Phytoglobin expression influences soil flooding response of corn plants

Abstract

Background and Aims Excess water is a limiting factor for crop productivity. Under conditions of full submergence or flooding, plants can experience prolonged oxygen depletion which compromises basic physiological and biochemical processes. Severe perturbations of the photosynthetic machinery with a concomitant decline in photosynthetic potential as a result of elevated levels of reactive oxygen species (ROS) are the major consequences of water excess. Phytoglobins (Pgbs) are ubiquitous proteins induced by several types of stress which affect plant response by modulating nitric oxide.

Methods Maize plants overexpressing or downregulating two Pgb genes were subjected to soil flooding for 10 d and their performance was estimated by measuring several gas exchange parameters including photosynthetic rate. Above-ground tissue was utilized to localize ROS and to measure the expression and activities of major antioxidant enzymes.

Key Results Relative to the wild type, flooded plants overexpressing Pgb genes retained a greater photosynthetic rate and enhanced activity of several antioxidant enzymes. These plants also exhibited high levels of ascorbic acid and reduced ROS staining. This was in contrast to flooded plants downregulating Pgb genes and characterized by the lowest photosynthetic rates and reduced expression and activities of many antioxidant enzymes.

Conclusions Induction of Pgb genes alleviates flooding stress by limiting ROS-induced damage and ensuring a sustained photosynthetic rate. This is achieved through improvements of the ascorbate antioxidant status including an enrichment of the ascorbate pool via de novo and recycling mechanisms, and increased activities of several ROS-scavenging enzymes.

INTRODUCTION

Depletion of oxygen, often experienced by plants grown in poorly drained soils or subjected to soil flooding, i.e. waterlogging, is a major cause of crop loss as it can result in stunted growth and severe injury. Under these conditions, roots are the first to experience stress and, while they are able to cope with anoxic or hypoxic conditions for a limited period of time, prolonged exposure can cause extensive and irreversible damage locally or to the above-ground organs (Bailey-Serres and Voesenek, 2008). Oxygen depletion interferes with several oxygen-requiring metabolic pathways, some of which are implicated in energy production. Hypoxia reduces the ATP pool and induces alcoholic fermentation in the roots (Perata and Alpi, 1993), disturbing basic physiological and biochemical processes in photosynthetic tissues including carbohydrate transport and assimilation, and hormonal balance co-ordinating root–shoot function (Vartapetian and Jackson, 1997). Substantial reductions in photosynthetic capacity and transpiration rate, common responses to soil flooding (Titarenko, 2000; Jackson, 2002), are attributable to a direct effect of stomatal density or stomatal closure linked to changes in the abscisic acid (ABA) level (Bai et al., 2013), and/or structural and physiological perturbations in the photosynthetic tissue. An increase in stomatal resistance, possibly due to an ABA-mediated signal originating from the oxygen-deficient roots, has been reported within the first few hours of soil flooding in association with a reduction in the mesophyll CO₂ level (Else et al., 2001). Depressed assimilation rates before and after flooding in soil-flooded tomato plants were not the result of reduced quantum yield or changes in photorespiratory activity but could be attributed to reductions in stomatal conductance (Bradford, 1983). Prolonged exposure to soil flooding can also result in perturbations of the photosynthetic machinery, with the decline in photosynthetic potential possibly due to reactive oxygen species (ROS) production (Chen et al., 2014). ROS can induce structural damage of chloroplast membranes, inhibiting the function of the photosystems (Titarenko, 2000).

Accumulation of ROS, including superoxide (${\text{O}}_{\text{2}}^{\text{–}}$) and hydrogen peroxide (H₂O₂), is an unavoidable consequence of metabolism, especially in oxygen-abundant chloroplasts. Produced under normal conditions, the imposition of several forms of stress can dangerously increase their level, leading to severe damage to nucleic acids, membranes and proteins (Bowler et al., 1992; Foyer et al., 1994). Independent evidence suggests that hypoxia is responsible for ROS-induced oxidative stress (Monk et al., 1987; Yu and Rengel, 1999), the symptoms of which include lipid peroxidation, electrolyte leakage, alterations in lipid composition and ultimately cell death (Hetherington et al., 1982; Crawford et al., 1994). In addition to increasing ROS synthesis, soil flooding also impairs mechanisms deployed by cells to cope with high ROS levels. These include the ability to augment the pool of antioxidants, such as phenolic compounds, ascorbate and glutathione, in an effort to maintain a redox homeostasis, and activate ROS-scavenging enzymes (Yordanova et al., 2004). Brassica napus seedlings grown in water-saturated soils showed a substantial decrease in the activities of superoxide dismutase (SOD) and catalase (CAT), enzymes involved in the removal of ${\text{O}}_{\text{2}}^{\text{–}}$ and H₂O₂ (Zhou et al., 1997). Similar enzymatic changes were reported in Oryza sativa seedlings germinated under water, which also showed severe perturbations in activities of enzymes of the ascorbate–glutathione cycle (Ushimaru et al., 1992), the major pathway by which ROS are degraded (Noctor and Foyer, 1998). In this ubiquitous pathway, ascorbate (AsA) is oxidized to monodehydroascorbate (MDAR) by the enzyme ascorbate peroxidase (APX) using H₂O₂ as a substrate. Monodehydroascorbate can be converted to dehydroascorbate (DHA) through non-enzymatic mechanisms, or reduced back to AsA by monodehydroascorbate reductase (MDHAR). Reduction of DHA to AsA can also occur by dehydroascorbate reductase (DHAR), in a reaction converting reduced glutathione (GSH) to its oxidized form (GSSG). The cycle is completed with the activity of glutathione reductase (GR). Co-ordinated enzymatic activity is therefore paramount for maintaining an adequate balance between the oxidized and reduced antioxidant pools and protecting against ROS toxicity.

