Response of soybean to soil waterlogging associated with iron excess in the reproductive stage

Abstract

Soil waterlogging is a common problem in some agricultural areas, including regions under soybean (Glycine max) cultivation. In waterlogged soils, soil O₂ depletion occurs due to aerobic microorganisms and plants, affecting the metabolic and physiological processes of plants after suffering anoxia in their root tissue. Another harmful factor in this situation is the exponential increase in the availability of iron (Fe) in the soil, which may result in absorption of excess Fe. The present study sought to evaluate the response mechanisms in soybean leaves ‘Agroeste 3680’ by physiological and biochemical analyses associating them with the development of pods in non-waterlogged and waterlogged soil, combined with one moderate and two toxic levels of Fe. Gas exchange was strongly affected by soil waterlogging. Excess Fe without soil waterlogging reduced photosynthetic pigments, and potentiated this reduction when associated with soil waterlogging. Starch and ureide accumulation in the first newly expanded trifoliate leaves proved to be response mechanisms induced by soil waterlogging and excess Fe, since plants cultivated under soil non-waterlogged soil at 25 mg dm⁻³ Fe showed lower contents when compared to stressed plants. Thus, starch and ureide accumulation could be considered efficient biomarkers of phytotoxicity caused by soil waterlogging and excess Fe in soybean plants. The reproductive development was abruptly interrupted by the imposition of stresses, leading to a loss of pod dry biomass, which was largely due to the substantial decrease in the net photosynthetic rate, as expressed by area (A), the blockage of carbohydrate transport to sink tissues and an increase of malondialdehyde (MDA). The negative effect on reproductive development was more pronounced under waterlogged soil.

Electronic supplementary material

The online version of this article (10.1007/s12298-020-00845-8) contains supplementary material, which is available to authorized users.

Introduction

Soybean is responsible for providing more than half of the world’s vegetable oil and two-thirds of it as protein meal, making it the most important grain legume crop (Divito et al. 2015). However, several regions under soybean cultivation are affected by soil waterlogging (Maekawa et al. 2011; Syed et al. 2015; Sartori et al. 2015), leading to a decrease in grain yield and economic losses (Linkemer et al. 1998; Yin and Komatsu 2017). According to Bailey-Serres et al. (2012) and Hirabayashi et al. (2013), climate change will increase the frequency and severity of this stressor, due to high rainfall and poor soil drainage, which may consequently negatively affect the cultivation of rain-fed plants in various regions of the world.

Excessive water affects the physical structure of soil (Rodríguez-Gamir et al. 2011). In addition, O₂ dissolved in the water-filled pore spaces has a low diffusion rate compared to air (Dat et al. 2004; Yin and Komatsu 2017). Consequently, one of the main effects of soil waterlogging is the scarcity of O₂, which is increased over time as a result of O₂ depletion by aerobic microorganisms and plants, leading to an anoxic condition (Mohanty 2003).

The root tissue of plants under anoxia switches the pathway of energy production from oxidative phosphorylation to glycolysis and fermentation (Mohanty 2003), resulting in a severe decline in the energy status of root cells. The overall yield of adenosine triphosphate (ATP) produced during fermentation is only two molecules of ATP per glucose molecule, rather than the 38 molecules produced during oxidative phosphorylation (Sairam et al. 2009). Martínez-Alcántara et al. (2012) determined that the metabolic and physiological effects of soil waterlogging on plants are mainly derived from this energy change experienced by plants. According to these authors, initially, there is a reduction in water flux from roots, followed by alterations in water use, nutrient uptake and dry weight partitioning. In addition, the transport of ureides via the xylem is inhibited in soybean plants (Puiatti and Sodek 1999), resulting from lower N₂-fixation activity. At the leaf level, one of the first detectable effects is decreased CO₂ availability due to stomatal limitation, which is considered the initial negative effect of waterlogging on photosynthesis, and a process instigated to prevent water loss to the environment (Zhang et al. 2016; Yan et al. 2018).

Another key point to consider is that the microorganisms present in anoxic soils are anaerobes and facultative anaerobes. These microorganisms utilise alternative electron acceptors to maintain their metabolism (Maranguit et al. 2017). In this context, Fe-rich soils under anaerobic conditions are made phytotoxic to plants when insoluble Fe³⁺ oxides are reduced to soluble Fe²⁺ forms and released into the water present in the soil (Frei et al. 2016; Maranguit et al. 2017). Excess Fe in plant tissue can displace the cell redox balance towards a pro-oxidant state, affecting different morphological, physiological, and biochemical aspects in plants, generating oxidative stress (De Oliveira Jucoski et al. 2013; Müller et al. 2017), mainly by the action of hydroxyl radicals (HO^∙) via the Fenton reaction (Hell and Stephan 2003; Pereira et al. 2013).

