Carbohydrate Polymers 86 (2011) 1089–1092

Contents lists available at ScienceDirect

Carbohydrate Polymers
journal homepage: www.elsevier.com/locate/carbpol

Short communication

Optimization of enzyme-assisted extraction of the Lycium barbarum

polysaccharides using response surface methodology
Jiao Zhang, Shaoyi Jia, Yong Liu ∗ , Songhai Wu, Jingyu Ran
School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, China

a r t i c l e i n f o a b s t r a c t

Article history: In this study, an efficient enzyme-assisted extraction procedure was developed and optimized for the
Received 14 April 2011 extraction of polysaccharide from Lycium barbarum. The concentrations of papain and cellulose were
Received in revised form 8 June 2011 optimized by an orthogonal experimental design and the optimal results were cellulose concentration
Accepted 9 June 2011
2.0% (wt.% of L. barbarum powder) and papain concentration 1.0%. The response surface methodology
Available online 16 June 2011
(RSM) based on Box Behnken Design (BBD) was employed to further optimize the extraction conditions
which included extraction time, extraction temperature and pH. The optimal conditions were as fol-
Keywords:
lows: extraction time 91 min, extraction temperature 59.7 ◦ C and pH 5.0. Under these conditions, the
Lycium barbarum polysaccharides
Enzyme-assisted extraction
experimental yield was 6.81 ± 0.10%, which was well matched with the value predicted by the model.
Response surface methodology © 2011 Elsevier Ltd. All rights reserved.
Optimization

1. Introduction further to optimize the extraction conditions for the yield of LBP by
response surface methodology (RSM). To the best of our knowledge,
Lycium barbarum L. in the family Solanaceae has been used as a there were no reports of systematic studies of enzyme-assisted
kind of traditional Chinese herbal medicine for thousands of years extraction process of LBP with compound enzymes.
(Li, Ma, & Liu, 2007). L. barbarum polysaccharide (LBP), the main
bioactive component of L. barbarum, has a large variety of bioactiv-
2. Materials and methods
ities, such as enhancing immunity, anti-tumor, protecting hepatic
function, anti-aging, antivirus, reducing neuronal damage, and low-
2.1. Materials
ering blood glucose (Chan et al., 2007; Wang et al., 2010). In order
to realize all its promising applications, there is a strong need to
Fruits of L. barbarum, family Solanaceae, were purchased from
supply the market with high-quality LBP. However, compared with
a herb market in Tianjin, China. Samples were ground and passed
the numerous studies of LBP biological activity, little attention was
through 100 mesh screen.
devoted to extraction of LBP.
Cellulase from trichoderma viride (11,000 U/mg), and Papain
Hot-water technology is the traditional extraction method for
(P3250, 0.5–2 U/mg) were supplied by Sigma Chemical Company.
polysaccharides, and has been widely investigated (Hou & Chen,
d-Glucose was obtained from Tianjin Zhong Ao Tian Yuan Company,
2008; Sun, Liu, & Kennedy, 2010; Wang, Luo, & Ena, 2007). Nonethe-
and the other chemicals were purchased from Tianjin Qian Cheng
less, this technology usually requires long extraction time and high
Wei Ye Company. All chemicals were reagent grade or better.
temperatures. Recently, enzyme-assisted extraction, which is con-
sidered as a mild, efficient and environmentally friendly method,
has been employed to extract different compounds from plants, 2.2. Methods
and has been proved to be effective in improving the yield of the
target component (Li, Smith, & Hossain, 2006; Moura et al., 2008; 2.2.1. Extraction procedure
Santamaria et al., 2000). The powder of L. barbarum (20 g) was double extracted with
The objective of this study was to investigate the extraction pro- petroleum ether at 90 ◦ C for 3 h each time to remove lipids, some
cess of LBP with compound enzymes (cellulase and papain), and colored materials, and oligosaccharides under reflux in the Soxhlet
set. After being vacuum dried at 60 ◦ C for 12 h, the defatted pow-
der (1 g) was extracted with 30 ml of compound enzymes solution
at the given concentration in a designed temperature, extraction
∗ Corresponding author. Tel.: +86 22 87401961; fax: +86 22 87401961. time and pH. The extract was concentrated, and then precipi-
E-mail address: jiaosherry@yahoo.cn (Y. Liu). tated by adding ethanol (12 h, 4 ◦ C) to give a final concentration of

