Chinese Herbal Medicines xxx (xxxx) xxx

Contents lists available at ScienceDirect

Chinese Herbal Medicines

journal homepage: www.elsevier.com/locate/chmed

Original Article

An integrated strategy toward comprehensive characterization and

quantification of multiple components from herbal medicine: An
application study in Gelsemium elegans
Meng-ting Zuo a,1, Yan-chun Liu a,1, Zhi-liang Sun a,b, Li Lin a, Qi Tang c, Pi Cheng c, Zhao-ying Liu a,b,c,⇑
a
College of Veterinary Medicine, Hunan Agricultural University, Changsha 410128, China
b
Hunan Engineering Technology Research Center of Veterinary Drugs, Hunan Agricultural University, Changsha 410128, China
c
Hunan Key Laboratory of Traditional Chinese Veterinary Medicine, Hunan Agricultural University, Changsha 410128, China

a r t i c l e i n f o a b s t r a c t

Article history: Objective: To develop a powerful integrated strategy based on liquid chromatography coupled with mass
Received 28 February 2020 spectrometry (LC-MS) systems for the comprehensive characterization and quantification of multiple
Revised 11 June 2020 components of herbal medicines.
Accepted 16 June 2020
Methods: Firstly, different mobile phase additives, analysis time, and MS acquisition modes were orthog-
Available online xxxx
onally tested with liquid chromatography coupled with quadrupole time-of-flight mass spectrometry
(LC-QTOF/MS) in order to detect as many components of Gelsemium elegans as possible with high peak
Keywords:
intensity. Secondly, several data mining strategies, including database searching, diagnostic ion filtering
Gelsemium elegans (Gardn. & Champ.) Benth
herbal medicine
and neutral loss filtering, were utilized to perform chemical profiling. Subsequently, this study focused on
LC-MS the quantification and validation of the performance of a liquid chromatography-triple mass spectrom-
mass spectrometry etry (LC-QqQ/MS) assay based on derivative multiple reaction monitoring (DeMRM).
Results: A total of 147 components from G. elegans were characterized, among them 116 nontarget com-
ponents were reported for the first time. A sensitive and reproducible LC-QqQ/MS method was success-
fully developed and validated for the simultaneous relative quantification of 41 components of G. elegans.
This LC-QqQ/MS method was then applied to compare the contents of components in the roots, stems
and leaves.
Conclusion: The present integrated strategy would significantly contribute to chemical studies on herbal
medicine, and its utility could be extended to other research fields, such as metabolomics, quality control,
and pharmacokinetics.
Ó 2020 Tianjin Press of Chinese Herbal Medicines. Published by ELSEVIER B.V. This is an open access
article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction in many different concentrations (Yan et al., 2013; Mustafa et al.,

2018; Tan et al., 2018). Due to their high diversity, the comprehen-
There is rising global interest in herbal medicines as a promis- sive characterization and quantification of multiple components of
ing source for new drug discovery and development (Shi et al., herbal medicines remain a great challenge and bottleneck.
2018; Wang, Shi, & Wang, 2018). The use of herbal medicines in Currently, liquid chromatography coupled with various types of
the clinic also requires the elucidation of as many components as mass spectrometers, especially hybrid mass spectrometers, has
possible to ensure improved quality control and clinical outcomes. been increasingly accepted as the predominant platform for the
A significant feature of herbal medicine is the number of chemical global analysis of complex herbal medicines. Among these mass
components, as each herbal medicine contains hundreds or even spectrometry techniques, time-of-flight (TOF) mass spectrometry
thousands of different components, depending on the number of and liquid chromatography coupled with hybrid tandem mass
genes. These components are small molecules (<1500 Da) among spectrometry (LC-QTOF/MS) are expected to be the most powerful
many different chemical substance classes, which can be present tools for structurally characterizing complex components of herbal
medicines due to the high resolution of both precursor and product
⇑ Corresponding author at: College of Veterinary Medicine, Hunan Agricultural ions with these methods (Lei, Li, Cheng, Wang, & Meng, 2018;
University, Changsha 410128, China. Zhang et al., 2017b; Zuo et al., 2018). Both types of information
E-mail address: [email protected] (Z.-y. Liu). are important and reliable for identifying components. Previously,
1
These authors contributed equally to this work.

https://doi.org/10.1016/j.chmed.2020.06.002
1674-6384/Ó 2020 Tianjin Press of Chinese Herbal Medicines. Published by ELSEVIER B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Please cite this article as: Meng-ting Zuo, Yan-chun Liu, Zhi-liang Sun et al., An integrated strategy toward comprehensive characterization and quantifi-
cation of multiple components from herbal medicine: An application study in Gelsemium elegans, Chinese Herbal Medicines, https://doi.org/10.1016/j.
chmed.2020.06.002
Meng-ting Zuo, Yan-chun Liu, Zhi-liang Sun et al. Chinese Herbal Medicines xxx (xxxx) xxx

most reported methodologies have been limited to target compo- components of herbal medicines. The general flowchart for this
nents and mainly depended on the use of reference components integral strategy is shown in Fig. 1. First, different mobile phase
and/or comparison with literature data (Huo, Du, Sun, Dong, & additives, analysis time and MS acquisition modes were orthogo-
Wang, 2018; Li et al., 2015a). Considering that the reference com- nally tested to detect as many components of herbal medicines
ponents that can be obtained are limited and most components as possible with high peak intensities based on LC-QTOF/MS. Then,
contained in herbal medicines are unknown (nontarget compo- various data mining techniques, including database searching,
nents), these methods are apparently insufficient for the compre- diagnostic ion filtering and neutral loss filtering, were used to char-
hensive detection and characterization of the complicated acterize nontarget components under the optimized conditions.
components of herbal medicines. Furthermore, to determine the relative concentrations of the char-
The characterization of nontarget components of herbal medici- acterized components, we quantified and validated the perfor-
nes is never an easy task (Chingin, Makarov, Denisov, Rebrov, & mance of an LC-QqQ/MS method based on derivative multiple
Zubarev, 2014; Xue et al., 2016; Samanipour, Reid, Bæk, & reaction monitoring (DeMRM). A single calibration curve was
Thomas, 2018). Great challenges remain in optimizing the detec- developed to calculate the relative contents of multiple compo-
tion of herbal medicines and processing and mining data from nents without authentic standards for each component.
the complex information obtained. Recently, corresponding strate- The proposed strategy was demonstrated on the herbal medi-
gies based on diagnostic fragment ions, key ion filtering, and mass cine Gelsemium elegans (Gardn. & Champ.) Benth, a species of flow-
defect filtering have been developed to characterize nontarget ering plant in Loganiaceae family, is known as a toxic plant. In
components from complex mixtures (Hao et al., 2008; Cai et al., China, it is known as Gouwen, Dachayao or Duanchangcao
2014; Zhang et al., 2016; Shi et al., 2017). However, it is worth not- (Ornduff, 1970) and has been used as a traditional Chinese medi-
ing that these studies mainly focused on qualitative aspects (Chen cine (TCM) for the treatment of rheumatoid arthritis, neuropathic
et al., 2016; Ma et al., 2016). Moreover, few studies on quantifying pain, spasticity, skin ulcers and cancer for many years. The bioac-
the amounts of multiple components or monitoring the chemical tive components of G. elegans have attracted much attention from
variations that are frequently present in different parts of herbal chemists, pharmacologists and toxicologists in recent years due to
medicines during different seasons have been conducted due to their multiple biological effects, such as anti-inflammatory,
the high diversity of components and lack of standards (Li et al., immunomodulating, analgesic, anxiolytic, antitumor, and neuro-
2015b; Yang et al., 2013). Therefore, the development of more pathic pain-relieving properties (Ling et al., 2014; Liu et al.,
comprehensive and effective strategies for the simultaneous char- 2013; Meyer, Boujedaini, Patte-Mensah, & Mensah-Nyagan, 2013;
acterization of nontarget components and quantification of multi- Y. Xu et al., 2012a; Y.-K. Xu et al., 2012b; Zhang et al., 2015b). To
ple components of herbal medicines in the absence of reference date, a total of 121 alkaloids, 25 iridoids and a number of other
substances is imperative. components from a wide spectrum of secondary metabolite classes
To address this challenge, the present study provides an inte- have been isolated from G. elegans and characterized (Jin et al.,
grated strategy to simultaneously identify target and nontarget 2014; Liu et al., 2017b; Yamada, Kitajima, Kogure,
components and monitors the concentrations of multiple Wongseripipatana, & Takayama, 2011; Zhang et al., 2017a).

Fig. 1. Workflow of the integrated strategy toward comprehensive characterization of nontarget components and derivative multiple reaction monitoring for multiple
components quantification of herbal medicine using LC-QTOF/MS and LC-QqQ/MS.

