High-pressure processing mitigates flavor loss in sauced duck legs during storage: Evidence from volatile profiling and myofibrillar protein–flavor compound interactions

Abstract

Maintaining flavor stability is a major quality concern for ready-to-eat meat products. This study evaluated the effects of high-pressure processing (HPP) at 400 MPa on the volatile profile of sauced duck legs during 14 days of 4 °C storage and explored the role of myofibrillar protein (MP)–flavor interactions. Results show that the volatiles of sauced duck legs were significantly affected by HPP and the total aldehyde concentration of HPP treated sample was lower on day 14. Anethole, linalool, and 1-octen-3-ol were selected as representative compounds. In an extracted-MP model, HPP reduced MP α-helix content and increased the apparent affinity coefficients and mass-normalized interaction capacities for all three compounds. Molecular docking and exploratory molecular dynamics simulations suggested HPP-dependent conformational changes and plausible protein–volatile contacts. These findings indicate that partial MP unfolding and enhanced apparent MP–flavor interactions may contribute to flavor retention after HPP.

Highlights

• •400 MPa HPP lowered total aldehydes in sauced duck legs on storage day 14.
• •HPP increased the retention of all three flavor compounds by extracted MP.
• •HPP reduced extracted-MP α-helix content from 28.6% to 18.1%.
• •HPP increased apparent interaction parameters for all three flavor compounds.

1. Introduction

With the growing consumer demand for convenient and diverse food options, ready-to-eat foods from a wide range of raw materials, including livestock, poultry, aquatic products, and grains, have developed rapidly (Huang et al., 2023). Among these products, ready-to-eat meat products, particularly those derived from livestock and poultry, are an important category. Spices, such as star anise, fennel, and basil, are commonly incorporated in meat products during processing to impart characteristic aromas and flavors (Melo et al., 2021; Murphy et al., 2024). However, short shelf life and flavor loss during storage remain two major challenges in the development of ready-to-eat meat products (Shimizu & Iwamoto, 2022). To address microbial spoilage and multiple quality changes during storage, natural preservatives, non-thermal processing technologies, and active packaging have attracted increasing attention as alternatives to conventional preservation approaches. Among these strategies, non-thermal processing technologies can reduce heat-induced quality deterioration while extending shelf life and modulating flavor characteristics (Ursachi et al., 2020).

High-pressure processing (HPP) is a non-thermal processing technology that can extend shelf life, reduce nutrient loss, and maintain food quality. Previous studies have mainly focused on its effects on microbial safety, shelf-life extension, and overall product quality. For example, Trejo et al. (2021) reported that HPP improved the safety of Spanish-style salami stored at different temperatures. Pimsannil et al. (2026) found that HPP combined with a chitooligosaccharide–catechin conjugate maintained the quality of precooked blue swimming crab. The reports of Liu et al. (2026) showed that HPP at 300–400 MPa improved the quality of Procambarus clarkii and reduced off-flavor volatiles. Our previous study also demonstrated that HPP at 400 MPa significantly extended the shelf life of sauced duck legs (Zhang et al., 2025). However, the effects of HPP on the changes in desirable and undesirable volatile compounds in sauced duck legs during storage remain unclear.

Flavor is an important quality attribute of meat products and directly affects consumer acceptance. Flavor loss can be characterized by an unfavorable shift in the volatile profile, including the reduction of desirable aroma compounds or the accumulation of undesirable storage-related volatiles. Such changes during storage represent a major form of quality deterioration in ready-to-eat meat products (Montero et al., 2021). Several strategies have been used to improve flavor stability, including the optimization of seasoning formulations, antioxidant systems, thermal processing conditions, packaging and storage conditions, and physical processing technologies such as ultrasound and HPP (Carrapiso et al., 2022; Sun et al., 2023; Zhang et al., 2025). These strategies may affect enzymatic reactions, protein and lipid oxidation, volatile compound retention, and protein–flavor interactions. Among these factors, protein–flavor interactions play an important role in the release and retention of volatile compounds (Ayed et al., 2014, Huang, Wang, Feng and Kan, 2022, Wang et al., 2022). Previous studies have shown that HPP can alter the conformation and functional properties of meat proteins. As the major structural proteins in muscle, myofibrillar proteins (MP) can interact with volatile flavor compounds and thereby influence their retention and release (Yang et al., 2017). However, it remains unclear whether HPP-induced conformational changes in MP contribute to the retention of key flavor compounds in ready-to-eat meat products during storage. Clarifying this relationship may provide a mechanistic basis for improving the flavor stability of HPP-treated meat products. Representative volatile compounds were systematically selected by combining multivariate importance, treatment-related statistical differences, and their relevance to the characteristic flavor of sauced duck legs (Wu et al., 2026).

Therefore, this study aimed to investigate the effects of HPP at 400 MPa on the volatile flavor profile of sauced duck legs during storage and to explore a potential mechanism underlying flavor retention. Gas chromatography–mass spectrometry (GC–MS) and multivariate analysis were first used to characterize the changes in volatile compounds and identify the representative flavor compounds in the whole-product system. MPs were then extracted from sauced duck legs and examined using circular dichroism, UV–visible absorption spectroscopy, and fluorescence spectroscopy to evaluate treatment-related structural differences and protein–flavor interactions. Apparent mass-normalized interaction parameters were used to characterize interactions in the extracted-MP model system. Molecular dynamics simulations and molecular docking were used as exploratory analyses of pressure-dependent conformational behavior and plausible protein–volatile contacts. These findings may guide for optimizing HPP conditions to improve flavor stability in ready-to-eat meat products.

