Phytochemicals-mediated production of hydrogen peroxide is crucial for high antibacterial activity of honeydew honey

Honeydew honey is increasingly valued due to its pronounced antibacterial potential; however, the underlying mechanism and compounds responsible for the strong antibacterial activity of honeydew honey are still unknown. The aim of this study was to investigate the inhibition of bacterial growth of 23 honeydew honey samples. Activity of bee-derived glucose oxidase (GOX) enzyme, the content of defensin-1 (Def-1) and hydrogen peroxide (H2O2), and total polyphenol content were determined in the 23 honey samples. Our results demonstrated that antibacterial activity of honeydew honey was equivalent to medical-grade manuka and kanuka honey and was abolished by catalase. Although H2O2 is an important factor in the inhibition of bacterial growth, polyphenolic compounds and their interaction with H2O2 are the key factors responsible for high antibacterial activity of honeydew honey. In addition, our results indicated that the antibacterial activity of honeydew honey is not dependent on GOX-mediated production of H2O2 or the presence of Def-1.

Comparison of GOX activity and H 2 O 2 production in honey samples. The GOX activity determined in 20% honey solutions ranged from 21 to 50 mU/ml in all honeydew honey samples, with an average value of 35.7 mU/ml ( Fig. 2A). The lowest enzymatic activity was monitored in both types of medical honey (<20 mU/ml),  suggesting that methylglyoxal, present in both samples, may have structurally modified the GOX enzyme. The concentration of H 2 O 2 in honeydew honey samples was determined in 40% honey solutions and ranged from 0.3 to 3.4 mM after 24 h incubation at 37 °C (Fig. 2B). The average value of H 2 O 2 concentration was 1.8 mM. Two honeydew honey samples (#16 and #21) showed a weak production of H 2 O 2 , with values below 1 mM. As expected, kanuka and manuka honey accumulated the lowest levels of H 2 O 2 . No significant correlations were found between the values of GOX activity and H 2 O 2 concentration among the honeydew honey samples (r = 0.26, P = 0.22). Therefore, we assumed that other compound(s) of botanical origin present in honeydew honey participate in the production of H 2 O 2 . Potential candidates responsible for the elevated levels of H 2 O 2 include phenolic compounds.
Total polyphenol and Def-1 content in honey samples. Def-1 is a key, but quantitatively variable, antibacterial factor in honey. Def-1 content was examined in the 2.5, 5 and 10% honey samples in PBS, using the newly developed competitive ELISA and the results were expressed as a percentage of total protein (Fig. 3A). The amount of Def-1 in each honey sample, including kanuka and manuka honey, is shown in Fig. 3A. The two honeydew honey samples #9 and #10 contained the highest amount of Def-1. Similar to the H 2 O 2 production capacity, kanuka and Manuka honey had the lowest content of Def-1 among all the tested samples. Overall content of Def-1 in the honeydew honey samples did not correlate with their antibacterial activity against S. aureus (r = 0.17, P = 0.45). Def-1, at the concentration observed in the honeydew honey samples, did not exert its antibacterial activity at the MICs of the honeydew honey samples. The obtained data also suggest that the antibacterial action of kanuka and manuka honey dependeds on methylglyoxal or other unknown botanical compounds rather than bee-derived antibacterial factors.
A direct antibacterial activity of phenolic compounds has not been shown since their concentrations in honey are too low to determine overall activity. However, phenolic compounds may act in synergy via pro-oxidative action by generating elevated levels of H 2 O 2 in honey, which mediates the inhibition of bacterial growth as a result of oxidative stress. Total phenolic content determined in the 20% (w/v) honeydew honey solutions is shown in Fig. 3B.
Contribution of bee-derived antibacterial compounds to the antibacterial activity of honeydew honey. To investigate the contribution of bee-derived antibacterial compounds to the overall antibacterial activity of honeydew honey, samples were treated with a proteolytic enzyme, proteinase K. Upon proteinase K treatment, all proteinous content in the 50% (w/v) honey solutions was completely digested (data not shown). Untreated and proteinase K-treated honeydew samples were subjected to immunoblot analysis of GOX and Def-1. As expected, GOX and Def-1 were present in all untreated samples (Fig. 4). On the other hand, no immunoreactive bands for GOX and Def-1 were observed in treated samples, confirming the proteolytic effectiveness of proteinase K.
MIC values were determined in untreated and proteinase K-treated honey samples. No significant changes were identified in the MIC values between untreated and proteinase-treated honey against both tested bacteria (Fig. 5), suggesting that bee-derived components, including GOX-mediated H 2 O 2 , are not responsible for the pronounced antibacterial activity of the honeydews honey. Although H 2 O 2 is a major antibacterial compound of honeydew honey, its source might be solely of botanical origin.

