2015 Volume 21 Issue 4 Pages 615-621
The present study evaluated the optimal parameters for Psidium guajava leaf extraction by ultrasound to obtain maximal anti-hyperglycemic activities in the extract. The response surface methodology was employed for empirical model building. The maximum inhibition of the two starch-digesting enzymes α-amylase and α-glucosidase was determined. The optimized extraction parameters of solvent-to-solid ratio, extraction time, and extraction temperature for obtaining the maximum inhibitory activity for 50 µg/mL extract were determined to be the following: 12.07 (v/w), 5.22 min, and 59.77°C for α-amylase (47.23% ± 0.76%) and 12 (v/w), 5.22 min, and 59.94°C for α-glucosidase (59.42% ± 0.65%), respectively. The F values for the lack-of-fit were not significant (p > 0.05) for both second-order models, indicating them to be appropriate for describing the response surface. In conclusion, our study presents the optimal conditions for ultrasound-assisted extraction of guava leaf as a potential substance for use in diabetes treatment.
Psidium guajava, commonly known as guava, is a member of the Myrtaceae family. It is an evergreen tree that grows wild in the Torrid Zone and in the subtropics and has been reported to be useful in the treatment of diabetes (Gutierrez et al., 2008; Keji, 1981). Past in vitro studies have confirmed the inhibitory effects of the aqueous extract of guava leaves on α-amylase and α-glucosidase activities, which are starch-digesting enzymes (Karthic et al., 2008; Wang et al., 2007). In a past in vivo study, the extract of guava leaves was found to inhibit the increase in blood sugar level in alloxan-induced diabetic rats (Maruyama et al., 1985). In a human study, oral administration of tea prepared from guava leaves significantly suppressed the increase in the postprandial blood glucose level (Deguchi et al., 1998).
Diabetes mellitus is a prevalent metabolic disorder affecting the global population, and it is estimated to be prevalent in 552 million individuals by the year 2030. In particular, non-insulin dependent diabetes mellitus accounts for more than 90% of diagnosed cases of diabetes. Non-insulin-dependent diabetes is expected to become an even more serious concern considering the changes in the life style of individuals, especially of those living in developed countries (Clark et al., 1987). Thus, the cure and prevention of non-insulin-dependent diabetes mellitus is important. In hyperglycemia, there is a rapid increase in the blood glucose level after eating, which is also the typical symptom of diabetes, especially of non-insulin-dependent diabetes mellitus (Deshpande et al., 2009). Retarding glucose digestion via the inhibition of carbohydrate-hydrolyzing enzymes decreases postprandial hyperglycemia, which may be the key strategy for controlling diabetes mellitus (Hirsh et al., 1997).
Methanolic extract of Tectona grandis flowers can inhibit the activities of α-amylase and α-glucosidase in vitro. In addition, application of this extract to type 2 diabetic rats significantly reduced their blood glucose level. The active polyphenols present in this extract impart a blood glucose-lowering effect via insulin sensitization as well as inhibition of α-amylase and α-glucosidase activities (Ramachandran and Rajasekaran, 2014). Commercial phaseolamin, which has high α-amylase inhibitory ability, can reduce the blood glucose levels in streptozotocin-induced diabetic rats; therefore, it possesses therapeutic potential for the treatment or prevention of the complications of diabetes (De Gouveia et al., 2014). According to these results, substances with the ability to inhibit the activity of α-amylase and/or α-glucosidase may also possess anti-hyperglycemic potential via alleviation of the high blood glucose status in diabetic individuals.
Several conventional methods such as heat reflux, soxhlet extraction, and maceration are commonly used for extraction purposes. Although these methods can effectively extract the active constituents, they are time-consuming, highly energy-consuming, and require large quantities of toxic organic solvent. Therefore, several environment-friendly extraction technologies have been developed, of which ultrasound is a commonly used method. Ultrasound-assisted extraction is an inexpensive, simple, and efficient extraction technique with several advantages, including high reproducibility, lesser time of operation, less solvent consumption, and low energy input. In recent years, ultrasound has been widely used in extractive procedures (Ghafoor et al., 2011; Jiao and Zuo, 2009; Wang and Zuo, 2011). Ultrasound-assisted extraction of guava leaves for its antioxidant activities has been evaluated in the past to indicate that ultrasonication is the most suitable method for guava leaf extraction, considering that it yielded extracts with significantly high total phenolic content and antioxidant activities as compared to those of extracts obtained by hot water extraction or soxhlet extraction (Nantitanon et al., 2010). However, no study has described the use of ultrasound-assisted procedures for guava leaf extraction for the purpose of evaluating the anti-hyperglycemic activity of the extract.
