Vol. 69, Nr. 8, 2004—JOURNAL OF FOOD SCIENCE C637Published on Web 10/7/2004

© 2004 Institute of Food TechnologistsFurther reproduction without permission is prohibited

C: Fo

od Ch

emist

ry &

Toxico

logy

JFS C: Food Chemistry and Toxicology

Characterization of Acid-soluble Collagen fromPacific Whiting Surimi Processing ByproductsJ.-S. KIM AND J. W. PARK

ABSTRACT: Three different solid byproducts (skin, frame, and refiner discharge) from Pacific whiting surimi manu-facturing are a good resource for collagen extraction according to their total protein concentrations and otherbiochemical properties. Denaturation temperature of acid-soluble collagens was 23.3 °C for refiner discharge,21.7 °C for skin, and 20.6 °C for frame. Based on the functional properties, acid-soluble collagen from refinerdischarge was the best and showed potential as an ingredient in processed food manufacturing.

Keywords: collagen, gelatin, surimi byproducts, refiner discharge

Introduction

Collagen, commonly manufactured from land animal by-products, is used in various food applications (clarification

agent, emulsifier, or whipping agent). Its usage extends even fur-ther to other industrial (shampoo and lipstick) and pharmaceuticalapplications (film-forming agent, microencapsulation, or tabletcoating) (Nagai and Suzuki 2000). Today’s health-conscious con-sumers, however, are reluctant to try collagen extracted from landanimals because of the recent outbreaks of bovine spongiform en-cephalopathy or foot-and-mouth disease. Therefore, raw materialsfrom fishery products receive new attention as a consumer-friendlycollagen resource.

Surimi byproducts are generated in a large quantity (approxi-mately 65% based on whole fish) during surimi manufacturing(Wendel 1999). However, the use of surimi solid byproducts as a hu-man food has not been widely studied yet. Most surimi solidbyproducts are conventionally used to produce fish meal and fer-tilizer or are directly discharged into estuaries, resulting in environ-mental pollution (Ciarlo and others 1997). New challenges must beattempted to find a way to upgrade the processing of waste to food-grade ingredients such as collagen.

Fish collagens are of interest in the food-processing industry fortheir use in the production of gelatin from the extracted collagen.Fish collagen, however, differs from animal collagen because of alower content of the imino acids (proline and hydroxyproline) (Park2000). To use collagen from fishery products in various foods orother industrial or pharmaceutical applications, functional proper-ties (water, oil absorption, and emulsion ability and stability) aswell as physicochemical properties must be optimized.

Some efforts have been made toward the preparation of col-lagens from fishery products, such as fish skin (Montero and others1990; Ciarlo and others 1997) and fish bone (Nagai and Suzuki2000). However, these studies were limited to fillet byproducts as acollagen resource and did not measure the functional properties ofcollagen for effective utilization.

Our objective was to examine various byproducts (skin, frame,and refiner discharge) from Pacific whiting surimi processing as acollagen resource by characterizing biochemical and functionalproperties of collagen.

Materials and Methods

Fish and surimi solid byproductsPacific whiting (Merluccius productus), 42 to 48 cm long, was

caught off the coast of Oregon by trawl. Pacific whiting and its surimisolid byproducts were obtained from a commercial surimi process-ing plant (Warrenton, Oreg., U.S.A.) in July 2002.

Skin and frame from Pacific whiting were generated from debon-ing and mincing steps, whereas refiner discharge was separatedfrom the refining step immediately before screw-press dewatering.The samples (fresh whole fish and solid byproducts of Pacific whit-ing surimi) were transferred on ice to the Oregon State Univ. Sea-food Laboratory within 30 min and kept frozen at –30 °C until usedfor collagen extraction.

Proximate composition, volatile basic nitrogen, andheavy metal

According to AOAC methods (AOAC 1990), moisture content wasquantified by oven drying at 105 °C, total protein by the Kjeldahlprocedure, and crude ash by incineration in a muffle furnace at550 °C. In addition, total lipid was extracted into a methanol-chlo-roform mixture and quantified according to the method of Blighand Dyer (1959). The concentration of volatile basic nitrogen (VBN)was determined using the method of Conway (1950). The mercu-ry content was determined by the combustion gold amalgamationmethod (KFDA 1999) using a mercury analyzer (Model SP-3A, Nip-pon Instrument Co., Tokyo, Japan). Other heavy metals, such as Pb,Cd, and Cr were determined by the wet ash method (Tsutagawaand others 1994), using an inductively coupled plasma spectropho-tometer (ICP, Atomscan 25, Thermo Electron Co., Waltham, Mass.,U.S.A.).

