Optimization of harvest and extraction factors by full factorial design for the improved yield of C-glucosyl xanthone mangiferin from Swertia chirata

Swertia chirata Buch.-Ham. ex C.B. Clarke is an important medicinal plant used in various herbal formulations as it shows significant biological activities such as hepatoprotective, hypoglycemic, anti-inflammatory, antimalarial, antioxidant and anti-parkinson. C-glucosyl xanthone glycoside (mangiferin) is known as bio-marker compound of genus Swertia L. Development of efficient extraction methods of C-glucosyl xanthone mangiferin from Swertia chirata was attempted by optimizing the pre-harvest, post-harvest and extraction techniques by full factorial design. Firstly, a full factorial design was implemented to evaluate the single and interactive effects of pre-harvest (growth stage and plant part), post-harvest (drying condition and storage periods) followed by selection of best extraction technique such as heat reflux extraction (HRE), microwave assisted extraction (MAE) and ultrasound assistant extraction (UAE) at different solvent types on mangiferin yield. HPTLC and HPLC techniques were used for the determination of mangiferin content in extracts generated from different plant samples. In addition, anti-oxidant and anti-diabetic properties were determined by using DPPH assay and percentage inhibition of α‑amylase enzyme. Substantial variation of mangiferin yield, ranged from 1.46 to 4.86% was observed, depending on the growth stage, plant part, drying condition, storage periods and extraction method. Results showed that drying of the leaves of Swertia chirata in the shade harvested at budding stage and stored for not more than 1 month was recommended for obtaining a higher mangiferin yield. Among different extraction techniques, MAE and UAE in 50% aqueous ethanol solvent were found to be efficient and cost-effective with better yield of mangiferin (4.82% and 4.86%, respectively) as compared to HRE (4.14%). Highest DPPH activity and percentage inhibition of α‑amylase was observed in the aqueous ethanol extract of S. chirata leaves harvested at bud-stage of plant followed by flowering stage. The study shows that optimization of various factors by full factorial design was found to be an effective procedure to improve mangiferin yield from Swertia chirata and can be used for extraction of mangiferin.

www.nature.com/scientificreports/ with tap water, cut into small pieces, shade dried and powdered separately in an electric grinder. High performance thin layer chromatography (HPTLC) method was to screen out the potent plant part of S. chirata containing mangiferin.
To study the effect of drying, three drying conditions i.e. sun drying, shade-drying and oven-drying were investigated. 50 g of fresh sample was spread over 1 m 2 of the white sheet and kept under varying drying condition i.e. Hot air oven-drying (24 h at 45 °C), sun-drying (5 days at 35-45 °C) and shade drying (5 days at 30 °C).
To study the effect of storage periods, dried samples of S. chirata were kept in polypropylene containers in dark conditions at room temperature for 1 and 6 months, respectively (Table 1).
All analyses are the mean of triplicate measurements ± standard deviation. The results were analyzed by one-way ANOVA followed by Duncan's Multiple Range Test (SPSS 16 was used for DMRT and Minitab 18 was used for full factorial design). Values with different superscript alphabet (a-d) within the same column are significantly different at p < 0.05.
Preliminary experiments for screening of solvent types and efficacy of extraction techniques. For better solubility and extraction of mangiferin compound, suitable solvents such as chloroform, ethyl acetate, methanol, aqueous methanol (25-75% v/v), ethanol, aqueous ethanol (25-75% v/v) and water was screened with heat reflux extraction method. Aliquots (1 g) powdered leaves were heated at 75 °C temperature with 50 mL of different solvents for 1 h, according to Chavan et al. 27 , with slight modifications. All three extraction methods viz. HRE (75 °C for 1 h), MAE (450 W for 2 min.) and UAE (200 W, 40 KH, 40 °C) used in the present work are summarized in Table 2.
