The PEG 6000 was purchased from Shanghai Sangon Biotechnology Co., Ltd, Shanghai, China. The plant materials were two Job’s tears cultivars with different drought resistance; yy18-1 is high resistance to drought stress with black seed coat and yy12-7 is susceptible to drought stress with white seed coat, which were approved by the Guizhou Crop Cultivar Approval Committee (Guiyang, Guizhou Province, China) in 2016. Seeds of Job’s tears cultivars were kindly provided by Institute of Upland Food Crops, Guizhou Academy of Agricultural Sciences, Guiyang, China.

The experiments were conducted as completely randomized design with three replications at the Institute of Upland Food Crops, Guizhou Academy of Agricultural Sciences, Guiyang, China (26.53˚N, 106.71˚E). Seeds of three Job’s tears cultivars were surface sterilized with 0.1% (w/v) HgCl₂ for about 8 min and then rinsed five times with sterile distilled water. Next, seed soaking with PEG 6000 solutions with 0 (CK), −0.05 (T1), −0.1 (T2), −0.15 (T3), and −0.2 (T4) MPa osmotic potentials were applied for 24 h before germinated. Osmotic potentials were produced by PEG6000 solutions of 0, 50, 80, 100, and 120 g∙L⁻¹ according to equation supplied by Michel and Kaufmann [12] . Fifty seeds were placed in sterilized plastic trays (24 cm × 24 cm × 12 cm) covered at the bottom with three layers of filter paper that had been autoclaved. The plastic trays were supplied with 5 mL of tested osmotic solutions (PEG 6000 solutions, osmotic potentials of 0, −0.05, −0.1, −0.15, and −0.2 MPa). Germination tests were carried out in a controlled growth chamber (RXZ-1000B, Ningbo Southeast Instrument Co., Ltd, Ningbo, China) at 25 ˚C ± 1˚C with 12 h light (340 μmol∙m⁻²∙s⁻¹) and 12 h dark photo-cycle. The relative humidity was maintained at 85% to prevent high evaporation of the medium. A further 1 mL of tested osmotic solutions was added daily to each plastic tray.

Seeds were considered germinated when the radicle was at least 1 mm length. The number of germinated seeds was recorded daily. The germination energy was calculated as N₄/N × 100% and the germination rate was calculated as N₈/N × 100%, where N₄ is the number of germinated seeds at 4 days, N₈ is the number of germinated seeds at 8 days, and N is the total number of tested seeds. The germination index was calculated as nd₂ + 0.75nd₄ + 0.50nd₆ + 0.25nd₈ according to Qin et al. [13] , where nd₂, nd₄, nd₆, and nd₈ are the germination rate at 2, 4, 6, and 8 days, respectively.

Plants were harvested after 10 days of drought stress treatments and five plants were selected randomly from each plastic tray. The seedling length refers to the distance from the seedling base to the top of seedling. The root length refers to the distance from the taproot base to the top of taproot. The diameters of seedling and root were measured with a digital caliper having accuracy of 0.001 mm. The fresh masses of seedling and root were the fresh weights of seedling and root, respectively. The dry masses of seedling and root were measured after drying in an oven at 80˚C for 24 h.

Five seedlings were collected randomly from each plastic tray after 10 days of drought stress treatments and the fresh mass (FM) was determined, followed by flotation on distilled water for 12 hat room temperature. The hydrated seedling tissues were weighed to determine the turgid mass (TM). The seedling tissues were subsequently dried in an oven at 80˚C for 24 h and weighed to determine the dry masse (DW). The RWC was calculated as [(FM − DM)/(TM − DM)] × 100% according to Smart and Bingham [14] .

H₂O₂ content was assayed according to the method of Gietler et al. [15] . 0.5 g of seedling were collected after 10 days of drought stress treatments and homogenized with a pestle in a mortar, which contained 5 mL of 50 mM phosphate buffer (pH 6.5). The homogenate was centrifuged at 6000× g for 25 min. The supernatant was mixed with 0.1% (w/v) titanium chloride and 20% (v/v) H₂SO₄, and then centrifuged at 6000× g for 15 min. The absorbance of the titanium peroxide complex at 410 nm was measured and H₂O₂ content was calculated using an extinction coefficient of 0.28 µM⁻¹∙cm⁻¹.

