A four-dimensional separation approach by offline 2D-LC/IM-TOF-MS in combination with database-driven computational peak annotation facilitating the in-depth characterization of the multicomponents from Atractylodis Macrocephalae Rhizoma (Atractylodes macrocephala)

2.1 Reagents and chemicals

In total, nineteen compounds (Fig. S1 and Table S1), involving five terpenoids, three phenylpropanoids, two volatile oils, one oligosaccharide, three flavonoids, and five others, were purchased from Shanghai Standard Biotech. Co., Ltd. (Shanghai, China) and Chengdu Desite. Co., Ltd. (Chengdu, China), and used as the reference compounds in this work. Acetonitrile, methanol (Fisher, Fair lawn, NJ, USA), formic acid (FA), and (ammonium acetate) (AA; Fisher, Fair lawn, NJ, USA) were LC-MS grade. Ultra-pure water was in-house prepared using a Milli-Q Integral 5 water purification system (Millipore, Bedford. MA, USA). The sample of A. macrocephala analyzed in this work was collected from Pan’an of Zhejiang Province. The voucher specimen of A. macrocephala was deposited in the authors’ lab in Tianjin University of Traditional Chinese Medicine (Tianjin, China).

2.2 Sample preparation

To the sample of A. macrocephalas, 6.0 g of the accurately weighed powder was soaked in 10 mL of 70% aqueous methanol (v/v), vortexed for 2 min, and extracted by ultrasound assistance for 40 min in a water bath at 25 °C. After a 10-min centrifugation at 3,219 g, the lost weight was compensated with 70% methanol. The supernatant was taken and further diluted in a 10-mL volumetric flask to the constant volume. Suffering from another centrifugation process at 11,481 g for 10 min, the supernatant obtained was used as the test solution (with the final concentration at 600 mg/mL).

2.3 Offline 2D-LC/IM-QTOF-MS

Combination of HILIC × RPLC was used to establish an offline 2D-LC system for separating the chemical components contained in A. macrocephalas. In detail, the total extract of A. macrocephalas was initially separated by an XBridge® Amide column (4.6 × 150 mm, 3.5 µm) on an Agilent 1260 HPLC system (Agilent Technologies, Waldbronn, Germany). The column temperature was maintained at 30 °C and the binary mobile phase consisting of 0.1% formic acid (A) and acetonitrile (B) at a flow rate of 1.0 mL/min was used following a gradient elution program: 0–2 min, 98%–90% (B); 2–6 min, 90%–85% (B); 6–8 min, 85%–80% (B); 8–10 min, 80%–75% (B); 10–11 min, 75%–70% (B); 11–12 min, 70%–67% (B); 12–14 min, 67%–65% (B); 14–14.5 min, 65%–50% (B); and 14.5–16 min, 50% (B). The injection volume was 40 µL. The injection was repeated for five times. The PDA detector monitored the signals at 282 nm. Eluate collection was based on a peak-dependent fractionation approach, which finally gave eight fractions (numbered as Fr. 1 to Fr. 8). Each of the dry residues corresponding to eight fractions was dissolved in 200 µL of 70% methanol (v/v) and further centrifuged at 11,481 g for 10 min, the resultant supernatants were the test solutions (with the final concentration at 1.0 mg/mL equal to the raw drug material) for the second-dimensional UHPLC separation in the RP mode. An Atlantis Premier BEH C18AX column (2.1 × 100 mm, 1.7 µm) maintained at 30 °C was configured on an ACQUITY UPLC I-Class/Vion IM-QTOF system (Waters Corporation, Manchester, UK) for the ²D separation. A binary mobile phase, containing 0.1% formic acid (A) and acetonitrile (B), ran consistent with the following gradient program: 0–3 min, 3%–30% (B); 3–6 min, 30%–50% (B); 6–10 min, 50%–60% (B); 10–13 min, 60%–65% (B); 13–16 min, 65%–75% (B); 16–19 min, 75%–85% (B); 19–24 min, 85%–93% (B); 24–24.5 min, 93%–98% (B); and 24.5–26.5 min, 98% (B). The flow rate was kept at 0.3 mL/min, and an injection volume of 10 µL was used.

