Background

As a non-essential heavy metal in plants, cadmium (Cd) is toxic and widespread present in the environment [1, 2]. Increasing of Cd levels in the arable soil severely limited the crop yield and harmed human health via the food chain [3]. Cd also exerted adverse impacts on various biochemical and physiological activities of plants, such as growth inhibition, oxidative stresses, protein inactivation and disturbance of nutrient uptake [4, 5]. However, plants have developed a sophisticated arsenal of metabolic “weapons” to combat the Cd-induced stresses, including restriction of intake and transportation, immobilization, chelation and sequestration of Cd in vacuoles, efflux Cd from the cytoplasm [6,7,8].

Accumulation of Cd in plant shoots is closely related to a myriad of physiological processes, including root uptake, vacuoles sequestration, and xylem and phloem translocation [9]. Generally, Cd can enter plants from soil through either apoplasmic pathway or symplasmic pathway [10]. Plants actively acquire Cd mainly through essential elements uptake systems, involving Fe2+, Ca2+, Zn2+, and Mn2+ [11,12,13]. It has also been suggested that high-affinity Cd transporters were involved in Cd uptake in Thlaspi caerulescens [14]. After entering the root, Cd is first delivered to the stele via the endodermis [14] and then transported to shoots via the xylem under the driving force of leaf transpiration [10, 15]. Phloem transportation is responsible for further seed or grain Cd accumulation [16, 17].

During the past decade, the continuing expansion of available transcriptional data has led to identification and characterization of the underlying genetic basis behind the above Cd accumulation physiological processes. Such advancement has tremendously enhanced our ability to explore Cd translocation and detoxification in Cd hyperaccumulating plants, such as Arabidopsis halleri [7, 18], Brassica juncea [19, 20], Sedum alfredii [21], and Noccaea caerulescens [22,23,24], as well as cultivated plants like the pea (Pisum sativum L.) [25], barley (Hordeum vulgare L.) [26, 27], and rice (Oryza sativa L.) [28, 29].

Soybean is a protein- and oil-enriched crop which is a main source of essential amino acids of plant food for humans and animals [30]. Generally, soybean is Cd-sensitive and accumulates Cd even at a low concentration of Cd in soils [31], the risk of Cd accumulation in soybean has raised great concerns [32]. Further, soybean genotypes differ greatly in Cd tolerance and accumulation [33, 34]. The Codex Committee on Food Additives and Contaminants proposed a safe upper limit of 0.2 mg/kg Cd in soybean seeds[35]. However, the Cd uptake, translocation, and accumulation processes in soybeans are still mostly unknown. In this research, we examined the Cd content in shoots and roots of two soybean varieties and found that Cd accumulation in these two varieties was significantly different. We employed comparative transcriptome analysis for the roots of the two soybean varieties before and after Cd treatment to elucidate the potential genetic reasons for the different physiological traits. Findings from this study may provide new insight into the molecular-assisted breeding methods for soybeans.

Results

The growth and cadmium accumulation of two soybean cultivars under cadmium treatment

After cadmium treatment for 72 h, HAS accumulated more cadmium compared with LAS not only in roots but also in shoots, the accumulated cadmium mainly existed in roots (Fig. 1a). To compare the effects of cadmium on two soybean varieties, we detected the growth parameters. Under both control and Cd treatment, the biomass of LAS was higher than HAS (Fig. 1b, c). The growth of two soybean varieties was all inhibited when treated with 50 µM CdSO4 for 72 h. The shoot biomass was significantly inhibited in HAS and LAS after cadmium treatment, decreased by 21.43% and 8.91% separately (Fig. 1b). The root biomass of HAS was more inhibited by cadmium treatment than LAS, decreased by 12.85% and 2.14% (Fig. 1c).

