## Introduction

Florida ranks first in acreage and total value of snap beans (Phaseolus vulgaris) production, but snap bean productivity is often greatly reduced by heavy rain events resulting from hurricanes and tropical storms. Snap bean is susceptible to flooding and suffers from hypoxic stresses, which results in growth suppression and yield reduction1. According to Holbrook and Zwieniecki2, the primary cause of hypoxic stress is oxygen deficiency in the soil; the oxygen diffusion coefficient is 10,000 times greater in the air (2.14 × 10–1 cm2 s−1) than in water (1.97 × 10–5 cm2 s−1). For most plant species3, oxygen deficiency occurs in the root zone soil with dissolved O2 concentrations less than 2 mg L−1. Thus, snap bean plants in flooded fields receive insufficient bioavailable O2 for normal root metabolism and suffer from waterlogging damage.

In hypoxia, anaerobic metabolism is activated, and oxidative phosphorylation is stopped. The ATP biosynthetic rate derived from glycolysis is low4,5,6,7. In the Krebs cycle, due to the lack of bioavailable O2 as the final electron acceptor, the intermediates build-up, NAD(P)+ levels decrease, pyruvate accumulates, and ATP concentrations decline. These changes negatively affect plant metabolism, including uptake of nutrients such as nitrogen (N) and phosphorus (P) and their assimilation3,8. In addition, plants suffer from ethanol accumulation and toxins formed in anaerobic metabolism9.

Alcohol dehydrogenase (ADH) is a well-studied enzyme in plants and is inducible in roots upon exposure to hypoxic or anaerobic conditions. When the oxygen bioavailability in the root zone is low, ADH activity increases significantly, improving the plant’s tolerance to hypoxia or anoxia10,11. ADH activity is considered essential for the survival of plants in hypoxic or anaerobic conditions12. Thus, ADH activity in root tips of flooded plants is used as an indicator of potential tolerance to flood and flooding severity.

Oxygen fertilization of the root zone is a potential method to mitigate the damage from flooding stress. There are two types of oxygen fertilizers: solid oxygen fertilizers (SOF), such as CaO2 and MgO2, and liquid oxygen fertilizers, such as hydrogen peroxide (H2O2)13,14. The solid oxygen fertilizers listed here release bioavailable oxygen slowly, i.e., slow-release oxygen fertilizer, whereas the liquid oxygen fertilizer, H2O2 liberates bioavailable oxygen rapidly and is considered as a fast oxygen fertilizer. Hydrogen peroxide can be decomposed biologically by the enzyme catalase (EC 1.11.1.16). The decomposition of H2O2 releases 0.5 mol of bioavailable O2 per mole H2O2 as shown in the equation15:

$${\text{H}}_{2} {\text{O}}_{2} \mathop{\longrightarrow}\limits^{Catalase}0.5{\text{O}}_{2} + {\text{H}}_{2} {\text{O}}.$$

In flooded soil, solid oxygen compounds (e.g., CaO2, MgO2) break down to H2O2, which then provides bioavailable oxygen to the rhizosphere, as shown in the equation1,13:

$$\begin{gathered} {\text{CaO}}_{{2}} + {\text{ 2H}}_{{2}} {\text{O }} \to {\text{ Ca}}\left( {{\text{OH}}} \right)_{{2}} + {\text{ 2H}}_{{2}} {\text{O}}_{{2}} \hfill \\ {\text{MgO}}_{{2}} + {\text{ 2H}}_{{2}} {\text{O }} \to {\text{ Mg}}\left( {{\text{OH}}} \right)_{{2}} + {\text{ 2H}}_{{2}} {\text{O}}_{{2}} \hfill \\ \end{gathered}$$

Thus, soil amended with insoluble solid oxygen fertilizers may potentially reduce hypoxic stress in the root zone caused by flooding.

The objectives of this study were to (1) quantify the oxygen bioavailability dynamics of hydroponic solution with H2O2 application, (2) determine the effects of oxygen fertilizer application on N and P uptake by flooded snap bean plants, and (3) evaluate the effect of slow-release solid oxygen fertilizers at different rates on growth and yield of flooded snap bean.

