TY - JOUR
AU - Yonekura,, L
AB - Abstract Background Lutein is gaining attention as a strong antioxidant contained in foods. It accumulates in the human blood and retina, and is considered to play an important role in the body, especially in the eyes. Objective A method to determine the lutein content of raw spinach (Spinacia oleracea L.) was developed with the aim of its enactment as a Japanese agricultural standard (JAS) measurement method for components beneficial to human health. Methods An interlaboratory study was conducted to evaluate an analytical method for the determination of lutein in spinach. The detection limit and quantification limit of lutein for this method were 0.2 and 0.7 mg/kg, respectively. Twelve participating laboratories independently analyzed test samples (five pairs of blind duplicates) using high-performance liquid chromatography (HPLC). Results After removal of a few outliers, the repeatability relative standard deviation (RSDr), reproducibility (RSDR), and predicted RSDR of the evaluated method were 3.4–7.5, 4.6–13, and 7.5–8.5%, respectively, in a concentration range from 64.9–150 mg/kg. Conclusions The HorRat values (RSDR/predicted RSDR) of the lutein concentration were calculated to be 0.61–1.6. Highlights The study results indicate the acceptable precision of this method. Introduction Lutein (Figure 1), a major carotenoid in vegetables, is considered to play an important role in the human body, especially in the eyes, because it exerts an antioxidant effect. Moreover, its consistent intake is considered to increase the concentration of blood and macular pigment (1–8). Spinach is a major source of lutein (9), which humans ingest at an approximate rate of 1.52 mg per day (10). However, the content of lutein in spinach is affected by season, variety, fertilization, and cultivation conditions (11–13). Therefore, it is important to quantify the lutein content of spinach to guarantee its value as a beneficial food. Figure 1. Open in new tabDownload slide Chemical structures of lutein and zeaxanthin. Figure 1. Open in new tabDownload slide Chemical structures of lutein and zeaxanthin. To determine the concentrations of effective components, reliable analytical methods are necessary. Some studies have reported analysis methods that simultaneously measure major carotenoids such as β-carotene, lutein, and lycopene (14, 15). These methods have been improved to shorten the measurement time by limiting the analysis target. However, it is yet to be reported that interlaboratory studies of lutein meet the International Union of Pure and Applied Chemistry (IUPAC) interlaboratory study protocol (16). To establish a Japanese agricultural standard (JAS), we conducted an interlaboratory study to assess an analytical method for lutein in spinach within an appropriate concentration range. Experimental Principle Ground spinach is first homogenized and then saponified with potassium hydroxide in the presence of an antioxidant, pyrogallol, dissolved in ethanol. The carotenoids in the spinach homogenate are extracted with hexane and ethyl acetate. The extract is dried and then dissolved in 0.1% (w/v) 2,6-di-tert-butyl-p-cresol (BHT) ethanol solution. After filtration of the extract, lutein is determined by high-performance liquid chromatograph (HPLC) with a UV-visible detector at 445 nm. Apparatus Apparatus for sample preparation Homogenizer.—Capable of 1500 revolutions per min. Requirements for interlaboratory study participants: Electronic balance.—Readability ± 0.1 mg. Centrifuge tubes.—Glass or polypropylene, 50 mL capacity, stoppered, sufficient space for adequate mixing and centrifugation at a 400 relative centrifugal force. Lid coated with or made of an appropriate material resistant to organic solvents and a strongly basic solution [e.g., polytetrafluoroethylene (PTFE)]. Shaker.—For bobbing or back-and-forth motion while holding tubes. Centrifuge.—Capable of a 400 relative centrifugal force. One-mark volumetric flasks.—To cover the volume range for standard dilution and sample extract dissolution, ISO1042, Class A. Single volume pipettes.—To cover the volume range for standard dilutions, ISO648, Class A. Piston pipettes.—To cover the volume range for standard dilutions, ISO 8655-2, Type A. Water bath.—Capable of being maintained at 70°C (±3°C) and of sufficient size to hold a centrifuge tube rack. Glass vessels.—Amber, to cover the volume range for lutein stock standard dilutions. Eggplant-shaped flask.—Of 100 mL capacity, ground neck, amber, usable for evaporation. Vacuum rotating evaporator.—With water bath and vacuum control; for evaporation of solvent, e.g., hexane, ethyl acetate, or ethanol. Membrane filters.—For organic solutions, PTFE material, pore size of <0.20 µm, filter and housing are a single-unit, housing material resistant to organic solvents. Vials.—Suitable for HPLC, amber, deactivated glass, deactivated insert vials (or other glass vials that have been determined to have no influence on measurement), septum of lid coated with PTFE or made of PTFE. Spectrometer.—Capable measuring at a wavelength of 445 nm, holding cells. Absorption cells.—Quartz glass or glass, optical path length 1 cm, stoppers (if multiple cells are used, they are guaranteed to have the same optical characteristics). HPLC.