Anthocyanins not just have various benefits in food industry but also have been used as natural colourants in cosmetic, coating products and as potential natural photosensitizers in solar cell. Thus, the main purpose of this study was to obtain information on the maximum yield of anthocyanin that can be recovered from
Anthocyanins are identified as water soluble compounds having molecular structure based on a C6-C3-C6 skeleton. Anthocyanins are the most conspicuous subset under flavonoid group, due to the wide range of colours resulting from their synthesis [
These valuable bioactive compounds also have been reported to have biological effects on the physiological functions of cells such as reducing oxidative cell damage and increasing high density lipoprotein (HDL) cholesterol level [
Extraction of the bioactive compound is influenced by various process parameters such as solvent composition, pH, temperature, extraction time, and solid to liquid ratio [
To the best of our knowledge, there were no studies focusing on optimizing the parameters for anthocyanin extraction from
The chemical reagents (potassium chloride, sodium acetate) used in this study were of analytical grade and obtained from Sigma Aldrich. Methanol (99.8%) and undenatured ethanol (99.8%) from Systerm were used as extraction solvents. Ethyl acetate (Systerm) was used in the separation process. Formic acid, ammonium formate, and acetonitrile were of HPLC grade obtained from Merck. Sephadex LH-20 and Amberlite XAD-7 from GE Healthcare were used as chromatography resin. Deionized water used in this study was purified at 18.2 MΩ.cm−1 (Barnstead RO & Deionized Systems).
Fresh mature fruits of
In the extraction procedure, 0.5 g of the fruit powder was mixed with various volumes of methanol acidified with 0.5% acetic acid or ethanol acidified with 0.5% acetic acid to give a solid to liquid ratio ranging from 0.5 : 5 to 0.5 : 35 (g/mL). Conical flask was used and covered with aluminium foil to prevent the evaporation of solvent. The flask containing sample powder along with solvent was incubated in thermostatic water bath at various temperatures (30–60°C) and various time intervals (60–120 min). After extraction for a period of selected time, the mixture was centrifuged for 10 min. The supernatant was then filtered and evaluated for the total anthocyanin content. Experiments were performed in randomized order to minimize the variability caused by nuisance factors. All the experiments were performed in triplicate and the average value was used for the determination of total anthocyanin content from
RSM was used to optimize the methanolic extraction and ethanolic extraction of anthocyanins from
Independent variables and their levels used for Box-Behnken design.
Variables |
Factors | Levels | ||
---|---|---|---|---|
|
−1 | 0 | 1 | |
Extraction temperature, (°C) |
|
30 | 45 | 60 |
Extraction time, (min) |
|
60 | 90 | 120 |
Solid to liquid ratio, (g/mL) |
|
0.5 : 5 | 0.5 : 20 | 0.5 : 35 |
The experimental design and statistical analysis were performed using Design-Expert software (version 8.0.7.1, Stat Ease Inc., Minneapolis, MN, USA). The model adequacies were checked in terms of the values of
The total anthocyanin content was determined according to the spectrophotometric pH differential method [
The crude anthocyanin extract was concentrated by using a rotary evaporator (40°C). The aqueous concentrates were then placed in a separating funnel and an equal volume of ethyl acetate was added to remove lipids, chlorophylls, and other nonpolar compounds from the mixture. The partitioned aqueous extract was further purified by using ion exchange chromatography (IEC) and size exclusion chromatography (SEC) using Amberlite XAD-7 resin and Sephadex LH-20 as separation matrixces, respectively. Anthocyanin content in the fractions collected using size exclusion chromatography was determined by pH differential method and fractions containing the highest content of anthocyanin were chosen for identification analyses.
Analytical ultra performance liquid chromatography (Perkin Elmer FX15) was used in this study. The anthocyanin fractions were then analyzed using AB Sciex 3200Q Trap, equipped with Phenomenex Aqua C18 reverse-phase column (50 mm
In this study, BBD was used for response surface optimization with three process variables (extraction temperature, extraction time, and solid to liquid ratio) at three levels. Designs using BBD are usually very efficient in terms of the number of required runs and therefore are less expensive to run compared to central composite design (CCD). The design points fall within a safe operating limit, within the nominal high and low levels, as BBD does not contain any points at the vertices of the cubic region. This could be advantageous when the factor-level combinations are prohibitively expensive or impossible to test because of the physical process constraints [
Two different tests, namely, sequential model sum of squares and model summary statistic were performed to check the adequacy of the models generated from the obtained data and the results are given in Table
Adequecy of model tested.
