Extraction, characterization and antioxidant activity evaluation of collagen from Monopterus albus bones

Abstract

BACKGROUND

Monopterus albus bones are a rich source of collagen, yet they are often discarded as processing byproducts, leading to resource waste. This study aimed to establish an efficient ultrasound‐assisted enzymatic extraction method for collagen from M. albus bones, characterize its structural properties, and evaluate its in vitro antioxidant activity.

RESULTS

The optimal extraction conditions were determined via single‐factor experiments and Box–Behnken response surface methodology as follows: solid–liquid ratio of 1:20 g g⁻¹, pepsin addition of 20 g kg⁻¹, enzymatic hydrolysis time of 2 h, ultrasonic time of 22 min, ultrasonic power of 375 W, and ultrasonic frequency of 40 kHz. Under these conditions, the collagen extraction yield reached 35.00 ± 0.48%. The extracted product was identified as type I collagen by ultraviolet spectroscopy and sodium dodecylsulfate–polyacrylamide gel electrophoresis. Scanning electron microscopy revealed a loose, porous fibrous network structure, while X‐ray diffraction indicated a combination of disordered flexibility and ordered active conformation. Differential scanning calorimetry and thermogravimetric analysis showed a thermal denaturation temperature of 92.70 ± 1.24 °C and good stability below 200 °C. In vitro antioxidant assays demonstrated that the collagen exhibited strong free radical scavenging activity, with IC₅₀ values of 6.01 ± 0.36 mg mL⁻¹ against DPPH radicals, 7.63 ± 0.24 mg mL⁻¹ against superoxide anion radicals, and 17.71 ± 0.19 mg mL⁻¹ against hydroxyl radicals.

CONCLUSION

This study provides a practical and sustainable ultrasound‐assisted enzymatic extraction method for obtaining high‐quality type I collagen from M. albus bones, which demonstrates excellent thermal stability and potent antioxidant activity. These findings support the high‐value utilization of M. albus processing byproducts and lay a foundation for the industrial development of natural antioxidant collagen. © 2026 Society of Chemical Industry.

INTRODUCTION

The Asian swamp eel (Monopterus albus) is classified under the genus Monopterus of the family Synbranchidae, within the order Synbranchiformes, and belongs to the class Actinopterygii. ¹ Renowned as a key freshwater aquaculture species in China, ² the M. albus aquaculture industry has a substantial scale, with its annual yield consistently exceeding 300 000 tons according to data from the China Fishery Statistical Yearbook. ³ It is valued for its succulent flesh, delicate taste, and rich protein content. ⁴ Its processed products include fresh eel slices, shredded eel, and various prepared foods, making it a highly valued aquatic product in the market. ⁵ However, with the rapid expansion of the breeding and processing scale of M. albus, a critical issue has become increasingly prominent: processing byproducts (including fish heads, fish bones, and other components), which account for 30–40% of the total weight of M. albus, have not been fully utilized. This not only results in resource waste, ⁶ but also poses a potential risk of environmental pollution. Currently, aquatic processing byproducts have garnered widespread attention as a novel source of collagen, ⁷ with advantages including mild extraction conditions, high collagen content, good biocompatibility, no risk of animal diseases, and environmental friendliness.8, 9, 10 However, studies on collagen derived from M. albus bones are scarce, and existing reports lack comprehensive and systematic characterization of extraction processes, structural properties, and functional activities. Therefore, systematically analyzing the nutritional composition and physicochemical properties of M. albus bones is of significant importance for promoting the high‐value utilization of its processing byproducts and reducing resource waste.

Collagen, as the primary structural protein in the connective tissue of terrestrial and marine animals, possesses a unique triple‐helix structure that is crucial for maintaining tissue homeostasis and biological integrity. This feature also endows it with wide application potential in areas including food, medicine, and cosmetics. ¹¹ Currently, traditional collagen extraction methods primarily include acid extraction, enzymatic extraction, alkaline extraction, salt extraction, and hot water extraction.12, 13 Among these, enzymatic extraction leverages the specific disruption of intermolecular hydrogen bonds in collagen by biological enzymes, enabling efficient extraction while preserving collagen structural integrity. ¹⁴ With its mild reaction conditions and environmental friendliness, this technique has now become the mainstream method for collagen extraction.15, 16 However, traditional enzymatic hydrolysis processes still have significant shortcomings: low substrate conversion rate, insufficient enzyme utilization, and limited hydrolysis efficiency. ¹⁷ To address this technical bottleneck, emerging pretreatment methods including ultrasonic, microwave, and high‐voltage pulsed electric field technologies have been progressively adopted. These technologies can effectively enhance hydrolysis efficiency and improve product functional properties by altering protein conformation and disrupting the microscopic structure of raw materials.18, 19 Among various emerging technologies, ultrasonic technology stands out for its advantages in collagen extraction. It induces cavitation effects via 20 kHz sound waves, generating high shear forces that disrupt collagen structure, expose enzyme cleavage sites, and increase enzyme–substrate contact, thereby accelerating hydrolysis.20, 21, 22 Ultrasound‐assisted methods have achieved higher collagen yields from sources such as fish scales, ²³ fish skin, ²⁴ and porcine bones, ²⁵ while preserving triple‐helix integrity and functional properties. As a green technology, ultrasonic pretreatment offers low energy consumption and no chemical residues, demonstrating strong industrial potential. ²⁶

In summary, the high‐value utilization of M. albus bones as processing byproducts has become an important research direction in the food industry. Although the antioxidant capacity of collagen provides a foundation for developing novel functional foods, research on the structural characteristics and biological activities of collagen derived from eel bone remains relatively limited. To address this, the study reported here employed ultrasound‐assisted enzymatic hydrolysis to extract collagen from eel bone and systematically characterized its structural properties and in vitro antioxidant activity, aiming to promote the high‐value utilization of eel bone byproducts and reduce resource waste. Additionally, it offers a green and sustainable processing method for the extraction of natural antioxidant collagen, laying a theoretical and technical foundation for the industrial application of eel bone collagen and the high‐value upgrading of the eel processing industry.

