Materials Science and Engineering C 33 (2013) 174–181

Contents lists available at SciVerse ScienceDirect

Materials Science and Engineering C journal homepage: www.elsevier.com/locate/msec

Studies on fish scale collagen of Pacific saury (Cololabis saira) Hideki Mori, Yurie Tone, Kouske Shimizu, Kazunori Zikihara, Satoru Tokutomi, Tomoaki Ida, Hideshi Ihara, Masayuki Hara ⁎ Department of Biological Science, Graduate School of Science, Osaka Prefecture University, 1-2 Gakuen-cho, Naka-ku, Sakai, Osaka 599-8570, Japan

a r t i c l e

i n f o

Article history: Received 25 April 2011 Received in revised form 14 July 2012 Accepted 12 August 2012 Available online 19 August 2012 Keywords: Fish scale Collagen Fibril Pacific saury Denaturation temperature Gel

a b s t r a c t We purified and characterized Type I collagen from the scales of the Pacific saury (Cololabis saira) and compared it with collagen from other organisms. Subunit composition of C. saira collagen (2α1+ α2) was similar to that of red sea bream (Pagrus major) and porcine collagen. C. saira collagen did not form a firm gel after neutralization of pH in solution. The temperature of denaturation (24–25 °C) of C. saira collagen was slightly lower than that of P. major collagen (26–27 °C). The contents of proline and hydroxyproline were lower in red sea bream and Pacific saury collagen than in porcine collagen. Circular dichroism spectra and Fourier-transformed infrared spectra showed that heat denaturation caused unfolding of the triple helices in all three collagens. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Although there is considerable progress in fabrication of artificial materials including various nanomaterials [1–3], materials from natural resource keep their values in practical applications [4–6]. The Pacific saury (Cololabis saira) is a popular edible fish caught and commonly eaten in Japan. Catch of this species in the Japanese fishery is approximately 200–300 million tons per year (http://www.maff.go. jp/j/tokei/pdf/gyogyou_seisan_10c.pdf, data were written in Japanese). A large portion of the fish catch comes from the area of the Oyashio Current near the northern part of Japan [7,8]. The scales of C. saira are discarded as waste in the fisheries industry because they can be easily removed from the body onboard the fishing vessel. However, these scales have a number of potential uses. We are interested in developing a novel way to use the proteins, polysaccharides, and other biological polymers contained in this bioresource that traditionally is discarded. In this study, we isolated collagen from the scales of Pacific saury. This species has not been as well studied as other edible fishes such as salmon, cod, flounder, and sea bream [9,10]. Fish scales are classified into several categories [11]. The scale is a very special tissue that exhibits distinct species- or group-specific heterogeneity in morphology and characteristics [11]. For example, scales of Chondrichthyes are placoid, and scales of Osteichthyes are classified into several types such as cosmoid, ganoid, leptoid (ctenoid,

⁎ Corresponding author. Tel./fax: +81 72 254 9842. E-mail address: [email protected] (M. Hara). 0928-4931/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.msec.2012.08.025

and cycloid). Red sea bream (Pagrus major) have ctenoid scales, whereas Pacific saury have cycloid scales. The surface of a fish scale is an osseous layer consisting of randomly oriented collagen fibrils with many hydroxyapatite (Ca10(O4)6(OH)2) crystals. Thin layers of the oriented collagen fibrils are piled up to form the fibrillary plate under the osseous layer [12]. The direction of the fibrils in each thin layer differs from each other. This plywood-like hierarchical supramolecular structure provides the high mechanical strength of scales, and it is of great interest to researchers in the field of biomaterials [12–17]. Fish collagen is present not only in scales but also in skin, bones, fins, and cartilage [9,10]. Scales are completely inedible, whereas fish skin is sometimes eaten. Therefore, fish scale collagen is the focus of studies aimed at using fish collagen for various purposes, including food production, industrial use, medical use, and promoting more effective use of marine bioresources. Although collagens from various fish species have been extracted from scales and skin and characterized [12–23], this study is the first to describe Type I collagen from the Pacific saury. We purified and characterized Type I collagen from the scales of C. saira and compared it with red sea bream and porcine collagen. 2. Materials and methods 2.1. Materials The Eishin Kasei Company, Rausu, Hokkaido Japan provided the scales of C. saira. Red sea bream were purchased from a local fish market in Sakai-city, Osaka, Japan. Porcine Type I collagen solution (0.6% (w/v), pH 3.0, Collagen BM, Nitta Gelatin, Osaka, Japan), Pepstatin A (M9T3556,