There are no previous reports on soil waterlogging associated with excess Fe in soybean crops, although tolerance is often a product of tolerance to anaerobiosis and to toxicities of the excessively available elements (Singh and Setter 2017). Therefore, the present study sought to evaluate the response mechanisms in soybean leaves ‘Agroeste 3680’ by physiological and biochemical analyses associating them with the development of pods in non-waterlogged and waterlogged soil, combined with one moderate and two toxic levels of Fe.

Materials and methods

Experimental site

The experiment was carried out under greenhouse conditions at the College of Agricultural and Technological Sciences, São Paulo State University (UNESP), São Paulo State, Brazil (21° 29′ S, 51° 2′ W; 396 m above sea level). The greenhouse used was an arch type, covered with a transparent light diffusing plastic film, with a thickness of 1000 microns. The sides of the greenhouse were covered by a Sombrite screen, with a 50% light-holding capacity.

Experimental design and treatments

The experimental design was completely randomised and arranged in a 2 × 3 factorial scheme, with two water levels in the soil (without waterlogging and with waterlogging) and three soil Fe levels (25, 125 and 500 mg dm⁻³). Each pot contained three plants, giving a final population of nine plants per treatment.

Cultivation conditions

The soil was a dystrophic Oxisol (Santos et al. 2018). The sample was collected at a depth of 0.0–0.2 m, forming a composite sample. It was crumbled, air-dried and sieved (4.0 mm), and shown to have the following chemical attributes: pH (CaCl₂) 4.6, organic matter 14 g dm⁻³, P (resin) 4 mg dm⁻³, K 2.1 mmol_c dm⁻³, Ca 7 mmol_c dm⁻³, Mg 5 mmol_c dm⁻³, S 5 mg dm⁻³, B 0.09 mg dm⁻³, Cu 0.6 mg dm⁻³, Fe 0.25 mg dm⁻³, Mn 16.7 mg dm⁻³, Zn 1 mg dm⁻³, potential acidity (H + Al) 18 mmol_c dm⁻³, Al 4 mmol_c dm⁻³, sum of bases 14.1 mmol_c dm⁻³, cation exchange capacity 32.1 mmol_c dm⁻³, and base saturation 44%.

Based on the chemical analysis, soil liming was carried out to raise the base saturation to 70% (Quaggio et al. 1985), by adding analytical reagent grade CaCO₃ and MgCO₃ at a ratio of 3:1. The soil containing carbonate salts was incubated for 30 days in pots at a humidity of 80% of the field capacity to allow it to equilibrate. The soil was then air-dried for seven days. The pots used were polypropylene with a capacity of 4 dm³, lined with a polystyrene blanket to prevent soil loss during the experiment.

Experimental process

The following fertilisation was carried out per pot: 10 mg dm⁻³ N as CO(NH₂)₂, 200 mg dm⁻³ P as Ca(H₂PO₄)₂·H₂O, 150 mg dm⁻³ K as K₂SO₄, 0.5 mg dm⁻³ B as H₃BO₃, 0.05 mg dm⁻³ Co as CoCl₂·H₂O), 1.0 mg dm⁻³ Cu as CuSO₄·5H₂O, 0.05 mg dm⁻³ Mo as H₂MoO₄, 0.05 mg dm⁻³ Ni as NiSO₄·7H₂O, 5.0 mg dm⁻³ Mn as MnSO₄, and 2.0 mg dm⁻³ Zn as ZnSO₄. The K supply was split into three equal applications, which were supplied before sowing, and at the V2 and R1 phenological stages (Fehr et al. 1971). Fe, as FeCl₃·6H₂O, was supplied at 0, 100 and 475 mg dm⁻³ to the soil to give one natural level of 25 mg dm⁻³ Fe and two high levels of 125 and 500 mg dm⁻³ Fe, respectively.

After four days, 10 soybean seeds were sown at a depth of 3 cm, having previously been inoculated with the N₂-fixing bacterium Bradyrhizobium japonicum, strains SEMIA 587 and 5019. At the V1 phenological stage (Fehr et al. 1971), the seedlings were thinned to three representative seedlings per pot, selecting those with greater vigour and homogeneity of size.

At the R3 phenological stage (Fehr et al. 1971), the soil was waterlogged for a period of 10 days, totalling 70 days of experimental conduction from the germination of soybean seeds. The pots undergoing a waterlogged soil treatment were placed in larger pots with non-draining bottoms, and the water level in these was maintained at 2 cm above the soil surface of the inner pot.

During the experimental conduction, the replenishment of evapotranspired water for the plots was achieved using suspended micro-sprinklers, which were activated twice a day (morning and afternoon). In this way, soil humidity was maintained at 80% of the field capacity, except during the 10 days of stress in the pots where the soil was waterlogged. In these days, all plants were irrigated manually, respecting the water levels of each plot.

Measurements of gas exchange

On the 10th day of stress imposition, gas exchange evaluations were carried out using a LCpro portable infrared gas analyser (ADC Bioscientific Ltd., Hoddesdon, United Kingdom) on the first newly expanded trifoliate leaf (counting from the apex) of two plants from each pot. Evaluations were performed on a clear day between 10:00 and 11:30 a.m. Photosynthetically active radiation (PAR) was standardised to an artificial saturating light of 1000 µmol m⁻² s⁻¹, 380 μmol CO₂ mol⁻¹ air and a chamber temperature of 28 °C, according to Reis et al. (2017).

