Generation of hydroxyl radicals by Fe-polyphenol-activated CaO2 as a potential treatment for soil-borne diseases

An Fe-polyphenol catalyst was recently developed using anhydrous iron (III) chloride and coffee grounds as raw materials. The present study aims to test the application of this Fe-polyphenol catalyst with two hydrogen peroxide (H2O2) sources in soil as a new method for controlling the soil-borne disease caused by Ralstonia solanacearum and to test the hypothesis that hydroxyl radicals are involved in the catalytic process. Tomato cv. Momotaro was used as the test species. The results showed that powdered CaO2 (16% W/W) is a more effective H2O2 source for controlling bacterial wilt disease than liquid H2O2 (35% W/W) when applied with an Fe-polyphenol catalyst. An electron paramagnetic resonance spin trapping method using a 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) assay and Fe-caffeic acid and Fe-chlorogenic acid complexes as models showed that these organometallic complexes react with the H2O2 released by CaO2, producing hydroxyl radicals in a manner that is consistent with the proposed catalytic process. The application of Fe-polyphenol with powdered CaO2 to soil could be a new environmentally friendly method for controlling soil-borne diseases.

Hydroxyl radical assay. The results of the electron paramagnetic resonance (EPR) experiments are shown in . The presence of the 5,5-dimethyl-1-pyrroline-N-oxide (DMPO)-OH radical was confirmed by the observed hyperfine coupling constants (hfcc) of aN = aH = 1.49 mT 17 . Figure 4 shows the spectra of the DMPO-OH radical after 30 s of reaction in the following systems: (a) CaO 2 , (b) Fe-CPP/CaO 2 and Fe-CPP, (c), Fe(II)/CaO 2 and Fe(II), and (d) Fe(III)/CaO 2 and Fe(III). Systems that not received liquid or powdered CaO 2 as H 2 O 2 no signals of DMPO-OH radical were detected. On the other hand, CaO 2 and Fe(III)/CaO 2 systems showed DMPO-OH radical signals among the treatments that received liquid or powdered CaO 2 as H 2 O 2 source. The signals characteristics of the DMPO-OH radical were also detected in the Fe-CA/CaO 2 and Fe-CGA/CaO 2 model systems (Fig. 5). When dimethyl sulfoxide (DMSO) was added to the reaction systems, DMPO-CH 3 (the spin adduct of methyl radical, hfcc: aN = 1.64 mT, aH = 2.35 mT) 18 was observed, and the intensity of the signals for the DMPO-OH radical decreased (Fig. 6). Figure 7 shows the EPR spectra and the yield of the DMPO-OH radical generated after 30 s of reaction in the powdered CaO 2 systems. Quantitative analysis revealed that the yields of the DMPO-OH radical generated by CPP-Fe/CaO 2 were 1.3-, 1.7-and 3.3-fold higher than those generated by the Fe-CA/CaO 2 , Fe-CGA/CaO 2 and Fe(III)/CaO 2 systems, respectively. However, no differences (p < 0.05) were found between the amounts of the DMPO-OH radical generated after 30 s of reaction time in the Fe-CPP/CaO 2 and the Fe(II)/CaO 2 systems. The amount of hydroxyl radical generated after 30 s of reaction time followed the order Fe-CPP = Fe(II) > Fe-CA > Fe-CGA >> Fe(III).