Phytoglobins (Pgbs), previously termed non-symbiotic haemoglobins, are haem-containing proteins found in all nucleated organisms and characterized by their ability to scavenge nitric oxide (NO) in response to stress conditions (Hill, 2012). Expression of Pgbs increases very rapidly under hypoxic conditions (Silva-Cardenas et al., 2003), and their expression level correlates with submergence tolerance (Campbell et al., 2015). When experimentally induced, Pgbs ameliorated the energy status and growth of both maize cells and alfalfa roots following imposition of hypoxic stress (Dordas et al., 2003; Igamberdiev and Hill, 2004). A similar result was also observed in arabidopsis where, relative to non-transformed plants, the constitutive expression of a Pgb gene enhanced survival of hypoxia (Hunt et al., 2002). These beneficial effects have been related to the protective role of Pgbs in maintaining cellular integrity and ultimately influencing survival (Huang et al., 2014). A link between Pgbs and ascorbate metabolism has been reported in root tissue under low oxygen tension (Igamberdiev et al., 2006b), and in relation to the NO-scavenging mechanisms mediated by Pgbs (Sullivan and Stern, 1982). Scavenging of NO occurs through the oxidation of oxyPgb to metPgb, which needs to be reduced by ascorbate to regenerate oxyPgb and sustain continual removal of NO. This reaction is the basis of Pgb function (Igamberdiev et al., 2006a).

One observed metabolic change as a result of Pgb manipulations is an increased expression of genes that are involved in regulating ROS (Igamberdiev et al., 2006b) that are known to be key components resulting in plant flooding damage (Blokhina and Fagerstedt, 2010). To determine whether Pgbs influence flooding tolerance by moderating generation of ROS, we overexpressed or downregulated two maize Pgb genes, ZmPgb1.1 and ZmPgb1.2, and assessed the performance of these plants under soil flooding conditions. Relative to wild-type (WT) plants, induction of the ZmPgb genes maintains an elevated photosynthetic rate and limits production of ROS and leaf injury tissue caused by flooding. This was in contrast to plants in which the ZmPgb genes were suppressed where flooding caused a substantial depression in the photosynthetic rate and a rise in ROS. These different behaviours were correlated to the Pgb mitigation of oxidative stress through the preferential regulation of key enzymes of the ascorbate–glutathione antioxidant system.

MATERIALS AND METHODS

Plant material and production of lines with altered levels of phytoglobins

Nineteen commercial genotypes of maize were kindly provided by Dr Lana Reid (Agriculture and Agri-Food Canada, Ottawa, Canada). Plant transformation with the two maize ZmPgb genes was executed as described by Huang et al. (2014). For construction of antisense and sense maize lines, ZmPgb1.1 (AF236080) and ZmPgb1.2 (DQ171946) were amplified from cDNA prepared from maize embryogenic tissue by reverse transcription–PCR (RT–PCR) using gene-specific primers, and subcloned into a pGEM T-Easy vector (Promega). Sequences were confirmed by DNA sequencing, and BamHI sites were added to both the 5′ and 3′ ends of both genes by PCR. The maize poly ubiquitin1 (ubi-1) promoter was isolated from the pUbiSXR 14 vector, and KpnI restriction sites were added to the 5′ and 3′ ends by PCR. The nopaline synthase (nos-t) terminator was isolated from pBI1218 with SacI restriction sites added to the 5′ and 3′ ends. The entire cassette was assembled through sequential ligation of the ubi-1 promoter, nos-t and ZmPgb1.1 or ZmPgb1.2 into pBluescript SK (–). Confirmed constructs were co-bombarded into immature maize embryos using the herbicide bar gene as a selective marker (Wang and Frame, 2009).

Plant regeneration from transformed embryogenic tissue obtained from the Plant Transformation Facility at Iowa State University (http://agron-www.agron.iastate.edu/ptf/) was performed exactly as described by Wang and Frame (2009). Briefly, tissue was cultured on a 2,4-D-containing maintenance liquid medium for 7 d, followed by a transfer onto the solid hormone-free development medium for 3 weeks. Glufosinate ammonia was used as the selective agent in both maintenance and development media. Fully mature somatic embryos were harvested from the development medium and plated onto a germination medium which induced formation of leaves and roots (Wang and Frame, 2009). After 2 weeks germination, uniform plants were selected and transplanted into pots containing Metro-Mix 900.

Soil flooding experiments and gas exchange measurements

Seedlings obtained from seeds of the 19 genotypes and from the regenerated somatic embryos were grown in 12 inch pots containing Metro-Mix 900. Uniform plants at the three-leaf stage of development were selected for soil flooding experiments. Briefly, pots were transferred into transparent Plexiglas containers and soil flooded for 10 d in a growth cabinet under a 16 h photoperiod of 22 °C light/20 °C dark. The water level was maintained 2 cm above the soil surface. A second set of control (unflooded) plants was grown under identical conditions and watered every 3 d to keep an adequate soil moisture level.

The photosynthetic rate, stomatal conductivity, internal CO₂ level, transpiration rate and water use efficiency (WUE) were measured on the fully expanded second youngest leaf of the plants subjected to 10 d of waterlogging using a gas exchange infrared gas analyser (IRGA; LI-6400, LI-COR, Inc., Lincoln, NE, USA). Experiments were performed between 1000 h and 1500 h. Gas exchange parameters were measured under photosynthetically active radiation of 400 μmol m⁻² s⁻¹, atmospheric CO₂ of 400 μmol mol⁻¹, relative humidity of 50 % and temperature of 22 °C. Measurement were performed on three biological replicates each consisting of at least five plants.

Leaf injury, reactive oxygen localization and gene expression studies

Leaf injury was estimated on the first three leaves of maize plants subjected to soil flooding for 10 d using a visual index scale (1–10) based on the percentage of damaged leaf area: 1, 10 % damage; 2, 20 % damage; 3, 30 % damage; 4, 40 % damage; 5, 50 % damage; 6, 60 % damage; 7, 70 % damage; 8, 80 % damage; 9, 90 % damage; 10, 100 % damage.

