Comprehensive chemical analysis of triterpenoids and polysaccharides in the medicinal mushroom Antrodia cinnamomea

Abstract

Antrodia cinnamomea is a precious medicinal mushroom used in adjuvant treatments of cancer. Triterpenoids (25R/S-ergostane epimers and Δ^7,9/Δ⁸-lanostanes) and polysaccharides are its major bioactive constituents. Quality control of this mushroom is difficult, due to a lack of effective analytical methods. In this study, a series of methods were established to analyze A. cinnamomea. The ergostane- and lanostane-type triterpenoids were extracted by 50% and 100% methanol, respectively, and were then analyzed by ultra-high performance liquid chromatography coupled with diode-array detection and quadrupole time-of-flight mass spectrometry (UHPLC/DAD/qTOF-MS). In total 49 compounds, including 34 ergostanes and 10 lanostanes, were characterized by comparing with 31 reference standards or by analyzing the UV and MS spectral data. Furthermore, the contents of 18 major triterpenoids, including 10 ergostanes and 8 lanostanes, were determined by UPLC/UV or supercritical fluid chromatography coupled with mass spectrometry (SFC/MS, for 25R/S-antcin A) within 16 min. This is the first time that the contents of pure optical forms for 5 pairs of 25R- and 25S-ergostane epimers (antcins A, B, C, H, and K) were determined. On the other hand, the polysaccharides in A. cinnamomea were analyzed by ion chromatography coupled with pulsed amperometric detection (IC/PAD) after acid hydrolysis, and the contents of 7 monosaccharides were determined. Finally, the established methods were applied to the analysis of 15 batches of A. cinnamomea. Samples derived from different cultivation techniques could be distinguished according to the contents of triterpenoids. The UPLC/UV, SFC/MS and IC/PAD methods established in this work provided powerful tools to evaluate the quality of A. cinnamomea.

---

1. Introduction

Antrodia cinnamomea (also known as Antrodia camphorata, polyporaceae family) is a rare and precious medicinal mushroom distributed in Taiwan. It is mainly used to treat cancer, inflammation, and intoxication by local residents.¹ According to a survey from the Taiwan Cancer Foundation, 12% of cancer patients in Taiwan use A. cinnamomea as an adjuvant therapeutic agent or nutrient during cancer treatment.² Clinical trials for A. cinnamomea have been initiated recently.³ However, few reports are available on quality control of A. cinnamomea, due to a lack of efficient chemical analytical methods.

A. cinnamomea contains a complicated array of tetracyclic triterpenoids, including ergostanes and lanostanes. They are considered as the major bioactive constituents of A. cinnamomea.¹ Thus far, at least 40 ergostanes and 13 lanostanes have been isolated from this herbal medicine, and they have been reported to possess anti-cancer, anti-inflammatory and hepatoprotective activities.^4–8 These triterpenoids have very similar structures, only differing in the number, location or stereochemistry of hydroxyl groups, which renders their separation difficult. As a result, the separation of a crude extract usually needs 60–100 min.^9–11 Moreover, the ergostanes in A. cinnamomea occur as C-25 epimeric pairs, and the resolution of epimers are even more difficult.¹² In most previous reports, the epimers are considered as one single compound due to poor chromatographic separation, and due to lack of pure reference standards. Recently, we have isolated the pure optical forms of eight pairs of ergostane epimers from A. cinnamomea, and fully established the stereochemistry of C-25.⁴ We have also developed an analytical SFC method to separate ergostane epimers, which was proved to be rapid and efficient for low-polarity epimers like 25R/S-antcin A.¹³ On the other hand, the lanostanes in A. cinnamomea include Δ^7,9(11) and Δ⁸ types. The Δ⁸ lanostanes have never been included in previous analytical studies due to their poor UV absorption. Therefore, new simple and efficient methods for qualitative and quantitative analyses of triterpenoids in A. cinnamomea are of significance.

Polysaccharides are another important class of bioactive chemicals in A. cinnamomea. They have been reported to exhibit anti-cancer, anti-inflammatory, and anti-hepatitis B virus activities.^1,14,15 Although a few polysaccharides have been purified,¹ little is known on the chemical composition of polysaccharides in A. cinnamomea. Several groups have analyzed the hydrolysis products of polysaccharides by gas chromatography.¹⁶ This method suffered from tedious chemical derivatization of the samples. Ion chromatography coupled with pulsed amperometric detection (IC/PAD) is a rapid, effective, and sensitive technique for the analysis of saccharides.¹⁷ More importantly, the samples can be directly analyzed without derivatization.

In the present study, we determined the contents of 18 major triterpenoids in A. cinnamomea by UPLC/UV (16 min analysis time) and SFC/MS (5 min), and characterized 31 minor compounds by comparing to reference standards or by UHPLC/DAD/qTOF-MS analysis. Particularly, the contents of pure optical forms for 5 pairs of 25R- and 25S-ergostane epimers, together with Δ⁸ lanostanes that have poor UV absorption, were determined for the first time. Moreover, polysaccharide composition of A. cinnamomea was profiled in acid-hydrolyzed water extracts by IC/PAD analysis. These rapid and efficient methods were applied to the analysis of 15 batches of A. cinnamomea samples.

2. Experimental

2.1. Chemicals and reagents

Acetonitrile, methanol (J. T. Baker, NJ, USA) and formic acid (Sigma-Aldrich, MO, USA) were of HPLC grade. De-ionized water was prepared by a Direct-Q3 ultrapure water system (Millipore, MA, USA). A total of 31 reference compounds (Fig. 1 and 1S†) were isolated from Antrodia cinnamomea by the authors.⁴ They include ergostanes (S)-antcin K (E1), (R)-antcin K (E2), camphoratin A (E3), antcamphin E (E7), antcamphin F (E8), antcamphin K (E10), antcamphin L (E11), antcin F (E13), (S)-antcin C (E16), (R)-antcin C (E17), antcamphin I (E20), antcamphin J (E21), (R)-antcin H (E22), (S)-antcin H (E23), (R)-antcin I (E27), (S)-antcin I (E28), (S)-antcin B (E29), (R)-antcin B (E30), (S)-antcin A (E33), (R)-antcin A (E34); lanostanes 3β,15α-dihydroxylanosta-7,9(11),24-triene-21-oic acid (L1), dehydrosulphurenic acid (L3), sulphurenic acid (L4), 15α-acetyldehydrosulphurenic acid (L5), dehydroeburicoic acid (L9), and eburicoic acid (L10); other compounds antcamphin A (O1), antcamphin B (O2), antcamphin C (O3), antcamphin D (O4), and 4-acetyl-antroquinonol B (O5). For all the ergostanes except for E3 and E13, their 25R/S configurations had been identified by modified Mosher's method.⁴ Reference sugar standards of D-galactose, D-glucose, L-fructose, L-arabinose, D-xylose, D-mannose, L-fucose, D-galacturonic acid and D-glucuronic acid were from Sigma-Aldrich with purities of above 97%. A total of 15 batches (S1–S15) of A. cinnamomea samples were analyzed in this study. The wild mushroom (natural growth) is now very rare, and most commercial products are derived from artificial cultivation.¹⁸ The fruiting bodies and mycelia samples of Antrodia cinnamomea YMT1002 (GenBank: KJ778613 and KJ704843) were cultivated in 2012–2014 and identified by Professor Yew-Min Tzeng at Chaoyang University of Technology, Taiwan. The dish culture samples were provided in June 2014 by Honest & Humble Biotechnology Co., Ltd, New Taipei City, Taiwan. The solid support culture and submerged fermentation of A. cinnamomea, as well as their products (lyophilized powder or capsules), were purchased from different stores in mainland China or Taiwan in 2013–2014. Voucher specimens were deposited at the authors' laboratory.

Fig. 1 Chemical structures of reference standards from A. cinnamomea.

2.2. Preparation of ergostanes- and lanostanes-enriched extract

For qualitative analysis, 100 mg of fruiting bodies (batch S10) was ultrasonicated in 5 mL of 50% methanol for 30 min and then filtered. The supernatant was used as the ergostanes-enriched extract, and was diluted by 5 folds before analysis. This step was repeated once (extract with 50% methanol for 30 minutes), and the supernatant was discarded. The fungal residue was then extracted with 5 mL of methanol to obtain the lanostanes-enriched extract, which was diluted by 2 folds before UHPLC/DAD/qTOF-MS analysis.