## Results

### Effects of H2O2 application on NH4+, NO3, and P uptake by flooded snap beans in hydroponic solution

In the hydroponic trial 1–1, “flooding” is defined as plants grown in a solution that receives neither aeration nor application of H2O2. In all the treatments, the concentrations of NO3 and P in the nutrient solution declined over time (Fig. 1), which indicated NO3 and P were taken up by the plant. There were differences in the nutrient uptake rates between the control and either aeration or H2O2 application. However, there was not any significant difference in the uptake rate of NO3 between the treatments of H2O2 and aeration (Fig. 2). The NO3 uptake rate was approximately 50% lower with flooding than with the aeration or H2O2 treatment. Phosphorus showed a similar trend as NO3; there was no significant difference between the H2O2 and aeration treatments, but the flooding control had significantly lower P uptake rate (Fig. 2) than each of the H2O2 and aeration treatments. The concentration of dissolved oxygen (DO) was not significantly different between the H2O2 and aeration treatments in the 10 h experiment (Fig. 3). The DO level of the control was significantly lower than each of the H2O2 and aeration treatments after 2 h. This result also indicated that the H2O2 application had a slightly better effect than aeration in oxygenating the hydroponic solution.

In trial 1–2, the plant took up NH4+ first in the H2O2 and aeration treatments, whereas it took up NO3 first in the flooding treatment. For the flooding treatment, the DO levels declined from 3 mg L−1 to approximately 0.01 mg L−1 from 0 to 10 h, the NO3 and NH4+ concentrations reduced with time (Fig. 4). The depletion of NO3 was faster than that of NH4+ from 0 to 8 h, and the NH4+ concentration remained greater than that of NO3. After 10 h, the NO3 and NH4+ concentrations were not significantly different. However, in the H2O2 and aeration treatments (Fig. 4), the concentration of NH4+ decreased faster than that of NO3 during 0 to 8 h, and the concentration of NO3 remained greater than that of NH4+. The DO levels of the H2O2 treatment remained greater than 8 mg L−1 during 0 to 10 h and declined to 3 mg L−1 at 20 h, which indicated that H2O2 had to be applied every 20 h to maintain a satisfactory DO concentration.

For trial 1–3, the DO level in the solution was 8 mg L−1 at the beginning, which was air saturated. The DO levels were consistently and significantly greater with H2O2 (Fig. 5) than with aeration (8 mg L−1) during 0 to 20 h. However, the DO level dropped to approximately 0.01 mg L−1 after 20 h, which indicated a hypoxic stress occurred before the beginning of day 2. At the beginning of day 2 before H2O2 was applied, the uptake rate of NO3 was significantly greater than that of NH4+ which was also observed in the flooding treatment in trial 1–2 (Fig. 4). Nitrate contains 3 atoms of oxygen and can provide oxygen served as the terminal electron acceptor in cellular respiration, but ammonium does not contain any oxygen. The reversal of this trend occurred 1 h after the H2O2 was applied at the beginning of day 2 and occurred 3 h after the H2O2 was applied on day 3.

### Effects of SOF application on flooded plants in containers

The flooded plants were significantly taller with CaO2 or MgO2 than without peroxide application. Plant height without flooding was significantly taller with oxygen fertilizer than without oxygen fertilizer. The non-flooded plants with 4 g CaO2 and 4 g MgO2 had the highest height (Table 1).

Shoot biomass of flooded plants was greater with CaO2 or MgO2 than without SOF (Table 1). In the non-flooding treatment, only those with 2 g MgO2 and without SOF had the lowest plant biomass (Table 1).

The pod yields of flooded snap bean were greater with SOF treatments than the control (Fig. 6). The 2 g CaO2 and 8 g MgO2 treatments had almost 50% greater yield than the control, which was significantly greater than that of rest of the treatments (Fig. 6).