MDA content was assayed according to the method of Akcay et al. [6] . 0.5 g of seedling were collected after 10 days of drought stress treatments and homogenized with a pestle in a mortar, which contained 1mLof 5% (w/v) trichloroacetic acid. The homogenate was centrifuged at 12,000× g for 15 min at room temperature. 0.2 mL of supernatant was mixed with 1 mL of 0.5% (w/v) thiobarbituric acid and 20% (w/v) trichloroacetic acid, and incubated in a water bath for 25 min at 96˚C. The reaction was stopped by placing the reaction tubes in an ice bath. Then the samples were centrifuged at 10,000× g for 5 min. Absorbance of supernatant was measured at 532 nm and the correction for non-specific turbidity was performed by subtracting the absorbance at 600 nm. MDA content was calculated using an extinction coefficient of 155 mM⁻¹∙cm⁻¹.

Proline content was assayed by a modification method of Bates et al. [16] . 0.5 g of seedling were collected after 10 days of drought stress treatments and homogenized with liquid nitrogen. The seedling powders were mixed with 1 mL of 3% (w/v) sulphosalicilic acid and centrifuged at 1000× g for 5 min at 4˚C. 0.1 mL of supernatant was mixed with 0.1 mL of 3% (w/v) sulphosalicilic acid, 0.2 mL of 96% (v/v) acetic acid and 0.2 mL acid ninhydrin, and incubated in a water bath at 96 ˚C. After 1 h, the mixtures were mixed with 1 mL toluene and centrifuged at 1000× g for 5 min at 4˚C. Absorbance of supernatant was measured at 520 nm and proline content was calculated using an extinction coefficient of 0.9986 mM⁻¹∙cm⁻¹.

Five seedlings were collected randomly from each plastic tray after 10 days of drought stress treatments and immediately frozen in liquid nitrogen for 30 min, then kept −60˚C for analysis of enzymes activities.

POD activity was assayed according to the method of Wang et al. [17] . 0.5 g of seedling samples were homogenized with a pestle in an ice-cold mortar, which contained 5 mL of 50 mM phosphate buffer (pH 5.5) and 0.2 g PVP. The homogenate was centrifuged at 3000× g for 10 min at 4˚C. 0.1 mL of supernatant was mixed with 1 mL of 50 mM guaiacol, 2.9 mL of 50 mM phosphate buffer (pH 5.5), and 1 mL of 2% (v/v) H₂O₂, and incubated in a water bath at 37˚C. After 5 min, the absorbance at 470 nm was measured by a spectrophotometer. One unit of enzyme activity (U) was defined as a change of 0.01 in absorbance per h.

CAT activity was assayed by a modification method of Gong et al. [18] . 0.5 g of seedling samples were homogenized with a pestle in an ice-cold mortar, which contained 1 mL 50 mM Tris-HCl (pH 7.8). The homogenate was centrifuged at 10,000× g for 10 min at 4˚C. 0.2 mL of homogenate was mixed with 3 mL of reaction mixture (50 mM potassium phosphate buffer, pH 7.7, 10 mM H₂O₂). After 30 min, the reaction was stopped by adding 20% (v/v) H₂SO₄ and the residual hydrogen peroxide was titrated against potassium permanganate. Absorbance at 240 nm was measured by a spectrophotometer. One unit of enzyme activity (U) was defined as a change of 0.01 in absorbance per h.

APX activity was assayed by a modification method of Nakano et al. [19] . 0.5 g of seedling samples were homogenized with a pestle in an ice-cold mortar, which contained 5 mL of buffer (50 mM potassium phosphate buffer, pH 7.0, 2 mM ascorbate, 5 mM EDTA, 0.2 g PVP). The homogenate was centrifuged at 16,000× g for 20 min at 4˚C. 30 µL of supernatant was mixed with 0.97 mL of 0.1 mM H₂O₂ and the absorbance at 290 nm was measured. One unit of enzyme activity (U) was defined as a change of 0.01 in absorbance per h.

GR activity was assayed according to the method of Akcay et al. [6] , with some modifications. 0.5 g of seedling samples were homogenized with a pestle in an ice-cold mortar, which contained 5 mL of buffer (100 mM potassium phosphate buffer, pH 7.8, 1% (w/v) PVP, 50% (v/v) Triton X-100, 0.1 mM EDTA). The homogenate was centrifuged at 10,000× g for 30 min at 4˚C. 0.2 mL of supernatant was mixed with 3 mL of reaction mixture (200 mM potassium phosphate buffer, pH 7.5, 1.5 mM MgCl₂, 0.2 mM EDTA, 25 µM NADPH, 0.25 mM GSSH). The absorbance at 340 nm was measured was measured. One unit of enzyme activity (U) was defined as a change of 0.01 in absorbance per h.

Analysis of variance was performed with SPSS 19.0 software (Statistical Package for Social Sciences, SPSS Institute Inc., Illinois), and data from each sampling data were analyzed separately. Means were tested by least significant difference at the 5% probability level (LSD_0.05).

