Received: 22 December 2019 Revised: 15 February 2020 Accepted: 16 February 2020

DOI: 10.1002/rcm.8760

RESEARCH ARTICLE

Structural characterization and discrimination of Paris

polyphylla var. yunnanensis by a molecular networking strategy
coupled with ultra-high-performance liquid chromatography
with quadrupole time-of-flight mass spectrometry

Yumei Wang1 | Qian Fan1 | Jun Xiang2 | Haibo Huang1 | Sheng Chen1 |
Bairu Liu1 | Aizhi Wu1 | Cuixian Zhang1 | Li Rong1

1
School of Pharmaceutical Sciences,
Guangzhou University of Chinese Medicine, Rationale: Paris polyphylla var. yunnanensis (Franch) Hand Mazz (PPY) is a traditional
Guangzhou, China
Chinese medicine with antitumor, antibacterial, hemostatic, and anthelmintic
2
Pharmacy Department, Second Affiliated
Hospital of Guangzhou University of activities. Identification of the chemical composition in PPY is helpful to discover its
Traditional Chinese Medicine, Guangzhou, active ingredients and can be used to establish its quality control protocols.
China
Methods: The composition of PPY was identified using ultra-high-performance liquid
Correspondence chromatography combined with quadrupole time-of-flight mass spectrometry
L. Rong, School of Pharmaceutical Sciences,
Guangzhou University of Chinese Medicine, (UHPLC/QTOF-MS/MS) coupled with a molecular networking strategy. First, the
Guangzhou 510006, China. UHPLC/QTOF-MS/MS approach was optimized for chemical compound profiling.
Email: 498701887@qq.com
Then, the MS data were processed using PeakView™ combined with an in-house
C. X. Zhang, School of Pharmaceutical database to quickly characterize the secondary metabolites. Finally, molecular
Sciences, Guangzhou University of Chinese
Medicine, Guangzhou, China. networking excavated new molecular weights to discover unknown or trace natural
Email: zhangcuixian@gzucm.edu.cn products based on the characteristics of each cluster.

Funding information Results: A total of 222 compounds, including 77 isospirostanols, 2 spirostanols,

National Natural Science Foundation of China, 19 furostanols, 10 pseudospirostanols, 6 cholesterols, 10 C21 steroids, 5 insect
Grant/Award Numbers: 81603269, 81741160;
The Fire Plan Project of the Guangzhou metamorphosis hormones, 3 plant sterols, 6 five-ring triterpenoids, 4 flavonoids,
University of Chinese Medicine, Grant/Award 8 fatty acids, 2 phenylpropanoids, and 8 other compounds, were characterized in
Number: XH20170110; Project on
Construction of High Level University of PPY by comparing their main fragmentation characteristics and pathways with the
Guangzhou University of Chinese Medicine, literature data, and 62 of them, 54 steroidals and 8 phenylpropanoids, were
Grant/Award Number: 81; Medical Research
Foundation of Guangdong Province, Grant/ discovered or tentatively identified for the first time.
Award Number: A2016334; Science and Conclusions: This study extended the application of a molecular networking strategy
Technology Planning Project of Guangdong
Province, Grant/Award Number: to traditional herbal medicines and developed a molecular networking based
2017A020217008; NSFC, Grant/Award screening approach with a significant increase in efficiency for the discovery and
Numbers: 81603269, 81741160
identification of trace novel natural products.
Peer Review
The peer review history for this article is
available at https://publons.com/publon/10.
1002/rcm.8760

Yumei Wang and Qian Fan contributed equally to this work as first authors.