2. Materials and methods

2.1. Materials and chemicals

Duck legs (Muscovy duck) were purchased from Hunan Jiatai Food Co., Ltd., China. The 1,2-Dichlorobenzene was purchased from Shanghai Macklin Biochemical Technology Co., Ltd. (≥99%), the n-alkanes standards were purchased from Supelco, Inc., USA (C7–C30), and the anethole (CAS: 4180-23-8, ≥98%), linalool (CAS: 78–70-6, ≥98%), and 1-octen-3-ol (CAS: 3391-86-4, ≥98%) were obtained from Beijing Solarbio Science & Technology Co., Ltd. Disodium hydrogen phosphate, sodium dihydrogen phosphate, magnesium chloride, ethylenediaminetetraacetic acid (EDTA), guanidine hydrochloride, urea, hydrochloric acid, 2,4-dinitrophenylhydrazine (DNPH), bromophenol blue, and ethylene glycol bis (2-aminoethyl ether) tetraacetic acid (EGTA) were analytical grade and purchased from Sinopharm Chemical Reagent Co., Ltd.

2.2. Preparation of sauced duck legs

Sauced duck legs were prepared according to the method described by Zhang et al. (2025), with minor modifications. Briefly, fresh duck legs (300 ± 20 g) were marinated with salt (1 g/100 g) and cooking wine (2 g/100 g) for 12 h. The duck legs were then simmered at 80 °C for 60 min in a mixture containing salt (10 g/L), chili peppers (15.6 g/L), Sichuan peppercorns (14.6 g/L), cinnamon (3 g/L), soy sauce (20 g/L), white sugar (20 g/L), and star anise (3 g/L). After cooking, the samples were cooled on a ventilated rack for 30 min and vacuum-packaged in 0.2-mm-thick bags (10 cm × 15 cm) at a vacuum level of 60 kPa. The packaged samples were randomly assigned to the untreated control and HPP-treated groups. Samples in the HPP-treated group were processed at 400 MPa for 15 min using a high-pressure processing system (L1–600-3, Tianjin Huataishengmiao Biotechnology Engineering Co., Ltd., Tianjin, China). Water was used as the pressure-transmitting medium, and the system was operated at a nominal temperature of 25 °C. The pressure was increased at approximately 200 MPa/min and released within 4 s after the holding period. The internal product temperature was not directly monitored during pressurization. The control samples underwent the same packaging and refrigerated-storage procedures without pressure treatment. All samples were stored at 4 °C and collected on days 0, 7, and 14 for subsequent analyses.

2.3. Analysis of volatile flavor compounds

Volatile compounds in sauced duck legs were analyzed using headspace solid-phase microextraction coupled with gas chromatography–mass spectrometry (HS-SPME–GC–MS), based on the method of Wang et al. (2026). A 1.0 g portion of the homogenized meat sample was transferred into a 20 mL headspace vial, and 5 μL of 1,2-dichlorobenzene solution (0.23 mg/L) was added as the internal standard. Before analysis, the 50/30 μm divinylbenzene/carboxen/polydimethylsiloxane fiber (Supelco, USA) was conditioned at 250 °C for 30 min. The vial was sealed and equilibrated at 60 °C for 5 min without agitation. Volatile compounds were then extracted at 60 °C for 40 min without agitation. Blanks were performed to assess background contamination, and no peaks interfering with compound identification or semi-quantification were observed.

After extraction, the fiber was inserted into the injection port of an Agilent 7890 gas chromatograph coupled to an Agilent 5977 mass-selective detector (Agilent Technologies, USA) and desorbed at 250 °C for 3 min in splitless mode. Volatile compounds were separated using an HP-5MS capillary column (30 m × 0.25 mm × 0.25 μm film thickness). Helium was used as the carrier gas at a constant flow rate of 1.0 mL/min. The oven temperature was initially maintained at 40 °C for 3 min, increased to 130 °C at 4 °C/min and held for 2 min, and then increased to 230 °C at 8 °C/min and held for 5 min. The mass spectrometer was operated in electron ionization mode at 70 eV. The ion-source, quadrupole, and transfer-line temperatures were 230, 150, and 280 °C, respectively. Mass spectra were recorded in full-scan mode over an m/z range of 35–550.

Partial least-squares discriminant analysis (PLS-DA) and variable importance in projection (VIP) analysis were performed using MetaboAnalyst. The validity of the PLS-DA model was assessed by 200 permutation tests using SIMCA version 14.1 (Umetrics, Umeå, Sweden). Volatile compounds with VIP values greater than 1.0 and significant treatment-related differences (p < 0.05) were considered representative differential compounds, following a screening strategy previously applied to duck-meat volatile profiles (Duan et al., 2023).

Volatile compounds were tentatively identified by comparing their mass spectra with the NIST 17 mass spectral library and their experimental linear retention indices with reference values in the NIST database. A C7–C30 n-alkane series was analyzed under the same chromatographic conditions. The linear retention indices (RIs) were calculated using a homologous series of n-alkanes according to the method of the Van Den Dool–Kratz equation (Babič et al., 2024) (Eq. 1):

$$\mathit{RI}=100\left[n+\frac{t_R\left(x\right)-t_R\left(n\right)}{t_R\left(n+1\right)-t_R\left(n\right)}\right]$$ (1)

where t_R(x) is the retention time of the analyte, t_R(n) and t_R(n + 1) are the retention times of the n-alkanes eluting immediately before and after the analyte, respectively, and n is the carbon number of the preceding n-alkane. Compounds with a mass spectral similarity score greater than 80% and an absolute difference of no more than 20 index units between the experimental and reference retention indices were retained as tentatively identified compounds. Semi-quantification was performed using the concentration of 1,2-dichlorobenzene and the peak-area ratio of each compound to the internal standard.