Discussion
In previous experiments, the results of an XPS survey revealed that both ferric iron (Fe 3+ ) and ferrous iron (Fe 2+ ) were present in the Fe-polyphenol catalyst but no zerovalent iron (nZVI) was present. Iron was present in the forms of Fe 2 O 3 /FeCl 2 and FeCl 3 14 . On the other hand, more than 98% of the iron released from the Fe-polyphenol catalyst was in the Fe 2+ form as detected by the phenanthroline method 15 . The results of in vitro experiments Figure 2. Effect of the treatments on the population of Ralstonia solanacearum in the soil after the growth period. CaO 2 = calcium peroxide (16% CaO 2 ) (2 g kg −1 dry soil); CPP = coffee polyphenols applied in the form of coffee grounds (2 g kg −1 dry soil); Fe-CPP = Fe-polyphenol catalyst developed using coffee grounds (2 g kg −1 dry soil); Fe(II) = iron (II) sulfate heptahydrate (0.18 g kg −1 dry soil); Fe(III) = anhydrous iron (III) chloride (0.36 g kg −1 dry soil  showed that the Fe-polyphenol catalyst can be used to supply iron to leaf vegetables 12 and rice 13 , and in vitro experiments under laboratory conditions showed that when applied in conjunction with liquid H 2 O 2 , this catalyst could disinfect pathogens such as Escherichia coli 14 and Ralstonia solanacearum (see Supplementary Figs S2, S3, and S4) or remove methylene blue from water systems 15 . We proposed a mechanism involving the generation of hydroxyl radicals by the reaction between the iron catalyst and H 2 O 2 . In the present study, the same Fe-polyphenol catalyst was prepared and applied with two H 2 O 2 sources with different H 2 O 2 release rates to suppress the bacterial wilt disease caused by R. solanacearum, which is one the most difficult soil-borne disease to control because the bacteria can survive in various environments 19,20 .  A chemiluminescence method based on luminol was used to verify the presence of ROS 21 by adding L-ascorbate, which can scavenge ROS, decreasing the chemiluminescence intensity. A high chemiluminescence intensity was found for all treatments, and the addition of L-ascorbate dramatically reduced the emission intensity of luminol, indicating the presence of radical species in all systems. The high luminol intensity observed in the Fe-CGA/CaO 2 and Fe-CA/CaO 2 model systems suggests that chlorogenic acid and caffeic acid may be associated with the generation of ROS in the Fe-CPP/CaO 2 system since these acids are the predominant polyphenols found in coffee grounds [22][23][24][25] . Luminol is a good indicator of the presence of ROS but cannot identify specific radicals because it emits chemiluminescence with all kinds of radicals, such as ·OH, ·O 2 − and 1 O 2 . Hydroxyl radicals are the most reactive and least selective ROS 26 , and they could play a role in the results of this experiment. To test this  hypothesis, a series of EPR experiments using DMPO as a spin trap were carried out. The results are shown in  No signals for DMPO-OH radicals were detected in the systems without an added H 2 O 2 source (Figs 4 and 5). In addition, as shown in Fig. 6, when DMSO was added to the reaction systems in which the DMPO-OH radical was detected, a signal for the DMPO-CH 3 radical was observed, and the intensity of the DMPO-OH radical signal decreased.
DMPO-CH 3 is produced through the oxidation of DMSO by hydroxyl radicals, indicating that the DMPO-OH radical signal detected by EPR analysis represents the generation of hydroxyl radicals 27,28 rather than the nucleophilic addition of water 29 . Thus, coffee grounds might contain polyphenols that can contribute to the generation of hydroxyl radicals when bound to iron as a catalyst in the Fenton process. In addition, the hydroxyl radicals generated by the modified Fenton system using the Fe-CPP catalyst might contribute to the lethal oxidative damage to the bacterial cells 30 occurring in the studied soil. These results show that hydroxyl radicals were the major ROS in the Fe-CPP/CaO 2 and Fe(II)/CaO 2 systems and agree with those showing that hydroxyl radicals are the major ROS in Fe(II)/CaO 2 systems 31 .
