Anaphylactic Rare Saponins Separated from Panax notoginseng Saponin and a Proteomic Approach to Their Anaphylactic Mechanism

In recent years, many traditional Chinese medicine injections based on Panax notoginseng saponin (PNS) have been reported to cause anaphylaxis. Previous studies on the anaphylactic saponins of PNS and their mechanism are inadequate. In this study, potential anaphylactic saponins were obtained by the separation of PNS and preparation of each individual component through comprehensive techniques, such as liquid chromatography, preparative chromatography, HPLC, NMR, and MS. The anaphylactic abilities of these saponins were tested using RBL-2H3 cells via a β-hexosaminidase release rate test. The results for the mechanism of anaphylaxis were obtained by a proteomic analysis using RBL-2H3 cells. The results indicate that, among all the saponins prepared, gypenoside LXXV and notoginsenoside T5 showed strong anaphylactic abilities and notoginsenoside ST-4 and ginsenoside Rk3 showed weak anaphylactic abilities. These 4 saponins can induce anaphylaxis via direct stimulation of effector cells. The gene oncology enrichment analysis results showed that, among these saponins, only gypenoside LXXV was related to organelles of the endoplasmic reticulum and Golgi apparatus and biological processes in response to organic cyclic compounds. Four proteins in RBL-2H3 cells with the accession numbers A0A0G2JWQ0, D3ZL85, D4A5G8, and Q8K3F0 were identified as crucial proteins in the anaphylactic process. This research will help traditional Chinese medicine injection manufacturers strengthen their quality control and ensure the safety of anaphylactic saponins.

ability, such as ginsenoside F2 [13] and Compound K [13]. e mechanism underlying the anaphylactoid reactions is distinct from that of IgE-mediated allergies, although they both have the same symptoms.
e complement system plays a key role in the process of anaphylaxis induced by many agents (e.g., radiocontrast media, liposomes, and TCM injections) [7]. Xu et al. [5,10] studied the anaphylactoid reaction induced by Xuesaitong injection and discovered that >10 kDa molecules could activate classical complement pathways through direct stimulation to cause anaphylaxis; Tween-80 can activate the complement system through classical and alternative pathways; and tannic acid can induce anaphylactoid reactions through coactivation of the complement system, the kallikrein-kinin system, and coagulation. Another mechanism is direct stimulation, which occurs through either direct G protein activation or opioid receptors [7]. For example, some TCM injections can also directly induce β-hexosaminidase and histamine release through mast cell degranulation [7].
Panax notoginseng saponin (PNS), a common material for TCM injections administered to cure cardiovascular diseases, is produced from Panax notoginseng (Burk.) Chen [5]. ADR reports indicated that the main TCM injections based on PNS as the crude material could induce more than 12% of the total TCM injection ADRs; moreover, anaphylactoid reactions mainly occurred within 30 min after the first administration [3,4]. However, previous studies on PNS, especially the crude PNS extract used in TCM injections, are insufficient.
To determine the specific anaphylactic constituents in PNS (from the crude extract of the root of Panax notoginseng (Burk.) F.H. Chen) and elucidate the mechanism, we prepared potential anaphylactic saponins via liquid chromatography, HPLC, and preparative chromatography and identified these saponins via ESI-MS and NMR. We further analyzed the cytotoxicity and β-hexosaminidase release rate (β-HexRR) of these saponins in vitro using RBL-2H3 cells. A proteomic analysis was performed to identify the possible mechanism underlying the anaphylactoid reactions.
is study will help manufacturers of TCM injections based on PNS ensure that those anaphylactic saponins are within safe levels and control the quality of TCM injections to avoid clinical anaphylactoid reactions.
All reagents were analytically or chromatographically pure.

