Optimized Aqueous Extraction of Saponins from Bitter Melon for Production of a Saponin-Enriched Bitter Melon Powder Sing P. Tan, Quan V. Vuong, Costas E. Stathopoulos, Sophie E. Parks, and Paul D. Roach
Bitter melon, Momordica charantia L. (Cucurbitaceae), aqueous extracts are proposed to have health-promoting properties due to their content of saponins and their antioxidant activity. However, the optimal conditions for the aqueous extraction of saponins from bitter melon and the effects of spray drying have not been established. Therefore, this study aimed to optimize the aqueous extraction of the saponins from bitter melon, using response surface methodology, prepare a powder using spray drying, and compare the powder’s physical properties, components, and antioxidant capacity with aqueous and ethanol freeze-dried bitter melon powders and a commercial powder. The optimal aqueous extraction conditions were determined to be 40 °C for 15 min and the water-to-sample ratio was chosen to be 20:1 mL/g. For many of its physical properties, components, and antioxidant capacity, the aqueous spray-dried powder was comparable to the aqueous and ethanol freeze-dried bitter melon powders and the commercial powder. The optimal conditions for the aqueous extraction of saponins from bitter melon followed by spray drying gave a high quality powder in terms of saponins and antioxidant activity.
Abstract:
E: Food Engineering & Physical Properties
Keywords: antioxidant activity, aqueous extraction, bitter melon, saponins, spray drying
This study highlights that bitter melon is a rich source of saponin compounds and their associated antioxidant activities, which may provide health benefits. The findings of the current study will help with the development of extraction and drying technologies for the preparation of a saponin-enriched powdered extract from bitter melon. The powdered extract may have potential as a nutraceutical supplement or as a value-added ingredient for incorporation into functional foods.
Practical Application:
Introduction Bitter melon, Momordica charantia L. (Cucurbitaceae), is a tropical fruit, which has been used for medicinal purposes (Palaniswamy 2001). Studies have associated bitter melon with a wide range of therapeutic effects, including antidiabetic (Clouatre and others 2011), anticancer (Yasui and others 2005), antiviral (Beloin and others 2005), antiinflammatory (Hsu and others 2012), and hypolipidemic and hypocholesterolemic (Nerurkar and others 2010) activities. It has been proposed that these therapeutic effects are linked to several bioactive compounds, especially the saponins, as they are also found associated with various extracts prepared from bitter melon (Han and others 2008; Keller and others 2011). For the extraction of bioactive components from bitter melon, organic solvents such as methanol, acetone, and ethanol have often been used, with 70% ethanol and 30% water being one of the most effective solvents (Ji and others 2010; Hu and others 2012b). However, even though ethanol is not as toxic to health as some of the other organic solvents, water is always preferable as an extraction solvent because it is nontoxic, environmentally friendly, and inexpensive compared with organic solvents (Vuong and others MS 20140087 Submitted 1/16/2014, Accepted 4/1/2014. Authors Tan, Vuong, Stathopoulos, Parks, and Roach are with School of Environmental and Life Sciences, Univ. of Newcastle, Ourimbah, NSW 2258, Australia. Author Parks is with Central Coast Primary Industries Centre, NSW Dept. of Primary Industries, Ourimbah, NSW 2258, Australia. Direct inquiries to author Tan (E-mail: [email protected]).
E1372
Journal of Food Science r Vol. 79, Nr. 7, 2014
2011). Importantly, water has been found to effectively extract saponins from other plant materials such as safed musli, teas, and jua (Barve and others 2010; Ramdani and others 2013; Ribeiro and others 2014). Furthermore, powdered bitter melon aqueous extracts have been shown to improve glycemic control, suggesting that it could be safely prescribed to diabetic patients (Virdi and others 2003; Wang and others 2011). As such, water is the solvent of choice for extracting and preparing saponin-enriched extracts from bitter melon. However, to the best of our knowledge, the optimal conditions for the aqueous extraction of saponins from bitter melon have not been established. Therefore, it is desirable to study the optimal conditions for the aqueous extraction of the saponins from bitter melon in order to maximize the medicinal efficacy of the extracts and to maximize the commercial viability of these preparations. Once active components are extracted from their original source, it is often desirable to remove the solvent in order to prepare dry powders, which are easier to handle and often more stable than solvent extracts. This is especially true for water extracts, which are susceptible to oxidative degradation. Aqueous extracts can be processed using several drying techniques, with freeze drying and spray drying being 2 of the most popular methods. Freeze drying can produce high quality powder products but it has extremely high production costs (Liu and others 2011). In comparison, spray drying can also often produce high quality powder products but at much lower production costs (Shahidi and Pegg 2007). Therefore, the aims of this study were to (1) optimize the conditions for the aqueous extraction of saponins from R C 2014 Institute of Food Technologists
doi: 10.1111/1750-3841.12514 Further reproduction without permission is prohibited
Extraction of bitter melon saponins . . .
