Heavy metals such as cadmium (Cd) are considered environmental pollutants that threaten all forms of organisms (Sytar et al., 2013; Yan et al., 2016) and are available naturally in the environment or through anthropic sources (Sabiha-Javied et al., 2008; Wuana & Okieimen, 2011). Many industrial activities are recognized as important sources of Cd contamination in agricultural soils (Bermudez et al., 2010; Salazar et al., 2012). Cultivated areas contaminated by Cd directly impacts the growth and seeds quality of large crops (Vollmann et al., 2015). This happens because heavy metals are persistent in the environment and may be transported through soils (Zhou et al., 2013). Plants accumulating heavy metals from soils in which they grow may either store these metals in root or translocate them to the biomass aboveground, entering in the food chain (Zhou et al., 2013; Vollmann et al., 2015).

In plants, the presence of Cd influences several physiological and biochemical processes. Excessive Cd levels results in an overproduction of reactive oxygen species (ROS) such as superoxide (O₂), hydrogen peroxide (H₂O₂), hydroxyl radicals (^·OH) and singlet oxygen (O₂ ^·). The overproduction of ROS can lead to lipid peroxidation, macromolecule deterioration, ions leakage, membrane degradation, and DNA chain cleavage (Trivedi & Ansari, 2015).

ROS in excessive concentrations are considered dangerous molecules for plant metabolism. However, it has been shown that ROS can also play an important role in plant defence against pathogens and the signalling of molecules, which contribute to the control of plant development and the perception of the external environment (Smirnoff, 2005). A complex enzymatic network of antioxidant molecules controls the concentration of ROS and the repair of oxidative damages. Antioxidant enzymes such as superoxide dismutase (SOD), catalase (CAT), ascorbate peroxidase (APX), glutathione peroxidase and glutathione reductase play an important role in reducing the production of ROS and protecting the cell from damage (Prasad & Strzalka, 2002; Smirnoff, 2005; Emamverdian et al., 2015).

Studies have shown that ROS production is directly related to the action of plant growth regulators (PGR), such as the hormone gibberellin, auxin and cytokinin (Xu et al., 2006; Krishnamurthy & Rathinasabapathi, 2013; Yan et al., 2016). PGR are well recognized for regulating processes related to growth and development, and their levels are rapidly altered by environmental fluctuation. In addition, the action of growth regulators has been questioned of being the connecting link regulating the level of ROS and its role in damage or in triggering signalling pathways in plants under stress (Krishnamurthy & Rathinasabapathi, 2013). The production of ROS under stress conditions may influence biosynthesis, transport, metabolism and signalling of auxin (Tognetti et al., 2012). Abscisic acid (ABA) may increase the activation of SOD and CAT enzymes (Jiang & Zhang, 2001; Agarwal et al., 2005) and inhibit the transport and translocation of Cd (Hsu & Kao, 2003). Further, ABA may act synergistically with cytokinin on the physiological regulation in plants under stress caused by Cd (Ghanem et al., 2008).

Soybean (Glycine max (L.) Merrill) is one of the most important protein crops in the world for the production of oil and food. The high demand of soybeans has resulted in this crop being cultivated in the vicinity of routes with heavy vehicular traffic and areas close to human activity such as industrial centres (Rodriguez et al., 2011). The soybeans cultivation in polluted areas and in soils contaminated with heavy metals can result in plants with alterations in growth, development and yield (Hasan et al., 2009; Piršelová et al., 2011), which can impact grain yield and quality and consequently, human health (Alam et al., 2003).

The exogenous application of natural and synthetic regulators is known for contributing to the development of plants under a variety of abiotic stresses (Alonso-Ramirez et al., 2009; Farooq et al., 2017). New alternatives which seek to increase crop tolerance to stresses caused by the presence of heavy metals need to be studied, in order to find workable solutions for productivity in contaminated soils. Therefore, the aim of this study was to evaluate the effects of different concentrations of Cd on the growth and physiology of soybean seedlings, and to verify the potential attenuating effect of oxidative stress by the exogenous application of PGR. It was hypothesized that seedling growth and antioxidant activity stressed by high concentrations of Cd are attenuated by the application of PGR.