Fig. 1
figure 1

Cd concentrations and dry weight in HAS and LAS soybean varieties under 50 µM CdSO4 treatment. a Cd concentrations in shoots and roots in two soybean varieties. b, c Dry weight of shoots and roots in two soybean varieties. Letters indicate significant differences at the p < 0.05 level of the LSD test

Full size image

RNA-Seq analysis

Approximately 18.76 million clean reads were obtained from the soybean root samples after quality control and sequence filtering (Supplementary Material File 1). The GC content of each sample was between 44.25–45.18%, and the average quality score ≥ 30 was 96.04%. The mapping efficiency of the eight samples to the soybean genome (Wm82.a2.v1) was ~ 84.01–86.01%, as shown in Supplementary Material File 2.

qRT-PCR validation

Twenty genes (ten genes from LAS, ten genes from HAS) with different expression patterns were randomly selected to preform RT-qPCR (Supplementary Material File 8). The gene expression levels measured by RT-qPCR were compared to those measured by RNA-seq methods, the results showed a highly significant correlation (R2 = 0.87) was observed between the RT-qPCR and RNA-seq data sets (Fig. 2), which confirmed that RNA-seq method generated reliable expression data.

Fig. 2
figure 2

Validation of RNA-Seq data by qRT-PCR of twenty genes in soybean roots after 50 µM CdSO4 treatment in LAS and HAS varieties

Full size image

Number of DEGs in different soybean varieties

In order to investigate the molecular mechanisms for different Cd accumulation in the two soybean varieties, transcriptional analysis was conducted. 72 h treatment with 50 µM CdSO4 was adopted to explore the response of soybeans to Cd stresses. Transcriptome libraries were created, and sequences from two biological replicates (each for HAS and LAS roots) under control and CdSO4 treatments were examined.

DEGs were identified by comparisons of the FPKM values for each gene between HAS and LAS (HAS-0 versus LAS-0 and HAS-50 versus LAS-50) or between Cd-treated and non-Cd-treated samples of each genotype (HAS-50 versus HAS-0 and LAS-50 versus LAS-0). The results showed that under control conditions, 907 genes (371 up and 536 down) were differentially expressed between HAS and LAS. After Cd treatment, DEGs between HAS and LAS were reduced to 778 (425 up and 353 down), among which 357 genes were common between the varieties, which indicate genetic differences between HAS and LAS (Fig. 3a, c). After Cd treatment, 739 DEGs (265 up and 474 down) were found in HAS; however, only 259 DEGs (88 up and 171 down) were observed in LAS, and 64 genes were common in these DGEs (Fig. 3b, d).

Fig. 3
figure 3

Summary of DEGs. a and b, number of DEGs between HAS and LAS under 0 or 50 µM CdSO4 conditions. c and d, Venn diagrams of DEGs in (a) and (c), respectively

Full size image

Gene ontology (GO) analysis of DEGs

To identify the major functional categories represented by the DEGs, GO enrichment analysis was performed (Fig. 4, Supplementary Material File 3). A total of over 21,000 annotated by GO annotation were assigned into three main GO functional categories -biological process, cellular component, and molecular function. For Cd-responsive DEGs, GO items of cellular process, metabolic process, response to stimulus and single-organism process in the biological process category, and cell, cell part and organelle part in the cellular component category, and binding and catalytic activity in the molecular function were enriched in both HAS and LAS(Fig. 4a, b).

Fig. 4
figure 4

GO enrichment analysis of all DEGs. Genes were assigned into three main categories: biological processes, cellular components or molecular functions. a The enriched GO terms of DEGs in LAS after cadmium treatment; b The enriched GO terms of DEGs in HAS after cadmium treatment. The y-axis indicates the percentage of DEGs numbers vs. background gene numbers in a given category. Detailed information of analysis of gene GO classification is illustrated in Supplementary Material File 3

Full size image

Pathway enrichment analysis of DEGs

KEGG pathway enrichment analysis was performed to classify the biological functions of the DEGs by mapping these genes to the reference pathways in the KEGG database. For DEGs between LAS and HAS, pathways of glycolysis/ gluconeogenesis, plant-pathogen interaction, phenylalanine metabolism and taurine and hypotaurine metabolism were enriched under no cadmium treatment (Fig. 5a, Supplementary Material File 4), after cadmium treatment, the DEGs between LAS and HAS were mainly enriched in glutathione metabolism and plant-pathogen interaction pathways (Fig. 5b, Supplementary Material File 4). In LAS, the DEGs were mainly enriched in taurine and hypotaurine metabolism, phenylpropanoid biosynthesis and phenylalanine metabolism pathways under Cd condition(Fig. 5c, Supplementary Material File 4), while in HAS under Cd condition, the DEGs were mainly enriched in biosynthesis of unsaturated fatty acids, glutathione metabolism, fatty acid metabolism and ABC transporters pathways (Fig. 5d, Supplementary Material File 4).