Materials and Methods Chemicals Ethanol, methanol, and hydrochloric acid (36%) were purchased from Fronine (Taren Point, NSW, Australia), Merck Pty. Ltd. (Kilsith, NSW, Australia), and Ajax Finechem (North Ryde, NSW, Australia), respectively. Folin-Ciocalteau (FC) reagent, 2,2 azinobis-(3-thylbenzothiozoline-6-sulfonic acid (ABTS), 2,2 diphenyl-1-picrylhydrazyl (DPPH), sodium carbonate, potassium persulfate, sodium nitrite, aluminium chloride, sodium hydroxide, vanillin, sodium acetate trihydrate, acetic acid, 2,4,6-tripyridyl-striazine (TPTZ), ferric (III) chloride hexahydrate, sulfuric acid, and standards (trolox, gallic acid, rutin, catechin, and aecsin) were all purchased from Sigma Pty. Ltd. (Castle Hill, NSW, Australia). Deionized water was prepared on the day of use with a Milli-Q Direct 16 water purification system (Millipore Australia Pty Ltd, North Ryde, NSW, Australia). Commercial bitter melon capsules were purchased from Nature’s Goodness Australia (Narellan, NSW, Australia). Nature’s Goodness claims that “each capsule contains 500 mg of bitter melon extract, which is equivalent to 5 g of fresh bitter melon.” Preparation of bitter melon The design of the experiments for preparing the bitter melon extracts and powders is described in Figure 1. To start off, fresh bitter melons (moonlight variety) were purchased from the Syd-
Freeze drying & Grinding
Filtering & Centrifugation Concentration
Spray drying
Aqueous Spraydried powder
Extraction of bitter melon To determine the optimal conditions for the aqueous extraction of the saponins from bitter melon, preliminary experiments using single factors were first done to identify important factors which could impact on the extraction of the saponins (data not shown). The extraction temperature (A, referred to as temperature), the length of the extraction (B, referred to as time), and the ratio of water (mL) to bitter melon (g; C, referred to as ratio) were identified as important parameters and relevant values for each were determined for use in the RSM as code variable levels. A RSM with the Box–Behnken system (Box and Behnken 1960) was then employed to design an experiment consisting of 15 runs with 12 factorial points and 3 central points. The independent parameters were temperature (A), time (B), and ratio (C) and the chosen code variable levels for each were based on the preliminary experiments (Table 1). To express the total saponin content as a function of the independent variables, a 2nd-order polynomial equation was generated as follows: Yi = a 0 + a 1 A + a 2 B + a 3 C + a 11 A2 + a 22 B 2 + a 33 C2 + a 12 AB + a 13 AC + a 23 BC
(1)
where Y is the dependent response (saponins); A, B, and C are the levels of the independent variables (temperature, time, and ratio, respectively), and a0 , ai , aii , and aij are the regression coefficients of the variables for the offset, linear, quadratic, and interaction terms, respectively.
Fresh bitter melon
Aqueous extraction (40°C, 15min, 20:1 mL/g)
ney Markets (Sydney, NSW, Australia) in December 2012 and were immediately stored at –20 °C until use. The whole fruit, including flesh, seeds, and arils, was used in this study. The frozen bitter melons were cut into slices (approximately 1 to 2 mm thick), placed in liquid nitrogen, and then dried using a FD3 freeze dryer (Rietschle Thomas, Seven Hills, NSW, Australia). Using a commercial blender (John Morris Scientific, Chatswood, NSW, Australia), the freeze-dried bitter melon slices were ground and passed through a 1 mm mesh sieve. The ground freeze-dried bitter melon preparation was then sealed and stored at room temperature (RT) until use.
70% Ethanol extraction (RT, 48 h, 20:1 mL/g) Filtering & Centrifugation Concentration
Freeze drying
Freeze drying
Aqueous Freezedried powder
Ethanol Freezedried powder
Figure 1–Experimental design for the production of bitter melon powders. Fresh bitter melon was first freeze dried and ground. This preparation was either extracted with water using the optimum conditions determine using RSM or with 70% ethanol. The aqueous extract was concentrated using a rotavapor and then either spray-dried or freeze-dried. The ethanol extract was concentrated using a rotavapor and freeze dried.
Preparation of powders from the bitter melon extracts To prepare powders from the aqueous bitter melon extract, the freeze-dried and ground bitter melon preparation described above was extracted at 40 °C for 15 min at a water-to-bitter melon ratio of 20:1 mL/g using a stirrer at 600 rpm (Figure 1); these were the conditions which gave the highest amount of saponins in the aqueous extract as determined using RSM as described above. The extract was then vacuum filtered through 2 layers of cheesecloth followed by a Whatman No. 1 filter paper (Lomb Scientific, Taren Point, NSW, Australia) to remove insoluble bitter melon residues. The extract was then centrifuged at 4350 × g for 10 min at 10 °C to remove insoluble fine particles using a Beckman JA-20 rotor in a Beckman J20MC Centrifuge (Beckman Instruments Inc., Palo Alto, Calif., U.S.A.). The aqueous extract was then concentrated using a rotavapor (BuchiRotavapor B-480, Noble Park, Vic, Australia) under reduced pressure (Figure 1). Finally, powders were prepared using spray- and freeze-drying methods (Figure 1). An aqueous spraydried powder was prepared by drying the concentrated extract from the rotavapor step using a spray drier (Buchi Mini Spray drier B-290, Noble Park Vic, Australia) with the inlet temperature set at 150 °C, the outlet temperature set at 90 °C, and the Vol. 79, Nr. 7, 2014 r Journal of Food Science E1373
E: Food Engineering & Physical Properties
bitter melon using response surface methodology (RSM), (2) prepare a saponin-enriched powder from the aqueous extract using spray drying, and (3) compare the aqueous spray-dried powder to freeze-dried aqueous and ethanol extracted powders as well as to a commercially available bitter melon powder. For comparison between the powders, their physicochemical properties were determined.
Extraction of bitter melon saponins . . . Table 1–Independent variables and their code variable levels used for the Box–Behnken design. Independent variables Coded variable levels +1 0 −1
Temperature (°C) (A)
Time (min) (B)
Ratio (mL/g) (C)
60 40 20
25 15 5
100 55 10
E: Food Engineering & Physical Properties
aspiration set at 100%. An aqueous freeze-dried powder was obtained by first freezing the concentrated extract obtained from the rotavapor step using liquid nitrogen and then drying it under vacuum using a FD3 freeze dryer with the temperature set at –40 °C for 48 h. For comparison, an ethanol freeze-dried powder was prepared by first extracting the ground freeze-dried bitter melon preparation using 70% ethanol and 30% water at RT for 48 h at a solvent to bitter melon ratio of 20:1 mL/g using a stirrer at 600 rpm (Figure 1). As described for the aqueous powders, the ethanol extract was then filtered, centrifuged, concentrated using the rotavapor, frozen using liquid nitrogen, and finally freeze-dried using a FD3 freeze drier to obtain the ethanol freeze-dried powder. All the powders were kept in sealed and labeled containers at 5 °C until use. In addition, commercial bitter melon capsules were used for comparison with the aqueous spray- and freeze-dried powders and the ethanol freeze-dried powder.