2.3. Sample preparation for quantitative analysis and method validation

2.3.1. Stock solutions. Reference standards E1, E2, E16, E17, E22, E23, E29, E30, E33, E34, L1, L3, L5 and L9 were dissolved separately in methanol to prepare individual stock solutions (1000 μg mL⁻¹ for E16, E17, E33, E34 and L9; 2000 μg mL⁻¹ for the other compounds). These stock solutions were mixed, dried under a gentle nitrogen flow at 35 °C, and reconstituted in methanol to obtain a mixed calibration stock solution (ranged from 160–480 μg mL⁻¹ for different compounds). L10 was dissolved in methanol to prepare a 400 μg mL⁻¹ single stock solution.

2.3.2. Calibration standard solutions. The mixed stock solution was then serially diluted to obtain calibration standard solutions (diluted by 2, 4, 8, 12, 18, 32, 48, 96, and 160 folds, respectively). The single stock solution of L10 was diluted by 1.33, 2, 4, 8, 20, 40, 80 and 100 folds. After being filtered through 0.22 μm nylon membranes, a 1 μL aliquot was injected for analysis.

2.3.3. Methanol extracts of crude drug samples. All crude drug materials were dried in the air and stored at 4 °C until use. To prepare A. cinnamomea extracts, the dried fruiting bodies or mycelia were pulverized into fine powder, and an aliquot of 100 mg was mixed with 10.0 mL of methanol. The mixture was weighed, ultrasonicated for 30 min (150 W, 40 kHz, 30–40 °C), and added with methanol to replenish the solvent loss. The supernatants were filtered through 0.22 μm membranes before UPLC/UV or SFC/MS analysis.

2.3.4. Quality control samples. QC samples were prepared at three concentration levels. Known amounts of the standards were spiked into a methanol extract of batch S10 (as matrix) to obtain high QC (HQC), middle QC (MQC), and low QC (LQC) samples. All the solutions were sealed and stored at 4 °C until use, and were kept at 15 °C during analysis.

2.4. Sample preparation for saccharide analysis

2.4.1. Water extract of crude drug samples. An aliquot of 1.0 g crude drug powder was refluxed in 50 mL water for 2 h, and then filtered to obtain the water extract.

2.4.2. Acid hydrolysis. Polysaccharides in the water extract were hydrolyzed into monosaccharides for analysis. The water extract (3 mL) was mixed with 4 mol L⁻¹ trifluoroacetic acid (to 3 mL water extract, add 1.3 mL of TFA), and heated in a water bath at 90 °C for 3 h without stirring. The mixture was evaporated to dryness, and then reconstituted in 6 mL of H₂O. The solution (1 mL) was filtered through a 0.45 μm membrane, and then applied to an OASIS HLB SPE column (6 cc, pre-eluted with 6 mL of MeOH and 6 mL of H₂O, successively). The column was eluted with 6 mL of H₂O, and the eluent was collected as the polysaccharide hydrolysis product. The obtained samples were diluted by different folds (9 folds for S2 and S7, 3 folds for the other batches) with H₂O before IC/PAD analysis.

2.4.3. Stock solution of monosaccharides. Reference monosaccharide standards were separately dissolved in water to prepare individual stock solutions (1000 μg mL⁻¹). These solutions were mixed to obtain a mixed stock solution (containing 40 μg mL⁻¹ each of L-arabinose, D-galactose, D-glucose, D-xylose, D-mannose, L-fructose, and L-fucose), which was then diluted by 2, 4, 8, 20, 40, 80, 200, and 400 folds to establish the calibration curves.

2.5. UHPLC/DAD/qTOF-MS analysis

For qualitative analysis, the high-accuracy mass spectral data were obtained on an Agilent 6530 qTOF mass spectrometer equipped with an Agilent series 1290 UHPLC instrument via an ESI interface (Agilent Technologies, Waldbronn, Germany). The UHPLC instrument was equipped with a binary pump, a diode-array detector, an autosampler, and a column compartment. Samples were separated on an Agilent Zorbax Eclipse Plus C₁₈ column (1.8 μm, 2.1 × 150 mm). The mobile phase consisted of acetonitrile (A) and water containing 0.1% formic acid (B). A linear gradient elution program was used as follows: 0 min, 30% A; 3 min, 53% A; 5 min, 53% A; 12 min, 90% A; 15 min, 95% A; 18 min, 95% A. The column temperature was 50 °C. The flow rate was 0.3 mL min⁻¹. The DAD detector scanned from 190 to 400 nm. A 0.2 μL aliquot of the methanol extract was injected for analysis. The ESI source was operated in the negative ion mode. High-purity nitrogen (N₂) was used as sheath gas (11 L min⁻¹) and nebulizing gas (50 psig). Other parameters were as follows: capillary voltage, 3500 V; nozzle voltage, 500 V; fragmentor voltage, 170 V (negative)/135 V (positive); skimmer voltage, 65 V; octopole 1 rf voltage, 750 V; data acquisition, 3 spectra/s. Collision-induced dissociation (CID) was used to obtain the MS/MS spectra. Data were analyzed by MassHunter software (Agilent).

2.6. UPLC/UV quantitative analysis

The ACQUITY H-Class UPLC system consisted of a quaternary solvent manager, an autosampler with a cooling system, a column heater, and a DAD detector (Waters, Milford, MA, USA). The organic mobile phase (A) was acetonitrile containing 1% of methanol and 0.2% of formic acid. The aqueous phase (B) was water containing 0.2% of formic acid. An Acquity UPLC HSS T3 column (1.8 μm, 2.1 × 150 mm) equipped with a VanGuard pre-column (1.8 μm, 2.1 × 5 mm) (Waters, MA, USA) was used for sample separation. A linear gradient elution program was used as follows: 0–2.5 min, 45% A; 3.0 min, 55% A; 9.0 min, 55% A; 11.5 min, 60% A; 12.0–15.0 min, 95% A. The flow rate was 0.4 mL min⁻¹, and the column temperature was 45 °C. The DAD detector scanned from 190 to 400 nm (extracted at 198 nm, 250 nm, and 270 nm). The sample tray temperature was maintained at 15 °C. A 1 μL aliquot of sample was injected for analysis. Data were collected and processed by Empower™ 3 software (Build 3471, Waters).

2.7. SFC/MS quantitative analysis

An Agilent 1260 Infinity analytical SFC (supercritical fluid chromatography) system was coupled to an Agilent 6120 quadrupole mass spectrometer to determine the contents of 25R- and 25S-antcin A. The SFC modules were the same as our recent report.¹³ The SFC eluent was introduced into the ESI interface with the aid of an isopump module (pumping methanol at 0.35 mL min⁻¹ into the ion source). The flow rate was 2 mL min⁻¹, the back pressure was maintained at 180 bar by the back pressure regulation system, and the column temperature was 37 °C. Samples were separated on a Chiralcel OJ-H column (5 μm, 4.6 × 250 mm, Daicel Industries, Tokyo, Japan) using a programmed elution: 0–5 min, 12%; 5.1–9.0 min, 20% MeOH in CO₂, v/v. To reduce the effect of sample solvent, we increased the sample concentration by extracting 200 mg material with 10 mL of methanol following section 2.3, and injected 5 μL for analysis. The eluent entered MS between 3.80 and 4.70 min for detection. The mass spectrometer was operated in (−)-ESI-SIM mode (m/z 453.3) to determine E33 and E34. The parameters were as follows: drying gas (N₂), 9.0 L min⁻¹; nebulizer pressure, 35 psig; drying gas temperature, 300 °C; capillary voltage, 3000 V; fragmentor voltage, 80 V. Data were processed by Openlab CDS Chemstation software (Agilent).

