Pigskin Collagen-Derived Antifreeze Peptides as Clean-Label Cryoprotectants: Inhibition of Ice Recrystallization and Suppression of Oxidative Deterioration in Heme-Rich Pork Sausages

Abstract

Frozen red meat products face dual challenges: structural damage from ice recrystallization and subsequent oxidative deterioration driven by endogenous pro-oxidants in heme-rich matrices. This study investigated the efficacy of pigskin collagen-derived antifreeze peptides (APPs) as clean-label cryoprotectants in pork sausages. In vitro characterization confirmed distinct thermal hysteresis (0.51 °C) and potent ice recrystallization inhibition. In the sausage system, 6% APPs overcame ionic screening effects, reducing thawing loss to 4.87% (vs. 7.51% in control) and cooking loss to 14.59%, significantly outperforming commercial phosphates in moisture retention. Low-field NMR revealed that APPs stabilized the actomyosin hydration shell, maintaining the immobilized water proportion (P₂₁) at 93.67% (vs. 88.50% in control) and inhibiting conversion of immobilized water into free water. Furthermore, APPs suppressed oxidative deterioration, limiting protein carbonyl content to 3.08 nmol/mg (vs. 3.71 nmol/mg in control), supporting a “physical–chemical cascade” mechanism whereby superior microstructural preservation mitigates downstream oxidative deterioration. Despite a trade-off in textural resilience (0.37 vs. 0.44), APPs function as specialized “Ice-Structure Stabilizers” offering a robust clean-label strategy for preserving heme-rich meat products.

1. Introduction

Freezing constitutes a cornerstone strategy in modern food preservation, effectively extending shelf-life and maintaining nutritional value by suppressing microbial growth and retarding biochemical reactions [1]. However, the freezing process is inherently a “double-edged sword.” During freezing and frozen storage, the phase transition of water—specifically the formation and recrystallization of ice crystals—inflicts irreversible physical and chemical damage on the food matrix. Temperature fluctuations (thermal shock) exacerbate this phenomenon by promoting the growth of large, irregular ice crystals via Ostwald ripening [2]. In meat products, these crystals mechanically rupture cell membranes and the myofibrillar network, leading to severe drip loss, textural softening, nutrient degradation, and sensory deterioration upon thawing [3].

To mitigate freeze–thaw induced damage, the food industry has long relied on commercial cryoprotectants such as sucrose, sorbitol, and phosphates blends (e.g., sodium tripolyphosphate) [4]. While these agents reduce the freezing point via colligative effects or modify protein hydration through ionic strength, their limitations face increasingly scrutiny. High-sugar additives result in excessive sweetness, caloric burden, and potential Maillard browning. More critically, excessive phosphate intake disrupts the human calcium-phosphorus metabolic balance, potentially increasing the risk of osteoporosis and cardiovascular diseases [5]. Consequently, there is an urgent industry-wide shift towards identifying natural, safe, and efficient “clean-label” alternatives [6].

Antifreeze peptides (AFPs) have emerged as a promising solution due to their ability to specifically control ice crystal growth. This is primarily achieved through two mechanisms: Thermal Hysteresis Activity (THA), which creates a non-colligative gap between the melting and freezing points to prevent ice growth; and Ice Recrystallization Inhibition (IRI), which prevents the Ostwald ripening process where larger ice crystals grow at the expense of smaller ones [7]. Among potential sources, pigskin—a major byproduct constituting 3–8% of total swine weight—presents a unique valorization opportunity [8]. Traditionally undervalued or discarded, pigskin is rich in type I collagen, characterized by unique repeating sequences of Glycine (Gly), Proline (Pro), and Hydroxyproline (Hyp) [9]. Recent biochemical studies indicate that enzymatic hydrolysis can release “high-activity” AFPs from this collagen matrix. Notably, novel peptides isolated from pigskin (e.g., AP-3) have demonstrated a THA of up to 5.28 °C [10], a value surpassing many fish-derived AFPs and rivaling insect antifreeze proteins. This suggests that pigskin-derived AFPs are not merely a waste-derived substitute, but a potent functional ingredient independent of genetic modification concerns.

Despite the proven efficacy of APPs in simple models (e.g., bacteria) [11] or specific food matrices such as frozen dough and ice cream [10,12], their application in processed meat products like sausages faces distinct challenges. In contrast to dough or ice cream (typically low-salt or high-sugar systems), sausages constitute a high-ionic-strength environment requisite for myofibrillar protein solubilization, a condition that may screen the electrostatic interactions of peptides [13,14]. Furthermore, heme-rich red meat systems are uniquely susceptible to rapid oxidative deterioration driven by endogenous pro-oxidants (e.g., heme iron) following freeze–thaw induced cellular rupture [15]. It remains unclear whether APPs can retain their IRI activity within this complex gel matrix to physically inhibit ice growth and thereby block the oxidative cascade.

