Asparagus cochinchinensis was obtained from Guangxi (China). All reagents were from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China) and of analytical grade unless otherwise indicated.

After drying at 45°C, AC roots were pulverized with a mortar in the laboratory and passed through a 40-mesh screen to generate a fine powder from which ACPs were subsequently extracted.

While certain polysaccharides can be extracted through the use of diluted acid solutions, others, particularly acidic polysaccharides or those polysaccharides that contain uronic acid, can be extracted more readily using alkaline solutions. For this study, an acid-assisted extraction procedure was employed. Briefly, 5.0 g of AC powder was refluxed twice with 85 mL of 0.1 moL/L HCl solution for 2.5 h at 85°C, after which these extraction solutions were neutralized, concentrated, and precipitated using a final 75% ethanol concentration for 12 h at 4°C. Next, 80 mL of distilled water was used to suspend 6 g of this crude polysaccharide preparation, followed by centrifugation (10 min, 10,000 xg) with subsequent separation of the supernatant using a DEAE Sepharose Fast Flow column and elution using water and solutions containing various concentrations of NaCl (0, 0.1, and 0.3 mol/L). Different eluates were concentrated, dialyzed, and lyophilized to generate the ACP-A, ACP-B, and ACP-C fractions, of which 100 mg of ACP-B was dissolved with 4 mL of 0.1 mol/L NaCl followed by centrifugation (10 min, 10,000 xg). The supernatant fraction was then separated with a Superdex™200 column and eluted using 0.1 mol/L NaCl to yield the major polysaccharide (ACP).

In total, 24 male SAMP8 mice and 6 senescence-resistant controls (SAMR1, NC) with a body weight of ~25.0 g were obtained from Beijing Weitong Lihua Laboratory Animal Co., Ltd. These animals were housed in a controlled environment (22°C, 50–70% humidity, 12 h light/dark cycle) with free food and water access. Animal experiments were performed as per the guidance of the Institute of Animal Care and User Committee (IACUC). Animals were randomized into wild-type control (SAMR1 mice, NC), control (SAMP8 mice), ACP (SAMP8 mice treated with 25, 50, or 100 mg/kg/d ACP), and positive control (SAMP8 mice treated with 1.667 mg/kg/d donepezil HCl) groups.

After a 7-week dosing period, murine cognitive function was assessed with the Morris water maze (MWM) test, using a slightly modified version of an approach reported previously (35). Briefly, mice were trained for 5 d using a circular basin (90 cm high, 100 cm in diameter) containing water at 22°C, with a hidden platform located 1 cm beneath the water surface. On each day of training, mice completed four trials, beginning each trial in a different quadrant. Mice were assessed to determine whether they reached the platform within 90 s, allowing the mouse to rest on the platform for 15 s after locating it prior to removing animals from the tank. When mice failed to reach the platform, they were manually guided to it and allowed to rest there for 15 s. After this 5-day training period, mice were placed opposite the location of the hidden platform, which was removed, and the number of platform crossings within 90 s as well as the proportion of time spent in that target quadrant were recorded with The Smart v3.0 Small Animal Behavioral Recording and Analysis System (Reward Corporation).

An E.Z.N.A.^® soil DNA Kit (Omega Bio-Tek, GA, United States) was used to extract total genomic DNA from fecal samples as directed, after which DNA quality and concentration were measured using 1.0% agarose gel electrophoresis and a NanoDrop^® ND-2000 spectrophotometer (Thermo Scientific Inc., United States), followed by storage at-80°C. The 338F (5′-ACTCCTACGGGAGGCAGCAG-3′) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′) primers targeting the V3–V4 hypervariable region of the 16S rRNA gene were used to amplify isolated DNA (36) with an ABI GeneAmp^® 9,700 PCR thermocycler (ABI, CA, United States). Each PCR reaction consisted of 4 μL 5x Fast Pfu buffer, 2 μ: 2.5 mM dNTPs, 0.8 μL of each primer (5 μM), 0.4 μL of Fast Pfu DNA polymerase, 0.2 μL of BSA, 10 ng of template DNA, and ddH₂O to 20 μL. Thermocycler settings were: 95°C for 3 min; 27 cycles of 95°C for 30 s, 55°C for 30 s, 72°C for 45 s; 72°C for 10 min, with a final resting incubation at 4°C. Sample amplification was conducted in triplicate, and PCR products were extracted following 2% agarose gel electrophoresis using an AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, CA, United States) as directed, with quantification then being performed with a Quantus™ Fluorometer (Promega, United States).

