2017 Volume 23 Issue 6 Pages 855-862
This paper described a modified HPLC method for the determination of eight acylated anthocyanins in purple-fleshed sweet potato tuber samples, including lyophilized powder and concentrated juice. Recoveries determined for two materials spiked with anthocyanin levels of 71 µg g−1 to 8.8 mg g−1 ranges 91.8–105.0%. For repeatability, the relative standard deviation and intermediate precision ranged from 0.8% to 5.0% and 1.9% to 8.2%, respectively. HorRat values ranged from 0.3 to 1.0, except for YGM-1a and YGM-1b in two samples, which were well within the limits of performance acceptability. The proposed method showed good intralaboratory reproducibility in the range from 0.10 ± 0.0027 (mean ± expanded measurement uncertainty, k = 2) to 0.51 ± 0.016, 0.10 ± 0.0084 to 0.21 ± 0.011, 0.40 ± 0.017 to 4.5 ± 0.12, 0.21 ± 0.0066 to 0.96 ± 0.021, 0.23 ± 0.014 to 0.57 ± 0.026, 0.15 ± 0.0087 to 1.3 ± 0.040, 0.94 ± 0.021 to 6.5 ± 0.19, and 0.63 ± 0.016 to 2.4 ± 0.061 mg g−1 for YGM-1a, YGM-1b, YGM-2, YGM-3, YGM-4b, YGM-5a, YGM-5b, and YGM-6, respectively.
Anthocyanins are naturally occurring pigments found in vegetables, fruits, and flowers, and are promising functional food materials owing to their beneficial effects on human health (Tsuda, 2012; Cerletti et al., 2016; Wallace et al., 2016). “Ayamurasaki” is the first sweet potato cultivar to have a purple-flesh color for processing purposes, and was developed by cross-breeding a line containing anthocyanins with a cultivar lacking β-amylase activity at NARO Kyushu Okinawa Agricultural Research Center (NARO/KARC) in Japan (Yoshinaga, 1995). The major anthocyanins in “Ayamurasaki” tubers are mono- and di-acylated forms of cyanidin and peonidin, as shown in Fig. 1 (Goda et al., 1997; Terahara et al., 1999). Among these anthocyanins, two (YGM-2 and YGM-5b) are mono-acylated with caffeic acid, while others are di-acylated with caffeic acid alone (YGM-1b and YGM-4b), caffeic acid and p-hydroxybenzoic acid (YGM-1a and YGM-5a), or caffeic acid and ferulic acid (YGM-3 and YGM-6). Other purple-fleshed sweet potato cultivars, “Murasakimasari” and “Akemurasaki”, of which the parent is “Ayamurasaki”, also contain these eight anthocyanins. “Purple sweet lord”, available on the market for table use, also contains these anthocyanins. Despite the complex chemical structure of anthocyanins, studies have confirmed that when anthocyanin-rich extract is administered to rats (Suda et al., 2002; Harada et al., 2004; Kano et al., 2005) or a beverage prepared from purple-fleshed sweet potato is consumed by humans (Harada et al., 2004; Kano et al., 2005; Oki et al., 2006), two components of anthocyanins, YGM-2 and YGM-5b, are rapidly absorbed into the body (detectable in the blood) and rapidly excreted in the urine. Moreover, beneficial effects on hypertension and hepatitis have been observed in both animal models (Suda et al., 1997; Kobayashi et al., 2005) and clinical trials (Suda et al., 1998; Suda et al., 2007; Oki et al., 2017). To date, acylated anthocyanins have been considered the most likely candidates for this efficacy.
Chemical structure of acylated anthocyanins from purple-fleshed sweet potato tubers, and their molecular weights and relative response factors.
Abbreviations: Me, methyl; Cy, cyanidin; Pn, peonidin; Caf, (E)-caffeic acid; Sop, sophoroside; PHB, p-hydroxybenzoic acid; Fer, (E)-ferulic acid; Glc, glucopyranoside; RRF, relative response factor.
RRFs were cited in our previous paper (Terahara et al., 2007).