All the extracts were filtered through Whatman No.1 filter paper, centrifuged for 4 min at 4 °C and supernatant was concentrated in Rota evaporator (Buchi model R-205). The residue (10 mg) was re-dissolved in 10 mL methanol and kept at 4 °C for further HPTLC and HPLC analysis. The stock solution of mangiferin reference compound was prepared in methanol solvent (10 mg per 100 mL) and then stock solution of the standard was stored at 4 °C. of Swertia chirata were conducted in two steps, firstly a triplicate completely randomized 3 × 5 × 2 full factorial design arrangement of preharvest condition evaluating (rosette, vegetative, bud, flower, fruit stage) at three drying method (shade, sun and oven) and two storage periods (first and sixth month) a total of 30 sets were created and implemented in a random order. Secondly, 3 × 5 factorial design of extraction techniques evaluating three extraction methods (HRE, MAE and UAE) at five different levels of aqueous ethanol (0, 25%, 50%, 75%, 100%) in 15 extraction sets were randomly generated. The mangiferin yield data were examined by analysis of variance (ANOVA) using the general linear model (GLM) approach of the MINITAB 18.0 software program (Minitab Inc, State College, PA, USA). The single and interactive effects of extraction factors on mangiferin yield were determined to examine the significance of differences at p < 0.05. The mean mangiferin yield was compared by Duncan's multiple range tests. Graphics (main effect plots and interaction plots) were created with Minitab 16 software.
Isolation of compound (flash chromatography, IR and NMR). Mangiferin isolation was performed on a Biotage One Flash Chromatography system (Sweden), with UV detection and an automatic collector. For isolation of bio-active compound, the crude ethanolic extract (1 mg mL −1 ) prepared from the leaves of Swertia chirata by MAE was used. The Flash Chromatography conditions were as follows: 40 g flash column packed with 60-120 mesh sized silica gel; elution system: ethyl acetate/methanol gradient; wavelength: 254 and 280 nm and flow rate: 20 drop min −1 . Before binding experiment, ethyl acetate was used to equilibrate the column former to isolation. Mangiferin was isolated in 40 min gradient program of solvent A (ethyl acetate) and solvent B (methanol). Out of seven fractions, mangiferin as the major peak was separated as first fraction (FA) after residue at wavelength 254 nm.
Structure of mangiferin compound in fractions was elucidated based on their spectral data (IR and 13 C NMR). Purity of mangiferin compound was confirmed by comparison with reference compound (procured from Sigma Aldrich, USA) using HPTLC (high performance thin-layer chromatography) method.

Quantification of marker compound by HPTLC and HPLC. HPTLC analysis. The HPTLC system
was CAMAG (Muttenz, Switzerland) having Linomat-5 automatic sample applicator furnished with a 100 μL Hamilton syringe (fixed 100 nL s −1 delivery rate). For analysis, a twin-trough glass tank (CAMAG) and UV cabinet was used. Chromatography was performed on stationary phase composed of 20 cm × 10 cm pre-coated silica gel 60 F 254 HPTLC plates (with 0.25 mm thickness). Samples were administered to the plates as 5 mm wide bands with Hamilton syringe. 3 µL plant samples were loaded on chromatographic plate. The pre-coated silica gel 60 F 254 TLC aluminium plates was developed in 20 cm × 10 cm twin trough glass chamber saturated (20 min pre-saturation) with mobile phase ethyl acetate-glacial acetic acid-formic acid-water 100:11:11:26 (v/v/v/v) up to distance 7.5 cm at constant temperature (25 °C ± 2 °C) and relative humidity (40 ± 2%). The developed plate was dried by hot air to evaporate the solvents from the plate; bands were visualized and photographed with CAMAG visualizer under UV 254 light. Scanning was completed with CAMAG TLC scanner-3 provided with CATS software (version: 1.4.4.6337) at λ = 254 nm. The determination of mangiferin in Swertia chirata samples were carried out by using high-performance thin layer chromatography (HPTLC) by following the method developed by Pandey et al. 28 , with slight modifications.
HPLC analysis. Analysis was performed with HPLC-PDA detector System (Waters) equipped with an auto sampler, a dual low-pressure gradient system, C-18 column. HPLC conditions were optimized with isocratic elution; 9 min time run and 1 mL min −1 flow rate. Working solutions of the samples were prepared at 100 ppm concentrations, and injection volume was set at 10 µL. Before injection, column was equilibrated for 10 min.
