Effects of plasma-activated lactic acid on the physicochemical properties and volatile organic compounds of refrigerated tilapia fillets

TL;DR

A new plasma-activated lactic acid method preserves tilapia fillets by slowing microbial growth and oxidation, extending shelf life by four days.

Contribution

A novel plasma-activated lactic acid treatment is introduced for preserving fish quality through multi-target mechanisms.

Findings

PALA reduced microbial counts to 6.41 Log CFU/g and TVB-N to 15.26 mg/100 g after 7 days of storage.

PALA suppressed myofibrillar protein oxidation, reducing carbonyls by 52% and increasing sulfhydryl content by 53%.

PALA improved flavor by reducing spoilage aldehydes while retaining fresh-flavor compounds.

Abstract

This study developed a novel plasma-activated lactic acid (PALA) strategy (650 W, 90 s, 0.0625 % lactic acid) to preserve refrigerated tilapia fillets. Mechanistically, PALA-generated reactive nitrogen species, particularly •NO₃/•NO₂, effectively stabilized pH and inhibited microbial proliferation, maintaining counts at 6.41 Log CFU/g and TVB-N at 15.26 mg/100 g after 7 days of storage. The synergy between acid and reactive oxygen nitrogen species significantly delayed lipid oxidation (TBARS reduced by >40 %) and protein degradation, preserving texture and color. PALA suppressed myofibrillar protein (MP) oxidation, reducing carbonyls and surface hydrophobicity by 52 % and 20 %, respectively, while increasing sulfhydryl content by 53 %, as confirmed by SDS-PAGE showing attenuated myosin/actin degradation. Volatile profiling revealed PALA reduced spoilage aldehydes (heptanal, nonanal) while retaining fresh-flavor aldehydes (octanal, dodecanal). This multi-target action extended the shelf life by 4 days, with nitrate/nitrite residues within safety limits, establishing PALA as an innovative, synergistic preservation technology for aquatic products.

Unlabelled Image

•PALA extended tilapia shelf life by 4 d via inhibiting microbial growth.•PALA delayed lipid/protein oxidation with lower TBARS and TVB-N.•PALA reduced MP carbonyls and surface hydrophobicity, preserving protein integrity.•PALA improved flavor by suppressing spoilage aldehydes and retaining fresh compounds.

PALA extended tilapia shelf life by 4 d via inhibiting microbial growth.

PALA delayed lipid/protein oxidation with lower TBARS and TVB-N.

PALA reduced MP carbonyls and surface hydrophobicity, preserving protein integrity.

PALA improved flavor by suppressing spoilage aldehydes and retaining fresh compounds.

Figures

The dynamic changes in pH, TBARS, TVB-N, and TVC of tilapia fillets during storage under different processing conditions. Different uppercase letters within the same storage time indicate significant differences (*P* < 0.05), while different lowercase letters within the same treatment denote significant differences (*P* < 0.05). CK, LA, PAW, and PALA represent control, lactic acid, plasma-activated water, and plasma-activated lactic acid treated fish fillets, respectively.Fig. 1

Changes in carbonyl (A), sulfhydryl groups (B), and surface hydrophobicity (C) of tilapia fillets during storage under different processing conditions. Different uppercase letters within the same storage time indicate significant differences (*P* < 0.05), while different lowercase letters within the same treatment denote significant differences (*P* < 0.05). CK, LA, PAW, and PALA represent control, lactic acid, plasma-activated water, and plasma-activated lactic acid treated fish fillets, respectively.Fig. 2

Comparative SDS-PAGE analysis of tilapia fillets treated with CK (A), LA (B), PAW (C), and PALA (D) during storage. CK, LA, PAW, and PALA represent control, lactic acid, plasma-activated water, and plasma-activated lactic acid treated fish fillets, respectively.Fig. 3

The pie chart of volatile organic compounds identified by GC/MS (A). The heat maps of volatile organic compounds detected in all different treatment groups of tilapia fillets identified in GC/MS (B). The correlation between physicochemical indicators and key flavor compounds (C). CK, LA, PAW, and PALA represent control, lactic acid, plasma-activated water, and plasma-activated lactic acid treated fish fillets, respectively. The dot size was proportional to the absolute value of the content. Red and blue colors indicate positive and negative correlations, respectively. (For interpretation of th

Schematic illustration of the proposed synergistic preservation mechanism of PALA treatment in tilapia fillets.Fig. 5

Full text

Introduction

1

The global seafood industry faces a dual challenge of significant post-harvest losses exceeding 20 % annually and growing consumer demand for natural, minimally processed products (International Organization, 1947). This disconnect underscores the urgent need for innovative preservation technologies that ensure safety and reduce waste (Yang, Zhang, Huang, et al., 2025). As a globally farmed staple, tilapia is particularly susceptible to rapid quality deterioration during refrigerated storage, driven by microbial proliferation, lipid oxidation, and protein degradation, which collectively compromise texture, color, flavor, and safety.

Non-thermal technologies have gained prominence, with cold plasma (CP) emerging as a promising solution due to its ability to generate reactive oxygen and nitrogen species (RONS) with potent antimicrobial and antioxidant properties (Mozzon et al., 2023). However, the inconsistent direct application of CP gases spurred the development of plasma-activated water (PAW) as a liquid-phase intervention (Zhao et al., 2025). While effective in model systems, PAW's practical application is constrained by the rapid decay of reactive species, dilution effects, and limited residual activity, hindering long-term shelf-life extension.

Consequently, research has evolved toward plasma-activated liquid agents (PALAs), where more complex, food-grade matrices enhance the stability and delivery of RONS (Oehmigen et al., 2010). A significant advance is the activation of organic acids like lactic acid (LA). This approach synergistically combines the antimicrobial effect of acidification with CP-derived oxidative stress, aligning with hurdle technology principles (Leistner, 2000; Oliveira et al., 2022). Preliminary studies indicate that PALA offers superior bactericidal efficacy compared to PAW or LA alone (Oliulla et al., 2024). Despite this promise, critical knowledge gaps persist regarding comprehensive quality preservation.

Despite this promise, critical knowledge gaps persist regarding comprehensive quality preservation. As summarized in Table 1, which compares key studies in this field, PAW treatment of chicken was effective for surface decontamination but its reactive species decay rapidly, with potential pro-oxidant effects on lipids and uncharacterized impacts on protein and flavor (Qian et al., 2022). For tilapia, PAW reduced surface Pseudomonas counts, yet its broader impact on product quality remained unevaluated (Liu et al., 2025). Most notably, a recent study applying PALA to tilapia fillets demonstrated a stronger initial bactericidal effect but did not extend the evaluation to systematic quality changes throughout storage (Cai et al., 2024). These works collectively highlight three critical, interconnected gaps: a lack of systematic multi-parametric evaluation that concurrently addresses microbial inhibition, lipid and protein oxidation, and volatile flavor dynamics throughout storage; insufficient direct quantitative evidence for synergy due to the absence of systematic comparisons between the combination treatment and all individual components (PAW and LA) under identical conditions; and a notable absence of studies correlating molecular-level oxidation markers with macro-scale quality deterioration and sensory-relevant volatile profiles to build a predictive, mechanism-based preservation model.Table 1. Comparative analysis of recent studies on plasma-activated solutions for meat and fish preservation.Table 1. Treatment agentTarget foodKey findingsIdentified limitations / Knowledge gapsReferencePAWChickenEffective surface decontamination; reduced specific pathogens.Reactive species decay rapidly; limited residual activity during storage; may promote lipid oxidation; effects on protein oxidation and flavor not fully characterized.(Qian et al., 2022)PALATilapia filletsSuperior bactericidal effect vs. PAW or LA alone; proposed acid-RONS synergy.Focused on initial microbial kill; comprehensive quality impact during storage not evaluated.(Cai et al., 2024)PAWTilapia filletsReduced the number of pseudomonas on the surface of fish slices.Clarified the bactericidal effect on bacteria; The impact on quality has not been evaluated.(Liu et al., 2025)This study: PALARefrigerated tilapia filletsSystematically evaluated microbial inhibition, lipid/protein oxidation, and volatile flavor dynamics; quantifies synergy via direct comparison with PAW and LA; correlated quality data.––

Therefore, to bridge these gaps, this study aimed to quantitatively elucidate the synergistic preservation mechanism of PALA through direct comparison with PAW and LA treatments across microbial, oxidative, and protein degradation; systematically evaluate its comprehensive effects on the physicochemical stability, protein integrity, and volatile flavor profile of refrigerated tilapia fillets; and establish mechanistic correlations between molecular-level oxidation events and macro-scale quality deterioration. We hypothesized that PALA would exhibit true synergy by outperforming additive effects in multi-target preservation, thereby providing a robust scientific basis for its industrial application and advancing the development of effective non-thermal strategies for aquatic product preservation.

