Inter-laboratory Validation Study of a Singlet Oxygen Absorption Capacity Assay Method for Determining the Antioxidant Capacities of Antioxidant Solutions and Food Extracts

Abstract

Singlet oxygen absorption capacity (SOAC) is an assay method used to evaluate the singlet oxygen quenching capacity. We evaluated the precision of the SOAC assay method in an inter-laboratory validation study using 3 antioxidant solutions and 3 food extracts. This study involved 14 participating laboratories, with data sets from 8 successfully participating laboratories subjected to statistical analysis. In accordance with the harmonized protocol, this study showed that the intermediate precision relative standard deviation (RSD_int) and the reproducibility relative standard deviation (RSD_R) ranged from 3.4 to 9.9%, and from 6.2 to 14.0%, respectively, and the HorRat values ranged from 0.87 to 1.97. Therefore, the SOAC assay method was validated by this inter-laboratory study in accordance with the internationally harmonized protocol.

Introduction

Carotenoids are natural pigments found mainly in fruits and vegetables. Mammals cannot synthesize carotenoids de novo, and therefore carotenoids must be obtained from the daily diet as vitamins. About 40 carotenoids are regularly identified in human serum and milk (Khachik et al., 1997). Carotenoids can quench singlet oxygen due to their chemical structure and this characteristic contributes to their beneficial effects. The dietary intake of carotenoid-rich fruits and vegetables has been associated with reduced risk of common diseases, including cancer, cardiovascular disease, age-related macular degeneration, and cataract formation (Rock et al., 1998). Hama et al. (2012) reported that astaxanthin prevented UV-induced skin damage by scavenging singlet oxygen production, and Li et al. (2010) showed that zeaxanthin and lutein prevented the light-induced production of singlet oxygen in human retinal pigment epithelium/choroid.

Singlet oxygen absorption capacity (SOAC) is an in vitro assay method used to evaluate the singlet oxygen quenching capacity of a sample and was developed by Mukai et al., 2015; Ouchi et al., 2010; Aizawa et al., 2011; Mukai et al., 2012; Iwasaki et al., 2015. This method measures absorption capacity for singlet oxygen induced by the thermal decomposition of 4-methyl-1,4-etheno-2,3-benzodioxin-1(4H)-propanoic acid (endoperoxide, EP) at 35°C using 2,5-diphenyl-3,4-benzofuran (DPBF) as a UV-Vis absorption probe. Time course changes in the absorption at 413 nm of DPBF are observed using a UV-Vis spectrophotometer equipped with a six-channel cell positioner. The protective effects of antioxidants are evaluated by assessing the half-life of the decay of DPBF or the first-order rate constant (k_Q) for the reaction of antioxidants with singlet oxygen. α-Tocopherol (α-Toc) is used as a standard antioxidant in the SOAC assay, and the antioxidant capacity of a sample is expressed as a value relative to that of α-Toc. This method allows evaluation of the singlet oxygen absorption capacity of both hydrophilic and lipophilic antioxidants in a variety of food extract such as from vegetables, fruits, and edible oils (Aizawa et al., 2011; Mukai et al., 2012; Iwasaki et al., 2015; Mukai et al., 2015). Moreover, the SOAC values of food extracts are nearly equal to theoretical values calculated from each SOAC value of antioxidants contained in these food extracts (Aizawa et al., 2011). These earlier reports suggest that the SOAC assay is a useful method for evaluating the singlet oxygen absorption capacity of food extracts without requiring analysis of the extract components.

Takahashi et al. (2016) modified the original SOAC assay method to allow the use of a 24-well glass microplate and a microplate reader because only a very limited number of laboratories have instruments comparable to a UV-Vis spectrophotometer equipped with a six-channel cell positioner. Moreover, the modified method allows higher throughput by measuring three samples simultaneously. However, the method's precision has not been evaluated in an inter-laboratory validation study. The precision of the method is needed when the SOAC values of antioxidants/food extracts refer to the reported papers and create an antioxidant database. Therefore, in this study, we conducted inter-laboratory tests in accordance with the internationally harmonized protocol to validate the modified SOAC assay method.