Preparation of collagen fractionsAll analyses were performed in a cold room (5 °C). Native col-

lagen was prepared as described by Sato and others (1986) andNagai and Suzuki (2000). Pacific whiting whole muscle and its suri-

MS 20040206 Submitted 4/4/04, Revised 5/8/04, Accepted 6/24/04. AuthorKim is with Div. of Marine Bioscience, Inst. of Marine Industry, GyeongsangNatl. Univ., Tongyeong 650-160, South Korea. Author Park is with SeafoodLaboratory & Dept. of Food Sci. and Tech., Oregon State Univ., 2001 MarineDrive #253, Astoria, OR 97103-3427. Direct inquiries to author Park (E-mail:jae.park@oregonstate.edu).

C638 JOURNAL OF FOOD SCIENCE—Vol. 69, Nr. 8, 2004 URLs and E-mail addresses are active links at www.ift.org

C: Food Chemistry & Toxicology

Collagen from surimi processing . . .

mi solid byproducts were 1st cut into small pieces before homoge-nizing with 5 volumes (v/w) of cold distilled water. The homogenatewas then centrifuged at 10000 × g for 20 min. To the residues, 20volumes (v/w) of 0.1 N NaOH was added to remove noncollagenousprotein. The homogenate was stirred overnight before centrifugingat 10000 × g for 20 min. Alkali-extractions were repeated 3 addition-al times. Final precipitate was washed thoroughly with cold distilledwater. For the frame sample, which contains a significant amount ofash (Table 1), alkali-insoluble residue was decalcified using 0.5 Methylenediaminetetraacetic acid (EDTA-4Na, pH 7.4) for 5 d bychanging the solution once a day and washing the residue thor-oughly with cold distilled water.

To all the residues, 10 volumes (v/w) of 0.5 M acetic acid wasadded. Suspensions were stirred for 3 d and then centrifuged at10000 × g for 20 min. This acid extraction process was repeatedonce more. The precipitates were then washed with cold distilledwater at 1:2 (w/v) ratio. The supernatant from acid extraction andthe filtrate from rinsing were combined and subjected to salting outby adding NaCl to 2.0 M before centrifuging at 20000 × g for 20 min.The precipitate was redissolved in 0.5 M acetic acid and then dia-lyzed (molecular weight cut-off 10000) against cold distilled waterto remove salt. Salting out and dialysis were repeated twice morefor further purification of collagen. The final dialyzed solution wasused as acid-soluble collagen fraction.

Separately from preparation of acid-soluble collagen, centrifugeresidues obtained after acid extraction were heated with 5 volumes(v/w) of distilled water in an autoclave at 120 °C for 1 h and thencentrifuged at 10000 × g for 20 min. The precipitates were rinsedusing hot distilled water at a 1:2 (w/v) ratio. The supernatant fromcentrifugation and the filtrate from rinsing were combined andused as acid-insoluble collagen fraction. The acid-insoluble frac-tion was used for calculation total Collagen N concentration.

Acid-soluble fraction was lyophilized and used for further anal-yses of collagen characteristics. Concentration and solubility ofcollagen were calculated, respectively, as (total collagen-N concen-tration, %/total-N concentration, %) × 100 and (acid-soluble col-lagen-N concentration, %/total collagen-N concentration, %) × 100.

Sodium dodecyl sulfate-polyacrylamide gelelectrophoresis

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis(SDS-PAGE) was performed by the method of Laemmli (1970) us-ing 7.5% gel containing 0.1% SDS at pH 8.8. SDS-PAGE was per-formed using a slab gel electrophoresis system (FB-VE 16-1, 16 × 14cm, Fisher Scientific, Pittsburgh, Pa., U.S.A.). Protein samples andprotein marker (M 4038; Sigma Chemical Co., St. Louis., Mo., U.S.A.)were heated at 100 °C for 3 min in 10 mM tris-HCl buffer (pH 6.8)containing 2% SDS, 1% 2-mercaptoethanol, 25% glycerol, and 0.1%bromophenol blue. The gels were stained for protein with 0.1%Coomassie brillant blue R-250 and destained in 10% methanol and10% acetic acid.

Amino acid compositionAmino acid composition was determined using an amino acid

analyzer (Biochrom 20, Pharmacia Biotech, Cambridge, Sweden)according to the method of Kimura (1988). Samples were hydro-lyzed in 6 N HCl in evacuated/sealed tubes at 110 °C for 16 h. Thehydrolysates were evaporated to dryness in a vacuum evaporatorat 40 °C and then diluted with Li-buffer for analyses of amino acidscontaining hydroxyproline and hydroxylysine.