Fig. 5
figure 5

KEGG analysis of DEGs in LAS and HAS without (a) or with (b) cadmium treatment and DEGs of LAS (c) or HAS (d) after cadmium treatment. Rich Factor refers to the ratio of the differentially expressed gene number and the number of genes annotated in this pathway and large Rich Factor indicates high degree of enrichment. The area of each colored circle is proportional to the number of genes involved in each pathway, the color indicated the p value, and the x-axis is the Rich Factor. Detailed information of KEGG classification is illustrated in Supplementary Material File 4

Full size image

Phenylalanine metabolism responding to cadmium stress in Low-cadmium-accumulating genotype (LAS)

KEGG analysis showed that phenylalanine metabolism pathway was enriched in LAS after cadmium treatment, suggested that genes involved in this pathway played important role in cadmium tolerance in LAS after cadmium exposure. We analysed the genes differently expressed in this pathway, and found that 10 of 12 DEGs in this pathway encoded one key enzyme- peroxidase, 1 of 126 DEGs encoded aspartate aminotransferase, the rest 1 DEG encoded an uncharacterized protein. (Fig. 5, Supplementary Material File 4).

ABC transporters responding to cadmium stress in High-cadmium-accumulating genotype (HAS)

Fourteen ABC transporters were found to enriched in HAS after cadmium according to the KEGG analysis. Only 1 ABC transporter was down-regulated, the others were upregulated after cadmium in HAS ((Fig. 5, Supplementary Material File 4).

Genes involved in heavy metal transport

There were more metal transporter genes differently expressed in HAS-0 versus HAS-50 than in LAS-0 versus LAS-50 (Fig. 6, Supplementary Material File 5), and most of these changes were variety specific. Fourteen ATP-binding cassette (ABC) transporter genes (13 upregulated and 1 downregulated) were found in HAS after Cd treatment. Two zinc-regulated transporter/iron-regulated transporter-like protein (ZIP) zinc transporters (Glyma.13G338300.Wm82.a2.v1 and Glyma.15G036200.Wm82.a2.v1) were found upregulated in both HAS and LAS after Cd exposure

Fig. 6
figure 6

Hierarchical clustering of the transporter genes in LAS (a) and HAS (b) under Cd exposure. Detailed information of differentially expressed transporter genes is illustrated in Supplementary Material File 5

Full size image

After Cd exposure, two sulfate transporters were upregulated in LAS, while in HAS, only one was upregulated and the other downregulated. Auxin transporters were all downregulated, one in LAS-0 versus LAS-50 and two in HAS-0 versus HAS-50 (Fig. 6, Supplementary Material File 5).

Differential expression of transcription factors after cadmium treatment

Transcription factors (TFs) play important roles in the Cd-responsive gene networks. Several Cd-induced TFs that belonged to different families were identified (Fig. 7a,b, Supplementary Material File 6). One WRKY TF was separately upregulated in LAS-0 versus LAS-50 and HAS-0 versus HAS-50. Three basic leucine zipper (bZIP) TFs were downregulated in HAS-0 versus HAS-50, whereas only one bZIP TF was identified in LAS-0 versus LAS-50(Fig. 7a, b). Four basic helix-loop-helix (bHLH) TFs were differentially expressed in HAS-0 versus HAS-50, two downregulated and two upregulated, while one downregulated bHLH TF was identified in LAS-0 versus LAS-50. Three ethylene-responsive TFs were upregulated in LAS-0 versus LAS-50, while two were identified in HAS-0 versus HAS-50, one upregulated and one downregulated. Further, one downregulated PosF21 probable TF were found not only in LAS-0 versus LAS-50, but also in HAS-0 versus HAS-50, while one upregulated SAC51-like TF, two downregulated E2F TFs, three downregulated GATA TFs, one downregulated EGL TF and one downregulated heat stress TF were found only in HAS (Table 1). Significant differences in metal transporter gene expression were found in LAS-0 versus LAS-50 and HAS-0 versus HAS-50 (Fig. 7a, b, Supplementary Material File 6).