Previously, we reported the determination method of eight major species of acylated anthocyanins in purple-fleshed sweet potato using high-performance liquid chromatography (Terahara et al., 2007). One characteristic of this method was that the relative response factor (RRF), which is determined by the analysis of standards and used to compute the true concentrations of compounds in samples, was used for acylated anthocyanins determination. RRF is a useful tool for reducing the cost of standard reagents, reducing experiment time, and simplifying experimental operations. For example, RRF has been used for the HPLC determination of catechins (ISO 14502-2:2005, 2005) and isoflavones (Collison, 2008) in tea and soybean products, respectively. Both methods were multi-laboratories validated, and used for quality control of tea and soybean products, displaying their contents on product labels and so on. Moreover, it was revealed that using RRF is effective for satisfactory precision in the analysis of multi-laboratory studies. However, an analytical method for quantification of acylated anthocyanins in purple-fleshed sweet potato has not yet been standardized. Therefore, analytical values cannot be compared and quality management has not been carried out. The objectives of this study were to optimize the extraction and HPLC conditions for quantification of eight species of acylated anthocyanins in sweet potato tubers and concentrated juice, and to determine intra-laboratory precision.
Reagents HPLC-grade acetonitrile and reagent-grade formic acid were used as the mobile phase. Deionized water with a specific resistivity of 18.2 Mω cm , monitored through a display on the apparatus (Milli-Q Advantage A10, Millipore, Billerica, MA, USA), was used. All other chemicals were of reagent grade. Peonidin (3-(6,6′-caffeylferulylsophoroside)-5-glucoside (YGM-6)) was isolated as its trifluoroacetic acid salt from purple-fleshed sweet potato cultivar, “Ayamurasaki”, tubers by HP-20 column chromatography, LH-20 column chromatography, preparative paper chromatography, preparative TLC, and preparative HPLC, as described previously (Goda et al., 1997; Terahara et al., 1999), and used as a standard. Based on HPLC analysis, the purity of YGM-6 was found to be more than 98.3%.
Plant materials Purple-fleshed sweet potato cultivars, “Akemurasaki”, “Murasakimasari”, and “Purple sweet lord”, used for tuber sampling were grown at the Miyakonojyo Branch of NARO Kyushu Okinawa Agricultural Research Center (131°1′E, 31°45′N). Tubers were harvested from a field in September 2011. Both end portions of the tubers were removed after peeling, then lyophilized and milled to a powder with a sterile Oster Blender (Osaka Chemical Co., Ltd., Osaka, Japan) and stored at −20°C until use.
Concentrated Juice Concentrated juice was prepared from purple-fleshed sweet potato cultivars, “Akemurasaki”, “Ayamurasaki”, and “Murasakimasari”, according to a previous report and outlined as follow (Suda et al., 2002). Purple-fleshed sweet potato tubers were washed with tap water, peeled, and boiled at 90°C for 10 min. The heated flesh was ground in the same weight of water using a Super Masscolloider (Masuko Sangyo Co., Ltd., Saitama, Japan) and centrifuged. The resultant sweet potato juice was treated with a composed-enzyme solution containing amylase, cellulase, hemicellulase, and pectinase at 50°C for 60 min and centrifuged. The supernatant was concentrated using a continuous evaporator and stored at −20°C until use.
Extraction of acylated anthocyanins Sweet potato tuber powder or concentrated juice (500 mg) was weighed into a 50 mL glass centrifuge tube, and 9 mL of extraction solvent (methanol/water/trifluoroacetic acid = 40:60:0.5, v/v) was added. After vigorously shaking the sample using a vortex mixer, the tube was capped, immersed in a water bath at 37°C, and sonicated for 5 min at an oscillating frequency of 40 kHz. The tubes were then incubated for 10 min in a water bath at 37°C and centrifuged at 1,870 ×g for 10 min. The supernatant was transferred into a 25 mL volumetric flask. This extraction procedure was repeated twice with 8 mL of the extraction solvent, and the final volume of the extract was made up to 25 mL
Optimization of extraction conditions Five kinds of materials (air-dried powder, concentrated juice, drum-dried powder, lyophilized powder, and paste) prepared from purple-fleshed sweet potato cultivar “Ayamurasaki” tubers were used. Two materials, lyophilized powder and concentrated juice, were prepared by the methods described above. Other materials were commercially available. These materials were extracted with various concentrations of acidified methanol solution (methanol concentrations of 0, 20, 40, 60, 85, or 100%, v/v). The extraction was repeated three times as described above, and the extracted solution was subjected to HPLC analysis. The methanol concentration in the extraction solvent was optimized by comparing the total content of eight species of acylated anthocyanins. Triplicate measurements were performed for each condition, and the mean and standard deviation were calculated.
