A new colorimetric method for determining antioxidant levels using 3,5-dibromo-4-nitrosobenzene sulfonate (DBNBS)

We describe here a novel assay that determines the total a+ntioxidative activities of known antioxidants and antioxidants in beverages. The method employs the substrate 3,5-dibromo-4-nitrosobenzene sulfonate (DBNBS) that yields the colored product 3,5,3’,5’-tetrabromoazobenzene sulfate sodium salt (azo-TBBS). The amounts of azo-TBBS are measured using HPLC and then used to calculate total antioxidative capacity (TAC) values. We first show that the TAC values measured using the new DBNBS system were significantly higher compared with the control. The assay was validated through further analysis of 56 compounds, including previously characterized antioxidants. The data are consistent with published values. Here we describe in detail the application of the DBNBS method to the measurement of the TAC values of eight beverages, including wines and fruit juices. The DBNBS assay employs a readily applicable protocol that sensitively determines the levels of antioxidants in foodstuffs. - A new DBNBS-mediated antioxidant assay system is compared with standard DPPH and ORAC assays - DBNBS traps hydrogen radicals to generate a readily measured colored reduction product that quantifies antioxidant levels

Agricultural and Biological Sciences More specific subject area; Analysis of natural products Method name; DBNBS assay Name and reference of original method; Not applicable Resource availability; Not applicable

Method details
Background Diseases such as cancer, diabetes, atherosclerosis, and Alzheimer's are closely associated with elevated levels of oxidative stress. Oxidative stress reflects an imbalance between the production of reactive oxygen species and detoxification capacities conferred by inherent cellular defense systems and exogenous antioxidants [1 , 2] . A state of overwhelming stress is etiologically associated with the aforementioned diverse pathologies [1 , 2] . A potential alternative to established chemical and biological therapeutics is the use of antioxidants contained in natural products. For example, continuous consumption of antioxidant-rich food may serve this purpose, particularly because of the adverse effects of synthetic antioxidants [2][3][4] . To achieve this goal, the food and pharmaceutical industries require access to detailed data regarding the levels of antioxidants contained in food to maximize the availability of antioxidants in the body [5 , 6] . Therefore, we have been conducting in vitro and in vivo studies to identify the potential beneficial therapeutic effects of antioxidants present in food and drinking water [7][8][9][10][11][12][13][14][15][16] .
For this purpose, among the 407 published antioxidant assays derived from 29 fundamental methods, we have mainly applied relatively common methods such as the oxygen radical absorbance capacity (ORAC) and 2, 2-diphenyl-1-picrylhydrazyl radical (DPPH) assays [17] . Analytical chemists and biochemists are thus confronted with an array of assays that must be evaluated for use in a specific system. As such, more than one assay may typically be employed.
Our laboratory focuses on establishing desirable methods to measure the TAC of methanolextracted research samples as well as commercial substances to guide further detailed characterization of the active compound(s). A literature review indicated that DBNBS may serve as a potential reagent for sensitively and specifically detecting antioxidants, because DBNBS traps hydrogen radicals to generate a readily measured colored reduction product that can quantify antioxidant levels.
Here, we describe a new DBNBS-mediated antioxidant assay system in comparison with standard DPPH and ORAC assays. The DBNBS assay was also used to evaluate antioxidant levels in selected beverages in comparison with these standard assays.

Beverages and sample preparations for assays
Beverages purchased from a local store in Fukuoka, Japan (NAFCO Corp. Japan [geographical coordinates; 33.604051, 130.4 4284 4]) were as follows: white wine, red wine, grape juice, orange juice, tomato juice, vegetable juice, oolong tea, and barley tea (Table S1). Samples were contained in glass (white wine, red wine, and orange juice) or plastic (grape juice, tomato juice, vegetable juice, oolong tea, and barley tea) bottles. Samples were extracted using methanol. Briefly, samples (50 ml) were freeze-dried, the residues were dissolved in the same volume of methanol, vortexed for 5 min, and centrifuged at 10,0 0 0 ×g for 5 min. The supernatant was collected and stored at 4 °C in a nitrogen atmosphere. To compare assay systems, three samples of each beverage were first assayed using the DBNBS method (Table S2).

