The chemical components, including total sugar, uronic acid, and protein, were determined according to the published methods (14–16).

The M_w of BSP was measured by gel permeation chromatography (WATERS-2695-2414, Waters Corporation, Milford, MA, United States) equipped with a TSKgelGMPWXL column (7.8 × 300 mm) (TOSOH Biotechnology Co., Ltd. Shanghai, China) and an RID-20 refractive index detector (Shimadzu, Japan). A 10-μL BSP solution (2 mg/mL) filtered through a 0.45-μm membrane was loaded on the instrument with deionized water (0.1% NaNO₃) as the eluent at 35°C. The standard curve was obtained based on T-series dextrans (T-5, T-10, T-40, T-70, and T-110) (lgM_w = −0.48 t + 8.91, R² = 0.9958).

Referencing the method in a prior study, the monosaccharide composition of BSP was analyzed (17). BSP was hydrolyzed with a 2 mol/L HCl solution and then derivatized using PMP before detection by high-performance liquid chromatography (WATERS-2695-2414, Waters Corporation, Massachusetts, United States). D-glucose, D-mannose, D-galactose, L-arabinose, D-xylose, and L-rhamnose were used as standards using the same assay method. The monosaccharide composition was determined by comparing retention time and peak area.

The UV spectrum of the BSP solution (1.0 mg/mL) was determined using a double-beam UV spectrophotometer (Helios Gamma, Thermo Corporation, Wales, England) in the range of 200–500 nm.

The dried BSP powder was mixed with KBr and ground into flakes. A Fourier transform infrared spectrometer (ALPHA-T, Bruker Co., DE, Germany) was used to record the spectra in the range of 400–4,000 cm⁻¹ with a resolution of 4 cm⁻¹ and an accumulation of 32 scans.

Methylation of BSP was carried out according to the method of Ciucanu and Kerek (18). Then, the methylated polysaccharide was hydrolyzed with 2 mol/L TFA at 100°C for 6 h, and excess acid was removed by evaporation. The hydrolysates were reduced with NaBH₄ for 24 h and acetylated with acetic anhydride-pyridine (1:1) at 100°C for 2 h. These acetates were analyzed with gas chromatography–mass spectrometry (GC–MS) (Varian-450-GC-320-MS, Agilent Co., Massachusetts, United States) using an HP-5 MS fused silica capillary column (30 m × 0.25 mm, 0.25 μm, Agilent, NY, United States). The parameters were set as follows: the injection temperature was 220°C, the initial temperature was 160°C for 2 min, followed by 5°C/min to 210°C for 1 min, and finally 10°C/min to 260°C for 10 min.

BSP was dissolved in D₂O with a concentration of 20 mg/mL. Proton nuclear magnetic resonance (¹H NMR) and carbon-13 nuclear magnetic resonance spectroscopy (¹³C NMR) spectra were determined using a Bruker DRX-600 NMR spectrometer (Bruker, Germany) at 25°C.

First, the BSP solution (2.0 mg/mL, 2.0 mL) and the Congo red solution (80 μmol/L, 2.0 mL) were mixed, and the NaOH solution (1.0 mol/L, 0.0–4.0 mL) was separately added into the mixture. The λmax of the mixed solution was then tested by a UV spectrometer in the range of 400–700 nm.

The BSP solution (1.0 mg/mL, 1.0 mL) was added to the iodine reagent (0.5 mL, containing 0.2% KI and 0.02% I₂, w/v), and the mixture was stirred for 10 min. The absorbance, from 300 to 700 nm, was determined using an Agilent 8453 G1103A ultraviolet–visible (UV–Vis) spectrophotometer (Beijing Purkinje General Instrument Co., Beijing, China).

Appropriate amounts of BSP powders were glued onto the sample stage and plated with a gold film using the sputtering method. The morphological features were observed using a scanning electron microscope (Thermo Fisher Scientific Inc., Waltham, MA, United States) at an accelerating voltage of 5.0 kV.

The BSP solution (5.0 μg/mL, 10.0 μL) was deposited onto the mica slice and allowed to dry naturally. Then, the atomic force microscopy (AFM) image was obtained using a Dimension FastScan (Bruker, Germany) at a scanning range of 5 μm × 5 μm and a frequency of 1 Hz.

First, each 1.0 mL BSP solution (0.0, 1.0, 2.0, and 4.0 mg/mL) was mixed with 1.0-mL α-amylase solution (0.0, 0.2, 0.4, 0.6, and 0.8 mg/mL) and stirred for 5 min at 37°C; then, the starch solution (1.0 mL, 2.0 mg/mL) was poured to the mixture. After a 30-min reaction, 2.5-mL of dinitrosalicylic acid was added to the mixture, and the reaction was terminated by incubation in boiling water for 5 min. During the reaction, the absorbance of the solution was continuously tested at 540 nm at 25°C. The reaction rate v (ΔA/min) is calculated according to the change value of the linear change of absorbance over time divided by time, with the enzyme concentration as the abscissa and the reaction rate as the ordinate to determine the inhibition type of polysaccharide on α-amylase activity.