High-accuracy MS data were acquired on a Waters Vion™ IM-QTOF mass spectrometer (Waters, Manchester, UK) coupled with the UPLC I-Class system via a Zspray™ ESI source. In order to obtain more reliable characterization results, high-definition MS^E in positive mode was utilized. The ESI source parameters were set as follows: capillary voltage, 3.0 kV; cone voltage, 60 V; source offset voltage, 80 V; source temperature, 120 °C; desolvation temperature, 500 °C; cone gas flow rate, 50 L/h; and desolvation gas flow rate, 800 L/h. For HDMS^E settings, the mass analyzer scanned over a mass range of 50–1400 Da in full scan with a scan time of 0.3 s under the low collision energy of 6 eV. MS² experiments were performed by collision-induced dissociation (CID) at the ramp collision energy (RCE) of 20–50 eV covering the same scan range as MS¹. Data calibration was performed using an external reference (LockSpray™) by the constant infusion of the leucine-enkephalin solution (LE; Sigma-Aldrich; 200 ng/mL) at a flow rate of 10 μL/min. Calibration of CCS was conducted according to the manufacture’s guidelines using a mixture of calibrants (Paglia et al., 2015).

2.4 Orthogonality and peak capacity

Orthogonality of the developed offline 2D-LC/IM-QTOF-MS system was assessed by the asterisk equations proposed by Camenzuli and Schoenmakers (Camenzuli and Schoenmakers, 2014). These equations are provided as Supplementary Materials. Retention time for each of the 54 index components in both two dimensions of separations was firstly transformed into the normalized retention time (t_R,norm(i)) following Eq. (1) (t_D: dead time; t_G: total gradient elution time). The separation space could be described by four crossing lines (S_z-, S_z+, S_z1, S_z2), and the spreading of all 54 index components around them was quantified by Eqs. (2)–(5) (σ stands for the standard deviation). Four Z parameters (Z_-, Z₊, Z₁, Z₂) were then generated in accordance with Eqs. (6)–(9), based on which orthogonality (A₀) was determined using Eq. (10).

2.5 Establishment of the in-house library of A. macrocephalas

An in-house library of A. macrocephalas was created by giving a comprehensive summary on the compounds ever isolated from this species. The literature from Web of Science (www.webofknowledge.com), CNKI (www.cnki.net), ChemSpider (www.chemspider.com), PubChem (http://pubchem.ncbi.nlm.nih.gov), and Chemicalbook (www.chemicalbook.com), etc., dealing with the phytochemical isolation of A. macrocephalas was retrieved. First, according to the information provided in the literature, the structure of each compound was prepared by the ChemDraw Professional software, which was saved as an .mol file. Second, the structural information of all known compounds was input into the EXCEL file in a prescribed format. The .mol file was named with the trivial name consistent with that in the EXCEL file. Third, the EXCEL file and all the structure files were incorporated into the UNIFI™ software through Import Library Item, which could be applied as a database for the subsequent peak annotation. A table detailing the in-house library of A. macrocephalas has been provided as Table S2.

2.6 Automatic peak annotation of the HDMS^E data by UNIFI^TM for structural elucidation

Structural elucidation of the multicomponents from A. macrocephalas was performed following an efficient computational workflow. The UNIFI software was applied to process and annotate the collected HDMS^E data (of eight fractionated samples) by searching the incorporated in-house library of A. macrocephalas. After corrected by reference to the LockMass at m/z 556.2766 (the theoretical m/z value for LE in the ESI+ mode), a list of the characterized components was generated, with the retention times and aligned characteristic fragments. Adduct ions filtering and MS² data analysis were further utilized to remove the false positive characterization results and confirm the identities.