As a bio-indicator, ADH activity is used to evaluate the effects of different amounts of SOF on the growth and development of flooded plants. The ADH activities were significantly lower with all the SOF treatments than with the flooding treatment. For example, ADH activity was in a range from 44 nmol NADH per minute per mg protein for the flooded snap bean plants without oxygen fertilization to 5 nmol NADH per minute per mg protein for those flooded plants with 4 g CaO2 (Fig. 6). The lower ADH activities indicated the SOF treatments had much less flooding stress compared to the flooding treatment. the treatments of 2 g CaO2, 4 g CaO2, and 8 g MgO2 had significantly lower ADH activities than the other treatments (Fig. 6), which indicated those treatments had even less or no flooding stress compared to the other treatments. This result provided physiologic evidence for evaluation of the potential of SOF on mitigating the flooding stress.

## Discussion

In trial 1, the H2O2 treatment had a positive effect on the N and P uptake of the plants in hypoxic stress. The lower uptake rate of flooded plants was possibly caused by ATP shortage, resulting from flooding-induced oxygen deficiency and anaerobic respiration. Thus, the application of H2O2, supplying O2 directly, was a potential method to alleviate the hypoxia damage. This finding is supported by other studies. Jampeetong and Brix16 found that Salvinia natans plants supplied with NO3 grew normally in a hypoxic solution but were greatly inhibited when only NH4+ was supplied as the nitrogen source.

In trial 2, the results showed that the application of SOFs enhanced the growth and development of snap bean plants in flooding stress. Lakitan17 reported that the total yield and survival of snap beans were susceptible to flooding imposed after reproductive development had begun (28 DAP), and flooding imposed at later growth (> 28 DAP) significantly reduced the total yields. Few plants survived, and there was no measurable yield when flooding was imposed at 36 DAP. More trials using SOF with different flood durations need to be conducted at various plants growth stages in the future.

After uptake, nitrogen assimilation follows a few steps including from the reduction of NO3 to NH4+ and the subsequent incorporation of NH4+ into amino acids18. NO3 reduction is catalyzed by the enzyme nitrate reductase (NR, EC 1.6.6.1-3). Nitrate reductase reduces NO3 to NO2, which is then transferred to the chloroplast to be reduced to NH4+ by nitrite reductase (NiR, EC 1.6.6.4). After that, NH4+ is used to biosynthesize glutamine by glutamine synthase (GS, EC 6.3.5.1)19. Nitrate reductase is a key enzyme tightly regulated by many environmental factors, such as flooding20,21. Anoxia and hypoxia increase NR activity, and NO3 without available molecular oxygen is used to oxidize organic molecules to obtain energy, a phenomenon called “nitrate respiration”22. That study may support the result that snap bean absorbs more NO3 than NH4+ in hypoxia because the increased NR and NiR activity would promote NO3 uptake and assimilation. Early trials used labeled15 NO3 to monitor nitrate assimilation during anaerobic germination of rice and showed that 15N was incorporated into amino acids23. The results confirmed that exogenous NO3 supplied during oxygen deficiency had been assimilated by plants23. Even if NO3 supply to flooded crops increases the percentage of survival, the biochemical and molecular mechanisms underneath its positive effects are not fully understood. Thus, future research will be needed to invest the mechanisms for plant surviving from flooding with NO3 amendment.

The results of this study showed the trend that the supplement of NO3 may enhance the NH4+ uptake (Fig. 4). For example, ample supply of NO3 was able to alleviate the consequences of root anoxia or hypoxia in barley9. It has been hypothesized that NO3 can serve as an alternative terminal electron acceptor in hypoxic or anoxic conditions8. Nitrate may alleviate the consequences of anoxia in other ways. As an ample supply of NO3 can directly induce nitrate reductase activity24, whereby NADH is diverted from the reduction of acetaldehyde to ethanol. The accumulation of injurious ethanol concentrations in the cells is delayed. Also, NO3 may influence rhizosphere redox potential and may act as an oxygen source through influx into O2-deficient roots or by uptake in partially aerated shallow roots9.