Chromatography The extract was filtered through a 0.45 µm polyvinylidene difluoride membrane, and 10 µL of the filtrate was injected into the HPLC system for analysis. The HPLC system consisted of a Model PU-2080 plus pump, a Model LG-2080-04 low-pressure gradient unit, a Model DG-2080-54 degasser, a Model AS-2057 plus auto injector, a Model CO-2060 plus column oven, and a Model UV-2070 plus detector with a conventional flow cell of 17 µL intrinsic volume (JASCO, Tokyo, Japan). Standard and sample solutions were stored at −20°C before HPLC analysis. A Cadenza CD-C18 column (250 mm × 4.6 mm i.d., 3 µm, Imtakt, Kyoto, Japan) was used and warmed to 30°C in the column oven. Mobile phase A consisted of water and formic acid in the ratio 100:0.6, v/v. Mobile phase B consisted of water, acetonitrile and formic acid in the ratio 50:50:0.6, v/v. Elution was conducted with a linear gradient as follows: raise from 20% to 50% B from 0 to 50 min, with a flow rate of 0.6 mL/min. Mobile phase A was used to wash the autosampler syringe. Target anthocyanins were monitored at a wavelength of 520 nm. The system was controlled by a ChromNAV (version 1.18.03) workstation (JASCO, Tokyo, Japan). If necessary, the extract was diluted with the same extraction solvent to adjust the target analyte concentration to within the calibration range.
Calculation of Individual Acylated Anthocyanin Amounts The concentration of individual acylated anthocyanins in the sample solution was calculated using the following equation (Eq.1):
 |
where Conc is the concentration of individual acylated anthocyanin expressed as dry base for lyophilized powder (mg g−1) or wet base for concentrated juice (mg g−1); Asample is the peak area of the acylated anthocyanin; Aintercept is the peak area of the calibration curve y-axis intercept; Vsample is the amount of sample extract (L); d is the dilution factor of the sample solution; S is the calibration curve slope; MW is the molecular weight of individual acylated anthocyanins; and Msample is the amount of sample (g). Relative response factors reported in our previous paper (Terahara et al., 2007) were used for determining the eight species of acylated anthocyanins in purple-fleshed sweet potato (Fig. 1).
Preparation of Stock Standard Solution and Working Standard Solutions To prepare the YGM-6 stock standard solution (nominal concentration of 50 µM), YMG-6 trifluoroacetic acid salt (1.24 mg) was dissolved in 20 mL of an acidified acetonitrile aqueous solution (acetonitrile/water/formic acid = 10:90:0.4, v/v). The net concentration of YGM-6 was determined using its molar extinction coefficient of 24,800 L mol−1 cm−1 at wavelength of 532 nm. The stock standard solution was stored at −20°C when not in use, and found to be stable for at least a month. Working standard solutions were freshly prepared by diluting the stock standard solution with the acidified acetonitrile aqueous solution described above. The concentrations of YGM-6 in the working standard solutions were 50, 37.5, 25, 12.5, 1.25, and 0.625 µM. Each standard solution was subjected to HPLC analysis under conditions described above in duplicate, and a six-point calibration graph was obtained for each sample by plotting peak area (y-axis, in µV sec) versus concentration (x-axis, in µM).
Stability Study of Stock Standard The stability of YGM-6 in a standard solution was evaluated. The standard solution was aliquoted into capped polypropylene tubes (safe-lock tubes, Eppendorf, Tokyo, Japan) and stored at −20°C. Aliquots of the standard solution were subjected to HPLC on 1, 3, 8, 14, 21, 29, and 35 days after sample preparation.
Recovery Test of the Proposed Method Lyophilized powder or concentrated juice prepared from sweet potato cultivar “Koganesengan” tuber was used as a matrix blank. This cultivar, having white flesh, contains no anthocyanins. Sunred YM (San-Ei Gen F.F.I., Inc., Osaka, Japan), which is a natural food colorant containing high concentrations of acylated anthocyanins in purple-fleshed sweet potato cultivar “Ayamurasaki”, was used as a material fortified to matrix. The fortifier was resolved in the extraction solvent and spiked in the lyophilized powder or concentrated juice (500 mg each) at three different levels of anthocyanin content. The total contents of eight acylated anthocyanins in fortifier added to the lyophilized powder were 29.1, 14.6, and 7.3 (mg g−1) at high-, middle-, and low-levels, respectively. The contents added to the concentrated juice were 11.5, 5.7, and 2.9 (mg g−1) at high-, middle- and low-levels, respectively. Fortified and unfortified samples were extracted as described above, the concentrations of acylated anthocyanins were determined, and the percentage recovery was calculated according to AOAC International guidelines.