The net photosynthetic rate (A, µmol CO₂ m⁻² s⁻¹), stomatal conductance (gs, mol H₂O m⁻² s⁻¹) and transpiration rate (E, mmol H₂O m⁻² s⁻¹) was obtained. Water use efficiency (WUE, μmol CO₂ mmol⁻¹ H₂O) and instantaneous carboxylation efficiency (EiC, mol air⁻¹) were estimated from Eqs. 1 and 2, respectively.

$$WUE=\left(\frac{A}{E}\right)$$ 1

$$EiC=\left(\frac{A}{\mathit{Ci}}\right)$$ 2

where Ci (µmol CO₂ mol ar⁻¹) is the internal CO₂ level in the substomatal chamber.

Material storage for biochemical analysis

At the end of the stresses imposition, the first newly expanded trifoliate leaves (counting from the apex) were collected and frozen in liquid N₂. The samples were packaged and labelled in aluminium foil and stored at −80 °C until analysed as described below.

Analysis of photosynthetic pigments

Methods for determining photosynthetic pigment levels (chlorophylls and carotenoids) were based on those described by Lichtenthaler and Wellburn (1983). Leaves from two plants in each pot were assayed. Fresh leaf tissue (0.5 g) was ground in liquid N₂ with a pestle and mortar. Five ml 80% acetone was added to the samples and they were then centrifuged at 1500 rpm for 5 min. Following extraction, the absorbance (ABS) of the solvent was read at wavelengths specific to the different pigments by spectrophotometry. Chlorophyll a (Chl a), chlorophyll b (Chl b), total chlorophyll (Tchl) and carotenoid (CAR) contents were expressed as μg ml⁻¹, and calculated according to Eqs. 3–6:

$$\text{Chl}a=\left(12.25\times {\text{ABS}}_{663.2}\right)-\left(2.79\times {\text{ABS}}_{646.8}\right)$$ 3

$$\text{Chl}b=\left(21.50\times {\text{ABS}}_{646.8}\right)-\left(5.10\times {\text{ABS}}_{663.2}\right)$$ 4

$$\text{Tchl}=\left(7.15\times {\text{ABS}}_{663,2}\right)+\left(18.71\times {\text{ABS}}_{646.8}\right)$$ 5

$$\text{CAR}=\left[\left(1000\times {\text{ABS}}_{470}\right)-\left(1.82\times \text{Chl}a-85.02\times \text{Chl}b\right)\right]/198$$ 6

Analysis of carbon and nitrogen compounds in leaves

Leaves from two plants of each pot were analysed. Fresh leaf tissue (0.5 g) was ground in liquid N₂ with a pestle and mortar. Based on the method of Bieleski and Turner (1966), compounds were extracted for 24 h with 0.5 ml MCW solution (60% methanol, 25% chloroform and 15% water). The samples were centrifuged at 3000 rpm for 15 min, and the supernatant and precipitate were collected.

One volume of chloroform and 1.5 volumes of water were added to 4 volumes of the resulting supernatant. After phase separation over 24 h, the aqueous (upper) phase was used for the quantification of compounds. The sucrose and total carbohydrate content (mg g⁻¹ fresh weight, FW) were determined by the anthrone methods of Umbreit et al. (1957) and Yemm and Willis (1954), respectively, using glucose as a standard. The total ureide, allantoate and allantoin contents were determined by the phenylhydrazine method of Vogels and Van Der Drift (1970), using allantoin as a standard. For the determination of allantoate, the initial alkaline hydrolysis step was omitted, and allantoin was calculated by the difference to the total ureides assay. Ammonia and nitrate were determined by the phenol-hypochlorite method of McCullough (1967) and the salicylic acid method described by Cataldo et al. (1975), respectively. The total ureides, allantoate and allantoin, nitrate and ammonia contents were expressed as µmol g⁻¹ FW.

From the residue obtained after centrifugation of the initial MCW extract (precipitate), the protein content was determined, as described by Bradford (1976), using bovine serum albumin as a standard. The starch content was also determined by the perchloric acid/anthrone method (Yemm and Willis 1954; Umbreit et al. 1957). Protein and starch contents were expressed as mg g⁻¹ FW.

Analysis of concentrations of hydrogen peroxide and lipid peroxidation

Hydrogen peroxide (H₂O₂) concentrations were determined by reaction with potassium iodide (KI), according to Alexieva et al. (2001). Leaves from two plants of each pot were assayed. Fresh leaf tissue (0.25 g) was ground in liquid N₂ with a pestle and mortar, to which 3 ml of 0.1% trichloroacetic acid (TCA) in 20% polyvinyl polypyrrolidone (PVPP) was added. After complete homogenisation, the samples were centrifuged at 10,000 rpm for 10 min at 4 °C in a refrigerated centrifuge. For the reaction, 0.2 ml of supernatant was added to 0.2 ml 100 mM potassium phosphate buffer pH 7.5 and 800 μl 1 M KI solution. The samples were kept on ice for 1 h, and then absorbance readings were taken at 390 nm. The H₂O₂ concentration was calculated based on a standard curve of H₂O₂, and the results were expressed in nmol g⁻¹ FW.

