Alleviation of structural deterioration in gluten and its components during freeze–thaw cycles by deacetylated konjac glucomannan

Abstract

The cryoprotective effects of deacetylated konjac glucomannan (DKGM) on gluten and its components in frozen dough remain unclear. This study aimed to investigate the impact of DKGM with varying degrees of deacetylation (DD) on the structural stability of gluten and its components during freeze–thaw (FT) cycles. Compared with konjac glucomannan, DKGM effectively alleviated the gluten structural “depolymerization–aggregation” process during FT cycles, with DK2 (DD, 50.21%) exhibiting the optimal cryoprotective effect. The DK2 group retained higher noncovalent interactions and disulfide bonds during FT cycles, thereby stabilizing the gluten structure. Studies on glutenin and gliadin suggest that glutenin is more susceptible to FT damage. DK2 provided the best protection for glutenin, whereas DK3 (DD, 66.61%) provided the most effective protection for gliadin. These distinct effects were likely attributable to differences in the particle size and steric hindrance of DKGM, as well as the inherent structural characteristics of the protein components.

Highlights

• •Gluten underwent a “depolymerization–aggregation” process during freeze–thaw cycles.
• •Deacetylated konjac glucomannan (DKGM) enhanced the structural stability of frozen gluten.
• •Glutenin is more vulnerable to freeze-thaw damage compared to gliadin.
• •DKGM with moderate DD (50.21%) optimally stabilizes the effect during freezing.

1. Introduction

The development of freezing technology has revolutionized the industrial production of dough products, enabling large-scale manufacturing and extended shelf life. However, temperature fluctuations can significantly degrade the quality of frozen dough during storage, primarily by affecting its gas retention and rheological properties (Jiang, Guo, Xing, & Zhu, 2023). This degradation is closely linked to the structural breakdown of gluten proteins, which serve as the “backbone” of the dough and are crucial for its quality (Zeng, Guan, Qin, Chen, & Liu, 2024). Gluten proteins mainly composed of glutenin and gliadin. During the formation of the gluten network, gliadin interacts with glutenin through hydrophobic interactions and disulfide bonds. In frozen storage, the gluten protein structure rearranges, and the covalent linkages are disrupted. High-molecular-weight glutenin tends to depolymerize, and gliadins exhibit the potential to disaggregate glutenin macropolymer. These changes thin the gluten film, impairing the quality of frozen dough(Wang et al., 2023).

Recent studies have proposed that adding hydrophilic polysaccharides is an effective, safe, and economical strategy to improve the cryostability of gluten proteins. Tang et al. (2025) revealed that low-ester pectin exhibited stronger inhibition of ice recrystallization than high-ester pectin, thereby mitigating gluten network deterioration during frozen storage. Similarly, Zhao et al. (2022) found that arabinoxylans with a higher arabinose/xylose ratio exhibited superior cryoprotective effects in dough. These results highlight the critical role of polysaccharide structural features in enhancing cryoprotective performance in gluten systems. Nevertheless, the influence of polysaccharides on the structural stability of glutenin and gliadin under freeze–thaw cycles remains insufficiently understood.

Konjac glucomannan (KGM), a neutral water-soluble polysaccharide, is widely recognized for its notable thickening, gelling, and emulsifying properties, which exhibit superior performance compared to conventional food gelling agents (He et al., 2020). However, KGM excessive hygroscopicity disrupted gluten network formation, predisposing dough to structural collapse during fermentation (He et al., 2020). Partial deacetylation of KGM yields deacetylated konjac glucomannan (DKGM), which exhibits reduced moisture sensitivity and improved mechanical strength (Wu et al., 2020). Recent work further revealed that DKGM showed promising cryoprotective potential in frozen dough, with efficacy dependent on the degree of deacetylation (DD) (Fan et al., 2025). Given that gluten proteins constitute the most critical component of dough, elucidating the interaction mechanisms between DKGM of varying DD and gluten proteins is of particular importance. Although previous studies reported that DKGM is associated with disulfide bond formation and enhanced ordered gluten structures (Cheng et al., 2024), research on the aggregation state of the gluten protein network and gluten component interactions with varying DD of DKGM remains limited.

This study investigated the effects of DKGM with varying DDs on the structural integrity of gluten and its components during freeze-thaw cycles, focusing on protein structure and polysaccharide-protein interactions. Our preliminary findings indicated that supplementation with 0.6% DKGM produced the greatest improvement in gluten protein stability, and this concentration was selected for further experiments. The work examined the influence of DKGM on the multidimensional structure and aggregation behavior of gluten proteins, and further elucidated interaction mechanisms by blending DKGM with glutenin or gliadin followed by freeze–thaw treatment. These findings provide new insights into the role of DKGM in modulating gluten structure and enhancing the cryostability of frozen gluten systems.

