Ultrasonics Sonochemistry 13 (2006) 511516 www.elsevier.

com/locate/ultsonch

Ultrasound-assisted extraction of xyloglucan from apple pomace

Fu Caili, Tian Haijun, Li Quanhong *, Cai Tongyi, Du Wenjuan
College of Food Science and Nutrition Engineering, China Agriculture University, Beijing 100083, China Received 9 May 2005; received in revised form 26 July 2005; accepted 26 September 2005 Available online 1 December 2005

Abstract Ultrasound-assisted extraction was evaluated as a simpler and more eective alternative to conventional extraction method for the isolation of xyloglucan from apple pomace. The apple pomace samples were extracted under indirect sonication in an ultrasound cleaning bath. The ultrasound extraction was compared with the conventional method, on the productions of xyloglucan. It was found that the ultrasound-assisted extraction of xyloglucan was about three times faster than the traditional extraction method. Response surface methodology was employed to study the eect of liquid:solid ratio, KOH concentration and ultrasound-assisted extraction time on the yield of xyloglucan from apple pomace. Regression analysis was performed on the data obtained. The most relevant variable was the quadratic of liquid:solid ratio. The coecient determination (R2) was good for the second-order model. The optimum combination was determined as a liquid:solid ratio of 34.4:1 (v/w), a KOH concentration of 3.3 M and an ultrasound-assisted extraction time of 2.5 h. 2005 Elsevier B.V. All rights reserved.
Keywords: Extraction; Xyloglucan; Apple pomace; Ultrasound; Response surface methodology; Optimization

1. Introduction Apple pomace is a major global waste product generated primarily during apple juice or cider processing. A great deal of eort has been expended to utilize this waste in a variety of ways such as biogas, energy production, food ingredients and speciality chemicals, including apple seed oils and polyphenols [16]. Traditionally, pectin is obtained from pomace [7]. Beside pectin the other major components of apple pomace are the cell wall polysaccharides cellulose and xyloglucan. Xyloglucan could be derivatized into a range of compounds, which might be useful as thickening agents or texture modiers, compared to those that can be prepared from cellulose (e.g. methylcellulose, hydroxypropylcellulose, carboxymethylcellulose) [8]. Another potential use for xyloglucan was as a source of biologically active oligosaccharides [9]. Xyloglucan oligosaccharides and itself were indicated to possess antitumor activity [10].
*

Corresponding author. Tel./fax: +086 10 82791623. E-mail address: quanhong_li@hotmail.com (Q. Li).

Because it binds tightly to cellulose microbrils through hydrogen bonds, cross-linking them into a cellulosexyloglucan network [1113], xyloglucan is often extracted by concentrated (1 or 4 M) KOH or NaOH. Xyloglucan from apple pomace was extracted by concentrated alkali [14] and the signicant (p < 0.05) factors were molarity of alkali and time of extraction. Xyloglucan extracted with 4 M KOH from the enzymically depectinated apple pomace was found to be lower in molecular weight than that from the non-depectinated pomace [8]. But there are few studies on ultrasound-assisted extraction of xyloglucan from apple pomace. The ultrasound treatment is a simple and rapid method in the fractionation of plant materials [15], two types of physical phenomena: diusion through the cell walls and washing out (rinsing) the cell contents once the walls are broken, involved in the extraction mechanism, are signicantly eected by ultrasound irradiation. Many polysaccharides such as hemicelluloses [16], polysaccharides from Salvia ocinalis L. [17], pectin [18] and immunologically active xylan from corn cobs [19] were highly eciently extracted with ultrasound without observation of substantial changes

1350-4177/$ - see front matter 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.ultsonch.2005.09.007