Rapid Commun Mass Spectrom 2020;34:e8760. wileyonlinelibrary.com/journal/rcm © 2020 John Wiley & Sons, Ltd. 1 of 17
https://doi.org/10.1002/rcm.8760
2 of 17
1 | I N T RO DU CT I O N
Paris polyphylla var. yunnanensis (Franch) Hand Mazz (PPY), called
Chong Lou, is a well-known, valuable traditional Chinese medicine
(TCM) that has been used to treat fractures, hemorrhages, snake
bites, and abscess for many years.1,2 PPY has been used as a main
ingredient for the preparation of many TCMs, such as Yunnan Baiyao,
Gongxuening capsules, and Jidesheng Sheyao tablets.3
Phytochemistry studies have indicated that the dried rhizomes of
Paris species contain steroidal saponins, triterpenoid saponins,
flavonoids, phenylpropanoids, ecdysteroids, phenolic glycosides, and
so on.4-6 The glucosyl groups of steroidal saponins are normally
attached to the steroidal aglycones through the hydroxyl groups at
the C-3 and/or C-26 positions (and sometimes the C-1, C-21, C-23,
6
and C-24 positions). Four types of steroidal sapogenins (spirostanol,
isospirostanol, furostanol, and pseudospirostanol) are considered
major aglycones, and typical sugars are arabinose, xylose, apiose,
fucose, rhamnose, glucose, and galactose.
Comprehensive identification of ingredients in PPY is critical to
understanding its biological mechanism and establishing quality
control protocols. Because of the complexity of these components,
their comprehensive analysis was extremely difficult. Since the mid-
2000s, because of its high separation resolution, enhanced separation
speed, unmatched sensitivity, and minimal sample preparation, liquid
chromatography/mass spectrometry (LC/MS) has been increasingly
used in natural products research.3,7-9 In addition, the low sample
quantity requirement makes LC/MS even more attractive to TCM
scientists.
Recently, Gerwick and co-workers10 developed Global Natural
Products Social Molecular Networking (GNPS) to accelerate the
discovery of compounds. GNPS is an emerging visual molecular
network that gives clear information on the material basis of
complex systems. It has allowed rapid comparison using MS
profiles of complex extracts and gained much attention in the
11
discovery of novel natural products. To date, the applications of
GNPS have led to the discovery of large numbers of natural
products such as spongonucleosides,12 aminopolyketide
derivatives,13 di- and tri-chlorinated acyl-amides,14 glycopeptides,15
monoterpene indole alkaloids,16 and cyclic lipopeptides,17 most of
which were metabolites from microbiota and rarely reported for
plant-derived natural products. Recently, structurally different
triterpene saponins from Eleutherococcus senticosus leaves and
phenolics in litchi pulp have been successfully revealed by
GNPS.18,19
This study implemented a systematic and universal
methodology based on the GNPS strategy coupled with ultra-high-
performance liquid chromatography with quadrupole time-of-flight
mass spectrometry (UHPLC/QTOF-MS/MS) to discriminate the
compounds in PPY. Based on the structures of the four steroidal
saponins, the major clusters, including steroidal saponins, were
successfully obtained by their MS/MS similarity, guiding the
discovery and structural characterization of more steroidal saponin
compounds from the extracts of PPY. In addition, the
WANG ET AL.
UHPLC/QTOF-MS/MS method using both positive- and negative-
ion modes was developed to quickly discriminate the compounds in
PPY. We summarized the fragmentation patterns to deduce
their steroidal saponin types. Moreover, 62 compounds were
discovered from the extracts of PPY for the first time—54
steroids (22 isospirostanol saponins, 16 furostanol saponins,
12 cholesterol saponins, and 4 insect metamorphosis hormones) and
8 phenylpropanoids. This study focuses on their chemical profile
analysis and will not only discover and identify trace novel natural
products, but also provide a basis for further research in
pharmacology and plant metabolism.
2 | M A T E R I A L S A N D M ET H O D S
2.1 | Chemicals
Reference standards of polyphyllin I (C-036-151122), polyphyllin II
(C-037-160623), polyphyllin VI (C-038-160126), and polyphyllin VII
(C-039-160623) were purchased from Chengdu Herbpurify Co., Ltd
(Chengdu, China). Their purities were above 98%. Methanol and
acetonitrile (LC/MS grade) were purchased from E. Merck (Darmstadt,
Germany). Formic acid with a purity of 99% (UHPLC grade) was
purchased from Anaqua Chemical Supply (ACS, Houston, TX, USA).
Deionized water (18.2 MΩ) was further purified using a Milli-Q
system (Millipore, Burlington, MA, USA). The other reagents were
analytical grade.
PPY samples were collected from the Wenshan region of Yunnan
Province, China, and were identified by Associate Professor Hai-Bo
Huang (Guangzhou University of Chinese Medicine, Guangzhou,
China), and the voucher of specimens was No. ChongLou-1. The
specimens were deposited in the Laboratory of Natural Products,
School of Pharmaceutical Sciences, Guangzhou University of Chinese
Medicine.
2.2 | Preparation of the sample solutions and
mixed standard solution
The decoction pieces of PPY rhizomes were crushed and ground
using a MM 400 mixer mill (Retsch GmbH, Haan, Germany) with
No. 40 mesh (pore size 0.250 mm). The PPY sample, 150 g of the
fine powder, was soaked in 600 mL of ethanol–H2O (VEtOH/
VH2O = 7:3) for 0.5 h at room temperature. It was then
ultrasonically extracted twice with ethanol–H2O (VEtOH/VH2O = 7:3,
w/v = 1:4), each for 0.5 h. The ethanol extracts were combined and
then filtered through three layers of filter paper. To the filtrate,
ethanol–H2O (VEtOH/VH2O = 19:1) was added in a filtrate–ethanol
ratio of 1:3 (v/v) to precipitate the polysaccharides and proteins.
This sample was kept overnight at 4
C, and after centrifugation
(4000 rpm for 20 min) in a DD-5M centrifuge (Bioridge Co.,
Shanghai, China), the supernatant ethanol filtrate was stored at
4
C. Next, the remaining dregs of the PPY sample were transferred
WANG ET AL.
into a round-bottomed flask (1000 mL), refluxed in a hot methanol
bath (600 mL) for 1.5 h, and filtered through four layers of gauze.
These operations were repeated twice. All the methanol extracts
were combined and then evaporated to dryness under reduced
pressure with a rotary evaporator at 50
C. The crude extracts were
reconstituted with an aqueous solution, and then petroleum ether
(PE) was added with a filtrate–PE ratio of 1:1 (v/v) to remove the
pigment and grease. After that, the supernatant ethanol filtrates
and aqueous solution after extraction with PE were combined and
evaporated to dryness at 50
C. The residue was suspended in H2O
and then successively partitioned with ethyl acetate (EtOAc) and n-
butanol (n-BuOH) (1:1, v/v, three times each), and the extraction
solutions were evaporated to dryness at 50
C. After the EtOAc
(0.9318 g) and n-BuOH (4.8143 g) extracts were obtained, they
were placed away from light and stored at 4
C (Figure S1,
supporting information). Afterward, the residue of the n-BuOH
extracts was dissolved in a methanol solution to obtain a
concentration of 10 mg mL−1. Through a 0.22-μm syringe filter,
2 μL of the sample solution was injected into the UHPLC/QTOF-
MS/MS system for analysis.
For the qualitative identification of the main constituents of PPY
extracts, the standard stock solutions of four reference standards
(polyphyllin I, polyphyllin II, polyphyllin VI, polyphyllin VII) were
dissolved in methanol. The appropriate amount of each standard
stock solution was then mixed and diluted to a concentration of
0.2 mg mL−1. Through a 0.22-μm syringe filter, 2 μL of the mixed
standard solution was injected into the UHPLC/QTOF-MS/MS
system for analysis.
2.3 | Liquid chromatographic and mass
spectrometric conditions
2.3.1 | Chromatographic separation
LC was performed on an LC-30A UHPLC system (Shimadzu, Kyoto,
Japan), consisting of an autosampler (Model SIL-30SD), an LC-30AD
binary pump, an online degasser (DGU-20A5R), and a temperature
controller module for the column (CTO-30A).
The separation was performed on a Shim-pack XR-ODSII column
(2.1 mm × 100 mm, 2.2 μm, Shimadzu). The column temperature was
set at 40
C, and the flow rate was 0.4 mL min−1. Mobile phase A
consisted of a 0.1% formic acid in acetonitrile, and mobile phase B
was 0.1% formic acid aqueous solution. The gradient elution program
was as follows: 0–1 min, 25% A, 75% B; 1–1.5 min, 25%–35% A,
75%–65% B; 1.5–3 min, 35%–40% A, 65%–60% B; 3–11 min, 40%–
42% A, 60%–58% B; 11–13 min, 42% A, 58% B; 13–16 min, 42%–
50% A, 58%–50% B; 16–24 min, 50%–55% A, 50%–45% B;
24–27 min, 55%–60% A, 55%–40% B; 27–30 min, 60%–65% A,
40%–35% B; 30–32 min, 65% A, 35% B; 32–34 min, 65%–70% A,
35%–30% B; 34–36 min, 70%–85% A, 30%–15% B; 36–38 min,
85%–98% A, 15%–2% B; 38–41 min, 98% A, 2% B. The sample
injection volume was 2 μL.
3 of 17
2.3.2 | Mass spectrometry
MS was performed on a Triple TOF 5600+ system (AB SCIEX, Concord,
Ontario, Canada) equipped with a TurboIonSpray interface, which has a
dynamic background subtraction function that triggers information-
dependent acquisition (IDA) MS/MS mode and could simultaneously
combine a TOF-MS survey scan with a maximum of eight
corresponding candidate MS/MS events. The MS method consisted of
a TOF-MS scan and IDA-fragmentation TOF-MS/MS scan with mass
ranges of m/z 50–1500 in positive and negative electrospray ionization
modes. The parameter settings of TOF-MS scan were optimized as
follows: ion spray voltage floating, +5500/−4500 V; turbo spray
temperature, 550
C; nebulizer gas (gas 1) pressure, 55 psi; heater gas
(gas 2) pressure, 55 psi; curtain gas pressure, 35 psi; declustering
potential, ±100 V; and collision energy (CE), ±5 eV. The parameter
settings of TOF-MS/MS scan were optimized as follows: declustering
potential, ±75 eV; CE, ±50 eV; and CE spread, 15 eV. The Triple TOF
system was equipped with an automated calibration system for mass
accuracy (calibrant delivery system), and recalibration was carried out at
intervals of 4 h. In addition, dynamic background subtraction and IDA
techniques were used to reduce the impact of matrix interference and
also increase the efficiency of analysis.
2.4 | Molecular networking
Molecular networking of the metabolites from the extracts of PPY
was created using the online workflow at GNPS (http://gnps.ucsd.
edu). The data were filtered by removing all MS/MS peaks within
±17 m/z units of the precursor ion m/z value. The MS/MS spectra
were window filtered by choosing only the top six peaks in the
±50 m/z units window throughout the spectrum. The data were then
clustered using MS-cluster with a precursor ion mass tolerance of
0.02 m/z units and an MS/MS product ion tolerance of 0.02 m/z units
to create consensus spectra. Consensus spectra that contained fewer