Lipid peroxidation was evaluated by determining the concentration of metabolites reactive to 2-thiobarbituric acid (TBA), mainly malondialdehyde (MDA), according to the method of Heath and Packer (1968). The initial procedures for MDA measurement were the same as those described above for H₂O₂ measurements. Following centrifugation, 0.25 ml of supernatant was added to 1 ml 20% TCA solution containing 0.5% thiobarbituric acid (TBA). The samples were kept in a dry bath at 95 °C for 30 min and then on ice for 20 min. Subsequently, the samples were centrifuged at 10,000 rpm for 5 min in order to separate residues formed during heating. Samples were read at two wavelengths, 535 nm and 600 nm, and MDA concentrations were calculated according to Eq. 7:

$$\text{MDA}(\text{nmol}{\text{g}}^{-1}\text{FW})=\left[({\text{ABS}}_{532}-{\text{ABS}}_{600})/155.000\right]\times {10}^3$$ 7

Fe accumulation in shoots and partitioned biomass production

At the end of the exposure to stresses, plants were separated into shoots and roots. The material was oven dried at 65 °C for 72 h, and the pods were harvested from the shoots in order to assess the deleterious effect of the stresses on their development. Subsequently, shoot (SDW), pod (PDW) and root (RDW) dry weights were determined and expressed as g plant⁻¹.

To determine the Fe accumulation in the shoot (FEAS), the samples were subjected to digestion with nitric-perchloric acid solution (3:1) at 200 °C (Malavolta et al. 1997) after being ground in a Wiley-type mill. Subsequently, Fe concentrations were determined by atomic absorption spectrophotometry. FEAS (µg plant⁻¹) was calculated by multiplying Fe concentrations by the SDW.

Data analysis

Data were initially subjected to the Shapiro–Wilk normality test (p < 0.05), followed by analysis of variance (ANOVA) using the F test (p ≤ 0.05). The traits were compared using the Tukey test (p < 0.05). Pearson’s correlation (p < 0.05) with multiple comparisons at 1% probability was calculated using the corrplot package. All statistical analysis of the data was performed using protocols developed in the R software (R Development Core Team 2019).

Results

Gas exchange

The traits of A, gs, E, WUE and EiC showed a significant effect of double interaction on ANOVA (Supplementary Table S1).

At Fe levels of 125 and 500 mg dm⁻³, plants cultivated without soil waterlogging showed reduced gs by 10 and 22% and E by 28 and 51%, respectively (Fig. 1II, III). The A and EiC traits were less sensitive and reduced only at 500 mg dm⁻³ Fe, decreasing by 14 and 17%, respectively (Fig. 1I, V). Levels of 125 and 500 mg dm⁻³ Fe resulted in similar reductions for WUE, by on average 46% (Fig. 1IV).

Fig. 1.

Net photosynthetic rate (I), stomatal conductance (II), transpiration rate (III), water use efficiency (IV) and instantaneous carboxylation efficiency (V) of soybean leaves based on the significance of ANOVA by factorial analysis (p ≤ 0.05), comprised of soils with two levels of water (without waterlogging and with waterlogging) and three iron levels (25, 125 and 500 mg dm⁻³). Different letters indicate significant differences according to the Tukey test (p < 0.05). Uppercase letters compare the conditions of waterlogging between different iron levels in the soil, while lowercase letters compare the conditions of soil waterlogging with the same iron levels. Vertical bars represent the standard deviation (n = 6 plants)

The increase of Fe levels associated with soil waterlogging showed no significant differences for A, gs, WUE and EiC (Fig. 1I, II, IV, V), while E decreased at 500 mg dm⁻³ Fe (Fig. 1III). The traits presented in Fig. 1 showed a marked reduction in soil waterlogging at all Fe levels relative to non-waterlogged soil (Fig. 1I–V).

Content of photosynthesis pigments

Chl a, Chl b, Tchl and CAR contents showed a significant effect of double interaction on ANOVA (Supplementary Table S1).

Similar decreases in Chl a, Chl b, Tchl and CAR contents were observed at 125 and 500 mg dm⁻³ Fe without soil waterlogging, resulting in average reductions of 24, 25, 25 and 20%, respectively. Similar behaviour was also found in soil waterlogging for contents of Chl b (− 24%), Tchl (− 14%) and CAR (− 20%; Fig. 2I–IV). In contrast, Chl a content was less sensitive, showing a reduction of 14% at 500 mg dm⁻³ Fe, while plants at 25 and 125 mg dm⁻³ Fe did not differ for this pigment (Fig. 2I).