Source | Sum of squares | df | Mean square |
|
Prob > |
Remarks |
---|---|---|---|---|---|---|
Sequential model sum of squares for acidified methanolic extraction | ||||||
Mean versus Total | 20893255.04 | 1 | 20893255.04 | |||
Linear versus Mean | 128877.78 | 3 | 42959.26 | 20.25 | <0.0001 | |
2FI versus Linear | 10748.42 | 3 | 3582.81 | 2.13 | 0.1600 | |
Quadratic versus 2FI | 12433.06 | 3 | 4144.35 | 6.60 | 0.0189 | Suggested |
Cubic versus Quadratic | 874.80 | 3 | 291.60 | 0.33 | 0.8045 | Aliased |
Residual | 3520.73 | 4 | 880.18 | |||
Total |
|
|
|
|||
| ||||||
Sequential model sum of squares for acidified ethanolic extraction | ||||||
Mean versus Total | 7196887.80 | 1 | 7196887.80 | |||
Linear versus Mean | 147605.80 | 3 | 49201.93 | 28.80 | <0.0001 | Suggested |
2FI versus Linear | 532.96 | 3 | 177.65 | 0.08 | 0.9683 | |
Quadratic versus 2FI | 13908.91 | 3 | 4636.30 | 4.18 | 0.0544 | Suggested |
Cubic versus Quadratic | 6069.16 | 3 | 2023.05 | 4.76 | 0.0829 | Aliased |
Residual | 1699.43 | 4 | 424.86 | |||
Total |
|
|
|
|||
| ||||||
Source | Std. Dev. |
|
Adjusted |
Predicted |
PRESS | Remarks |
| ||||||
Model summary statistics for acidified methanolic extraction | ||||||
Linear | 46.058 | 0.824 | 0.783 | 0.654 | 54157.757 | |
2FI | 41.023 | 0.892 | 0.828 | 0.542 | 71629.308 | |
Quadratic | 25.059 | 0.972 | 0.936 | 0.875 | 19498.004 | Suggested |
Cubic | 29.668 | 0.977 | 0.910 | + | Aliased | |
| ||||||
Model summary statistics for acidified ethanolic extraction | ||||||
Linear | 41.334 | 0.869 | 0.839 | 0.741 | 43972.027 | Suggested |
2FI | 46.559 | 0.872 | 0.796 | 0.406 | 100897.766 | |
Quadratic | 33.314 | 0.954 | 0.895 | 0.413 | 99761.992 | Suggested |
Cubic | 20.612 | 0.990 | 0.960 | + | Aliased |
+ Case(s) with leverage of 1.0000: PRESS statistic not defined.
Summary of analysis of variance (ANOVA) for the selected quadratic polynomial model for methanolic extraction and ethanolic extraction was listed in Table
ANOVA for response surface quadratic model.
Source | Coefficient estimate | Sum of squares | df | Mean square |
|
|
Remarks |
---|---|---|---|---|---|---|---|
Acidified methanolic extraction | |||||||
Model | 1114.595 | 152059.256 | 9 | 16895.473 | 26.906 | 0.0001 | Significant |
|
103.359 | 85464.893 | 1 | 85464.893 | 136.105 | <0.0001 | |
|
26.475 | 5607.274 | 1 | 5607.274 | 8.930 | 0.0203 | |
|
68.744 | 37805.617 | 1 | 37805.617 | 60.206 | 0.0001 | |
|
−45.922 | 8435.296 | 1 | 8435.296 | 13.433 | 0.0080 | |
|
22.196 | 1970.579 | 1 | 1970.579 | 3.138 | 0.1198 | |
|
9.254 | 342.543 | 1 | 342.543 | 0.546 | 0.4842 | |
|
24.728 | 2574.683 | 1 | 2574.683 | 4.100 | 0.0825 | |
|
−48.608 | 9948.234 | 1 | 9948.234 | 15.843 | 0.0053 | |
|
11.160 | 524.442 | 1 | 524.442 | 0.835 | 0.3912 | |
Residual | 4395.537 | 7 | 627.934 | ||||
Lack of fit | 874.804 | 3 | 291.601 | 0.331 | 0.8045 | Not significant | |
Pure error | 3520.733 | 4 | 880.183 | ||||
Cor total | 156454.793 | 16 | |||||
| |||||||
Acidified ethanolic extraction | |||||||
Model | 664.170 | 162047.676 | 9 | 18005.297 | 16.224 | 0.0007 | Significant |
|
118.806 | 112918.450 | 1 | 112918.450 | 101.747 | <0.0001 | |
|
57.681 | 26616.436 | 1 | 26616.436 | 23.983 | 0.0018 | |
|
31.763 | 8070.915 | 1 | 8070.915 | 7.272 | 0.0308 | |
|
7.237 | 209.468 | 1 | 209.468 | 0.189 | 0.6770 | |
|
−3.966 | 62.917 | 1 | 62.917 | 0.057 | 0.8186 | |
|
−8.071 | 260.580 | 1 | 260.580 | 0.235 | 0.6428 | |
|
31.220 | 4103.958 | 1 | 4103.958 | 3.698 | 0.0959 | |
|
−11.849 | 591.175 | 1 | 591.175 | 0.533 | 0.4892 | |
|
−48.100 | 9741.404 | 1 | 9741.404 | 8.778 | 0.0210 | |
Residual | 7768.593 | 7 | 1109.799 | ||||
Lack of fit | 6069.165 | 3 | 2023.055 | 4.762 | 0.0829 | Not significant | |
Pure error | 1699.428 | 4 | 424.857 | ||||
Cor total | 169816.269 | 16 |
The coefficient of determination (
Box-Behnken design arrangement and responses.