Hydroxylation (%) of proline (Pro) and lysine (Lys) was calculated

on the basis of the amino acid composition according to the follow-ing equations:

Hydroxylation of Pro (%) = (Hyp × 100)/(Pro + Hyp)Hydroxylation of Lys (%) = (Hyl × 100)/(Pro + Hyl)

HyP = hydroxyproline, Hyl = hydroxylysine

Denaturation temperatureThermal denaturation temperature was measured using 5 mL of

collagen solution (30 mg collagen dissolved in 100 mL acetic solu-tion, 0.1 M) according to the method of Zhu and Kimura (1991). Thecontrol sample was 0.1 M acetic acid. All sample solutions were keptin water bath (8 °C) before measuring specific viscosity using anOstwald viscometer. Specific viscosity was measured at 8 °C, 15 °Cto 31 °C at every 2 °C, and at 45 °C. We assumed that collagen heli-cal conformation was undenatured at 8 °C, whereas breakdown wascompleted at 45 °C.

Denaturation temperature of collagen solution was defined asthe temperature at which the change in viscosity reached by 50%.Fraction change was calculated as follows:

[�2/C) – (�3/C)]/[(�1/C) – (�3/C)]

C = collagen concentration (mg/mL); �1 = specific viscosity at 8 °C;�2 = specific viscosity at measured temperature (°C); �3 = specificviscosity at 45 °C.

Functional propertiesWater and oil absorption capacities were determined by the

method of Beuchat (1981). Freeze-dried collagen (0.3 g) was mixedwith 10 mL distilled water for water absorption measurement andwith 10 mL vegetable oil for oil absorption measurement. Mixingwas done at fast speed using a vortex mixer for 30 s. Samples werethen allowed to stand at room temperature (22 °C) for 30 min be-fore centrifuging at 5000 × g for 30 min. The volume of supernatantwas measured in a graduated cylinder (10 mL), and the value wasexpressed on a dry weight basis.

Emulsifying activity and cooking stability were determined bythe method of Wang and Kinsella (1976). Freeze-dried collagen (0.2g) and Tween-80 (0.2 g, Fisher Scientific, Pittsburg, Pa., U.S.A.) wereadded to 20 mL of 0.1 N acetic acid, respectively, and the mixturewas set at room temperature for 2 min using a PT 10/35 polytronhomogenizer at setting 3 (Kinematica, Luzern, Switzerland). Twen-ty milliliters of vegetable oil (Mazola corn oil, CPC Intl. Inc., Engle-wood Cliffs, N.J., U.S.A.) were added before mixing for 3 min usinga PT 10/35 polytron homogenizer at high speed (setting 5). The

Table 1—Proximate composition of various solidbyproducts and whole musclea,b

Solid byproducts (g/100 g)

Whole RefinerComponents muscle Skin Frame discharge

Moisture 80.5 70.5 69.7 81.4Crude lipid 1.5 (7�7) 3.1 (10�5) 1.0 (3�3) 1.9 (10�2)Crude protein 16.8 (86�2) 26.1 (88�5) 11.7 (38�6) 15.4 (82�8)Crude ash 1.0a (5�1) 0.1 (0�3) 17.4 (57�4) 1.0a (5�4)aBased on the comparison of refiner discharge with other raw materials (wholemuscle, skin, and frame), no superscripts in the same row indicate asignificant difference at P < 0.05.bNumbers in parentheses indicate values based on dried weight.

Vol. 69, Nr. 8, 2004—JOURNAL OF FOOD SCIENCE C639URLs and E-mail addresses are active links at www.ift.org

C: Fo

od Ch

emist

ry &

Toxico

logy

Collagen from surimi processing . . .

resulting emulsion was centrifuged at 1300 × g for 15 min. Emulsi-fying activity was expressed as the following equation:

(Height of emulsified layer/height of total contents in the tube) × 100%

Cooking stability was determined similarly to the emulsifyingactivity except that the emulsion in the centrifuge tube (height ×inner dia, 11.5 ×3.0 cm) was initially heated in a water bath (80 °C)for 30 min and subsequently cooled to 15 °C before centrifugation.Cooking stability was measured using the following equation:

(Height of emulsified layer after centrifugation/height of total contents in the tube) × 100%

Statistical analysisStatistical analysis was done using the ANOVA (analysis of vari-

ance) test. Significant differences between means were performedusing Systat Version 7.5K (SPSS, Inc. Richmond, Va., U.S.A.) atP < 0.05 (Steel and Torrie 1980).