2. Materials and methods

2.1. Materials

Konjac glucomannan (glucomannan content greater than 90%) was sourced from Fairy Konjac Products Co., Ltd. (Shiyan, China). Gluten was purchased from Midaner Food Technology Co., Ltd. (Zhengzhou, China). Ethanol, dichloromethane and NaOH were acquired from Chuandong Chemical Co., Ltd. (Chongqing, China). 8-Anilinonaphthalene-1-sulfonic acid, acetic acid, Tris and glycine were purchased from Aladdin Biochemical Co., Ltd. (Shanghai, China). EDTA, NaCl, urea and bromophenol blue dye were obtained from Macklin Biochemical Co., Ltd. (Shanghai, China). All chemicals and solvents were of analytical reagent grade or higher.

2.2. Preparation of DKGM

The KGM (12.0 g) was blended with 30 vol% ethanol (360.0 mL) in a water bath stirrer for 30 min at 50 °C. Then, 12.0 mL, 20.0 mL, and 28.0 mL NaOH solutions (0.1 mol/L) were added to the mixture for 30 min. Subsequently, 30 vol% and pure ethanol were added to the mixture separately. After evaporating ethanol, the DKGM powder was obtained by vacuum freeze-drying.

The DD of DKGM was determined as described by Li, Cai, Liu, and Liu (2023). Three DKGM of different DD were obtained: DK1 (DD, 31.31%), DK2 (DD, 50.24%), and DK3 (DD, 66.61%), respectively.

2.3. Extraction of wheat glutenin and gliadin

The extraction method of glutenin and gliadin was according to Wang et al. (2014). Firstly, gluten (20.0 g) was stirred with dichloromethane (250.0 mL) twice for 1 and 1.5 h at room temperature to defat. After adding dichloromethane, the defatted gluten was mixed with 60 vol% ethanol (300.0 mL) twice. Then, the mixture was centrifuged at 4510 ×g for 10 min at room temperature. The sediment was glutenin, combined supernatants after vacuum rotary evaporation at 40 °C were obtained as gliadin. Finally, all samples were lyophilized and sieved with 100 sieves.

The moisture contents of glutenin and gliadin were 7.81% and 7.63% (dry basis, w/w), respectively. The crude protein contents of glutenin and gliadin were 79.14% and 89.59% (dry basis, w/w), respectively.

2.4. Preparation of different gluten samples

The preparation of different gluten samples was described by Li et al. (2024) with necessary modifications. Gluten, glutenin, and gliadin samples containing 0.6% (w/w) KGM or DKGM were dissolved in water at a mass ratio of 2:3. A freeze-thaw (FT) cycle was set as follows: the samples were frozen at −18 °C for 20 h and thawed at 30 °C for 4 h. The samples underwent 0, 1, 2, and 3 FT cycles, labeled as FT-0, FT-1, FT-2, and FT-3. The sample without KGM or DKGM was named the CK group.

2.5. Determination of the intrinsic fluorescence spectra of gluten protein

According to the method of Chen et al. (2023), an F-4700 fluorescence spectrometer (Hitachi, Tokyo, Japan) was used to measure the intrinsic fluorescence spectra of gluten. Gluten samples were diluted to 0.8 mg/mL with deionized water. The parameters were set as follows: excitation wavelength, 280 nm; emission wavelength, 300–500 nm; slit width, 5 nm.

2.6. Measurement of surface hydrophobicity of gluten protein

According to reported methods (Cui, Chen, & Zhang, 2023), a fluorescent probe, 8 mM 8-anilinonaphthalene-1sulfonic acid (ANS), was used to determine the surface hydrophobicity of gluten. An 80 mg sample was dispersed in 10 mL of 0.01 M phosphate buffer (pH 7.0). The dispersion was centrifuged at 1760 ×g for 10 min. The supernatant was subjected to a series of dilutions with phosphate buffer. Then, 4 mL of the diluted supernatant was mixed with 20 μL of ANS solution and incubated for 15 min. Using a Microplate reader (Epoch2, BioTek Co, Ltd., VT, USA) at excitation and emission wavelengths of 390 and 470 nm to record the fluorescence intensity. A curve of fluorescence intensity versus protein concentration was plotted, and the initial slope of the curve was taken as the surface hydrophobicity. Every sample was determined in four replicates.

2.7. Atomic force microscopy (AFM) of gluten protein

Gluten powder was dissolved in 0.05 mol/L acetic acid to obtain a protein concentration of 1.5 μg/mL. The solution was shaken at 180 rpm (SHA-BA, LangYue Co., Ltd., Changzhou, China) for 12 h at 25 °C, then the supernatant was removed by centrifugation (3000 ×g, 10 min). In tapping mode, the scanning range and speed of AFM (Bruker Co, Ltd., MA, USA) were 10 × 10 μm and 1 Hz, respectively.