than two spectra were discarded. A network was then created, where
edges were filtered to have a cosine score of above 0.7 and more
than six matched peaks. Further edges between two nodes were
retained in the network only if each of the nodes appeared in each
other's respective top 10 most similar nodes. The spectra in the
network were then searched against GNPS spectral libraries. The
library spectra were filtered in the same manner as the input data. All
matches kept between network spectra and library spectra were
required to have a score above 0.7 and at least six matched peaks.
The molecular network was visualized using Cytoscape (www.
cytoscape.org; version 3.6.1).
2.5 | Data analysis strategy
PeakView software (version 1.2, AB SCIEX) was used for the
qualitative identification of PPY. First, a database of chemical
compounds belonging to PPY, which includes names, molecular
4 of 17
formulae, and chemical structures, was established by searching
relevant reported literature and chemistry database websites. The
database was then imported into the XIC Manager module in
PeakView to build a list of target compounds for performing the
function of XIC calculation. The parameters of the method were set
as follows: intensity > 300 counts, S/N > 10, isotope ratio%
difference ≤ 20, and mass error ≤ 10 ppm. After carrying out the
calculation, the compounds that met the requirements of the
database were extracted and highlighted. The MS fragmentation
profiles of each compound were then obtained based on the relative
intensity of precursor and product ions. By comparison with standard
literature information, such as retention time and mass spectrometry
product ions, the compounds were finally identified as those existing
in the extracts of PPY.
3 | RESULTS AND DISCUSSION
3.1 | Optimization of chromatographic separation
and mass spectrometric detection
The extracts of PPY were a complex mixture containing a range of
compounds. Thus, positive- and negative-ion mode MS were
necessary to obtain as comprehensive spectral information as possible
for these extracts. CEs of ±20, ±30, ±40, and ± 50 eV were
investigated to maximize the production of appropriate product ions.
Finally, taking standards of polyphyllin I, polyphyllin II, polyphyllin VI,
and polyphyllin VII as references, ±50 eV was selected as the optimal
CE because of the adequate number of product ions for structure
analysis. Other MS parameters, such as turbospray temperature,
nebulizer gas pressure, curtain gas pressure, heater gas pressure, ion
spray voltage, and declustering potential, were also optimized to
obtain a better response.
3.2 | Identification of compounds in PPY
With the aid of an in-house chemical library of the genus Paris
(Table S1, supporting information), 160 compounds—77
isospirostanols, 2 spirostanols, 19 furostanols, 10 pseudospirostanols,
6 cholesterols, 10 C21 steroids, 5 insect metamorphosis hormones,
3 plant sterols, 6 five-ring triterpenoids, 4 flavonoids, 8 fatty acids,
2 phenylpropanoids, and 8 other compounds—were successfully
characterized from the extracts of PPY based on the common
fragmentation pathways, reference standards, and comparison with
the data in the literature.5,20-38 Their structures are shown in
Figure S2 (supporting information). All three batches of PPY
samples were analyzed with the developed UHPLC/QTOF-MS/MS
method. The base peak chromatograms of the PPY sample in
negative- and positive-ion mode obtained using UHPLC/QTOF-
MS/MS are shown in Figures 1A and 1B. The detailed information
for these reported components is summarized in Table S2
(supporting information). The constituents identified in PPY were
WANG ET AL.
mainly classified into four types: steroids, triterpenoid saponins,
flavonoid glycosides, and other compounds (fatty acids,
anthraquinones, phenylpropanoids, etc.). Steroidal saponins were
the most important components. Many studies have reported the
fragmentation mechanism of steroidal glycosides in the rhizomes of
genus Paris,4,20-22 and all these were very helpful for identifying
components in the extracts of PPY.
In this investigation, we summarized the characteristic
fragmentation patterns of compounds listed in the library constructed
from literature searches. The exact molecular weight and elemental
composition of steroidal glycosides can be obtained from the [M − H]−,
[M + HCOO]−, [M + H]+, and [M + Na]+ ions. The masses of pairs of ions
separated by 22 m/z units ([M + HCOO]− – [M + Na]+), 24 m/z units
([M + Na]+ – [M−H]−), and 46 m/z units ([M + HCOO]− – [M − H]−)
provide diagnostic information about the precursor ions. The classes
and number of sugar units can be inferred from the loss of 162 Da
(Glc/Gal), 146 Da (Rha/Fuc), and 132 Da (Ara/Xyl) neutral
fragments.23,24 In the MS/MS ESI (+) mode, the diosgenin-type
aglycone showed diagnostic ions at m/z 415, 397, 283, 271, and
253, whereas the detection of the diagnostic ions at m/z 431, 413, and
395 and an obvious neutral loss of 144 Da (C8H16O2) indicated the
presence of a pennogenin-type aglycone, which is attributed to a
substituent on the C-20 or C-27 position of the aglycone. Meanwhile,
if the loss of a glucose unit is detected with the diagnostic ions at
m/z 431, 413, 271, and 253 as well as an obvious neutral loss of
160 Da (C8H16O3), these saponins are usually nuatigenin-type
aglycones. However, the presence of diagnostic ions at m/z 253, 271,
or 287, 269, 251 indicates that the saponin is rarely present in the
furost-type aglycones typically observed in spirostanol saponins in the
MS/MS ESI (+) mode.22,25 A strategy for characterizing steroidal
saponin structures in P. polyphylla was then proposed (Figure S3,
supporting information) and further verified by studying the reference
standards of four diosgenin saponins. The identification procedures of
four major types of compounds are described in more detail in the
following sections.
3.2.1 | Identification of steroids
Polyphyllin I, polyphyllin II, polyphyllin VI, and polyphyllin VII were the
major active compounds that belong to isospirostanol saponins
existing in PPY in huge amounts, and they normally generated strong
signals in the positive-ion mode. Experimental results suggested that
their fragmentation mainly occurred in the lactone ring in ESI+ and
produced a complex product ion spectrum. Polyphyllin II was taken as
an example to illustrate the process of identification. Peak 201
exhibited a [M + H]+ ion at m/z 1015.5354 in the positive-ion mode
with a retention time of 30.229 min and generated a series of
characteristic product ions at m/z 869.4810, 723.4269, 577.3706,
and 415.3194 (Figures 2A and 2B) due to the successive loss of three
Rha units and one Glc unit. The diagnostic ions at m/z 397.3091 and
283.2410 were formed by the successive elimination of one water
molecule (18 Da) and a neutral fragment C6H10O2 (114 Da) from the
WANG ET AL.
F I G U R E 1 Base peak chromatograms (BPCs) of the Paris polyphylla var. yunnanensis (PPY) extracts obtained using UHPLC/Q-TOF-MS/MS in
A, ESI [−] mode and B, ESI [+] mode [Color figure can be viewed at wileyonlinelibrary.com]
ion at m/z 415.3206, respectively. The other diagnostic ions at m/z
271.2049 and 253.1948 resulted from the successive elimination of a
neutral fragment C8H16O2 (144 Da) and H2O (18) from m/z
415.3206. By comparing with a standard, peak 201 was definitely
+
identified as polyphyllin II. Similarly, the compounds with [M + H]
ions at m/z 855.4658 at 31.228 min, m/z 739.4234 at 26.014 min,
and m/z 1031.5347 at 24.574 min were identified as polyphyllin I,
polyphyllin VI, and polyphyllin VII, respectively. Peaks 21, 25, 30, 31,
39, 43, 65, 67, 71, 72, 75, 76, 78, 82, 85, 94–97, 99, 101, 103, 107,
110, 113, 117, 121, 122, 128, 130, 137, 138, 141, 143–148,
150–158, 160, 162, 168, 169, 171, 179, 182–187, 189, 190, 193,
196, 198, 202–206, 210, 211, 213, and 214 were similarly identified
and reported (references in Table S2, supporting information). Peak
+
37 showed a [M + H] ion at m/z 1063.4966, and product ions at m/z
1045.4831 [M + H–H2O]+, 1009.4503 [M + H–3H2O]+, 899.4241
[M + H–Xyl–CH3OH]+, 881.4151 [M + H–Xyl–CH3OH–H2O]+,
+
767.3807 [M + H–Xyl–Rha–H2O] , 621.3242 [M + H–Xyl–Rha–Fuc–
+ +
H2O] , and 459.2745 [M + H–Xyl–Rha–Fuc–Glc–H2O] were also
observed, together with diagnostic ions m/z 441.2637, 423.2526, and
5 of 17
393.2410. Based on the product ion spectrum, peak 37 was
tentatively identified as parisyunnanoside H, and this assignment was
validated by comparison with the relevant literature.26 Peak 51
showed product ions at m/z 1193.5893, 1047.5361, 901.4766,
739.4248, 593.3680, and 413.3063, and it was similarly identified as
an isomer of parisyunnanoside G.27
The furostanol, pseudospirostanol, and cholesterol steroid
saponins were three of the main active substances found in PPY. In
the positive-ion mode, peak 127 produced a [M + H]+ ion at m/z
1177.5808 at 18.122 min, six major MS2 product ions at m/z
1031.5368 [M + H–Rha] + , 885.4800 [M + H–2Rha] + , 869.4844
[M + H–Glc–Rha] + , 723.4275 [M + H–Glc–2Rha] + , 577.3704
+
[M + H–Glc–3Rha] , and 415.3190 [M + H–2Glc–3Rha]+, and
diagnostic ions at m/z 271.2066 and 253.1953. Based on this
information, peak 127 was tentatively identified as pseudoproto-Pb,
and this assignment was validated by comparison with the relevant
literature.27 Similarly, peaks 34, 46, 49, 56, 58, 60, 73, 74, 80, 86, 88,
93, 98, 100, 108, 127, 131, and 161 could be tentatively identified
(references in Table S2, supporting information). In addition, peak 114
6 of 17
F I G U R E 2 A, Proposed fragmentation pathways of peak 201 (polyphyllin II). B, Characteristic ions of peak 201 (polyphyllin II) [Color figure
can be viewed at wileyonlinelibrary.com]
produced a [M + H]+ ion at m/z 1047.5351 at 16.956 min, generating
+
prominent product ions at m/z 885.4834 [M + H–Glc] , 739.4248
[M + H–Glc–Rha]+, 721.4291 [M + H–Glc–Rha–H2O]+, 593.3687
[M + H–Glc–2Rha]+, and 431.3160 [M + H–2Glc–2Rha]+, with
diagnostic ions at m/z 413.3054, 395.2946, 293.1232, 271.2056,
and 253.1949, respectively. Thus, peak 114 was tentatively
identified as 26-O-glc-nuatigenin-3-O-rha(1 ! 2)[rha(1 ! 4)]-glc
WANG ET AL.
by comparison with the literature data.28 Moreover, peaks 82, 96, 102,
106, 109, 134, 164, 167, and 199 were similarly identified. Peak 69,
with a [M + H]+ ion at m/z 1191.5763 (Rt = 10.765 min), generated
prominent product ions at m/z 1011.5098 ([M + H–Glc–H2O]+,
865.4550 [M + H–Glc–Rha–H2O]+, 719.3968 [M + H–Glc–2Rha–H2O]+,
573.3411 [M + H–Glc–3Rha–H2O]+, and 411.2886 [M + H–2Glc–3Rha–H2O]+,
and diagnostic ions at m/z 393.2784, 285.1853, and 267.1747.
WANG ET AL.
Thus, peak 69 was tentatively identified as parispseudoside C by