After 10 d of soil flooding, ROS localization was performed on the second youngest leaf using nitroblue tetrazolium (NBT) and diaminobenzidine (DAB) to stain ${\text{O}}_{\text{2}}^{\text{–}}$ and H₂O₂, respectively, exactly as described by Campbell et al. (2015).

For gene expression studies, the above-ground tissue of plants subjected to soil flooding and control (unflooded ) plants was harvested at day 1, 2 and 10, frozen in liquid nitrogen and stored at –80 °C. TRI Reagent Solution was used to extract RNA following the manufacturer’s instruction (Invitrogen), and a High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) was used for cDNA synthesis. Gene expression studies were performed by quantitative PCR. All primers used for gene expression studies are listed in Supplementary Data Table S1. The relative gene expression level was analysed with the 2^−ΔΔCT method (Livak and Schmittgen, 2001) using actin as the reference gene.

Activities of antioxidant enzymes and quantitation of ascorbic acid and glutathione

The activities of ascorbate free radical reductase (AFRR), APX, DHAR and GR were analysed following homogenization of tissues at 4 ºC in medium containing 50 mm Tris–HCl, pH 7·2, 0·3 m mannitol, 1 mm EDTA, 0·1 % bovine serum albumin (BSA), 0·05 % cysteine and 2 % (w/v) polyvinylpyrrolidone. The homogenate was then centrifuged at 4 °C for 20 min at 16 000 g, and the supernatant was collected for the analysis of enzymatic activities (Zhang and Kirkham, 1996).

The activity of APX was estimated by measuring the H₂O₂-dependent oxidation of AsA (extinction coefficient 2·8 mm⁻¹ cm⁻¹) following the decrease in absorbance at 265 nm. The reaction mixture contained 50 μm AsA, 90 μm H₂O₂ and 50 mm potassium phosphate buffer, pH 6·5.

The activity of AFRR was measured following the rate of NADH oxidation at 340 nm (extinction coefficient 6·0 mm⁻¹ cm⁻¹). The reaction mixture contained 0·2 mm NADH, 1 mm AsA and 0·1 m Tris–HCl, pH 7·2. The reaction was initiated with the addition of 0·5 U of ascorbate oxidase.

The DHAR activity was determined by following the GSH-dependent production of AsA at 265 nm (extinction coefficient 6·2 mm⁻¹ cm⁻¹). The reaction mixture contained 1 mm DHA, 2 mm GSH and 100 mm potassium phosphate buffer, pH 6·3.

The activity of GR was determined following the NADPH-dependent oxidation of GSSG at 340 nm (extinction coefficient 14 mm⁻¹ cm⁻¹). The reaction mixture contained 0·1 m Tris–HCl, pH 7·8, 2 mm EDTA and 0·5 mm GSSG. The reaction was initiated with the addition of 50 μM NADPH.

The activities of CAT and SOD were analysed following homogenization of tissues at 4 ºC in 50 mm sodium phosphate buffer (pH 7·0) containing 0·2 mm EDTA and 1 % (w/v) polyvinylpyrrolidone. The homogenate was then centrifuged at 4 ºC for 20 min at 16 000 g, and the supernatant was collected for the analysis of enzymatic activities.

The activity of CAT was measured according to Zhang and Kirkham (1996), by following the decomposition of H₂O₂ (decline in absorbance at 240 nm) for 1 min. The reaction mixture (3 mL) contained 50 mm phosphate buffer (pH 7·0), 15 mm H₂O₂ and 0·1 mL of enzyme extract.

The activity of SOD was determined by measuring the ability of the enzyme to inhibit the reduction of NBT in a reaction mixture containing 13 mm methionine, 75 μm NBT, 0·1 mm EDTA and 2 μm riboflavin in 50 mm phosphate buffer (pH 7·8). The reaction was conducted exactly as reported by Zhang and Kirkham (1996).

Spectrophotometric quantitation of AsA, MDHA, GSH and GSSG was performed exactly as previously described (Stasolla et al., 2001).

Statistical analysis

Data were analysed by one-way analysis of variance (ANOVA) using the SPSS program (Version 19.0). Treatment means were compared by Duncan’s test (α = 0·05) to differentiate the significance of difference between various parameters. Linear regression analyses were performed using SigmaPlot 10 (Systat Software, Inc., San Jose, CA, USA).

RESULTS

Maize flooding, photosynthesis and Pgb expression

The photosynthetic rates of the 19 commercial maize genotypes varied over almost a 3-fold range (Supplementary Data Table S2). Soil flooding reduced the photosynthetic rate, stomatal conductivity and transpiration rate, while it increased the level of internal CO₂ and had little effect on WUE. Maintenance of the photosynthetic rate during flooding (the ratio between the photosynthetic rate of flooded vs. non-flooded plants) showed variations among genotypes. Five genotypes, AAFC-3, CO441, CO442, CO447 and CO452, showed the highest retention of the photosynthetic rate (Supplementary Data Fig. S1).

To assess potential relationships between photosynthesis during flooding and Pgbs, the ability of the genotypes to sustain photosynthesis after flooding was compared with the level of expression of the two maize phytoglobins genes ZmPgb1.1 and ZmPgb1.2 measured prior to flooding (Supplementary Data Fig. S2A). There was a wide range of ZmPgb1.1 and ZmPgb1.2 expression among the genotypes, with CO447 having a particularly high level of expression of ZmPgb1.1. A weak correlation (R² = 0·248) was measured between the expression of the two ZmPgb genes prior to flooding and retention of the photosynthetic rate after flooding.