Results and Discussion

Proximate compositionA significant difference (P < 0.05) was found for all parameters

between samples except between whole muscle and refiner dis-charge for ash content (Table 1). Moisture of refiner discharge(81.4%) was similar to that of whole muscle (80.5%) but much high-er than that of other samples. The highest total protein contentwas obtained from skin sample, followed by whole muscle, refinerdischarge, and frame. The highest lipid was also found in skin sam-ples probably. The highest ash content was 17.4% for frame, prob-ably because of the calcium content derived from the bone (Kimand others 2002). The trend of proximate composition on the driedweight basis among samples was similar to that based on the wetweight. For effective utilization of the solid byproducts for collagen,foreign components, such as lipid and ash, should be removedfrom the solid byproducts. Therefore, frame did not appear to be agood raw material for collagen extraction because of its high ashcontent.

Volatile basic nitrogen (VBN) and heavy metalcontents

Volatile basic nitrogen contents in solid byproducts were 3.9 mg/100 g in skin and frame and 2.1 mg/100 g in refiner discharge (Table2). The concentrations of VBN were much lower than 20 mg/100 g,which are believed to be an acceptable limit for marine products(Kim and others 2002). Mercury was not detected in all solidbyproducts. Lead was also not detected in skin and refiner dis-charge except for frame (0.020 mg/kg). Cadmium and chromiumranged from 0 to 0.019 mg/kg and 0.002 to 0.010 mg/kg in all solidbyproducts, respectively. According to Codex Code (2004), theheavy metal safety values were 0.2 to 1.0 mg/kg for cadmium, 0.2 to0.4 mg/kg for lead, and zero for mercury and cadmium. Because theconcentrations of VBN and heavy metals in solid byproducts werebelow these reported safety limits, all solid byproduct samples ap-peared safe as a raw material for collagen.

Collagen contentCollagen contents of solid byproducts of Pacific whiting surimi

were in the descending order of skin (65.6%), refiner discharge(55.3%), and frame (30.2%) (Figure 1). Whole muscle showed a sig-

nificantly low concentration (2.4%). A significant difference in col-lagen content between 2 samples (whole muscle and refiner dis-charge) containing similar moisture content was probably becausethe refining process is primarily to separate connective tissues fromwashed mince. The collagen content of fish frame (bone) was quitelow compared with fish skin and refiner discharge, probably be-cause of its high ash content (17.4%). To use the frame as a collagensource, the excessive ash, foreign material of collagen, should beremoved before collagen extraction (Nagai and Suzuki 2000).Therefore, the frame did not appear to be a good raw material forcollagen extraction because of economic values and complicatedprocessing. These results suggested that skin and refiner dischargecould be used for collagen extraction based on the concentration ofcollagen.

Collagen solubilityThe protein solubility of collagen was 77.4% for skin, 66.9% for

refiner discharge, 66.4% for whole muscle, and 24.5% for frame (Fig-ure 2). There was, however, no difference (P > 0.05) in solubilitybetween refiner discharge and whole muscle, probably becauserefiner discharge was primarily generated from washed mince. Lowsolubility of frame collagen compared with the others might bebecause of a different structure by covalent cross-link in collagen,especially telopeptide region (Takahashi and others 1989). Monte-ro and others (1990) and Yamaguchi and others (1976) reported thatcollagen solubility of hake (Merluccius merluccius L.) was higher in

Table 2—Volatile basic nitrogen (VBN) and heavy metalcontents of various solid byproducts from Pacific whitingsurimi processing

Solid byproducts

RefinerComponents Skin Frame discharge

VBN (mg/100 g) 3.9 ����� 0.0 3.9 ����� 0.0 2.1 ����� 0.0

Heavy Hg ND ND NDmetal Pb ND 0.020 � 0.010 ND(mg/kg) Cd 0.002 � 0�002 0.019 � 0.004 ND

Cr 0.002 � 0.001 0.010 � 0.000 0.002 � 0.002

Cd = cadmium; Cr = chromium; Hg = mercury; Pb = lead; ND = not detected.

Figure 1—Collagen contents of various solid byproducts andwhole muscle. Different letters on the bars indicate a sig-nificant difference at P < 0.05.