Determination of physical properties and composition of the powders Physical properties. Moisture content was determined by weight difference after drying at 80 °C for 24 h in a vacuum oven drier (Thermoline Scientific, Wetherill Park, NSW, Australia). Water activity was measured using a water activity meter (AquaLabPawkit, Pullman, Wash., U.S.A.).The water solubility index of the powders was determined using the method of Anderson and others (1969) with some modifications. The powder (2.5 g) was dispersed in 25 mL of deionized water and vigorously agitated using a vortex mixer for 10 min at RT. The solution was transferred to a centrifuge tube and centrifuged at 4350 × g for 10 min in a Beckman J20 MC Centrifuge. Finally, the supernatant was vacuum oven dried at 80 °C for 24 h. The water solubility index was calculated by the difference between the weight of the powder obtained after drying and the initial weight of powder used in the solubility test, and expressed as a percentage. To determine the pH of the powders, 2.5 g of powder was dissolved in 25 mL of deionized water and the pH was measured using a pH meter (MP220, Mettler Toledo Ltd, Port Melbourne, Vic, Australia) calibrated at 20 °C with standard pH 4 and 7 buffers. The color of the powders was measured using a Minolta chroma meter (CR-400 Chroma meter, Konica Minolta Sensing, Sakai, Osaka, Japan). The meter was calibrated with a white standard tile before measurement. The samples were packed into a polyethylene pouch for measurement and the results were expressed as Hunter color values of lightness, chroma, and hue angle. Chroma indicates color intensity or saturation and was calculated by the formula (a∗2 + b∗2 )1/2 . The hue angle was calculated by the formula: arctan (b∗ /a∗ ). Total saponin content. The total saponin content of the extracts from the bitter melon preparation was measured according to the method of Hiai and others (1976) with slight modifications. For the extracts, 0.3 mL was mixed with 0.3 mL of 8% E1374 Journal of Food Science r Vol. 79, Nr. 7, 2014
(w/v) vanillin solution and 3 mL of 72% (v/v) sulphuric acid. The mixture was mixed and incubated at 60 °C for 15 min and then cooled on ice for 10 min. The absorption of the mixture was measured at 560 nm using a spectrophotometer (Carry 50 Bio, Varian Pty. Ltd., Mulgrave, Vic, Australia) against a reagent blank. Aecsin was used as a standard and results were expressed aecsin equivalents (AE) per gram of the dry weight of the bitter melon preparation used in the extraction (mg AE/g). For the powders, 1 mg was dissolved in 1 mL of deionized water and 300 μL was used in the assay as described for the extracts above and the results were expressed as AE per gram of powder (mg AE/g). Total phenolic content. The total phenolic content of the powders was determined using the method of Cicco and others (2009) with some modifications. Briefly, 1 mg was dissolved in 1 mL of deionized water and 0.3 mL was mixed with 0.3 mL of FC reagent. The solution was mixed well and incubated at RT for 2 min to equilibrate. Then, 2.4 mL of a 5% (w/v) sodium carbonate solution was added and the solution mixed and incubated at RT for 2 h. The absorption of the solution was recorded at 765 nm using a spectrophotometer against a reagent blank. Gallic acid was used as a standard and results were expressed as gallic acid equivalents (GAE) per gram of powder (mg GAE/g). Total flavonoid content. A method described by Wu and Ng (2008) with some modifications was employed to determine the total flavonoid content of the powders. Briefly, 1 mg was dissolved in 1 mL of deionized water and 0.5 mL was mixed with 2 mL of deionized water followed by the addition of 0.15 mL of 5% (w/v) sodium nitrite solution. After 6 min, 0.15 mL of 10% (w/v) aluminum chloride was added and the mixture incubated for another 6 min. Then, 2 mL of 4% (w/v) sodium hydroxide was added. The solution was then made up to 5 mL with deionized water, mixed and placed in the dark at RT for 15 min. The absorption of the solution was measured at 510 nm using a spectrophotometer against a reagent blank. Rutin was used as a standard and results were expressed as rutin equivalents (RE) per gram of powder (mg RE/g). Total proanthocyanidin content. The total proanthocyanidin content of the powders was determined as described by Li and others (2006). Briefly, 1 mg was dissolved in 1 mL of deionized water and 0.5 mL was mixed with 0.3 mL of 4% (w/v) vanillin solution and 1.5 mL of HCl (36%) and incubated at RT for 15 min. The absorption of the solution was measured at 500 nm using a spectrophotometer against a reagent blank. Catechin was used as a standard and the results were expressed as catechin equivalents (CE) per gram of powder (mg CE/g).
Determination of antioxidant capacity ABTS assay. The ABTS assay was as described by Re and others (1999) with some modifications. Stock solutions of 7.4 mM ABTS and 2.6 mM potassium persulfate were prepared and kept at 4 °C until use. The working solution was prepared by mixing the 2 stock solutions in equal quantities and incubating them for
Extraction of bitter melon saponins . . .
Results and Discussion Fitting the response surface methodology model Based on the preliminary extraction experiments (data not shown), temperature, time, and ratio were identified as important parameters, which could impact on the extraction of saponins from bitter melon and relevant values for each parameter were determined for use in the RSM as code variable levels (Table 1): 20, 40, and 60 °C for the extraction temperature (A), 5, 15, and 25 min for the length of the extraction (B), and 10, 55, and 100 for the ratio (C) of water (mL) to bitter melon.
Then, to determine the optimal conditions for the aqueous extraction of the saponins from bitter melon, a predictive equation was used for fitting the experimental data to the Box–Behnken design model. The experimental data for the total saponin content of the aqueous bitter melon extracts was analyzed using multiple regression and response surface analysis to fit the generated 2nd-order polynomial equation for the 3 independent variables and their code variable levels. The predictive equation for the total saponin content, with coefficient values for each independent variable, was as follows: Y = 50.083 − 2.059A + 1.524B + 5.205C − 2.590A2 − 2.295B 2 − 4.333C2 − 0.868AB + 2.025AC − 1.6BC (2) where the equation represents the correlation between the response Y (total saponin content) and the independent variables, temperature (A), time (B), and ratio (C). The statistical analysis showed that the 2nd-order polynomial model adequately represented the true behavior of the system and that the model could be used for predicting the experimental data (Table 2 and 3). There was an excellent correlation between the experimental data and the predicted data obtained from the Box–Behnken design (Table 2); based on the coefficient of determination (R2 = 0.96), a total of 96% of the variation in the total saponin content could be explained while only 4% of the variation could not be explained by the model (Table 3). In addition, the lack of fit for the total saponin content was nonsignificant (P = 0.395), indicating that the 2nd-order polynomial model was adequate to describe the effects of the variables in the model.