2.4. Extraction of myofibrillar proteins

MP were extracted from sauced duck legs according to the method described by Huang et al. (2026), with slight modifications. Briefly, 10 g of sauced duck leg sample was homogenized in an ice bath with four volumes of phosphate buffer at a sample-to-buffer ratio of 1:4 (w/v). The buffer contained 10 mM NaH₂PO₄/Na₂HPO₄, 0.1 M NaCl, 2 mM MgCl₂, and 1 mM EGTA and was adjusted to pH 7.0. Homogenization was performed at 10,000 rpm for 80 s. The homogenate was centrifuged at 3000 ×g for 15 min at 4 °C, and the supernatant was discarded. This washing and centrifugation cycle was performed three times in total.

The resulting pellet was filtered through four layers of gauze to remove connective tissue. It was then resuspended in four volumes of 0.1 M NaCl solution and homogenized at 10,000 rpm for 80 s. The suspension was centrifuged at 2500 ×g for 10 min at 4 °C, and the supernatant was discarded. This washing and centrifugation cycle was performed four times in total. MP was independently extracted from three separate day-0 duck-leg samples for each treatment. Because the initial protein concentration varied among the independently prepared extracts, the protein concentration of each extract was determined using the Biuret method and adjusted to 0.5 mg/mL before subsequent analyses. The three independently prepared MP extracts were analyzed separately and were not pooled.

2.5. Interactions between MP and flavor compounds

Following the screening procedure described in Section 2.3, anethole, linalool, and 1-octen-3-ol were selected as representative compounds for subsequent MP–flavor interaction analyses.

2.5.1. Measurement of flavor retention

The retention of the three compounds by MP was evaluated based on their release into the headspace. The MP solutions were diluted to 0.5 mg/mL using 20 mM phosphate buffer containing 0.6 M NaCl (pH 6.25). Anethole, linalool, and 1-octen-3-ol were separately added to the MP solutions to obtain final concentrations of 0, 0.02, 0.04, 0.06, 0.08, and 0.10 mM. The mixtures were transferred into 15 mL headspace vials, sealed, vortexed, and equilibrated at room temperature for 24 h. The corresponding blank samples contained the same concentrations of flavor compounds but no MP. The volatile compounds released into the headspace were analyzed by HS-SPME–GC–MS under the conditions described in Section 2.3. Flavor retention was calculated according to the description of Guo et al. (2020) (Eq. 2):

$$\mathit{Flavor\; retention\; rate}\left(\%\right)=\left(1-\frac{S_{\mathit{MP}}}{S_0}\right)\times 100\%$$ (2)

where S_MP and S₀ represent the peak areas of the target volatile compound in the headspace of the protein-containing sample and the corresponding blank sample without MP, respectively.

2.5.2. Estimation of apparent mass-normalized interaction parameters

The apparent affinity coefficient (K) and apparent mass-normalized interaction capacity (n) were estimated from the flavor-retention data using an operational Klotz-type model according to Gu et al. (2020) (Eq. 3):

$$\frac{1}{v}=\frac{1}{\mathit{nK}\left[L\right]}+\frac{1}{n}$$ (3)

in:

$$v=\frac{\left(\frac{\left[\mathit{HS}\right]c-\left[\mathit{HS}\right]p}{\left[\mathit{HS}\right]c}\right)\times C_T}{C_p}$$ (4)

in:

$$\left[L\right]=\frac{\left[\mathit{HS}\right]p}{\left[\mathit{HS}\right]c}\times C_T$$ (5)

where [HS]_c and [HS]_p are the headspace concentrations of the target flavor compound in the blank control and protein-containing sample, respectively, C_T is the total flavor-compound concentration expressed in mM, C_p is the standardized MP mass concentration of 0.5 mg/mL, and [L] is the apparent free flavor-compound concentration expressed in mM. No conversion of MP mass concentration to molar protein concentration was performed because the extracted MP was a heterogeneous protein fraction without a defined molecular mass. Accordingly, ν and n were interpreted as apparent mass-normalized retention capacities, expressed in mmol g⁻¹, whereas K was interpreted as an apparent concentration-dependent affinity coefficient, expressed in mM⁻¹. These empirical, model-dependent parameters were used only for relative comparisons between the control and HPP-treated groups and were not interpreted as absolute thermodynamic binding constants or molecular binding-site numbers.

2.6. Circular dichroism spectroscopy

Circular dichroism (CD) spectroscopy was performed according to the method described by León-Heredia et al. (2025) with appropriate modifications. MP solutions were diluted to 0.5 mg/mL using 20 mM phosphate buffer containing 0.6 M NaCl (pH 6.25). Anethole, linalool, or 1-octen-3-ol was separately added to the MP solutions to obtain final concentrations of 0, 0.02, 0.04, 0.06, 0.08, and 0.10 mM. The mixtures were incubated at room temperature for 30 min before spectral acquisition. CD spectra were recorded using a spectropolarimeter (J-1500, JASCO Corporation, Tokyo, Japan) equipped with a quartz cuvette with a path length of 0.1 cm. The corresponding phosphate buffer was used as the background. Spectra were collected from 190 to 250 nm at 1 nm intervals under continuous nitrogen purging. Each independent sample was scanned four times, and the four scans were averaged after background subtraction. The proportions of α-helix, β-sheet, β-turn, and random coil structures were estimated using the BeStSel online server (https://bestsel.elte.hu/index.php).