—Equipped to perform binary pump elution (or the solvent line can be switched manually), degassing unit, thermostatically controlled column compartment, UV-visible detector set at 445 nm, data collection/integration system. Chromatographic column.—Reverse-phase C30 (Triacontyl), 4.6 mm internal diameter, 250 mm length, 5 µm particle size, chemically bonded C30 base in a stainless steel column. Reagents Requirements for interlaboratory study participants [except for reagent (c)]: Water.—Conforming to grade A2, A3, or A4 of Japanese industrial standard (JIS) K 0557. Lutein.—Of minimum mass fraction (C40H56O2) ≥95.0%, purchased from ChromaDex Inc., CA. Zeaxanthin.—Of minimum mass fraction (C40H56O2) ≥95.0%, purchased from ChromaDex Inc. Ethanol.—Of minimum mass fraction (C2H6O) ≥99.5%, according to JIS K 8101. Pyrogallol.—Purity ≥99.0%, according to JIS K 8780. Potassium hydroxide.—Of minimum mass fraction (KOH) ≥85.0%, according to JIS K 8574. Sodium chloride.—Of minimum mass fraction (NaCl) ≥99.5%, according to JIS K 8150. Hexane.—Purity ≥96.0%, according to JIS K 8848. Ethyl acetate.—Of minimum mass fraction (C4H8O2) ≥99.5%, according to JIS K 8361. Methanol.—HPLC grade. Acetonitrile.—HPLC grade. Ethanol (HPLC).—HPLC grade. Ammonium acetate.—According to JIS K 8359. Nitrogen.—Of volume fraction (N2) ≥99.5%. BHT.—Of minimum mass fraction (C15H24O) ≥98.0%. Pyrogallol ethanol solution.—Dissolve 30 g of pyrogallol per 1.0 L of ethanol; prepare as necessary. Potassium hydroxide (water) solution.—Dissolve 60 g of potassium hydroxide per 100 mL of water. Sodium chloride (water) solution.—Dissolve 10 g of sodium chloride per 1.0 L of water. Hexane–ethyl acetate mixture.—Mix 9 parts per volume of hexane with 1 part per volume of ethyl acetate. BHT ethanol solution.—Dissolve 1.0 g of BHT per 1.0 L of ethanol. HPLC mobile phase A (Ammonium acetate methanol–acetonitrile mixture).—Dissolve 5 g of ammonium acetate per 1.0 L of methanol, mix 20 parts per volume of this solution with 75 parts per volume of acetonitrile. HPLC mobile phase B.—Ethanol (HPLC). Lutein stock standard solution (corresponding to approximately 100 µg/mL).—Dissolved lutein in ethanol at a concentration of approximately 100 µg/mL. Transfer the lutein solution into a labeled bottle with a stopper and store at −20°C or lower temperature in light-shielded condition. Lutein is purified as follows: Add an acetone solution of lutein to a small amount of silica gel (Wakogel C-100, FUJIFILM Wako Pure Chemical Corp., Osaka, Japan). Evaporate the solvent and transfer it to a chromatographic tube packed with silica gel. Conduct stepwise elution using a hexane–ethyl acetate mixture (10:0 to 6:4, v/v). Recover lutein from the hexane–ethyl acetate fraction (6:4, v/v). Evaporate the organic solvent and dissolve the lutein in ethanol to approximately 100 ppm concentration. Store at −20°C or lower temperature. Correct concentration is determined by absorbance measurement. Standard stock solution for absorbance measurement.—Before the absorbance measurement, return the lutein stock solutions to room temperature. Dilute the lutein stock solution 50-fold with ethanol. Calculate the lutein concentration of the stock standard solution, ρ (µg/mL), using the formula: ρ=A×V1×10000ε×V2, where A is the absorbance of standard solution at 445 nm (ethanol, 1 cm cell); ε is the absorption coefficient of lutein at a concentration of 1% and optical pathlength of 1 cm (=2550) (17); V1 is volumetric flask capacity (mL) (=10 mL); and V2 is the volume (mL) of stock standard solution (=0.200 mL). (y)Dilution series of stock standard solution.—Dilution procedure is conducted on the same day as absorbance measurement. Prepare a dilution series of lutein stock standard solutions for HPLC measurement using the same single-stock standard solution as that used for the absorbance measurement. Place certain amounts of lutein stock standard solutions into vials and gently evaporate the solvent under a stream of nitrogen. Re-dissolve the residues in BHT ethanol to concentrations of approximately 1.0, 2.0, 5.0, 10.0, and 20.0 µg/mL. The lutein concentration of the dilute standard solution, ρi, (µg/mL), is given by the formula: ρi=ρ×V3V4, where ρ is the concentration of lutein stock standard solution (µg/mL), V3 is the volume of lutein stock standard solution (mL), and V4 is the volume of BHT ethanol (mL). Preliminary Study HPLC analytical column Because C18 columns did not separate lutein and zeaxanthin well (data not shown), C30 columns, which have commonly been used for the separation of carotenoid isomers, were used. As shown in Figure 2, the columns could separate standards of lutein and zeaxanthin under the conditions described in the following section. The resolution [as defined in the United States Pharmacopeia (18)] of the two carotenoids peaks between lutein and zeaxanthin was >1.5. Figure 2. Open in new tabDownload slide Separation of lutein and zeaxanthin standards on C30 stationary phase. Figure 2. Open in new tabDownload slide Separation of lutein and zeaxanthin standards on C30 stationary phase. Eluent conditions Acetonitrile, methanol, and ethanol were selected as eluents because they are relatively