Role of H 2 O 2 in honeydew honey antibacterial activity.
To determine the role of H 2 O 2 in honeydew honey antibacterial activity, 50% (w/v) honey solutions were treated with catalase for 2 h and MICs were determined. The MIC values of honey samples with or without catalase treatment against both tested bacteria are shown in Fig. 5. Catalase-treated samples had lower antibacterial activity with average MIC values of 30%. Furthermore, in seven samples, the MIC values were greater than 30% following catalase treatment. The effect of catalase on the MIC values in all honeydew honey samples was highly significant (NumDF = 2, DenDF = 268.43, F = 2136.38, P < 0.001) and predicted values were three times larger ( Fig. 6) compared with untreated and protease K-treated samples. Interestingly, statistically significant changes (NumDF = 1, DenDF = 268.27, F = 8.52, P = 0.004)) were observed following catalase treatment between S. aureus and P. aeruginosa (Fig. 6). S. aureus was more resistant to the honey solution after the removal of H 2 O 2 compared with P. aeruginosa.
Interestingly, the MIC of kanuka honey following catalase treatment was also changed, suggesting that the antibacterial activity of kanuka honey is, at least in part, due to the generation of H 2 O 2 . As expected, no remarkable changes in antibacterial activity were observed in manuka honey. Surprisingly, the antibacterial activity of manuka honey against P. aeruginosa following catalase treatment was even higher than without enzyme treatment. Despite the fact that the antibacterial activities of honeydew honey samples were comparable, H 2 O 2 production was not uniform. Statistical analysis did not show any correlation between antibacterial activity and the level of H 2 O 2 in the different honey samples (r = −0.38, P = 0.08 and r = −0.03, P = 0.89). The honey samples with the highest level of H 2 O 2 were not the most active.