The changes in expression (flooded/unflooded) of each ZmPgb gene were also compared with the retention of the photosynthetic rate (Supplementary Data Fig. S2B), and a positive correlation (R² = 0·546 for ZmPgb1.1 and R² = 0·745 for ZmPgb1.2) was observed (Fig. 1). This trend suggests the existence of a close relationship between the ability to enhance Pgb expression, especially that of ZmPgb1.2, during flooding and the capacity to maintain the photosynthetic rate as a result of flooding in the maize genotypes.

Fig. 1.

Correlation between the expression ratio (flooded/unflooded) of ZmPgb1.1 and ZmPgb1.2 in relation to the ability of the genotypes to retain photosynthetic rates after 10 d of flooding. Values were taken from Supplementary Data Fig. S2.

Collectively, these results prompted the examination of the effect of altering maize Pbg gene expression on parameters related to photosynthesis and other known physiological responses to flooding to assess the merit of the above observations. Maize lines overexpressing (S) or downregulating (A) ZmPgb1.1 or ZmPgb1.2 were generated through somatic embryogenesis (Fig. 2). Under normoxic conditions, no morphological differences were observed among the lines, which showed similar growth morphology at all stages of development (Supplementary Data Fig. S3A). In both WT and transformed lines, the photosynthetic rate was greatly reduced by flooding (Table 1). This decline was more pronounced in lines in which the ZmPgb genes were downregulated [ZmPgb1.1(A) and ZmPgb1.2(A)] where the photosynthetic rate dropped from about 15 μmol CO₂ m⁻² s⁻¹ to < 4 μmol CO₂ m⁻² s⁻¹. This decrease was accompanied by a reduction in shoot dry weight (Supplementaey Data Fig. S3). Relative to the WT, lines in which the ZmPgb genes were upregulated [ZmPgb1.1(S) and ZmPgb1.2(S)] retained higher photosynthetic rates after flooding, exhibited the highest shoot dry weight values and had a shorter root system (Table 1; Fig. S3B, C). A general decline in stomatal conductivity was observed in all lines subjected to soil flooding, especially in the WT and the lines in which the ZmPgb genes were downregulated (Table 1). Internal CO₂ levels were higher in the ZmPgb1.1(A) line under control (unflooded) conditions and increased in all lines after 10 d of flooding. Relative to the WT, the transpiration rate was severely inhibited in flooded lines in which the ZmPgb genes were downregulated, whereas both ZmPgb1.1(S) and ZmPgb1.2(S) lines retained the highest transpiration rate. No major differences among lines were found for WUE, which did not fluctuate significantly as a result of flooding (Table 1).

Fig. 2.

Characterization of the transgenic lines with altered expression of ZmPgb1.1 (A) or ZmPgb1.2 (B). The left panel shows the insertion of the transgene in the genomic DNA. Lane 1, WT; lane 2, transformed line; line 3, DNA ladder. Primers used in (A): ZmUb1 (F) and ZmPgb1.1(R) for ZmPgb1.1(S) or ZmPgb1.1(F) for ZmPgb1.1(A). Primers used in (B): ZmUb1 (F) and ZmPgb1.2(R) for ZmPgb1.2(S) or ZmPgb1.2(F) for ZmPgb1.2(A). Right panels show the expression levels of ZmPgb.1·1 or ZmPgb1.2 in leaves of WT plants and plants in which the respective ZmPbg gene was downregulated or upregulated. Primers used are ZmPgb1.1(F) and (R) for the top graph and ZmPgb1.2(F) and (R) for the bottom graph. Primer sequences are listed in Table S1.

Table 1.

Gas exchange parameters measured on the second leaf of the WT line and lines in which ZmPgb1.1 and ZmPgb1.2 were overexpressed (S) or downregulated (A)

Genotype	Photosynthetic rate (μmol CO2 m−2 s−1)	Stomatal conductivity (mmol m−2 s−1)	Internal CO2 (μmol CO2 mol−1 air)	Transpiration rate (mmol H2O m−2 s−1)	WUE (kg m−3)
Unflooded
WT	15·27 ± 1·03	0·09 ± 0·01	98·48 ± 10·81	2·60 ± 0·32	4·48 ± 0·30
ZmPgb1.1(S)	15·61 ± 0·67	0·09 ± 0·01	90·94 ± 7·57	3·29 ± 0·21	4·94 ± 0·48
ZmPgb1.1(A)	15·07 ± 1·33	0·08 ± 0·02	137·57 ± 6·21	2·93 ± 0·36	5·38 ± 0·32
ZmPgb1.2(S)	16·78 ± 1·11	0·11 ± 0·01	97·95 ± 8·44	3·59 ± 0·31	4·86 ± 0·28
ZmPgb1.2(A)	15·83 ± 0·91	0·11 ± 0·01	105·90 ± 10·25	3·93 ± 0·47	3·48 ± 0·22
Soil flooded
WT	5·47 ± 0·47*	0·04 ± 0·00*	183·22 ± 11·69*	1·60 ± 0·207*	3·68 ± 0·24
ZmPgb1.1(S)	9·90 ± 0·59*	0·07 ± 0·01*	139·78 ± 10·43*	2·41 ± 0·11*	4·15 ± 0·25
ZmPgb1.1(A)	2·94 ± 0·44*	0·03 ± 0·00*	167·00 ± 12·38	0·85 ± 0·13*	4·01 ± 0·60
ZmPgb1.2(S)	10·51 ± 0·80*	0·06 ± 0·00*	151·15 ± 10·60*	2·75 ± 0·25	4·11 ± 0·36
ZmPgb1.2(A)	3·93 ± 0·47*	0·04 ± 0·00*	167·08 ± 9·10*	1·13 ± 0·12*	3·48 ± 0·22

Measurements were performed on unflooded (control) plants and plants subjected to soil flooding for 10 d.

Values are means ± s.e. of at least three biological replicates with each replicate consisting of > 5 plants.

*Statistically significant values (P ≤ 0·05) from the respective unflooded value.

