Food Chemistry 217 (2017) 91–97 Contents lists available at ScienceDirect Food Chemistry journal homepage: www.elsevier.com/locate/foodchem Salting-out extraction of allicin from garlic (Allium sativum L.) based on ethanol/ammonium sulfate in laboratory and pilot scale Fenfang Li b, Qiao Li b, Shuanggen Wu b, Zhijian Tan a,⇑ a b Institute of Bast Fiber Crops and Center of Southern Economic Crops, Chinese Academy of Agricultural Sciences, Changsha 410205, China College of Chemistry and Chemical Engineering, Central South University, Changsha 410083, China a r t i c l e i n f o Article history: Received 13 June 2016 Received in revised form 24 August 2016 Accepted 24 August 2016 Available online 25 August 2016 Keywords: Salting-out extraction Allicin Purification Biological activities a b s t r a c t Salting-out extraction (SOE) based on lower molecular organic solvent and inorganic salt was considered as a good substitute for conventional polymers aqueous two-phase extraction (ATPE) used for the extraction of some bioactive compounds from natural plants resources In this study, the ethanol/ammonium sulfate was screened as the optimal SOE system for the extraction and preliminary purification of allicin from garlic Response surface methodology (RSM) was developed to optimize the major conditions The maximum extraction efficiency of 94.17% was obtained at the optimized conditions for routine use: 23% (w/w) ethanol concentration and 24% (w/w) salt concentration, 31 g/L loaded sample at 25 °C with pH being not adjusted The extraction efficiency had no obvious decrease after amplification of the extraction This ethanol/ammonium sulfate SOE is much simpler, cheaper, and effective, which has the potentiality of scale-up production for the extraction and purification of other compounds from plant resources Ó 2016 Elsevier Ltd All rights reserved Introduction Garlic (Allium sativum L.) is a well-known edible and medicinal plant since ancient China Allicin (diallylthiosulfinate) is an organosulfur compound, and it is one major biological active substance in garlic (Tyagi, Pradhan, Srivastava, & Mehrotra, 2014) Actually, allicin is converted from alliin after crushing of the garlic clove under the action of alliinase (Amagase, Petesch, Matsuura, Kasuga, & Itakura, 2001; Ankri & Mirelman, 1999) Allicin was first studied for its antibacterial properties in the middle of 20th century (Cavallito & Bailey, 1944), then its other pharmacological actions of anti-oxidant, antifungal, antihypertensive, antiinflammatory, and inhibition of tumor were also found (ElKashef, El-Kenawi, Suddek, & Salem, 2015; Hirsch et al., 2000; Liu, Ren, Wang, Yao, & He, 2015) Up to now, the solvent extraction using water or ethanol aqueous solution (Arzanlou & Bohlooli, 2010; Bocchini, Andalo, Pozzi, Galletti, & Antonelli, 2001; Wang et al., 2014) and supercritical fluid extraction (SFE) (del Valle, Glatzel, & Martinez, 2012; Liang, Qiao, Bi, Zou, & Zheng, 2012; Rybak, Calvey, & Harnly, 2004) are the most widely used methods for the extraction of allicin from garlic in laboratory, pilot or large scale However, the solvent extraction can obtain the crude extract ⇑ Corresponding author E-mail address: tanzhijiantgzy2010@aliyun.com (Z Tan) http://dx.doi.org/10.1016/j.foodchem.2016.08.092 0308-8146/Ó 2016 Elsevier Ltd All rights reserved and the samples need further purification; SFE requires sophisticated instrument and high cost Aqueous two-phase extraction (ATPE) was first introduced by Albersson, and the most commonly used two aqueous two-phase systems (ATPSs) were PEG/salt and PEG/dextran (Albertsson, 1986) ATPS based on low molecular organic solvents (e.g methanol, ethanol, acetone, and n-propanol) and inorganic salts had been developed in recent years, which can also be called salting-out extraction (SOE) system (Dong et al., 2016) Compared with polymer ATPS, SOE has the advantages of lower cost, lower viscosity, quicker phase separation time, relatively lower environmental toxicity and easier to scale up (Amid, Shuhaimi, Sarker, & Manap, 2012; Fu, Yang, & Xiu, 2015; Liu, Zou, Gao, Gu, & Xiao, 2014; Ooi et al., 2009; Wang, Han, Xu, Hu, & Yan, 2010) SOE systems had been used to extract various bioactive compounds from different plants resources, such as anthocyanins from grape juice (Wu et al., 2014), alkaloids from Sophora flavescens Ait (Zhang et al., 2015), phenolic compounds from Ficus carica L (Feng et al., 2015), rutin from acerola waste (Reis et al., 2014), lignans from Zanthoxylum armatum (Guo, Su, Huang, Wang, & Li, 2015), and polysaccharides from Semen Cassiae (Chen et al., 2016) The objective of this study is to use SOE for the extraction and preliminary purification of allicin from garlic powder The allicin is extracted into alcohol-rich phase, while