The results of this study indicated that the peroxides, CaO2 and MgO2 have the potential to mitigate the flooding damage in snap beans. Several researchers have observed that prolonged flooding causes a cessation of root and shoot growth, wilting, decreased nutrient uptake, and often plant death25. According to Lakitan17 flooding on different growth stages of snap bean had different effects on growth and yield, and there is evidence that longer flooding duration would result in lower survival and yields17. Thus, further studies are warranted to assess the effects of solid oxygen fertilizations on reducing flooding stress, and survival rates of snap beans at different growth stages and with different duration of flooding treatments.

## Conclusions

Snap bean is susceptible to hypoxic stress and often suffers from flooding, particularly in the hurricane and tropical storm seasons. We conducted both hydroponic and pot trials to investigate how oxygen fertilization affects uptake rates of NH4 and NO3 and P, growth, and yields of flooded snap bean. Three peroxides, including H2O2, CaO2, and MgO2, were evaluated. The findings from this study indicated oxygen fertilization had the potential to minimize flooding damage to flooded snap bean. Hydrogen peroxide application significantly increased the dissolved oxygen level and the uptake rate of NH4+, NO3 and P by snap bean. Snap bean had an uptake preference for of NH4+-N with sufficient bioavailable oxygen. However, NO3-N is preferred in flooding conditions. The right dosage of solid oxygen fertilizer of 2 g CaO2 or 8 g MgO2 had the greatest yield among all flooded snap bean plants. Thus, both CaO2 and MgO2 oxygen fertilizations had the potential to keep absorbing nutrients such as N and P by flooded snap bean plants, reduce flooding damage, and minimize the yield loss. These novel methods of oxygen fertilization may mitigate economic loss and benefit agriculture often facing different hypoxic stresses.

## Materials and methods

### Plant materials and soil

Snap bean (P. vulgaris. cv. ‘Bronco’) is a primary vegetable crop and widely cultivated in USA. The ‘Bronco’ seeds are commercially available for research and breeding and were purchased from a seed company, HOSS TOOLS located in Norman Park, GA (GPS coordinates: Latitude: 31.2462 and Longitude -83.6549) before the study was started in spring 2015. Miracle-Gro® garden soil was procured from Lowes in Gainesville, FL.

In Trial 1, snap bean seeds were soaked in 0.15% H2O2 for 24 h and rinsed with deionized (DI) water for 5 times, and then were placed into a 12-cm petri dish with moistened filter paper and sealed inside a zip-lock plastic bag and incubated at 33 °C overnight. After seedling emergence, each seedling was transplanted into a 5.1-cm diameter plastic mesh basket. The baskets (Net Pots Net Cups, Heavy Duty Plastic Net Pot with Wide Rim Design) were placed in an aeroponic system to allow for rapid root growth and development. The aeroponic system consisted of a tank half-filled with 10% Hoagland solution26, a misting pump, tubing, and three sprinklers that continuously misted the roots. After two weeks, the seedlings with three true leaves were transferred from the aeroponic system to a hydroponic system where the roots were submerged in a 1000 mL 0.2 mM CaSO4 solution for measurement use. The use of 0.2 mM CaSO4 solution was to protect the membranes, since calcium is essential for the intactness and selectivity of biological membranes27.

In trial 2, two sizes of pots were used for different purposes. The smaller pots with 7.6 cm in diameter × 8.9 cm in height was for seed germination. The larger pots with 15 cm in diameter × 16.5 cm in height for production. Miracle-Gro® garden soil was used, and the fertilizers were applied per pot as follows: NH4NO3, 2.92 g; triple superphosphate, 1.65 g; potassium chloride, 0.44 g per pot.