Chromatographic separation was carried out using Sunfire C-18 column (4.6 mm × 250 mm; 5 µm particle size). The mobile phase was composed of acetonitrile (A) and 0.1% formic acid in water (B) using the following gradient program: 20-46% A (0-1. Method validation was performed on the parameters such as linearity, limit of sensitivity, specificity, precision, accuracy, recovery and robustness. Anti-oxidant activity assay. 1,1-Diphenyl-2-picrylhydrazyl (DPPH) assay. The DPPH assay was used in order to evaluate the free radical scavenging activity of leaf parts of S. chirata harvested at bud-and flower-stages of plant. Extracts of these plant samples were prepared as aqueous, methanolic, ethanolic and aqueous ethanolic (25-75% v/v) extracts. The DPPH radical scavenging activity was determined using the method reported by Roy et al. 29 . DPPH was dissolved in methanol at a concentration of 50 μM. The DPPH solution (2.0 mL) was mixed with 2.0 mL of various concentrations (10,20,40, and 80 μg mL −1 ) of S. chirata extracts and immediately absorbance was measured at 517 nm using a spectrophotometer (UV2550, Shimadzu, Japan). Samples were incubated in a dark room for 16 min at 27 °C. After incubation, decrease in absorbance was again measured for all samples. Ascorbic acid was employed as a positive reference.
The scavenging activity was calculated using Eq. (1): www.nature.com/scientificreports/ where Ae and Ac are absorbance of extract and control, respectively. All experiments were performed using three replicates.

Anti-diabetic assay (in-vitro).
In-vitro anti-diabetic activity of leaf parts of S. chirata samples harvested at bud-and flower-stages of plant was evaluated. Extracts of all plant samples were prepared by using different solvents viz. aqueous, methanolic, ethanolic and aqueous ethanolic (25-75% v/v). In-vitro anti-diabetic activity was assayed by means of method Roy et al. 29 . In this, inhibitory activity of α-amylase enzyme was determined to check in vitro anti-diabetic assay which involves the breakdown of starch into glucose. 1.0 mL of each S. chirata extracts (100 μg mL −1 of aqueous, methanolic, ethanolic and 25-75% aqueous ethanolic) were checked individually and into each test tube 1.0 mL of α-amylase enzyme (Sigma, St. Louis, USA) was added and incubated at 37 °C for 10 min. After pre-incubation, 1.0 mL of 1% starch solution was added into each test tube. The reaction mixtures were again incubated at 37 °C for 15 min. Then the reaction was stopped with 2.0 mL of 3,5-dinitrosalicylic acid (Sigma, St. Louis, USA) color reagent. The test tubes were then incubated in a boiling water bath for 5 min and then allow it to cool at room temperature, the absorbance was measured at 546 nm using a spectrophotometer. The control (buffer in place of sample extract) represents the 100% of enzyme activity. The % age inhibition of α-amylase enzyme activity was calculated by Eq. (2): All experiments were performed in triplicates and data were the mean ± SD. Statistical differences among groups were analyzed by one-way analysis of variance (ANOVA) followed by Duncan's multiple range test (DMRT) by using SPSS 16. Groups were considered statistically significant at the significance level of p < 0.05. Ethics approval. The plant material for this study was collected from a nursery commercially which comply with relevant institutional, national, and international guidelines and legislation.

Results
In this study, the effects of plant parts, growth stage, drying method, storage periods, solvent type and the extraction method on the mangiferin content were investigated. Quantitative estimation of mangiferin in different extracts generated from different batches of Swertia chirata were identified by HPTLC fingerprinting. Although the HPTLC fingerprinting showed that mangiferin was present only in leaves part of the plant, it was decided to consider only the leaf part while excluding the stem and root parts (Fig. 1). In addition, HPLC-PDA analysis was used to confirm the HPTLC results.
In addition, flash chromatography was used to isolate the mangiferin compound from Swertia chirata leaf extract and out of seven fractions; mangiferin as the major peak was isolated as first fraction (FA) after residue at wavelength 254 nm. Fraction A containing mangiferin was confirmed and characterized for purity through IR and 13 C NMR spectra (Fig. 2) and comparison with reference compound by using HPTLC (High performance thin-layer chromatography) and UV (Fig. 3).
Analytical characteristics of method validation parameters for mangiferin compound are presented in Table 3.
