Food Technology and Biotechnology

Collagen characterization

Bovine collagen fibre (d_particle=1.80–1.90 mm) and collagen powder (d_particle=0.45–0.6 mm) were supplied by Novaprom Food Ingredients Ltda (Guaiçara, Brazil). Hydroxyproline and total protein content in the collagen samples were determined and protein fractions were identified by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE), morphological structure using scanning electron microscope (SEM) technique, chemical composition with X-ray photoelectron spectroscopy (XPS), and crystallinity through X-ray diffraction (XRD).

Total protein and hydroxyproline

The protein and hydroxyproline contents in collagen samples were quantified according to AOAC method 981.10 (11) and AOAC method 991.20 (12), respectively.

Gel strength

Initially, fibre and powdered collagen gels were prepared at a 1:6 (m/V) ratio, heated and homogenized until reaching 72 °C. After that, the solution was cooled down to 8 °C for at least 12 h. Gel strength was then established using a TA.XT2 texture analyser (Stable Micro Systems Ltd, Godalming, Surrey, UK) with a 10-kg load cell, and pre-test, test, and post-test speeds of 1, 1 and 10 mm/s, respectively, as well as a 0.5-inch diameter spherical probe.

SDS-PAGE electrophoresis

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed according to the method proposed by Laemmli (13) and Bustamante-Vargas et al. (14). First, the powdered and fibre collagen samples were prepared at a concentration of 5.0 and 2.5 mg/mL, respectively. For sample preparation, 40 µL of 60% trichloroacetic acid (m/V) (Labsynth, Diadema, Brazil) were added to 100 µL of raw samples, placed in Eppendorf tubes and stored overnight in a freezer at –15 °C. The samples were subsequently centrifuged (model 5403; Eppendorf, Hamburg, Germany) at 10 000×g and 4 °C for 30 min, and then the supernatant was removed.

Scanning electron microscopy

Fibre and powder collagen morphology has been analysed by scanning electron microscopy (SEM microscope model JSM-6510; JEOL, Austin, TX, USA). The sample surfaces were coated with a gold layer (approx. 20 nm) using a sputter coater (model BAL-TEC SCD 050; BAL-TEC AG, Balzers, Liechtenstein).

X-ray photoelectron spectroscopy

Surface chemical composition analysis was done using X-ray photoelectron spectroscopy (XPS model Escalab 250Xi; Thermo Fisher Scientific, Lafayette, CO, USA) attached to a scanning electron microscope (SEM model JSM-6510; JEOL). In addition, XPS mapping approach was used to study chemical element dispersion on the electrode surface.

X-ray diffraction

Collagen physical structure (crystalline and/or amorphous) was characterised by X-ray diffraction (XRD) using a diffractometer (model XRD-6000; Shimadzu, San Diego, CA, USA) with CuKα radiation, from 10° to 80° (2θ) at 2°/min.

Hamburger formulation preparation

Elaboration and characterization of the hamburgers was carried out entirely in a large industrial unit (Santa Catarina, Brazil). Mixed meat (chicken and pork) hamburger formulations were prepared according to the parameters set by the agribusiness industry, following legal standards (1). The formulations were made by varying the mass fractions of fibre and powdered collagen (0.75–1.50%) and flocculated soy protein (Grupo Bremil, Arroio do Meio, Rio Grande do Sul, Brazil) (2.5–4.0%): P (standard formulation)=4.0% flocculated soy protein, T1 formulation=3.25% flocculated soy protein and 0.75% collagen fibre, T2 formulation=2.50% flocculated soy protein and 1.50% collagen fibre, T3 formulation=3.25% flocculated soy protein and 0.75% powdered collagen, and T4 formulation=2.50% flocculated soy protein and 1.50% powdered collagen.

Only chicken and pork meat (undeclared quantities; industrial formulation) were used, using cuts with no obvious fat and with minimum visible connective tissue (pork loin and chicken breast). The fat used was removed from the pork loins and the cuts of meat were kept frozen (maximum 0 °C).

First, the meat was cut using a mini cutter (Incomaf Indústria Ltda., São Paulo, Brazil) for 30 s, and then mixed with other ingredients. Then, the meat with soy protein and collagen was homogenized, hydrated with water for 15 min, and subsequently mixed using a blender (Risco, São Paulo, Brazil) for 3 min. A second grounding using a meat grinder (Seydelmann, Stuttgart, Germany) with a 5-mm disc was performed in order to standardize particle size.

After formulation preparation and grinding, the hamburger mix was submitted to a moulding stage. The process was done using a manual moulding equipment producing 90 g patties, after which the samples were frozen at –9 °C for a period of 90 days.