Impact of the extraction parameters on the total saponin content of the extracts The correlations of the total saponin content of the extracts with the independent variables, temperature, time, and ratio, as determined using the t ratio and the P value (Table 3), were also significant (P < 0.05) for 2 of the 3 1st-order linear effects (A was inversely related and C was directly related but B was not related) and 2 of the 3 2nd-order quadratic effects (A2 and C2 were inversely related but B2 was not related). However, the correlations were not significant for the interactive effects (AB, AC, and BC). Based on their regression coefficients, the independent variables with significant effects could also be listed from the largest to the smallest effect on the total saponin content of the extract (Table 3) as follows: linear ratio (C), quadratic ratio (C2 ), quadratic temperature (A2 ), and linear temperature (A).Therefore, the ratio (C) had a bigger influence on the total saponin content of the extracts than temperature (A) and time (B), with the latter having no effect (Table 3). The influence of the 3 independent variables, temperature, time, and ratio on the total saponin content of the extracts as obtained from the predictive equation is presented in the 3D surface and 2D contour plots of Figure 2(A–C). Figure 2(A and B) shows that the total saponin content is predicted to change significantly in the range of temperatures from 20 to 60 °C; the saponin content is predicted to increase and reach a peak at 40 °C but then dramatically decrease when the temperature exceeds 40 °C. The latter finding indicates that the saponins extracted from bitter melon are sensitive to heat and tend to degrade when exposed to temperatures higher than 40 °C. This finding is consistent with a recent Vol. 79, Nr. 7, 2014 r Journal of Food Science E1375
E: Food Engineering & Physical Properties
15 to 16 h in the dark at RT. Then, 1 mL of the working solution was mixed with 60 mL of methanol to obtain an absorbance of 1.1 ± 0.02 units at 734 nm using a spectrophotometer. Fresh working solution was prepared for each assay. For the powders, different concentrations were prepared in deionized water (0.25, 0.5, 0.75, and 1 mg/mL) and 0.15 mL was mixed with 2.85 mL of the working solution and incubated for 2 h in the dark at RT. The absorption of the solution at 734 nm was measured using a spectrophotometer. Trolox was used as a standard and results were expressed as trolox equivalents (TE) per gram of powder (μmole TE/g). DPPH assay. The DPPH assay was as described by BrandWilliams and others (1995) with some modifications. A stock solution of 0.6 M DPPH in methanol was prepared and kept at –20 °C until use. The working solution was prepared by mixing 10 mL of stock solution with 45 mL of methanol to obtain an absorbance of 1.1 ± 0.02 units at 515 nm using a spectrophotometer. For the powders, different concentrations were prepared in deionized water (0.25, 0.5, 0.75, and 1 mg/mL) and 0.15 mL was mixed with 2.85 mL of working solution. The sample was allowed to stand for 30 min and the absorption at 515 nm was then measured using a spectrophotometer. Trolox was used as a standard and results were expressed as TE per gram of powder (μmole TE/g). FRAP assay. The FRAP assay was as described by Benzie and Strain (1996) with some modifications. The stock solutions of 300 mM acetate buffer (pH 3.6), 10 mM TPTZ in 40 mMHCl, and 20 mM ferric (III) chloride hexahydrate in deionized water were prepared and kept at 4 °C until use. The fresh working solution was prepared by mixing 100 mL of acetate buffer, 10 mL of TPTZ, and 10 mL of ferric (III) chloride hexahydrate in a ratio of 10:1:1. The working solution was then incubated at 37 °C before use. For the powders, different concentrations were prepared in deionized water (0.25, 0.5, 0.75, and 1 mg/mL) and 0.15 mL was mixed with 2.85 mL of the working solution for 30 min in the dark at RT. The absorption of the solution at 593 nm was measured using a spectrophotometer against a reagent blank. Trolox was used as a standard and results were expressed as TE per gram of powder (μmole TE/g). Statistical analyses. The adequacy of the RSM second-order polynomial model was determined based on the lack of fit and the coefficient of determination (R2 ). The three-dimensional (3D) surface and two-dimensional (2D) contour plots of the total saponin content were drawn using the JMP software version 10.0. Results were expressed as mean values with standard deviations and significant differences between treatments were tested using analysis of variance and the Bonferroni posthoc test with a 5% significance level (P < 0.05). Correlations among data were calculated using Pearson’s correlation coefficient and expressed as R2 . The SPSS software version 19.0 statistical package was used for all analyses.