The present study demonstrated that the generation of hydroxyl radicals by the reaction of CaO 2 with an Fe-polyphenol catalyst developed using coffee grounds was associated with the observed bactericidal effects. Hydroxyl radicals have the highest oxidation potential (2.76 V) among ROS and are generated in the reaction between iron (II) as a catalyst and H 2 O 2 as an oxidant 32 . The disease incidence was drastically reduced by the Fe-CPP/CaO 2 treatment compared to the Fe-CPP/H 2 O 2 treatment. This effect remained until the fruiting stage (see Supplementary Fig. 4S). These results agree with recent studies suggesting that CaO 2 is a more effective source of H 2 O 2 than liquid H 2 O 2 for in situ chemical oxidation [33][34][35] . The chemical oxidation capacity of CaO 2 is dependent on the generation of H 2 O 2 (equation (1)) and the subsequent production of hydroxyl radicals from the released H 2 O 2 (equation (2)) 36,37 . The advantage of this reaction is that the concentration of released H 2 O 2 is autoregulated by the rate of CaO 2 dissolution, which reduces the disproportionation of H 2 O 2 in the media since not all the H 2 O 2 is available at once, as is the case with liquid H 2 O 2 38 . In our experiments, the lower efficacy of liquid H 2 O 2 compared with that of powdered CaO 2 as a source of H 2 O 2 was obvious and could be explained through the rapid decomposition of liquid H 2 O 2 that occurs in soils. These factors limit the applicability of the modified Fenton process for in situ chemical oxidations 35 . The most important limitation of the conventional Fenton reagent is the instability of the large amount of hydroxyl radicals instantaneously produced from liquid H 2 O 2 34,35 . The excess H 2 O 2 could act as a scavenger and compete for hydroxyl radicals 39,40 , inhibiting the oxidation of bacterial cells. In this study, the release of H 2 O 2 was autoregulated by the rate of CaO 2 dissolution, which prevented all the H 2 O 2 from being available at once, as it is when liquid H 2 O 2 is used as the reagent 34 . As a result, the bactericidal effect of the H 2 O 2 reaction with Fe-polyphenol increased when CaO 2 was used. On the other hand, the amount of hydroxyl radicals produced by the Fe-polyphenol-activated CaO 2 was estimated to be much higher than that generated by the Fe(II) or Fe(III) catalysts, which was verified by EPR spectroscopy (Fig. 7).
In our experiment, the failure of the Fe(II) and Fe(III) catalysts to reduce the incidence of wilt disease when applied with either source of H 2 O 2 was studied (Fig. 1). These results can be explained by the lower total radical concentration produced by the Fe(III)/CaO 2 and Fe(II)/CaO 2 systems than that produced by the Fe-CPP/CaO 2 treatment. The weak effect of the Fe(III)/CaO 2 treatment on wilt disease could be attributed to the low reactivity of Fe(III) with H 2 O 2 , which results in a lower content of OH radicals produced. Compared to other catalysts, the Fe(III) catalyst produced a lower yield of hydroxyl radicals when reacted with the same amounts of H 2 O 2 and powdered CaO 2 (Fig. 7). The Fe(III)-activated CaO 2 exhibited several limitations, such as precipitation of the iron as ferric hydroxide (Fe(OH) 3 ), which does not readily redissolve and inhibits the oxidation process 41 . The addition of chelating agents such as citric acid, tartaric acid, oxalic acid, and glutamic acid has been proposed as a way to overcome these drawbacks 41,42 . We believe that the caffeic acid and chlorogenic acid present in coffee grounds probably contributed to the Fenton process by reducing Fe 3+ to Fe 2+ and/or served as electron donors binding Fe 2+ to maintain the activity of Fe in the reduced state in the Fenton cycle.
A single application of H 2 O 2 to the soil did not reduce the disease incidence. Usually, a solution containing 588 to 3529.4 mmol L −1 H 2 O 2 is used in the in situ chemical oxidation process 43 , but the half-life of H 2 O 2 at these concentrations is only minutes to hours. These degradation rates are much higher than that of the 1.5 mmol L −1 H 2 O 2 solution used in this experiment.
For the in situ chemical oxidation process, iron can be added as Fe 2+ or Fe 3+ salts 44 or as native iron-containing minerals such as goethite and ferrihydrite 45,46 . The low solubility of Fe 3+ at neutral pH necessitates the use of chelators to increase the Fe 3+ concentration in the aqueous phase 47,48 . Citric acid, oxalic acid, ethylenediaminetetraacetic acid, 1,4-benzenedicarboxylic acid, N,N-dimethyl formamide and tartaric acid have been successfully applied as Fe 3+ chelating agents for the Fenton process 36,49 . If insufficient Fe 2+ is added or if only Fe 3+ is present initially, Fe 2+ is regenerated through various reactions 50 .