Materials and methods

2

Materials and chemicals

2.1

Fresh tilapia (600–800 g) was obtained from the agricultural market in Haikou (Hainan, China). Fish were transported on ice (0–4 °C) to maintain freshness and processed within 2 h of procurement. Upon arrival, fish were immediately rinsed with sterile 0.90 % (w/v) physiological saline to remove surface impurities. Processing commenced within 2 h of procurement. Using sterilized utensils, fish were decapitated, eviscerated, skinned, and filleted. All muscle tissues were then rinsed three times with the aforementioned sterile saline and patted dry with sterile absorbent paper. The prepared fish fillets were uniformly cut into portions of 3.0 cm × 3.0 cm × 1.0 cm (approximately 10 ± 0.5 g per piece) for subsequent treatments. DL-LA (85–90 % purity) was purchased from YiEn Chemical Technology Co., Ltd. (Shanghai, China). All experimental procedures were conducted using analytical-grade reagents to ensure methodological consistency.

Preparation of plasma-activated solution

2.2

According to previous research (Cai, Wang, et al., 2022), a plasma-activated solution was prepared using an atmospheric-pressure plasma jet device (PG-1000ZD, Nanjing Suman Plasma Engineering Research Institute Co., Ltd., China). The operational parameters were set as follows: power output of 650 W, treatment duration of 90 s, solution volume of 150 mL, and a working gas flow rate of 0.9 L/min. The working gas used was ambient air, with the humidity controlled at approximately 50 % relative humidity (RH) and ambient temperature maintained at 25 °C throughout the treatment process. These experiments systematically evaluated power levels (550–800 W), treatment durations (30–180 s), and LA concentrations (0.0325–0.25 %) to identify the optimal condition that maximized antimicrobial efficacy while preserving fish quality. For PALA, a LA solution was diluted with sterile deionized water and added to achieve a final concentration of 0.0625 % (v/v) in the pre-treatment solution. PAW was prepared under identical conditions using only sterile deionized water. The resulting solutions were sealed in sterile beakers and stored at 4 °C for 1 h. All preparation steps were conducted under aseptic conditions to ensure reproducibility and avoid contamination.

Sample preparation and treatment

2.3

The pre-processed tilapia fish fillets (Section 2.1) were cut into portions weighing approximately 25 g each. The fish fillets were then soaked in PAW, LA, and PALA for 20 min. For the control group, fillets were treated identically using sterile distilled water (DW). After treatment, the samples were aseptically removed, placed on sterile stainless-steel grids, and air-dried for 30 min in a laminar flow biosafety cabinet (25 °C, ≤40 % RH) until no visible surface moisture remained. All treated fillets were then individually packaged in pre-sterilized polyethylene self-sealing bags under aseptic conditions. The packaged samples were stored in a dark, temperature-controlled refrigerator at 4 °C and analyzed at predetermined intervals over a 7 d storage period. Three independent biological replicates (n = 3), each consisting of three technical sample portions, were prepared for each treatment group and sampling time point.

Determination of physicochemical properties

2.4

Determination of pH

2.4.1

Sample (1 g) was mixed with 10 mL of potassium chloride solution and homogenized at 15,000 rpm for 2 min. The homogenate was allowed to stand at room temperature for 10 min to allow full equilibration of hydrogen ions. The pH was then measured using a pH meter (PHS-3C, Shanghai Yidian Scientific Instruments Co., Ltd., China).

Determination of color

2.4.2

The color changes (L*, a*, b*) of fish fillets (4 × 4 × 2 cm) were measured using a colorimeter (CR-410, Konica Minolta, Japan). Each fillet was measured at five different locations, and the average value was calculated to represent the color of the fillet.

Determination of texture

2.4.3

Fish samples were cut into uniform fillets (2 × 2 × 1 cm) and placed flat on the measurement platform of texture analyzer (TA.XT Plus, Stable Micro Systems, UK) (Yang, Zhang, Huang, et al., 2025). A flat-bottomed cylindrical probe (P/36R) was used under the following conditions: pre-test speed of 30 mm/min, test speed of 60 mm/min, and post-test speed of 60 mm/min. The samples were compressed to a compression ratio of 50 %, with a trigger force of 5 g. Key texture parameters, including hardness, cohesiveness, springiness, gumminess, and chewiness, were determined. After each measurement, the probe was cleaned to avoid cross-contamination. Between sample measurements, the probe and platform were thoroughly cleaned with 75 % ethanol and dried with a lint-free cloth to prevent carryover. Each fish fillet was analyzed in triplicate to ensure reproducibility.

Determination of lipid oxidation

2.4.4

Lipid oxidation was determined following the method of Wang et al. (2023) with minor modifications. Briefly, 5 g of sample was homogenized with 15 mL of 2-thiobarbituric acid reactive substances (TBARS) solution. The mixture was heated in a water bath at 90 °C for 30 min, followed by cooling to room temperature. Immediately after heating, tubes were cooled in an ice-water bath for 10 min to stop the reaction. Subsequently, the mixture was centrifuged at 5000 ×g for 10 min at 4 °C (Centrifuge 5424 R, Eppendorf, Germany). The supernatant was carefully filtered through a 0.45 μm nylon syringe filter. The absorbance of the resulting pink chromogen was measured at 532 nm against a reagent blank using a UV–Vis spectrophotometer (UV-1800PC, Mapada, China). TBARS values were expressed as mg malondialdehyde (MDA) equivalents per kg sample (mg MDA/kg).

Determination of total volatile basic nitrogen (TVB-N)

2.4.5

The TVB-N content was quantified using a fully automated Kjeldahl nitrogen analyzer (K9860, HaiNeng Future Technology Group Co., Ltd., China). Briefly, 2 g of fish was homogenized with 15 mL of sterile distilled water and allowed to stand for 30 min. The mixture was then alkalinized with 1 g of MgO in the distillation tube and distilled for 3 min. The released ammonia was collected in 30 mL of boric acid absorption solution (20 g/L) containing a mixed indicator (methyl red and bromocresol green in ethanol at 1:5). The absorbed solution was titrated with 0.01 M standardized HCl until the endpoint color changed from green to pink (Yang, Zhang, Wang, et al., 2025). TVB-N values were calculated and expressed as as mg/100 g fish.

Determination of microbiology

2.5

For microbial enumeration, 5 g aliquots of fish tissue were aseptically homogenized in 45 mL sterile water. The homogenate was then serially diluted (10-fold) in 9 mL of 0.9 % sterile physiological saline (sodium chloride, analytical grade). From appropriate dilution tubes, 150 μL aliquots were aseptically pipetted and spread evenly onto the surface of pre-dried Plate Count Agar (PCA) plates. Following 48 h incubation at 37 °C, colonies were counted and TVC values reported as Log CFU/g wet weight.