Leaf injury and ROS production in maize plants under soil flooding conditions

Leaf injury in leaves of transformed maize plants subjected to soil flooding was estimated on the first three leaves using a visual index scale (see the Materials and Methods and Supplementary Data Fig. S4). Foliar injury was apparent in all lines upon flooding, and the extent of damage increased with leaf age (Fig. 3). Overexpressing either ZmPgb1.1 or ZmPgb1.2 reduced the degree of leaf damage compared with the WT at all leaf ages during flooding. Downregulating either ZmPgb1.1 or ZmPgb1.2 produced results which did not differ greatly from the WT (Fig. 3).

Fig. 3.

Leaf injury measured on the three youngest leaves of WT plants and plants in which ZmPgb1.1 and ZmPgb1.2 were overexpressed (S) or downregulated (A) after 10 d of soil flooding. Values ± s.e. are means of three biological replicates. See the Materials and Methods for information relative to the visual scale. An asterisk indicates statistically significant values (P ≤ 0·05) compared with the respective WT value.

Leaf injury, decreased photosynthetic rate and production of ROS are frequently associated with flooding responses (Blokhina and Fagerstedt, 2010; Campbell et al., 2015). Superoxide and H₂O₂ were visually scored in leaf # 2, the youngest leaf first to show substantial injury symptoms as a result of flooding, and the expression of several ROS marker genes in the above-ground tissue of the lines was determined. In the WT line, accumulation of superoxide, estimated by NBT staining, was scattered throughout the leaf blade (Fig. 4A). Leaves from lines in which the ZmPgb genes were upregulated showed a reduced NBT signal. Downregulation of ZmPgb1.1 or ZmPgb1.2 caused heavy accumulation of superoxide. A less defined staining pattern was observed when DAB was used to localize H₂O₂. Relative to the WT, the DAB signal appeared to be more concentrated in specific areas of leaves in which the ZmPgb genes were downregulated, especially along the main vein (Fig. 4A).

Fig. 4.

Generation of reactive oxygen species (ROS) in WT plants and plants in which ZmPgb1.1 and ZmPgb1.2 were overexpressed (S) or downregulated (A) after 10 d of soil flooding. (A) Localization of superoxide (${\text{O}}_{\text{2}}^{\text{–}}$) and hydrogen peroxide (H₂O₂) on the second leaf of soil-flooded plants. (B) Relative expression of the Respiratory Burst Oxidase Homolog (RBOH) genes on above-ground tissue of plants subjected to soil flooding. Values ± s.e. are means of three biological replicates and are normalized to the WT value (set at 100) of the respective day of soil flooding. An asterisk indicates statistically significant values compared with the WT value of the respective day of soil flooding. (C) Relative expression of WRKY6 on above-ground tissue of plants subjected to soil flooding. Values ± s.e. are means of three biological replicates and are normalized to the WT value (set at 100) of the respective day of soil flooding. An asterisk indicates statistically significant values compared with the WT value of the respective day of soil flooding.

Production of ROS in plants is mainly controlled by NADPH oxidases, a family of membrane-bound enzymes contributing to the oxidative burst in a variety of responses (Sagi and Fluhr, 2006). The transcript levels of four Respiratory Burst Oxidase Homolog (RBOH) genes, encoding subunits of the NADPH machinery, were affected relative to the WT by varying the expression of the ZmPgb genes (Fig. 4B). Overexpressing ZmPgb1.1 and ZmPgb1.2 resulted in decreased expression of RBOH(A) during the first 2 d of flooding, while suppression of the Pgb genes increased the expression of RBOH(A). An overall rise of transcript levels in the lines in which the ZmPgb were downregulated was also observed for RBOH(C) and (D) (Fig. 4B).

Another marker of ROS production is the WRKY6 transcription factor gene (Campbell et al., 2015), the expression of which was greatest in the lines in which the ZmPgb genes were downregulated [ZmPbg1.1(A) and ZmPgb1.2(A)] after 2 d of soil flooding (Fig. 4C).

The data indicate that, during flooding, downregulating the ZmPgb genes, an action known to increase cellular NO levels (Hill et al., 2013), results in increased production of ROS, such as ${\text{O}}_{\text{2}}^{\text{–}}$ and H₂O₂, and increased expression of genes associated with ROS production. Constitutively expressing the ZmPgb genes, thus reducing cellular NO, results in the reduction of the effects described above.

Antioxidant response in maize plants during soil flooding

Prevention of ROS formation and oxidative damage is a crucial defence strategy operating during flooding stress (Blokhina and Fagerstedt, 2010). Maintenance of ROS homeostasis is mediated by conserved antioxidant mechanisms regulated by the levels of AsA and GSH. Relative to the WT line, the AsA content was augmented in the above-ground tissue of the ZmPgb1.1(S) and ZmPgb1.2(S) lines, reaching a peak 2 d after flooding (Fig. 5). This was in contrast to the ZmPgb1.1(A) and ZmPgb1.2(A) lines, which had the lowest content of AsA. DHA, an oxidized form of ascorbate, displayed an opposite trend, with the highest levels in the ZmPgb1.1(A) and ZmPgb1.2(A) lines during the initial days of soil flooding before dropping to WT values by day 10. Tissues overexpressing the two ZmPgb genes had the lowest DHA content. GSH and its oxidized form, GSSG, did not fluctuate markedly during the experiment and showed no major differences among lines (Fig. 5).

Fig. 5.

Cellular levels of AsA, DHA, GSH and GSSG in the above-ground tissue of WT plants and in which ZmPgb1.1 and ZmPgb1.2 were overexpressed (S) or downregulated (A) during different days of soil flooding. Values ± s.e. are means of three biological replicates.

