Rapid Characterization of Chemical Components in Edible Mushroom Sparassis crispa by UPLC-Orbitrap MS Analysis and Potential Inhibitory Effects on Allergic Rhinitis

Sparassis crispa is a kind of edible fungus widely grows in the north temperate zone, which shows various medicinal properties. Due to the complexity of chemical constitutes of this species, few investigations have acquired a comprehensive configuration for the chemical profile of it. In this study, a strategy based on ultra-high performance liquid chromatography (UPLC) combined with Orbitrap mass spectrometer (MS) was established for rapidly characterizing various chemical components in S. crispa. Through the summarized MS/MS fragmentation patterns of reference compounds and systematic identification strategy, a total of 110 components attributed to six categories were identified for the first time. Moreover, allergic rhinitis (AR) is a worldwide inflammatory disease seriously affecting human health, and the development of drugs to treat AR has been a topic of interest. It has been reported that the extracts of S. crispa showed obvious inhibitory effects on degranulation of mast cell- and allergen-induced IgE and proinflammatory mediators, but the active components and specific mechanism were still not clear. Src family kinases (SFKs) participate in the initial stage of allergy occurrence, which are considered the targets of AR treatment. Herein, on the basis of that self-built chemical database, virtual screening was applied to predict the potential SFKs inhibitors in S. crispa, using known crystal structures of Hck, Lyn, Fyn, and Syk as receptors, followed by the anti-inflammatory activity evaluation for screened hits by intracellular calcium mobilization assay. As results, sparoside A was directly confirmed to have strong anti-inflammatory activity with an IC50 value of 5.06 ± 0.60 μM. This study provides a useful elucidation for the chemical composition of S. crispa, and demonstrated its potential inhibitory effects on AR, which could promote the research and development of effective agents from natural resources.


Introduction
Sparassis crispa, namely "cauliflower mushroom", is a brown-rot fungus that widely grows on the stumps of coniferous trees in the north temperate zone, as northeast of China and Japan. As a kind of valuable edible mushroom, S. crispa shows various medicinal properties, such as anti-tumor [1], anti-inflammation [2], immunoregulation [3], hypoglycemic effect [4], and ameliorating skin conditions [5]. A number of previous studies have reported the isolation and structural determination of more than thirty compounds within different categories in this species, including AR and various other allergic disease, and has already entered the clinical trial stage [21]. In the field of natural products, a large number of studies have also found that some natural compounds can regulate mast cell degranulation by inhibiting SFKs, so as to play an anti-allergic role. Lu et al. [22] reported that nujiangexanthone A isolated from garcinia could inhibit IgE/Ag-induced mast cell activation by inhibiting Syk enzyme activity, thereby inhibiting degranulation and the production of eicosanoids. In addition, some other natural compounds including atractylode lactone III [23], piceatannol [23], rosmarinic acid [24], emodin [25], and resveratrol [26] all could inhibit mast cell degranulation by inhibiting SFKs. Herein, on the basis of that self-built chemical substance database of S. crispa, virtual screening based molecular docking was firstly applied to predict the potential SFKs inhibitors in S. crispa, using known crystal structures as targeted receptors including Hck, Lyn, Fyn, and Syk, followed by the anti-inflammatory activity evaluation for screened hits by intracellular calcium mobilization assay. As results, some compounds were fished out by virtual screening as the potential SFKs inhibitors. Among them, sparoside A was directly confirmed to show strong anti-inflammatory activity for the first time, with an IC 50 value of 5.06 ± 0.60 µM. In conclusion, this study provides a useful elucidation for the chemical composition of S. crispa, and demonstrated its potential inhibitory effects on AR, which could promote the development of effective agents from natural resources.