Precision Test of the Proposed Method Lyophilized powders and concentrated juices of purple-fleshed sweet potatoes were used as test samples to evaluate the precision of the proposed method. The lyophilized powders were prepared from three cultivars, “Akemurasaki”, “Murasakimasari”, and “Purple sweet lord”. The concentrated juices were prepared from three cultivars, “Akemurasaki”, “Ayamurasaki”, and “Murasakimasari”. These test samples were extracted as described above, and concentrations of acylated anthocyanins were determined. For lyophilized powder, two operators participated, each repeating duplicate measurements on three different days. For the concentrated juice, one operator repeated duplicate measurements on five different days. Repeatability relative standard deviation (RSDr) and intermediate precision (relative standard deviation; RSDint) were obtained from this experiment schedule according to AOAC International guidelines. In this study, we used the Horwitz ratio (HorRat) as an index of method performance with respect to precision. HorRat is the ratio of observed RSDint to corresponding predicted relative standard deviation of reproducibility (PRSDR). PRSDR is calculated from the Horwitz equation (Eq.2):
 |
where C is the commonly measured concentration of the analyte in the sample, expressed as a mass fraction.
Statistical Analysis The repeatability and intermediate standard deviations for the acylated anthocyanin contents were calculated by one-way analysis of variance (ANOVA), and the stability of YGM-6 in the standard stock solution was evaluated with a linear regression model using Microsoft Excel 2013 (Microsoft Office 2013, Microsoft Co., WA, USA). Variations in anthocyanin levels extracted from various materials were analyzed by one-way ANOVA for six groups (methanol concentration of 0, 20, 40, 60, 85, and 100%, v/v) using SAS Add-In for Microsoft Office 5.1 (SAS Institute Inc., Cary, NC, USA). Within the groups, Tukey multiple-comparison correction was performed to compare the mean extracted anthocyanin levels because ANOVA testing revealed a significant effect of methanol concentration (p < 0.05) in all five materials.
Optimization of Extraction Efficacy Anthocyanins are typically extracted using an acidified aqueous solution containing alcohol. Our previous study demonstrated that using 40% (v/v) methanol aqueous solution containing 0.5% (v/v) trifluoroacetic acid was superior at extracting monomeric anthocyanins from the seed coat of black soybean and bran of black rice (Sawai et al., 2012). In this study, the extraction efficacy for anthocyanins in five kinds of matrixes prepared from purple-fleshed sweet potato was examined using an acidified aqueous solution containing various concentrations of methanol. Among the extraction solvents tested, solvent with a methanol concentration of 100% extracted by far the highest amount of total anthocyanins from lyophilized powder (100%, 1.58 ± 0.03 mg g−1). However, total anthocyanins were extracted from concentrated juice efficiently by solvents with methanol concentrations ranging from 20% to 85% (data not shown). Moreover, anthocyanins in paste, drum-dried powder, and air-dried powder were extracted efficiently by solvents with methanol concentrations ranging from 20 to 100%, 20 to 60%, and 20 to 40%, respectively (data not shown). Thus, the optimum methanol concentration for anthocyanin extraction varied according to the material. This was probably due to conditions in the matrix containing analytes being affected by different manufacturing processes, despite all materials being prepared from the same ingredient, namely purple-fleshed sweet potato. Furthermore, it was found that using lower methanol concentrations in the extraction solvent, namely 0 to 20%, led to a muddy extract due to the presence of starch extracted together with the anthocyanins. Conversely, an un-mixed lump of flour was observed in a few matrices during the extraction procedure when higher methanol concentrations, namely 85 to 100%, were used. Taking these results into consideration, we decided on the composition of the solvent (methanol/water/trifluoroacetic acid = 40:60:0.5, v/v) for extracting anthocyanins from various materials made from purple-fleshed sweet potato, although this solvent extracted 7% less total anthocyanins than solvent with 100% methanol concentration from lyophilized powder only (40%, 1.47 ± 0.02 mg g−1).