(2) % age inhibition of α−amylase = Enzyme activity of control− Enzyme activity of extract Enzyme activity of control ×100   www.nature.com/scientificreports/ The ANOVA analysis of the full factorial design data confirmed the expected significant effects of growth stages, storage periods and drying conditions on mangiferin yield. The ANOVA analysis further revealed that interaction of growth stage × storage periods was significant while other interactions viz. drying methods × growth stage and drying methods × growth stage × storage conditions were not significant. Table 1 shows the content of mangiferin in different samples of S. chirata at different growth stages, drying methods and storage periods. The Duncan's multiple range test for mangiferin yield mean values and standard deviation (Table 1) showed that mangiferin yield at bud stage was significantly higher than the other growth stages of S. chirata plant.
HPTLC fingerprints obtained under UV 254 light, 3-D Densitogram, overlay spectra and Densitometric-HPTLC chromatograms of S. chirata test samples compared with standard compound (mangiferin) are shown in Figs. 4 and 5. Figure 6 presents HPLC chromatogram of mangiferin: standard compound with HPLC chromatograms obtained from S. chirata samples prepared with ultrasound-assisted extraction (UAE), microwave-assisted extraction (MAE) and heat reflux extraction (HRE), respectively by using 50% EtOH solution. Interaction and main plots of growth stage, drying methods and storage periods for mangiferin yield from S. chirata leaves are presented in Figs. 7 and 8. www.nature.com/scientificreports/  (Figs. 7, 8). The relative proportions of mangiferin of shade-dried leaves were higher than the other two drying methods. It could be concluded that drying of leaves of S. chirata in the shade drying was more suitable than oven and sun drying and is recommended for obtaining higher yield of mangiferin. The effect of the sun drying was more pronounced with a large decrease (4.663%) of mangiferin as compared to oven drying (4.717%) of mangiferin in 1-month old stored leaves harvested at bud-stage (4.73%).
Effect of post-harvest (storage periods) on mangiferin content. An attempt has been made to check the degradation of mangiferin compound under different storage periods after harvesting. Powdered samples of S. chirata leaves were stored in polypropylene containers and kept in dark at room temperature. HPTLC reports revealed that there were no significant variations of mangiferin content in plant samples immediately after drying and storing up to 1 month in our studies but subsequently there was declined in the mangiferin content in dried leaves. The comparison of the two-storage time revealed that total mangiferin contents of 1-month old plant sample are higher than those of 6-month-old samples. The results displayed that mangiferin content in the 1-month-old leaves showed 4.73% mangiferin as compared to 4.34% in 6 months old S. chirata leaves powder (Figs. 7, 8) showing that there was significant variation of mangiferin content in all the plant samples stored under controlled conditions for 1 month and 6 months. These results are according with those reported by Manika et al. 18 at the same conditions.   Table 2). Ethanol is recommended as non-toxic solvent and also yielded highest content of mangiferin in S. chirata leaves. Mangiferin content in S. chirata at room temperature by methanol solvent was found to be 2.16%. Hence, ethanol instead of methanol could be used as an effective solvent for extracting the significant phytochemical marker-'Mangiferin' for obtaining the highest content from S. chirata leaves. Effect of different compositions of aqueous ethanol on mangiferin yield from S. chirata leaves are presented with Densitometric-HPTLC chromatograms (Fig. 5).

Optimization of extraction conditions by full factorial design. Selection of effective extraction
method is a key concern for the recovery and isolation of desired bio-active compounds from a mixture of crude matrix. The present work attempts to compare the heat reflux (HRE), microwave assisted (MAE) and ultrasound assisted extraction (UAE) techniques to determine the efficient method to extract mangiferin from S. chirata leaves. The extraction efficacy of the different techniques has been varied between 1.46 and 4.86% (Table 2).