512

C. Fu et al. / Ultrasonics Sonochemistry 13 (2006) 511516

in their structural, molecular properties and biological activities. When many factors and interactions aect desired response, response surface methodology (RSM), a statistical-mathematical method, which uses quantitative data from an appropriate experimental design to determine and simultaneously solve multivariate equation [20], is an eective tool to optimize the process. Response surface experiments attempt to identify the response that can be thought of as a surface over the explanatory variables experimental space. It usually uses an experimental design such as central-composite rotatory design (CCRD) to t an empirical, full second-order polynomial model. A central-composite experimental design coupled with a full second-order polynomial model is a very powerful combination that usually provides an adequate representation of most continuous response surfaces over a relatively broad factor domain [21]. In this paper, at rst, the ultrasound eect on production of xyloglucan from apple pomace compared with the conventional method was studied. In the second part, the eect of liquid:solid ratio, KOH concentration and ultrasound time on the ultrasound-assisted extraction of xyloglucan from apple pomace were evaluated by response surface methodology in order to nd a optimal use of ultrasound for the extraction of xyloglucan. 2. Experimental 2.1. Materials Apple pomace was obtained from China Yantai North Andre Juice Co., Ltd. All pomace had been powdered to a particle size less than 0.5 mm before the experiment. The ultrasound cleaning bath SK5200H (China Shanghai Kudos ultrasound instrument Co., Ltd) with 30 24 15 cm internal dimensions, a volume of 10.8 l and 4 transducers was used in the experiment, operating at a frequency of 59 kHz with input power of 160 W. The extraction vessel was put at the center of the bath in the process of the experiment. 2.2. Ultrasound-assisted extraction Xyloglucan was extracted sequentially as was described by Watt et al. elsewhere [8]. Typically alcohol insoluble residue was suspended in 0.1 M KOH solution (with 8 mM NaBH4) for 4 h, then the slurry was centrifuged. Then 3 g residue was extracted by KOH (with 26 mM NaBH4) with an ultrasound cleaning bath at sonic power of 160 W. The soluble extracts were collected by centrifugation (15,000g) and adjusted pH 5.0 by addition of glacial acetic acid followed by dialyse, concentrate in a rotary evaporator and then freeze-dried to obtain the crude xyloglucan. The extraction conditions for dierent extracts referred to the experiment design and all experiments were duplicated.

2.3. Colorimetric analysis of xyloglucan from apple pomace [22] The xyloglucan content was estimated by an iodine-staining assay. One milliliter 5% xyloglucan fractions were mixed with Gramstain (750 ll, 6.6 g of KI and 3.8 g of iodine (I2) per litre of water) and then sodium sulfate (5 ml, 0.2 g/ml) added. The mixture was kept for 1 h at room temperature and then the absorbance at 620 nm was measured. 2.4. The eect of ultrasound-assisted extraction (UAE) time on the extraction of xyloglucan from apple pomace (XAP) In order to study the ultrasound eect compared with the conventional method and determine the optimization levels of ultrasound-assisted extraction time, 3 g apple pomace was ultrasound-assisted extracted for 0.25, 0.5, 1, 2, 3, 4 and 5 h by 100 ml 4 M KOH (with 26 mM NaBH4), respectively. Three gram apple pomace was also extracted for 16 h without ultrasound-assisted by 100 ml 4 M KOH (with 26 mM NaBH4). 2.5. Response surface methodology The combination eect of liquid:solid ratio, KOH concentration and ultrasound-assisted extraction time was studied by response surface methodology. For this purpose, a central-composite rotary design (CCRD) with three variables at ve levels was used to study the response pattern and to determine the optimum combination of variables. The eect of the independent variables X1 (liquid:solid ratio, R), X2 (KOH concentration, C) and X3 (ultrasound time, t) at ve variation levels (Table 1) in the extraction process was shown in Table 2. Six replicates (treatment 1520) at the center of the design were used to allow for estimation of a pure error sum of squares. Experiments were randomized in order to maximize the eects of unexplained variability in the observed responses due to extraneous factors. The variables were coded according to Eq. (1): xi X i X i =DX i ; where xi Xi Xi DXi 1

dimensionless value of an independent variable; real value of an independent variable; real value of an independent variable at the center point; step change.