Cell
Culture. RBL-2H3 cells were maintained in RPMI 1640 media with 10% FBS and 1% penicillin-streptomycin solution (100x) at 37°C and 5% CO 2 . Single-cell suspensions were plated at a density of 5-8 × 10 5 cells/mm 2 on a culture dish (D � 100 mm). When the cells reached over 80% confluence, they were harvested and passaged at a ratio of 1 : 3.

Anaphylactic Properties of the Subfractions.
Five milliliters of each subfraction was concentrated and freezedried in a Heidolph Rotary Evaporator (Heizbad Hei-VAP, Heidolph Instruments GmbH & CO. KG, Germany) and Biocool Freeze Dryer (Biocool FD-1C-50, Boyikang, Beijing), respectively. e obtained powders were weighed and added to the RPMI 1640 culture media with 1% DMSO, and the final concentration of each solution was 200 μg/mL.
Exponentially growing RBL-2H3 cells in 6-well dishes were washed with PBS 3 times. Two milliliters of the tested supernatant was added to the corresponding wells, including the negative control group (RPMI 1640 culture media with 1% DMSO) and the positive control group (C 48/80; 20 μg/mL). Each group was made in triplicate. e 6well dishes were incubated in an incubator (37°C, 5% CO 2 ) for 30 min before the supernatants were discarded. e 2 Evidence-Based Complementary and Alternative Medicine dishes were placed under an inverted microscope (Nikon Eclipse TS100-F, Japan) for observation. e groups with high degranulation rates were considered to have the potential constituents of the anaphylactoid reaction.

Preparation and
Identification of the Potential Saponins of the Anaphylactoid Reaction. Isocratic elution and preparative separation of Fr. 3, Fr. 5, and Fr. 7-8 were conducted using preparative chromatography with mobile phases of 68% methanol/water, 74% methanol/water, and 80% methanol/water, respectively. e eluents were subsequently concentrated and freeze-dried into powders before analysis by ESI-MS and NMR for molecular weight and 13 C-NMR information.

IC 50 Test.
Each saponin was dissolved in RPMI 1640 media into solution or dispersed suspension. Exponentially growing RBL-2H3 cells in a 96-well dish were washed with PBS before they were incubated with the corresponding saponins. e control group only contained the culture medium and RBL-2H3 cells, while the blank group only contained the culture medium. Cells were incubated for 24 h, and the supernatants were discarded. After washing with PBS 3 times, 100 μL culture medium was added to each well, and then MTS reagents were added for an incubation of approximately 30 min. e dish was subsequently placed on a microplate reader to record the absorbance.

β-Hexosaminidase Release Rate.
Exponentially growing RBL-2H3 cells in 24-well dishes were washed with RPMI 1640 media 3 times. en, 100 μL of RPMI 1640 medium was added to each well before incubation for 10 min. Another 100 μL of each tested constituent at different concentrations was added to the corresponding wells. Incubations of 15 min, 30 min, 1 h, and 2 h were conducted before an ice bath for 10 min to end the reaction. Fifty microliters of the supernatant in each well together with the substrate (β-D-hexosamine; 4 mM) were added to a 96-well dish before incubation for 1 h at 37°C. en, 150 μL of stop solution (glycine buffer; pH 10.7; 200 mM) was added to each well to stop the reaction. e absorbance (absorbance of supernatant, AoS) at 405 nm was measured for all wells. e rest of the supernatant in the 24-well dishes was removed, and 200 μL of 0.5% Triton X-100 was added to each well to lyse the cells. e lysate of each well was centrifuged at 3000 rpm for 1 min, and 50 μL of the supernatant in each well was transferred to a 96-well dish. β-D-Hexosamine (4 mM) as the substrate was added to each well at the same time. e 96-well dish was incubated at 37°C for 1 h before 150 μL of stop solution was added to each well. e absorbance (absorbance of lysate, AoL) at 405 nm was measured for all wells. β-HexRR can be calculated using the following equation: (1)

