Exploring the Impact of Non-Enzymatic Reactions on the Structural Characteristics of Wolfberry Polysaccharides During Lactiplantibacillus plantarum NCU137 Fermentation

Abstract

Plant-based lactic acid bacteria fermentation is known to alter polysaccharide structures by enzymatic reactions, whereas the role of non-enzymatic reactions remains unclear. To investigate how non-enzymatic reactions impact the structural modification of wolfberry polysaccharides during fermentation by Lactiplantibacillus plantarum NCU137, a series of structural characterization techniques with metabolomics analysis were applied. The results showed fermentation induced marked degradation of wolfberry polysaccharides, as evidenced by decreased proportion of →5)-Araf-(1→(Af₅) residues. Untargeted metabolomics identified 242 differential metabolites, suggesting that fermentation-induced polysaccharide remodeling was associated with changes in the acidic and redox microenvironments. Accordingly, organic acids, vitamins, fatty acids, and amino acids were selected for targeted quantification based on variations in metabolite type and abundance. Notably, lactic acid accumulated sharply to 4.8 mg/mL, whereas vitamin C decreased significantly from 2847.3 to 2672.0 μg/g, indicating that lactic acid accumulation and vitamin C-associated redox changes may act as one of key non-enzymatic contributors to polysaccharide structural remodeling. Finally, in vitro simulation further supported the synergistic effects of lactic acid and vitamin C on the structure modification. These findings offer a theoretical perspective on the functional enhancement of polysaccharide during fruit and vegetable fermentation.

Graphical abstract

Keywords

Lactic acid bacteria fermentation / Non-enzymatic reactions / Wolfberry juice / polysaccharides / Structural characterization / Metabolomics analysis

Highlight

	● Fermentation induced degradation and structural remodeling of wolfberry polysaccharides.
	● Fermentation-induced polysaccharide remodeling was accompanied by substantial metabolic reprogramming.
	● Fermentation caused substantial accumulation of lactic acid and depletion of vitamin C.
	● In vitro simulation verified lactic acid- and vitamin C-mediated non-enzymatic polysaccharide modification.

1 Introduction

Lactic acid bacteria (LAB) fermentation is widely used in the deep processing of fruits and vegetables, due to its mild condition, environmentally friendly characteristics, flavor enhancement and shelf-life extension [1]. Lactiplantibacillus plantarum, a common LAB species found in food fermentation, has been shown to improve the functional attributes of food products, particularly antioxidant activities [2]. Wolfberry (Lycium barbarum), a typical medicine–food homologous plant widely used in China, is well-known for diverse bioactivities, including antioxidant [3], anti-inflammatory [4], antitumor [5], anti-aging [6], antidiabetic [7], and immunomodulatory effects [8]. It has been shown that fermentation could enhance the bioactivities of wolfberry by reshaping chemical constituents profiles [9]. Polysaccharides as one of the key bioactive substances responsible for the biological activities of wolfberry, such as antioxidant, antitumor, and antihyperglycemic properties, [10,11] have garnered increasing attention in recent years from researchers, especially for the fermentation-induced structure modification and the consequent enhanced health benefits [12].

Accumulating evidence has demonstrated that polysaccharide structures could be modified through various enzymatic reactions during fermentation. For instance, the decrease in the molecular weight (MW) of okra polysaccharides has been associated with the increased synthesis of carbohydrate-active enzymes (CAZymes) [13]. Similarly, the increase in sugar content in fermented lychee pulp was linked to the presence of three endoglucanases, while the decreased relative abundances of galactose and galacturonic acid in polysaccharides were partially linked to the activities of multiple galactosidases [14]. Notably, the fermentation process is also accompanied by the occurrence of non-enzymatic reactions, which result from alterations in ingredient profile induced by fermentation. For instance, Peng et al. [15] found that tartaric acid derived from LAB can hydrolyze the RG-I side chains of pectin. Moreover, free radicals generated by oxidation may cause the fragmentation of polysaccharides [16]. For example, Kamlow et al. [17] published that vitamins such as vitamin C can decrease the molar mass of galactomannan. However, the mechanism underlying the impact of non-enzymatic reactions on wolfberry polysaccharides during fermentation remains unclear.

Previous studies have demonstrated that L. plantarum NCU137 exhibited excellent potential for fermenting wolfberries [18]. After fermentation, the viable cell count in wolfberry juice reached 8.8 log CFU/mL, with the pH dropping from 5.0 to 3.8. As reported by Liu et al. [2], this process involved the secretion of various CAZymes, such as glycoside hydrolases and carbohydrate esterases, which consequently modify the structures of polysaccharides. Additionally, the changes in chemical constituents during fermentation theoretically may also have an impact on polysaccharides structure, however, the underlying mechanisms remain unclear.

This study primarily aims to investigate the effects of chemical factors rather than enzymes that change the structure of wolfberry polysaccharides during fermentation with L. plantarum NCU137. Therefore, from a metabolomic perspective, the research focused on elucidating the impact of non-enzymatic reactions, such as the generation of organic acids and the oxidative degradation of vitamins and fatty acids, on the structural integrity of polysaccharides, which could not only contribute to elucidating the intrinsic relationships among microbial metabolism, non-enzymatic reactions, and polysaccharide molecular reconstruction, but also can furnish a theoretical foundation for the precise development of fermented functional foods.

2 Materials and Methods

2.1 Reagents and bacterial strain

Wolfberries from Ningxia were purchased from Tmall Supermarket. L. plantarum NCU137 powder was provided by the State Key Laboratory of Food Science and Resources, Nanchang University. Monosaccharide standards were purchased from Sigma-Aldrich (Shanghai, China), with xylose (Xyl) was obtained from J&K Scientific Ltd. (Beijing, China). Standards of organic acids, vitamins, and fatty acids were supplied by Yuanye Bio-Technology Co., Ltd. (Shanghai, China). Amino acid standards were supplied by Sykam (Beijing, China). Trifluoroacetic acid (TFA), NaBD₄, thermostable α-amylase, KBr, deuterium oxide (D₂O), methanol, and protease were obtained from Aladdin Biochemical Technology Co., Ltd. (Shanghai, China). Sulfuric acid and sodium nitrate (NaNO₃) were sourced from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Phenol and ethanol were acquired from Xilong Scientific Co., Ltd. (Guangzhou, China). Reagents not explicitly mentioned were of analytical grade.