H. Mori et al. / Materials Science and Engineering C 33 (2013) 174–181

Nacalai tesque, Kyoto, Japan), a protein assay kit (500-0113,4,5, BioRad, Hercules, CA, USA), phenol (V0G3204, Nacalai tesque), ninhydrin (TFN5028, Wako, Osaka, Japan), buffer solution for amino acid analysis (TFN5028, Wako), standard solution for amino acid analysis (EPM1075, Wako), hydroxylysine (5-hydroxy-DL-lysine hydrochloride)(KWH4600, Wako), hydroxyproline (L-hydroxyproline)(KLL4255, Wako), cellulose tubes (5015-19, Membrane Filtration Products, Inc., Braine-l'Alleud, Belgium), molecular weight markers (LC5801.pps, Invitrogen, Carlsbad, CA, USA), and other chemicals and salts of specific grade were used. Isotonic phosphate buffered saline (PBS) consisted of 0.8 g/l NaCl, 0.02 g/l KCl, 0.02 g/l KH2PO4 (anhydrous), and 0.115 g/l Na2PO4 (anhydrous). 2.2. Scanning electron microscope (SEM) observation of the fish scales Scales from both fish species were washed thoroughly in distilled water and lyophilized in a freeze dryer (FD-1, Eyela, Tokyo, Japan). Samples were coated with platinum using ion-sputtering equipment (E1010, Hitachi, Tokyo, Japan). Fine structures on the outer and inner sides of the scales were observed using a Hitachi SU1510 SEM [24,25]. 2.3. Extraction of collagen from the scales Collagen was extracted from the scales at 4 °C to minimize protein denaturation. Scales were washed and lyophilized as described above, weighed, and then incubated for 1 week in a solution consisting of 50 mM Tris HCl (pH 7.5) and 20 mM sodium ethylenediaminetetraaceticacid (Na-EDTA). Scales were washed again with distilled water and incubated for 2 days in another solution consisting of 500 mM acetic acid and 0.5% (w/v) pepsin. NaCl was added to the solution to generate a final concentration of 0.7 M. The solution was incubated overnight. After addition of pepstatin A (5 μl/ml), the mixture was incubated overnight and then centrifuged for 30 min at 10,000 rpm (CX-210, Tomy, Tokyo, Japan). The precipitate was collected and dissolved in 500 mM acetic acid solution to obtain a collagen solution. Protein concentration was measured by the Bradford method using a protein assay kit to estimate the yield of the preparation.

175

redissolution in distilled water. The sample then was dissolved in 20 mM HCl solution, and the concentration was adjusted to 0.1 nmol/μl. The sample was filtered to remove dust, introduced into a vial, and used for the amino acid analysis, which was performed using an Amino Acid Analyzer (L-8500A, Hitachi). The ninhydrin reaction was used to measure the concentration of amino acids in the samples. Briefly, ninhydrin reagent (500 μl), the buffer solution (500 μl), and the sample solution (5 μl) were mixed and heated for 3 min at 100 °C. Absorbance at 570 nm was measured to estimate the concentration of amino acids using a UV–visible spectrophotometer (U-2000, Hitachi). A standard amino acid solution was used to calibrate the ΔA570 value. Porcine collagen was also hydrolyzed by the same procedure, and then used as a control. 2.6. Circular dichroism (CD) spectra CD spectra of the collagen samples were measured using a CD spectrophotometer (J-720, Jasco, Tokyo, Japan) [24]. Collagen was diluted in 0.2 M sodium phosphate buffer (pH 7.4) to a concentration of 0.1% (w/v). The sample was placed in a quartz cuvette (light path length: 1 mm; S10-UV-1, GL Science, Tokyo, Japan). The temperature was raised in 2 °C intervals from 10 °C to 40 °C, followed by incubation for 5 min before each measurement. An incubator (LP-3100, Advantec, Tokyo, Japan) was used to control the temperature. 2.7. Fourier-transformed infrared (FTIR) spectra Native collagen solution was incubated for 30 min at 80 °C to prepare heat-denatured collagen. Aliquots (1–4 ml) of native collagen solution (0.1–0.6% (w/v)) and heat-denatured collagen were dried in a culture dish (φ = 3 cm) for several days at 10 °C to prepare dry films. The dry films were peeled off the dish and used to measure FTIR spectra at 22 °C using a spectrophotometer (NICOLET 8700, Thermo Scientific, Waltham, MA) with a detector (DTGS KBr) and a sample window of BaF2. The scan was repeated 1024 times to average the data in the range from 400 to 4000 cm −1 with a resolution of 4 cm −1 (data interval 1.928 cm −1). 3. Results

2.4. SEM observation of collagen gels 3.1. Structure of the fish scale surface Collagen solution of either P. major or C. saira (10 ml) in 500 mM acetic acid was dialyzed for 1 week at 4 °C in a cellulose tube against 1 l of 1 mM HCl solution to exchange the solvent. The collagen solution was mixed with a 1 mM HCl/10-fold concentrated PBS/distilled water (5:1:4 volume ratio) to neutralize the pH, and it was incubated at 20 °C to form collagen fibrils. The collagen gel in the test tube was incubated for 3 h at 4 °C in PBS containing 2.5% (w/v) glutaraldehyde to crosslink the protein. The sample was washed for 1 h at 4 °C in PBS, incubated for 1 h in ethanol at 4 °C, and then incubated twice for 20 min each at 37 °C in t-butyl alcohol. The gel was wrapped in plastic film, frozen for 10 min at − 80 °C in a deep freezer, and lyophilized in a freeze dryer overnight. The sample then was coated with platinum and observed by SEM as described above [24,25]. 2.5. Amino acid analysis Red sea bream collagen, and Pacific saury collagen were extracted as described above, lyophilized for 3–4 h, and then weighed. An acid solution consisting of 6 M HCl and 0.1% (w/v) phenol was added to the sample (3 mg/1 ml) in a glass vial with a mininert valve. After the atmosphere in the vial was degassed and twice replaced with nitrogen gas, the sample was hydrolyzed under nitrogen for 24 h at 110 °C in an oil bath. The hydrolyzed sample was dissolved in 0.1 M HCl solution and transferred to a vacuum flask. Acid was removed by five repeated cycles of evaporation in vacuum followed by

