, in terms of root tip number and root surface area, the roots of BXGZ were significantly less affected compared to MYQZ, which may be attributed to the difference in Cd tolerance between the cultivars

Lignin is produced in a combination of three lignin monomers (H-, G-, and S-lignin) through the processing of phenylalanine catalyzed by multiple enzymes. The lignin content of eggplant roots was affected by Cd treatment, cultivar (p<0.01) and treatment × cultivar (p<0.05) ( Table S8 , supporting information). The results of lignin content of eggplant roots were in agreement with the expression levels of genes related to lignin synthesis as well, which were significantly higher in MYQZ than in BXGZ (p<0.05) ( Figures 6 ). The results of two-way ANOVA showed that the PME activity in eggplant roots was affected by Cd treatment, cultivar (p<0.01) and treatment × cultivar (p<0.05) ( Table S9 , supporting information). As shown in Figure 6 , PME activity in MYQZ was significantly higher than that in BXGZ under both CK and Cd treatments (p<0.05). PME activity in BXGZ showed almost no effect under Cd stress (p>0.05), while its activity in MYQZ was significantly elevated after Cd treatment (p<0.05).

The characteristics of the functional groups of root cell walls were determined by the absorbance values at different wavelengths by FTIR spectroscopy. As shown in Figure 6 , the FTIR wave peaks of eggplant root cell walls were most significant at 3397, 1655 and 1046 cm^-1, representing the O-H of polysaccharides (Regvar et al., 2013), C=O of protein and C-C or C-O of pectin and polysaccharides respectively (Gao et al., 2021). In addition, the elevated absorbances were observed at 2902, 1720, 1513, 1238 and 1157 cm^-1, which were assigned to the C-H of cellulose and pectin, C=O of methyl-esterified pectin, aromatic ring of lignin, amide III and C-C or C-O of lignin and polysaccharides respectively (Regvar et al., 2013; Lahlali et al., 2017). In terms of relative absorbance values, the content of functional groups associated with pectin and lignin was significantly higher in MYQZ cell walls than in BXGZ (p<0.05) under either CK or Cd treatment. Moreover, Cd stress led to a significant elevation in the content of relevant groups in the cell wall of MYQZ, while these groups of BXGZ was only slightly increased at 3397, 2902 and 1238 cm^-1 ( Figure 6 ).

The sulfur metabolism pathway is an important contributor of thiol groups (-SH) in plants, which can chelate with Cd ions and reduce the toxicity of Cd to plants. The results of two-way ANOVA showed that cysteine content was significantly affected by cultivar (p<0.01) and treatment × cultivar (p<0.05) ( Table S10 , supporting information). The GSH content was only affected by both Cd treatment and cultivar (p<0.01) ( Table S11 , supporting information). As for GSSG and POD contents, they were significantly affected by Cd treatment, cultivar and treatment × cultivar ( Table S12, S13 , supporting information). In addition, the content of cysteine in the roots of MYQZ was significantly increased (p<0.05) and significantly higher than that of BXGZ (p<0.05) ( Figure 7 ). The contents of GSH and GSSG in BXGZ and MYQZ were also significantly increased under Cd stress (p<0.05), but the content in MYQZ was remarkably higher than BXGZ (p<0.05) ( Figure 7B ). The content of peroxidase (POD) was consistent with the expression level of POD (PER3, 52, 53 and 72), both of which exhibited significantly higher in MYQZ than in BXGZ under Cd stress (p<0.05) ( Figure 7 ).

The root cell wall is an important organ for plants to achieve self-protection against environmental stresses (Guo et al., 2021). Increased content of cell wall components bound to Cd ions has been observed in plants such as Oryza sativa L. (rice), Zea mays L. (maize), as an effective strategy for plants to cope with Cd stress (Vatehová-Vivodová et al., 2018; Riaz et al., 2021). The stronger the root cell wall fixation of Cd, the lower the amount of Cd that can enter the root cells, and the less can be transported to the shoot and fruit (Huybrechts et al., 2019).

Among the components of the cell wall, pectin is considered to be the main Cd binding molecular, owing to the strong ion exchange activity (Loix et al., 2017; Guo et al., 2021). The Cd ions (Cd²⁺) binding capacity of pectin requires the catalytic release of its free carboxyl groups (-COOH) group by PME (Wu et al., 2020). It has been widely demonstrated that higher PMEs expression levels and PMEs activity enhance Cd binding of cell walls and impede Cd²⁺ entry into cells (Wu et al., 2021a; Huang et al., 2022). In this study, roots of MYQZ under Cd treatment exhibited lower PMEI and higher PME expression levels and activity, and FTIR absorbance also revealed that root cell wall of MYQZ contained more pectin, which ultimately led to stronger fixation of Cd by the MYQZ root cell wall.