2.2 Preparation of fermented wolfberry juice

Samples were prepared according to Liu et al. [2] In brief, dried wolfberries were extracted with boiling water at a solid-to-liquid ratio of 1:3 (w/v) for 30 min, filtered through an 80-mesh sieve. Subsequently, the extracted wolfberry juice was sterilized at 105 °C for 20 min, cooled down, inoculated with 0.02% L. plantarum NCU137, and fermented at 37 °C for 48 h under shaking conditions. The unfermented and fermented wolfberry juice were designated NWJ and FWJ, respectively.

2.3 Structural characteristics of polysaccharides

2.3.1 Extraction of polysaccharides from NWJ and FWJ

Polysaccharides were prepared following the previous protocol [19], with some modifications. Specifically, NWJ and FWJ were mixed with water at a ratio of 1:1 (w/w) and maintained at 95 °C for 2 h to facilitate extraction. The extracts were filtered through 200-mesh gauze and sequentially treated with thermostable α-amylase (95 °C, 1.5 h) and protease (60 °C, 2 h). Following a 15 min boiling treatment, the mixture was subjected to centrifugation, and the resulting supernatant was precipitated overnight with 80% ethanol. The precipitate was separated by centrifugation (2290×g, 10 min), redissolved in water, concentrated under reduced pressure, and dialyzed using membranes with a molecular weight cut-off of 8000–14000 Da for 48 h. The retained fraction was then centrifuged to remove insoluble matters, and was freeze-dried to obtain polysaccharide samples. The polysaccharides isolated from NWJ and FWJ were designated NWP and FWP, respectively.

2.3.2 Physicochemical properties

The contents of neutral sugar, uronic acids, and proteins were determined using phenol-sulphuric acid method [20], sulfate-carbazole method [21] and Coomassie brilliant blue method [22], successively.

2.3.3 Monosaccharide compositions

The monosaccharide determination was performed according to previous report [23], with minor modification. Briefly, samples were dispersed with 12 M H₂SO₄ on ice bath, and then diluted and hydrolyzed at 100 °C for 2 h. The resulting hydrolysates were further diluted and filtered through a 0.22 μm nylon membrane. The released monosaccharides were separated and detected using Dionex™ICS-5000 high-performance anion exchange chromatography (Thermo Fisher, USA). A CarboPac™ PA20 analytical column (3 mm × 150 mm) fitted with a matching guard column (3 mm × 30 mm) was used for the analysis.

2.3.4 Fourier transform-infrared (FT-IR)

Dried NWP and FWP were thoroughly ground with KBr powder at a 1:50 mass ratio and pressed into transparent sheets. FTIR spectra of NWP and FWP were obtained using Nicolet 5700 spectrometer (Thermo Fisher Scientific, USA) within the wavenumber range of 4000–400 cm⁻¹.

2.3.5 Scanning electron microscope (SEM)

Samples were attached to conductive silicone rubber and sputter-coated with gold. The morphology of NWP and FWP was then observed by field emission SEM JSM-6701F (JEOL, Japan) at magnifications of 500× and 3000×.

2.3.6 Methylation analysis

Glycosidic linkages of NWP and FWP were determined by methylation analysis coupled with GC–MS according to previously reported methods [24,25]. Briefly, freeze-dried samples (2–3 mg) were dissolved in DMSO (2 mL) overnight, followed by methylation with anhydrous NaOH (30 mg) and CH₃I (1 mL). The methylated products were extracted with CH₂Cl₂, hydrolyzed with 4 M TFA at 100 °C for 6 h, reduced with NaBD₄ overnight, and acetylated with acetic anhydride at 120 °C for 1 h. The resulting partially methylated alditol acetates (PMAAs) were dissolved in CH₂Cl₂, filtered, and analyzed by GC–MS using an SP-2330 column (30 m × 0.25 mm, 0.2 μm film thickness). The GC oven was initially set at 160 °C, increased to 210 °C at 2 °C/min and then to 240 °C at 5 °C/min, and held for 20 min. The injector temperature was maintained at 250 °C. Mass spectral data were acquired in electron ionization (EI) mode with an ion source temperature of 230 °C, an electron energy of 70 eV, and an m/z range of 35–400. PMAAs were identified by matching their EI–MS spectra against the CCRC Spectral Database for PMAAs.

2.3.7 ¹H NMR analysis

NWP and FWP were dissolved in D₂O and lyophilized three times to ensure complete hydrogen-deuterium exchange. Subsequently, the deuterated samples (0.6 mL) were transferred to an NMR tube. ¹H NMR spectra were recorded at 298 K and 323 K by Bruker AVANCE 400 MHz NMR spectrometer (Bruker, Switzerland) equipped with a cryogenic probe (Bruker, Switzerland).

2.4 Untargeted metabolite analysis

As described previously [26], NWJ and FWJ (100 mg) were extracted with 80% (v/v) aqueous methanol (1 mL), vortexed on ice for 30 s, mechanically disrupted at 60 Hz for 180 s, ultrasonicated for 30 min, and vortexed again for 30 s. Proteins were precipitated overnight, followed by centrifugation at 15,100×g for 20 min. The supernatants were collected and analyzed using UPLC-Triple-TOF/MS 6600 (AB SCIEX, USA) equipped with a Shim-pack GIST C18 column (2.1 mm × 75 mm, 2 µm).

Chromatographic separation was performed using 0.1% (v/v) formic acid in water (A) and methanol (B) as the mobile phases. The gradient was programmed as follows: 5% B (0–1 min), 5%–20% B (1–3 min), 20%–40% B (3–5 min), 40%–60% B (5–10 min), 60%–85% B (10–15 min), 85%–95% B (15–16 min), 95% B (16–17 min), and 95%–5% B (17–19 min), followed by equilibration at 5% B until 20 min. The flow rate, injection volume, and column temperature were 0.25 mL/min, 3 μL, and 35 °C, respectively. MS detection was performed with an electrospray ionization (ESI) source in positive and negative ion modes. The ESI parameters were as follows: interface heater temperature, 650 °C; ion spray voltage, +4.5/−4.5 kV; GS I, 50 psi; GS II, 50 psi; CUR, 35 psi; CE, 30 eV; and DP, 60 V. The m/z range was 40–1000.

