Chemical constituents from shells of Xanthoceras sorbifolium

Article abstract of DOI:10.1016/j.phytochem.2020.112288

Three undescribed triterpenes and four previously unreported saponins, along with two known ones, were isolated from shells of Xanthoceras sorbifolium (Sapindaceae). Their structures were elucidated by the interpretation of 1D and 2D NMR data. The nitric oxide (NO) assay revealed that 28-O-isobutyryl-21-O-angeloyl-R₁-barrigenol and 3-O-β-D-6-O-methylglucuronopyranosyl-21,22-di-O-angeloyl-R₁-barrigenol possessed stronger inhibitory effects on LPS-induced NO overproduction (IC₅₀ = 18.5 ± 1.2 and 28.2 ± 1.8 μM, respectively) than the positive drug minocycline (IC₅₀ = 30.1 ± 1.3 μM) in activated BV2 cells. Western blot, RT-qPCR, and docking experiments further validated that the regulation of iNOS and IL-1β expressions was involved in the anti-neuroinflammatory effects of these two compounds.

Full text of DOI:10.1016/j.phytochem.2020.112288

Phytochemistry172(2020)112288
Contents lists available at ScienceDirect
Phytochemistry
journal homepage: www.elsevier.com/locate/phytochem
Chemical constituents from shells of Xanthoceras sorbifolium
Gang Chena, Yumeng Xiea, Di Zhoua, Yanqiu Yangb, Jingyu Liub, Yue Houb, Maosheng Chengc,
Yang Liuc,∗∗, Ning Lia,∗
a School of Traditional Chinese Materia Medica, Shenyang Pharmaceutical University, Shenyang, 110016, Liaoning, China
b School of Life Science and Biopharmaceutics, Shenyang Pharmaceutical University, 103 Wenhua Road, Shenyang, China
c Key Laboratory of Structure-Based Drug Design & Discovery of Ministry of Education, School of Pharmaceutical Engineering, Shenyang Pharmaceutical University,
Shenyang, 110016, China
A R T I C L E I N F O
A B S T R A C T
Keywords:
Three undescribed triterpenes and four previously unreported saponins, along with two known ones, were
isolated from shells of Xanthoceras sorbifolium (Sapindaceae). Their structures were elucidated by the inter-
pretation of 1D and 2D NMR data. The nitric oxide (NO) assay revealed that 28-O-isobutyryl-21-O-angeloyl-R1-
barrigenol and 3-O-β-D-6-O-methylglucuronopyranosyl-21,22-di-O-angeloyl-R1-barrigenol possessed stronger
Xanthoceras sorbifolium
Sapindaceae
Triterpene saponins
Neuroinflammation
Docking
inhibitory effects on LPS-induced NO overproduction (IC50 = 18.5
1.2 and 28.2
1.8 μM, respectively)
than the positive drug minocycline (IC50 = 30.1
1.3 μM) in activated BV2 cells. Western blot, RT-qPCR, and
Microglia
docking experiments further validated that the regulation of iNOS and IL-1β expressions was involved in the
anti-neuroinflammatory effects of these two compounds.