2.7. Fluorescence spectroscopy

MP solutions were diluted to 0.5 mg/mL using 20 mM phosphate buffer containing 0.6 M NaCl (pH 6.25). Anethole, linalool, or 1-octen-3-ol was added separately to obtain final concentrations of 0, 0.02, 0.04, 0.06, 0.08, and 0.10 mM. After incubation at room temperature for 30 min, the samples were analyzed using a multimode microplate reader (Varioskan LUX, Thermo Fisher Scientific Oy, Vantaa, Finland). The excitation wavelength was set at 280 nm, and fluorescence emission spectra were recorded from 300 to 500 nm. No inner-filter correction was applied to the fluorescence intensities.

The fluorescence-quenching behavior of the flavor compounds toward MP was evaluated using the Stern–Volmer equations (Eq. 6):

$$\frac{F_0}{F}=1+K_q{\tau }_0\left[Q\right]=1+K_{\mathit{SV}}\left[Q\right]$$ (6)

in:

$$K_q=\frac{K_{\mathit{SV}}}{{\tau }_0}$$ (7)

where F₀ and F are the fluorescence intensities in the absence and presence of the quencher, [Q] is the quencher concentration, K_SV is the Stern–Volmer quenching constant, K_q is the bimolecular quenching constant, and τ₀ is the average fluorescence lifetime of the fluorophore in the absence of the quencher, which was assumed to be 10⁻⁸ s.

2.8. UV–visible absorption spectroscopy

The MP solution was diluted to 0.5 mg/mL with 0.01 M phosphate-buffered saline (PBS, pH 7.0). A 1.0 mL aliquot of the MP solution was mixed with 200 μL of anethole, linalool, or 1-octen-3-ol solution prepared in phosphate buffer containing 0.6 M NaCl (pH 6.25). The mixtures were incubated with shaking at 500 rpm for 30 min. The UV–visible absorption spectra were recorded from 200 to 400 nm using the absorbance measurement mode of a spectropolarimeter (J-1500, JASCO Corporation, Tokyo, Japan).

2.9. Molecular dynamics simulation

Molecular dynamics simulations were performed using GROMACS 2019.6. The initial three-dimensional structure was the AlphaFold-predicted model of LLGL scribble cell polarity complex component 2 (LLGL2) from mallard (Anas platyrhynchos platyrhynchos). The model was obtained from the AlphaFold Protein Structure Database. Its model ID was AF-U3IQ23-F1, and the corresponding UniProt accession was U3IQ23. The structure was predicted by AlphaFold from the corresponding protein sequence in UniProtKB. The mean predicted local distance difference test (pLDDT) score was 82.5, indicating generally high confidence in the predicted local structure.

The GROMOS54A7 force field and SPC/E water model were used. The protein was placed in a cubic periodic box. The minimum distance between the protein surface and the box boundary was 1.0 nm. NaCl was added to a final concentration of 0.6 M. Counterions were added to neutralize the system.

The system was first subjected to energy minimization. It was then equilibrated for 400 ps under the canonical (NVT) ensemble at 298.15 K. This was followed by 400 ps under the isothermal–isobaric (NPT) ensemble at 298.15 K and 0.1 MPa. The control system was maintained at 0.1 MPa. For the high-pressure system, pressure was increased stepwise to 400 MPa under NPT conditions. The Berendsen barostat was used for rapid pressure adjustment at each step. The Parrinello–Rahman barostat was then used for constant-pressure equilibration. Production simulations were conducted for 150 ns at 298.15 K and the corresponding pressure. Covalent bonds were constrained using the linear constraint solver (LINCS) algorithm. Long-range electrostatic interactions were calculated using the particle mesh Ewald method. The van der Waals cutoff was set to 1.0 nm.

2.10. Molecular docking

The final frames of the 0.1 and 400 MPa production trajectories were used as the control and high-pressure receptor conformations, respectively. Both trajectories had reached relative stability according to the root mean square deviation (RMSD) and radius of gyration (Rg) profiles. The unrefined AlphaFold model was not used directly for docking.

Molecular docking was performed using AutoDock Vina with an exhaustiveness value of 8. Separate search spaces were used for the two receptor conformations. For the control receptor, the grid box was centered at x = 2.501, y = 1.062, and z = 3.908 Å, with dimensions of 85.616 × 81.331 × 111.723 Å³. For the 400 MPa receptor, the grid box was centered at x = 74.893, y = 58.530, and z = 88.169 Å, with dimensions of 33.147 × 33.147 × 33.147 Å³. The top-ranked pose with the lowest predicted binding energy within each search space was selected for interaction analysis and visualization using Discovery Studio and PyMOL. Because the search spaces differed, the docking results were used only for qualitative visualization and not for quantitative comparison of predicted binding energies or definitive inference of pressure-induced binding-site migration.

2.11. Statistical analysis

For volatile profiling, three independent samples (n = 3) were analyzed for each treatment and storage time. For MP-based analyses, MP was independently extracted from three day-0 samples per treatment (n = 3), and each extract was standardized to 0.5 mg/mL before analysis. Each CD spectrum represented the average of four instrumental scans of one independent sample. Results are expressed as the mean ± standard deviation. Analysis of variance followed by Duncan's multiple-range test was performed using SPSS Statistics version 27.0 (IBM Corp., Armonk, NY, USA). Differences were considered significant at p < 0.05. Figures were prepared using Origin 2021 (OriginLab Corporation, Northampton, MA, USA). For each volatile compound, differences among storage times within the same treatment and differences between treatments at the same storage time were evaluated using one-way ANOVA followed by Duncan's multiple-range test.

