Limosilactobacillus reuteri Fermented Brown Rice: A Product with Enhanced Bioactive Compounds and Antioxidant Potential

Oxidative stress has been postulated to play a role in several diseases, including cardiovascular diseases, diabetes, and stress-related disorders (anxiety/depression). Presently, natural plant-derived phytochemicals are an important tool in reducing metabolomic disorders or for avoiding the side effects of current medicinal therapies. Brown Rice (Oryza sativa L.) is an important part of Asian diets reported as a rich source of bioactive phytonutrients. In our present study, we have analyzed the effect of different lactic acid bacteria (LABs) fermentation on antioxidant properties and in the enhancement of bioactive constituents in Korean brown rice. Therefore, the antioxidant activities and phytochemical analysis were investigated for raw brown rice (BR) and different fermented brown rice (FBR). BR fermented with Limosilactobacillus reuteri, showed the highest antioxidant activities among all samples: DPPH (121.19 ± 1.0), ABTS (145.80 ± 0.99), and FRAP (171.89 ± 0.71) mg Trolox equiv./100 g, dry weight (DW). Total phenolic content (108.86 ± 0.63) mg GAE equiv./100 g, DW and total flavonoids content (86.79 ± 0.83) mg catechin equiv./100 g, DW was also observed highest in Limosilactobacillus reuteri FBR. Furthermore, phytochemical profiling using ultra-high-performance liquid tandem chromatography quadrupole time-of-flight mass spectrometry (UHPLC-QTOF/MS) and cell antioxidant assay (CAA) revealed L. reuteri FBR as a strong antioxidant with an abundance of bioactive compounds such as gamma-aminobutyric acid, coumarin, cinnamic acid, butanoic acid, ascorbic acid, nicotinic acid, and stearic acid. This study expanded current knowledge on the impact of fermentation leading to the enhancement of antioxidant capacity with an abundance of health-related bioactive compounds in BR. The results obtained may provide useful information on functional food production using fermented brown rice.


Introduction
Oxidative stress is a condition that is caused by an imbalance between antioxidants and free radicals of living organisms. This imbalance occurs due to the excessive production of reactive oxygen species (ROS) or antioxidant deficiency that leads to the damage of aerobic organisms as well as chronic inflammation; referred to as oxidative stress [1]. Lower ROS concentration is important for normal cellular signaling, while excess ROS can cause oxidative damage to DNA, lipids, proteins, and is associated with several chronic diseases [2,3]. The current definition of oxidative stress includes metabolic stressrelated pathways that participate in both cellular and extracellular metabolic events. The biology of oxidative stress is extremely complex, with multiple mechanisms at work [4]. Regardless of the mechanism, oxidative stress causes the onset of many diseases including cardiovascular diseases, diabetes, and anxiety or depression which are considered a major public health issue worldwide. As a result, consuming antioxidants to prevent oxidative

Rice Samples
BR (Oryza sativa L. Var. Japonica) used in this experiment were obtained from the local market of Chuncheon, Gangwon-do, South Korea. Using an electric mill, raw BR and brown rice after processing (fermentation) were pulverized into a fine powder and sifted through a 40 mesh sieve. Before further extraction, samples were kept at −20 • C.

Microorganisms
Limosilactobacillus reuteri AKT1 and all other lactic acid bacterial strains used for fermentation in our study were obtained from the Department of Food Science and Biotechnology, Kangwon National University, Korea. LABs were chosen for fermentation in the current study because they demonstrated high GABA (inhibitory neurotransmitter) content and fermentation efficiency in our last study (data not shown) [2]. The bacteria stock culture was kept at −80 • C in MRS broth (Difco), which contained 20% glycerol (v/v).

Sample Preparation 2.4.1. Brown Rice Fermentation
The LAB's growth medium consists of sterilized rice powder in distilled water. Before inoculation with lactic acid bacteria, the growth media were autoclaved for 15 min at 121 • C. Different bacterial strains used for fermentation (L. reuteri (AKT1), L. fermentum (AKT2), and L. plantarum (FMP2)) (2 × 10 7 cfu/mL) were transferred from 12 h (overnight) incubated culture to 100 mL of autoclaved growth media. The media was then incubated for 48 h at 37 • C with 150 rpm agitation, then centrifuged for 10 min at 10,000× g and the supernatant was freeze-dried and stored at −20 • C for further study.