Determination of protein oxidation

2.6

Extraction of MP

2.6.1

The MP extraction was conducted according to the method of Lu et al. (2024) with modifications. Briefly, 5 g of fish muscle was finely chopped and homogenized (15,000 rpm, 1 min) with 50 mL of ice-cold extraction buffer (0.01 M PBS, pH 7.4, containing 5 mg/mL NaN₃). The buffer volume represents a 10:1 (v/w) ratio relative to the sample weight. The homogenate was centrifuged (5000 ×g, 10 min, 4 °C), and the resulting pellet was washed three times with 50 mL of the same buffer. Subsequently, the washed pellet was resuspended in 50 mL of 0.6 M KCl-PBS buffer (0.01 M PBS, pH 7.4, containing 0.6 M KCl). The suspension was placed on a roller mixer (SCILOGEX MX-T6-S) and extracted overnight (for 12 h) at 4 °C with gentle agitation set at 30 rpm. Following extraction, the suspension was centrifuged again (5000 ×g, 10 min, 4 °C). The final supernatant, containing the solubilized MP, was carefully collected, filtered through two layers of cheesecloth to remove any lipid or connective tissue debris, and stored on ice for immediate use or at −80 °C for subsequent analysis.

Measurement of carbonyl content

2.6.2

The protein carbonyl content using the method improved by Zhang et al. (2025). Specifically, MP extracts were prepared and adjusted to a precise concentration of 2 mg/mL using 0.01 M phosphate buffer (pH 7.4). Subsequently, the protein was precipitated by adding 0.6 mL of pre-chilled 20 % (w/v) trichloroacetic acid (TCA), followed by incubation on ice for 15 min and centrifugation (10,000 ×g, 10 min, 4 °C). The resulting precipitate was washed three times with 2 mL of ethanol: ethyl acetate (1:1, v/v) solution to remove unreacted DNPH. The protein precipitate was then dissolved in 2 mL of 6 mol/L guanidine hydrochloride solution and incubated at 37 °C for 20 min with shaking. After final centrifugation (10,000 ×g, 10 min, 4 °C), the supernatant absorbance was measured at 370 nm. A blank control was prepared by replacing the protein solution with 0.3 mL of 2 mol/L HCl and processed identically. All measurements were performed in triplicate for each sample group. The calculation was based on Eq. (1) below:

$$[eqn]$$

where A represents the absorbance at 370 nm, B represents the dilution factor, ρ is the protein concentration (mg/mL), D is the cuvette light path length (cm), and 22,000 is the molar absorption coefficient for protein carbonyls (L·mol^−1^·cm^−1^).

Measurement of total sulfhydryl groups

2.6.3

The total sulfhydryl content was determined as follows: A 0.5 mL aliquot of protein solution (2 mg/mL) was mixed with 4 mL of 0.05 M phosphate buffer (containing 0.6 M NaCl, 1 mM EDTA, and 8 M urea, pH 8.0). Subsequently, 0.5 mL of freshly prepared 10 mM 5,5′-dithiobis (2-nitrobenzoic acid) (DTNB) solution was added and thoroughly mixed. The reaction mixture was incubated in a water bath at 40 °C for 30 min, with all steps performed under light-protected conditions. The absorbance was measured at 412 nm using a spectrophotometer, and the thiol concentration was calculated using the molar extinction coefficient of 13,600 L·mol^−1^·cm^−1^ (Du et al., 2020). Calculations were conducted using Eq. (2):

$$[eqn]$$

Where A was the absorbance at 412 nm, D was the dilution ratio, which was the ratio of the mother liquor to the present concentration, C represents the concentration of MP (mg/mL); and ε was the molar absorbance coefficient (13,600 L·mol^−1^·cm^−1^).

Measurement of surface hydrophobicity

2.6.4

The surface hydrophobicity was determined using the bromophenol blue (BPB) binding method with modifications (Wang, Zhou, et al., 2021). 1 mL aliquot of protein solution (5 mg/mL) was mixed with 0.2 mL of BPB solution (1 mg/mL) and vortexed thoroughly. After 10 min of dark incubation, the mixture was centrifuged at 4500 rpm for 15 min at room temperature. A blank control was prepared by mixing 0.2 mL of the BPB solution with 1 mL of the phosphate buffer (pH 7.0) without protein, which was processed identically. For the blank control, 0.2 mL of BPB solution was mixed with 1 mL of phosphate buffer (pH 7.0) and processed identically. The supernatants from both experimental and control groups were then 10-fold diluted with deionized water, and the absorbance was measured at 595 nm. The surface hydrophobicity was calculated as follows:

$$[eqn]$$

Where 200 was the mass of bromophenol blue (μg), A_1_ was the absorbance of the sample, and A_0_ was the absorbance of the blank control.

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE)

2.7

Protein samples were adjusted to 1 mg/mL using PBS buffer (0.6 M KCl) and mixed with 4× loading buffer at a 4:1 (v/v) ratio (Lu et al., 2024). The mixtures were denatured by boiling for 5 min in a water bath. Subsequently, 10 μL of each sample, along with 4 μL of protein marker, were loaded onto a discontinuous polyacrylamide gel consisting of a 5 % stacking gel and 12 % separating gel. Electrophoresis was performed at constant voltage (80–120 V) until the dye front reached the gel bottom. Following electrophoresis, the gel was carefully removed from the cassette. Protein bands were visualized by staining with 0.1 % (w/v) Coomassie Brilliant Blue G-250 in a solution of 10 % (v/v) acetic acid and 40 % (v/v) methanol for 2 h with gentle orbital agitation at room temperature. Destaining was carried out using a solution of acetic acid: ethanol: water (2:3:5, v/v/v) until clear protein bands were visible against a transparent background. Gel images were captured using a digital imaging system for subsequent analysis.

Determination of nitrate and nitrite content

2.8

For nitrite quantification, 1 mL of filtrate was mixed with 2 mL of Griess reagent (containing 1 % sulfanilamide in 5 % phosphoric acid and 0.1 % N-(1-naphthyl)-ethylenediamine dihydrochloride) (Qian et al., 2020). After 20 min of dark incubation at room temperature, absorbance was measured at 540 nm. A standard calibration curve was established using NaNO₂ solutions at six concentrations ranging from 0.05 to 2.0 mg/L. The method detection limit (MDL) was determined to be 0.02 mg/L based on seven replicate analyses of the lowest standard. Nitrate analysis involved reduction through a cadmium column activated with 2 % HCl and equilibrated with 0.1 mol/L NH₄Cl solution. A 5 mL aliquot of filtrate was passed through the column at 1 mL/min flow rate to convert NO₃^−^ to NO₂^−^. Total NO₂^−^ was determined by the Griess method (St. Louis, MO, USA), using the same calibration procedure and detection limit as for nitrite quantification. Nitrate content was calculated by subtracting the original nitrite concentration from the total nitrite concentration obtained after reduction.

Determination of volatile organic compounds (VOCs)

2.9

The VOCs were analyzed according to the method of Wang et al. (2024) and improved using solid-phase microextraction (SPME) combined with gas chromatography–mass spectrometry (GC–MS). Approximately 3.0 g of minced fish sample was precisely weighed and transferred into a 15 mL headspace vial containing a magnetic stir bar. The vial was immediately sealed with a PTFE/silicone septum-cap (Supelco, Bellefonte, PA, USA) to prevent volatilization loss. All samples were prepared in triplicate from independent biological replicates. A 65 μm PDMS/DVB fiber was exposed 4 mm below the vial septum and extracted at 80 °C for 20 min, followed by thermal desorption at 250 °C for 5 min in the GC injection port. Separation was achieved on a DB-WAX capillary column (30 m × 0.25 mm, 0.25 μm) with helium as carrier gas. Mass spectral data were acquired in electron ionization (EI) mode at 70 eV, scanning m/z 29–550. Compounds were identified by comparing both mass spectra with NIST 14 database and Kovats retention indices (RI). Semi-quantitative analysis was performed using 2,4,6-trimethylpyridine (TMP) as internal standard.