Optimization of UPLC and MS Conditions
In order to acquire chromatograms with intense peak response and high resolution, the mobile phase compositions were firstly optimized. Compared with methanol/water, the acetonitrile/water system showed higher baseline stability and lower pressure, as well as stronger elution and isolation abilities for investigated components. When a small amount of formic acid was added into the water phase, the shapes of most peaks were improved apparently. Therefore, it was finally decided that acetonitrile/0.1% formic acid aqueous solution was used as the mobile phase. The column temperature was set at 40 • C to reduce the pressure, and flow rate was constant at 0.4 mL/min.
To acquire high sensitivity for most analytes, some parameters of heated electrospray ionization (HESI) source were also optimized by multiple experiments, including sheath gas flow, auxiliary gas flow, spray voltage, source heater temperature, capillary temperature, capillary voltage, and tube lens voltage. These parameters directly contributed little to total ion current chromatogram (TIC) but were key for MS/MS fragmentation. The optimal conditions were set as follows: sheath gas flow, 50 arb; auxiliary gas flow, 10 arb; spray voltage, 4 kV/3.5 kV (positive/negative); probe heater temperature, 350 • C; capillary temperature, 380 • C; S-lens RF level, 55. Because the molecular mass of all known compounds in S. crispa was distributed in the range of 100-1,500 Da, in full scan mode the mass spectra were acquired in the m/z range of 50-1,500 Da, and the resolution was empirically set as 70,000. Moreover, the sizes of collision-induced dissociation (CID) energy were also considered. After some attempts, the MS/MS energy was finally set as 30 V as stepped normalized collision energy (NCE), under which more abundant fragment ions with appropriate mass could be produced at the resolution of 35,000.

UPLC-Orbitrap MS Analysis of S. crispa and Component Identification
The optimized UPLC and Orbitrap MS conditions were applied for characterization of chemical components in S. crispa extracts. The TIC in positive and negative ESI modes were shown in Figure 1. The reported compounds in S. crispa could be classified into seven types on the grounds of their chemical structures: alkaloids, organic acid, sterols, sesquiterpenes, sterols, phthalides, and others. Except organic acids, most compounds showed strong response and typical fragmentation in the positive ESI mode. Thus, the targeted MS/MS experiments for citric acid (2) were conducted in negative mode, and others were in positive mode. The identification of components in S. crispa started from the recognition of part known compounds by importing the data into the Compound Discoverer 2.1 loaded with OTCML database. Then, the unidentified most peaks were processed through an established systematic strategy based on high-resolution MS [14]. First of all, the chemical elemental composition for each targeted peak was deduced by the accurate mass spectra of designated protonated/deprotonated molecular ions or adduct ions using a formula predictor, as well as their corresponding isobaric molecular ions. The proposed molecular formulas were also approved by additional judgements such as nitrogen rule, elemental composition of fragment ions and general formula features of natural compounds. Then the formulas were searched in self-built chemical database of S. crispa to match the known structures. For those formulas not included in the self-built database, they could be input into the SciFinder database for screening possible compounds, and the hits were refined in the genus of Sparassis. The next process was to verify components after learning the knowledge of characteristic product ions and fragmentation rules of various types of compounds, and the MS/MS fragmentation patterns of six reference compounds were sufficiently investigated ( Table 1). Those components owned the identical retention time, mass and fragment ions with the reference compounds were firstly identified undoubtedly. Other components could be identified via comparing the fragmentation patterns with those known analogous compounds and referring reported structures in literatures. Finally, a total of 110 compounds in S. crispa extracts were identified or tentatively identified. The retention time, m/z values of adduct ions and MS/MS fragment ions in positive/negative ESI modes, mass error, accurate molecular mass, formula, and confidence levels of identity [27] of all the identified compounds were completely summarized in Table S1. negative mode, and others were in positive mode. The identification of components in S. crispa started from the recognition of part known compounds by importing the data into the Compound Discoverer 2.1 loaded with OTCML database. Then, the unidentified most peaks were processed through an established systematic strategy based on high-resolution MS [14]. First of all, the chemical elemental composition for each targeted peak was deduced by the accurate mass spectra of designated protonated/deprotonated molecular ions or adduct ions using a formula predictor, as well as their corresponding isobaric molecular ions. The proposed molecular formulas were also approved by additional judgements such as nitrogen rule, elemental composition of fragment ions and general formula features of natural compounds. Then the formulas were searched in self-built chemical database of S. crispa to match the known structures. For those formulas not included in the self-built database, they could be input into the SciFinder database for screening possible compounds, and the hits were refined in the genus of Sparassis. The next process was to verify components after learning the knowledge of characteristic product ions and fragmentation rules of various types of compounds, and the MS/MS fragmentation patterns of six reference compounds were sufficiently investigated ( Table 1). Those components owned the identical retention time, mass and fragment ions with the reference compounds were firstly identified undoubtedly. Other components could be identified via comparing the fragmentation patterns with those known analogous compounds and referring reported structures in literatures. Finally, a total of 110 compounds in S. crispa extracts were identified or tentatively identified. The retention time, m/z values of adduct ions and MS/MS fragment ions in positive/negative ESI modes, mass error, accurate molecular mass, formula, and confidence levels of identity [27] of all the identified compounds were completely summarized in Table S1.     (100) a The bracketed bold figures shows the serial number of corresponding reference compounds. b The bracketed figures following m/z shows the relative abundance (%) of each fragment ion.
To further understand the interaction between drugs and targeted enzymes, the confirmations of small molecules bonded into the "active cavity" of proteins and their interaction patterns with amino acid residues of protein was simulated via Discovery Studio Visualizer 4.0. Taking the active compound sparoside A (Figure 2a) as an example, as shown in Figure 2b, the small molecule was "curling up" in the protein active site of Fyn, which was unambiguously presented in the graphical molecular ribbon model. It was found to form five hydrogen bonds with five residues in chain X, including with Leu17 (4.6 Å) on hydroxyl, with Ser89 (3.9 Å) on carboxyl, with Thr82 (4.0 Å), Glu83 (4.3 Å), and Met85 (4.0 Å) on glucosyl group (Figure 2c). The other intermolecular interactions were also depicted, which were included but not confined to electrostatic force and van der Waals force. It was that all these interactions contributed to the high anti-inflammatory activity of sparoside A.  (Figure 2c). The other intermolecular interactions were also depicted, which were included but not confined to electrostatic force and van der Waals force. It was that all these interactions contributed to the high anti-inflammatory activity of sparoside A.