2.5 Quantification of key metabolites of NWJ and FWJ

2.5.1 Organic acids

The determination of organic acid levels was carried out on an Agilent 1260 Infinity II HPLC system (Agilent Technologies, USA), following the protocol reported by He et al. [27] with slight optimization. In brief, samples (2.0 mL) were subjected to centrifugation for 10 min at 11,300×g under 4 °C conditions. The obtained supernatant was subsequently clarified using a 0.22 μm membrane filter before injection. The analysis was performed using 6 mM sulfuric acid as the mobile phase at 0.8 mL/min. The column oven was set to 35.0 °C, and analytes were detected by UV–Vis absorbance at 210 nm after injecting 20 μL of the prepared sample.

2.5.2 Vitamins

Following the method of Liu et al. [26], samples (1.0 g) were homogenized with 0.1 mM HCl (1 mL) and ultrapure water (4 mL), ultrasonically extracted for 15 min, and centrifuged at 4472×g for 10 min. The supernatants were collected and analyzed using an Agilent 1260 Infinity II HPLC system (Agilent Technologies, USA).

Chromatographic analysis was carried out using a Diamonsil C18 column (4.6 mm × 250 mm, 5 μm), with the column temperature controlled at 30 °C and the detection wavelength set at 254 nm. The elution system was composed of methanol as phase A and 0.1 mol/L ammonium formate as phase B, operating at a constant flow rate of 1 mL/min. For gradient elution, the proportion of phase B was set at 80% for the first 2 min, changed to 77% over 2–10 min, and then adjusted to 75% from 10 to 20 min.

2.5.3 Fatty acids

Fatty acids were analyzed according to a previously reported method with slight modifications [28]. Briefly, the fatty acid methyl ester standard mixture (GLC-463) and methyl heneicosanoate (C21:0, internal standard; 5 mg/mL) were prepared in n-heptane and n-hexane, respectively. Samples (10 mg) were mixed with internal standard (10 μL), n-heptane (2 mL), and methanolic KOH (0.1 mL), vortexed for 1 min, and centrifuged at 1016×g for 5 min. The supernatants were filtered and analyzed using an Agilent 8890N gas chromatograph (Agilent Technologies, USA) with an injection volume of 1 μL. The injector and detector temperatures were maintained at 250 °C, with high-purity hydrogen as the carrier gas at a linear velocity of 26 cm/s and a split ratio of 10:1. The oven temperature was programmed from 60 °C (5 min) to 170 °C at 11.5 °C/min (25 min), then to 200 °C at 5 °C/min (5 min), and finally to 215 °C at 2 °C/min (20 min).

2.5.4 Free amino acids (FAA)

Determination of FAA in wolfberry juice following the method described by Liu et al. [29] In specific, the supernatant was mixed with 2% of 5-sulfosalicylic acid at a 1:9 (v/v) ratio, followed by centrifugation (1006×g, 10 min), vortexed (15 min), and analysis on a fully automated amino acid analyzer S-433 (Sykam, Germany).

2.6 In vitro simulation validation

To assess the synergistic effects of fermentation-associated organic acids and vitamin oxidation on the changes of polysaccharide structure, an in vitro simulation experiment was conducted using NWP. Briefly, the polysaccharide sample (100 mg) was solubilized in distilled water (10 mL), lactic acid solution (pH 4.0), distilled water containing vitamin C (15 mg), and lactic acid solution (pH 4.0) containing vitamin C (15 mg), which represented the blank control (NWP-K), lactic acid treatment (NWP-Acid), vitamin C treatment (NWP-VC), and combined lactic acid–vitamin C treatment (NWP-Acid+VC), respectively. All samples were incubated at 37 °C for 24 h, after which a molecular weight cut-off of 8000–14000 Da and lyophilization. The resulting polysaccharides were analyzed for physicochemical properties, monosaccharide composition using the methods described above.

The MW of the resulting polysaccharides were determined by HPLC. Samples were dissolved with mobile phase (0.1M NaNO₃) into 1 mg/mL and filtered through 0.22 μm nylon membrane. The MW were measured by Waters e2695 high performance liquid chromatography (Waters, USA) equipped with a Waters Ultrahydrogel™ guard column (6 mm × 40 mm) and a Waters Ultrahydrogel™ gel column (7.8 mm × 300 mm).

2.7 Statistical analysis

Minitab (v.20) was used to run one-way ANOVA with Tukey’s multiple comparisons test. The data were shown as mean ± SD (n = 3), and P < 0.05 was considered significant. MetaboAnalyst 6.0 was used to analyze untargeted metabolomics.

3 Results and Discussion

3.1 Structural characterization of NWP and FWP

3.1.1 Compositional changes

The physicochemical properties and monosaccharide compositions of NWP and FWP were shown in Table 1. NWP contained (40.4 ± 0.4)% of neutral sugars and (10.5 ± 0.6)% of uronic acids, which was in line with published reports [30]. After fermentation, the neutral sugar content increased to (44.8 ± 1.5)%, whereas the uronic acid content decreased to (9.4 ± 0.3)% in FWP. This could be associated with the metabolic activities of LAB, including the secretion of CAZymes and the production of metabolites [31].

Monosaccharide compositions results further showed that NWP and FWP shared similar monosaccharide types, including rhamnose (Rha), arabinose (Ara), galactose (Gal), glucose (Glc), xylose (Xyl), mannose (Man), galacturonic acid (GalA) and glucuronic acid (GlcA), while their proportions differed between each other. Although the monosaccharide types were similar to that reported previously [32], the Rha content increased from (2.5 ± 0.1)% to (3.2 ± 0.2)%, similar results had been reported in L. plantarum fermented Lvjian okra polysaccharides [13]. The decrease in Ara and Glc contents may be related to the modification of polysaccharide side chains during fermentation. As Ara- and Glc-rich residues are commonly distributed in branched regions of polysaccharides, the L. plantarum NCU137 fermentation may have induced structural rearrangements and partial degradation of these side chains, resulting in changes in their relative abundance.