2
Meng-ting Zuo, Yan-chun Liu, Zhi-liang Sun et al. Chinese Herbal Medicines xxx (xxxx) xxx

Previous studies have focused mainly on the isolation and purifica- gradient analysis times and MS acquisition mode (see Supporting
tion of these components. Only our laboratory has recently pub- information, Table S2). The gradient elution for 40 min was as fol-
lished an analytical strategy for the characterization and lows: 0–2 min, 10% B; 2–7 min, 10% B to 15% B; 7–20 min, 15% B to
structural analysis of target components from G. elegans by using 35% B; 20–30 min, 35% B to 90% B; 30–33 min, 90% B; 33.01–
LC-QTOF/MS based on the use of accurate mass databases com- 40 min, 10% B. The gradient elution for 50 min was as follows:
bined with MS/MS spectra (Liu et al., 2017b; Xiao, Huang, Sun, & 0–2 min, 10% B; 2–7 min, 10% B to 15% B; 7–30 min, 15% B to
Liu, 2017; Yang et al., 2018a). However, many nontarget compo- 45% B; 30–40 min, 45% B to 90% B; 40–43 min, 90% B; 43.01–
nents in G. elegans have not been characterized. There are few 50 min, 10% B. The gradient elution for 60 min was as follows:
reports on quantification methods for G. elegans formulations, 0–2 min, 10% B; 2–7 min, 10% B to 15% B; 7–40 min, 15% B to
and the number of analytes quantified in these reports is at most 55% B; 40–50 min, 55% B to 90% B; 50–53 min, 90% B; 53.01–
three (Hu et al., 2017; Liang et al., 2010; Wang, Wen, & Meng, 60 min, 10% B.
2018; Yang et al., 2018b; Zhang et al., 2015a). We developed a Mass spectrometric detection was performed in positive elec-
practical and reliable high-performance liquid chromatography- trospray ionization (ESI) mode. The operating parameters were as
ultraviolet detector (HPLC-UV) method for fingerprint analysis follows: fragmentor voltage, 30 V; capillary voltage, 3500 V; dry
using two major components, gelsemine and koumine. The results gas temperature, 300 °C; sheath gas temperature, 350 °C; dry gas
showed that at least seven relatively major components present in (N2) flow rate, 9 L/min; sheath gas flow rate, 11 L/min; nebulizer,
G. elegans may be useful for its quality control (Liu et al., 2017a). 35 psi; VCap, 4000; nozzle voltage, 1000 V; fragmentor, 175; Skim-
Using the methodology established in the study, a total of 147 mer1, 65. Data were acquired in a mass range of m/z 50–1000. Auto
components were characterized from G. elegans, and among these MS/MS mode was utilized to obtain abundant structural informa-
components, 116 nontarget components were reported for the first tion without knowledge of the sample. The collision energy (CE)
time. The simultaneous quantification of 41 components of G. ele- was set at 10, 20, and 30 V.
gans was achieved using DeMRM mode. The instrument performed internal mass calibration automati-
cally using an automated calibrant delivery system. The calibration
solution contained internal reference masses of m/z 121.0508 and
2. Materials and methods
922.0098 in positive mode. All of the data acquisition was con-
trolled by Agilent MassHunter workstation software (version
2.1. Chemicals and materials
B.01.03 Build 1.3.157.0 2).
The reference components gelsemine (>98%), koumine (>98%),
koumidine (>98%) and gelsenicine (>98%) were purchased from 2.4. LC-QqQ/MS and MS/MS conditions
Shanghai Jiwei Biochemical Technology Co., Ltd. (Shanghai, China).
HPLC-grade acetonitrile and methanol were purchased from Merck Analysis was performed on an Agilent series 1290 Infinity HPLC
(Darmstadt, Germany). Ammonium formate, ammonium acetate, instrument coupled with an Agilent 6460 QqQ/MS mass spectrom-
formic acid and acetic acid were purchased from SIGMA- eter (Agilent Technologies, Santa Clara, CA, USA). Samples were
ALDRICH (USA). Deionized water was purified by a Milli-Q water separated on a Waters C18 column (3.5 lm, 4.6 mm  150 mm).
purification system (Bedford, MA, USA). The other chemicals used The flow rate was 0.3 mL/min, and the column temperature was
were of analytical grade. All reference components were weighed, maintained at 30 °C.
dissolved in methanol and then diluted to appropriate concentra- Auto MS/MS analysis was performed in both scan mode and
tions for LC-QTOF/MS and LC-MS/MS analysis. multiple reaction monitoring (MRM) mode. The MS/MS switching
time and scan speed were set at < 2 ms and 5200 amu/s, respec-
2.2. Plant materials and sample preparation tively. The operating parameters were set as follows: scan range,
m/z 50–1000; capillary temperature, 350 °C; capillary voltage,
A total of nine samples including G. elegans roots, stems and 3000 V; dry gas (N2) flow rate, 12 L/min; nebulizer pressure, 40
leaves from three different periods were collected from Guangxi psi. Full mass spectra were recorded at a mass resolving power
Province in China (see Supporting information, Table S1). The sam- of 0.7 Da (full width at half maximum).
ples were authenticated by the author Qi Tang and stored at Hunan
Key Laboratory of Traditional Chinese Veterinary Medicine. The 2.5. Relative quantitation of multiple components
samples were dried and then ground into powder. According to
our optimized method, a 1 g aliquot of the powder was extracted In the present study, DeMRM was used to obtain the mass
twice by ultrasonication with 80% ethanol (1:25) for 0.5 h at responses of targeted analytes, and a koumidine calibration curve
60 °C. The extraction solution was combined for filtration, and was employed instead of authentic standards to determine the
1 mL of the filtered solution was evaporated and dissolved in concentrations of multiple components of G. elegans. The method
1 mL of acetonitrile-ammonium acetate (1:4, volume percent). was first validated with four representative components (gelsem-
Then, the solution was filtered through a 0.22 lm membrane ine, koumine, koumidine and gelsenicine) in terms of linearity, lim-
before use. A 5 lL aliquot was injected for analysis. its of detection and quantification (LOD and LOQ), and intraday and
interday precision. LOD and LOQ were determined at a signal-to-
2.3. LC-QTOF/MS conditions noise ratio of approximately 3 and 10, respectively. Intraday preci-
sion was determined by analyzing six replicates of each represen-
Analysis was performed on an Agilent series 1290 Infinity HPLC tative component at 10 ng/mL and 100 ng/mLl within one day,
instrument coupled with an Agilent 6530 Q-TOF mass spectrome- while interday precision was assessed on three consecutive days.
ter (Agilent Technologies, Santa Clara, CA, USA). Samples were sep- Then, the DeMRM method was carried out with six G. elegans sam-
arated on a Waters C18 column (3.5 lm, 4.6 mm  150 mm). The ples to demonstrate that the whole method was reliable and
flow rate was 0.3 mL/min, and the column temperature was main- acceptable for multiple component quantification. Six samples
tained at 30 °C. To detect as many chemical components as possi- were prepared separately according to the above method and ana-
ble with high peak intensities, an L9 (3) orthogonal array was used lyzed to measure repeatability. To assess stability, the samples
to examine the influence of difference mobile phase additives, were exposed to ambient temperature and then analyzed at 0, 2,
3
Meng-ting Zuo, Yan-chun Liu, Zhi-liang Sun et al. Chinese Herbal Medicines xxx (xxxx) xxx

4, 8, and 24 h. The relative concentrations of multiple components analytical strategy was to extract components from raw acquisi-
of G. elegans were calculated from the koumidine calibration curve. tion data with the ‘‘Find by Auto MS/MS” function. The mass match
tolerance was set to 0.05 m/z, and the peak abundance was set to
1000. Following this step, the target components were matched
3. Results and discussion
with personal database searching and characterized according to
our published analytical approach (Liu et al., 2017b).
3.1. Optimization of LC-QTOF/MS conditions
The precise mass of precursor ions (within ± 5 ppm) found to
match data in the MS database of Gelsemium would give the exact
This study examined different mobile phase additives, analysis
element compositions and the possible known structure of the
times and MS acquisition modes, which could help to significantly
components. Structural characterization of the components was
improve the information obtained regarding the content of nontar-
achieved on the basis of determining the accurate mass and frag-
get components in herbal medicines
mentation behavior of the product ions. A total of 31 components
The results of the detection of all components in G. elegans are
were matched with the personal database and characterized. Thir-
shown in Fig. 2. These results showed that there were great differ-
teen of them were gelsedine-type alkaloids, which was the largest
ences in the number of G. elegans components detected under differ-
class in G. elegans. They were components 28, 40, 75, 83, 94, 100,
ent test conditions. As shown in Fig. 2A, among the experimental
109, 110, 118, 124, 126, 131 and 134. Components 99, 117, 119,
groups, the number of components detected in the D and I groups
128 and 143 were target sarpagine-type alkaloids obtained after
was the smallest, and the number of components detected in the C
matching. The [M + H]+ ions of components 133 and 137 were
and E groups was the highest. Relative to the number detected with
m/z 355.2025 and 371.1974, respectively, which corresponded to
the other two mobile phases, the total number of compounds
humantenine-type alkaloids. Components 53 and 95 had masses
detected in groups A, B and C with a 0.1% formic acid–water mobile
of m/z 325.1917 and 307.1812, respectively, and were character-
phase was significantly large. The number of compounds detected in
ized as koumine-type alkaloids. Components 63 and 72 were target
the corresponding groups C, F, and I with a consistent analysis time
gelsemine-type alkaloids obtained after matching. Four kinds of
of 60 min was not significantly different from that detected in
iridoid components were detected by matching with Gelsemium
groups B, E, and H, corresponding to an analysis time of 50 min. As
database: 30, 57, 91 and 92. Only two phenolic acids were detected
shown in Fig. 2A, the component peaks were mainly concentrated
in G. elegans after matching with the MS database.
in the first 30 min of analysis, so the number of components did
not change by extending the analysis time over 50 min. Fig. 2B
showed that among the experimental groups, the peak intensities
3.3. Characterization of nontarget components in G. elegans
of the components in G. elegans in groups B, D, and I were the highest,
and the intensities of the components in groups E and F were the
The second step was to characterize the nontarget components
lowest. These results suggested that compared to the other two
when a component fails to match the information in our personal
acquisition modes, the auto MS/MS mode not only had higher peak
database. Based on the fragmentation of the target components,
intensities but also was more efficient. What’s more, compared with
common fragment ions and neutral loss ions could be classified
ammonium acetate and ammonium formate as additives, an aque-
into families. According to this idea, some post-acquisition data
ous solution of 0.1% formic acid provided higher peak intensities.
mining procedures, including key ion filtering, diagnostic ion filter-
Thus, we recommend the application of a 50 min gradient, MS/MS
ing, neutral loss filtering and online database (Metlin and HMDB
mode and 0.1% formic acid to ensure better results.
public database) searching, were performed in this study (Qiao
et al., 2016). The structure of the nontarget components could be
3.2. Characterization of target components in G. elegans characterized by accurate MS/MS spectra and fragmentation com-
parisons. As a result, a total of 116 nontarget components were
The optimal LC-QTOF/MS conditions were used to acquire characterized. Table 1 summarized the exact mass, fragment ion,
information about G. elegans sample No. S1. The first step of this and retention time for the characterized components of Gelsemium

Fig. 2. Effect of mobile phases, analysis time and mass spectrometry modes on number of detected features (A) and sum of intensities (B) for G. elegans components. A, 0.1%
FA, 40 min of gradient separation and MS mode; B, 0.1% FA, 50 min of gradient separation and auto MS/MS mode; C, 0.1% FA, 60 min of gradient separation and target MS/MS
mode; D, 5 mmol/L NH4Ac, 40 min of Gradient separation and auto MS/MS mode; E, 5 mmol/L NH4Ac, 50 min of gradient separation and target MS/MS mode; F, 0.1% 5 mmol/L
NH4Ac, 60 min of Gradient separation and MS mode; G, 10 mmol/L NH4F, 40 min of Gradient separation and target MS/MS mode; H, 10 mmol/L NH4F, 50 min of gradient
separation and MS mode; I, 10 mmol/L NH4F, 60 min of gradient separation and auto MS/MS mode.

4
 Table 1

Meng-ting Zuo, Yan-chun Liu, Zhi-liang Sun et al.

Retention time, accurate mass, fragment ions of target and non-target compounds.