Data are derived from three or more independent experiments and are presented as means ± standard deviation (x ± s). All analyses were performed using GraphPad Prism 8.0. When comparing data with a normal distribution and homogenous variance among multiple groups, one-way ANOVAs with the Student Newman–Keuls (SNK) multiple comparison test were used. p < 0.05 was regarded as the threshold for significance.

Here, an acid-assisted extraction approach was used to isolate ACP, with a yield of 3.67% of the dry weight of the raw AC input material. Following ACP deproteinization and decolorization, samples were fractionated and purified with DEAE-52 and Sephadex G-200 chromatography columns (Figures 1A,B). Initial DEAE-52 column purification yielded three fractions, including deionized water, 0.1 mol/L NaCl, and 0.3 mol/L NaCl fractions with respective yields of 23.9, 10.24 and 8.9%. ACPs-B was subject to further Sephadex G-100 chromatography column purification, yielding a single symmetrical ACP peak with a 91.2% yield (not shown). In view of the previous product yield and the experimental effect in animals, we finally selected ACPs-2 (ACP). ACP had a calculated molecular weight of 15,580 Da, and its polydispersity index was ~1.14, with this value being close to 1 such that this ACP fraction was regarded as likely being homogenous (Table 1).

Infrared spectral analyses are commonly used to detect functional groups including O–H, C–H, and C=O. ACP samples exhibited FT-IR signals at 3330, 2935, 1,413, and 1,055 cm⁻¹ that are characteristic of polysaccharides (Figure 1C) (30). The peak at 3411 cm⁻¹ was consistent with O–H hydrogen bond vibrations, while the peak at 2933 cm⁻¹ was consistent with C–H vibrations, and the peaks at 3411 cm⁻¹ and 2,933 cm⁻¹ were indicative of the presence of a polysaccharide sample (24). Absorption bands in the 1,500–400 cm⁻¹ region were sensitive to changes, with the peaks at 1036 cm⁻¹ and 1,109 cm⁻¹ confirming that C–O–C and C–O–H stretching vibrations were present, indicative of a pyranose ring. Signals were also observed at 529 cm⁻¹ and 1,100–1,420 cm⁻¹.

ACP was found to primarily consist of glucose, galactose, L-fucose, and fructose at a 82.14:12.23:2.61:2.49 ratio, with trace levels of xylose, arabinose, and rhamnose at a 0.48:0.04:0.02 ratio (Table 2). In other reports, ACP was indicated to include fructose and glucose at a 93.3:6.7 molar ratio (37), with other studies having also described ACP preparations containing xylose, arabinose, glucose, rhamnose, mannose, galactose, glucuronic acid, and galacturonic acid (38). These differences in monosaccharide composition and molar ratios are likely related to differences in ACP sources or the methods employed for extraction and purification.

Methylation analyses can offer significant insight regarding polysaccharide structural characteristics. As it contained little uronic acid, the direct and complete methylation of all free OH groups in ACP was observed. Following the hydrolysis, reduction, and acetylation of permethylated polysaccharides to isolate PMAAs, these samples were subjected to GC–MS analysis (38). Peak areas in the GC chromatogram were compared to compute the molar percentage ratios for various sugar residue types, ultimately revealing the presence of Glcp-(1→, →2)-Glcp-(1→, →6)-Glcp-(1→, →4)-Glcp-(1→, →3, 4)-Glcp-(1→, →2,4) -Galp-(1→, →4,6)-Galp-(1→, and →3,4,6)-Galp-(1 → sugar linkages in ACP at molar percent ratios of 23.70:1.30:3.55:50.77:6.91:1.10:11.50:1.18 (Table 3). Based on the monosaccharide and methylation linkage analysis results, ACP was primarily composed of glucose (50.77%), consistent with the →4)-Glcp-(1 → residues in the backbone of the ACP structure.