2.8. IC/PAD analysis

IC/PAD analysis (ion chromatography coupled with pulsed amperometric detection) was carried out on an ICS3000 ion chromatography instrument (Thermo-Dionex Inc., USA) equipped with a DP-5 gradient pump, and an AS-1 autosampler. Samples were separated on a Dionex CarboPac PA20 column (3.0 × 150 mm) protected with a CarboPac PA20 guard column (3.0 × 30 mm). An ED-3000 electrochemical detector with an Au working electrode and a pH-Ag/AgCl reference electrode was used. The waveform was as follows: E1 = 0.10 V, 0–0.4 s; E2 = −2.00 V, 0.41–0.42 s; E3 = 0.60 V, 0.43 s; and E4 = −0.10 V, 0.44–0.50 s (E1 was the detection potential, and E2–E4 were potentials to clean and restore the electrode for subsequent detection). The linear gradient elution program was as follows: 0–17 min, 2 mM NaOH; 17.1–24 min, 15 mM NaOH + 150 mM NaOAc; 24.1–26 min, 200 mM NaOH; 26.1–34 min, 2 mM NaOH. The flow rate was 0.45 mL min⁻¹, and the temperature was 30 °C. The injection volume was 5 μL. Data were processed by Chromeleon 2.2 software (Thermo-Dionex).

3. Results and discussion

3.1. Separation of ergostane- and lanostane-type triterpenoids by solvent extraction

Antrodia cinnamomea contains lanostane and ergostane type triterpenoids. These two types of compounds are usually co-eluted on RP-HPLC, which renders their separation and identification difficult. To facilitate their analysis, we established a stepwise extraction procedure to separate ergostanes and lanostanes, based on our previous study.¹³ Most ergostanes were readily soluble in aqueous methanol, whereas the lanostanes could only be dissolved in pure methanol due to their low polarity. Therefore, we used 50% methanol (50 folds) to extract the ergostanes, and then extracted again with 100% methanol to obtain the lanostanes-enriched extract. The two types of triterpenoids were basically separated by this simple method (Fig. 2). Furthermore, 7-ΔO ergostanes (∼274 nm), 7-OH/-H ergostanes (∼254 nm), Δ^7,9(11) lanostanes (∼242 nm), and Δ⁸ lanostanes (∼198 nm) could be differentiated according to their maximal UV absorptions. These diagnostic characteristics were used for structural characterization of unknown triterpenoids in A. cinnamomea.

Fig. 2 Maximum UV absorption for ergostanes (A) and lanostanes (B), and their preliminary separation from Antrodia cinnamomea using a stepwise extraction (C).

3.2. Characterization of ergostanes and lanostanes by UHPLC/DAD/qTOF-MS

Antrodia cinnamomea contains characteristic 4α-methylergostanes and 24-methylene-lanostanes. Although Yang et al. had reported the fragmentations of Ganoderma lanostanes, little is known on the fragmentations of these special types of Antrodia triterpenoids.^19,20 In this work, we studied the (−)-ESI CID fragmentations of 26 pure lanostanes and ergostanes on a qTOF mass spectrometer, and tentatively characterized 18 unknown minor triterpenoids from A. cinnamomea (Fig. 3). We also identified 5 other compounds, including norergostanes, ergostane aldehydes, and cyclohexenones, in (+)-ESI mode. In total, 49 compounds were characterized from A. cinnamomea by UHPLC/DAD/qTOF-MS, including 34 ergostanes (E1–E34), 10 lanostanes (L1–L10), and 5 other compounds (O1–O5). Among them, 31 compounds were further confirmed by comparing with reference standards.

Fig. 3 Extracted ion chromatograms of 25R/S-ergostanes (A) and Δ^7,9/Δ⁸-lanostanes (B). Chromatograms A and B were obtained from the ergostanes-enriched extract and lanostanes-enriched extract, respectively. Epimeric pairs and analogous pairs were marked with blue and red arrows, respectively.

The ergostanes and lanostanes yielded intensive [M − H]⁻ ions in (−)-ESI mode. The lanostanes carry fewer hydroxyl groups than ergostanes, and are more difficult to dissociate in MS/MS analysis upon collision-induced dissociation. Therefore, a CID collision energy 40 eV was used for ergostanes, and 55 eV was used for lanostanes. Collision induced dissociation of the [M − H]⁻ ions included cleavages at the side chain and the tetracyclic backbone.^19–21 For ergostanes, the 25R- and 25S-epimers showed identical fragments. Meanwhile, we found that subtle changes in the triterpenoid structures could lead to remarkable differences in tandem mass spectrometry fragmentations. For instance, isomers E16/E27 (m/z 469) and E7/E22 (m/z 485) showed very different MS/MS spectra. Therefore, it is important to study authentic standards with different substitution patterns, and to summarize their fragmentation pathways.

All ergostanes investigated in this study have a COOH group at C-25. They could produce [M − 44]⁻ ions from [M − H]⁻ in their MS/MS spectra. The loss of 43.9895 Da for E34 was assigned to CO₂ (calcd 43.9898) cleavage of the side chain (Fig. 4). MS/MS spectra of E34, E29 and E27 were dominated by [M − 60]⁻ ions. The loss of 60.0212 Da was assigned to C₂H₄O₂ (calcd 60.0211). This fragment was also derived from the side chain, as shown in Fig. 4.^19,22 The abundant [M − 60]⁻ ion could be a diagnostic signal for 7-ΔO/7-H ergostanes.

Fig. 4 (−)-ESI-MS/MS spectra of representative 7-H, 7-ΔO and 7-OH ergostanes (E10, E16, E27, E29, and E34).

Compounds E16 and E10 are 7-OH ergostanes. They produced abundant m/z 247, 259 and 301 fragments. As depicted in Fig. 4, following the neutral loss of CO₂, 7-OH of E16 could trigger B-ring RDA fragmentation to produce m/z 301.2166 (C₂₀H₂₉O₂, Δ = −2.2 ppm) and m/z 123.0812 (C₈H₁₁O, Δ = 3.0 ppm). It could also lead to B/C-ring cleavage to yield m/z 259.2062 (C₁₈H₂₇O, Δ = 1.9 ppm) and m/z 247.2071 (C₁₇H₂₇O, Δ = −1.6 ppm).

Fragments m/z 247, 259 and 301 could be considered as diagnostic ions for 7-OH ergostanes. They were also observed in two other 7-OH compounds E7 and E1. The product ion at m/z 83 should be derived from RDA cleavage of A-ring. Ergostanes E21 and E22 were substituted by –OH at C-12 position. Their MS/MS spectra were dominated by fragments [M − H − 44 − 18]⁻, [M − H − 44 − 28]⁻, and [M − H − 44 − 58]⁻. For instance, E21 ([M − H]⁻ m/z 483) produced three major product ions at m/z 421, 411, and 381. The loss of 61.9988, 71.9837, and 101.9947 was assigned to CO₂ + H₂O, CO₂ + CO, and CO₂ + CO + CH₂O (calcd 62.0004, 71.9847 and 101.9953), respectively. The fragmentation pathway was proposed in Fig. 5. Yang et al. have also reported that 12-OH could trigger the neutral loss of CH₂O in lanostane-type triterpenoids.¹⁹ Their report was consistent with our finding. Compound E22 is a 3-OH analogue of E21. It also showed abundant [M − CO₂ − H₂O]⁻, [M − CO₂ − CO]⁻ and [M − CO₂ − CO − CH₂O]⁻ fragments.

Fig. 5 (−)-ESI-MS/MS spectra of representative 12-OH, 4-OH, and Δ¹⁴⁽¹⁵⁾ ergostanes (E1, E7, E13, E21, and E22).

Compounds E1 and E7 are 4-OH ergostanes. They could readily produce [M − 62]⁻ ions in their MS/MS spectra, which was due to the loss of CO₂ and H₂O. For example, ergostane E7 could produce m/z 423.2905 (C₂₈H₃₉O₃, Δ = −0.07 ppm). Similarly, E1 could eliminate CO₂ and H₂O to produce m/z 425, or CO₂ + 2H₂O to produce m/z 407. A fragment from B-ring cleavage was also observed for E7 (m/z 137).

Antcin F (E13) contains a Δ¹⁴ double bond. Its MS/MS spectrum was different from other ergostanes. It produced m/z 233 due to CO₂ elimination and C-ring cleavage. Upon a successive A-ring cleavage, m/z 149 and 83 could also be observed.