C640 JOURNAL OF FOOD SCIENCE—Vol. 69, Nr. 8, 2004 URLs and E-mail addresses are active links at www.ift.org

C: Food Chemistry & Toxicology

Collagen from surimi processing . . .

skin collagen (93.1% and 65.2%, respectively) than in muscle col-lagen (74.4% and 60.0%, respectively). A difference in solubilitybetween our results and reported values was believed to be due toa difference in age of fish, fish species, and/or different methodol-ogy (Sikorski and Borderias 1994).

Amino acid compositionGlycine was the most abundant amino acid present in all col-

lagen samples (Table 3). Alanine and proline compositions werealso rich in all collagen samples. However, cysteine, methionine,isoleucine, tyrosine, phenylalanine, hydroxylysine, and histidineshowed a significantly low concentration. Similar patterns of ami-no acid composition of collagen samples were also found in skinand muscle of various species, such as squid mantle (Ando andothers 2001), hake (M. merluccis L.), and trout (Montero and others1990), carp, eel, common mackerel, saury and chum salmon (Kimu-ra and others 1988), and filefish (Kim and Cho 1996).

Hydroxylation ratio of proline and lysineThe hydroxylation ratio of proline in collagens from solid

byproducts were quite high, with a descending order of refiner dis-charge (39.4%), skin (36.9%), and frame (32.8%) (Figure 3). Therewas, however, no significant difference (P > 0.05) in the hydroxyla-tion ratio of lysine among collagens from solid byproducts andwhole muscle. Ando and others (2001) and Montero and others(1990) reported that hydroxyproline plays a role in stabilizing the

triple helix, whereas hydroxylysine contributes to the formationand stabilization of cross-links of nonhydrolyzable bonds. Kimuraand others (1988) and Zhu and Kimura (1991) also reported thatthe hydroxylation ratio of proline was higher in muscle collagen thanin skin collagen. Based on our results and the literature, the ther-mal stability of collagens from solid byproducts is likely to be in thedescending order of refiner discharge, skin, and frame. The hydrox-ylation ratio of proline in collagen from refiner discharge was 39.4%and was similar to that of collagens from muscle of common horsemackerel (38%), yellow sea bream (40%), and tiger puffer (39%)(Yata and others 2001). However, it was much lower than that fromshark and carp (Yata and others 2001) and land animals and otherhigher vertebrates (Park and others 1995).

SDS-PAGE patternAccording to SDS-PAGE pattern (Figure 4), 3 distinctive chains,

2 � bands (�1, upper; �2, lower) with their molecular weight atabout 100 kDa and their �-cross-linked components, with a molec-ular weight of 200 kDa, were clearly detected in all solid byproducts.As a whole, our electrophoretic patterns of collagens from byprod-ucts were almost identical to those of the corresponding calf skintype I collagen in mobility of chains (data not shown) and similar tothose obtained with collagens from the skin and muscle of otherspecies (hake [Merluccius hubbsi, Ciarlo and others 1997; M. mer-luccis L., Montero and others 1990], trout [Montero and others1990], and Alaska pollock [Kimura and Ohno 1987]). There was nodifference in the relative mobility of �1 and �2 chains among acid-soluble collagens from all solid byproducts. Based on our resultsand the literature, the collagens obtained in this experiment werefound to be free of noncollagenous proteins.

High molecular weight cross-links (> 205 kDa) were formed andshown on the top of each lane with more evidence in whole muscleand frame than refiner discharge or skin. These large molecules areyet to be identified.

Thermal denaturationThermal denaturation temperature of collagens from solid

byproducts was 23.3 °C for refiner discharge, 21.7 °C for skin, and

Figure 2—Solubility of collagen from various solidbyproducts and whole muscle. Different letters on the barsindicate a significant difference at P < 0.05.

Figure 3—Hydroxylation ratio of proline (HDP) and hydroxy-lation ratio of lysine (HDL) in collagen from various solidbyproducts and whole muscle. Different letters on the barsindicate a significant difference at P < 0.05.

Figure 4—Sodium dodecyl sulfate-polyacrylamide gel elec-trophoretic pattern of acid-soluble collagen from varioussolid byproducts and whole muscle (B = soluble collagenfrom frame ; M = soluble collagen from whole muscle; MP= marker protein; RD = soluble collagen from refiner dis-charge; S = soluble collagen from skin).

Vol. 69, Nr. 8, 2004—JOURNAL OF FOOD SCIENCE C641URLs and E-mail addresses are active links at www.ift.org

C: Fo

od Ch

emist

ry &

Toxico

logy

Collagen from surimi processing . . .