The type of reversible inhibition of α-amylase activity was determined. The α-amylase solution (0.6 mg/mL) was mixed with a 1.0-mL BSP solution (0.0, 1.0, 2.0, and 4.0 mg/mL) and then incubated at 37°C for 5 min. Next, a 1.0-mL starch solution (0.5, 1.0, 1.5, 2.0, and 2.5 mg/mL) was added to the mixture. Using the method described above, the absorbance was tested at 540 nm every 2 min. The Lineweaver–Burk curve, using 1/[S] as the abscissa and 1/v as the ordinate, determines the type of reversible inhibition of α-amylase and calculates the enzyme inhibition constant K_i.

The BSP solution (40 μL; 0.0, 0.5, 1.0, and 2.0 mg/mL) and pNPG solution (20 μL, 2.0 mmol/L) were added sequentially to 96-well plates, shaken for 1 min, and incubated at 37°C for 5 min. Then, the α-glucosidase solution (40 μL, 0.0–0.16 U/mL) was added to the mixture. Next, the 96-well plate was put into a microplate reader (Spark 10 M, Tecan Group Ltd., Switzerland). The initial velocity (v) was determined by monitoring the absorbance at 405 nm every 1 min. The inhibition (reversible or irreversible type) was confirmed by plotting v against [E].

Reversible inhibition was further studied. The BSP solution (40 μL, 0.0–2.0 mg/mL) and the α-glucosidase solution (40 μL, 0.16 U/mL) were added to 96-well plates, shaken for 1 min, and incubated at 37°C for 5 min. The pNPG solution (20 μL, 0.25–4.0 mmol/L) was added to the mixture. Following the same operation as described above, the absorbance was recorded every 1 min, and the Lineweaver–Burk curves were obtained.

The L. plantarum and L. acidophilus strains were activated using modified MRS medium base according to the reported methods (20). Carbohydrate-free MRS broth supplemented with 0.1% (w/v) BSP was used as the basal medium to investigate the proliferation of bacterial strains. Glucose was used as the positive control. The active L. plantarum and L. acidophilus strains were separately transferred to the MRS broth medium at a concentration of 1 × 10⁶ CFU/mL and then incubated in an anaerobic chamber at 37°C for 48 h under anaerobic conditions (85% N₂, 10% CO₂, and 5% H₂).

The proliferation of the two strains in the media was determined by measuring the optical density at 600 nm at intervals of 12, 18, 24, and 30 h during the incubation period. The pH of the medium was determined using a pH meter (Thermo Fisher Scientific (China) Co., Ltd, Beijing, China) at 6 h intervals.

After 30 h incubation in BSP medium, the probiotics were centrifuged, washed twice with 0.9% NaCl, and then fixed with 2% glutaraldehyde. The morphology of the probiotics was observed using SEM at a 20.0 kV acceleration voltage.

After 30 h incubation, the levels of short-chain fatty acids (SCFAs) in the medium were analyzed for acetic acid, propionic acid, and butyric acid using GC–MS (7890 GC system equipped with 7000C Triple Quad MSD, Agilent) equipped with an Agilent DB-FFAP column (30 m × 0.25 μm). Helium was used as the carrier gas. The inlet temperature was 230°C. The column temperature was improved from 90°C to 120°C at 8°C/min, then to 150°C at 5°C/min, and finally to 250°C at 25°C/min and held for 5 min.

Figure 1D showed the FT-IR spectrum of BSP, exhibiting the typical characteristics of polysaccharides in the range of 4,000 – 500 cm⁻¹. The strong absorption at 3,433 cm⁻¹ was derived from the O−H stretching vibrations (21). The characteristic peaks at 2,919 and 1,323 cm⁻¹ were attributed to the stretching and bending vibrations of the C−H bond, respectively (22). The band at 1,635 cm⁻¹ was attributed to the −OH flexural vibrations (23). The peaks between 1,200 and 1,000 cm⁻¹ reflected the asymmetric glycosidic band vibrations of C−O−C (24). The characteristic absorption bands at 904 and 888 cm⁻¹ suggested that both α- and β-configurations existed in BSP (25).

The methylated BSP was acid-hydrolyzed and acetylated for the GC–MS analysis. The individual peaks were identified by comparison with the mass spectrum. As shown in Table 1, the results showed the presence of four liberations of 2,3,4,6-Me₄-Manp (residue: 1-linked Manp), 2,3,6-Me₃-Manp (residue: 1,4-linked Manp), 2,3,6-Me₃-Glcp (residue: 1,4-linked Glcp), and 2,3,4-Me₃-Galp (residue: 1,6-linked Galp). The molar ratio of each monosaccharide (galactose: mannose: glucose) was consistent with the results of HPLC (3.2:5.4:1.0).