3. Results and discussion

3.1. Effects of HPP on volatile profiles and selection of representative compounds in sauced duck legs

As shown in Table S1, aldehydes and alcohols were the major volatile classes detected in sauced duck legs during storage, together with hydrocarbons, esters, ketones, and other volatile compounds. The effects of HPP on the volatile profile varied with storage time. Total aldehyde concentrations in the HPP-treated and control groups were 2.04 and 1.44 μg/kg on day 0, 2.33 and 1.88 μg/kg on day 7, and 0.88 and 1.47 μg/kg on day 14, respectively. Thus, total aldehydes were higher in the HPP-treated group during the early and middle storage stages but lower than those in the control group at the end of storage. However, aldehydes should not be regarded as a uniformly undesirable volatile class because their sensory effects depend on their chemical identity and abundance. Some aldehydes, such as heptanal, nonanal, and unsaturated aldehydes, may contribute fruity, fatty, citrus-like, or meaty notes, whereas hexanal is commonly associated with grassy, pungent, and rancid odors (Zhang et al., 2025). Among the individual aldehydes, hexanal content increased during the early storage stage but was lower in the HPP-treated group than in the control group on day 14. Hexanal is commonly formed through linoleic acid oxidation. Its lower concentration at the end of storage may indicate reduced formation of some lipid-oxidation-derived volatiles, but it does not prove that overall lipid oxidation was inhibited (Azarbad & Jeleń, 2015). Therefore, the lower hexanal and total aldehyde concentrations on day 14 indicate a shift in volatile composition but do not by themselves demonstrate reduced overall flavor deterioration.

The total alcohol concentration was higher in the HPP-treated group on days 0 and 14 but slightly lower on day 7. Linalool remained one of the most abundant alcohols throughout storage. On day 7, its concentration was significantly lower in the HPP-treated group (0.47 ± 0.03 μg/kg) than in the control group (0.58 ± 0.03 μg/kg, p < 0.05). In contrast, the concentration of 1-octen-3-ol on day 7 was significantly higher in the HPP-treated group (0.10 ± 0.01 μg/kg) than in the control group (0.04 ± 0.02 μg/kg, p < 0.05). Total hydrocarbon concentrations were higher in the HPP-treated group on days 0 and 14 but were similar between the two groups on day 7. Anethole was among the most abundant individual volatile compounds. On day 0, its concentration was significantly higher in the HPP-treated group (2.09 ± 0.40 μg/kg) than in the control group (0.76 ± 0.18 μg/kg, p < 0.05). Ketones and esters were present at relatively low concentrations and showed smaller treatment-related changes. Overall, HPP altered the time-dependent distribution of volatile compounds rather than producing a uniform increase or decrease in any single chemical class.

Multivariate analysis further characterized the differences in volatile profiles among the samples. The PLS—DA score plot showed separation among the samples, with Components 1 and 2 accounting for 37.2% and 41.0% of the modeled variation, respectively, and 78.2% cumulatively (Fig. 1a). The PLS—DA model yielded R² = 0.919 and Q² = 0.708. The negative Q² intercept obtained from the 200-permutation test suggested that the model was not obviously overfitted (Fig. 1b). The overall VIP plot showed that anethole had the highest VIP score, while linalool and 1-octen-3-ol also had VIP values greater than 1.0 (Fig. 1c). The time-specific VIP plots further showed that these compounds contributed to the discrimination between the control and HPP-treated samples at different storage stages (Fig. 1d–f). Anethole and linalool were likely derived mainly from the spices used during braising, particularly star anise, and are associated with anise-like and floral notes, respectively. By contrast, 1-octen-3-ol is commonly formed through the oxidation of unsaturated fatty acids and has a characteristic mushroom-like odor. Depending on its concentration and the food matrix, 1-octen-3-ol may contribute to characteristic duck-meat aroma or earthy off-notes (Zhong et al., 2024). Because all three compounds had VIP values greater than 1.0 and each showed a significant treatment-related difference at least one storage time point (p < 0.05), they were selected as representative compounds for the subsequent MP–flavor interaction analyses. Therefore, these compounds represented both spice-derived characteristic aroma compounds and a lipid-derived volatile associated with storage-related flavor changes (Zhong et al., 2026).

Fig. 1.

Multivariate analysis of volatile profiles in control and HPP-treated sauced duck legs during storage. (a) PLS-DA score plot; (b) 200-permutation test of the PLS-DA model; (c) overall VIP scores; and VIP scores on days 0 (d), 7 (e), and 14 (f). Compounds with VIP > 1.0 were considered important variables. C-0, C-7, and C-14 represent untreated controls stored for 0, 7, and 14 days, respectively; 4–0, 4–7, and 4–14 represent samples treated at 400 MPa and stored for the corresponding periods.

3.2. Effect of HPP on flavor retention by MP

As shown in Fig. 2, the apparent retention rates of anethole, linalool, and 1-octen-3-ol decreased progressively as their concentrations increased from 0.02 to 0.10 mM. At each tested concentration, MP extracted from the HPP-treated samples exhibited significantly higher retention rates than MP extracted from the control samples (p < 0.05). Among the three compounds, linalool generally showed the highest retention rate, particularly at higher concentrations, whereas the retention of 1-octen-3-ol decreased more markedly with increasing concentration. These compound-dependent differences may be related to differences in molecular structure and interactions with MP.

Fig. 2.

Apparent retention rates of flavor compounds by MP extracted from control and HPP-treated sauced duck legs. Apparent retention rates of anethole (a), linalool (b), and 1-octen-3-ol (c) at concentrations of 0.02–0.10 mM. Values are expressed as mean ± SD of three independent samples. Different lowercase letters indicate significant differences among concentrations within the same treatment, whereas different uppercase letters indicate significant differences between treatments at the same concentration (p < 0.05).

