Bioresource Technology 101 (2010) 62076214

Contents lists available at ScienceDirect

Bioresource Technology
journal homepage: www.elsevier.com/locate/biortech

Optimization of physico-chemical properties of gelatin extracted from sh skin of rainbow trout (Onchorhynchus mykiss)
H. Shahiri Tabarestani a,*, Y. Maghsoudlou a, A. Motamedzadegan b, A.R. Sadeghi Mahoonak a
a b

Department of Food Science and Technology, Gorgan University of Agricultural Sciences and Natural Resources, Gorgan, Iran Department of Food Science, Sari Agricultural Sciences and Natural Resources University, PO Box 578, Sari, Mazandaran, Iran

a r t i c l e

i n f o

a b s t r a c t
Physico-chemical properties of gelatin extracted from rainbow trout (Onchorhynchus mykiss) skin were optimized using response surface methodology (RSM). Central rotatable composite design was applied to study the combined effects of NaOH concentration (0.010.21 N), acetic acid concentration (0.01 0.21 N) and pre-treatment time (13 h) on yield, molecular weight distribution, gel strength, viscosity and melting point of gelatin. Regression models were developed to predict the variables. Predict values of multiple response at optimal condition were that yield = 9.36%, a1 =a2 chain ratio = 1.76, b chain percent = 32.81, gel strength = 459 g, viscosity = 3.2 mPa s and melting point = 20.4 C. The optimal condition was obtained using 0.19 N NaOH and 0.121 N acetic acid for 3 h. The results showed that the concentration of H+ during pre-treatment had signicant effect on molecular weight distribution, melting point and gel strength. The concentration of OH had signicant effect on viscosity and for extraction yield, pretreatment time was the critical factor. 2010 Elsevier Ltd. All rights reserved.

Article history: Received 15 November 2009 Received in revised form 1 February 2010 Accepted 16 February 2010 Available online 25 March 2010 Keywords: Gelatin Fish skin Rainbow trout Physico-chemical properties RSM

1. Introduction Gelatin is functional gelling protein bio polymer, widely used in many industrial elds, such as food, material, pharmacy and photography, especially in the food and pharmaceutics industries for its unique chemical and physical properties (Yang et al., 2007). About 95% of commercial gelatins are derived from mammalian sources, mainly pig skin and cow hide and the remaining part, about 5% comes from bones of porcine and bovine, but for many socio-cultural reasons alternative sources, like sh skins, are highly demanded especially in halal and kosher food markets. Use of sh skin as an excellent raw material is interesting in the preparation of high-protein foods, beside helping to eliminate harmful environmental aspects and improve quality in sh processing. Since gelatin is derived from denatured collagen, its properties for a particular application is greatly inuenced not only by the species or tissue from which it is extracted, but also by the extraction process, which may depend on pH, temperature, and time during both pre-treatment and extraction process (Gmez-Guilln and Montero, 2001). So far sh gelatin has known limited application because the gels formed, tend to be less stable and to show weak rheological properties compared to gelatins extracted from land mammals. This has been largely related to a lower content of the imino acids proline and hydroxyproline in cold-water species, than
* Corresponding author. Tel./fax: +98 171 4426432. E-mail address: hoda.shahiri@yahoo.com (H.S. Tabarestani). 0960-8524/$ - see front matter 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2010.02.071

in warm blooded animals (Norland, 1990). Although collagen from cold-water sh species normally contains less imino acids than collagen from warm water sh species and mammalians, the contents of amino acids as well as other properties such as molecular weight distribution and gelatin viscosity, seem to be species specic (Gmez-Guilln et al., 2002; Gudmundsson, 2002; JohnstonBanks, 1990). However, sh gelatin from cold-water sh species is still preferred due to the greater availability of by-products (i.e. skin and bone) from which it can be manufactured (Haug et al., 2004). Many sh species have been investigated as raw materials for gelatin extraction and the properties of gelatin obtained from these sources have also been examined. There are, however, limited studies using formal optimization procedures for gelatin extraction, which is an important tool for understanding how processing conditions affect the nal product and for being able to get products with desired characteristics. Rainbow trout skin was chosen as the raw material for this project for several reasons. First, very few studies have been done on gelatin extraction from trout species. Second, the total world production of trout and it is commercial harvest is increasing in parts of the Caspian Sea. Thus, trout species especially rainbow trout are harvested in sufcient quantity for them to have the commercial potential for gelatin production. Rainbow trouts are usually processed into skinless llets (mainly for gelte sh, a European Jewish sh ball-like product), so there is an abundant quantity of raw skins available. Here, optimization of the gelatin extraction procedures and a study on the properties of rainbow trout skin gelatin could be helpful in