less harmful than the alternatives. The mixing ratio was 15:4:1, respectively. The effect of ammonium acetate on the sensitivity of the lutein peak was evaluated. Consistent with previous reports (19, 20), ammonium acetate improved peak sensitivity (data not shown). The linearity of the calibration curve was also improved (Figure 3). Therefore, HPLC mobile phase A included ammonium acetate. In this method, HPLC analysis was performed using a mixture of mobile phases A and B. Figure 3. Open in new tabDownload slide Difference between two lutein calibration curves. Ammonium acetate was added (filled squares) or not added (filled triangles) to mobile phase A. The correlation coefficients of the calibration curves were 1.000 and 0.997, respectively. Figure 3. Open in new tabDownload slide Difference between two lutein calibration curves. Ammonium acetate was added (filled squares) or not added (filled triangles) to mobile phase A. The correlation coefficients of the calibration curves were 1.000 and 0.997, respectively. Saponification In this study, saponification was performed because there was a concern that an unknown peak overlapping with the lutein peak appeared with a certain C30 column and interfered with lutein quantification. After saponification, the unknown peak, which is considered to likely have been chlorophyll based on a previous report (21), disappeared (Figure 4). Another C30 column, with which chlorophyll peaks did not appear, was used to confirm the effect of saponification on lutein quantification regardless of the presence of interference peaks. A typical HPLC chromatogram of spinach extract is shown in Figure 5. The major peak in the spinach extract corresponds to lutein (all-trans form). The limits of detection and quantification (LoQ) for lutein were estimated as 0.2 and 0.7 mg/kg spinach, respectively, which were calculated from the signal-to-noise ratios of 3 and 10, respectively, in the lutein chromatogram, assuming the highest values among the three columns. A trace peak of zeaxanthin appeared around 17 min, but it was below the lower limit of detection. A previous report (22) assigned the two small peaks around 10 min to the 13-cis and 13′-cis forms of lutein. However, in the present study, these peaks were too small to confirm a lutein-specific spectrum with the photodiode array detector of HPLC. In addition, standard products of 13-cis and 13′-cis-lutein are not commercially available, increasing the difficulty in identifying these two peaks. If these peaks indeed correspond to 13-cis and 13′-cis-isomers, they appeared below the lower limit of quantification. The amount of 13-cis-lutein in spinach is reportedly very low, constituting only 1.9% of all-trans-lutein (23). Therefore, only the all-trans form of lutein was analyzed in the present interlaboratory study. Figure 4. Open in new tabDownload slide Change of peak pattern in test sample before (upper) and after (lower) saponification. Figure 4. Open in new tabDownload slide Change of peak pattern in test sample before (upper) and after (lower) saponification. Figure 5. Open in new tabDownload slide Typical HPLC chromatogram of spinach extract. Figure 5. Open in new tabDownload slide Typical HPLC chromatogram of spinach extract. Recovery test To verify trueness, a test material that included 88.0 mg/kg lutein was used for recovery tests. Spiked samples were prepared by adding 0.5 and 0.25 mL of 60 μg/mL standard solutions to 2 g of test sample. The average recoveries were 107 and 105% (n = 3), respectively. The correlation coefficient was 1.0000. The marginal recoveries of lutein were within the acceptable range specified in the Codex criteria (90–107% at concentrations of 100 mg/kg) (24). Sample stability Before initiation of the interlaboratory study, the stability of lutein in the homogenized test materials stored at −20°C or less for 4 days and 6 months was tested. After storage, the average concentrations of lutein (n = 3) were 78.4 mg/kg (Sr = 0.68 mg/kg) at 4 days and 77.3 mg/kg (Sr = 1.4 mg/kg) at 6 months. No significant difference was observed between these groups (p > 0.05), as shown in Figure 6, indicating that lutein in the test samples was stable during the interlaboratory study under these conditions. In addition, lutein in the test solutions was stable at −20°C or less for 2 weeks. If interlaboratory participants stored the solutions in a freezer after lutein extraction, the solutions were allowed to reach room temperature before HPLC measurement. Figure 6. Open in new tabDownload slide Stability of lutein in test sample at 6 months. The data are presented as the mean ± SD (n = 3). Figure 6. Open in new tabDownload slide Stability of lutein in test sample at 6 months. The data are presented as the mean ± SD (n = 3). Interlaboratory Study Test materials In total, 39 fresh spinach plants were purchased from local food markets and provided to researchers from July 2016 to April 2017. These plants were used as test materials in this study. The concentrations of lutein in these spinach plants were 64.9–150 mg/kg, and five test materials covering the same concentration range were selected for the interlaboratory study. Other test materials were selected to verify the trueness and stability