Preparation of Extracts
Extraction was done by the method used by Pradeep et al. [14] with some modification. To remove lipids, samples were defatted with hexane using soxhlet equipment. Grounded samples were defatted three times in an orbital shaker at room temperature with hexane (1:5, w/v, 2 h). Defatted flours were kept at −20 • C till further use. Soluble phenolics from defatted samples (5 g) were extracted in an orbital shaker (RK-2D, DAIHAN scientific, Wonju, Korea) for 1 h at 50 • C with 50% ethanol (1:20 w/v). The extracts were centrifuged (Union 32R plus, Hanil Science Industrial, Incheon, Korea) at 4000 rpm for 10 min and supernatants were collected and this process was repeated until the third extraction. The supernatants were evaporated at 50 • C and freeze-dried. Before being reconstituted in ethanol, the lyophilized solids were stored at −20 • C. The stock solution of samples was prepared at a concentration of 1 mg/mL that was used throughout the analysis.

Determination of Total Phenolic Content (TPC)
Folin-Ciocalteu colorimetric method was used to measure the TPC reported by Pradeep et al. [15] with few modifications. In brief, 100 µL of the sample extract or standard (gallic acid solution) or 95% (v/v) methanol as blank was treated with 200 µL of Folin-Ciocalteu reagent for a short duration of 6 min. The mixture was then alkalized with 1 mL of Na 2 CO 3 700 mM. After being kept in dark conditions for 90 min, the SpectraMax i3 plate reader (Molecular Devices Korea, LLC, Seoul, Korea) was used to measure the absorbance at 760 nm wavelength. Gallic acid was used as a standard to calculate the TPC and results were expressed as milligrams of gallic acid equivalent per 100 g of sample (mg GAE equiv./100 g, DW).

Determination of Total Flavonoid Content (TFC)
TFC was analyzed using the method of Pradeep et al. [15] with few modifications. Briefly, a 200 µL sample extract was combined with 75 µL of NaNO 2 (50 gL −1 ) followed by the addition of 1 mL of distilled water. Then, the reaction mixture was allowed to settle for 5 min and then 75 µL of AlCl 3 (100 gL −1 ) was added. After 6 min, 600 µL of distilled water followed by 500 µL of 1 M NaOH were added. The SpectraMax i3 plate reader (Molecular Devices Korea, LLC) was used to measure the absorbance at 510 nm wavelength. Catechin was used as a standard and results were expressed as milligram catechin equivalents per 100 g of sample (mg catechin equiv./100 g, DW).

2.7.
Determination of Antioxidant Activities of BR 2.7.1. DPPH Radical Scavenging Activity DPPH activity was determined by the methods of Chang et al. [16] after slight modifications. In short, 100 µL of the sample extract or standard (Trolox) or blank (methanol) was mixed with freshly prepared 100 µL of 500 µM DPPH solution (dissolved in methanol) in a 24-well microplate and incubated at room temperature for 30 min. The absorbance was measured at 515 nm wavelength. The Trolox concentration plot with DPPH radical scavenging activity was used as a baseline curve. DPPH values were expressed as mg Trolox equivalent per 100 g of sample (mg Trolox equiv./100 g, DW) using the following formula: where A e represents the absorbance of the extract or standard and A c represents the absorbance value of the blank sample.

ABTS Radical Scavenging Activity
ABTS assay was carried out as described by Chang et al. [16] with little modifications. ABTS stock solution was prepared by mixing 2.45 mmol/L of potassium persulfate with 7 mmol/L of ABTS solution (1:1, v/v) and kept in the dark for 12-16 h at room temperature. The ABTS + reagent was constantly diluted with methanol until 0.700 ± 0.020 absorbance at 734 nm wavelength. Afterwards, 100 µL of extracts or standards were mixed with ABTS + solution (1 mL) and absorbance was measured at 734 nm. The per cent inhibition of ABTS was measured using the same formula as for the DPPH assay (mentioned above). ABTS values were expressed as mg Trolox equiv./100 g, DW using the Trolox standard curve.