Of the three AsA biosynthetic genes analysed, i.e. GDP-l-galactose phosphorylase (GGP), l-galactose-1-phosphate phosphorylase (GPP) and l-galactose dehydrogenase (GDH), the relative expression of GPP and GDH was particularly sensitive to variation in either ZmPgb1.1 or ZmPgb1.2 during flooding (Fig. 6). Constitutive expression of the ZmPgb genes induced the two genes, while downregulation of ZmPgb1.2 had a repressive effect. There were no significant differences in expression of GGP in any of the treatments (Supplementary Data Fig. S5).

Fig. 6.

Relative expression levels of l-galactose-1-phosphate phosphatase (GPP), l-galactose dehydrogenase (GDH), two ascorbate peroxidase (APX) genes and four monodehydroascorbate reductase (MDHAR) genes on above-ground tissue of WT plants and in which ZmPgb1.1 and ZmPgb1.2 were overexpressed (S) or downregulated (A) during soil flooding. Values ± s.e. are means of three biological replicates and are normalized to the WT value (set at 100) of the respective day of soil flooding. An asterisk indicates statistically significant values compared with the WT value of the respective day of soil flooding.

Two key enzymes involved in the recycling of AsA are APX and MDHAR. Of the three APX genes measured, the relative levels of APX3 transcripts were augmented significantly in the ZmPgb1.1(S) and ZmPgb1.2(S) lines and declined (especially APX3) in the lines in which the ZmPgb genes were downregulated during flooding (Fig. 6). Variations in ZmPgb levels did not produce any pronounced change in the expression of APX1 (Supplementary Data Fig. S6).

The four MDHARs that regenerate AsA showed almost identical expression patterns. Relative to the WT, overexpression of either ZmPgb1.1 or ZmPgb1.2 induced the expression of MDHAR1– MDHAR4, especially on day 2, while the downregulation of the ZmPgb genes repressed the levels of MDHAR gene transcripts (Fig. 6). The expression of GR was not greatly affected by alterations in ZmPgb gene levels (Supplementary Data Fig. S7).

The transcript levels of four genes (SOD2, SOD3, SOD4 and SOD9) encoding the superoxide-scavenging SOD were measured in the maize transgenic lines. Relative to the WT, the transcript levels of SOD3 and SOD4 increased in ZmPgb1·hx00B7;1(S) and ZmPgb1.2(S) tissues, in particular 2 d after soil flooding (Fig. 7). An increase also occurred for SOD9 in the ZmPgb1.2(S) tissue at day 10. Downregulation of the ZmPgb genes repressed the expression of the SOD genes on different days during flooding (Fig. 7).

Fig. 7.

Relative expression levels of four superoxide dismutase (SOD) genes and three catalase (CAT) genes on above-ground tissue of WT plants and in which ZmPgb1.1 and ZmPgb1.2 were overexpressed (S) or downregulated (A) during soil flooding. Values ± s.e. are means of three biological replicates and are normalized to the WT value (set at 100) of the respective day of soil flooding. An asterisk indicates statistically significant values compared with the WT value of the respective day of soil flooding.

The relative transcript abundance of CAT genes, encoding the enzyme responsible for the removal of H₂O₂, was also influenced by the levels of the ZmPgb genes. Lines overexpressing ZmPgb1.1 and 2 had the highest expression of CAT2 and CAT3 2 d after flooding (Fig. 7). This was in contrast to lines in which the two ZmPgb genes were suppressed, showing repression of CAT2.

To better interpret the transcription studies, we also measured the activities of the antioxidant enzymes participating in AsA and GSH metabolism during soil flooding. In the lines in which the ZmPgb genes were upregulated, the activity of SOD increased significantly after 2 d (Fig. 8). This was in contrast to the ZmPgb1.1(A) and ZmPgb1.2(A) lines where SOD activity was generally reduced. Similar activity profiles were also observed for APX and especially MDHAR, which showed pronounced differences among the lines, in particular after 2 d of flooding. No distinctive and consistent changes in activity were observed for CAT, DHAR and GR. The only pronounced difference was noticed for the activity of DHAR, which was highest in the ZmPgb1.1(S) line at the beginning of flooding, and GR, which was lowest in the ZmPgb1.2(A) line on the same day (Fig. 8).

Fig. 8.

Enzymatic activity of the major antioxidant enzymes (superoxide dismutase, SOD; catalase, CAT; ascorbate peroxidase, APX; dehydroascorbate reductase, DHAR; monodehydroascorbate reductase, MDHAR; and glutathione reductase, GR) measured during different days of soil flooding on the above-ground tissue of WT plants and in which ZmPgb1.1 and ZmPgb1.2 were overexpressed (S) or downregulated (A). Values ± s.e. are means of three biological replicates.

Therefore, the effect of varying maize Pgb expression on the antioxidant response associated with soil flooding was to alter the expression of genes that protect the plant against ROS. Overexpression of the ZmPgb genes increased the antioxidant capacity and reduced the ROS signal, while their downregulation limited the plant’s capability to protect against ROS

DISCUSSION

The ability of plants to tolerate periods of hypoxia is one of the critical factors that influence plant adaptation to habitat. It is also important in maximizing yield in dry land crops experiencing periods of soil flooding or submergence. The negative effect of flooding stress on plant growth and development is mainly ascribed to the reduced diffusion rate of gasses in water (Voesenek et al., 2006). The consequences of oxygen deprivation include diminished ATP levels, impaired water and mineral transport, an imbalance in nutrient allocation and disturbance of hormone homeostasis (Vartapetian and Jackson, 1997). Periodic cycles of normoxia/hypoxia occur during the day/night transition as a result of prolonged root flooding or submergence. The ROS generated at each transition (Bailey-Serres and Voesenek, 2008) can result in programmed cell death (PCD) and reduced photosynthetic capacity (Chen et al., 2014). Pgb expression influences the plant’s capacity to regulate ROS levels (Igamberdiev et al., 2006b), while the two maize Pgb genes, ZmPgb1.1 and ZmPgb1.2, affect PCD in a cell-specific fashion during maize somatic embryogenesis (Huang et al., 2014). The results presented here indicate that, during maize root flooding, Pgb expression affects genes that minimize cell ROS levels that can affect leaf injury and photosynthesis. In almost all of the foliar parameters measured, overexpression of Pgb during flooding had a positive effect, while downregulation was deleterious with respect to regulating ROS.