partial impurities are extracted into the salt-rich phase The extraction conditions were 92 F Li et al / Food Chemistry 217 (2017) 91–97 optimized at laboratory scale, SOE was scaled up to the pilot scale under the optimized conditions The phase-forming components of ethanol and ammonium sulfate were recycled and reused Materials and methods 2.1 Materials and reagents The garlic samples were originated in Shandong province and bought from the Vanguard supermarket in Changsha City, Hunan Province The allicin standard was purchased from National Institutes for Food and Drug Control (Beijing, China) with HPLC purity larger than 98% HPLC grade acetonitrile was purchased from TEDIA Company, Inc (Fairfield, OH, USA) The analytical reagents of different organic solvents (ethanol, n-proanol, isopropanol, acetone and acetonitrile) and salts (ammonium sulfate, sodium dihydrogen phosphate, sodium sulfate, and potassium phosphate) were provided by Sinopharm Chemical Reagent Co., Ltd (Shanghai, China) 2,2-diphenyl-1-picpicrylhydrazyl (DPPH), ferrous sulfate, salicylic acid hydrogen peroxide were purchased from Alladin Reagent Co., Ltd (Shanghai, China) Escherichia Coli DH5a was provided by Tiangen Biotech Co., Ltd (Beijing, China) All other reagents used in this study were analytical grade and no further treatments were processed for them 2.2 Preparation of crude extract The garlic was peeled, then smashed into powder In order to make complete converting of alliin to allicin, a grinding time of 30 was employed Then the garlic powder was extracted by absolute ethanol with ultrasonic assisted extraction for 20 in an ultrasonic bath (model KQ-5200 DE, Kunshan Ultrasound Co Ltd., Kunshan, China) The mixture was centrifugated and filtrated to remove the insoluble substance The supernate was evaporated to remove the ethanol, the crude extract of allicin was obtained 2.3 Phase diagram The phase diagrams indicate the information of the concentration of each component is required to form the ATPS, which were drawn by the similar turbidimetric titration reported in our previous work (Tan, Li, & Xu, 2013) Firstly, organic solvent of known mass was added into a tube Then, a salt solution of known mass fraction was added dropwise, and the solution in the tube was well mixed The solution became turbid after adding of the salt solution then separated into two phases The mass fraction of each added component was calculated Lastly, a few drops of water was added to make the mixture clear again The above procedures were repeated to obtain sufficient data to construct the phase diagrams 2.4 Salting-out extraction SOE system was formed by adding salt aqueous solution into a tube containing a certain amount of organic solvents, then crude allicin extract was added ATPS was used for the SOE of allicin after complete stirring Allicin was extracted into the top phase (organic solvent-rich phase), while some hydrophilic substances, such as alliin, lysine, glutamic, etc., tended to partition in the bottom phase (salt-rich phase) (Jiang, Lu, Tan, & Cui, 2014) Centrifugation was performed to accelerate the phase-forming and achieve complete extraction, then two clear phases formed and the volume of each phase was noted down HPLC analysis was used for the quantitative determination of the allicin concentration The extraction efficiency (E/%) of allicin in top phase was calculated by the Eq (1) E¼ Ct V t 100% Cov o ð1Þ where Ct was the allicin concentration in top phase, and Vt was the volume of top phase, Co and Vo were the concentration and volume, respectively of crude allicin extract added into the system 2.5 HPLC conditions The allicin samples were analyzed using a Dionex UltiMate 3000 HPLC system (Dionex, Sunnyvale, California, USA) coupled with a LPG-3400 pump, a VWD-3400 UV–vis detector, and a TCC-3000 column compartment A Promosil C18 chromatographic column (250 4.6 mm i.d., lm particle size) was used to analyze the samples A PC coupled with a Chameleon software was used to collect and analyze the data The mobile phase was composed of acetonitrile, H2O, and acetic acid (75:25:0.6; v/v/v) The flow rate was 1.0 mL/min with isocratic elution and the effluent was monitored at the wavelength of 240 nm The oven temperature was set at 30 °C All the samples of 20 lL were injected into the HPLC for analysis The mobile phase and samples solution were filtered through microfiltration membrane (0.45 lm) before analysis The standard curve for analysis of allicin is A = 292.67C + 0.34 (R2 = 0.9995), where A is the peak area and C is the allicin concentration Standard solutions of allicin were diluted in the range from 0.024 to 0.114 mg/mL 2.6 Antioxidant activity of allicin obtained by SOE The free-radical scavenging activity of allicin was assessed using DPPH radical