Extraction of bitter melon saponins . . . Table 2–The experimental and predicted values for the total saponin content of the aqueous extracts measured or predicted for the 15 run Box–Behnken design. Total saponin content(mg AE/g) Run
E: Food Engineering & Physical Properties
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Pattern
Temperature(˚C)
Time(min)
Ratio(mL/g)
Experimental
Predicted
000 ++0 +0− −−0 +−0 000 −0− 0−− 0+− −+0 0++ 0−+ +0+ 000 −0+
40 60 60 20 60 40 20 40 40 20 40 40 60 40 20
15 25 15 5 5 15 15 5 25 25 25 5 15 15 15
55 55 10 55 55 55 10 10 10 55 100 100 100 55 100
51.6 43.8 32.8 44.9 44.2 48.5 42.7 34.5 42.4 48.0 49.2 47.7 47.7 50.2 49.5
50.1 43.8 33.9 44.9 42.5 50.1 42.1 35.1 41.4 49.7 48.6 48.7 48.4 50.1 48.4
Table 3–The regression coefficients and their significance values generated from the fitted 2nd order equation for the Box–Behnken design. Total saponin content (mg AE/g) Regression coefficienta
Regression coefficient values
t ratio
P value
50.1∗
47.71
Bitter melon, Momordica charantia L. (Cucurbitaceae), aqueous extracts are proposed to have health-promoting properties due to their content of saponins and their antioxidant activity. However, the optimal conditions for the aqueous extraction of saponins from bitter melon and the effects of spray drying have not been established. Therefore, this study aimed to optimize the aqueous extraction of the saponins from bitter melon, using response surface methodology, prepare a powder using spray drying, and compare the powder’s physical properties, components, and antioxidant capacity with aqueous and ethanol freeze-dried bitter melon powders and a commercial powder. The optimal aqueous extraction conditions were determined to be 40 °C for 15 min and the water-to-sample ratio was chosen to be 20:1 mL/g. For many of its physical properties, components, and antioxidant capacity, the aqueous spray-dried powder was comparable to the aqueous and ethanol freeze-dried bitter melon powders and the commercial powder. The optimal conditions for the aqueous extraction of saponins from bitter melon followed by spray drying gave a high quality powder in terms of saponins and antioxidant activity.
Abstract:
E: Food Engineering & Physical Properties
Keywords: antioxidant activity, aqueous extraction, bitter melon, saponins, spray drying
This study highlights that bitter melon is a rich source of saponin compounds and their associated antioxidant activities, which may provide health benefits. The findings of the current study will help with the development of extraction and drying technologies for the preparation of a saponin-enriched powdered extract from bitter melon. The powdered extract may have potential as a nutraceutical supplement or as a value-added ingredient for incorporation into functional foods.
Practical Application:
Introduction Bitter melon, Momordica charantia L. (Cucurbitaceae), is a tropical fruit, which has been used for medicinal purposes (Palaniswamy 2001). Studies have associated bitter melon with a wide range of therapeutic effects, including antidiabetic (Clouatre and others 2011), anticancer (Yasui and others 2005), antiviral (Beloin and others 2005), antiinflammatory (Hsu and others 2012), and hypolipidemic and hypocholesterolemic (Nerurkar and others 2010) activities. It has been proposed that these therapeutic effects are linked to several bioactive compounds, especially the saponins, as they are also found associated with various extracts prepared from bitter melon (Han and others 2008; Keller and others 2011). For the extraction of bioactive components from bitter melon, organic solvents such as methanol, acetone, and ethanol have often been used, with 70% ethanol and 30% water being one of the most effective solvents (Ji and others 2010; Hu and others 2012b). However, even though ethanol is not as toxic to health as some of the other organic solvents, water is always preferable as an extraction solvent because it is nontoxic, environmentally friendly, and inexpensive compared with organic solvents (Vuong and others MS 20140087 Submitted 1/16/2014, Accepted 4/1/2014. Authors Tan, Vuong, Stathopoulos, Parks, and Roach are with School of Environmental and Life Sciences, Univ. of Newcastle, Ourimbah, NSW 2258, Australia. Author Parks is with Central Coast Primary Industries Centre, NSW Dept. of Primary Industries, Ourimbah, NSW 2258, Australia. Direct inquiries to author Tan (E-mail: [email protected]).
E1372
Journal of Food Science r Vol. 79, Nr. 7, 2014
2011). Importantly, water has been found to effectively extract saponins from other plant materials such as safed musli, teas, and jua (Barve and others 2010; Ramdani and others 2013; Ribeiro and others 2014). Furthermore, powdered bitter melon aqueous extracts have been shown to improve glycemic control, suggesting that it could be safely prescribed to diabetic patients (Virdi and others 2003; Wang and others 2011). As such, water is the solvent of choice for extracting and preparing saponin-enriched extracts from bitter melon. However, to the best of our knowledge, the optimal conditions for the aqueous extraction of saponins from bitter melon have not been established. Therefore, it is desirable to study the optimal conditions for the aqueous extraction of the saponins from bitter melon in order to maximize the medicinal efficacy of the extracts and to maximize the commercial viability of these preparations. Once active components are extracted from their original source, it is often desirable to remove the solvent in order to prepare dry powders, which are easier to handle and often more stable than solvent extracts. This is especially true for water extracts, which are susceptible to oxidative degradation. Aqueous extracts can be processed using several drying techniques, with freeze drying and spray drying being 2 of the most popular methods. Freeze drying can produce high quality powder products but it has extremely high production costs (Liu and others 2011). In comparison, spray drying can also often produce high quality powder products but at much lower production costs (Shahidi and Pegg 2007). Therefore, the aims of this study were to (1) optimize the conditions for the aqueous extraction of saponins from R C 2014 Institute of Food Technologists
doi: 10.1111/1750-3841.12514 Further reproduction without permission is prohibited
Extraction of bitter melon saponins . . .