Anti-inflammatory Activity Confirmation of Two Predicted Components
To investigate the anti-inflammatory effects of those predicted hits, Ca 2+ mobilization in Furo-2AM loaded RBL-2H3 cells was measured. As shown in Figure 3a, compared to the control vehicle, sparoside A (100 μM) significantly decreased intracellular Ca 2+ concentration, while linoleic acid slightly increased it. This result suggested that sparoside A was a positive inhibitor of SFKs, considering by combining the aforementioned docking results. But linoleic acid was probably an agonist of SFKs, although it owned strong affinity with SFKs as the inhibitors. The inhibitors and agonists couldn't be distinguished by virtual screening. In the next experiment, it was found that sparoside A dose-dependently decreased intracellular Ca 2+ mobilization, and its IC50 value was 5.06 ± 0.60 μM in six replicates (Figure 3b).

Anti-inflammatory Activity Confirmation of Two Predicted Components
To investigate the anti-inflammatory effects of those predicted hits, Ca 2+ mobilization in Furo-2AM loaded RBL-2H3 cells was measured. As shown in Figure 3a, compared to the control vehicle, sparoside A (100 µM) significantly decreased intracellular Ca 2+ concentration, while linoleic acid slightly increased it. This result suggested that sparoside A was a positive inhibitor of SFKs, considering by combining the aforementioned docking results. But linoleic acid was probably an agonist of SFKs, although it owned strong affinity with SFKs as the inhibitors. The inhibitors and agonists couldn't be distinguished by virtual screening. In the next experiment, it was found that sparoside A dose-dependently decreased intracellular Ca 2+ mobilization, and its IC 50 value was 5.06 ± 0.60 µM in six replicates (Figure 3b). Molecules 2019, 24, x FOR PEER REVIEW 9 of 17

Reagents and Materials
S. crispa was cultured by Shanxi Agricultural University, China, and its pileus was collected for the current investigation. Voucher specimens were preserved at the authors' laboratory.

Sample Preparation
After accurately weighed and grounded, 1.0 g air-dried pileus of S. crispa was extracted with 20 mL methanol in a 50 mL erlenmeyer flask by ultrasonic extraction for 30 min. After cooling down, the lost volume of methanol was complemented. Then 5.0 mg of six reference compounds were dissolved into 5 mL methanol to get six standard solutions, respectively. Finally, the above herb extracts solution and all standard solutions were filtered through a 0.22 μm membrane as the samples.

UPLC Separation
UPLC separation was carried out on a Thermo Vanquish Flex Binary RSLC platform (Thermo Fisher Scientific, Waltham, MA, USA) equipped with a diode array detector (DAD). The chromatographic column used was a Thermo Accucore aQ C18 (150 × 2.1 mm, 2.6 μm; Thermo Fisher

Reagents and Materials
S. crispa was cultured by Shanxi Agricultural University, China, and its pileus was collected for the current investigation. Voucher specimens were preserved at the authors' laboratory.