according to the method with moderate modification (Athukorala, Kim, & Jeon, 2006) Two milliliter allicin solution was added into 2.0 mL DPPH ethanol solution (40 lg/mL) Another two samples were prepared with 2.0 mL ethanol being added into 2.0 mL DPPH (as blank) and 2.0 mL allicin concentration solution being added into 2.0 mL ethanol (as control), respectively The mixture was shaken vigorously and left to stand for 40 at ambient temperature in darkness Absorbance was determined at 517 nm by a UV–Vis spectrophotometer (UV-2100, Unico, USA), vitamin c (Vc) was used as a comparison to allicin The results were expressed as percentage inhibition of the DPPH radical on Eq (2) Scavenging capacity %ị ẳ ẵ1 As Ac ị=Ao ð2Þ where Ac, As and Ao represent the absorbance of control, sample and blank, respectively The antioxidant capacity of test compounds was expressed as IC50, the concentration necessary for 50% reduction of DPPH The scavenging ability of hydroxyl radical (OH) was evaluated according to the modified Fenton method (Wang & Jiao, 2000) To a colorimetric tube, 4.0 mmol/L FeSO4, 10.0 mmol/L H2O2, 5.0 mmol/L salicylic acid ethanol solution and a certain concentration of allicin sample were added The proportion of each component was indicated as follow: Ac (1.0 mL FeSO4 + 0.5 mL salicylic acid + 7.5 mL H2O + 1.0 mL H2O2), As (1.0 mL FeSO4 + 0.5 mL salicylic acid + 7.5 mL allicin sample + 1.0 mL H2O2), and Ao (1.0 mL FeSO4 + 0.5 mL H2O + 7.5 mL allicin sample + 1.0 mL H2O2) Each mixture was vigorously shaken and left to stay for 30 at ambient temperature in darkness The absorbance was determined at 510 nm by a UV–Vis spectrophotometer (UV-2100, Unico, USA) and the scavenging radical activity was calculated according to the identical equation (5) used in DPPH determination The antioxidant capacity of test compounds was expressed as IC50, the concentration necessary for 50% reduction of OH 93 F Li et al / Food Chemistry 217 (2017) 91–97 2.8 Statistical analysis Design Expert 7.0 (DE, Stat-Ease, Inc., Minneapolis, USA) was used to analyze the experimental data and obtain the response models The lack of fit and coefficient of determination (R2) were used to judge the suitability of model The analysis of variance (ANOVA) was carried out with the comparison of the actual and predicted values, and the optimized conditions for three major variables were acquired by statistical analysis The statistical analysis for screening the SOE system and single factor experiments was done using the Duncan’s multiple range test (Granato, Calado, & Jarvis, 2014) 90 Mass fraction of organic solvents (%) An agar-well diffusion method was used to assess the antibacterial ability of allicin obtained by SOE (Shan, Cai, Brooks, & Corke, 2007) The concrete procedures were as follows: firstly, E coli was selected as a strain and subcultured, and allicin samples of different concentration ranges were prepared Secondly, inoculated suspension was spread onto the surface of every plate equally Thirdly, three wells with diameter of mm were cut from the agar using a hole puncher Fourthly, 50 lL allicin solution was pipetted into each well Ethanol and deionized water acted as the negative controls Lastly, the inoculated plates were left undisturbed and incubated at 35 °C for 24 h Following these processes, antibacterial activity was estimated by measuring the diameter of growth inhibition zone using a Vernier caliper Each sample was tested in triplicate and all the equipments and tested reagents used in this experiment were sterilized (a) iso-propanol acetone n-propanol acetonitrile ethanol 80 70 60 50 40 30 20 10 10 11 Mass fraction of ammonium sulfate (%) (b) 90 Mass fraction of ethanol (%) 2.7 Antibacterial property of allicin obtained by SOE Potassium phosphate Sodium sulfate Sodium dihydrogen phosphate Ammonium sulfate 80 70 60 50 40 30 20 10 11 12 13 14 15 Mass fraction of salt (%) Fig Binodal curves for the different organic solvent and salt ATPS at 298.15 K Results and discussion 3.1 Selection of the optimum SOE system To screen the optimal SOE system for the extraction and isolation of allicin, various organic solvents (ethanol, n-propanol, isopropanol, acetone and acetonitrile) and salts (ammonium sulfate, sodium dihydrogen phosphate, sodium sulfate, and potassium phosphate) were considered as the phase-forming components The binodal curves were shown in Fig 1, which imply that the closer the binodal curves to the coordinates, the less salt (organic solvent) required for forming two phases under the same organic solvent (salt) concentration It can be seen in Fig 1(a) that the phase-forming ability of all the organic solvents in the organic solvents/ammonium sulfate ATPSs followed the order: acetonitrile > npropanol > isopropanol > ethanol > acetone