Another 14 unknown ergostanes were detected from the ergostanes-enriched extract. Their structures were tentatively characterized by analyzing their spectral data (Table 1). Six ergostanes were detected in the extracted ion chromatogram (EIC) of m/z 485. Among them, E7/E8 and E22/E23 were identified by comparing with reference standards. E6 and E12 showed identical MS and MS/MS spectra. The diagnostic fragments at m/z 149 and 233 were the same as those of E13, indicating there was a Δ¹⁴ double bond in the molecule, and that the structures for A/B/C ring were the same as that of E13. The molecular weights of E6 and E12 were 18 Da (H₂O) greater than E13. Thus, E6 and E12 were tentatively characterized as hydrated antcin F. The hydration may possibly take place at the Δ²⁴⁽²⁸⁾ double bond. E14 showed identical mass spectra to that of E13, and was identified as the C-25 epimer of E13. E31 and E32 showed [M − H]⁻ ions at m/z 511.3057 and m/z 511.3061, respectively, establishing their formula as C₃₁H₄₄O₆. Their UV spectra showed a maximum absorption at 250 nm, which was consistent with 7-H or 7-OH ergostanes (Fig. 2). The MS/MS spectra showed a major fragment at m/z 59.0141 (C₂H₃O₂, Δ = −4.12 ppm), indicating the presence of an acetoxyl (OAc) group. The above information was consistent with the previously reported antcin G.²³ The fragments at m/z 391 and 407 were assigned to [M − H − HOAc − CO₂]⁻ and [M − H − HOAc − C₂H₄O₂]⁻, respectively. Although we had found that free 7-OH could trigger the A/B ring cleavage, acetylation of 7-OH resulted in a MS/MS fragmentation pattern similar to that for 7-H ergostanes. Given their similar HPLC retention times, E31 and E32 were tentatively characterized as 25R/S-epimers of antcin G. Their fragmentation pathways were proposed (Fig. 2S†).

Table 1 Characterization of triterpenoids in Antrodia cinnamomea by UHPLC/DAD/qTOF-MS^a

No.^a		tR (min)	λmax (nm)	Formula	HRMS^b (m/z)		Δm (ppm)	MS/MS (m/z)	Identification
					Calcd	Meas
a *Compounds identified by comparing with reference standards; w, UV signals were weak.b Measured and predicted [M − H]^− for E1–E34, L1–L10, [M + H]^+ for O1–O4, and [M + Na]^+ for O5.
*	E1	4.083	254	C29H44O6	487.3065	487.3072	−1.41	147, 247, 301	Antcin K, S
*	E2	4.175	254	C29H44O6	487.3065	487.3070	−1.00	147, 247, 301	Antcin K, R
*	E3	4.559	w	C29H44O6	487.3065	487.3074	−1.82	245, 391, 407, 425	Camphoratin A
	E4	4.767	259	C29H42O7	501.2858	501.2867	−1.84	149, 245, 395, 423, 439	12-OH antcamphin E/F
	E5	5.235	259	C29H42O7	501.2858	501.2859	−0.24	149, 245, 395, 423, 439	12-OH antcamphin E/F
	E6	4.834	w	C29H42O6	485.2909	485.2904	0.95	149, 233	Hydrated antcin F (Δ^14)
*	E7	5.360	w	C29H42O6	485.2909	485.2901	1.57	137, 247, 301, 423	Antcamphin E, S
*	E8	5.627	w	C29H42O6	485.2909	485.2899	1.98	137, 247, 301, 423	Antcamphin F, R
	E9	5.460	w	C29H40O7	499.2701	499.2703	−0.35	107, 245, 301	New ergostane (tetra-O-substituted)
*	E10	5.652	257	C29H44O5	471.3116	471.3109	1.48	83, 121, 247, 259, 301	Antcamphin K, S
*	E11	5.952	257	C29H44O5	471.3116	471.3111	1.05	83, 121, 247, 259, 301	Antcamphin L, R
	E12	5.860	261	C29H42O6	485.2909	485.2903	1.16	149, 233	Hydrated antcin F (Δ^14)
*	E13	6.936	300	C29H40O5	467.2803	467.2798	1.06	149, 233	Antcin F
	E14	7.245	w	C29H40O5	467.2803	467.2810	−1.50	149, 233	Antcin F epimer
	E15	7.111	w	C29H40O7	499.2701	499.2704	−0.55	83, 123, 207, 249	New ergostane (tetra-O-substituted)
*	E16	7.177	255	C29H42O5	469.2959	469.2954	1.17	83, 123, 247, 259, 301	Antcin C, S
*	E17	7.437	255	C29H42O5	469.2959	469.2958	0.31	83, 123, 247, 259, 301	Antcin C, R
	E18	7.236	w	C29H40O6	483.2752	483.2744	1.68	83, 123, 245, 379, 395, 439	New ergostane (3-ΔO, 7-OH, 12-OH, Δ)
	E19	7.320	w	C29H40O6	483.2752	483.2745	1.47	83, 123, 245, 379, 395, 439	New ergostane (3-ΔO 7-OH, 12-OH, Δ)
*	E20	7.604	w	C29H40O6	483.2752	483.2746	1.26	381, 411, 421, 439	Antcamphin I, R
*	E21	7.704	w	C29H40O6	483.2752	483.2748	0.85	381, 411, 421, 439	Antcamphin J, S
*	E22	7.670	275	C29H42O6	485.2909	485.2904	0.95	383, 413, 423, 441	Antcin H, R
*	E23	7.837	275	C29H42O6	485.2909	485.2903	1.16	383, 413, 423, 441	Antcin H, S
	E24	8.071	w	C29H40O6	483.2752	483.2746	1.26	83, 379, 395, 421, 423, 439	Antcin D
	E25	8.638	w	C29H40O7	499.2701	499.2708	−1.34	97, 207, 411	New ergostane (tetra-O-substituted)
	E26	8.805	w	C29H40O7	499.2701	499.2706	−0.95	97, 207, 411	New ergostane (tetra-O-substituted)
*	E27	9.138	272	C29H42O5	469.2959	469.2959	0.10	409, 425	Antcin I, R
*	E28	9.213	272	C29H42O5	469.2959	469.2966	−1.39	409, 425	Antcin I, S
*	E29	9.397	270	C29H40O5	467.2803	467.2802	0.21	407, 423	Antcin B, S
*	E30	9.489	270	C29H40O5	467.2803	467.2800	0.64	407, 423	Antcin B, R
	E31	9.772	w	C31H44O6	511.3065	511.3057	1.59	59, 391, 407	Antcin G (25R/S epimer)
	E32	9.856	w	C31H44O6	511.3065	511.3061	0.81	59, 391, 407	Antcin G (25R/S epimer)
*	E33/34	10.982	255	C29H40O4	453.3010	453.3008	0.51	309, 393	Antcin A, S + R
*	L1	8.146	244	C30H46O4	469.3323	469.3315	1.77	83, 97	3β,15α-Dihydroxylanosta-7,9(11),24(28)-triene-21-oic acid
	L2	8.363	198	C30H48O4	471.3480	471.3488	−1.73	83, 97	3β,15α-Dihydroxylanosta-8,24(28)-diene-21-oic acid
*	L3	9.022	244	C31H48O4	483.3480	483.3477	0.59	83, 97	Dehydrosulphurenic acid
*	L4	9.247	198	C31H50O4	485.3636	485.3635	0.27	83, 97, 181	Sulphurenic acid
*	L5	11.532	w	C33H50O5	525.3585	525.3592	−1.24	59	15α-Acetyldehydrosulphurenic acid
	L6	11.858	w	C33H52O5	527.3742	527.3743	−0.19	59	Versisponic acid D
	L7	13.956	w	C30H46O3	453.3374	453.3372	0.48	97, 325	3β-Hydroxylanosta-7,9(11),24(28)-triene-21-oic acid
	L8	14.360	w	C30H48O3	455.3531	455.3530	0.15	97, 210, 275	3β-Hydroxylanosta-8,24(28)-diene-21-oic acid
*	L9	14.644	244	C31H48O3	467.3531	467.3533	−0.49	97, 337	Dehydroeburicoic acid
*	L10	15.069	198	C31H50O3	469.3687	469.3688	−0.17	97, 339	Eburicoic acid
*	O1	4.129	w	C24H36O5	405.2636	405.2635	0.13		Antcamphin A
*	O2	6.306	w	C26H36O4	413.2686	413.2691	−1.13		Antcamphin B
*	O3	12.412	w	C29H42O3	439.3207	439.3210	−0.75		Antcamphin C, trans
*	O4	12.645	w	C29H42O3	439.3207	439.3212	−1.21		Antcamphin D, cis
*	O5	9.919	w	C26H38O7	485.2519	485.2510	−2.00		4-Acetyl-antroquinonol B