3.1.2 FT-IR spectra and SEM analysis

FT-IR and SEM analysis were further performed to evaluate the chemical structure and microstructure of NWP and FWP (Figs. 1A and 1B). The broad band at 3000–3500 cm⁻¹ was assigned to O–H stretching vibrations [33], while the absorption peak at 2931 cm⁻¹ corresponded to C–H stretching vibrations of alkyl groups [34]. The bands at 1611, 1604, and 1402 cm⁻¹ were attributed to asymmetric and symmetric stretching vibrations of carboxylate groups, indicating the presence of uronic acids [35]. Absorption bands at 1068 and 1055 cm⁻¹ were associated with C–OH stretching vibrations of pyranose rings, suggesting the presence of pyranosyl residues in the polysaccharide chains [35]. The peaks at 652 and 645 cm⁻¹ were assigned to O–H out-of-plane bending vibrations [33]. Spectral analysis indicated that NWP and FWP were neutral polysaccharides, which was in line with the results of monosaccharide composition. Moreover, the stronger absorption at 1068 cm⁻¹ in FWP may indicate an increased relative contribution of pyranose ring-related vibrations after fermentation.

SEM images revealed obvious morphological differences between NWP and FWP, NWP exhibited an irregular, rough morphology with obvious lamellar streaks, whereas FWP showed a comparatively smoother surface. This finding aligns with previous reports on polysaccharides from fermented Polygonatum kingianum [36]. Together with the FT-IR results, these observations indicated that fermentation may induce changes in the surface morphology and microstructural organization of polysaccharides.

3.1.3 Linkage patterns and ¹H NMR analysis

The glycosidic linkage configurations of NWP and FWP were further investigated through methylation analysis, and the relative percentage of peak area was performed in Table 2. Considering the relatively low uronic acid content and the limited changes observed in monosaccharide composition and physicochemical characteristics after fermentation, uronic acid residues were not further considered in the methylation analysis. Based on fragmentation patterns, it was found that NWP and FWP shared the same 19 glycosidic linkage types, although the relative proportions differed. In NWP, the predominant residues were Araf-(1→ (Af_t), →2)-Araf-(1→ (Af₂), Glcp-(1→ (Gl_t), →5)-Araf-(1→ (Af₅), and →3,6)-Galp-(1→ (G_3,6), accounting for (15.7 ± 2.2)%, (12.0 ± 0.5)%, (8.7 ± 0.3)%, (11.2 ± 0.2)%, and (15.2 ± 0.2)%, respectively. This result was consistent with previous finding [37] and indicated that NWP possessed a highly branched arabinogalactan-like structure. After fermentation, the major glycosidic linkage types remained unchanged, whereas the relative abundance of Af₅ decreased to (8.1 ± 0.5)%, indicating partial degradation of arabinose-containing side chains during fermentation. Overall, the methylation results were consistent with the monosaccharide composition analysis, and the slight increase in pyranose-type residues after fermentation was in accordance with the enhanced FT-IR absorption at 1068 cm⁻¹.

Fig. 1C presented the ¹H NMR spectra of NWP and FWP, showing both polysaccharides mainly had Araf and Galp residues in α- and β-configurations, respectively. Specifically, the signals at 5.24 ppm, 5.09 ppm, 4.73 ppm, 4.53 ppm, and 5.44 ppm were assigned to Af_t, Af₅, →3)-Galp-(1→ (G₃), G_3,6, and Gl_t residues, respectively. These assignments were consistent with the methylation results, which showed abundant Araf and Galp linkages in both NWP and FWP. Particularly, G_3,6 might serve as an important branching residue, whereas Af_t, and Af₅ were likely involved in arabinan-like side chains. Based on earlier reports [38–40], NWP could be characterized as a highly branched arabinogalactan-like polysaccharide with a β-Galp-rich backbone and Araf/Galp-containing branches. After fermentation, the reduced proportion of Af₅ might be due to the partial degradation of arabinan-like side chains. Detailed information on the glycosidic linkages was provided in Supplementary Material Table S1.

Taken together, fermentation induced partial depolymerization and side-chain remodeling of wolfberry polysaccharides, especially the decreases in Af₅ residues. These structural changes were consistent with the recognized effects of CAZymes secreted by LAB. Nevertheless, enzymatic hydrolysis alone may not fully account for the observed structural modifications, as fermentation also alters the nutrient and metabolite composition of wolfberry juice, thereby generating acidic or redox-active microenvironments that may induce non-enzymatic polysaccharide modification. To further investigate this possibility, untargeted metabolomics was performed on NWJ and FWJ to identify differential metabolites and potential non-enzymatic factors associated with polysaccharide structural remodeling.

3.2 Untargeted metabolomics analysis

3.2.1 OPLS-DA analysis and differential metabolites identification

In total, 4724 and 6811 metabolites were detected in negative and positive ion modes, respectively. A supervised multivariate statistical approach, OPLS-DA, was employed to distinguish metabolite profiles between samples and identify variables contributing to group separation [41]. The OPLS-DA score plots showed clear separation between unfermented and fermented wolfberry juice within the 95% confidence regions (Supplementary Material Figs. S1A and S1B), indicating marked metabolic differences during fermentation and offering a theoretical basis for further identification of differential metabolites.

Based on VIP values ≥ 1, fold change (FC) ≥ 2 or ≤ 0.5, and P < 0.05, 242 metabolites with significant differences were identified, including 139 and 103 metabolites detected in positive and negative ion modes, respectively. The differential metabolites were classified into 18 categories (Supplementary Material Fig. S1C), i.e., lipids and lipid-like molecules (20.25%), organoheterocyclic compounds (16.12%), benzenoids (11.16%), amino acids, peptides, and analogues (10.74%), carbohydrates and carbohydrate conjugates (8.26%), organic acids and derivatives (7.44%), steroids and steroid derivatives (4.96%), flavonoids (3.31%), nucleosides, nucleotides, and analogues (2.89%), phenolic acids (2.89%), phenols (2.48%), vitamins and derivatives (2.07%), and others. Among them, lipids and lipid-like molecules, organoheterocyclic compounds, benzenoids, amino acids, peptides, and analogues, and organic acids and derivatives represented the predominant metabolite classes, consistent with previous findings on L. plantarum-fermented loquat [42].