The optimal extraction conditions for the maximal extraction of total phenolic content from guava leaves have been established in our previous study. In addition, the chromatographic profile and the concentration of 7 major bioactive components in extracts with anti-hyperglycemic activities, including gallic acid (0.87%), chlorogenic acid (0.62%), catechin (2.25%), caffeic acid (0.11%), epicatechin (1.45%), epigallocatechin gallate (0.47%), and quercetin (0.83%), were analyzed by HPLC (Liu et al., 2014; 2015). The aim of the present study was to establish the optimal extraction conditions for guava leaves via ultrasound assistance to obtain maximal anti-hyperglycemic activity in the extract. The anti-hyperglycemic activity was estimated by determining the ability of the extract to inhibit α-amylase and α-glucosidase. The results of the present study provide important information for promoting the utilization of agricultural residues by maximizing their bioactivities and for establishing an environment-friendly extraction procedure.
Materials and chemicals Psidium guajava leaves of Jen Ju Pa (Myrtaceae family) were harvested in the Jing-cin Farm (Tianzhong Township, Changhua County, Taiwan) at the stage of appearance of flower buds. The plant materials were taxonomically identified and the data were deposited at the Fengshan Tropical Horticultural Experiment Branch, Taiwan Agricultural Research Institute Council of Agriculture, Executive Yuan (FTHA000282). All chemicals used in this study were purchased from Sigma-Aldrich (St. Louis, MO, USA).
Guava leaf extraction The guava leaves were washed, minced, immersed in distilled water, and then extracted by using an ultrasound probe (Branson, Danbury, CT, USA), and the ultrasonic power was fixed at 1100 W. Distilled water was used as the only extraction solvent in this study. Different experimental parameters were considered to optimize the extraction process, such as solvent-to-solid ratio, extraction temperature, and extraction time. After extraction, the liquid extract was filtered through Whatman No. 1 filter paper and concentrated using a rotary vacuum evaporator (Eyela N-N series; Tokyo Rikakikai, Tokyo, Japan), freeze dried (Eyela FD-5N; Tokyo Rikakikai, Tokyo, Japan), and finally crushed into a powder form. The powder was stored at −80°C for further use in the in vitro anti-hyperglycemic activity assay. For all experiments performed in this study, individual extraction was performed using 800 mL of solvent.
Optimization of the experimental design A five-level, three-variable central composite rotatable design (CCRD-RSM) was used to determine the optimum combinations of extraction conditions for the inhibition of α-amylase and α-glucosidase activities by guava leaves (Myers, 2002). Different ranges of the three independent variables including solvent-to-solid ratio (X1), extraction temperature (X2), and extraction time (X3) were determined by the single factor test. Based on the results of our previous study (Liu et al., 2014), the parameters that gave the highest total phenols, including solvent-to-solid ratio (12 v/w), extraction time (5 min), and extraction temperature (60°C), were used to optimize the experimental design in the present study. Eight factorial points were used, including six axial points and five central points, resulting in a total of 19 experimental runs (Table 1).
| Treatment#a | Coded level of variable | ||||
|---|---|---|---|---|---|
| v/w ratio |
Temperature (°C) |
Time (min) | α-Amylase activity inhibition (%) |
α-Glucosidase activity inhibition (%) |
|
| 1 | 10(−1) | 50(−1) | 4(−1) | 39.12 | 51.42 |
| 2 | 10(−1) | 50(−1) | 6(1) | 41.30 | 53.05 |
| 3 | 10(−1) | 70(1) | 4(−1) | 39.69 | 52.02 |
| 4 | 10(−1) | 70(1) | 6(1) | 42.69 | 53.17 |
| 5 | 14(1) | 50(−1) | 4(−1) | 39.98 | 50.76 |
| 6 | 14(1) | 50(−1) | 6(1) | 40.45 | 52.42 |
| 7 | 14(1) | 70(1) | 4(−1) | 39.48 | 50.38 |
| 8 | 14(1) | 70(1) | 6(1) | 40.64 | 52.63 |
| 9 | 15.4(1.68) | 60(0) | 5(0) | 40.49 | 51.24 |
| 10 | 8.6(−1.68) | 60(0) | 5(0) | 37.01 | 49.43 |
| 11 | 12(0) | 76.8(1.68) | 5(0) | 37.36 | 49.85 |
| 12 | 12(0) | 43.2(−1.68) | 5(0) | 39.63 | 50.47 |
| 13 | 12(0) | 60(0) | 6.7(1.68) | 42.40 | 54.09 |
| 14 | 12(0) | 60(0) | 3.3(−1.68) | 40.33 | 50.69 |
| 15 | 12(0) | 60(0) | 5(0) | 45.63 | 58.40 |
| 16 | 12(0) | 60(0) | 5(0) | 46.16 | 58.63 |
| 17 | 12(0) | 60(0) | 5(0) | 47.16 | 59.51 |
| 18 | 12(0) | 60(0) | 5(0) | 48.20 | 59.92 |
| 19 | 12(0) | 60(0) | 5(0) | 47.63 | 59.10 |
v/w ratio = solvent-to-solid ratio; Temperature = extraction temperature; Time = extraction time.