At low concentrations, a relatively large proportion of the flavor molecules could interact with available binding sites on MP. As the flavor-compound concentration increased, these sites became progressively occupied, leaving a larger proportion of the compounds unbound. Consequently, the apparent retention rate decreased. A similar concentration-dependent decrease in flavor retention was observed in the interaction between bacon MP and 2,3-butanedione, which was attributed to progressive saturation of the available binding sites (Huang et al., 2026). In addition, processing-induced unfolding of MP was reported to expose additional binding sites and enhance its interaction with furan flavor compounds (Yu et al., 2024).

3.3. Apparent mass-normalized interaction parameters

As shown in Fig. 3, an operational mass-normalized Klotz-type analysis was used to describe the apparent concentration-dependent interaction behavior between MP and the three flavor compounds within the tested concentration range. Compared with the control group, the fitted regression lines for anethole, linalool, and 1-octen-3-ol shifted downward after treatment at 400 MPa, indicating that HPP altered the apparent concentration-dependent interaction behavior of MP within the operational analysis. As shown in Table 2, the apparent affinity coefficient (K) and apparent mass-normalized interaction capacity (n) for all three flavor compounds were significantly higher in the HPP-treated group than in the control group (p < 0.05). Specifically, the K value for anethole increased from 16.77 ± 0.85 to 18.83 ± 0.75 mM⁻¹, and its n value rose from 2.96 ± 0.03 to 3.59 ± 0.04 mM g⁻¹. For linalool, the K value increased significantly from 13.29 ± 0.70 to 42.32 ± 2.15 mM⁻¹, and the n value rose from 2.99 ± 0.03 to 3.63 ± 0.04 mM g⁻¹. Linalool therefore showed the greatest increase in K after HPP treatment. Similarly, the K and n values of 1-octen-3-ol were also significantly increased after HPP treatment.

Fig. 3.

Linear relationships between flavor-compound concentration and reciprocal apparent MP retention capacity. Interactions of MP with anethole (a), linalool (b), and 1-octen-3-ol (c) in the control group, and with anethole (d), linalool (e), and 1-octen-3-ol (f) in the HPP-treated group.

Table 2.

Apparent mass-normalized interaction parameters derived from the operational Klotz-type analysis.

Flavor compounds	Treatment	K (mM−1)	n (mM g−1)	R2
Anethole	Control	16.77 ± 0.85B	2.96 ± 0.03B	1.00
	400 MPa	18.83 ± 0.75 A	3.59 ± 0.04 A	0.99
Linalool	Control	13.29 ± 0.70B	2.99 ± 0.03B	0.98
	400 MPa	42.32 ± 2.15 A	3.63 ± 0.04 A	0.98
1-Octen-3-ol	Control	13.90 ± 0.75B	3.02 ± 0.03B	0.97
	400 MPa	32.87 ± 1.78 A	3.64 ± 0.04 A	0.99

Different letters between groups indicate significant differences (p < 0.05).

These results suggest that HPP-induced structural changes may have increased the accessibility of potential interaction regions on MP, thereby strengthening its apparent interactions with the three flavor compounds. The degree of enhancement differed among the compounds, indicating compound-specific binding responses. Zhao et al. (2022) found that treatment of pork MP at 200 MPa exposed intrinsic tryptophan residues and promoted the partial conversion of α-helix to β-sheet. Similarly, Xu et al. (2024) reported that increasing the pressure to 400 MPa caused progressive α-helix unfolding and β-sheet formation in whitebait MP. These structural changes are consistent with the decreased α-helix content observed in the present study and support the interpretation that HPP rearranged MP structure and increased the accessibility of potential interaction regions.

3.4. Secondary structure analysis

As shown in Fig. 4, HPP treatment at 400 MPa altered the secondary structure distribution of MP. In the absence of added flavor compounds, the α-helix content decreased from 28.6% to 18.1%, while the β-sheet, β-turn, and random coil contents increased from 12.1%, 9.9%, and 49.4% to 16.2%, 13.3%, and 52.4%, respectively. These changes suggest that HPP induced partial unfolding and structural rearrangement of MP. The addition of anethole, linalool, and 1-octen-3-ol further decreased the α-helix content, while the changes in β-sheet, β-turn, and random coil contents depended on the type and concentration of the flavor compound. The addition of anethole decreased the α-helix content and increased the β-sheet content. Linalool produced relatively smaller changes, although a decrease in α-helix content and an increase in β-turn content were observed. Similarly, 1-octen-3-ol decreased the α-helix content and maintained a relatively high random coil content.

Fig. 4.

Effects of HPP and flavor compounds on the circular dichroism spectra and secondary structure composition of MP. Circular dichroism spectra of MP extracted from control samples in the presence of anethole (a), linalool (b), and 1-octen-3-ol (c), and of MP extracted from HPP-treated samples in the presence of anethole (d), linalool (e), and 1-octen-3-ol (f). Secondary structure compositions of control MP interacting with anethole (g), linalool (h), and 1-octen-3-ol (i), and of HPP-treated MP interacting with anethole (j), linalool (k), and 1-octen-3-ol (l). The stacked bars represent the α-helix, β-sheet, β-turn, and random coil contents. Flavor-compound concentrations ranged from 0 to 0.10 mM, and 0 mM represents MP without added flavor compounds.

These results indicate that HPP induced partial unfolding and structural rearrangement of MP. The further decrease in α-helix content after adding flavor compounds suggests that protein–flavor interactions also affected MP conformation. The different responses to the three compounds indicate that the extent of structural change depended on their molecular characteristics and the initial conformation of MP. The decrease in α-helix content and the increase in other secondary structures are consistent with previous reports that HPP can rearrange the secondary structure of MP (Xu et al., 2024). Flavor molecules can also alter MP conformation through non-covalent interactions (Qian et al., 2024). Therefore, the concentration-dependent decrease in α-helix content may be related to interactions between the flavor compounds and accessible regions of MP (Sun et al., 2023). The stronger effect of anethole suggests that different flavor compounds interact with MP to different extents, which is consistent with the compound-dependent structural responses reported by Vatansever et al. (2024).