Ferric Reducing Antioxidant Power (FRAP)
FRAP assay was analyzed using the method as documented by Xiang et al. [17] with little modifications. In short, 0.1 mL of extracts were combined with a FRAP reagent of 3.9 mL that was prepared using acetate buffer (50 mL, 0.3 M, pH 3.6), Tripyridyl Triazine (5 mL, TPTZ) solution (10 mM of TPTZ in 40 mM of HCl) and FeCl 3 · 6H 2 O (5 mL, 20 mM) and kept for 10 min at 37 • C, then absorbance was taken at 593 nm wavelength. These findings were expressed as mg Trolox equiv./100 g, DW.

Identification of BR Bioactive Compounds Using UHPLC-Q-TOF-MS/MS
The bioactive compounds of BR samples were analyzed using UHPLC Q-TOF-MS/MS (SCIEX Exion LC AD system, Framingham, MA, USA) according to the protocol previously conducted in our laboratory by Tyagi et al., and Daliri et al. [2,18]. The mass spectrometric analysis was conducted in both positive (ESI+) and negative (ESI−) ion modes. In summary, UHPLC Q-TOF-MS/MS system was fitted with different components including an autosampler, photodiode array detector, and controller. The analytical column used in the analysis was 100 mm × 3 mm Accucore C18 column (Thermo Fisher Scientific, Waltham, MA, USA). Later, the sample (10 µL) was injected by autosampler and eluted into the column with a binary mobile phase consisting of solvent A (water with 0.1% of formic acid) and solvent B (methanol). A flow rate (0.4 mL/min) with a linear gradient programmed for 25 min was used in this analysis. Under these conditions, the scanning time was approximately 1 s. The bioactive compounds of BR were identified by using a metabolomics workbench.

Cell Viability Assay
After performing antioxidant assays and untargeted metabolomics of raw BR and different LABs treated BR samples, the best (L. reuteri FBR) sample was selected for further cell line analysis. To test the viability in Caco-2 cell lines, the L. reuteri FBR sample was analyzed using a laboratory EZ-cytox assay. WST of EZ-cytox exists in the respiratory chain of mitochondria and is active only in living cells. Briefly, Caco-2 cells in the growth medium were seeded on a 96-well plate at a density of 4 × 10 4 cells/well. The growth medium was removed and the cells were washed using PBS after 24 h of incubation at 37 • C with 5 per cent CO 2 . Then, 100 µL of growth medium with various sample extract concentrations was applied. In the control group, a medium without sample extract was added. After 12 h of incubation at 37 • C with 5 per cent CO 2 , 10 µL WST-1 solution was added to each sample. The 96-well plate was left for 10 min at 37 • C and the absorbance was measured at 455 nm by SpectraMax i3 plate reader. If a concentration of sample extract reduced the cell viability by >10%, then the extract at this concentration was cytotoxic.
cell viability (%) = mean absorbance in test well mean absorbance in control well × 100

Cellular Antioxidant Activity (CAA)
The formation of ROS within cells was investigated using oxidation sensitive DCFH-DA probes by the method of Ti et al. [19] with slight modifications. In brief, Caco-2 cells were grown overnight at a density of 6 × 10 4 cells per well in black 96-well microplates. Later, PBS (50 µL) was used to wash the cells after 2 h of pretreatment with various concentrations of sample extracts (0.5-5 mg/mL) and 100 µL of DCFH-DA (25 µmol/L). Then, 100 µL of DMEM medium (composed of 600 µmol L −1 ABAP) was inoculated in each well excluding the blank well, which received 100 µL of DMEM medium without ABAP. The fluorescence was measured using a plate reader SpectraMax i3 plate reader (Seoul, South Korea) and wavelengths of 485 nm excitation and 538 nm emission for 13 cycles (5 min each). The formula used for calculating the CAA unit was as follows: where SA and CA denote the integral areas under the sample and control time-fluorescence value curves, respectively.

Statistical Analysis
GraphPad Prisma 8.0 was used to analyze the data. Using the SPSS program and GraphPad Prism 8.0, differences in mean values between brown rice samples of phenolic extracts were calculated using one-way variance analysis (ANOVA) followed by a Tukey test at p < 0.05 significance stage. The findings were referred to as mean standard deviation (SD).