The uncorrected HDMS^E data in Continuum were further processed using UNIFI 1.9.3.0 (Waters). The UNIFI software performed data correction, peak picking, and peak annotation efficiently. Key parameters set in UNIFI were as follows. Find 4D Peaks: High-energy intensity threshold, 500.0 counts; low-energy intensity threshold, 1000.0 counts. Target by mass: Target match tolerance: 10.0 ppm; screen on all isotopes in a candidate, generate predicted fragments from structure, and fragment match tolerance, 10.0 ppm. Adducts: Positive adducts including +H, +Na. Lock Mass: Combine width, 3 scans; mass window, 0.5 m/z; reference mass, 556.2766; reference charge, +1.

3.1 Development of an offline 2D-LC/IM-QTOF-MS system enabling the four-dimensional separation for the multicomponent profiling and characterization of A. macrocephala

Good chromatographic separation is a prerequisite to comprehensively identify the multicomponents in a complicated chemical system like an herbal extract. Conventional LC/MS approaches, due to the utilization of a single mechanism of separation (such as the most preferable RPC) may suffer from severe co-eluting which largely restrains the exposure of those minor ingredients. Herein, we propose to establish a four-dimensional separation approach, by offline 2D-LC/IM-QTOF-MS, to tackle the insufficiency encountered in the comprehensive characterization of TCM components. By this method, 2D-LC can enable orthogonal chromatographic separation with significantly improved peak capacity, and IM-QTOF-MS renders orthogonal IM and MS separations providing the CCS and high-resolution MS¹/MS² information.

The offline 2D-LC system was optimized, in the first step, by screening the stationary phase and optimizing the column temperature and gradient elution program in each dimension of chromatography (Fu et al., 2018; Qiu et al., 2015; Yao et al., 2018). Considering the integration of different separation mechanisms can improve the selectivity and promote the exposure of the minor components from TCM, efforts were made to achieve large orthogonality of two-dimensional separations (Stavrianidi, 2020). Meanwhile, taking into account the high selectivity and the compatibility with the mass spectrometer of the RP mode, RPC was selected as the second-dimensional (²D) chromatography coupled with MS. In this section, eight candidate sub-2 µm particles packed columns featured by different silica gel core, different bonding groups and bonding technologies, including BEH C18, Zorbax Extend C18, Zorbax SB C18, Atlantis Premier BEH C18AX, HSS T3, Zorbax Eclipse Plus C18, Kinetex XB-C18, and Luna Omega Polar C18, were tested using a total extract of A. macrocephala. In general, the resolution of the multicomponents from A. macrocephala by these columns showed significant difference. Comparatively, BEH C18 (the resolvable ions: 10184), Zorbax SB-C18 (10188), Atlantis Premier BEH C18AX (10173), Zorbax Eclipse Plus C18 (10187), and Luna Omega Polar C18 (10184), were able to resolve more peaks than the other three (Fig. 1A), showing better selectivity. By further comparison of these five columns mentioned, we could find that Atlantis Premier BEH C18AX exhibited much better peak symmetry and column efficiency than the others, which was thus considered as the best choice in the ²D chromatography. It is noted that, Atlantis Premier BEH C18AX is a mixed mode column newly released by Waters, which enables the separation of complex ingredients simultaneously by the RP and ion exchange modes. The combined separation modes lead to good resolution of the multicomponents from A. macrocephala. In the next step, by using 19 compounds as the index components, the linearity regression correlation coefficient (R²) was used as a parameter to compare the separation difference between another six candidate columns with Atlantis Premier BEH C18AX that had been selected in the ²D chromatography (Fu et al., 2018). Notably, these candidate columns have different separation mechanisms (RP and HILIC) or are bonded with different functional groups. As exhibited in the right part of Fig. 1A, we could primarily deduce the separation difference with the Atlantis Premier BEH C18AX mixed mode column was consistent with the order: Zorbax HILIC Plus > BEH HILIC > XBridge Amide > COSMOCORE PBr > Exsil Plus 100 C18-PFP > XCharge C18, based on the increasing R² values. However, the Zorbax HILIC Plus and BEH HILIC columns were considered unsuitable for separating the multicomponents of A. macrocephala because of the poor retention ability. As a result, the XBridge Amide column was finally selected as the first-dimensional (¹D) column. The column temperature and gradient elution program for each dimension of chromatography were optimized as well, which resulted in the establishment of the offline 2D-LC system achieving sufficient resolution of the multicomponents from A. macrocephala.