Proteomic Analysis. RBL-2H3 cells of different groups
were collected into the corresponding microtubes and centrifuged at 4000 rpm for 4 min before the supernatant was discarded. e cells were washed with PBS and centrifuged once again, and the supernatant was discarded. RIPA lysis buffer and protease inhibitor were added before treatment in an ice bath for 30 min. e cells were then lysed via ultrasonication. Protein concentrations were measured using the bicinchoninic acid (BCA) method.
Urea (8 M) and DTT (final concentration 10 mM) were added to the microtubes containing the cell lysate at volume ratios of 1 : 1 and 1 : 2, respectively. e mixture was incubated at 37°C for 4 h before IAA was added (IAA : DTT � 5 : 1, mol/mol). e mixture was further incubated in the dark for 1 h at room temperature. NH 4 HCO 3 (50 mM) was added to the mixture to dilute urea to below 1 M. Sequencing grade modified trypsin was added (trypsin: protein � 1 : 50, w/w). e mixture was incubated in a water bath at 37°C for 10-16 h. e mixture was desalted before the peptide concentration was measured via a NanoDrop ( ermo Fisher, Waltham, MA, USA). LC-MS/MS analysis was performed on a Q-Exactive HF Hybrid Quadrupole-Orbitrap mass spectrometer ( ermo Fisher Scientific) coupled online to a nanoflow LC system (EASY-nLC 1000, ermo Fisher Scientific). Peptides were delivered onto a 2 cm self-packed trap column (100 μm inner diameter, 3 μm resin, ReproSil-Pur C18-AQ, Dr Maisch GmbH) in solvent A (0.1% formic acid FA in HPLC grade water). After loading and washing, the peptides were transferred to a 12 cm column (150 μm inner diameter, 1.9 μm resin, ReproSil-Pur C18-AQ, Dr Maisch GmbH) and separated over 78 min nonlinear gradients from 6% to 95% solvent B (0.1% FA in ACN) at a flow rate of 600 nL/min. e exact elution gradients were as follows: 0-16 min, 6%-10% B; 16-51 min, 10%-24% B; 51-71 min, 24%-34% B; 71-72 min, 34%-95% B; 72-78 min, 95% B. e peptides were ionized using a 2.0 kV spray voltage and a capillary temperature of 320°C. e data acquisition was performed in the OT-IT mode. Full MS scans (300 to 1,400 m/z) were performed at a resolution of 120,000, a maximum injection time of 80 ms, and an AGC target value of 3e6. Tandem mass spectra were generated for up to 20 precursors by HCD with a normalized collision energy of 27%. e dynamic exclusion was set to 12 s. e MS2 spectra were read out in the Orbitrap at a resolution of 15,000 with an AGC target value of 5e4 and a maximum injection time of 19 ms.
e parameters and settings were as follows: precursor mass tolerance, 15 ppm; fragment mass tolerance, 20 mmu; enzyme name, trypsin; max missed cleavage sites, 2; static modification, carbamidomethyl; dynamic modifications, acetyl and oxidation. e confidence level was set to 95%.

Evidence-Based Complementary and Alternative Medicine
Differentially expressed proteins were determined by the ratio of protein expression of the adjacent groups. e expression of upregulated proteins in the high dose group/ medium dose group, medium dose group/low dose group, and low dose group/control group was all above 1.2; downregulated proteins in the high dose group/medium dose group, medium dose group/low dose group, and low dose group/control group were all below 1/1.2. Cluster heatmaps were drawn based on the upregulated proteins and downregulated proteins. Gene Oncology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were based on the upregulated proteins, downregulated proteins, and the DAVID (Database for Annotation, Visualization and Integrated Discovery) tool.

Anaphylactic Properties of the Subfractions.
Photographs of RBL-2H3 cells incubated with each subfraction are presented in Figure 1. e cells in the positive control groups were strongly degranulated and showed blurred cell membranes and obvious granules. In contrast, few cells in the negative control groups showed degranulation. Fr. 3, Fr. 5, and Fr. 7-8 were shown to induce degranulation of RBL-2H3 cells (>50% degranulated cells) while Fr. 1-2, Fr. 4, Fr. 6, Fr. 9-10, and Fr. 11 showed no anaphylactic properties.