Fig. 7
figure 7

Hierarchical clustering of the transcriptional factors in LAS (a) and HAS (b) under Cd exposure. Detailed information of differentially expressed transcription factor genes is illustrated in Supplementary Material File 6

Full size image
Table 1 Differentially expressed genes of the transcription factor family
Full size table

Discussion

After cadmium treatment, two soybean varieties accumulated greatly amount cadmium especially in roots and HAS accumulated more cadmium than LAS not only in shoots but also in roots (Fig. 1a). LAS seems more tolerant to cadmium than HAS. The dry shoot weight of HAS decreased more than the decreased biomass of LAS dry shoot (Fig. 1b). The dry root weight of HAS decreased significantly while there was no influence on the dry root weight of LAS (Fig. 1c). All these results implies the difference mechanism in different soybean cultivars responding to cadmium.

RNA-Seq was conducted to dissect the molecular mechanism behind differential response to Cd stress between these two soybean varieties. Significant genetic differences were found by comparing Cd-responsive DEGs between them. We identified more DEGs in HAS (265 up and 474 down) compared to LAS (88 up and 171 down), indicating a faster response to Cd stress in HAS as compared with LAS. This finding is consistent with the reports of Qian Zhou and colleagues [36] which showed that in pak choi, after Cd treatment, more DEGs were identified in the high-Cd-accumulating genotype compared to the low-Cd-accumulating genotype.

Plants employ an important strategy in Cd detoxification, like antioxidative enzymes. Plant peroxidases (EC 1.11.1.7) are oxidoreductases and have been suggested to be involved in various metabolic steps such as the synthesis of lignin [37, 38], Lignins are particularly important in the formation of cell walls [39], Many studies had reported that heavy metals altered the activity of peroxidases, and the activity of peroxidases which was associated with lignification as a defencing response of roots to Cd [40, 41]. Our results showed that after cadmium treatment, 8 DEGs of peroxidase were high expressed in LAS, which suggested to be crucial for resistance to Cd in LAS.

Among various strategies employed by plants to detoxify Cd, chelation and sequestration are two effective ways of restricting the transportation and circulation of free Cd ions in the cytosol [36]. Reduced glutathione (GSH) plays a vital role in Cd detoxification by conjugation with Cd, a reaction catalyzed by glutathione S-transferases (GST) [2, 42]. After Cd treatment, four GST-encoding genes were upregulated in HAS (Supplementary Material File 7), suggesting that the Cd resistance of HAS depends heavily on the GST-Cd conjugation process. Moreover, GSH could serve as precursors to phytochelatins (PCs), which play important roles in Cd detoxification by forming the PC-Cd complex in plants [36]. Enhancement of PC generation could increase Cd accumulation [43].

It has been suggested that genes involved in Cd efflux and transportation played starring roles in influencing the Cd accumulation ability of different genotypes [43]. These genes are generally categorized into several families based on subcellular location, their sequence specificity, and their carried metals [6]. ABC transporters help transport various substances involved in the response of plants to different environmental stress [44]. For example, it was reported that in rice and Arabidopsis, a PDR-type ABC transporter, encoded by OsPDR9 and AtPDR8, was induced by Cd stress and acted as an efflux pump for Cd or Cd conjugates [45, 46]. In our experiment, we found more than ten (mostly upregulated) ABC transporter genes only in the HAS-0 versus HAS-50; this expression may be responsible for the high uptake of Cd in roots and subsequently contribute to the high accumulation of Cd in HAS shoots.

Cd can enter root cells in the form of Cd2+ through ZIP transporters, which are capable of transporting a variety of divalent cations, including Cd2+, Fe2+, Zn2+, and Mn2+ [47, 48]. In our study, two ZIP transporters (Glyma.15G036200.Wm82.a2.v1 and Glyma.13G338300.Wm82.a2.v1) were upregulated in both genotypes, which indicated that ZIP transporters played vital roles in the uptake of Cd in two soybean genotypes.

Our results revealed that Cd also affected sulfur assimilation. Two sulfate transporters were upregulated in LAS-0 versus LAS-50 only, suggesting that LAS absorbed higher sulfate than HAS under Cd stress. Sulfur assimilation correspondingly activated pathways involved in GSH biosynthesis [49]. GSH functions as an effective antioxidant to mitigate Cd-induced oxidative stresses and plays a vital role in synthesizing PC [42]. Apparently, the sulfur assimilation pathway and GSH synthesis are two important ways of ameliorating the Cd-induced phytotoxicity in LAS.