Statistical analysis

2.10

Statistical analysis was conducted using SPSS version 27.0 (IBM Corp., Armonk, NY, USA). Normality and homogeneity of variance were first verified using the Shapiro-Wilk test and Levene's test. To evaluate significant differences among treatment groups, one-way analysis of variance (ANOVA) was applied, followed by Duncan's multiple range test for post-hoc comparisons. A probability value of P < 0.05 was considered statistically significant for all tests. All data shown are based on three independent biological replicates, with each replicate measured in triplicate (n = 3). Results are expressed as mean ± standard deviation (SD). Figures were generated using Origin 2022 (OriginLab, Northampton, MA, USA).

Results and discussion

3

Physicochemical properties changes in tilapia fillets

3.1

pH

3.1.1

The pH dynamics of tilapia fillets during storage were investigated across four treatment groups (CK, LA, PAW, PALA) in response to the acidification effect observed in plasma-activated liquids. All groups exhibited a characteristic biphasic trend: pH initially decreased, reaching minima by day 1 (CK: 6.18, LA: 6.22, PAW: 6.26, PALA: 6.21), and then progressively increased thereafter (Fig. 1A). The subtle but statistically significant differences in the initial pH minima among groups suggest distinct underlying drivers of early acidification. Early acidification was primarily driven by postmortem glycogenolysis and lactate accumulation, as supported by Yang, Zhang, Wang, et al. (2025). Exogenous LA addition contributed to the initial pH drop in the LA group, while PALA and PAW treatments induced more pronounced acidification through long-lived reactive nitrogen species (NO₃^−^/NO₂^−^) generated by CP, consistent with findings by Cai et al. (2024). The fact that PALA, despite containing LA, did not result in the lowest day-1 pH indicates a complex interplay where its active species might modify metabolic pathways or initial substrate availability. That PALA, despite containing LA, did not result in the lowest day-1 pH indicated a complex interplay where its active species might modify metabolic pathways or initial substrate availability. In the later storage stage, the pH rise was mainly attributed to alkaline compounds (ammonia and amines) produced by microbial and enzymatic proteolysis. The CK group showed the most significant pH increase, indicating severe spoilage. In contrast, both PAW and PALA treatments effectively inhibited microbial growth, thereby slowing protein degradation and reducing alkaline accumulation. Although the pH in PALA-treated samples remained above 6.0 throughout storage, this mild acidification had no detrimental impact on fillet quality, consistent with observations in PAW-treated Asian sea bass (Wang et al., 2023).Fig. 1. The dynamic changes in pH, TBARS, TVB-N, and TVC of tilapia fillets during storage under different processing conditions. Different uppercase letters within the same storage time indicate significant differences (P < 0.05), while different lowercase letters within the same treatment denote significant differences (P < 0.05). CK, LA, PAW, and PALA represent control, lactic acid, plasma-activated water, and plasma-activated lactic acid treated fish fillets, respectively.Fig. 1

Color

3.1.2

Color serves as a critical freshness indicator in fish products, directly influencing consumer acceptance. Prolonged storage resulted in a gradual decrease in L* values across all groups (Table 2), indicating loss of surface glossiness attributable to lipid and protein oxidation, myoglobin degradation, and microbial spoilage (Cai, Zhong, et al., 2022). The superior preservation of L* values in the PALA group likely resulted from synergistic actions: the acidic environment stabilized heme proteins while reactive species inhibited spoilage enzymes produced by Pseudomonas spp. Both PAW and PALA treatments significantly alleviated this decrease, as CP has been reported to maintain higher L* values by delaying oxidative color changes in fish such as golden carp (Gao et al., 2024). A similar declining trend was noted in a* values, most markedly in the CK group, which fell significantly from −2.22 to −4.84 (P < 0.05). This reduction was associated with myoglobin denaturation and metabolic activity of Pseudomonas aeruginosa (Zhao et al., 2024), a phenomenon also supported by studies on sea bass fillets where ROS-mediated myoglobin oxidation under PAW treatment reduced a* values (Olatunde et al., 2019). The observed increase in b* values during storage likely reflected the accumulation of secondary oxidation products from lipids and proteins, consistent with reports in tilapia (Yang, Zhang, Huang, et al., 2025). Overall, CP-based treatments effectively delayed color deterioration in refrigerated fish products. The superior color stability achieved by PALA over PAW indicated a multi-targeted mechanism involving acid-mediated pigment stabilization, targeted suppression of spoilage bacteria, and enhanced antioxidant activity, thereby providing a more comprehensive defense against the complex biochemical pathways of fish discoloration.Table 2. Color changes in tilapia fillets under different treatments during storage.Table 2. GroupStotage time (d)CKLAPAWPALAL065.95 ± 0.02^aB^65.91 ± 0.01^aA^66.83 ± 0.01^aC^66.93 ± 0.01^aD^163.70 ± 0.01^bA^64.09 ± 0.01^bB^65.58 ± 0.01^bC^65.92 ± 0.02^bD^360.87 ± 0.02^cA^61.50 ± 0.01^cB^63.43 ± 0.022^cC^64.36 ± 0.01^cD^555.43 ± 0.01^dA^57.84 ± 0.02^dB^60.49 ± 0.01^dC^61.32 ± 0.01^dD^752.32 ± 0.01^eA^55.24 ± 0.01^eB^58.81 ± 0.01^eC^59.26 ± 0.01^eD^a0−2.22 ± 0.01^aB^−2.25 ± 0.01^aC^−2.28 ± 0.01^aD^−2.13 ± 0.01^aA^1−3.00 ± 0.02^bC^−2.96 ± 0.01^bBC^−2.90 ± 0.01^bB^−2.69 ± 0.01^bA^3−3.82 ± 0.01^cD^−3.25 ± 0.01^cC^−3.10 ± 0.01^cB^−2.85 ± 0.01^cA^5−4.65 ± 0.01^dD^−3.72 ± 0.01^dC^−3.47 ± 0.01^dB^−3.04 ± 0.01^dA^7−4.84 ± 0.01^eD^−3.86 ± 0.01^eC^−3.59 ± 0.01^eB^−3.32 ± 0.01^eA^b*00.43 ± 0.01^cB^0.43 ± 0.01^eB^0.49 ± 0.01^eA^0.43 ± 0.01^eB^10.62 ± 0.18^cA^0.62 ± 0.01^dA^0.65 ± 0.01^dA^0.63 ± 0.01^dA^30.89 ± 0.01^bA^0.86 ± 0.01^cB^0.83 ± 0.01^cC^0.82 ± 0.0^1cC^50.93 ± 0.01^bA^0.92 ± 0.01^bA^0.90 ± 0.01^bB^0.89 ± 0.01^bB^71.32 ± 0.01^aA^1.11 ± 0.01^aB^1.08 ± 0.01^aC^0.98 ± 0.05^aD^Different uppercase letters indicate significant differences (P < 0.05) between sample groups with the same storage time, while different lowercase letters indicate significant differences (P < 0.05) within the same group with different storage times.