3.5. UV–visible absorption spectroscopy

As shown in Fig. 5, HPP at 400 MPa changed the UV absorption profile of MP. Without added flavor compounds, MP extracted from the HPP-treated samples showed lower absorbance than MP extracted from the control samples, especially at 270–290 nm. After the addition of flavor compounds, the absorbance increased with increasing concentrations of anethole and linalool, whereas the spectra obtained at different 1-octen-3-ol concentrations showed substantial overlap. These results suggest that HPP changed the local environment of aromatic amino acid residues in MP. Changes at 270–290 nm are commonly associated with changes in the environment of tyrosine and tryptophan residues (Wei et al., 2024). However, separate flavor-compound blanks were not subtracted. Therefore, part of the increased absorbance may have resulted from direct absorption by the added compounds. The UV spectra were thus interpreted only as supporting evidence together with the CD, fluorescence, and retention results. Similar UV spectral changes have been linked to altered exposure and solvent accessibility of aromatic side chains during protein–flavor interactions (Li et al., 2022). The different responses may be related to the molecular structures of the compounds. Anethole may interact mainly through hydrophobic and aromatic interactions. Linalool and 1-octen-3-ol may also form hydrogen bonds through their hydroxyl groups (Qian et al., 2024; Xue et al., 2022).

Fig. 5.

Effects of HPP and flavor compounds on the UV–visible absorption spectra of MP. Spectra of MP extracted from control samples in the presence of anethole (a), linalool (b), and 1-octen-3-ol (c), and of MP extracted from HPP-treated samples in the presence of anethole (d), linalool (e), and 1-octen-3-ol (f). The insets show enlarged spectra in the wavelength range of 270–290 nm. Flavor-compound concentrations ranged from 0 to 0.10 mM, and 0 mM represents MP without added flavor compounds.

3.6. Fluorescence quenching spectroscopy

Fluorescence quenching spectroscopy was used to examine the interactions between MP and the three flavor compounds. As shown in Fig. 6, before the flavor compounds were added, MP extracted from the HPP-treated samples showed lower fluorescence intensity than MP extracted from the control samples. However, the maximum emission wavelength remained near 329 nm. After anethole, linalool, or 1-octen-3-ol was added, the fluorescence intensity of MP decreased progressively with increasing compound concentration in both groups. A slight red shift in the maximum emission wavelength was also observed. Among the three compounds, 1-octen-3-ol caused the greatest decrease in fluorescence intensity and the largest wavelength shift, followed by linalool and anethole. As shown in Table 1, the K_SV values of all three compounds were significantly higher in the HPP-treated group than in the control group (p < 0.05). These results indicate that HPP changed the tertiary structure of MP and increased its fluorescence-quenching response to the flavor compounds. The decrease in fluorescence intensity suggests that the flavor compounds interacted with fluorescent residues or regions close to these residues. The slight red shift indicates that the local environment around the fluorophores became more polar after interaction (Zhou et al., 2025). The different responses among the three compounds also suggest that their interactions with MP depended on their molecular structures. Zhao et al. (2022) reported that HPP disrupted hydrogen bonds and hydrophobic interactions in porcine MP. These changes altered the tertiary structure and increased the exposure of tryptophan residues to the aqueous environment. The apparent K_q values in the present study exceeded the commonly cited diffusion-controlled limit of 2.0 × 10¹⁰ L mol⁻¹ s⁻¹. This result is consistent with static quenching being the predominant mechanism and may reflect the formation of protein–flavor complexes (Xue et al., 2022). HPP-induced structural rearrangement may have exposed additional interaction regions and increased the accessibility of aromatic residues. This could explain the higher K_SV values observed after HPP treatment. However, inner-filter effects were not corrected in the present study. Therefore, K_SV and K_q were treated as apparent quenching parameters rather than definitive proof of a specific quenching mechanism. In addition, these results were obtained using extracted MP and model flavor compounds. They support a possible role of MP–flavor interactions in flavor retention but do not directly confirm the complete retention mechanism in sauced duck legs.

Fig. 6.

Effects of HPP and flavor compounds on the fluorescence emission spectra of MP. Spectra of MP extracted from control samples in the presence of anethole (a), linalool (b), and 1-octen-3-ol (c), and of MP extracted from HPP-treated samples in the presence of anethole (d), linalool (e), and 1-octen-3-ol (f). The excitation wavelength was 280 nm. Flavor-compound concentrations ranged from 0 to 0.10 mM, and 0 mM represents MP without added flavor compounds.

Table 1.

Stern–Volmer quenching parameters for interactions between MP and flavor compounds in different treatment groups.

Flavor compounds	Treatment	KSV (103 L mol−1)	Kq (1011 L mol−1 s−1)	R2
Anethole	Control	2.35 ± 0.27B	2.35 ± 0.27B	0.96
	400 MPa	3.51 ± 0.21 A	3.51 ± 0.21 A	0.99
Linalool	Control	1.27 ± 0.15B	1.27 ± 0.15B	0.96
	400 MPa	8.03 ± 1.15 A	8.03 ± 1.15 A	0.94
1-Octen-3-ol	Control	17.53 ± 3.54B	17.53 ± 3.54B	0.89
	400 MPa	20.57 ± 2.99 A	20.57 ± 2.99 A	0.94

Different letters between groups indicate significant differences (p < 0.05).