Therefore, this study hypothesizes that APPs could stabilize frozen pork sausages through a dual mechanism: physically inhibiting ice recrystallization to preserve the gel network and secondarily suppressing oxidation by maintaining cellular integrity. Specifically, the study aimed to: (1) verify the THA and IRI activities of APPs; (2) determine the effective concentration (2–6%) for replacing phosphates in sausages; and (3) elucidate the underlying mechanism using Low-Field NMR (water dynamics) and microstructural analysis, thereby establishing a theoretical basis for utilizing APPs as clean-label phosphate alternative.

2. Materials and Methods

2.1. Materials

Fresh pork meat and back fat were supplied by Fujian Yuguan Food Co., Ltd. (Fuzhou, China). Composite phosphate additives (composed of sodium tripolyphosphate, sodium pyrophosphate, and sodium hexametaphosphate) were obtained from Quanzhou Dachuan Biotechnology Co., Ltd. (Quanzhou, China). Pigskin gelatin was purchased from a local supermarket (Fuzhou, China). Alcalase was obtained from Notlas biotechnology Co., Ltd. (Beijing, China). 2,4-dinitrophenylhydrazine (DNPH) and all other analytical-grade chemicals were purchased from Macklin Biochemical Co., Ltd. (Shanghai, China) or Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China).

2.2. Preparation of Antifreeze Peptides (APPs)

The preparation of APPs strictly followed the standardized in-house protocol previously established by our research group, as detailed by Wang et al. [11]. Because the preparation was conducted in the same laboratory using identical material sources and enzymatic conditions, the hydrolysates are considered to maintain the previously characterized structural profile. Briefly, pigskin gelatin was hydrolyzed with Alcalase at 37 °C and pH 9.0 for 3 h, using an enzyme-to-substrate ratio of 0.02 (w/w). The reaction was terminated by heating at 100 °C for 5 min, followed by centrifugation at 10,000× g for 20 min. The supernatant was collected and lyophilized using a freeze-dryer (Alpha 2–4, Martin Christ, Osterode am Harz, Germany). The obtained powder, designated as antifreeze peptides (APPs), was stored at −20 °C until further use. According to the comprehensive HPLC analysis previously conducted by Wang et al. [11] on APPs produced under these exact specific conditions, the molecular weight fractions are mostly distributed between 150 and 2000 Da, accounting for approximately 80% of the hydrolyzate.

2.3. Preparation of Sausages and Freeze–Thaw Treatment

The sausage formulation consisted of 80% (w/w) lean pork and 20% (w/w) back fat. Sodium chloride (1.1%, w/w) and ice water (10%, w/w) were added to the meat during chopping to form a homogeneous pork mince. This mince was then divided into five treatment groups based on the mince weight: a blank control (no additives); a positive control treated with 0.2% (w/w) commercial composite phosphate; and three experimental groups supplemented with APPs at concentrations of 2.0%, 4.0%, and 6.0% (w/w). These concentrations were selected based on preliminary experiments and previous literature [11], which indicated that levels below 2.0% offered limited cryoprotection, while levels exceeding 6.0% negatively impacted texture and cost-effectiveness. Additives were thoroughly mixed into the mince to ensure homogeneity. The mixtures were stuffed into casings and cooked using a two-stage steaming process: 55 °C for 30 min, followed by 75–80 °C for 35 min, ensuring a final core temperature above 72 °C. After cooling to room temperature, the sausages were sectioned for analysis. Freshly cooked samples (0 freeze–thaw cycles) were analyzed immediately. The remaining samples were subjected to five repeated freeze–thaw cycles, with each cycle consisting of freezing at −18 °C for 24 h and thawing at 4 °C for 12 h. Analyses were conducted after the 1st, 3rd, and 5th cycles, with all measurements performed in triplicate.

2.4. Determination of THA

The THA of APPs was assessed using differential scanning calorimetry (DSC 214, NETZSCH, Selb, Germany) following the procedure of Zhang et al. [16]. Prior to analysis, the DSC instrument was calibrated for temperature and heat flow using high-purity metal standards, including indium (In), tin (Sn), lead (Pb), and zinc (Zn). A 5 µL aliquot of APPs solution (20 mg/mL) was loaded and hermetically sealed in an aluminum crucible. The sample was cooled from room temperature to −20 °C at 5 °C/min, held for 2 min, and then heated to 20 °C at the same rate to determine the enthalpy of fusion (ΔH_m). For THA measurement, the sample was cooled to −20 °C, held for 2 min, and then warmed to a target isothermal temperature (T_h) where ice crystals coexisted with the unfrozen solution. After equilibration at T_h for 5 min, the sample was cooled again to −20 °C at a rate of 1 °C/min. This cycle was repeated at various T_h temperatures. The THA value and ice crystal fraction (Ф) were calculated using Equations (1) and (2), respectively.

$$\mathrm{T}\mathrm{H}\mathrm{A}=T_h-T_o$$

(1)

$$\begin{array}{c}\Phi \left(\%\right)=\left(1-{\displaystyle \frac{\mathsf{\Delta }{H}_r}{\mathsf{\Delta }{H}_m}}\right)\times 100\%\end{array}$$

(2)

where THA is thermal hysteresis activity (°C); T_h is holding temperature (°C); T_o is onset temperature (°C); Φ is the amount of ice crystals (%); ΔH_r is exothermic enthalpy of freezing (J/g); ΔH_m is melting enthalpy (J/g).