2.8. Confocal laser scanning microscopy (CLSM) of gluten protein

Using CLSM (LSM780, Zeiss Co., Ltd., Oberkochen, Germany) to characterize the microstructure of gluten (Liu et al., 2023). The samples were sliced into 30 μm thick sections, and then the samples were stained with Rhodamine B (0.1%, w/v) and fluorescent brightener (0.1%, w/v) for 15 min. Rhodamine B and fluorescent brightener were excited at wavelengths of 559 and 405 nm, respectively. The average pore size and porosity were measured using Image J version Fiji (National Institutes of Health, MD, USA).

2.9. Interaction force test

2.9.1. Determination of free sulfhydryl (SH_free) content in gluten and its components

The SH_free content of the sample was determined using the method of Zhou et al. (2022) with slight modifications. Firstly, two kinds of solutions were prepared and marked as solution A and B. Solution A: 0.086 mol/L Tris, 0.09 mol/L glycine, 0.005 mol/L EDTA, and adjusted to pH 8. Solution B: 4 mg DTNB (5,5′-dithiobis-2-nitrobenzoic acid) solution was dissolved in 1 mL of solution A. Then, 400 mg of samples were suspended in 4 mL of solution A and 40 μL of solution B. The mixture was stirred at 25 °C for 1 h, followed by centrifugation at 10142 ×g for 10 min. Finally, the supernatants were determined by the Microplate reader (Epoch2, Biotech Co, Ltd., VT, USA) at 412 nm absorbance. SH_free content was calculated using the Eq. (1).

$${\mathrm{SH}}_{\text{free}}\left(\text{μmoL}/\mathrm{g}\right)=73.53\times A_{412}\times D/C$$ (1)

where 73.53 is the coefficient; A₄₁₂ is the absorbance at 412 nm; D is the dilution factor; C is the concentration (mg/mL). Each sample was conducted at least three times.

The increased rate of SH_free was calculated using the Eq. (2).

$$\text{The increased rate of}{\mathrm{SH}}_{\text{free}}\left(\%\right)=\left({\mathrm{SH}}_{\text{free}3}-{\mathrm{SH}}_{\text{free}0}\right)/{\mathrm{SH}}_{\text{free}0}\times 100$$ (2)

where ${\mathrm{SH}}_{\text{free}3}$ and ${\mathrm{SH}}_{\text{free}0}$ were the SH_free content at FT-3 and FT-0, respectively.

2.9.2. Determination of non-covalent interactions in gluten and its components

The non-covalent interactions include ionic bonds, hydrogen bonds, and hydrophobic interactions, which were determined by Kuang et al. (2023). Three solvents were prepared and marked as S₁, S₂, and S₃: 0.05 mol/L NaCl (S₁), 0.6 mol/L NaCl (S₂), 0.6 mol/L NaCl +1.5 mol/L urea (S₃). Different gluten samples (0.2 g) were suspended in 10 mL of specific solvents for 1 h. Subsequently, the solution was centrifuged (3000 ×g, 5 min) and the solubility of proteins in the supernatant was quantified. Five replicates were carried out for each determination.

2.10. Fourier-transform infrared spectra of different gluten components

The Fourier-transform infrared spectra were evaluated with Spectrum 100 (PerkinElmer Co., Ltd., MA, USA). In the 600–4000 cm⁻¹ range, 14 scans were taken with a resolution of 4 cm⁻¹ (Chen et al., 2023). Secondary structures were resolved through second-derivative deconvolution of the amide I band (1600–1700 cm⁻¹). The amide I bands were assigned as follows: α-helix, 1650–1660 cm⁻¹; β-sheet, 1600–1640 cm⁻¹; β-turn, 1660–1700 cm⁻¹; and random coil,1640–1650 cm⁻¹ (Fan et al., 2025). All samples were conducted at least three times.

2.11. Subunit distribution analysis of gluten protein and its components

Subunit distribution of proteins was conducted following the previous method (Li et al., 2024). Briefly, the extracted samples (10 mg, obtained after 12 h of extraction at 25 °C) were mixed with 4 mL of SDS-PAGE sample buffer (0.125 mol/L), followed by boiling for 5 min. After centrifugation, the supernatant was collected and combined with bromophenol blue dye at a 1:4 (v/v) ratio for subsequent protein electrophoretic analysis. Consequently, 20 μL of supernatant was loaded into subsequent pockets. After electrophoresis, gel images were collected by a gel imager (Fusion Solo S, Vilber Co., Ltd., Paris, France). The band gray intensity was analyzed by ImageJ version Fiji.

2.12. Docking simulation between DKGM and protein

The structures of KGM and DKGM with high, medium, and low DD (designated HDKGM, MDKGM, and LDKGM, respectively) were constructed using GaussView 5.0.9 software. The structures of glutenin (PDB ID: P10387) and gliadin (PDB ID: P04726) were retrieved from the PubChem database (http://pubchem.ncbi.nlm.nih.gov/). AutoDock Tools were employed to prepare the protein and KGM/DKGM structures: hydrogen atoms were added to the proteins, and water molecules were removed. Molecular docking was conducted using AutoDock Vina 1.1.2 software to explore interactions between the proteins and DKGM. The optimal docking conformation was determined by comparing the scores of the docking results. Finally, PyMOL and Discovery Studio 2019 software were used to visualize the results, generating interaction analysis diagrams depicting the ligand and key residues.