Our results are consistent with those of other studies 27,51 . The detected EPR signals together with the results of the scavenging tests with L-ascorbate indicated that hydroxyl radicals were the major ROS in the Fe-CPP/CaO 2 , Fe-CA/CaO 2 , Fe-CGA/CaO 2 and Fe(II)/CaO 2 systems but not in the Fe(III)/CaO 2 system, as no DMPO-OH radical signal was detected in this system. The peaks of O· 2 − were not confirmed in the EPR analyses of all the treatments, indicating that low concentrations of O· 2 − were generated in the systems studied.  Figure 8 shows the proposed mechanism for the treatment of soil-borne disease by the CAF-Fe activation of powdered CaO 2 . First, Fe 3+ is reduced to Fe 2+ , and then the Fe 2+ forms a complex with the coffee polyphenols. The Fe 2+ -polyphenol species react with the H 2 O 2 from the calcium peroxide to generate •OH radicals. Finally, the •OH radicals oxidize the bacterial cells in the soil. We proposed that the coffee polyphenols such as chlorogenic acid and caffeic acid used in our study reduced and chelated the iron, creating conditions that favour the oxidation of bacterial cells in the soil environment by the Fenton process. Generally, hydroxyl radicals are generated from electron transfer between the complex of H 2 O 2 and iron sites. The electron-rich organic ligands could donate electrons to the Fe ions 51 . Coffee polyphenols probably contributed to the Fenton process by reducing Fe 3+ to Fe 2+ and/or served as electron donors to maintain the activity of Fe in its reduced state in the Fenton cycle. Reduction of Fe 3+ generates Fe 2+ , which can participate in the Fenton reaction and generate ROS 52,53 .
Regardless of the investigated Fe-polyphenols and CaO 2 as an advancement in soil-borne disease control, further investigations are required to evaluate the injection mode of these particles in soils. The developed method could reduce the dependence on high-risk chemicals for disease management, and this method is ecologically sound and environmentally friendly. Evaluating the effectiveness of CPP-Fe/CaO 2 for controlling soil-borne disease on a large scale is difficult because few controlled studies on the rate of dissolution of CaO 2 and the yield of H 2 O 2 in different types of soil and on the stability of the CPP-Fe material in soil have been reported. The efficiency of the treatment will significantly depend on the contact between the bacteria and the catalyst with the CaO 2 particles. Therefore, particles with a high mobility must easily reach the contaminated target soil layers. Other factors such as soil pH, natural scavengers, soil texture, and water content could alter the effectiveness of Fe-polyphenol-activated CaO 2 for controlling soil-borne disease in field conditions. The release rate of H 2 O 2 from CaO 2 is autoregulated by the rate of CaO 2 dissolution, which can be controlled by adjusting the pH 54 . Carbonate and bicarbonate buffer species act as radical scavengers in the Fenton process 55 . Thus, the soil pH could certainly alter the effectiveness of the CPP-Fe/CaO 2 treatment. The carboxylate or phenolic functional groups in natural organic substances could act as a ligand for Fe(II), scavenge hydroxyl radicals, or reduce ferric oxides altering the effectiveness of Fenton or Fenton-like reactions 56 . Humic acid can act as a free-radical scavenger, as a radical chain promoter, and as a catalytic site inhibitor 56,57 . Fenton oxidation and ·OH production were enhanced in the presence of peat by one or more peat-dependent mechanisms 58 . The Fe concentration and availability in the peat, the reduction of Fe 3+ to Fe 2+ by the organic matter, and the reduction of organic-complexed Fe 3+ to Fe 2+ were probable causes of this enhancement. In addition, microbial activity may also be responsible for hydrogen peroxide decomposition 59 .
The presence of inorganic components in the soil could affect the generation of ·OH. Ammonium sulfate and monobasic sodium phosphate have been used to stabilize hydrogen peroxide 60 . Of the four inorganic stabilizers (i.e., monobasic potassium phosphate, dibasic potassium phosphate, sodium tripolyphosphate, and silicic acid) for hydrogen peroxide, monobasic phosphate was found to propagate hydrogen peroxide over the longest distance in soil columns 61 ; however, monobasic phosphate was depleted by adsorption and may also function as a radical scavenger 60 . Those stabilizers could increase the effectiveness of the CPP-Fe/CaO 2 treatment.