1. Introduction
2019). Due to the potent in vivo effect, the pharmaceutical research has
been conducted to increase the water solubility of xanthoceraside by
Xanthoceras sorbifolium Bunge (Sapindaceae), the nuts of which has
been used to treat infantile enuresis for generations in China, is a
monotypic genus of Xanthoceras, and was included in the pharmacopoeia
(1977 edition) of China (Pharmacopoeia Committee of the People's
Republic of China, 1977). Phenols from X. sorbifolium have been reported
to display anti-tumor, radical-scavenging, and neuroinflammation-in-
hibiting effects (Li et al., 2016a; Yang et al., 2016). Previous chemical
studies have showed that oleanene-type and barrigenol-like triterpenes
are the main constituents of X. sorbifolium. (Wang et al., 2017; Xiao et al.,
2013; Yu et al., 2012; Zhang et al., 2019). In recent years, more attention
has been paid to X. sorbifolium due to the neuroprotective effects that it
exhibits in Alzheimer Diseases (AD)-related models. The xanthoceraside
from X. sorbifolium has been reported to elicit protective effects on
Aβ25–35-induced learning and memory impairment in mice (Chi et al.,
2009). Extracts of X. sorbifolium also has been reported to show similar
effects by attenuating dendritic spine deficiency and cognitive decline
through the up-regulation of BDNF (Li et al., 2016b). A latest research
has demonstrated that compared to the positive drug (huperzine A),
more targets, including those related to neuroinflammation, energy
metabolism and antioxidant ability, are involved in the cognitive dis-
order protective effects of the husk extract of X. sorbifolium (Rong et al.,
constructing xanthoceraside hollow gold nanoparticles aiming at pro-
moting the bioavailability (Meng et al., 2016).
To our best knowledge, so far there is no well-acknowledged target
for AD as almost all the candidate drugs failed the phase III clinical trial.
However, it has been well acknowledged that neuroinflammation is a key
factor in the pathogenesis of neurodegeneration in AD (Buckwalter and
Wyss-Coray, 2004). In addition to the aforementioned result that anti-
neuroinflammation is one of the in vivo mechanisms of the husk extract of
X. sorbifolium (Rong et al., 2019), exploring more anti-neuroin-
flammatory constituents from X. sorbifolium will contribute to the dis-
covery of natural anti-AD agents. Therefore, in this paper, the isolation
and identification of seven previously undescribed (1–7) and two known
triterpene compounds (8–9) from X. sorbifolium and their in vitro anti-
neuroinflammatory effects in activated microglia were reported.
2. Results and discussion
2.1. Phytochemical investigation
Seven previously undescribed (1-7) and two known triterpene
compounds 21-O-(3′,4′-di-O-angeloyl)-β-D-fucopyranosyl barrigenol C
∗ Corresponding author.
∗∗ Corresponding author.
E-mail addresses: y.liu@syphu.edu.cn (Y. Liu), liningsypharm@163.com (N. Li).
https://doi.org/10.1016/j.phytochem.2020.112288
Received 14 September 2019; Received in revised form 25 January 2020; Accepted 27 January 2020
0031-9422/©2020ElsevierLtd.Allrightsreserved.

G. Chen, et al.
Phytochemistry172(2020)112288
(8) (Chen et al., 1984) and 28-O-acetyl-3-O-[α-L-arabinofuranosyl-
suggested that except for NMR signals of the epoxyangeloyl, all the
NNR resonances of 1 ascribable to the R1-barrigenol and angeloyl
moieties were identical to those of 2 (Tables 1 and 2), indicating a
different substituent linked to C-21 in 2 compared with that of 1. The
remaining carbon signals in 2 [δC 174.5 (C-1′), 77.4 (C-2′), 70.5 (C-3′),
21.4 (C-4′), 16.8 (C-5′)] resembled those of the 2, 3-dihydroxy-2-me-
thylbutyryl (Torres-Valencia et al., 1998), which was further confirmed
by the HMBC correlations from H-5′ to C-1'/C-3′, and from H-4′ to C-2'
(Fig. 3). Finally, based on the correlation between H-21 and C-1′ in the
HMBC spectrum, the structure of 2 was determined as 21-O-(2, 3-di-
hydroxy-2-methylbutyryl)-22-O-angeloyl-R1-barrigenol.
(1
→ 3)]-[β-D-galactopyranosyl-(1 → 2)]-β-D-6-O-methylglu-cur-
onopyranosyl-21-O-(3′,4′-di-O-angeloyl)-β-D-fucopyranosyl barrigenol
C (9) (Wang et al., 2016) were isolated from the 70% ethanol extract of
shells of X. sorbifolium (see Fig. 1).