Texture properties

3.1.3

The texture of fish tissue is a critical quality attribute that significantly influences consumer preference, as moderate firmness is favored while softening indicates quality deterioration. Textural parameters including hardness, springiness, cohesiveness, gumminess, and chewiness exhibited significant progressive declines during storage (P < 0.05) (Table 3). These changes were mainly attributed to proteolytic degradation of the tissue microstructure and water loss or muscle relaxation induced by microbial activity (Yang et al., 2022). Among all groups, the CK exhibited the most pronounced reduction in hardness and elasticity. In contrast, all treatments (PALA, PAW, LA) effectively mitigated texture degradation, with the PALA group demonstrating the most effective preservation. The superior performance of PALA resulted from synergy rather than simple addition. While PAW inactivated surface microbes and LA induced mild protein denaturation, PALA's reactive species likely penetrated deeper due to LA-induced structural loosening. This protective effect could be ascribed to the synergistic actions of CP-derived reactive species and LA, which collectively suppressed protease activity and microbial proliferation, thereby reducing protein hydrolysis and moisture migration (Mahto et al., 2014). The improved hardness and springiness in CP-treated groups align with reports in Atlantic salmon, where protease inhibition helped maintain MP integrity (Zhu et al., 2023). PALA uniquely suppressed both microbial and muscle-derived proteases, a nuanced effect not observed with PAW or CP alone. Comparable texture stabilization in CP-treated Asian sea bass further demonstrated the broad-spectrum antibacterial efficacy of CP in delaying texture deterioration during refrigerated storage (Wang et al., 2023).Table 3. Texture changes in tilapia fillets under different treatments during storage.Table 3. GroupStorage time (d)01357Hardness (g)CK207.30 ± 0.49^aA^171.61 ± 0.39^bD^143.10 ± 0.30^cC^134.60 ± 0.21^dB^93.20 ± 0.20^eD^LA202.01 ± 0.31^aC^176.20 ± 0.56^bC^146.30 ± 0.16^cB^136.21 ± 0.91^dA^112.81 ± 0.29^eC^PAW198.70 ± 0.17^aD^177.92 ± 0.19^bB^146.11 ± 0.41^cB^134.12 ± 0.45^dB^116.80 ± 0.02^eB^PALA203.50 ± 0.25^aB^195.91 ± 0.29^bA^167.01 ± 0.25^cA^135.92 ± 0.23^dA^120.21 ± 0.42^eA^Springiness (mm)CK3.71 ± 0.03^aA^3.38 ± 0.01^bC^3.13 ± 0.01^cD^3.08 ± 0.01^dC^3.05 ± 0.01^dC^LA3.62 ± 0.04^aA^3.51 ± 0.02^bB^3.41 ± 0.01^cC^3.16 ± 0.08^dB^3.11 ± 0.02^eB^PAW3.67 ± 0.01^aA^3.51 ± 0.01^bB^3.48 ± 0.01^cB^3.24 ± 0.01^dB^3.11 ± 0.0.02^eB^PALA3.68 ± 0.01^aA^3.67 ± 0.02^aA^3.61 ± 0.01^bA^3.49 ± 0.02^cA^3.27 ± 0.01^dA^CohesivenessCK0.33 ± 0.01^aC^0.32 ± 0.01^bD^0.31 ± 0.01^cB^0.30 ± 0.02^dD^0.29 ± 0.01^eD^LA0.34 ± 0.01^aB^0.32 ± 0.01^bC^0.32 ± 0.03^bA^0.31 ± 0.01^cC^0.30 ± 0.01^dC^PAW0.34 ± 0.01^aB^0.34 ± 0.03^bB^0.31 ± 0.01^dB^0.32 ± 0.02^cB^0.31 ± 0.01^eB^PALA0.36 ± 0.01^aA^0.33 ± 0.01^bA^0.32 ± 0.01^cA^0.32 ± 0.01^cdA^0.32 ± 0.01^dA^Gumminess (g)CK71.96 ± 0.35^aA^61.48 ± 0.05^bB^52.60 ± 0.01^cD^48.23 ± 0.23^dA^36.85 ± 0.07^eD^LA71.50 ± 0.09^aA^58.38 ± 0.23^bD^55.40 ± 0.09^cB^48.67 ± 0.03^dA^48.20 ± 0.01^eA^PAW66.20 ± 0.09^aC^59.00 ± 0.35^bC^53.06 ± 0.14^cC^45.47 ± 0.25^dC^41.13 ± 0.12^eC^PALA68.72 ± 0.02^aB^63.30 ± 0.05^bA^56.70 ± 0.06^cA^47.47 ± 0.38^dB^44.37 ± 0.07^eB^Chewiness (mJ)CK2.30 ± 0.01^aC^1.85 ± 0.02^bC^1.78 ± 0.01^cC^1.50 ± 0.01^dC^1.18 ± 0.01^eD^LA2.33 ± 0.01^aB^2.01 ± 0.02^bB^1.84 ± 0.01^cB^1.60 ± 0.02^dB^1.25 ± 0.02^eC^PAW2.38 ± 0.03^aA^2.05 ± 0.01^aB^1.71 ± 0.01^bD^1.60 ± 0.01^bB^1.30 ± 0.01^bB^PALA2.38 ± 0.01^bA^2.94 ± 0.01^aA^2.13 ± 0.01^cA^1.73 ± 0.01^dA^1.43 ± 0.01^eA^Different uppercase letters indicate significant differences between sample groups (P < 0.05), while different lowercase letters indicate significant differences within the same group for different storage times (P < 0.05).

TBARS

3.1.4

As a key indicator of secondary lipid oxidation, TBARS accumulation is widely recognized as a critical factor in meat quality deterioration. The PAW-treated samples exhibited significantly higher initial TBARS values (0.24 mg/kg) compared to the CK, LA, and PALA groups (P < 0.05), which showed similar initial levels (∼0.20 mg/kg) (Fig. 1B). This immediate elevation confirmed that RONS generated during plasma treatment directly induce lipid oxidation upon initial contact, underscoring an intrinsic trade-off between the immediate antimicrobial action of PAW and its pro-oxidant effect. Although all groups experienced significant TBARS increases throughout storage (P < 0.05), the PALA-treated fish consistently demonstrated the lowest levels among all groups. This inhibitory effect was attributed to the dynamic redox chemistry of PALA, where subsequent radical scavenging and lipid stabilization collectively shift the redox balance toward protection, counteracting the initial pro-oxidant tendency of RONS (Wang et al., 2023). Notably, an initial TBARS elevation was observed in both PAW- and PALA-treated samples during early storage (0–1 d). The increase in the PAW group was driven by RONS that promote primary oxidation (Liu et al., 2022), whereas the transient TBARS increase following PALA treatment was linked to NO• activity, which liberated free iron from myoglobin-bound iron. This free iron served as a known catalyst for lipid oxidation (Qian et al., 2020). However, NO• subsequently scavenged hydroxyl radicals (OH•) via nitrite formation, thereby attenuating peroxidation (Wang, Zhou, et al., 2021). While PAW's RONS drive continuous oxidation, PALA's specific species establish a self-limiting redox cycle, conferring superior oxidative stability. Previous studies established that TBARS levels exceeding 0.60 mg/kg would compromise sensory and flavor profiles (Li et al., 2023). Overall, PALA treatment effectively inhibited lipid oxidation and delayed spoilage, maintaining TBARS within acceptable limited throughout refrigerated storage.