System Selectivity Using HPLC conditions optimized in our previous study, the two peaks corresponding to mono-acylated anthocyanins, such as YGM-2 and YGM-5b, were broad. In order to improve peak signal to noise ratio and lower the detection limit, we tried to modify these HPLC conditions. Changing the formic acid concentration from 0.4 to 0.6% in both mobile phases resulted in sharper peaks than previously. The shape of two peaks, the peaks of 1 and 4, were still board; however, higher concentration of formic acid led to the insufficient peak separation of target anthocyanins. Both peaks were mono-acylated anthocyanins, while other six peaks were di-acylated anthocyanins. The reason was unclear; however, the shape of mono-acylated anthocyanins seemed to be broader than that of di-acylated anthocyanins in the proposed HPLC conditions. Under these conditions, acylated anthocyanins in extracts of purple-fleshed sweet potato eluted in the order YGM-2, YGM-1b, YGM-1a, YGM-5b, YGM-3, YGM-4b, YGM-5a, and YGM-6 with retention times of 30.8, 32.5, 33.1, 34.5, 35.9, 37.4, 38.3, and 40.9 min, respectively (Fig. 2). An adequate resolution of more than 1.5 between two adjacent peaks was observed for all target analytes.
Typical reversed-phase HPLC chromatogram of purple-fleshed sweet potato cultivar tuber extract.
Peak 1, YGM-2 (30.8 min); peak 2, YGM-1b (32.5 min); peak 3, YMG-1a (33.1 min); peak 4, YGM-5b (34.5 min); peak 5, YGM-3 (35.9 min); peak 6, YGM-4b (37.4 min); peak 7, YGM-5a (38.3 min); and peak 8, YGM-6 (40.9 min).
Linearity of Calibration Curve Standard YGM-6 solutions of six different concentrations were randomly injected into the HPLC system. The YGM-6 peak areas were plotted, and a linear regression model was used to determine the slope and y-axis intercept. HPLC analyses for constructing the calibration graph were performed eight times on separate days, with the resulting correlation coefficients always greater than 0.9999 in the nominal concentration range of 0.625 to 50 µM (data not shown). Changes in the slope were not observed in any replicates, and the coefficient of variance was 0.56%. Furthermore, the linear regression analysis of eight calibration curves for YGM-6 demonstrated that a 95% confidence interval for the y-axis intercept included zero. Hence, it was concluded that the y-axis intercept did not differ from zero at the 0.05 level of significance. However, the origin of the coordinates was not used to construct the calibration graph plot in this study.
Stability of Standard Solution Stability of the standard solution was assessed after 1, 3, 4, 8, 14, 21, 29, and 35 days of storage at −20°C. Analysis of the peak area at a nominal concentration of 25 µM using a linear regression model (x-axis, day; y-axis, peak area) revealed that the slope p-value was 0.229 and, thus, the slope was indistinguishable from zero (95% confidence interval of the slope: −100 to 377). Thus, it was concluded that there was no significant change in the concentration of the stock standard solution during a storage period of at least 35 days at −20°C.
Accuracy of the Proposed Method A recovery study based on spiking a food colorant from purple-fleshed sweet potato cultivar “Ayamurasaki” into negative control materials (lyophilized powder and sweet potato tuber concentrate) at 200, 100, and 50% of the expected level was performed in triplicate. The spike recovery results are shown in Table 1. The average recoveries of acylated anthocyanins ranged from 92.8 to 105.0%, 91.8 to 104.2%, and 94.2 to 104.7% for low-, middle-, and high-level additions, respectively. There were differences of recoveries between lyophilized powder and concentrated juice, and almost all recovery data from concentrated juice were over 100%. Lower recovery observed in lyophilized powder was accounted for by the absorption of anthocyanins into matrixes. On the other hand, the reason why the recovery of over 100% was frequently observed in concentrated juice was unclear in this study. The AOAC International guidelines for single-laboratory validation describes the expected recovery as a function of analyte concentration ranges from 90 to 108% and from 92 to 105% for analyte concentrations of 0.1 and 1%, respectively (AOAC Int., Appendix K, 2012). These results clearly demonstrate the satisfactory accuracy of our method for quantifying the contents of acylated anthocyanins in freeze-dried powder and concentrated juice of sweet potato tubers.