6208

H.S. Tabarestani et al. / Bioresource Technology 101 (2010) 62076214

rationalizing the use of sh residuals. The severity and duration of both acidic and alkaline pre-treatment requires a monitoring process to yield gelatin with better quality. This study was designed to optimize the pre-treatment process to obtain the highest physicochemical characteristics (yield, gel strength, viscosity and melting point) and molecular properties for gelatin production from rainbow trout skin. The effect of each factor has been considered very carefully to design a process giving high quality gelatin with desired characteristics for a particular application. 2. Methods 2.1. Rainbow trout skin Cultured rainbow trout (Onchorhynchus mykiss) were delivered on ice to process plant, Kian Maahi Khazar Co., Ltd. (Babolsar, Iran) and freezed at 20 C till skinning. Skins were taken off at 20 C and kept frozen till use. Proximate composition of the skin was 41.6 0.06% moisture, 13.12 0.02% crude lipid, 5.45 0.1% crude ash and 41.12 0.1% crude protein. Content of collagen, which is a precursor of gelatin, was about 30%. All reagents used, were of analytical grade. 2.2. Gelatin extraction Gelatin was prepared following the procedure described by Zhou and Regenstein (2005) with slight modications. Freezing skin was thawed overnight at 4 C and then the residues of adhering tissues were removed manually. Thawing skin was cut into 1 2 cm squares using scissors and then washed in cold-water (about 8 C). The cleaned skin was treated with three volumes (v/w) of cold alkali solution (0.010.21 N, NaOH) at 7 C to remove the non-collagen protein and subcutaneous tissue after they were swollen. Alkaline-treated skins were then washed out with tap water until neutral or faintly basic pHs of wash water were obtained. The skins were then soaked in cold acid (0.010.21 N, acetic acid at 7 C) with a skin/solution ratio of 1:3 (w/v). The acid solutions were drained off, and then the samples were washed out with cold-water. Each soaking treatments was repeated three times with a total time of 13 h (according to RSM, Table 1) for each treatment. All reactions happened in a batch glass container. For hot-water extraction, three volumes (v/w) of distilled water were added to sample and heated at 50 C for 16 2 h in a water bath (Memmert, WB14, Germany). Solubilized gelatin was separated from residual skin fragments by using two layer lter cloth. The gelatin extract was then concentrated in a rotavapor (50 C, 30 min) (Bchi, R-205, Germany) before drying as a thin lm in a freeze-dryer (Operon, FDU-7012, South Korea). 2.3. Chemical composition The sh skin was analyzed by standard method 934.01, 988.05, 920.39 and 942.05 for moisture, protein (N 6.25), fat and ash, respectively (Horwitz, 2000). Hydroxyproline content was estimated by the method of Woessner (1961), using a conversion factor of 11.42. The analyses were repeated three times.

2.4. Yield The concentration of soluble protein after pre-treatment was determined by the Biuret method (Yang et al., 2007) measuring the absorbance with a spectrophotometer at 540 nm (PG Instruments, T80+UVVIS, UK) and using a liquid protein solution containing 6 g/dl bovine serum albumin (Zistchimi, Tehran, Iran) as a standard protein. The yield of extracted gelatin was calculated using the following equation:

Yield % C V=M 100

where C = protein concentration in extracted solution (g/ml); V = volume, M = weight of sample (g) used for extraction. 2.5. SDSpolyacrylamide gel electrophoresis (SDSPAGE) The gelatin solution was dialyzed 2448 h against distilled water using 12 kDa cut-off dialysis tube. The concentration of gelatin solutions were adjusted to 2 mg/ml with distilled water and then a 3-fold-concentrated loading buffer containing b-mercaptoethanol was added at 60 C. Protein samples were heat-denatured for 5 min at 90 C and analyzed by SDSpolyacrylamide gel electrophoresis (SDSPAGE) according to Laemmli (1970) with a 5% (w/v) stacking gel and 15% (w/v) separating gel. Vertical slab gel electrophoresis was carried out in a mini electrophoresis unit (Cleaver scientic Ltd.) at 25 2 C. Equal amounts of proteins were loaded onto each lane of gel and run for 3 h at 50 mA in a 1 mm thick gel. After electrophoresis, the gel was stained with Coomassie Blue R-250 dye in methanol:acetic acid:water solution (5:1:4, by volume) and destained in methanol:acetic acid:water solution without dye (1:1:8, by volume). The gels were densitometred by obtaining volumograms on a photo documentation system (GENIUS Tech., USA) by using the version 6.0.8 GeneSnap software. Approximate molecular weight of sample was determined using wide-molecular-weight marker kit obtained from Fermentas (Fermentas Company, Germany). 2.6. Gel strength The gel strength was determined on 6.67% gel (w/v), formed by dissolving the dry gelatin powder in distilled water at 60 C, cooling the solution in a refrigerator to 7 C, and matured at refrigerated temperature for 1618 h according to Fernndez et al. (2001). The gel strength at 7 C was determined using a texture analyzer (Hounseld, H5KS, England) equipped with a 1.27 cmdiameter at-faced cylindrical Teon plunger with a load cell of 5 kN and cross-head speed 1 mm/s. The dimensions of the sample were 3.3 cm diameter and 6 cm height. Gel strength was expressed as maximum force (in g), taken when the plunger had penetrated 4 mm into the gels. 2.7. Viscosity Gelatin solutions at a concentration of 6.67% (w/v) were made by dissolving the dry powder in distilled water and heating to 60 C. The viscosity (V, mPa s) of 18 ml of gelatin solutions was