Germination energy and germination rate of both cultivars were affected considerably by PEG treatments, and yy18-1 showed higher germination energy and germination rate than yy12-7 in the same PEG treatment (Table 1). It was observed that in all of cultivars there were decreases in germination energy and germination rate due to drought stress increment. Germination energy and germination rate of both cultivars significantly decreased when they were subjected to PEG6000 solution with below −0.1 MPa osmotic potential, but no significant changes were observed at the osmotic potential of −0.05 MPa. Retardations of germination energy and germination rate were clearer in yy12-7 when compared to yy18-1.

CK, T1, T2, T3, and T4 are polyethylene glycol 6000 solutions with 0, −0.05, −0.1, −0.15, and −0.2 MPa osmotic potentials, respectively. Data are presented as the mean ± standard error of three replicates. Values followed by different letters within a column are significantly different at the 5% probability level.

Germination index decreased with the decrease of the osmotic potential of PEG 6000 solution in both cultivars, and yy18-1 showed higher germination index than yy12-7 in the same PEG treatment (Table 1). Besides, there were significant differences between control and drought stress treatments except for the osmotic potential of −0.05 MPa for germination index in both cultivars. Compared with the control, the T1, T2, T3, and T4 treatments decreased the germination index by 6.77%, 30.66%, 45.67%, and 51.37%, respectively for yy18-1, and decreased that by 9.25%, 31.34%, 49.25%, and 60.30%, respectively for yy12-7.

Root length and root diameter of yy18-1 were higher than those of yy12-7 in the same PEG treatment, which showed decrease trends with the decrease of the osmotic potential of PEG 6000 solution in both cultivars (Table 2). There were significant differences between control and drought stress treatments except for the osmotic potential of −0.05 MPa for root length and root diameter in both cultivars. In addition, at the same drought stress treatment, compared with the control treatment, the decreasing amplitudes of root length and root diameter of yy12-7 were higher than those of yy18-1.

Root fresh mass and root dry mass of both cultivars were affected considerably by PEG 6000 solution treatments, and yy18-1 showed higher root fresh mass and root dry mass than yy12-7 in the same PEG treatment (Table 2). It was observed that in all of cultivars there were decreases in root fresh mass and root dry mass due to drought stress increment. Root fresh mass and root dry mass of both cultivars significantly decreased when they were subjected to PEG 6000 solution with below −0.1 MPa osmotic potential, but no significant changes were

CK, T1, T2, T3, and T4 are polyethylene glycol 6000 solutions with 0, −0.05, −0.1, −0.15, and −0.2 MPa osmotic potentials, respectively. Data are presented as the mean ± standard error of three replicates. Values followed by different letters within a column are significantly different at the 5% probability level.

observed at the osmotic potential of −0.05 MPa. Retardations of root fresh mass and root dry mass were clearer in yy12-7 when compared to yy18-1.

Seedling length and seedling diameter decreased with the decrease of the osmotic potential of PEG 6000 solution in both cultivars, and yy18-1 showed higher seedling diameter and seedling diameter than yy12-7 in the same PEG treatment (Table 3). There were significant differences between control and drought stress treatments except for the osmotic potential of −0.05 MPa for seedling diameter and seedling diameter in both cultivars. Besides, at the same drought stress treatment, compared with the control treatment, the decreasing amplitudes of seedling length and seedling diameter of yy12-7 were higher than those of yy18-1.

Seedling fresh mass and seedling dry mass of yy18-1 was higher than those of yy12-7 in the same PEG treatment, which showed decrease trends with the decrease of the osmotic potential of PEG 6000 solution in both cultivars (Table 3). There were significant differences between control and drought stress treatments except for the osmotic potential of −0.05 MPa for seedling fresh mass and seedling dry mass in both cultivars. In addition, at the same drought stress treatment, compared with the control treatment, the decreasing amplitudes of seedling fresh mass and seedling dry mass of yy12-7 were higher than those of yy18-1.

Under no PEG treatment, both cultivars showed almost identical RWC (91.87% in yy18-1 and 90.32% in yy12-7). However, drought stress caused the deleterious effects in RWC of both yy18-1 and yy12-7 (Table 4). RWC of both cultivars significantly decreased when it was subjected to PEG 6000 solution with below −0.1

CK, T1, T2, T3, and T4 are polyethylene glycol 6000 solutions with 0, −0.05, −0.1, −0.15, and −0.2 MPa osmotic potentials, respectively. Data are presented as the mean ± standard error of three replicates. Values followed by different letters within a column are significantly different at the 5% probability level.