Compound 1 was obtained as a white amorphous powder, showing
the pseudo-molecular ion peak m/z 685.4316 [M-H]- in the HRESI-MS
spectra. Therefore, the molecular formula of 1 was determined as
C
40H62O9 (calcd. 685.4340 for C40H61O9). The 1H NMR spectra gave
seven angular methyl signals at δH 0.67 (Me-24), 0.80 (Me-29), 0.88
(Me-25), 0.89 (Me-26), 0.89 (Me-23), 1.00 (Me-30), and 1.30 (Me-27);
one olefinic proton signal at δH 5.37 (1H, br s, H-12); signals of four
oxygen-substituted methines at δH 3.63 (1H, m, H-15), 3.71 (1H, m, H-
16), 5.54 (1H, d, J = 10.1 Hz, H-22), and 5.79 (1H, d, J = 10.1 Hz, H-
21). Accordingly, the 13C NMR spectra also revealed seven methyl
signals at δC 15.5 (Me-25), 16.1 (Me-24), 17.0 (Me-26), 19.4 (Me-30),
20.3 (Me-27), 28.2 (Me-23), and 29.1 (Me-29), along with four oxygen-
substituted methine signals at δC 66.3 (C-15), 72.9 (C-16), 71.5 (C-22),
and 79.8 (C-21) as determined by the HSQC spectra. All these NMR data
were characteristic resonances of barrigenol-like triterpenes (Wang
et al., 2017). Through comparing NMR signals assignable to the tri-
terpene skeleton with those in the literature (Wang et al., 2017), the
triterpene skeleton of 1 was elucidated as R1-barrigenol (5α-olean-12-
ene-3β,15α,16α,21β,22α,28-hexol). The relative configuration of the
R1-barrigenol skeleton was determined based on both NOESY correla-
tions (H-24/H-25/H-26/H-15/H-28/H-22/H-30, H-3/H-23/H-5/H-7/
H-27, and H-29/H-21) (Fig. 2) and the multiplicity of H-21/22
(J = 10.1 Hz) that suggested the axial orientation for H-21 and H-22. In
addition, one angeloyl and one oxyangeloyl were also deduced due to
chemical shifts of [δC 15.5 (C-5″), 20.4 (C-4″), 127.6 (C-3″), 138.4 (C-
2″), 166.1 (C-1″) for angeloyl] and [δC 13.1 (C-5′), 19.2 (C-4′), 58.8 (C-
3′), 59.3 (C-2′), 168.7 (C-1′) for oxyangeloyl (Li et al., 2005)]. In the
NOESY spectra, the correlation of δH 3.05 (H-3′)/δH 1.42 (H-5′) sug-
gested the cis-configuration of H-3'/H-5'. Subsequently, HMBC corre-
lations from δH 5.73 (H-21) to δC 168.7 (C-1′), and from δH 5.54 (H-22)
to δC 166.1 (C-1″) determined the location of the angeloyl and the
epoxyangeloyl (Fig. 3). Therefore, the structure of 1 was established as
21-O-epoxyangeloyl-22-O-angeloyl-R1-barrigenol.
Compound 3 possessed the molecular formula of C39H62O8 de-
termined by the pseudo-molecular ion peak m/z 681.4322 [M+Na]+
(calcd. 681.4342 for C39H62NaO8) provided by the HRESI-MS. The
NMR data of 3 were quite similar to those of 1 and 2 regarding the R1-
barrigenol moiety (Tables 1 and 2), suggesting that 3 had the same
triterpene skeleton as 1 and 2. Apparent difference in the 13C NMR data
of 3, compared with those of 1 and 2, were observed at C-22 (δC 69.1
for 3 vs δC 71.5 and δC 71.4 for 1 and 2, respectively) and C-28 (δC 63.9
for 3 vs δC 62.1 for both 1 and 2), indicating different substitution
patterns of C-22 and C-28 for 3. Besides those assignable to the R1-
barrigenol moiety, one group of angeloyl [δC 15.4 (C-4′), 20.6 (C-5′),
128.6 (C-2′), 135.3 (C-3′), 167.3 (C-1′)] and isobutyryl [δC 18.7 (C-4″),
18.9 (C-3″), 33.5 (C-2″), 175.6 (C-1″)] signals were also observed. In
the HMBC spectra, correlations from H-21 to C-1′ of the angeloyl, and
from H-28 to C-1″ of the isobutyryl allowed the determination of the
location for the angeloyl and the isobutyryl, respectively (Fig. 3).