3.7. Molecular dynamics simulation

Molecular dynamics (MD) simulations can provide molecular-level information on pressure-dependent protein conformational behavior (Kou et al., 2023). As shown in Fig. 7a–d, RMSD, Rg, RMSF, and SASA were analyzed for the systems simulated at 0.1 and 400 MPa. The RMSD values of both systems fluctuated during the initial stage and subsequently reached relatively stable plateaus. The 400 MPa system showed a lower RMSD plateau than the 0.1 MPa system. In contrast, its Rg and SASA values were slightly higher. The RMSF profiles were generally similar between the two systems, although several local regions showed lower fluctuations at 400 MPa. These results indicate that both trajectories reached relatively stable conformational states. The lower RMSD at 400 MPa reflects a smaller average deviation from the reference structure. The higher Rg and SASA values indicate that the protein adopted a slightly more expanded conformation with greater solvent exposure. Therefore, the 400 MPa system did not simply undergo greater structural instability. Instead, it appeared to stabilize in a slightly expanded conformational state. The lower RMSF values in some regions also indicate reduced local flexibility. High pressure can alter protein hydration and non-covalent interactions, thereby shifting the conformational equilibrium of proteins (Hata et al., 2020). Similar changes in protein compactness and solvent exposure have been reported in spectroscopic and MD studies of high-pressure-treated proteins (Yang et al., 2024). The results are also consistent with experimental evidence that HPP can induce partial unfolding and structural rearrangement of MP (Chapleau et al., 2004). Nevertheless, the MD simulation was based on an LLGL2 model from mallard, whereas the experimental system comprised a heterogeneous MP fraction extracted from Muscovy duck. Because LLGL2 is not a constituent of myofibrillar proteins, the simulation results are model-specific and should be regarded only as qualitative evidence of pressure-dependent protein conformational behavior. They should not be directly extrapolated to the experimental MP system.

Fig. 7.

Molecular dynamics characteristics of the LLGL2 model at 0.1 and 400 MPa and representative docking poses of three flavor compounds. The root mean square deviation (RMSD) (a), radius of gyration (Rg) (b), root mean square fluctuation (RMSF) (c), and solvent-accessible surface area (SASA) (d) of the LLGL2 model. Representative top-ranked docking poses of anethole, linalool, and 1-octen-3-ol with the final stable LLGL2 conformations extracted from the 0.1 MPa (e) and 400 MPa (f) production trajectories.

3.8. Molecular docking

Molecular docking was used to visualize plausible interaction modes between the LLGL2 model and the three flavor compounds. As shown in Fig. 7e and f, the top-ranked poses suggested that anethole, linalool, and 1-octen-3-ol could form non-covalent contacts with both the 0.1 and 400 MPa receptor conformations. These predicted contacts included hydrophobic interactions, hydrogen bonds, and van der Waals interactions. In the 0.1 MPa receptor, the top-ranked pose of anethole involved Lys717 at approximately 2.1 Å, whereas that of linalool involved His597 at approximately 3.2 Å. The predicted pose of 1-octen-3-ol involved Leu793 and Glu791 at approximately 2.5 and 2.8 Å, respectively. In the 400 MPa receptor, the corresponding poses involved Lys514 for anethole, His392 for linalool, and Arg655 for 1-octen-3-ol. The predicted contact distances were approximately 2.3, 2.0, and 2.5 Å, respectively. These results indicate that all three compounds could interact with the LLGL2 model, but their predicted interaction patterns differed between the two receptor conformations. The differences may be related to both ligand structure and pressure-dependent receptor conformation. Anethole is relatively hydrophobic and may mainly favor hydrophobic contacts. In contrast, the hydroxyl groups of linalool and 1-octen-3-ol may also participate in hydrogen bonding. Similar studies have shown that the functional groups of flavor compounds and the local conformation of food proteins can affect predicted contact residues and non-covalent interaction patterns (Wongprasert et al., 2024). However, separate search spaces were used for the two receptor conformations. Therefore, the differences in contact residues cannot be interpreted as direct evidence of pressure-induced binding-site migration. Molecular docking also provides static and model-dependent predictions. The results were therefore used only to visualize plausible interaction modes and were not used to compare binding energies quantitatively.

4. Conclusion

HPP at 400 MPa changed the volatile composition of sauced duck legs during storage, and the treated samples showed a lower total aldehyde concentration than the control on day 14. The anethole, linalool, and 1-octen-3-ol were identified as representative compounds affected by HPP. HPP decreased the α-helix content and increased the apparent affinity coefficients and mass-normalized interaction capacities to maintain the flavor quality of sauced duck legs. The findings of our work could provide some guidance for the development of high-quality and long flavor keep meat products. Nevertheless, the relationship between flavor keeping and fat oxidation requires further verification through direct lipid-oxidation measurements, aroma-activity assessment, sensory analysis, and validation in future work.

CRediT authorship contribution statement

Qingwu Shen: Supervision, Resources, Project administration, Funding acquisition, Data curation, Conceptualization. Changhai Ding: Writing – review & editing, Writing – original draft, Visualization, Software, Methodology, Formal analysis. Zhixuan Tang: Data curation, Conceptualization. Wei Quan: Project administration, Funding acquisition. Aihua Lou: Resources, Project administration. Yan Wang: Resources, Project administration. Yan Liu: Project administration, Funding acquisition. Weipeng Zhang: Writing – review & editing, Visualization, Software. Lei Zhou: Writing – review & editing, Writing – original draft, Visualization.

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.

Appendix A. Supplementary data

Supplementary material 1

Changes in Volatile Components of Sauced Duck Legs after HPP Treatment.

Data availability

Data will be made available on request.