The mobility of the Fe-CPP and CaO 2 particles in soils (i.e., saturated and unsaturated zones) should be investigated prior to in situ applications. The effect of Fe-CPP/CaO 2 treatment on soil quality and native microbiota should be investigated. Prior to field or in situ applications, feasibility studies are necessary to determine the extent and rate of bacterial oxidation on a batch scale.

Conclusion
From the results obtained in this work, we conclude that the polyphenols in coffee, such as caffeic acid and chlorogenic acid, play an important role in the generation of hydroxyl radicals in the Fe-polyphenol catalyst developed using coffee grounds. The developed catalyst is low-cost, has a low toxicity and could be used as an environmentally friendly method for suppressing the incidence of soil-borne diseases. However, the feasibility of this method on the field scale needs to be verified. Material and Methods Chemicals. 2,3,5-Triphenyl tetrazolium chloride and 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) were purchased from Tokyo Chemical Industry Co., Japan. Anhydrous iron (III) chloride was obtained from Kanto Chemical, Japan. H 2 O 2 (35% W/W), agar (powder), chloramphenicol, crystal violet, cycloheximide, polymyxin B sulfate, calcium peroxide (CaO 2 ), caffeic acid (CA, 3,4-dihydroxycinnamic acid), chlorogenic acid (CGA), iron(II) sulphate and phosphate buffer (pH 7.4) were purchased from Wako Pure Chemical Industries, Japan. Casamino acids, peptone and dextrose were purchased from Becton Dickinson and Co., Sparks, United States. Other chemicals were of reagent grade and were used as received without further purification. Coffee grounds were collected from a coffee beverage company (AGF Co., Suzuka, Japan).

Synthesis of iron catalysts.
Eighty-eight grams of coffee grounds was mixed with 12 g of anhydrous iron (III) chloride (Fe(III)) and 300 mL of water. The mixture was heated to 98 °C for 24 hours and then dried at 82 °C for 48 hours. The coffee grounds-iron mixture was subsequently ground before the experiments 15 .
CA and CGA, which are the main polyphenols in coffee [22][23][24][25] , were reacted with iron and used as Fe-polyphenol models to clarify the role of these Fe-polyphenol complexes in the activation of CaO 2 and the generation of hydroxyl radicals. The Fe-CGA and Fe-CA complexes were prepared with deionized water. A total of 252.2 mg of CA and 496.0 mg of CGA were individually mixed with 227.1 mg L -1 of anhydrous iron (III) chloride (Fe(III)).
Soil-borne disease assessment. Tomato cv. Momotaro was used as the test specie. For the inoculum, Ralstonia solanacearum MAFF301487 4 (see Supplementary Fig. S5) was cultured in 1 L of casamino acid-peptone-glucose medium (CPG medium) (0.1% casamino acid, 1% peptone, and 0.5% glucose, pH 7.0) in a sealed 500 mL Erlenmeyer flask at 32 °C for 3 days in the dark with continuous shaking. All treatments, except for the negative control treatment ((−) CNT), were inoculated with this bacterial solution. Two hundred and fifty grams of previously sterilized gardening soil (NIPPI, Nihon Hiryo Co., Tokyo, Japan) was placed in a polyethylene plant pot (9.2 cm × 8.2 cm, Asahikasei, Tokyo, Japan) and inoculated with the bacterial solution to a final R. solanacearum population of 5.0 log CFU g −1 dry soil. Then, the following treatments were applied: no inoculation of an R. solanacearum treatment: 1. negative control: no application of any material ((−) CNT); inoculation of R. solanacearum treatments: 2. positive control: no application of any material ((+) CNT); 3. 