Therefore, 3 was elucidated as 28-O-isobutyryl-21-O-angeloyl-R1-bar-
rigenol.
Compound 4 was obtained as a white amorphous powder, and its
molecular formula was elucidated to be C42H66O13 as deduced by the
[M-H]- m/z: 777.4445 (calcd. 777.4425 for C42H65O13) observed in the
HRESI-MS spectra. The proton signals of δH 4.29 (d, J = 7.8 Hz) with
the corresponding carbon signal at δC 105.5 determined by the HSQC
spectra suggested the presence of an anomeric proton of a sugar moiety
in 4. The NMR data assigned to the aglycone moiety of 4 were closely
similar to those of 3, expect for the absence of signals of the isobutyryl
located at C-28 of 3 (Tables 1 and 2). And the high-field shifted C-28
carbon signal in 4 (δC 63.9 for 3 and δC 62.5 for 4) further confirmed
that the C-28 hydroxyl was unbound. In the HMBC spectra, the corre-
lation from H-21 to C-1′ of the angeloyl group was detected (Fig. 3),
leading to the determination of the aglycone of 4 as 21-O-angeloyl-R1-
barrigenol. Furthermore, by comparing the NMR data of the sugar
moiety [δC 105.5 (C-1″), 73.7 (C-2″), 75.9 (C-3″), 71.7 (C-4″), 75.3 (C-
Compound 2 was also obtained as a white amorphous powder with
the molecular formula of C40H64NaO10 as indicated by the pseudo-
molecular ion peak m/z 727.4417 [M+Na]+ (calcd. 727.4397 for
C
40H64NaO10) in the HRESI-MS spectra. The 1H NMR data was asso-
ciated with the 13C NMR data by the HSQC spectra as summarized in
Tables 1 and 2. Through comparing NMR data of 1 and 2, it was
Fig. 1. Structures of previously unreported compounds 1–7 from X. sorbifolium. Ang: angeloyl. Ang: angeloyl.
2

G. Chen, et al.
Phytochemistry172(2020)112288
Fig. 2. Key NOE correlations of 1.
5″), 169.7 (C-6″), 51.9 (-OCH3)] to those in the literature (Nakamura
et al., 2010) and the HMBC correlation from δH 3.66 (-OCH3) to δC
169.7 (C-6″), the β-D-6-O-methylglucuronopyranosyl was also eluci-
dated. The absolute configuration of the sugar moiety was also de-
termined by the hydrolysis experiment (see experimental section) of the
alkaline hydrolysis-afforded product (Brenner, 2013) of 4, which
showed the presence of the D-glucuronic acid. Combined with the
correlation between δH 4.29 (H-1″) and δC 88.1 (C-3) in the HMBC
spectra, 4 was established as 3-O-β-D-6-O-methylglucuronopyranosyl-
21-O-angeloyl-R1-barrigenol.
barrigenol (5) and 3-O-β-D-6-O-methylglucuronopyranosyl-21,22-di-O-
angeloyl-R1-barrigenol (6), respectvely.