of the data. Sample preparation Because of the large individual differences in the components of fresh vegetables, the test samples were homogenized and subdivided to confirm their homogeneity before distributing them to the laboratory participants. The test material was prepared by removing the spinach roots, mincing the remaining major components, and weighing them into a homogenizer cup. Pyrogallol ethanol of three times the weight of each minced sample was added to the cup and then the spinach was pulverized using a homogenizer for 5 min (5000 rpm). At room temperature, the pulverized materials were portioned into 6 mL amber containers, ensuring each container contained 2.5 g of spinach homogenate. Each spinach homogenate was portioned into 60 such containers. Ten containers were selected randomly and used for a homogeneity test. Another 24 containers were used for the interlaboratory test (12 laboratories received a blind pair of test samples), and the remaining 26 containers were used as spares. All containers were labeled and preserved at −20°C or less before distribution to participating laboratories. Homogeneity The 10 containers used for the homogeneity test were selected randomly. Two 1 g test samples were taken from each container and their homogeneity was tested according to the international harmonized protocol of IUPAC proficiency testing (25; Table 1). After the homogeneity was checked, five pairs of bottles for each of the five test materials were distributed to each laboratory as blind duplicates. Samples were frozen during transportation. In brief, each laboratory received 10 blind test samples. Table 1. Lutein contents of raw spinach determined by proposed method in the homogeneity test for the interlaboratory study . Materiala A . Material B . Material C . Material D . Material E . Container number 1 2 1 2 1 2 1 2 1 2 1 64.5 65.3 68.1 75.2 90.2 94.7 121 120 144 148 2 67.5 64.8 71.6 71.9 90.9 96.5 114 122 147 140 3 62.3 63.3 74.9 73.0 90.4 84.1 115 120 148 148 4 64.9 67.1 73.5 75.4 80.7 88.6 124 118 148 151 5 64.4 66.9 79.5 74.0 90.5 87.6 122 120 148 147 6 65.3 66.1 73.4 75.0 81.9 86.1 117 118 144 148 7 62.6 65.4 74.9 76.0 92.0 83.4 123 120 150 150 8 63.3 64.7 78.4 77.9 89.8 98.5 113 123 150 149 9 63.6 65.1 75.8 77.8 94.7 93.3 119 123 151 153 10 62.3 62.8 74.7 78.1 87.3 81.4 118 111 148 153 n 20 20 20 20 20 Grand mean 64.6 75.0 89.1 119 148 Standard deviation σp (Horwitz equation) 5.52 6.26 7.25 9.28 11.2 Largest value in Cochran testb 0.238 0.469 0.207 0.328 0.405 San2 1.65 5.38 18.3 15.3 6.05 Ssam2 0.841 2.00 8.14 0 3.49 σ2all 2.74 3.53 4.74 7.74 11.2 Critical value (F1σ2all + F2 San2c) 6.82 12.1 27.4 30.0 27.2 Ssam2 ≤ F1 σ2all + F2 San2 Passd Pass Pass Pass Pass . Materiala A . Material B . Material C . Material D . Material E . Container number 1 2 1 2 1 2 1 2 1 2 1 64.5 65.3 68.1 75.2 90.2 94.7 121 120 144 148 2 67.5 64.8 71.6 71.9 90.9 96.5 114 122 147 140 3 62.3 63.3 74.9 73.0 90.4 84.1 115 120 148 148 4 64.9 67.1 73.5 75.4 80.7 88.6 124 118 148 151 5 64.4 66.9 79.5 74.0 90.5 87.6 122 120 148 147 6 65.3 66.1 73.4 75.0 81.9 86.1 117 118 144 148 7 62.6 65.4 74.9 76.0 92.0 83.4 123 120 150 150 8 63.3 64.7 78.4 77.9 89.8 98.5 113 123 150 149 9 63.6 65.1 75.8 77.8 94.7 93.3 119 123 151 153 10 62.3 62.8 74.7 78.1 87.3 81.4 118 111 148 153 n 20 20 20 20 20 Grand mean 64.6 75.0 89.1 119 148 Standard deviation σp (Horwitz equation) 5.52 6.26 7.25 9.28 11.2 Largest value in Cochran testb 0.238 0.469 0.207 0.328 0.405 San2 1.65 5.38 18.3 15.3 6.05 Ssam2 0.841 2.00 8.14 0 3.49 σ2all 2.74 3.53 4.74 7.74 11.2 Critical value (F1σ2all + F2 San2c) 6.82 12.1 27.4 30.0 27.2 Ssam2 ≤ F1 σ2all + F2 San2 Passd Pass Pass Pass Pass a All results are in units of mg/kg. b Critical value of Cochran test statistic for duplicates =0.718 (10 samples were measured in duplicate, 99% confidence). c F1 and F2 are the factors in the test for sufficient homogeneity. When 10 samples were measured in duplicate, we obtained F1 = 1.88 and F2 = 1.01. d Homogeneity test passed. Open in new tab Table 1. Lutein contents of raw spinach determined by proposed method in the homogeneity test for the interlaboratory study . Materiala A . Material B . Material C . Material D . Material E . Container number 1 2 1 2 1 2 1 2 1 2 1 64.5 65.3 68.1 75.2 90.2 94.7 121 120 144 148 2 67.5 64.8 71.6 71.9 90.9 96.5 114 122 147 140 3 62.3 63.3 74.9 73.0 90.4 84.1 115 120 148 148 4 64.9 67.1 73.5 75.4 80.7 88.6 124 118 148 151 5 64.4 66.9 79.5 74.0 90.5 87.6 122 120 148 147 6 65.3 66.1 73.4 75.0 81.9 86.1 117 118 144 148 7 62.6 65.4 74.9 76.0 92.0 83.4 123 120 150 150 8 63.3 64.7 78.4 77.9 89.8 98.5 113 123 150 149 9 63.6 65.1 75.8 77.8 94.7 93.3 119 123 151 153 10 62.3 62.8 74.7 78.1 87.3 81.4 118 111 148 153 n 20 20 20 20 20 Grand mean 64.6 75.0 89.1 119 148 Standard deviation σp (Horwitz equation) 5.52 6.26 7.25 9.28 11.2 Largest value in Cochran testb 0.238 0.469 0.207 0.328 0.405 San2 1.65 5.38 18.3 15.3 6.05 Ssam2 0.841 2.00 8.14 0 3.49 σ2all 2.74 3.53 4.74 7.74 11.2 Critical value (F1σ2all + F2 San2c) 6.82 12.1 27.4 30.0 27.2 Ssam2 ≤ F1 σ2all + F2 San2 Passd Pass Pass Pass Pass . Materiala A . Material B . Material C . Material D . Material E . Container number 1 2 1 2 1 2 1 2 1 2 1 64.5 65.3 68.1 