Seeds of soybean cultivar RK 7214 IPROˈ were sterilized in 2% sodium hypochlorite solution for 5 min and then washed in distilled water 5 times. After that, the seeds were soaked for 12 h at room temperature, in distilled water (control treatment) and in solution with plant growth regulator (PGR treatment) in dose recommended by the manufacturer (5 mL L^-1). The pre-treatments followed the method of Liu TT et al. (2011) and Liu L et al. (2018). The PGR used in this study was a commercial product composed of three plant regulators in the following concentration: 0.005% of indolebutyric acid (IBA, auxin analogue), 0.009% kinetin (cytokinin) and 0.005% gibberellic acid (GA, gibberellin). After 12 h, the seeds were germinated in two layers of rolled filter paper moistened with distillated water and placed in transparent plastic bags. The plastic bags were conditioned in chamber type Mangelsdorf with temperature controlled at 25 °C in the dark, and left to grow until the radicle reached 3-8 mm in length (Piršelová et al., 2011). Uniformly germinated seeds were selected from control and PGR treatments, collected and transferred to in acrylic boxes (Gerbox®) with fresh Germitest® paper moistened with four concentrations of CdCl₂.H₂O (0, 100, 500 and 900 mg L^-1). These concentrations were previously defined based on dose-response curve plotted from pilot experiments, conducted prior to the study, using a wide range of Cd levels (5 mg L^-1 to 1 g L^-1) (Neumann et al., 1998; Kuriakose & Prasad, 2008). The seedlings were grown for seven days in chamber type Mangelsdorf with temperature controlled at 25 °C in the day/night, photoperiod of 12/12 h and white light of 670 lux. After that the growth and biochemical parameters of the seedlings were analysed.

The experimental design was completely randomized composed of 4 doses of Cd (0, 100, 500 and 900 mg L^-1) and two treatments, one with distilled water (control) and another with application of PGR.

The results indicated that the dry mass of the seedlings, both shoot and roots, were not significantly changed by different Cd levels and PGR treatment (Fig. 1). Considering the root morphology, the seedlings showed a high sensitivity to the Cd concentrations and PGR treatment. In the control treatment, the length, volume and area of the roots, when the seedlings were under the dose of 100 mg L^-1, tended to increase compared to 0 mg L^-1. However, at dose of 900 mg L^-1, these same parameters were significantly reduced (p<0.05) when compared to the dose of 100 mg L^-1 (61%, 58% and 64% reduction, respectively, for root length, volume and area). Under the PGR treatment, the length, volume and area of roots decreased at dose of 900 mg L^-1 compared to the dose of 100 mg L^-1 (64%, 67% and 74% of reduction, respectively).

Without the application of Cd (dose 0 mg L^-1) the root length increased 54% in PGR, compared to the control (Fig. 1) while the roots volume and area were not affected by the same treatment (p>0.05). However, under the Cd dose of 500 mg L^-1 the values of length, volume and root area decreased significantly (p<0.05) in the presence of the bioregulator compared to the control treatment (71%, 64% and 67% reduction, respectively).

The lipid peroxidation of the seedlings expressed by the MDA content under control treatment was increased in the shoot at the dose of 500 mg L^-1 compared to the dose 0 mg L^-1, the roots however, showed no significant changes of MDA content to increase of Cd treatment (p>0.05). On the other hand, in the PGR treatment, the MDA in the shoot increased progressively as the level of Cd enhanced, while the roots, increased significantly (p<0.05) in the doses of 100 mg L^-1 and 500 mg L^-1 compared to the dose 0 mg L^-1.

Regarding the response of seedlings under control and PGR treatment, the MDA content in the shoot increased by PGR treatment when compared to the control, especially in the Cd doses of 100 and 900 mg L^-1 (Fig. 2). The MDA in the roots was increased by PGR when compared to the control treatment, but this increase was only significant at doses 100 and 500 mg L^-1 (p<0.05).

The proline content was increased as the level of Cd enhanced (p<0.05) in both treatments (control and PGR) (Fig. 2). Regardless of the Cd doses tested, the proline was significantly higher under PGR treatment compared to control (p<0.05).

It was observed that the increase in the levels of Cd under the control treatment did not influence the SOD enzyme activity in the shoot. However, under the PGR application, the enzyme activity was reduced concomitantly to the increase of Cd level. In the roots under control treatment, the activity of SOD was significantly reduced as the Cd doses increased, whereas under PGR, the activity of this enzyme did not show significant changes as the level of Cd increased (p>0.05).

Compared to the control treatment, the PGR did not influence significantly the activity of SOD in the shoot of the seedlings (p>0.05) in all concentrations of Cd tested (Fig. 3). On the other hand, the SOD activity in the roots was reduced by the PGR in 2.8 times in the dose 100 mg L^-1 of Cd, whereas in the dose of 900 mg L^-1 it increased 2.2 times when compared to the control treatment.