TVB-N

3.1.5

TVB-N serves as a critical indicator of fish freshness, primarily resulting from microbial and enzymatic degradation of proteins into volatile nitrogenous compounds such as ammonia, dimethylamine, trimethylamine, putrescine, and cadaverine (Zhu et al., 2023). Established industry standards define TVB-N values below 15 mg/100 g as premium-quality fish, with 30 mg/100 g as the maximum acceptable limit (Cheng et al., 2016). All three treatments significantly inhibited TVB-N accumulation compared to the CK group (P < 0.05), with the synergistic PALA treatment showing the most pronounced suppression (Fig. 1C). By day 7 of storage, TVB-N values reached 37.13, 20.30, 23.41, and 15.26 mg/100 g in the CK, LA, PAW, and PALA groups. PALA's superior efficacy stemmed from a multi-target mechanism: the synergistic antimicrobial action of plasma-generated reactive species and LA suppressed proteolytic spoilage bacteria, the resulting acidic environment inhibited endogenous fish muscle proteases and decarboxylases, and the overall disruption of microbial metabolism curtailed the generation of volatile nitrogenous bases. In contrast to reports where single dielectric barrier discharge (DBD) treatment increased TVB-N in chicken breast (Qian et al., 2024), the PALA combination effectively suppressed its generation, likely because synergistic components mitigated protein damage and reduced spoilage substrates. The observed TVB-N reduction aligned with the suppressed pH rise during storage, as alkalinization from amine accumulation was minimized. Collectively, PALA preserved freshness by dual mechanisms: inhibiting spoilage microbes and blocking proteolytic and amino acid catabolic pathways.

Microbial analysis

3.2

Microbial growth is a critical indicator of freshness deterioration in refrigerated fish products, with a total viable count (TVC) ≥ 7.00 Log CFU/g representing the accepted spoilage threshold that signifies imminent deterioration (Cai, Zhong, et al., 2022). As shown in Fig. 1D, TVC increased progressively across all groups during storage, but samples treated with PALA consistently exhibited the lowest microbial counts. The CK group rapidly exceeded the microbial limit after 3 d, aligning with the typical growth curve of endogenous spoilage flora in an unrestricted environment. LA and PAW treatments surpassed the threshold by day 5, demonstrating their limited and transient inhibitory effects. In contrast, PALA treatment effectively suppressed microbial growth to 6.41 Log CFU/g even after 7 d, extending the microbial shelf-life by 4 d compared to conventional treatments. This pronounced and sustained antimicrobial effect could be mechanistically deconstructed into a synergistic, multi-stage process that went beyond simple additive effects. The initial LA-induced pH reduction created a potent acidic stress that specifically disrupted transmembrane proton gradients and inhibited key metabolic enzymes in bacteria, weakening their homeostatic capacity (Wu et al., 2023). This acidic matrix within PALA also played a crucial stabilizing role, significantly prolonging the half-life of RONS such as NO₃^−^, ·OH, and ONOO^−^ (Liu et al., 2023). The extended persistence of these radicals enabled a continuous, low-level oxidative assault rather than a single burst, which prevented microbial adaptation. Most critically, the data suggest a targeted action of these stabilized RONS, likely peroxidizing unsaturated membrane lipids and cross-linking peptidoglycan to disrupt membrane integrity, increase permeability, and lead to the observed leakage of proteins and nucleic acids (Jyung et al., 2023). The observed suppression of microbial growth was consistent with the reduced TVB-N levels, further confirming the preservation efficacy of PALA. These findings align with those of Cai et al. (2024), who similarly reported a 4 d shelf-life extension in PALA-treated fish samples. All other treatment groups exceeded the spoilage threshold of 7.00 Log CFU/g by 7 d, indicating advanced deterioration. The eventual microbial recovery in these groups might be attributed to the depletion of antibacterial agents and the availability of nutrients derived from ongoing protein and lipid degradation, which ultimately accelerate spoilage (Chanioti et al., 2023). The outstanding performance of PALA underscored its potential as a novel and effective preservation strategy for aquatic products.

Protein oxidation analysis

3.3

Changes in carbonyl content

3.3.1

Protein carbonyl groups, derived from oxidized amino acid side chains, serve as sensitive biomarkers for assessing protein oxidative damage in aquatic products, reflecting modifications to both side-chains and the polypeptide backbone (Yu et al., 2023). These carbonyls were primarily generated through the oxidation of susceptible residues including proline, threonine, lysine, and arginine (Rahim & Abdullah, 2021). As shown in Fig. 2A, carbonyl content increased significantly across all treatment groups during storage (P < 0.05), indicating progressive oxidative deterioration of MP. This increase correlated strongly with rising TBARS values, indicating linked lipid-protein oxidation pathways in which lipid-derived secondary products covalently bind to protein side chains, forming carbonyl adducts and amplifying oxidative damage. Notably, PAW treatment initially accelerated protein oxidation, resulting in significantly higher carbonyl levels than the CK, LA, and PALA groups in the early storage phase. This was likely attributed to its reactive species overwhelming the cellular antioxidant defenses, a phenomenon consistent with PAW-induced oxidation reported in golden pompano (Gao et al., 2024). In contrast, PALA treatment effectively suppressed MP oxidation by concurrently inhibiting lipid oxidation product formation and protecting vulnerable amino acid residues and peptide bonds from oxidative attack.Fig. 2. Changes in carbonyl (A), sulfhydryl groups (B), and surface hydrophobicity (C) of tilapia fillets during storage under different processing conditions. Different uppercase letters within the same storage time indicate significant differences (P < 0.05), while different lowercase letters within the same treatment denote significant differences (P < 0.05). CK, LA, PAW, and PALA represent control, lactic acid, plasma-activated water, and plasma-activated lactic acid treated fish fillets, respectively.Fig. 2

Changes in total sulfhydryl groups

3.3.2

Sulfhydryl groups in MP serve as critical markers of myosin structural integrity, as protein oxidation primarily involves sulfhydryl reduction and disulfide bond formation (Zhang et al., 2022). All treatment groups exhibited a continuous decrease in total sulfhydryl content throughout storage (Fig. 2B). Both PAW and PALA treatments resulted in higher initial (0 d) sulfhydryl content than the LA and CK groups. The PAW group maintained significantly higher sulfhydryl levels than the CK group (P < 0.05), possibly because ROS/RNS preferentially target free amino groups rather than sulfhydryl groups (Chaijan et al., 2021). This targeting suggested that PAW's antimicrobial effect delayed the onset of spoilage-associated oxidative stress. PAW-treated samples still retained more sulfhydryl groups than the CK group (P < 0.05) by day 3, indicating a sustained protective effect against oxidation during early-to-mid storage. (Gao et al., 2024). The final total sulfhydryl contents were 38.07 nmol/mg (CK), 42.55 nmol/mg (LA), 47.62 nmol/mg (PAW), and 58.19 nmol/mg (PALA), corresponding to reductions from day 0 of 44.32 %, 39.23 %, 27.40 %, and 21.78 %, respectively, by day 7 of storage. These results demonstrate that PALA treatment most effectively preserved MP sulfhydryl groups during refrigerated storage.

Changes in surface hydrophobicity

3.3.3

The conformational changes of MP were elucidated by evaluating surface hydrophobicity dynamics, a sensitive indicator of hydrophobic group distribution. All treatment groups exhibited a significant increase in surface hydrophobicity over the storage period (P < 0.05) (Fig. 2C). Initial values were approximately 21.00 for all groups, rising to 58.21, 52.02, 51.86, and 46.70 in the CK, LA, PAW, and PALA groups, respectively, by day 7. The PALA group demonstrated a significantly slower rate of increase (P < 0.05), strongly suggesting that this treatment uniquely mitigated MP unfolding and the exposure of internal hydrophobic regions. This effect might result from the combined action of the acidic environment from LA altering protein charge states and plasma-derived reactive species modulating MP surface properties, a synergy not achieved by either component alone. Reactive nitrogen species from CP could modulate protein conformation (Cheng et al., 2022), whereas CP alone typically promoted MP unfolding and aggregation (Shakeri et al., 2023). Although specific structural details remain uncharacterized, the significantly reduced surface hydrophobicity provided direct functional evidence that PALA treatment modified the MP surface to favor dispersion and inhibit aggregation.

