The research team, based at Jianghan University and collaborating institutions in Hubei Province, started from cowpea (Vigna unguiculata L.), a legume prized for its nutritional value, environmental adaptability, and contribution to food security. Cowpea seeds were washed, soaked, sterilized, and inoculated with Bacillus subtilis natto (CICC 10263), a strain selected for its exceptional nattokinase production capacity, exceeding 5000 FU/g under optimized conditions of 37°C and pH 7.0, and for its thermostable fermentation adaptability. Solid-state fermentation proceeded at 37°C for 24 hours, followed by after-ripening at 4°C and freeze-drying. Polysaccharides were then extracted from both raw cowpea seeds and the fermented cowpea natto by hot water extraction, degreasing with ethanol, acidic precipitation with trichloroacetic acid, and ethanol precipitation, yielding crude extracts named JD and JD-F, respectively.

Fermentation measurably improved polysaccharide recovery. The extraction yield of JD-F from fermented cowpea was 6.17% ± 0.33%, significantly higher than the 4.55% ± 0.34% obtained from raw seeds. The researchers attribute this increase to fermentation-induced changes in the cowpea matrix: microbial growth and enzyme secretion disrupt plant cell-wall networks, release bound polysaccharides, and weaken interactions between polysaccharides and other macromolecular components, thereby facilitating polysaccharide release during subsequent extraction. Chemical composition analysis reinforced the enrichment. JD contained 33.58% ± 2.82% total carbohydrate, 8.47% ± 0.43% protein, and 1.01% ± 0.49% uronic acid. After fermentation and crude extraction, total carbohydrate in JD-F rose to 44.03% ± 2.15% and uronic acid to 8.22% ± 1.54%, while protein fell below the detection limit, consistent with both fermentation-associated improvements in carbohydrate extractability and trichloroacetic-acid-mediated deproteinization.

JD-F was selected for detailed investigation because of its relatively high polysaccharide purity, abundant uronic acid content, and low protein content. Fractionation by ion-exchange chromatography on DEAE-Sepharose Fast Flow yielded four carbohydrate-containing fractions eluted with distilled water and 0.2, 0.5, and 1.0 mol/L NaCl. Molecular-weight analysis indicated that the water-eluted fraction, JDF-E1, consisted mainly of polysaccharides, whereas the 0.2 mol/L NaCl fraction contained predominantly oligosaccharides and other low-molecular-weight carbohydrate components. JDF-E1 was therefore further purified by size-exclusion chromatography on Sephacryl S-300 HR, affording the homogeneous fraction JDFE1-F. High-performance gel permeation chromatography showed a single predominant polysaccharide peak corresponding to a weight-average molecular weight of 5.175 kDa, with a polydispersity index of 1.36, indicating a relatively narrow molecular-weight distribution. Total carbohydrate content of JDFE1-F reached 85.12% ± 3.09%, with uronic acid at 9.16% ± 0.77% and no detectable protein.

Structural characterization proceeded through a battery of complementary techniques. Fourier-transform infrared spectroscopy revealed a broad O-H stretching band at 3405 cm⁻¹, C-H stretching at 2930 cm⁻¹, bound-water bending at approximately 1640 cm⁻¹, and prominent C-O-C and C-O stretching bands at 1154, 1081, and 1020 cm⁻¹ characteristic of glycosidic linkages and pyranose rings. An absorption band at 855 cm⁻¹, commonly associated with α-glycosidic configuration, supported the predominance of α-linked glucopyranose residues. Monosaccharide composition analysis by HPLC after trifluoroacetic acid hydrolysis and PMP derivatization showed that JDFE1-F was a glucose-rich polysaccharide containing more than 98% glucose, with only trace amounts of mannose (0.024%), galactose (0.557%), and arabinose (1.199%).

Methylation analysis followed by gas chromatography–mass spectrometry identified four glycosidic linkage types: 1,4-Glcp (81.023%), t-Glcp (12.001%), 1,4,6-Glcp (5.575%), and 1,5-Araf (1.400%). The dominant 1,4-Glcp residues form the main backbone, while the smaller proportion of 1,4,6-Glcp residues act as branching points substituted at O-6, and terminal glucose units cap the minor side chains. Nuclear magnetic resonance spectroscopy, including one-dimensional ¹H and ¹³C spectra and two-dimensional COSY, HSQC, HSQC-TOCSY, HMBC, and NOESY experiments in D₂O, confirmed these assignments. A prominent anomeric cross-peak at δ 5.24/99.35 ppm was assigned to →4)-α-D-Glcp-(1→, and another at δ 5.21/99.32 ppm to →4,6)-α-D-Glcp-(1→. Long-range HMBC correlations and spatial NOESY cross-peaks established consecutive →4)-α-D-Glcp-(1→ and →4,6)-α-D-Glcp-(1→ residues linked through O-4, with the O-6 position of the branched residues substituted by either terminal α-D-Glcp-(1→ or →5)-α-L-Araf-(1→ units, the latter further bearing terminal glucose at O-5. Reducing-end α- and β-glucopyranose residues were also identified from signals at δ 5.05/91.73 and 4.48/95.54 ppm.