Sample Preparation
After accurately weighed and grounded, 1.0 g air-dried pileus of S. crispa was extracted with 20 mL methanol in a 50 mL erlenmeyer flask by ultrasonic extraction for 30 min. After cooling down, the lost volume of methanol was complemented. Then 5.0 mg of six reference compounds were dissolved into 5 mL methanol to get six standard solutions, respectively. Finally, the above herb extracts solution and all standard solutions were filtered through a 0.22 µm membrane as the samples.

UPLC Separation
UPLC separation was carried out on a Thermo Vanquish Flex Binary RSLC platform (Thermo Fisher Scientific, Waltham, MA, USA) equipped with a diode array detector (DAD). The chromatographic column used was a Thermo Accucore aQ C 18 (150 × 2.1 mm, 2.6 µm; Thermo Fisher Scientific, Waltham, MA, USA), which conducted in 40 • C. The mobile phase was composed of 0.1% formic acid aqueous solution (A) and acetonitrile (B), and the gradient elution program was as follows: 5%-100% B at 0-20 min; 100% B at 20-23 min. The flow rate was constant at 0.4 mL/min. The injection volume was set at 2 µL.

Orbitrap MS Analysis and Data Processing
The Q Exactive Plus mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) was coupled to the UPLC by a HESI interface. The specific parameters were set as abovementioned. The mass spectrometer calibration was conducted before each experiment. In the MS/MS experiments, data-dependent scanning was adopted to trigger the second stage fragmentation, which was to select the strongest four parent ions in each scanning point of MS 1 as targeted precursor ions for the further fragmentation. The dynamic exclusion function was utilized to prevent the repetitive ion scans and save the analysis time. The software Xcalibur 4.1 (Thermo Fisher Scientific, Waltham, MA, USA) and Compound Discoverer 2.1 (Thermo Fisher Scientific, Waltham, MA, USA) loaded with OTCML database 1.0 (Thermo Fisher Scientific, Waltham, MA, USA) were employed to process the UPLC-MS data. To ensure the reliability of the identification results, those peaks with intensity over 10 5 in TIC were selected for identification. The formulas of all parent and fragment ions in selected peaks were generated according to their accurate mass using a formula predictor. The maximal mass accuracy error was confined to ±3 ppm. Considering the possible elemental compositions of existed compounds in S. crispa, the number of four types of atoms were limited as follows: C ≤ 50, H ≤ 100, O ≤ 20, and N ≤ 10.

Virtual Screening for Potential SFKs Inhibitors
To predict the SFKs inhibitors in S. crispa, molecular modelling and virtual screening based on docking were performed using Surflex-Dock GeomX (SFXC) program [102] interfaced with SYBYL-X 2.1.1 (Tripos, USA) on Dell Precision T5500 workstation. SFXC is a fast and automated docking program that considers ligand conformational flexibility by an incremental fragment placing technique. It was used to dock the small molecules into the active site of the protein and fished out the best-fit compounds. The 3D coordinates of the active site of four kinds of SFKs were taken from the reported X-ray crystal structure of the protein catalytic core in complex with a cocrystallized natural ligand from RCSB Protein Databank (https://www.rcsb.org). The corresponding PDB codes of Hck, Lyn, Fyn and Syk were 5H0B [103], 5XY1 [104], 2DQ7 [105], and 6HM7 [106], respectively. They were picked as the queries to screen the self-built 3D chemical database including 110 identified compounds derived from S. crispa. As the positive control, the natural ligands were also docked via the identical procedure. The "protomol" for docking was defined as all amino acids within 6.5 Å proximity of the natural ligands, and the other parameters were set as default. Finally, the Total scores were calculated to denote the matching degree between the conformers of each compound and the four enzymes, along with other reference fit values including Crash, Polar, Similarity, D score, PMF score, G score, Chem score, and C score; namely, a higher Total score indicated a better match. The docking results were visualized and analyzed with Discovery Studio Visualizer 4.0 (Accelrys, USA).