The second-order polynomial model used in the response surface analysis was as follows:
 |
Where, Y represents the response function; bi are the regression coefficients of the linear terms; bik are the regression coefficients of the interactive terms; bii are the regression coefficients of quadratic terms; and Xi represent the coded independent variables.
The results of the experimental design and data were analyzed by statistical software package (SAS, 2002). The fit of the models was evaluated by the determination coefficients (R2).
The optimal extraction conditions, including v/w ratio, extraction temperature, and extraction time, for maximizing the anti-hyperglycemic activities were estimated through three-dimensional (3D) response surface analysis of independent and response variables. The inhibited levels of α-amylase and α-glucosidase activities under the optimum conditions were determined for five dependent extracts. The results were compared with the predicted value.
Evaluation of anti-hyperglycemic activity in vitroInhibition of α-amylase The assay was performed according to a previously described method (Ali et al., 2006) with some modifications. One-hundred microliters of 0.02 M phosphate-buffered saline (PBS; pH 6.9) containing porcine pancreatic α-amylase solution (5 unit/mL) was mixed with an equal volume of guava leaf extract (GvEx) and pre-incubated at 37°C for 10 min. Then, 800 µL of 1% starch solution in 0.02 M PBS (pH 6.9) was added to the mixture. The reaction mixture was incubated at 37°C for 3 min, and the reaction was terminated by the addition of 200 µL of 5% dinitrosalicylic acid reagent and heating at 90°C for 10 min, followed by cooling to room temperature. The absorbance of the mixture was measured at 540 nm by us ing a spectrophotometer (Agilent Varian Cary 50; Santa Clara, CA, USA). In the control group, 100 µL of the buffer solution was used instead of the sample solution. The percent inhibitory activity was calculated as follows: % inhibition = [(AControl − ASample)/(AControl)] × 100.
Inhibition of α-glucosidase α-Glucosidase from Saccharomyces cerevisiae was used for the inhibitory assay, which was performed as described by Shim et al. (Shim et al., 2003) with some modification. Briefly, enzyme and substrate solutions were prepared from α-glucosidase and p-nitrophenyl-α-d-glucopyranoside dissolved in 0.1 M PBS (pH 6.8) to obtain a final concentration of 1 unit/mL and 0.53 mM, respectively. Then, a mixture containing 20 µL of the enzyme solution, 100 µL of the GvEx solution at various concentrations in 0.1 M PBS (pH 6.8), and 380 µL of the substrate solution were mixed and incubated at 37°C for 20 min. The GvEx solution was replaced with 0.1 M PBS in the control group. After incubation, the reaction was terminated by the addition of 500 µL of 0.1 M sodium carbonate, and the absorbance was recorded at 400 nm. The percent inhibitory activity was calculated by the following formula: % inhibition = [(AControl − ASample)/(AControl)] × 100.
Statistical analysis The response surface regression procedure of the statistical software package was used to analyze the experimental data (SAS, 2002). Statistical analysis of the model was performed by analysis of variance (ANOVA). p < 0.05 was considered statistically significant.
Optimization of parameters by RSM The experimental results of the inhibitory effect of GvEx on α-amylase and α-glucosidase activities are shown in Table 1. The values ranged from the maximum of 48.2% and 59.92% (treatment #18) to the minimum of 37.01% and 49.43% (treatment #10) for α-amylase and α-glucosidase, respectively. The experimental data was analyzed by multiple regression, and the response and test variables related to the second-order polynomial equations are presented in Table 2. The qualities of fit of the models were expressed by the R2 correlation coefficient. Calculated models were used to explain 91.6% and 95.1% of the results with respect to α-amylase and α-glucosidase activities, respectively, suggesting a close agreement between the observed and predicted values (Vuataz, 1986). In addition, coefficients of variation were found to be 1.4% for α-amylase and 1.1% for α-glucosidase, indicating the reproducibility of the models. The results of ANOVA for quadratic models fitted to the response surface are shown in Table 3. The results were found to be significant (α-amylase, p = 0.001; α-glucosidase, p < 0.001), which attested to the goodness of the fit of both the models. F values, which indicate lack-of-fit, were insignificant in all cases (α-amylase, p = 0.206; α-glucosidase, p = 0.069), indicating that the second-order model was appropriate for describing the response surface (Myers, 2002).