comparison with the literature data.29 Furthermore, peaks 136,
139, 140, 192, and 200 were similarly identified (references in
Table S2, supporting information).
The C21 steroids, insect metamorphosis hormones, and plant
sterols were the other three types of important steroid substances
distributed in PPY. However, these types of compounds were
susceptible to the successive neutral loss of one or more H2O
molecules in the MS/MS spectrum. In the negative-ion mode, peak 22
produced a [M − H]− ion at m/z 977.4322 at 4.383 min and
−
four major MS ions at m/z 845.3916 [M − H–Xly] , 797.3685
2
[M − H–Glc–H2O]−, 665.3218 [M − H–Xly–Glc–H2O]−,
519.2618 [M − H–Xly–Rha–Glc–H2O]−. Therefore, peak 22 was
tentatively identified as parisyunnanoside J by comparison with
the literature data,26 and peaks 26, 40, 63, 105, 115, 120, 123,
176, and 178 were also similarly identified (references in
Table S2, supporting information). In addition, peak 12 produced
the adduct ion [M + HCOO]− at m/z 541.3051 and
−
deprotonated molecule [M − H] at m/z 495.2990. In the
MS/MS spectrum, two product ions were detected at m/z
477.2875 [M − H–H2O]− and 441.2674 [M − H–3H2O]−. Thus, peak
12 was tentatively identified as turkesterone, and this assignment was
validated by comparison with the relevant literature. 30
Furthermore,
peaks 19, 20, 27, and 42 were similarly identified. Peak 219 produced
the adduct ion [M + HCOO]− at m/z 945.5511 and a [M − H]− ion at
m/z 899.5475 and yielded relatively major product ions at m/z
737.4913 [M − H–Glc]−, 719.4819 [M − H–Glc–H2O]−, and 575.4369
[M − H–2Glc]−. Thus, peak 219 was tentatively identified as
pariposide F, and this assignment was validated by comparison with the
relevant literature.31 Meanwhile, peaks 221 and 222 were similarly
identified (references in Table S2, supporting information).
3.2.2 | Identification of triterpenoid saponins
Six triterpenoid saponins were detected in the extracts of PPY
providing spectra in both positive- and negative-ion modes. They
were susceptible to successive loss of sugar groups, the neutral loss
of H2O (18 Da) and HCOOH (46 Da), and, finally, the formation of
oleanane aglycones. In the negative-ion mode, peak 215 produced the
adduct ion [M + HCOO]− at m/z 797.4371 and a [M − H]− ion at
m/z 751.4265, and it generated product ions at m/z 705.2944
− −
[M − H–HCOOH] and 573.3747 [M − H–HCOOH–Ara] .
Therefore, peak 215 could be assigned as 3β,23-diol-oleane-
12-en-28-oic acid-3-O-glc(1 ! 4)-ara by comparison with the
literature reports.5 Furthermore, peaks 68, 92, 208, 209, and 220 were
similarly assigned (references in Table S2, supporting information).
3.2.3 | Identification of flavonoid glycosides
Only four flavonoid glycoside compounds were detected in the
extract of PPY. These compounds can yield very strong [M − H]
7 of 17
ions in negative-ion ESI-MS. In the MS/MS spectrum, they were
susceptible to continuous neutral losses of 18 Da (H2O), 32 Da
(CH3OH), 146 Da (Rha), and 162 Da (Glc). In the negative-ion mode,
peak 17 produced the adduct ion [M + HCOO]− at m/z 685.1668
and the deprotonated molecule [M − H]− at m/z 639.3429, and it
generated MS2 ions at m/z 477.2855 [M − H–Glc]−, 459.2754
[M − H–Glc–H2O]−, and 315.0502 [M − H–Glc–Glc]. Therefore,
peak 17 was tentatively identified as isorhamnetin-3-O-
gentiobioside, and these data were consistent with the literature
reports.32,33 Moreover, peaks 9, 16, and 23 were similarly
assigned as flavonoid glycoside compounds (references in
and Table S2, supporting information).
3.2.4 | Identification of other compounds
In the negative-ion mode, eight fatty acid compounds, eight other
the compounds, and two phenylpropanoids were detected in the extract
of PPY, some of which were isomers. They were susceptible to
successive MS/MS losses of 18 Da (H2O), n × 14 Da (−(CH2)n),
and 46 Da (HCOOH). In particular, peak 173 produced a prominent
[M − H]− ion at m/z 343.2486, which generated MS2 ions at m/z
297.2410 [M − H–HCOOH]− and 279.2310 [M − H–HCOOH–H2O]−.
Therefore, peak 173 was tentatively identified as methyl-
(9S,10R,11S)-triol-12(Z)-octadecenoate by comparison with the
relevant literature.34 Moreover, peaks 133, 149, 174, 195, 212, 217,
and 218 were tentatively assigned as fatty acid compounds
(references in Table S2, supporting information). Similarly, peaks
1–5, 47, 64, and 118 were also tentatively identified.35,36
Moreover, peaks 32 and 45 were assigned as flavonoid glycoside
phenylpropanoids.37,38
3.3 | MS-based molecular networking of the PPY
extracts
Molecular networking of the PPY extracts based on the MS/MS
spectral similarity was generated by GNPS, which led to the presence of
precursor ions visualized as nodes in the molecular map. A total of
77 known compounds and 62 new compounds were identified and
dereplicated through the GNPS library and in-house database. More
details can be found on the GNPS website (https://gnps.ucsd.edu/
ProteoSAFe/status.jsp?task = a35dedca42924592bd2b07653bb5d753
and https://gnps.ucsd.edu/ProteoSAFe/status.jsp?task=65a864e5e6c3
4d2287677ee8d39a2a87). The results revealed the presence of five
distinct steroid clusters and one phenylpropanoid cluster in the
negative-ion mode and three obvious steroid clusters in the positive-
ion mode (Figures S4A and S5, supporting information).Meanwhile,
the unreported chemical structures of each representative cluster are
also shown in Figure 3. This study indicated that isospirostanol
saponins were the major steroid compounds in PPY,39 and they
−
were selected for further investigation.
8 of 17
FIGURE 3 Structures of the new compounds identified from the extracts of PPY
3.3.1 | Identification of steroidal saponins in
molecular networking
In the negative-ion mode, cluster I contains 57 nodes, and precursor
ions from m/z 897 to 1133 were laid out directly through the
Cytoscape software. These were attributed to the presence of a
series of isospirostanol, furostanol, and cholesterol saponins, as
shown in Figure 4A. In particular, by comparison with the authentic
standard compound (Table S3, supporting information, and Figure 2B),
the m/z 1013.52 [M − H]− and 899.469 [M + HCOO]− nodes were
identified as the known compounds polyphyllin II (peak 201) and
polyphyllin I (peak 207, Figure S4B[a], supporting information),
respectively. Meanwhile, as shown in Figure 4A, the ions at m/z
897.456, 903.502, and 919.488 correlated with the standard
compound ion at m/z 899.469 (peak 207), indicating that they were
structurally correlated, and they had the same aglycone as
polyphyllin I, which originated from a series of losses of sugar units
[−Ara−Rha−Glc] (MS/MS in Figures S4B[b–d], supporting
information, and Table 1). Based on the structural features of PPY in
30
the relevant literature and the precursor ions that were 2 m/z units
lower than m/z 899.469, the m/z 897.456 node was explained as the
C26–C27 double bond formation product of PPY-58 (peak 188).
According to the structural features of PPY in related literature6 and
their precursor ions 4 m/z units higher than m/z 899.469, the m/z
903.502 node was elucidated as the C16–C22 and C22–C26 cleavage
product of peak 207, which was tentatively identified as PPY-52
(peak 166). Finally, based on the structural feature of PPY in the
relevant literature40 and the precursors 16 m/z units higher than m/z
903.502, the m/z 919.488 node was determined as a 16-hydroxyl
derivative of PPY-52 (PPY-53). In addition, other series of m/z
WANG ET AL.
1079.54 and 1093.585 ions were clustered. By comparison with the
structures of reported compounds22 and the database (Tables S1 and
S3, supporting information), the m/z 1093.585 node was identified as
parisaponin I (peak 80). Moreover, the precursor ions 14 m/z units
lower than m/z 1093.585 and the m/z 1079.54 nodes produced the
[M + HCOO]− ion at m/z 1079.5352 (peak 87) and the major MS2
ions at m/z 901.4863 [M − H–Ara]−, 887.4711 [M − H–Rha]−,
755.4258 [M − H–Rha–Ara]−, and 593.3712 [M − H–Ara–Rha–Glc]−
(MS/MS in Figure S6, supporting information). These implied that the
14 m/z unit difference in the precursor ions was caused by a change
from Rha to Ara on the sugar chain at C-3 (Table 1 and Figure 4A).
Therefore, the m/z 1079.54 node was definitely identified as PPY-35
(peak 87). With the aid of GNPS, 30 steroids with 2–9 degrees of
polymerization were identified by cluster I of PPY (Table 1 and
Figure 4A). Among them, 18 steroids were discovered from PPY for
the first time. Meanwhile, 11 steroids with 2–8 degrees of
polymerization were identified by cluster II of PPY (Figure 4B; and
Figure S4C, supporting information), and five steroids among them
were discovered from PPY for the first time (Table 1; and Figure S6,
supporting information). Furthermore, cluster III in the positive-ion
mode was attributed to the presence of a series of isospirostanol,
furostanol, and cholesterol saponins, mostly containing tri-, tetra-, and
pentasaccharides, as shown in Figure S5A (supporting information). As
a result, 18 steroid saponins were discovered from PPY for the first
time (Table 1; and Figure S6, supporting information).
In the positive-ion mode, cluster IV contains a total of 11 nodes,
ranging from m/z 739 to 804, which were due to the presence of a
series of isospirostanol and furostanol saponins (see Figure S5B,
supporting information). In particular, by comparison with the
authentic standard compound (Table S3, supporting information), the
WANG ET AL.
F I G U R E 4 Total of molecular networking of the extracts of PPY for A, cluster I and B, cluster II [Color figure can be viewed at
wileyonlinelibrary.com]
m/z 756.45 [M + NH4]+ and 739.424 [M + H]+ nodes were both
confidently identified as the known compound polyphyllin VI (peak
177). Moreover, by comparison with the structures of reported
compounds23 and the database (Tables S1 and S3, supporting
information), the m/z 758.43 and 742.434 nodes were identified as
(3β,17α,25S)-spirost-5-ene-3,17,27-triol-3-O-ara(1 ! 4)-glc (peak
107) and pennogenin-3-O-ara(1 ! 4)-glc23 (peak 190), respectively.
9 of 17
Meanwhile, as shown in Figure S5B (supporting information), the
unknown ion at m/z 788.439 correlated with the known ion at m/z
758.43, indicating that they were structurally related. Based on the
relevant literature and the structural features of peak 107,31 the m/z
788.439 node was determined to be a 7-methoxy derivative of peak
107 (PPY-21 or PPY-44). Finally, based on the precursor ion 2 m/z
units higher than m/z 788.439, the m/z 790.456 node was elucidated
T A B L E 1 New compounds identified using ultra-high-performance liquid chromatography combined with quadrupole time-of-flight mass spectrometry (UHPLC/QTOF-MS/MS) from Paris
polyphylla var. yunnanensis (Franch) Hand Mazz (PPY) extracts
10 of 17