To further identify metabolites potentially associated with non-enzymatic polysaccharide modification, volcano plot and heatmap analyses were performed based on the differential metabolites. Under stricter thresholds of FC > 2.5 and P < 0.001, 80 markedly altered metabolites were screened from the 242 differential metabolites, including 53 upregulated and 27 downregulated metabolites (Supplementary Material Fig. S1D). Among the upregulated metabolites, lactic acid, tropic acid, (S)-(-)-2-hydroxyisocaproic acid, LysoPC (16:0), allomaltol, and guanosine showed the most pronounced increases, suggesting enhanced organic acid production and lipid-related metabolic remodeling during fermentation. In contrast, DL-3,4-dihydroxyphenyl glycol, L-phenylalanine, (2R,3S)-piscidic acid, physalin B, pyrocatechol, and m-coumaric acid were markedly decreased, indicating changes in amino acid and redox-related metabolism. Similar metabolic shifts had been reported in fermented Lycium barbarum and Platycodon grandiflorus [43], supporting that fermentation can reshape acidic and redox-related metabolic profiles and thereby contribute to non-enzymatic polysaccharide structural modification.

Heatmap analysis further revealed clear metabolic differences between NWJ and FWJ, particularly in organic acids, vitamins and derivatives, fatty acids, and amino acid-related metabolites (Supplementary Material Fig. S1E). Several organic acids, including lactic acid, veratric acid, tropic acid, (2S,3S)-3-hydroxy-3-methylbutanoic acid, and (S)-(-)-2-hydroxyisocaproic acid, increased after fermentation, whereas alterations in LysoPCs and vitamin-related metabolites suggested potential lipid oxidation and vitamin degradation during fermentation. The volcano plot and heatmap analyses collectively indicated that LAB fermentation reshaped metabolites profile, with organic acids and vitamin-related metabolites showing particularly clear responses.

3.2.2 Enrichment analysis

KEGG-based enrichment analysis was applied to reveal the metabolic pathways involved in wolfberry juice fermentation. The top 25 enriched pathways (Supplementary Material Figs. S1F and S1G) were primarily associated with amino acid, lipid, carbohydrate, and vitamin metabolism. Based on P values, phenylalanine metabolism, phenylalanine, tyrosine and tryptophan biosynthesis, and tyrosine metabolism were the three most significantly enriched pathways, consistent with previous findings on LAB-fermented fruit and vegetable matrices [43,44]. In addition, pathway enrichment analysis similarly showed that fermentation changed the oxidative microenvironment of wolfberry juice. Based on the volcano plot, heatmap results, and KEGG enrichment analysis, organic acids, vitamins, fatty acids, and amino acids were selected for targeted quantification to clarify the metabolite-level changes underlying fermentation-induced acidic and redox microenvironmental shifts.

3.3 Targeted quantification of key metabolites in NWJ and FWJ

3.3.1 Changes in the acidic microenvironment

Based on untargeted metabolomic results, six representative organic acids, including oxalic acid, lactic acid, citric acid, DL-malic acid, acetic acid, and succinic acid, were quantified in NWJ and FWJ (Fig. 2A). Fermentation sharply increased lactic acid to (4.8 ± 0.1) mg/mL, while citric acid, DL-malic acid, acetic acid, and succinic acid decreased. The pronounced accumulation of lactic acid suggested the establishment of an acidified microenvironment during L. plantarum NCU137 fermentation. This interpretation was further supported by the concomitant depletion of DL-malic acid, as malic acid conversion to lactic acid was a typical feature of L. plantarum-mediated malolactic fermentation [45]. Considerable depletion of malic acid with concurrent lactic acid accumulation had also been observed in L. plantarum-fermented apple juice [46]. Given that acidic conditions have been reported to promote partial polysaccharide hydrolysis and reduce MW [47,48], the lactic acid-enriched microenvironment may have contributed to the non-enzymatic depolymerization of wolfberry polysaccharides. This interpretation was consistent with the altered monosaccharide composition detected after fermentation.

3.3.2 Changes in the redox microenvironment

Analysis above suggested that fermentation also altered metabolites associated with the redox microenvironment. Therefore, vitamins, fatty acids, and free amino acids in NWJ and FWJ were quantified (Figs. 2B–2D). Among these metabolites, vitamin C was considered as the primary redox-active candidate due to its abundance in wolfberry juice and showed direct relevance to reactive oxygen species scavenging and redox-state regulation in fruit matrices [49].

As shown in Fig. 2B, the significant decrease in vitamin C from (2847.3 ± 15.6) to (2672.0 ± 30.6) μg/g suggested a redox-shifted microenvironment during fermentation. Free radical-mediated reactions have been reported to cleave polysaccharide glycosidic linkages mainly through hydrogen abstraction, leading to polysaccharide depolymerization and MW reduction [50]. Therefore, vitamin C depletion may have reflected oxidative consumption during fermentation and potentially contributed to the non-enzymatic structural remodeling of wolfberry polysaccharides.

Fatty acid changes provided additional evidence for redox-related remodeling. As shown in Fig. 2C, seven long-chain fatty acids were further quantified using an internal standard method. C16:0 (palmitic acid), 9cC18:1 (oleic acid), and 9c12cC18:2 (linoleic acid) were the predominant fatty acids in wolfberry juice, representing the major saturated, monounsaturated, and polyunsaturated fatty acids, respectively, which was in line with previously reported results [51]. After fermentation, the total content of quantified fatty acids decreased from (65.9 ± 11.4) to (22.5 ± 6.8) mg/g. Both saturated and unsaturated fatty acids decreased, from (29.6 ± 4.6) to (12.7 ± 3.6) mg/g and from (38.4 ± 7.0) to (11.2 ± 3.5) mg/g, respectively. Notably, the relative proportion of unsaturated fatty acids decreased from (58.1 ± 0.6)% to (49.5 ± 1.1)%, indicating a more pronounced decrease in unsaturated fatty acids than in saturated fatty acids. This greater reduction may have been associated with the higher susceptibility of C=C bonds to microbial transformation and oxidative degradation [52,53]. Given that radical-mediated reactions can induce polysaccharide depolymerization [50], the decrease in unsaturated fatty acids was considered auxiliary evidence for redox-related changes during fermentation. However, fatty acid depletion found in the current study mainly reflected complex lipid metabolism and oxidative transformation, rather than a direct and controllable polysaccharide modification pathway.