Table 1 Independent variables and their levels in the 3-factor, 5-level central composite rotatable design for production of rainbow trout (O. mykiss) skin gelatin. Independent variables Symbols Range and levels 1.414 Concentration of NaOH (N) Concentration of acetic acid (N) Pretreatment time (h) X1 X2 X3 0.01 0.01 1 1 0.04 0.04 1.29 0 0.1 0.1 2 +1 0.18 0.18 2.71 +1.414 0.21 0.21 3

H.S. Tabarestani et al. / Bioresource Technology 101 (2010) 62076214

6209
2 3 X X i1 ji1

determined using a Brookeld LVDV-II viscometer (Brookeld Engineering Laboratories Ltd., Middleboro, MA) with small sample adapter equipped with a No. 00 spindle at 90 rpm.

Y b0

3 X i1

bi X i

3 X i1

bii X 2 i

bij X i X j

2.8. Melting point Melting point of gelatin was determined by the BS755 method (BSI, 1975). Solutions containing 6.67% (w/v) gelatin were prepared in screw cap test tubes. The test tubes were lled and kept in refrigerator (7 C) for 1618 h, and then were transferred into a cold-water bath (4 C). Measurement was performed at a heating rate of 0.2 C/min, using oil-drop method. 0.5% methylredchloroform was used as indicator of melting point. The temperature, at which the gel melted, allowing the chloroform to start falling, was recorded as the melting point.

where Y is the dependant variable (yield, molecular weight distribution, gel strength, viscosity and melting point), b0 is the constant, bi ; bii and bij are regression coefcients and Xi, Xj are levels of the independent variables. After the multi factor analysis of variance and the second-order model prediction determinations, the optimal pre-treatment condition was obtained by the desirability function approach using Minitab software (Version 14; Minitab Inc., State College, PA, USA). The response surface plots were developed using MATLAB software (Version 6.5, The Math Works Inc., Natick, Mass, USA) and represented a function of two independent variables while keeping the other independent variable at the optimal value. 3. Results and discussion 3.1. Optimization of gelatin extraction 3.1.1. Response surface model Experimental results of the 3-factors, 5-level central composite design are presented in Table 2. The RSREG procedure for SAS software was employed to t the quadratic polynomial equation to the experimental data. All the coefcients of linear (X1, X2, X3), quadratic (X11, X22, X33) and interaction were calculated for signicant differences using t-statistic test and the estimated coefcients of each model are presented in Table 3. To develop the tted response surface model equations, all insignicant terms (P P 0.05) were eliminated and the tted models are shown in Table 4. The values of R2 suggest that the quadratic models can explain most of variabilities in the observed data. Thus, the analysis of variance showed that predicted response surface models were statistically signicant (P < 0.01). 3.1.2. Analysis of variance The statistical signicance of the quadratic polynomial model equation was evaluated by the analysis of variance. Analysis of variance for the quadratic polynomial models is presented in Table 5. The results showed that the quadratic term for the dependant variables of Y2, Y4 and Y5, and cross product term for most dependant variables (Y2, Y3, Y4, Y6) were not signicant (P > 0.05). Whereas linear term (X1, X2, X3) and total regression model were highly

2.9. Response surface methodology (RSM) Central composite rotatable design (CRCD, Myers and Montgomery, 2002) was adopted in the optimization of gelatin extraction from rainbow trout skin. CRCD in the experimental design consists of 23 factorial points, six axial points (a 1:414) and four replicates of the central point (Table 1). Pre-treatment is one of the most important steps in gelatin extraction process. Concentration of NaOH (N, X1), concentration of acetic acid (N, X2) and pre-treatment time (h, X3) were chosen as independent variables. The range and central point values of these three independent variables were based on the results of preliminary experiments (Table 1). The yield of gelatin (%, Y1), molecular weight distribution (a1 =a2 chain ratio, and b chain percent, Y2), gel strength (Bloom, Y3), viscosity (mPa s, Y4) and melting point (C, Y5) were selected as the dependant variables for the combination of the independent variables (Table 2). Experimental runs were randomized to minimize the effects of unexpected variations in the observed responses.

2.10. Analysis of data The response surface regression (RSREG) procedure of the Statistical Analysis System software (Version 9.1, SAS Institute Inc., USA) was used to t the following quadratic polynomial equation.