No. t/R(min) [M + H]+ Fragment ions Formula Tentative characterization Classification

1 5.4 150.0921 132.0807,117.0574,77.0383 C9H11NO Unknown Amino acid
2 5.5 521.2131 490.1948,459.1777,441.1624,403.1448,359.1604,328.1423,297.1249,279.1138,269.1286,256.0980,108.0806 C25H32N2O10 Unknown Gelsedine-type
3 5.6 537.2074 506.1900,489.1864,475.1775,457.1617,375.1521,344.1369 C25H32N2O11 Unknown Gelsedine-type
4 5.8 229.1551 170.0785,142.0858,132.0788,114.0550,96.0812,70.0659,58.0667 C11H20N2O3 Pro Leu or Pro Ile Amino acid
5 6.0 236.1323 218.1163,150.0921,132.0811,117.0593,87.0449 C13H17NO3 Unknown Amino acid
6 6.4 195.0626 177.0483,149.0596,121.0621,91.0552,77.0366 C10H10O4 Ferulic acid Phenolic acids
7 6.4 213.0735 195.0627,177.0517,165.0518,149.0568,131.0466,121.0621,103.0519,91.0523,77.0370,69.0317,57.0323 C10H12O5 Gelsemide Iridoids
8 6.4 231.0841 216.1015,195.0604,177.0518,149.0568,121.0625,107.0465,93.0672,91.0521,77.0365,57.0321 C10H14O6 Geleganoid A/GRIR-1 Iridoids
9 6.6 229.1536 195.0628,177.0517,142.0854,121.0626,70.0645 C11H20N2O3 Pro Leu or Pro Ile Amino acid
10 6.8 199.0949 135.0774,121.0622,117.0865,109.0641,91.0533,79.0540,67.0542,55.0540 C10H14O4 9-DeoxyGRIR-2 Iridoids
11 6.8 217.1057 163.0705,135.0786,109.0640,91.0536,79.0538,67.0535,55.0175 C10H16O5 7-Hydroxygelsemiol/9-Hydroxygelsemiol Iridoids
12 6.8 239.0877 177.0541 C12H14O5 Hydroxyl of ferulic acid ethyl ester Phenolic acids
13 7.0 408.1505 213.0748,195.0648,151.0738 C17H21N5O7 Unknown Nucleoside
14 7.3 187.0959 109.0639,105.0698,95.0854,91.0542,81.0701,79.0544,77.0385,67.0553 C9H14O4 7-Deoxygeleganoid D/9-Deoxygeleganoid D Iridoids
15 7.7 217.1073 135.0802,117.0701,109.0650,91.0540,79.0549,67.0547,55.0186 C10H16O5 Isomer of 7-Hydroxygelsemiol/9- Iridoids
Hydroxygelsemiol
16 7.8 239.0866 177.0534,84.0794,58.0655 C12H14O5 Hydroxyl of ferulic acid ethyl ester Phenolic acids
17 8.2 422.1295 325.0941,277.0637,213.0759,151.0761,133.0635 C17H19N5O8 Unknown Nucleoside
18 8.3 236.1285 150.0906,132.0791,117.0553,87.0447 C13H17NO3 Unknown Amino acid
19 8.6 213.0759 195.0636,177.0543,149.0583,131.0477,121.0630,91.0546,77.0398,57.0344 C10H12O5 Isomer of GEIR-1 Iridoids
20 8.8 187.0968 123.0804,105.0708,95.0855,91.0564,81.0691,77.0396,67.0550,55.0558 C9H14O4 7-Deoxygeleganoid D /9-Deoxygeleganoid D Iridoids
21 8.9 217.1059 135.0848,117.0683,91.0559,81.0576 C10H16O5 Isomer of 7-Hydroxygelsemiol/9- Iridoids
Hydroxygelsemiol
22 8.9 407.1820 376.1611,345.1455,331.1300,166.0879 C20H26N2O7 Tihydroxygelsemicine Gelsedine-type
5

23 9.2 422.1300 395.1598,213.0757,195.0638,175.0232,151.0750,129.0183,73.0295 C17H19N5O8 Unknown Nucleoside

24 9.4 187.0966 123.0786,109.0650,105.0707,95.0852,91.0536,81.0698,77.0397,67.0548,55.0548 C9H14O4 Isomer of 7-Deoxygeleganoid D /9- Iridoids
Deoxygeleganoid D
25 9.6 199.0937 135.0776,117.0675,115.0515,105.0667,91.0521,79.0522,67.0519,55.0162 C10H14O4 9-Deoxygeleganoid F Iridoids
26 9.6 217.1045 163.0717,135.0781,117.0676,109.0624,105.0669,91.0521,79.0525, 67.0526,55.0164 C10H16O5 7-Hydroxygelsemiol/9-Hydroxygelsemiol Iridoids
27 9.6 239.0855 177.0447,105.0674 C12H14O5 Hydroxyl of ferulic acid ethyl ester Phenolic acids
28 9.8 359.1578 328.1395,311.1366,297.1212,279.1105,269.1259,256.0960,228.0718,124.0740,108.0791,96.0790 C19H22N2O5 11,14-Dihydroxygelsenicine Gelsedine-type
29 9.9 195.0642 177.0546,121.0623.91.0534,77.0375 C10H10O4 Isomer of ferulic acid Phenolic acids
30 9.9 213.0754 195.0644,177.0511,149.0579,121.0626,91.0532,79.0529,77.0372,57.0326,55.0171 C10H12O5 GEIR-1 Iridoids
31 10.1 375.1553 344.1359,327.1335,313.1175,295.1072,279.1139,267.1137,124.0756 C19H22N2O6 11,14,15-Trihydroxygelsenicine Gelsedine-type
32 10.7 187.0969 133.0695,123.0822,109.0638,105.0706,95.0847,91.0544,81.0702,79.0539,77.0386,67.0542,55.0541 C9H14O4 Isomer of 7-Deoxygeleganoid D /9- Iridoids
Deoxygeleganoid D
33 10.7 417.1649 386.1485,355.1260,166.0854 C21H24N2O7 Isomer of 14-Acetoxy-dihydroxygelsenicine Gelsedine-type
34 10.8 213.0738 177.0499,149.0581,121.0628,103.0522,91.0529,77.0374,67.0533,55.0532 C10H12O5 Isomer of gelsemide Iridoids
35 10.9 395.1603 364.1396,347.1355,333.1212,307.1060,277.0950,246.1106,228.0630,132.0789 C22H22N2O5 Unknown Gelsedine-type
36 11.4 373.1768 C20H24N2O5 14,15-Dihydroxyrankinidine Humantenine-

Chinese Herbal Medicines xxx (xxxx) xxx

342.1581,327.1357,311.141,271.1448,180.1024,122.0965,108.0810
type
37 11.5 339.1707 293.1616,236.1063,195.0677 C20H22N2O3 Gelsemine oxide Gelsemine-type
38 11.6 408.1510 213.0765,195.0636,177.0518,167.0732,149.0606 C17H21N5O7 Unknown Nucleoside
39 11.8 325.1913 307.1808,281.1644,238.1268,220.1179,194.0997,150.0908,130.06481,70.0655 C20H24N2O2 19-(S)-Hydroxydihydrokoumine/19-(R)- Coumine-type
Hydroxydihydrokoumine
40 11.9 329.1499 311.1387,269.1300,168.1021,148.0381,122.0949,108.0809 C18H20N2O4 Nb-Methylgelsedilam Gelsedine-type
41 11.9 391.1869 360.1639,329.1497,311.1414,168.1018 C20H26N2O6 Dihydroxy gelsemicine Gelsedine-type
42 12.2 361.1763 330.1564,301.1184,283.1101,162.0545,84.0812 C19H24N2O5 11,14-Dihydroxy-dihydrogelsenicine Gelsedine-type
43 12.3 422.1293 213.0759,195.0642,177.0538,149.0578,121.0653 C17H19N5O8 Unknown Nucleoside

(continued on next page)

 Meng-ting Zuo, Yan-chun Liu, Zhi-liang Sun et al.
Table 1 (continued)

No. t/R(min) [M + H]+ Fragment ions Formula Tentative characterization Classification

44 12.4 323.1757 293.1646,279.1496,198.0908,138.0907,70.0606 C20H22N2O2 Isomer of gelsemine Gelsemine-type
45 12.4 377.1717 346.1527,317.1131 ,299.1021,271.1068,232.1196,203.0812,138.0552 C19H24N2O6 11,14,15-Trihydroxy-dihydro gelsenicine Gelsedine-type
46 12.7 339.1709 323.1703,252.1005,70.0657 C20H22N2O3 Gelsemine oxide Gelsemine-type
47 12.9 311.1760 293.1654,252.1387 C19H22N2O2 Hydroxy of koumidine Sarpagine-type
48 12.9 385.1769 354.1602,339.1356,311.1336,134.0956,122.0995 C21H24N2O5 15-Hydroxyhumantenox-enine Humantenine-
type
49 13.2 571.2292 540.2101,509.1940,481.1919,359.1604,328.1422,297.1245, C29H34N2O10 Diydroxyl of gelseiridone/gelseganine D Gelsedine-type
50 13.4 527.2389 508.2282,496.2177,479.2188,339.1347 C28H34N2O8 Unknown Gelsedine-type
51 13.5 341.1866 323.1742,311.1156,297.1663,238.1224 C20H24N2O3 19S-Hydroxydihydrogelsemine Gelsemine-type
52 13.5 391.1872 360.1683,329.1499,281.1284,238.0866,148.0396 C20H26N2O6 Dihydroxy gelsemicine Gelsedine-type
53 13.7 325.1916 307.1810,281.1661,238.1226,220.1116,194.0995,70.0660 C20H24N2O2 19-(S)-Hydroxydihydrokoumine/19-(R)- Koumine-type
Hydroxydihydrokoumine
54 13.9 323.1750 307.1801,238.1236,220.1118,194.1006,150.0913,122.0960 C20H22N2O2 Koumine N-oxide Koumine-type
55 14.1 422.1300 197.0816,179.0702,161.0589,153.0889,135.0796,108.0656 C17H19N5O8 Nucleoside
56 14.2 339.1712 325.1546,297.1603,252.1007,210.0910,70.0657 C20H22N2O3 Gelsemine oxide Gelsemine-type
57 14.4 197.0817 151.0752,133.0639,115.0534,105.0703,91.0547,81.0705,79.0543,77.0383,67.0543 C10H12O4 7-Deoxygelsemide /9-Deoxygelsemide Iridoids
58 14.4 375.1558 344.1735,326.1625,313.1450,311.1388,299.1051,298.1672,122.0965,108.0809 C19H22N2O6 11,14,15-Trihydroxygelsenicine Gelsedine-type
59 14.5 321.1604 291.1520,236.1068,210.0932,178.0846,136.0753 C20H20N2O2 Gelebolines A Other types of
alkaloids
60 14.7 327.1711 311.1756,297.1581,291.1470,238.1229,135.0806 C19H22N2O3 3-Hydroxykoumidine Oxide Sarpagine-type
61 14.8 527.2396 509,2273,496.2207,479.2195,468.2315,339.1342 C28H34N2O8 Unknown Gelsedine-type
62 14.9 355.1031 328.1448,309.1591,163.0393,145.0286,135.0442,117.0338,89.0400 C16H18O9 1-O-Caffeoylquinic acid/4-O-Caffeoylquinic Phenolic acids
acid
63 15.1 341.1869 323.1759,311.1751,297.1605,281.1452,238.1221,158.0595,70.0656 C20H24N2O3 19R-Hydroxydihydrogelsemine Gelsemine-type
64 15.2 197.0807 179.0702,151.0745,133.0646,115.0539,105.0697,91.0544,81.0699,79.0547,77.0385,67.0550 C10H12O4 7-Deoxygelsemide /9-Deoxygelsemide Iridoids
65 15.2 373.1759 C20H24N2O5 11,14-Dihydroxyrankinidine Humantenine-
6

