Silver carp skin was delivered on ice by a local fishery (Tulcea, Romania). Pepsin (EC 3.4.23.1), proteinase K (EC 3.4.21.64), Sephadex G-75, bicinchoninic acid (BCA) 2,4,6-trinitrobenzenesulfonic acid (TNBS), 2,2’-azino’-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) and 1,1-diphenyl-2-picrylhydrazyl (DPPH) were purchased from Sigma-Aldrich, Merck (Taufkirchen, Germany). All other chemicals of analytical grade used in experiments were purchased from Sigma-Aldrich, Merck, unless specified otherwise. A culture cell line of mouse fibroblasts NCTC clone L929 and European Collection of Authenticated Cell Cultures (ECACC) maintained on Eagle’s minimum essential medium (MEM) were purchased from Sigma-Aldrich, Merck.

Collagen was extracted from silver carp skin using the enzymatic method with pepsin in acetic acid according to a previously described protocol (21). Briefly, skin was thoroughly washed in cold tap water, minced and extracted in 0.5 M acetic acid containing pepsin (2500 U/mg), in an enzyme/tissue mass ratio 1:10, at 4 °C, with gentle stirring, for 24 h. The process was repeated twice. The gel was separated from the undigested tissue, purified by precipitation with 0.7 M NaCl, centrifuged at 5000×g for 20 min and dialyzed against distilled water. The resulting carp skin collagen gel mass fraction was 1.2% with pH=5.5 and it was stored at -18 °C for 2 months.

Enzymatic hydrolysis of carp skin collagen was carried out under specific conditions of pH and temperature (Fig. 1). The treatment protocol consisted of incubation of diluted carp skin collagen in 0.05 M Tris buffer, pH=8, containing 0.15 M NaCl and 5 mM CaCl₂ at 45 °C in a water bath for 30 min. Then, proteinase K (10 U/mL) in the same buffer, pH=8, supplemented with 0.5% sodium dodecyl sulfate (SDS), as enzyme activator, was added and incubation continued at 55 °C for 6 h. The pH of the reaction mixture was maintained at an optimum value of 8 during incubation. Finally, the solution was heated at 95 °C for 5 min to inactivate the enzyme, then cooled at room temperature and centrifuged at 8000×g and 4 °C for 15 min. The supernatant representing the enzymatic preparation of carp skin collagen was subjected to centrifugal ultrafiltration using units with cellulose membrane at a molecular mass cut-off of 3 kDa (Amicon, Taufkirchen, Germany). Thepermeate containing carp skin collagen peptides was collected and stored at -18 °C until analysis. Each experiment was performed on 10 samples of carp skin collagen. All samples were analyzed for protein content by BCA assay and a standard curve was built with pig skin gelatin in the range of 0.1-3 mg/mL in order to calculate the process yield. Hydroxyproline (Hyp) content was determined in alkaline hydrolysate using chloramine-T and Ehrlich’s reagent, as previously described (22).

The concentration of free ɛ-amino groups was determined using TNBS assay, as previously described (23), recording the absorbance (A) of the resulting complex at 346 nm using a UV-Vis spectrophotometer V650 (Jasco, Tokyo, Japan). The degree of hydrolysis (DH) was calculated as a percentage relative to total hydrolysate prepared by carp skin collagen (CSC) incubation in 6 M HCl at 110 °C for 18 h, using the following equation:

Carp skin collagen obtained by pepsin hydrolysis was analyzed by SDS-polyacrylamide electrophoresis, according to the method of Laemmli (24). The sample was denatured at 50 °C for 30 min and loaded on 7.5% SDS-polyacrylamide (PAA) resolving gel provided with 5% stacking gel. Migration was carried out at 30 V for 4 h using a Biometra source (Jena, Germany). A sample of rat tail tendon collagen was migrated under the same conditions. The high molecular mass marker (55−250 kDa) served to estimate the molecular mass of separated bands based on their retention factor (Rf).

The enzymatic preparations of carp skin collagen and the carp skin collagen peptides were denatured at 37 °C for 5 min and migrated in 10-20% gradient tricine SDS-PAA gel (Invitrogen, Waltham, MA, USA) at 90 V for 2.5 h (25). Low molecular mass marker (1.7−40 kDa) was migrated under the same conditions. The migration patterns were visualized by gel staining with Coomassie Brilliant Blue.

