First, we examined the Cd phenotype by using Arabidopsis wild-type and ABA mutants (Figure 1A). There were no phenotypic differences between the two under control conditions. However, when plants were cultivated for 4 weeks under control conditions and then exposed for 3 days to 20 μM Cd, Col-0 showed more resistance to Cd, while bglu10 and bglu18 mutants displayed more sensitivity to Cd (Figure 1A).

Leaf chlorophyll is an important indicator of plant tolerance to Cd (DalCorso et al., 2008). We observed that after Cd stress, chlorophyll degradation rate in Col-0 was 12%, while the corresponding rates in bglu10 and bglu18 were both 20%, which was significantly higher than that of Col-0. This finding demonstrated that Col-0 was more tolerant to Cd than either of the ABA mutants (Figure 1B).

Proline and malondialdehyde (MDA) are also important indicators of stress tolerance. Proline was able to maintain the stability of the membrane structure and to eliminate reactive oxygen species. The accumulation of proline is positively correlated with plant stress tolerance. As for MDA, it is one of the most important products of membrane lipid peroxidation; it is cytotoxic, because it promotes cross-link polymerization of living macromolecules, such as proteins and nucleic acids. After Cd stress, proline concentration in roots of Col-0 was significantly higher than in roots of either ABA mutant. In contrast, root MDA was significantly lower in Col-0 than in ABA mutants (Figures 1C,D). Our data suggest that after Cd stress, Col-0 showed higher Cd tolerance when compared to either of the ABA mutants tested.

In view of the phenotypic differences shown in Figure 1, because the materials are ABA mutants, we determined the ABA distribution and content differences in root cells under Cd stress (Figures 2A,B). We found that, compared with bglu10 and bglu18, the ABA content in Col-0 root vacuoles accounted for 77.0% of protoplast ABA content, which is much lower than the ABA contents found in the ABA mutants, which were 91.9 and 88.5%, respectively (Figure 2A). Therefore, we conclude that the amount of ABA in the cytoplasm of Col-0 root cells was significantly higher than that in either the bglu10 or the bglu18 ABA mutant (Figure 2B).

We took the Arabidopsis roots that were grown under control conditions for 4 weeks and then treated them with 200 μM Cd for 6 h. We then tested for the gene expression of NRT1.5 and NRT1.8. The expression of NRT1.5 was significantly inhibited after Cd treatment, regardless of the material. In contrast, the expression of NRT1.8 was significantly induced (Supplementary Figures 1a,b). However, fold change of NRT1.5 down-regulation and NRT1.8 up-regulation in the wild-type and the mutants was different after exposure to Cd stress. In this case, fold change of NRT1.5 down-regulation in Col-0 was significantly higher than fold change in bglu10 or bglu18. On the other hand, although the expression of NRT1.8 was induced, there was almost no difference in fold change of NRT1.8 up-regulation between the wild-type and the ABA mutants (Figures 2C,D). These results indicated that NRT1.5, but not NRT1.8, responded to ABA signaling.

After 3 days of 20 μM Cd treatment, there was a significant difference in proton pump activity between the wild-type and the ABA mutants tested. V-ATPase (Figure 3A) and V-PPase (Figure 3B) activities were significantly higher in Col-0 than in bglu10 or bglu18. This suggests an increased ability of Col-0 plants to transport Cd²⁺ into the vacuole. The distribution of Cd²⁺ in vacuoles and protoplasts is shown in Figure 3C. As the ratio of vacuolar to protoplasmic Cd²⁺ is higher in Col-0 than that in bglu10 or bglu18, the Cd²⁺ remaining in the cytoplasm in Col-0 is significantly lower than in bglu10 or bglu18 (Figure 3D). At the same time, the proton pump activity also influenced the distribution of NO3− in the cells. Additionally, the ratio of vacuolar NO3− to protoplasmic NO3− in Col-0 was higher than in bglu10 or bglu18 (Figure 3E), therefore, NO3− remaining in the cytoplasm in Col-0 was significantly lower than in bglu10 or bglu18 (Figure 3F).