Compared to the control treatment, the application of PGR significantly reduced (p<0.05) the activity of CAT in the roots of seedlings under the doses 0 and 100 mg L^-1 of Cd, and in the shoot of the seedlings under the 0 mg L^-1 dose (Fig. 3). On the other hand, at the dose of 100 mg L^-1 the PGR increased the CAT activity in the shoot (62% higher compared to the control treatment) whereas, in the dose 500 mg L^-1 the enzyme activity decreased 83% in this same treatment.

Under the control treatment, the dose of 900 mg L^-1 reduced by 39% the CAT activity in the shoot compared to the dose of 0 mg L^-1, while the CAT in the roots was reduced by 85% in the dose of 500 mg L^-1 compared to the 0 mg L^-1 dose (Fig. 3). In the PGR treatment, the activity of this enzyme was reduced by the dose of 500 mg L^-1 in the shoot in relation to 0 mg L^-1, while in the roots, the CAT activity was reduced in doses ≥100 mg L^-1.

The activity of APX in the shoot was reduced by PGR, compared to the control treatment in all levels of Cd tested (Fig. 3). The roots showed a reduction of approximately 90% in the activity of APX under PGR when compared to the control at dose 0 mg L^-1.

Considering the influence of the Cd levels on the APX activity of the shoot, it was observed that in both treatments (control and PGR) there was a significant decrease in the enzymatic activity in doses ≥100 mg L^-1, whereas in the root, the enzymatic activity was significantly reduced (p<0.05) in the control treatment. Under the PGR the APX activity showed no alteration in all the concentrations evaluated (p>0.05).

Principal component analysis (PCA) showed different responses in seedlings when evaluating the Cd levels and treatments (control and PGR). The PCA was performed considering all morphological and biochemistry parameters represented 58.3% of the total variation of the original data (Fig. 4). The component one (PC1) explained 41.5% of the variance, while the component two (PC2) accounted for 16.7%. The variables that contributed most to distinguish the groups in PC1 were proline, root length and APX activity in roots (eigenvector value >0.81), whereas for PC2 they were MDA content in roots, root area and CAT activity in roots (eigenvector values>0.55). The spatial ordination of the replicates allowed the visualization of a group composed by the control treatment seedlings at the dose 0 mg L^-1 in the top-right quadrant and another group composed by PGR seedlings at doses 500 and 900 mg L^-1 of Cd in the left quadrant. In the PC2 the replicates clustered in the bottom quadrant representing the PGR treatment at the dose Cd of 100 mg L^-1. Seedlings under control treatment at Cd doses of 100 mg L^-1, 500 mg L^-1 and 900 mg L^-1 and the seedlings of the PGR treatment at Cd doses of 0 mg L^-1 were overlaid on the graph.

The exposure of soybean plants to Cd was documented in previous studies. These responses may include decreases in plant length and biomass, increases in lipid peroxidation, leaf proline content, activity of antioxidant enzymes hydrogen peroxide and nitric oxide (Pérez-Chaca et al., 2014; Nasser Alyemeni et al., 2017; Marques et al., 2019). Nasser Alyemeni et al. (2017) showed that the biomass of soybean plants was reduced when grown under concentrations below 150 mg L^-1 of Cd supplied to pots with sand and ver-micompost. However, our results reveal that dry mass of seedling was not influenced by doses above 100 mg L^-1 and by the PGR. Heavy metal-plant interactions are complex and depend on many factors. Among them are the cultivar sensibility, the stage of plant development, the chemistry of the metal, including its concentration (Prasad & Strzalka, 2002; Liu L et al., 2018). The lack of biomass decreases on our study suggests that soybean in seedling stage is tolerant to the presence of Cd, but it is important take into account that the growing medium (gerbox) may have been another factor influencing this biomass response.

Although seedling dry mass was not reduced by exposure to Cd, an increase tendency in root length, volume and area at the Cd dose of 100 mg L^-1 compared to the 0 mg L^-1 was observed in the control treatment. When the seedlings were exposed to 900 mg L^-1 these same parameters were reduced compared to 100 mg L^-1. Other studies have documented this phenomenon, named as “hormesis”, defined as low dose stimulation and high dose inhibition, and it has been observed in several organisms (Kumar & Prasad, 2004; Jain et al., 2007; Cornu et al., 2016) including soybean (Liu et al., 2018). In the present work, the hormesis phenomenon also occur in seedlings under PGR treatment, however the inhibition by the Cd occurred in 500 mg L^-1.

The application of PGR increased the development of the roots when they were not exposed to the Cd (0 mg L^-1). The use of plant growth regulators (as used in our study) were documented for increasing the biomass accumulation and the root system of tomato plants (Cato et al., 2013), cotton (Vieira & Santos, 2005) and the productivity of soybeans (Albrecht et al., 2012). However, when seedlings of our study were exposed to the 500 mg L^-1of Cd, the root development was decreased compared to the control, suggesting that the physiological responses to the PGR and the oxidative stress caused by heavy metal were synergistic.