Materials and Methods

Reagents and chemicals    EP and (±)-α-Toc (CAS number: 10191-41-0) were purchased from Wako Pure Chemicals Industries (Tokyo, Japan). DPBF was purchased from Tokyo Kasei Chemicals (Tokyo, Japan). Capsanthin, astaxanthin, and zeaxanthin were purchased from Extrasynthese (Genay, France). All other chemicals were reagent grade.

Instruments    A 24-well glass microplate (FU-24, NGS Precision, Tokyo, Japan) was initially used (laboratories 1-3); however, because the manufacture of this glass microplate has been discontinued, a 24-well microplate made of polytetrafluoroethylene (PTFE; iTEC SCIENCE Co, Ltd., Ibaraki, Japan) was developed and also used (laboratories 4-14). A custom-made quartz glass plate (128 × 85 × 2 mm) was used to cover the 24-well glass microplate.

Test samples    Capsanthin, astaxanthin, and zeaxanthin were dissolved at 5 mg/L in ethanol/chloroform/D₂O (50:50:1, v/v/v) solvent to provide the antioxidant solutions. Carrot, tomato, and red paprika (bell pepper) were purchased from retail stores in Ibaraki, Japan, in December 2014, and were cut into small pieces, snap-frozen in liquid nitrogen, and lyophilized; samples were then pulverized using a grinder mill (GM-200; Retsch, Haan, Germany). Carrot powder (9.9 g dry weight (DW)), tomato powder (8 gDW), and red paprika powder (7 gDW) were divided into several batches (about 1 gDW each) and extracted with the ethanol/chloroform/D₂O solvent three times using an accelerated solvent extractor (ASE-350; Dionex, San Jose, CA, USA). The extracts from each vegetable were pooled and diluted to 1 L with ethanol/chloroform/ D₂O solvent. Aliquots (about 20 mL) of each antioxidant solution or extract were portioned into 50 glass vials; 5 vials were used for the homogeneity test, and the remainder were stored at _20°C until transport to the participants.

Homogeneity of the test samples    Five vials (units) were chosen at random for each test material and each unit was split into 2 equal parts (unit subsamples). The SOAC values of each unit subsample were measured as described below under repeatability conditions, i.e., the same method on identical test items in the same laboratory by the same operator using the same equipment within a short time scale.

The within- and between-unit standard deviations for the SOAC values were calculated by one-way analysis of variance (ANOVA) and by applying the F-test at the 95% confidence level. All p values of the test materials for the validation study were greater than 0.05, indicating that the samples were homogeneous (data not shown).

Stability of the test samples    The stabilities of the antioxidant solutions and food extracts were verified prior to the inter-laboratory study. Antioxidant solutions and food extracts were prepared in a manner similar to that described above, and about 20-mL aliquots of each antioxidant solution or food extract were placed in each of 16 vials. No significant changes in the SOAC values were observed by t-tests in the test materials during storage at −20°C for 7 weeks (data not shown).

Consequently, after preparation, all test samples were stored at −20°C until distribution. All test samples were packed in insulated boxes along with a cooling bag and sent to participants in chilled containers. Upon receipt of the test samples, participants were asked to immediately store the test samples in a freezer (−20°C) until use. In addition, samples were to be analyzed within 7 weeks, ensuring adequate stability.