| Recover (%) | ||||||
|---|---|---|---|---|---|---|
| Lyophilized power | Concentrated juice | |||||
| High-level | Middle-level | Low-level | High-level | Middle-level | Low-level | |
| YGM-1a | 98.5 | 92.7 | 94.4 | 101.6 | 104.2 | 101.1 |
| YGM-1b | 99.8 | 91.8 | 92.9 | 104.7 | 98.0 | 105.0 |
| YGM-2 | 94.2 | 95.6 | 97.6 | 100.7 | 101.6 | 101.8 |
| YGM-3 | 99.1 | 93.0 | 94.9 | 103.3 | 100.8 | 100.3 |
| YGM-4b | 101.2 | 92.9 | 94.3 | 101.8 | 100.3 | 100.2 |
| YGM-5a | 100.3 | 93.9 | 92.8 | 101.5 | 100.8 | 100.4 |
| YGM-5b | 99.4 | 93.4 | 94.2 | 104.0 | 100.2 | 99.5 |
| YGM-6 | 101.0 | 93.0 | 95.5 | 101.4 | 102.6 | 101.7 |
For lyophilized powder, the contents of spiked anthocyanin at middle level were follows: YGM-1a, 0.37 mg g−1; YGM-1b, 0.43 mg g−1; YGM-2, 1.0 mg g−1; YGM-3, 0.78 mg g−1; YGM-4b, 2.0 mg g−1; YGM-5a, 2.0 mg g−1; YGM-5b, 4.4 mg g−1; YGM-6, 3.6 mg g−1. For concentrated juice, the contents of spiked anthocyanin at middle level were follows: YGM-1a, 0.14 mg g−1; YGM-1b, 0.17 mg g−1; YGM-2, 0.45 mg g−1; YGM-3, 0.30 mg g−1; YGM-4b, 0.77 mg g−1; YGM-5a, 0.76 mg g−1; YGM-5b, 1.7 mg g−1; YGM-6, 1.4 mg g−1. At high level, the content was twice higher than at middle level. At low level, the content was one half lower than at middle level.
Precision of the Proposed Method Precision was evaluated by performing sample extraction and HPLC analysis on lyophilized powders and concentrated juices of purple-fleshed sweet potato tubers. Table 2 summarize the precision results obtained from repeated measurements of three kinds of powder samples and three kinds of juice samples, respectively. A statistical treatment revealed that RSDr and RSDint for the determination of acylated anthocyanin contents in purple-fleshed sweet potato tuber samples were 0.8 to 5.0% and 1.9 to 8.2%, respectively. In order to indicate the acceptability of the precision for this method, HorRat can be used as normalized performance parameter (Horwitz and Albert, 2006). As shown in Table 2, the HorRat values of target analytes ranged from 0.3 to 1.0. According to the single-laboratory validation guidelines of AOAC International (AOAC Int., Appendix F, 2012), acceptable HorRat values are 0.3 to 1.3, indicating satisfactory precision for the proposed method.
| Sample | Cultivar | YGM-1a | YGM-1b | YGM-2 | YGM-3 | YGM-4b | YGM-5a | YGM-5b | YGM-6 | ||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Lyophilized powder | Akemurasaki | Mean | (mg g−1) | 0.51 | 0.21 | 4.5 | 0.96 | 0.51 | 1.3 | 6.5 | 1.9 |
| RSDr | (%) | 2.9 | 4.1 | 1.3 | 0.82 | 1.6 | 1.3 | 1.3 | 1.8 | ||
| RSDint | (%) | 3.1 | 5.2 | 2.6 | 2.2 | 4.4 | 3.2 | 2.9 | 2.5 | ||
| HorRat | 0.50 | 0.72 | 0.58 | 0.39 | 0.71 | 0.58 | 0.69 | 0.49 | |||
| Murasakimasari | Mean | (mg g−1) | ND | ND | 1.2 | 0.36 | 0.57 | 0.84 | 5.3 | 2.4 | |
| RSDr | (%) | N/A | N/A | 1.3 | 0.82 | 1.6 | 1.3 | 1.3 | 1.8 | ||
| RSDint | (%) | N/A | N/A | 2.6 | 2.2 | 4.4 | 3.2 | 2.9 | 2.5 | ||
| HorRat | N/A | N/A | 0.54 | 0.59 | 0.81 | 0.52 | 0.57 | 0.42 | |||
| Purple sweet lord | Mean | (mg g−1) | ND | ND | 0.42 | 0.22 | 0.23 | 0.15 | 0.94 | 0.63 | |