342.1570,325.1550,311.1394,293.1285,108.0808
type
66 15.3 375.1919 357.1798,326.1625,311.1388,297.1584,283.1448,198.1100,122.0965,108.0809 C20H26N2O5 14,15-dihydroxy-19,20dihydrorankinidine Humantenine-
type
67 15.4 387.1913 369.1850,356.1697,341.1448,325.1548,311.1734,194.1165,138.0899 C21H26N2O5 Dihydrogelegamine A Humantenine-
type
68 15.5 394.1343 341.0486,295.0510,197.0804,179.0705,161.0593,153.0910,135.0804,108.0658 C16H19N5O7 Unknown Nucleoside
69 15.6 343.1647 312.1465,295.1437,281.1281,264.1019,240.1012,212.0746,159.0673,108.080714,71.0734 C19H22N2O4 11-Hydroxygelsenicine Gelsedine-type
70 15.6 505.2169 474.1980,443.1818,425.1636,339.1511,371.1530,341.0567,325.1311,297.1600,240.1194 C25H32N2O9 Unknown Gelsedine-type
71 15.9 359.1606 328.1422,285.1235,95.0732 C19H22N2O5 Hydroxyl of gelseziridine Gelsedine-type
72 16.2 323.1753 293.1627,262.1222,236.1072,70.0658 C20H22N2O2 Gelsemine Gelsemine-type
73 16.2 424.1452 3397.0429,341.00501,197.0804,179.0698,161.0959,135.0805 C17H21N5O8 Unknown Phenolic acids
74 16.7 359.1600 328.1420,311.1234,297.1238,279.1125,251.1168,225.1020 C19H22N2O5 14,15-Dihydroxygelsenicine Gelsedine-type
75 16.8 343.1641 312.1473,281.1289,263.1181,253.1336,240.1033,212.0764,139.0990,124.0758,108.0811 C19H22N2O4 14-Hydroxygelsenicine Gelsedine-type
76 17.0 406.1351 378.1419,343.1626,197.0796,179.0695,161.0591,153.0888,135.0792,107.0848,81.0729 C17H19N5O7 Unknown Phenolic acids
77 17.2 339.1707 323.1508,312.1470,295.1451,281.1279,251.1339,108.0812 C20H22N2O3 Gelsemine N-oxide Gelsemine-type

Chinese Herbal Medicines xxx (xxxx) xxx

78 17.3 409.1756 378.1576,361.1555,347.1380,162.0904 C23H24N2O5 Unknown Gelsedine-type
79 17.4 375.1911 344.1728,313.1547,295.1430,253.1331,168.1019 C20H26N2O5 11-Hydroxygelsemicine Gelsedine-type
80 17.4 423.1558 392.1369,375.1340,361.1186,335.1069,305.0927,188.0714,176.0707,151.0633,107.0731 C23H22N2O6 Unknown Gelsedine-type
81 17.6 379.1655 348.1469,331.1440,317.1286,291.1147,289.1335,261.1034,212.0706,132.0807,107.0734 C22H22N2O4 Unknown Gelsedine-type
82 17.8 345.1818 314.1617,285.1230,267.1147,168.1011 C19H24N2O4 14-Hydroxygelsedine Gelsedine-type
83 17.8 401.1712 370.1526,339.1340,311.1398,295.1131,166.0864,154.0849 C21H24N2O6 14-Acetoxy-15-hydroxygelsenicine Gelsedine-type
84 17.9 361.1758 330.1571,301.1181,283.1078,271.1045,138.0544 C19H24N2O5 Isomer of 11,14-Dihydroxy-dihydro Gelsedine-type
gelsenicine
85 18.0 307.1802 293.1654,277.1697,251.1539,220.1113,138.0905,108.0806 C20H22N2O Isomer of koumine Koumine-type
86 18.0 343.1651 312.1460,295.1495,281.1281,165.1330,108.0806,95.0730 C19H22N2O4 Hydroxyl of gelsenicine Gelsedine-type
 Meng-ting Zuo, Yan-chun Liu, Zhi-liang Sun et al.
Table 1 (continued)

No. t/R(min) [M + H]+ Fragment ions Formula Tentative characterization Classification

87 18.0 449.1078 431.0937,413.0829,395.0740,377.0659,353.0643,329.0654,299.0547 C21H20O11 Orientine/iso-orientine Flavone
88 18.1 305.1653 277.1680,220.1117,162.0904,130.0650,120.0803,70.0655 C20H20N2O Dehydrokoumine Koumine-type
89 18.2 311.1755 293.1656,269.1647,251.1546,138.0918,108.0816 C19H22N2O2 Hydroxy of koumidine Sarpagine-type
90 18.2 428.2539 398.1746,339.0895,299.0594,118.0864 C24H33N3O4 Unknown Oil
91 18.3 183.1009 137.0908,119.0842,107.0851,93.071,91.0541,77.0387,67.0545 C10H14O3 GSIR-1 Iridoids
92 18.3 201.1118 147.0788,135.0788,119.0845,107.0843,93.0691,91.0536,81.0692,79.0538,77.0382, 67.0541,55.0180 C10H16O4 Gelsemiol Iridoids
93 18.3 369.1802 338.1611,323.1733,307.1432,178.1215,122.0952 C21H24N2O4 Gelsevirine N-oxide Gelsemine-type
94 18.5 373.1767 342.1570,313.1564,311.1394,293.1285,283.1432,270.1129,189.0778,108.0808 C20H24N2O5 GS-2 Gelsedine-type
95 18.7 307.1810 277.1701,233.1229,220.1122,176.1074,130.0654,70.0658 C20H22N2O Koumine Koumine-type
96 19.0 359.1611 328.1391,311.1351,283.1451,271.1076,254.0819,190.0723,150.0907,138.0909 C19H22N2O5 19,20-Dihydroxygelsenicine Gelsedine-type
97 19.0 449.1084 413.0855,383.0752,353.0650,329.0650,299.0547,209.1642 C21H20O11 Orientine/iso-orientine Flavone
98 19.4 371.1966 340.1781,313.1542,311.1404,295.1437,277.1331 C21H26N2O4 15-hydroxyhumantenine Humantenine-
type
99 19.5 295.1802 277.1697,222.1269,156.0802,144.0806,138.0908,108.0807 C19H22N2O Koumidine Sarpagine-type
100 19.5 375.1916 344.1723,313.1540,299.1388,265.1323,257.1268,198.1108,132.0441 C20H26N2O5 Hydroxylation of Gelsemicine Gelsedine-type
101 19.6 359.1606 328.1405,311.1387,299.1414,281.0969,185.0702 C19H22N2O5 14,19-Dihydroxygelsenicine Gelsedine-type
102 19.8 417.1655 386.1469,368.1498,341.1498,329.1128,323.13838,311.1122,283.1073,194.0804 C21H24N2O7 14-Acetoxy-dihydroxygelsenicine Gelsedine-type
103 19.9 391.1868 360.1680,331.1283,329.1478,313.1176,217.0959,138.0546 C20H26N2O6 Dihydroxy gelsemicine Gelsedine-type
104 20.1 369.1816 325.1542,311.1405,295.1418 C21H24N2O4 Humantenoxenine Humantenine-
type
105 20.2 433.1975 402.1784,371.1601,343.1628,311.1386,283.1443,150.0915 C22H28N2O7 11-Hydroxy-14-acetoxygelselegine Gelsedine-type
106 20.3 371.1977 340.1774,325.1548,323.1762,138.0913 C21H26N2O4 19(R)-Hydroxydihydrogelsevirine/19(S)- Gelsemine-type
hydroxydihydrogelsevirine
107 20.9 371.1973 340.1772,323.1750,212.0715,122.0961 C21H26N2O4 19(R)-Hydroxydihydrogelsevirine/19(S)- Gelsemine-type
hydroxydihydrogelsevirine
7

108 21.1 295.1816 277.1700,247.1239,156.0806,144.0812,138.0918,120.0813,108.0814 C19H22N2O Isomer of koumidine Sarpagine-type

109 21.1 405.2028 374.1823,343.1652,329.1502,325.1525 C21H28N2O6 11-Methoxy-19-(R)-hydroxygelselegine Gelsedine-type
110 21.3 327.1720 296.1534,265.1357,225.1055,108.0825,95.0747,71.0750 C19H22N2O3 Gelsenicine Gelsedine-type
111 21.6 371.1966 340.1782,325.1552,311.1410,178.1229,122.0965 C21H26N2O4 6-hydroxyhumantenine Humantenine-
type
112 21.9 357.1809 326.1632,311.1397,297.1263,269.1285,178.1228,164.1073,122.0967,108.0816 C20H24N2O4 14-Hydroxyrankinidine Humantenine-
type
113 22.1 353.1865 323.1722,322.1675,291.1491,164.1067,108.0809 C21H24N2O3 Gelsevirine Gelsemine-type
114 22.5 343.1663 312.1480,281.1265,255.1127,238.0867,210.0915,174.0783,136.0785,118.0653 C19H22N2O4 Hydroxyl of gelsenicine Gelsedine-type
115 22.6 325.1914 307.1796,281.1658,243,1500,158.0617,136.1124 C21H26N2O2 Gardnerine Sarpagine-type
116 22.8 363.1712 332.1515,301.1133,261.1032,212.0702,144.0803,121.0855 C22H22N2O3 Unknown Gelsedine-type
117 22.9 311.1761 293.1674,267.1490,249.1398,229.1338,158.0605,138.0900,122.0964,108.0815 C19H22N2O2 3-Hydroxykoumidine Sarpagine-type
118 22.9 357.1811 326.1625,297.1428,295.1442,278.1185,254.1174,213.0919,108.0813,71.0740 C20H24N2O4 4,20-Dehydrogelsemicine Gelsedine-type
119 22.9 383.1970 365.1857,341.1872,321.1592,180.1017,172.0753,138.0913 C22H26N2O4 Gelsempervine A Sarpagine-type
120 23.7 339.1709 308.1521,277.1339,225.1019,176.1067,148.1119,114.0918 C21H26N2O2 Nb-Demethylgelsevirine Gelsemine-type

Chinese Herbal Medicines xxx (xxxx) xxx

121 23.7 533.2496 515.2388,502.2278,484.2206,467.2177,453.2030,381.1824,353.1859,339.1698 C27H36N2O9 Unknown Gelsedine-type
122 23.9 309.1963 291.1840,265.1713,220.1113,178.1226,122.0960 C20H24N2O Dihydrokoumine Koumine-type
123 23.9 353.1870 322.1677,295.1721,291.1488,239.1178,121.0883 C21H24N2O3 19-(Z)-Akuammidine Sarpagine-type
124 23.9 359.1964 328.1778,297.1602,279.1484,222.0943,182.1161 C20H26N2O4 Gelsemicine Gelsedine-type
125 23.9 417.2016 399.1909,376.1818,368.1717,357.1786,326.1627,298.1686,269.1284,163.0985 C22H28N2O6 Unknown Gelsedine-type
126 24.1 385.1764 354.1573,323.1407,311.1390,295.1444,263.1179,237.1045,121.0880,108.0812,95.0732 C21H24N2O5 14-Acetoxygelsenicine Gelsedine-type
127 24.3 517.1342 488.1965,443.1667,163.0386,145.0282,117.0338 C25H24O12 1,3-dicaffeoylquinic acid Phenolic acids
128 24.9 369.1812 323.1412,309.1539,307.1439,265.1079,172.1063,148.1113,122.0964,107.0742 C21H24N2O4 19E-16-epi-Voacarpine Sarpagine-type
129 25.0 329.1859 298.1671,269.1279,257.1167,152.1059,84.0808 C19H22N2O3 Gelsedine Gelsedine-type
130 25.1 417.2016 386.1832,355.1651,343.1652,295.1433,225.1015,150.0907 C22H28N2O6 14-Acetoxygelselegine Gelsedine-type
131 25.5 389.2069 358.1881,327.1703,309.1577,284.1405 C21H28N2O5 11-Methoxygelselegine Gelsedine-type

(continued on next page)

Meng-ting Zuo, Yan-chun Liu, Zhi-liang Sun et al. Chinese Herbal Medicines xxx (xxxx) xxx

elegans. They could be divided into seven groups according to their

structural types and MS/MS fragmentation pathways.