TPC and TFC
The TPC and TFC of all four samples are shown in (Table 1) as mean ± SD of triplicate analyses with statistically significant differences (Tukey and Duncan test p ≤ 0.05). TPC ranged between 16.08 ± 0.49 to 108.86 ± 0.63 mg GAE equiv./100 g, DW. TPC content was found lowest in raw BR samples 16.08 ± 0.49 and highest in L. reuteri FBR 108.86 ± 0.63. Hydrolysis by enzymes during fermentation usually increases total phenolic content, as observed in this study. TPC increases with fermentation, L. fermentum FBR (75.00 ± 0.017 mg GAE/100 g, DW), L. plantarum FBR (96.87 ± 0.94 mg GAE equiv./100 g, DW), and L. reuteri FBR (108.86 ± 0.63 mg GAE equiv./100 g, DW), compared with raw BR. This study shows higher TPC content of BR than reported by [21,22]. TFC was found higher in L. reuteri FBR 86.79 ± 0.83 mg catechin equiv./100 g, DW, followed by L. plantarum FBR and L. fermentum FBR (66.28 ± 0.71 and 54.77 ± 1.02 mg catechin equiv./100 g, DW) ( Table 1). TFC content was lower than TPC in BR samples. TFC in our study was higher than previously reported by Huang et al. [23], similar to TPC, but both TPC and TFC were found lower than reported by Gong et al. [24]. These differences in BR values of different researchers could be because of genotype, cultivation landscape, and climate conditions. In addition, it is worth noting that the phenolic content can be significantly influenced by different extraction solvents and procedures.

Antioxidant Assay (DPPH, ABTS, FRAP)
Various methodologies including reducing capacity, free radical scavenging, lipid peroxidation inhibition, and metal ion chelation have been studied to explain how rice extracts have shown effective antioxidant potential [25]. In recent years, the fermentation process is thought to be an effective method for increasing antioxidant activity in cereals. The antioxidant activity of raw and different LABs fermented brown rice samples were assessed using the DPPH, ABTS, and FRAP assays in the current study. The antioxidant values of DPPH, ABTS, and FRAP of phenolic extracts of BR and differently treated BR samples were presented in Table 1, respectively.
Similarly, ABTS is considered an important method for determining radical scavenging activity in grains and plant materials. Furthermore, FRAP assay was originally designed to assess plasma antioxidant ability but has also been commonly used in a wide variety of pure compounds and biological samples to determine antioxidant capacity. It measures absorption changes caused by the formation of blue iron (II) from colorless iron oxide (III). In the present research, the same trend as DPPH was observed in ABTS and FRAP assays. ABTS activity was measured highest in L. reuteri FBR (145.80 ± 0.99 mg Trolox equiv./100 g, DW) ( Table 1). In ABTS, the lowest activity was also by raw BR. Similarly, FRAP was found highest in FBR (171.89 ± 0.71 mg Trolox equiv./100 g, DW) followed by L. plantarum and L. fermentum FBR. As a result of different antioxidants assays, L. reuteri FBR showed the highest activity among all samples. These findings were higher than earlier reports by Lin et al. [26] and IIowefah et al. [21] in fermented BR. Furthermore, other studies have reported the high antioxidant activity of BR [27].

Untargeted Metabolomics Using UHPLC Q-TOF-MS/MS in Brown Rice Samples
UHPLC Q-TOF-MS/MS detection is considered a gold standard technique for the precise detection and quantification of a wide variety of components. Therefore, in this study, we have used this detection technique for the identification of phenolic compounds in brown rice.

Phenolic Compounds
In the present research, the phenolic compositions of BR treated with different fermentation bacteria were selected and positively or tentatively identified by UHPLCQ-TOF-MS/MS. Phenolic identification and characterization were achieved by comparing our results with mass spectral literature evidence and cross-referencing it with other available spectral databases, such as Metlin and Metabolomics Workbench. A total of 15 phenolic compounds were tentatively found from our soluble extracts of raw BR, L. reuteri FBR, L. fermentum FBR, and L. plantarum FBR respectively, as shown in Table 2   Results showed that the highest phenolic compounds were detected in the L. reuteri FBR sample. Because phenolic compounds are not readily available, they typically occur in cereals in esterified linkages to the cereal wall matrix [28]. Fermentation is considered to be a possible strategy to release insoluble or bound phenolic compounds and thus leading to improve the poor bioavailability of grain phenolics. Comparing different fermenting bacteria in the present study we found that L. reuteri fermentation releases most of the phenolic compounds compared with other bacterial strains and thus improves the bioavailability and bioaccessibility of cereal grains such as brown rice phenolics [29]. Many phenolic compounds detected in the current study such as p-coumaric acid [30], ascorbic acid [31], cinnamic acid [32], and vanillic acid [33] are already reported in the literature for their strong antioxidant capacities.