Gel filtration chromatography of the enzymatic preparations of the carp skin collagen and the carp skin collagen peptides was performed on a glass column (230 mm×15 mm) packed with swollen Sephadex G-75 gel, equilibrated with distilled water, as previously described (26). Samples (2 mg/mL) were loaded, eluted at a flow rate of 0.1 mL/min and monitored at 220 nm using a UV-Vis spectrophotometer V-650 (Jasco). Bovine serum albumin (66 000 Da), cytochrome c (12 384 Da), insulin (5777 Da) and bacitracin (1423 Da) were used as standards to build a molecular mass calibration curve for the calculation of the molecular mass of separated fractions.

In vitro cytocompatibility of carp skin collagen peptides was evaluated according to the international standard ISO 10993-5:2009 (29) for medical devices by direct contact method. Briefly, mouse fibroblasts from L929 cell line were seeded in 96-well culture plate, at a cell density of 4·10⁴ cell/mL, and cultivated in MEM culture medium supplemented with 10% foetal bovine serum and 1% mixture of antibiotics. The plates were incubated at 37 °C in 5% CO₂ humidified atmosphere. After 24 h, the medium was replaced with a fresh medium containing different concentrations of carp skin collagen peptides (10−400 µg/mL) and the incubation continued under standard conditions for 24 h. Cells cultivated in MEM served as negative control, while cells treated with 0.5 mM H₂O₂ served as positive control. Cells were also treated with 15 µg/mL ascorbic acid, a known agent for stimulation of fibroblast proliferation. Cell viability was evaluated by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay, as previously described (30). The absorbance values read at 570 nm were proportional to the number of viable cells, which reduced MTT to formazan salt. Cell viability was expressed as percentage relative to control, considered 100%. The experiments were performed in triplicate.

Qualitative analysis consisted of light microscopy observations of cell morphology, under similar conditions as described above. Thus, L929 fibroblasts were cultivated in the presence of different concentrations of carp skin collagen peptides and incubated under standard conditions for 24 h. After medium harvesting, cells were washed, fixed in methanol and stained with Giemsa solution. Micrographs were acquired with an Axio Observer D1 optical microscope equipped with a digital camera (Carl Zeiss, Oberkochen, Germany).

The effect of carp skin collagen peptides on cell migration during wound healing has been evaluated in an experimental model of scratch-wound achieved in a monolayer of fibroblasts, as previously described (31). Briefly, L929 fibroblasts were cultivated in 35-mm Petri dishes in MEM culture medium until confluence. Then, a mechanical scratch was made on the monolayer using a pipette tip and fresh medium containing 100 µg/mL carp skin collagen peptides was added. Untreated cells served as negative control and cells treated with ascorbic acid (15 µg/mL) served as positive control. After 24 h of cultivation under standard conditions, cells were photographed using an AxioStar Plus phase contrast microscope (Carl Zeiss). The rate of wound repair was calculated after image analysis using ImageJ software (32). Thus, the density of migrated cells within the wounded area was calculated by automatic measurement of cells as particles in the captured images and then reported as percentage of the initial open area of the scratch.

The effect of carp skin collagen peptides on ROS production was evaluated in an experimental model of oxidative stress induced in the fibroblast culture, as previously described (33). L929 cells were seeded in a 12-wells culture plate at a density of 5·10⁴ cell/mL and cultivated in MEM under standard conditions for 24 h. Adhered cells were incubated with fresh medium containing 100 µg/mL CSCP at 37 °C in humidified atmosphere with 5% CO₂ for 24 h (pretreatment), and then the cells were treated with 50 µM tert-butyl hydroperoxide (t-BHP) for 30 min.

The assay used to measure ROS production at cellular level was based on the mechanism of hydrogen atom transfer using a cell-permeable fluorogenic probe 2’,7’-dichlorofluorescein diacetate (DCFH-DA). Upon reaction with free radicals, the formation of a fluorescent product was monitored. The cells treated as above were incubated with 10 µM DCFH-DA for 30 min and analyzed using a flow cytometer LSR II (Becton, Dickinson and Company, Franklin Lakes, NJ, USA). Acquired histograms were processed to calculate ROS production (%) in correlation with the fluorescence intensity using FlowJo v. 10.6.1 (34) and BD FACSDiva v. 6.1.3 (35) software. The cells incubated in normal medium and the cells pretreated with 12 µM ascorbic acid were processed under similar conditions and served as controls.

All experiments were carried out in triplicate. The results were expressed as mean value±standard deviation (S.D.) for three experiments. Statistical analysis of the data was performed on each pair of interest using two-tailed, paired Student’s t-test. Differences were considered statistically significant at p<0.05.