0144-8617/$ – see front matter © 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carbpol.2011.06.027
1090 J. Zhang et al. / Carbohydrate Polymers 86 (2011) 1089–1092

Table 1 Table 3
Results and analysis of orthogonal experiment. Box Behnken design matrix (in coded level of three variables) and response values
for the yield of LBP.
No. (A) Cellulose (B) Papain Polysaccharide yield (%)
concentration concentration Number x1 x2 x3 Polysaccharide yield (%)
(wt.% of L. (wt.% of L.
1 −1 −1 0 5.89
barbarum powder) barbarum powder)
2 −1 1 0 5.57
1 1 (1.5%) 1 (0.5%) 4.24 3 1 −1 0 6.03
2 1 2 (1.0%) 5.17 4 1 1 0 5.37
3 1 3 (1.5%) 4.29 5 −1 0 −1 4.24
4 2 (2.0%) 1 6.03 6 −1 0 1 5.97
5 2 2 6.64 7 1 0 −1 4.53
6 2 3 6.05 8 1 0 1 6.06
7 3 (2.5%) 1 5.47 9 0 −1 −1 4.56
8 3 2 5.68 10 0 −1 1 5.65
9 3 3 5.38 11 0 1 −1 3.54
K1 13.7 15.74 12 0 1 1 6.16
K2 18.72 17.49 13 0 0 0 6.64
K3 16.53 15.72 14 0 0 0 6.74
R 5.02 1.77 15 0 0 0 6.51
16 0 0 0 6.74
17 0 0 0 6.69
Table 2
Independent variables and their levels used in the response surface design.

Independent variables Factor levels

yk was the response function; bk0 , bki , bkii , and bkij represented the
constant regression coefficients of the model, and xi and xj were the
−1 0 1
independent variables.
X1 : extraction time (min) 60 90 120
X2 : temperature (◦ C) 50 60 70
X3 : pH 3.6 4.6 5.6
3. Results and discussion

3.1. Orthogonal analysis of compound enzymes concentrations

80% (v/v). The precipitate was collected and dried to obtain crude
polysaccharides. The polysaccharides content was measured by In the enzyme-assisted extraction process, various parameters
phenol–sulfuric acid method using d-glucose as a standard (Dubois, played a role in increasing extraction yield of LBP. In order to
Gilles, Hamilton, Rebers, & Smith, 1956). The percentage LBP extrac- learn more about a range of factors which affect the process and
tion yield (%) was calculated with the formula of y (%) = c/w × 100%, reduce number of experimental trials, an orthogonal test design
where c was the polysaccharides content of extraction, and w rep- was applied to obtain the optimal compound enzymes concentra-
resented dried sample weight (1 g). tions, and then extraction time, temperature and pH were further
optimized using RSM.
Orthogonal analysis of results of L9 (3)4 was showed in Table 1.
2.2.2. Orthogonal test design of compound enzymes
The results indicated the influence (R) of extraction factors on the
concentrations
extraction yield was RA > RB . The optimal combination parameters
An orthogonal L9 (3)4 test design was applied to investigate the
were A2B2, namely, cellulose concentration (2.0%) and papain con-
optimal concentrations of cellulase and papain with the following
centration (1.0%). In these conditions, the maximum extraction
extraction conditions: pH 4.6, temperature 60 ◦ C, and extraction
yield of 6.64% was achieved.
time 90 min. As seen from Table 1, the extraction experiment was
carried out with 2 factors and 3 levels, and the range of each factor
was based on the results of preliminary single-factor experiments. 3.2. Optimization of the procedure by RSM
The yield (%) of LBP was the dependent variable and was obtained
following the method in Section 2.2.1. The design matrix and the corresponding results of RSM experi-
ments were shown in Table 3. The mathematical model describing
the extraction yield of LBP (y) as a function of the coded indepen-
2.2.3. Box Behnken design and statistical analysis
dent variables in the selected ranges was given by the following
On the basis of orthogonal test of compound enzymes concen-
equation:
trations, RSM was used to further optimize the enzyme-assisted
extraction conditions of LBP. A BBD with three independent vari- y = 6.664 + 0.04x1 − 0.18625x2 + 0.87125x3 − 0.36325x12
ables (X1 , extraction time; X2 , temperature; X3 , pH) at three levels
was performed. The coded and uncoded (actual) levels of the inde- − 0.085x1 x2 − 0.58575x22 − 0.05x1 x3 + 0.3825x2 x3
pendent variables are presented in Table 2, which were based on the − 1.10075x32 (2)
results of preliminary experiments. As seen from Table 3, the whole
design consisted of 17 experimental points, and the experiment
was carried out in random order. All trials were performed in trip- After the response surface regression procedure, the value of
licate. The extraction yield was treated as response. SAS (Version the determination coefficient (R2 = 0.9923) showed a good fit to the
9.2, USA) software package was used to generate the experimental model and indicated that only 0.77% of the total variations were
design, statistical analysis and regression model. not explained by the model. The value of the adjusted determina-
Data from BBD were analyzed by multiple regressions to fit the tion coefficient (Adj. R2 = 0.9692) also confirmed that the model was
following quadratic polynomial model: highly significant. A very low value of coefficient of the variation
(C.V.) (2.24%) clearly indicated a very high degree of precision and