Fig. 2.

Chlorophyll a (I), chlorophyll b (II), total chlorophyll (III) and carotenoid (IV) contents of soybean leaves based on the significance of ANOVA by factorial analysis (p ≤ 0.05), comprised of soils with two levels of water (without waterlogging and with waterlogging) and three iron levels (25, 125 and 500 mg dm⁻³). Different letters indicate significant differences according to the Tukey test (p < 0.05). Uppercase letters compare the conditions of waterlogging between different iron levels in the soil, while lowercase letters compare the conditions of soil waterlogging with the same iron levels. Vertical bars represent the standard deviation (n = 6 plants)

At each Fe level, plants showed a lower photosynthetic pigment content under soil waterlogging. At a natural level of Fe, Chl a, Chl b, Tchl and CAR showed a reduction of 50, 51, 50 and 43%, respectively, while the average reduction was 41%, 51%, 46% and 34%, respectively, with an excess of Fe (125 and 500 mg dm⁻³), as shown in Fig. 2I–IV.

Carbon compounds in leaves

Starch, sucrose and total carbohydrate contents showed a significant effect of double interaction on ANOVA (Supplementary Table S1).

An increase in starch content was observed in non-waterlogged soil at 500 mg dm⁻³ Fe and waterlogged soil at 125 and 500 mg dm⁻³ Fe, with increases of 41, 16 and 16%, respectively. In addition, soil waterlogging provided higher starch contents at all Fe levels (Fig. 3I). In contrast, total carbohydrate and sucrose contents decreased in plants cultivated at 500 mg dm⁻³ Fe under both soil water conditions, and were lower under soil waterlogging, with a tendency towards less variation with increasing Fe levels compared to plants in non-waterlogged soils. These variations led to a reduction in total carbohydrate and sucrose contents by 49 and 45% (without waterlogging), and 41 and 34% (with waterlogging), respectively, between levels of 25 and 500 mg dm⁻³ Fe (Fig. 3II, III).

Fig. 3.

Starch (I), total carbohydrate (III) and sucrose (II) contents of soybean leaves based on the significance of ANOVA by factorial analysis (p ≤ 0.05), comprised of soils with two levels of water (without waterlogging and with waterlogging) and three iron levels (25, 125 and 500 mg dm⁻³). Different letters indicate significant differences according to the Tukey test (p < 0.05). Uppercase letters compare the conditions of waterlogging between different iron levels in the soil, while lowercase letters compare the conditions of soil waterlogging with the same iron levels. Vertical bars represent the standard deviation (n = 6 plants)

Nitrogen compounds in leaves

As shown by ANOVA (Supplementary Table S1), ammonia content was not affected by any factor, and its content in leaf tissue was null (0 nmol g⁻¹ FW). Total ureides, allantoate and allantoin contents showed a significant single effect for soil waterlogging conditions and Fe levels. On the other hand, nitrate and protein contents showed a significant effect of double interaction.

Regardless of Fe levels, total ureides, allantoate and allantoin contents increased in plants on soil waterlogging, by 24, 18 and 31%, respectively. At 500 mg dm⁻³ Fe, total ureides, allantoate and allantoin contents showed increases of 28, 18 and 19%, respectively, compared to plant averages at 25 and 125 mg dm⁻³ Fe (Fig. 4I–III), regardless of the waterlogging conditions.

Fig. 4.

Total ureides (I), allantoate (II), allantoin (III), nitrate (IV) and protein (V) contents of soybean leaves based on the significance of ANOVA by factorial analysis (p ≤ 0.05), comprised of soils with two levels of water (without waterlogging and with waterlogging) and three iron levels (25, 125 and 500 mg dm⁻³). Different letters indicate significant differences according to the Tukey test (p < 0.05). Isolated uppercase letters compare the conditions of soil waterlogging regardless of iron levels, while isolated lowercase letters compare iron levels regardless of soil waterlogging. Uppercase letters compare the conditions of waterlogging between different iron levels in the soil, while lowercase letters compare the conditions of soil waterlogging with the same iron levels. Vertical bars represent the standard deviation (n = 6 plants)

Nitrate content was not changed in plants cultivated under increasing Fe levels without soil waterlogging. On the other hand, plants at 125 and 500 mg dm⁻³ Fe in waterlogged soils showed a similar reduction of 33% on average (Fig. 4IV). At 25 mg dm⁻³ Fe, nitrate content showed an increase of 38% for waterlogged soil, compared to plants in soil without waterlogging, while levels in plants at 125 and 500 mg dm⁻³ Fe did not differ between the two soil water conditions (Fig. 4IV).

Protein content showed an increase of 86% in non-waterlogged soil at 500 mg dm⁻³ Fe. On the other hand, in waterlogged soil, an increase in protein content occurred in plants cultivated at 25 mg dm⁻³ Fe, showing a 57% increase compared to higher levels of Fe (Fig. 4V). At 25 mg dm⁻³ Fe, protein content increased substantially on soil waterlogging relative to non-waterlogging soil, while it did not differ between the two water conditions at levels of 125 and 500 mg dm⁻³ Fe (Fig. 4V).