Free amino acid profiling revealed a depletion-dominated pattern after fermentation, with seven of the thirteen detected amino acids significantly decreased and three significantly increased (Fig. 2D). This result indicated extensive amino acid remodeling during L. plantarum fermentation and corresponded with the enrichment of amino acid-related pathways observed in untargeted metabolomics. Notably, tyrosine, a phenolic amino acid with redox-related properties, decreased from (0.12 ± 0.0) mg/mL to below the detection limit, suggesting that specific redox-active amino acids were strongly affected by fermentation. Similar amino acid remodeling has been reported in lactic acid bacteria-fermented wolfberry juice, supporting the reliability of this metabolite response phenomenon [54]. However, free amino acid profile was closely associated with fermentation performance such as microbial nitrogen utilization, precursor conversion, and cell growth. Therefore, the changes in free amino acids were considered as the supporting evidence for fermentation-induced redox changes, rather than directly responsible for polysaccharide structural modification.

In summary, quantitative analysis revealed that lactic acid accumulation was the predominant factor contributing to acidification of the microenvironment, while vitamin C depletion served as a representative indicator of redox-related changes during fermentation, given its central role in cellular redox balance. In contrast, changes in fatty acids and free amino acids were considered supporting indicators due to their more complex involvement in microbial metabolism, nutrient utilization, and lipid transformation. Therefore, lactic acid and vitamin C were identified as representative non-enzymatic factors.

3.4 In vitro simulation validation

To further validate whether or not the metabolomics-inferred acidic and redox microenvironmental changes could directly contribute to polysaccharide structural remodeling, an in vitro simulation was conducted using NWP as the substrate, and lactic acid and vitamin C were selected as representative non-enzymatic factors for subsequent simulation experiments based on the above quantitative analysis.

3.4.1 Physicochemical properties and monosaccharide composition of NWP under different treatments

The corresponding results were summarized in Table 3. As described above, LAB fermentation induced compositional remodeling of wolfberry polysaccharides, including increased neutral sugar content, decreased uronic acid and protein contents, and altered monosaccharide proportions. Compared with NWP-K, NWP-Acid showed an increase in neutral sugar content from (43.0 ± 0.7)% to (47.4 ± 1.7)%, accompanied by decreases in Ara and GlcA and increases in Rha, Gal, and Glc. These changes suggested that lactic acid-mediated acidification promoted partial cleavage of polysaccharide fractions, thereby altering the relative abundance of sugar residues. NWP-VC exhibited a different compositional response, with decreased neutral sugar and protein contents and an increase in GalA from (7.1 ± 0.7)% to (11.4 ± 0.5)%, indicating that vitamin C-associated redox changes affected polysaccharide-associated components through a pathway different from acid treatment. Notably, NWP-Acid+VC showed a more fermentation-like shift, characterized by increased neutral sugar content, decreased uronic acid content, increased Rha, and decreased Ara. Similar non-enzymatic treatments, such as ultrasound-assisted H₂O₂/VC treatment, have also been reported to alter the relative molar percentages of Rha, Ara, Gal, and Glc without changing monosaccharide types [55]. Therefore, the compositional changes induced by Acid+VC treatment suggested that lactic acid accumulation and vitamin C-associated redox changes could jointly participate in the non-enzymatic structural modification of wolfberry polysaccharides during fermentation.

3.4.2 The MW of NWP under different treatments

The MW distributions were shown in Fig. 3. Compared with NWP-K, NWP-Acid and NWP-VC showed only slight changes in the elution profiles, indicating that lactic acid or vitamin C alone had limited effects on the MW of wolfberry polysaccharides. In contrast, NWP-Acid+VC induced a more pronounced redistribution of MW fractions. Specifically, the peak corresponding to the relatively high-molecular-weight fraction decreased, whereas the peak corresponding to the relatively low-molecular-weight fraction increased. This shift indicated that the combined acidic and redox conditions promoted partial depolymerization of NWP and accumulation of lower-molecular-weight polysaccharide fractions. Mechanistically, lactic acid may create a mildly acidic environment favorable for glycosidic bond hydrolysis, while vitamin C-mediated redox reactions may generate hydroxyl radicals that contribute to polysaccharide chain scission and structural degradation [17].

4 Conclusions

In this study, the impact of fermentation on the structure of wolfberry polysaccharides, as well as the non-enzymatic reaction factors were systematically investigated. It was found that fermentation could decrease the uronic acid content, induce a smoother surface morphology, and affect the side-chain structure, particularly by reducing the relative abundance of Af₅ residues, indicating partial structural remodeling of wolfberry polysaccharides by the LAB fermentation, with fundamental arabinogalactan-like structure largely preserved. Notably, non-enzymatic microenvironmental factors, such as organic acids and vitamin C, were demonstrated to contribute to the structural modification of wolfberry polysaccharides during fermentation by non-enzymatic reactions, based on untargeted metabolomics, targeted quantification and in vitro simulation test.

Collectively, this study explored the potential contribution of non-enzymatic reactions to fermentation-induced structural changes in wolfberry polysaccharides. Integrating metabolomics with molecular-level structural characterization established a mechanistic link between microenvironmental changes and polysaccharide modification, providing a theoretical basis for precision regulation of bioactive polysaccharides in fermented fruit and vegetable products. Further studies are needed to quantitatively distinguish enzymatic and non-enzymatic contributions and clarify the cleavage patterns of key glycosidic linkages.