Discussion
Honey is increasingly valued for its antibacterial and wound healing activity and effectiveness as a treatment of chronic wounds and burns. However, every natural honey exhibits a certain antibacterial activity and only those with a strong and relatively stable antibacterial activity have been selected for medical and clinical use. In this study, we focused on honeydew honey as a source of potential medical-grade honey to elucidate its antibacterial activity and to investigate the role of bee-derived antibacterial components on its overall antibacterial activity.
A plethora of studies has demonstrated the important role of Def-1 and H 2 O 2 in the antibacterial activity of honey.
Def-1 is an antibacterial peptide belonging to the insect defensin group that is composed of 51 amino acids with a molecular weight of 5.52 kDa. Bee Def-1 is effective against Gram-positive bacteria [19][20][21] ; however, some studies using recombinant Def-1 have also reported its activity against Gram-negative bacteria including P. aeruginosa and Salmonella choleraesuis 22,23 . According to our recent study 24 , Def-1 is a common but quantitatively variable component in honey and its levels in different blossom honey samples are correlated with antibacterial activity against S. aureus. Recently, Def-1 has also been shown to be an immunomodualtor in wound healing where it positively contributes to cutaneous wound closure 25 . In the present study, the amount of Def-1 in honeydew honey did not correlate with the overall antibacterial activity of samples against S. aureus, suggesting that its contribution at low dilutions (5-10%) is negligible.
The most important antibacterial factor in honey is H 2 O 2 , which is produced by GOX-mediated conversion of glucose to gluconic acid under aerobic conditions in diluted honey 26 . Therefore, GOX plays an important role in the generation of H 2 O 2 15 . Although H 2 O 2 at concentrations found in honey does not kill bacteria, it is able to interact with bacterial cell proliferative signals, and thus affects bacterial growth even at high dilutions of honey 27 . Some researchers have suggested that the levels of H 2 O 2 in honey may differ between types of honey regardless of botanical and geographical origin 14 . A few studies have attempted to determine the concentration of H 2 O 2 in We assume that plant-derived polyphenolic compounds, often reported in honey samples, may contribute to or modulate antibacterial activity. In the presence of transition metal ions (Cu and Fe) and peroxides, polyphenols, well-known dietary antioxidants, can act as pro-oxidants by accelerating hydroxyl radical formation and oxidative strand breakage in DNA 31 . In fact, polyphenols work in two ways to promote antibacterial activity: by directly producing H 2 O 2 , and by reducing Fe (III) to Fe (II), which triggers the Fenton reaction to create more potent reactive oxygen species such as hydroxyl radicals. A key factor in determining whether polyphenolic compounds exhibit antioxidative or antibacterial properties is pH value 32 . In weakly alkaline conditions (pH 7.0-8.0), polyphenols can exhibit pro-oxidative activity and inhibit bacterial growth. In the present study, the determination of honey MIC values and H 2 O 2 concentration was carried out at pH 7.3 and 7.0, respectively. In these experimental conditions, it is obvious that the polyphenols in honeydew honey may act in synergy and inhibit bacterial growth. As mentioned, we measured the accumulation of H 2 O 2 in diluted honey samples, but found no correlation between the concentration of H 2 O 2 and antibacterial activity. It is likely that besides accumulation of H 2 O 2 , polyphenol-mediated production of hydroxyl radicals in honeydew honey significantly affects the overall antibacterial activity. Moreover, it has been shown that the chemical interaction of honey polyphenols with H 2 O 2 results in the generation of products responsible for the degradation of bacterial DNA 33 . Interestingly, H 2 O 2 alone is not able to induce DNA breaks but is somehow involved in this process. Although H 2 O 2 is necessary for accelerating the auto-oxidation of polyphenols via hydrolysis and serves as a source of hydroxyl radicals, polyphenol concentration is a critical factor for antibacterial activity.
Total polyphenol content was reported to be high in other studies of honeydew honey samples 7,34,35 . In the present study, the total polyphenol content did not differ substantially between the tested honeydew honey samples (with the exception of sample #15), with an average value of 50 mg GAE/100 g of honey. It will be important to determine the critical concentration limit of total polyphenols for their behavior as pro-oxidant agents.
Several studies have attempted to identify biologically-active polyphenols and flavonoids from honeydew honey of different botanical and geographical origin 6,36,37 . Ferulic acid, quercetin, kaempherol and particular p-coumaric acid were the most abundant polyphenols in honeydew honeys. In a very recent studies 38 , combination of p-coumaric acid with bacteriocin showed synergistic effects against planktonic cells of food-borne bacteria. Similarly, combination of plant polyphenols and H 2 O 2 was shown to be more effective against bacteria than H 2 O 2 alone and can induce oxidative stress-related responses in bacteria 32 . It has been hypothesized that antibacterial effect of plant polyphenols is linked to polyphenols-induced H 2 O 2 and an increase of endogenous reactive oxygen species generation 39 .
Until now, no study had been conducted to determine the role of bee-derived antibacterial compounds in honeydew honey. Here, we showed that the antibacterial activity of honeydew honey is not dependent on GOX-mediated production of H 2 O 2 or the presence of Def-1. This is an interesting observation in the light of current knowledge regarding the antibacterial properties of honey. The antibacterial potential of honeydew honey is therefore not associated with the production capacity of GOX and Def-1 in particular bee colonies, which has a great impact on the biological properties of honey.
In conclusion, the tested honeydew honey samples showed equivalent or, in some cases, higher antibacterial activity compared with medical-grade kanuka and manuka honey. This strong antibacterial activity was abolished by treatment with catalase. Although H 2 O 2 is an important factor in the inhibition of bacterial growth, phytochemicals such as polyphenolic compounds and their interaction with H 2 O 2 are the key factors responsible for the antibacterial activity of honeydew honey. In addition, our results indicate that the antibacterial activity of honeydew honey is not dependent on GOX-mediated production of H 2 O 2 or the presence of Def-1.

Materials and Methods
Honey samples. Honey samples (n = 28) collected in 2016 were received from beekeepers from several regions of Slovakia. The samples were selected according to their electrical conductivity value (Codex Alimentarius, Bogdanov 2002). Five samples were excluded from the study, as they did not meet the criteria for honeydew honey. Overall, 23 honeydew honey samples were used for further analysis. The honey samples were also compared to medical grade honey, Manuka Honey MGO 550 purchased from Manuka Health (New Zealand), and raw kanuka honey (HLKN1, No. 5054), kindly provided by Dr. Shaun Holt (honeylab, New Zealand).