Materials and Methods Chemicals Ethanol, methanol, and hydrochloric acid (36%) were purchased from Fronine (Taren Point, NSW, Australia), Merck Pty. Ltd. (Kilsith, NSW, Australia), and Ajax Finechem (North Ryde, NSW, Australia), respectively. Folin-Ciocalteau (FC) reagent, 2,2 azinobis-(3-thylbenzothiozoline-6-sulfonic acid (ABTS), 2,2 diphenyl-1-picrylhydrazyl (DPPH), sodium carbonate, potassium persulfate, sodium nitrite, aluminium chloride, sodium hydroxide, vanillin, sodium acetate trihydrate, acetic acid, 2,4,6-tripyridyl-striazine (TPTZ), ferric (III) chloride hexahydrate, sulfuric acid, and standards (trolox, gallic acid, rutin, catechin, and aecsin) were all purchased from Sigma Pty. Ltd. (Castle Hill, NSW, Australia). Deionized water was prepared on the day of use with a Milli-Q Direct 16 water purification system (Millipore Australia Pty Ltd, North Ryde, NSW, Australia). Commercial bitter melon capsules were purchased from Nature’s Goodness Australia (Narellan, NSW, Australia). Nature’s Goodness claims that “each capsule contains 500 mg of bitter melon extract, which is equivalent to 5 g of fresh bitter melon.” Preparation of bitter melon The design of the experiments for preparing the bitter melon extracts and powders is described in Figure 1. To start off, fresh bitter melons (moonlight variety) were purchased from the Syd-
Freeze drying & Grinding
Filtering & Centrifugation Concentration
Spray drying
Aqueous Spraydried powder
Extraction of bitter melon To determine the optimal conditions for the aqueous extraction of the saponins from bitter melon, preliminary experiments using single factors were first done to identify important factors which could impact on the extraction of the saponins (data not shown). The extraction temperature (A, referred to as temperature), the length of the extraction (B, referred to as time), and the ratio of water (mL) to bitter melon (g; C, referred to as ratio) were identified as important parameters and relevant values for each were determined for use in the RSM as code variable levels. A RSM with the Box–Behnken system (Box and Behnken 1960) was then employed to design an experiment consisting of 15 runs with 12 factorial points and 3 central points. The independent parameters were temperature (A), time (B), and ratio (C) and the chosen code variable levels for each were based on the preliminary experiments (Table 1). To express the total saponin content as a function of the independent variables, a 2nd-order polynomial equation was generated as follows: Yi = a 0 + a 1 A + a 2 B + a 3 C + a 11 A2 + a 22 B 2 + a 33 C2 + a 12 AB + a 13 AC + a 23 BC
(1)
where Y is the dependent response (saponins); A, B, and C are the levels of the independent variables (temperature, time, and ratio, respectively), and a0 , ai , aii , and aij are the regression coefficients of the variables for the offset, linear, quadratic, and interaction terms, respectively.
Fresh bitter melon
Aqueous extraction (40°C, 15min, 20:1 mL/g)
ney Markets (Sydney, NSW, Australia) in December 2012 and were immediately stored at –20 °C until use. The whole fruit, including flesh, seeds, and arils, was used in this study. The frozen bitter melons were cut into slices (approximately 1 to 2 mm thick), placed in liquid nitrogen, and then dried using a FD3 freeze dryer (Rietschle Thomas, Seven Hills, NSW, Australia). Using a commercial blender (John Morris Scientific, Chatswood, NSW, Australia), the freeze-dried bitter melon slices were ground and passed through a 1 mm mesh sieve. The ground freeze-dried bitter melon preparation was then sealed and stored at room temperature (RT) until use.
70% Ethanol extraction (RT, 48 h, 20:1 mL/g) Filtering & Centrifugation Concentration
Freeze drying
Freeze drying
Aqueous Freezedried powder
Ethanol Freezedried powder
Figure 1–Experimental design for the production of bitter melon powders. Fresh bitter melon was first freeze dried and ground. This preparation was either extracted with water using the optimum conditions determine using RSM or with 70% ethanol. The aqueous extract was concentrated using a rotavapor and then either spray-dried or freeze-dried. The ethanol extract was concentrated using a rotavapor and freeze dried.
Preparation of powders from the bitter melon extracts To prepare powders from the aqueous bitter melon extract, the freeze-dried and ground bitter melon preparation described above was extracted at 40 °C for 15 min at a water-to-bitter melon ratio of 20:1 mL/g using a stirrer at 600 rpm (Figure 1); these were the conditions which gave the highest amount of saponins in the aqueous extract as determined using RSM as described above. The extract was then vacuum filtered through 2 layers of cheesecloth followed by a Whatman No. 1 filter paper (Lomb Scientific, Taren Point, NSW, Australia) to remove insoluble bitter melon residues. The extract was then centrifuged at 4350 × g for 10 min at 10 °C to remove insoluble fine particles using a Beckman JA-20 rotor in a Beckman J20MC Centrifuge (Beckman Instruments Inc., Palo Alto, Calif., U.S.A.). The aqueous extract was then concentrated using a rotavapor (BuchiRotavapor B-480, Noble Park, Vic, Australia) under reduced pressure (Figure 1). Finally, powders were prepared using spray- and freeze-drying methods (Figure 1). An aqueous spraydried powder was prepared by drying the concentrated extract from the rotavapor step using a spray drier (Buchi Mini Spray drier B-290, Noble Park Vic, Australia) with the inlet temperature set at 150 °C, the outlet temperature set at 90 °C, and the Vol. 79, Nr. 7, 2014 r Journal of Food Science E1373
E: Food Engineering & Physical Properties
bitter melon using response surface methodology (RSM), (2) prepare a saponin-enriched powder from the aqueous extract using spray drying, and (3) compare the aqueous spray-dried powder to freeze-dried aqueous and ethanol extracted powders as well as to a commercially available bitter melon powder. For comparison between the powders, their physicochemical properties were determined.
Extraction of bitter melon saponins . . . Table 1–Independent variables and their code variable levels used for the Box–Behnken design. Independent variables Coded variable levels +1 0 −1
Temperature (°C) (A)
Time (min) (B)
Ratio (mL/g) (C)
60 40 20
25 15 5
100 55 10
E: Food Engineering & Physical Properties
aspiration set at 100%. An aqueous freeze-dried powder was obtained by first freezing the concentrated extract obtained from the rotavapor step using liquid nitrogen and then drying it under vacuum using a FD3 freeze dryer with the temperature set at –40 °C for 48 h. For comparison, an ethanol freeze-dried powder was prepared by first extracting the ground freeze-dried bitter melon preparation using 70% ethanol and 30% water at RT for 48 h at a solvent to bitter melon ratio of 20:1 mL/g using a stirrer at 600 rpm (Figure 1). As described for the aqueous powders, the ethanol extract was then filtered, centrifuged, concentrated using the rotavapor, frozen using liquid nitrogen, and finally freeze-dried using a FD3 freeze drier to obtain the ethanol freeze-dried powder. All the powders were kept in sealed and labeled containers at 5 °C until use. In addition, commercial bitter melon capsules were used for comparison with the aqueous spray- and freeze-dried powders and the ethanol freeze-dried powder.