75.2 90.2 94.7 121 120 144 148 2 67.5 64.8 71.6 71.9 90.9 96.5 114 122 147 140 3 62.3 63.3 74.9 73.0 90.4 84.1 115 120 148 148 4 64.9 67.1 73.5 75.4 80.7 88.6 124 118 148 151 5 64.4 66.9 79.5 74.0 90.5 87.6 122 120 148 147 6 65.3 66.1 73.4 75.0 81.9 86.1 117 118 144 148 7 62.6 65.4 74.9 76.0 92.0 83.4 123 120 150 150 8 63.3 64.7 78.4 77.9 89.8 98.5 113 123 150 149 9 63.6 65.1 75.8 77.8 94.7 93.3 119 123 151 153 10 62.3 62.8 74.7 78.1 87.3 81.4 118 111 148 153 n 20 20 20 20 20 Grand mean 64.6 75.0 89.1 119 148 Standard deviation σp (Horwitz equation) 5.52 6.26 7.25 9.28 11.2 Largest value in Cochran testb 0.238 0.469 0.207 0.328 0.405 San2 1.65 5.38 18.3 15.3 6.05 Ssam2 0.841 2.00 8.14 0 3.49 σ2all 2.74 3.53 4.74 7.74 11.2 Critical value (F1σ2all + F2 San2c) 6.82 12.1 27.4 30.0 27.2 Ssam2 ≤ F1 σ2all + F2 San2 Passd Pass Pass Pass Pass a All results are in units of mg/kg. b Critical value of Cochran test statistic for duplicates =0.718 (10 samples were measured in duplicate, 99% confidence). c F1 and F2 are the factors in the test for sufficient homogeneity. When 10 samples were measured in duplicate, we obtained F1 = 1.88 and F2 = 1.01. d Homogeneity test passed. Open in new tab Collaborating laboratories Twelve laboratories participated in the study. Each of the 12 laboratories analyzed 10 test samples (5 pairs of blind duplicates) according to the method described above. In addition to the test samples, each laboratory received a stock standard solution. All items were stored at −20°C or less until measurement. Before measurement, samples were allowed to reach room temperature. Each laboratory also prepared a dilution series of the stock standard solution. Procedure The test sample, i.e., spinach homogenate portioned into a container (2.00 g), was incubated with 10 mL of pyrogallol ethanol and 1 mL of potassium hydroxide solution in a centrifuge tube at 70°C for 30 min and then cooled to room temperature. To extract the lutein-containing carotenoids following saponification, 20 mL of 1% salt water and 12 mL of hexane–ethyl acetate (9:1 mixture) were added to the centrifuge tube, followed by shaking for 5 min. After centrifugation (400 relative centrifuge force) for 5 min at room temperature, the supernatant was transferred to a 100 mL eggplant-shaped flask using a pipette. The extraction procedure was then repeated from the hexane–ethyl acetate addition step to the supernatant transfer step. The organic solvent was dried using a vacuum rotary evaporator at <40°C and a subsequent nitrogen stream. The residue in the eggplant-shaped flask was dissolved in BHT ethanol solution and then transferred to a 10 mL volumetric flask. The volumetric flask was filled with BHT ethanol solution. The test solution resulting after filtration through a membrane filter was transferred to a vial and then analyzed using HPLC, as described below. Dilution series of standard solutions and test solutions were injected into the HPLC and lutein was detected via a photo diode array. The amounts of lutein in the test solutions were calculated based on their peak areas using an absolute calibration method. The following HPLC operating conditions were used: flow rate, 1.0 mL/min; column oven temperature, 40°C; injection volume, 10 µL; and detector wavelength, 445 nm. When a binary pump system was used (Table 2), 95% mobile phase A and 5% mobile phase B were used for 15 min after the injection. Then, mobile phase B was increased to 95% to rapidly elute remaining analytes from the column for 15 min. The ratio was then restored to 95% mobile phase A and 5% mobile phase B and allowed to equilibrate for 10 min before the next injection. When using manual solvent line switching, mobile phase A and mobile phase B were mixed beforehand. Table 2. Elution conditions of the binary pump system Time, min . Mobile phase, % . Objects . Aa . Bb . 0–15 95 5c Elution of lutein 15–30 5 95 Washingd 30–40 95 5 Column equilibration Time, min . Mobile phase, % . Objects . Aa . Bb . 0–15 95 5c Elution of lutein 15–30 5 95 Washingd 30–40 95 5 Column equilibration a A= Acetonitrile–methanol (75:20, v/v) containing 0.11% ammonium acetate. b B= Ethanol. c The corresponding channel for these mobile phases was set in the range of 5–95% to avoid crossflow (internal leakage) between the two channel ports of the pump. d Elution of compounds such as β-carotene (not analyzed here). Open in new tab Table 2. Elution conditions of the binary pump system Time, min . Mobile phase, % . Objects . Aa . Bb . 0–15 95 5c Elution of lutein 15–30 5 95 Washingd 30–40 95 5 Column equilibration Time, min . Mobile phase, % . Objects . Aa . Bb . 