The reasons for influencing the phase-forming ability of organic solvents are complicated, some literatures reported that the hydrophobicity, solubility, boiling-point, and kosmotropicity of salts of organic solvents were attributed to this phenomenon (Nemati-Knade, Shekaari, & Jafari, 2012; Ooi et al., 2009) Fig 1(b) showed the phase-forming ability of the four types of salt in the ethanol/salt ATPSs followed the order: potassium phosphate > sodium sulfate > ammonium sulfate > sodium dihydrogen phosphate The phase-forming ability (salting-out ability) of salt influencing different ATPSs is mainly related to the ions’ Gibbs free energy of hydration D(Ghyd) (Bridges, Gutowski, & Rogers, 2007; Tan, Li, & Xu, 2012) The phase diagrams can be used for selecting the organic solvent and salt concentrations in the RSM experiments The extraction efficiencies using different SOE systems were shown in Fig As it can be seen that two ATPSs formed by ammonium sulfate and sodium sulfate had relatively higher extraction efficiency than other two ATPSs formed by sodium dihydrogen phosphate and potassium phosphate It can also be seen that the alkaline system (ethanol/potassium phosphate) was not suitable for the allicin extraction Considering to the reasons that ammonium sulfate can lead to acidic condition and has better solubility than sodium sulfate, ammonium sulfate was chosen as the phaseforming salt As to the systems formed by different organic solvents and ammonium sulfate, the SOE system based on ethanol had the highest extraction efficiency Therefore, ethanol/ammonium sulfate was chosen as the optimum SOE system 3.2 Single factor experiments There are many factors, such as ethanol and ammonium sulfate concentration, loaded sample, temperature, pH etc., can influence the SOE Therefore, it is necessary to optimize these conditions to obtain the maximum extraction efficiency In the single factors experiments, when one factor was investigated, the other factors were set at fixed values 3.2.1 Effect of ethanol and ammonium sulfate concentration The ethanol and ammonium sulfate concentration can be considered as the most influencing factors in majority SOE systems In order to optimize the ethanol concentration, the concentration ranges of 18%–26% (w/w) were studied The results in Fig 3(a) showed that the maximum extraction efficiency was obtained at the ethanol concentration of 24% (w/w) The change of salt concentration was behaved as the change of salting-out effect To study the ammonium sulfate concentration, the salt concentration in the range of 17%–25% (w/w) was chosen The results in Fig 3(b) showed that the extraction efficiency increased with increasing of salt concentration, then decreased with further addition of salt, and the maximum extraction 94 F Li et al / Food Chemistry 217 (2017) 91–97 Fig SOE of allicin using different ATPSs formed by 25% (w/w) organic solvents and 12% (w/w) salts at room temperature, the loaded sample was 40 g/L with system pH being not adjusted Different letters in the same series indicate significant difference at P < 0.05 level Fig Single factor experiments for studying the effect of (a) ethanol concentration (the ammonium sulfate concentration was 25% and the loaded sample was 40 g/L at room temperature with pH being not adjusted), (b) ammonium sulfate concentration (the ethanol concentration was 20% and the loaded sample was 40 g/L at room temperature with pH being not adjusted), (c) loaded sample (the ethanol concentration was 20% and the ammonium sulfate concentration was 25% at room temperature with pH being not adjusted), (d) temperature (the ethanol concentration was 20%, the ammonium sulfate concentration was 25%, and the loaded sample was 40 g/L with pH being not adjusted), and (e) pH (the ethanol concentration was 20%, the ammonium sulfate concentration was 25%, and the loaded sample was 40 g/L at room temperature) on the SOE of allicin Different letters in the same series indicate significant difference at P < 0.05 level efficiency was obtained at the ammonium sulfate concentration of 23% (w/w) In fact, the increasing of salt concentration can increase the salting-out ability (Liu, Mu, Sun, Zhang, & Chen, 2013), which is beneficial for the extraction, however the salt concentration will reach saturation and more water will be pulled into salt-rich phase (Dong et al., 2016), which can probably make the alcohol-rich phase become not good solvent for dissolving allicin 3.2.2 Effect of loaded sample After the extraction experiments being designed, as much allicin as possible will be added into the system in order to achieve biggest economic benefit As it was reported that the crude load can alter the partition behavior of target molecules (Amid et al., 2012) The crude extract was dissolved