| Response | Model equations | R2 |
|---|---|---|
| α-Amylase activity inhibition (%) | Y= 46. 92 + 0.452X1 − 0.266X2 + 1.274X3 − 7.475X12 − 7.714X22 − 4.86X32 − 0.81X1X2 + 0.539X2X3 − 1.282X1X3 | 0. 916* |
| α-Glucosidase activityinhibition (%) | Y= 59. 07 − 0.048X1 − 0.061X2 + 1.538X3 − 7.946X12 − 8.102X22 − 5.891X32 − 0.318X1X2 + 0.039X2X3 + 0.408X1X3 | 0. 951* |
X1 = solvent-to-solid ratio; X2 = extraction temperature; X3 = extraction time.
| α-Amylase | |||||
| Source | Degree of freedom | Sum of squares | Means of squares | F-value | Prob > F |
| Model | 9 | 195.56 | 21.73 | 10.94 | 0.001 |
| Linear | 3 | 9.06 | 3.02 | 1.52 | 0.275 |
| Quadratic | 3 | 183.99 | 61.33 | 30.89 | <0.001 |
| Interaction | 3 | 2.50 | 0.84 | 0.42 | 0.743 |
| Total error | 9 | 17.87 | 1.99 | ||
| Lack of fit | 5 | 13.44 | 2.69 | 2.42 | 0.206 |
| Pure error | 4 | 4.44 | 1.11 | ||
| α-Glucosidase | |||||
| Source | Degree of freedom | Sum of squares | Means of squares | F-value | Prob > F |
| Model | 9 | 225.88 | 25.09 | 19.59 | <0.001 |
| Linear | 3 | 11.31 | 3.77 | 2.94 | 0.091 |
| Quadratic | 3 | 214.31 | 71.44 | 55.76 | <0.001 |
| Interaction | 3 | 0.26 | 0.09 | 0.07 | 0.976 |
| Total error | 9 | 11.53 | 1.28 | ||
| Lack of fit | 5 | 9.98 | 1.99 | 5.15 | 0.069 |
| Pure error | 4 | 1.55 | 0.39 | ||
The coefficients of the regression model were used to estimate the optimization parameter of three quadratic terms (X12, X22, and X32) that significantly affected the 3D response surface plots of enzyme inhibition (p < 0.05). 3D response surface plots of α-amylase and α-glucosidase inhibition are presented in Fig. 1 and Fig. 2, respectively, which represent the interactions between two factors, with the third factor taken as a constant at its middle level. Fig. 1A shows the effect of v/w ratio and the extraction temperature on α-amylase inhibition, with the maximum inhibition being observed at a v/w ratio in the range of 11 to 13 and at an optimal temperature of approximately 60°C. Fig. 1B shows the effect of v/w ratio and the extraction time on α-amylase inhibition. The maximum inhibition was achieved with an extraction time of 4.5 – 5.5 min and a v/w ratio of 11 – 13. Fig. 1C shows the effect of extraction temperature and time on α-amylase inhibition; under the optimal conditions, maximum inhibition was obtained at an extraction temperature of 55°C – 65°C and an extraction time of 4.5 – 5.5 min. The effect of v/w ratio and extraction temperature on α-glucosidase inhibition is shown in Fig. 2A. The extraction conditions of a v/w ratio in the range of 11 – 13 and a temperature of 55°C – 65°C achieved the maximum inhibition. The effect of extraction time and v/w ratio on α-glucosidase inhibition is shown in Fig. 2B; under optimal conditions, maximum inhibition was obtained with a v/w ratio of 11 – 13 and an extraction time of 4.5 – 5.5 min. Fig. 2C shows the effect of extraction temperature and time on α-glucosidase inhibition; under optimal conditions, maximum inhibition was obtained at an extraction temperature of 55°C – 65°C and an extraction time of 4.5 – 5.5 min. The optimal conditions for maximum inhibition of α-amylase and α-glucosidase activities were as follows: v/w ratio of 12.07 and 12, extraction temperature of 59.77°C and 59.94°C, and extraction time of 5.22 min and 5.22 min, respectively. The maximum predicted values of enzyme inhibition were 47.01% for α-amylase and 59.17% for α-glucosidase, under the optimal conditions (Table 4). As described by Deguchi et al. (Deguchi et al., 1998), the concentrations of hot water extract of dried guava leaves required to inhibit 50% of pancreatic α-amylase, intestinal maltase, and sucrase activities were 600, 2100, and 3600 µg/mL, respectively. In another study, 1500 µg/mL extracts of dried guava leaves extracted using ethanol, dichloromethane, ethyl acetate, n-butanol, or water inhibited 31.7%, 9.3%, 43.9%, 54.4%, and 29.3% of pancreatic α-amylase activity, respectively. In addition, these extracts also inhibited 38.3%, 18.3%, 46.3%, 63.5%, and 34.5% of sucrase activity and 33.4%, 15.2%, 40.6%, 47.7%, and 27.3% of maltase activity, respectively (Wang et al., 2010). Thus, our results suggest that GvEx obtained from the defined optimum conditions possessed excellent anti-hyperglycemic activities.