Sarpagine-type

Sarpagine-type
Gelsedine-type

Gelsedine-type

Gelsedine-type
Gelsedine-type
Humantenine-

Humantenine-

Humantenine-

Humantenine-

Humantenine-

Humantenine-

Humantenine-
Classification

Yohimbane
3.3.1. Characterization of gelsedine-type alkaloids

Flavone
A total of 52 components were recognized as gelsedine-type
type

type

type

type

type

type

type

Oil
alkaloids. Based on the fragmentation behavior of these compo-
nents, they contained a Na-methoxy group. On the one hand, there
was also a methoxy group on the C-11 position for some compo-
nents; on the other hand, some of the components had a CH2OH

Na-Methoxy-19(Z)anhydrovobasinediol
Isomer of 19,20-Dihydrohumantenine
group at this position. Therefore, the main fragmentation pattern
of components could be a neutral loss of 62 Da (two OCH3 or
Acetyl of 14-Hydroxygelsedine
Na-Desmethoxyhumante-nine
OCH3 plus CH2OH). The neutral loss chromatogram (pNLC) of m/z
Rankinidine (Gelsemamides)
19,20-Dihydrohumantenine

62 was shown in Fig. 3. For example, components 74 and 124 dis-

Phthalic acid dibutyl ester

Tentative characterization

11-Methoxyhumantenine
11-hydroxyhumantenine

played [M + H]+ ions at m/z 359.1600 and 359.1964, respectively.

Hydroxyl of gelsenicine
Hydroxyl of gelsenicine
Both of them showed a fragment loss of 62 Da. However, for com-
Gelseoxazolidinine

ponent 74, a fragment of m/z 279 (M + H-80 Da) was lost, indicat-
Oxokoumidine
Humantenine

Sempervirine

ing that component 74 had more than one hydroxyl group;

Unknown
moreover, component 74 was an isomer of 28. Therefore, compo-
nents 74 and 124 were characterized as 14,15-
dihydroxygelsenicine and gelsemicine, respectively.
The fragment ions of component 58 were m/z 344, 327, and 267,
C21H28N2O3

C21H26N2O3

C23H28N2O6

C21H26N2O4

C22H28N2O4

C21H28N2O3

C21H28N2O4

C21H26N2O2
C19H22N2O4
C19H22N2O4
C19H20N2O2
C20H24N2O3

C20H22N2O2

which were each 16 Da higher than the corresponding fragment ions

C19H16N2

C19H18O6
C16H22O4
Formula

m/z 328, 311, and 251 of component 96, respectively. Therefore,

component 58 was the hydroxylated product of component 96.
Components 58 and 45 provided the same fragment ion at m/z 299
by loss of a C2H3 or C2H5 group, respectively, indicating that compo-
nent 45 was a reduction product of component 58. Components 41
and 52 were 16 Da and 32 Da higher in molecular mass than compo-
nents 100 and 124, respectively, suggesting that they were dihy-
droxylation products of gelsemicine. The fragmentation pathways
of the gelsedine-type alkaloids were summarized in Fig. 4.
308.1862,293.1644,279.1541,178.1221,164.1068,148.1117,136.1116,122.0961,108.0815

312.1448,295.1285,281.1294,265.1071,255.1113,238.0862,210.0991,174.0773,136.0750
311.1749,310.1712,295.1483,281.1355,178.1266,164.1109,122.1003,108.0850,96.0849

3.3.2. Characterization of sarpagine-type alkaloids

Compared with the target components, component 99 was
312.1450,281.1317,265.1067,255.1137,238.0862,210.0917,174.0872,118.0860

characterized as koumidine. Furthermore, component 108 was an

isomer of 99. According to the fragmentation pathway of koumi-
dine (99), component 117 was characterized as a hydroxylated
325.1911,324.18556,311.1703,310.1658,309.1625,178.1253122.0992

398.1839,385.1536,353.1862,339.1695,222.1110,166.0849,122.0946

derivative of component 99. Components 89 and 117 were charac-

terized as 3-hydroxykoumidine and hydroxylated koumidine,
respectively. Component 60 was the N-oxide form of 3-
326.1984,311.1773,298.2032,239.1178,181.1452,124.1115

295.1466,281.1322,138.0939,132.0478,120.0813,108.0847

340.1764,325.1531,311.1396,164.1055,122.0946,108.0797

265.0186,219.1034,149.0232,121.0285,93.0342,57.0706

hydroxykoumidine. The sarpagine-type components could first

lose the group at the Na position, C-3 position or C-16 position.
The diagnostic ion of these alkaloids was m/z 138. Components
326.1968,311.1726,297.1587,181.1431,124.1223
354.1920,339.1716,311.1720,178.1228,122.0972

123 and 113 had the same molecular formula, but component
113 had a diagnostic ion at m/z 323. The most abundant fragment
327.0868,313.0706,299.0907,282.0881

ion of component 123 was at m/z 295. Therefore, component 123

was characterized as 19-(Z)-akuammidine. The fragmentation
pathway of the sarpagine-type alkaloids was shown in Fig. 5.
294.1728, 279.1519, 164.1066
342.1929,327.1680,313.1549

3.3.3. Characterization of humantenine-type alkaloids

The key filter ions of humantenine-type alkaloids were m/z
257.1064,254.1068

311.17 (components 67, 132, 133, 136, 138 and 139) and 311.14
Fragment ions

(components 36, 48, 65, 66, 98, 104, 111, 112 and 137). The results
indicated that humantenine-type alkaloid components could be
found by filtering m/z 311 in the extracted-ion chromatogram
(EIC) MS/MS spectrum. The EIC MS/MS spectrum of m/z 311 was
shown in Fig. 6. These components could lose H2O, hydroxymethyl,
357.2166

341.1896

371.1951

385.2134

357.2162

373.2127
325.1911

273.1388

343.1654
343.1655
343.1184
279.1598
309.1616
429.2018

339.2068

methoxy, methyl, or methylene groups to form the fragment ion at

[M + H]+

355.204

m/z 311. For example, component 112 could lose OCH3 and CH3
groups to form m/z 311, and component 137 could lose CH2 after
losing 46 Da (OCH3 plus CH3) to form m/z 311. The molecular for-
Table 1 (continued)

t/R(min)

mulas of compounds 98, 111, and 137 were calculated as

30.2
30.6

40.7
25.7

25.9

25.9
26.1
26.1

27.4

27.5

27.8

28.9
29.7

33.6

40.0
36.0

C21H26N2O4 based on their measured accurate mass of m/z

371.196, which suggested that components 98, 111, and 137 were
132

133

134
135
136

137

138

139

140
141

142
143
144
145
146
147
No.

isomers. Components 36 and 65 were 16 Da higher in molecular

weight than component 112, which showed they were formed by
8
Meng-ting Zuo, Yan-chun Liu, Zhi-liang Sun et al. Chinese Herbal Medicines xxx (xxxx) xxx

Fig. 3. EIC spectrum (A), MS/MS spectrum (B) of component 94 (GS-2), pNLC spectrum of gelsedine type alkaloids with neutral loss ion at m/z 62 (C) and pNLC spectrum of G.
elegans compounds (D).

dihydroxylation of component 112. As a result, components 112 yielded a diagnostic ion at m/z 220. Koumine-type alkaloids could
and 137 were characterized as 14-hydroxyrankinidine and 11- gradually lose the groups at the C-19 position, Nb position, C-16
hydroxyhumantenine, respectively. Through online database and C17 positions. The fragmentation pathways of the koumine-
searching, component 104 was tentatively characterized as type alkaloids are shown in Fig. S2. By comparison to a koumine
humantenoxenine, which belonged to the humantenine-type alka- standard, component 95 was characterized as koumine. Compo-
loid family. The fragmentation pathways of the humantenine-type nent 95 was 2 Da higher in molecular weight than component
alkaloids were shown in Fig. 6. 88, and components 95 and 88 had the same fragment ions, which
showed that component 88 was the dehydrogenation product of
component 95. Therefore, component 88 was named dehydrok-
3.3.4. Characterization of gelsemine-type alkaloids
oumine. Components 39 and 53 also had the same fragment ions
The precursor ions of 11 gelsemine-type alkaloids, including
and were characterized as 19-(S)-hydroxydihydrokoumine and
compounds 46, 44, 51, 56, 63, 72, 77, 93, 106, 107, and 113 were
19-(R)-hydroxydihydrokoumine by literature searching (Kitajima,
screened out by using the diagnostic ion at m/z 323. The fragmen-
Kobayashi, Kogure, & Takayama, 2010), respectively.
tation pathways of the gelsemine-type alkaloids were shown in
Fig. S1. These components would first lose the group at the Na posi-
tion, the methyl at the Nb position and the group at the C-19 posi- 3.3.6. Characterization of iridoids
tion to form the diagnostic ion at m/z 323. For example, Iridoids were filtered by a neutral loss of 46 Da (CH2O2) and a
components 106 and 107 could lose OCH3 (31 Da) to form the pro- diagnostic ion at m/z 91. This filtering could be applied to EIC
duct ion at m/z 340 and then could lose OH (17 Da) at the C-19 MS/MS and pNLC spectra due to the structure of iridoids. For exam-
position to form m/z 323. Based on the fragmentation data and ple, components 10 and 25 had the same formula, but component
accurate mass values, components 106 and 107 were a pair of iso- 25 could lose 82 Da (2H2O plus CH2O2), whereas component 10
mers and were characterized as 19(R)-hydroxydihydrogelsevirine could lose only 64 Da (H2O plus CH2O2), which proved that com-
or 19(S)-hydroxydihydrogelsevirine, respectively. Component 120 pared to component 10, component 25 had an additional hydroxyl
was 14 Da lower in molecular weight than component 113 (loss group. Components 11 and 26 were 16 Da higher in molecular
of a methyl group at the Nb position), and component 93 was weight than component 92 and were characterized as 7-
16 Da higher in molecular weight than component 113 (Nb posi- hydroxygelsemiol or 9-hydroxygelsemiol, respectively. Through
tion). Thus, components 120, 113, and 93 were characterized as the combination of target ion and database searching, the nontar-
Nb-demethylgelsevirine, gelsevirine and gelsevirine N-oxide. Fur- geted components were characterized quickly. The fragmentation
thermore, components 37, 46, 56, and 77 were the oxidized form pathways of the iridoids are shown in Fig. S3.
of components 44 or 72.
3.3.7. Characterization of phenolic acids
3.3.5. Characterization of koumine- type alkaloids Only two types of phenolic acids were detected in G. elegans
A total of seven components, compounds 39, 53, 54, 85, 88, 95, after matching with the MS database, and their structures were
and 122 were classified as koumine-type alkaloids since they all determined based on the MS/MS spectra. Component 62 produced
9
Meng-ting Zuo, Yan-chun Liu, Zhi-liang Sun et al. Chinese Herbal Medicines xxx (xxxx) xxx

Fig. 4. Mass spectral fragmentation pathways of gelsedine-type alkaloids.

a sodium-adduct molecular ion [M + Na]+ at m/z 377.0866 and a the protonated molecule. The ion at m/z 145.0291 was formed by
protonated molecular ion [M + H]+ at 355.1024, and the enriched the loss of a molecule of H2O (18 Da) from m/z 163.0394. The min-
fragment ion at m/z 163 formed by losing glucuronic acid from imum ion at m/z 89.0403 was generated by the neutral loss of two

10
Meng-ting Zuo, Yan-chun Liu, Zhi-liang Sun et al. Chinese Herbal Medicines xxx (xxxx) xxx

Fig. 5. Mass spectral fragmentation pathways of sapargine-type alkaloids.