2.13. Statistical analysis

Results were expressed as mean ± standard deviation. Statistical differences (p < 0.05) were analyzed using one-way analysis of variance (ANOVA) and Duncan multiple range tests with SPSS trial version software (IBM Co. Ltd., NY, USA).

3. Results and discussions

3.1. Analysis of intrinsic fluorescence spectra

The intrinsic fluorescence spectra of gluten proteins are primarily attributed to tryptophan residues (Sadat, Corradini, & Joye, 2019). As shown in Fig. 1 and Table 1, compared with CK, adding DKGM induced a red shift in the maximum emission wavelength (λ_max). This shift indicated that DKGM promoted a more hydrophilic microenvironment around tryptophan residues, thereby facilitating the transfer of tryptophan from a hydrophobic to a polar environment (Chen et al., 2023). However, the magnitude of the shift decreased with FT progression (Table 1), which was likely due to ice recrystallization disrupting the gluten microenvironment. With the addition of DKGM, the maximum intrinsic fluorescence intensity (I_max) of gluten decreased initially and then increased with increasing DD, with the lowest value in the DK2 group. At FT-0, all DKGM groups exhibited lower I_max than the CK group. This reduction was attributed to polysaccharide–protein interactions through covalent and noncovalent bonds, which promoted aggregate formation, partially masked tryptophan residues, and stabilized the gluten network (Cui et al., 2023). After FT-3, DK2 and DK3 displayed higher I_max values, which suggested that moderate and high DD levels better preserved the tryptophan microenvironment, reducing fluorescence quenching caused by FT.

Fig. 1.

The effect of DKGM with different DDs on intrinsic fluorescence spectra of frozen gluten protein samples. A, B, and C represent the samples that underwent FT-0, FT-1, FT-2, and FT-3, respectively.

Table 1.

The effect of DKGM with different DDs on intrinsic fluorescence spectra of maximum emission wavelength (λ_max) and maximum intrinsic fluorescence intensity (I_max). Lowercase letters represent a significant difference in λ_max and I_max of gluten protein samples (p < 0.05).

λmax	FT-0	FT-1	FT-2	FT-3
CK	343.67 ± 0.31b	343.53 ± 0.46b	344.00 ± 0.72a	344.67 ± 0.83a
KGM	344.47 ± 1.89ab	343.6 ± 0.35b	344.47 ± 0.64a	344.53 ± 0.76a
DK1	344.07 ± 0.50ab	343.93 ± 0.83ab	344.27 ± 0.42a	344.47 ± 0.50a
DK2	344.80 ± 1.40ab	344.67 ± 0.46a	344.40 ± 0.20a	344.87 ± 0.46a
DK3	345.27 ± 0.81a	344.00 ± 0.60ab	344.67 ± 0.23a	344.80 ± 0.35a
Imax
CK	567.37 ± 21.43a	432.47 ± 40.27b	399.47 ± 11.66b	347.33 ± 18.31c
KGM	547.10 ± 31.80ab	477.20 ± 12.50ab	411.90 ± 5.07a	364.33 ± 4.11b
DK1	514.93 ± 22.56b	450.87 ± 23.24ab	405.93 ± 23.31ab	364.43 ± 8.76b
DK2	499.00 ± 35.07b	500.03 ± 42.29a	415.63 ± 12.41a	376.77 ± 0.55a
DK3	547.67 ± 8.21ab	453.57 ± 20.4ab	413.40 ± 12.32a	383.67 ± 15.75a

3.2. Analysis of surface hydrophobicity

Surface hydrophobicity is an important indicator of protein tertiary structure (Rahaman, Vasiljevic, & Ramchandran, 2016). The effect of FT cycles on the surface hydrophobicity of gluten protein with DKGM is shown in Fig. 2. The addition of DK1, DK2, and DK3 reduced the surface hydrophobicity of gluten proteins by 8.85%, 14.03%, and 13.69%, respectively, compared with that in the KGM group, consistent with intrinsic fluorescence results (Fig. 1). This reduction suggested that DKGM altered the spatial conformation of gluten proteins, increasing internal hydrophobicity and structural stability (Hu et al., 2023). As FT progressed, the surface hydrophobicity decreased in all groups and stabilized after FT-2. At FT-3, the DKGM groups exhibited higher surface hydrophobicity than the KGM group, with an increase of 0.76%, 4.83%, and 4.59% for DK1, DK2, and DK3, respectively. This indicated that DKGM prevented excessive aggregation of the gluten protein conformation during FT cycles, corroborating findings of Yang, Xu, Zhou, Wu, and Xu (2022), who reported that excessive aggregation of gluten proteins negatively impacted the formation of stable secondary structures. Section 3.7 further illustrates the changes in the secondary structure of gluten proteins.