---

E4/E5 had a molecular formula of C₂₉H₄₂O₇, which was new for A. cinnamomea ergostanes. Thus, they were detected from this mushroom for the first time. The MS/MS spectra of [M − H]⁻ at m/z 501 were similar to those of E7, showing a fragment ion at m/z 439 due to the loss of 62 Da (CO₂ + H₂O). Thus, they should contain hydroxyl groups at C-4 and C-7. The neutral loss of 28 Da from m/z 423 to 395 was similar to E21/E22, and indicated the presence of a hydroxyl group at C-12 (Fig. 2S†). Thus, the structures for E4/E5 were tentatively characterized as 12-hydroxyl antcamphin E/F. E18/E19/E24 all had the molecular formula C₂₉H₄₀O₆ ([M − H]⁻ m/z 483). E24 could fragment into m/z 423 due to the loss of 60 Da (C₂H₄O₂), and m/z 421 due to the loss of 62 Da (CO₂ + H₂O). This information was consistent with the known antcin D.²³ E18/E19 showed similar MS/MS spectra to E4 and E5, except that m/z 83 and m/z 123 were observed, and the [M − CO₂ − H₂O]⁻ ion was absent. Thus, their C-4 was not substituted with hydroxyl group. They were identified as 3-ΔO, 7-OH, 12-OH, Δ-ergostanes (Table 1). E9, E15, E25 and E26 had the same molecular formula C₂₉H₄₀O₇ ([M − H]⁻ m/z 499), which was also new for A. cinnamomea. They should be ergostanes with tetra-O-substitutions, though the structures could not be established due to limited structural information.

Lanostanes exhibited only a few fragment ions upon CID at a collision energy of 50 eV, probably due to their low level of oxygenation (Fig. 3S†). Common fragment ions included m/z 83 (C₆H₁₁) or m/z 97 (C₇H₁₃) derived from the side chain,²¹ together with m/z 59 from the acetyl groups. Thus, we could only determine the molecular formula of lanostanes according to the high-resolution mass spectral data, and little structural information could be obtained from the tandem mass spectra. Fortunately, we found the lanostanes in A. cinnamomea were present in pairs, containing either Δ^7,9(11) or Δ⁸ double bond. For each pair, the two compounds usually had very similar HPLC retention times. Moreover, the characteristic maximal UV absorption (around 242 and 198 nm for Δ^7,9(11) and Δ⁸ lanostanes, respectively) could also assist in identifying their structures. A total of 10 lanostanes were characterized, and 6 of them (L1, L3, L4, L5, L9, L10) were further confirmed by comparing with reference standards. L2, L7 and L8 were reported from A. cinnamomea for the first time. L7 had been reported from Poria cocos (Fu-Ling), which is also a medicinal mushroom from polyporaceae family.²⁴ Although L2 and L8 were present in A. cinnamomea in noticeable amounts, they might be neglected in previous phytochemical studies due to their poor UV absorption.

Two nor-ergostanes O1 and O2, as well as two ergostane aldehydes O3 and O4, were identified from A. cinnamomea by LC/MS. We had reported these novel ergostanes from the fruiting bodies very recently.⁴ O1 and O2 contained fewer carbon atoms (24C and 26C, respectively) than normal ergostanes (29C). O3 and O4 were the first ergostane aldehydes reported from A. cinnamomea. Because they do not contain a carboxyl group, their ionization was poor in the (−)-ESI mode. Thus, they were analyzed in the positive ion mode (Fig. 1S†). We also detected a cyclohexenone (O5), which was identified by comparing with a reference standard.

3.3. Quantitative analysis of triterpenoids in A. cinnamomea

3.3.1. Optimization and validation of the UPLC/UV method. Lin et al. and Du et al. had reported HPLC methods to analyze triterpenoids in A. cinnamomea.^9,12 However, these methods required a long analysis time (≥100 min) due to similar retention behaviors of analogous compounds. Moreover, the 25R/S-epimers of ergostanes were usually not baseline-separated. In this study, we used ultra-performance liquid chromatography to improve chromatographic resolution and to reduce analysis time.

We tested different types of reversed-phase columns including Acquity CSH C₁₈ (2.1 × 100 mm, 1.7 μm, Waters, MA, USA), Zorbax SB-C₈ (2.1 × 150 mm, 1.8 μm, Agilent), Zorbax SB–Phenyl (2.1 × 150 mm, 1.8 μm, Agilent); Zorbax Eclipse Plus Phenyl–Hexyl (2.1 × 100 mm, 1.8 μm, Agilent); Zorbax Eclipse Plus C₁₈ (2.1 × 150 mm, 1.8 μm, Agilent); HSS T3 (1.8 μm, 2.1 × 150 mm, Waters). HSS T3 showed the best resolution to ergostane epimers (Fig. 4S†), and was chosen for quantitation. Meanwhile, we found that a small portion of methanol in the mobile phase could alter the distribution of major ergostanes.²⁵ Thus, acetonitrile containing 1% methanol was used as the organic phase. A column temperature at 45 °C could significantly improve the resolution than 40 °C, and was finally used (Fig. 5S†). Different flow rates (0.2, 0.3, 0.4 mL min⁻¹) showed little contribution to the resolution, and the flow rate of 0.4 mL min⁻¹ was used. The column pressure before injection was 9250 psi.

According to their UV absorption maxima, different types of triterpenoids were detected at varied wavelengths. The 7-OH/H ergostanes (E1, E2, E22, E23) and Δ^7,9(11) lanostanes (L1, L3, L5, L9) were detected at 250 nm, and 7-ΔO ergostanes (E16, E17, E29, E30) were monitored at 270 nm (Fig. 6A). The 12 major peaks were well separated within 16 minutes.

Fig. 6 LC/UV chromatograms and SFC/MS selected ion chromatograms to determine 25R/S-ergostanes and Δ^7,9(11)-lanostanes (A), 25R/S-antcin A (B), and Δ⁸-lanostanes (C) in Antrodia cinnamomea. The chromatogram of batch S2 was used for A, and batch S10 was used for B and C.

Linearity of the 12 analytes were r² > 0.9990 in a dynamic range of no less than 40 folds (1–8.33 μg mL⁻¹ to 40–480 μg mL⁻¹, Table 2). Precision was determined by calculating the relative standard deviation (RSD) of peak areas at three concentration levels measured in the same day (n = 5) and on three consecutive days. Intra- and inter-day precision was 0.70–1.80% and 0.22–2.12%, respectively, suggesting acceptable variation of the method. Accuracy was evaluated by standard addition test at three concentration levels (n = 3 for each level). QC solutions were spiked into a sample matrix as described in section 2.3. The recovery rates were calculated to represent accuracy of the method (recovery = amount found/amount spiked × 100%). Recoveries of the 12 triterpenoids ranged from 93.18–112.06% (Table 3). The limits of quantitation (LOQ, S/N = 10, assigned as the lowest concentration of the calibration curve) ranged from 1–8.33 μg mL⁻¹, while the limits of detection (LOD, S/N = 3) were lower than 0.3 μg mL⁻¹ for the analytes. Repeatability was evaluated by preparing triplicate samples at three different concentrations, and calculating the RSD values of peak areas. Variation introduced by sample preparation was lower than 4.51%. Stability was determined by analyzing samples after 2, 4, 8, 12, 24, 36 h storage in the sample tray. Variation resulted from storage never exceeded 4.88% (Table 1S†).