SOAC measurement using a microplate reader    The SOAC values of the test samples were measured according to the method described by Takahashi et al. (2016). Briefly, 6-mL, 3-mL, and 2-mL aliquots of each test sample were transferred into 10-mL brown volumetric flasks and diluted with ethanol/chloroform/D₂O solvent. The undiluted and three diluted test samples, 1.5 mM α-Toc solution, and ethanol/chloroform/D₂O solvent (300 µL each) were added to the wells of a 24-well microplate using a digital analytical syringe (eVol XR; SGE Analytical Science, Victoria, Australia) on ice. DPBF solution (0.19 mM, 300 µL) and EP solution (1.2 mM, 300 µL) were added next to the wells of the same 24-well microplate using a digital analytical syringe, but were not added to the baseline wells. The final volume of each well was adjusted to 900 µL with ethanol/chloroform/D₂O solvent. Immediately after completion of these additions, the 24-well microplate was tightly sealed with a quartz glass cover and transferred to the microplate reader, which had been pre-incubated to 35°C. The absorbance at 413 nm of each well was measured every 30 s for 120 min at 35°C.

Calculation of the average relative SOAC value    As described in previous reports (Ouchi et al., 2010; Aizawa et al., 2011; Mukai et al., 2012; Iwasaki et al., 2015), the relative SOAC value is defined as:   

where t_1/2 is the half-life of DPBF, and [α-Toc] and [Sample] are the concentration of the α-Toc and test sample (mol/L; antioxidant solutions, or gDW/L; food extracts), respectively. A pseudo-first-order condition is satisfied in the time range 40 < t < 70 min in this method (Takahashi, et al., 2016). Therefore, t_1/2 was calculated using Equation 2:   

where k is the slope of the ln (absorbance) versus t plot of the disappearance of DPBF in the time range 40 < t < 70 min. Therefore, the relative SOAC values were calculated from each t_1/2 value of the diluted test samples, and these values were averaged. If the difference between the t_1/2 values of the test sample and the blank (t_1/2^Sample − t_1/2^blank) was smaller than 5 min, this value was excluded from the average relative SOAC value.

Inter-laboratory studies of SOAC measurement    The inter-laboratory study was performed in accordance with an internationally harmonized protocol (Horwiz, 1995) as recommended by CAC/GL 28-1995 and involved 14 participating laboratories analyzing 3 antioxidant solutions and 3 food extracts. Participants received a shipment containing 12 containers of test samples comprising two sets of test samples and each set consisted of 6 different solutions. Each sample was provided in duplicate and was randomly labeled. Participants were also provided with a method protocol and an electronic evaluation and reporting sheet (MS Excel format). Participants were requested to follow the method protocol exactly.

Statistical analyses of the inter-laboratory study data    Data could not be used from laboratories in which i) the SOAC measurements were unsuccessful due to leakage of solution from the 24-well glass or PTFE microplates, ii) the protocol was not followed, or iii) an extremely low absorbance of the DPBF solution was measured. All other data were subjected to the statistical tests described in the internationally harmonized protocol (Horwiz, 1995) recommended by CAC/GL 28-1995. The Cochran test and single and double Grubbs tests were conducted to detect outlying variances and outlying data set averages, respectively. The one-tail test at a probability value of 2.5% was applied to the Cochran test, and a probability value of 1.25% to the single and double Grubbs tests.

Results and Discussion

Measurements were unsuccessful at four of the laboratories due to instrumentation problems. Laboratories 11-13 reported that the solvent leaked from the bottom of the 24-well PTFE microplate, and laboratory 14 reported that the solvent spilled between the glass plate lid and 24-well microplate due to the swinging motion of the microplate reader. Therefore, further development of the 24-well microplate is required for the SOAC assay method. The values obtained for the various samples by the remaining 10 participating laboratories are given in Table 1. Laboratory 9 did not follow the method protocol and used too high a concentration of α-Toc. The data from laboratory 10 showed low absorption at 413 nm of DPBF at time zero, suggesting that the DPBF may have degraded prior to the measurement. Thus, the results from these laboratories were also excluded from the statistical analysis. In this study, differences in the microplates did not produce an observable effect.