3 
3 
3
a good deal of reliability for the experimental values.
yk = bk0 + bki xi + bkii xi2 + bkij xi xj (1)
The significance of each coefficient of Eq. (2) was determined
i=1 i=1 i<j=2 using P-value (Table 4). The smaller was the value of P, the more
 J. Zhang et al. / Carbohydrate Polymers 86 (2011) 1089–1092 1091

Fig. 1. Response surface plots (A, C, and E) and contour plots (B, D and F) showing the effects of variables (X1 : extraction time, min; X2 : temperature, ◦ C; X3 : pH) on extraction
yield of LBP (y).

significant was the corresponding coefficient (Hou & Chen, 2008). the contour plots indicated different interactions between the vari-
It can be seen that the linear coefficients (x1 , x2 , x3 ), quadratic term ables. Elliptical contours were obtained when there was a perfect
coefficients (x12 , x22 , x32 ) and cross product coefficient (x2 × x3 ) were interaction between the independent variables while circular con-
significant, with very small P values (P < 0.05). The results indicated tour plot indicated otherwise (Guo, Zou, & Sun, 2010). As shown in
that the extraction time, temperature and pH were all significantly Fig. 1(A) and (B), the extraction yield of LBP increased with increase
correlated with the LBP extraction yield. of temperature (x2 ) from 50 to 59.66 ◦ C, but beyond 59.66 ◦ C,
The graphical representations of the regression Eq. (2), and the extraction yield of LBP decreased with increasing temperature as
contour plots are presented in Fig. 1. Two variables within the higher temperatures reduced the enzyme activity. When temper-
experimental range were depicted in one 3D surface plot while the ature was set, extraction yield of LBP was also found to increase
other variable was kept constant at zero level. Different shapes of with increase of extraction time (x1 ) from 60 to 90.97 min, and then
1092 J. Zhang et al. / Carbohydrate Polymers 86 (2011) 1089–1092