m/z [M Error Component

No. Rt (min) Formula Mass (Da) + HCOO]− (ppm) (−) MSE m/z (+) MSE m/z name Structure type Cluster
6 2.729 C24H28O13 524.4750 523.1260 0.63 487.1482, 443.2574, 419.1929, 525.1095 PPY-1 Phenylpropanoids VI
341.1055, 307.0793, 297.1133,
248.1001, 181.0499, 145.0288
7 2.787 C24H24O11 488.1319 533.1549 −0.28 487.1484, 443.2526, 371.2310, 489.1594 PPY-2 Phenylpropanoids VI
341.1115, 307.0843, 217.0499,
179.0561, 163.0395, 145.0291
8 2.889 C25H26O12 518.1424 563.1648 1.09 527.1303, 503.1239, 473.1031, 536.1963 PPY-3 Phenylpropanoids VI
455.1072, 425.0821, 395.0820,
383.0792, 353.0721, 323.0621
10 3.123 C56H90O26 1224.5411 1269.5521 0.78 1177.5878, 1045.5347, 1031.5293, 1241.5747, 1061.4801, 1043.4662, PPY-4 Isospirostanol II
899.4720, 883.4783, 737.4138, 881.4160, 749.3729, 603.3150,
557.3484 459.2733, 441.2625, 423.2528
11 3.163 C52H84O24 1092.5353 1137.5106 −0.95 1091.5040, 959.4609, 945.4473, 1110.5485 PPY-5 Cholesterol I
813.3990, 747.3589, 651.3386
13 3.462 C52H80O24 1088.5040 1133.5016 0.50 1087.4556, 1033.5344, 901.4898, 1106.5279 PPY-6 Cholesterol I
887.4727, 755.4273, 737.4158
14 3.537 C33H54O12 642.3615 687.3653 −0.55 641.3600, 623.3493, 479.3043, 643.3075, 589.3335, 533.2740, PPY-7 Insect IX
461.2932, 319.1924, 301.1814 481.3155, 463.3046, 445.2945, metamorphosis
371.2218 hormone
15 3.658 C51H82O24 1078.5196 1123.4935 2.73 1077.4921, 945.2283, 931.4344, 1101.5013 PPY-8 Isospirostanol I
799.3860, 733.3509, 637.3303
18 3.868 C27H44O9 512.2985 557.2888 −2.55 511.2901, 493.2810, 415.2609, 513.3057 PPY-9 Insect IX
351.1821, 333.1711, 265.1402 metamorphosis
hormone
24 4.757 C39H62O15 770.4089 815.4122 7.57 769.4106, 623.3481, 605.3356, 788.4156, 735.3955, 589.3350, PPY-10 Isospirostanol IV
461.2924, 445.3014, 407.1873 445.2963, 427.2846, 391.2653
28 5.138 C50H80O22 1032.5141 1033.5193 −0.64 1077.5239, 10131.5165, 899.4722, 1015.5134, 853.4532, 835.4469, PPY-11 Furostanol III
881.4627, 753.4131, 737.4160, 721.4132, 575.3566, 431.3153,
687.3836, 591.3560, 573.3485, 413.3046, 395.2945, 279.1079
437.1319, 307.1021
29 5.214 C31H36O16 664.2003 709.2022 6.71 663.1994, 595.1582, 547.2357, 687.1892 PPY-12 Phenylpropanoids VI
517.1603, 499.1480, 487.1484,
403.1388
33 5.466 C30H34O15 634.1898 679.1930 0.38 487.1485, 469.1296, 453.1089, 652.3597, 520.3392, 472.2886, PPY-13 Phenylpropanoids VI
341.1095, 307.08, 235.07, 454.2803, 340.1862
145.0284
35 5.864 C50H80O22 1032.5141 1050.5458 −3.99 PPY-14 Furostanol III
(Continues)
WANG ET AL.
TABLE 1 (Continued)

m/z [M Error Component

No. Rt (min) Formula Mass (Da) + HCOO]− (ppm) (−) MSE m/z (+) MSE m/z name Structure type Cluster
WANG ET AL.

1077.5444, 1031.5282, 899.4758, 1015.5067, 883.4675, 853.4533,

881.4655, 753.4153, 735.4084, 835.4427, 721.4128, 575.3564,
619.3506, 591.3579 413.3037, 395.2937, 271.2048,
263.1943

36 5.930 C44H70O19 902.4511 947.4472 −1.05 901.4537, 859.3698, 815.3749, 885.4408, 753.4015, 727.3519, PPY-15 Isospirostanol I
769.4070, 755.3928, 727.3223, 607.3445, 589.3326, 445.2934,
623.3470, 557.3118, 491.2667, 427.2830, 409.2731, 391.2604
447.2781, 403.2121
38 6.136 C31H36O16 664.2003 709.2004 4.12 663.1968, 517.1578, 499.1454, 687.1892 PPY-16 Phenylpropanoids VI
487.1440, 337.0915, 265.0719
41 6.585 C28H42O9 522.2829 567.2833 0.01 521.2841, 475.2713, 403.2489, 540.3212,452.1517, 378.2851, PPY-17 Insect IX
385.2366, 315.1596, 243.1360 260.0801 metamorphosis
hormone
44 6.690 C32H38O17 694.2109 739.2144 −1.80 693.2120, 517.1596, 499.1496, 717.1987, 627.2108, 589.3308, PPY-18 Phenylpropanoids VI
337.0928, 319.0730, 265.0699, 555.1469, 537.1362, 522.1009,
175.0393 451.1080, 409.2803, 379.1023,
343.0773
48 7.351 C32H38O17 694.2109 739.2141 8.25 693.2114, 667.3263, 517.1611, 712.3681, 515.1541, 303.0863, PPY-19 Phenylpropanoids VI
499.1490, 473.2457, 353.0530 285.0749, 243.0623, 177.0545
50 39.382 C27H42O7 478.2931 523.2936 1.06 477.2867, 459.2771, 443.2169, 479.2998 PPY-20 Insect IX
361.1910, 317.1750, 274.1208, metamorphosis
199.0982 hormone
52 8.348 C39H62O15 770.4089 815.4082 2.74 769.4075, 751.3974, 637.3266, 788.4156, 753.4054, 735.3955, PPY-21 Isospirostanol IV
607.3539, 589.3444, 507.2668, 589.3350, 573.3379, 445.2963,
491.2672, 427.2888, 383.1148, 427.2846, 391.2653
329.2343
53 8.414 C56H90O25 1162.5800 1163.5804 0.38 1161.5819, 1015.5205, 997.5107, 1127.5642, 1031.5465, 1001.5302, PPY-22 Furostanol III
883.4798, 865.4658, 737.4185, 869.4783, 723.4299, 577.3719,
671.3908 415.3195
54 8.423 C45H72O19 916.4730 917.4723 −0.79 869.4603, 737.4159, 591.3561, 899.4511, 755.4204, 593.3698, PPY-23 Furostanol III
525.3241, 289.0926, 247.0819 575.3581, 431.3162, 413.3053,
395.2942
55 8.471 C38H60O15 756.3932 774.4295 7.29 801.3994, 755.3943, 737.3823, 712.3681, 515.1541, 303.0863, PPY-24 Isospirostanol Individual
734.9970, 623.3618, 605.3387, 285.0749, 243.0623, 177.0545 node
587.3240, 443.2811, 425.2699
57 8.695 C50H82O20 1002.5399 1020.5313 1.24 1001.5338, 837.4630, 723.4305, 984.5140, 858.4795, 712.3581, PPY-25 Cholesterol III
577.3674 562.3673, 416.3240, 398.3128,
272.2090
(Continues)
11 of 17
TABLE 1 (Continued)

m/z [M Error Component

12 of 17

No. Rt (min) Formula Mass (Da) + HCOO]− (ppm) (−) MSE m/z (+) MSE m/z name Structure type Cluster
60 8.792 C57H92O27 1208.5826 1226.6129 2.39 1161.5819, 1015.5205, 883.4798, 1209.5847, 1191.5735, 1029.5227, PPY-26 Furostanol III
865.4658. 737.4185 883.4653, 737.4073, 591.3508,
429.2990, 411.2893
61 9.303 C50H80O22 1032.5141 1077.5214 −1.23 1031.5216, 899.4758, 881.4655, 1015.5067, 853.4533, 835.4427, PPY-27 Isospirostanol III
867.4514, 753.4152, 735.4084, 721.4128, 575.3564, 413.3037
591.3579
62 9.388 C51H82O22 1046.5298 1091.5363 −1.98 1045.5340, 899.4757, 881.4674, 1011.5120, 885.4789, 739.4239, PPY-28 Furostanol III
753.4131, 735.3996, 591.3553, 723.4291, 577.3715, 415.3195,
527.3510 397.3099, 271.2053
66 10.026 C39H64O15 772.4245 790.4561 0.16 817.4285, 771.4227, 609.3666, 755.4185, 737.4102, 611.3810, PPY-29 Furostanol IV
493.2508, 447.3145 593.3665, 575.3585, 449.3251,
431.3153, 413.3044, 395.2944,
377.285
70 10.784 C57H90O26 1190.5720 1235.5821 1.70 1189.5802, 1171.5661, 1043.5198, 1173.5687, 1011.5098, 865.4550, PPY-30 Isospirostanol II
1025.5046, 897.4584, 861.4420, 719.3968, 573.3411, 411.2886,
751.3988, 715.3774, 571.3303, 393.2784, 293.1228
393.1410, 315.1304
77 13.553 C45H72O18 900.4719 945.4698 0.87 899.4696, 869.4246, 767.3950, 901.4749, 883.4581, 739.4245, PPY-31 Isospirostanol III
753.4121, 737.4208, 621.3360, 577.3719, 433.2590, 395.2941,
573.3473, 555.3013, 459.2797 287.2007, 269.1898, 251.1797
79 14.612 C38H60O15 756.3932 774.4221 −0.12 755.3850, 623.3384, 477.2778, 755.3850, 623.3384, 593.3722, PPY-32 Isospirostanol Individual
357.2634 477.2778, 357.2634 node
81 14.835 C45H72O18 900.4719 945.4706 −1.52 899.4755, 813.4112, 767.3950, 901.4740, 883.4685, 755.4235, PPY-33 Isospirostanol III
753.4130, 735.3992, 607.3529, 739.4265, 721.4136, 593.3680,
541.3130, 505.2762, 459.2741 431.3152, 413.3060, 395.2943,
309.1184, 271.2051, 253.1939
83 14.929 C49H80O22 1020.5141 1065.5117 0.42 1019.5545, 887.5091, 873.4896, 1041.5269 PPY-34 Furostanol I
741.4485, 723.4399
87 15.168 C50H82O22 1034.5298 1079.5352 −3.10 1033.5313, 901.4863, 887.4711, 1018.5231, 885.4810, 855.4688, PPY-35 Furostanol I
755.4258, 739.4323, 593.3712, 711.3570, 577.3728, 415.3200,
575.3603, 527.3378, 431.3170, 397.3101, 271.2057
413.3039, 397.1301
90 15.591 C45H72O18 900.4719 945.4765 7.92 899.4735, 809.4216, 753.4141, 901.4757, 883.4626, 739.4225, PPY-36 Isospirostanol III
699.3919, 631.3496, 607.3517, 721.4132, 593.3662, 579.3134,
541.3228, 506.9627, 445.3096 431.3149, 395.2943, 271.2050,
253.1947
91 15.659 C44H74O18 890.4875 935.4911 −2.03 889.4864, 757.4419, 727.4288, 908.5201 PPY-37 Cholesterol I
595.3826, 577.3721, 433.3314,
305.0867
(Continues)
WANG ET AL.
TABLE 1 (Continued)

m/z [M Error Component

No. Rt (min) Formula Mass (Da) + HCOO]− (ppm) (−) MSE m/z (+) MSE m/z name Structure type Cluster
WANG ET AL.

104 16.624 C56H90O26 1178.6084 1196.6053 4.46 1177.5879, 1045.5346, 1031.5293, 1179.5950, 1047.5327, 1017.5318, PPY-38 Isospirostanol III
883.4782, 737.4139, 557.3483 885.4830, 855.4713, 739.4245,
723.4289
111 16.938 C50H82O22 1034.0600 1035.0551 −4.70 1033.5313, 901.4863, 887.4711, 1017.5195, 855.4677, 723.4270, PPY-39 Furostanol III
755.4258, 739.4323, 593.3712 577.3690, 415.3176, 397.3087
112 16.942 C50H80O21 1016.5500 1034.5509 0.90 1015.5221, 883.4774, 865.4671, 1017.5175, 885.4783, 855.4662, PPY-40 Furostanol III
737.4164, 557.3501 723.4261, 577.3691, 559.3604,
415.3175, 397.3071
116 17.096 C57H90O25 1174.5771 1175.5804 0.37 1173.5827, 1027.5228, 881.4636, 1029.5255, 1013.5286, 867.4721, PPY-41 Isospirostanol III
735.4038, 717.3876, 393.1366, 721.4140, 575.3562, 557.3454,
351.1301 413.3049
119 17.531 C44H70O19 902.4511 947.4571 9.31 901.4572, 835.9271, 801.3868, 920.4629, 885.4408, 753.4015, PPY-42 Isospirostanol I
769.4099, 739.3978, 696.3916, 712.3953, 607.3445, 589.3326,
607.3539, 571.3271, 541.3226, 573.3407, 445.2934, 427.2830,
525.3247, 459.2839, 445.2984, 409.2731, 295.1011, 271.2046
401.2696
124 17.809 C51H80O22 1044.5141 1045.5208 0.75 1043.5187, 751.3830, 617.2333, 1027.5031, 899.4593, 881.4537, PPY-43 Isospirostanol VII
539.1993, 453.1620, 393.1401, 753.4038, 735.3935, 607.3465,
351.1293, 309.1187 589.3364, 445.2950, 427.2845,
409.2743
125 18.087 C39H62O15 770.4089 815.4092 3.92 769.4102, 711.3609, 637.3338, 788.4318, 753.4039, 595.3109, PPY-44 Isospirostanol IV
607.3538, 491.2622, 323.1030 429.3000, 411.2130, 393.2782,
271.2056, 253.1953
126 18.112 C57H96O27 1212.6139 1211.5707 1.27 1175.6004, 1029.5362, 883.4831, 1184.6008 PPY-45 Cholesterol II
737.4178, 671.3806, 435.1539,
289.0894
129 18.227 C50H78O24 1062.5247 1061.5245 −0.75 1015.5221, 883.4774, 869.4620, 901.4765, 755.4208, 593.3679, PPY-46 Isospirostanol I, IV, VII
737.4164, 557.3501 431.3156, 413.3043, 271.2052,
253.1954
132 18.463 C51H80O21 1028.5192 1073.5234 −1.84 1027.5229, 617.2318, 539.1988, 1029.5228, 883.4620, 867.4708, PPY-47 Cholesterol III
423.1515, 411.1508, 393.1405, 721.4023, 575.3569, 557.3474,
351.1291, 309.1184 413.3046, 343.1735
135 18.655 C57H90O25 1174.5771 1219.5826 2.11 1173.5827, 1027.5228, 1009.5122, 1175.5820, 1013.5316, 867.4748, PPY-48 Furostanol II
881.4636, 863.4601, 735.4038, 721.4154, 575.3579, 413.3039
573.3434, 351.1301
142 19.656 C45H66O17 878.4300 923.4265 −0.67 877.3932, 745.3496, 731.3325, 896.4607 PPY-49 Isospirostanol I
599.2894, 533.2554, 437.2379,
393.2133
(Continues)
13 of 17
TABLE 1 (Continued)