Table 2 Central composite rotatable design and responses of dependent variables for gelatin extraction from rainbow trout (O. mykiss) skin to independent variables.a Run no. Coded levels of variable X1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 1 1 1 1 1 1 1 1 0 0 0 0 1.414 1.414 0 0 0 0 X2 1 1 1 1 1 1 1 1 0 0 1.414 1.414 0 0 0 0 0 0 X3 1 1 1 1 1 1 1 1 1.414 1.414 0 0 0 0 0 0 0 0 Response Y1 13.6 10.7 15.7 8.1 11.3 13.1 14.6 9.9 14.3 13.5 15.2 12.8 14.4 16.3 16.2 17.7 16.9 16.6 Y2 1.69 1.33 0.68 0.96 0.96 1.41 0 0 1.03 0.94 1.87 0.93 1.65 0 1.18 1.18 1.18 1.18 Y3 25.89 7.86 0 0 4.53 2.73 0 0 0 0 26.29 0 33.82 0 3.57 3.57 3.57 3.57 Y4 444 362 224 142 245 423 100 85 230 137 423 281 327 128 255 255 255 252 Y5 3.53 2.85 1.8 1.82 1.69 1.47 1.32 1.26 1.97 1.97 1.79 1.38 2.71 1.59 2.3 2.34 2.3 2.28 Y6 23 20.5 16.5 13 14 16.5 4 4 15 16 22 13.5 18 10.5 19 18.5 19.2 19

a Y1 (yield, %), Y2 (a1/a2), Y3 (b chain percent), Y4 (gel strength, Bloom), Y5 (viscosity, mPa s), Y6 (melting point, C). X1 (concentration of NaOH, N), X2 (concentration of acetic acid, N), X3 (pretreatment time, h).

6210

H.S. Tabarestani et al. / Bioresource Technology 101 (2010) 62076214

Table 3 Regression coefcients for the response surface models in terms of coded units.a Term Intercept X1 X2 X3 X11 X22 X33 X21 X31 X32 Y1 1.71E+01 2.81E01NS 3.14E01NS 1.21E+00 HS 1.13E+00 S 1.8E+00HS 1.83E+00HS 5.83E02NS 9.71E01S 1.40E+00HS
HS

Y2 1.21E+00 3.85E01HS 4.23E01HS 2.02E02NS 2.23E01NS 6.46E02NS 1.43E01NS 1.22E01NS 6.63E02NS 2.38E02NS
HS

Y3 2.20E+01 7.31E+00HS 6.43E+00HS 7.28E02NS 2.06E+00NS 3.98E+00NS 9.99E+00NS 1.82E+00NS 2.76E+00NS 2.27E+00HS
HS

Y4 2.55E+02 5.00E+01HS 9.37E+01HS 1.10E+01NS 1.38E+01NS 4.85E+01S 3.58E+01NS 5.38E+00NS 4.09E+01NS 2.41E+01NS
HS

Y5 2.25E+00 4.87E01HS 3.27E01HS 7.83E02NS 1.17E02NS 2.71E01S 7.83E02NS 2.73E01S 4.75E02NS 1.08E01NS
HS

Y6 1.90E+01HS 3.76E+00HS 4.04E+00HS 1.74E01NS 2.45E+00HS 7.04E01NS 1.83E+00HS 1.06E+00NS 1.06E+00NS 4.38E01NS

HS, high signicant: P < 0.01; S, signicant: 0.01 < P < 0.05; NS, non-signicant: P > 0.05. a Y1 (yield, %), Y2 (a1/a2), Y3 (b chain percent), Y4 (gel strength, Bloom), Y5 (viscosity, mPa s), Y6 (melting point, C). X1 (concentration of NaOH, N), X2 (concentration of acetic acid, N), X3 (pretreatment time, h).

Table 4 Response surface models for processing conditions of gelatin from rainbow trout (O. mykiss) skina. Responses Y1 Y2 Y3 Y4 Y5 Y6 Quadratic polynomial model Y 17:135 1:206X 3 1:126X 1 2 1:797X 2 1:827X 2 0:970X 3 X 1 1:398X 3 X 2 2 3 Y 1:21 0:385X 1 0:423X 2 0:223X 2 1 Y 24:022 7:306X 1 6:433X 2 9:988X 2 3 Y 254:52 50:037X 1 93:65X 2 48:47X 2 2 Y 2:245 0:487X 1 0:326X 2 0:27X 2 0:272X 2 X 1 2 Y 19:002 3:759X 1 4:043X 2 2:454X 2 1:829X 2 1 3 R2 0.9045 0.9067 0.8945 0.8969 0.9249 0.9662 P-value 0.0032 0.0029 0.0046 0.0042 0.0013 <0.0001

a Y1 (yield, %), Y2 (a1/a2), Y3 (b chain percent), Y4 (gel strength, Bloom), Y5 (viscosity, mPa s), Y6 (melting point, C). X1 (concentration of NaOH, N), X2 (concentration of acetic acid, N), X3 (pretreatment time, h).