HPLC analysis
DBNBS-or DBNBS-derived reduction products were separated using an HPLC system equipped with a multidiode array detector (Waters 2095 and 996, Waters Co., Tokyo, Japan). The mobile phases were mixtures of H 2 O and CH 3 CN, and the stationary phase was a C 18 reverse-phase column (Sunniest C18, 5 μm, 250 × 4.6 mm; ChromaNik Technologies Inc., Osaka, Japan). Typically, 80 μl of sample was injected into the column equilibrated with 10% CH 3 CN in 0.1% trifluoroacetic acid (TFA), 10 mM ammonium acetate, or 5 mM ammonia/ammonium acetate. DBNBS-or DBNBS-derived reduction products were separated by linearly increasing the CH 3 CN concentration from 2% to 50% over 40 min at 1 ml/min.

DBNBS assay
The product of the colorimetric DBNBS assay was measured using a UV/Vis spectrometer (UV-2450, Shimadzu Co., Tokyo, Japan) or a 96-well microplate reader (Infinite F200 pro, Tecan Japan Co., Ltd., Kanagawa, Japan). The standard procedure was as follows: Reaction mixtures containing 200 μl of 30 mM DBNBS, 200 μl of 500 mM borate buffer (pH 9.0), and 600 μl of sample were incubated at room temperature (RT) and then at 40 °C or 70 °C for 1-24 h. Under alkaline conditions, DBNBS reactions develop a yellow color, which was initially monitored at 450 nm and subsequently at 460 nm, because HPLC analysis detected one absorption peak at 460 nm. The absorbance values of yellow samples were corrected by subtracting the absorbance values of unreacted samples from those of reacted samples. Samples were stored at 4 °C.

ORAC assay
ORAC assays were conducted according to a published method [18] . Briefly, 15 μM fluorescein (20 μl) containing 100 mM phosphate buffer (pH 7.0) was added to the wells of a black 96-well plate, followed by the addition of 25 μl each of the blank or the antioxidant standard (Trolox 5-40 μM). Freshly prepared 30 mM 2,2'-azobis (2-methylpropionamidine) dihydrochloride (15 μl) was added to all wells. Fluorescence (485-nm excitation and 528-nm emission) was monitored at 10-min intervals for 180 min. The net area under the reaction kinetics curve was calculated by subtracting the integrated value of the area under the curve of the blank from those of the test sample and Trolox. The TAC value of an individual test sample was calculated by subtracting the net Trolox value.

DPPH assay
A modified published DPPH assay was used to measure TAC [19] . Briefly, 500 μM (DPPH prepared in ethanol or acetonitrile) was added to reaction mixtures comprising 50 μl each of a blank, Trolox (standard), or test sample. Acetate butter (10 μl, pH 5.5) and DPPH (40 μl) were mixed and protected from light for 30 min. Absorbance values (515 nm) were measured using a microplate reader, and a standard curve was generated using 50-250 μM Trolox.

Data analysis
Microsoft Excel 2016 was used to calculate the mean value and standard deviation, or coefficient of determination, of triplicate experiments. One-way ANOVA (Statmate III for Windows) was used to evaluate the significance of differences between datasets.