2.5. IRI Activity

IRI activity was evaluated using a polarized light microscope (Leica DM4P, Leica Microsystems, Wetzlar, Germany) equipped with a programmable thermal stage [17]. A working solution of APPs (5 mg/mL) was prepared in 23% (w/v) sucrose, with pure 23% sucrose serving as the control. A 30 μL droplet of each solution was placed on a glass slide and covered with a coverslip. The thermal program involved rapidly cooling the sample to −25 °C at 20 °C/min, holding for 20 min, and then warming to −16 °C at 5 °C/min. Subsequently, the sample underwent five annealing cycles between −10 °C and −16 °C (1 °C/min), with a 1 min hold at −16 °C at the end of each cycle. Ice crystal morphology was recorded, and the mean grain size (MGS) was quantified using ImageJ software version 1.54g (National Institutes of Health, Bethesda, MD, USA). MGS was expressed as a percentage relative to the control; a lower MGS value indicates superior IRI activity.

2.6. Determination of Thawing Loss

Thawing loss was determined according to the procedure described by Kosim et al. [18] with slight modifications. Briefly, frozen sausage samples from designated freeze–thaw cycles were weighed (m₁), then thawed at 4 °C until the core temperature reached 4 °C. Surface exudate was carefully removed by blotting with filter paper, and samples were weighed again (m₂). Thawing loss was calculated using the following Equation (3):

$$\mathit{Thawing} \mathit{loss} (\%) = {\displaystyle \frac{m_1 - m_2}{m_1}} \times 100\%$$

(3)

2.7. Determination of Cooking Loss

Cooking loss was determined following the method of Adzitey et al. [19]. Briefly, thawed sausage samples were blotted dry to remove surface moisture and weighed (m₃), and cooked in a water bath at 100 °C until its core temperature reached 75 °C. After cooking, the samples were cooled to room temperature (approximately 25 °C), blotted dry again to remove any exudate, and re-weighed (m₄). Cooking loss was calculated using the following Equation (4):

$$\mathit{Cooking} \mathit{loss} (\%)={\displaystyle \frac{m_3-m_4}{m_3}}\times 100\%$$

(4)

2.8. Determination of Water-Holding Capacity (WHC)

WHC was measured according to the method of Mehta et al. [20] with minor modifications. Briefly, sausage samples were cut into 4 mm thick slices, weighed (m₅), and wrapped in filter paper. The wrapped samples were then centrifuged at 6000× g for 10 min at 4 °C. After centrifugation, the filter paper was removed, and the sample was weighed again (m₆). WHC was calculated using the following Equation (5):

$$W_{HC}=\left(1-{\displaystyle \frac{m_5-m_6}{m_5}}\right)\times 100\%$$

(5)

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

Water distribution and mobility in the sausages were analyzed using LF-NMR according to the method of Tang et al. [21]. Measurements were performed with a 21.3 MHz LF-NMR analyzer (PQ001, Shanghai Niumag Analytical Instrument Co., Shanghai, China). Cooked sausage samples were cut into uniform cylinders (0.5 cm diameter × 1.5 cm height) and placed into NMR tubes. The transverse relaxation time (T₂) was measured using the Carr–Purcell–Meiboom–Gill (CPMG) pulse sequence with the following acquisition parameters: time waiting (TW) = 4000 ms, echo time (TE) = 0.6 ms, number of echoes (NECH) = 1600, and number of scans (NS) = 8. The obtained data were analyzed using the Niumag Multi-ExpInv Analysis software version 1.0 (Niumag Co., Ltd., Shanghai, China).

2.10. Texture Profile Analysis (TPA)

TPA was performed using a texture analyzer (Model TA.XTC-18, Bosin, Shanghai, China) equipped with a TA/36 cylindrical probe, following a method adapted from Vilcapoma et al. [22]. Prior to analysis, the casings of the sausage samples were removed, and the samples were cut into uniform cylinders (10 mm in height). A two-cycle compression test was conducted to 30% of the original sample height. Pre-test and post-test speeds were 3.0 mm/s, while the test speed was 1.0 mm/s, with a trigger force of 5 g. Five replicates were measured per group.

