Ginseng powder (400 g) was boiled in deionized water (4 L) for 3 h. After filtration, the solid phase was boiled in deionized water at the same ratio (g/ml) two more times. The filtrate was combined, centrifuged, and concentrated. The concentrated solution was mixed with anhydrous ethanol at a volume ratio of 1:3 and left overnight at 4°C. Discard the supernatant by centrifugation and the precipitate was successively washed with 75% ethanol, 95% ethanol, anhydrous ethanol, and ether. The crude ginseng polysaccharides were obtained after drying and subsequently redissolved in deionized water. After removal of proteins using a Sevage reagent, a 3 KD molecular sieve was used to remove small molecules to obtain WGP. The content of sugar was tracked and monitored by the phenol sulfuric acid method (Wang et al., 2020).

WGP (5 mg/ml) was filtered via 0.45 μm microfiltration membrane. The filtered sample (20 μL) was loaded into a TSK-Gel G4000PWXL column (Tosoh, Shanghai branch, China) controlled by LC-10Avp system (Shimadzu, Shanghai branch, China). High performance gel permeation chromatography (HPGPC) was performed using 0.2 M NaCl as mobile phase at flow rate of 0.5 ml/min. T-series Dextran standards were used for reference standards.

Monosaccharide composition analysis was performed as previously described (Wang et al., 2020). Briefly, WGP (200 g) was hydrolysed in anhydrous methanol solution (1 ml) containing hydrochloric acid in nitrogen. Then, the sample was dried and hydrolysed in 2 M trifluoroacetate acid. After dried, the sample was dissolved using 0.3 M sodium hydroxide and added an equal volume of 0.5 M PMP (1-phenyl-3-methyl-5-pyrazolone) with thoroughly blending using pipettor. Placed the mixture (0.2 ml) for 30 min at 70°C, added 0.1 ml hydrochloric acid and 0.7 ml dichloromethane for extraction. The aqueous phase was filtered via 0.22 μm organic membrane. Conversion of monosaccharides with PMP were detected via high-performance liquid chromatography (HPLC). The sample (10 μL) was injected into a 4.6 mm × 250 mm COSMOSIL 5C18-PAQ column (Nacalai Tesque, Shanghai branch, China) controlled by an LC-20AT system (Shimadzu, Shanghai branch, China). HPLC was performed using a mobile phase composed of 19.5% Acetonitrile and 80.5% 0.1 M PBS (pH 7.0) at flow rate of 1 ml/min. The absorbance values at wavelength 245 nm were compared with those of monosaccharide standards including arabinose, fucose, galactose, galacturonic acid, glucose, glucuronic acid, mannose, rhamnose, and xylose to determine the monosaccharide composition of WGP.

Fourier transform-infrared spectroscopy (FT-IR) spectra of WGP was acquired via Tenor 27 spectrophotometer (Shimadzu, Shanghai branch, China). WGP was ground with KBr powder at a mass ratio of 3:1 and compressed into a pellet. The FT-IR spectra were recorded in range of 400–4,000 cm⁻¹.

The lyophilized WGP was dissolved in D₂O. The ¹H and ¹³C nuclear magnetic resonance (NMR) spectrum were performed on AV-500, 600, and 800 instruments (Bruker, Germany) using tetramethylsilane as the internal standard.

Human hepatic cell line L-O2 was purchased from the American Type Culture Collection (Manassas, VA, United States). Cells were cultured using DMEM media containing 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM L-glutamine (Thermo Fisher Scientific, Shanghai branch, China) and 10% (V/V) fetal bovine serum (Abwbio, Guangzhou, China) and incubated in HERAcell150i incubator (Thermo Fisher Scientific) set to 5%CO2/95% air and 37°C. Cells were tested for mycoplasma monthly by the PCR method described by Uphoff and Drexler (Uphoff and Drexler, 2005).

The L-O2 cells were treated with a series of equal ratio gradient concentrations of WGP for 72 h in a 96-well plate. MTT [3-(4, 5-dimethyl-thiazol-2yl)-2, 5-diphenyl tetrazolium bromide] (Sigma-Aldrich, Shanghai branch, China) was added at the final concentration of 0.5 mg/ml and incubated for 4 h in incubator. Then, the cells were lysed in 10% SDS containing 10 mM HCl overnight. The absorbance values at wavelength 590 nm were measured via a microplate reader.