Two-way ANOVA tables were generated for the determination of significance of individual and interactive factors for mangiferin yield from S. chirata leaves. On the basis of p values (≤ 0.05) the significance of all the factors were judged statistically (Table 4). Main effect and interaction plots were constructed to study the most influential factors affecting the mangiferin yield (Figs. 9, 10). The ANOVA analysis of the full factorial design data confirmed the expected significant effects of extraction type (HRE, MAE, UAE) and also revealed the www.nature.com/scientificreports/ significant impact of aqueous ethanol solvent composition on mangiferin yield from S. chirata ( Table 2). The Duncan's multiple range test for mangiferin yield mean values showed for S. chirata leaf extracted with UAE and MAE was better than HRE. Heat refluxing with different solvent compositions such as absolute ethanol, 25% aqueous ethanol, 50% aqueous ethanol, 75% aqueous ethanol and aqueous for 1 h at 75 °C yielded 1.46%, 1.77%, 4.14%, 3.137% and 2.75%, respectively. Subsequently, microwave assisted extraction (MAE) and ultra-sonicator assisted extraction (UAE) methods were applied to extract mangiferin from S. chirata leaf using same solvent compositions as that of reflux method and these methods were found to be better than the reflux extraction.
In the experimentation, UAE has shown highest mangiferin content 4.86% as compared to MAE (4.82%) and HRE (4.14%) methods.  Table 2. The results showed that mangiferin yield varies with the aqueous ethanol composition by keeping the extraction time 1 h. An increasing trend in yield of mangiferin with respect to increase in polarity of solvent was observed from 100% ethanol to 50% aqueous ethanol after that there was a decrease in the yield of mangiferin at 75% aqueous ethanol and 100% aqueous (Figs. 9, 10). The high temperature facilitates the dissolution and displacement of saturation equilibrium constant and increases efficiency of extractable compounds.
Effect of MAE and aqueous ethanol on mangiferin yield. MAE resulted in higher yield of mangiferin as compared to HRE method, ranging from 2.147 ± 0.09 to 4.820 ± 0.19 ( Table 2). The pattern observed (highest at 50% aqueous ethanol) is similar to that observed in HRE. The results of the effect of aqueous ethanol on yield of mangiferin are presented in Figs. 9 and 10.
Effect of UAE and aqueous ethanol on mangiferin yield. Content yield vary from 2.22 ± 0.091 to 4.860 ± 0.19 (Table 2). Ultra sonicator exposure for 30 min showed best mangiferin yield at 50% aqueous ethanol. Effect of aqueous ethanol on yield of mangiferin is depicted in Figs. 9 and 10.  (Table 5), with aqueous ethanol (50%) extracts having greater DPPH and percentage inhibition of α-amylase activity than the methanolic, ethanolic and aqueous extracts. The DPPH activity was in the range of 62.5-82.1% with the highest and lowest DPPH activity observed in the aqueous ethanol (50%) extracts of the leaves harvested at bud-stage and ethanol extracts of the leaves harvested at flowering stage, respectively. Different concentrations of DPPH (10-80 μg mL −1 ) were used to calculate the anti-oxidant activity of plant samples, and it was observed that percentage increases sharply with increased concentration of DPPH. At a concentration of 80 μg mL −1 , the highest DPPH activity was observed in the aqueous ethanol (50%) extract of S. chirata leaves harvested at bud-stage (82%) of plant followed by flowering stage (78%). Moreover, percentage inhibition activity of α-amylase was observed in the range of 69.5-77.6% with the maximum and least in-vitro activity observed in the aqueous ethanol (50%) extracts of the leaves harvested at bud stage and ethanol extracts of the leaves harvested at flowering stage, respectively (Table 5). Among all the plant samples, highest percentage inhibition of α-amylase was observed in the aqueous ethanol (50%) extract of S. chirata leaves harvested at bud-stage (77%) followed by flowering stage (74%) plant.  www.nature.com/scientificreports/ All analyses are the mean of three replicates measurements ± standard deviation. The results were analyzed by one-way ANOVA followed by Duncan's Multiple Range Test. Values with different alphabets (a-c) and (a-b) within the same column (DPPH %) and (% α-amylase activity) are significantly different at p < 0.05 respectively.