Determination of physical properties and composition of the powders Physical properties. Moisture content was determined by weight difference after drying at 80 °C for 24 h in a vacuum oven drier (Thermoline Scientific, Wetherill Park, NSW, Australia). Water activity was measured using a water activity meter (AquaLabPawkit, Pullman, Wash., U.S.A.).The water solubility index of the powders was determined using the method of Anderson and others (1969) with some modifications. The powder (2.5 g) was dispersed in 25 mL of deionized water and vigorously agitated using a vortex mixer for 10 min at RT. The solution was transferred to a centrifuge tube and centrifuged at 4350 × g for 10 min in a Beckman J20 MC Centrifuge. Finally, the supernatant was vacuum oven dried at 80 °C for 24 h. The water solubility index was calculated by the difference between the weight of the powder obtained after drying and the initial weight of powder used in the solubility test, and expressed as a percentage. To determine the pH of the powders, 2.5 g of powder was dissolved in 25 mL of deionized water and the pH was measured using a pH meter (MP220, Mettler Toledo Ltd, Port Melbourne, Vic, Australia) calibrated at 20 °C with standard pH 4 and 7 buffers. The color of the powders was measured using a Minolta chroma meter (CR-400 Chroma meter, Konica Minolta Sensing, Sakai, Osaka, Japan). The meter was calibrated with a white standard tile before measurement. The samples were packed into a polyethylene pouch for measurement and the results were expressed as Hunter color values of lightness, chroma, and hue angle. Chroma indicates color intensity or saturation and was calculated by the formula (a∗2 + b∗2 )1/2 . The hue angle was calculated by the formula: arctan (b∗ /a∗ ). Total saponin content. The total saponin content of the extracts from the bitter melon preparation was measured according to the method of Hiai and others (1976) with slight modifications. For the extracts, 0.3 mL was mixed with 0.3 mL of 8% E1374 Journal of Food Science r Vol. 79, Nr. 7, 2014
(w/v) vanillin solution and 3 mL of 72% (v/v) sulphuric acid. The mixture was mixed and incubated at 60 °C for 15 min and then cooled on ice for 10 min. The absorption of the mixture was measured at 560 nm using a spectrophotometer (Carry 50 Bio, Varian Pty. Ltd., Mulgrave, Vic, Australia) against a reagent blank. Aecsin was used as a standard and results were expressed aecsin equivalents (AE) per gram of the dry weight of the bitter melon preparation used in the extraction (mg AE/g). For the powders, 1 mg was dissolved in 1 mL of deionized water and 300 μL was used in the assay as described for the extracts above and the results were expressed as AE per gram of powder (mg AE/g). Total phenolic content. The total phenolic content of the powders was determined using the method of Cicco and others (2009) with some modifications. Briefly, 1 mg was dissolved in 1 mL of deionized water and 0.3 mL was mixed with 0.3 mL of FC reagent. The solution was mixed well and incubated at RT for 2 min to equilibrate. Then, 2.4 mL of a 5% (w/v) sodium carbonate solution was added and the solution mixed and incubated at RT for 2 h. The absorption of the solution was recorded at 765 nm using a spectrophotometer against a reagent blank. Gallic acid was used as a standard and results were expressed as gallic acid equivalents (GAE) per gram of powder (mg GAE/g). Total flavonoid content. A method described by Wu and Ng (2008) with some modifications was employed to determine the total flavonoid content of the powders. Briefly, 1 mg was dissolved in 1 mL of deionized water and 0.5 mL was mixed with 2 mL of deionized water followed by the addition of 0.15 mL of 5% (w/v) sodium nitrite solution. After 6 min, 0.15 mL of 10% (w/v) aluminum chloride was added and the mixture incubated for another 6 min. Then, 2 mL of 4% (w/v) sodium hydroxide was added. The solution was then made up to 5 mL with deionized water, mixed and placed in the dark at RT for 15 min. The absorption of the solution was measured at 510 nm using a spectrophotometer against a reagent blank. Rutin was used as a standard and results were expressed as rutin equivalents (RE) per gram of powder (mg RE/g). Total proanthocyanidin content. The total proanthocyanidin content of the powders was determined as described by Li and others (2006). Briefly, 1 mg was dissolved in 1 mL of deionized water and 0.5 mL was mixed with 0.3 mL of 4% (w/v) vanillin solution and 1.5 mL of HCl (36%) and incubated at RT for 15 min. The absorption of the solution was measured at 500 nm using a spectrophotometer against a reagent blank. Catechin was used as a standard and the results were expressed as catechin equivalents (CE) per gram of powder (mg CE/g).
Determination of antioxidant capacity ABTS assay. The ABTS assay was as described by Re and others (1999) with some modifications. Stock solutions of 7.4 mM ABTS and 2.6 mM potassium persulfate were prepared and kept at 4 °C until use. The working solution was prepared by mixing the 2 stock solutions in equal quantities and incubating them for
Extraction of bitter melon saponins . . .
Results and Discussion Fitting the response surface methodology model Based on the preliminary extraction experiments (data not shown), temperature, time, and ratio were identified as important parameters, which could impact on the extraction of saponins from bitter melon and relevant values for each parameter were determined for use in the RSM as code variable levels (Table 1): 20, 40, and 60 °C for the extraction temperature (A), 5, 15, and 25 min for the length of the extraction (B), and 10, 55, and 100 for the ratio (C) of water (mL) to bitter melon.
Then, to determine the optimal conditions for the aqueous extraction of the saponins from bitter melon, a predictive equation was used for fitting the experimental data to the Box–Behnken design model. The experimental data for the total saponin content of the aqueous bitter melon extracts was analyzed using multiple regression and response surface analysis to fit the generated 2nd-order polynomial equation for the 3 independent variables and their code variable levels. The predictive equation for the total saponin content, with coefficient values for each independent variable, was as follows: Y = 50.083 − 2.059A + 1.524B + 5.205C − 2.590A2 − 2.295B 2 − 4.333C2 − 0.868AB + 2.025AC − 1.6BC (2) where the equation represents the correlation between the response Y (total saponin content) and the independent variables, temperature (A), time (B), and ratio (C). The statistical analysis showed that the 2nd-order polynomial model adequately represented the true behavior of the system and that the model could be used for predicting the experimental data (Table 2 and 3). There was an excellent correlation between the experimental data and the predicted data obtained from the Box–Behnken design (Table 2); based on the coefficient of determination (R2 = 0.96), a total of 96% of the variation in the total saponin content could be explained while only 4% of the variation could not be explained by the model (Table 3). In addition, the lack of fit for the total saponin content was nonsignificant (P = 0.395), indicating that the 2nd-order polynomial model was adequate to describe the effects of the variables in the model.