Table 2 Calibration curves of triterpenoids for UPLC/UV and SFC/MS quantitation

	Detection	Calibration equation	r^2	Range (μg mL^−1)
a UV 198 nm data were collected after subtraction of a blank run (methanol).
E1	UV 250 nm	y = 1642.2x + 1654.6	0.9999	3.33–320
E2	UV 250 nm	y = 1430.6x + 1806.0	0.9999	5.00–480
E16	UV 250 nm	y = 2053.6x + 674.62	0.9997	3.13–150
E17	UV 250 nm	y = 2045.9x + 1841.4	0.9996	3.13–150
E22	UV 270 nm	y = 2628.5x + 2458.6	0.9998	3.33–160
E23	UV 270 nm	y = 2119.3x + 3644.5	0.9999	5.00–480
L1	UV 250 nm	y = 1958.6x + 2181.5	0.9999	6.67–320
L3	UV 250 nm	y = 2435.4x − 701.63	0.9998	5.00–480
E29	UV 270 nm	y = 2256.1x + 2242.2	0.9998	4.17–200
E30	UV 270 nm	y = 2546.2x + 1878.8	0.9998	3.33–160
L5	UV 250 nm	y = 2252.3x + 268.98	0.9994	1.00–40
L9	UV 250 nm	y = 3129.7x + 4595.2	0.9999	8.33–400
L10	UV 198 nm^a	y = 4.6725x − 2.5000	0.9999	4.00–400

---

Table 3 Accuracy of the UPLC/UV method (n = 3)

	E1	E2	E16	E17	E22	E23	L1	L3	E29	E30	L5	L9	L10
Matrix (μg)	20.26	38.51	19.91	12.70	23.21	42.72	33.14	49.36	26.21	15.41	2.05	29.18	44.04
RSD (%)	0.41	0.41	2.29	0.07	0.54	0.52	0.54	0.53	0.40	0.57	0.06	0.48	0.90
Low QC
Spiked (μg)	10.67	24.00	10.00	3.67	11.00	20.00	16.00	20.00	17.33	7.00	0.73	13.33	20.00
Found (μg)	10.61	23.24	9.81	3.81	11.48	20.03	17.16	20.73	19.42	6.85	0.76	13.31	20.77
Accuracy (%)	99.50	96.82	98.07	101.75	104.32	100.13	107.24	103.63	112.06	97.92	103.07	98.29	103.87
RSD (%)	3.20	3.49	4.31	3.51	3.33	3.39	3.64	2.99	2.12	2.63	4.09	3.50	1.16
Medium QC
Spiked (μg)	21.33	48.00	20.00	7.33	22.00	40.00	32.00	40.00	34.67	14.00	1.47	26.67	40.00
Found (μg)	21.00	45.89	18.64	7.78	22.65	39.58	33.88	41.22	38.60	13.56	1.48	26.29	40.21
Accuracy (%)	98.43	95.61	93.18	106.40	102.94	98.95	105.89	103.06	111.34	96.88	100.69	98.75	100.51
RSD (%)	0.98	0.84	2.85	1.04	1.17	1.29	1.06	1.32	0.93	1.75	1.31	1.13	0.18
High QC
Spiked (μg)	32.00	72.00	30.00	11.00	33.00	60.00	48.00	60.00	52.00	21.00	2.20	40.00	60.00
Found (μg)	31.18	68.31	27.96	11.30	34.48	58.85	50.41	61.18	57.52	20.10	2.19	39.18	58.67
Accuracy (%)	97.42	94.88	93.18	102.22	104.50	98.08	105.03	101.96	110.61	95.71	99.34	97.77	97.79
RSD (%)	0.70	0.78	2.47	0.89	1.58	0.49	0.70	0.40	0.57	0.18	0.81	0.31	0.95

---

3.3.2. Quantitation of 25R/S-antcin A by SFC/MS. 25S- and 25R-antcin A (E33 and E34) are a pair of low-polarity ergostane epimers that could not be separated by RP-HPLC. In our recent study, we found that they could be well separated by SFC.¹³ UV detection was difficult for quantitative analysis of these two compounds, due to their low amounts in A. cinnamomea and the interference of neighboring peaks. Therefore, we used SFC/MS to determine the contents of E33 and E34 in this study. E33 and E34 are the only two compounds with a molecular weight of 454 Da in the extract. They were detected at m/z 453.3 in the selected ion monitoring (SIM) mode by using a singular quadrupole mass spectrometer. As shown in Fig. 6B, the epimers could be well separated within 5 minutes on SFC. Their concentrations and peak areas fit the quadratic model in the range of 1.95–62.5 μg mL⁻¹. The calibration curves for E33 and E34 were y = −1276.6x² + 154713x − 71 554 (r² = 0.9909) and y = −1114.6x² + 144244x − 12 2729 (r² = 0.9933), respectively, where y represented the peak area, x the analyte concentration in μg mL⁻¹, and r the correlation coefficient.

3.3.3. Quantitation and semi-quantitation of Δ⁸ lanostanes. Δ⁸ lanostanes have poor UV absorption with a maximum absorption wavelength at 198 nm. Their contents in A. cinnamomea have never been clarified, thus far. Our UHPLC/qTOF-MS analysis revealed the presence of a number of Δ⁸ lanostanes in A. cinnamomea. They could be clearly observed in the chromatogram of 198 nm, after background subtraction (Fig. 6S†). However, we obtained only one reference standard L10, as the other compounds were neglected during chromatographic separation due to their poor UV absorption. Therefore, we determined the content of L10 at 198 nm by UHPLC/UV, and calculated the contents of three other analogues (L2, L4, and L6) by semi-quantitation using the calibration curve of L10 (Fig. 6C). Method validation data for L10 were shown in Table 3 and 1S†.

3.3.4. Contents of 18 triterpenoids in 15 batches of A. cinnamomea. The contents of 10 ergostanes (including 5 epimeric pairs of 25R/S-antcin A, B, C, H, and K) and 8 lanostanes (3 of them were semi-quantified) were tested in 15 batches of A. cinnamomea samples. These samples were derived from different cultivation techniques, including cutting wood culture (5 batches of fruiting bodies and 4 batches of mycelia), solid support culture (2 batches), submerged fermentation (2 batches), and dish culture (2 batches). The results are given in Table 4 and Fig. 7. The 18 compounds accounted for 118.2 ± 28.2, 89.4 ± 30.8, and 116.5 ± 1.1 mg g⁻¹ in wood-cultured fruiting bodies, wood-cultured mycelia, and dish-cultured samples, respectively. However, no triterpenoids were detected in the 4 batches of solid support cultivation or submerged fermentation samples. It should be pointed out that these 4 batches of samples were purchased from food stores or supermarkets, and were claimed to be derived from these cultivation techniques. Whether these cultivation methods could produce triterpenoids needs to be verified by more batches of samples.

Table 4 Contents of 18 triterpenoids in Antrodia cinnamomea samples^b

No.	Ergostanes (mg g^−1)											Lanostanes (mg g^−1)									Total
	E1	E2	E16	E17	E22	E23	E29	E30	E33	E34	Sum	L1	L2^a	L3	L4^a	L5	L6^a	L9	L10	Sum
a Semi-quantified by using the calibration curve of L10; ND, not detected.b E1/E2, 25S/R-antcin K; E16/E17, 25S/R-antcin C; E22/E23, 25R/S-antcin H; E29/E30, 25S/R-antcin B; E33/E34, 25S/R-antcin A; L1, 3β,15α-dihydroxylanosta-7,9(11),24-triene-21-oic acid; L3, dehydrosulphurenic acid; L5, 15α-acetyldehydrosulphurenic acid; L9, dehydroeburicoic acid.
Wood culture, fruiting bodies
S1	11.4	19.8	3.4	1.8	5.9	13.3	7.9	4.6	0.5	0.2	68.8	7.8	3.9	9.7	23.8	1.6	3.5	2.0	13.9	66.2	135.0
S2	7.6	10.8	4.5	3.5	4.3	8.3	2.4	1.3	0.3	0.2	43.2	10.3	3.4	12.8	19.6	0.7	ND	8.1	12.2	67.1	110.3
S3	12.6	21.8	11.1	9.3	11.2	20.1	11.4	6.4	0.3	0.2	104.4	9.8	4.1	8.0	10.2	2.0	1.7	4.1	2.4	42.3	146.7
S10	6.8	12.9	6.8	4.4	7.8	14.2	8.5	5.0	0.3	0.2	66.9	10.5	4.0	15.6	13.7	0.7	0.7	9.4	3.7	58.3	125.2
S11	5.1	9.8	4.6	4.5	4.5	7.8	6.0	3.2	0.5	0.3	46.3	1.6	1.4	4.0	11.6	0.6	1.0	3.8	3.5	27.5	73.8
Wood culture, mycelia
S4	1.2	2.3	ND	ND	0.8	1.0	1.7	1.2	1.1	0.6	9.9	2.2	3.4	11.5	19.6	0.5	ND	20.9	12.2	70.3	80.2
S5	1.5	3.2	0.7	0.4	1.2	2.2	2.6	1.8	0.7	0.3	14.6	6.5	3.9	23.5	23.8	1.1	0.9	29.0	13.9	102.6	117.2
S6	0.6	1.3	ND	ND	0.5	0.6	1.3	0.9	0.6	0.3	6.1	7.1	3.8	26.6	20.9	1.3	0.8	30.1	13.2	103.8	109.9
S12	3.3	3.9	ND	ND	0.6	1.0	ND	ND	ND	ND	8.8	5.0	2.9	10.8	13.2	0.3	ND	6.7	2.3	41.2	50.0
Dish culture
S13	23.8	29.4	7.8	5.7	3.9	7.0	2.2	1.1	0.4	0.3	81.6	1.6	2.1	5.9	18.2	0.8	1.5	2.7	2.8	35.6	117.2
S14	23.8	29.7	9.2	6.5	3.6	6.6	1.6	0.8	0.4	0.3	82.5	1.6	2.1	5.9	16.1	0.9	1.4	2.7	2.6	33.3	115.8
Solid support culture (commercial)
S9	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND
S15	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND
Submerged fermentation (commercial)
S7	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND
S8	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND

---

Fig. 7 UPLC/UV chromatograms at 250 nm for triterpenoids (A) and IC/PAD chromatograms for polysaccharide hydrolysis products (B) in A. cinnamomea samples S3, S5, S7, S13 and S15. gal, D-galactose; glu, D-glucose; fru, L-fructose; ara, L-arabinose; xyl, D-xylose; man, D-mannose; fuc, L-fucose. nC stands for 1 × 10⁻⁹ Coulomb. E1/E2, 25S/R-antcin K; E16/E17, 25S/R-antcin C; E22/E23, 25R/S-antcin H; E29/E30, 25S/R-antcin B; E33/E34, 25S/R-antcin A; L1, 3β,15α-dihydroxylanosta-7,9(11),24-triene-21-oic acid; L3, dehydrosulphurenic acid; L5, 15α-acetyldehydrosulphurenic acid; L9, dehydroeburicoic acid.

It is generally considered that fruiting bodies of A. cinnamomea are better in quality than the mycelia. Thus, the products derived from fruiting bodies in the market are much more expensive than those from mycelia. Our results showed the triterpenoids in fruiting bodies and mycelia were remarkably different. The fruiting bodies contained a lot more ergostanes than mycelia (65.9 ± 24.4 vs. 9.9 ± 3.6 mg g⁻¹), whereas the amounts of lanostanes were similar (52.3 ± 17.1 vs. 79.5 ± 29.9 mg g⁻¹). The ratio of ergostanes to lanostanes was roughly 5 : 4 in fruiting bodies, and 1 : 8 in mycelia, as shown in Fig. 8. We had discovered that the ergostanes had better oral bioavailability than lanostanes in our pharmacokinetic study on A. cinnamomea.²⁶ From this point of view, it is reasonable that the fruiting bodies have higher prices than the mycelia.

Fig. 8 Contents of major (25R)-ergostanes, (25S)-ergostanes, Δ^7,9(11) lanostanes, Δ⁸ lanostanes (A) and monosaccharides (B) in 15 batches of A. cinnamomea samples. gal, D-galactose; glu, D-glucose; fru, L-fructose; ara, L-arabinose; xyl, D-xylose; man, D-mannose; fuc, L-fucose.

The contents of single triterpenoids in the fruiting bodies or mycelia samples varied remarkably. For instance, the amount of E29 was 2.4 mg g⁻¹ in batch S2, and 11.4 mg g⁻¹ in S3. As a result, the triterpenoid profiles of different batches of sample showed big variations. This may be due to different culture conditions and cultivation ages.⁹

Dish culture is a relatively new and efficient cultivation technique for the manufacturing of A. cinnamomea. Our results showed the dish culture samples had very similar triterpenoid profiles to the fruiting bodies. Ergostanes were also more abundant than lanostanes. Particularly, the amount of antcin K (E1 and E2) was as high as >50 mg kg⁻¹. Given the significant bioactivities of antcin K,²⁷ dish culture samples may be a good source to purify this important compound. The chemical variation for dish culture samples was relatively small, though only two batches of samples from the same manufacturer were analyzed.

The amounts of 25R/S-epimers of 5 pairs of ergostanes (antcins A, B, C, H, and K) in A. cinnamomea were determined in pure optical form for the first time. Interestingly, the total amounts of 25S-forms were almost identical with those of 25R-forms, for all the wood culture or dish culture samples. The S/R ratio was 1.14, 1.03, and 1.02 for fruiting bodies, mycelia, and dish culture samples, respectively (Fig. 8). However, the ratio was different for each pair of epimers. The S-form was more abundant than the R-form for antcin C (S/R ratio 1.4 ± 0.3), antcin H (1.8 ± 0.3), antcin B (1.7 ± 0.2), and antcin A (1.8 ± 0.3). The situation was different for antcin K, which had an S/R ratio of 0.6 ± 0.1. These results were consistent with our semi-quantitative data by SFC analysis.¹³

Similar to 25R/S ergostanes, Δ^7,9(11) and Δ⁸ lanostanes were also present in pairs. The total amounts of Δ^7,9(11) lanostanes (29.8 ± 18.7 mg g⁻¹, L1 + L3 + L5 + L9) were almost identical with those of Δ⁸ lanostanes (29.1 ± 10.4 mg g⁻¹, L2 + L4 + L6 + L10). However, we could not draw a clear conclusion on the ratio for each pair of lanostanes.

3.4. Chemical profiling of polysaccharides

Polysaccharides are another important type of bioactive chemicals in A. cinnamomea.^14,15 To elucidate the polysaccharide composition, they were hydrolyzed into monosaccharides by using trifluoroacetic acid, and were then determined by IC/PAD.^28,29 The seven monosaccharides (glucose, galactose, fructose, fucose, mannose, xylose, arabinose) showed good linearity (r² > 0.998) in a dynamic range of 1.0–40.0 μg mL⁻¹, and the concentration of S/N = 10 (signal-to-noise ratio) ranged from 0.02–0.31 μg mL⁻¹ (Table 2S†). The hydrolysis products were diluted (9 folds for S2 and S7; 3 folds for the other samples), and were analyzed within the dynamic ranges. The results are shown in Fig. 8B.

The polysaccharides in A. cinnamomea were mainly composed of 7 monosaccharides, including glucose (glu), galactose (gal), fructose (fru), fucose (fuc), mannose (man), xylose (xyl), and arabinose (ara) (Fig. 7). Other sugars like rhamnose, galacturonic acid or glucuronic acid were not detected in the hydrolyzed products. Glu and gal were the major monosaccharides. For the wood culture samples, glu and gal accounted for 87.24 ± 11.40% among the seven sugars (Fig. 8B). This result was consisted with previous reports.^14,16 This ratio was lower for the other samples, which contained noticeable amounts of xyl, man, fru, and ara. Interestingly, the xyl : man : fru : ara ratio was similar for solid culture (10 : 7 : 5 : 8) and submerged fermentation samples (10 : 8 : 10 : 8), and was remarkably different from the dish culture samples (1 : 3 : 15 : 0). This index may be useful to identify the cultivation techniques.

The contents of total polysaccharides varied from 3.2–416.7 mg g⁻¹ among the 15 batches of samples (Table 3S†). Even for the same cultivation technique, the contents varied remarkably between the different batches. For instance, the contents in fruiting bodies ranged from 3.2–143.8 mg g⁻¹. Although the two batches of dish culture samples contained similar amounts of triterpenoids, the polysaccharide contents varied remarkably (120.0 and 62.3 mg g⁻¹, respectively). No triterpenoids were detected in solid culture and submerged fermentation samples. However, they generally contained high amounts of polysaccharides. For batch S7, the content of polysaccharide was as high as 416.7 mg g⁻¹.