SDS-PAGE

3.4

SDS-PAGE analysis assessed the influence of different treatments on the molecular weight distribution of MP in tilapia fillets. The electrophoretic profiles exhibited characteristic bands corresponding to myosin heavy chain (MHC, 250 kDa), actin (45 kDa), tropomyosin (TM, 30–40 kDa), and myosin light chain (MLC, 15–20 kDa) (Lu et al., 2024). As demonstrated in Fig. 3, MHC and actin displayed the most intense staining across all groups, reflecting their abundance in the MP fraction. No significant differences in MHC band intensity were observed initially (P > 0.05) between CP-treated (PAW/PALA) and non-CP groups (CK/LA), indicating that plasma-generated reactive species did not induce substantial primary peptide backbone cleavage. This critical observation indicated that plasma treatment, under the applied conditions, did not act as a non-specific protein denaturant but exerted its protective effect through alternative pathways (Ekezie et al., 2018).Fig. 3. Comparative SDS-PAGE analysis of tilapia fillets treated with CK (A), LA (B), PAW (C), and PALA (D) during storage. CK, LA, PAW, and PALA represent control, lactic acid, plasma-activated water, and plasma-activated lactic acid treated fish fillets, respectively.Fig. 3

Progressive degradation of MP was observed in all groups throughout refrigerated storage. By day 3, broadening of MHC bands indicated the accumulation of modified protein isoforms, suggesting non-enzymatic post-translational modifications that altered electrophoretic mobility without causing complete fragmentation (Yu et al., 2024). Both PAW and PALA treatments retained significantly higher MHC band intensity than CK and LA (P < 0.05), demonstrating CP's ability to mitigate myosin degradation and suppress aggregation, likely through radical scavenging or oxidative microenvironment modification. Actin bands in the CK and LA groups showed more pronounced broadening relative to CP-treated samples, implying enhanced proteolysis of actin. This differential degradation between myosin and actin supported previous observations that these proteins were primary targets of oxidative modification during chilled storage (Koddy et al., 2020). The attenuated protein degradation in CP-treated samples was likely attributable to suppressed microbial protease activity and inhibition of cross-linking between protein fragments (Situ et al., 2022). Collectively, these results demonstrate that CP treatment effectively preserved MP integrity and retarded proteolytic deterioration in refrigerated tilapia fillets.

Changes in nitrate and nitrite content

3.5

CP treatment increased nitrate and nitrite content in tilapia fillets (Table 4). Initial nitrate concentrations reached 73.29 mg/kg for PAW and 82.35 mg/kg for PALA, with corresponding nitrite levels of 1.58 mg/kg and 0.92 mg/kg. Nitrate concentrations decreased significantly during storage from 73.29 to 2.92 mg/kg (PAW) and from 82.35 to 1.84 mg/kg (PALA), attributable to metabolic conversion by nitrate-reducing bacteria (Qian et al., 2020). Concurrently, nitrite levels declined from initial values to below the detection limit and remained well below the 30 mg/kg safety threshold for meat products (Nader et al., 2021). This reduction reflects two synergistic mechanisms: (i) the higher activity of bacterial nitrite reductase relative to nitrate-reducing metabolism, facilitating nitrite conversion to NO (Kapil et al., 2020), and (ii) the intrinsic antimicrobial properties of nitrite that suppress spoilage microorganisms. These processes collectively underpin the efficacy of PALA in enhancing fillet preservation and extending shelf life. The residual nitrite levels in PALA-treated fillets remained consistently within regulatory limits, ensuring product safety and preserved sensory quality.Table 4. Changes in nitrate and nitrite levels in tilapia fillets under different treatments during storage.Table 4. GroupCKLAPAWPALANO_2_^−^NO_3_^−^NO_2_^−^NO_3_^−^NO_2_^−^NO_3_^−^NO_2_^−^NO_3_^−^0 d––––0.92^aB^73.29^aB^1.58^aA^82.35^aA^1 d––––0.47^bB^36.56^bB^1.41^bA^65.99^bA^3 d–––––2.93^cB^0.40^cA^3.99^cA^5 d–––––2.92^cB^–3.39^cA^7 d–––––2.92^cB^–1.84^dA^Different uppercase letters indicate significant differences (P < 0.05) between sample groups with the same storage time, while different lowercase letters indicate significant differences (P < 0.05) within the same group with different storage times.“-”, Not detected.

VOCs analysis

3.6

This study investigated the VOCs in tilapia fillets over 7 days of refrigerated storage using GC–MS, identifying a total of 63 VOCs, including aldehydes, esters, alcohols, ketones, acids, benzodiazepines, heterocyclic compounds, and hydrocarbons (Fig. 4A). Significant dynamic changes in VOC concentrations throughout storage indicated that refrigeration temperature substantially influenced the volatile profile of tilapia.Fig. 4. The pie chart of volatile organic compounds identified by GC/MS (A). The heat maps of volatile organic compounds detected in all different treatment groups of tilapia fillets identified in GC/MS (B). The correlation between physicochemical indicators and key flavor compounds (C). CK, LA, PAW, and PALA represent control, lactic acid, plasma-activated water, and plasma-activated lactic acid treated fish fillets, respectively. The dot size was proportional to the absolute value of the content. Red and blue colors indicate positive and negative correlations, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)Fig. 4

Among these, 11 aldehydes were identified as key degradation products derived from lipid oxidation and protein breakdown (Fig. 4B). While fresh samples contained only trace levels of heptanal, nonanal, and decanal, both the diversity and abundance of aldehydes increased significantly during storage (Benet et al., 2014). The content of heptanal and nonanal by oxidation products of oleic acid and linoleic acid, exhibited particularly pronounced changes. Previous studies have established heptanal as a reliable biomarker for fish flavor deterioration, while nonanal specifically interacts with tryptophan residues in MP, suggesting a direct molecular mechanism for odorant-protein interactions in aroma perception (Karbsri et al., 2024). Hexanal, nonanal, octanal, and 3-octanal were identified as major contributors to the characteristic fishy odor of tilapia. The PALA group exhibited significantly lower heptanal and nonanal levels compared to LA and PAW treatments by day 5, while retaining higher contents of fresh-flavor aldehydes like octanal and dodecanal, demonstrating effective flavor modulation.

Although possessing low flavor thresholds, alcohols contribute delicate aromas and oily notes that enhance fish flavor (Fang et al., 2022). 9 volatile alcohols were detected during storage, with 1-octen-3-ol (mushroom-like odor) and 1-hexanol showing the most notable variations, primarily arising from the oxidative degradation of oleic and palmitic acid (Kunyaboon et al., 2021). The strong positive correlation between 1-octen-3-ol and TBARS values underscores its link to lipid oxidation (Sang et al., 2024). PALA treatment consistently suppressed the accumulation of this compound, likely due to the scavenging of alkoxy radicals by plasma-generated reactive species and acidification-driven suppression of hydroperoxide lyase activity.

Despite their high odor thresholds, ketones help neutralize fishy odors (Zhao et al., 2021). 2-Heptanone, derived from linoleic acid oxidation, was among the ketones detected. The PALA group exhibited moderate ketone accumulation in later storage without excessive oxidative degradation. This implied that PALA might direct alkoxy radical β-scission toward ketone over aldehyde formation, a pathway modulated by its protic environment and reactive nitrogen species that stabilize specific radical intermediates.

Hydrocarbon compounds demonstrated relatively low abundance and high flavor thresholds, contributing negligibly to the overall flavor profile of tilapia. These compounds primarily originate from homolytic cleavage of fatty acid alkoxy radicals (Moretti et al., 2016). All experimental groups during storage correlated with lipid oxidation-mediated degradation of long-chain alkanes into smaller molecular species, indicating that hydrocarbons serve as transient intermediates. PALA likely accelerated their secondary oxidation into more polar volatiles, thereby altering the final VOC profile.