Fig. 2.

The effect of DKGM with different DDs on surface hydrophobicity of frozen gluten protein samples. Different lowercase letters indicate significantly different (p < 0.05) in different groups.

3.3. Analysis of gluten chains

The mechanical properties of the gluten protein network are significantly influenced by the conformation of molecular chains (Liang et al., 2022). The two-dimensional and three-dimensional atomic force microscopy images of gluten protein molecular chains are presented in Fig. 3. Wheat gluten exhibited a network structure composed of island-like and strip-like aggregates (bright regions). At FT-0, the proportion of bright regions increased following the addition of DKGM, indicating that DKGM promoted the aggregation of gluten proteins. This observation aligned with the findings of Hu et al. (2023), who suggested that such effects were attributed to polysaccharides modulating disulfide bonds and promoting aggregate formation through hydrophobic and hydrogen bond interactions. Subsequent analysis of SH_free groups (Table 2) and intermolecular forces (Fig. 5) further supported this hypothesis. After FT-3, the CK group exhibited markedly broader bright regions and increased polymer height, indicating the aggregation of gluten chains during FT. Liang et al. (2022) speculated that FT caused mechanical damage to gluten proteins, thereby increasing their flexibility and enhancing their tendency to aggregate through the formation of intermolecular interactions. In contrast, the medium- and high-DD DKGM groups (DK2 and DK3) displayed a more uniform chain distribution in FT-3, indicating the formation of a more stable structure that alleviated gluten aggregation.

Fig. 3.

AFM images of gluten proteins with or without DKGM added during FT-0 and FT-3. A₁-E₁ and F₁-J₁ were the corresponding cross-sections of gluten proteins with or without varying DDs of DKGM during FT-0 and FT-3, respectively. A₂-E₂ and F₂-J₂ were the corresponding three-dimensional topography of gluten proteins with or without varying DDs of DKGM during FT-0 and FT-3, respectively.

Table 2.

Effects of different DDs of DKGM on the SH_free contents of gliadin, glutenin and gluten during freeze-thaw cycles. Different lowercase letters indicate significantly different (p < 0.05) in different groups, and different uppercase letters indicate significantly different (p < 0.05) in freeze-thaw cycles.

Samples		FT-0	FT-1	FT-2	FT-3
Gluten	CK	5.25 ± 0.08aC	5.51 ± 0.13aBC	5.66 ± 0.19aAB	5.88 ± 0.25aA
	KGM	5.32 ± 0.17aB	5.37 ± 0.19bAB	5.54 ± 0.04abAB	5.64 ± 0.08abA
	DK1	5.37 ± 0.19aB	5.42 ± 0.04abAB	5.47 ± 0.08abAB	5.54 ± 0.11bA
	DK2	5.33 ± 0.04aB	5.34 ± 0.08bB	5.39 ± 0.04bB	5.44 ± 0.07cA
	DK3	5.34 ± 0.04aB	5.39 ± 0.11bAB	5.44 ± 0.13abA	5.47 ± 0.08cA
Glutenin	CK	5.42 ± 0.04aD	5.81 ± 0.19aC	6.13 ± 0.11aB	6.52 ± 0.04aA
	KGM	5.37 ± 0.07aC	5.61 ± 0.04aB	5.81 ± 0.07bB	6.10 ± 0.19bA
	DK1	5.39 ± 0.04aD	5.66 ± 0.07aC	5.83 ± 0.15bB	6.00 ± 0.04bA
	DK2	5.42 ± 0.11aB	5.61 ± 0.04aA	5.71 ± 0.08cA	5.76 ± 0.04cA
	DK3	5.39 ± 0.15aC	5.64 ± 0.11aBC	5.86 ± 0.08bAB	6.05 ± 0.19bA
Gliadin	CK	2.33 ± 0.26aC	2.42 ± 0.19abB	2.66 ± 0.15aA	2.75 ± 0.13aA
	KGM	2.40 ± 0.17aC	2.48 ± 0.11bC	2.57 ± 0.11bB	2.70 ± 0.11abA
	DK1	2.50 ± 0.07aC	2.62 ± 0.13aB	2.64 ± 0.07aB	2.71 ± 0.07abA
	DK2	2.33 ± 0.04aC	2.45 ± 0.04bB	2.53 ± 0.04cA	2.58 ± 0.19cA
	DK3	2.48 ± 0.04aB	2.52 ± 0.15aB	2.59 ± 0.15abA	2.61 ± 0.19cA

Fig. 5.