The first step was the extraction of undenatured collagen from fish skin using pepsin treatment under acidic conditions, as previously applied to mammalian tissues (21). After purification by salt precipitation, the pepsin-soluble collagen preparation with intact triple helix molecules was characterized by SDS-PAA electrophoresis under denaturing conditions (Fig. 2a). The pattern of carp skin collagen indicated the characteristic structure of type I collagen, consisting of three α chains, [(α1)₂α2], similar to that of the collagen extracted, in a similar way, from rat tail tendon. The observed doublet, corresponding to two α1 chains and one α2 chain, had calculated molecular mass of 130 and 116 kDa, respectively. Typical bands corresponding to β-dimers and γ-trimers were also observed in both collagen extracts. Previous studies reported similar electrophoresis pattern of pepsin-soluble collagen from grass carp (36) and marine giant croaker (37). The minor variation of molecular mass values for α-, β- and γ-bands was observed within a narrow domain and correlated with the living environment (warm/cold water, salinity) and extraction method (acid, enzymatic), which could influence their structural characteristics (37). The chemical and structural characteristics of collagen extracted from fish species show their better rheological properties than mammalian collagen, which recommends it for cosmetic applications (38).

The second step of peptide preparation consisted of carp skin collagen hydrolysis by proteinase K under specific and controlled conditions of pH, temperature and time of incubation in order to maximize the process. After a short incubation to denature its triple helical structure, the enzymatic hydrolysis of carp skin collagen followed the mechanism of action characteristic for an endopeptidase, cleaving the amino acid chain at a specific site containing peptide bonds adjacent to an aliphatic or aromatic amino acid. Thus, this enzymatic process of fish collagen hydrolysis was more controlled than a previously applied thermal treatment, which resulted in gelatin preparations with high MM (39).

The degree of carp skin collagen hydrolysis was 30.35% after proteinase K treatment for 6 h. Close values of 28.85 and 30.27% were obtained after 2 and 4 h of incubation, respectively, indicating that the enzymatic process reached a plateau. Previous studies have shown that the degree of hydrolysis varied between 11.50 and 33% for capelin protein hydrolysates and influenced their physicochemical properties and antioxidant activity (40).

After ultrafiltration, the obtained peptide preparation was preliminary characterized, showing an average of 87.5% protein, 8.6% hydroxyproline and 2.1% dry mass. The yield of peptide obtained from carp skin on dry mass basis of initial tissue using the two-step enzymatic process was 8%, a significantly increased value over previously reported 1.7−3.4% (16), indicating optimal hydrolysis parameters.

Fig. 2b presents the electrophoretic migration pattern of the enzymatic preparations of collagen, before and after ultrafiltration in tricine-SDS-gradient gel. Collagen preparation presented several bands distributed throughout the gel, indicating an efficient hydrolysis of collagen to fragments with M<40 kDa. Among them, a dense extended group of bands had calculated molecular mass corresponding to a range of 16−22 kDa and two narrow bands were observed at 4 and 1 kDa, respectively. Collagen peptide preparation obtained after ultrafiltration showed a more pronounced band at around 10 kDa, while faint blue coloured bands below 10 kDa indicated the presence of low molecular mass oligopeptides and peptides, considered bioactive compounds. Similar enzymatic hydrolysis method applied to grass carp skin using protease at 60 °C, pH=9, for 80 min resulted in the separation of a fraction containing peptides with M<1 kDa, which proved to be very useful for functional food development (41).

The enzymatic preparations of collagen and its peptides were also subjected to gel filtration chromatography in order to separate peptide fractions according to their molecular mass. Fig. 2c shows the chromatographic profile of samples obtained before and after ultrafiltration. The E-CSC preparation showed two distinct peaks, one corresponding to low elution volumes containing high molecular mass oligopeptides and proteins ranging between 28 and 112 kDa with a maximum intensity at 70 kDa. The second peak had higher intensity than the first peak, indicating high degree of CSC hydrolysis by proteinase K. CSCP preparation obtained after ultrafiltration presented a single peak, corresponding to peptides with a molecular mass ranging between 1.4 and 18 kDa with a maximum intensity at 4.5 kDa. This confirmed the separation of low molecular mass peptides with bioactive potential. Previous mathematical modelling showed that collagen was cleaved by proteinase K to small fragments of oligopeptides and bioactive peptides with low molecular mass, which could exert significant biomedical activity (18).

The collagen peptide preparation containing small and intermediate size peptides was further investigated for in vitro antioxidant activity and biological activity in cell culture experimental models.