With the structure established, the team comparatively evaluated the α-amylase inhibitory activity of JD, JD-F, and JDFE1-F. All three preparations inhibited the enzyme in a concentration-dependent manner. At 4 mg/mL, inhibition rates reached 44.43% ± 1.17% for JD, 50.80% ± 0.78% for JD-F, and 52.55% ± 1.37% for JDFE1-F. The IC₅₀ values of JD-F and JDFE1-F were 3.509 ± 0.311 and 2.668 ± 0.281 mg/mL, respectively, both significantly lower than that of JD at 7.286 ± 0.992 mg/mL (p < 0.05). Acarbose served as the positive control. The stronger activity of JD-F relative to JD likely reflects fermentation-induced changes in polysaccharide extractability and molecular structure, including alterations in molecular-weight distribution, monosaccharide proportions, branching, and functional-group accessibility. The numerically lower IC₅₀ of JDFE1-F than JD-F may partly reflect its higher polysaccharide purity, since purification increased total carbohydrate content from 44.03% to 85.12%, providing a greater effective amount of polysaccharide at the same sample mass.

The mechanism of α-amylase inhibition remains a subject of careful interpretation. Polysaccharides may inhibit the enzyme through direct binding, adsorption onto starch, or restriction of enzyme-substrate contact. JDFE1-F is structurally distinct from high-molecular-weight hydroxyl-rich polymers such as konjac glucomannan, whose viscosity-mediated diffusion barriers around starch granules have been proposed as an inhibitory mechanism. Given its relatively low molecular weight of 5.175 kDa, such a barrier is unlikely to be the principal explanation. Instead, the glucose-rich α-(1→4)-linked glucan structure may allow JDFE1-F to interact with the carbohydrate-binding subsites of α-amylase through substrate-like recognition and hydrogen bonding, thereby interfering with enzyme-substrate association. However, because α-(1→4)-linked glucans can also serve as substrates for α-amylase, the present results cannot distinguish direct enzyme inhibition from alternative-substrate competition or related assay-dependent effects, and a specific kinetic inhibition type cannot be assigned from the current data.

Bile acid-binding capacity was assessed through a simulated gastrointestinal digestion protocol involving pepsin and pancreatin treatment, followed by incubation with sodium glycodeoxycholate or sodium taurodeoxycholate and quantification of unbound bile acids by absorbance at 387 nm after sulfuric acid dehydration. All three preparations bound more than 50% of both bile acids at 4 mg/mL, with JDFE1-F binding 58.09% ± 2.58% of sodium glycodeoxycholate and 60.79% ± 2.87% of sodium taurodeoxycholate. The concentration required for 50% sodium glycodeoxycholate binding was 1.711 ± 0.574 mg/mL for JDFE1-F, significantly lower than that of JD at 3.431 ± 0.898 mg/mL (p < 0.05), whereas JDFE1-F and JD-F did not differ significantly. No significant differences among the three preparations were detected for sodium taurodeoxycholate binding. The statistically supported improvement was therefore restricted to glycodeoxycholate binding by JDFE1-F relative to JD, rather than representing a general enhancement toward both bile acids. Differences in conjugated groups, charge distribution, steric configuration, and hydrophobicity between the two bile acids may result in different affinities for the same α-glucan structure.

The authors acknowledge several limitations. Because no specific de-starching treatment was applied, a contribution from starch-derived α-glucans to JDFE1-F cannot be completely excluded. The structural changes occurring during fermentation were not fully resolved, and the α-amylase inhibitory and bile acid-binding activities were evaluated only in vitro. Future studies should clarify the structural origin, structure-activity relationships, and physiological relevance of JDFE1-F through comparative structural analysis, mechanistic studies, simulated digestion, and in vivo validation. Nevertheless, the work extends current knowledge of legume-derived polysaccharides and establishes a basis for further evaluation of fermented cowpea glucans as functional food ingredients, linking a precisely defined molecular architecture with measurable in vitro functional performance.

Subject of Research: Structural characterization and in vitro functional properties of a glucan from fermented cowpea natto

Article Title: Structural Characterization and In Vitro α‐Amylase Inhibitory and Bile Acid‐Binding Properties of a Glucan Isolated From Fermented Cowpea Natto

Article References: Yu, J., Xu, M., Li, H., Yan, E., Niu, H., Liu, Q., Zhang, H., & Wang, H. (2026). Structural Characterization and In Vitro α ‐Amylase Inhibitory and Bile Acid‐Binding Properties of a Glucan Isolated From Fermented Cowpea Natto. Food Science & Nutrition, 14(10), Article e72456. https://doi.org/10.1002/fsn3.72456

Image Credits: AI Generated

DOI: 10.1002/fsn3.72456

Keywords: Structural, Characterization, Vitro, Amylase, Inhibitory, Bile, Acid, Binding, Properties, Glucan, Isolated, Fermented