Hydrogen peroxide and lipid peroxidation concentrations

H₂O₂ and MDA concentrations showed a significant single effect for soil waterlogging conditions and Fe levels on ANOVA (Supplementary Table S1).

Regardless of Fe levels, H₂O₂ and MDA concentrations increased by 94 and 23%, respectively, in plants in waterlogged soil. The greatest increase for H₂O₂ was observed at 500 mg dm⁻³ Fe, while a similar increase was found for MDA at 125 and 500 mg dm⁻³. The increases were 17 and 23% for H₂O₂ and MDA concentrations, respectively (Fig. 5I, II), regardless of the waterlogging conditions.

Fig. 5.

Hydrogen peroxide (a) and malondialdehyde (b) concentrations of soybean leaves based on the significance of ANOVA by factorial analysis (p ≤ 0.05), comprised of soils with two levels of water (without waterlogging and with waterlogging) and three iron levels (25, 125 and 500 mg dm⁻³). Different letters indicate significant differences according to the Tukey test (p < 0.05). Isolated uppercase letters compare the conditions of soil waterlogging regardless of iron levels, while isolated lowercase letters compare iron levels regardless of soil waterlogging. Vertical bars represent the standard deviation (n = 6 plants)

Fe accumulation in shoots and partitioned biomass production

The traits of SDW, PDW, RDW and FEAS showed a significant effect of double interaction on ANOVA (Supplementary Table S1).

Plants cultivated in non-waterlogged soil showed severe decreases in SDW and RDW with increasing Fe levels. Upon soil waterlogging, these traits showed less variation, being reduced only at 500 mg dm⁻³ Fe. In addition, SDW and RDW showed the lowest values in waterlogged soil compared to non-waterlogged soil, especially for RDW, and both SDW and RDW showed a tendency towards less variation with increasing Fe levels (Fig. 6I, III).

Fig. 6.

Shoot dry weight (I), pod dry weight (II), root dry weight (III) and iron accumulation in the shoots (IV) of soybean leaves based on the significance of ANOVA by factorial analysis (p ≤ 0.05), comprised of soils with two levels of water (without waterlogging and with waterlogging) and three iron levels (25, 125 and 500 mg dm⁻³). Different letters indicate significant differences according to the Tukey test (p < 0.05). Uppercase letters compare the conditions of waterlogging between different iron levels in the soil, while lowercase letters compare the conditions of soil waterlogging with the same iron levels. Vertical bars represent the standard deviation (n = 6 plants for FEAS; n = 9 plants for dry weights)

BDW decreased with increasing Fe levels in both water conditions. This effect was more pronounced in non-waterlogged soil, resulting in reductions of 44 and 94% at 125 and 500 mg dm⁻³ Fe, respectively. In waterlogged soil, BDW showed a similar reduction, averaging 60% at both these Fe levels. At each Fe level, plants showed lower values in waterlogged soils, recording decreases of 94 (25 mg dm⁻³), 95 (125 mg dm⁻³) and 65% (500 mg dm⁻³) compared to non-waterlogged soils (Fig. 6II). FEAS showed a 91% decrease at 500 mg dm⁻³ Fe (without waterlogging), and a 25 and 74% decrease at 125 and 500 mg dm⁻³ Fe (with waterlogging), respectively. The highest values were observed in waterlogged soil (Fig. 6IV).

In Fig. 7, most traits showed significant correlations. However, nitrate, protein and FEAS were the traits that had the smallest relationship with the others.

Fig. 7.

Pearson’s correlation (p < 0.05) with multiple comparisons at 1% probability between traits. The squares that received the white colour belong to the category of non-significant correlative values. Net photosynthetic rate (A), stomatal conductance (gs), transpiration rate (E), water use efficiency (WUE), instantaneous carboxylation efficiency (EiC), Chlorophyll a (Chl a), chlorophyll b (Chl b), total chlorophyll (Tchl), carotenoids (CAR), hydrogen peroxide (H₂O₂), malondialdehyde (MDA), shoot dry weight (SDW), pod dry weight (PDW), root dry weight (RDW), Fe accumulation in the shoot (FEAS)

Discussion

Gas exchange was affected by soil waterlogging and excess Fe

Photosynthesis is a primary physiological process for plant survival and growth (Yan et al. 2018). In sensitive species, soil waterlogging and excess Fe are stresses that directly affect photosynthesis, the measurements of which can determine the degree of tolerance shown by species. Reductions in gs can limit the influx of CO₂ for photosynthesis. In addition, limitations may also occur in the carboxylation reactions, which leads to reduced photosynthesis, and which are commonly used to infer the cumulative effects of biochemical limitations (Müller et al. 2017).