20.6 °C for frame (Figure 5). These results might be due to the dif-ference of the hydroxylation ratio of proline, which is highly corre-lated to thermal stability (Ando and others 2001). Thermal denatur-ation temperature of collagen from refiner discharge was similar tothat of collagen from whole muscle, 23.5 °C. However, the thermaldenaturation temperature of collagens from refiner discharge wasmuch lower than collagens obtained from shark (Hamada 1990),carp (Miyauchi and Kimura 1990), and land animal byproducts(Voigt and Botta 1989). This result suggested that the thermal sta-bility of collagens from surimi byproducts must be improved to bemore effectively used as a food or industrial ingredient.

Water and oil absorption capacitiesWater absorption capacity of acid-soluble collagens from refiner

discharge was 15.92 (mL/g), followed by 9.83 mL/g for frame and8.58 mL/g for skin (Figure 6). There was, however, no significantdifference (P > 0.05) in water absorption capacity between col-

Table 3—Amino acid composition of collagen from various solid byproducts and whole musclea

Amino Cod skin Whole Solid byproducts (residues/1000 residues)

acids gelatinb muscle Skin Frame Refiner discharge

Asp 51 32.5 � 0�7a 34.4 � 0�7ab 36.9 � 1�2b 32.5 � 1�0aHyp 53 68.7 � 0�4c 60.6 � 0�3b 50.6 � 1�1a 68.0 � 0�7cThr 25 28.8 � 0�5a 28.8 � 0�5a 33.0 � 1�0b 28.9 � 0�9aSer 69 41.0 � 0�5a 40.9 � 0�5a 43.1 � 0�5a 40.0 � 1�7aGlu 75 76.7 � 0�5a 79.9 � 1�7a 85.2 � 0�5b 77.9 � 0�3aPro 102 104.8 � 1�5a 103.7 � 1�7a 103.9 � 1�8a 104.6 � 1�5aGly 345 350.3 � 1�3b 349.6 � 4�1b 322.0 � 3�2a 350.7 � 3�4bAla 107 130.2 � 1�1a 128.1 � 3�5a 124.6 � 1�2a 130.3 � 2�5aCys <1 NDc NDc NDc NDc

Val 19 19.1 � 0�1a 20.6 � 1�1a 24.7 � 1�2b 19.1 � 0�8aMet 13 14.5 � 0�3a 15.2 � 0�6a 17.1 � 0�5b 14.5 � 0�4aIle 11 11.2 � 0�3a 12.3 � 0�3a 16.5 � 0�4b 11.2 � 0�7aLeu 23 20.0 � 0�7a 21.4 � 0�8a 30.1 � 0�5b 20.0 � 1�0aTyr 4 1.7 � 0�0a 2.1 � 0�2a 3.2 � 0�2b 1.7 � 0�2aPhe 13 14.7 � 0�3a 13.5 � 1�3a 16.3 � 0�5a 14.7 � 0�4aHyl 6 6.9 � 0�2a 7.1 � 0�4a 7.0 � 0�3a 6.9 � 0�3aLys 25 26.4 � 0�4a 26.9 � 1�3a 30.9 � 1�3a 26.4 � 1�5aHis 8 6.8 � 0�3a 7.8 � 0�6ab 8.6 � 0�3b 6.8 � 0�3aArg 51 45.8 � 0�6a 47.1 � 0�6a 46.3 � 0�5a 45.8 � 1�0aTotal 1000 1000.0 1000.0 1000.0 1000.0aDifferent letters in the same row indicate a significant difference at P < 0.05.bVoigt and Botta (1990).cND = not detected.

lagens from refiner discharge and whole muscle (16.12 mL/g). Waterabsorption capacity of acid-soluble collagens from refiner dischargewas also higher than those of the other protein from vegetables,such as soy protein isolate (3 to 8 mL/g, Yim and Lee 2000), alfalfaleaf protein (1.85 to 3.58 mL/g, Wang and Kinsella 1976), and beanprotein (1.67 to 5.93 g/g, Sathe and Salunkhe 1981). This differ-ence may be due to the high swelling ability of collagen (Sadowskaand Rudzki 1987) along with the differences of size, shape, hydro-philic-hydrophobic balance of amino acids in the protein moleculeand the physicochemical environment such as pH, ionic strength,and temperature (Sathe and Salunkhe 1981).

Oil absorption capacity of acid-soluble collagens from refiner dis-charge was 26.10 mL/g, followed by 12.93 mL/g for skin and 11.82mL/g for frame (Figure 6). There was no difference (P > 0.05) in oilabsorption capacity for collagens from refiner discharge and whole

Figure 5—Effects of incubation temperature on fractionalchange of acid-soluble collagen from various solidbyproducts and whole muscle

Figure 6—Water absorption capacity (WAC) and oil absorp-tion capacity (OAC) of acid-soluble collagen from varioussolid byproducts and whole muscle. Different letters onthe bars within the same properties indicate a significantdifference at P < 0.05.

C642 JOURNAL OF FOOD SCIENCE—Vol. 69, Nr. 8, 2004 URLs and E-mail addresses are active links at www.ift.org

C: Food Chemistry & Toxicology

Collagen from surimi processing . . .

muscle (26.52 mL/g). Oil absorption capacity of acid-soluble collagenfrom refiner discharge was also higher than that of soy protein isolate(2 to 10 mL/g, Yim and Lee 2000), alfalfa leaf protein (1.75 to 4.30 mL/g, Wang and Kinsella 1976), Atlantic salmon muscle hydrolysates(2.86 to 7.07 mL/g, Kristinsson and Rasco 2000), and bean proteinconcentrates (1.00 to 4.12 g/g, Sathe and Salunkhe 1981).

Emulsion activity and cooking stabilityAcid-soluble collagen from refiner discharge had the highest

emulsifying activity (51.9%) among acid-soluble collagens fromsolid byproducts (Figure 7). This value was also superior to a com-mercial emulsifier, Tween-80 (45.9%) (P8192; Sigma Chem., St. Lou-is, Mo., U.S.A.). However, there was no significant difference (P >0.05) in emulsion activity of collagens from refiner discharge andwhole muscle. The emulsion activity of collagen from byproducts isprimarily determined by the orientation at the interface betweenthe 2 phases where a monomolecular film is formed around the col-loidal particles (Yang 1997).

Acid-soluble collagen from refiner discharge also had the high-est cooking stability (45.4%) among acid-soluble collagens from solidbyproducts (Figure 7). This value was superior to a commercialemulsifier, Tween-80 (45.9%). However, there was no significantdifference (P > 0.05) in cooking stability between collagens fromrefiner discharge and whole muscle. Kim and others (1996) report-ed that emulsifying properties of gelatin from cod bone were sim-ilar to those of a commercial emulsifier, such as Tween-60 andTween-80.

Conclusions

Acid-soluble collagens from refiner discharge of surimi manu-facturing demonstrated the highest functional properties

compared with the other byproducts, such as skin and frame, andcould be used as a functional ingredient for food and industrial ap-plications. In addition, if the thermal stability of the acid-solublecollagens is improved, collagens from surimi byproducts could bemore effectively used.

References[AOAC] Assn. of Official Analytical Chemists. 1990. Official methods of analysis.

15th ed. Washington, D.C.: AOAC.Ando M, Ando M, Makino M, Tsukamasa Y, Makinodan Y, Miyosh M. 2001. Inter-

dependence between heat solubility and pyridinoline contents of squid man-tle collagen. J Food Sci 66:265–9.

Beuchat LR. 1981. Functional and electrophoretic characteristics of succinylatedpeanut flour proteins. J Agric Food Chem 46:71–5.

Bligh EG, Dyer WJ. 1959. A rapid method of lipid extraction and purification. CanJ Biochem Physiol 37:911–7.

Ciarlo AS, Paredi ME, Fraga AN. 1997. Isolation of soluble collagen from hakeskin (Merluccius hubbsi). J Aquatic Food Prod Technol 6:65–77.

Codex Code. 2004. European community comments for the Codex committee onfood additives and contaminants—agenda item 15(a) and 16(f ). Rome, Italy:FAO.

Conway EJ. 1950. Microdiffusion analysis and volumetric error. London: CosbyLochwood and Son.

Hamada H. 1990. Effects of the preparation conditions on the physical propertiesof shark-skin gelatin gels. Nipp Suis Gakkai 56:671–7.

Kim JS, Cho SY. 1996. Screening for raw material of modified gelatin in marineanimal skins caught in coastal offshore water in Korea. Agric Chem Biotech-nol 39:134–9.

Kim SK, Jeon YJ, Lee BJ, Lee CK. 1996. Purification and characterization of thegelatin from the bone of cod, Gadus macrocephalus. Korean J Life Sci 6:14–26.

Kim JS, Kim IS, Heu MS, Kong CS, Lee TG, Yeum DM, Kang HG. 2002. The principleand application of canned food. 1st ed. Seoul: Hyoil Publishing Co. p 205–12.

Kimura S, Ohno Y. 1987. Fish type I collagen tissue specific existence of 2 molec-ular forms, (�1)2, �2 and �1 �2 �3, in Alaska pollock. Comp Biochem Physiol88b:409–13.

Kimura S, Zhu XP, Matsui R, Shijoh M, Takamizawa S. 1988. Characterization offish muscle type I collagen. J Food Sci 53:1315–8.

[KFDA] Korean Food and Drug Administration. 1999. 1999 food code. Seoul:Moonyeoung Publishing. p 70–2.

Kristinsson HG, Rasco BA. 2000. Biochemical and functional properties of At-lantic salmon (Salmo salar) muscle proteins hydrolyzed with various alkalineproteases. J Agric Food Chem 48:657–66.

Laemmli VK. 1970. Cleavage of structural proteins during the assembly of theheads of Bacteriophage T4. Nature 227:680–5.

Miyauchi Y, Kimura S. 1990. Characterization of a �3 chain from carp skin type Icollagen. Nipp Suis Gakkai 56:1509–14.

Montero P, Borderias J, Turnay J, Leyzarbe MA. 1990. Characterization of hake(Merluccius merluccius L.) and trout (Salmo irideus Gibb) collagen. J AgricFood Chem 38:604–9.

Nagai T, Suzuki N. 2000. Preparation and characterization of several fish bonecollagens. J Food Biochem 24:427–36.

Park JW. 2000. Surimi and surimi seafood. 1st ed. New York, Marcel Dekker. p 127–65.

Park YH, Chang DS, Kim SB. 1995. Seafood processing and its utilization. 1st ed.Seoul: Hyungsul Publishing. p 106–13.

Sadowska M, Rudzki J. 1987. The chemical and functional properties of meatcollagen. Lebnsm-Wiss Technol 20:171–3.

Sathe SK, Salunkhe DK. 1981. Functional properties of the Great Northern Bean(Phaseolus vulgaris L.) proteins: emulsioning, foaming, viscosity, and gela-tion properties. J Food Sci 46:71–4.

Sato K, Yoshinaka R, Sato M, Ikeda S. 1986. A simplified method for determiningcollagen in fish muscle. Bull Jpn Soc Sci Fish 52:889–93.

Sikorski ZE, Borderias J. 1994. Collagen in the muscles and skin of marine an-imals. In: Sikorski ZE, Pan BS, Shahidi F, editors. Seafood proteins. New York:Chapman & Hall, p 58–70.

Steel RGD, Torrie JH. 1980. Principle and procedures of statistics. 1st ed. Tokyo:McGraw-Hill Kogakusha. p 187–221.

Takahashi K, Suzuki A, Wada K. 1989. Gelatinization of pig bone insoluble col-lagen. Nippon Shokuhin Kogyo Gakkaishi 36:538–42.

Tsutagawa Y, Hosogai Y, Kawai H. 1994. Comparison of mineral and phosphoruscontents of muscle and bone in the wild and cultured horse mackerel. J FoodHygenic Soc Jpn 34:315–8.

Voigt MN, Botta JR. 1990. Advances in fisheries technology and biotechnologyfor increased profitability. 1st ed. Lancaster, Pa.: Technomic Publishing Co. p325–33.

Wang JC, Kinsella JE. 1976. Functional properties of novel proteins: alfalfa leafprotein. J Food Sci 41:286–92.

Wendel AP. 1999. Recovery and utilization of Pacific whiting frame meat for suri-mi production [MSci thesis]. Oregon State Univ., Corvallis, Oreg.

Yamaguchi K, Lavety J, Love RM. 1976. The connective tissue of fish. 8. Compar-ative studies of hake, cod and catfish collagens. J Food Technol 11:389–99.

Yang JS. 1997. Applications of gelatin in food and biotechnology. J Food Sci Nutr2:263–8.

Yata M, Yoshida,C, Fujisawa S, Mizuta S, Yoshinaka R. 2001. Identification andcharacterization of molecular species of collagen in fish skin. J Food Sci 66:247–51.

Yim MH, Lee JH. 2000. Functional properties of fractionated soy protein isolatesby proteases from meju. Food Sci Biotechnol 9:253–7.

Zhu XP, Kimura S. 1991. Thermal stability and subunit composition of muscle andskin type I collagens from skipjack. Nipp Suis Gakkai 57:755–60.

Figure 7—Emulsion activity and cooking stability of acid-soluble collagen from various solid byproducts and wholemuscle. Different letters on the bars within the same prop-erties indicate a significant difference at P < 0.05.