Effect of DKGM on the intermolecular forces of the different samples. Gluten, glutenin, and gliadin are represented by A, B, and C. Ionic bonds, hydrogen bonds, and hydrophobic interactions are represented by 1, 2, and 3. Different lowercase letters indicate significantly different (p < 0.05) in different groups, and different uppercase letters indicate significantly different (p < 0.05) in freeze-thaw cycles.

3.4. Microstructure of gluten

Fig. 4A shows the microstructure of gluten with DKGM of varying DDs before and after FT. The red region represents gluten protein, and the blue region represents polysaccharides. The results showed that the polysaccharides were dispersed within the pores of the gluten protein network, exhibiting a “filling effect” (Li et al., 2023). Large polysaccharide aggregates were observed in the DK3 groups. Huang, Takahashi, Kobayashi, Kawase, and Nishinari (2002) found that high-DD DKGM tends to self-aggregate through intramolecular hydrogen bonding among hydroxyl groups, thereby influencing the formation of a stable gluten network.

Fig. 4.

CLSM images of gluten blended with DKGM of different DDs (A) and gluten (B). In Fig. 4B, the values on the bottom left corner represent the average pore size, and the bottom right corner represent the porosity.

Fig. 4B illustrates the gluten network structure after FT treatment, where black regions represent pores. DKGM addition increased the average pore size of the gluten network. This was likely due to polysaccharide–gluten interactions that promoted protein aggregation, as also supported by intrinsic fluorescence and surface hydrophobicity results (Fig. 1, Fig. 2). After FT cycles, the number and size of pores in the gluten network increased due to mechanical damage from ice crystal recrystallization (Dai, Gao, Zeng, & Liu, 2022).

3.5. Analysis of SH_free group

Disulfide bonds and SH_free groups within high-molecular-weight glutenin subunits are crucial for the formation of the gluten protein network. Meanwhile, the cysteine residues in α-, and γ-gliadins interact with glutenin through disulfide bonds (Zhou et al., 2020). The SH_free content, which inversely reflects disulfide bond levels, is shown in Table 2. During FT, the SH_free content increased due to the protein unfolding and refolding, leading to disulfide bond cleavage. This increase was attenuated by DKGM supplementation. Compared with the KGM group, DKGM supplementation attenuated the increase in SH_free content in gluten, glutenin, and gliadin during FT. The glutenin exhibited higher SH_free content than gliadin, consistent with previous findings (Du, Dang, Jia, Xu, & Li, 2022), mainly because of their distinct structural characteristics. However, FT had a more significant effect on glutenin than gliadin, as the increased rate of SH_free content was higher in glutenin than in gliadin (Table 3). Wang, Jin, and Xu (2015) indicated that gliadin confined water more effectively, thereby reducing mechanical damage caused by ice crystals. Differences in disulfide bond types also contributed to this phenomenon: glutenin was stabilized by both intra- and inter-chain disulfide bonds, whereas gliadin contained only intra-chain disulfide bonds due to the spherical structure. The inter-chain disulfide bonds were more easily cleaved (Ma, Han, Li, Zheng, & Wang, 2019).

Table 3.

The sulfhydryl contents increase rate after three freeze-thaw cycles.

Samples	Gluten	Glutenin	Gliadin
CK	12.15%	20.36%	18.03%
KGM	5.99%	13.70%	12.50%
DK1	3.20%	11.36%	8.40%
DK2	2.30%	6.33%	10.73%
DK3	2.76%	12.27%	5.24%

3.6. Analysis of non-covalent bonds

Low-molecular-weight glutenin subunits form noncovalent bonds among themselves; ω-gliadins also interact through noncovalent bonds, further promoting the formation of the protein network (Minnan et al., 2023). As shown in Fig. 5, the ionic bonds and hydrophobic interactions predominated among the noncovalent forces in proteins. The proportions of covalent and noncovalent interactions differ among gluten protein components. Glutenin contains higher levels of hydrogen and ionic bonds, whereas gliadin exhibits a greater proportion of hydrophobic interactions. This distinction arises because polar amino acid residues in glutenin are more exposed on the surface, thereby facilitating the formation of hydrogen and ionic bonds with surrounding molecules. In contrast, most polar groups in gliadin are buried within the molecular interior, resulting in a predominance of hydrophobic regions (Shewry & Belton, 2024).