Hamburger characterization

Protein, moisture, fat, hydroxyproline, mass loss, instrumental texture (hardness), as well as histological and sensorial characteristics of the hamburger formulations were characterized on the first storage day.

Statistical analysis

The results (N=3) were analysed by analysis of variance (ANOVA), followed by the Tukey’s test to compare the average differences, using the Statistica v. 5.0 software (22), with a 95% confidence level. In addition, the Pearson correlation analysis and principal component analysis (PCA) were performed using XLSTAT software (23).

Hydroxyproline and protein contents, and gel strength of collagen preparations

Table 1 shows the characteristics of fibre and powdered collagen samples. It can be noted that there was a significant difference (p<0.05) in the hydroxyproline mass fraction between the samples, with powdered collagen having higher value (2.06 g/100 g) than collagen fibre. Collagen fibre had (p<0.05) a slightly higher protein content (97.81 g/100 g) than the powdered one (96.87 g/100 g).

Variations in protein and hydroxyproline contents are due to the raw material, collagen extraction process and origin. According to Gómez-Guillén et al. (24), high temperatures cause protein solubilization and greater fragmentation of the collagen structure.

Gauza-Włodarczyk et al. (25) state that whatever their origin, collagen contains 19 amino acids; two of them, hydroxyproline and hydroxylysine, are practically absent from other proteins. The same authors verified that the hydroxyproline content in bovine Achilles tendon collagen (8.15 g per 100 g of protein) is 30% higher than the hydroxyproline content of collagen from fish skin and concluded that fish skin collagen is less stable than the bovine Achilles tendon.

The gel strength of collagen fibre was significantly different (p<0.05) from the powdered one with a higher value (146 315 Pa), demonstrating that fibre allows greater solvent entrapment than powdered collagen gels. According to Prestes (26), gel resistance (Bloom force) depends on concentration and molar mass, where a higher Bloom value is correlated to collagen molar mass, hence a high Bloom value (about 300 g) results in firmer gels. The obtained data show that powdered and fibre collagen are different products with distinctive characteristics, possibly due to the extraction process used.

SDS-PAGE of collagen preparations

Fig. 1 shows SDS-PAGE for the tested collagen samples. On the electrophoresis gel, four distinct bands were identified for powdered collagen with molecular masses varying from 50 to 100 kDa, whereas a 100-kDa band for collagen fibre was found.

Fig. 1

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) for tested collagen. Legend: from left to right – column 1=molecular mass standards, column 2=collagen powder, column 3=collagen fibre

According to Oechsle et al. (27), SDS-PAGE gel with bovine telopeptide-poor collagen showed two distinct bands of approx. 123 kDa, indicating α1(I) and α2(I) chain monomers of type I collagen. Furthermore, the slightly larger band of α1(III) chain was observed above, indicating a type III collagen. The mass spectrometry analysis noted the presence of α1(I), α2(I) and α1(III) chains with 133, 129 and 138 kDa, respectively.

Scanning electron microscopy of collagen preparations

The SEM micrographs obtained for collagen fibre and powdered collagen are shown in Fig. 2. These results show that both materials had quite a different microstructure, with the powdered collagen (B1 and B2 in Fig. 2) having lost its fibrous characteristic during the milling of the collagen fibre (A1 and A2 in Fig. 2). Collagen fibre showed an internal axis with several thin ramifications connecting them to smaller external particles, whereas powdered collagen showed a major trend to form agglomerates. The thin filaments observed in the collagen fibre may be important in the eventual interaction between the polymeric matrix and fillers.

Fig. 2

Micrographic images of collagen fibre (A1 and A2) and powder (B1 and B2). Magnification: 1=100× and 2=1500×

The greatest difference between fibre and powdered collagen is that the fibre physical structure retains water chemically, either through protein matrix or hydrogen bonding with water (26). As such, fibre swells in contact with water, blocking both moisture and fat from exiting the system.

X-ray photoelectron spectroscopy of collagen preparations

The main elements found in both powdered and collagen fibre were carbon, nitrogen and oxygen (Fig. S1 and Fig. S2). Elemental mapping showed that such elements were homogeneously distributed on both powdered and collagen fibre surfaces. The presence of other elements, such as aluminium, magnesium, sodium and fluorine, was also noted in both collagen samples, while iron was found only in the powdered one. These differentiations could have been due to collagen production methods.