The specic codes are x1 R 35=5 and x2 C 3:0=0:5 3 2

C. Fu et al. / Ultrasonics Sonochemistry 13 (2006) 511516 Table 1 Independent variable values of the process and their corresponding levels Independent variables Symbol Uncodied Liquid:solid ratio (v/w) KOH concentration (M) Ultrasound time (h) X1 X2 X3 Codied x1 x2 x3 Levels 2 25:1 2.0 2.0 1 30:1 2.5 2.5 0 35:1 3.0 3.0 1 40:1 3.5 3.5 2 45:1 4.0 4.0

513

and x3 th 3:0=0:5 2.6. Statistical analyses The absorbance at 620 nm was measured, after xyloglucan-containing fractions were mixed with the Gramstain. A second-order polynomial equation was used to t the experimental data given in Table 2. The model proposed for the response (Yi) was given below: Y i b0 b1 x1 b2 x2 b3 x3 b11 x2 b22 x2 b33 x2 1 2 3 b12 x1 x2 b13 x1 x3 b23 x2 x3 ; where Yi is the predicted response, b0 is the value of the tted response at the center point of the design, bi, bii and bij are the linear, quadratic and cross-product terms, respectively. The proportion of variance explained by the polynomial models obtained was given by the multiple coecient of determination, R2. The signicance of each coecient was determined using the Student t-test and p-value. The behavior of the surface was investigated for the response function (Yi) using the regression equation (5). A graphic 4

technique was used [23,24] to deduce workable optimum conditions by xing one variable at predetermined optimum condition. The optimum condition was veried by conducting experiments under these conditions. Responses were monitored and results were compared with model predictions. In order to visualize the relationship between the response and experimental levels of each factor and to deduce the optimum conditions, the tted polynomial equation was expressed as surface and contour plots [25].The computer software used for this study was STATISTICA, version 6.0, by Statsoft Inc. 3. Results and discussion 3.1. The eect of ultrasound-assisted extraction (UAE) time on the extraction of xyloglucan from apple pomace (XAP) Fig. 1 shows the eect of UAE time on the extraction of XAP. The results indicated that the extraction of XAP was increased with the increase of UAE time. UAE reached a high point in 4 h. If UAE time was longer than 4 h, the extraction of xyloglucan was decreased with the increase of time. It was found that the ultrasound-assisted extraction of xyloglucan was about three times faster than the traditional extraction method. 3.2. Response surface methodology In the second part of the study, the eect of liquid:solid ratio, KOH concentration and ultrasound time on the ultrasound-assisted extraction of xyloglucan from apple pomace were evaluated by response surface methodology in order to nd a optimal use of ultrasound for the extraction of xyloglucan.

Table 2 CCRD three variables with the observed responses and predicted values Treat Variable levels x1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 1 1 1 1 1 1 1 1 2 2 0 0 0 0 0 0 0 0 0 0 x2 1 1 1 1 1 1 1 1 0 0 2 2 0 0 0 0 0 0 0 0 x3 1 1 1 1 1 1 1 1 0 0 0 0 2 2 0 0 0 0 0 0 Experimental (Y0) 1.964 1.625 1.939 1.488 2.225 1.817 1.761 1.922 1.203 1.275 2.041 2.233 1.986 1.438 1.942 1.933 1.938 1.951 1.937 1.939 Predicted (Yi) 1.985 1.468 1.903 1.615 2.005 1.760 1.824 1.808 1.179 1.392 2.201 2.167 2.025 1.492 1.956 1.956 1.956 1.956 1.956 1.956 Y0 Yi 0.021 0.157 0.036 0.127 0.220 0.057 0.063 0.114 0.024 0.117 0.160 0.066 0.039 0.054 0.014 0.023 0.018 0.005 0.019 0.017

Fig. 1. Eect of UAE time on the extraction of xyloglucan.