Auxin, a key plant hormone, which is reported not only to regulate plant growth and development [50] but also in response to cadmiu Cd stress [51]. Auxin transporters, which mediated the transport of auxin, were found to responding to Cd stress not only in LAS but also in HAS. Treatment with the auxin transporter inhibitors increased the Cd sensitivity of WT rice [52]. Our results found that the expression of auxin transporters were decreased after Cd treatment not only in LAS but also in HAS, which suggested that auxin transporter played important roles in soybean responding to Cd stress.

In this research, differential expression of many transporter genes between the two soybean genotypes in response to Cd stress. Together, the overall findings indicate that transporters may play important but different roles in Cd transport and accumulation in LAS and HAS, thus resulting in the corresponding differential Cd tolerance.

TFs participate actively in a wide range of plant stress signaling processes. They belong to different families as diverse as bZIP, WRKY, NAC, ERF (ethylene-responsive factor), and MYB, and are essentially involved in regulating the specific stress-related gene expression under Cd stress [53]. Furthermore, differential expression of several TF genes (WRKY, MYB, ERF, bHLH, and bZIP) was observed in the two soybean genotypes in response to Cd stress, suggesting an important role of TFs in Cd stress response in the soybean, especially in HAS, which exhibited a greater number of differentially expressed TF genes. To decipher the networks of the whole differing Cd-stress-responsive pathway, further analysis of the differentially expressed TF genes is needed. Their potential role in heavy metal tolerance is currently studied by our team.

In summary, we identified two soybean varieties (LAS and HAS) that differentially accumulate Cd and then prepared and sequenced cDNA libraries from untreated and Cd-treated roots. Numerous DEGs were identified in the two soybean varieties under Cd stress. Transcription dynamics of Cd response genes and their related major biological functions were characterized based on GO and KEGG categories. Gene expression analysis suggests that the differential expression of TF and transporter genes is mainly responsible for controlling the contrasting Cd accumulation of the HAS and LAS soybean varieties. Further, research involving in gene function validation will be conducted to clarify the mechanism of its response to cadmium stress.

Conclusions

Based on the transcriptome sequencing of two soybean varieties (su8, high-Cd-accumulating (HAS) and su7, low Cd-accumulating (LAS)) grown with 0 or 50 μM CdSO4, a total of 18.76 million clean reads from the soybean root samples were obtained. More differentially expressed genes were found in HAS than LAS after cadmium treatment. Differentially expressed genes were mainly distributed in “Plant-pathogen interaction”, “Phenylpropanoid biosynthesis”, “Phenylalanine metabolism”, “Carbon metabolism”, “Starch and sucrose metabolism”, “Glutathione metabolism”, and “Protein processing in endoplasmic reticulum” pathways. DEG clustering and enrichment analysis showed several identified biological processes for coping with Cd stress. Some metal transporters and transcription factors were differently expressed in the two soybean varieties after cadmium treatment. In general, this study revealed new insights on the underlying molecular mechanisms after cadmium treatment, which provides a foundation for further function identification of genes in soybean.

Methods

Plant material and Cd treatment

Seeds of two soybean cultivars-su8 (high-Cd-accumulating (HAS)) and su7 (low Cd-accumulating (LAS)) which were selected from Jiangsu Acadmy Agricultural Sciences and suitable for planting in Jiangsu Province were surface sterilized by 2% H2O2 for 10 min and fully rinsed with deionized water. Thereafter, the sterilized seeds were sown in sterile vermiculite for germination under constant temperature (25 ± 1 °C) and a fixed photoperiod (14:10 h light:dark cycle). After 1 week, similarly sized healthy seedlings of each genotype were transplanted to a half-strength modified Hoagland nutrient solution [54] in a greenhouse under controlled temperature (25 − 28 °C) and a fixed photoperiod (14:10 h light:dark cycle).

When the primary leaves were fully opened, soybean seedlings of the two genotypes were treated with fresh medium with or without CdSO4 (final Cd concentration of 50 μM) for 72 h. These conditions represent mild stresses and would not cause visual toxic symptoms for the two genotypes. The experiment was completely randomized with three replicate vessels each with ten seedlings.