Often photosynthetic capacity can be reduced under stressful conditions, which results in an increase in the electron transport rate of photosystems to O₂ (Souza et al., 2004; Yan et al., 2012). This response, in turn, may triggers the increase in the production of ROS resulting to lipid peroxidation and an enhance in the lipid peroxidation products, as MDA (Zayneb et al.; 2015; Yan et al., 2016). MDA has cytotoxic action and when found at high levels indicate an increase in oxidative stress at the cellular level. Our results indicated that the increment in Cd levels increased the lipid peroxidation in the shoot until the dose of 500 mg L^-1 in both treatments (control and PGR) (Fig. 2). However, under the 900 mg L^-1 dose, seedlings in the control treatment decreased the MDA content, whereas in PGR treatment this same parameter increased. Additionally, it was observed that the MDA content in both shoots and roots was higher in the PGR treatment. These results indicate that lipid peroxidation, besides being increased by the increment in Cd levels, was also potentiated by the PGR.

Plants are able to develop different survival strategies when exposed to heavy metals (Monni et al., 2001a, b). They may be protected against these elements by form a mutualistic symbiosis with mycorrizal fungi that immobilize the metal in the root system (Rask et al., 2019), or can tolerate high concentrations through specific physiological mechanisms such as the antioxidant enzymes SOD, CAT and APX (Prasad & Strzalka, 2002; Emamverdian et al., 2015). Among these enzymes, SOD and APX play an important role in the process of cleaning superoxide radicals. However, SOD activity has been recognized to be more important in the regulation of plant cell function on oxidative stress (Nasser Alyemeni et al., 2017). Our results showed that under the control treatment, the SOD activity in the roots was inhibited by the exposure to Cd, which did not happen in the shoots, once the reduction of the activity in this organ was observed with the increase of Cd (Fig. 3). On the other hand, under PGR the shoots showed a reduction of the SOD activity with the increment of the Cd doses, whereas in the roots, the activity of the enzyme was not influenced by the metal. These results indicate that plant growth regulators influenced the organ (shoot and root) and the level of SOD activity in response to the Cd levels tested. The reduction of SOD activity in response to the increase of Cd has already been documented in pine roots (Schützendubel et al., 2002) and common beans (Somashekaraiah et al., 1992).

The reduction in CAT and APX activities on shoots with the increasing of Cd levels (900 mg L^-1in control treatment and ≥500 mg L^-1 in PGR), and the lower values of CAT and APX of roots under the treatment with the PGR, suggest that the plant growth regulator acted antagonistically to the antioxidant activity. These results corroborate the negative influences of the plant growth regulator on the responses to oxidative stress caused by heavy metal as described above in the root morphology and MDA content. The CAT and APX activity were reported to reduce in wheat (Milone et al., 2003) and Brassica juncea (Nouairi et al., 2006) treated with Cd.

Previous studies have shown that exogenous application of plant growth hormones (also called as phytohormones or plant hormones) can improve protection against heavy metal toxicity (Masood et al., 2012; Zhu et al., 2012; Agami & Mohamed, 2013; Khan et al., 2018). Exogenous applications of both natural and synthetic auxins under controlled concentrations show enhanced antioxidant system activity (Piotrowska-Niczyporuk & Bajguz, 2014). Some recent approaches showed that interaction between heavy metal and auxin can be used as a protective mechanism against toxicity in crop plants or as a useful tool in phytoremediation programs for detoxification of polluted areas (Bucker-Neto et al., 2017). Tandon et al. (2015) showed that exogenous application of natural and synthetic auxins increased phytoremediation efficiency in waste-water treatment. Pandey & Gupta (2015) demonstrated that exogenous application of auxin in rice seedlings reduced the stress caused by selenium (Se) over the morphological and biochemical characteristics. The toxicity to Cd manifested as a negative effect on seed germination and seedling growth is reported to be relieved by exogenous application of cytokinin which results in an increase of antioxidant enzymes activity (Gangwar et al., 2010, 2014; Khan et al., 2018). On the other hand, exogenous application of GA in low concentrations counteracts some of the adverse effects of chromium phytotoxicity with the increased levels of antioxidants, whereas high concentrations of GA showed apparently reverse effects under Cr phytotoxicity (Gangwar et al., 2011). These studies reveal that the exogenous application of plant hormones to prevent growth inhibition and increase heavy metal tolerance rely on several processes, and are still poorly understood, especially when these substances are applied concomitantly.