Impact of the extraction parameters on the total saponin content of the extracts The correlations of the total saponin content of the extracts with the independent variables, temperature, time, and ratio, as determined using the t ratio and the P value (Table 3), were also significant (P < 0.05) for 2 of the 3 1st-order linear effects (A was inversely related and C was directly related but B was not related) and 2 of the 3 2nd-order quadratic effects (A2 and C2 were inversely related but B2 was not related). However, the correlations were not significant for the interactive effects (AB, AC, and BC). Based on their regression coefficients, the independent variables with significant effects could also be listed from the largest to the smallest effect on the total saponin content of the extract (Table 3) as follows: linear ratio (C), quadratic ratio (C2 ), quadratic temperature (A2 ), and linear temperature (A).Therefore, the ratio (C) had a bigger influence on the total saponin content of the extracts than temperature (A) and time (B), with the latter having no effect (Table 3). The influence of the 3 independent variables, temperature, time, and ratio on the total saponin content of the extracts as obtained from the predictive equation is presented in the 3D surface and 2D contour plots of Figure 2(A–C). Figure 2(A and B) shows that the total saponin content is predicted to change significantly in the range of temperatures from 20 to 60 °C; the saponin content is predicted to increase and reach a peak at 40 °C but then dramatically decrease when the temperature exceeds 40 °C. The latter finding indicates that the saponins extracted from bitter melon are sensitive to heat and tend to degrade when exposed to temperatures higher than 40 °C. This finding is consistent with a recent Vol. 79, Nr. 7, 2014 r Journal of Food Science E1375
E: Food Engineering & Physical Properties
15 to 16 h in the dark at RT. Then, 1 mL of the working solution was mixed with 60 mL of methanol to obtain an absorbance of 1.1 ± 0.02 units at 734 nm using a spectrophotometer. Fresh working solution was prepared for each assay. For the powders, different concentrations were prepared in deionized water (0.25, 0.5, 0.75, and 1 mg/mL) and 0.15 mL was mixed with 2.85 mL of the working solution and incubated for 2 h in the dark at RT. The absorption of the solution at 734 nm was measured using a spectrophotometer. Trolox was used as a standard and results were expressed as trolox equivalents (TE) per gram of powder (μmole TE/g). DPPH assay. The DPPH assay was as described by BrandWilliams and others (1995) with some modifications. A stock solution of 0.6 M DPPH in methanol was prepared and kept at –20 °C until use. The working solution was prepared by mixing 10 mL of stock solution with 45 mL of methanol to obtain an absorbance of 1.1 ± 0.02 units at 515 nm using a spectrophotometer. For the powders, different concentrations were prepared in deionized water (0.25, 0.5, 0.75, and 1 mg/mL) and 0.15 mL was mixed with 2.85 mL of working solution. The sample was allowed to stand for 30 min and the absorption at 515 nm was then measured using a spectrophotometer. Trolox was used as a standard and results were expressed as TE per gram of powder (μmole TE/g). FRAP assay. The FRAP assay was as described by Benzie and Strain (1996) with some modifications. The stock solutions of 300 mM acetate buffer (pH 3.6), 10 mM TPTZ in 40 mMHCl, and 20 mM ferric (III) chloride hexahydrate in deionized water were prepared and kept at 4 °C until use. The fresh working solution was prepared by mixing 100 mL of acetate buffer, 10 mL of TPTZ, and 10 mL of ferric (III) chloride hexahydrate in a ratio of 10:1:1. The working solution was then incubated at 37 °C before use. For the powders, different concentrations were prepared in deionized water (0.25, 0.5, 0.75, and 1 mg/mL) and 0.15 mL was mixed with 2.85 mL of the working solution for 30 min in the dark at RT. The absorption of the solution at 593 nm was measured using a spectrophotometer against a reagent blank. Trolox was used as a standard and results were expressed as TE per gram of powder (μmole TE/g). Statistical analyses. The adequacy of the RSM second-order polynomial model was determined based on the lack of fit and the coefficient of determination (R2 ). The three-dimensional (3D) surface and two-dimensional (2D) contour plots of the total saponin content were drawn using the JMP software version 10.0. Results were expressed as mean values with standard deviations and significant differences between treatments were tested using analysis of variance and the Bonferroni posthoc test with a 5% significance level (P < 0.05). Correlations among data were calculated using Pearson’s correlation coefficient and expressed as R2 . The SPSS software version 19.0 statistical package was used for all analyses.
Extraction of bitter melon saponins . . . Table 2–The experimental and predicted values for the total saponin content of the aqueous extracts measured or predicted for the 15 run Box–Behnken design. Total saponin content(mg AE/g) Run
E: Food Engineering & Physical Properties
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Pattern
Temperature(˚C)
Time(min)
Ratio(mL/g)
Experimental
Predicted
000 ++0 +0− −−0 +−0 000 −0− 0−− 0+− −+0 0++ 0−+ +0+ 000 −0+
40 60 60 20 60 40 20 40 40 20 40 40 60 40 20
15 25 15 5 5 15 15 5 25 25 25 5 15 15 15
55 55 10 55 55 55 10 10 10 55 100 100 100 55 100
51.6 43.8 32.8 44.9 44.2 48.5 42.7 34.5 42.4 48.0 49.2 47.7 47.7 50.2 49.5
50.1 43.8 33.9 44.9 42.5 50.1 42.1 35.1 41.4 49.7 48.6 48.7 48.4 50.1 48.4
Table 3–The regression coefficients and their significance values generated from the fitted 2nd order equation for the Box–Behnken design. Total saponin content (mg AE/g) Regression coefficienta
Regression coefficient values
t ratio
P value
50.1∗
47.71