Table 4 extraction temperature 59.7 ◦ C and pH 5.0. Under the optimal con-
Test of significance for regression coefficients.
ditions, the experiment extraction yield of LBP was 6.81 ± 0.10%,
Effect Estimate Stand error t-Ratio P-value which was agreed closely with the predicted value.
x1 0.04 0.020353 4.85 0.0019a
x2 −0.18625 0.082443 6.48 0.0003a Acknowledgment
x3 0.87125 0.714863 12.38 <0.0001a
x1 × x1 −0.36325 0.000068995 −5.85 0.0006a This research was financially supported by the Construction of
x1 × x2 −0.085 0.000212 −1.33 0.2239
x1 × x3 −0.05 0.002124 −0.78 0.4583
Technical Platform for Pharmaceutical Separation and Refinement
x2 × x2 −0.58575 0.000621 −9.43 <0.0001a (No. 2009ZX09301-008).
x2 × x3 0.3825 0.006371 6.00 0.0005a
x3 × x3 −1.10075 0.062095 −17.73 <0.0001a References
a
Significant coefficient (P < 0.05).
Chan, H. C., Chang, R. C. C., Ip, A. K. C., Chiu, K., Yuen, W. H., Zee, S. Y., et al. (2007).
Neuroprotective effects of Lycium barbarum Lynn on protecting retinal ganglion
there was a little decrease when the extraction time continued to cells in an ocular hypertension model of glaucoma. Experimental Neurology, 203,
increase. From Fig. 1(C) and (D), it indicated that extraction yield of 269–273.
Dubois, M., Gilles, K. A., Hamilton, J. K., Rebers, P. A., & Smith, F. (1956). Colorimetric
LBP increased rapidly when the pH (x3 ) increased from 3.6 to 4.99, method for determination of sugars and related substances. Analytical Chemistry,
and then began to decrease. In Fig. 1(E) and (F), maximum extrac- 28, 350–356.
tion yield of LBP (6.83%) was achieved when extraction temperature Guo, X., Zou, X., & Sun, M. (2010). Optimization of extraction process by response
surface methodology and preliminary characterization of polysaccharides from
and pH were 59.66 ◦ C and 4.99, respectively. Phellinus igniarius. Carbohydrate Polymers, 80, 344–349.
It could be concluded that the optimal extraction conditions Hou, X. J., & Chen, W. (2008). Optimization of extraction process of crude polysac-
of LBP were extraction time 90.97 min, extraction temperature charides from wild edible BaChu mushroom by response surface methodology.
Carbohydrate Polymers, 72, 67–74.
59.66 ◦ C, pH 4.99, and the model predicted a maximum response of
Li, B. B., Smith, B., & Hossain, M. M. (2006). Extraction of phenolics from citrus peels
6.84%. Among the three extraction parameters studied, pH was the II. Enzyme-assisted extraction method. Separation and Purification Technology,
most significant factor to affect the extraction yield of LBP, followed 48, 189–196.
Li, X. M., Ma, Y. L., & Liu, X. J. (2007). Effect of the Lycium barbarum polysaccharides
by temperature and extraction time.
on age-related oxidative stress in aged mice. Journal of Ethnopharmacology, 111,
The suitability of the model equation for predicting the opti- 504–511.
mum response values was tested by confirmation experiments. Moura, J. M. L. N., Campbell, K., Mahfuz, A., Jung, S., Glatz, C. E., & Johnson, L. (2008).
The experiment yield of LBP was 6.81% (±0.10%), which was Enzyme-assisted aqueous extraction of oil and protein from soybeans and
cream de-emulsification. Journal of the American Oil Chemists Society, 85, 985–
not significantly different from the predicted value of 6.84%. 995.
These data proved that the model designed in this study was Santamaria, R. I., Reyes-Duarte, M. D., Barzana, E., Fernando, D., Gama, F. M., Mota,
valid. M., et al. (2000). Selective enzyme-mediated extraction of capsaicinoids and
carotenoids from chili guajillo puya (Capsicum annuum L.) using ethanol as sol-
vent. Journal of Agricultural and Food Chemistry, 48, 3063–3067.
4. Conclusion Sun, Y. X., Liu, J. C., & Kennedy, J. F. (2010). Application of response surface method-
ology for optimization of polysaccharides production parameters from the roots
of Codonopsis pilosula by a central composite design. Carbohydrate Polymers, 80,
Enzyme-assisted extraction was a green and efficient extrac- 949–953.
tion technique that could be used to improve the extraction yield Wang, J. M., Hu, Y. L., Wang, D. Y., Zhang, F., Zhao, X. N., Abula, S., et al. (2010). Lycium
of LBP. The optimal compound enzymes concentrations (cellulose barbarum polysaccharide inhibits the infectivity of Newcastle disease virus to
chicken embryo fibroblast. International Journal of Biological Macromolecules, 46,
concentration 2.0%, papain concentration 1.0%) were obtained by 212–216.
an orthogonal test design. Then extraction time, temperature and Wang, Z. J., Luo, D. H., & Ena, C. (2007). Optimization of polysaccharides extrac-
pH were chosen to further optimize LBP extraction by RSM. These tion from Gynostemma pentaphyllum Makino using uniform design. Carbohydrate
Polymers, 69, 311–317.
optimal conditions were as the following: extraction time 91 min,