MATERIALS AND METHODS

Materials

The bones of M. albus were purchased from Hubei Yuntaifang Food Co. Ltd, and divided into portions. The samples were subsequently kept at −20 °C for storage.

The experimental procedures utilized analytical‐grade reagents throughout. Sodium hydroxide, isopropanol, EDTA–2Na (ethylenediaminetetraacetic acid disodium salt), citric acid, hydrochloric acid, glacial acetic acid, methanol, perchloric acid, and sodium chloride were sourced from Sinopharm Chemical Reagent Co. Ltd. Pepsin (3000 U mg⁻¹) was acquired from Shanghai Macklin Biochemical Co. Ltd.

Pretreatment of M. albus bones

The pretreatment of M. albus bones was performed with reference to the modified method proposed by Yu et al. ²⁷ The procedure was as follows: frozen M. albus bone samples were allowed to thaw under running tap water and rinsed thoroughly with distilled water until no residual blood remained. Then, the bone was cut into sections approximately 4 cm in length and mixed with 50 g kg⁻¹ NaOH solution at a solid‐to‐liquid ratio of 1:20 (g/g), and the resultant mixture was incubated at 4 °C for 3 h. The NaOH solution was renewed at hourly intervals to remove non‐collagenous proteins. Subsequently, the treated bone was rinsed thoroughly with distilled water until a neutral pH was achieved. Demineralization was then carried out using a 0.5 mol L⁻¹ EDTA‐2Na solution (pH 8.0) at a solid‐to‐liquid ratio of 1:20. This process was conducted at 4 °C over 72 h, with the solution being renewed at 12 h intervals. Finally, the sample was rinsed thoroughly with distilled water. Defatting was performed by first immersing the material in a 100 g kg⁻¹ isopropanol solution (1:10 solid‐to‐liquid ratio) and then storing it at 4 °C for 48 h. Throughout this period, the solution was replaced every 12 h, followed by rinsing with distilled water. The treated M. albus bone was subjected to freeze‐drying for 48 h, then pulverized and sieved through a 40‐mesh sieve, and was finally preserved in airtight containers at −20 °C for subsequent experiments.

Composition determination of M. albus bone powder

Ash content was determined by the high‐temperature ashing method in accordance with the national standard GB 5009.4‐2016. Moisture content was measured via the direct drying method following the national standard GB 5009.3‐2016. Protein content was quantified using the Kjeldahl nitrogen determination method with reference to the national standard GB 5009.5‐2016. Fat content was analyzed by the Soxhlet extraction method pursuant to the national standard GB 5009.6‐2016.

Preparation of collagen

Extraction technology of collagen

By adapting the method reported by Shaik et al. ²⁸ with minor adjustments, collagen was isolated from M. albus bones through the following procedure. Bone powder was mixed thoroughly with distilled water at a solid‐to‐liquid ratio of 1:20 (g/g). After adjusting the pH to 2.0, 20 g kg⁻¹ pepsin was added. This mixture was initially sonicated for 20 min at 37 °C, 250 W, and 20 kHz, and then transferred to a 37 °C constant‐temperature shaker for further hydrolysis for 100 min. After hydrolysis, the reaction mixture was incubated in an 85 °C water bath for 10 min to deactivate. Following cooling, it was centrifuged at 10 000 × g for 15 min, and the supernatant was collected for further use. NaCl was incorporated into the supernatant to a final concentration of 0.9 mol L⁻¹, followed by incubation at 4 °C overnight. This mixture was subsequently centrifuged at 6000 × g for 30 min at 4 °C, producing a precipitate that was harvested as the crude collagen extract. Following redissolution in 0.5 mol L⁻¹ acetic acid, the crude collagen was dialyzed sequentially against 0.1 mol L⁻¹ acetic acid and distilled water at 4 °C. The dialysis buffer was refreshed every 6 h, with the entire dialysis process lasting for 3 days. Following dialysis, the dialyzed sample underwent freeze‐drying to yield the purified collagen product, which was subsequently preserved at −80 °C for subsequent experiments.

Determination of collagen content

Hydroxyproline is a characteristic amino acid of collagen, and the hydroxyproline content of different bone collagen exhibits a stable correlation with collagen content. With reference to NY/T 3608‐2020, the hydroxyproline content M. albus bone meal was determined via spectrophotometry. An aliquot (1 mL) of the enzymatically hydrolyzed M. albus bone meal solution was transferred to a hydrolysis tube, mixed with 20 mL of 6 mol L⁻¹ HCl, and hydrolyzed at 105 °C for 20 h. Upon completion, the hydrolysate was transferred to a 100 mL volumetric flask, diluted to volume with distilled water, and filtered through a 0.45 μm membrane. A 4 mL aliquot of the filtrate was then mixed with 2 mL of chloramine T reagent and left at room temperature for 20 min. Subsequently, 2 mL of color development reagent was added, and the mixture was incubated in a water bath at 60 °C for 20 min. After cooling for 30 min, the absorbance was measured at 558 nm. A standard curve was prepared using serial dilutions of hydroxyproline standard solutions (0.5–5.0 μg mL⁻¹), yielding an R ² value of 0.999, and was used to calculate the hydroxyproline content in the samples. Collagen content was then calculated using Eqn (1):

$$\text{Collagen content}\left(\mathrm{mg}{\mathrm{g}}^{-1}\right)=\begin{array}{l}\text{Hydroxyproline content}\left(\mathrm{mg}{\mathrm{g}}^{-1}\right)\\ \times 11.1\end{array}$$ (1)

where 11.1 represents the conversion factor for aquatic animal collagen. ²⁹

Optimization of extraction technology for collagen

Single‐factor experiments

In the initial stage, a single‐factor experimental design was adopted. The variables and their tested levels were as follows: ultrasonic time (0, 20, 40, 80, 100 min), ultrasonic power (0, 125, 250, 375, 500 W), and ultrasonic frequency (20, 40, 60, 20/40, 20/60, 40/60, and 20/40/60 kHz, with the slash representing combined frequencies). The effect of each factor on the extraction yield was evaluated to establish its optimal operational range.