Fig. 1A shows a red sea bream scale (left) and a Pacific saury scale (right); the former is colorless (white) and the latter is blue. SEM images of the outer surface (Fig. 1B, C) and inner surface (Fig. 1D, E) of a red sea bream scale and of the outer surface (Fig. 1F, G) and inner surface (Fig. 1H, I) of a Pacific saury scale also are shown. For both fish species, characteristic growth rings (annual ring-like structures) are visible on the outer side of the scale (Fig. 1B, C, F, G) and parallel collagen fibril-like structures without growth rings are visible on the inner side of the scale (Fig. 1D, E, H, I). Many sawtooth-like projections (ctenii) are arrayed in a line at the edge of the growth rings in the ctenoid scale of the red sea bream (Fig. 1C), but they are absent in the cycloid scale of the Pacific saury (Fig. 1G). This is the typical morphological difference between a ctenoid scale and a cycloid scale. 3.2. Extraction of collagen Yield of extracted collagen from the scales was relatively variable depending on the batch of preparation: 7–26% in red sea bream and 4–15% in Pacific saury. Purity of the Type I collagen was high, as indicated by the SDS-PAGE [26] results shown in Fig. 2. A thick α1 subunit protein band and another closely associated α2 subunit protein band of less thickness were observed in collagen from both species of fish and in porcine collagen. These bands represent the typical composition of

176

H. Mori et al. / Materials Science and Engineering C 33 (2013) 174–181

A

B

bar : 20 µm

C

bar : 2 µm

E

bar : 6 µm

F

bar : 20 µm

bar : 8 mm

D

bar : 20 µm

G

bar : 2 µm

H

bar : 20 µm

I

bar : 2 µm

Fig. 1. (A) Photograph of an isolated scale from the red sea bream (Pagrus major, left) and from the Pacific saury (Cololabis saira, right). SEM image of the outer surface (B, C) and inner surface (D, E) of a red sea bream scale and of the outer surface (F, G) and inner surface (H, I) of a Pacific saury scale.

subunits in Type I collagen (2α1 + α2). A covalently crosslinked dimer of α subunits (β-band) was observed in all three collagens, but the trimer of α subunits (γ-band) was observed only in the porcine collagen. 3.3. Preparation of collagen gel For both types of fish collagen, we prepared collagen gels containing collagen fibrils by pH neutralization in solution. Fibril formation began

A

C

B γ β

260 kDa 160 kDa 110 kDa

D β α1 α2

α1 α2

β α1 α2

60 kDa 50 kDa

30 kDa

Molecular weight markers

Porcine collagen

Red sea bream collagen

Pacifc saury collagen

Fig. 2. SDS-PAGE results for the molecular weight markers (A), porcine collagen (B), scale collagen from red sea bream (Pagrus major) (C), and scale collagen from Pacific saury (Cololabis saira) (D) on a 7.5% (w/v) polyacrylamide gel. Arrows show the α1 subunit, α2 subunit, β band (covalently crosslinked dimer of α subunits), and γ band (covalently crosslinked trimer of α subunits).

within 3–4 min after neutralization, and a firm hydrogel was formed after overnight incubation in the case of red sea bream collagen (Fig. 3A). The neutralized solution of Pacific saury collagen did not form a firm gel; the solution became only slightly turbid due to partial gel formation, even after overnight incubation (Fig. 3D). Mechanical strength of the scale collagen gels seemed to be far lower than that of the porcine collagen gel, because the scale collagen gels were rather soft and fragile. Firm hydrogel was formed when the porcine collagen was used in the same experimental procedure [24,25]. The collagen gels of the red sea bream (Fig. 3B, C) and those of Pacific saury (Fig. 3E, F) were observed by SEM after the necessary pretreatment. Both types of gel contained collagen fibrils, but those from Pacific saury collagen were thinner and less dense. 3.4. Amino acid composition Table 1 summarizes the results of the amino acid analysis of the porcine, red sea bream, and Pacific saury collagen. The values of percentage (%) in the contents of amino acids in the table were obtained after acidic hydrolysis of protein for 24 h; hydrolysis for either 12 h or 48 h generated similar results (data not shown). Proline (Pro) and hydroxyproline (Hyp) contents were lower in both red sea bream collagen (Hyp: 1.19, Pro: 10.19) and Pacific saury collagen (Hyp: 1.33, Pro: 9.10) than in porcine collagen (Hyp: 1.82, Pro: 12.5). Content of phenylalanine (Phe), tyrosine (Tyr), histidine (His), and threonine (Thr) were higher in both types of fish collagen than in porcine collagen. 3.5. Secondary structure of collagen and heat denaturation Fig. 4 shows the CD spectra of the red sea bream (Fig. 4A) and Pacific saury (Fig. 4C) collagen measured after increasing the temperature

H. Mori et al. / Materials Science and Engineering C 33 (2013) 174–181

177

A

B

bar: 10 µm

C

bar:1 µm

D

E

bar:10 µm

F

bar:1 µm

Fig. 3. Photographs (A, D) and SEM images of the collagen gels (B, C, E, F) formed from collagen from red sea bream (Pagrus major) (A, B, C) and Pacific saury (Cololabis saira) (D, E, F).