514

C. Fu et al. / Ultrasonics Sonochemistry 13 (2006) 511516

The following regression equation, which is an empirical relationship between the absorbance and the test variable in coded unit as given in Eq. (5), was obtained with the application of RSM. Y 1:9556 0:0533x1 0:0086x2 0:1333x3 0:0246x1 x2 0:0679x1 x3 0:0571x2 x3 0:1333x2 0:0570x2 0:0492x2 . 1 2 3 5

For each of the observed values, Y0 was compared with the predicted value, Yi calculated from the model. The multiple coecients of correlation R = 0.9453 indicated a close agreement between experimental and predicted values of the xyloglucan yield. The signicance of each coecient was determined using the Student t-test and p-value in Table 3. The corresponding variables will be more signicant if the absolute t-value becomes larger and the p-value becomes smaller [26]. It can be seen that the variable with the largest eect was the quadratic of liquid:solid ratio (x1x1), followed by the linear term of ultrasound time (x3), and the quadratic of KOH concentration, (x2x2). The factor t-test value (6.6075) and p-value (p = 0.0001) correspond to x1x1, while the t-test values for x3 and x2x2 are smaller at 4.1960 and 2.2503, respectively, but the p values are still signicant at p = 0.0018 and 0.0482, respectively. The total determination coecient R2 = 89.35% implied that the sample variations of 89.35% for the yield of xyloglucan were attributable to the independent variables, viz. liquid:solid ratio, KOH concentration and ultrasoundassisted extraction time. The 3D surface curves were drawn to illustrate the main and interactive eects of the independent variables on the dependent one. These graphs were drawn by imposing a constant value (i.e., the central points) to one independent variable [27]. The eect of liquid:solid ratio, KOH concentration and ultrasound time on response are depicted (Table 3) by the coecients of second-order polynomials. The response surfaces based on these coecients are shown in Figs. 24 with one variable kept at optimum level and the other two varied within the experimental range.
Table 3 Signicance of regression coecient for the yield of xyloglucan Variables Constant x1 (Liquid:solid ratio) x2 (KOH concentration) x3 (Ultrasound time) x1x1 x2x2 x3x3 x1x2 x1x3 x2x3 R, 0.9453; R2, 89.35%. Regression coecient 1.9556 0.0533 0.0086 0.1333 0.1675 0.0570 0.0492 0.0246 0.0679 0.0571

Fig. 2. 3D graphic surface optimization of xyloglucan yield versus liquid:solid ratio and KOH concentration.

Fig. 2 shows the eect of KOH concentration and liquid:solid ratio on xyloglucan yield, a quadratic eect for both liquid:solid ratio and KOH concentration on the response could be observed. The eect of KOH concentration and ultrasound-assisted extraction time is depicted with Fig. 3, KOH concentration exerted a quadratic eect on xyloglucan production, whereas ultrasound-assisted extraction time had a linear eect approximately. Fig. 4 depicts the eect of liquid:solid ratio and ultrasoundassisted extraction time, both also exerted a quadratic eect. Interaction between two varies could be observed in Figs. 24. Then a numerically feasible combination of factors for limiting response values (constraints) was determined simultaneously by using Newtons method [28]. In the application of the method, central-composite rotary design data were taken into consideration and a formulation of the conditions were found as a liquid:solid ratio of 34.4:1

Standard error 0.0507 0.0318 0.0318 0.0318 0.0253 0.0253 0.0253 0.0449 0.0449 0.0449

Computed t-value 38.5791 1.6780 0.2695 4.1960 6.6075 2.2503 1.9419 0.5481 1.5106 1.2714

Signicance level, p-value 0.0000 0.1243 0.7930 0.0018 0.0001 0.0482 0.0808 0.5957 0.1618 0.2324