The antioxidant capacity of low molecular mass peptide fraction was determined as scavenging potential against the free radicals ABTS, DPPH and ^•OH, in comparison to that of the undenatured collagen (CSC). The results are given in Table 1. The scavenging potential of CSCP against free ABTS (68.6%) and DPPH (14.0%) radicals was moderate, compared to Trolox, the synthetic analogue of vitamin E. However, the values were significantly higher (p<0.05) than those of CSC (7.0 and 0.5%, respectively). Strong antiradical activity of collagen peptides incubated in the presence of ^•OH radicals (71.0%) was recorded, representing highly encountered ROS within in vivo systems. Thus, our results showed that small peptides and oligopeptides were more efficient antioxidant agents than collagen, having more hydrogen atom- and unpaired electron-rich moieties to react with free stable radicals (42).

Similar studies analyzed the free radical scavenging potential of fish bioactive peptides according to the fish species, enzyme used and hydrolysis parameters, showing different values between 30 and 90%, but higher than those of undenatured fish collagen (43). The great variability was also due to the wide spectrum of tested concentrations and assays, from ABTS, DPPH, ^•OH and superoxide to lipid peroxidation or metal chelating capacity (44). Previous studies on papain hydrolysate of catfish skin collagen showed an increase of DPPH scavenging capacity in the first 2.5 h of hydrolysis, but lower value (63.06%) than of the hydrolysate of bone tissue (71.55%) (45). Fractions of marine fish oligopeptides with molecular mass between 5 and 10 kDa showed higher antioxidant activity against DPPH radicals (55%) than the fraction of low M<1 kDa peptides (37%) (46). The fractions with M>10 kDa had good antihypertensive activity due to the inhibition of ACE activity (7), while those found in grass carp skin hydrolysates with M<1 kDa (Gln-Pro, Trp-Pro-Pro) were good anti-ageing agents due to the ^•OH radical scavenging and regulating redox processes (47). Also, Alcalase-treated silver carp muscle proteins containing low molecular mass peptides showed strong metal chelating capacity and lipid peroxidation inhibition (15). However, a recent study reported that no significant differences in ABTS scavenging potential were observed between peptide fractions separated according to molecular mass from the Alcalase-treated fish skin gelatin (48). No studies on the antioxidant activity of proteinase K fish hydrolysate were found.

As previously reported in databases storing collagen peptide sequences, proteinase K favoured the release of fish peptides with C-terminal hydrophobic amino acids and increased the number of amino and carboxyl groups in the collagen peptide preparation, favouring the free radical scavenging (49). In contrast, the long polymeric chains of fish skin collagen have limited functional groups, show abundance of hydrophobic amino acids (Ala, Gly, Pro, Val, Leu, Ile, Met, Phe) and a smaller proportion of hydrophilic amino acids with charged side chains (Glu, Asp, Arg, Lys) and polar amino acids (Hyp, Thr, Ser) than hydrogen bonds (50).

The effect of collagen peptides on the processes of skin wound healing and redox balance under oxidative stress conditions was investigated using in vitro experimental models developed in fibroblast culture. In this regard, the optimal concentration was first selected for testing.

Cell viability of fibroblasts cultivated in the presence of different concentrations of collagen peptides was evaluated by MTT assay. The results are presented in Fig. 3a. The values of fibroblast cell viability varied between 112 and 119%, in the range of peptide concentrations of 10−200 µg/mL. The maximum value of 119.2% was recorded at a concentration of 100 µg/mL peptides and it was close to that of ascorbic acid (108.7%). Higher concentrations of peptides significantly (p<0.05) decreased the cell viability as low as 6.8%, similar to hydrogen peroxide (10.5%), used as positive control. These data indicated that peptides were cytocompatible with fibroblast culture in a wide range of concentrations and had a similar effect to that of ascorbic acid, a known antioxidant agent involved in the upregulation of fibroblast proliferation.

The results of in vitro cytocompatibility were confirmed by light microscopy observations of cell morphology in fibroblasts cultured in the presence of the same carp skin collagen peptide concentrations. The acquired micrographs showed that peptide treatment maintained the normal cell morphology at concentrations between 10 and 200 µg/mL (Fig. 3b). The fibroblasts treated with carp skin collagen peptides had characteristic fusiform phenotype, euchromatic nuclei and clear cytoplasm, similar to that of untreated cells (control). Similar or higher cell density than in control culture was observed and the cells were homogeneously distributed in the culture plate. Cell treatment at higher peptide concentrations revealed morphological changes, degeneration and cell lysis, similar to the cells treated with hydrogen peroxide, while the cell density decreased.

Quantitative and qualitative data indicated that carp skin collagen peptides were cytocompatible with cultured fibroblasts at a concentration of 100 µg/mL and stimulated cell metabolism, and were subsequently tested in experimental models in vitro.