Compound 7 afforded the quasi-molecular ion peak of [M+Na]+ m/
z: 1013.5473 in the HR-ESI spectra, suggesting the molecular formula of
C53H82NaO17 (calcd 1013.5450 for C53H82NaO17). Two sets of anomeric
NMR signals were shown at δC 105.5 (δH 4.29, d, J = 7.7 Hz) and δC
104.4 (δH 4.40, d, J = 7.7 Hz), which means there were two mono-
saccharides in 7. Furthermore, the NMR data of 7 indicated the presence
of 3-O-β-D-6-O-methylglucuronopyranosyl-R1-barrigenol moiety via
comparing to NMR data of 5 and 6 (Tables 1 and 2). The remaining NMR
data of 7 were identical to those of 3,4-di-O-angeloyl-β-D-fucopyranosyl
in the literature (Chen et al., 1984). HMBC correlations from H-3'' (δH
5.03) to C-1‴ (δC 166.3) and from H-4'' (δH 5.23) to C-1‴' (δC 166.6)
further confirmed the existence of the 3,4-di-O-angeloyl-β-D-fucopyr-
anosyl moiety. Additionally, compared with 5 and 6, a significant low-
field shift occurred at C-21 of 7 (δC 90.7 for 7; δC 77.9 and 77.6 for 6 and
5, respectively), indicating that the 3,4-di-O-angeloyl-β-D-fucopyranosyl
was linked to C-21 of 7. HMBC correlation from H-21 to C-1″ of the 3,4-
di-O-angeloyl-β-D-fucopyranosyl confirmed the deduced linkage (Fig. 3).
Therefore, 7 was elucidated as 3-O-β-D-6-O-methylglucuronopyranosyl-
21-O-(3,4-di-O-angeloyl-β-D-fucopyranosyl) barrigenol C.
Compounds 5 and 6 were both 6-O-methylglucuronopyranosides
with 21-O-angeloyl-R1-barrigenol as the triterpene skeleton as deduced
through comparing the NMR data of 5 and 6 with those of 4. In addition
to the NMR signals identical to those of 4, chemical shifts for C-22 of 5
and 6 were both δC 71.9, which was apparently high-field shifted
compared with that of 4 (δC 69.7), suggesting that hydroxyls of C-22 in
5 and 6 were substituted. Additionally, NMR data of 5 afforded a group
of isobutyryl signals [δC 175.4 (C-1″), 33.6 (C-2″), 18.7 (C-3″), and 18.9
(C-4″)], while the 13C NMR spectra of 6 revealed a group of angeloyl
sigals [δC 166.6 (C-1″), 128.2 (C-2″), 135.7 (C-3″), 15.2 (C-4″), and 20.3
(C-5″)], which were consistent with the HR-ESI MS-indicated molecular
formulas of C46H72NaO14 for 5 and C47H72NaO14 for 6 based on the
quasi-molecular ion peaks ([M+Na]+ m/z: 871.4838 for 5; [M+Na]+
m/z: 883.4843 for 6). In conjunction with HMBC correlations from δH
5.44 (H-22) to δC 175.4 (C-1″) for 5, and from δH 5.54 (H-22) to δC
166.6 (C-1″) for 6, the structures of 5 and 6 were determined as 3-O-β-
D-6-O-methylglucuronopyranosyl-21-O-angeloyl-22-O-isobutyryl-R1-
2.2. Inhibitory effect on LPS-activated microglia
In the central nervous system (CNS), microglia is the main resident
immunocompetent and phagocytic cell, and activation of it would in-
duce the production of neurotoxic factors, such as nitric oxide (NO) and
Fig. 3. Key HMBC correlations of compounds 1–7. Ang: angeloyl.
3

G. Chen, et al.
Phytochemistry172(2020)112288
Table 1
1H NMR data of 1–7 (DMSO‑d6).
No.