2.11. Lipid Oxidation

Lipid oxidation was evaluated using the thiobarbituric acid reactive substances (TBARS) assay according to the method of Lin et al. [23] with slight modifications. Briefly, 5 g of minced sausage sample was homogenized with 10 mL of 20% (w/v) trichloroacetic acid (TCA) for 1 min and centrifuged at 3212× g for 15 min at 4 °C. The supernatant was collected and filtered through qualitative filter paper. Subsequently, 5 mL of the filtrate was mixed with 5 mL of 0.02 M thiobarbituric acid (TBA) solution. A reagent blank was prepared by mixing 5 mL of 20% TCA with 5 mL of the TBA solution. All mixtures were incubated in a boiling water bath (100 °C) for 20 min, cooled, and absorbance was measured at 532 nm. TBARS values were calculated using the following Equation (6):

$$TBARS\left(\mathrm{m}\mathrm{g}/\mathrm{k}\mathrm{g}\right)=\left(A_{532}+0.002\right)\times 2.587$$

(6)

2.12. Determination of Carbonyl Content in Myofibrillar Protein (MP)

MP extraction was performed based on the method described by Chen et al. [24], and carbonyl content was determined according to the procedure of Zhang et al. [25]. MP was extracted from minced sausage samples by homogenization in 6 volumes (v/w) of ice-cold phosphate-buffered saline (PBS, pH 7.0), followed by centrifugation at 2000× g for 15 min at 4 °C. The pellet was washed twice with PBS. Connective tissue was removed by filtration through four layers of gauze, followed by a final wash with 0.1 M NaCl. Proteins were precipitated by adjusting the pH to 6.0 with 0.1 M HCl and centrifuged at 8000× g for 8 min. The final pellet was resuspended in PBS. For the carbonyl assay, 0.5 mL of the MP solution was mixed with 0.5 mL of 10 mmol/L DNPH in 2 mol/L HCl. Control samples used 2 mol/L HCl without DNPH. All mixtures were incubated in the dark at room temperature for 1 h. Subsequently, proteins were precipitated by adding 0.5 mL of 20% (w/v) trichloroacetic acid (TCA) and centrifuged at 13,000× g for 5 min at 4 °C. The pellets were washed three times with a 1:1 (v/v) ethanol-ethyl acetate mixture to remove unreacted DNPH. The final pellets were dissolved in 1.5 mL of 6 mol/L guanidine hydrochloride (prepared in 2 mol/L HCl), and incubated at 37 °C for 30 min. After centrifugation (13,000× g, 5 min), the absorbance of the supernatant was measured at 370 nm. Protein carbonyl content was calculated using the following Equation (7) with a molar extinction coefficient of 22,000 L·mol⁻¹·cm⁻¹:

$$Carbonylcontent\left(\mathrm{n}\mathrm{m}\mathrm{o}\mathrm{l}/\mathrm{m}\mathrm{g} \mathrm{p}\mathrm{r}\mathrm{o}\mathrm{t}\mathrm{e}\mathrm{i}\mathrm{n}\right)={\displaystyle \frac{A_{370}-A_{control}}{\mathrm{22,000}\times C}}\times {10}^6$$

(7)

where C is the protein concentration (mg/mL) of the MP solution, which was determined using the bicinchoninic acid (BCA) assay.

2.13. Determination of Total Volatile Basic Nitrogen (TVB-N)

TVB-N content was determined using a modified procedure based on the method of Lu et al. [26]. Briefly, 10 g of minced sausage sample was homogenized with 75 mL of distilled water and allowed to stand for 30 min. Subsequently, the homogenate was transferred to a distillation tube containing 1 g of magnesium oxide (MgO). Steam distillation was carried out using an automatic Kjeldahl nitrogen analyzer (Model SKD-1000, Peiou Analytical Instrument Co., Ltd., Shanghai, China). The liberated ammonia was distilled and absorbed into a boric acid solution. The absorbed ammonia was then titrated with a standardized HCl solution. The TVB-N value was calculated based on the volume of HCl consumed during titration and expressed as milligrams of nitrogen per 100 g of sample using the following Equation (8):

$$TVB-N\left(\mathrm{m}\mathrm{g}/100 \mathrm{g}\right)={\displaystyle \frac{{(V}_1-V_2)\times C\times 14}{m}}\times 100$$

(8)

where V₁ is the volume of HCl consumed by the sample (mL); V₂ is the volume of HCl consumed by the blank (mL); C is the concentration of HCl (mol/L); and m is the mass of the sample (g)

2.14. Statistical Analysis

With the exception of the progressive THA threshold determination (which was conducted as a single continuous profiling assay), all other measurements were performed at least in triplicate, and the data are presented as mean ± standard deviation (SD). Statistical analysis was conducted using SPSS version 27.0 (IBM Corp., Armonk, NY, USA). Prior to the analysis, the normality of the data distribution and the homogeneity of variances were verified using the Shapiro–Wilk test and Levene’s test. Multiple group comparisons were performed using one-way analysis of variance (ANOVA), followed by Tukey–Kramer post hoc test to determine significant differences among the treatment groups. Statistical significance was defined at p < 0.05.

3. Results and Discussion

3.1. In Vitro Antifreeze Activity: THA and IRI Evaluation

To quantify the intrinsic cryoprotective potential, the THA and IRI activities of APPs were characterized in vitro. Differential Scanning Calorimetric (DSC) analysis (Figure 1) revealed a distinct thermal hysteresis gap. As shown in

Table 1. The T_o, ΔH_m, ΔH_r, Φ and THA of differential scanning calorimetry (DSC) curves of APP solutions at various holding temperatures (T_h).