One possible explanation for the variations in antioxidants enzymes activity, in response to Cd increment in our study, may be related to the exposure time of the seedlings to the metal. Studies have shown that antioxidant activity may play an effective role in short-term stress, but not in the long term (Dazy et al., 2008; Yan et al., 2016). Stroiński & Kozłowska (1997) observed that SOD, CAT and APX activity can be reduced after one hour of exposure to Cd. The decreases observed at high Cd concentrations can be interpreted as a cytotoxicity sign due to overproduction of ROS (Dazy et al., 2008). Another possible explanation is the excess of superoxide radicals that disturb the signal transduction and triggers the genes related to the production of antioxidant enzymes (Lamb & Dixon, 1997). On the other hand, Romero-Puertas et al. (2004) showed that the CAT transcripts may be inducted by Cd while de CAT activity is decreased. They assumed that Cd induce posttranslational modifications of CAT.

A third possible explanation for the antioxidant activity reduction is the binding of heavy metal ions to the enzyme active center. The Cd may can substitute Fe ions in active site of SOD and CAT decreasing the enzyme activity (Stroiński & Kozłowska, 1997; Štolfa et al., 2015). In addition, Cd probably interact with the thiol group in SOD structure, inactivating them (Shah & Nongkynrih, 2007; Štolfa et al., 2015; Nasser Alyemeni et al., 2017). The limited antioxidant activity caused by different concentrations of Cd has been observed in Petroselinum crispum seedlings (Ulusu et al., 2017), wheat (Dey et al., 2007) and roots of Pinus sylvestris (Schützendubel et al., 2002).

Once the enzymatic antioxidants can be activated under short-term Cd stress, non-enzymatic processes may have a long-term effect in scavenging for ROS (He et al., 2011). The increase of compatible organic osmolytes in the cytosol has been recognized as a tolerance response of plants to environmental stresses by mediating the maintenance of the cell water content (Tester, 2003). High proline content plays a key role in osmotic adjustment, cellular sub-structures stabilization (e.g., membranes and proteins), scavenging free radicals and buffering cellular redox potential (Kavi Kishor et al., 2005; Hayat et al., 2012). As a response to stress and oxidative damage, the proline content in our study was increased as Cd levels enhanced which, in turn, may have contributed to reduce the lipid peroxidation of shoot under control treatment and root under PGR treatment. The accumulation of proline in plants tolerant to high doses of Cd has already been documented for other species including soybean (He et al., 2011; Nasser Alyemeni et al., 2017). However, the role of free proline in the maintenance of metabolic activities of soybean seedlings under Cd stress must be further investigated once this amino acid is associated with the ROS scavenger and as a heavy metal chelator (Hayat et al., 2012).

Principal components analysis performed using all growth parameters reinforces the synergistic effect of the bioregulator and oxidative stress caused by Cd. The PCA indicated that PGA treatment in seedlings which were not exposed to Cd and exposed to the minimum dose of Cd (100 mg L^-1) promoted responses similar to seedlings of intermediate and maximum levels of Cd (500 mg L^-1 and 900 mg L^-1) in the control treatment. While the seedlings under PGR treatment in intermediate and maximum doses of Cd were those that showed the highest levels of toxicity to this heavy metal.

Bioregulators play an important role in regulating the physiology of stress. However, plant tolerance to Cd appears to be regulated not only by a single hormone, but by the complex combination of multiple plant hormones (Yan et al., 2016). In addition, the components of the bioregulator tested in this study (IBA, gibberellin and cytokinin) may have interacted with each other, and also with other endogenous hormones produced by the seedlings. Our results indicated that the combination of these substances may have affected synergistically the oxidative stress caused by Cd exposure, although proline levels were increased with the PGR treatment. Signaling of plant growth regulators to stress responses is possibly linked to redox reactions (Jiang & Zhang, 2001), but the relationship of these substances with ROS and antioxidant enzymes in signal-transduction cascades is unclear.

This study showed that exposure of Cd at different concentrations influenced the morphological and physiological parameters of soybean seedlings. Although some of the parameters have shown “hormesis” responses to Cd increase (antioxidant activity and root morphology parameters), the results indicated that the increase in proline content may have played a key role in the maintenance of metabolic activities and maintenance of biomass. Variations in root development (length, volume and area) indicated a structural reorganization of the tissues according to the concentration of Cd. In addition, the results showed that the use of the bioregulator behaved synergistically to the toxicity responses caused by the Cd doses, invalidating our hypothesis. Stress indicators such as MDA content and antioxidant activity in different organs (root and shoot) of soybean seedlings may present a high complexity regarding the use of products based on plant growth hormones in environments contaminated with Cd.