As FT cycles prolonged, gluten, glutenin, and gliadin exhibited a similar trend in noncovalent bonds; the proportions of ionic and hydrogen bonds decreased, whereas the proportion of hydrophobic interactions increased, consistent with previous findings (Liang et al., 2024). The increased trend in hydrophobic interactions during the FT process (Fig. 5 A₃, B₃, and C₃) was attributed to the exposure of hydrophobic groups, which aggregated via hydrophobic interactions (Shang, Guo, Liu, Wu, & Zhou, 2022). Such aggregation indicated protein denaturation and gradual loss of function. As a neutral polysaccharide, DKGM may form physical barriers between protein molecules, thereby reducing polar interactions and consequently lowering ionic bond levels (Fig. 5 A₁, B₁, and C₁). Furthermore, DKGM hydroxyl groups competed for polar sites, forming hydrogen bonds and markedly increasing their proportion (Fig. 5 A₂, B₂, and C₂) (Zhang, Ma, Yang, Li, & Sun, 2022). After FT-3, the DKGM groups exhibited a higher proportion of ionic and hydrogen bonds but lower hydrophobic interactions than the KGM. This suggested reduced steric hindrance in DKGM, facilitating the formation of more stable polysaccharide–protein complexes and preserving protein structure during FT. However, DKGM molecules with high DD were prone to aggregation, thereby increasing the distance between protein molecules and reducing the formation of chemical interactions among proteins (Yan et al., 2021). These observations aligned with the results of the SH_free group (Table 2).

3.7. Analysis of Fourier transform infrared spectroscopy (FTIR)

The FTIR spectra of the samples (Fig. 6 A_1,2, B_1,2, and C_1,2) indicated that DKGM addition with different DDs slightly altered the characteristic protein absorption peaks without generating new bands. DKGM resulted in a slight blue shift in the peaks compared with the CK group, and this trend persisted under FT cycles. The shift of the absorption peak at 3280 cm⁻¹ was associated with the formation of hydrogen bonds between the hydroxyl groups in DKGM and the amino and carboxyl groups in the protein (Saengsuk et al., 2022). The peak change around 2930 cm⁻¹ likely reflected alterations in the protein microenvironment impacting the C—H stretching vibrations (Nawrocka, Miś, & Niewiadomski, 2017). However, the shift near 1635 cm⁻¹ was associated with changes in the secondary structure of the protein, resulting in a peak shift (Nawrocka et al., 2017). The shift at around 1520 cm⁻¹ peak might be due to hydrogen bonding between DKGM and protein C—N or N—H groups (Kong & Yu, 2007).

Fig. 6.

FT-IR spectra of different samples. A₁, B₁, and C₁ are represented by gluten, glutenin, and gliadin without FT cycles. A₂, B₂, and C₂ are represented by gluten, glutenin, and gliadin undergoing FT-3 cycles. A₃, B₃, and C₃ represent the secondary structure content of frozen gluten, glutenin, and gliadin with varying DDs of DKGM. Different letters indicate significantly different (p < 0.05) in different groups.

The secondary structures of frozen gluten, glutenin, and gliadin are shown in Fig. 6 A₃, B₃, and C₃. The results revealed that α-helix and random coil were the predominant secondary structures, with relatively low β-sheet content. The α-helix content was higher in gliadin than in glutenin. After FT-3, the α-helix content decreased in all groups, the β-sheet and random coil content increased, and the β-turn content remained essentially unchanged. This suggested that FT promoted the transformation of α-helix into β-sheet and random coil, thereby increasing protein network disorder. DK2 and DK3 exhibited the most effective alleviation of α-helix loss in glutenin (Fig. 5 B3) and gliadin (Fig. 5 C3), respectively. This difference might be related to protein structure: glutenin, a rigid linear protein, readily formed robust networks, but DKGM with high DD tended to aggregate and thus dispersed poorly within the protein network. In contrast, gliadin, a loosely structured globular protein, allowed better interaction with DK3 (Wang et al., 2015).

3.8. Molecular weight distribution

The major protein subunits within different molecular weights are labeled in Fig. 7 according to Li et al. (2024). In all samples, the number and position of protein bands remained essentially unchanged, indicating that DKGM addition did not alter the composition of protein subunits or generate new ones. After FT-3, the band gray intensity of high-molecular-weight glutenin subunits (HMW-GS) decreased, whereas the intensity of bands located at the low-molecular-weight glutenin subunits (LMW-GS) region increased, indicating depolymerization of gluten. Previous studies have attributed this depolymerization mainly to the disruption of disulfide bonds between glutenin subunits and noncovalent interactions between glutenin and gliadin subunits (Guo et al., 2021; Li & Ma, 2025). The rate of change in gray intensity was used as an indicator of protein structural alterations during FT. With DKGM addition, the rate of change in band intensity was reduced, suggesting that DKGM effectively protected internal disulfide and noncovalent bonds within gluten proteins (3.5, 3.6). A pronounced increase in the band gray intensity of ω-gliadins was also observed after FT-3. The KGM group exhibited the lowest rate of increase, likely due to its strong water-binding capacity, which synergistically enhanced gliadin hydration and mitigated its aggregation. In contrast, the increased rate declined with increasing DD in the DKGM groups, probably because reduced steric hindrance enabled DKGM to disperse more effectively within the loosely structured spherical gliadins and facilitate interactions.

Fig. 7.