Response surface methodology experimental design

Preliminary single‐factor trials guided the development of a three‐factor, three‐level response surface experiment rooted in Box–Behnken design principles, with the experimental framework constructed via Design‐Expert 13 software. Ultrasonic time (A), power (B), and frequency (C) were chosen as the independent variables, with the extraction yield of M. albus bone collagen (Y) serving as the response indicator.

Characterization of physicochemical properties of collagen

UV absorption spectroscopy

The procedure followed the protocol outlined by Zhang et al., ³⁰ with minor adjustments. A suitable amount of the lyophilized M. albus bone collagen sample was dissolved into a 0.5 mol L⁻¹ acetic acid solution to formulate a 0.1 mg mL⁻¹ sample solution. With 0.5 mol L⁻¹ acetic acid solution designated as the blank control, scanning was conducted across the 200–400 nm wavelength range using a UV–visible spectrophotometer. The UV absorption spectrum was recorded, and the characteristic absorption peaks were analyzed.

Sodium dodecylsulfate–polyacrylamide gel electrophoresis (SDS‐PAGE) analysis

Referring to the method of Ahmad et al. with slight modifications, ³¹ a lyophilized sample was redissolved in 50 g kg⁻¹ SDS solution to yield a 10 mg mL⁻¹ working solution, followed by 1 h water bath incubation at 85 °C. This mixture was then centrifuged at 8000 × g for 10 min. The resulting supernatant was combined with 5× loading buffer at a 1:4 (v/v) volume ratio, prior to a 15 min water bath incubation at 85 °C. The sample and 10 μL of protein marker were loaded onto an 80 g kg⁻¹ gradient polyacrylamide gel, with electrophoresis carried out at a constant voltage of 120 V lasting 60 min. The gel was stained with 1 g kg⁻¹ Coomassie Brilliant Blue R‐250 for 1 h. Following this, it was destained using a solution of 300 g kg⁻¹ methanol and 100 g kg⁻¹ acetic acid until the protein bands became clearly visible. Image acquisition was performed using a gel imaging analysis system.

Scanning electron microscopy (SEM)

Lyophilized M. albus bone collagen samples were cut into thin slices, fixed on an SEM sample stage with conductive tape, purged with high‐purity nitrogen gas, and sputter‐coated with gold. SEM was used to observe the microstructural characteristics at magnifications of ×300, ×500, and ×1000.

X‐ray diffraction (XRD)

With minor modifications to the protocol described by Malison et al., ³² a freeze‐dried M. albus bone collagen sample was placed uniformly on the instrument's sample stage. The apparatus was configured with the following operating parameters: copper target X‐ray source (λ = 0.154 nm), with a 2θ scan range from 5° to 90° and a scanning rate of 5° min⁻¹. The XRD pattern of the M. albus bone collagen sample was thus acquired.

Amino acid profile analysis

The amino acid composition and content of M. albus bone collagen were determined using an automatic amino acid analyzer according to the method described by Shen et al. ³³ A freeze‐dried sample was dissolved in 10 mL of 6 mol L⁻¹ hydrochloric acid solution. The hydrolysis tube was then frozen, evacuated, and purged with nitrogen gas before undergoing hydrolysis at 110 °C for 22 h, followed by cooling. The hydrolyzate was diluted to a 50 mL volume. A 1.0 mL aliquot of the filtrate was dried under reduced pressure, then dissolved in a pH 2.2 sodium citrate buffer. The mixture was passed through a 0.22 μm filter membrane to prepare the test solution for amino acid composition analysis.

Rheological experiments

The rheological behaviors of M. albus bone collagen were characterized via a rotational rheometer. A collagen solution (10 mg mL⁻¹) was prepared by dissolving the sample in 0.1 mol L⁻¹ acetic acid. Approximately 1 mL of this solution was subjected to rheological measurement using a 15 mm parallel‐plate geometry at a constant temperature of 25 °C. During frequency sweep tests at a constant strain of 0.01, both the storage modulus (G′) and loss modulus (G″) were recorded across the applied frequency range, with values expressed in pascals (Pa).

Thermogravimetric analysis (TGA)

The method of Lin et al. was followed with minor adjustments. ³⁴ A measured quantity of the sample was transferred to an alumina crucible for subsequent analysis. Thermal analysis was conducted with a simultaneous thermal analyzer, employing a nitrogen purge (10 mL min⁻¹), a heating rate of 10 °C min⁻¹, and a temperature range of 30–1000 °C. A blank crucible served as the reference, and the variations in the sample's thermodynamic properties throughout the heating process were documented.

Differential scanning calorimetry (DSC)

An assessment of the thermal stability of M. albus bone collagen was performed using DSC, with modifications based on the method of Frosi et al. ³⁵ For DSC analysis, a precisely weighed lyophilized collagen sample was loaded into an aluminium pan and subsequently hermetically sealed. A blank aluminium pan was used as the reference. Under a continuous nitrogen atmosphere, the sample was subjected to heating from 30 to 140 °C at 10 °C min⁻¹. The peak temperature observed in the DSC thermogram was recorded as the thermal denaturation temperature of M. albus bone collagen.

Antioxidant activity of M. albus bone collagen in vitro

Determination of DPPH radical scavenging activity

The DPPH radical scavenging assay was conducted following the protocol described by Zhang et al. ³⁶ Briefly, 1 mL aliquots of sample solutions at distinct concentration levels were mixed with 1 mL of 0.2 mmol L⁻¹ DPPH fully mixed. The mixture was incubated in a dark environment at 26 °C for 30 min, after which its absorbance was determined at a wavelength of 517 nm. Each sample was tested in triplicate. Distilled water functioned as the blank, while an ascorbic acid solution was used as the positive control.