300 mL of 1.5 mmol L −1 liquid H 2 O 2 (H 2 O 2 ); 4. powdered CaO 2 (16% W/W); 5. coffee polyphenols from coffee grounds (CPP); 6. Fe-polyphenol catalyst developed using coffee grounds (Fe-CPP); 7. Fe-CPP and liquid H 2 O 2 (Fe-CPP/H 2 O 2 ); 8. Fe-CPP and powdered CaO 2 (Fe-CPP/CaO 2 ); 9. iron (II) sulfate heptahydrate and liquid H 2 O 2 (Fe(II)/H 2 O 2 ); 10. iron (II) sulfate heptahydrate and CaO 2 (Fe(II)/CaO 2 ); 11. anhydrous iron (III) chloride and liquid H 2 O 2 (Fe(III)/H 2 O 2 ); and 12. anhydrous iron (III) chloride and powdered CaO 2 (Fe(III)/CaO 2 ). Both the liquid H 2 O 2 (35% W/W) and powdered CaO 2 (16% W/W) treatments were applied at the same final concentrations (4.42 mmol H 2 O 2 kg −1 dry soil). The catalysts Fe-CPP, iron (II) sulfate heptahydrate (Fe(II)) and iron (III) chloride anhydrous (Fe(III)) were applied at the same final concentrations (1.5 mmol Fe kg −1 dry soil) in their respective treatments. Each treatment was repeated three times (twelve pots per replicate) with one plant per pot. The disease incidence was assessed by counting the wilting plants at weekly intervals for 42 days postinoculation. The populations of R. solanacearum in the soils at the end of the experiment were estimated using a selective medium 16 . Tomato seeds were sown in a tray, and the seedlings were transplanted when they reached 10 cm in height. The soil moisture level does not affect Ralstonia solanacearum populations except in instances of severe drought. To minimize the effect of drought on the bacterial populations, water was continuously provided by placing the pots in a tray in which the water level was maintained at 5 mm from the bottom by frequent watering.
Reactive oxygen species (ROS) assay. A chemiluminescence assay 62 was carried out to determine the total amount of ROS generated in the reaction of CaO 2 with the Fe-CPP, Fe(II), Fe(III), Fe-CA and Fe-CGA catalysts. Fifty microlitres of each iron catalyst solution containing 1.5 mmol L −1 of Fe was transferred to a tube and placed in a luminometer (AB 2270, ATTO, Tokyo, Japan), and then, 50 μL of a solution containing 0.13 mol L −1 of NaOH, 4.42 mmol L −1 H 2 O 2 in the form of CaO 2 and 2.8 mmol L −1 luminol was injected into the system via a pump through the upper injection port. Fifty microlitres of 10 mmol L −1 L-ascorbate was added to the reaction to verify the presence of radicals. The intensities of the signals were recorded for 120 s.
The H 2 O 2 in the samples was analysed by a spectroscopic method 63 using a UV spectrophotometer (UV-1800, Shimadzu, Tokyo, Japan).
Hydroxyl radical (·OH) assay. An EPR assay was carried out to identify the presence of hydroxyl radicals in the systems. To follow the hydroxyl radical generation in the modified Fenton reaction using the iron catalysts, a spin trapping method using DMPO was employed. In the spin trapping experiment, 400 µL of phosphate buffer (pH 7.4) was mixed with 200 µL of 220 mmol L −1 DMPO, 100 µL of 4.42 mmol L −1 H 2 O 2 in the form of liquid H 2 O 2 (35% W/W) or CaO 2 (16% W/W) and 100 µL of 1 mmol L −1 Fe in the form of Fe-CPP, Fe(III) and Fe(II). To investigate whether the observed DMPO-OH radical originated from hydroxyl radical generation, an additional assay was performed in which 100 μL of 14 mol L −1 DMSO, an authentic hydroxyl radical scavenger, was added to each reaction system. Furthermore, the reactions of Fe-CGA and Fe-CA with CaO 2 were performed as models. The EPR spectra were recorded 30 s after the addition of the respective iron catalyst using an X-band EPR spectrometer (MS 5000, Magneteck, Berlin, Germany). The measurement conditions for EPR were as follows: magnetic field, 337.5 mT; field modulation frequency, 100 kHz; field modulation width, 0.16 mT; sweep time, 60 s; microwave frequency, 9.463 GHz; and microwave power, 5 mW. Statistical analyses. Completely randomized designs were used in all the experiments. Statistical significance (p < 0.05) for the wilt disease assay, population of R. solanacearum in the soil and total ROS generated were each assessed by one-way analysis of variance (ANOVA) followed by Fisher's least significant difference (LSD) post hoc test for multiple comparisons at a significance level of p < 0.05. Data availability. All data generated or analysed during this study are included in this published article (and its Supplementary Information files). The data sheets generated and/or analysed in the current study are available from the corresponding author on reasonable request.