For RAN sequencing, after 72 h Cd treatment, shoots and roots of each genotype were separately harvested, followed by wash three times with deionized water. For Cd accumulation determination, washed shoots and roots were oven-dried at 70 °C to a constant weight. Fresh root tissues were frozen in liquid nitrogen (N2) and stored in − 80 °C refrigerator for future RNA extraction and subsequent de novo library construction. Roots from four seedlings of each genotype with or without Cd treatment were randomly selected for RNA sequencing. The experiments were performed with three biological replicates.

Determination of Cd concentration and dry weight

Shoots and roots from three plants of each genotype with or without Cd treatment were oven-dried (70 °C) to a constant weight and weighted, then digested with a solution of extra pure HNO3 and HClO4 (87:13, v:v) in a microwave. After cooling down, the digester was measured for Cd concentration using FAAS (HITACHI Z-5300, Japan) in accordance with the manufacturer’s instructions. A Certified Reference Material (CRM; GBW-07603, provided by the National Research Center for CRM, China) was used to assess the precision of the analytical procedures for plant material. Data were analyzed using SPSS statistics 17.0 for the analysis of variance (ANOVA) test. Significant differences in Cd concentration between the two genotypes were determined by the least significant difference (LSD) test at P < 0.05.

RNA isolation, RNA-Sequencing (RNA-Seq) library preparation, and sequencing

Total RNA was isolated from frozen and ground root tissue using a plant RNA kit (OmegaBio-Tek, Norcross, GA, USA) according to the manufacturer’s instructions. RNA concentration was assessed using a ND-8000 spectrophotometer (Nanodrop Technologies, Inc., Wilmington, DE, USA), a 2100-Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA), and agarose gel electrophoresis. RNA samples with no observable smearing (a 260/280 ratio above 2.0 and a RNA integrity number greater than 8.0), were collected for subsequent analysis.

The de novo transcriptome analysis was performed by combining three replicate root samples into a single total RNA sample (In total, two samples were prepared for each treatment), which were then sent to the biological company for sequencing.

De novo transcriptome assembly and annotation

We performed sequencing using an Illumina GAll according to the manufacturer’s protocol with an average sequencing depth of 5.34X. Adapter sequences were eliminated from the raw sequence reads using a FASTX-toolkit. Then, sequence quality was analyzed, and low-quality sequences were removed accordingly using FastQC to obtain clean reads to increase sequence confidence. Clean reads were then aligned to the soybean genome (Glycine max Wm82.a2.v1) using Tophat v. 2.0.10. Subsequently, transcriptome assemblies were conducted using Cufflinks, while gene expression levels were calculated as reads per kilobase of exon model per million mapped reads (FPKM) [55]. The differentially expressed genes (DEGs) were identified and screened by DESeq software [56]. In this method, the adjusted P-values used the false discovery rate (FDR) of < 0.01 and |log2 (fold change)|> 1 or < -1 as the thresholds for differential gene expression. Meanwhile, the screening process adopted fold changes of the expression levels between different samples as the criteria.

Validation of gene expression

A set of 20 randomly selected DEGs from the transcriptome analysis (ten genes from LAS, ten genes from HAS) were validated by quantitative Real-time polymerase chain reaction (qRT-PCR) using the same RNA samples used for transcriptome analysis according to [36]. The sequences of the corresponding target genes were obtained from the National Center for Biotechnology Information (NCBI). Primers were designed using Primer Premier 5.0 (Supplementary Material File 8) for qRT-PCR. And genomic DNA was removed using the RNase-free DNase I Set (Omega, USA) according to the manufacturer’s instructions after RNA extraction. First-strand complementary DNA (cDNA) was synthesized from approximately 1 µg of RNA using a reverse transcription kit (BioTeke, China). qRT-PCR was performed using SYBR Green (Bio-Rad) in a LightCycle480 system (Roche).The relative quantification was normalized to the GmActin reference gene. The 2–ΔΔCt method was used for data analysis. Each PCR reaction, including the control reaction, was performed in triplicate.

Statistical analysis

Statistical analyses were performed in Excel and SPSS v17.0 (link/cite SPSS.) The significance threshold between samples was p < 0.05, and all results of expression data were indicated as averages ± standard deviations (SDs).