Assessment of superoxide anion radical scavenging capacity

Superoxide anion scavenging activity was assessed following the protocol of Zeng et al. ³⁷ For the assay, the reaction system consisted of 1 mL of sample (at varying concentrations), 4.5 mL of 50 mmol L⁻¹ Tris–HCl buffer (pH 8.2), and 0.1 mL of 3 mmol L⁻¹ pyrogallol solution. The mixture was shaken well and incubated in a 25 °C water bath for 10 min. To terminate the reaction, 1 mL of 8 mmol L⁻¹ HCl solution was added, after which the absorbance was recorded at 320 nm. All samples were analyzed in triplicate. Distilled water acted as the blank control, while ascorbic acid solution served as the positive control.

Determination of hydroxyl radical scavenging activity

Hydroxyl radical scavenging capacity was quantified using the protocol reported by Dong et al. ³⁸ with slight modifications. In brief, to 2 mL aliquots of sample solutions with varying concentration levels, 2 mL of 9 mmol L⁻¹ ferrous sulfate, 2 mL of a 9 mmol L⁻¹ salicylic acid–ethanol solution (salicylic acid dissolved in ethanol), and 2 mL of 9 mmol L⁻¹ hydrogen peroxide (H₂O₂) solution were added in sequence. The mixture was thoroughly vortexed and then incubated in a 37 °C water bath for 30 min. Following incubation, it was cooled to room temperature and finally centrifuged at 5000 × g for 10 min. The absorbance of the resulting supernatant was determined at a wavelength of 510 nm. All determinations were carried out in three independent replicates. Distilled water served as the blank control, while an ascorbic acid solution was employed as the positive control.

Statistical analysis

All experiments were conducted in triplicate, and results are expressed as mean ± standard deviation. Experimental data were analyzed by one‐way analysis of variance (ANOVA) via IBM SPSS Statistics 27, where the statistical significance threshold was set at P < 0.05. Figures were plotted via Origin 2021 software. Meanwhile, response surface optimization was conducted utilizing a Box–Behnken design established with Design‐Expert 13, and the experimental data were subjected to analysis via a quadratic polynomial regression model to evaluate the model's statistical significance.

RESULTS AND DISCUSSION

Nutritional composition of M. albus bone meal

The results of nutritional component analysis of M. albus bone meal (Table 1) showed that the contents of fat (81.9 ± 4.9 g kg⁻¹), moisture (66.2 ± 0.4 g kg⁻¹), and ash (4.1 ± 0.3 g kg⁻¹) in the sample were relatively low, while the crude protein content reached 828.0 g kg⁻¹, accounting for the highest proportion. Since most proteins in bones are collagen, it is inferred that collagen can be extracted from M. albus bones.

Table 1.

Nutrients of Monopterus albus bone powder

M. albus bone powder	Ash content (g kg−1)	Moisture content (g kg−1)	Crude protein content (g kg−1)	Fat content (g kg−1)
Measured value	4.1 ± 0.3	66.2 ± 0.4	828.0 ± 23.6	81.9 ± 4.9

Single‐factor experiment

Effect of different ultrasonic times on extraction yield of collagen

Figure 1(A) shows that the extraction yield was the highest in the group with 20 min of ultrasonication, and when the ultrasonication time exceeded 20 min, the extraction rate decreased significantly. This indicates that although the ultrasonic cavitation effect can promote protein release, excessively long treatment time may induce collagen decomposition or denaturation, thereby reducing the extraction efficiency. ³⁹ Therefore, 20 min was determined as the optimal ultrasonication time parameter in this study.

Figure 1.

Impacts of various ultrasonic processing parameters on collagen extraction yield from Monopterus albus bones. (A) Ultrasound treatment duration. (B) Ultrasound power intensity. (C) Ultrasound frequency.

Effect of different ultrasonic powers on extraction yield of collagen

As shown in Fig. 1(B), the control group yielded a significantly higher extraction yield compared to treatments using ultrasonic power below 250 W. This is because the mechanical forces generated by cavitation bubble collapse at low ultrasonic power levels are relatively weak, which insufficiently disrupts the three‐dimensional structure of the sample. Additionally, it reduces the penetration efficiency of the enzymatic solution into the substrate, leading to fewer effective contacts between the enzyme and collagen, thereby resulting in a decreased extraction rate. Studies have shown that under low‐power conditions, the number of cavitation bubbles induced by ultrasonication is small, and ultrasonication has limited ability to disrupt the sample's cell walls and release polysaccharides. ⁴⁰ With an increase in ultrasonication power, the extraction rate increased significantly, reaching a peak value of 27.56 ± 3.21% at 375 W. Once the power surpassed 375 W, the collagen extraction yield exhibited a significant reduction (P < 0.05). This indicates that the ultrasonic cavitation effect can effectively disrupt the cell membrane structure, promote substance release, and enhance enzyme–substrate binding, thereby improving the extraction efficiency. Görgüç et al. have demonstrated that ultrasonication power significantly affects the protein extraction rate. ⁴¹ However, excessively high power may cause protein hydrolysis, denaturation, and aggregation, or damage the structure of proteases, thereby reducing the extraction rate. Xu et al. have confirmed that excessively high power can impair the structural and functional properties of proteins. ⁴² Therefore, 375 W was determined as the optimal ultrasonication power parameter in this experiment.

Effect of different ultrasonic frequencies on extraction yield of collagen

Figure 1(C) shows that during single‐frequency processing, the extraction rate initially increased and then decreased with increasing frequency, with the 40 kHz single‐frequency ultrasound treatment group having the highest extraction yield (32.52 ± 0.99%). Ultrasound frequency is closely related to cavitation effects. Low‐frequency ultrasound has a higher energy density, which more easily triggers intense cavitation effects and mechanical shear forces, thereby enhancing the solubilization of biomacromolecules. Consequently, the cavitation phenomenon and mechanical forces induced by low‐frequency ultrasound (20 and 40 kHz) exerted a potent disruptive effect on the sample's tissue structure while facilitating collagen solubilization. ⁴³ The collagen extraction yield in both dual‐ and triple‐frequency ultrasound groups was found to be notably lower than that of the single‐frequency ultrasound group (P < 0.05). It is speculated that the high‐intensity noise signals and cavitation bubble pulsations caused by multi‐frequency ultrasound damaged the structure of the protease, ⁴⁴ thereby reducing the extraction efficiency. Likewise, Guo et al. ⁴⁵ demonstrated that dual‐ and triple‐frequency ultrasound modes exert a more pronounced inactivation effect on horseradish peroxidase relative to the single‐frequency counterpart. Therefore, considering both extraction efficiency and enzyme activity preservation, this experiment determined 40 kHz as the optimal ultrasound frequency parameter.