CO molecules (28 Da) at m/z 145.0291. The structure of the pheno- Second, higher precursor ion and product ion responses were
lic acids was based on a benzene ring and numerous hydroxyl obtained through adequate optimization of instrumental parame-
groups, so its fragment ions were formed by losses of H2O and ters by repeatedly testing four standards (gelsemine, koumine, kou-
CO. Component 6 was established as ferulic acid, and it was midine and gelsenicine) in reference multiple reaction monitoring
proposed as an isomer of component 29. Component 62 was (RMRM) mode. Finally, multiple components of G. elegans were opti-
characterized tentatively as 1-O-caffeoylquinic acid or 4-O- mized in terms of ion pairs and CE in LC-QqQ/MS (Table 2).
caffeoylquinic acid. Through an online database search, compo- To avoid interference with some low-concentration compo-
nents 12, 16, 27, 73, and 127 were classified as phenolic acids. In nents, ion pairs were set to several segments to improve sensitiv-
addition, components 12, 16, and 27 were hydroxylated deriva- ity. The multiple components need to be optimized in terms of the
tives of ethyl ferulic acid. analysis time period by adjusting the LC conditions to ensure good
peak shape without tailing and drift. In addition, it is observed that
compounds with similar structures are often assigned similar MRM
3.4. MRM transitions derived from MS2 parameters and transitions. Therefore, we developed a DeMRM
method to monitor multiple components in herbal medicines, even
With a combination of LC-QTOF/MS and LC-QqQ/MS, we devel- those present at a trace level. In this study, G. elegans was selected
oped a quantitative approach for analysis in the absence of stan- as an example to demonstrate our approach. The retention times,
dards. The present LC-MS/MS strategy enabled the separation of monitored ion pairs, and related voltage parameters of multiple
41 components in the method validation protocol. Many parameters components in G. elegans were shown in Table 2.
must be optimized during the development of a DeMRM method,
which is always laborious and time-consuming. Therefore, a simple
standard operation procedure was recommended in the present 3.5. Method validation of proposed method
study.
First, the precursor ions and corresponding product ions of each Table S3 summarized the validation results of the four repre-
component were obtained by LC-QTOF/MS and were used to form sentative components (gelsemine, koumine, koumidine and gelse-
an ion pair; in general, each precursor ion had 2–3 product ions. nicine) for RMRM. The correlation coefficients of the four
11
Meng-ting Zuo, Yan-chun Liu, Zhi-liang Sun et al. Chinese Herbal Medicines xxx (xxxx) xxx

Fig. 6. EIC spectrum (A) and MS/MS spectrum (B) of component 133 (humantenine); MS/MS spectrum extracted by extracting diagnostic ions at m/z 311.17, and 311.14 (C).
Mass spectral fragmentation pathways of humantenine-type alkaloids (D).

12
Meng-ting Zuo, Yan-chun Liu, Zhi-liang Sun et al. Chinese Herbal Medicines xxx (xxxx) xxx

Table 2
Monitor ion pairs, CE, segment, repeatability and stability of multiple compounds in G. elegans samples.

No. Analytes t/R Ion pair CE/ Seg Repeatability (mg/g ± RSD Stability/RSD
(min) eV %) (%)
1 7-Deoxygelsemide or 9-deoxygelsemide 6.2 197.1 ? 105.1 30 2 0.001 ± 3.31 3.02
2 11,14-Dihydroxygelsenicine 11.4 359.2 ? 108.1 30 3 0.086 ± 2.88 1.63
3 14,15-Dihydroxygelsenicine 11.5 359.2 ? 328.1 30 3 0.066 ± 2.96 2.56
4 Unknown (375) 15.6 375.2 ? 311.1 30 4 0.029 ± 1.42 1.98
5 Gelsemine 16.5 323.2 ? 70.1 33 4 3.787 ± 0.94 1.58
6 11-Hydroxygelsenicine 16.5 343.2 ? 281.1 30 4 0.017 ± 1.27 3.42
7 Gelsemicine 17.1 359.2 ? 108.1 30 4 0.002 ± 2.79 2.12
8 GSIR-1 17.4 183.1 ? 91.1 30 4 0.001 ± 2.76 4.68
9 14-Hydroxygelsemicine or other hydroxylation of gelsemicine 17.5 375.2 ? 313.2 35 4 0.237 ± 2.36 1.56
10 14-Hydroxygelsedine 17.8 345.2 ? 285.1 30 4 0.034 ± 2.13 2.46
11 14-Hydroxygelsenicine 18.0 343.2 ? 108.1 30 5 0.730 ± 3.92 1.63
12 Koumine 19.1 307.2 ? 180.0 53 5 0.830 ± 1.83 2.07
13 14-Dehydroxygelsefuranidine or other dehydroxylgelsefuranidine 19.4 405.2 ? 343.1 30 5 0.071 ± 2.70 1.86
(2)
14 11-methoxy-14-hydroxygelsenicine 19.5 373.2 ? 108.1 30 5 0.142 ± 2.16 2.69
15 Unknown (295) 19.8 295.2 ? 138.1 30 5 0.058 ± 1.78 3.03
16 Hydroxyl of gelsedine 19.9 345.2 ? 285.1 30 5 0.001 ± 4.46 4.20
17 Gelsemoxonine 20.1 359.2 ? 311.1 30 6 0.942 ± 3.38 2.10
18 One of other 5 hit compounds 21.1 371.2 ? 323.1 30 6 0.067 ± 2.84 1.64
19 14-Dehydroxygelsefuranidine or other dehydroxylgelsefuranidine 21.1 405.2 ? 374.2 30 6 0.049 ± 2.92 4.24
(1)
20 Koumidine 21.2 295.2 ? 144.1 30 6 0.042 ± 2.01 2.06
21 14-Acetoxygelsedilam or other acetoxyl of gelsedilam (1) 21.4 373.2 ? 342.2 30 6 0.005 ± 4.13 2.06
22 11-Hydroxyhumantenine 21.5 371.2 ? 325.2 30 6 0.312 ± 1.77 2.26
23 Gelsevirine 22.0 353.2 ? 291.2 30 7 0.191 ± 2.30 1.39
24 Gelsenicine 22.2 327.2 ? 265.1 30 7 0.958 ± 1.75 1.41
25 12b-hydroxy-pregn-4,16-diene-3,20-dione 22.2 329.2 ? 97.1 30 7 0.002 ± 4.96 3.00
26 Koumicine 23.8 353.2 ? 166.1 30 7 0.342 ± 3.81 1.70
27 Gelsedine or Nb-Methylgelsedilam 23.4 329.2 ? 298.2 30 7 0.004 ± 1.54 2.01
28 Gelseoxazolidinine 23.9 429.2 ? 339.2 30 7 0.003 ± 2.53 2.99
29 16-epi-voacarpine or gelsevirine N-oxide 24.0 369.2 ? 166.1 30 7 0.823 ± 1.20 1.84
30 Humantenoxenine 24.3 369.2 ? 108.1 30 7 0.009 ± 3.48 3.31
31 Iso-Gelsedine or Nb-Methylgelsedilam 24.7 329.2 ? 269.3 30 7 0.014 ± 4.55 2.48
32 1-O-Caffeoylquinic acid or 4-O-Caffeoylquinic acid 25.0 355.1 ? 135.0 30 8 0.117 ± 1.24 1.39
33 Gelseziridine 25.7 343.2 ? 108.1 30 8 0.001 ± 2.57 2.27
34 Na-Desmethoxyhumantenine 25.8 325.2 ? 136.1 35 8 0.020 ± 2.63 1.55
35 14-Acetoxygelsedilam or other acetoxyl of gelsedilam (2) 26.6 373.2 ? 342.2 30 8 0.033 ± 2.14 1.95
36 14-Acetoxygelsedilam or other acetoxyl of gelsedilam (3) 27.3 373.2 ? 342.2 30 8 0.033 ± 2.48 2.57
37 19R-Hydroxydihydrogelsevirine or 19S-Hydroxydihydrogelsevirine 26.5 371.2 ? 164.1 30 8 4.526 ± 4.24 4.18
38 11-Methoxyhumantenine 26.8 385.2 ? 339.2 30 8 0.061 ± 4.33 4.53
39 iso-12b-Hydroxy-5a-pregn-16-ene-3,20-dione 31.2 331.2 ? 109.1 30 11 0.003 ± 4.42 4.02
40 Gelse-norursane E 31.7 471.2 ? 217.1 30 10 0.008 ± 4.28 3.99
41 12b-Hydroxy-5a-pregn-16-ene-3,20-dione 31.8 331.2 ? 97.1 30 11 0.038 ± 4.24 4.18

Fig. 7. Differences contents of four standards between roots (A), stems (B) and leaves (C) from G. elegans.

compounds were higher than 0.99 in the concentration range of retention time of gelsemine was 16.5 min, and that of koumidine
10–200 ng/mL. The LODs of gelsemine, koumine, koumidine and was 22.2 min, which indicated the specificity of this analytical
gelsenicine were 2.5, 2, 5 and 1.5 ng/mL, respectively. The LOQs method. Table 2 summarizes the relative concentrations of 41 com-
of gelsemine, koumine, koumidine and gelsenicine were 5, 5, 10 ponents in the G. elegans sample and relative standard deviations
and 5 ng/mL, respectively. (RSDs) of the concentrations. The intraday and interday precision
The DeMRM method was examined in terms of specificity, accu- were expressed as the RSD. The RSDs of the four representative com-
racy, and stability. As presented in Fig. S4, no interfering peaks were ponents at the two tested concentrations were all within 10%. More-
observed at the retention times of the 41 components in the DeMRM over, all the RSDs of the DeMRM method were within the accepted
chromatograms of G. elegans samples. Most componets had no the variable limits. The results support that the DeMRM method has rea-
phenomenon of trailing, incomplete peak of the sample and the sonable accuracy and stability and is applicable to the quantitative
retention time of the sample was suitable. For example, the analysis of complex herbal medicines.