0–15 95 5c Elution of lutein 15–30 5 95 Washingd 30–40 95 5 Column equilibration a A= Acetonitrile–methanol (75:20, v/v) containing 0.11% ammonium acetate. b B= Ethanol. c The corresponding channel for these mobile phases was set in the range of 5–95% to avoid crossflow (internal leakage) between the two channel ports of the pump. d Elution of compounds such as β-carotene (not analyzed here). Open in new tab Determination Quantitative determination was performed using the external standard method with integration of peak area, which was then related to the corresponding value for the standard substance. The lutein concentration (µg/mL) was calculated in each series of standard solutions. Linear calibration graphs were constructed using the peak areas of the standards. The correlation coefficient of the linear calibration was required to be >0.995. The concentration (µg/mL) of lutein in an individual sample solution was calculated using the linear calibration. The lutein content, Wi, expressed as a mass percentage of the spinach sample, was given by the formula: Wi =C×V5M×MspMsp+Met, where C is concentration (µg/mL) of lutein in the test solution; V5 is constant volume (mL) at the time of dissolution of recovered extracted carotenoids in the interlaboratory tests (=10); M is the mass (g) of the sample test portion; Msp is the mass (g) of spinach sample in the preparation; and Met is the mass (g) of pyrogallol ethanol in the preparation. Results and Discussion System Suitability and Evaluation of Chromatograms System suitability and chromatograms were assessed during this study. Analytical dates and chromatograms were provided by all participants. The correlation coefficients of the calibration curves ranged from 0.996–1.000, which exceeded the minimum requirement of 0.995. Chromatograms were checked for consistency in terms of lutein peak retention time. The peak shape and resolution of the lutein were also examined. Laboratories 7 and 10 provided outlier results for Material A (Table 3); however, their chromatograms showed good resolution and the retention time of lutein was consistent. This indicated an absence of chromatographic issues. Table 3. Raw data of lutein contentsa Laboratory . Materialb A . Material B . Material C . Material D . Material E . 1 63.8 66.7 71.6 68.5 96.3 85.9 123 121 161 147 2 71.4 62.3 31.4c 74.8c 92.3 109 128 130 150 161 3 65.7 64.6 73.5 76.7 105 92.2 126 123 155 151 4 66.0 65.9 79.0 77.5 105 94.6 117 122 152 155 5 64.5 62.1 69.9 74.8 87.8 78.7 118 115 150 152 6 61.0 67.2 77.1 67.6 84.0 95.9 122 122 152 155 7 46.9d 63.4d 67.7 59.6 71.5 63.5 99.8 115 160 146 8 58.9 65.8 68.0 72.1 79.9 88.7 107 119 146 144 9 63.3 64.5 73.5 70.7 92.5 95.4 123 116 142 143 10 60.6d 54.4d 65.2 68.9 72.9 71.7 112 116 140 144 11 64.5 66.0 65.6 71.7 84.1 88.2 115 106 140 140 12 69.8 64.5 80.5 79.8 98.3 96.7 126 128 161 151 Laboratory . Materialb A . Material B . Material C . Material D . Material E . 1 63.8 66.7 71.6 68.5 96.3 85.9 123 121 161 147 2 71.4 62.3 31.4c 74.8c 92.3 109 128 130 150 161 3 65.7 64.6 73.5 76.7 105 92.2 126 123 155 151 4 66.0 65.9 79.0 77.5 105 94.6 117 122 152 155 5 64.5 62.1 69.9 74.8 87.8 78.7 118 115 150 152 6 61.0 67.2 77.1 67.6 84.0 95.9 122 122 152 155 7 46.9d 63.4d 67.7 59.6 71.5 63.5 99.8 115 160 146 8 58.9 65.8 68.0 72.1 79.9 88.7 107 119 146 144 9 63.3 64.5 73.5 70.7 92.5 95.4 123 116 142 143 10 60.6d 54.4d 65.2 68.9 72.9 71.7 112 116 140 144 11 64.5 66.0 65.6 71.7 84.1 88.2 115 106 140 140 12 69.8 64.5 80.5 79.8 98.3 96.7 126 128 161 151 a All results are in units of mg/kg. b Each material column involves blind duplicate results. c Removed as abnormal figures because they did not comply with the test protocol. d Outliers in first-paired Grubbs tests. Open in new tab Table 3. Raw data of lutein contentsa Laboratory . Materialb A . Material B . Material C . Material D . Material E . 1 63.8 66.7 71.6 68.5 96.3 85.9 123 121 161 147 2 71.4 62.3 31.4c 74.8c 92.3 109 128 130 150 161 3 65.7 64.6 73.5 76.7 105 92.2 126 123 155 151 4 66.0 65.9 79.0 77.5 105 94.6 117 122 152 155 5 64.5 62.1 69.9 74.8 87.8 78.7 118 115 150 152 6 61.0 67.2 77.1 67.6 84.0 95.9 122 122 152 155 7 46.9d 63.4d 67.7 59.6 71.5 63.5 99.8 115 160 146 8 58.9 65.8 68.0 72.1 79.9 88.7 107 119 146 144 9 63.3 64.5 73.5 70.7 92.5 95.4 123 116 142 143 10 60.6d 54.4d 65.2 68.9 72.9 71.7 112 116 140 144 11 64.5 66.0 65.6 71.7 84.1 88.2 115 106 140 140 12 69.8 64.5 80.5 79.8 98.3 96.7 126 128 161 151 Laboratory . Materialb A . Material B . Material C . Material D . Material E . 1 63.8 66.7 71.6 68.5 96.3 85.9 123 121 161 147 2 71.4 62.3 31.4c 74.8c 92.3 109 128 130 150 161 3 65.7 64.6 73.5 76.7 105 92.2 126 123 155 151 4 66.0 65.9 79.0 77.5 105 94.6 117 122 152 155 5 64.5 62.1 69.9 74.8 87.8 78.7 118 115 150 152 6 61.0 67.2 77.1 67.6 84.0 95.9 122 122 152 155 7 46.9d 63.4d 67.7 59.6 71.5 63.5 99.8 115 160 146 8 58.9 65.8 68.0 72.1 79.9 88.7 107 119 146 144 9 63.3 64.5 73.5 70.7 92.5 95.4 123 116 142 143 10 60.6d 54.4d 65.2 68.9 72.9 71.7 112 116 140 144 11 64.5 66.0 65.6 71.7 84.1 88.2 115 106 140 140 12 69.8 64.5 80.5 79.8 98.3 96.7 126 128 161 151 a All results are in units of mg/kg. b Each material column involves blind duplicate results. c Removed as abnormal figures because they did not comply with the test protocol. d Outliers in first-paired Grubbs tests. Open in new tab Precision The results of the study are presented in Table 3. For Material B, the results from Laboratory 2 were disregarded before the outlier tests, because they did not comply with the test protocol. Based on the first-paired Grubbs tests for Material A, the results of Laboratories 7 and 10 were considered outliers and removed. No other results were removed by the outlier tests. The laboratories that submitted outlier results were found to have complied with the test protocol; no reason for the outliers was identified. As