Th (°C)	To (°C)	△Hm (J/g)	△Hr (J/g)	Φ (%)	THA (°C)
−0.34	−0.59	302.5	128.7	57.45	0.25
−0.14	−0.51	302.5	200.4	33.75	0.37
0.06	−0.45	302.5	266	12.07	0.51
0.26	-	302.5	-	-	-

T_h, holding temperature; T_o, onset temperature; ΔH_m, melting enthalpy; ΔH_r, exothermic enthalpy of freezing; Ф, the amount of ice crystal; THA, thermal hysteresis activity.

, elevating the holding temperature (T_h) from −0.34 °C to 0.06 °C resulted in a sharp decline in ice crystal content (Φ) within the APPs solutions, decreasing from 57.45% to 12.07%. This reduction was accompanied by a pronounced expansion of the recrystallization exotherm. At a T_h of 0.06 °C, the APPs solution demonstrated a THA of 0.51 °C. This magnitude notably exceeds the typical activity range of many plant-derived antifreeze proteins and aligns with the functional scope reported for moderately active fish AFPs [16], thereby classifying APPs as an effective bio-sourced cryoprotective agent.

Polarized light microscopy provided direct visual evidence of ice recrystallization inhibition. In the control group (23% sucrose solution), ice crystals underwent considerable coarsening through Ostwald ripening [17], developing into large, irregular structures. Quantitative analysis indicated that after five annealing cycles, the MGS increased by approximately 79.9% relative to the first cycle (Figure 2A,C,E). In contrast, samples supplemented with APPs maintained a fine and homogeneous dispersion of micro-crystals throughout the thermal cycling process. The growth of ice crystals was effectively restrained; the MGS after the fifth cycle was limited to approximately 60.3% larger than that measured after the initial cycle (Figure 2B,D,E). Moreover, ice crystals in the APPs-treated group exhibited a distinct rounded, non-faceted morphology with thinner crystal boundaries compared to the control. This implies that APPs adsorb onto specific ice planes via the Kelvin effect, increasing the energy barrier for water molecule incorporation and thereby effectively altering the inherent Ostwald ripening kinetics [27].

3.2. Water-Holding Capacity and Moisture Mobility in Frozen Sausages

To ascertain whether the robust in vitro antifreeze activity translates to functional quality preservation in complex food matrices, a systematic evaluation was conducted on frozen pork sausages, with an initial focus on myofibrillar integrity and moisture stability.

Freeze–thaw cycling induced severe moisture loss in control samples, where thawing loss escalated from 3.36% to 7.51% over five cycles (Figure 3A). This degradation is attributed to the mechanical rupture of muscle cells by recrystallized ice, leading to exudation [28]. Conversely, the inclusion of APPs demonstrated a pronounced, concentration-dependent protective effect. The 6% inclusion level was identified as the most effective concentration among this tested, significantly suppressing thawing loss to 4.87% by the fifth cycle, outperforming both the positive control (0.2% composite phosphates, 5.61%) and other APPs concentrations (Figure 3A, p < 0.05). This indicates that APPs retain robust ice-binding capability even within the high-ionic-strength environment of processed sausages (approx. 1.1% salt), effectively overcoming electrostatic screening effects that typically deactivate conventional peptides. Consequently, 6% concentration of APPs provides superior mitigation of cellular damage.

This protective trend extended to thermal stability and gel integrity. Cooking loss, a direct indicator of product juiciness [29], was most effectively minimized by 6% APPs (14.59% vs. 22.82% in controls after 5 cycles; Figure 3B). Similarly, WHC analysis (Figure 3C) confirmed that while all groups experienced a decline, the 6% APPs treatment maintained the highest retention (72.47% after 5 cycles), significantly higher than the control (64.74%) and phosphate groups (69.81%). These results collectively suggest that APPs stabilizes the myofibrillar network, enhancing the entrapment of capillary water during both frozen storage and heat-induced shrinkage [30]. Notably, within the tested concentration range (2–6%), the protective efficacy of APPs exhibited a clear positive correlation with peptide concentration, culminating in the most effective concentration in this test at 6%. This dose-dependent behavior may be attributed to the molecular nature of APPs as bioactive peptides. It is plausible that higher peptide concentrations increase surface coverage on ice crystal or enhance interactions with myofibrillar proteins, thereby more effectively inhibiting ice recrystallization and stabilizing the water-protein network.

LF-NMR provided mechanistic insight into these macroscopic observations by tracking water populations: tightly bound (T_2b, 0.1–10 ms), immobilized (T₂₁, 10–100 ms), and free water (T₂₂, 100–1000 ms) (Figure 4A–D). In control samples, repeated freeze–thaw cycles drove a significant shift towards longer relaxation times in water distribution: the proportion of immobilized water (P₂₁) dropped from 94.56% to 88.50%, while free water (P₂₂) increased from 4.56% to 11.50% (p < 0.05) (

Table 2. T₂ relaxation times (T_2b, T₂₁, T₂₂) and corresponding water population proportions (P_2b, P₂₁, P₂₂) in sausage samples treated with different concentrations of APPs at different freeze–thaw stages.

