Optimization of aqueous enzymatic extraction of oil from shrimp processing by-products using response surface methodology

Several studies have investigated to obtain carotenoids (Sowmya & Sachindra, 2012; Sachindra & Mahendrakar, 2005), astaxanthin (Pu et al., 2010; Handayani et al., 2008), polyunsaturated fatty acids (Treyvaud Amiguet et al., 2012), protein (Ferrer et al., 1996), peptides (Zhao et al., 2013; Huang et al., 2011), antioxidants (Seymour et al., 1996), chitin biopolymers (Pinelli Saavedra et al., 1998) and salt-fermented shrimp sauce (Kim et al., 2005) from shrimp processing by-products. However, to our knowledge, there are few studies focused on shrimp oil extraction from shrimp processing by-products.


Introduction
Shrimps are processed in seafood export processing units, generating large number of by-products, such as shrimp head and shrimp shells.Utilization of such large quantities of shrimp processing discards, such as for oil recovery, would not only reduce the disposal problems associated with these wastes, but also enhance the economy of shrimp processing.
Generally, conventional methods of producing edible oil from oilseeds were to use expeller pressing and organic solvent extraction.Mechanical extraction can result in low oil recovery and denatured proteins, and the use of organic solvents may cause solvent residue and non-friendly to environment.The AEE for oil extraction as an emerging technology, enables simultaneous recovery of oil and protein from most oil materials (Li et al., 2013a).Enzymes are useful for the extraction of oil due to their high efficiency and specificity.Enzyme preparations hydrolyse and rupture the cell wall constituents and improve the release of intracellular contents.AEE technology has been applied for oils extraction from seed crops, such as pumpkin (Jiao et al., 2014), oil palm fruit (Teixeira et al., 2013), bayberry (Zhang et al., 2012), peanut (Li et al., 2011;Jiang et al., 2010), watermelon (Sui et al., 2011), sesame (Latif & Anwar, 2011), flax (Long et al., 2011), wheat (Li et al., 2010), and soybean (Towa et al., 2010).
Large-scale production required consumption of massive fresh raw shrimp, and at the same time, produced a large number of shrimp waste.Therefore, effective method of oil extraction from shrimp processing by-products has higher economic value.
However, to the best of our knowledge, there are few studies about the AEE of oil from shrimp processing by-products.Therefore, the objective of this study was to investigate the optimal AEE conditions of oil from shrimp processing by-products by response surface methodology (RSM).And the fatty acid compositions of AEE extracted shrimp processing by-products oil was studied.

Materials
Shrimp processing by-products were from Yingzhou Seafood Corporation (Zhoushan city, Zhejiang province, China).After dried by vacuum at 75 °C, they were ground in a powder, and dispersed through 60 mesh sieve.The powders were sealed in plastic containers and stored in a refrigerator at 4 °C until extraction.

Aqueous enzyme extraction
The shrimp processing by-product powders were commixed with 0.2 M phosphate buffer solution at a designed ratio.Enzyme was added and the mixtures were incubated in a water bath at proper temperature for suitable time.The suspension was centrifuged (4 °C, 10,000 g) for 20 min.The layers of free oil and emulsion phase were collected separately.Total free oil and the emulsion phase was then demulsified by foam suppressor HS-508 and separately further centrifuged to get free oil.Total oils were collected and weighed.

Shrimp oil extraction yield calculation
The content of crude fat of shrimp processing by-products was measured by Soxhlet extraction method (Abdulkarim et al., 2005) (Equation 1).

( )
the weight of the shrimp oil shrimp oil extraction yield % = ×100% the weight of the crude fat (1)

GC-MS analyses of fatty acid compositions
The fatty acid compositions of shrimp oils were analysed by GC-MS.Prior to injection, the obtained oil was converted to its fatty acid methyl esters (FAME) through alkaline transmethylation by using KDH in methanol as a methylating agent (Li et al., 2013b).GC-MS analysis was performed using agilent 5975B GC-MS gas chromatography/mass spectrometer, equipped with an HP-5 silica capillary column (30 m×0.32mmO.D.; film thickness 0.2 μm).The detail operating conditions were carried out as followingI: helium gas flow rate 3 mL/min; split ratio 1I:10; injector temperature 250 °C; injection volume 1 μL; oven temperature progress from 110 to 230 °C at the rate of 15 °C/min; detector temperature 280 °C; ion source temperature 220 °C; ionisation mode used at electronic impact 70 eV; mass range 50-500 m/z.Odentification of chemical constituents of shrimp oils was based on the comparisons of their retention indices and mass spectra with publish data and computer matching the mass spectra fragmentation patterns with those stored in mass spectral library NOST05 provided by the software of GC-MS system.Relative percentage compositions of oils were calculated from the total ion chromatograms by a computerized integrator.

Experimental design and statistical analysis
The preliminary range of the extraction variables were determined through single factor experiments.Response surface methodology (RSM) based on central composite rotatable design (CCRD) was applied to evaluate the effects of four independent variables, enzyme amount (X 1 ), liquid/solid ratio (X 2 ), hydrolysis time (X 3 ), hydrolysis temperature (X 4 ), and their interaction on the measured response, extraction yield (Y).The independent variables were coded at five levels (-2, -1, 0, +1, +2), and the complete design consisted of 31 experimental points including 7 replications of the centre points.The coded levels of the independent variables used in the RSM design were listed in Table 1.
The second-order polynomial model proposed for the response surface analysis of the designed experiment was explained by Equation 2I: Where Y is the extraction yield; β 0 , β i , β ii and β ij are the coefficients of intercept, linear, quadratic, and interactive terms respectively; while X i and X j are the coded values of the four independent variables.To analyze the multiple regression and variance, a regression equation between variables and response, and numerical optimum the procedure, the SAS software program (version 9.0, SAS Onstitute Onc., Cary, NC, USA) was employed.