Anti-inflammatory Activity Evaluation by Intracellular Calcium Mobilization Assay
Extracellular Ca 2+ influx was an essential process in activating mast cells and other related immune cells to induce allergic reactions [107]. Thus, the intracellular calcium mobilization assay was employed to evaluate the anti-inflammatory activity of those hits in virtual screening. Due to the unavailability of pure substances, only linoleic acid and sparoside A were applied in this research. The two kind of drugs were dissolved in DMSO, and 0.25% DMSO were used as the control vehicle. The final concentration of DMSO in each well did not exceed 0.25% for all of the tested drugs. Rat basophilic leukemia (RBL)-2H3 cells [108] were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA) and cultured at 37 • C in DMEM supplemented with 10% FBS and antibiotics (100 U/mL penicillin, 100 µg/mL streptomycin) in a 5% CO 2 incubator. The intracellular Ca 2+ level ([Ca 2+ ] i ) was measured using Fura-2AM loading by monitoring the fluorescence intensity. First, cells were pretreated with drug solutions or vehicle for 1 h at 37 • C, and thereafter, washed with Ringer's Solution (155 mM NaCl, 4.5 mM KCl, 2 mM MgCl 2 , 10 mM dextrose, 5 mM HEPES, pH 7.4), supplemented with 1 mM CaCl 2 . Then they were loaded with 1 µM Fura-2AM at a concentration of 10 7 cells/mL for 1 h in the dark. At last, the cells were washed, resuspended in Ca 2+ supplemented Ringer's Solution, and the [Ca 2+ ] i of 5 × 10 5 cells was monitored on Quanta master Spectrofluorometer (Photon Technology International, Birmingham, NJ, USA) at room temperature. Ca 2+ mobilization was expressed as the ratio (Relative Fluorescence, RF) of Fura-2AM fluorescence at 510 nM caused by the two excitation wavelengths (340 nm/380 nm). The IC 50 value was calculated by RF. Each datum represents the mean ± standard deviation in six replicates.

Data Process and Analysis
All data acquired were processed using one-way analysis of variance (ANOVA), followed by Student's t-test to find the differences between group means in GraphPad Prism v7.0 (GraphPad Software, La Jolla, CA, USA). The level of significance was set at less than 5% (p < 0.05).

Conclusions
In conclusion, an UPLC coupled with Orbitrap MS method was firstly developed and applied for rapid separation and characterization of chemical components in S. crispa. Based on reference compounds, optimized UPLC and MS conditions, and systematic fragment ions-based identification strategy, a total of 110 compounds of interest were detected and identified or tentatively identified. The MS/MS fragmentation patterns of all the characterized compounds in positive/negative ion modes were also explored. Furthermore, it was found that in UPLC eluted with gradient acetonitrile and 0.1% formic acid aqueous solution, the retention time of different types of chemical constituents in S. crispa was roughly as the following order: organic acids < phthalides < sterols/sesquiterpenes, and the occurrence of alkaloids covered the entire separation time. To our knowledge, this is the first study to systematically establish the chemical composition profile of S. crispa by UPLC-MS, and the method presented here has been demonstrated as an effective pathway for the analysis of the components in a complex sample from natural resource. In addition, in order to explore the inhibitory effects of S. crispa on AR, virtual screening was conducted to predict the inhibitors of SFKs based on the self-built chemical substance database. Two available compounds sparoside A and linoleic acid were applied in intracellular calcium mobilization assay to evaluate their anti-inflammatory activity. Finally, it was confirmed directly for the first time that sparoside A showed obvious anti-inflammatory activity, which could dose-dependently decrease intracellular Ca 2+ mobilization with IC 50 value 5.06 ± 0.60 µM. This result was an important improvement over the previous data from Bang et al. [12]. As to whether sparoside A could treat AR, further pharmacological studies are needed in the future. This study could provide essential reference for the medicinal and edible research and development on this kind of mushroom.
Supplementary Materials: The following can be found at online. Table S1: All the identified components from Sparassis crispa extracts and their UPLC-MS/MS data; Table S2: The virtual screening results for self-built 3D chemical database including 110 compounds in S. crispa with four kinds of SFKs (Hck, Lyn, Fyn and Syk) as the receptors; Figure S1: MS spectra and proposed fragment ions of riboflavin (1) in positive ion mode; Figure S2: MS spectra and proposed fragment ions of citric acid (2) in negative ion mode; Figure S3: MS spectra and proposed fragment ions of ainsliatone A (3) in positive ion mode; Figure S4: MS spectra and proposed fragment ions of ergosterol (4) in positive ion mode; Figure S5: MS spectra and proposed fragment ions of fraxinellone (5) in positive ion mode; Figure S6: MS spectra and proposed fragment ions of mannitol (6) in positive ion mode; Figure S7: Comparison of total ion current chromatograms (TIC) of S. crispa extracted with different solvents in positive and negative ion modes; Figure S8: Chemical structures and available raw MS 2 spectra of some components identified from S. crispa.  Acknowledgments: Assistance of Yuxin Zhang in Beijing University of Chinese Medicine for the activity evaluation was also acknowledged.

Conflicts of Interest:
The authors declare no conflict of interest.