Close

Fig. 1

Selection of the stationary phase in establishing an offline 2D-LC system (A) and optimization of two key source parameters (capillary voltage and cone voltage) of the Vion IM-QTOF mass spectrometer (B) dedicated to well separating the multicomponents from Atractylodes macrocephala. Left column of section A presents the base peak intensity (BPI) chromatograms of eight candidate columns in the ²D chromatography together with the number of resolvable peaks; right column is the scattering plots of index components based on the relative retention time (ranging from 0 to 1) determined on six candidate columns in ¹D chromatography (x-axis) and the selected Atlantis Premier BEH C18AX column of ²D separation (y-axis). R² stands for the linearity regression correlation coefficient, and a smaller value can indicate large separation difference between the two columns. The histograms of section B are plotted by the peak areas of the precursor ions for five representative compounds determined at five different levels.

Export to PPT

Key parameters of the Vion IM-QTOF instrument operating in the positive ESI mode, involving two key ion source parameters (capillary voltage and cone voltage) and the ramp collision energy (RCE), were further optimized by comparing the variations of peak areas of the precursor ions for five representative components (involving β-sitosterol, daphnetin, atractylenolide II, scopoletin, and atractylenolide I). Capillary voltage and cone voltage can affect the ion response and may induce in-source decay. For capillary voltage varying between 1.0 and 3.0 kV, the peak areas for most of the index components positively correlated with the increasing of capillary voltage, and the setting of 3.0 kV could yield the best ion response and good repeatability through the triplicate operations. Histograms showing the variation tendencies of the ion response for five index components among five different levels of cone voltage (20–100 V) are given in Fig. 1B. A clear “increasing and decreasing” trend was observed, and the best ion response was obtained at cone voltage of 40 V or 60 V. As the highest precursor ion peak areas for more components could be obtained at cone voltage 60 V, in this work, we set the cone voltage in the positive mode at 60 V. In addition, the quality of the MS² spectra and the number of characterized components largely depend on the collision energy (Saurina and Sentellas, 2019). The Vion IM-QTOF mass spectrometer enables RCE, which can benefit the obtaining of more balanced MS² spectra (Jia et al., 2019; Zhang et al., 2019; Zuo et al., 2019) than the single collision energy. We set four groups of RCEs, including 10–40 eV, 20–50 eV, 30–60 eV, and 40–70 eV with an increase of 30 eV in each collision ramp, to compare the CID-MS² data of atractylenolide I, atractylenolide II, atractylon, biepiasterolide, and 8-β-ethoxyasterolid, as the representative compounds (Fig. S2). We could discern a remarkable tendency that increasing of RCE enabled more fragmentation of the precursors yielding more diversified product ions with medium mass, and then more low-mass product ions with higher intensity were obtained when RCE further ascended. To simultaneously retain the precursor ions and obtain rich fragments, we considered the choice of RCE 20–50 eV in this study.

3.2 Orthogonality, peak capacity and method validation for the established offline 2D-LC/IM-QTOF-MS system

Peak capacity and orthogonality are the core indicators for evaluating the separation performance of a 2D-LC system (Pirok, Stoll and Schoenmakers, 2019). In this work, orthogonality evaluation was based on a series of asterisk equations established by Camenzuli and Schoenmakers by using 54 index components (Camenzuli and Schoenmakers, 2014). Distribution degree of these compounds in a two-dimensional plane could be depicted by four lines (Fig. S3). The results showed the selected compounds were relatively evenly distributed around these four lines, with Z_-, Z₊, Z₁, and Z₂ calculated at 0.99, 0.93, 0.92, and 0.98, respectively, based on which orthogonality (A₀) of the established 2D-LC system was 0.91. In addition, the average peak widths at the baseline in ¹D (HILIC) and ²D (mixed mode of IEC and RPC) chromatography were 0.30 min and 0.19 min, respectively, and accordingly, we could deduce the peak capacity of 66 for ¹n_c and 157 for ²n_c. Thereby, the theoretical peak capacity of this 2D-LC system (n_2D) was 10362. The high orthogonality between the selected two dimensions of chromatography and largely improved peak capacity can testify the potency of the elaborated 2D-LC system, and thus lay foundation for the separation and identification of more components from A. macrocephala, in particular those minor ones.