| RSDr | (%) | N/A | N/A | 2.6 | 2.3 | 3.5 | 4.4 | 2.1 | 1.7 | ||
| RSDint | (%) | N/A | N/A | 2.9 | 3.3 | 6.1 | 5.6 | 2.2 | 2.5 | ||
| HorRat | N/A | N/A | 0.44 | 0.47 | 0.87 | 0.75 | 0.39 | 0.41 | |||
| Concentrated juice | Akemurasaki | Mean | (mg g−1) | 0.32 | 0.16 | 1.7 | 0.31 | 0.45 | 1.2 | 3.8 | 0.99 |
| RSDr | (%) | 2.5 | 5.0 | 0.8 | 1.5 | 1.5 | 1.3 | 1.0 | 1.2 | ||
| RSDint | (%) | 2.6 | 5.2 | 2.7 | 2.3 | 2.4 | 2.3 | 2.9 | 1.9 | ||
| HorRat | 0.39 | 0.69 | 0.51 | 0.33 | 0.38 | 0.42 | 0.64 | 0.33 | |||
| Ayamurasaki | Mean | (mg g−1) | 0.15 | 0.1 | 0.4 | 0.21 | 0.54 | 1.1 | 2.1 | 1.3 | |
| RSDr | (%) | 4.2 | 3.9 | 1.4 | 1.6 | 2.2 | 2.0 | 1.7 | 1.8 | ||
| RSDint | (%) | 4.4 | 8.2 | 4.2 | 3.2 | 2.9 | 3.2 | 3.4 | 2.9 | ||
| HorRat | 0.58 | 1.0 | 0.65 | 0.44 | 0.46 | 0.57 | 0.68 | 0.53 | |||
| Murasakimasari | Mean | (mg g−1) | 0.10 | 0.15 | 0.65 | 0.24 | 0.53 | 0.60 | 2.78 | 1.2 | |
| RSDr | (%) | 2.5 | 5.0 | 0.79 | 1.5 | 1.5 | 1.3 | 0.96 | 1.2 | ||
| RSDint | (%) | 2.62 | 5.16 | 2.67 | 2.25 | 2.39 | 2.32 | 2.95 | 1.88 | ||
| HorRat | 0.39 | 0.69 | 0.51 | 0.33 | 0.38 | 0.42 | 0.64 | 0.33 |
Abbreviations: N/A, not applicable; ND, not determined, when the peak area fell below the calibration range despite being observed in the chromatogram. Contents were as follows: YGM-1a < 33 µg g−1 and YGM-1b < 39 µg g−1.
An HPLC method was modified to determine eight acylated anthocyanins in purple-fleshed sweet potato tuber samples, including lyophilized powder and concentrated juice. Anthocyanins were extracted using aqueous methanol containing trifluoroacetic acid (methanol/water/trifluoroacetic acid = 40:60:0.5, v/v) and acylated anthocyanins in purple-fleshed purple sweet potato were determined by HPLC with visible absorption using a relative response factor to one of the anthocyanins (YGM-6). Recoveries determined for two materials spiked with anthocyanin levels of 71 µg g−1 to 8.8 mg g−1 ranges 91.8–105.0%. For repeatability, the relative standard deviation and intermediate precision ranged from 0.8% to 5.0% and 1.9% to 8.2%, respectively. HorRat values ranged from 0.3 to 1.0, except for YGM-1a and YGM-1b in two samples, which were well within the limits of performance acceptability. The proposed method showed good intralaboratory reproducibility in the range from 0.10 ± 0.0027 (mean ± expanded measurement uncertainty, k = 2) to 0.51 ± 0.016, 0.10 ± 0.0084 to 0.21 ± 0.011, 0.40 ± 0.017 to 4.5 ± 0.12, 0.21 ± 0.0066 to 0.96 ± 0.021, 0.23 ± 0.014 to 0.57 ± 0.026, 0.15 ± 0.0087 to 1.3 ± 0.040, 0.94 ± 0.021 to 6.5 ± 0.19, and 0.63 ± 0.016 to 2.4 ± 0.061 mg g−1 for YGM-1a, YGM-1b, YGM-2, YGM-3, YGM-4b, YGM-5a, YGM-5b, and YGM-6, respectively. This method is applicable to the analysis of eight species of acylated anthocyanins in purple-fleshed sweet potato tuber samples, including lyophilized powder and concentrated juice.
Acknowledgement We wish to thank Dr. Tetsufumi Sakai for providing samples of sweet potato tubers. Also, we would like to thank Editage (www.editage.jp) for English language editing.