Discussion
The effects of plant parts, growth stage, drying method, storage periods, solvent type and the extraction method on the mangiferin content were investigated. Earlier reports revealed the significant variation of mangiferin yield in Swertia spp. leaves amongst rosette and fruiting stage of plant that were harvested between June and November months. Yang et al. 30 reported the highest content of mangiferin in a bud-stage of Swertia mussoti (a potent Chinese Swertia species) herb but showed no variation between flowering and fruiting stage. It was already reported that content of bio-active compound varies with growth stage of plant, seasonal variation/and proper harvesting time period 31,32 . In other medicinal plants, some researchers also observed the gradual seasonal variation in various bio-active compounds 18,31,[33][34][35] . Many previous researchers reported the importance of drying methods and their circumstances on the amount and quality of the phytochemicals. All these studies show the superiority of shade drying over the oven drying, sun drying and other drying methods of the plant samples 36 . As far as selection of solvent were concern ethanol is recommended as non-toxic solvent and also yielded highest content of mangiferin in Swertia chirata leaves (Figs. 3,4,5,8,9,10). Hence, ethanol instead of methanol could be used as an effective solvent to extract the significant phytochemical marker-'Mangiferin' to obtain the highest content of these compounds from S. chirata leaves 19 . Mangiferin content in Swertia chirata at room temperature by methanol solvent was found to be 2.16% with the method described by Pandey et al. 28 .
The high temperature facilitates the dissolution and displacement of saturation equilibrium constant and increases efficiency of extractable compounds. Ruiz-Montanez et al. 37 extracted mangiferin by using ethanol-water (8:2 v/v) by soxhlet from mango peel. When comparing the time needed to achieve the complete extraction of mangiferin, HRE takes longer time (60 min) followed by UAE (30 min) and MAE (2 min) at 50% aqueous ethanol. Similar mangiferin yield was obtained in a plant sample by using method of Pandey et al. 28 at room temperature by soaking leaf powder in methanol for 24 h. This is because an increase in temperature in HRE facilitates the dissolution of mangiferin and penetration of aqueous ethanol solvent in plant matrix more than methanol solvent in short duration 27 . Microwave assisted extraction (MAE) uses microwave radiations for better extraction by directly affecting the molecules by dipole polarization and thus rapidly heats the solvent 22 . www.nature.com/scientificreports/ UAE uses ultrasound waves to rupture the cell wall due to the micro-cavities in the plant material and thus resulting in the efficient extraction of bio-active compounds 23 . When comparing the time needed to achieve the complete extraction of mangiferin, HRE takes longer time (60 min) followed by UAE (30 min) and MAE (2 min) at 50% aqueous ethanol.
In consequence UAE and MAE methods resulted in the highest yield of bio-active compounds and also supposed to be a cost effective, rapid and green extraction technology. In future, UAE and MAE methods can be efficiently used to get the maximum mangiferin content from S. chirata leaves.
The study found that binary solvent mixture (aqueous ethanol) resulted in better yield of polyphenols 38 . Ethanol breaks the bond between solute and plant materials, reduces surface tension of the medium and increases the mass transfer of bioactive compounds into the solvent; on the other hand water causes cell swelling to increase the surface area 39 . In the present study, aqueous ethanol was also found to give maximum anti-oxidant and anti-diabetic activity, which is higher than water, absolute ethanol and methanol. Mangiferin, as xanthonoid is the major phytochemical compound of S. chirata, which is well known to exhibit anti-oxidant and anti-diabetic activity. As the present study concluded that only aqueous ethanol produced the highest yield of mangiferin in Figure 10. Main plots of extraction methods and aqueous ethanol composition for mangiferin yield from Swertia chirata leaves. Dot is the mean value of mangiferin content. www.nature.com/scientificreports/ S. chirata plant samples, consequently it was coherent that the plant leaves extracted with the same solvent had the highest anti-oxidant and anti-diabetic activities.

Conclusions
The experimental design approach using full factorial design was successfully applied in the optimization of mangiferin yield from Swertia chirata leaves by using HPTLC and HPLC. Firstly, full factorial design for optimization of pre-harvest and post-harvest factors was evaluated and improved mangiferin yield (4.73%) was only found in S. chirata test samples harvested at bud-stage and shade dried for 1 month. Furthermore, optimum mangiferin yield (4.86%) was obtained by choosing an extraction method, UAE and 50% aqueous ethanol solvent solution.
In this study, appropriate growth stage at budding, leaf part, shade drying and storage period for 1 month, UAE and 50% aqueous ethanol were found to be significant factors to achieve the highest yield of mangiferin from Swertia chirata.

Data availability
The datasets used and/or analyzed during the current study available from the corresponding author on reasonable request.