To render a powerful analytical approach potentially with wide application, method validation was performed in terms of the intra-day/inter-day precision, stability, repeatability, and limit of detection (defined for an analyte at the lowest concentration the fragments generated can be used for structural characterization), considering its qualitative analysis property (Fu et al., 2018; Yang et al., 2016). Precision was evaluated by six consecutive injections on the first day and another three injections separately on the second and third days. The intra-day and inter-day precision varied 0.98–3.22% and 1.91–5.03% for the ¹D separation, and 1.01–5.00% and 3.82–13.03% for the ²D separation, respectively. Repeatability was assessed by analyzing six reduplicative samples of the sub-sample Fr. 1 (the first fractionation sample after the ¹D HILIC separation), and repeatability was less than 15% among six samples (Table S3–S5). LOD for six reference compounds (atractylenolide I, atractylenolide II, luteolin, scopoletin, β-sitosterol, apigenin) varied between 0.33 and 0.44 ng. The obtained results could indicate the sensitive and stabilized performance of the configured 2D-LC system suitable for the qualitative analysis of the multicomponents from A. macrocephala.

3.3 Comparison between HDMS^E and MS^E

The use of HDMS^E on the Vion IM-QTOF mass spectrometer offers high-definition full-scan spectrum. We compared the performance between the IM-enabled HDMS^E and IM-disabled MS^E in characterizing the multicomponents from the extract of A. macrocephala. Consequently, the differentiations in the data obtained by these two methods can be reflected in two aspects. First, the ion response for the components of A. macrocephala was significantly different, and the absolute abundance recorded by MS^E was much higher than that determined by HDMS^E (Fig. S4). For instance, the peak area of atractylenolide II determined by MS^E was 7.3 folds of that obtained by HDMS^E. We are aware that the additional IM separation can separate the ions with the same m/z but different charge state or shape, which may result in the reduction in ion response recorded by HDMS^E. Second, because of the enhanced separation ability facilitated by HDMS^E, the composition of ion species in full-scan MS¹ spectra became simpler (dubbed high-definition spectra), in contrast to MS^E. As shown in Fig. 2B, three components (t_R at 6.96, 10.95, and 12.99 min) got better resolved by HDMS^E showing much less interference. HDMS^E enabled the additional IM separation providing one new dimension of drift time, compared with MS^E. Accordingly, the CID-MS² spectra of single compounds obtained by HDMS^E (especially for atractylenolide II) seemed cleaner, making the data interpretation work easier to implement (Fig. S5). Because of these merits, we selected HDMS^E to acquire the fragmentation information of the components from A. macrocephala.

Close

Fig. 2

Comparison between the IM-enabled HDMS^E and IM-disabled MS^E in the positive ESI mode in profiling and charactering the multicomponents from Atractylodes macrocephala, using atractylenolides I (m/z 231.13), II (m/z 233.15), and III (m/z 249.15) as representatives. A-The base peak chromatograms recorded by MS^E (up) and HDMS^E (down); B-the full-scan MS¹ spectra of atractylenolides I–III recorded by MS^E (up) and HDMS^E (down); C-the 3D plot recorded by MS^E showing three peaks of atractylenolides I–III; D-2D-drift scope plot showing the additional separation of isomers by drift time.