While most of the limitation in CO₂ assimilation under flooding conditions is attributed to the low stomatal conductivity (Jackson, 2002), part of it can be the result of undesirable physiological and biochemical reactions, including the production of ROS, impairing the photosynthetic machinery (Yordanova et al., 2004). In all the maize genotypes and transformed lines subjected to soil flooding, the decline in photosynthetic capacity, stomatal conductivity and transpiration rate, without a corresponding reduction in internal CO₂ level (Table 1; Table S2), can be interpreted as a non-stomatal effect on the photosynthetic processes. That is, the closure of stomata might not be limiting CO₂ diffusion in the mesophyll tissue and ultimately photosynthesis.

Altered expression of ZmPgb genes has profound effects on gas exchange parameters during soil flooding. Relative to their WT counterpart, flooded lines overexpressing ZmPg genes are able to retain an elevated photosynthetic rate, stomatal conductivity and transpiration rate, parameters significantly reduced in tissue where the ZmPgb genes have been repressed (Table 1). Thus experimental elevation of Pgb expression can be utilized to enhance plant performance in response to soil flooding, an observation consistent with the correlation between high levels of induced Pgb expression and retention of the photosynthetic rate observed in the flooded genotypes (Fig. 1). This correlation could be exploited by using Pgb genes as molecular markers to predict plant behaviour in response to soil flooding.

Independent studies have integrated Pgbs in plant responses to low oxygen levels. Phytoglobins are rapidly expressed during hypoxia (Silva-Cardenas et al., 2003), and their expression level returns to control values following the reinstatement of normoxic conditions (Dordas et al., 2003). While their function as oxygen sensors and carriers has been discounted (Hill, 1998), Pgbs influence the behaviour of hypoxic cells by altering several physiological functions. Overexpression of Pgb genes in cells exposed to low oxygen tension enhances their survival by maintaining a high energy status (Sowa et al., 1998) and altering the cellular redox state, leading to a lower NAD(P)H/NAD(P) ratio (Igamberdiev and Hill, 2004). These effects are mediated by NO, which is rapidly scavenged by Pgbs (reviewed by Hill, 2012). A more recent study revealed the direct role of Pgbs in modulating cell death and survival through the regulation of oxidative stress and ROS (Huang et al., 2014). Of interest, overproduction of ROS during hypoxia has been linked to injury of tissue resulting from active lipid peroxidation (Crawford et al., 1994), alterations in lipid content (Hetherington et al., 1982) and membrane damage in maize leaves subjected to soil flooding (Yan et al., 1996). It is apparent, however, that the magnitude of these effects is related to the expression of Pgb genes in the tissue, as they are more pronounced in lines in which the ZmPgb genes are downregulated and reduced in those where the ZmPgb genes have been induced. This is suggestive of a role played by Pgbs in attenuating oxidative stress damage in photosynthetic tissue through the limitation of ROS accumulation, a condition compromising plant performance. Transcriptional regulation of the ZmRBOH genes, encoding the subunits of the maize NADPH oxidase, the major generator of ROS (Sagi and Fluhr, 2006), by altered ZmPgb expression (Fig. 4B) suggests that this may be another mechanism by which Pgbs regulate ROS production. The expression of some of the four subunits is repressed in tissue overexpressing the ZmPgb genes and exhibiting limited ROS signal, and induced in those in which the ZmPgb genes are downregulated where ROS accumulation is highest. A similar Pgb regulation of RBOH expression was also described in maize cultured cells (Huang et al., 2014). Another molecular marker of ROS production in maize is WRKY6 (Campbell et al., 2015), the expression of which does not seem to be affected greatly by the ZmPgb genes. This gene is highly induced in senescing tissues often accumulating ROS (Robatzek and Somssich, 2002). This unexpected result might suggest specificity of Pgbs in the regulation of different ROS-producing pathways.

Collectively, these observations suggest that expression of Pgb genes attenuates the deleterious effect of flooding on the photosynthetic rate and plant performance, and that this protective role might be exercised by limiting ROS accumulation in photosynthetic tissue.

The extent to which ROS accumulate depends not only on their synthesis, but also on the efficiency of the ROS-scavenging system which operates through the synergistic activity of key enzymes centred around the ascorbate–glutathione cycle. Sustained transcription and activity of SOD, which catalyses the disproportion of ${\text{O}}_{\text{2}}^{\text{–}}$ to H₂O₂, a crucial response in flooding tolerance (Yordanova et al., 2004), requires ZmPgbs. In lines overexpressing the ZmPgb genes and characterized by a limited ${\text{O}}_{\text{2}}^{\text{–}}$ signal, the transcript levels of SOD3 and 4, as well as SOD activity are highest during flooding. This is in contrast to tissue in which the ZmPgb genes were downregulated where the reduced transcription and activity of SOD correlates with an elevated ${\text{O}}_{\text{2}}^{\text{–}}$ signal. These results support the idea that the varying activity of SOD in different species correlates with their ability to sustain growth in oxygen-limited environments (Monk et al., 1987).