An NMR approach was next used to gain further insight into the structural characteristics of ACP, including the glycosidic bond connections and configurations present therein. Virtually all protons in the ACP ¹H NMR spectra wherein the 3.00–5.30 ppm range (Figure 2A), as is normal for polysaccharides. The strong signal peak at 4.70 ppm corresponds to the D₂O solvent peak. Glycosidic bond configurations can be determined based on allocephalic hydrogenation shifts, with α-and β-configuration polysaccharide molecules generally exhibiting these shifts in the 5.0–5.5 ppm and 4.5–5.0 ppm ranges, respectively. The majority of ACP heterohydrogenation shifts fell in the 2.15–2.75 and 3.15–4.25 ppm ranges (Figure 2A). Heteropolytic proton signals at 4.50, 4.55, 4.59, 4.61, and 4.86 ppm were assigned to β-pyranose units, while signals at 5.13, 5.18, and 5.19 ppm were assigned to α-pyranose units. These data were consistent with large numbers of glycosidic bonds in the β configuration together with a relatively limited number in the α configuration in ACP.

Chemical shifts in the 1-dimensional carbon NMR spectrum were wider than those for the proton NMR spectrum, with slightly higher resolution (Figure 2B). These carbon spectral data can be employed to determine the positions and molecular conformation of groups. Polysaccharide heterocephalic carbon signals are generally observed in the δ_C 90–113 ppm range, with α-type glycosidic bonds exhibiting shifts < δ_C 102 ppm. ACP exhibited bonds in the β-configuration, with seven signal peaks in the δ_C 90–113 ppm range including signals at δ_C 92.10, 95.93, at 98.10 ppm, supporting the presence of both α-and β-type bonds within ACP. Furanose C3 and C5 signals were present in the δ 82–84 ppm range, while pyranose C3 and C5 signals were < 80 ppm, enabling differentiation between the two.

For residue A, anomeric hydrogen and carbon chemical shifts (4.61/95.93 ppm) were consistent with glucose in the β-configuration. In the COSY map, H2 (3.2 ppm) of residue A was determined based on cross peak 4.61/3.2 ppm, and H3 (3.85 ppm) of residue A was determined based on cross peak 3.2/3.85 ppm. H4 (3.45 ppm) of residue A was determined based on cross peak 3.85/3.45 ppm, and H5 (4.04 ppm) of residue A was determined based on cross peak 3.45/4.04 ppm. The H6 (3.83, 3.69 ppm) of residue A was determined from the cross peak 4.04/3.83, 3.69 ppm, which can be attributed to the chemical shift of hydrogen on the complete sugar ring. Then, the chemical shift of C on the sugar ring is attributed by HSQC signal. The chemical shift of C1 of residue A is 95.93 ppm, the chemical shift of C2 of residue A is 74.16 ppm, the chemical shift of C3 of residue A is 69.53 ppm, and the chemical shift of C4 of residue A is 75.66 ppm. The C5 chemical shift of residue A is 73.48 ppm, the C6 chemical shift of residue A is 60.75 ppm, and the chemical shift of C1 and C4 is shifted to the low field, indicating that the residue is replaced at the O-1, O-4 position of the sugar ring. Combined with the results of methylation analysis and literature reports, it is inferred that the residue A was concluded to correspond to →4)-β-D-Glcp-(1 →.

For residue B, anomeric hydrogen and carbon chemical shifts (5.19/92.11 ppm) were consistent with glucose in the α-configuration. In the COSY map, H2 (3.49 ppm) of residue B was determined based on cross peak 5.19/3.49 ppm, and H3 (3.7 ppm) of residue B was determined based on cross peak 3.49/3.7 ppm. H4 (3.95 ppm) of residue B was determined based on cross peak 3.7/3.95 ppm, and H5 (4.2 ppm) of residue B was determined based on cross peak 3.95/4.2 ppm. The H6 (3.88, 3.72 ppm) of residue B was determined from the cross peak 4.2/3.88, 3.72 ppm, which can be attributed to the chemical shift of hydrogen on the complete sugar ring. Then, the chemical shift of C on the sugar ring is attributed by HSQC signal. The chemical shift of C1 of residue B is 92.11 ppm, the chemical shift of C2 of residue B is 71.47 ppm, the chemical shift of C3 of residue B is 71.37 ppm, and the chemical shift of C4 of residue B is 69.27 ppm. The C5 chemical shift of residue B is 71.44 ppm, the C6 chemical shift of residue B is 61.24 ppm, and the chemical shift of C1 is shifted to the low field, indicating that the residue is replaced at the glycocyclic O-1 position. Combined with the results of methylation analysis and literature reports, it is inferred that the carbohydrate residue B may be α-D-Glcp-(1→).