X-ray diffraction patterns of collagen preparations

Fig. 3 shows great XRD pattern similarities of both fibre and powdered collagen samples. Collagen fibre preponderant peaks were obtained with 2θ at approx. 7°, 25° and 30°, and powdered collagen preponderant peaks were obtained with 2θ approx. 25° and 30°, in line with reported reticulated collagen values (28), typical for amorphous material. The results indicated that powdered collagen is more amorphous than the fibre.

Fig. 3

X-ray diffraction (XRD) patterns for collagen powder (grey line) and fibre (black line)

Physicochemical and sensorial characteristics of hamburger formulations

Average mass fractions and the respective standard deviations of protein, fat, moisture and hydroxyproline of different hamburger formulations on the first day of storage are shown in Table 2. The mass fraction of protein varied from 16.73 (T3) to 17.8% (T4). Formulations T1 and T3, with 0.75% fibre and powdered collagen respectively were not significantly different (p<0.05) from the standard one. Formulations T2 and T4, with 1.5% collagen fibre and powdered collagen respectively had higher protein content (p<0.05). The results for protein content suggest that the addition of 1.5% collagen (fibre or powdered) increased protein content. Such results are in accordance with Prestes et al. (29), who also reported total protein content increase in products containing bovine collagen.

Regarding lipid content, a variation in the formulations was noted, with values ranging from 8.22% (standard) to 10.2% (T3), given that T3 had statistically higher (p<0.05) lipid content than the other formulations. Variations in lipid content were possibly due to the changes in the raw materials, i.e. the fraction of fat removed from the meat.

Formulations T2 and T4, containing 1.5% fibre and powdered collagen respectively, were significantly different (p<0.05) from the standard and formulations T1 and T3 (Table 2) and had slightly higher moisture values. Positive correlation was confirmed by the principal component analysis (Fig. 4). Such results are explained by the high water retention capability of the collagen, which reduces water loss during freezing. According to Pietrasik and Janz (30), non-meat proteins exhibit similar behaviour to meat proteins, promoting water retention, higher binding, and occupying the interstitial spaces in the gel matrix.

Fig. 4

Principal component analysis (PCA) of the hamburger formulations on the first day of storage. Standard=4% soy protein, T1=0.75% collagen fibre and 3.25% soy protein, T2=1.5% collagen fibre and 2.5% soy protein, T3=0.75% collagen powder and 3.25% soy protein, and T4=1.5% collagen powder and 2.5% soy protein

Hydroxyproline content served as a parameter to establish collagen amount in meat and meat products (3, 25). Significant differences (p<0.05) in hydroxyproline mass fraction were observed among the developed formulations (Table 2). All formulations with added collagen showed higher hydroxyproline mass fractions than the standard sample. However, formulations T1 and T3, with 0.75% collagen fibre and powdered collagen respectively, were not significantly different (p<0.05) from each other.

It was also noted that formulation T4, containing 1.5% powdered collagen, had the highest hydroxyproline content (0.77 g/100 g), 4.8 times higher than the standard formulation. The results found are in accordance with the ones reported by Prestes et al. (29), where formulations of chicken ham containing a mixture of collagen had higher values of hydroxyproline. Formulation T2 (1.5% collagen fibre) showed increased (p<0.05) hardness (85.7 N) (Table 2 and Fig. 5). When only collagen fibre was added, it resulted in a higher compressive strength and higher shear force in the samples due to the physical structure and larger particle size, which could be related to how high collagen contents in emulsions increase hardness and rigidity, while reducing the mass stability (31). In addition, by retaining water chemically through the protein matrix and swelling when in contact with water, collagen fibre alters the texture and cohesion of the hamburger mix, increasing the final product firmness (32). This behaviour was also observed by Li (33) when adding collagen to the preparation of cooked ham. In that case, collagen caused an increase in hardness from 11.96 to 16.91 N, suggesting that small size proteins affected the texture of the ham.

Fig. 5

Photomicrographs of hamburger formulations with a detail to the right. Legend: A and B=standard, C and D=formulation T1, E and F=formulation T2, G and H=formulation T3, I and J=formulation T4. Muscle tissue (arrows), adipose tissue (ta), conjunctive tissue (tc), tissue disorganization (*), and empty spaces (ev). Magnification: 10× (A, C, E, G and I) and 25× (B, D, F, H and J)

Mass loss of the hamburger samples prepared in the conventional oven (Table 2) on the first day of storage was on the whole higher than of the ones prepared on the grill (except T3). It must be pointed out that mass loss rate of T3 formulation was lower, which may be better visualized in the multivariate analysis in Fig. 4. Lower mass loss after freeze-thaw and reheating process was found of samples with powdered collagen (Table 2), a phenomenon explained by its greater interaction with the ingredients and additives present in the formulations, creating a cohesive mass. The addition of collagen to meat products as a binder is advantageous; at low levels functional collagen proteins promote an increase in cooking yield due to their gelling and water-binding properties (34). According to Pietrasik (35), the higher the percentage of added collagen, the lower the release of water due to a greater number of bonds between the polypeptide chains during cooking (formation of a dense protein matrix).