shown in Table 4, the average lutein concentration range in the test materials was 64.9–150 mg/kg; this range was selected to cover the concentration range for domestic commercial spinach. Therefore, the HPLC analysis procedure is an appropriate method for the measurement of lutein in domestically cultivated spinach. The ranges of repeatability standard deviation (Sr), repeatability (RSDr), reproducibility standard deviation (SR), and reproducibility (RSDR) were 3.3–6.7 mg/kg, 3.4–7.5%, 3.6–12 mg/kg, and 4.6–13%, respectively. In the SR calculation for Material A, ubb [ISO Guide 35 7.9 Equation (6) (26)], was used as an estimate of pure between laboratory standard deviation (SL) because SR was smaller than Sr and it was necessary to reduce the risk of SR underestimation. Predicted RSDR values derived from the Horwitz equation (27) were 7.5–8.5%, and HorRat values derived from the ratio of RSDR to predict RSDR were 0.61–1.6. In food analysis, the Horwitz equation has been widely applied as the criterion for assessing deviation in chemical analysis methods; the appropriate range for the HorRat ratio is considered to be 0.5–2.0. If a measured HorRat ratio is within this range, its SR value is considered to be adequate (28). All HorRat values of the test materials in the present study were within this range. Within the lutein concentration range of this study, 64.9–150 mg/kg, the reproducibility of the analysis method was acceptable. Table 4. Interlaboratory study results for the determination of lutein by proposed method Material . No. of labsa (outlier) . Average, mg/kg . sr . 2.8 × sr . sR . 2.8 × sR . RSDr, % . RSDR, % . Predicted RSDRb, % . HorRatc . A 10 (2) 64.9 3.3 9.2 3.6 10 5.1 5.6 8.5 0.66 B 11 (0) 71.8 3.6 10 5.4 15 5.0 7.5 8.4 0.89 C 12 (0) 88.8 6.7 19 12 33 7.5 13 8.1 1.6 D 12 (0) 119 4.8 13 7.5 21 4.1 6.4 7.8 0.82 E 12 (0) 150 5.1 14 6.8 19 3.4 4.6 7.5 0.61 Material . No. of labsa (outlier) . Average, mg/kg . sr . 2.8 × sr . sR . 2.8 × sR . RSDr, % . RSDR, % . Predicted RSDRb, % . HorRatc . A 10 (2) 64.9 3.3 9.2 3.6 10 5.1 5.6 8.5 0.66 B 11 (0) 71.8 3.6 10 5.4 15 5.0 7.5 8.4 0.89 C 12 (0) 88.8 6.7 19 12 33 7.5 13 8.1 1.6 D 12 (0) 119 4.8 13 7.5 21 4.1 6.4 7.8 0.82 E 12 (0) 150 5.1 14 6.8 19 3.4 4.6 7.5 0.61 a Number of laboratories retained after outliers removed. b Predicted RSDR = 2 × (Average × 10−6)−0.1505. c HorRat = RSDR/Predicted RSDR. Open in new tab Table 4. Interlaboratory study results for the determination of lutein by proposed method Material . No. of labsa (outlier) . Average, mg/kg . sr . 2.8 × sr . sR . 2.8 × sR . RSDr, % . RSDR, % . Predicted RSDRb, % . HorRatc . A 10 (2) 64.9 3.3 9.2 3.6 10 5.1 5.6 8.5 0.66 B 11 (0) 71.8 3.6 10 5.4 15 5.0 7.5 8.4 0.89 C 12 (0) 88.8 6.7 19 12 33 7.5 13 8.1 1.6 D 12 (0) 119 4.8 13 7.5 21 4.1 6.4 7.8 0.82 E 12 (0) 150 5.1 14 6.8 19 3.4 4.6 7.5 0.61 Material . No. of labsa (outlier) . Average, mg/kg . sr . 2.8 × sr . sR . 2.8 × sR . RSDr, % . RSDR, % . Predicted RSDRb, % . HorRatc . A 10 (2) 64.9 3.3 9.2 3.6 10 5.1 5.6 8.5 0.66 B 11 (0) 71.8 3.6 10 5.4 15 5.0 7.5 8.4 0.89 C 12 (0) 88.8 6.7 19 12 33 7.5 13 8.1 1.6 D 12 (0) 119 4.8 13 7.5 21 4.1 6.4 7.8 0.82 E 12 (0) 150 5.1 14 6.8 19 3.4 4.6 7.5 0.61 a Number of laboratories retained after outliers removed. b Predicted RSDR = 2 × (Average × 10−6)−0.1505. c HorRat = RSDR/Predicted RSDR. Open in new tab Comments from Interlaboratory Participants In all cases, participants regarded the method description as adequate. Few laboratories commented that it was difficult to remove a portion of the test sample as it became stuck to solid objects. It was also noted by several laboratories that it was difficult to distinguish the interface between the organic solvent layer and the aqueous layer in the amber centrifuge tubes. However, such operational difficulty did not affect the results; indeed, no outliers were detected in the data from these laboratories. Conclusions An interlaboratory study of a lutein measurement method for spinach was implemented. Five concentration test materials were used and 12 laboratories participated in the study. After abnormal values and outliers were removed, the data were used for an evaluation of precision. Within the lutein concentration range of 64.9–150 mg/kg, the reproducibility was 4.6–13%. The HorRat values were 0.61–1.6. The results indicate that the measurement method has satisfactory precision and accuracy. The method was published as an official JAS method in 2019. Acknowledgments The planning and validation of the results for this interlaboratory study were implemented in Validation Exploratory Committee in Food and Agricultural Materials Inspection Center (FAMIC). We sincerely thank all members of the Committee and the following collaborators for their participation in the interlaboratory study. Kouji Iwamaru, FAMIC Nagoya Regional Center, Aichi, Japan; Masataka Kumagai, FAMIC Kobe Regional Center, Hyogo, Japan; Hayato Maeda, Hirosaki University, Aomori, Japan; Mayu Miyamoto, FAMIC Headquarters, Saitama, Japan; Kenji Mizuta, FAMIC Sendai Regional Center, Miyagi, Japan; Mayumi Morioka, FAMIC Fukuoka Regional Center, Fukuoka, Japan; Tatsuya Sugawara, Yuuki Manabe, Kyoto University, Kyoto, Japan; Fumihito Takahashi, Japan Food Research Laboratories, Tokyo, Japan; Hiroshi Ueda, Institute of Fruit Tree and Tea Science of National Agriculture and Food Research Organization, Mie, Japan; Eriko Watanabe, FAMIC Yokohama Office, Kanagawa, Japan; Ryosuke Yamoto, FAMIC Sapporo Regional Center, Hokkaido, Japan; Lina Yonekura, Kagawa University, Kagawa, Japan; We also would like to thank Enago (www.enago.jp) for the English language review. Conflict of Interest The authors declare no conflicts of interest. References 1 Bernstein P.S. ( 2002 ) Pure Appl. Chem . 