Figures 2A,B showed the ¹H NMR and ¹³C NMR spectra of BSP; in the ¹H NMR spectrum, most chemical shifts were concentrated within 3–5 ppm, presenting the typical characteristic of polysaccharides (26). In the ¹³C NMR spectrum, the chemical shifts from 60 to 78 ppm were attributed to the signals of C2–C6 carbons. Moreover, the chemical shifts of anomeric protons (δ_H 4.0–5.3 ppm) and anomeric carbons (δ_C 90.0–110.0 ppm) proved that both α- and β-configurations were present in BSP (27). The peaks accurately occurred at 101.29, 103.19, 103.53, and 103.64 ppm (anomeric carbons) and 5.29, 5.02, 4.96, and 4.53 ppm (anomeric protons), revealing that BSP contained four types of monosaccharide residues. The presence of →4)-β-D-Manp -(1 → was proven by the chemical shifts at δ 103.64, 76.95, 71.14, 81.82, 72.55, and 61.03 ppm (C-1–C-6) (19, 28). The main sugar residues of BSP also included →4)-α-D-Glcp-(1→, which was confirmed by the signals of δ_H/C 5.29/101.29, 3.59/72.55, 3.87/72.19, 3.54/78.96, 3.51/69.17, and 3.55/60.85 ppm (29). Similarly, the →6)-β-D-Galp-(1 → and β-D-Manp-(1 → were identified based on the corresponding signals. Chen et al. (11) obtained a polysaccharide named pFSP from the fibrous roots of B. striata with the M_w of 9.1 × 10⁴ Da, and pFSP consisted of three monosaccharides, including D-Glc, D-Gal, and D-Man. The main sugar residues in pFSP were →4)-α-D-Glcp-(1→, →4)-β-D-Manp-(1 → and →3,6)-β-D-Manp-(1→, moreover, →6)-β-D-Galp-(1 → and β-D-Manp-(1 → were also found in pFSP. Therefore, BSP might have a structure similar to the pFSP, but their M_ws were different.

In the present study, Congo red and I₂-KI tests were used to determine the spatial configuration of BSP. As shown in Figure 1E, the λmax of BSP mixed with the Congo red solution was the same as that of the Congo red solution alone, and the λmax of two solutions both decreased with the NaOH concentration in a dose-dependent manner. Under different NaOH concentrations, there was no red shift of the λmax in BSP mixed with the Congo red solution, suggesting that BSP had no triple-helix conformation (30).

In addition, the maximum absorption peak of BSP was at 360 nm, there was no absorption peak around 565 nm (Figure 1F), suggesting that long and branched chains were contained in the backbone of BSP (31). This was in accordance with the result of the Congo red test.

SEM was used to characterize the microscopic surface morphology of BSP (presented in Figures 3A,B), and the surface exhibited fragmented sheet-like aggregation with a rough surface. Many polysaccharides also showed this feature after lyophilization (19, 32).

AFM exhibited the planar and three-dimensional images of BSP. As shown in Figure 3C, it is clear that BSP was plate-shaped with tight structure, but the lump height (0.5–30.9 nm) was obviously higher than that of individual polysaccharide molecule (<1.0 nm), indicating that the molecules intertwined with each other and formed aggregates due to the hydroxyl groups in the polysaccharide chains.

Using BSP as the sole carbon source, the growth of L. plantarum and L. acidophilus was monitored, and the pH change of the medium was also tested. As shown in Figure 6A, similar growth curves for L. plantarum in BSP and glucose media were obtained. The bacteria numbers increased with the incubation time in two media (p < 0.05); the glucose medium had a slightly better effect on bacteria proliferation than the BSP medium. Moreover, the pH values of both culture media decreased with the incubation time (p < 0.05), indicating an increase in metabolic activity of the bacteria (41). The increased populations of L. acidophilus are shown in Figure 6B, and both glucose and BSP stimulated their proliferation. Based on the pH changes of the media and the increase in bacteria numbers, it could be seen that BSP had a similar effect on the proliferation compared to the glucose at the same concentration.

The proliferation abilities of L. plantarum and L. acidophilus in BSP media were characterized using SEM, as shown in Figures 6C,D. Both bacterial strains exhibited full and complete morphology, indicating a good reproductive state.

The intestinal microbiota produces a variety of small molecular acids through the fermentation of dietary fiber and other substrates, mainly including SCFAs and some other metabolites. SCFAs, such as acetic acid, propionic acid, and butyric acid, are not only essential for intestinal health but also have a profound impact on systemic health (42).

Table 2 shows the production of SCFAs in two media. It was found that acetic acid, propionic acid, and butyric acid were generated after a 30-h incubation; this well-explained the decrease in pH values. The concentrations of total SCFAs in glucose and BSP media were very similar, indicating the proliferation of probiotics and the conversion of BSP into SCFAs. After considering L. plantarum as an example, the concentrations of acetic acid, propionic acid, and butyric acid in the BSP medium were 122.64 ± 0.14, 88.15 ± 0.46, and 175.32 ± 0.34 μmol/L, respectively. In contrast, these concentrations were 134.72 ± 0.68, 85.66 ± 0.36, and 172.85 ± 0.18 μmol/L in the glucose medium. Lee et al. (43) studied the prebiotic activity of mucilage polysaccharide from molokhia leaves (MPF) with an M_w of 51.2 kDa. They discovered that MPF increased the contents of acetic acid, propionic acid, and butyric acid, which correlates with our findings. These results further demonstrated that BSP exhibited prebiotic activity and could serve as a carbon source to stimulate the reproduction of probiotics.