from 10 to 40 °C. As reported previously [24], intact collagen molecules with a triple helical structure have a positive peak (maximum) at 221 nm and a negative peak (minimum) at 191 nm in the molecular ellipticity of the CD spectrum. Those peaks diminish when the triple helices are destroyed by heat denaturation. Thus, the molecular ellipticity at 200 nm was plotted versus temperature to estimate the temperature at which denaturation occurs. The half-denaturation temperature was approximately 26–27 °C for red sea bream collagen (Fig. 4B) and 24–25 °C for Pacific saury collagen (Fig. 4D). Therefore, denaturation temperature for Pacific saury collagen was slightly lower (by approximately 2 °C) than that of red sea bream collagen and far lower than that of the porcine collagen (~40 °C, see Fig. 7 of [24]). 3.6. FTIR spectra Fig. 5A–F shows the FTIR spectrum of native porcine collagen, denatured porcine collagen, native red sea bream collagen, denatured

Table 1 Percentage (%) in the content of amino acids in porcine collagen, red sea bream collagen, and Pacific saury collagen after hydrolysis for 24 h at 110 °C. Amino acid

Porcine collagen

Red sea bream scale collagen

Pacific saury scale collagen

Asp Thr Ser Glu Gly Ala Cys Val Met Ile Leu Tyr Phe Hlys Lys His Arg Hpro Pro Total

4.82 1.66 2.62 7.5 36.7 12.13 0.53 3.33 1.1 1.29 2.58 0.53 1.5 0.52 2.92 0.88 5.05 1.82 12.5 100

5.3 2.71 3.79 7.97 35.31 13.65 0.32 2.74 1.12 1.2 2.32 0.59 1.6 0.44 3.35 1.07 5.14 1.19 10.19 100

6.14 3.02 5.18 7.08 33.54 11.59 0.48 3.43 1.49 1.62 2.99 1.14 1.77 0.63 3.34 1.31 4.81 1.33 9.1 100

red sea bream collagen, native Pacific saury collagen, and denatured Pacific saury collagen, respectively. All of the spectra appear similar, and they also are similar to the spectra reported by other groups [17–25]. Table 2 lists the wave number (cm −1) that corresponds to the major peak in the FTIR spectra for each type of collagen. These values come directly from the measured spectra and not from the deconvolved peak. Ratios between major peaks in height from the FTIR spectra are summarized in Table 3. The Amide A band at 3315–3330 cm −1 and Amide B band at 3077–3083 cm −1 were observed, as were the Amide I band at 1656–1660 cm −1, Amide II band at 1550–1554 cm −1, and Amide III band at 1238–1240 cm −1. The band representing the CH2-asymmetric vibration mode at 2958–2980 cm−1 and that of the CH3-asymmetric vibration mode also were observed [27]. Another major peak at 1450–1456 cm−1 was observed between Amide II and Amide III (written as 1454 cm−1 in Table 2). Bryan et al. reported that a peak at 1654 cm−1 broadened and shifted to 1640 cm−1 when the triple helix structure was unfolded by heating [28]. Videl and Mello showed that a peak at 1660 cm−1 shifted to 1630 cm−1 when the native collagen was heat denatured [29]. The triple helical structure of Type I collagen consists of three alpha subunits (2α1 + α2) [30]. Those subunits in the polyprolinelike conformation are associated with each other in a supercoiled structure [29–32]. Polyproline itself also exhibits the peak at 1630 cm − 1, just like the denatured collagen [28]. The data summarized in Table 2 show that all three collagens exhibited a small shift of Amide I peaks from 1660–1658 cm − 1 to 1652–1656 cm − 1 after heat denaturation, and this finding is in accordance with the reports described above [29]. The Amide I band derives mainly from C_O stretching. The peak of Amide A (NH stretching), Amide II (NH bending, CN stretching), Amide III (NH bending in a plane, CN stretching, CH2 vibration of the glycine backbone and that of the proline side chain), and the peak at 1451–1450 cm − 1 (pyrrolidine vibration of proline and that of hydroxyproline) have been described previously in the literature [33,34]. The data in Fig. 5 and Tables 2 and 3 show that the Amide II/Amide I ratio and the Amide III/Amide I ratio decreased and that a shoulder peak in the Amide A (left side) also decreased (became less clear) when the collagen was denatured. Rabotyagova et al. reported previously that the shoulder peak at around 3430–3500 cm−1 corresponds to the hydrogen bonds in the native collagen molecule [27]. Thus, the decrease in the shoulder peak observed herein likely reflects the deformation of

178

H. Mori et al. / Materials Science and Engineering C 33 (2013) 174–181

0 40°C 30°Ca 28°C

-5,000 -10,000

26°C

-15,000

24°C 10°C

-20,000 -25,000 190

200

210

220

230

240

Wavelength (nm)

Θ [deg cm2 dmol-1]

C

Θ [deg cm2 dmol-1]

B 5,000

0 -5000 -10000 -15000 -20000 0

10

20

30

40

50

40

50

Temperature (°C)

D 5,000 0 -5,000 26°C 24°C

-10,000 -15,000

40°C 28°C

22°C 10°C

-20,000 -25,000 190

200

210

220

230

240

Wavelength (nm)

Θ [deg cm2 dmol-1]

Θ [deg cm2 dmol-1]

A

0 -5000 -10000 -15000 -20000

0

10

20

30

Temperature (°C)

Fig. 4. CD spectra of the collagen from red sea bream (Pagrus major) (A) and Pacific saury (Cololabis saira) (C), accompanied by a graph showing the heat-denaturation curve estimated from the temperature-dependent change in molecular ellipticity at 200 nm (B and D, respectively).