In non-waterlogged soil, gs showed a progressive decrease with increasing Fe levels (Fig. 1II); however, changes in A and EiC were minor under these conditions (Fig. 1I, IV), suggesting that the behaviour observed in A was more closely associated with biochemical than stomatal limitations. Similar results were obtained with rice (Oryza sativa) by Müller et al. (2017). The decrease in gs in the current study was probably linked to the increase in FEAS (Fig. 6a), which was corroborated by the negative correlation recorded in Fig. 7. According to Dufey et al. (2009) and Pereira et al. (2013), stomatal closure is a late response to increased Fe uptake, rather than a defence mechanism in rice. Consequently, with the reduction of gs, E was reduced (Fig. 1II, III); however, it was not enough to optimise the WUE (Fig. 1V).

In waterlogged soil, gas exchange was substantially reduced (Fig. 1I–V). The natural level of Fe in the soil provided values similar to the two highest levels of Fe for traits associated with gas exchange, except for a reduction in E detected at 500 mg dm⁻³ Fe (Fig. 1III). These responses showed that soil waterlogging strongly affected the gas exchange of soybean at the R3 phenological stage. However, it is important to emphasise that the effects of waterlogging are complex, and vary depending on genotype, environmental conditions, growth stage and the duration of waterlogging (Tian et al. 2019).

Soil waterlogging and excess Fe reduced photosynthetic pigments

Damage to PSI and PSII can be inferred by the contents of Chl a and Chl b, respectively, since chlorophyll content varies within the photosystems (Wientjes et al. 2017). Under excess Fe in non-waterlogged soil, the contents of Chl a and Chl b (Fig. 2A, B) behaved similarly to that of Tchl (Fig. 4III), indicating that decreases recorded in these situations were similar between chlorophylls. However, the reduction of chlorophyll content at 125 mg dm⁻³ Fe did not reduce A (Fig. 1I), indicating that even with lower electron flow, A was maintained.

Soil waterlogging associated with excess Fe enhanced damage to Chl b content, which was more affected than Chl a content, mainly at 125 mg dm⁻³ Fe. In addition, the decline in Chl b content was not compensated for by Tchl content (Fig. 2I–III). Thus, it was assumed that a combination of these stresses impaired light energy utilisation and dissipation, and caused a reduction in antenna size relative to reaction centres, indicating greater photooxidative and oxidative damage to photosystems (Xu et al. 2015). On the other hand, soybean plants cultivated under natural levels of Fe in waterlogged soil showed a similar decrease between Chl a and Chl b content. In addition, the observed reductions in chlorophyll content did not determine a change in A (Fig. 1I), since this trait did not show differential behaviour with increasing Fe levels. Previous studies on soil waterlogging have reported notably greater damage to Chl b content in Helianthus tuberosus (Yan et al. 2018), Vigna radiata (Sairam et al. 2009) and Gossypium hirsutum (Kuai et al. 2014).

Carotenoids participate in light-absorption and protect the antenna complex from oxidative damage, due to their ability to extinguish reactive oxygen species (ROS) that are formed naturally during the photosynthetic process (Pinto et al. 2003). The variations observed in chlorophyll and carotenoid contents (Fig. 2I–IV) were possibly related to oxidative stress, as verified by MDA and H₂O₂ concentrations (Fig. 5I, II), which was corroborated by their significant correlation (Fig. 7).

Carbon and nitrogen compounds: starch and ureides accumulated in leaves

Starch is the main storage product in plants (Kuai et al. 2014). Its accumulation (Fig. 3I) suggested a blockage of carbohydrate transport to sink tissues and the decreased export of triose phosphates to the cytoplasm, resulting in lower contents of sucrose and total carbohydrates (Fig. 3II, III). The lower dry weight of partitioned biomass under stresses (Fig. 6II–IV) supported this hypothesis. Previous studies have shown similar results; for example, Wample and Davis 1983 found that sucrose translocation from the leaves to roots was blocked in waterlogged soil, resulting in an accumulation of starch in the leaf tissue of sunflower (Helianthus annuus). Kuai et al. (2014) reported that sucrose transport from the subtending leaves to the boll in cotton (Gossypium hirsutum) was irreversibly blocked after three days of soil waterlogging.

Ureides accumulation in soybean leaves under water deficit has previously been described (Vadez et al. 2000; King and Purcell 2005). These authors suggested that the accumulation of ureides could have a regulatory role, participating in an N-feedback inhibition of nitrogenase. However, a study by Gil-Quintana et al. (2013) using a split-root system rejected this hypothesis, since ureide content increased in the stressed nodules later than the inhibition of N₂-fixation. It remains to be determined whether the accumulation of ureides in leaf tissue under soil waterlogging is a signal for the inhibition of N₂-fixation, which opens up avenues for further research in this segment. The influence of Fe levels on ureides accumulation in soybean leaves may be related to Mn²⁺ availability. Izaguirre-Mayoral and Sinclair (2005) found that high Fe levels resulted in decreased Mn levels in soybean leaves, and proposed that the inhibition of allantoate amidohydrolase activity is caused by deficiency of its Mn²⁺ co-factor. This enzyme hydrolyses allantoate to ureideglycolate (Watanabe et al. 2014). Thus, allantoate accumulation may prevent stored allantoin in peroxisomes from moving to the endoplasmic reticulum, where ureides catabolism occurs.