Table 5 Analysis of variance (ANOVA) for the response surface model.a Sources DF Y1 Y2 Y3 Y4 Y5 Y6 SS PV SS PV SS PV SS PV SS PV SS PV Regression 9 105.49 0.0032 4.692 0.0029 2353.76 0.0046 185566 0.0042 5.541 0.0013 464.62 <0.0001 Linear 3 19.58 0.0358 3.936 0.0003 1137.37 0.0033 136757 0.0008 4.2 0.0002 366.11 <0.0001 Square 3 62.69 0.0012 0.594 0.079 958.87 0.0057 30556 0.0575 0.636 0.0591 78.91 0.0019 Interaction 3 23.21 0.0234 0.162 0.484 257.51 0.135 18253 0.156 0.704 0.0471 19.59 0.0828 Residual 8 11.13 0.482 277.53 21327 0.45 16.246 Lack-of-t 5 10.04 0.0952 0.487 <0.0001 277.53 <0.0001 21320 <0.0001 0.448 0.0009 15.978 0.0071 Pure error 3 1.09 0 0 6.75 0.002 0.267 Total 17 116.6 5.174 2631 206893 5.991 480.8

a Y1 (yield, %), Y2 (a1/a2), Y3 (b chain percent), Y4 (gel strength, Bloom), Y5 (viscosity, mPa s), Y6 (melting point, C). X1 (concentration of NaOH, N), X2 (concentration of acetic acid, N), X3 (pretreatment time, h).

signicant for all the dependant variables (P < 0.05). The P-values for the lack-of-t test in Table 5 were large and show that the models are adequate for predicting gelatin pre-treatment condition. However, the lack-of-t test for Y1 did not show signicant P-value (P = 0.0952) at 95% probability level. The check of model adequacy was performed by a normality test (AndersonDarling normality test) for error terms using residuals of all dependant variables. The error terms of all dependant variables had the normal distribution in the AndersonDarling normality test. Therefore, response surface model represented as quadratic polynomial equation was statistically signicant. 3.1.3. Multiple response optimization According to preliminary study three factors including concentration of NaOH (X1), concentration of acetic acid (X2), and pretreatment time (X3), were identied as critical variables that had signicant effects on different characteristics of gelatin extracted

from rainbow trout skin. Optimal conditions including coded and uncoded values of each dependant variable are presented in Table 6. Predict values of Y1, Y2, Y3, Y4, Y5 and Y6 were 17.34 (%), 1.96 (%), 34.38 (%), 487 (Bloom), 3.22 (mPa s) and 23.7 (C), respectively. There was no signicant difference among the critical values of dependant variables and actual values except for yield. To optimize six dependant variables (Y1Y6) simultaneously, desirability function of MINITAB statistical software were dened as following conditions; goal (target), target (Y1 = 10, Y2 = 1.96, Y3 = 35, Y4 = 450, Y5 = 3.5, Y6 = 23). Coded values of the independent variables were concentration of NaOH, X1 = 0.908; concentration of acetic acid, X2 = 1.414 and pre-treatment time, X3 = 1.414. Actual values of independent variables against coded values were X1 = 0.19 (N), X2 = 0.121 (N) and X3 = 3 (h). Predict values of multiple response optimal conditions were Y1 = 9.36 (%), Y2 = 1.763, Y3 = 32.81, Y4 = 459.313, Y5 = 3.198, Y6 = 20.416 with 0.38 of the value of composite desirability function.

H.S. Tabarestani et al. / Bioresource Technology 101 (2010) 62076214 Table 6 Optimal conditions of gelatin processing from rainbow trout (O. mykiss) skin. Dependent variable Independent variable Critical value Coded Y1 (yield) X1 X2 X3 X1 X2 X3 X1 X2 X3 X1 X2 X3 X1 X2 X3 X1 X2 X3 X1 X2 X3 0.059 0.079 0.412 0.251 1.390 0.042 1.210 0.706 0.173 0.208 1.254 0.044 1.235 0.626 0.282 0.373 1.363 0.005 0.908 1.414 1.414 Uncoded 0.105 0.093 2.412 0.125 0.210 2.042 0.210 0.170 2.173 0.12 0.21 1.956 0.21 0.162 2.282 0.137 0.21 1.995 0.190 0.121 3 17.334 Maximum Predicted value

6211

Stationary point

Y2 (a1/a2)

1.96

Saddle point

Y3 (b chain percent)

34.38

Maximum

Y4 (gel strength)

487

Saddle point

Y5 (viscosity)

3.22

Maximum

Y6 (melting point)