1
2
3
4
5
6
7
1
2
3
0.91/1.53 m
1.45 m
0.90/1.53 m
1.45 m
0.88 m/1.51 m
1.44 m
0.88 m/1.52 m
1.52 m/1.62 m
3.06 m
1.52 m/1.88 m
1.53 m/1.64 m
3.06 m
0.90/1.54 m
1.54 m
0.89 m/1.51 m
1.63 m
3.01 m
3.00 m
2.99 m
3.07 (dd, 4.3,
11.5)
3.07 (dd, 4.2, 11.3)
5
0.67 overlap
1.29 m
0.67 overlap
1.44 m
0.66 m
0.70 (d, 11.9)
1.43 m
0.70 (d, 11.8)
1.28 m/1.43 m
1.59 m/1.68 m
1.47 m
0.70 (d, 11.8)
1.22 m/1.44 m
1.60 m/1.70 m
1.49 m
0.72 (d, 11.8)
1.31 m/1.48 m
1.26 m
6
1.30/1.43 m
1.60/1.71 m
1.44/1.79 m
1.79 m
7
1.60/1.70 m
1.48 (d, 8.9)
1.83 (dd, 3.3, 8.9)
5.37 br s
3.63 m
1.59/1.70 m
1.47 (d, 8.6)
1.82 (dd, 3.4, 8.6)
5.37 br s
3.62 m
1.59 m/1.69 m
1.46 (d, 9.0)
1.80 (dd, 5.3, 9.0)
5.35 br s
3.60 m
9
1.56 m
11
12
15
16
18
19
21
22
23
24
25
26
27
28
29
30
1.81 m
1.82 m
1.80 m
5.29 br s
5.35 br s
3.62 m
5.36 br s
3.63 m
5.20 br s
1.18 (d, 9.1)/1.58 m
4.02 (d, 9.1)
2.40 m
3.65 m
3.71 m
3.71 m
3.77 m
3.74 m
3.72 m
3.73 m
2.43 m
2.43 m
2.40 (dd, 3.7, 14.4)
1.07/2.53 m
5.56 (d, 10.0)
3.77 (d, 10.0)
0.90 s
2.43 m
2.43 m
2.43 m
1.13 (dd, 3.1, 11.7)
5.79 (d, 10.1)
5.54 (d, 10.1)
0.89 s
1.09 overlap
5.73 (d, 10.1)
5.52 (d, 10.1)
0.90 s
1.05/2.43 m
5.52 (d, 9.9)
3.86 (d, 9.9)
0.97 s
1.12 m/2.52 m
5.77 (d, 10.0)
5.44 (d, 10.0)
0.97 s
1.13 m/2.53 m
5.81 (d, 10.1)
5.54 (d, 10.1)
0.97 s
0.96 m
4.03 (d, 9.4)
3.75 (d, 9.4)
1.89 s
0.67 s
0.67 s
0.67 s
0.75 s
0.75 s
0.75 s
0.83 s
0.88 s
0.88 s
0.87 s
0.89 s
0.88 s
0.89 s
0.88 s
0.89 s
0.90 s
0.90 s
0.90 s
0.89 s
0.90 s
0.75 s
1.30 s
1.30 s
1.28 s
1.27 s
1.30 s
1.31 s
1.35 s
2.87/3.09 m
0.80 s
2.87/3.09 m
0.78 s
3.80 (d, 10.6)/3.65 (d, 10.6)
0.77 s
2.87/3.09 m
0.75 s
2.87 m/3.09 m
0.77 s
2.90/3.12 (dd, 3.7, 10.0)
2.87/3.09 m
0.96 s
0.78 s
0.99 s
1.00 s
098 s
0.95 s
0.91 s
0.97 s
0.91 s
21-
21
21-Ang
21-Ang
21-Ang
21-Ang
3-O-Glc A
1′
–
–
–
–
–
–
4.29 (d, 7.7)
3.01 m
2′
–
–
–
–
–
–
3′
3.05 (q, 5.4)
1.05 (d, 5.4)
1.42 s
3.65 (q, 6.4)
0.96 (d, 6.4)
1.08 s
6.01 (q, 7.0)
1.91 (d, 7.0)
1.86 s
6.00 (br q, 7.2)
1.91 (br d, 7.2)
1.86 br s
6.07 (br q, 7.1)
1.88 (br d, 7.1)
1.76 br s
5.94 (br q, 7.1)
1.85 (d,7.1)
1.76 s
3.16 (d, 9.7)
3.30 m
4′
5′
3.72 (d, 9.7)
3.66 s
OMe
–
–
–
–
–
–
22-Ang
22-Ang
28-
3-O-Glc A
22-
22-Ang
21-O-Fuco
1″
–
–
–
4.29 (d, 7.8)
2.99 m
3.16 m
3.28 m
3.72 m
–
–
–
4.40 (d, 7.7)
3.57 m
2″
–
–
2.54 m
2.36 m
–
3″
6.10 (br q, 7.3)
6.02 (br q, 7.1)
1.09 (d, 7.0)