13
Meng-ting Zuo, Yan-chun Liu, Zhi-liang Sun et al. Chinese Herbal Medicines xxx (xxxx) xxx

3.6. Sample analysis when authentic standards are not available and/or the absolute
quantity is not needed.
The validated DeMRM method was subsequently applied to
determine the relative concentrations of multiple components in
different tissues of G. elegans during different periods. The quanti- 4. Conclusion
tative performance of the DeMRM method was examined by com-
paring the experimental values of the four representative The present work contributed to the development of a powerful
components obtained by DeMRM and RMRM. The DeMRM results integrated strategy based on liquid chromatography coupled with
were expressed as the relative content of herb based on a koumi- mass spectrometry (LC-MS) systems. The results demonstrated the
dine calibration curve, while the RMRM results were converted significant advantages of this strategy over other strategies. First,
to the herb contents by calculating the absolute amounts of the the number of components detected with high peak intensity
components in the herb. Taking koumine in the root as an example, was successfully maximized by comprehensively optimizing the
as shown in Fig. 7 (Root), the RMRM results indicated that the kou- LC-QTOF/MS method, and 31 target components were character-
mine content obtained from chemical standards was the highest in ized through matching analysis with our established personal Gel-
December, followed by September, and was the lowest in Novem- semium database. Second, various data mining techniques,
ber, which is the result of RMRM. In Table 3, the relative contents including database searching, diagnostic ion filtering and neutral
of koumine obtained from the koumidine calibration curves were loss filtering, were implemented to fully and systematically clarify
1.376, 1.089, and 1.772 mg/g in September, November, and Decem- the structure of various chemical components in G. elegans. A total
ber, respectively, which indicated that the results obtained by of 147 components were characterized from G. elegans, and among
DeMRM were consistent with the above trends observed by them, 116 nontarget components were reported for the first time.
RMRM. The same trend proved that the results of the relative A sensitive and reproducible LC-QqQ/MS method was successfully
quantification of 41 components calculated by DeMRM in G. ele- developed and validated for the simultaneous relative quantifica-
gans were reliable. Although some deviation exists, the approach tion of 41 components of G. elegans. This method was effective in
proposed herein still offered a direct and rapid method for semi- identifying a variety of nontarget components and provided a tech-
quantitative determination with reasonable accuracy in cases nical reference for the characterization of other chemical

Table 3
Multi-compounds contents in Gelsemium elegans samples.

No. Analytes S1 S2 S3 S4 S5 S6 S7 S8 S9
1 7-Deoxygelsemide or 9-Deoxygelsemide – – 0.004 – – 0.001 – – 0.001
2 11,14-Dihydroxygelsenicine 0.043 0.041 0.114 0.012 0.058 0.193 0.009 0.148 0.472
3 14,15-Dihydroxygelsenicine 0.032 0.032 0.087 0.009 0.045 0.148 0.009 0.114 0.365
4 Unknown (375) 0.037 0.007 0.005 0.032 0.015 0.015 0.043 0.015 0.009
5 Gelsemine 3.036 4.145 4.679 2.208 3.014 3.565 3.956 4.678 3.19
6 11-Hydroxygelsenicine 0.048 0.032 0.08 0.008 0.039 0.009 0.002 0.003 0.006
7 Gelsemicine – – – 0.001 0.001 0.002 0.003 0.004 0.004
8 GSIR-1 0.001 0.001 0.003 – – 0.001 – – 0.001
9 14-Hydroxygelsemicine or other hydroxylation of gelsemicine 0.424 0.145 0.129 0.273 0.187 0.13 0.417 0.136 0.088
10 14-Hydroxygelsedine 0.048 0.005 0.007 0.066 0.005 0.004 0.132 0.008 0.005
11 14-Hydroxygelsenicine 0.383 0.171 0.576 0.704 0.407 0.552 0.951 1.198 0.925
12 Koumine 1.376 0.644 0.693 1.089 0.562 0.51 1.772 0.333 0.197
13 14-Dehydroxygelsefuranidine or other dehydroxylgelsefuranidine (2) 0.076 0.037 0.106 0.064 0.041 0.143 0.117 0.03 0.097
14 11-Methoxy-14-hydroxygelsenicine 0.064 0.02 0.099 0.133 0.069 0.091 0.224 0.182 0.169
15 Unknown (295) 0.077 0.039 0.023 0.071 0.027 0.016 0.152 0.025 0.008
16 Hydroxyl of gelsedine 0.001 – – – – – 0.003 – –
17 Gelsemoxonine 0.158 0.04 0.455 0.167 0.207 1.994 0.617 1.066 3.185
18 One of other five hit compounds 0.08 0.028 0.065 0.069 0.022 0.071 0.184 0.017 0.039
19 14-Dehydroxygelsefuranidine or other dehydroxylgelsefuranidine (1) 0.059 0.013 0.034 0.06 0.01 0.046 0.111 0.019 0.03
20 koumidine 0.022 0.04 0.02 0.028 0.028 0.022 0.107 0.028 0.012
21 14-Acetoxygelsedilam or other Acetoxyl of gelsedilam (1) 0.01 0.01 0.008 0.001 0.005 0.003 0.004 0.003 0.003
22 11-Hydroxyhumantenine 0.987 1.069 0.963 0.186 0.541 0.185 0.077 0.097 0.1
23 Gelsevirine 0.257 0.108 0.062 0.389 0.08 0.02 0.631 0.03 0.007
24 Gelsenicine 0.944 0.228 0.823 1.085 0.805 0.55 1.508 0.901 0.488
25 12b-Hydroxy-pregn-4,16-diene-3,20-dione 0.002 – 0.001 0.002 0.001 0.001 0.002 0.001 0.001
26 Koumicine 0.326 0.237 0.349 0.219 0.217 0.446 0.441 0.199 0.327
27 Gelsedine or Nb-Methylgelsedilam 0.007 0.001 0.002 0.006 0.002 0.001 0.01 0.001 0.001
28 Gelseoxazolidinine 0.015 0.005 0.01 0.002 – 0.003 0.002 – 0.001
29 16-epi-Voacarpine or gelsevirine N-oxide 0.853 0.514 0.812 0.57 0.645 1.13 0.934 0.817 1.033
30 Humantenoxenine 0.006 0.008 0.02 0.003 0.004 0.013 0.009 0.009 0.011
31 Iso-Gelsedine or Nb-Methylgelsedilam 0.012 0.002 0.002 0.042 0.001 0 0.042 0.001 –
32 1-O-Caffeoylquinic acid or 4-O-Caffeoylquinic acid 0.176 0.044 0.072 0.165 0.059 0.071 0.268 0.064 0.045
33 Gelseziridine 0.001 – – 0.003 – 0.001 0.004 – –
34 Na-Desmethoxyhumantenine 0.028 0.005 0.012 0.028 0.01 0.014 0.046 0.012 0.009
35 14-Acetoxygelsedilam or other Acetoxyl of gelsedilam (2) 0.039 0.008 0.019 0.068 0.019 0.02 0.069 0.019 0.019
36 14-Acetoxygelsedilam or other Acetoxyl of gelsedilam (3) 0.028 0.009 0.023 0.087 0.017 0.019 0.09 0.018 0.018
37 19R-Hydroxydihydrogelsevirine or 19S-Hydroxydihydrogelsevirine 3.436 1.261 1.335 9.918 2.4 2.677 10.567 1.584 2.629
38 11-Methoxyhumantenine 0.105 0.08 0.04 0.081 0.034 0.034 0.272 0.016 0.011
39 iso-12b-Hydroxy-5a-pregn-16-ene-3,20-dione 0.004 0.001 0.004 0.002 – 0.006 0.008 – 0.005
40 Gelse-norursane E 0.012 0.001 0.002 0.006 0.001 0.007 0.036 0.001 0.003

Note: -, nondetected.