Levels of Amino Acid in Brown Rice
In the growth and development of organisms, amino acids play an important role and can also improve the taste of food. In our present study, a total of 18 amino acids were detected in raw and differently fermented BR samples (Figure 2 and Table 3) which shows statistically significant differences from each other after comparing their levels. Raw BR contained the least number of amino acids, which may be due to more bound molecules with the parent, whereas fermentation leads to an increase in amino acid content. The levels of amino acids were detected highest in the L. reuteri FBR sample which might strain-specific as fermentation microorganisms produce enzymes that lead to the formation of several metabolites and bioactive compounds from the food matrix [34]. In the ethanol extract, we found levels of some essential amino acids (tryptophan, lysine, methionine, and histidine), as well as certain conditionally essential amino acids (arginine, ornithine, serine, and glutamine), increased drastically after fermentation ( Figure 2 and Table 3). The identification was done by comparing with mass spectral libraries, XCMS online (Metlin) and Metabolomics Workbench. In amino acids, L. reuteri FBR also shows the highest number of amino acids content as observed in phenolic compounds. Nd-not detected, BR-brown rice, and FBR-fermented brown rice.

Level of Fatty Acid in Brown Rice Samples
In particular, fermentation has been proposed as a tool for enhancing foods' nutritional values, both in terms of enhanced bioavailability of bioactive components as well as the production of health-promoting end-products. Due to their proven benefit, in the last decade, short-chain fatty acids (SCFAs) have emerged as some of the most researched compounds. In the present study, 13 fatty acids were detected in raw and different LABs fermented BR samples (Table 4) and fatty acid levels were found to be significantly different in all samples. The results show that the highest levels of fatty acids were found in L. reuteri FBR. Heat map analysis was used for separating fatty acids based on the different concentrations, represented in different shades of green (dark to light) in decreasing order (Figure 3). mation of several metabolites and bioactive compounds from the food matrix [34]. In the ethanol extract, we found levels of some essential amino acids (tryptophan, lysine, methionine, and histidine), as well as certain conditionally essential amino acids (arginine, ornithine, serine, and glutamine), increased drastically after fermentation (Figure 2 and Table 3). The identification was done by comparing with mass spectral libraries, XCMS online (Metlin) and Metabolomics Workbench. In amino acids, L. reuteri FBR also shows the highest number of amino acids content as observed in phenolic compounds.    ND-not detected, BR-brown rice, and FBR-fermented brown rice.

Cell Viability Assay
Cytotoxicity is regarded as an important step in determining the suitability and further applications of any food extract. Using the Ez cytox assay kit, the cytotoxic effect of L. reuteri FBR extracts at 0.3-10 mg/mL concentrations was investigated using Caco-2 cell lines. Figure 4 depicts the cell viability results of the extract after 12 h of incubation. It was

Cell Viability Assay
Cytotoxicity is regarded as an important step in determining the suitability and further applications of any food extract. Using the Ez cytox assay kit, the cytotoxic effect of L. reuteri FBR extracts at 0.3-10 mg/mL concentrations was investigated using Caco-2 cell lines. Figure 4 depicts the cell viability results of the extract after 12 h of incubation. It was observed that cell viability was not much decreased after increasing the concentration up to 10 mg/mL. No significant differences were observed in cytotoxicity assay by using 0.3-10 mg/mL concentrations (Figure 4). The extract was observed to be non-toxic after 12 h assay as extract still shows about 97 per cent of cell viability. Our results were found similar to the results presented by Yue et al. [35].
observed that cell viability was not much decreased after increasing the concentration up to 10 mg/mL. No significant differences were observed in cytotoxicity assay by using 0.3-10 mg/mL concentrations (Figure 4). The extract was observed to be non-toxic after 12 h assay as extract still shows about 97 per cent of cell viability. Our results were found similar to the results presented by Yue et al. [35].