Whole cell lysates were obtained via ultrasonic cell disruption and subjected to SDS-polyacrylamide gel electrophoresis. Proteins were electrophoretically transferred onto PVDF (polyvinylidene difluoride) membrane (Thermo Fisher Scientific), and subsequently immunoblotted utilizing anti-β-actin (Proteintech, Rosemont, IL, United States) and -C4 (Abcam, Cambridge, MA, United States) antibodies. Visualization of immunoreactive proteins was conducted using the Odyssey Infrared Imaging System (LiCor, Lincoln, NE, United States). Densitometry measurements were performed using Odyssey V3.0 software (LiCor).

Total RNA was extracted by TRIZOL method and used to make cDNAs via reverse transcription PCR Kit (Thermo Fisher Scientific), as described previously (Edwards et al., 2009). The LightCycler 480 real-time PCR meter (Roche, Indianapolis, IN, United States) and TaqMan probe Hs00246758_m1 (Thermo Fisher Scientific) were used to quantify C4 transcripts. Comparative Ct method was used to calculate the fold changes (Livak and Schmittgen, 2001). C4 transcripts were normalized to GAPDH transcripts measured by TaqMan probe (Hs02786624_g1).

The C4 promoter region was PCR amplified from human genomic DNA via full length-forward and reverse primers (Table 1). Then, the PCR product was cloned into pGEM-T-Easy vector (Promega, Madison, WI, United States). A single clone of C4 promoter was identified and digested by XhoI and HindIII (Promega). The C4 promoter was then subcloned into reporter gene vector pGL4.19 basic (Promega) at the XhoI and HindIII restriction sites to generate pC4-1,007/+44. To generate the 5’-deletion constructs, pC4-119/+44, pC4-102/+44, pC4-92/+44, pC4-72/+44, and pC4-48/+44, genomic DNA fragments were PCR amplified from the pGL4.19 plasmid using forward (segment 1 to segment 5) and reverse primers (Table 1). Then, the PCR product was digested by HindIII and XhoI, and subcloned into pGL4.19 basic.

Statistical analyses were performed using GraphPad Prism (GraphPad Software, San Diego, CA, United States). Differences were compared via the non-pair-wised two-sample t-test.

WGP were obtained via hot-water extraction and alcohol precipitation from ginseng powder. The yield of WGP was 0.85%. Carbohydrate content of WGP was over 98% and protein impurity was less than 0.1%. HPGPC was used for molecular weight analysis of WGP. We obtained two overlapping peaks and a single symmetric peak indicating that the molecular weight range of the WGP was from 1 kD to 79.4 kD (Figure 1A). The peak at 1 kD is due to incomplete interception of molecular sieve. Then, monosaccharide compositions analysis of WGP hydrolysate was performed using HPLC and nine monosaccharide reference standards. Monosaccharide profile was obtained by comparing retention times of nine standards under the same analytical conditions (Figures 1B,C). Chromatographic results demonstrated that WGP was composed of galacturonic acid, galactose, glucose, arabinose, rhamnose, glucuronic acid, and mannose in molar proportions of 28.9: 24.4: 23.0: 13.9: 6.7: 2.6: 0.6, respectively. Xylose and fucose were not detected. Consistent with published studies, glucose, galacturonic acid, and galactose are the most common monosaccharide compositions of WGP (Shi et al., 2017; Shin et al., 2017; Chen et al., 2018; Song et al., 2018; Zhao et al., 2019; Park et al., 2020). However, the specific proportion of these monosaccharides has varied widely across these studies, which may be due to the different origins of ginseng.

The FT-IR spectrum of WGP showed characteristic hydroxyl and C-H stretching vibration peak at 3,416.51 and 2,929.93 cm⁻¹, respectively. The absorption peaks at 1,625.99 and 1735.68 cm⁻¹ were caused by bound water and C=O stretching vibration of uronic acid, respectively. The absorption peaks at 1,414.96 and 1,241.87 cm⁻¹ represented C-H angular vibrations of carbohydrate. In addition, 1,153.18, 1,080.52, and 1,023.91 cm⁻¹ were assigned to C-O-H and C-O-C stretching vibration peaks of pyran, indicated that WGP was connected by α-pyranoside bond. The absorption peaks at 936.61 and 851.02 cm⁻¹ were the characteristic absorption peaks of α-Glcp (Figure 2A).