For residue C, anomeric hydrogen and carbon chemical shifts (5.13/94.1 ppm) were consistent with galactose in the α-configuration. In the COSY map, H2 (3.89 ppm) of residue C was determined according to the cross peak 5.13/3.89 ppm, and H3 (3.6 ppm) of residue C was determined according to the cross peak 3.89/3.6 ppm. H4 (4.06 ppm) of residue C was determined according to the cross peak 3.6/4.06 ppm, H5 (3.8 ppm) of residue C was determined according to the cross peak 4.06/3.8 ppm, and H5 (3.8 ppm) of residue C was determined according to the cross peak 3.8/3.74. 3.88 ppm determines the H6 of residue C (3.74, 3.88 ppm), which can be attributed to the chemical shift of hydrogen on the complete sugar ring. Then, the chemical shift of C on the sugar ring is attributed by HSQC signal. The chemical shift of C1 on residue C is 94.1 ppm, the chemical shift of C2 on residue C is 70.4 ppm, the chemical shift of C3 on residue C is 72.76 ppm, and the chemical shift of C4 on residue C is 75.94 ppm. The C5 chemical shift of residue C is 71.52 ppm, and the C6 chemical shift of residue C is 67.56 ppm. The chemical shifts of C1, C4, and C6 are shifted to the lower field, indicating that the residue is replaced at the positions of O-1, O-4, and O-6 in the sugar ring. Combined with the results of methylation analysis and literature reports, it is inferred that the sugar residue C may be In the COSY map, H2 (3.89 ppm) of residue C was determined according to the cross peak 5.13/3.89 ppm, and H3 (3.6 ppm) of residue C was determined according to the cross peak 3.89/3.6 ppm. H4 (4.06 ppm) of residue C was determined according to the cross peak 3.6/4.06 ppm, H5 (3.8 ppm) of residue C was determined according to the cross peak 4.06/3.8 ppm, and H5 (3.8 ppm) of residue C was determined according to the cross peak 3.8/3.74. 3.88 ppm determines the H6 of residue C (3.74, 3.88 ppm), which can be attributed to the chemical shift of hydrogen on the complete sugar ring. Then, the chemical shift of C on the sugar ring is attributed by HSQC signal. The chemical shift of C1 on residue C is 94.1 ppm, the chemical shift of C2 on residue C is 70.4 ppm, the chemical shift of C3 on residue C is 72.76 ppm, and the chemical shift of C4 on residue C is 75.94 ppm. The C5 chemical shift of residue C is 71.52 ppm, and the C6 chemical shift of residue C is 67.56 ppm. The chemical shifts of C1, C4, and C6 are shifted to the lower field, indicating that the residue is replaced at the positions of O-1, O-4, and O-6 in the sugar ring. Combined with the results of methylation analysis and literature reports, it is inferred that the sugar residue C may be →4,6)-α-d-Galp -(1 → .

For residue D, anomeric hydrogen and carbon chemical shifts (4.86/93.69 ppm) were consistent with glucose in the α-configuration. In the COSY map, H2 (3.37 ppm) of residue D was determined based on cross peak 4.86/3.37 ppm, and H3 (4.07 ppm) of residue D was determined based on cross peak 3.37/4.07 ppm. H4 (3.48 ppm) of residue D was determined based on cross peak 4.07/3.48 ppm, and H5 (3.86 ppm) of residue D was determined based on cross peak 3.48/3.86 ppm. The H6 (3.98, 3.77 ppm) of residue D was determined from the cross peak 3.86/3.98, 3.77 ppm, which can be attributed to the chemical shift of hydrogen on the complete sugar ring. Then, the chemical shift of C on the sugar ring is attributed by HSQC signal. The chemical shift of C1 of residue D is 93.69 ppm, the chemical shift of C2 of residue D is 69.66 ppm, the chemical shift of C3 of residue D is 75.66 ppm, and the chemical shift of C4 of residue D is 75.94 ppm. The C5 chemical shift of residue D is 72.77 ppm, and the C6 chemical shift of residue D is 63.38 ppm. The chemical shifts of C1, C3, and C4 are shifted to the low field, indicating that the residue is replaced at the positions of O-1, O-3, and O-4 in the sugar ring. Combined with the results of methylation analysis and literature reports, it is inferred that the sugar residue D may be →3,4)-α-D-Glcp-(1 → .