into the ethanol to prepare concentration ranges of 20–60 g/L in this study, and then salt was added to form ATPS for SOE of allicin The results in Fig 3(c) showed that when 40 g/L allicin sample was prepared, the maximum extraction efficiency was obtained, and then decreased with increasing of loaded allicin That can be interpreted the excessive addition of allicin will make the saturation of extraction 3.2.3 Effect of temperature and pH The extraction temperature and system pH are also key factors affecting the extraction to some extent The effect of temperature was studied in the range of 25–50 °C The results in Fig 3(d) showed that the extraction efficiencies decreased with increasing of temperature In our previous studies, it was also found that the SOE was better to be operated around the room temperature because this procedure was spontaneous and exothermic (Tan et al., 2013, 2014) Therefore, the extraction experiments were done at room temperature without heating F Li et al / Food Chemistry 217 (2017) 91–97 The change of pH can influence the electrical charge of target compound, thus to influence its hydrophobicity/hydrophilicity and furtherly influence its partition in SOE (Wang et al., 2010) The system pH was adjusted by hydrochloric acid and sodium hydroxide aqueous solution The effect of pH was investigated at pH 4.0–8.0 It can be seen from Fig 3(e) that the pH 6.0 was most suitable for the extraction The aqueous solutions of allicin lead to a pH of approximately 6.5 (Cavallito & Bailey, 1944; Freeman & Kodera, 1995), therefore, the system pH around this value is favored This optimal ethanol/ammonium sulfate ATPS can generate an acidic system with a pH of approximately 6.5, and the alkaline condition was not suitable for SOE, thus it is unnecessary to adjust the system pH at the extraction procedure 3.3 Optimization of the SOE conditions by RSM The SOE procedure is very complex, the interactions between influencing factors should also be considered Therefore, it’s better to study the interaction between the three independent process variables (ethanol concentration, ammonium sulfate concentration, and loaded sample), RSM based on Box–Behnken design (BBD) was developed to optimize the extraction conditions Table showed the obtained results by the RSM models The ANOVA in Table showed a P-value was 0.0006, which indicates regression model was significant The Model F-value of 16.74 implies that the model is significant There is a 0.06% probability that a ‘‘Model F-Value” occurs due to noise The model terms can be considered to be significant if the P-values are less than 0.05 According to this principle, the model terms of X1, X2, X2X3, X21, X22 can be considered as significant terms 95 The ‘‘Lack of Fit F-value” of 5.23 implies the ‘‘Lack of Fit” is not significant relative to the pure error and that there is a 7.2% chance that a ‘‘Lack of Fit F-value” this large could occur due to noise The regression model was generated by the software, which was present in Eq (3): Y ẳ 2:24 ỵ 0:57X 0:14X 0:060X ỵ 0:091X X ỵ 0:26X X 0:081X X 0:59X 21 0:10X 22 0:23X 23 R2 ẳ 0:9556ị 3ị where Y is the extraction efficiency (%), X1 is the ethanol concentration (%, w/w), X2 is the ammonium sulfate concentration (%, w/w), and X3 is the loaded sample (g/L) The coefficient of determination (R2) is 0.9556, implying that more than 95.56% of the variations in the process efficiency could be explained by the model The response surfaces for the effects of the independent variables on the average extraction efficiency of allicin were shown in Fig Based on the quadratic model, the calculated optimum conditions for the extraction of allicin are an ethanol concentration of 22.57% (w/w), salt concentration of 24.11% (w/w), and the loaded sample of 30.96 g/L The yield, averaged over triplicate runs, was 94.17% under the optimized conditions [23% (w/w) ethanol concentration, 24% (w/w) salt concentration, and 31 g/L loaded sample], which was very close to the predicted value of 94.68% This demonstrates that the model is adequate for predicting the expected optimization The purity of the obtained allicin under the optimized conditions was determined by quantitative analyses using HPLC The result indicated that the purity of allicin reached to 68.4%, having noteworthy increase compared with the crude extract (purity of 31.8%) 3.4 Antioxidant activity and antibacterial property Table Experimental results for the three-factor/three-level BBD and analysis of variance (ANOVA) for the quadratic response surface model Run Factor X1: ethanol concentration (%, w/w) Factor X2: salt concentration (%, w/w) Factor X3: loaded sample (g/L) Extraction efficiency (%) 10 11 12 13 14 15 16 17 18 26 18 26 18 26 18 26 22 22 22 22 22 22 22 22 22 21 21 25 25 23 23 23 23 21 