References

[1] Zhang Y , Sun M X , Feng Y . et al. Integrating metabolomics, proteomics, transcriptomics, and genomics to enhance the flavor-modulating capacity of lactic acid bacteria in fermented beverages. Trends in Food Science & Technology, 2026, 171: 105646

[2] Liu S , He Y X , He W W . et al. Exploring the biogenic transformation mechanism of polyphenols by Lactobacillus plantarum NCU137 fermentation and its enhancement of antioxidant properties in wolfberry juice. Journal of Agricultural and Food Chemistry, 2024, 72(22): 12752–12761

[3] Hu S K , Xu Y T , Shadrack S M . et al. Lycium barbarum (goji berry): a comprehensive review of chemical composition, bioactive compounds, health-promoting activities, and applications in functional foods and beyond. Food Chemistry, 2025, 496: 146588

[4] Lan T , Zhou K X , Duan G Z . et al. Lycium L. flavonoids: extraction, purification, signal transduction pathways, and interactions with intestinal microbiota. npj Science of Food, 2026, 10(1): 128

[5] Miranda M R , Vestuto V , Amodio G . et al. Antitumor mechanisms of Lycium barbarum fruit: an overview of in vitro and in vivo potential. Life, 2024, 14(3): 420

[6] Niu Y H , Liao J L , Zhou H T . et al. Flavonoids from Lycium barbarum leaves exhibit anti-aging effects through the redox-modulation. Molecules, 2022, 27(15): 4952

[7] Zhou W T , Yang T T , Xu W Q . et al. The polysaccharides from the fruits of Lycium barbarum L. confer anti-diabetic effect by regulating gut microbiota and intestinal barrier. Carbohydrate Polymers, 2022, 291: 119626

[8] Li D , Jin J . Regulatory effects of Lycium barbarum polysaccharides on immune function and their application prospects. Frontiers in Immunology, 2025, 16: 1671971

[9] Xie H , Gao P Y , Lu Z M . et al. Changes in physicochemical characteristics and metabolites in the fermentation of goji juice by Lactiplantibacillus plantarum. Food Bioscience, 2023, 54: 102881

[10] Wang B , Yang J , Tao L J . et al. Structural-activity relationship of Lycium barbarum polysaccharides in immunomodulation: integrating molecular insights with target identification for therapeutic development. Frontiers in Immunology, 2026, 17: 1730418

[11] Chen Y , Li L , Zhang R Y . et al. A review of Lycium barbarum polysaccharides: structural characteristics, antitumor molecular mechanism and transformation application. International Journal of Biological Macromolecules, 2025, 333: 148743

[12] Tian X J , Liang T S , Liu Y L . et al. Extraction, structural characterization, and biological functions of Lycium barbarum polysaccharides: a review. Biomolecules, 2019, 9(9): 389

[13] Wang X J , Hu K D , Chen Y . et al. Effect of Lactiplantibacillus plantarum fermentation on the physicochemical, antioxidant activity and immunomodulatory ability of polysaccharides from Lvjian okra. International Journal of Biological Macromolecules, 2024, 257: 128649

[14] Huang G T , Su D X , Lee Y K . et al. Accumulation of water-soluble polysaccharides during lychee pulp fermentation with Lactiplantibacillus plantarum involves endoglucanase expression. Journal of Agricultural and Food Chemistry, 2025, 73(6): 3669–3679

[15] Peng J , Bu Z B , Ren H Y . et al. Physicochemical, structural, and functional properties of wampee (Clausena lansium (Lour. ) Skeels) fruit peel pectin extracted with different organic acids. Food Chemistry, 2022, 386: 132834

[16] Yuan E D , Nie S Y , Liu L Y . et al. Study on the interaction of Hericium erinaceus mycelium polysaccharides and its degradation products with food additive silica nanoparticles. Food Chemistry: X, 2021, 12: 100172

[17] Kamlow M A , Marks J , Reid J E S J . et al. Ascorbic acid mediated hydrolysis of galactomannans. Food Hydrocolloids, 2026, 172: 111996

[18] Liu Z G , Cheng H , Li D Y . et al. Optimizing the fermentation conditions of fermented goji using sensory analysis and the biomass of Lactiplantibacillus plantarum NCU137. Journal of Food Processing and Preservation, 2022, 46(10): e16828

[19] Shen W T , He W W , Hu X Y . et al. Kidney protective effect of fermented celery polysaccharides and Eubacterium rectale in diabetic rats. Food Bioscience, 2026, 80: 108902

[20] DuBois M , Gilles K A , Hamilton J K . et al. Colorimetric method for determination of sugars and related substances. Analytical Chemistry, 1956, 28(3): 350–356

[21] Bitter T , Muir H M . A modified uronic acid carbazole reaction. Analytical Biochemistry, 1962, 4(4): 330–334

[22] Grintzalis K , Georgiou C D , Schneider Y J . An accurate and sensitive Coomassie Brilliant Blue G-250-based assay for protein determination. Analytical Biochemistry, 2015, 480: 28–30

[23] Tang W , Liu C C , Liu J J . et al. Purification of polysaccharide from Lentinus edodes water extract by membrane separation and its chemical composition and structure characterization. Food Hydrocolloids, 2020, 105: 105851

[24] Wan Y J , Hong T , Shi H F . et al. Probiotic fermentation modifies the structures of pectic polysaccharides from carrot pulp. Carbohydrate Polymers, 2021, 251: 117116

[25] Tang W , Liu D , Li Y . et al. Structural characteristics of a highly branched and acetylated pectin from Portulaca oleracea L. Food Hydrocolloids, 2021, 116: 106659

[26] Liu S , Peng Y J , He W W . et al. Metabolomics-based mechanistic insights into antioxidant enhancement in mango juice fermented by various lactic acid bacteria. Food Chemistry, 2025, 466: 142078

[27] He Y X , Hu M W , He W W . et al. Volatile compound dynamics during blueberry fermentation by lactic acid bacteria and its potential associations with bacterial metabolism. Food Bioscience, 2024, 59: 103639

[28] Xu Z H , Liu S S , Shen M Y . et al. Evaluation of trans fatty acids, carbonyl compounds and bioactive minor components in commercial linseed oils. Food Chemistry, 2022, 369: 130930

[29] Liu S , Li Y H , Song X X . et al. Changes in volatile and nutrient components of mango juice by different lactic acid bacteria fermentation. Food Bioscience, 2023, 56: 103141