F/T	Samples	T2b (ms)	T21 (ms)	T22 (ms)	P2b	P21	P22
0	Control	1.10 ± 0.11 a	31.44 ± 0.00 a	391.92 ± 15.89 c	0.88% ± 0.46% a	94.56% ± 0.53% d	4.56% ± 0.12% a
	2% APPs	1.05 ± 0.07 a	30.03 ± 1.22 b	471.38 ± 0.00 b	0.49% ± 0.12% a	94.67% ± 0.20% d	4.84% ± 0.27% a
	4% APPs	1.02 ± 0.11 a	29.33 ± 0.00 b	554.57 ± 22.48 a	1.24% ± 0.50% a	95.97% ± 0.44% c	2.79% ± 0.06% b
	6% APPs	0.73 ± 0.11 b	26.75 ± 1.06 c	494.78 ± 40.54 b	0.96% ± 0.48% a	97.67% ± 0.46% a	1.36% ± 0.14% c
	0.2% Phosphates	0.96 ± 0.17 a	27.36 ± 0.00 c	402.37 ± 41.90 c	0.58% ± 0.34% a	96.72% ± 0.47% b	2.70% ± 0.13% b
1	Control	1.30 ± 0.16 a	33.70 ± 0.00 a	450.30 ± 18.25 a	0.32% ± 0.32% d	92.67% ± 0.46% e	7.01% ± 0.26% a
	2% APPs	1.29 ± 0.00 a	33.70 ± 0.00 a	472.13 ± 32.76 a	0.68% ± 0.56% cd	93.43% ± 0.28% d	5.90% ± 0.40% b
	4% APPs	1.22 ± 0.37 a	33.70 ± 0.00 a	366.82 ± 39.48 b	2.01% ± 0.26% a	95.28% ± 0.41% c	2.70% ± 0.43% c
	6% APPs	1.12 ± 0.08 a	31.44 ± 0.00 b	472.13 ± 32.76 a	1.53% ± 0.54% ab	97.27% ± 0.59% a	1.20% ± 0.07% d
	0.2% Phosphates	1.18 ± 0.18 a	32.95 ± 1.30 a	357.65 ± 24.81 b	1.17% ± 0.24% bc	96.10% ± 0.39% b	2.73% ± 0.28% c
3	Control	1.52 ± 0.06 a	38.72 ± 0.00 a	341.11 ± 13.83 bc	1.25% ± 1.15% ab	90.23% ± 1.21% d	8.52% ± 0.20% a
	2% APPs	1.48 ± 0.62 a	38.72 ± 0.00 a	357.08 ± 0.00 b	0.85% ± 0.52% ab	91.33% ± 0.64% cd	7.82% ± 0.13% b
	4% APPs	1.48 ± 0.50 a	37.85 ± 1.50 a	326.22 ± 26.73 cd	1.61% ± 0.70% a	93.17% ± 0.82% b	5.22% ± 0.31% c
	6% APPs	1.30 ± 0.25 a	36.12 ± 0.00 b	303.84 ± 12.03 d	0.85% ± 0.60% ab	95.05% ± 0.50% a	4.10% ± 0.17% d
	0.2% Phosphates	1.48 ± 0.00 a	36.12 ± 0.00 b	410.27 ± 0.00 a	0.07% ± 0.03% b	92.02% ± 0.17% bc	7.91% ± 0.18% b
5	Control	-	44.49 ± 0.00 a	410.27 ± 0.00 a	-	88.50% ± 0.28% d	11.50% ± 0.28% a
	2% APPs	-	41.50 ± 0.00 b	252.35 ± 0.00 c	-	90.13% ± 0.06% c	9.87% ± 0.06% b
	4% APPs	1.61 ± 0.62 a	41.50 ± 0.00 b	241.07 ± 9.77 c	0.08% ± 0.11% a	92.37% ± 0.39% b	7.55% ± 0.43% c
	6% APPs	1.54 ± 0.31 a	39.65 ± 1.61 c	200.66 ± 16.44 d	0.22% ± 0.20% a	93.67% ± 0.28% a	6.11% ± 0.10% d
	0.2% Phosphates	1.61 ± 0.28 a	39.65 ± 1.61 c	333.66 ± 23.15 b	0.20% ± 0.29% a	90.62% ± 0.16% c	9.18% ± 0.31% b

Means without a common letter indicate significant differences between different treatment groups within the same freeze–thaw cycle (p < 0.05). F/T, freeze–thaw.

). This shift signifies severe structural damage, where water originally entrapped within the myofibrillar lattice is liberated due to ice crystal growth [31].