Structural Identification of the Potential Saponins of the
Anaphylactoid Reaction. Compounds 1, 2, and 3 were obtained from the eluents of Fr. 3. Compounds 2, 3, and 4 were obtained from the eluents of Fr. 5. Compounds 5, 6, and 7 were obtained from the eluents of Fr. 7-8. All compounds were white amorphous powders that showed a red color on thin-layer plates when reacting with 10% ethanol sulfate solution. e 13 C-NMR spectra are shown in Figure S1, and the data are listed in Supplementary Materials. e data for compounds 1 and 4 were in accordance with the data in [14], and they were identified as notoginsenoside T5 (noto-T5) and notoginsenoside T4 (noto-T3), respectively. e data for compounds 2 and 3 were in accordance with the data in [15]; therefore, they were identified as ginsenoside Rk3 (Rk3) and ginsenoside Rh4 (Rh4), respectively. e data for compounds 5, 6, and 7 were in accordance with the data in [16][17][18], respectively; therefore, they were identified as ginsenoside Rg3 (Rg3), gypenoside LXXV (gyp-LXXV), and notoginsenoside ST-4 (noto-ST-4), respectively. Notably, all of the above saponins are rare saponins in PNS.

Basic Physical and Chemical Information of Each Saponin.
Saponins on a microgram scale were weighed and mixed with 1 mL of RPMI 1640 media with 1% DMSO at room temperature. e chemical structure, chemical formula, molecular weight, and solubility of each saponin are shown in Table 1.
Noto-T3 was not involved in the following studies because it presented flocculent precipitate and aggregation in the solution and was hard to be solved or shattered into small pieces, which was not suitable for the follow-up experiments.

IC 50 Test.
To prepare an appropriate concentration gradient of each saponin, an IC 50 test was performed based on RBL-2H3 cells using the MTS assay. e average cell viability of each group is shown in Table S1.
e IC 50 value of each saponin was obtained by linear interpolation. According to the IC 50 value of each saponin, the concentrations of the low, medium, and high dose groups were set geometrically, following the rules that the dose of high dose groups was close to but lower than the IC 50 value ( Table 2). According to the experiments, C 48/80 could induce abnormal absorbance in the IC 50 test. erefore, the low, medium, and high dose groups of C 48/80 were set based on [12].

β-Hexosaminidase Release Rate.
e β-hexosaminidase release rate is the gold standard to quantitatively evaluate the level of anaphylactoid reactions [7,10,19]. According to clinical reports, anaphylactoid reactions induced by TCM injections containing PNS mainly occur within 30 min after administration, and few ADR cases still occur within 30 min to 2 h after administration [20]. erefore, 4 periods of time (15 min, 30 min, 1 h, and 2 h) were set to investigate the β-HexRR of the control group, C 48/80 group, and 6 saponin groups (Figures 2(a)-2(d)).
e results indicated that, compared with the control and C 48/80 groups, Rg3 had no anaphylactic ability, gyp-LXXV and noto-T5 had strong anaphylactic abilities, and noto-ST-4 and Rk3 had weak anaphylactic abilities (Figure 2(e)). For Rh4, the results cannot support the conclusion that it possesses anaphylactic ability. Surprisingly, its high dose group with an incubation time of 1 h even induced a significantly lower β-HexRR.
ese results also demonstrated that gyp-LXXV, noto-T5, noto-ST-4, and Rk3 can directly activate β-hexosaminidase release and degranulation from RBL-2H3 cells, which suggested that these saponins can induce anaphylaxis via direct stimulation of effector cells.