Previous research demonstrated that stress decouples nitrate assimilation from photosynthesis through stress-initiated nitrate allocation to roots (SINAR), which is mediated by nitrate transporters NRT1.8 and NRT1.5, and functions to promote stress tolerance (Li et al., 2010; Chen et al., 2012). Here, we showed that ABA produced by Arabidopsis wild-type and ABA mutants differed in response to Cd stress. The cytoplasmic ABA levels in Col-0 plants were significantly higher than those in bglu10 or bglu18, which resulted in a much higher degree of inhibition of expression of NRT1.5 in the former, whereas the level of expression of NRT1.8 differed slightly between wild-type and mutants (Figure 2). The function of NRT1.5 is to load the xylem nitrate into the shoot, thus, after Cd stress, Col-0 had more nitrate in the root than bglu10 or bglu18 (Figure 4A). Concomitantly, due to the difference in the activity of the proton pump, the amount of nitrate remaining in the cytoplasm in Col-0 was reduced, as was the nitrate transported to the shoot (Figures 3E,F). The overall result of this was more nitrate left in the roots in Col-0, thereby reflecting wild-type plant resistance to Cd.

On the other hand, due to the difference in the activity of the proton pump, the content of Cd in the cytoplasm of Col-0 was lower than in bglu10 or bglu18 (Figures 3C,D); thus, more Cd accumulated in the roots (Figure 4B), the net result of which was that Cd-induced damage was not as severe in Col-0 plants as in either of the ABA mutants.

In summary, the combined effects of nitrate and proton pump activity increased the resistance of Col-0 plants to Cd. Furthermore, the resistance of Col-0 to Cd was higher than that of bglu10 or bglu18, but the NUE was significantly lower in Col-0 plants than in either bglu10 or bglu18 (Figure 4C). In order to verify the anti-Cd mechanism in plants, we treated B. napus with exogenous ABA and arrived at the following results.

After treatment with exogenous ABA, the cotyledons of B. napus showed more severe yellowing than under Cd treatment alone due to the joint effects of both, ABA and Cd. ABA accelerated senescence of cotyledons, while Cd stress promoted cotyledon yellowing in. However, in this case the new leaves showed no trace of Cd poisoning, while the new leaves of B. napus under Cd-treatment alone showed obvious yellowing. Cd poisoning mainly affected new leaves; thus, the addition of exogenous ABA increased the anti-Cd ability of B. napus (Figures 5A,C). Further, after the addition of exogenous ABA, the proline concentration of B. napus was significantly higher than under Cd treatment alone (Figure 5B), whereas MDA concentration was significantly lower (Figure 5D). This confirmed that the addition of exogenous ABA enhanced Cd resistance of B. napus.

After the addition of exogenous ABA, the expression level of BnNRT1.5 was significantly downregulated (sixfold), compared to Cd treatment alone (Figure 6A). However, there was no difference in the expression level of BnNRT1.8 (Figure 6B). Under CK (normal culture) and ABA treatments, we arrived at the same conclusion: BnNRT1.5 responded to ABA signal and the expression level was downregulated, while BnNRT1.8 did not respond. Further, NO3− content in the shoots and roots under Cd treatment alone was significantly higher than in the case of Cd treatment followed by ABA addition (A+C) (Figure 6C). However, in the (A+C) treatment, the NO3− concentration ratio between root and shoot was significantly higher than under the Cd treatment alone (Figure 6D). This indicated that the addition of exogenous ABA caused a greater proportion of NO3− to be distributed in the root of B. napus seedlings, thereby enhancing their resistance to Cd.

A number of studies have reported that the addition of exogenous ABA inhibited Cd absorption and increased Cd resistance in Arabidopsis and rice (Hsu and Kao, 2003; Uraguchi et al., 2009; Fan et al., 2014). Similarly, here we observed that after the addition of exogenous ABA, the absorption of Cd was also inhibited in B. napus, and that shoots and roots of B. napus were significantly lower in Cd content than under Cd treatment alone (Figures 7A,B).

Arabidopsis thaliana wild-type Columbia-0 (Col-0) was used as the control for ABA conjugate hydrolysis mutants (bglu10 and bglu18). The functions of BGLU10 and BGLU18 have been confirmed in the reports of Wang and Lee. BGLU10, a member of the β-glucosidase family in Arabidopsis, is localized in vacuoles, where it hydrolyzes ABA-GE to produce active ABA; additionally, BGLU18 is localized in the ER, also hydrolyzing ABA-GE to produce active ABA (Lee et al., 2006; Wang et al., 2011). Mutants bglu10 and bglu18 used are BGLU10 and BGLU18 gene-deletion mutants, respectively. These were a gift from Zhang Jianhua, from the Chinese University of Hong Kong. B. napus (814) was provided by the Hunan Branch of Improvement Center of National Oil Crops, Hunan, China.