WGP ¹H NMR spectrum exhibited seven anomeric proton signals at δ5.32, 5.31, 5.14, 5.13, 5.06, 5.02, and 5.00 ppm, suggesting that the analyte was made up of seven monosaccharides. Intense signals within δ3.10–4.20 ppm represented CH-O and CH₂-O groups of carbohydrate. The chemical shift from δ3.15–4.13 ppm was contributed by H-2 to H-6 protons. Meanwhile, no signal was observed at δ5.50 ppm indicating that WGP contained glucopyranose, consistent with the FT-IR result (Figure 2B). The structure of WGP was further analyzed by ¹³C NMR spectroscopy. β-1,4-Linked Gal residues exhibited six signals at δ105.15, 70.70, 73.05, 75.44, 74.05, and 61.58 ppm, corresponding to their C-1 to C-6. Signals at δ105.13 and 82.84 ppm were attributed to C-1 and C-3 of β-1,3-Gal, respectively. Furthermore, δ101.42 ppm was the heterocephalic carbon position of α-Glcp. Peak at δ78.04 ppm indicated that C-4 of α-Glcp had been replaced. δ73.07, 74.81, 72.54, and 62.83 ppm were the positions of C-2, C-3, C-5, and C-6 of α-1,4-Glcp, respectively. The anomeric signals at δ108.58 and 83.72 ppm were due to C-1 and C-3 carbons of α-1,3,5-Ara (Figure 2C). These results were consistent with the monosaccharide composition analysis and FT-IR spectrum of WGP.

C4 is mostly synthesized in the liver (Galibert et al., 1997). Therefore, we chose human normal hepatocyte L-O2 cells as model. First, we investigated the cytotoxic effect of WGP on L-O2 cells by treating the L-O2 cells with WGP for 72 h. The results obtained from MTT assays showed that WGP treatment had minimal effect on viable cells, with the inhibition rate of viable cells less than 14% (Figure 3). Then, we determined the effect of WGP on the protein levels of C4 by western blotting. As shown in Figure 4, WGP increased C4 protein levels as early as 24 h and lasted for 72 h in a dose dependent way. These results show that WGP increase C4 production in hepatocytes.

To determine if WGP enhance C4 production through transcriptional mechanisms, we treated L-O2 cells with WGP for 24, 48, and 72 h, and then measured C4 mRNA levels by real-time PCR. Treatment of L-O2 cells with WGP for 48 and 72 h significantly increased levels of C4 mRNA, indicating WGP enhances C4 gene transcription (Figure 5A). To determine if WGP truly enhances C4 gene transcription, pGL4.19 reporter plasmid consisting of −1,007 to +44 region of the C4 promoter (designated FL) was used to confirm the effect of WGP on C4 transcription. Transient transfection of pC4-1,007/+44 into L-O2 cells was performed first, then the cells were treated with WGP for 72 h. The p-C4-1,007/+44 construct showed significant reporter activity compared to pGL4.19 basic in the absence or presence of WGP. Interestingly, treatment of the cells with WGP for 72 h significantly enhanced the C4 promoter reporter gene activity (Figure 5B). Taken together, these results demonstrate that WGP increase the production of C4 by promoting C4 gene transcription.

C4 promoter contains three E-boxes (positions −137 to −132, −98 to −93, and −78 to −73; designated E-box1, E-box2, and E-box3), one NF1 binding site (positions −110 to −97), and one Sp1 (positions −57 to −49) binding element (Figure 6A). To determine which Cis-acting element(s) is (are) vital to the enhancing effect of WGP, a series of 5’ deletion constructs (designated S1-S5; Figure 6A) were generated by using the −1,007 to +144 region of the C4 promoter as the template. As shown in Figures 6B,C, deletion from −1,007 to −119 significantly decreased C4 promoter activity and significantly decreased the enhancing effect of WGP on C4 promoter activity compared to the FL construct, indicating the E-box1 element in this region plays an important role in mediating the enhancing effect of WGP on C4 promoter. In contrast, further deletions from −119 to −102, −102 to −92, and −92 to −72 did not significantly affect the enhancing effect of WGP on C4 promoter. Interestingly, deletion from −72 to −48 completely abolished the enhancing effect of WGP, indicating the Sp1 element in this region also plays an important role in mediating the WGP effect on C4 promoter. Taken together, results from our deletion analyses suggest that the E-box1 and Sp1 cis-elements play vital roles in mediating the enhancing effect of WGP on C4 gene transcription.