Acidic substances in fish primarily originated from protein hydrolysis and lipid oxidation during storage, with their content progressively increasing over time. Residual LA (2-hydroxypropanoic acid) was detected in all experimental groups, showing a gradual decline during refrigerated storage. Its progressive decrease, coupled with the concurrent increase in esters, implied an acid-catalyzed esterification reaction during storage. PALA likely enhanced this process through its combined acidic and oxidative environment, promoting the formation of ethyl formate and other esters that mask off-flavors, effectively reducing sensory thresholds while enhancing fruity and sweet aromatic notes (Qiu et al., 2024).

A total of 13 esters were identified, imparting fruity and creamy aromas that mask off-odors and bitterness. The PALA group exhibited a significantly higher content of characteristic esters (ethyl formate) compared to the control. This increase was mechanistically linked to plasma-induced oxidative decarboxylation of organic acids and subsequent esterification with ethanol produced by microbial activity or lipid oxidation. The enhanced ester formation in the PALA group directly demonstrated its role in redirecting metabolic and oxidative pathways toward favorable flavor compounds. Beyond odor-masking effects, these esters exhibit synergistic interactions with residual LA, collectively enhancing the fruity notes of the product (Wang et al., 2022).

PALA treatment also effectively suppressed the formation of benzene-derived and nitrogen-containing heterocyclic compounds associated with Maillard reaction and spoilage (Zheng et al., 2023). This suppression likely resulted from multiple mechanisms: the antimicrobial action of PALA limited precursor availability through reduced microbial deamination; its acidic environment inhibited Strecker degradation; and plasma-generated reactive species scavenged key intermediates. These findings further demonstrate that PALA actively redirected biochemical pathways away from spoilage-related volatiles toward fresher or neutral aroma compounds, substantiating its efficacy in flavor preservation.

Correlation analysis

3.7

To elucidate the molecular mechanism of changes in the quality of refrigerated tilapia fillets, multivariate statistical analysis examined the correlation between quality parameters and flavor compounds (Fig. 4C). Pearson correlation analysis revealed significant associations between the L* value and multiple quality parameters, including hardness, TBARS, TVB-N, TVC, carbonyl content, sulfhydryl content, and surface hydrophobicity (P < 0.05). This pattern positions L* not merely as a colorimetric index, but as a sensitive, integrative proxy for underlying deterioration pathways. Specifically, the negative correlation with surface hydrophobicity and carbonyl content indicated that protein unfolding and oxidation expose hydrophobic residues and introduce carbonyl groups, thereby directly reducing light scattering and lowering L* values. Conversely, the positive link with hardness and sulfhydryl content suggested that structural integrity and native protein conformation contribute to higher lightness. Therefore, the preservation of L* in the PALA group reflected its concurrent ability to inhibit protein oxidation, maintain myofibrillar structure, and suppress microbial and lipid-derived deteriorative reactions. Key aldehydes (heptanal, hexanal, nonanal) and ketones (2-heptanone) showed strong inter-correlations, confirming their shared origin in the oxidation of n-3/n-6 polyunsaturated fatty acids and suggesting a cascade where their decomposition products might further exacerbate oxidation and off-flavor. These patterns reflect the progressive biochemical deterioration in refrigerated tilapia: protein oxidation led to myofibrillar disruption and texture loss; microbial growth promoted protein degradation; and lipid oxidation products accumulate, reducing sensory quality. The PALA group exhibited more stable trends in all quality parameters, demonstrating its comprehensive preservation effects.

Collaborative preservation mechanism of PALA treatment

3.8

The superior preservation efficacy of the PALA combination, compared to individual LA or PAW treatments, stemmed from a multi-target synergistic mechanism across chemical, microbiological, and biochemical levels (Fig. 5). The acidic environment generated by LA served as the foundation for this synergy, directly lowering fillet surface pH to inhibit spoilage microorganism proliferation while also stabilizing PAW-derived CP-generated RONS to prolong antimicrobial activity. This critical synergistic antibacterial action severely damaged bacterial cell integrity through membrane lipid peroxidation and peptidoglycan degradation, culminating in cellular content leakage and cell death. Effective microbial suppression consequently retarded spoilage biochemistry, with the strong negative correlation between TVC and the key fresh odorant uniquely observed in the PALA group directly linking microbial inhibition to flavor preservation. PALA treatment also played a key role in modulating oxidative processes. RNS specifically scavenged highly reactive ·OH radicals to interrupt radical chain reactions, mitigating lipid peroxidation as reflected in lower TBARS values. Concurrently, RONS facilitated the structural rearrangement of MP, promoting hydrophobic group burial to inhibit excessive protein aggregation and surface hydrophobicity increase, thereby aiding in protein integrity maintenance, texture improvement, and carbonyl reduction. Ultimately, the coordinated inhibition of microbial activity and oxidative pathways by PALA significantly delayed lipid and protein decomposition. This process effectively slowed color deterioration and texture softening while minimizing off-odor generation and promoting pleasant aroma compound retention. Overall, PALA achieved comprehensive preservation through a tripartite synergy of proton sensitization, targeted RONS biochemistry, and redirected oxidative pathways, maintaining physicochemical, microbiological, and sensory qualities to extend shelf life.Fig. 5. Schematic illustration of the proposed synergistic preservation mechanism of PALA treatment in tilapia fillets.Fig. 5

Stability and performance under real food conditions

3.9

The superior preservation efficacy of PALA under controlled laboratory conditions provided a strong foundation, yet its industrial-scale application required careful consideration of stability and performance in real-world scenarios. Key challenges included the inherent instability of reactive species with short half-lives in complex organic matrices like fish exudates, which could reduce effective antimicrobial concentration and compromise treatment consistency on heterogeneous surfaces. Additionally, the protein-rich tissue could buffer PALA's acidic component, attenuating the crucial proton-mediated sensitization, particularly in thicker cuts. Variable industrial conditions such as temperature fluctuations during distribution also threatened residual activity by accelerating microbial recovery and active species decay, while scaling up posed engineering hurdles in maintaining uniform treatment coverage and potency in continuous processing. Despite these limitations, PALA's hurdle-based design offered distinct advantages through its ability to leave safe, food-grade residues that provided sustained antimicrobial and antioxidant effects during storage, a valuable asset for supply chain logistics. Future research should therefore prioritize pilot-scale validation across diverse fish species, development of stabilization strategies for reactive components, and comprehensive cost-benefit analyses to transition PALA from a promising laboratory innovation into a commercially viable tool for sustainable seafood preservation.

Conclusion

4

This study established PALA as a novel, eco-friendly preservation technology that effectively extended the shelf life of refrigerated tilapia fillets. The core mechanism involved synergistic acid-induced membrane sensitization and targeted oxidative damage by reactive nitrogen species, which collectively inhibited microbial growth and decelerated lipid and protein oxidation. This multi-target action was evidenced by the maintained physicochemical quality, suppressed spoilage volatiles, and preserved protein integrity in PALA-treated samples. PALA achieved a 4 day shelf-life extension using food-grade components, with safe nitrate/nitrite residues, offering a commercially viable, additive-free solution for the seafood industry. From a societal and environmental perspective, PALA implementation could reduce post-harvest losses, enhance food security, and improve resource efficiency. By potentially replacing energy-intensive or chemically based methods, it might lower the carbon footprint of aquatic product storage and align with cleaner production principles. A current limitation is the need to validate efficacy across diverse species and industrial conditions, including a full lifecycle assessment. Future research should prioritize scaling the technology, optimizing energy and cost parameters, and conducting comprehensive real-world evaluations of safety, sensory quality, and environmental impact to realize its full sustainable potential.

CRediT authorship contribution statement

Tingting Yang: Writing – original draft, Formal analysis. Wentao Deng: Methodology, Formal analysis. Li Liu: Visualization. Guanghua Xia: Writing – review & editing. Liming Zhang: Writing – review & editing, Funding acquisition. Jiamei Wang: Supervision, Project administration.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.