Removal of H₂O₂, the product of the SOD reaction, is often mediated by a wide range of enzymes including CAT and APX. Although it altered the transcript levels of several CAT genes, ZmPgb gene expression did not influence CAT activity, an observation consistent with the unchanged activity of this enzyme following flooding in some species (Hurng and Kao, 1994). It is noteworthy, however, that both transcription and activity of the other H₂O₂-removing enzyme, APX, are highly influenced by the ZmPgb genes. The transcript level of APX3 and 6 (Fig. 6), as well as APX activity (Fig. 8), are increased in tissue overexpressing the ZmPgb genes and diminished to some extent in tissues in which the ZmPgb genes were downregulated. In barley, the activity of both thylakoid-bound and soluble APX in chloroplasts was induced by flooding where it protects the photosynthetic machinery from photogenerated H₂O₂ (Yordanova et al., 2004). During soil flooding, sustained activity of APX by the ZmPgb genes might be facilitated by the high level of its substrate, AsA, produced de novo and/or through recycling mechanisms. A positive correlation exists between endogenous AsA content and ZmPgb gene expression, and this might be in part due to the transcriptional regulation of the AsA biosynthetic enzymes. Both GPP and GDH, key regulators of AsA production (Sanahuja et al., 2013), are induced by the overexpression of the ZmPgb genes and repressed by their downregulation during flooding (Fig. 6).

Recycling mechanisms, essential for relieving ROS accumulation during different forms of stress, including flooding, also contribute significantly to the AsA pool. Of the two ascorbate recycling enzymes, i.e. DHAR and MDHAR, MDHAR seems to be the most influenced by ZmPgbs. During flooding, the transcription of MADHR1–MADHR4 and the activity of MDHAR are induced in tissue overexpressing the ZmPgb genes, but are repressed in those tissues where the ZmPgb genes have been suppressed. A similar Pgb regulation was also documented in alfalfa roots cultured under low oxygen tension (Igamberdiev et al., 2006b). It is worth noting that the observed fluctuations in MDHAR activity might regulate the pool of MDHA required to maintain Pgb in its reduced form and sustain the removal of NO (Igamberdiev and Hill, 2004). Scavenging of NO occurs through the conversion of oxyPgb to metPgb, and the latter is reduced back to oxyPgb in a reaction requiring AsA and producing MDHA (Igamberdiev et al., 2006a). Thus removal of MDHA by MDHAR might accelerate the reduction rate of metPgb.

Compared with MDHAR, the activity profile of DHAR, the other AsA-recycling enzyme, was less defined, with higher values measured in the two ZmPgb gene-overexpressing lines. This observation does not agree with previous work showing elevated DHAR activity in tissue in which Pgb is suppressed (Igamberdiev et al., 2006a), a discrepancy suggestive of different ascorbate regulatory mechanisms operating across species and/or organs of the same species.

Collectively, these results indicate that plants overexpressing the ZmPgb genes are able to retain a larger pool of reduced ascorbate (AsA) through the transcriptional induction of biosynthetic enzymes and enhanced recycling mechanisms where the high activity of MDHAR, and to a lesser extent DHAR, reduces DHA content in favour of AsA. Elevated levels of AsA might sustain the high APX activity needed to attenuate ROS toxicity in photosynthetic tissue. These physiological adjustments are precluded in tissue in which the ZmPgb genes are suppressed. Unlike ascorbate, alterations of the glutathione redox state, often occurring in response to ROS-producing stress (reviewed by Noctor and Foyer, 1998), do not seem to be relevant in our system. Both expression and activity of GR, as well as content of GSH and GSSG are not influenced by the expression of the ZmPgb genes.

The results of this work have implications relative to the development of new genotypes with enhanced ability to tolerate hypoxic stress. The relationship between the ability to retain photosynthetic activity relative to the capability of the genotype to increase Pgb expression (Fig.1) may be valuable as a tool to screen for improved hypoxic stress tolerance in maize or other germplasms. The protocol provides methodology to screen lines reasonably rapidly at an early stage of plant growth in a non-destructive fashion. To produce genetically modified lines, however, there is need for considerably more research before it will be possible to generate successfully transgenic plants with increased tolerance to stress or to develop lines with specific desirable characteristics. Considering the complexity of the potential interactions of NO with Pgbs (Hill, 2012), constitutive expression of Pgbs in commercial genotypes is likely to create as many problems as would be solved. In addition, the cell specificity of Pgb expression (Huang et al., 2014) suggests that cell-targeted expression of the gene may be essential for some traits. These potential sorts of problems have already been demonstrated with respect to barley (Hebelstrup et al., 2014).

In conclusion, this study demonstrates an intimate link between Pgbs and flooding response. Elevated expression of Pgb genes correlates with enhanced performance of plants subjected to flooding. Within the photosynthetic tissue, Pgb genes might limit ROS-induced damage and ensure a sustained photosynthetic rate. This is achieved through improvements of the ascorbate antioxidant status including the enrichment of the AsA pool via de novo and recycling pathways, and increased activities of several ROS-scavenging enzymes.

SUPPLEMENTARY DATA

Supplementary data are available online at www.aob.oxfordjournals.org and consist of the following. Table S1: list of primers used for gene expression studies. Table S2: gas exchange parameters measured on the second leaf of 19 commercial maize genotypes. Figure S1: ratio of the photosynthetic rate (flooded/unflooded) of the 19 corn genotypes. Figure S2: photosynthetic rates and ZmPgb1.1 and ZmPgb1.2 expression in 19 commercial maize genotypes. Figure S3: synchronous emergence and similar growth morphology under normoxic conditions; fresh and dry weight of the unflooded or soil-flooded corn lines; and morphology of soil-flooded plants showing the short root system in plants overexpressing ZmPgb genes. Figure S4: leaf injury in the visual scale used to generate Fig. 2. Figure S5: relative expression levels of GGP on above-ground tissue of WT plants and plants overexpressing or downregulating ZmPgb1.1 and ZmPgb1.2 during soil flooding. Figure S6: relative expression levels of APX1 on above-ground tissue of WT plants and plants overexpressing or downregulating ZmPgb1.1 and ZmPgb1.2 during soil flooding. Figure S7: relative expression levels of GR on above-ground tissue of WT plants and plants overexpressing or downregulating ZmPgb1.1 and ZmPgb1.2 during soil flooding.