Table 1. Inter-laboratory study results for the determination of the average relative SOAC values of the test samples using the SOAC assay method

	Laboratory a
	1	2	3	4	5	6	7	8	9	10
Capsanthin (mol α-Toc/mol)	121.3	129.3	139.7	133.6	118.9	129.9	136.6	131.7	128.7	104.2
	126.8	144.6	114.6	123.2	138.5	120.1	134.2	130.5	177.7	153.2
Zeaxanthin (mol α-Toc/mol)	124.7	161.3	119.3	138.7	131.3	142.0	109.3	132.7	129.3	176.0
	106.0	136.7	118.0	130.7	133.3	130.0	148.7	133.3	168.7	163.3
Astaxanthin (mol α-Toc/mol)	139.6	146.2	144.0	170.3	141.1	161.5	125.7	134.5	141.8	193.7
	114.8	144.0	141.8	158.6	145.5	149.1	133.0	152.0	173.2	195.2
Tomato (µmol α-Toc/gDW)	148.1	140.6	163.8	145.6	142.5	151.9	163.8	145.6	153.1	216.9
	155.0	133.8	135.6	149.4	133.1	156.3	161.9	146.3	178.1	150.6
Red paprika (µmol α-Toc/gDW)	265.0	236.4	234.3	267.1	257.1	274.3	265.0	245.7	265.0	294.3
	259.3	263.6	260.7	284.3	262.9	258.6	302.1	258.6	302.1	259.3
Carrot (µmol α-Toc/gDW)	123.7	123.7	116.2	118.2	124.2	132.3	126.3	122.7	121.7	152.5
	107.6	114.1	113.1	118.2	118.7	131.8	128.3	117.2	157.6	137.9

a.  Results of laboratories 1-8 were used for the statistical analyses. Results of laboratories 9 and 10 were excluded, because laboratory 9 did not follow the protocol, and absorbance of DPBF solution of laboratory 10 was beyond the tolerable ranges. Laboratories 11–14 failed to measure SOAC values of test samples.

The data sets of the 8 remaining participating laboratories were subjected to the statistical tests described in the Protocol for the Design, Conduct and Interpretation of Method Performance Studies (Horwitz, 1995), using the Cochran test to identify outlying variances and the single and double Grubbs tests to detect outlying data set averages. No data sets were removed as Cochran outliers or single Grubbs outliers (Table 1). The intermediate precision and reproducibility were calculated as defined by the protocol (Horwiz, 1995). The RSD_int and RSD_R of the test samples ranged from 4.3 to 9.9%, and from 7.5 to 14.0%, respectively (Table 2).

Table 2. Statistical analysis of data from the inter-laboratory study for the determination of the relative SOAC values

	Material
	Capsanthin	Zeaxanthin	Astaxanthin	Tomato	Red paprika	Carrot
Mean	129.6	131.0	143.9	148.3	262.2	121.0
Number of laboratory	8	8	8	8	8	8
Sinta	9.7	13.0	9.0	7.9	15.0	5.2
SRb	9.7	15.7	20.2	13.5	19.3	9.6
RSDintc	7.5	9.9	6.3	5.4	5.7	4.3
RSDRd	7.5	11.9	14.0	9.1	7.4	7.9
HorRate	1.05	1.68	1.97	1.27	1.03	1.11

a.  S_int = Intermediate precision standard deviation, mol α-Toc/mol (antioxidant solutions) or µmol α-Toc/gDW (food extracts).

b.  S_R = Reproducibility standard deviation, mol α-Toc/mol (antioxidant solutions) or µmol α-Toc/gDW (food extracts).

c.  RSD_int = Intermediate precision relative standard deviation, %.

d.  RSD_R = Reproducibility relative standard deviation, %.

e.  HorRat = RSD_R /predicted RSD_R. Predicted RSD_R was calculated from the concentration of α-Toc standard solution (0.5 mM) to be 7.13% using the Horwitz equation.