#### Experimental design

The trials were all conducted in the Horticultural Sciences Department, University of Florida/IFAS, Gainesville, Florida. The hydroponic trials had three sub-trials. Trial 1–1 and 1–2 had the same treatments with four replicates. The trial 1–1 was conducted in an environmental-controlled growth chamber with a temperature range of 23–27 °C, 16-h daylight and 8-h dark. The trial 1–2 and trial 1–3 were conducted in a standard greenhouse with a temperature range of 23–27 °C. The trial 2 was conducted in a high tunnel under a natural atmosphere. The treatments were imposed by (1) flooding control without bubbling or applying H2O2 after the pre-set time (approximately 100 μM DO level at the beginning, no oxygen supplement), (2) aeration (continuous oxygen supply with an air pump), and (3) H2O2 (application of 529 μM H2O2 at 0 h). The growth medium was 1 L DI water with 0.22 g L−1 KH2PO4, 0.373 g L−1 KNO3, 0.02 g L−1 CaSO4, 0.012 gL−1 MgSO4, 0.29 mg L−1 H3BO3, 0.18 mg L−1 MnCl2∙4H2O, 0.02 mg L−1 ZnSO4∙7H2O, 0.008 mg L−1 CuSO4∙5H2O, 0.002 mg L−1 H2MoO4∙H2O, 0.65 mg L−1 Ethylenediaminetetraacetic acid (EDTA), and 0.62 mg L−1 FeSO4∙7H2O. For trial 1–1, 2 mL samples were collected from the center of pots every 2 h for 10 h. The samples were analyzed for ortho-P (EPA Method 365.3), and NO3 (EPA Method 352.1) using a Seal AQ-2 discrete analyzer (2006 SEAL Analytical Ltd. Mequon, WI) (Fig. 1).

For trials 1–2, Neulog™ Dissolved oxygen sensors, Neulog NO3 sensors, and Neulog NH4+ sensors were set up for each container to monitor concentration changes (Fig. 7). Those three sensors were combined and installed in each container, and aluminum foil was used to hold the sensors and plants and block the sunlight. The data were monitored every 10 min for 20 h and recorded with NeuLog™ Dissolved Oxygen sensor, Nitrate logger sensors, and Ammonium logger sensors (Neulog brand dissolved oxygen sensor, EISCO Scientific, Rochester, NY, USA).

Trials 1–3 used the same sensors and settings as trials 1–2 to monitor the consistently dynamic nitrogen changes for 96 h (96 h). The growth medium was 1 L DI water with the same ingredients as trial 1–1 but used 0.57 g L−1 NH4H2PO4 and 1.8 g L−1 KNO3 instead of 0.22 g L−1 KH2PO4 and 0.373 g L−1 KNO3. Additionally, 0.57 g L−1 NH4H2PO4 and 1.8 g L−1 KNO3 were applied every 24 h to provide 5 mg L−1 of NO3 and 5 mg L−1 of NH4+. The data were recorded every 30 min for 4 days (96 h).

In trial 2, a completely randomized design was used with 4 replications. There were two SOFs at different rates, i.e., CaO2 at 0, 1, 2, and 4 g per pot and MgO2 at 0, 2, 4, and 8 g per pot. The seeds were planted in 7.6 cm pots for two weeks and were transferred to the 15.2-cm pots (internal volume: 2,622 cm3) after germination. The SOF, CaO2 or MgO2 was mixed with the soil (approximate 2,000 cm3 per pot), the mixed soil was added into the 15.2-cm pots, and then the seedlings were transplanted into the pots, respectively. One seedling was planted per pot. The fertilizer rates were based on the vegetable production handbook of Florida28. The N, P and K application rates were 112 kg ha−1, 134 kg ha−1 and 134 kg ha−1 for the whole season. After one week, both groups with CaO2 and MgO2 were each divided into two subgroups. One subgroup was individually flooded in 18.9 L buckets filled with water to simulate the flooding conditions for two days, while another subgroup was not flooded.

The data of leaf greenness representing leaf chlorophyll content determined by a SPAD meter (Knoica-Minolta, Osaka, Japan), plant height, ADH activities, and shoot biomass were measured when the flooded plants started wilting. To collect yield data, this trial was identically repeated and the yields were measured 55 DAP.