Purple Perilla seeds were sterilized with 1% (m/v) hydrogen peroxide for 60 min and rinsed thoroughly with distilled water. Then, purple Perilla seeds were initially grown on artificial, non-polluted soil for 2–3 weeks until the seedlings developed two healthy, tender leaves. These uniform seedlings were used as research subjects for hydroponic experiments in a greenhouse located at Shenyang Agricultural University, Shenyang, Liaoning, China. The hydroponic experiment was carried out in a greenhouse at temperatures of 25/15°C (day/night), a 16-h photoperiod of approximately 300 mE/m²/s intensity, and 65% average relative humidity.

The uniform seedlings were transplanted in to 1 L pots containing 400 ml of Hoagland’s solution (half strength) of the following composition: 2.5 mmol/L Ca(NO₃)₂, 2.5 mmol/L KNO₃, 0.5 mmol/L KH₂PO₄, 0.5 mmol/L MgSO₄, 25 μmol/L H₃BO₃, 2.25 μmol/L MnCl₂, 1.9 μmol/L ZnSO₄, 0.15 μmol/L CuSO₄, 0.05 mmol/L (NH₄)₆Mo₇O₂₄, and 5 μmol/L Fe-EDTA. A cadmium salt, CdCl₂ 2.5H₂O (guaranteed reagent) was used as the pollutant source. The Cd salt was added to the hydroponic culture. The pH of the nutrient solution was maintained at 6.0 by the addition of 0.1 mol/L NaOH. Four treatments were used: control check (CK), 50% shade treatment (T1), Cd treatment (T2), and 50% shade + Cd treatment (T3). The hydroponic solution was renewed every 4 days. After 16 days, the plants were harvested. The dry tissues were used for Cd concentration analysis with three biological replicates for each treatment. Meanwhile, fresh leaves were frozen in liquid nitrogen and stored at −80°C for metabolite analysis, with six biological replicates for each stress treatment.

Each harvested plant was divided into root and stem sections. Individual plant parts (roots and stems) were first washed with tap water and then rinsed three times with deionized water. The plant parts used for the determination of Cd concentrations were oven-dried at 110°C for 30 min, and then dried at 70°C for 96 h to constant weight.

Roots and stems were separately ground in a mill, and then digested in flasks at 80°C on an electric heating plate with concentrated HNO₃. The temperature was then increased to 110°C, and the mixture was maintained at that temperature until the solution became clear (Zhang et al., 2016). The sample volume was adjusted to 25 ml with ultrapure water. The Cd concentrations in each sample were measured with an inductively coupled plasma mass spectrometer (ICP-MS) (Agilent 7500a, United States). Standard reference material for bush twigs and leaves (GBW07603: GSV-2) was used to monitor the heavy metal recovery from the plant samples (recovery: 90 ± 10%).

The chemical analysis was carried out at the Analytical and Testing Center, Shenyang Agricultural University, Shenyang, Liaoning, China.

The plant tissues (80 mg of leaves) were quickly frozen in liquid nitrogen and ground into fine powder with a mortar and pestle. One thousand microliters of methanol/acetonitrile/H₂O (2:2:1, v/v/v) were added to homogenized solution for metabolite extraction. The mixture was centrifuged for 15 min (14,000 g, 4°C) and the supernatant was dried in a vacuum centrifuge. For LC-MS analysis, the samples were re-dissolved in 100 μL acetonitrile/water (1:1, v/v) solvent.

Analyses were performed using ultra-high performance liquid chromatography (UHPLC) (1290 Infinity LC, Agilent Technologies, United States) coupled to a quadrupole time-of-flight (AB Sciex TripleTOF 6600, United States).

For HILIC separation, samples were analyzed using a 2.1 mm × 100 mm ACQUIY UPLC BEH 1.7-µm column (Waters, Ireland). In both electrospray ionization (ESI)-positive and ESI-negative modes, the mobile phase contained A = 25 mm ammonium acetate and 25 mm ammonium hydroxide in water and B = acetonitrile. The gradient was 85% B for 1 min and was linearly reduced to 65% in 11 min. It was then reduced to 40% in 0.1 min and maintained for 4 min, after which it increased to 85% in 0.1 min, with a 5-min re-equilibration period.

The data of Cd concentrations detected by ICP-MS were analyzed using ANOVA testing and SPSS 18.0 software (SPSS, Chicago, IL, United States). Data, presented as mean values, were compared by Duncan’s test. A p-value less than 0.05 was considered to indicate a significant difference.

For the metabolite data, the processed data were analyzed using the R package (ropls) after being normalized to total peak intensity. The data were then subjected to multivariate data analysis, including Pareto-scaled principal component analysis (PCA) and orthogonal partial least-squares discriminant analysis (OPLS-DA). Seven-fold cross-validation and response permutation testing were used to evaluate the robustness of the model. The variable importance in the projection (VIP) value of each variable in the OPLS-DA model was calculated to indicate its contribution to the classification. Metabolites with VIP values > 1 were further applied to the Student’s t-test at the univariate level to measure the significance of each metabolite, and p-values less than 0.05 were considered to be statistically significant.

The Cd accumulation in the roots and stems of purple Perilla grown in each treatment is shown in Figure 1. Compared with the normal sunlight treatment (T2), the shade stress (T3) increased the Cd concentrations in the roots and reduced the Cd concentrations in the stem. In addition, shade stress caused the translocation rate from roots to shoots to improve relative to the normal sunlight treatment.

To investigate the purple Perilla metabolism in response to shade and Cd stresses, the metabolites in leaves were measured using UHPLC-MS. In total, 247 metabolites were detected in the leaves, including 134 metabolites detected by the positive mode, and 113 metabolites detected by the negative mode. These metabolites could be mainly classified into three groups, including organic acids and derivatives, organic oxygen compounds, and lipids and lipid-like molecules (Supplementary Figure S1).

With the OPLS-DA model, the samples from the control and all stress treatments could be clearly separated for the examined metabolites in leaves, indicating that the metabolite profiles were changed markedly by any stress treatment (Supplementary Figure S2). Variable importance for the projection (VIP) obtained from the OPLS-DA model can measure the explanatory ability of the expression patterns of metabolites for the classification and discrimination of each group of samples. Generally, metabolites with VIP values >1 were considered to have made a significant contribution to model interpretation. In this experiment, the metabolites with VIP values >1 and p-values < 0.05 were considered to show significant differences. These metabolites are displayed in Supplementary Table S1. Overall, we identified 18 increased metabolites and 55 decreased metabolites in the various stress treatments.

To evaluate the effects of shade and Cd stresses on metabolite profiles, the ratios of metabolite levels under each stress and those in the control treatment were calculated (Supplementary Table S1). Notably, meso-tartaric acid was increased by 2.30-fold under shaded condition relative to the control. In the Cd treatment, 2-hydroxyadenine showed a 2.16-fold increase relative to the control. Additionally, DL-indole-3-lactic acid increased dramatically by 9.20-fold when exposed to shade and Cd stresses compared to shaded conditions alone.

Under shade and Cd stresses, the metabolites belonging to different classifications accumulated differently (Figure 2). In comparison with normal sunlight, the organic nitrogen compounds were significantly increased (Figure 2B), while the sharp decline of organic acids and derivatives in addition to lipids and lipid-like molecules occurred (Figures 2A,B) under the shade condition. With Cd stress, the organic acids and derivatives showed a marked increase (Figures 2C,D). However, the organic nitrogen compounds, organic oxygen compounds, lipids, and lipid-like molecules decreased markedly (Figure 2C). Relative to the shade treatment, the benzenoids and organic nitrogen compounds increased significantly, while the majority of the organic oxygen compounds showed a marked reduction when exposed to shade and Cd stresses (Figures 2E,F).

Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis was performed to identify metabolic pathways that were altered in response to abiotic stresses (Figure 3). Twenty pathways were the most significantly varied (p < 0.05, VIP >1). In comparison with the control, glyoxylate and dicarboxylate metabolism and the citrate cycle (TCA cycle) were highly changed (Figure 3A). Under exposure to Cd stress, we found that galactose metabolism and the TCA cycle were strongly influenced (Figure 3B). Moreover, shade and Cd stresses together generated the biosynthesis of amino acids and ascorbate and aldarate metabolism with striking differences relative to the shade treatment (Figure 3C).

There are four major processes involved in Cd transport from roots to stems, including the symplast pathway and apoplast pathway in roots, xylem loading, the symplast pathway and apoplast pathway in leaves, and phloem loading (Mendoza-Cózatl et al., 2011; Wei et al., 2018). Our study found that shade stress inhibited the transport capability from the roots toward the stems relative to normal sunlight. Shade stress reduced photosynthesis and transpiration, which increased Cd accumulation in roots and decreased Cd accumulation in stems (Figure 1). Photosynthesis is an important physiological process for plant productivity that directly provides energy for plant growth (Chen et al., 2020). The reduction of photosynthesis decreased the energy supply for Cd transport in the symplast pathway and the phloem. Simultaneously, we found that the TCA cycle as well as glyoxylate and dicarboxylate metabolism, as the energy metabolic pathways, were importantly influenced by shade stress (Figure 4A). Even more importantly, the inhibition of transpiration reduced Cd transport in the xylem, which is the important pathway from roots toward stems.

When exposed to various abiotic stresses, plants can regulate their metabolisms in order to alleviate the stress effect (Sicher et al., 2012). Our study showed that all cis-(6,9,12)-linolenic acid accumulated significantly under both types of stress (Supplementary Table S1). Moreover, thiamine was upregulated importantly in Cd stress, whether under normal sunlight or shade stress (Supplementary Table S1). Similarly, Zhao et al. (2021) also found that some stresses (including osmotic stress and metal stress) have similar effects on changes in the metabolites in barley plants. It is thus deduced that more accumulation of all cis-(6,9,12)-linolenic acid or thiamine confers purple Perilla with higher tolerance to abiotic stresses. Previous studies have demonstrated the protective role of thiamine in plants grown under various stress conditions, such as salt (Kaya and Ashraf, 2015), copper (Al-Hakimi and Hamada 2011), and Cd stresses (Kaya and Aslan 2020). Our results also showed that the increase of thiamine improved the tolerance to Cd stress in plants (Supplementary Table S1). Previous research revealed that thiamine improved the antioxidant defense system, eliminated the reactive oxygen species (ROS), reduced the Cd content, and improved the Ca and K contents, inducing enhanced tolerance to Cd stress in plants (Kaya and Aslan 2020).

The majority of the metabolites with significant differences were downregulated under shade and Cd stresses (Supplementary Table S1). In the shade and Cd treatments, the organic acids and derivatives (l-malic acid and isocitrate), lipids and lipid-like molecules (3-hydroxy-3-methylglutaric acid), and phenylpropanoids and polyketides (hydroxyphenyllactic acid) decreased. Whether under normal sunlight or shaded conditions, under exposure to Cd, the organic acids and derivatives (pyruvate), the organic oxygen compounds (galactinol, d-galactarate, shikimate, glyceric acid, and d-lactose), and lipids and lipid-like molecules (oleanolic acid and palmitic acid) were downregulated. However, previous studies found that some of above-mentioned metabolites increased under Cd stress, suggesting that they were involved in a defense mechanism in response to abiotic stress. For instance, galactinol has been considered an ROS scavenger under abiotic stress, protecting cells from oxidative stress (Keunen et al., 2013). Thus, it may be assumed that different changes in these metabolites may be influenced by plant species and different forms of abiotic stress. This should be clarified in future work.

Metabolic pathways were analyzed on the basis of the important metabolites listed in Supplementary Table S1. Compared with single shade stress, Cd stress inhibited the synthesis of pyruvate, whether under sunlight or shaded conditions (Figure 4). Therefore, it may be suggested that the decrease of pyruvate is a unique event for purple Perilla suffering from Cd stress. Moreover, the reduction of pyruvate, an important intermediate involved in various metabolic pathways, changed purple Perilla metabolism (Figures 4B,C). Thus, the decrease of pyruvate could be regarded as the metabolite marker for plants under Cd stress. Gil-Monreal et al. (2018) confirmed that pyruvate as a signal participated in the regulation of ethanol fermentation in plants treated with herbicides. To the authors’ knowledge, our study describes for the first time the significance of pyruvate involved in metabolic pathway under Cd stress.

Under exposure to different abiotic stresses (Figure 4), five common metabolic pathways were identified in the top 20 pathways, including the TCA cycle; glyoxylate and dicarboxylate metabolism; the glucagon signaling pathway; phenylalanine, tyrosine, and tryptophan biosynthesis; and carbon metabolism. Moreover, the TCA cycle, glyoxylate and dicarboxylate metabolism, and carbon metabolism are associated with energy metabolism. Zhao et al. (2021) found that cadmium treatment resulted in the upregulation of the TCA cycle, with some intermediates showing significant increases in roots, which may be attributed to enhanced respiration under abiotic stress. Therefore, the energy metabolism is changed, which may be a general phenomenon under different forms of abiotic stress.