The precision data obtained in this inter-laboratory study were compared with the predicted levels of precision obtained from the Horwitz equation,   

where C is the commonly measured concentration of the analyte in the sample, expressed as a mass fraction. In the case of oxygen radical absorbance capacity methods, the predicted RSD_R of each sample was calculated relative to the concentration of Trolox, the standard antioxidant (Watanabe et al., 2012 and 2016). In this method, α-Toc is used as a standard antioxidant, and the antioxidant capacity of a sample is expressed as the α-Toc equivalent. Therefore, the predicted RSD_R of each sample was calculated from the final concentration of α-Toc. A mass fraction of 0.5 mM for α-Toc was calculated by multiplying the volume concentration by the molecular weight of α-Toc, and the predicted RSD_R was calculated to be 7.13% according to the Horwitz equation. The HorRat value, the ratio of RSD_R (measured) to the predicted RSD_R (Horwitz), gives a comparison between the actual precision and the precision predicted by the Horwitz equation. The calculated HorRat values can be used as a performance parameter, indicating the acceptability of the precision of the method. A HorRat value of <2 usually indicates satisfactory reproducibility, whereas a value of >2 usually indicates unsatisfactory performance of the method. The HorRat values of the test samples ranged from 0.96 to 1.97 (Table 2). All HorRat values of the test samples were less than 2, indicating that the SOAC measurement was validated by an inter-laboratory test in accordance with the internationally harmonized protocol (Horwitz, 1995).

In this study, we demonstrated the precision of the SOAC assay method by inter-laboratory validation involving 14 participating laboratories using 3 antioxidant solutions and 3 food extracts. However, the measurements were unsuccessful at four of the laboratories due to instrumentation problems and at two laboratories due to technical issues. Therefore, the development of a 24-well microplate is required for the SOAC assay method and protocol method.

Acknowledgements    This work was supported by a grant-in-aid, “A Scheme to Revitalize Agriculture and Fisheries in Disaster Area through Deploying Highly Advanced Technology” from the Ministry of Agriculture, Forestry and Fisheries of Japan. We express our appreciation to the following collaborators for their participation in this study: Yuji Takahashi (Japan Food Research Laboratories), Fumiaki Itou (Ehime Research Institute of Agriculture, Forestry and Fisheries), Katsuki Ohtani (Asahikawa Medical University), Kouta Isono (Kyoto Prefectural Agriculture, Forestry and Fisheries Technology Center), Ayako Fukunaga (Western Region Agricultural Research Center, National Agriculture and Food Research Organization), Nobuaki Urata (Nippon Paper Industries Co., Ltd.), Takashi Ishida (Kyowa Hakko Bio Co., Ltd.), Takashi Matsuyama (Vegetech Co., Ltd.), Kanako Fujino (Kumamoto Industrial Research Institute), Keiichi Negoro (Industrial Technology Center of Wakayama Prefecture), Masao Sasaki (Thermo Scientific Co.), Yoshiko Toyama (National Food Research Institute, National Agriculture and Food Research Organization), and Ayako Konishi (Kyoto Prefectural Agriculture, Forestry and Fisheries Technology Center). We would like to thank Prof. Kazuo Mukai (Ehime University) and Prof. Junji Terao (Tokushima University) for their comments.