C. Fu et al. / Ultrasonics Sonochemistry 13 (2006) 511516

515

on ultrasound-assisted extraction of xyloglucan needs to progress in a pilot scale. 4. Conclusions The application of ultrasound during the preparation of xyloglucan from apple pomace had positive eect on the extraction yield. It was found that the ultrasound-assisted extraction of xyloglucan from apple pomace was about three times faster than the traditional extraction method. The production of xyloglucan from apple pomace was optimized using Statsoft Statistica version 6.0 software. The three independent variables involved in the optimization are liquid:solid ratio (x1), KOH concentration (x2) and ultrasound-assisted extraction time (x3). The Student t-test, p-value indicated that the variable with the largest eect was the quadratic of liquid:solid ratio which was followed by the linear term of ultrasound time, and the quadratic of KOH concentration. A liquid:solid ratio of 34.4:1 (v/w), a KOH concentration of 3.3 M and an ultrasound-assisted extraction time of 2.5 h were found to be optimal for xyloglucan extraction from apple pomace. Acknowledgement We gratefully acknowledge the nancial support received in the form of a research grant (Project No. 6041003) from Beijing Natural Science Foundation, the Peoples Republic of China. References
[1] K.J. Carson, J.L. Collins, M.P. Peneld, Journal of Food Science 59 (1994) 1213. [2] Y.D. Hang, Food Technology 41 (1987) 115. [3] M.J. Kennedy, Australasian Biotechnology 4 (1994) 43. [4] L. Yeap Foo, Y. Lu, Food Chemistry 64 (1999) 511. [5] Y. Lu, L. Yeap Foo, Food Chemistry 68 (2000) 81. [6] Y. Lu, L.Y. Foo, Food Chemistry 59 (1997) 187. [7] C.D. May, Carbohydrate Polymers 12 (1990) 79. [8] D.K. Watt, D.J. Brasch, D.S. Larsen, L.D. Melton, Carbohydrate Polymers 39 (1999) 165. [9] S. Aldington, S.C. Fry, Advances in Botanical Research 19 (1993) 1. [10] Y. Kato, J. Uchida, S. Ito, Y. Mitsuishi, International Congress Series 1223 (2001) 161. [11] N.C. Carpita, D.M. Gibeaut, Plant Journal (1993) 31. [12] W.D. Bauer, K.W. Talmadge, K. Keegstra, P. Albersheim, Plant Physiology 51 (1973) 174. [13] T. Hayashi, M.P.F. Marsden, D.P. Delmer, Plant Physiology 83 (1987) 384. [14] C.M.G.C. Renard, C. Lemeumier, J.-F. Thibaut, Carbohydrate Polymers 28 (1995) 209. [15] M. Vinatoru, Ultrasonics Sonochemistry 8 (2001) 303. [16] Z. Hromadkova, A. Ebringerova, Ultrasonics Sonochemistry 10 (2003) 127. [17] Z. Hromadkova, A. Ebringerova, P. Valachovic, Ultrasonics Sonochemistry 5 (1999) 163. [18] I.N. Ranchev, N.A. Kirtchev, C.G. Kratchanov, International Journal of Food Science and Technology 23 (1988) 337. [19] Z. Hromadkova, J. Kovacikova, A. Ebringerova, Industrial Crops and Products 9 (1999) 101.

Fig. 3. 3D graphic surface optimization of xyloglucan yield versus KOH concentration and ultrasound time.

Fig. 4. 3D graphic surface optimization of xyloglucan yield versus liquid:solid ratio and ultrasound time.

(v/w), a KOH concentration of 3.3 M and an ultrasoundassisted extraction time of 2.5 h. The calculated value was 2.122 in agreement with the experimental mean value (2.131) of ve determinations. A lot of reports on the eort to utilize apple pomace, a global major low value waste, were documented. It was of interest to test the eect of ultrasound on extracting xyloglucan from apple pomace, which, until now, has not been investigated. Further work

516

C. Fu et al. / Ultrasonics Sonochemistry 13 (2006) 511516 [24] M. Giovanni, Food Technology 37 (11) (1983) 41. [25] R. Triveni, T.R. Shamala, N.K. Rastogi, Process Biochemistry 36 (2001) 787. [26] N.A.S. Amin, D.D. Anggoro, Fuel 83 (2004) 487. [27] Q. Li, C. Fu, Food Chemistry 92 (4) (2005) 701. [28] H. Ozoglu, A. Bayindirli, Food Control 13 (2002) 213.

[20] K. Theodore, T. Panda, Enzyme and Microbial Technology 11 (17) (1995) 1043. [21] S.N. Deming, Quality by designPart 5, Chemtech 20 (1990) 118. [22] T. Doco, P. Williams, M. Pauly, M.A. ONeillb, P. Pellerin, Carbohydrate Polymers 53 (2003) 253. [23] J.D. Floros, M.S. Chinnan, Food Technology 42 (1988) 71.