Response surface methodology experiment

Response surface methodology‐based experimental design and outcomes

Drawing on the findings from single‐factor trials, a response surface methodology experimental design was developed, as presented in Table 2. The associated experimental outcomes are summarized in Table 3. Using Design‐Expert 13 software, the collagen extraction yield from M. albus bone powder was fitted with a quadratic polynomial model: Y = −147.32521 + 2.03710A + 0.650832B + 1.88780C + 0.000800AB − 0.001250AC − 0.000600BC − 0.052500A ² − 0.000832B ² − 0.022500C ².

Table 2.

Factors and levels of Box–Behnken test design

Level	A: Ultrasonic time (min)	B: Ultrasonic power (W)	C: Ultrasonic frequency (kHz)
−1	10	313	20
0	20	376	40
1	30	438	60

Table 3.

Results of response surface test design

Run no.	A: Ultrasonic time (min)	B: Ultrasonic power (W)	C: Ultrasonic frequency (kHz)	Extraction yield (%)
1	−1	−1	0	24 ± 1.23
2	1	−1	0	27 ± 0.84
3	−1	1	0	25 ± 2.03
4	1	1	0	30 ± 1.34
5	−1	0	−1	22 ± 0.74
6	1	0	−1	26 ± 1.15
7	−1	0	1	16 ± 1.05
8	1	0	−1	19 ± 0.79
9	0	−1	−1	24 ± 1.34
10	0	1	1	28 ± 1.08
11	0	−1	1	19 ± 1.04
12	0	1	0	20 ± 1.23
13	0	0	0	36 ± 2.21
14	0	0	0	35 ± 2.09
15	0	0	0	34 ± 1.57
16	0	0	0	35 ± 1.35
17	0	0	0	35 ± 2.04

ANOVA and tests of significance for the model's coefficients were conducted, with the findings compiled in Table 4. In response surface analysis, the F‐value quantifies the significance of each variable on the response. The overall model F‐value of 233.27 (P < 0.0001) demonstrates that the regression model is highly significant. A nonsignificant lack‐of‐fit F‐value of 3.86 (P = 0.9136 > 0.05) further confirms the model's good fit to the experimental data. Based on P, all linear terms (A, B, C) and quadratic terms (A ², B ², C ²) exhibited a highly significant influence on collagen extraction yield, and the interaction term BC was also statistically significant. Based on the F‐values, the impact of each factor on the collagen extraction yield was ranked in the following order: ultrasonic frequency (C) > ultrasonic time (A) > ultrasonic power (B). The adjusted coefficient of determination (R ² _Adj) was 0.9924, and the coefficient of determination (R ²) was 0.9967, indicating a strong agreement between the predicted and experimental values. Overall, the regression model demonstrates a good fit and adequately reflects the impact of each variable on the extraction yield, making it suitable for optimizing and predicting the extraction process of collagen from M. albus bone.

Table 4.

Analysis of variance of regression model equation

Source of variation	Sum of squares	Degrees of freedom	Mean square	F	P	Significance
Model	674.81	9	74.98	233.27	<0.0001	**
A: Ultrasonic time (min)	28.13	1	28.13	87.50	<0.0001	**
B: Ultrasonic power (W)	10.13	1	10.13	31.50	0.0008	**
C: Ultrasonic frequency (kHz)	84.50	1	84.50	262.89	<0.0001	**
AB	1.0000	1	1.0000	3.11	0.1211
AC	0.2500	1	0.2500	0.7778	0.4071
BC	2.25	1	2.2500	7.00	0.0331	*
A 2	116.05	1	116.05	361.05	<0.0001	**
B 2	44.47	1	44.47	138.36	<0.0001	**
C 2	341.05	1	341.05	1061.05	<0.0001	**
Residual	2.25	7	0.3214
Lack of fit	0.2500	3	0.0833	0.1667	0.9136
Pure error	2.00	4	0.5000
Total	667.06	16
R 2	0.9967
R 2 adj	0.9924

^*Indicates a statistically significant difference (P < 0.05).

^**Signifies an extremely significant difference (P < 0.01).

Analysis of response surfaces for interaction effects of different factors

Employing Design‐Expert 13 software, a multiple regression model was established for extraction yield from M. albus bones, and response surface analysis was conducted to explore the interaction effects of the three factors across three levels. The results demonstrate that the relationship between the factor interactions and extraction rate was not strictly linear. As shown in Fig. 2, the response surface curves for the interaction between ultrasonic time and ultrasonic power, as well as between ultrasonic time and ultrasonic frequency, are relatively flat, with contour lines approximating circular shapes. This suggests that the impacts of these two pairs of interactions on the extraction yield are insignificant. In contrast, the contour lines for the interaction between ultrasonic power and ultrasonic frequency exhibit an elliptical distribution, and the three‐dimensional surface shows a noticeable vertical span variation. This verifies that this particular interaction exerts a significant influence on the extraction yield. Based on the response surface optimization results and using extraction rate as the indicator, the optimal ultrasonic parameters were determined, providing a theoretical basis for subsequent process optimization.

Figure 2.

Three‐dimensional response surface diagrams of critical parameter interactions on collagen extraction yield from M. albus bones. (A, D) Interaction between ultrasound duration and power intensity. (B, E) Interaction between ultrasound duration and frequency. (C, F) Interaction between ultrasound power and frequency.