References

•  Aizawa,  K.,  Iwasaki,  Y.,  Ouchi,  A.,  Inakuma,  T.,  Nagaoka,  S.,  Terao,  J., and  Mukai,  K. (2011). Development of singlet oxygen absorption capacity (SOAC) assay method. 2. Measurements of the SOAC values for carotenoids and food extracts. J. Agric. Food Chem., 59, 3717-3729.
•  Hama,  S.,  Takahashi,  K.,  Inai,  Y.,  Shiota,  K.,  Sakamoto,  R.,  Yamada,  A.,  Tsuchiya,  H.,  Kanamura,  K.,  Yamashita,  E., and  Kogure,  K. (2012). Protective effects of topical application of a poorly soluble antioxidant astaxanthin liposomal formulation on ultraviolet-induced skin damage. J. Pharm. Sci., 101, 2909-2916.
•  Horwitz,  W. (1995). Protocol for the design, conduct and interpretation of method-performance studies: Revised 1994 (Technical Report). Pure Appl. Chem., 67, 331-343.
•  Iwasaki,  Y.,  Takahashi,  S.,  Aizawa,  K., and  Mukai,  K. (2015). Development of singlet oxygen absorption capacity (SOAC) assay method. 4. Measurements of the SOAC values for vegetable and fruit extracts. Biosci. Biotechnol. Biochem., 79, 280-291.
•  Khachik,  F.,  Spangler,  C. J.,  Smith,  J. C. Jr,  Canfield,  L. M.,  Steck,  A., and  Pfander,  H. (1997). Identification, quantification, and relative concentrations of carotenoids and their metabolites in human milk and serum. Anal. Chem., 69, 1873-1881.
•  Li,  B.,  Ahmed,  F., and  Bernstein,  P. S. (2010). Studies on the singlet oxygen scavenging mechanism of human macular pigment. Arch. Biochem. Biophys., 504, 56-60.
•  Mukai,  K.,  Ishikawa,  E.,  Abe,  T.,  Ouchi,  A.,  Nagaoka,  S.,  Murata,  K.,  Miyazawa,  T., and  Nakagawa,  K. (2015). Kinetic study of the quenching reaction of singlet oxygen by seven rice bran extracts in ethanol solution. Development of a singlet oxygen absorption capacity (SOAC) assay method. Biosci. Biotechnol. Biochem., 79, 2063-2072.
•  Mukai,  K.,  Ouchi,  A.,  Takahashi,  S.,  Aizawa,  K.,  Inakuma,  T.,  Terao,  J., and  Nagaoka,  S. (2012). Development of singlet oxygen absorption capacity (SOAC) assay method. 3. Measurements of the SOAC values for phenolic antioxidants. J. Agric. Food Chem., 60, 7905-7916.
•  Ouchi,  A.,  Aizawa,  K.,  Iwasaki,  Y.,  Inakuma,  T.,  Terao,  J.,  Nagaoka,  S., and  Mukai,  K. (2010). Kinetic study of the quenching reaction of singlet oxygen by carotenoids and food extracts in solution. Development of a singlet oxygen absorption capacity (SOAC) assay method. J. Agric. Food Chem., 58, 9967-9978.
•  Rock,  C. L.,  Lovalvo,  J. L.,  Emenhiser,  C.,  Ruffin,  M. T.,  Flatt,  S. W., and  Schwartz,  S. J. (1998). Bioavailability of beta-carotene is lower in raw than in processed carrots and spinach in women. J. Nutr., 128, 913-916.
•  Takahashi,  S.,  Iwasaki-Kino,  Y.,  Aizawa,  K.,  Terao,  J., and  Mukai,  K. (2016). Development of singlet oxygen absorption capacity (SOAC) assay method using a microplate reader. J. AOAC Int., 99, 193-197.
•  Watanabe,  J.,  Oki,  T.,  Takebayashi,  J.,  Yada,  H.,  Wagaki,  M.,  Takano-Ishikawa,  Y., and  Yasui,  A. (2016). Improvement and interlaboratory validation of the lipophilic oxygen radical absorbance capacity: determination of antioxidant capacities of lipophilic antioxidant solutions and food extracts. Anal. Sci., 32, 171-175.
•  Watanabe,  J.,  Oki,  T.,  Takebayashi,  J.,  Yamasaki,  K.,  Takano-Ishikawa,  Y.,  Hino,  A., and  Yasui,  A. (2012). Method validation by interlaboratory studies of improved hydrophilic oxygen radical absorbance capacity methods for the determination of antioxidant capacities of antioxidant solutions and food extracts. Anal. Sci., 28, 159-165.