74 , 1419 – 1425 . doi:10.1351/pac200274081419 Crossref Search ADS 2 Loskutova E. , Nolan J., Howard A., Beatty S. ( 2013 ) Nutrients 5 , 1962 – 1969 . doi:10.3390/nu5061962 Crossref Search ADS PubMed 3 Kopsell D.A. , Lefsrud M.G., Kopsell D.E., Wenzel A.J., Gerweck C., Curran-Celentano J. ( 2006 ) J. Agric. Food Chem. 54 , 7998 – 8005 . doi:10.1021/jf0614802 Crossref Search ADS PubMed 4 Graydon R. , Hogg R.E., Chakravarthy U., Young I.S., Woodside J.V. ( 2012 ) Br. J. Nutr. 108 , 334 – 342 . doi:10.1017/S0007114511005599 Crossref Search ADS PubMed 5 Tanito M. , Obana A., Gohto Y., Okazaki S., Gellermann W., Ohira A. ( 2012 ) Jpn. J. Ophthalmol. 56 , 488 − 496 .doi:10.1007/s10384-012-0157-0 6 Landrum J. , Bone R., Mendez V., Valenciaga A., Babino D. ( 2012 ) Acta Biochim. Pol. 59 , 167 – 169 . doi:10.18388/abp.2012_2198 Crossref Search ADS PubMed 7 Berrow E.J. , Bartlett H.E., Eperjesi F. ( 2016 ) Doc. Ophthalmol. 132 , 123 – 135 . doi:10.1007/s10633-016-9532-3 Crossref Search ADS PubMed 8 Johnson E.J. , Chung H.Y., Caldarella S.M., Snodderly D.M. ( 2008 ) Am. J. Clin. Nutr . 87 , 1521 – 1529 . doi:10.1093/ajcn/87.5.1521 Crossref Search ADS PubMed 9 Mangels A.R. , Holden J.M., Beecher G.R., Forman M.R., Lanza E. ( 1993 ) J. Am. Diet Assoc . 93 , 284 – 296 . doi:10.1016/0002-8223(93)91553-3 Crossref Search ADS PubMed 10 Nagai N. , Izumi-Nagai K., Suzuki M., Shinoda H., Koto T., Uchida A., Mochimaru H., Tomita Y., Miyake S., Kobayashi S., Sasaki M., Tsubota K., Ozawa Y. ( 2015 ) Retina 35 , 820 – 826 . doi:10.1097/IAE.0000000000000382 Crossref Search ADS PubMed 11 Lester G.E. , Makus D.J., Hodges D.M., Jifon L.J. ( 2013 ) J. Agric. Food Chem. 61 , 7019 – 7027 . doi:10.1021/jf401461z Crossref Search ADS PubMed 12 Mercadante Z.A. , Rodriguez-Amaya D.B. ( 1991 ) J. Agric. Food Chem. 39 , 1094 – 1097 . doi:10.1021/jf00006a018 Crossref Search ADS 13 Reif C. , Arrigoni E., Neuweiler R., Baumgartner D., Nystrom L., Hurrell R.F. ( 2012 ) J. Agric. Food Chem. 60 , 5819 – 5824 . doi:10.1021/jf301114p Crossref Search ADS PubMed 14 Sander C. , Sharpless K.E., Craft N.E., Wise S.A. ( 1994 ) Anal. Chem. 66 , 1667 – 1674 . doi:10.1021/ac00082a012 Crossref Search ADS PubMed 15 Schierle J. , Pietsch B., Ceresa A., Fizet C., Waysek E.H. ( 2004 ) J. AOAC Int . 87 , 1070 – 1082 Crossref Search ADS PubMed 16 Horwitz W. ( 1995 ) Pure Appl. Chem . 67 , 331 – 343 . doi:10.1351/pac199567020331 Crossref Search ADS 17 Britton G. , Liaaen-Jensen S., Pfander H. ( 2004 ) Carotenoids Handbook , Birkhauser Verlag , Basel , pp 133 Google Scholar Crossref Search ADS Google Scholar Google Preview WorldCat COPAC 18 United States Pharmacopeia ( 2019 ) 42nd Ed. The United States Pharmacopeial Convention , Rockville, MD , pp 6788 19 Kaiser P. , Surmann P., Fuhrmann H. ( 2009 ) J. Sep. Sci. 32 , 34 – 43 . doi:10.1002/jssc.200800408 Crossref Search ADS PubMed 20 Ito H. ( 2014 ) JARQ . 48 , 111 – 120 . doi:10.6090/jarq.48.111 Crossref Search ADS 21 Larsen E. , Christensen L.P. ( 2005 ) J. Agric. Food Chem. 53 , 6598 – 6602 . doi:10.1021/jf050622+ Crossref Search ADS PubMed 22 Aman R. , Biehl J., Carle R., Conrad J., Beifuss U., Schieber A. ( 2005 ) Food Chem . 92 , 753 – 763 . doi:10.1016/j.foodchem.2004.10.031 Crossref Search ADS 23 Chitchumroonchokchai C. , Schwartz J.S., Failla L.M. ( 2004 ) J. Nutr . 134 , 2280 – 2286 . doi:10.1093/jn/134.9.2280 Crossref Search ADS PubMed 24 Guidelines for Establishing Numeric Values for Method Criteria and/or Assessing Methods for Compliance Thereof, Codex Procedural Manual 26th Ed., Codex Alimentarius Commission. http://www.fao.org/3/i8608en/I8608EN.pdf (accessed 2019) 25 Thompson M. , Ellison S.L.R., Wood R. ( 2006 ) Pure Appl. Chem . 78 , 145 – 196 . doi:10.1351/pac200678010145 Crossref Search ADS 26 ISO Guide 35: 2006 ( 2006 ) Reference Materials-General and Statistical Principles for Certification , International Organization for Standardization , Geneva Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 27 Horwitz W. , Kamps L.R., Boyer K.W. ( 1980 ) J. Assoc. Off. Anal. Chem. 63 , 1344 – 1354 PubMed 28 Official Methods of Analysis ( 2019 ) 21st Ed., AOAC INTERNATIONAL , Rockville, MD , Appendix D Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC © AOAC INTERNATIONAL 2020. All rights reserved. For permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
TI - Validation of a Method for Quantification of Lutein in Spinach Using High-Performance Liquid Chromatography: Interlaboratory Study
JF - Journal of AOAC International
DO - 10.1093/jaoacint/qsaa014
DA - 2020-07-01
UR - https://www.deepdyve.com/lp/oxford-university-press/validation-of-a-method-for-quantification-of-lutein-in-spinach-using-NpEI6YXf0L
SP - 1073
EP - 1080
VL - 103
IS - 4
DP - DeepDyve
ER -