In contrast, APPs effectively mitigated this water migration. The 6% APPs group demonstrated the most potent stabilization effect, maintaining P₂₁ at 93.67% and restricting P₂₂ to only 6.11% even after five cycles. Its water distribution profile most closely resembled the fresh state and was significantly superior to all other treatments (p < 0.05). Furthermore, regarding relaxation times, the tightly bound water signal (T_2b), which became undetectable in both the control and the 2% APPs group by the fifth cycle, was still preserved in samples treated with 4% and 6% APPs. This critical distinction highlights the superior capability of higher APPs concentrations in protecting the protein-water interface throughout extensive freeze–thaw stress, thereby effectively inhibiting the migration of immobilized water (T₂₁) to free water (T₂₂) [32]. These LF-NMR findings confirm that APPs specifically stabilize the hydration shell of actomyosin against freeze-concentration induced desorption. Unlike phosphates, which primarily rely on ionic strength to chemically induce myofibrillar swelling and osmotic water retention [33], APPs function as physical ‘Ice-Structure Stabilizers’ preserving the protein network by directly inhibiting ice recrystallization at the molecular interface.

3.3. Mitigation of Freeze–Thaw Induced Textural Deterioration

TPA quantitatively assessed the structural integrity of the sausage matrix under freeze–thaw stress. In the control group, hardness declined from 3713.20 g to 2916.73 g after five cycles, with resilience decreasing from 0.48 to 0.44 (

Table 3. Effects of APPs on hardness, springiness, and resilience of sausages treated with different concentrations of APPs at different freeze–thaw stages.

F/T	Samples	Hardness	Springiness	Resilience
0	Control	3713.2 ± 304.56 b	0.86 ± 0.01 ab	0.48 ± 0.02 ab
	2% AFPs	3858.16 ± 128.22 ab	0.86 ± 0.01 ab	0.47 ± 0.02 ab
	4% AFPs	3729.59 ± 174.96 b	0.82 ± 0.02 b	0.45 ± 0 ab
	6% AFPs	3913.57 ± 166.63 ab	0.86 ± 0.02 ab	0.44 ± 0.04 b
	0.2% Phosphates	4214.75 ± 114.94 a	0.89 ± 0.03 a	0.49 ± 0.03 a
1	Control	3499.77 ± 346.92 ab	0.85 ± 0.03 b	0.43 ± 0.01 b
	2% AFPs	3113.89 ± 468.68 b	0.83 ± 0.02 b	0.43 ± 0 b
	4% AFPs	3648.94 ± 325.86 ab	0.84 ± 0.03 b	0.42 ± 0.01 b
	6% AFPs	3227.31 ± 291.62 ab	0.83 ± 0.02 b	0.38 ± 0.01 c
	0.2% Phosphates	3920.45 ± 252.45 a	0.9 ± 0.01 a	0.45 ± 0.02 a
3	Control	2967.24 ± 311.58 a	0.85 ± 0.02 a	0.43 ± 0.01 ab
	2% AFPs	3045.33 ± 440.77 a	0.84 ± 0.03 a	0.41 ± 0.01 b
	4% AFPs	3180.59 ± 204.12 a	0.83 ± 0.03 a	0.42 ± 0.01 b
	6% AFPs	2900.97 ± 174.07 a	0.86 ± 0.04 a	0.38 ± 0.02 c
	0.2% Phosphates	3383.66 ± 15.27 a	0.89 ± 0.01 a	0.45 ± 0.02 a
5	Control	2916.73 ± 265.58 a	0.86 ± 0.02 abc	0.44 ± 0.01 a
	2% AFPs	2731.88 ± 216.09 a	0.87 ± 0.01 ab	0.40 ± 0.01 bc
	4% AFPs	2572.25 ± 286.50 a	0.83 ± 0.03 bc	0.38 ± 0.01 cd
	6% AFPs	2947.49 ± 79.47 a	0.83 ± 0.02 c	0.37 ± 0.03 d
	0.2% Phosphates	2881.4 ± 299.01 a	0.88 ± 0.01 a	0.43 ± 0.01 ab

Means without a common letter indicate significant differences between different treatment groups within the same freeze-thaw cycle (p < 0.05). F/T, freeze-thaw.

). This degradation reflects the mechanical disruption of the myofibrillar network by ice crystals [34].

The textural influence of APPs was concentration-dependent and parameter-specific. At Cycle 5, none of the APPs treatments (2–6%) showed a significant advantage in hardness over the control. However, a clear dose-dependent reduction was observed for resilience, which decreased with increasing APPs concentration, with the 6% group recording the lowest value (0.37). This dose-dependent decrease in resilience implies a “plasticizing effect” of the small-molecule peptides [35]. While APPs excel as “Ice-Structure Stabilizers” to maintain microstructural integrity (as evidenced by WHC), their intercalation within the myofibrillar matrix may weaken the protein-protein cross-links (e.g., disulfide bonds) required for elastic recovery [36].

The phosphate group exhibited a distinct mechanistic profile. It demonstrated superior springiness and resilience throughout the study, consistent with its chemical role in enhancing protein solubilization and gel elasticity [33]. However, its initial advantage in hardness was not sustained; by Cycle 5, its hardness (2881.40 g) was comparable to the control and APPs groups. This highlights a fundamental mechanistic divergence: phosphates rely on chemical gelation to reinforce elasticity, whereas APPs provide physical defense against ice-mediated structural damage. The inability of phosphates to sustain hardness suggests their limited efficacy in controlling ice growth compared to the direct IRI activity of APPs.