Based on these results from one-dimensional NMR (¹H and ¹³C) and two-dimensional NMR (HSQC and HMBC) approaches, the fine structure of ACP was ultimately determined, as shown in Figure 2 and Table 4.

To further assess the structure of the backbone of ACP, 1D-and 2D-NMR spectra were analyzed, assigning ¹H and ¹³C NMR signals based on correlations in the HMBC and NOESY spectra and values that have been reported in the literature. The HMBC and NOESY spectra can effectively reveal glycosylic linkages between sugar residues, and they can also reveal intra-residue connections, as shown in ACP. The NOESY spectrum was used to analyze connections in ACP, as it exhibited lower signal intensity at the cross peaks of the HMBC spectrum. Some cross peaks were observed in the NOESY spectrum (Figure 2D), including peaks corresponding to H1 of sugar residue A has A cross peak 4.61/3.45 ppm with H4 of sugar residue A, the H1 of sugar residue A has a cross peak 4.61/4.36 ppm with H4 of sugar residue E, and the H1 of sugar residue A has a cross peak 4.61/4.07 ppm with H3 of sugar residue D. The H1 of sugar residue B has A cross peak 5.19/3.45 ppm with H4 of sugar residue A, the H1 of sugar residue C has a cross peak 5.13/3.45 ppm with H4 of sugar residue A, and the H1 of sugar residue D has a cross peak 4.86/3.88 ppm with H6 of sugar residue C.

Therefore, based on one-dimensional and two-dimensional NMR information and methylation analysis, it is concluded that the polysaccharide is mainly composed of →4)-β-D-Glcp-(1 → and a small amount→4,6)-α-D-Galp-(1 → and →3,4)-α-D-Glcp-(1 → and so on. Branched chain is mainly composed of α-D-Glcp-(1 → 4)-β-D-Glcp-(1 → connected to the sugar residues α-D-Glcp-(1 → 4)-β-D-Glcp-(1 → O-4 position or sugar residues of α-D-Glcp-(1 → 4)-β-D-Glcp-(1 → O-3 position.

SEM approaches are often employed to assess the surfaces and microstructural characteristics of polysaccharides, offering insight into macromolecule morphology, shape, size, and porosity (39). SEM imaging revealed that all ACP samples presented with block-like structures after acid and water treatment (Figure 3). This may be attributable to cavitation activity, turbulence shearing, and instantaneous high pressures. Acid-assisted extraction can also disrupt cellular structures, increasing the contact area between the liquid and raw material phases. This may explain the block-like surface characteristics of ACP. Indeed, extraction and purification strategies have been confirmed to influence polysaccharide shape and surface topological characteristics.

The impact of ACP administration on spatial memory in SAMP8 model mice was assessed with the MWM, monitoring the swimming paths of mice during testing (Figure 4A). Relative to the NC group, SAMP8 model mice exhibited fewer platform crossings, while tighter paths and more platform crossings were observed for mice in the ACP50 and ACP100 groups as compared to the SAMP8 group. A similar improvement was also evident in the donepezil group. Compared to NC controls, SAMP8 mice also exhibited significantly increased Target Zone (%) and Fast Time in the Target Zone (s) values together with decreased Mean Speed in Target Zone and Latency 1st Entrance to Zone (s)-Target values, consistent with impaired spatial learning and memory (Figures 4B–E). A significant increase in the Latency 1st Entrance to Zone (s)-Target was also observed in the ACP25 group (p < 0.05), while significantly increased Distance in Target Zone (%) (p < 0.001) and Fast Time in Target Zone (s) (p < 0.05) were observed for the ACP50 group.