m/z [M Error Component

14 of 17

No. Rt (min) Formula Mass (Da) + HCOO]− (ppm) (−) MSE m/z (+) MSE m/z name Structure type Cluster
159 23.196 C50H82O21 1018.5349 1063.5320 0.00 1017.5378, 885.4837, 867.4844, 1041.5269 PPY-50 Cholesterol I
739.4210, 577.3796, 559.3660
165 24.837 C45H68O17 880.4457 925.4421 −0.66 879.4099, 747.3656, 733.3554, 903.4374 PPY-51 Isospirostanol I
601.3052, 535.2688, 439.2469,
395.2207
166 24.842 C44H74O16 858.4875 903.5015 −1.75 857.4997, 733.3541, 725.4536, 855.4596, 779.3912, 723.4136, PPY-52 Cholesterol I
711.4273, 579.3949, 247.0814 577.3629, 413.3047, 395.2927
170 25.416 C44H74O17 874.4926 919.4787 1.44 919.4667, 873.4737, 741.4291, 897.4781, 693.3785, 653.3442, PPY-53 Cholesterol I
595.3709, 527.3242, 309.1115 599.2691, 507.2902, 453.2097,
413.1772
172 25.762 C48H80O19 960.4137 1005.4408 −0.18 959.4361, 941.4156, 827.3928, 1017.5244, 855.4659, 737.4080, PPY-54 Furostanol I
813.3734, 681.3315, 663.3184, 591.391, 577.3734, 429.3009,
615.2970, 519.2739, 475.2465 411.2890, 271.2060, 253.1954
175 25.957 C37H58O12 694.3928 739.3809 −0.12 693.3743, 415.1452, 397.1350, 717.4172 PPY-55 Cholesterol V
295.2271
180 26.710 C44H68O17 868.4457 913.4501 −1.40 867.4462, 735.4046, 589.3427, 869.4537, 737.4086, 721.3772, PPY-56 Isospirostanol I
457.1572, 439.1462, 397.1350, 591.3511, 447.2378, 429.2993,
379.1031, 233.0658 411.2886, 393.2767, 285.1853
181 27.024 C51H88O24 1084.5666 1129.5544 0.83 1083.5898, 1029.5358, 883.4809, 1085.5701 PPY-57 Cholesterol II
737.4138, 591.3598, 393.1377,
342.9952, 289.0921
188 27.948 C44H68O16 852.4507 897.4555 0.25 851.4524, 727.3541, 719.4076, 853.4514, 835.4449, 721.4108, PPY-58 Isospirostanol I
573.3453, 247.0828 575.3578, 395.2943, 251.1788
191 28.294 C51H82O20 1032.5141 1033.5727 2.66 1031.5105 1015.5412, 869.4846, 833.4633, PPY-59 Isospirostanol III
723.4265, 577.3708, 415.3187,
397.3086
194 29.075 C44H68O18 884.5080 902.5075 −0.50 883.4431, 751.3970, 737.4144, 885.4807, 723.4289, 705.4210, PPY-60 Furostanol III
605.3371, 457.1566, 439.1467, 577.3716, 543.3670, 415.3187,
379.1248, 307.1026 397.3087, 379.2990, 271.2056,
253.1951
197 29.729 C48H80O18 944.4188 989.4467 −0.29 943.4451, 811.4000, 797.3844, 989.4470 PPY-61 Furostanol I
665.3397, 599.3064, 503.2837,
459.2564, 349.1155
216 34.937 C37H60O11 680.4136 725.4012 1.03 679.3942, 415.1460, 397.1349, 681.4255 PPY-62 Cholesterol V
323.0988, 305.0866, 281.2482 ,
235.0817
WANG ET AL.
WANG ET AL.
as the C22–C26 cleavage product27 of PPY-21 or PPY-44, which was
subsequently identified as PPY-24 (Table 1; and Figure S6, supporting
information). With the aid of GNPS, four steroids were discovered
from PPY for the first time. Furthermore, cluster V was also attributed
to the presence of a series of saponins, and the negative-ion mode
MS indicated that these compounds mostly contain disaccharides, as
shown in Figure S5C (supporting information). As a result, two steroid
saponins were discovered from PPY for the first time (Table 1; and
Figure S6, supporting information). In addition, the MS/MS ions of
another class of PPY (cluster VI) isospirostanol saponins were
clustered into a distinct group of nodes, and the aglycone C-3 of
isospirostanol saponins mostly contains a di-, tri-, or tetrasaccharide
(Figure S5B, supporting information). Eight steroids, including
1 steroid (peak 129) that was identified for the first time, were
identified. Similarly, cluster VII was also constituted by nodes of
isospirostanol saponins, and these compounds mostly contained
tetrasaccharides, as shown in the positive-ion spectra. Two steroids
(peaks 124 and 129) were identified for the first time (Table 1; and
Figure S5E, supporting information).
3.3.2 | Identification of phenylpropanoids in
molecular networking
In the negative-ion mode, the 15 nodes from m/z 470 to 739 were
grouped as cluster VIII and were attributed to the presence of a
series of phenylpropanoids, as shown in Figure S5F (supporting
information). In particular, the nodes of m/z 709.205 [M + HCOO]−
and 663.198 [M−H]− generated MS2 ions at m/z 517.1578
[M − H–Rha]−, m/z 499.1454 [M − H–Rha–H2O]−, and 487.1440
−
[M − H–Rha–CH2O] , which indicated the presence of sugar units
[Rha]. By comparing the structures of reported parispolyside G
(peak 45)22 and the fragmentation patterns (Table S3, supporting
information), the m/z 709.205 and 663.198 nodes were both
confidently identified as analogues of parispolyside G (PPY-12 or
PPY-16). Meanwhile, the m/z 739.214 node correlated with m/z
709.205, indicating that they were structurally related. Based on the
relevant literature and structural features of PPY,41 the m/z
788.439 node was determined as a 3-methoxy derivative of PPY-12
or PPY-16 (PPY-18 or PPY-19). Based on the analysis of GNPS,
eight new phenylpropanoids were identified from PPY (Figure S5F,
supporting information).
3.3.3 | Identification of insect metamorphosis
hormones in molecular networking
In the negative-ion mode, cluster IX contains a total of 17 nodes,
ranging from m/z 479 to 637, which were attributed to the presence
of a series of insect metamorphosis hormones, as shown in
Figure S5G (supporting information). In particular, by comparison with
the structures of reported compounds30,42 and the database
(Tables S1 and S3, supporting information), the m/z 525.309 and
15 of 17
541.304 nodes were identified as β-ecdysone42 (peak 19) and
turkesterone (peak 12 or 20),30 respectively. Meanwhile, as shown in
Figure S5G (supporting information), the unknown ion at m/z 557.3
correlated with the known ion at m/z 541.304, indicating that they
were structurally related. Based on their fragmentation (Table S3,
supporting information), structural features of turkesterone,30 and the
precursors 16 m/z units higher than m/z 541.304, the m/z 557.3 node
was determined to be a 24-hydroxyl derivative of turkesterone (PPY-
18). Moreover, based on the precursor ions 2 m/z units lower than m/z
525.309, the m/z 523.293 node was elucidated as the C22–C23
(or C14–C15) double bond formation product of β-ecdysone,42,43
which was subsequently identified as PPY-20 (Figures S5G and S6,
supporting information; and Table 1). In addition, other series of
unknown ions with m/z 687.366 and 567.284 nodes were clustered
and tentatively identified. With the aid of GNPS, four moulting
hormones were discovered from PPY for the first time (Figure S5G,
supporting information).
4 | CONC LU SIONS
Molecular networking of the extracts of PPY was generated based on
the MS/MS similarities among the compounds. By implementing
MS/MS molecular networking, we successfully analyzed nine clusters
associated with isospirostanol, furostanol saponins, moulting
hormones, and phenylpropanoids. Moreover, an effective data
analysis method using a molecular networking strategy coupled with
UHPLC/QTOF-MS/MS was developed to globally profile compounds
and rapidly discover new compounds. In our study, a total of
222 compounds from PPY were characterized, among which 62 were
discovered for the first time. These discoveries remarkably expanded
our understanding of the chemical material basis of PPY, and the
present proposed strategy could be applied to other herbal extracts or
biological samples for the rapid and efficient matching of minor novel
compounds.
ACKNOWLEDG MENTS
The authors acknowledge funding from NSFC (Nos. 81741160 and
81603269), Science and Technology Planning Project of Guangdong
Province (No. 2017A020217008), the Fire Plan Project of the
Guangzhou University of Chinese Medicine (No. XH20170110),
Medical Research Foundation of Guangdong Province
(No. A2016334), and Project on Construction of High Level
University of Guangzhou University of Chinese Medicine (No. 81).
OR CID
Qian Fan https://orcid.org/0000-0003-3630-5538
RE FE RE NCE S
1. Song Y, Wang S, Ding Y, et al. Chloroplast genomic resource of Paris
for species discrimination. Sci Rep. 2017;7(1):3427-3434.
2. Wei J, Gao W, Yan X, Wang Y, Jing S, Xiao P. Chemical constituents
of plants from the genus Paris. Chem Biodivers. 2014;11(9):1277-1297.
16 of 17
3. Zhang T, Liu H, Liu X, Xu D, Chen X, Wang Q. Qualitative and
quantitative analysis of steroidal saponins in crude extracts from Paris