According to Oliveira et al. (2013), nitrate uptake by soybean plants is lower under hypoxia than under normoxia, and the location at which it is metabolised is changed, occurring in the leaves under hypoxia and in the roots under normoxia. The nitrate content found in soybean leaves in this study did not differ between the two water conditions at levels of 125 and 500 mg dm⁻³ Fe; however, there was an increase of its content in plants exposed to soil waterlogging compared to those in non-waterlogged soil at 25 mg dm⁻³ Fe. FEAS (Fig. 6IV) showed a significant correlation with nitrate (Fig. 7), suggesting the influence on nitrate dynamics in soybean plants. In addition, protein increase in this cultivation situation (non-waterlogged soil at 25 mg dm⁻³ Fe) may be related to the formation of chelators or other substances to attenuate ROS.

Toxicity and partitioned biomass weights

Plants cultivated in waterlogged soil had higher FEAS values (Fig. 6IV), which were potentially toxic to plants, as this catalyses H₂O₂ decomposition, generating HO^∙ (Fenton reaction). Fe³⁺ produced by this reaction can be reduced to Fe²⁺ by O^∙−₂ (Haber–Weiss reaction), allowing Fe²⁺ to participate in the Fenton’s reaction again (Becana et al. 1998; Hell and Stephan 2003). In addition, waterlogged soil can affect the dynamics of the functioning of the photosynthetic apparatus, leading to the leakage of ions from the electron transport chain to O₂, inducing the over-production of ¹O₂ and H₂O₂ in PSII (Barbosa et al. 2014; Zheng et al. 2017). In addition, the reduced ferredoxin electron can be transferred to O₂, instead of to NADP, generating O^∙−₂ at the acceptor side of PSI (Barbosa et al. 2014; Yan et al. 2018). These events might explain the higher H₂O₂ and MDA concentrations in soybean leaves due to soil waterlogging and excess Fe (Fig. 5I, II), which was associated with a loss of dry biomass (Fig. 7).

Soybean plants had severely decreased dry biomass with stress imposition (Fig. 6I–III). In particular, soil waterlogging was what most impacted PDW (Fig. 6III), being extremely harmful to soybeans at this phenological stage. This progressive loss of biomass and plant performance was a result of the physiological and biochemical changes documented in the current study (Figs. 1, 2, 3, 4, 5); however, other factors may have also contributed: (I) a reduction in soil redox and consequently a reduction in uptake capacity tending towards a decrease in plant N and P concentrations (Dat et al. 2004); (II) a change from oxidative phosphorylation to glycolysis and fermentation due to the absence of O₂ in the root tissue (Sairam et al. 2009; Martínez-Alcántara et al. 2012); and (III) the ability of plants to mitigate excessive Fe uptake through exclusion mechanisms and/or by storing/remobilising absorbed Fe (Müller et al. 2017).

Conclusion

Gas exchange was strongly affected by soil waterlogging. Excess Fe without soil waterlogging reduced photosynthetic pigments, and potentiated this reduction when associated with soil waterlogging.

Starch and ureide accumulation in the first newly expanded trifoliate leaves proved to be response mechanisms induced by soil waterlogging and excess Fe, since plants cultivated under soil non-waterlogged soil at 25 mg dm⁻³ Fe showed lower contents when compared to stressed plants. Thus, starch and ureide accumulation could be considered efficient biomarkers of phytotoxicity caused by soil waterlogging and excess Fe in soybean plants.

The reproductive development was abruptly interrupted by the exposure to stresses, leading to a loss of pod dry biomass, which was largely due to the substantial decrease in A, the blockage of carbohydrate transport to sink tissues, and an increase in MDA. The negative effect on reproductive development was more pronounced under waterlogged soil.

Future studies should seek to understand the role of starch and ureide accumulation in soybean leaves induced by these stresses. In addition, new studies should be directed to elucidate the strategies used by soybean plants to avoid Fe toxicity under anoxia.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

ATP

Adenosine triphosphate

A

Net photosynthetic rate

gs

Stomatal conductance

E

Transpiration rate

EiC

Instantaneous carboxylation efficiency

WUE

Water use efficiency

Chl a

Chlorophyll a

Chl b

Chlorophyll b

Tchl

Total chlorophyll

CAR

Carotenoids

PSI

Photosystem I

PSII

Photosystem II

HO^∙

Hydroxyl

¹O₂

Singlet oxygen

H₂O₂

Hydrogen peroxide

O^∙−₂

Superoxide anion

ROS

Reactive oxygen species

MDA

Malondialdehyde

SDW

Shoot dry weight

PDW

Pod dry weight

RDW

Root dry weight

FEAS

Fe accumulation in the shoots

ABS

Absorbance

ANOVA

Analysis of variance