Run |
|
|
|
Anthocyanin yield* (mg/100 g) | Anthocyanin yield* (mg/100 g) | ||||
---|---|---|---|---|---|---|---|---|---|
(acidified methanolic extraction) | (acidified ethanolic extraction) | ||||||||
|
|
% sd |
|
|
% sd | ||||
1 | 45 | 90 | 20 | 1132.18 | 1114.60 | 3.40 | 654.60 | 664.17 | 2.65 |
2 | 60 | 120 | 20 | 1164.47 | 1110.16 | 1.32 | 897.29 | 867.26 | 3.46 |
3 | 30 | 60 | 20 | 925.12 | 1114.60 | 1.10 | 484.27 | 514.29 | 8.36 |
4 | 45 | 90 | 20 | 1154.45 | 1114.60 | 2.84 | 633.44 | 664.17 | 1.85 |
5 | 45 | 60 | 5 | 982.45 | 1114.60 | 2.75 | 537.70 | 506.71 | 2.19 |
6 | 45 | 90 | 20 | 1094.33 | 1000.58 | 1.74 | 685.77 | 664.17 | 3.13 |
7 | 30 | 120 | 20 | 1065.39 | 1213.52 | 4.05 | 612.29 | 615.18 | 4.41 |
8 | 45 | 120 | 5 | 1021.42 | 1025.63 | 3.10 | 642.07 | 638.21 | 2.08 |
9 | 45 | 90 | 20 | 1079.86 | 1162.90 | 4.65 | 676.86 | 664.17 | 4.25 |
10 | 60 | 90 | 5 | 1177.27 | 914.96 | 2.87 | 704.42 | 738.30 | 3.95 |
11 | 45 | 120 | 35 | 1190.35 | 1181.62 | 4.29 | 654.60 | 685.59 | 1.79 |
12 | 45 | 60 | 35 | 1114.37 | 991.18 | 4.28 | 582.51 | 586.38 | 7.79 |
13 | 30 | 90 | 5 | 999.15 | 1344.78 | 2.65 | 491.78 | 492.76 | 1.18 |
14 | 60 | 60 | 20 | 1207.89 | 1114.60 | 3.10 | 740.32 | 737.43 | 9.90 |
15 | 30 | 90 | 35 | 1079.31 | 1093.67 | 7.93 | 598.10 | 564.21 | 8.31 |
16 | 60 | 90 | 35 | 1346.21 | 1059.75 | 5.70 | 794.87 | 793.89 | 0.00 |
17 | 45 | 90 | 20 | 1112.15 | 1174.63 | 3.16 | 670.18 | 664.17 | 3.32 |
% sd < 10 is considered significant.
As both models showed a satisfactory fit, normal probability plot of the residuals were generated to check the normality of the residuals (Figure
Normal probability plots of residuals for (a) methanolic extraction and (b) ethanolic extraction.
The significance of each coefficient was determined by Fisher’s
For methanolic extraction, the anthocyanin yield can be increased with the increase of extraction temperature as shown in Figure
Contour plot showing the effects of variables for (a–c) methanolic extraction and (d-f) ethanolic extraction.
Table
Pertubation plot shows how a function of a certain factor responded as the level of that factor changes, when the other factors are fixed at their optimum levels [
Perturbation plots for (a) methanolic extraction at temperature 60°C; time 120 min; solid to liquid ratio 23.1 g/mL and (b) ethanolic extraction at temperature 60°C; time 86.89 min; solid to liquid ratio 35 g/mL.
Optimization of anthocyanin extraction from
Experiment confirmation of predicted value at optimal extraction condition.
Optimal levels |
Anthocyanin yield (mg/100 g) | |||||
---|---|---|---|---|---|---|
|
Experimental value | Mean* | Relative errora (%) | |||
1 | 2 | 3 | ||||
Acidified methanolic extraction | 1345.320 | 1297.503 | 1437.774 | 1303.348 | 1346.208 | 0.066 |
|
||||||
|
||||||
|
||||||
Acidified ethanolic extraction | 869.290 | 878.355 | 886.060 | 878.355 | 880.923 | 1.321 |
|
||||||
|
||||||
|
Mean is average value from triplicate of experimental run.
Desirability ramp of optimization for (a) methanolic extraction and (b) ethanolic extraction.
A triplicate experiment was set up to validate the optimized condition. As shown in Table
Figure
Mass spectra of anthocyanin-rich extract after gel filtration using Sephadex LH-20: (a–c) positive mode and (d-e) negative mode.
The experimental design approach using RSM was successfully applied in the optimization of anthocyanins from
Response surface methodology
Ultra performance liquid chromatography electrospray tandem mass spectrometry
Box-Behnken design.
The authors gratefully acknowledge The Institute of Research Management and Monitoring (IPPP), University of Malaya, Kuala Lumpur, Malaysia, for the research Grants PV082/2012A, BK024-2011A, and RP003B-13AFR and The Institute of Biological Sciences, Faculty of Science, University of Malaya, Kuala Lumpur, Malaysia, for providing the facilities to carry out this research.