Cellular Antioxidant Activity (CAA)
The effect of pretreatment of Caco-2 cells with L. reuteri fermented extract of brown rice on intracellular reactive oxygen species (ROS) was determined using a cell-based assay. The fluorescent probe DCFH-DA is used as an indicator of ROS and oxidative stress in our study. The nonionic and nonpolar DCFH-DA probe diffuses passively into cells before being hydrolyzed by intracellular esterases to form nonfluorescent 2′,7′-dichlorofluorescein (DCFH). Later in the presence of ROS, DCFH that is trapped inside cells is oxidized into fluorescent 2′,7′-dichlorofluorescein (DCF) [36]. When the cellular antioxidant defense system fails to compensate for ROS production, oxidative stress occurs. This reaction can be slowed down using bioactive compounds, preventing the generation of DCF. Following the uptake of antioxidant compounds can be accomplished on the cell membrane surface or within the cell [37]. We evaluated the effect of our L. reuteri FBR extract against oxidative stress in Caco-2 cells. In our study, ABAP was chosen as an intracellular oxidizing agent to simulate oxidative stress in cells. 600 μmol L −1 ABAP was chosen as the optimal concentration to induce oxidation. As represented in Figure 5A, CAA values in L. reuteri FBR extract were observed to be 5.7 times higher than the raw BR sample at a concentration of 1mg/mL. Our results indicate that extracts reduced ROS levels at rest in a dose-dependent manner ( Figure 5B); CAA values were increased with concentration (0.5 mg/mL to 5 mg/mL) from 49.50 ± 1.67% to 72.49 ± 1.23%. The strength of inhibition strongly followed a curvilinear pattern as L. reuteri FBR extract concentrations increased. A similar effect was previously observed in the study of Grauzdytė et al., where they observed Phyllanthus phillyreifolius extracts in HEK-293 cells [38], and the study of Kellett et al. [39] in Caco-2 cells.

Cellular Antioxidant Activity (CAA)
The effect of pretreatment of Caco-2 cells with L. reuteri fermented extract of brown rice on intracellular reactive oxygen species (ROS) was determined using a cell-based assay. The fluorescent probe DCFH-DA is used as an indicator of ROS and oxidative stress in our study. The nonionic and nonpolar DCFH-DA probe diffuses passively into cells before being hydrolyzed by intracellular esterases to form nonfluorescent 2 ,7 -dichlorofluorescein (DCFH). Later in the presence of ROS, DCFH that is trapped inside cells is oxidized into fluorescent 2 ,7 -dichlorofluorescein (DCF) [36]. When the cellular antioxidant defense system fails to compensate for ROS production, oxidative stress occurs. This reaction can be slowed down using bioactive compounds, preventing the generation of DCF. Following the uptake of antioxidant compounds can be accomplished on the cell membrane surface or within the cell [37]. We evaluated the effect of our L. reuteri FBR extract against oxidative stress in Caco-2 cells. In our study, ABAP was chosen as an intracellular oxidizing agent to simulate oxidative stress in cells. 600 µmol L −1 ABAP was chosen as the optimal concentration to induce oxidation. As represented in Figure 5A, CAA values in L. reuteri FBR extract were observed to be 5.7 times higher than the raw BR sample at a concentration of 1mg/mL. Our results indicate that extracts reduced ROS levels at rest in a dose-dependent manner ( Figure 5B); CAA values were increased with concentration (0.5 mg/mL to 5 mg/mL) from 49.50 ± 1.67% to 72.49 ± 1.23%. The strength of inhibition strongly followed a curvilinear pattern as L. reuteri FBR extract concentrations increased. A similar effect was previously observed in the study of Grauzdytė et al., where they observed Phyllanthus phillyreifolius extracts in HEK-293 cells [38], and the study of Kellett et al. [39] in Caco-2 cells.

Conclusions
In our study, we discovered that L. reuteri FBR had higher antioxidant activity as well as a higher concentration of phenolics and flavonoids among all LABs used for the study. This shows the ability of L. reuteri as a promising fermentation strain to increase the bioavailability of cereals or grains in producing health-promoting functional materials. L. reuteri fermentation improves phenolic constituents and antioxidant activity of BR, improves Figure 5. In Caco-2 cells, peroxyl radical-induced oxidation of DCFH to DCF and ROS inhibition by raw BR and L. reuteri FBR extract (A,B) showing the effect of dose-dependent inhibition of L. reuteri FBR extracts (0.5-5 mg/mL). Data were represented as means ± standard deviations (n = 3) with one way ANOVA. The columns with different letters (a-d) show significant differences using Tukey's test at p < 0.05.