25 21 25 23 23 23 23 23 40 40 40 40 30 30 50 50 30 30 50 50 40 40 40 40 40 71.79 86.43 79.06 90.74 85.24 94.32 83.45 92.31 78.15 89.53 90.44 83.78 95.50 92.68 94.03 96.04 95.13 Source Sum of squares Degree of freedom Mean square F-value P-value prob > F Model X1 X2 X3 X1X2 X1X3 X2X3 X21 X22 X23 Residual Lack of fit Pure error Cor total 757.04 244.87 33.21 0.94 2.19 0.012 81.36 91.36 270.35 5.94 35.17 28.02 7.15 792.21 1 1 1 1 16 84.12 244.87 33.21 0.94 2.19 0.012 81.36 91.36 270.35 5.94 5.02 9.34 1.79 16.74 48.73 6.61 0.19 0.44 2.408 103 16.19 18.18 53.81 1.18 0.0006 0.0002 0.0037 0.6786 0.5302 0.9622 0.0050 0.0037 0.0002 0.3128 5.23 0.072 The DPPH and OH assays presented different antioxidant capacity It can be seen from Fig 5(a) that the radical scavenging capacity of allicin for the DPPH assay was poorer than that of Vc, the IC50 values for allicin and Vc were approximate 5.0 lg/mL and 15 lg/mL, respectively Fig 5(b) appeared the radical scavenging ability for the Fenton reaction assay, it can be seen that the radical scavenging ability of OH increased with increasing of allicin concentration The IC50 for allicin was approximate 10 lg/mL, while that for Vc was approximate 80 lg/mL, indicating that the allicin had very strong scavenging ability of OH 3.5 Antibacterial property Allicin obtained by SOE was used to test the antibacterial ability The result in Fig 5(c) showed that allicin exhibited significantly antibacterial ability based on the measurement of the diameter of inhibition zone Allicin obtained by SOE had similar antibacterial ability compared with the standard allicin at the same concentration (seeing plates A and C) The data showed that diameter of inhibition zone increased significantly from 1.80 to 2.72 cm with increasing of allicin concentration from 0.185 to 0.923 mg/mL The controls of ethanol solution and pure water indicated that they had no antibacterial ability 3.6 Amplification of SOE and recycling of the phase-forming components SOE was scaled-up from 25 g to kg, 10 kg, and 20 kg of the total mass under the optimized conditions for routine application (23% (w/w) ethanol concentration, 24% (w/w) salt concentration, and 31 g/L loaded sample) using a continuous stirred tank reactor (Model GR-20, Zhengzhou Greatwall Scientific Industrial and Trade Co, Ltd, Zhengzhou, China) Salt was directly added into the crude 96 F Li et al / Food Chemistry 217 (2017) 91–97 (a) Scavenging capacity of DPPH (%) 100 90 80 70 allicin Vc 60 50 40 30 20 10 10 20 30 (b) 100 Scavenging capacity of OH (%) 40 50 60 70 80 Concentration (±J/mL) allicin Vc 90 80 70 60 50 40 30 20 10 10 20 30 40 50 60 70 80 90 100 Concentration (±g/mL) Fig (a) Test the antioxidant capacity of allicin obtained by SOE using DPPH; (b) Test the antioxidant capacity of allicin obtained by SOE using OH; (c) Test the antibacterial ability of allicin (A—allicin standard; B—ethanol solution; C—allicin obtained by SOE; D—pure water) Fig Response surface curves for the extraction of allicin from garlic showing interaction between (a) ethanol concentration and salt concentration, (b) ethanol concentration and loaded sample, and (c) salt concentration and loaded sample extract after removing the sediments using a centrifuge (Model TL-1000, Jiangsu Taizhou Taida Mechanical Industrial Co, Ltd, Changzhou, China) in the pilot scale The experimental results indicated that the extraction efficiencies had no obvious reduction from laboratory scale to pilot scale, the extraction efficiency was 91.5% at 20 kg scale, showing minor difference with 94.17% at 25 g scale The study of recycling of ethanol and ammonium sulfate was done using the reported method (Li, Teng, & Xiu, 2010; Tan et al., 2013) in the laboratory scale The two phases were separated by a separating funnel, the ethanol in top phase can be recycled by evaporation, and the concentration of recycled ethanol was 84.6% Ammonium sulfate in salt-rich phase can be recycled by dilution crystallization, the recycling efficiency can reach to 91.64% using methanol at a volume ratio of 2:1 to salt aqueous solution Conclusions In this work, ethanol/ammonium sulfate ATPS was successfully applied for the SOE of allicin from garlic in laboratory and pilot scale The factors influencing the SOE were investigated in detail The obtained optimum results were as follows: 24% (w/w) ethanol concentration, 23% (w/w) ammonium sulfate concentration, 40 g/L loaded sample at room temperature with pH being not adjusted The major three influencing factors were further optimized by RSM The optimized conditions were 22.57% (w/w) ethanol concentration, 24.11% (w/w) ammonium sulfate concentration, and 30.96 g/L loaded sample, 94.17% of allicin can be obtained in alcohol-rich phase under this condition The purity of allicin F Li et al / Food