4. Conclusions

A series of new methods were established for qualitative and quantitative analyses of triterpenoids and polysaccharides in A. cinnamomea. A total of 49 compounds, including 34 ergostanes and 10 lanostanes, were characterized by UHPLC/DAD/qTOF-MS, and 31 of them were confirmed by comparing with reference standards. The (−)-ESI-MS/MS fragmentation pathways of ergostanes with different substitution patterns (7-H/7-ΔO, 7-OH, 12-OH, 4-OH, Δ¹⁴) were proposed. In total 18 major triterpenoids were determined by UPLC/UV or SFC/MS (for 25R/S-antcin A). The contents of pure optical forms for 5 pairs of 25R- and 25S-ergostane epimers (antcins A, B, C, H and K), together with Δ⁸ lanostanes, were determined for the first time. Moreover, polysaccharide composition was profiled in acid-hydrolyzed water extracts of A. cinnamomea by IC/PAD analysis, and the contents of 7 component sugars were determined. These methods were used to analyze 15 batches of A. cinnamomea samples derived from different cultivation techniques. They could be powerful tools to evaluate the quality of A. cinnamomea and its products.

Acknowledgements

This work was supported by National Natural Science Foundation of China (no. 81222054, no. 81303294), the Program for New Century Excellent Talents in University from Chinese Ministry of Education (no. NCET-11-0019), and the State Key Laboratory of Drug Research (no. SIMM1403KF-16). We thank Drs Rong An, Lang Li, and Ying Meng (Agilent Technologies) for their technical help in establishing the SFC/MS system. We also wish to thank Mr Jen-Yu Lo (Honest & Humble Biotechnology Co., Ltd) for kindly providing the dish culture samples.

Notes and references

1. M. Geethangili and Y. M. Tzeng, J. Evidence-Based Complementary Altern. Med., 2011, 212641.
2. Formosa Cancer Foundation, 2005, Internet at http://www.canceraway.org.tw/uploads/200509.pdf, Accessed 28 Feb 2014.
3. Chang Gung Memorial Hospital, 2011, http://clinicaltrialsfeeds.org/clinical-trials/show/NCT01287286, Accessed 15 Apr 2014.
4. Y. Huang, X. Lin, X. Qiao, S. Ji, K. Liu, C. T. Yeh, Y. M. Tzeng, D. Guo and M. Ye, J. Nat. Prod., 2014, 77, 118–124.
5. Z. H. Ao, Z. H. Xu, Z. M. Lu, H. Y. Xu, X. M. Zhang and W. F. Dou, J. Ethnopharmacol., 2009, 121, 194–212.
6. C. C. Liaw, Y. C. Chen, G. J. Huang, Y. C. Tsai, S. C. Chien, J. H. Wu, S. Y. Wang, L. K. Chao, P. J. Sung, H. C. Huang and Y. H. Kuo, J. Nat. Prod., 2013, 76, 489–494.
7. G. J. Huang, J. S. Deng, S. S. Huang, C. Y. Lee, W. C. Hou, S. Y. Wang, P. J. Sung and Y. H. Kuo, Food Chem., 2013, 141, 3020–3027.
8. C. T. Yeh, Y. K. Rao, C. J. Yao, C. F. Yeh, C. H. Li, S. E. Chuang, J. H. T. Luong, G. M. Lai and Y. M. Tzeng, Cancer Lett., 2009, 285, 73–79.
9. T. Y. Lin, C. Y. Chen, S. C. Chien, W. W. Hsiao, F. H. Chu, W. H. Li, C. C. Lin, J. F. Shaw and S. Y. Wang, J. Agric. Food Chem., 2011, 59, 7626–7635.
10. E. Majid, K. B. Male, Y. M. Tzeng, J. O. Omamogho, J. D. Glennon and J. H. T. Luong, Electrophoresis, 2009, 30, 1967–1975.
11. Y. K. Rao, M. Geethangili and Y. M. Tzeng, Anal. Methods, 2013, 5, 5724–5730.
12. Y. C. Du, T. Y. Wu, F. R. Chang, W. Y. Lin, Y. M. Hsu, F. T. Cheng, C. Y. Lu, M. H. Yen, Y. T. Tsui, H. L. Chen, M. F. Hou, M. C. Lu and Y. C. Wu, J. Pharm. Biomed. Anal., 2012, 58, 182–192.
13. X. Qiao, R. An, Y. Huang, S. Ji, L. Li, Y. M. Tzeng, D. Guo and M. Ye, J. Chromatogr. A, 2014, 1358, 252–260.
14. J. J. Cheng, C. C. Chang, C. H. Chao and M. K. Lu, Carbohydr. Polym., 2012, 90, 134–139.
15. I. H. Lee, R. L. Huang, C. T. Chen, H. C. Chen, W. C. Hsu and M. K. Lu, FEMS Microbiol. Lett., 2002, 209, 63–67.
16. C. C. Chen, Y. W. Liu, Y. B. Ker, Y. Y. Wu, E. Y. Lai, C. C. Chyau, T. H. Hseu and R. Y. Peng, J. Agric. Food Chem., 2007, 55, 5007–5012.
17. Z. Zhang, N. M. Khan, K. M. Nunez, E. K. Chess and C. M. Szabo, Anal. Chem., 2012, 84, 4104–4110; Y. Cai, J. Liu, Y. Shi, L. Liang and S. Mou, J. Chromatogr. A, 2005, 1085, 98–103.
18. M. C. Lu, M. El-Shazly, T. Y. Wu, Y. C. Du, T. T. Chang, C. F. Chen, Y. M. Hsu, K. H. Lai, C. P. Chiu, F. R. Chang and Y. C. Wu, Pharmacol. Ther., 2013, 139, 124–156.
19. M. Yang, X. Wang, S. Guan, J. Xia, J. Sun, H. Guo and D. Guo, J. Am. Soc. Mass Spectrom., 2007, 18, 927–939.
20. C. Cheng, M. Yang, Z. Wu, Y. Wang, F. Zeng, W. Wu, S. Guan and D. Guo, Rapid Commun. Mass Spectrom., 2011, 25, 1323–1335.
21. X. Wu, R. Luo, X. W. He, P. Yu and G. Yuan, Eur. J. Mass Spectrom., 2011, 17, 297–304.
22. X. Liu, S. L. Li, Y. Zhou, J. Z. Song, Y. F. Zheng, G. P. Peng and H. X. Xu, Rapid Commun. Mass Spectrom., 2010, 24, 1514–1522.
23. I. H. Cherng, D. P. Wu and H. C. Chiang, Phytochemistry, 1996, 41, 263–267.
24. A. Kanematsu and S. Natori, Chem. Pharm. Bull., 1970, 18, 779–783.
25. X. Qiao, M. Ye, Y. Liang, W. Yang and D. Guo, J. Sep. Sci., 2011, 34, 169–175.
26. X. Qiao, Q. Wang, S. Ji, Y. Huang, K. Liu, Z. Zhang, T. Bo, Y. Tzeng, D. Guo and M. Ye, J. Pharm. Biomed. Anal., 2015, 111, 266–276.
27. Y. C. Shen, Y. H. Wang, Y. C. Chou, C. F. Chen, L. C. Lin, T. T. Chang, J. J. Tien and C. J. Chou, Planta Med., 2004, 70, 310–314; M. Geethangili, S. H. Fang, C. H. Lai, Y. K. Rao, H. M. Lien and Y. M. Tzeng, Food Chem., 2010, 119, 149–153.
28. J. Da, W. Y. Wu, J. J. Hou, H. L. Long, S. Yao, Z. Yang, L. Y. Cai, M. Yang, B. H. Jiang, X. Liu, C. R. Cheng, Y. F. Li and D. A. Guo, J. Chromatogr. A, 2012, 1222, 59–70.
29. W. Song, L. L. Si, S. Ji, H. Wang, X. M. Fang, L. Y. Yu, R. Y. Li, L. N. Liang, D. M. Zhou and M. Ye, J. Nat. Prod., 2014, 77, 1632–1643.

---

Footnote

† Electronic supplementary information (ESI) available: Method validation and optimization data. See DOI: 10.1039/c5ra04327a

---