Validation experiment of optimal conditions from response surface methodology

Response surface optimization analysis showed that the theoretically optimal process for extracting collagen from M. albus bones by ultrasound‐assisted enzymatic hydrolysis was as follows: ultrasonication time of 22 min, power of 389 W, and frequency of 36 kHz, with a theoretical extraction rate of 34.31%. Considering practical feasibility, the process parameters were adjusted to: ultrasonication time of 22 min, power of 375 W, and frequency of 40 kHz. Three independent validation experiments were performed, affording an extraction yield of 35.00 ± 0.48%, which closely aligned with the theoretical prediction. This demonstrates that the response surface model is capable of accurately forecasting the enzymatic extraction process for M. albus bone collagen.

Compared with traditional collagen extraction methods, the ultrasound‐assisted enzymatic approach used in this study offers superior extraction yield. Yu et al. achieved an acid extraction yield of 24.50% from turtle skin ²⁷ ; Srinivasan reported yields of 12.86% (acid) and 14.23% (enzymatic) from sea bass scales, with ultrasound assistance only improving the yield to 15.36% ⁴⁶ ; Scheja et al. obtained 18–25% from lumpfish skin at pilot scale. ⁴⁷ These values are all markedly lower than the yield of 35.00 ± 0.48% achieved in this study. This comparison suggests that ultrasonic cavitation can efficiently break down bone tissue, expose cleavage sites, and promote enzyme–substrate interaction, thus accelerating hydrolysis and improving extraction efficiency.

Characterization of M. albus bone collagen

UV absorption characteristics

Proteins can absorb UV light at specific wavelengths. UV absorption spectroscopy records the curve of protein absorbance in a test solution as a function of the incident light wavelength using a UV spectrophotometer. Aromatic amino acids, notably tyrosine (Tyr) and tryptophan (Trp), display strong absorption peaks at 280 and 251 nm, a characteristic attributable to their benzene ring structures. ⁴⁸ Unlike most proteins, collagen contains relatively low levels of aromatic amino acids. Its typical absorption peak is mainly attributed to C=O bonds, carboxyl (COOH), and amide (CO—NH₂) groups, usually appearing near 230 nm. ⁴⁹ As shown in Fig. 3(A), the spectrum of isolated M. albus bone collagen exhibited a prominent absorption peak at 236 nm; this aligns with the known absorption maximum of type I collagen and authenticates the sample. Additionally, a weak and broad peak was observed around 280 nm, likely resulting from a small amount of exposed tyrosine residues.

Figure 3.

Characterization of M. albus bone collagen. (A) UV absorption spectrum. (B) Electrophoresis image. (C) XRD pattern. (D) SEM images (a: ×300; b: ×500c: ×1000). (E) Mechanical strength. (F) TGA curve. (G) DSC curve.

Molecular weight and subunit composition

SDS‐PAGE serves as a commonly employed method for ascertaining the molecular weight distribution of proteins and their subunit composition. ⁵⁰ Figure 3(B) shows the SDS‐PAGE profile of M. albus bone collagen, with lane 1 corresponding to the protein marker and lane 2 to the extracted collagen sample. Two distinct bands were observed at approximately 120 kDa, which correspond to the α ₁ and α ₂ chains. Furthermore, two bands around or exceeding 270 kDa were identified as β and γ components, in line with prior studies on type I collagen. ⁵¹ The absence of significant low‐molecular‐weight impurity bands in the gel indicates that the obtained collagen possesses high purity.

Microstructure

Microstructural characteristics of M. albus bone collagen were visualized via SEM. After freeze drying, the sample appeared as a white sponge‐like material with a delicate texture and good elasticity. Figure 3(D) shows SEM images at different magnifications (×300; ×500; ×1000). It can be seen that the surface pores of the collagen are evenly distributed, the collagen fibers are densely interwoven, forming a typical fibrous network structure, indicating the presence of stable crosslinking between collagen molecules.

Triple‐helical structure characteristics

XRD is a widely used classic technique for analyzing the spatial structure of materials, based on the scattering of X‐rays by atoms or ions in crystals. ⁵² The XRD pattern of M. albus bone collagen (Fig. 3(C)) was analyzed using the Bragg equation: d (Å) = λ/(2 sin θ) (λ = 0.154 nm). This value agrees with the characteristic spacing between collagen molecular chains, reflecting the aggregation state of the macromolecular chains. ⁵³ The second diffraction peak at 2θ = 22.58° gave d = 3.94 Å, a feature attributed to ordered scattering from the multilayer stacking of collagen fibrils, indicating regularity in the supramolecular structure. ⁵⁴ The detected peaks align with the characteristic spectral profile of collagen with a triple‐helical conformation.

Amino acid composition

The nutritional value and physicochemical functionalities of collagen are associated with its amino acid composition. ⁵⁵ As presented in Table 5, M. albus bone collagen was rich in amino acid variety and content, comprising a total of 18 amino acids, among which 8 are essential amino acids (accounting for 302.6 g kg⁻¹). Among them, glycine (141.0 ± 12.4 g kg⁻¹), glutamic acid (106.5 ± 10.8 g kg⁻¹), and proline (89.7 ± 16.7 g kg⁻¹) are the most abundant, which corresponds to the amino acid compositional features of type I collagen. ⁵⁶ Proline and hydroxyproline play a key role in maintaining the molecule's thermal stability and preserving the integrity of its triple‐helical conformation. ⁵⁷ Methionine, tyrosine, isoleucine, and histidine were present at minimal levels in the collagen, and cysteine accounted for only 1.6 ± 1.2 g kg⁻¹. This profile matches the documented amino acid composition of eel skin collagen. ⁵⁸

Table 5.

Amino acid composition and content of collagen in M. albus bones

Compound name	Content relative to total amino acids (g kg−1)
Asp	81.3 ± 2.3
Thr	41.9 ± 10.2
Ser	46.2 ± 3.4
Glu	106.5 ± 5.8
Gly	141.0 ± 12.4
Ala	57.8 ± 2.6
Cys	1.6 ± 1.2
Val	37.0 ± 2.2
Met	15.1 ± 5.6
Ile	30.8 ± 6.7
Leu	54.0 ± 11.3
Tyr	23.5 ± 8.2
Phe	31.9 ± 4.7
His	47.6 ± 10.2
Lys	44.3 ± 11.5
Arg	75.9 ± 20.1
Pro	89.7 ± 16.7
Hyp	73.8 ± 3.5

Essential amino acids: Thr, Val, Met, Ile, Leu, Phe, His, Lys; hydrophobic: Ala, Val, Met, Pro, Ile, Leu, Tyr, Phe; aromatic: Phe, Tyr; electron‐/hydrogen‐donating‐related: Glu, Met; positively charged: His, Arg, Lys.

Research indicates that hydrophobic and aromatic amino acids can act as free radical scavengers by donating hydrogen atoms and electrons, ⁵⁹ Basic and acidic amino acids possess hydrogen‐donating capacity, enabling them to participate in free radical scavenging, ⁶⁰ thereby exerting antioxidant activity. ⁶¹ As presented in Table 5, M. albus bone collagen contains a relatively high content of hydrophobic amino acids (339.8 g kg⁻¹), with aromatic amino acids accounting for 55.4 g kg⁻¹, along with histidine (47.6 ± 10.2 g kg⁻¹) and aspartic acid (81.3 ± 2.3 g kg⁻¹). These amino acid residues may directly participate in the capture and neutralization of free radicals through their active groups, thereby endowing the collagen with significant antioxidant capacity.

Rheological properties

Rheological analysis (Fig. 3(E)) indicates that within the measurement range, the storage modulus (G′) of the M. albus collagen solution consistently exceeded the loss modulus (G″). This demonstrates that the collagen exhibits strong energy storage and deformation resistance, coupled with weak viscous dissipation. Collagen's triple‐helix conformation, coupled with intermolecular forces (encompassing hydrogen bonds and van der Waals interactions), constructs a stable three‐dimensional network, endowing M. albus bone collagen with superior mechanical properties and elasticity.

Thermodynamic property analysis

TGA reflects a material's thermal stability by monitoring changes in sample mass as temperature varies. The TGA curve of M. albus collagen protein (Fig. 3(F)) shows a stepwise decrease in mass with increasing temperature, with a total weight loss of 69.82%. The primary mass loss consists of sample moisture, volatile impurities, and broken collagen polymer peptide chains. After thermal decomposition is complete, the remaining mass is primarily inorganic salts and carbonization products. The temperature range of 200–760 °C exhibits the most significant degradation of M. albus bone collagen protein (65.64 ± 2.12% weight loss), including peptide chain breakage and carbonization of organic matter. M. albus bone collagen exhibits outstanding thermal stability below 200 °C, largely due to its triple‐helix structure. The thermal stability characteristics of M. albus bone collagen, reflected in its decomposition behavior, lay a theoretical foundation for its prospective application in thermally processed foods and high‐temperature biomedical fields, such as pharmaceutical preparations.

Thermal stability

The intramolecular hydrogen bonds in collagen's triple‐helical α‐chain conformation are essential to its resistance to high temperatures. ⁶² DSC was employed to assess the thermal stability of the collagen. The results (Fig. 3(G)) showed that M. albus bone collagen exhibited an endothermic peak at 92.7 ± 1.24 °C, indicating that collagen underwent thermal denaturation at this temperature. TGA confirmed that no significant weight loss occurred at this temperature, ruling out thermal decomposition. ⁶³ Previous studies reported denaturation temperatures (Tₘₐₓ) of acid‐soluble and pepsin‐soluble collagen from giant salamander bone at 45.78 and 46.51 °C, respectively, ⁶⁴ while collagen from common marine and freshwater fish typically denatures at approximately 15.2–41.8 °C. ⁶⁵ In contrast, certain amphibian collagens show higher thermal stability than that of many marine or freshwater species; for example, collagen from Asian bullfrog skin exhibits denaturation temperatures ranging from 64.93 to 110.12 °C. ⁶⁶ The superior thermal stability of M. albus bone collagen arises from a combination of structural determinants. The contents of proline (89.7 ± 16.7 g kg⁻¹) and hydroxyproline (73.8 ± 3.5 g kg⁻¹) are markedly elevated relative to those in common freshwater fish, ⁶⁷ and their pyrrolidine rings confer rigidity to the triple‐helical structure. ⁶⁸ Such high amino acid levels promote the formation of a dense network of intramolecular hydrogen bonds, further stabilized by water‐bridge interactions.69, 70 In addition, XRD patterns indicate a well‐ordered crystalline organization and closely packed fibril arrangement, which facilitate intermolecular crosslinking.71, 72 Together, these features – increased chain rigidity, an extensive hydrogen‐bonding network, and a highly ordered supramolecular architecture – account for its exceptional thermal stability.

Antioxidant activity of M. albus bone collagen

DPPH free radical scavenging activity

As shown in Fig. 4(A), the DPPH free radical scavenging potential of bone collagen and ascorbic acid both exhibited a marked concentration‐dependent increase. When the collagen concentration reached 10 mg mL⁻¹, the scavenging rate was 73.00 ± 0.82%, and the corresponding IC₅₀ value was calculated to be 6.01 ± 0.36 mg mL⁻¹. Compared with other collagen and collagen‐derived products reported in the literature, the M. albus bone collagen extracted in this study displayed stronger DPPH scavenging activity. A significantly lower IC₅₀ value was observed compared to that of giant salamander skin collagen (9.96 mg mL⁻¹), ⁷³ tilapia skin collagen (56.31 mg mL⁻¹), ⁷⁴ and deer bone collagen peptides (21.57 mg mL⁻¹). ⁷⁵ This potent DPPH scavenging activity can be attributed to the relatively high content of aromatic amino acids (55.4 g kg⁻¹) in the collagen (Table 5), particularly tyrosine and phenylalanine, which act as effective hydrogen donors to stabilize DPPH radicals. ⁵⁹ In comparison, ascorbic acid exhibited a significantly lower IC₅₀ value of 0.18 ± 0.02 mg mL⁻¹ (P < 0.01 versus M. albus bone collagen), confirming its potent antioxidant capacity as a positive control. In summary, the collagen from M. albus bones has strong antioxidant activity.

Figure 4.

Scavenging activities of M. albus bone collagen (2, 4, 6, 8, 10 mg mL⁻¹) and ascorbic acid (Vc, vitamin C; 0.2, 0.4, 0.6, 0.8, 1.0 mg mL⁻¹) against (A) DPPH radical, (B) superoxide anion radical, and (C) hydroxyl radical. The X‐axis is shared for visual comparison; the actual concentration ranges of collagen and Vc are specified in the legend. The IC₅₀ values of M. albus bone collagen were 6.01 ± 0.36 mg mL⁻¹ (DPPH), 7.63 ± 0.24 mg mL⁻¹ (superoxide anion), and 17.71 ± 0.19 mg mL⁻¹ (hydroxyl radical). The IC₅₀ values of ascorbic acid were 0.18 ± 0.02 mg mL⁻¹ (DPPH), 0.34 ± 0.03 mg mL⁻¹ (superoxide anion), and 0.28 ± 0.02 mg mL⁻¹ (hydroxyl radical). Data are expressed as mean ± standard deviation (n = 3).

Scavenging activity of superoxide anions

Superoxide anion radicals exhibit strong oxidative properties that can induce various oxidative damages, leading to cellular and tissue injuries. ⁷⁶ Antioxidant proteins can scavenge superoxide anion radicals produced during the autoxidation of pyrogallol, thereby inhibiting its autoxidation. ⁷⁷ Figure 4(B) shows that the ability of M. albus bone collagen and ascorbic acid to scavenge superoxide anion radicals is positively correlated with their concentration. M. albus bone collagen at 10 mg mL⁻¹ achieves a superoxide anion scavenging rate of 62.05 ± 0.47%, and its IC₅₀ value is calculated to be 7.63 ± 0.24 mg mL⁻¹. Compared with the previously reported Andrias davidianus skin collagen (IC₅₀ = 7.79 mg mL⁻¹) ⁷³ and tilapia skin collagen peptides (IC₅₀ = 49.10 mg mL⁻¹), ⁷⁴ the M. albus bone collagen extracted in this study exhibited stronger superoxide anion radical scavenging activity. The superoxide anion scavenging capacity may be related to the presence of basic and acidic amino acids such as histidine (47.6 g kg⁻¹) and aspartic acid (81.3 g kg⁻¹) (Table 5), which can donate protons or interact electrostatically with the negatively charged superoxide radical. ⁶⁰ By contrast, the IC₅₀ value of ascorbic acid against superoxide anion radicals was 0.34 ± 0.03 mg mL⁻¹, which was significantly lower than that of M. albus bone collagen (P < 0.01).

Hydroxyl radical scavenging capacity

In the process of aerobic metabolism, the body produces highly oxidative hydroxyl radicals. The antioxidant activity of rice field M. albus bones collagen can be analyzed by measuring its ability to scavenge hydroxyl radicals. ⁷⁸ The results in Fig. 4(C) show that the hydroxyl radical scavenging rates of M. albus bone collagen and ascorbic acid increased with increasing sample concentration. Over the concentration interval of 2–10 mg mL⁻¹, the scavenging activity against hydroxyl radicals rose from 18.19 ± 0.43% to 41.68 ± 0.27%, with the corresponding IC₅₀ value determined as 17.71 ± 0.19 mg mL⁻¹. The ability to scavenge hydroxyl radicals is likely due to the metal‐chelating potential of histidine residues (47.6 g kg⁻¹; Table 5), which can inhibit Fenton reaction‐mediated hydroxyl radical generation, as well as the hydrogen‐donating capacity of other amino acids such as aromatic residues. ⁶¹ The IC₅₀ value of ascorbic acid against hydroxyl radicals was determined to be 0.28 ± 0.02 mg mL⁻¹, significantly lower than that of M. albus bone collagen (P < 0.01). These findings indicate that M. albus bone collagen is a promising hydroxyl radical scavenger.

CONCLUSIONS

This study successfully achieved the high‐value conversion of Monopterus albus bone processing byproducts through ultrasound‐assisted enzymatic hydrolysis. Under optimized process conditions, the collagen extraction yield reached 35.00 ± 0.48%. The obtained type I collagen exhibited a loose and porous network structure, with a hydrophobic amino acid content of 33.98 g kg⁻¹, demonstrating excellent mechanical properties and thermal stability. Its IC₅₀ values against DPPH radicals, superoxide anion radicals, and hydroxyl radicals were 6.01 ± 0.36, 7.63 ± 0.24, and 17.71 ± 0.19 mg mL⁻¹, respectively, indicating significant antioxidant activity. Based on its high thermal stability (92.7 °C) and strong antioxidant activity, M. albus bone collagen shows application potential in heat‐processed food processing, sterilizable biomedical materials, and functional cosmetics, with its antioxidant performance being significantly superior to that of various previously reported collagens. This research not only provides a practical pathway for the resource utilization of M. albus bone byproducts, effectively reducing bone resource waste, but also establishes an environmentally friendly and sustainable processing method, offering important practical significance for promoting the high‐value utilization of aquatic processing byproducts.

AUTHOR CONTRIBUTIONS

Ruichang Gao: Conceptualization, supervision, writing – review and editing, funding acquisition, supervision, formal analysis, methodology. Yingxi Jia: Investigation, formal analysis, methodology, data curation, writing – original draft, writing – review and editing. Jiayu Lu: Investigation. Zitao Guo: Formal analysis, methodology, writing – review and editing.

CONFLICT OF INTEREST

There is no conflict of interest among all the authors.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.