Table 3 shows the results of sensorial evaluation of the hamburger formulations. A significant difference (p<0.05) could be noticed among the formulations. On the whole, formulation T3, made with 0.75% powdered collagen, was the one that received the highest scores (in all attributes) compared to the other formulations. It also had the highest general acceptability of 81% (Table 3).

In general, the panellists positively accepted the replacement of soy protein with collagen in hamburgers. It is assumed that such substitutes enhance the acceptability of the hamburgers, as well as help to improve their physical properties, especially in the case of formulation T3. In contrast, the standard formulation obtained the lowest scores. In terms of flavour, it was noted that all formulations containing collagen received higher scores and differed statistically (p<0.05) from the standard sample. These results agree with Sousa et al. (32), who verified higher texture scores of frankfurter-type sausages containing different collagen mass fractions (25 to 75%) and attributed this effect to the gelatinization property of collagen.

Table S1 and Fig. 4 show, respectively, the Pearson correlation and principal component analysis (PCA) for the physi- cochemical and sensorial variables of the hamburger samples on the first day of storage. The variables are shown as vectors (Fig. 4); the longer the vector, the greater the sample variability. The samples were represented by triangles, where each vertex represented a repetition (N=3). It was observed that there was good discrimination among the formulations. The first (PC1) and second (PC2) principal components corresponded to 65.31% of total variance. PC1 contributed with 43.91%, and PC2 contributed with 21.40%.

The values obtained with the Pearson correlation validated the correlation among the parameters observed in the PCA (Fig. 4), with protein presenting a positive correlation (Table S1) with hydroxyproline content (0.666), mass loss during preparation on the grill (0.521) and moisture (0.556). Formulation T4 was the closest to these vectors, also confirmed by the values shown in Table 2, i.e. formulation T4 had the highest protein (17.8 g/100 g) and hydroxyproline (0.77 g/100 g) contents. Formulation T3 (0.75% powdered collagen and 3.25% soy protein) received the best sensorial scores (Table 3 and Fig. 4). Positive correlations (Table S1 and Fig. 4) were also verified between flavour and fat content, and between texture (>0.70) and fat, appearance, colour, odour and flavour. However, with increased hardness (instrumental texture), there was a decrease in oven mass loss.

Due to its properties, such as extender, emulsifier, texture improver, and its nutritional value, collagen has great application potential in the industry of restructured and emulsified meat products, providing better technological performance and economic results. Collagen beneficially participates in meat emulsions in the range from 15 to 18%, mainly aiding the texture and stability of the mass (36), reducing water loss in defrosting and cooking.

Histological characteristics of hamburger formulations

Fig. 5 shows the photomicrographs of hamburger formulations. The use of histological methods allowed for the qualitative analysis of muscular, adipose and conjunctive tissues. The standard sample (A and B in Fig. 5) shows muscular tissue (arrows) associated with dense conjunctive tissue (tc). Empty spaces (ev) can be noticed between the muscular tissues and/or spaces with adipose tissue. It achieved a cellular organization classified as “good”, but without consistency. It was the sample with the worst texture according to sensorial evaluation (Table 3 and Fig. 4). Formulation T1 (C and D in Fig. 5) shows tissue disorganization (*), i.e. empty spaces (ev) dispersed among muscular cells (arrows), possibly adipose tissue, as well as dense conjunctive tissue (tc). The bad cellular organization and added collagen fibre caused an increase in the hamburger hardness (Table 2). Formulation T2 (E and F in Fig. 5) has a muscular cell organization with peripheral nuclei (arrows) and some adipose cells (ta). It shows a non-emulsified conjunctive tissue with spaces between the conjunctive tissues, with cellular organization. The non-emulsion of the connective tissue caused the highest values of hardness due to the higher mass fraction of collagen fibre (1.5%). Formulation T3 (G and H in Fig. 5) showed intense tissue disorganization (*) since there was not any difference among muscle (arrows), conjunctive and adipose tissues, technologically qualified as emulsified material, providing to the better hamburger texture (Table 2) and sensory (Table 3) characteristics. Formulation T4 (I and J in Fig. 5) showed intense tissue cohesion, but due to little difference among adipose, muscle and conjunctive (*) tissues, it also had greater mass loss during heat treatment (Table 2 and Fig. 4).