23.7

Saddle point

Multiple response optimization

3.2. Effect of pre-treatment condition on physico-chemical properties of gelatin The effect of pre-treatment condition on different physicochemical properties of gelatin were evaluated using RSM. The pre-treatment conditions were NaOH concentration (0.01 0.21 N), acid concentration (0.010.21 N) and pre-treatment time (13 h). The important physico-chemical properties of gelatin were extraction yield, molecular weight distribution, gel strength, viscosity and melting point. 3.2.1. Molecular weight distribution The properties of gelatin depend on the molecular weight distribution of collagenous components and on the a1 =a2 ratio (GmezGuilln et al., 2002). To evaluate the molecular weight distribution including the a1 =a2 molar ratio and b chain amount, and to study the effect of pre-treatment conditions on gelatin fractions, gelatin preparations were analyzed by SDSPAGE (Fig. 1). After densitometric analysis of the Coomassie blue-stained gels, the a chain ratio (a1 =a2 ) and b components ($200 kDa) of extracted gelatin were determined. According to the electrophoresis pattern and densitometry results, the values of a1 =a2 chain ratio obtained for the gelatin samples in different pre-treatment conditions ranged among 01.87 (Table 2). Using the multiple regression model, it could explain 90.67% of the variations in a1 =a2 chain ratio (Table 4). Gelatin pre treated with lower concentration of alkaline and acidic

solution presented a clear different band pattern compared to gelatin pretreated with higher concentration of NaOH and acetic acid. Gelatin pretreated with lower concentration of NaOH and acetic acid showed very low a1 =a2 chain ratio. Thus pre-treatment of skin in solution with high concentration of NaOH and acetic acid increased the a1 =a2 molar ratio in extracted gelatin (Fig. 2a). Pretreatment time had no signicant effect (P > 0.05) on the regression model of a1 =a2 chain ratio. On the other hand, all three pre-treatment variables employed during the preparation of gelatin had signicant effects on production of gelatin with higher b components (covalently linked a chain dimers). Fig. 2b shows that there is positive correlation between alkali concentration and quantity of b components in extracted gelatin. However, the correlation between acid concentration and number of b components is positive up to pre-treatment time of 2 h and the correlation become negative after 2 h acidic pre-treatment. This nding indicates that probably low concentration of OH and H+, provide the proper pH condition for activation of collagenase, which its optimum pH is among 68, therefore partial loss of a chains and b chain amount could be occurred. Generally acidic pre-treatment conditions had the signicant effect on extraction of high-molecular weight proteins from rainbow trout sh skin. These results can be conrmed by the ndings of Zhou and Regenstein (2005) who reported that alkaline extraction caused some polypeptide chains of collagen to break into small pieces, while with neutral or weak acid conditions, gelatin fractions were mainly a chain, b chain and high oligomers. Gmez-Guilln et al. (2002) reported a characteristic a1 =a2 chain ratio of around two for sole and megrim and much lower a chain ratio ($1) for cod and hake skin gelatin. In best pre-treatment conditions in our studies, the observed a1 =a2 chain ratio and b chain amount for rainbow trout skin gelatin were 1.96% and 35%, respectively.

Fig. 1. Electrophoresis pattern of rainbow trout gelatins extracted in different pretreatment conditions. Numbers dene the coded level of RSM design.

3.2.2. Yield The multiple regression model that developed to predict yield of rainbow trout gelatin, indicate the adequacy of the selected model to explain the variations in yield (R2 = 9045). The signicant variable affecting yield was the linear term of pre-treatment time and quadratic terms of all three variables (Table 4). Signicant interactions were found between pre-treatment time and concentration of NaOH as well as between acid concentration and