polyphylla var. yunnanensis and P. polyphylla var. chinensis by high
performance liquid chromatography coupled with mass spectrometry.
J Pharm Biomed Anal. 2010;51(1):114-124.
4. Wu X, Wang L, Wang GC, et al. Triterpenoid saponins from rhizomes
of Paris polyphylla var. yunnanensis. Carbohydr Res. 2013;368(7):1-7.
5. Feng L, Jun LL, Zeper A, et al. Structural characterization of steroidal
saponins by electrospray ionization and fast-atom bombardment
tandem mass spectrometry. Rapid Commun Mass Spectrom. 2002;16:
1168-1173.
6. Liu Y, Tian X, Hua D, et al. New steroidal saponins from the rhizomes
of Paris delavayi and their cytotoxicity. Fitoterapia. 2016;111:
130-137.
7. Bolleddula J, Fitch W, Vareed S, Nair MG. Identification of
metabolites in Withania sominfera fruits by liquid chromatography
and high-resolution mass spectrometry. Rapid Commun Mass
Spectrom. 2012;26(11):1277-1290.
8. Ma B, Yang S, Li J, et al. A four-step filtering strategy based on ultra-
high- performance liquid chromatography coupled to quadrupole-
time-of-flight tandem mass spectrometry for comprehensive profiling
the major chemical constituents of Akebiae Fructus. Rapid Commun
Mass Spectrom. 2019;33(18):1464-1474.
9. Liu G, Xu Y, Sang J, Zhu B, Wang J. Characterization of a new
component and impurities in josamycin by trap-free two-dimensional
liquid chromatography coupled to ion trap time-of-flight mass
spectrometry. Rapid Commun Mass Spectrom. 2019;33(12):1058-1066.
10. Yang J, Sanchez L, Rath CM, et al. Molecular networking as a
dereplication strategy. J Nat Prod. 2013;76(9):1686-1699.
11. Quinn RA, Nothias LF, Vining O, Meehan M, Esquenazi E,
Dorrestein PC. Molecular networking as a drug discovery, drug
metabolism, and precision medicine strategy. Trends Pharmacol Sci.
2017;38(2):143-154.
12. Bertin MJ, Schwartz SL, Lee J, et al. Spongosine production by a
Vibrio harveyi strain associated with the sponge Tectitethya crypta.
J Nat Prod. 2015;78(3):493-499.
13. Liaw CC, Chen PC, Shih CJ, et al. Vitroprocines, new antibiotics
against Acinetobacter baumannii, discovered from marine Vibrio
sp. QWI-06 using mass-spectrometry-based metabolomics approach.
Sci Rep. 2015;5(1):12856-12866.
14. Kleigrewe K, Almaliti J, Tian IY, et al. Combining mass spectrometric
metabolic profiling with genomic analysis: A powerful approach for
discovering natural products from cyanobacteria. J Nat Prod. 2015;
78(7):1671-1182.
15. Nasir W, Toledo AG, Noborn F, et al. Sweetnet: A bioinformatics
workflow for glycopeptide MS/MS spectral analysis. J Proteome Res.
2016;15(8):2826-2840.
16. Fox Ramos AE, Alcover C, Evanno L, et al. Revisiting previously
investigated plants: A molecular networking-based study of
geissospermum laeve. J Nat Prod. 2017;80(4):1007-1014.
17. Nguyen DD, Melnik AV, Koyama N, et al. Erratum: Indexing the
pseudomonas specialized metabolome enabled the discovery of
poaeamide B and the bananamides. Nat Microbiol. 2017;2(1):16197-
17010.
18. Ge YW, Zhu S, Yoshimatsu K, Komatsu K. MS/MS similarity
networking accelerated target profiling of triterpene saponins in
Eleutherococcus senticosus leaves. Food Chem. 2017;227:444-452.
19. Lyu Q, Kuo TH, Sun C, Chen K, Hsu CC, Li X. Comprehensive
structural characterization of phenolics in litchi pulp using tandem
mass spectral molecular networking. Food Chem. 2019;282:9-17.
20. Li R, Zhou Y, Wu Z, Ding L. ESI-Q-TOF-MS/MS and APCI-IT-MS/MS
analysis of steroid saponins from the rhizomes of Dioscorea panthaica.
J Mass Spectrom. 2006;41(1):1-22.
21. Man S, Gao W, Zhang Y, et al. Characterization of steroidal saponins
in saponin extract from Paris polyphylla by liquid chromatography
WANG ET AL.
tandem multi-stage mass spectrometry. Anal Bioanal Chem. 2009;
395(2):495-505.
22. Kang LP, Yu K, Zhao Y, et al. Characterization of steroidal glycosides
from the extract of Paris Polyphylla var. Yunnanensis by UPLC/Q-TOF
MSE. J Pharm Biomed Anal. 2012;62:235-249.
23. Kang LP, Huang YY, Zhan ZL, et al. Structural characterization and
discrimination of Paris polyphylla var. yunnanensis and Paris
vietnamensis based on metabolite profiling analysis. J Pharm Biomed
Anal. 2017;142:252-261.
24. Wang Y, Liu E, Li P. Chemotaxonomic studies of nine Paris species
from China based on ultra-high performance liquid chromatography
tandem mass spectrometry and Fourier transform infrared
spectroscopy. J Pharm Biomed Anal. 2017;140:20-30.
25. Zhu LL, Zhao Y, Xu YW, et al. Comparison of ultra-high performance
supercritical fluid chromatography and ultra-high performance liquid
chromatography for the separation of spirostanol saponins. J Pharm
Biomed Anal. 2016;120:72-78.
26. Zhang T, Liu H, Liu XT, Chen XQ, Wang Q. Steroidal saponins from
the rhizomes of Paris delavayi. Steroids. 2009;74(10):809-813.
27. Zhao Y, Kang LP, Liu YX, et al. Steroidal saponins from the rhizome of
Paris polyphylla and their cytotoxic activities. Planta Med. 2009;75(4):
356-363.
28. Cheng CX, Zhang YT, Zhou J. The glycosides of aerial Parts polyphylla
Var. yunnanensis. Acta Botanica Yunnanica. 1995;17(01):473-478.
29. Shao B, Guo H, Cui Y, Ye M, Han J, Guo D. Steroidal saponins from
Smilax China and their anti-inflammatory activities. Phytochemistry.
2007;68(5):623-630.
30. Kristina JS, Nadin K, Karsten S, Sven J, Gerd W, Matthias FM.
Chemical composition and biological activity of Paris quadrifolia L. Z
Naturforsch Chem. 2012;67(11):565-570.
31. Wu X, Wang L, Wang GC, et al. New steroidal saponins and sterol
glycosides from Paris polyphylla var. yunnanensis. Planta Med. 2012;
78(15):1667-1675.
32. Wang Y, Gao WY, Zhang TJ, Guo YQ. A novel phenylpropanoid
glycosides and a new derivation of phenolic glycoside from Paris
Polyphylla var. yunnanensis. Chin Chem Lett. 2007;18(5):548-550.
33. Sudarshana RB, Pavankumar P, Sridhar L, Saha S, Narahari SG,
Prabhakar S. Differential cationization of fatty acids with monovalent
cations studied by ESI-MS/MS and computational approach. Rapid
Commun Mass Spectrom. 32(14):1126-1134. https://doi.org/10.
1002/rcm.8143
34. Wu X, Zhang YB, Wang GC, Feng YF, Li YL. Study on the
identification of two hydroxy fatty acid from Paris Polyphylla var.
yunnanensis and their antiproliferative activities against nasopharyngeal
carcinoma cells. J Guangdong Pharma Univ. 2014;30:698-701.
35. Qin XJ, Yu MY, Ni W, et al. Steroidal saponins from stems and leaves
of Paris polyphylla var. yunnanensis. Phytochemistry. 2016;121:20-29.
36. Liu XX, Wang L, Long Y, Sun L, Wang Q. Chemical constituent from
Paris mairei. Chin J Chin Mater Med. 2014;39:3107-3110.
37. Wang Y, Gao WY, Zhang TJ, Guo YQ. A novel phenylpropanoid
glycosides and a new derivation of phenolic glycoside from Paris
Polyphylla var. yunnanensis. Chin Chem Lett. 2007;18(5):548-550.
38. Yan L, Gao W, Zhang Y, Wang Y. A new phenylpropanoid glycosides
from Paris polyphylla var. yunnanensis. Fitoterapia. 2008;79(4):
306-307.
39. Chen P, Jin HY, Sun L, Ma SC. Multi-component determination and
chemometric analysis of Paris polyphylla by ultra high performance
liquid chromatography with photodiode array detection. J Sep Sci.
2016;39(18):3550-3557.
40. Sun CL, Ni W, Yan H, et al. Steroidal saponins with induced platelet
aggregation activity from the aerial parts of Paris verticillata. Steroids.
2014;92:90-95.
41. Yan LL, Gao WY, Zhang YJ, Wang Y. A new phenylpropanoid
glycosides from Paris polyphylla var. yunnanensis. Fitoterapia. 2008;
79(4):306-307.
WANG ET AL.
42. Wu X, Wang L, Wang H, Dai Y, Ye WC, Li YL. Steroidal saponins from
Paris polyphylla var. yunnanensis. Phytochemistry. 2012;81:133-143.
43. Liu XX, Wang L, Long Y, Sun L, Wang Q. Chemical constituent from
Paris mairei. Chin J Chin Mater Med. 2014;39(16):3107-3110.
SUPPORTING INFORMATION
Additional supporting information may be found online in the
Supporting Information section at the end of this article.
17 of 17
How to cite this article: Wang Y, Fan Q, Xiang J, et al.
Structural characterization and discrimination of Paris
polyphylla var. yunnanensis by a molecular networking strategy
coupled with ultra-high-performance liquid chromatography
with quadrupole time-of-flight mass spectrometry. Rapid
Commun Mass Spectrom. 2020;34:e8760. https://doi.org/10.
1002/rcm.8760