Absorbance

0.4

3

2

2700

1700

Wave number

0 700

3700

(cm-1)

2700

1700

Wave number

D

2700

1700

0 700

Wave number (cm-1)

3700

2700

1700

Wave number

0 700

(cm-1)

F

2.5

2.5

2

Absorbance

1.5 1

2 1.5 1

0.5

0.2

3700

0.2

0 700

0.8

0.4

0.4

(cm-1)

E

1

0.6

0.6

1

0.2

3700

0.8

Absorbance

0.8 0.6

C

4

3700

2700

1700

0 700

Wave number (cm-1)

Absorbance

B

1

Absorbance

A

ratio decreased after denaturation for all three collagens (Table 3). Assuming that the Amide I band did not decrease much due to heat denaturation, the results indicate that the Amide II and Amide III bands and the peak at 1454 cm−1 clearly decreased after heat denaturation. The Amide III/Amide I ratio and the Amide III/1454 cm−1 ratio also clearly decreased after denaturation in the two fish collagens. This shift likely

Absorbance

the triple helical structure that is supported by inter- and intramolecular hydrogen bonds. The Amide III/(peak at 1454 cm−1) ratio, which reflects the content of the triple helix [33], also decreased after denaturation. Sylvester et al. reported that the 1235 cm−1/1454 cm−1 ratio was 1.04–1.13 in collagen film but only 0.59 in gelatin film [35]. Gelatin is a denatured form of collagen. In our study, the 1235 cm−1/1454 cm−1

0.5

3700

2700

1700

0 700

Wave number (cm-1)

Fig. 5. FTIR spectra of porcine (A, B), sea bream (C, D), and Pacific saury (E, F) collagen. Native collagen (A, C, E) and heat-denatured collagen (B, D, F) were used to prepare the dry film as described in the Materials and methods section. Wave numbers of the major peak and ratios of the height of the major peaks are summarized in Tables 2 and 3, respectively.

H. Mori et al. / Materials Science and Engineering C 33 (2013) 174–181 Table 2 Wave number (cm−1) corresponding to the major peaks in the FTIR spectra. Porcine collagen

Red sea bream scale collagen

Pacific saury scale collagen

Native

Denatured

Native

Denatured

Native

Denatured

3330 3090 2980 2960 1660 1550 1450 1240

3315 3077 2958 2940 1652 1544 1454 1241

3326 3077 2977 2935 1658 1552 1456 1240

3322 3077 2977 2939 1656 1548 1454 1240

3324 3081 2958 2940 1658 1554 1456 1238

3318 3079 2958 2939 1656 1550 1454 1240

Table 3 Ratio of height of major peaks in the FTIR spectra. Ratio of (peak/peak)

Amide A Amide B CH2-vibration CH3-vibration Amide I Amide II

179

Amide II/Amide I Amide III/Amide I Amide III/ 1454 cm−1 1454 cm−1/ Amide I

Porcine collagen

Red sea bream scale collagen

Pacific saury scale collagen

Native

Denatured

Native

Denatured

Native

Denatured

0.679 0.355 1.15

0.463 0.204 0.765

0.705 0.417 1.21

0.599 0.256 0.858

0.667 0.349 1.23

0.583 0.279 1.03

0.309

0.266

0.344

0.298

0.284

0.270

Amide III

reflects unfolding of the triple helices. Overall, the FTIR data indicate that all three collagen exhibited unfolding of the triple helices due to heat denaturation at 80 °C for 30 min.

4. Discussion 4.1. Heat stability of the collagens In the quest to understand the relationship between protein structure and stability, Type I collagen is one of the most extensively studied proteins [31,32]. Various forces and bonds contribute the heat stability of the collagen molecule. Three subunits (2α1 + α2) bind to each other in a triple helical structure via intersubunit (intramolecular) hydrogen bonding contributed by hydroxyproline. In aqueous solution at neutral pH, collagen molecules attach to each other on their lateral surfaces by intermolecular hydrophobic interactions. Collagen molecules are arrayed in parallel in a quarter-staggered manner to form collagen fibrils with a D-period (i.e., a stripe with an interval of approximately 67 nm). If the collagen molecules have some negative or positive net charge on their surface, electrostatic repulsion will occur and may inhibit the hydrophobic interaction. Koshimizu et al. reported that mammalian Type I collagen molecules in collagen fibrils have higher heat stability (denaturation temperature of 50–60 °C) than collagen in solution (denaturation temperature of 40–45 °C) [25]. The amino acid sequence in collagen subunits contains repeated triplet of -Gly-X-Y- (X is a variable and frequently is Pro, and Y is variable and frequently is Hyp). The side residue of Gly (-H) is small and always compactly oriented in the central axis of the triple helix. Therefore, it is difficult to replace the Gly with another amino acid. Side chains of amino acids in the X position and those in the Y position are oriented outward of the triple helix. Thus, amino acids with relatively bulky side chains are acceptable in these positions. If many bulky amino acids are in the X or Y position, the surface of the collagen molecule has some roughness, which may inhibit fibril formation by disturbing intermolecular adhesion based on the hydrophobic interaction [32]. If the amino acids in the X position and Y position are not so bulky and hydrophobic, they will be able to enhance intermolecular adhesion and contribute to fibril formation. These general rules about amino acid sequence and heat stability have been experimentally evaluated by comparing the ratio of each amino acid in the X position and the Y position with the heat stability of whole molecules. Our results suggest that the lower content of Pro and Hyp in the collagens of both fishes contributed to their lower heat stability compared with that of porcine collagen. The lower content of these amino acids likely reduced the number of the intersubunit (intramolecular) hydrogen bonds and water bridges [27,28]. The higher content of bulky amino acids such as Phe, Tyr, and His in the collagens of both fishes also might have contributed to the lower heat stability of these samples by inhibiting the intermolecular interaction.

Another important factor to be considered is the electrostatic interaction between charged amino acids. Thus, we compared the content of charged amino acids (negatively charged Asp and Glu, positively charged Lys and Arg) in the three sample types (Table 1). No big differences among the three collagen types were observed. Therefore, we concluded that electrostatic interaction did not contribute much to the differences in fibril formation or heat stability observed in this study. Forces and molecular interactions affecting the stability of collagen as described above are in major triple helical region of the intact collagen molecule. Besides those non-covalent interactions influencing the stability of collagen, the melting temperature of collagen in fibrils depend on both inter- and intramolecular covalent crosslink between collagen subunits in the telopeptide region, too. The result of SDS-PAGE in Fig. 2D showed the presence of a β band (covalently crosslinked dimer of α subunits) in Pacific saury collagen, as similar to the case of other fishes [19]. If the collagen was extracted in cold acid solution without pepsin, the yield in extraction (0.18–1.7%) was far lower than the present value (7–26%) described in this report. Those results strongly suggested the presence of natural crosslink between the subunits of collagen in Pacific saury scales. 4.2. Collagens from other fishes and organisms Type I collagen has been extracted and purified from various fish species, including salmon [21], sardine [20] cod, shark, flounder, red sea bream [13,14], tilapia [23], Lates calcarifer [16], Corvina fish [36], Arapaima gigas [37], carp [15], spotted golden goatfish [33], Labeo rohita [22], Catla catla [22], skipjack tuna [19], Japanese sea bass [19], and horse mackerel [19], and has been characterized biochemically [9,10]. Collagens from various invertebrates, including jelly fish [38], seastars [39], sea cucumber [40,41], demosponges [42], and sea anemones [42], also have been studied. Recently, comparative studies about collagens from animals belonging to various phyla have become popular among biologists studying the evolution of collagen molecules and connective tissues that mechanically support the body [43]. The higher order in the content of hydroxyproline of Type I collagen (porcine collagen> tilapia collagen > red sea bream collagen) was the same as the higher order of heat stability depending on the average temperature of the habitat except the case of mammals and birds, according to the literature [23,9]. Recently it was also reported by another group that a deep-sea fish had collagen of extremely low melting temperature [44]. Based on the literature, it is reasonable that Pacific saury collagen had a slightly lower denaturation temperature than that of red sea bream. This result seems to correspond to the average temperature of the habitat of Pacific saury (i.e., the Oyashio Current). Most previous studies of scale collagens focused on ctenoid scales, although Pati et al. recently reported that heat-stable collagen was extracted from the cycloid scale of a fresh-water fish from a tropical area [22]. Generally, collagen fibrils are more stable than the extracted collagen molecule because both inter- and intramolecular interaction stabilized the triple helical structure against the thermal perturbation. Melting temperature of collagen fibrils is generally higher than that of extracted collagen as exemplified previously about porcine

180

H. Mori et al. / Materials Science and Engineering C 33 (2013) 174–181

collagen [25]. Therefore the melting temperature of the collagen fibrils in Pacific saury scale is supposed to be higher than that of extracted collagen at 24–25 °C in this report estimated by CD spectral data, and higher than the temperature of sea water in the habitat of Pacific saury, although we have not proved by measuring the melting temperature by DSC yet in this report. We have not carried out TGA of the Pacific saury scale collagen because the equipment was not available. Dried collagen fibrils are generally more stable than the wet ones because the thermal perturbation was suppressed and also because sometimes dehydrothermal crosslinking will be formed in dry state [45], although a controversial data was recently reported that collagen from scales of A. gigas was more stable with higher water content from thermal gravimetry analysis (TGA) and differential scanning calorimetry (DSC) [37]. Collagen sheet prepared from L. calcarifer scales showed temperature-dependent weight loss in three steps: loss of water around 52 °C, decomposition of protein around 293 °C, and loss of inorganic phase around 592 °C [46]. It was recently reported that when the fish collagen solution and porcine collagen solution were mixed and then incubated at neutral pH, hybrid fibrils of both collagens were formed and showed the intermediate melting temperature between the melting temperature of porcine collagen and that of fish one [47]. If the strategy of constructing hybrid fibrils is universally applicable to many species, suitable melting temperature of collagen fibrils will be controllable by mixing the collagen from different species. 4.3. Application of scale collagens for food, medical and industrial use Fish collagens have a number of potential applications. For example, use of fish collagen and fish gelatin as alternatives to mammalian collagen and gelatin in the food and pharmaceutical industries is gaining attention [9,10]. One reason for this interest is concern about the potential for contracting infectious diseases from mammalian products (e.g., Creutzfeld–Jakob disease). Another reason is the public's interest in reducing the use of materials from mammals, even when safety has been proven experimentally and use has been approved by the regulatory agencies of the government. A third reason for interest in fish collagen is that some people do not eat porcine or bovine proteins for religious reasons (e.g., Muslims, Jews, and Hindus). Generally speaking, extracted native mammalian collagen, which is rather expensive to extract, is used in biomedical devices and as a scaffold in tissue engineering [48]. Denatured collagen and gelatins from various animal sources are widely used in the food, pharmaceutical, and chemical industries (e.g., as jelly, glue, thickener, capsule, and film) [9,10]. Collagen peptides, which are partially digested collagens, are used as food ingredients, neutraceuticals, etc. [9,10]. Native collagen, denatured one (gelatin), and digested one (collagen peptide) were extracted from fish scale, skin, fin, bone, etc., and then used for various purposes in those applications as described above. There is another strategy in application of fish scale collagen. Fish scale is a natural composite material consisting of hydroxyapatite crystals and collagen fibrils, like a plywood-like hierarchical supramolecular structure as described previously in Introduction [12–17]. Therefore it can be used as transparent biomaterials of both elasticity and high tensile strength after removal of calcium phosphate (hydroxyapatite). Krishnan et al. compared the property of decalcified scale, that of cornea, and that of human amniotic membrane as possible scaffolds in tissue engineering [16]. The amniotic membranes are also mechanically strong, elastic, biocompatible sheets consisting of orthogonally oriented collagen fibrils, and are good scaffolds especially for epidermal tissues including artificial cornea in tissue engineering for ophthalmology [49]. Moura et al. reported that glutaraldehyde-crosslinked fish scales could be used as adsorbent for heavy metal ions like a toxic chromium ion (Cr(VI)) [36]. That is another strategy in utilization of fish scale for environmental protection. Lower cost and less effort are necessary for

quality control of the product made of fish scale in such an environmental technology when compared to that in producing implantable medical devices [48,49], as described in former paragraph. To date, Type I collagen from the scales of the Pacific saury is an unutilized material, and it has potential for further application. It is not suitable for preparing collagen gel for biomedical devices or as a scaffold for cell culture because of its low heat stability and low capacity for fibril formation. However, Type I collagen from the scales of Pacific saury can be processed into gelatin (i.e., heat-denatured collagen) or collagen peptides (digested collagen) for use in food products. It might also be used as a thickener in food processing and as industrial glue. Catching of Pacific saury in domestic fisheries is in large amount (200–300 million tons/year) as described in the Introduction, because it is a very popular edible fish, caught in the ocean, in Japan. Therefore the scales will be relatively easy to collect in large amounts as raw materials. However it has a disadvantage in quality control of scales as raw materials in terms of tracking the growth condition of fish in its habitat, and in the constant supply of the same amount in every year, in comparison with the cultured fishes like Tilapia or sea bream [14,23]. With these situations and a relatively low melting temperature, we consider that the possible application of the Pacific saury scale collagen is more suitable for food additives such as a thickener of the solution, glues for industrial use, or as an adsorbent for toxic substances or metal ions, rather than for the construction of scaffold in tissue engineering, although we have not tried the real industrial application yet. 5. Conclusion We purified Type I collagen from the scales of the Pacific saury, and characterized it as a first report. Its subunit composition (2α1 + α2) was similar to the red sea bream collagen and porcine collagen. The Pacific saury collagen did not become a firm gel after the neutralization of pH. The temperature of denaturation (24–25 °C) was slightly lower than that of the red sea bream collagen (26–27 °C). Both of these fish collagens contain a lower amount of proline and hydroxyproline than the porcine collagen. Acknowledgment This research was partially supported by the Asahi Beer Science Promotion Foundation and the Sentan Kagaku Kyoudou Kenkyu Project of Osaka Prefecture University (OPU). The authors would like to thank the members of the Eishin Kasei Company and Prof. Emer. Yoshihisa Nakano of OPU for their support of the project. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16]

K. Ariga, A. Vinu, Y. Yamauchi, Q. Ji, J.P. Hill, Bull. Chem. Soc. Jpn. 85 (2012) 1. G. Wang, L. Zhang, J. Zhang, Chem. Soc. Rev. 41 (2012) 797. M. Perfézou, A. Turner, A. Merkoçi, Chem. Soc. Rev. 41 (2012) 2606. P.L. Hamilton, D.P. Arya, Nat. Prod. Rep. 29 (2012) 134. D.J. Newman, G.M. Cragg, J. Nat. Prod. 75 (2012) 311. J.W. Blunt, B.R. Copp, R.A. Keyzers, M.H.G. Munro, M.R. Prinsep, Nat. Prod. Rep. 29 (2012) 144. W.-B. Huang, Fish. Sci. 76 (2010) 21. Y. Tian, T. Akamine, M. Suda, Fish. Res. 60 (2003) 439. A.A. Karim, R. Bhat, Food Hydrocoll. 23 (2009) 563. M.C. Gomez-Guillen, B. Gimenez, M.E. Lopez-Caballero, M.P. Montero, Food Hydrocoll. 25 (2011) 1813. C.D. Roberts, Bull. Mar. Sci. 52 (1993) 60. M. Okuda, N. Ogawa, M. Takeguchi, A. Hashimoto, M. Takaya, S. Chen, N. Hanagata, T. Ikoma, Microsc. Microanal. 17 (2011) 788. H.S. Youn, T.J. Shin, J. Struct. Biol. 168 (2009) 332. H. Ikoma, H. Kobayashi, J. Tanaka, D. Walsh, S. Mann, J. Struct. Biol. 142 (2003) 327. A.M.C. Garrano, G. La Rosa, D. Zhang, L.-N. Niu, F.R. Tay, H. Majd, J. Mech. Behav. Biomed. Mater. 7 (2012) 17. S. Krishnan, S. Sekar, M.F. Katheem, S. Krishnakumar, T.P. Sastry, Artif. Organs (2012), http://dx.doi.org/10.1111/j.1525-1594.2012.01452.x.

H. Mori et al. / Materials Science and Engineering C 33 (2013) 174–181 [17] F. Pati, P. Datta, B. Adhikari, S. Dhara, K. Gosh, P. Kumar, D. Mohapatra, J. Biomed. Mater. Res. A 100A (2012) 1068. [18] T. Nagai, M. Izumi, M. Ishii, Int. J. Food Sci. Technol. 39 (2004) 239. [19] T. Nagai, N. Suzuki, Food Chem. 68 (2000) 277. [20] Y. Nomura, H. Sakai, Y. Ishii, K. Shirai, Biosci. Biotechnol. Biochem. 60 (1996) 2092. [21] S. Yunoki, T. Suzuki, M. Takai, J. Biosci. Bioeng. 96 (2003) 575. [22] F. Pati, B. Adhikari, S. Dhara, Bioresour. Technol. 101 (2010) 3737. [23] T. Ikoma, H. Kobayashi, J. Tanaka, D. Walsh, S. Mann, Int. J. Biol. Mol. 32 (2003) 199. [24] N. Inoue, M. Bessho, M. Furuta, T. Kojima, S. Okuda, M. Hara, J. Biomater. Sci. Polym. Ed. 17 (2006) 837. [25] N. Koshimizu, M. Bessho, S. Suzuki, Y. Yuguchi, S. Kitamura, M. Hara, Biosci. Biotechnol. Biochem. 73 (2009) 1915. [26] U.K. Laemmli, Nature 227 (1970) 680. [27] O.S. Rabotyagova, P. Cebe, D.L. Kaplan, Mater. Sci. Eng. C28 (2006) 1420. [28] M.A. Bryan, J.W. Brauner, G. Anderle, C.R. Flach, B. Brodsky, R. Mendelsohn, J. Am. Chem. Soc. 129 (2007) 7877. [29] B.D.C. Videl, M.L.S. Mello, Micron 42 (2011) 283. [30] K. Beck, B. Brodsky, J. Struct. Biol. 122 (1998) 17. [31] J.P.R.O. Orgel, A. Miller, T.C. Lrving, R.F. Fischetti, A.P. Hammersley, T.J. Wess, Structure 9 (2001) 1061. [32] A.V. Persikov, J.A.M. Ramshaw, A. Kirkpatrick, B. Broadsky, Biochemistry 39 (2000) 14960. [33] K. Matmaroh, S. Benjakul, T. Prodpran, A.B. Encarnacion, H. Kishimura, Food Chem. 129 (2011) 1179.

[34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49]

181

A.M.D. Plepis, G. Goissis, D.K. Das-Guputa, Polym. Eng. Sci. 36 (1996) 2932. M.F. Sylvester, I.V. Yannas, E.W. Salzman, M.J. Forbes, Thromb. Res. 55 (1989) 135. K.O. Moura, E.F.S. Viera, A.R. Cestari, J. Appl. Polym. Sci. 124 (2012) 3208. F.G. Torres, O.P. Troncoso, E. Amaya, Mater. Sci. Eng. C (2012), http: //dx.doi.org/10.1016/j.msec.2012.06.003. S. Miura, S. Kimura, J. Biol. Chem. 260 (1985) 15352. S. Kimura, Y. Omura, M. Ishida, H. Shirai, Comp. Biochem. Physiol. 104B (1993) 663. J.A. Trotter, G. Lyons-Levy, F.A. Thurmound, T.J. Koob, Comp. Biochem. Physiol. 112A (1995) 463. J.-Y. Exposite, U. Valcourt, C. Cluzel, C. Lethias, Int. J. Mol. Sci. 11 (2010) 407. J.-Y. Exposito, C. Larroux, C. Cluzel, U. Valcourt, C. Lethias, B.M. Degnanm, J. Biol. Chem. 283 (2008) 28226. A.L. Rychel, S.E. Smith, H.T. Shimamoto, B.J. Swalla, Mol. Biol. Evol. 23 (2006) 541. L. Wang, X. An, F. Yang, Z. Xin, L. Zhao, Q. Hu, Food Chem. 108 (2008) 616. I.V. Yannas, A.V. Tovolsky, Nature 215 (1967) 509. S. Sankar, S. Sekar, R. Mohan, S. Rani, J. Sundaraseelan, T.P. Sastry, Int. J. Biol. Macromol. 42 (2008) 6. S. Chen, T. Ikoma, N. Ogawa, S. Migita, H. Kobayashi, N. Hanagata, Sci. Technol. Adv. Mater. 11 (2010) 035001. S.F. Badylak, in: A. Atala, R.P. Lanza (Eds.), Methods in Tissue Engineering, Academic Press, San Diego, CA, 2002, p. 505. A.K. Riau, R.W. Beuerman, L.S. Lim, J.S. Mehta, Biomaterials 31 (2010) 216.