14
Meng-ting Zuo, Yan-chun Liu, Zhi-liang Sun et al. Chinese Herbal Medicines xxx (xxxx) xxx

components. The present integrated strategy would significantly Model of Diabetic Neuropathy. Biological|Pharmaceutical Bulletin, 37(5),
858–864.
contribute to chemical studies on herbal medicine, and its utility
Liu, M., Huang, H.-H., Yang, J., Su, Y.-P., Lin, H.-W., Lin, L.-Q., Liao, W.-J., & Yu, C.-X.
could be extended to other research fields, such as metabolomics, (2013). The active alkaloids of Gelsemium elegans Benth. are potent anxiolytics.
quality control and pharmacokinetics. Psychopharmacology (Berl), 225(4), 839–851.
Liu, Y.-C., Lin, L. i., Cheng, P. i., Sun, Z.-L., Wu, Y., & Liu, Z.-Y. (2017a). Fingerprint
analysis of Gelsemium elegans by HPLC followed by the targeted identification of
Declaration of Competing Interest chemical constituents using HPLC coupled with quadrupole-time-of-flight mass
spectrometry. Fitoterapia, 121, 94–105.
Liu, Y.-C., Xiao, S. a., Yang, K., Ling, L. i., Sun, Z.-L., & Liu, Z.-Y. (2017b). Comprehensive
The authors declare that they have no known competing finan-
identification and structural characterization of target components from
cial interests or personal relationships that could have appeared Gelsemium elegans by high-performance liquid chromatography coupled with
to influence the work reported in this paper. quadrupole time-of-flight mass spectrometry based on accurate mass databases
combined with MS/M: Structural characterization of Gelsemium elegans
components by LC-QqTOF MS. J. Mass Spectrom., 52(6), 378–396.
Acknowledgments Ma, X.-D., Fan, Y.-X., Jin, C.-C., Wang, F., Xin, G.-Z., Li, P., & Li, H.-J. (2016). Specific
targeted quantification combined with non-targeted metabolite profiling for
quality evaluation of Gastrodia elata tubers from different geographical origins
This work was supported by National Key R&D Program of and cultivars. Journal of Chromatography A, 1450, 53–63.
Intergovernmental Key Projects (Grant No: 2018YFE0101700) Meyer, L., Boujedaini, N., Patte-Mensah, C., & Mensah-Nyagan, A. G. (2013).
and National Natural Science Foundation of China (Grant No. Pharmacological effect of gelsemine on anxiety-like behavior in rat.
Behavioural Brain Research, 253, 90–94.
31972737).
Mustafa, A. M., Eldahmy, S. I., Caprioli, G., Bramucci, M., Quassinti, L., Lupidi, G.,
Beghelli, D., Vittori, S., & Maggi, F. (2018). Chemical composition and biological
activities of the essential oil from Pulicaria undulata (L.) C. A. Mey. Growing
Appendix A. Supplementary data
wild in Egypt. Natural Product Research, 1–5.
Ornduff, R. (1970). The systemetics and breeding system of Gelsemium
Supplementary data to this article can be found online at (Loganiaceae). Journal of the Arnold Arboretum, 51(1), 1–17.
Qiao, X., Li, R. u., Song, W., Miao, W.-J., Liu, J., Chen, H.-B., Guo, D.-a., & Ye, M. (2016).
https://doi.org/10.1016/j.chmed.2020.06.002.
A targeted strategy to analyze untargeted mass spectral data: Rapid chemical
profiling of Scutellaria baicalensis using ultra-high performance liquid
References chromatography coupled with hybrid quadrupole orbitrap mass spectrometry
and key ion filtering. Journal of Chromatography A, 1441, 83–95.
Samanipour, S., Reid, M. J., Bæk, K., & Thomas, K. V. (2018). Combining a
Cai, H., Xu, Q., Chang, C., Song, C., Yong, X., Yi, X., ... Jiang, H. (2014). The rapid
Deconvolution and a Universal Library Search Algorithm for the Nontarget
discovery and identification of physalins in the calyx of Physalis alkekengi L. var.
Analysis of Data-Independent Acquisition Mode Liquid Chromatography High-
Franchetii (Mast.) Makino using ultra-high performance liquid
Resolution Mass Spectrometry Results. Environmental Science and Technology, 52
chromatography–quadrupole time of flight tandem mass spectrometry
(8), 4694–4701.
together with a novel three-step. Journal of Chromatography A, 1361, 139–152.
Shi, X. J., Yang, W. Z., Qiu, S., Yao, C. L., Shen, Y., Pan, H. Q., ... Guo, D. A. (2017). An in-
Chen, J., Shi, Z., Song, Y., Guo, X., Zhao, M., Tu, P., & Jiang, Y. (2016). Source
source multiple collision-neutral loss filtering based nontargeted metabolomics
attribution and structure classification-assisted strategy for comprehensively
approach for the comprehensive analysis of malonyl-ginsenosides from Panax
profiling Chinese herbal formula: Ganmaoling granule as a case. Journal of
ginseng, p.Quinquefolius, and p.Notoginseng. Analytica Chimica Acta, 952,
Chromatography A, 1464, 102–114.
59–70.
Chingin, K., Makarov, A., Denisov, E., Rebrov, O., & Zubarev, R. A. (2014).
Shi, W., Zhang, C., Zhao, D., Wang, L., Li, P., & Li, H. (2018). Discovery of Hepatotoxic
Fragmentation of Positively-Charged Biological Ions Activated with a Beam of
Equivalent Combinatorial Markers from Dioscorea bulbifera tuber by
High-Energy Cations. Analytical Chemistry, 86(1), 372–379.
Fingerprint-Toxicity Relationship Modeling. Scientific Reports, 8(1). https://doi.
Hao, H., Cui, N., Wang, G., Xiang, B., Liang, Y., Xu, X., Zhang, H., Yang, J., Zheng, C.,
org/10.1038/s41598-017-18929-z.
Wu, L., Gong, P., & Wang, W. (2008). Global Detection and Identification of
Tan, J., Zheng, M., Duan, S., Zeng, Y., Zhang, Z., Cui, Q., Zhang, J., Hong, T., Bai, J., & Du,
Nontarget Components from Herbal Preparations by Liquid Chromatography
S. (2018). Chemical profiling and screening of the marker components in the
Hybrid Ion Trap Time-of-Flight Mass Spectrometry and a Strategy. Analytical
fruit of cassia fistula by hplc and uhplc/ltq-orbitrap ms(n) with chemometrics.
Chemistry, 80(21), 8187–8194.
Molecules, 23(7), 1501–1507.
Hu, Y., Chen, M., Wang, Z., Lan, Y., Tang, L., Liu, M., Zhao, J., Hu, M., Zhang, L., & Ye, L.
Wang, H., Shi, S., & Wang, S. (2018). Can highly cited herbs in ancient Traditional
(2017). Development of a validated UPLC-MS/MS method for determination of
Chinese medicine formulas and modern publications predict therapeutic
humantenmine in rat plasma and its application in pharmacokinetics and
targets for diabetes mellitus?. Journal of Ethnopharmacology, 213, 101–110.
bioavailability studies. Biomedical Chromatography, 31(12), e4017. https://doi.
Wang, L., Wen, Y., & Meng, F. (2018). Simultaneous determination of gelsemine and
org/10.1002/bmc.v31.1210.1002/bmc.4017.
koumine in rat plasma by UPLC-MS/MS and application to pharmacokinetic
Huo, J.-H., Du, X.-W., Sun, G.-D., Dong, W.-T., & Wang, W.-M. (2018). Identification
study after oral administration of Gelsemium elegans Benth extract. Biomedical
and characterization of major constituents in Juglans mandshurica using ultra
Chromatography, 32(6), e4201. https://doi.org/10.1002/bmc.v32.610.1002/
performance liquid chromatography coupled with time-of-flight mass
bmc.4201.
spectrometry (UPLC-ESI-Q-TOF/MS). Chinese Journal of Natural Medicines, 16
Xiao, S. a., Huang, Y.-J., Sun, Z.-L., & Liu, Z.-Y. (2017). Structural elucidation of
(7), 525–545.
koumine metabolites by accurate mass measurements using high-performance
Jin, G.-L., Su, Y.-P., Liu, M., Xu, Y., Yang, J., Liao, K.-J., & Yu, C.-X. (2014). Medicinal
liquid chromatography/quadrupole-time-of-flight mass spectrometry:
plants of the genus Gelsemium (Gelsemiaceae, Gentianales)—A review of their
Structural elucidation of koumine metabolites by HPLC/QqTOF-MS. Rapid
phytochemistry, pharmacology, toxicology and traditional use. Journal of
Communications in Mass Spectrometry, 31(3), 309–314.
Ethnopharmacology, 152(1), 33–52.
Xu, Y., Qiu, H. Q., Liu, H., Liu, M., Huang, Z. Y., Yang, J., & Yu, C. X. (2012a). Effects of
Kitajima, M., Kobayashi, H., Kogure, N., & Takayama, H. (2010). New oxindole and
koumine, an alkaloid of gelsemium elegans benth., on inflammatory and
indole alkaloids from Gelsemium rankinii. Tetrahedron, 66(32), 5987–5992.
neuropathic pain models and possible mechanism with allopregnanolone.
Lei, X.-Q., Li, G., Cheng, L., Wang, X.-L., & Meng, F.-Y. (2018). Identification of
Pharmacology Biochemistry and Behavior, 101(3), 504–514.
Ligustici Rhizoma et Radix and its adulterants based on their chemical
Xu, Y.-K., Liao, S.-G., Na, Z., Hu, H.-B., Li, Y., & Luo, H.-R. (2012b). Gelsemium
constituents by UHPLC-Q/TOF-MS combined with data mining. Journal of
alkaloids, immunosuppressive agents from Gelsemium elegans. Fitoterapia, 83
Pharmaceutical and Biomedical Analysis, 154, 123–137.
(6), 1120–1124.
Li, W. L., Sun, X. M., Song, H., Ding, J. X., Bai, J., & Chen, Q. (2015a). Hplc/q-tof-ms-
Yamada, Y., Kitajima, M., Kogure, N., Wongseripipatana, S., & Takayama, H. (2011).
based identification of absorbed constituents and their metabolites in rat serum
Seven new monoterpenoid indole alkaloids from Gelsemium elegans. Chemistry -
and urine after oral administration of cistanche deserticola extract. Journal of
An Asian Journal, 6(1), 166–173.
Food Science, 80(9), H2079–2087.
Yan, Y., Chai, C.-Z., Wang, D.-W., Yue, X.-Y., Zhu, D.-N., & Yu, B.-Y. (2013). HPLC-
Li, Z., Xiao, S., Ai, N. i., Luo, K., Fan, X., & Cheng, Y. (2015b). Derivative multiple
DAD-Q-TOF-MS/MS analysis and HPLC quantitation of chemical constituents in
reaction monitoring and single herb calibration approach for multiple
traditional Chinese medicinal formula Ge-Gen Decoction. Journal of
components quantification of traditional Chinese medicine analogous
Pharmaceutical and Biomedical Analysis, 80, 192–202.
formulae. Journal of Chromatography A, 1376, 126–142.
Yang, K., Huang, Y.-J., Xiao, S. a., Liu, Y.-C., Sun, Z.-L., Liu, Y.-S., & Liu, Z.-Y. (2018a).
Liang, Y., Hao, H., Kang, A. n., Xie, L., Xie, T., Zheng, X., Dai, C., Wan, L., Sheng, L., &
Identification of gelsemine metabolites in rat liver S9 by high-performance
Wang, G. (2010). Qualitative and quantitative determination of complicated
liquid chromatography/quadrupole-time-of-flight mass spectrometry. Rapid
herbal components by liquid chromatography hybrid ion trap time-of-flight
Communications in Mass Spectrometry, 32(1), 19–22.
mass spectrometry and a relative exposure approach to herbal
Yang, K., Long, X.-M., Liu, Y.-C., Chen, F.-H., Liu, X.-F., Sun, Z.-L., & Liu, Z.-Y. (2018b).
pharmacokinetics independent of standards. Journal of Chromatography A,
Development and in-house validation of a sensitive LC–MS/MS method for
1217(30), 4971–4979.
simultaneous quantification of gelsemine, koumine and humantenmine in
Ling, Q., Liu, M., Wu, M.-X., Xu, Y., Yang, J., Huang, H.-H., & Yu, C.-X. (2014). Anti-
porcine plasma. Journal of Chromatography B, 1076, 54–60.
allodynic and Neuroprotective Effects of Koumine, a Benth Alkaloid, in a Rat

15
Meng-ting Zuo, Yan-chun Liu, Zhi-liang Sun et al. Chinese Herbal Medicines xxx (xxxx) xxx

Yang, Y., Wang, H.-J., Yang, J., Brantner, A. H., Lower-Nedza, A. D., Si, N., Song, J.-F., the determination of gelsemine in rat plasma and tissue: Application to a
Bai, B., Zhao, H.-Y., & Bian, B.-L. (2013). Chemical profiling and quantification of pharmacokinetic and tissue distribution study: Liquid Chromatography. Journal
Chinese medicinal formula Huang-Lian-Jie-Du decoction, a systematic quality of Separation Science, 38(6), 936–942.
control strategy using ultra high performance liquid chromatography combined Zhang, W., Zhang, S.-Y., Wang, G.-Y., Li, N.-P., Chen, M.-F., Gu, J.-H., & Ye, W.-C.
with hybrid quadrupole-orbitrap and triple quadrupole mass spectrometers. (2017a). Five new koumine-type alkaloids from the roots of Gelsemium elegans.
Journal of Chromatography A, 1321, 88–99. Fitoterapia, 118, 112–117.
Zhang, X., Chen, Y. i., Gao, B. o., Luo, D., Wen, Y., & Ma, X. (2015a). Apoptotic Effect of Zhang, Y. u., Cheng, Y., Liu, Z., Ding, L., Qiu, T., Chai, L., & Chen, X. (2017b). Systematic
Koumine on Human Breast Cancer Cells and the Mechanism Involved. Cell screening and characterization of multiple constituents in Guizhi Fuling capsule
Biochemistry and Biophysics, 72(2), 411–416. and metabolic profiling of bioactive components in rats using ultra-high-
Zhang, J. Y., Wang, Z. J., Li, Y., Liu, Y., Cai, W., Li, C., ... Qiao, Y. J. (2016). A strategy for performance liquid chromatography/quadrupole-time-of-flight mass
comprehensive identification of sequential constituents using ultra-high- spectrometry. Journal of Chromatography B, 1061-1062, 474–486.
performance liquid chromatography coupled with linear ion trap-orbitrap Zuo, M.-T., Liu, S.-S., Lin, L., Wang, Z.-Y., Bai, X., Sun, Z.-L., & Liu, Z.-Y. (2018).
mass spectrometer, application study on chlorogenic acids in flos Lonicerae Characterization of N -methylcanadine and N -methylstylopine metabolites in
japonica. Talanta, 147(62), 16–27. rat liver S9 by high-performance liquid chromatography/quadrupole time-of-
Zhang, S., Hu, S., Yang, X., Shen, J., Zheng, X., Huang, K., & Xiang, Z. (2015b). flight mass spectrometry. Rapid Communications in Mass Spectrometry, 32(23),
Development of a liquid chromatography with mass spectrometry method for 2047–2054.

16