Export to PPT

3.4 Systematic characterization of the multicomponents from A. macrocephala based on the positive HDMS^E data of eight fractionated samples by UNIFI-driven automatic peak annotation

Traditional methods used for the structural elucidation of TCM components based on the MSⁿ data rely on manual interpretation, yielding the results closely correlated to the professional knowledge and largely different among the researchers. Therefore, we sought to establish highly efficient methods to analyze the MSⁿ data that can generate reproducible characterization results. UNIFI^TM is a multifunctional bioinformatics platform enabling the automatic annotation of the high-resolution MS² data obtained by MS^E, DDA, HDMS^E, and HDDDA, etc. (Radchenko et al., 2020; Zhang et al., 2019). Standardized workflows by applying UNIFI to annotate the positive HDMS^E data were established in this work for the primary structural elucidation generating the lists of “Identified Components” and “Unknown Components”, while a confirming process based on the MS² fragmentation pathways and retention characteristics was conducted which generated the final characterization list. The first-dimensional HILIC separation of the A. macrocephala extract yielded eight fractionated samples, and from Fig. 3 we were able to deduce that the multicomponents of A. macrocephala got enriched and numerous minor peaks were newly exposed. By analyzing the HDMS^E data of these eight fractionated samples, as a result, we could identify or tentatively characterize a total of 251 components as given in Table S6, including 115 sesquiterpenoids, 90 polyacetylenes, 11 flavonoids, 9 benzoquinones, 12 coumarins, and 14 others. We emphasized and discussed the structural elucidation of the sesquiterpenoids, coumarins, flavonoids, and polyacetylenes, four subclasses of the major bioactive ingredients of A. macrocephala (Zhu et al., 2018).

Close

Fig. 3

Illustration for the eluate collection by the first-dimensional HILIC separation of the total extract of Atractylodes macrocephala and the fractionation effect by comparing the positive-mode HDMS^E spectra of the total extract and eight fractionated samples.

Export to PPT

3.4.1

3.4.1 Sesquiterpenoids

Sesquiterpenoids are one of the most abundant and major bioactive constituents in A. macrocephala (Yang et al., 2016). The sesquiterpenoids (e.g. atractylenolides) have been extensively investigated and considered to have promising medicinal potential. Atractylenolides I–III are three representative compounds of this type, which are derived from atractylones by oxidative ring opening and dehydration. In this section, atractylenolide I (compound 190#, t_R 12.99 min) was utilized as a reference to discuss the CID-MS² fragmentation of sesquiterpenoids. In the MS¹ spectrum, it gave the protonated precursor ion at m/z 231.1385, and the main cleavages occurred on rings A and C (Fig. 4). Fragmentation on the protonated molecule could open the lactone ring by eliminating H₂O producing a fragment at m/z 213.1278, which was able to lose CO giving the product ion observed at m/z 185.1329. What’s more, the fragments of m/z 155.0859, 141.0703, 128.0624, and 115.0547 were obtained due to the additional cleavages of ring A on the ion m/z 185. The left section of ring B also generated the product ions at m/z 91.0543, 77.0387, and 65.0388, in the low-mass spectral region. The compound 248# (t_R 22.37 min) produced a protonated precursor ion at m/z 463.2854 (C₂₀H₃₉O₄⁺, mass error: 1.90 ppm), the CID-MS² of which could produce a series of abundant product ions at m/z 447.2905, 231.1388, 185.1335, 163.0762, 143.0859, and 105.0705, which may be due to the eliminations of CH₄, bimolecular fragmentation, and the loss of C₂H₄ after the breaking of ring C and ring A. These CID features were in accordance with the characteristic fragmentation pathways because of the lactone cleavage, and accordingly, we speculated compound 248# was dimerized by two atractylenolide II, tentatively characterized as biatractylenolide/biepiasterolide (Fig. 4) (Ma, Dou and Tian, 2020). In addition, compounds 105# (t_R 6.96 min), 168# (t_R 10.95 min), 191# (t_R 13.00 min), and 227# (t_R 18.04 min), were identified as atractylenolide III, atractylenolide II, atractylodin, and atractylon, respectively, as a result of the comparison with their retention time, precursor ions, and the fragmentation behavior with those of the reference compounds. Characterization of the other unknown sesquiterpenoids from A. macrocephala was mainly based on the positive CID-MS² data analysis and searching against the established in-house library.