According to the studies of Brassica chinensis L. and Vicia sativa, lignin content was involved in the differences of Cd tolerance and Cd accumulation characteristics between high- and low-Cd cultivars (Rui et al., 2016; Wang et al., 2020). In Brassica chinensis L., approximately 14.1% of Cd was adsorbed by lignin in the high-Cd cultivar, while 14.5% of Cd was adsorbed by the lignin of the low-Cd cultivar (Wang et al., 2020). The lignin synthesis-related genes (CADs and PODs), lignin content as well as FTIR analyses of lignin (peaks at 1238 and 1157 cm^-1) in MYQZ were significantly higher (p<0.05) than that of BXGZ in this study, which are consistent with our previous study (Shen et al., 2021), revealing that the distribution of Cd in the root cell wall is significantly higher in MYQZ than in BXGZ. Collectively, differences in lignin should also be an important factor in determining the Cd accumulation characteristics of eggplant.

Sulfur metabolites such as GSH and PCs containing thiol groups (-SH) are crucial for plant growth in response to environmental stress (Cui et al., 2020). Plants have been shown to minimize the toxicity of Cd by elevating the GSH content and promoting the complexation of GSH and Cd²⁺ with the help of GST (Zhang et al., 2021). In this study, the sulfate transporter (SULTR1;3 and SULTR4;2) as well as the enzymes (APS1, APR2, RCS3, GST1 and 3) involved in sulfur metabolism and chelation of Cd were with significantly higher (p<0.05) expression in MYQZ than in BXGZ, and the contents of GSH, GSSG and their precursor cysteine were much higher (p<0.05) in MYQZ under Cd stress. Although MYQZ is Cd intolerant in terms of the biomass results, the enhanced sulfur metabolism should contribute to the mitigation of Cd toxicity.IRT1 is an important metal transporter that contributes to Cd uptake by roots (Barberon et al., 2014). The overexpression and knockdown of IRT1 have been revealed to cause dynamical Cd contents increases and decreases in Brassica chinensis and Arabidopsis thaliana (Connolly et al., 2002; Wu et al., 2021b). Here we observed that the expression of IRT1 was significantly higher (p<0.05) in MYQZ than in BXGZ under Cd stress. In addition, ABC, CAX, and MTP are key transporters responsible for transporting Cd²⁺ from the cytoplasm to the vacuole, which should contribute to the retaining of Cd²⁺ in root vacuoles for the sake of eliminating the Cd translocation and accumulation in shoot (Yang et al., 2022). The DEGs analysis showed that ABC, CAX, and MTPs were also expressed at significantly higher levels (p<0.05) in MYQZ than in BXGZ. In this study, although the higher expression level of IRT1 in MYQZ could certainly promote the Cd uptake, it was very likely that the presence of ABCs, CAXs and MTPs constrained Cd in the vacuole and reduced the translocation of Cd to the shoot.

The root tip is the most active root region for Cd uptake (Lux et al., 2011). It was observed that Cd accumulation in rice was positively correlated with the number of root tips, with fewer root tips leading to less Cd accumulation (Huang et al., 2019). Comparative studies of high- and low-Cd cultivars in potato and hot pepper revealed that high-Cd cultivar possessed longer root length, more root tips, larger root surface area and root volume, which should contribute to its higher Cd uptake and translocation (Huang et al., 2015; Xin et al., 2017). Our previous study revealed that the Cd net uptake via root of BXGZ under low and high Cd treatment was were significantly higher than those of MYQZ (Shen et al., 2021). In this study, with the significantly higher expression of root development genes (ARF19, GATA4, 5, 11 and PIN5), BXGZ showed a more developed root system of BXGZ under Cd treatment, indicting the root development of eggplant should be a key factor affecting the Cd net uptake via roots.

Transpiration pull is the main driving force for Cd translocation from roots to shoots in plants (Khanna et al., 2022). Studies in Arabidopsis and grapevine uncovered a positive correlation between aquaporin expression levels and plant transpiration (Pou et al., 2013; Macho-Rivero et al., 2018). Moreover, the density and aperture of stomata are also the factors affecting the transpiration of plants (Carins Murphy et al., 2014). In Sedum alfredii, ABA regulated the transpiration rate by depressing the expression of root aquaporin (SaPIP), stomatal density and size, resulting in reduced Cd translocation from roots to shoots (Tao et al., 2021). Studies demonstrated that PCaP1 promoted the stomatal closure, while CYCA2-1 contributed to the stomatal development, both of which can affect transpiration rate in plants (Nagata et al., 2016; Qu et al., 2018). The higher expression level of aquaporin and CYCA2-1, and stomatal density along with the higher transpiration rate, which inevitably lead to a stronger driving force for translocation of Cd to the shoot of BXGZ. Conversely, the higher expression of PCaP1 and the corresponding lower density and transpiration rate would limit the transport of Cd to the shoot in MYQZ.