Proteomic Analysis.
To investigate the mechanism of the anaphylactoid reaction induced by the potential anaphylactic saponins in PNS, differentially expressed proteins of each saponin group with an incubation time of 30 min were studied. According to the filter conditions, proteins were screened and classified into upregulated proteins and downregulated proteins (Table 3). Based on the upregulated and downregulated proteins, cluster heatmaps were drawn ( Figure 3); GO and KEGG enrichment analyses were also carried out via the DAVID tool. GO enrichment analyses of the biological process (BP) (p < 0.005), cellular component (CC) (p < 0.005), and molecular function (MF) (p < 0.005) categories are displayed in Figure 4. e KEGG enrichment pathway analysis results are shown in Figure S2.   Evidence-Based Complementary and Alternative Medicine     Table 2. Each group was repeated 3 times, and the data are expressed as the mean ± SEM (n � 3). * p < 0.05 and * * p < 0.01, compared with the corresponding control group (t-test). e β-HexRR of the high dose group of Rh4 with an incubation time of 1 h was significantly lower than that of the control group, while all the other significant differences were significantly higher. Table 3: Number of upregulated and downregulated proteins and total identified proteins of each saponin group with an incubation time of 30 min.

Discussion
e β-HexRR results were in accordance with a previous study showing that Rg3 could inhibit mast cell-mediated allergies by blocking degranulation [21].
Further analysis demonstrated that there were 4 proteins in common in the differentially expressed proteins of the strong anaphylactic groups (gyp-LXXV groups and noto-T5 groups), namely, A0A0G2JWQ0, D3ZL85, D4A5G8, and Q8K3F0, all of which were downregulated proteins and not included in the differentially expressed proteins of the other 4 groups (Figure 5(b)). is result indicated that these 4 proteins had a higher possibility of playing a key role in the anaphylactoid reactions of RBL-2H3 cells.
For gyp-LXXV, KEGG enrichment pathway analysis results were not obtained after data filtration. For the noto-T5 and other 4 groups, the KEGG enrichment pathway analysis result was not consistent with RBL-2H3 cells and anaphylactoid reactions ( Figure S1). e results indicate that the molecular mechanism of the anaphylactoid reactions induced by gyp-LXXV and noto-T5 requires further in-depth study.
Although gyp-LXXV and noto-T5 were identified as strong anaphylactic saponins and noto-ST-4 and Rk3 were identified as weak anaphylactic saponins, these findings do not necessarily mean that every TCM injection based on PNS will induce anaphylaxis. In fact, incomplete clinical reports indicated that anaphylactoid reactions induced by several main TCM injections based on PNS had an overall incidence rate of less than 1% and varied according to the batch [20,22], which also indicates that anaphylactic saponins are most likely to be rare saponins and that the occurrence of anaphylaxis depends on the total content of these rare saponins. Other related factors include the dripping speed, age, sex, and body constitution [3,4]. Here we only discuss the potential anaphylactic abilities of these saponins and offer a guide for TCM injection manufacturers.

Conclusions
In conclusion, among the 7 saponins identified, gyp-LXXV and noto-T5 had strong anaphylactic abilities, noto-ST-4 and Rk3 had weak anaphylactic abilities, and Rg3 had no anaphylactic ability. Anaphylactic saponins can induce anaphylaxis via direct stimulation of effector cells. e results of proteomic studies indicated that the anaphylactoid reaction of gyp-LXXV was related to the biological processes in the response to organic cyclic compounds and the cellular components of the endoplasmic reticulum and Golgi apparatus. Four proteins were most likely to play key roles in the anaphylactoid reactions of RBL-2H3 cells, and their accession numbers were A0A0G2JWQ0, D3ZL85, D4A5G8, and Q8K3F0. To avoid clinical anaphylactoid reactions, manufacturers should strictly control the ingredients of TCM injections based on PNS and ensure that gyp-LXXV, noto-T5, noto-ST-4, and Rk3 are within safe levels.

Data Availability
All data used to support the findings of this study are available from the corresponding author upon request.