Response surface plots showing the effects on α-amylase activity inhibition, including (A) solvent-to-solid (v/w) ratio and extraction temperature, (B) v/w ratio and extraction time, and (C) extraction temperature and time.
Response surface plots showing the effects on α-glucosidase activity inhibition, including (A) solvent-to-solid (v/w) and extraction temperature, (B) v/w ratio and extraction time, and (C) extraction temperature and time.
| Response variables | Optimal conditions | Values | |||
|---|---|---|---|---|---|
| v/w | Temperature (°C) |
Time (min) | Experimentala | Predicted | |
| α-Amylase activity inhibition (%) | 12.07 | 59.77 | 5.22 | 47.23 ± 0.76 | 47.01b |
| α-Glucosidase activity inhibition (%) | 12.00 | 59.94 | 5.22 | 59.42 ± 0.65 | 59.17c |
v/w = solvent-to-solid ratio; Temperature = extraction temperature; Time = extraction time.
Our previous study illustrated the effect of extraction parameters on the extraction of total phenols from guava leaves (Liu et al., 2014). In the present study, we clarified the influence of extraction parameters on anti-hyperglycemic activities of guava leaves. In addition, we found highly positive correlation between total phenolic content and anti-hyperglycemic activities (for α-amylase, R2 = 0.98, p < 0.001, Pearson's correlation; for α-glucosidase, R2 = 0.99, p < 0.001, Pearson's correlation), suggesting the direct influence of total phenolic content from guava leaves on anti-hyperglycemic activities. In addition, the yield of the extracts, total soluble solids in guava leaf aqueous extract after freeze-drying, obtained by the optimal extraction conditions for α-amylase and α-glucosidase inhibition were 2.37% and 2.41% of the original weight of guava leaves, respectively. These freeze-dried powders contained 260 mg and 262 mg gallic acid equivalents/g phenolic compounds.
Verification of the model The verification experiment was performed under the optimal conditions in five dependent extractions. The mean inhibited values of α-amylase and α-glucosidase activities were 47.23% and 59.42%, respectively. The values from optimal extraction conditions were achieved within a 95% confidence interval of experimental values, demonstrating the reliability of the RSM models (Table 4).
The optimum extraction conditions that resulted in obtaining maximum anti-hyperglycemic activities from guava leaves were investigated by using RSM models. The second-order polynomial models suitably described the experimental data. Extraction temperature, time, and v/w ratio showed significant quadratic effects on the inhibition of α-amylase and α-glucosidase activities. For α-amylase inhibition, the suggested optimal extraction conditions were v/w ratio of 12.07, extraction temperature of 59.7°C, and extraction time of 5.22 min. For α-glucosidase inhibition, the recommended extraction conditions were v/w ratio of 12, extraction temperature of 59.94°C, and extraction time of 5.22 min. The predicted values of enzyme inhibition were highly fitted with the experimental data (95% confidence interval), indicating the success of extraction quadratic models. In conclusion, the establishment of optimal extraction conditions to obtain maximum anti-hyperglycemic activities from guava leaves will promote the utilization of guava leaves in diabetes treatment. In addition, the optimization study offers an additional choice for the prevention and control of hyperglycemia and/or non-insulin-dependent diabetes mellitus. On similar lines, in vivo studies on the effects of GvEx on the inhibition of carbohydrate-hydrolyzing enzymes and the maintenance of postprandial glucose levels are underway.
Acknowledgements The authors are grateful for the financial support for this research by the Tunghai University of Taiwan R.O.C., under the project of “Global Research and Education on Environment and Society (GREEnS)” and Grand No. GREEnS 4-3.