Chemistry 217 (2017) 91–97 obtained by SOE was 68.4%, which was much higher than that of crude extract (31.8% purity) The bioactive tests indicated allicin obtained by SOE had good antioxidant activity and antibacterial property Finally, this SOE was scaled up in pilot scale under the optimized conditions, no obvious decrease of extraction efficiencies was observed The scale-up and recycling experiments demonstrated this ethanol/ammonium sulfate SOE had the potentiality for use in industrial production of allicin Acknowledgement This work was financially supported by the National Natural Science Foundation of China (No 21406262) References Albertsson, P A (1986) Partition of cell particles and macromolecules (3rd ed.) New York: Wiley Amagase, H., Petesch, B L., Matsuura, H., Kasuga, S., & Itakura, Y (2001) Intake of garlic and its bioactive components The Journal of Nutrition, 131(3), 955S–962S Amid, M., Shuhaimi, M., Sarker, M Z I., & Manap, M Y A (2012) Purification of serine protease from mango (Mangifera Indica Cv Chokanan) peel using an alcohol/salt aqueous two phase system Food Chemistry, 132(3), 1382–1386 Ankri, S., & Mirelman, D (1999) Antimicrobial properties of allicin from garlic Microbes and Infection, 1(2), 125–129 Arzanlou, M., & Bohlooli, S (2010) Introducing of green garlic plant as a new source of allicin Food Chemistry, 120(1), 179–183 Athukorala, Y., Kim, K N., & Jeon, Y J (2006) Antiproliferative and antioxidant properties of an enzymatic hydrolysate from brown alga, Ecklonia cava Food and Chemical Toxicology, 44(7), 1065–1074 Bocchini, P., Andalo, C., Pozzi, R., Galletti, G C., & Antonelli, A (2001) Determination of diallyl thiosulfinate (allicin) in garlic (Allium sativum L.) by high-performance liquid chromatography with a post-column photochemical reactor Analytica Chimica Acta, 441(1), 37–43 Bridges, N J., Gutowski, K E., & Rogers, R D (2007) Investigation of aqueous biphasic systems formed from solutions of chaotropic salts with kosmotropic salts (salt-salt ABS) Green Chemistry, 9(2), 177–183 Cavallito, C J., & Bailey, J H (1944) Allicin, the antibacterial principle of Allium sativum I Isolation, physical properties and antibacterial action Journal of the American Chemical Society, 66(11), 1950–1951 Chen, Z., Zhang, W., Tang, X., Fan, H., Xie, X., Wan, Q., Wu, X., & Tang, J Z (2016) Extraction and characterization of polysaccharides from Semen Cassiae by microwave-assisted aqueous two-phase extraction coupled with spectroscopy and HPLC Carbohydrate Polymers, 144, 263–270 del Valle, J M., Glatzel, V., & Martinez, J L (2012) Supercritical CO2 extraction of allicin from garlic flakes: Screening and kinetic studies Food Research International, 45(1), 216–224 Dong, Y., Pang, B., Yu, F., Li, L., Liu, W., & Xiu, Z (2016) Extraction and purification of IgG by hydrophilic organic solvent salting-out extraction Journal of Chromatography B, 1012–1013, 137–143 El-Kashef, D H., El-Kenawi, A E., Suddek, G M., & Salem, H A (2015) Protective effect of allicin against gentamicin-induced nephrotoxicity in rats International Immunopharmacology, 29(2), 679–686 Feng, Y C., Li, W L., He, F M., Kong, T T., Huang, X W., Gao, Z H., Lu, N H., & Li, H L (2015) Aqueous two-phase system as an effective tool for purification of phenolic compounds from fig fruits (Ficus carica L.) Separation Science and Technology, 50(12), 1785–1793 Freeman, F., & Kodera, Y (1995) Garlic chemistry: stability of S-(2-propenyl)-2propene-1-sulfinothioate (allicin) in blood, solvents, and simulated physiological fluids Journal of Agricultural and Food Chemistry, 43(9), 2332–2338 Fu, H., Yang, S T., & Xiu, Z (2015) Phase separation in a salting-out extraction system of ethanol-ammonium sulfate Separation and Purification Technology, 148, 32–37 Granato, D., Calado, V M D., & Jarvis, B (2014) Observations on the use of statistical methods in food science and technology Food Research International, 55, 137–149 Guo, T., Su, D., Huang, Y., Wang, Y., & Li, Y.-H (2015) Ultrasound-assisted aqueous two-phase system for extraction and enrichment of Zanthoxylum armatum lignans Molecules, 20(8), 15273–15286 97 Hirsch, K., Danilenko, M., Giat, J., Miron, T., Rabinkov, A., Wilchek, M., Mirelman, D., Levy, J., & Sharoni, Y (2000) Effect of purified allicin, the major ingredient of freshly crushed garlic, on cancer cell proliferation Nutrition and Cancer, 38(2), 245–254 Jiang, X., Lu, Y., Tan, C., & Cui, B (2014) Combination of aqueous two-phase extraction and cation-exchange chromatography: New strategies for separation and purification of alliin from garlic powder Journal of Chromatography B, 957, 60–67 Li, Z G., Teng, H., & Xiu, Z L (2010) Aqueous two-phase extraction of 2,3butanediol from fermentation broths using an ethanol/ammonium sulfate system Process