Close

Fig. 4

Illustration for the annotation of the positive CID-MS² spectra of representative sesquiterpenoids (compounds 190# and 248#) and coumarins (compounds 74# and 114#).

Export to PPT

3.4.4

3.4.4 Polyacetylenes

Polyacetylene is one class of characteristic components in A. macrocephala, for which the double bond is mainly of the trans-configuration. For polyacetylenes, CID-MS² could produce multiple product ions by the neutral eliminations of CO, H₂O, HCOOH, and CH₃COOH, etc., together with the cleavages between the unsaturated olefinic and acetylenic bonds (Li et al., 2017). A total of 90 polyacetylene compounds were tentatively characterized, and most of them contained the 14-C skeleton of 2,8,10-trien-4,6-diyn-1,12,14-triol. Moreover, the hydroxyl group at C-12 /C-14 of the core structure was further esterified with acetic acid, 3-methylbut-2-enoic acid, and 3-methylbutanoic acid, etc.. In this study, we illustrated the characterization of four compounds (81#, 127#, 155#, and 189#) belonging to this structure subclass. In detail, the protonated ion of compound 81# (t_R 4.87 min) was observed at m/z 233.1178, which, upon CID-MS², yielded abundant product ions at m/z 215.1086, 187.0757, 152.0628, 141.0705, 128.0626, 115.0546, 105.0703, and 91.0544, respectively (Fig. 5), which were attributed to the loss of OH, C₂H₄O, C₂H₅O₃, C₃H₇O₃ C₄H₈O₃, C₅H₉O₃, C₆H₇O₃, and C₇H₈O₃, from the precursor ion. Similarly, in the case of compound 189# (t_R 12.74 min), in addition to the protonated molecule at m/z 401.1971, characteristic ions at m/z 328.1673 ([M + H–C₃H₄O₂]⁺), 313.1440 ([M + H–C₄H₇O₂]⁺), 267.1024 ([M + H–C₆H₁₃O₃]⁺), 205.0502 ([M + H–C₁₂H₁₉O₂]⁺), 165.0707 ([M + H–C₁₀H₁₉O₆]⁺) and 91.0546 ([C₇H₇]⁺), were observed. Based on the MS information, compounds 81# and 189# were tentatively characterized as 12-hydroxytetradeca-trien-4,6-diyn-1-ol (or its isomer) and 12-methylbutyryltetradeca-trien-4,6-diyne-1,14-diacetate (or its isomer), respectively (Chen, 1987; Yao and Yang, 2014). The protonated precursor ion of compound 155# (t_R 9.97 min) at m/z 399.1603 was indicative of one molecule of H₂ less than compound 189#. Very similar fragmentation pathways were observed for compound 155# producing diverse product ions of m/z 339.1603, 327.1588, 265.1236, 165.0703, and 91.0545. Therefore, compound 155# was tentatively characterized as 12-senecioyloxytetradeca-trien-4,6-diyne-1,14-diacetate (or its isomer). The MS/MS fragment ions of compound 127# (t_R 8.34 min, [M + H]⁺ at m/z 359.1861) involved m/z 298.0890 ([M + H–C₄H₁₂]⁺), 267.1388 ([M + H–C₅H₁₁O₂]⁺), 231.1385 ([M + H–C₆H₇O₃]⁺), and 181.1019 ([C₇H₁₃O₅]⁺). Based on the structure of compound 189# and the fragment information, it could be inferred that the acetyl group at C-1 of compound 127# should be replaced by OH. Therefore, we were able to putatively characterize compound 127# as 14-acetoxy-12-methylbutyryltetradeca-trien-4,6-diyn-1-ol (Yao and Yang, 2014).

Close

Fig. 5

Illustration for the annotation of the positive CID-MS² spectra of representative polyacetylenes (compounds 81# and 189#).

Export to PPT

MS-based rapid chemical elucidation could primarily delineate the chemical composition of the medicinal plants. However, due to the lack of sufficient reference compounds, a major proportion of these compounds were only putatively characterized. We shall continue to perform the phytochemical research aiming to further validate the compounds identification results and support the bioactivity screening.