CK, T1, T2, T3, and T4 are polyethylene glycol 6000 solutions with 0, −0.05, −0.1, −0.15, and −0.2 MPa osmotic potentials, respectively. RWC, H₂O₂, and MAD stands for relative water content, hydrogen peroxide, and malondialdehyde, respectively. Data are presented as the mean ± standard error of three replicates. Values followed by different letters within a column are significantly different at the 5% probability level.

MPa osmotic potential, but no significant change was observed at the osmotic potential of −0.05 MPa. Besides, at the same drought stress treatment, compared with the control treatment, the decreasing amplitude of RWC of yy18-1 was lower than that of yy12-7.

Similarly, H₂O₂ content of both cultivars in the no PEG treatment was almost identical and independent of their drought resistance (Table 4). It was observed that in all of cultivars there was an increase in H₂O₂ content due to drought stress increment. There were significant differences between control and drought stress treatments except for the osmotic potential of −0.05 MPa for H₂O₂ content in both cultivars. Compared with the control, the T1, T2, T3, and T4 treatments increased the H₂O₂ content by 5.64%, 19.97%, 26.98%, and 40.85%, respectively for yy18-1, and increased that by 21.12%, 42.08%, 53.09%, and 88.08%, respectively for yy12-7.

MDA content was affected considerably by PEG 6000 solution treatments, and yy18-1 showed lower MDA content than yy12-7 in the same PEG treatment (Table 4). It was observed that in all of cultivars there was an increase in MDA content due to drought stress increment. MDA content of both cultivars significantly increased when it was subjected to PEG 6000 solution with below −0.1 MPa osmotic potential, but no significant change was observed at the osmotic potential of −0.05 MPa. Compared with the control, the T1, T2, T3, and T4 treatments increased the MDA content by 37.52%, 73.44%, 92.25%, and 122.97%, respectively for yy18-1, and increased that by 39.35%, 80.85%, 134.45%, and 195.03%, respectively for yy12-7.

Proline content of yy18-1 was higher than that of yy12-7 in the same PEG treatment, which showed an increase trend with the decrease of the osmotic potential of PEG 6000 solution in both cultivars (Table 4). There were significant differences between control and drought stress treatments except for the osmotic potential of −0.05 MPa for proline content in both cultivars. In addition, the increasing amplitude of proline content was clearer in yy18-1 when compared to yy12-7.

POD activity was affected considerably by PEG 6000 solution treatments, and yy18-1 showed higher POD activity than yy12-7 in the same PEG treatment (Table 5). POD activity increased and then decreased with the decrease of osmotic potential of PEG 6000 solution in both cultivars and the maximum value appeared in the −0.1 MPa osmotic potential treatment. Compared with the control, the T1 and T2 treatments increased the POD activity by 3.55% and 12.48% (mean of both cultivars), but the T3 and T4 treatments decreased that by 4.87% and 19.13% (mean of both cultivars), respectively.

CAT activity of yy18-1 was higher than that of yy12-7 in the same PEG treatment, which exhibited single peak curves with a peak at the −0.1 MPa osmotic potential treatment (Table 5). Compared with the control, the T1 and T2 treatments increased the CAT activity by 1.17% and 5.72% (mean of both cultivars), but the T3 and T4 treatments decreased that by 2.69% and 10.61% (mean of both cultivars), respectively.

APX activity was affected considerably by PEG 6000 solution treatments, and yy18-1 showed higher APX activity than yy12-7 in the same PEG treatment

CK, T1, T2, T3, and T4 are polyethylene glycol 6000 solutions with 0, −0.05, −0.1, −0.15, and −0.2 MPa osmotic potentials, respectively. POD, CAT, APX, and GR stands for peroxidase, catalase, ascorbate peroxidase, and glutathione reductase, respectively. Data are presented as the mean ± standard error of three replicates. Values followed by different letters within a column are significantly different at the 5% probability level.

(Table 5). It was observed that in all of cultivars there was an increase and then decrease in APX activity due to drought stress increment and the maximum value appeared in the −0.1 MPa osmotic potential treatment. Compared with the control, the T1 and T2 treatments increased the APX activity by 1.55% and 6.41% (mean of both cultivars), but the T3 and T4 treatments decreased that by 5.37% and 9.34% (mean of both cultivars), respectively.

GR activity of yy18-1 was higher than that of yy12-7 in the same PEG treatment, which showed an increase and then decrease trend with the decrease of the osmotic potential of PEG 6000 solution in both cultivars and the maximum value appeared in the −0.1 MPa osmotic potential treatment (Table 5). Compared with the control, the T1 and T2 treatments increased the GR activity by 1.72% and 12.10% (mean of both cultivars), but the T3 and T4 treatments decreased that by 8.35% and 13.61% (mean of both cultivars), respectively.