DBNBS assay of known antioxidants
To validate the DBNBS assay system, we tested different concentrations of six polyphenols with antioxidant activity ( Fig 1 ). The plots of each sample were linear, with coefficient of determination values > 0.99 ( Fig 1 ), allowing the calculation of TAC DBNBS values. We then calculated TAC DBNBS values and expressed each antioxidant as the Trolox equivalent (TE, Table 1 ). The values measured by the ORAC and DPPH assays are shown as TAC ORAC and TAC DPPH for comparison.    Table 2 , and the relative TAC values were calculated by defining white wine as 100%. Data are expressed as the mean ± SD ( n = 3). * indicates a significant difference with respect to each DBNBS samples (p < 0.01).
We tested the commonly used standards gallic acid and chlorogenic acid as well as other antioxidants arranged in the order of the phenolic OH numbers as follows: kaempferol, 1-OH; (-)epicatechin, 2-OH; and rosmarinic acid, 4-OH. The TAC DBNBS values did not reflect the number of phenolic OH moieties, while ORAC and DPPH values were in the expected order [20 , 21] . The TE values of these three independent assay systems were consistent, supporting the conclusion that the DBNBS assay is comparable to existing TAC calculable assay systems. The results of the DBNBS and ORAC assays closely correlate, suggesting that the consumption of reactive hydrogen atoms and/or electrons by the radical generation system used in the ORAC assay are similar to the hydrogen atoms and/or electrons transfer reaction in the DBNBS assay. The hydrogenation reactions are influenced by solvents, and both methods employ an aqueous environment [20 , 22] . Moreover, hydrogen-donating ability depends on the structure of each polyphenol, which confers hydrogenation specificity [21][22][23][24] . Thus, the order of the antioxidant values for these six compounds determined by the DBNBS assay is consistent with those of the ORAC assay ( Table 1 ).
When we measured other antioxidants and non-antioxidants, we found that the results closely correlated with those of the ORAC assay (Table S3) [21][22][23][24] . This suggests that calculating relative values (e.g. white wine as 100% in Fig. 2 ) allows comparisons with other methods. Further, TAC DBNBS values are consistent with those acquired using conventional methods such as the ORAC assay. Moreover, the DBNBS assay was superior to the other two methods, because DBNBS was not affected by organic solvents or millimolar levels of metal ions (Table S3). However, DBNBS reacts with certain amino acids [25 , 26] , consistent with the present data (Table S3). Other improvements in assay conditions will be required to avoid this complication.

Determination of TAC values of beverage samples using the DBNBS assay
We next used the DBNBS assay system to determine the TAC values of several commonly consumed beverages compared with those using the ORAC and DPPH assays (TAC ORAC and TAC DPPH ) ( Table 2 ). Further, we used the DBNBS method to analyze triplicate samples of eight types of beverages (Table S2). The TAC DBNBS values of barley tea was undetectable. The TAC DBNBS values of other samples measured were: grape juice > red wine > orange juice > vegetable juice > white wine > tomato juice > oolong tea. Comparison of total antioxidant values revealed discrepancies among the results of the three assays, and significant differences were observed between the DBNBS and the other two methods. The TAC DBNBS values were approximately as much as 10 times higher compared with those of the other two methods ( Table 2 ). Differences between TAC DBNBS and TAC DPPH values were as much as 33 times higher for tomato and vegetable juices, indicating that the DBNBS assay may not reflect actual antioxidant levels in each sample.
We calculated relative ratios by defining the white wine values for each assay system as 100% ( Fig 2 ), which revealed that the TAC DBNBS and TAC ORAC results were similar. The DPPH assay yielded higher values for white and red wines and grape juice compared with those of the other two assays, while the values for orange juice and vegetable juice were lower compared with those of the other two assays. The main reason for the higher values obtained from the DBNBS assays compared to other two assays seems to be that DBNBS reacts with amino group-containing compounds (Table S3). The Folin-Ciocalteu reagent is occasionally used for the measurement of total phenolics [21] in present studies, and this method is known to react with nitrogen-containing compounds as well as reducing substances [27] , as in the DBNBS assays. The Folin-Ciocalteu reagent is also known to react with various inorganic ions [27] . In contrast, the DBNBS assay is expected to be non-reactive with metal ions, and is considered useful in certain sample applications. For example, while high magnification concentrations are required to detect extremely low amounts (part per-trillion level) of reductants in natural water, DBNBS assays are expected to detect such samples without considering the salt concentration. It is suggested that there are extremely low amounts of reduction products in natural water providing beneficial effects, therefore, our assay could be useful for such applications [28] .

Appendix A. Supplemental material
Supplemental data related to this article can be found in the online version. The data include detailed chemical analyses of intermediates produced by the DBNBS-atomic hydrogen reaction and TAC DBNBS values for selected samples, including various minerals.

Declaration of interests
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: