Effects of CO₂ and O₂ in Modified Atmosphere Packaging on Water Retention, Protein Stability, and Microbial Growth in Atlantic Salmon Fillets

Abstract

Modified atmosphere packaging (MAP) is commonly used to prolong the shelf life and maintain the quality of perishable food. However, it may contribute to more severe juice loss and texture changes in salmon. To explore the reasons why, this study designed different ratios of O₂ (to inhibit anaerobic bacteria), CO₂ (to inhibit Gram-negative bacteria), and N₂ (to maintain the packaging shape) in order to investigate the effects of MAP on the properties, structure, and oxidation of salmon proteins. The experiments’ results showed that MAP with about 60% CO₂ could slow bacterial growth effectively, as well as the accumulation of total volatile basic nitrogen and cooking loss. The carbonyl content decreased with increasing CO₂ contents but increased with high contents of O₂. A low concentration of O₂ (10%) was also beneficial for the inhibition of oxidation and degradation of proteins, and the lowest carbonyl content was found in 60%CO₂/10%O₂/30%N₂ conditions, with 2.01 μmol/g protein on day 12. Overall, we report that MAP with 60%CO₂ and 10%O₂ is properly able to limit structure changes in the myofibrils of salmon fillets during cold storage.

1. Introduction

Salmon (Salmo salar) is rich in water, which is generally found in the interstitial spaces of myofibrillar proteins or bound to biological macromolecules, endowing it with taste qualities such as hardness and springiness [1]. Myosin and actin, as the primary components of myofibrillar proteins, maintain the water-holding capacity and texture of fish muscle by interacting with water molecules to form a stable network structure [2]. During postmortem storage, microbial activity and proteolytic enzymes progressively degrade proteins and induce oxidative modifications, compromising the muscle’s structural integrity. This dual process manifests through distinct mechanisms: Proteolytic degradation widens intermyofibrillar spaces, resulting in measurable drip loss. Concurrently, oxidative damage exposes hydrophobic amino acid residues, facilitates covalent cross-linking between proteins, and modifies surface charge characteristics. These cumulative biochemical alterations significantly impair the muscle tissue’s water-holding capacity and ultimately diminish product quality [3,4,5]. These changes in muscle and its related water distribution and water-holding properties can be measured by differential scanning calorimetry (DSC), low-field nuclear magnetic resonance (LF-NMR), and cooking loss [6,7,8].

Modified atmosphere packaging (MAP) has gained considerable attention as an effective method to slow down the deterioration process and improve the water distribution of salmon fillets [9]. CO₂, O₂, and N₂ are important components of MAP. It has been reported that CO₂ can inhibit Gram-negative bacterial metabolism by partially decreasing intracellular pH, depolarizing cell membranes, lowering proton motive force, changing lipophilic nature, and reducing ATP synthesis [10]. O₂ can inhibit the growth of anaerobic bacteria and combine with myoglobin to give meat a fresh red appearance [11]. Additionally, nitrogen is usually used to maintain the packaging shape as an inert gas [12].

The effects of CO₂ and O₂ on the bacteria of aquatic products and other meat products are well studied. However, their effect on salmon water distribution and protein characteristics remains unclear. This paper offers a theoretical basis for the application of MAP in salmon preservation by analyzing changes in microbial growth, protein degradation and oxidation, and water migration.

2. Materials and Methods

2.1. Sample Preparation and Modified Atmosphere Packaging

Fresh farmed whole salmons (approximately 6 kg per fish, aged 2–3 years, harvested and sold within 2 days) were obtained from RT-MART (Pudong District, Shanghai, China) and complied with the Aquaculture Stewardship Council (ASC) standards. The salmons were placed in foam boxes with ice slurry and delivered to the laboratory within 30 min. Upon arrival, the samples were cut into pieces, and each piece of salmon weighed approximately 120 g (mainly the dorsal muscle of the fish from its mid part). The fillets were washed in cold water and drained. Then, they were placed into polyamide/polyethylene (PA/PE) pouches, randomly (Xilong Packaging Co., Ltd., Shijiazhuang, China; size: 32 cm × 22 cm; thickness: 0.32 mm; oxygen permeability: 7 cm³/(m²·24 h,·0.1 MPa)). The pouches were separated into six different batches, settled as follows: (1) CK: air, control; (2) 100C: 100%CO₂; (3) 60C40N: 60%CO₂/40%N₂; (4) 20C80N: 20%CO₂/80%N₂; (5) 60C10O30N: 60%CO₂/10%O₂/30%N₂; (6) 60C40O: 60%CO₂/40%O₂.

Firstly, the MAP pouches were treated using gas flushing equipment (Model BQ-360W, Shanghai Qingba Food Packaging Machinery Co., Ltd., Shanghai, China) at a ratio of about 3:1 (gas volume to product). The samples were stored at 4 °C in a refrigerator for a duration of 12 days, and sampling was carried out every 2 days.

2.2. Microbiological Analysis

Microbiological analysis was conducted, including the total mesophilic bacterial count (TMB) and the total psychrophilic bacteria (TPB), which were measured according to the method contained in [13] with a slight modification. A sample (25 g) was taken for analysis. The total mesophilic bacterial count (TMB) was determined after cultivation on Plate Count Agar at 30 °C for 72 h, and the total psychrophilic bacteria (TPB) were determined after cultivation at 4 °C for 10 d. The colonies were counted, and the results were expressed as log cfu/g. The experiment was tested in triplicate.

2.3. Total Volatile Basic Nitrogen (TVB-N) Analysis

The total volatile basic nitrogen (TVB-N) content in the salmon was determined using an Automatic Kjeldahl Apparatus [13]. A 5.0 g portion of salmon flesh was ground and mixed with 50 mL of distilled water, then filtered. The sediment was then heated with steam after the addition of MgO. The TVB-N values are reported as mg N/100 g salmon. The experiment was performed in triplicate.

2.4. Cooking Loss Analysis

Cooking loss referrers to the loss of weight related to heat treatment, as a percentage [8]. The raw salmon fillets were uniformly diced into 8 cm³ cubic specimens (2 cm × 2 cm × 2 cm), with their initial weight precisely documented as W₁. These fish cubes were heated at 80 °C for 10 min. After absorbing the extra water on the surface, their weight after cooking was recorded as W₂. The experiment was performed in triplicate. The cooking loss was calculated with the following formula:

$$Cooking loss (\%)={\displaystyle \frac{W_1{-W}_2}{W_1}}\times 100\%$$

(1)

2.5. Protein Oxidation Determination

Carbonyl content was determined according to the method used by Zhang, et al. [14]. Briefly, 1 mL of a protein solution was mixed with 10 mM DNPH and incubated in the dark for 1 h. After adding 1 mL of 20% trichloroacetic acid (TCA), the sample was centrifuged at 10,000× g for 5 min. The sediment was washed three times by ethyl acetate/ethanol (1:1, v/v), and then dissolved in 6 M guanidine hydrochloride (in 2 M HCl) and incubated at 37 °C for 15 min. The absorbance at 370 nm was measured, and the carbonyl content was calculated using a molar extinction coefficient of 22,000 M⁻¹ cm⁻¹. The experiment was tested in triplicate.

Total sulfhydryl group content was measured using a sulfhydryl group determination kit (A063-1, Nanjing Jiancheng Bioengineering Institute, Nanjing, China) according to its protocol. Briefly, 0.1 mL of myofibrillar protein (MP) was reacted with the chemical agents, and the absorption was measured at 412 nm. Each sample was tested in triplicate.

2.6. SDS–Polyacrylamide Gel Electrophoresis (SDS-PAGE)

Sarcoplasmic and myofibrillar proteins were extracted one at a time using the method described by Niamnuy, et al. [15] at 4 °C. The extracted sample solutions were mixed with a loading buffer at a ratio of 1:10 and heated at 90 °C for 5 min. A 5% stacking gel and 10% separating gel were utilized, and electrophoresis was performed at a voltage of 120 V for approximately 20 min, followed by a constant voltage of 100 V. The proteins were stained with 0.125% Coomassie brilliant blue R-250 and destained in 10% acetic acid. The experiment was tested in triplicate.

2.7. Fourier Transform Infrared Spectroscopy (FTIR) Analysis

A Fourier transform infrared (FTIR) spectrometer (Nicolet IS50; Thermo Scientific Inc., Waltham, MA, USA) was used to detect the secondary structures of salmon myofibrillar proteins in the spectra spanning 1600–1700 cm⁻¹. The content of secondary structures was calculated by PeakFit 4.12 software (SPSS Inc., Chicago, IL, USA).

2.8. Low-Field Nuclear Magnetic Resonance (LF-NMR) Analysis

The states of moisture in the salmon fillets were determined according to the method of Nikoo Mehdi, Benjakul Soottawat, Ahmadi Gavlighi Hassan, Xu Xueming, and Regenstein Joe M. [6], with some modifications. The samples were cut into 2 cm × 2 cm × 2 cm fish pieces and placed in a cylindrical glass tube in an LF-NMR analyzer (Niumag, Ltd., Shanghai, China). The transverse relaxation time T₂ was obtained by the analysis software (NMI20-030H-1 NMR analyzer: Suzhou Niumag Analytical Instrument Co., Suzhou, China).

2.9. Statistical Analysis

The statistical analysis was carried out for multiple comparisons by SPSS 20.0 software (SPSS Version 20.0, Inc., Chicago, IL, USA) and significance was determined at the 0.05 level, followed by Duncan’s test. The diagrams were designed in the Origin2018 software (OriginLab, Northampton, MA, USA).

3. Results

3.1. Microbial Growth

The growth of mesophilic bacteria in the salmon during storage in MAP is shown in Figure 1a. The initial TMBs were 3.59 log cfu/g. The TMBs increased during storage in all packaging conditions, continuously. On day 6, the TMB of the CK group exceeded 7 log cfu/g (threshold). At the end of storage, the TMBs of the 60C40N and 60C10O30N groups were the lowest (p > 0.05).

The changes in the TPB of the salmon are shown in Figure 1b. On day 12, the TPB counts of the CK, 100C, 60C40N, 20C80N, 60C10O30N, and 60C40O groups increased from 1.87 log cfu/g to 8.02 log cfu/g, 6.45 log cfu/g, 5.31 log cfu/g, 6.33 log cfu/g, 4.94 log cfu/g and 6.76 log cfu/g, respectively.

3.2. Changes in TVB-N Content

As illustrated in Figure 1c, the initial TVB-N count was 12.79 mgN/100 g. At the end of storage, the TVB-N counts of the CK, 100C, 60C40N, 20C80N, 60C10O30N, and 60C40O groups significantly increased, to 24.09 mgN/100 g, 24.55 mgN/100 g, 18.63 mgN/100 g, 22.93 mgN/100 g, 17.13 mgN/100 g, and 23.04 mgN/100 g, respectively (p < 0.05).

3.3. Changes in Cooking Loss

As shown in Figure 1d, the initial value of cooking loss in the CK group was 19.48%. The cooking loss increased during storage, but the values for the 60C40O and 60C10O30N groups were significantly lower than for the other groups (p < 0.05).

3.4. Changes of Carbonyl Content and Total Sulfhydryl Content

The carbonyl content of fresh the salmon fillets was 1.25 μmol/g protein (Figure 2a). The carbonyl content increased during storage, which for the 100C, 60C10O30N, and 60C40O groups were significantly lower than other groups on day 12 (p < 0.05).

As shown in Figure 2b, the sulfhydryl content of the fresh salmon was 68.36 μmol/g protein. The total sulfhydryl content of all groups decreased, with the CK group reaching the lowest level (45.29 μmol/g protein) and the 60C10O30N group reaching the highest (p < 0.05).

3.5. SDS-PAGE

The muscle proteins in aquatic products are mainly composed of myofibrillar and sarcoplasmic proteins. As shown in Figure 3, the main bands on the gel represent myosin heavy chain (MHC, ~200 kDa), paramyosin (PM, 100~135 kDa), actin (35~48 kDa), and the myosin light chain (MLC, 14.3~27 kDa) proteins. In the first six days of storage, the optical density of the MHC proteins in all the samples slightly decreased, while the optical density of the PM increased. On day 6 and 12, the MHC intensity in the 60C10O30N group was heavier than in the other groups.

3.6. Secondary Structure

As shown in Figure 4, the amide I region (1700–1600 cm⁻¹) of FTIR spectra (Figure 4a,b) can be used to study the secondary structure of proteins (random coil: 1640–1650 cm⁻¹; α-helix: 1650–1660 cm⁻¹; β-turn: 1660–1680 cm⁻¹; and β-sheet: 1680–1700 cm⁻¹). The quantitative analysis of the four secondary structures of myofibrillar proteins extracted from salmon are shown in Figure 4c,d. The initial ratios of α-helix, β-sheet, β-turn, and random coil were 18.2%, 30.8%, 28.2%, and 17.8%, respectively. The ratios of α-helix and β-sheet in the control group decreased to 11.6% and 27% and the ratio of random coil increased to 42.7% on day 12. At the end of storage, the ratios of α-helix and β-sheet in the CK and 20C80N groups were lower than in the other groups.

3.7. LF-NMR

LF-NMR can characterize the water distribution in fish samples by analyzing changes in the peak area of relaxation time under different MAP conditions (Figure 5). We considered that the moisture in the salmon consisted of three parts: A₂₁ (bond water, tightly bound to protein, Figure 5a), A₂₂ (immobilized water, within the protein-dense myofibrillar network, Figure 5b), and A₂₃ (free water, in the extra myofibrillar space, Figure 5c). At the end of storage, the 60%CO₂/10%O₂/30%N₂ group and 60%CO₂/40%N₂ group maintained a higher bound water content compared to the other MAP conditions (Figure 5a). Figure 5b showed that the fresh salmon had the highest content of immobilized water, and all the MAP groups had higher immobilized water contents compared to the control group (CK), except for 100CO₂, during storage time. The content of free water increased continuously during storage (Figure 5c).

3.8. Correlation Analysis

As shown in Figure 6, the TMB and TPB levels exhibited significant positive correlations with TVB-N values, cooking losses, A₂₃, carbonyl, and random coils (p < 0.05), but significant negative correlations with A₂₁, A₂₂, α-helixes, and β-turns (p < 0.05). The changes in β-sheet and β-turn had little relation to the other indicators.

4. Discussion

TMB and TPB are important indicators of fish spoilage, closely associated with the degradation of myofibrillar proteins, and have significant impacts on the water-holding capacity and water distribution of fish [16,17]. The 100% CO₂ condition initially inhibit bacterial activity during the first 2 days of storage (Figure 1a), which should be due to a reduction in the pH of fish flesh through carbonic acid formation [18]. However, its suppressive effect on acid-tolerant and anaerobic bacteria remains limited [19]. It was found that the inhibitory effect of 20C80N conditions on mesophilic bacteria was close to that of 100C but less efficient than the 60C40%N and 60C10O30N conditions, which was consistent with the results of a previous study [20]. This was attributed to the fact that a moderate amount of oxygen helps to inhibit the growth of anaerobic bacteria [21]. Therefore, excessive CO₂ might induce muscle protein dissolution and membrane structural damage, triggering leakage of intracellular substances. These leaked components serve as substrates for microbial metabolism, paradoxically accelerating spoilage [22]. The count in the 60C10O30N group was lower than the 60C40O group, indicating that a low level of O₂ (10%) could slow the growth of anaerobic bacteria [23]. Compared to the 60C40O, 60C40N, and 60C10O30N groups, the TMB of the 60C10O30N group was lowest during storage, and that of the 60C40N group was the highest. The increasing tendency of the TPBs counts in each group was similar to that of TMB (Figure 1b), and the 60C10O30N group had significantly lower TPB counts than the other groups (p < 0.05).

The TVB-N value, a critical indicator for fish spoilage, is the result of the decomposition of peptides and proteins by microorganisms to produce basic nitrogenous substances, with the generally accepted threshold for the spoilage of marine fish being 20 mg N/100 g [24]. The TVB-N values of all the groups increased during storage and displayed a similar pattern to bacterial growth (Figure 1c). The rapid increase in TVB-N value in the CK group was due to the microbial degradation of proteins. The TVB-N values of the 100C, 20C80N and 60C40O groups were close to that of the CK group, and those of the 100C group were even higher than the CK group after 2 days of storage (p < 0.05), which exceeded the threshold on day 6 (21.03 mg N/100 g). The adverse effect of 100C conditions on TVB-N was in agreement with our previous study [20]. It is hypothesized that a high ratio of CO₂ induces the degradation of proteins, leading to the rapid multiplication of anaerobic bacteria. Therefore, proper proportions of CO₂ and O₂ are essential for inhibiting the accumulation of TVB-N. The 60C10O30N conditions had the lowest TVB-N content (p < 0.05), which was consistent with the microbial results.

Cooking loss reflects the water-holding capacity of fish tissue, originating from the heat-induced denaturation of myofibrillar proteins that compromises muscle structural integrity [25]. This phenomenon is related to three-dimensional protein network alterations, including oxidative modification, polypeptide chain degradation, and conformational changes [26]. The cooking loss in 100C conditions was higher than that in the CK group, reaching 29.43% on day 12. This might be attributed to inadequate microbial suppression during late storage, which in turn produced exogenous enzymes that degraded the myofibrillar protein—a finding consistent with the observations of Sun, et al. [27]. Those in 20C80N and 60C40O conditions were also as high as the CK group (p > 0.05). The 60C40N and 60C10O30 conditions had the best inhibition effect (p < 0.05) (Figure 1d). These two groups should exert dual preservation effects through significant microbial growth inhibition (p < 0.05) that slows exogenous proteolytic activity and the effective suppression of protein carbonyl formation (p < 0.05) that maintains structural conformation.

Carbonyl and sulfhydryl groups are commonly used to evaluate the oxidation level of proteins and exhibit a negative correlation with the moisture content of fish [28]. The CK group had a higher content of carbonyl groups compared to the other groups (Figure 2a), which was due to the oxidation of amino acid residues in the proteins into carbonyl derivatives such as aldehydes [29], followed by the 20C80N and the 100C groups. As shown in Figure 2b, the sulfhydryl content of the CK group was lower than the other groups because of the oxidative exchange of sulfhydryl or disulfide bonds [30]. The sulfhydryl content of the MAP-treated samples was significantly higher than that of the control group. The highest total sulfhydryl content was observed in the 60C40N and 60C10O30N groups on day 12 (p < 0.05), possibly due to their low oxygen concentrations [31]. Therefore, a gas mixture with a low O₂ level was beneficial to the inhibition of protein oxidation, and a CO₂ level of about 60% was also better than 100%.

SDS-PAGE profiles are often used to reflect the degree of protein degradation [32]. During the first 6 days of storage, there was a slight decrease in MHC optical density in all the samples, while the PM optical density increased (Figure 3). This confirmed that the myofibrillar proteins underwent continuous degradation during storage. The thickness bands were clearly observed in the 60C40N, 60C40O, and 60C10O30N groups, suggesting that 60% CO₂ effectively maintained protein integrity, consistent with the results of TMB, TPB, and TVB-N.

The stability of protein is generally positively related to the ratio of α-helix and β-sheets and is negatively related to the random coils [33]. The proportions of α-helixes and β-sheets in the myofibrillar protein decreased (Figure 4c,d), and the proportion of random coils increased due to the oxidation of protein amino acid residues and the changes in non-covalent interactions [34]. At the end of storage, the ratios of α-helixes and β-sheets in the CK and 20C80N groups were lower than in the other groups. It could be found that a very high or low level of CO₂ was not proper to maintain its relatively stable structure.

As indicated by the LF-NMR data, the intensity of bound water and immobilized water in the CK group decreased, while the intensity of free water increased during storage (Figure 5a–c). This phenomenon should be attributed to protein degradation and oxidation, which caused changes in protein conformation and converted the immobilized water to free water [35]. The water distribution of the fish was maintained in all the MAP groups except the 100% CO₂ group, while the 60C10O30N and 60C40N groups had higher contents of bound and immobilized water and a lower content of free water. This was in agreement with the cooking loss results.

Correlation analysis showed that the growth of bacteria was responsible for the quality deterioration of salmon, including the increases in TVB-N, cooking loss, A₂₃, carbonyl, and random coils.

5. Conclusions

This study demonstrated that MAP application significantly influences salmon fillet quality parameters during storage, including a suppression of TMB and TPB, reduction in TVB-N, and mitigation of cooking loss. However, 100% CO₂ conditions proved to induce some deleterious effects on TVB-N, cooking loss, and free water increasement in salmon fillets, and 40% O₂ conditions also had adverse effects on TVB-N and cooking loss. Notably, the 60% CO₂ treatment demonstrated optimal efficacy in preserving both the water-holding capacity and secondary structure integrity of salmon myofibrillar proteins. Thus, MAP under 60% CO₂/10% O₂/30% N₂ conditions was identified as the optimal gas composition. Future experiments should explore how gases affect protein degradation through bacterial and endogenous enzymatic activities to clarify the underlying mechanisms.