6212

H.S. Tabarestani et al. / Bioresource Technology 101 (2010) 62076214

chain and the amount of b component. The gel strength after overnight maturation at 7 C varied considerably depending on the conditions used for pre-treatment processing in rainbow trout sh skin. According to the results of gel strength, there is positive covariance among gel strength, alkaline and acid concentration. However, the later is more affective. Fig. 3b shows the effect of independent variables of NaOH and acid concentration on gel strength. The gelatin samples produced in pre-treatment solutions contained higher concentration of NaOH and acetic acid, showed higher gel strength. This condition possibly is due to the presence of high-molecular weight fragments in gelatin preparation, because of employed pre-treatment condition. According to Badii and Howell (2006) gelatin with high-molecular weight distribution shows better gelling properties compared to low molecular weight gelatin. Gelatin with high-molecular weight distribution allows better stabilising interactions to form more organized triple helical structures (Sims et al., 1997). The gel strength of gelatin obtained from skin of hake, cod and megrim varied from 100 to 210 g (Gmez-Guilln and Montero, 2001; Fernndez et al., 2001; Gudmundsson and Hafsteinsson, 1997). The gelatin extracted from channel catsh skins showed higher gel strength of 276 g compared to that of cod and hake (Yang et al., 2007). However, in this study, the gelatin extracted from rainbow trout skin has had gel strength of 459 g in multiple response optimized pre-treatment condition which is remarkable for gel strength of sh gelatin but slightly lower than that reported by Zhou and Regenstein (2005) of 460 g for Pollock skin gelatin. Cho et al. (2005) reported gel strength of 426 g for gelatin extracted from yellow n tuna skin. 3.2.4. Viscosity Viscosity is the second important physical property of the gelatin. The shear viscosity values obtained for rainbow trout gelatin samples (6.67% w/v) ranged among 1.263.53 mPa s, depending on the conditions used in the pre-treatment of raw material (Table 2). The multiple regression model for predicting viscosity could explain 92.49% of the variations in this factor. NaOH concentration found to be the most important processing variable inuencing the viscosity of the gelatin produced from rainbow trout skin (Fig. 3c). Alkaline concentration has had synergistic effect on viscosity with acetic acid concentration. This increase possibly is due to complete opening up of polypeptide chain to random chain and intermolecular hydrodynamic interaction leading to increase viscosity. Gudmundsson and Hafsteinsson (1997) reported that the viscosity of the gelatin solutions is mainly controlled by molecular weight and polydispersity. The viscosity of rainbow trout skin gelatin is in mid range since the viscosity of commercial gelatins have been reported to be from 2 to 7 mPa s for most cases and up to 13mPa s for specialized ones (Johnston-Banks, 1990). 3.2.5. Melting point The model developed for predicting rainbow trout gelatin melting point could explain 96.62% of the variation observed in this functional property (Table 4). The effects of acid concentration and pre-treatment time on the melting point of gelatin are shown in Fig. 3d. The three dimensions plot shows that there has been direct relationship between acid concentration and melting point of extracted gelatin. The same trend has been observed with all NaOH concentrations. This trend is conrmed by the results of SDSPAGE of rainbow trout gelatins (Fig. 1). It could be observed that in high alkaline and acid concentration the resultant gelatin consist of high-molecular weight fragments. According to Johnston-Banks (1990) there was a positive correlation between melting point and molecular weight of peptides in gelatin. Moreover, the electrostatic interactions and stability will increase in gelatin gel network as inorganic matters remove effectively with increase in acid and alkaline concentration. Fish gelatins especially those extracted

Fig. 2. The response surface plots of rainbow trout gelatin extraction, responses a and b refers to contour plots of a1/a2 and b chain percent, respectively.

pre-treatment time (P < 0.05). According to covariance coefcient between independent and dependant variables, pre-treatment time and acid concentration are two important factors affecting gelatin yield (P < 0.01). The response surface plot (Fig. 3a) shows that yield of extraction has increased by increasing in acid concentration from 0.01 to 0.18 N and decreased in acid concentration of higher than 0.018 N. The result is in accordance with Alaska pollock, where high yields can only be obtained using neutral and acid pre-treatments (Zhou and Regenstein, 2005). On the other hand, increase in pre-treatment time lead to increase in yield of gelatin extraction. These ndings supported by results of Johnston-Banks (1990), Poppe (1992), Gudmundsson and Hafsteinsson (1997), Zhou and Regenstein (2005) and Ladislaus et al. (2007) who reported that sufcient number of cross-links must be broken down to convert the collagen into a suitable form for extraction. 3.2.3. Gel strength Gel strength is one of the most important functional properties of gelatin. Fish gelatins typically show less gel strength compared to gelatins obtained from mammalian sources (Gilsenan and Ross-Murphy, 2000). Gel strength is a function of complex interactions determined by amino acid composition and the ratio of a

H.S. Tabarestani et al. / Bioresource Technology 101 (2010) 62076214

6213

Fig. 3. The response surface plots of rainbow trout gelatin extraction, responses a, b, c and d refer to contour plots of yield, gel strength, viscosity and melting point, respectively.

from cold-water species have lower melting points compared to mammalian gelatins. This is due to the lower imino acid content of sh gelatin, which in turn reduces the propensity for intermolecular helix formation (Choi and Regenstein, 2000). The melting point of gelatin extracted from rainbow trout sh skin was 23 C in optimum pre-treatment condition, and was far higher than those reported values for cod skin (810 C) (Gudmundsson and Hafsteinsson, 1997), hake (14 C), sole (19.4 C), megrim (18.8 C) (Gmez-Guilln et al., 2002) and grass carp (19.5 C) (Ladislaus et al., 2007). The melting point of the gelatins obtained from bovine, porcine and yellow n tuna skin were 23.8, 25.6 and 24.3 (C), respectively (Cho et al., 2005). High gelling and melting points expand the range of gelatin application. However, Choi and Regenstein (2000) reported that the lower melting point of cold-water sh gelatin enhances avor release, fruit aroma, and melt rate in water gel desserts. 4. Conclusion This study suggests that Rainbow trout skins might be successfully used in gelatin production, giving relatively high physicochemical properties. According to the model, optimum extraction

conditions were found to be 0.19 N concentration of OH, 0.121 N concentration of H+, and 3 h pre-treatment time giving a predicted set of independent variables with reasonable yield and good physico-chemical properties. The concentration of H+ during pre-treatment showed signicant effect on molecular weight distribution, melting point and gel strength; also the concentration of OH had signicant effect on viscosity and for yield of extraction, pre-treatment time was the critical factor.