Biochemistry, 45(5), 731–737 Liang, G., Qiao, X., Bi, Y., Zou, B., & Zheng, Z (2012) Studies on purification of allicin by molecular distillation Journal of the Science of Food and Agriculture, 92(7), 1475–1478 Liu, D., Zou, X., Gao, M., Gu, M., & Xiao, H (2014) Hydrophilic organic/saltcontaining aqueous two-phase solvent system for counter-current chromatography: A novel technique for separation of polar compounds Journal of Chromatography A, 1356, 157–162 Liu, S G., Ren, P Y., Wang, G Y., Yao, S X., & He, X J (2015) Allicin protects spinal cord neurons from glutamate-induced oxidative stress through regulating the heat shock protein 70/inducible nitric oxide synthase pathway Food & Function, 6(1), 320–329 Liu, X., Mu, T., Sun, H., Zhang, M., & Chen, J (2013) Optimisation of aqueous twophase extraction of anthocyanins from purple sweet potatoes by response surface methodology Food Chemistry, 141(3), 3034–3041 Nemati-Knade, E., Shekaari, H., & Jafari, S A (2012) Thermodynamic study of aqueous two phase systems for some aliphatic alcohols plus sodium thiosulfate plus water Fluid Phase Equilibria, 321, 64–72 Ooi, C W., Tey, B T., Hii, S L., Mazlina, S., Kamal, M., Lan, J C W., Ariff, A., & Ling, T C (2009) Purification of lipase derived from Burkholderia pseudomallei with alcohol/salt-based aqueous two-phase systems Process Biochemistry, 44(10), 1083–1087 Reis, I A., Santos, S B., Pereira, F D., Sobral, C R., Freire, M G., Freitas, L S., Soares, C M., & Lima, A S (2014) Extraction and recovery of rutin from acerola waste using alcohol-salt-based aqueous two-phase systems Separation Science and Technology, 49(5), 656–663 Rybak, M E., Calvey, E M., & Harnly, J M (2004) Quantitative determination of allicin in garlic: Supercritical fluid extraction and standard addition of alliin Journal of Agricultural and Food Chemistry, 52(4), 682–687 Shan, B., Cai, Y.-Z., Brooks, J D., & Corke, H (2007) The in vitro antibacterial activity of dietary spice and medicinal herb extracts International Journal of Food Microbiology, 117(1), 112–119 Tan, Z., Li, F., & Xu, X (2013) Extraction and purification of anthraquinones derivatives from Aloe vera L using alcohol/salt aqueous two-phase system Bioprocess and Biosystems Engineering, 36(8), 1105–1113 Tan, Z., Li, F., & Xu, X (2012) Isolation and purification of aloe anthraquinones based on an ionic liquid/salt aqueous two-phase system Separation and Purification Technology, 98, 150–157 Tan, Z., Wang, C., Yi, Y., Wang, H., Li, M., Zhou, W., Tan, S., & Li, F (2014) Extraction and purification of chlorogenic acid from ramie (Boehmeria nivea L Gaud) leaf using an ethanol/salt aqueous two-phase system Separation and Purification Technology, 132, 396–400 Tyagi, G., Pradhan, S., Srivastava, T., & Mehrotra, R (2014) Nucleic acid binding properties of allicin: Spectroscopic analysis and estimation of anti-tumor potential Biochimica et Biophysica Acta (BBA)-General Subjects, 1840(1), 350–356 Wang, H., Li, X., Liu, X., Shen, D., Qiu, Y., Zhang, X., & Song, J (2014) Influence of pH, concentration and light on stability of allicin in garlic (Allium sativum L.) aqueous extract as measured by UPLC Journal of the Science of Food and Agriculture, 95(9), 1838–1844 Wang, S Y., & Jiao, H (2000) Scavenging capacity of berry crops on superoxide radicals, hydrogen peroxide, hydroxyl radicals, and singlet oxygen Journal of Agricultural and Food Chemistry, 48(11), 5677–5684 Wang, Y., Han, J A., Xu, X H., Hu, S P., & Yan, Y S (2010) Partition behavior and partition mechanism of antibiotics in ethanol/2-propanol-ammonium sulfate aqueous two-phase systems Separation and Purification Technology, 75(3), 352–357 Wu, Y., Wang, Y., Zhang, W., Han, J., Liu, Y., Hu, Y., & Ni, L (2014) Extraction and preliminary purification of anthocyanins from grape juice in aqueous twophase system Separation and Purification Technology, 124, 170–178 Zhang, W., Zhu, D., Fan, H., Liu, X., Wan, Q., Wu, X., Liu, P., & Tang, J Z (2015) Simultaneous extraction and purification of alkaloids from Sophora flavescens Ait by microwave-assisted aqueous two-phase extraction with ethanol/ ammonia sulfate system Separation and Purification Technology, 141, 113–123 ... capacity of allicin obtained by SOE using DPPH; (b) Test the antioxidant capacity of allicin obtained by SOE using OH; (c) Test the antibacterial ability of allicin (A? ?allicin standard; B? ?ethanol. .. excessive addition of allicin will make the saturation of extraction 3.2.3 Effect of temperature and pH The extraction temperature and system pH are also key factors affecting the extraction to some... calculated optimum conditions for the extraction of allicin are an ethanol concentration of 22.57% (w/w), salt concentration of 24.11% (w/w), and the loaded sample of 30.96 g/L The yield, averaged