Choice of enzyme types
As presented in Figure 1, more oil was recovered using hydrolytic enzymes (66.4-76.1%),when compared with the control (40.6%, no enzyme).The recovery rate of shrimp oil was higher significantly when using flavor protease, which might be attributed to breakdown of the protein networks of oleosin-based membranes that surround lipid bodies, and in turn it liberated more oils.Ot had also proved that the use of protease resulted in higher oil yield than without enzyme treatment (Mat Yusoff et al., 2016;Latif et al., 2008).

Optimization of AEE by RSM
Aqueous enzymatic extraction conditions of oil from shrimp processing by-products were optimized by RSM.Dn the basis of the experimental results of CCD (Table 1) and regression analysis, a second-order polynomial equation was established to estimate the relationship between the oil extraction yield and variables.The model could be expressed as (Equation 3)I: Where, the X 1 , X 2 , X 3 and X 4 correspond to the coded values of the four independent variables (enzyme amount, liquid/solid ratio, hydrolysis time and temperature).
The results of the analysis of the models were summarized in Table 2 and Table 3.The determination coefficient (R 2 = 0.9810) was showed by ANDVA of the quadratic regression model, indicating that only 1.9% of the total variations were not explained by the model.At the same time, a very low value coefficient of 2.45 of the variation (CV) clearly indicated a very high degree of precision and a good deal of reliability of the experimental values.The model was found to be adequate for prediction within the range of experimental variables.
Three-dimensional (3D) response surface plots presenting the effects of the four independent variables on the response were shown in Figure 2.Each of the plots in Figure 2 was drawn to illustrate two of the variables and their interaction affecting the dependent variable with another two variables fixed (0 level).
They provide a means of visualizing relationship between the responses and experimental levels of each variable and the type of interactions between the two test variables.The shapes of the contour plots, circular or elliptical, indicate whether the mutual interactions between the variables are significant or not.Circular contour plots indicate that the interactions between the corresponding variables are negligible, while elliptical contour plots indicate that the interactions between the corresponding variables are significant.Through these 3D response surface and their respective contour plots, it is very easy and convenient to understand the interactions between two variables and to locate their optimum ranges.
As shown in Figure 2a, enzyme amount and liquid/solid ratio affected significantly extraction yield, and the oil extraction yield reached the highest when enzyme amount around at 2.0% and liquid/solid ratio at 8 ml/g.As shown in Figure 2b, extraction yield increased with rising temperature as the rate of enzyme-catalyzed reactions.However, at higher temperatures (50-70 °C), the yield decreased sharply, because the enzymes would be denatured at higher Figure 2C describes the interactive effect of hydrolysis time and enzyme amount.Hydrolysis time exhibited an important effect on the oil yield.As shown in Figure 2C and D, extraction yield increased with extended time at a given enzyme amount and liquid/solid ratio in an early stage of extraction.With the increasement of enzyme amount, the extraction yield rose at first, then decreased slightly when the enzyme amount reached its high levels.
Based on the mathematical predicted model, the optimal experimental conditions were as followingI: enzyme amount of 2.04% (w/w), liquid/solid ratio of 8.66 ml/g, hydrolysis time of 2.58 h and temperature at 48.97 °C.Considering the actual operation, enzyme amount, liquid/solid ratio, time and temperature were modified to 2.0%, 9.0 ml/g, 2.6 h and 50 °C, respectively.
The reliability of the theoretical model was verified under optimal parameters.A yield of 88.9 ± 1.0% was obtained from these experiments, which was a good fit for the value forecasted (89.6%) by the regression model.Therefore, the oil extraction conditions achieved by RSM were reliable and practical.
The types and contents of saturated fatty acids in our results were similar to the previously reported values (Sánchez-Camargo et al., 2012), but obtained different content in the EPA and DHA when compared to the krill oil (Colombo-Hixson et al., 2011;Phleger et al., 2002).The difference of fatty acids composition may due to the Farmed or wild, region, climatic conditions, species, and processing treatment (Sampaio et al., 2006;Harlioglu et al., 2016)

Conclusion
On this study, the oil was firstly extracted from shrimp processing by-products by aqueous enzymatic method.The extraction conditions were optimized by RSM.We concluded that extraction yield was dependent on all the linear terms (enzyme amount, liquid/solid ratio, time and temperature), and all the quadratics (enzyme amount, liquid/solid ratio, time and temperature), and cross product (the interactions between enzyme amount and liquid/solid ratio, liquid /solid ratio and temperature, liquid /solid ratio and time, enzyme amount and temperature).A polynomial regression model was used to describe the experimental results, and based on the proposed model, whose availability and accuracy was verified by validation experiments, the optimal extraction conditions for extraction yield were enzyme amount, liquid/solid ratio, time, temperature as 2.0% (w/w), 9.0 ml/g, 2.6h and 50 °C.Under the optimum conditions, the experimental extraction yield was 88.9%.Furthermore, the GC analysis showed that the shrimp oil composited by eleven fatty acids.

Table 1 .
Analytical factors and levels for RSM.

Table 2 .
The ANDVA results of RSM.

Table 3 .
Regression coefficients of the predicted quadratic polynomial model.