In summary, APPs and phosphate employ complementary preservation strategies. Phosphate excels in chemically preserving elasticity, whereas the potential of APPs lies in physical ice inhibition. The concentration-dependent trade-off observed with APPs highlights the need for optimized application strategies to balance protection with textural quality.

3.4. Suppression of Lipid Oxidation and Protein Degradation

The chemical stability of the sausage matrix was evaluated by quantifying TBARS (lipid oxidation), protein carbonyls (protein oxidation), and TVB-N (proteolytic degradation). Following five freeze–thaw cycles, control samples exhibited significant oxidative degradation. TBARS values increased sharply from 0.19 to 3.71 mg MDA/kg, while protein carbonyl levels rose from 0.81 to 3.71 nmol/mg (Figure 5A,B). This progression is attributed to the physical disruption of cellular compartments (e.g., mitochondria and lysosomes) by ice crystals. In this heme-rich red meat matrix, such rupture releases free iron and heme pigments, acting as potent pro-oxidants to accelerate the oxidative cascade [37,38,39,40].

Supplementation with APPs attenuated oxidative damage in a concentration-dependent manner, though its efficacy relative to phosphate varied by indicator. For lipid oxidation, APPs treatment consistently yielded lower TBARS values than the control at Cycle 5, with the 6% group (3.34 mg MDA/kg) showing the strongest effect among APPs concentrations (Figure 5A). However, the phosphate group demonstrated superior inhibition, maintaining the lowest TBARS value (2.70 mg MDA/kg). A clearer advantage for APPs was observed in mitigating protein carbonylation. The 6% APPs formulation limited carbonyl content to 3.08 nmol/mg at Cycle 5, significantly lower than the control (3.71 nmol/mg) (Figure 5B). This suggests peptides of APPs may effectively scavenge radicals involved in protein oxidation [41].

Interpretation of TVB-N values necessitates distinguishing between exogenous nitrogen input and endogenous spoilage. The elevated baseline TVB-N in APPs-treated groups is attributed to the interference of the added nitrogenous peptides with the detection assay, rather than spoilage. Critically, the rate of TVB-N accumulation served as the true indicator of freshness preservation. In contrast to the control group (71.9% increase), the 6% APPs group significantly retarded this progression (46.4% relative increase), demonstrating that APPs effectively suppressed the generation of new volatile bases by inhibiting freeze–thaw induced proteolysis and microbial growth.

The antioxidative efficacy of APPs operates via a “Physical–chemical Cascade.” Primarily, by physically IRI, APPs maintain membrane integrity and compartmentalize endogenous pro-oxidants (heme iron), effectively cutting off the oxidation pathway at its source. Secondarily, the hydrophobic amino acid residues in APPs may provide direct radical scavenging activity. This proves that stabilizing the physical microstructure is a prerequisite for chemical stability in frozen meat [42]. While certain bioactive peptides are known to possess inherent antioxidant capacity, the contribution of such a direct chemical mechanism in the present system requires further verification [43]. In contrast, the superior performance of phosphate in inhibiting lipid oxidation may stem from its strong water-binding and ionic effects, which limit the mobility and reactivity of pro-oxidative substrates in the aqueous phase. In summary, APPs demonstrated a clear, concentration-dependent capacity to inhibit oxidative damage, especially against protein carbonylation. Its effect on lipid oxidation was significant but less pronounced than that of phosphate. The elevated TVB-N values were largely attributable to the peptide additive itself, and when accounting for this baseline, APPs still showed a mitigating effect on proteolytic progression during storage.

4. Conclusions

This study confirms that APPs possess intrinsic cryoprotective activity, characterized by distinct thermal hysteresis and potent ice recrystallization inhibition in vitro. In the high-ionic-strength matrix of frozen pork sausages, an inclusion level of 6% APPs was identified as the most effective concentration among those tested. At this concentration, APPs effectively stabilized the actomyosin hydration shell, significantly reducing thawing and cooking losses by inhibiting the migration of immobilized water (P₂₁) to free water (P₂₂). Furthermore, APPs validated a “physical–chemical cascade” mechanism by effectively suppressing protein oxidation. While a trade-off in textural resilience was observed due to a plasticizing effect, the superior moisture retention distinguishes APPs from the chemical gelation mechanism of phosphates. However, this study has certain limitations. While the inclusion of white, odorless APPs effectively maintained the physicochemical and microstructural stability of the sausages without causing noticeable visual degradation, formal sensory panel evaluations (e.g., precise flavor and color scoring) were not conducted. Future research should focus on formal sensory evaluation and the molecular interaction mechanisms between APPs and meat proteins. Collectively, these findings classify APPs as specialized “Ice-Structure Stabilizers” that function primarily via physical ice inhibition, offering a promising clean-label strategy for mitigating heme-iron-catalyzed oxidation and moisture loss in processed meat products.