Microorganisms.
The antibacterial activity of honey samples was assessed against the isolates Pseudomonas aeruginosa CCM1960 and Staphylococcus aureus CCM4223, obtained from the Department of Medical Microbiology, Slovak Medical University (Bratislava, Slovakia). These bacterial isolates are the two most frequently isolated pathogens from wounds.
Determination of honey antibacterial activity. The antibacterial efficacy of the honey samples was evaluated with a minimum inhibitory concentration (MIC) assay as described by Bucekova et al. (2014). Briefly, overnight bacterial culture was suspended in phosphate-buffered saline (PBS), pH 7.2, and the turbidity of the suspension was adjusted to 10 8 colony-forming unit (CFU)/ml and diluted with Mueller-Hinton broth (MHB) medium (pH 7.3 ± 0.1) to a final concentration of 10 6 CFU/ml. Then, 10-μl aliquots of suspension were inoculated into each well of sterile 96-well polystyrene plates (Sarstedt, Germany). The final volume in each well was 100 μl, consisting of 90 μl of sterile medium or diluted honey and 10 μl of bacterial suspension. After 18 h of incubation at 37 °C, bacterial growth inhibition was determined by monitoring the optical density at 490 nm. The MIC was defined as the lowest concentration of honey inhibiting bacterial growth. All tests were performed in triplicate and repeated three times.
Besides MIC determination, MBC was evaluated in honey samples 40 . The viability of bacteria in wells with no turbidity was determined by spreading 100 µl onto an MHB agar plate and incubating at 37 °C for 24 h. The lowest honey that resulted in no survival of viable bacteria was recorded as MBC.
Enzymatic treatment of honey samples with catalase and proteinase K. The 50% (w/v) honey solutions were treated with catalase (2000-5000 U/mg protein; Sigma-Aldrich, UK) at a final concentration ranging from 1000-2500 U/ml at room temperature for 2 h or proteinase K (30 U/mg; Promega, WI, USA) at a final concentration of 50 μg/ml at 37 °C for 30 min. Catalase-and proteinase K-treated honey samples were then used in the antibacterial assay to determine MIC values against S. aureus and P. aeruginosa.
To determine the effectiveness of both enzymes, the accumulation of H 2 O 2 was measured in catalase-treated samples after 24 h incubation at 37 °C and the presence of GOX and Def-1 in proteinase K-treated samples were determined by immunoblot analysis (see 2.9.).
Determination of GOX activity. The bee-derived GOX activity was determined by with a Megazyme GOX assay kit (Megazyme International Ireland Ltd., Bray, Ireland), which is based on the oxidative catalysis of β-D-glucose to D-glucono-δ-lactone, with the concurrent release of H 2 O 2 . The resultant H 2 O 2 reacts with p-hydroxybenzoic acid and 4-aminoantipyrine in the presence of peroxidase to form a quinoneimine dye complex, which has a strong absorbance at 510 nm. The enzyme activity was determined in 20% (w/v) honey solutions in a 96-well microplate according to the manufacturer's instructions. Quantification of Def-1. Bee-derived peptide Def-1 was quantified as described by Valachova et al. (2016).

Determination
Briefly, serial dilutions of each honey sample were prepared from a 50% (w/v) solution of honey, resulting in final concentrations of 2.5, 5, and 10%. An enzyme-linked immunosorbent assay (ELISA) was performed using a rabbit polyclonal anti-honeybee Def-1 antibody raised against a synthetic peptide corresponding to the C-terminus of bee Def-1 (CRKTSFKDLWDKRFG) 24 . The amount of Def-1 measured was expressed as a percentage of total proteins, which were measured using the Quick Start Bradford Protein Assay (Bio-Rad, CA, USA) according to the manufacturer's instructions. All measurements were performed in triplicate and repeated three times.
Total phenolic content. Total phenolic content was determined with a Folin Ciocalteu Phenolic Content Quantification Assay Kit (BioQuoChem, Spain) in a 20% (w/v) honey solution in a 96-well microplate according to the manufacturer's instructions. Gallic (GAE) acid was used as the reference standard and results were expressed as GAE equivalents (mg/ml). Absorbance was measured at 700 nm at 37 °C.

Statistical analysis.
A general linear mixed model was used to analyse the differences in MIC values following treatment with catalase and proteinase K as well as untreated samples. As every sample was tested for each treatment and both P. aeruginosa and S. aureus, the identity of samples was controlled by including a random part of the model with the sample ID. This approach allowed for the exclusion of inter-sample variability. An R programming and statistical environment was used for statistical analyses (www.R-project.org). The general linear mixed model was analysed with the lme4 package 41 . Correlations among all parameters were tested with the Hmisc package (R package version 4.1-1) using non-parametric Spearmans rank correlations. Correlations with correlation coefficient r > 0.7 or <−0.7 were considered important.