Arabidopsis plants were grown in a nutrient solution in plastic pots as described in Gong et al. (2003) and Han et al. (2016). The solution was changed every 3 days, with pH adjusted to 5.8 and 0.5 g L^-1 MES (2- (4-Morpholino) ethanesulfonic acid) was added. Pots were arranged in a completely randomized design with six biological replications. The nutrient solution consisted of 1.25 mM KNO₃, 0.625 mM KH₂PO₄, 0.5 mM MgSO₄, 0.5 mM Ca (NO₃)₂⋅4H₂O, 0.025 mM Fe-EDTA, 0.25 ml L^-1 micronutrients (stock solution concentrations were the following: 70 mM H₃BO₃, 14 mM MnCl₂, 1 mM ZnSO₄, 0.5 mM CuSO₄, and 0.2 mM NaMoO₄).

Soaked B. napus seeds were sown onto gauze fixed to an enamel pan, and soaked with deionized water. After 6-days, seedlings were transplanted into 2-L black plastic pots containing nutrient solution. The experiment was laid in a completely randomized block design with six replicates. The nutrient solution consisted of 5.0 mM KNO₃, 1.0 mM KH₂PO₄, 2.0 mM MgSO₄⋅7H₂O, 5.0 mM Ca(NO₃)₂⋅4H₂O, 0.05 mM Fe-EDTA, 9 μM MnCl₂⋅4H₂O, 0.8 μM ZnSO₄⋅7H₂O, 0.3 μM CuSO₄⋅5H₂O, 0.1 μM NaMoO₄⋅2H₂O, and 50 μM H₃BO₃ (Zhang D. et al., 2014). The experiments were conducted at Hunan Agricultural University in a phytotron set at 70% relative humidity, 16 h/8 h light/dark cycle (A. thaliana) or 14 h/10 h light/dark cycle (B. napus), at constant temperature (22°C). The nutrient solution for Arabidopsis plants was changed every 3-days and, after 4 weeks of cultivation, they were treated for 3-days with 20 μM Cd. The nutrient solution for B. napus plants was changed every 5-days and, after 10-days of cultivation, they were treated for 4-days with either 10 μM Cd or 10 μM Cd added with 5 μM ABA. B. napus and Arabidopsis were analyzed separately.

Leaves (approximately 0.15 g) of A. thaliana were sampled and extracted in 10 ml 1:1 absolute ethanol: acetone for 24 h. Absorbance was then measured at 663, 645, and 652 nm to determine chlorophyll concentration. Chlorophyll loss (a) was calculated as the chlorophyll concentration under the control conditions (b) minus chlorophyll concentration under Cd stress, and (c) divided by concentration under control conditions, i.e., a = (b-c)/c^∗100. MDA and proline were measured in root tissues. For MDA, 0.5 g of root tissue was ground in 5 ml 5% TCA, then centrifuged at 925 × g for 10 min. The supernatant was collected and used for determination of MDA concentration using the modified thiobarbituric acid–malondialdehyde (TBA–MDA) assay (Song et al., 2014). Proline was assayed according to the method described in Bates et al. (1973) and Sharma and Dubey (2005). Briefly, root tissues (0.5 g) were sampled and ground in 5 ml of 3% sulfosalicylic acid, then centrifuged at 22000 × g for 5 min. The supernatant was collected and used for determination of proline concentration by reaction with acidic ninhydrin (Chen et al., 2012).

Root tissues of A. thaliana (0.5 g) were collected to isolate intact protoplasts and vacuoles as described in Robert et al. (2007), with minor modifications as outlined in Huang et al. (2012) and Han et al. (2016). Purified protoplasts and vacuoles were subsampled and used to determine NO3− and Cd²⁺ concentrations (Vögeli-Lange and Wagner, 1990) and for enzyme activity assays (Ma et al., 2005). NO3− concentration in protoplasts and vacuoles were measured by a continuous flow auto-analyzer (Auto Analyzer 3, Bran and Luebbe, Norderstedt, Germany) as described previously (Han et al., 2016). The activities of acid phosphatase (ACP) and cytochrome oxidase (COX) were determined using plant ACP colorimetry and COX assay kits (GenMedSci, Inc., Shanghai, China) following the instructions by the manufacturer. ACP activity specific to vacuoles was determined and used to normalize NO3− accumulation. We measured NO3− in the protoplast outside the vacuole, which includes the cytosol and organelles, e.g., mitochondria and Golgi Apparatus (Robert et al., 2007). As most NO3− in the protoplast outside the vacuole is located in the cytosol (Krebs et al., 2010), we refer to NO3− distribution between vacuoles and cytosol rather than vacuole versus protoplast; Cd²⁺ concentrations in protoplasts and vacuoles were measured by inductively coupled plasma-mass spectrometry (ICP-MS, ELAN DRC-e, PerkinElmer, Shelton, United States) as described in Huang et al. (2012), with the corresponding modification.

V-ATPase and V-PPase activities within microsomal membranes collected from the root tissues of A. thaliana were colorimetrically determined as Pi release after an incubation period of 40 min at 28°C (Zhu et al., 2001; Krebs et al., 2010; Han et al., 2015). Reactions were terminated by adding 40 mM citric acid. For the blank value, 10 μg of bovine serum albumin was used instead of tonoplast vesicles. The V-ATPase assay medium contained 25 mM Tris-MES (pH 7.0), 4 mM MgSO₄⋅7H₂O, 50 mM KCl, 1 mM NaN₃, 0.1 mM Na₂MoO₄, 0.1% Brij 35, 500 μM NaVO₄, and 2 mM Mg-ATP. Activity was expressed as the difference in Pi release measured in the absence and in the presence of 100 nM concanamycin A. V-PPase was assayed in a reaction medium containing 25 mM Tris-MES (pH 7.5), 2 mM MgSO₄ × 7H₂O, 0.1 mM Na₂MoO₄, 0.1% Brij 58, and 0.2 mM K₄P₂O₇. V-PPase activity was calculated as the difference in Pi release measured in the absence and the presence of 50 mM KCl.

Nitrate was extracted from tissue samples (shoot: 1 g; root: 0.5 g) in deionized water for 30 min in a boiling water bath; next, 0.1 ml of the sample solution was taken, 0.4 ml of 5% salicylic acid-sulfuric acid solution was added, and mixed. After cooling, the mixture was cooled at room temperature for 20 min, and then 9.5 ml of an 8% sodium hydroxide solution was added. The sample was then allowed to cool to room temperature and spectrophotometrically determined for nitrate at 410 nm (Cataldo et al., 1975).

Whole, hydroponically grown seedlings of B. napus and A. thaliana were sampled, oven-dried to constant weight, first at 105°C for 30 min, followed by 70°C. N concentration was determined as described by Han et al. (2016) (N data is used to calculate NUE). For the Cd²⁺ assay, shoots and roots were sampled separately, dried, and weighed; Cd²⁺ concentration was then determined by ICP-MS, after digesting with 4:1 HNO₃: HClO₄ (Huang et al., 2012).

Endogenous ABA was extracted from the isolated vacuoles and protoplasts of each sample using 0.5 mL of homogenizing buffer (70% methanol, 0.1% formic acid); 2 ng ABA-d6 (Olchemim, Olomouc, Czechia) were added to the extracts as an internal standard (Balcke et al., 2012). The mixture was diluted twice using deionized water, and the ABA concentration of a 50-μL dilution of each sample was determined using the UPLC-TripleTOF 5600+ system (Sciex, Concord, ON, Canada).

Root samples were ground in liquid nitrogen. Total RNA was extracted with TRIzol (Ambion, United States). The first-strand cDNA was synthesized using the total RNA by PrimeScript reverse transcription (RT) reagent kit (TaKaRa, Shiga, Japan). The qRT-PCR assays for the detection of relative gene expression were performed using SYBR^® Premix Ex TaqTM II (Tli RNaseH Plus) (TaKaRa, Shiga, Japan) with an Applied Biosystems StepOneTM Plus Real-time PCR System (Thermo Fisher Scientific, Waltham, MA, United States). The thermal cycles were as follows: 95°C for 3 min, followed by 40 cycles of 95°C for 10 s, then 60°C for 30 s. Melt curve analysis to ensure the primer gene-specificity was conducted as follows: 95°C for 15 s, 60°C for 1 min, 60–95°C for 15 s (+0.3°C per cycle). The gene-specific primers for qRT-PCR assays are listed in Supplementary Table 1 (Bustin et al., 2009; Wang et al., 2014).

We used the SPSS software (IBM SPSS Statistic 19) for ANOVA and mean separation of main effects and interactions using LSD’s multiple range test at P < 0.05. Data are means and SE of three or six replicates from three independent experiments. Different letters associated with specific data (e.g., at the top of histogram bars in figures) indicate significant differences at P < 0.05.