Acknowledgements The authors thank Mr. Gashtasb, manager of Kian Mahi Khazar Fisheries processing plant Co.,Ltd in Babolsar for providing the sh skin used as experimental materials.

References
Badii, F., Howell, N.K., 2006. Fish gelatin: structure, gelling properties and interaction with egg albumen proteins. Journal of Food Hydrocolloids 20 (5), 630640. BS 755, 1975. British Standards Institution Specication for Gelatin. Pentonville Rd, London, UK.

6214

H.S. Tabarestani et al. / Bioresource Technology 101 (2010) 62076214 Horwitz, W., 2000. Ofcial Methods of the Association of Ofcial Analytical Chemists International, 17th ed. AOAC International, Gaithersburg. Johnston-Banks, F.A., 1990. Gelatin. In: Harris, P. (Ed.), Handbook of Food Gels. Elsevier, Applied Science, London, pp. 233289. Ladislaus, M., Kasankala, Y.X., Weilong, Y., Sun, D.H., Qian, H., 2007. Optimization of gelatine extraction from grass carp (Catenopharyngodon idella) sh skin by response surface methodology. Bioresource Technology 98, 33383343. Laemmli, U.K., 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680685. Myers, R.H., Montgomery, D.C., 2002. Response Surface Methodology. John Wiley and Sons, Inc., New York. 798p. Norland, R.E., 1990. Fish Gelatin. Advances in Fisheries and Biotechnology for Increased Protability. Technomic, Lancaster, PA. pp. 325333. Poppe, J., 1992. Gelatin. In: Imeson, A. (Ed.), Thickening and Gelling Agents for Food. Blackie Academic and Professional, Glasgow, UK, pp. 98123. Sims, T.J., Bailey, A.J., Field, D.S., 1997. The chemical basis of molecular weight differences in gelatins. The Imaging Science 45, 171177. Woessner, J.F., 1961. The determination of hydroxyproline in tissue and samples containing small proportions of this imino acid. Biochemistry and Biophysics 93, 440447. Yang, H., Wang, Y., Jiang, M., Oh, J.H., Herring, J., Zhou, P., 2007. 2-Step optimization of the extraction and subsequent physical properties of channel catsh (Ictalurus punctatus) skin gelatin. Journal of Food Science 72, C188C195. Zhou, P., Regenstein, J.M., 2005. Effects of Alkaline and Acid Pretreatments on Alaska Pollock Skin Gelatin Extraction. Journal of Food Science 70, 392 396.

Cho, S.M., Gu, Y.S., Kim, S.B., 2005. Extracting optimization and physical properties of yellown tuna(Thunnus albacares) skin gelatin compared to mammalian gelatins. Journal of Food Hydrocolloids 19, 221229. Choi, S.S., Regenstein, J.M., 2000. Physicochemical and sensory characteristics of sh gelatin. Journal of Food Science 65 (2), 194199. Fernndez, M.D., Montero, P., Gmez-Guilln, M.C., 2001. Gel properties of collagens from skins of cod (Gadus morhua) and hake (Merluccius merluccius) and their modication by the coenhancers magnesium sulfate, glycerol and transglutaminase. Journal of Food Chemistry 74, 161167. Gilsenan, P.M., Ross-Murphy, S.B., 2000. Rheological characterisation of gelatins from mammalian and marine sources. Journal of Food Hydrocolloids 12, 191 195. Gmez-Guilln, M.C., Montero, P., 2001. Extraction of gelatin from megrim (Lepidorhombus boscii) skins with several organic acids. Journal of Food Science 66, 213216. Gmez-Guilln, M.C., Turnay, J., Fernndez-Daz, M.D., Ulmo, N., Lizarbe, M., Montero, P., 2002. Structural and physical properties of gelatin extracted from different marine species: a comparative study. Journal of Food Hydrocolloids 16, 2534. Gudmundsson, M., 2002. Rheological properties of sh gelatins. Journal of Food Science 67, 21722176. Gudmundsson, M., Hafsteinsson, H., 1997. Gelatin from cod skins as affected by chemical treatment. Journal of Food Science 62, 3739. Haug, I.J., Draget, K.I., Smidsrd, O., 2004. Physical and rheological properties of sh gelatin compared to mammalian gelatin. Journal Food Hydrocolloids 18, 203 213.