[30] Zhang M , Tang X L , Wang F . et al. Characterization of Lycium barbarum polysaccharide and its effect on human hepatoma cells. International Journal of Biological Macromolecules, 2013, 61: 270–275

[31] Wang Z C , Zheng Y , Zhou X Y . et al. Effect of Lactobacillus fermentation on the structural feature, physicochemical property, and bioactivity of plant and fungal polysaccharides: a review. Trends in Food Science & Technology, 2024, 148: 104492

[32] Cao C , Wang Z F , Gong G P . et al. Effects of Lycium barbarum polysaccharides on immunity and metabolic syndrome associated with the modulation of gut microbiota: a review. Foods, 2022, 11(20): 3177

[33] Hu T G , Zhu W L , Yu Y S . et al. The variation on structure and immunomodulatory activity of polysaccharide during the longan pulp fermentation. International Journal of Biological Macromolecules, 2022, 222: 599–609

[34] Hu T G , Wu H , Yu Y S . et al. Preparation, structural characterization and prebiotic potential of mulberry leaf oligosaccharides. Food & Function, 2022, 13(9): 5287–5298

[35] Pei F Y , Cao X B , Wang X M . et al. Structural characteristics and bioactivities of polysaccharides from blue honeysuckle after probiotic fermentation. LWT, 2022, 165: 113764

[36] Yang J J , Zhang X , Dai J F . et al. Effect of fermentation modification on the physicochemical characteristics and anti-aging related activities of Polygonatum kingianum polysaccharides. International Journal of Biological Macromolecules, 2023, 235: 123661

[37] Wu J X , Chen T , Wan F Q . et al. Structural characterization of a polysaccharide from Lycium barbarum and its neuroprotective effect against β-amyloid peptide neurotoxicity. International Journal of Biological Macromolecules, 2021, 176: 352–363

[38] Liang X F , Liu M Q , Guo S . et al. Structural elucidation of a novel arabinogalactan LFP-80-W1 from Lycii fructus with potential immunostimulatory activity. Frontiers in Nutrition, 2023, 9: 1067836

[39] Huang W Q , Zhao M W , Wang X Y . et al. Revisiting the structure of arabinogalactan from Lycium barbarum and the impact of its side chain on anti-ageing activity. Carbohydrate Polymers, 2022, 286: 119282

[40] Wang Z F , Liu Y , Sun Y J . et al. Structural characterization of LbGp1 from the fruits of Lycium barbarum L. Food Chemistry, 2014, 159: 137–142

[41] Xu N , Lai Y H , Shao X F . et al. Different analysis of flavors among soft-boiled chicken: based on GC-IMS and PLS-DA. Food Bioscience, 2023, 56: 103243

[42] Meng F B , Lei Y T , Li Q Z . et al. Effect of Lactobacillus plantarum and Lactobacillus acidophilus fermentation on antioxidant activity and metabolomic profiles of loquat juice. LWT, 2022, 171: 114104

[43] Zhang J G , Zhang Y , Zhang W W . et al. Differential metabolites analysis in Lycium barbarum and Platycodon grandiflorus fermented wine by untargeted metabolomics. Applied Food Research, 2024, 4(2): 100631

[44] An X , Li T T , Hu J X . et al. Evaluation of physicochemical characteristics, bioactivity, flavor profile and key metabolites in the fermentation of goji juice by Lacticaseibacillus rhamnosus. Food Chemistry: X, 2024, 23: 101755

[45] Markkinen N , Pariyani R , Jokioja J . et al. NMR-based metabolomics approach on optimization of malolactic fermentation of sea buckthorn juice with Lactiplantibacillus plantarum. Food Chemistry, 2022, 366: 130630

[46] Liang J R , Deng H , Hu C Y . et al. Vitality, fermentation, aroma profile, and digestive tolerance of the newly selected Lactiplantibacillus plantarum and Lacticaseibacillus paracasei in fermented apple juice. Frontiers in Nutrition, 2022, 9: 1045347

[47] Lupo C , Boulos S , Gramm F . et al. A microcalorimetric and microscopic strategy to assess the interaction between dietary fibers and small molecules. Carbohydrate Polymers, 2022, 287: 119229

[48] Chen S K , Wang X , Guo Y Q . et al. Exploring the partial degradation of polysaccharides: structure, mechanism, bioactivities, and perspectives. Comprehensive Reviews in Food Science and Food Safety, 2023, 22(6): 4831–4870

[49] Arabia A , Munné-Bosch S , Muñoz P . Ascorbic acid as a master redox regulator of fruit ripening. Postharvest Biology and Technology, 2024, 207: 112614

[50] Chen X Y , Sun-Waterhouse D , Yao W Z . et al. Free radical-mediated degradation of polysaccharides: mechanism of free radical formation and degradation, influence factors and product properties. Food Chemistry, 2021, 365: 130524

[51] Skenderidis P , Lampakis D , Giavasis I . et al. Chemical properties, fatty-acid composition, and antioxidant activity of goji berry (Lycium barbarum L. and Lycium chinense Mill.) fruits. Antioxidants, 2019, 8(3): 60

[52] Iorizzo M , Di Martino C , Letizia F . et al. Production of conjugated linoleic acid (CLA) by Lactiplantibacillus plantarum: a review with emphasis on fermented foods. Foods, 2024, 13(7): 975

[53] Hennebelle M , Villeneuve P , Durand E . et al. Lipid oxidation in emulsions: new insights from the past two decades. Progress in Lipid Research, 2024, 94: 101275

[54] Duan W H , Guan Q J , Zhang H L . et al. Improving flavor, bioactivity, and changing metabolic profiles of goji juice by selected lactic acid bacteria fermentation. Food Chemistry, 2023, 408: 135155

[55] Cai W D , Luo Y C , Xue J Y . et al. Effect of ultrasound assisted H₂O₂/Vc treatment on the hyperbranched Lignosus rhinocerotis polysaccharide: structures, hydrophobic microdomains, and antitumor activity. Food Chemistry, 2024, 450: 139338

Rights & permissions

The Author(s) 2026. This article is published by Higher Education Press.