Effects of different DDs of DKGM on the molecular weight distributions of gluten (A), glutenin (B), and gliadin (C), and effect of varying DDs of DKGM on the HMS-GS (D), LMS-GS (E) and ꞷ-gliadin (F) gray intensity for FT-0 and FT-3. Increased rate means the percentage increase in gray intensity of FT-3 compared with FT-0 in the specified area.

3.9. Molecular docking

The binding energy obtained from molecular docking reflects the interaction strength between ligands and receptors, with values below −4.25 kcal/mol indicating some binding activity and those below −5.0 kcal/mol suggesting strong binding (Tang, Ye, et al., 2025). As shown in Fig. 8, DKGM with different DD exhibited favorable binding to components of gluten proteins. The amino acids involved in these interactions were similar across DDs, mainly Arg, Glu, Gln, and Ser in glutenins. These amino acids are hydrophilic and readily form hydrogen bonds with surrounding water molecules. DKGM may interact with these residues via hydrogen bonding, thereby stabilizing the hydrogen bond network and conformation of gluten proteins during FT. DKGM with moderate DD showed the strongest binding affinity to glutenins, with a binding energy of −6.3 kcal/mol. In contrast, high-DD DKGM tended to self-aggregate, and the dispersed distribution of residues weakened its binding affinity for glutenins. Meanwhile, DKGM was bound to gliadins through Glu and Gln, consistent with the findings of Tang, Ye, et al. (2025). The dense distribution of Glu and Gln residues in gliadins enabled high-DD DKGM to form tighter interactions (Ewart, 1964). However, the docking results can only provide limited auxiliary information because the rigid crystal structures retrieved from the PDB may not fully capture the dynamic conformational ensembles within the complex gluten matrix.

Fig. 8.

Molecular docking result of DKGM with varying degrees of deacetylation (DD) and different gluten components.

3.10. Schematic model

The schematic model illustrating the effects of DKGM on frozen gluten proteins and their components is shown in Fig. 9. The experimental results revealed that gluten proteins underwent a “depolymerization–aggregation” process during FT cycles. First, the disruption of disulfide bonds led to the depolymerization of HMW-GS. Then, the hydrophobic groups of glutenin were exposed, and the content of low-molecular-weight gluten proteins increased. These proteins subsequently aggregated via hydrophobic interactions, manifesting as an aggregated state on the glutenin chains. These results highlighted the structural change complexity of gluten during FT. DKGM appeared to form stable complexes primarily by establishing hydrogen bonds with hydrophilic amino acids (Fig. 8). This alleviated the breakage of ionic, hydrogen, and disulfide bonds and restrained excessive hydrophobic interactions, thereby delaying structural deterioration during FT. The impact of DKGM with different DDs on gluten protein and its components varied. The formation of inter-gluten disulfide bonds was restricted due to the high viscosity of KGM and DK1 (Table 2) (Yan et al., 2021). DK3 inherent tendency to aggregate (Fig. 3) reduced its cryoprotective efficacy in the more rigid glutenin. In contrast, DK2, characterized by moderate water-binding capacity and particle size, this may made DKGM was uniformly dispersed within the gluten network and effectively enhanced the cryostability of glutenin (Yan et al., 2021). Glutenin is more susceptible to freeze-induced structural disruption. DK2 demonstrated the optimal overall cryoprotective performance for gluten proteins.

Fig. 9.

The schematic model of the effect of DKGM on the different gluten components after FT cycles.

4. Conclusion

This study explored the effects of DKGM with different DDs on the structure of frozen gluten and its components. Compared with KGM, DKGM more effectively stabilized the secondary and tertiary structures of gluten proteins and mitigated structural deterioration during FT cycles. This enhanced stability might be primarily attributed to DKGM promoting the formation of disulfide bonds and α-helices, besides strengthening noncovalent interactions within the gluten network. The analysis of glutenin and gliadin revealed that glutenin was more susceptible to FT-induced structural disruption and that DKGM primarily improved the cryostability of gluten by enhancing the structural integrity of glutenin. DKGM with a moderate degree of deacetylation (50.21%) exhibited the most pronounced cryoprotective effect, likely because it could closely associate with glutenin to form more stable polysaccharide–protein complexes. Collectively, these findings provide novel mechanistic insights into the cryoprotective role of DKGM in frozen dough systems and highlight its potential for broader application in frozen food preservation. Future studies should focus on elucidating the effects of DKGM on water mobility in frozen gluten and employing molecular simulation approaches to further clarify the underlying interaction mechanisms.

CRediT authorship contribution statement

Jianwei Fan: Writing – review & editing, Writing – original draft, Methodology, Formal analysis, Data curation, Conceptualization. Haoyuan Wang: Investigation. Zhilong Zeng: Writing – review & editing. Yijia Li: Writing – review & editing. Xiaoli Qin: Writing – review & editing, Supervision, Conceptualization. Yao Li: Writing – review & editing. Xiong Liu: Writing – review & editing, Supervision, Conceptualization.

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.

Data availability

Data will be made available on request.