1.01 (d,7.0)
6.00 (dq, 7.1)
5.03 (dd, 3.5, 10.1)
5.23 (d, 3.5)
4.08 m
4″
1.92 (d, 7.1)
1.90 (d, 7.1)
1.10 (d, 7.0)
0.96 (d,7.0)
1.81 (dd, 7.1)
5″
1.78 br s
1.79 br s
–
–
–
–
–
–
1.76 s
6″
–
–
–
–
–
–
1.04 (d, 6.3)
–
OMe
3.66 s
3-O-Glc A
3-O-Glc A
3″-O-Ang
1‴
4.29 (d, 7.7)
3.00 m
4.29 (d,7.8)
3.01 m
3.17 m
3.30 m
3.73 m
3.66 s
–
2‴
–
3‴
3.16 m
6.10 (q, 7.2)
1.91(d, 7.2)
1.75 br s
–
4‴
3.28 m
5‴
3.72 m
OMe
3.66 s
4″-O-Ang
3‴'
4‴'
5‴'
6.20 (q, 7.3)
1.88 (d, 7.3)
1.86 br s
IL-1β, that mediates inflammatory processes in CNS. Moreover, the
activated microglia has also been indicated in pathological lesions in
neurological diseases. Thus, examining the NO production in LPS-
treated microglia has been used as a in vitro model to screen the anti-
neuroinflammtory agents (Li et al., 2015; Wu et al., 2007).
6, the inducible nitric oxide synthase (iNOS) and IL-1β expression were
tested via Western blot and RT-qPCR methods. The results showed that
both 3 and 6 potently suppressed the LPS-induced ectopic up-regulation
of iNOS (Fig. 5A) and IL-1β (Fig. 5B), suggesting the anti-neuroin-
flammatory effects of 3 and 6.
In this paper, all these nine isolates (1–9) were subjected to the in
vitro evaluation on the NO production in LPS-activated BV2 cells.
Though 1–9 are structurally similar, only compounds 3 and 6 exhibited
Moreover, inhibitory effects on iNOS of 1–3 were also investigated
through docking study for these three aglycones elicited different ef-
fects on NO production. The software of Glide was used to examine the
binding ability of 1–3 to the protein iNOS (PDB:4UX6). The docking
results showed that 3 perfectly occupied the active pocket of iNOS
(Fig. 6A–C), while 1 and 2 failed to dock into the same site on iNOS,
being consistent with the in vitro result of the NO assay.
potent inhibitory effects (IC50 = 18.5
1.2 and 28.2
1.8 μM for 3
and 6, respectively) (Fig. 4) compared with the positive drug minocy-
cline (IC50 = 30.1 1.3 μM), while the remaining compounds showed
no effects (IC50 > 100 μM). To further understand the effect of 3 and
4

G. Chen, et al.
Phytochemistry172(2020)112288
Table 2
13C NMR data of 1–7 (DMSO‑d6).
No.
1
2
3
4
5
6
7
1
38.5
27.1
76.8
38.3
54.6
18.2
35.6
40.6
46.3
36.2
23.1
124.6
142.5
46.6
66.3
72.9
46.9
39.8
46.0
35.6
79.8
71.5
28.2
16.1
15.5
17.0
20.3
62.1
29.1
19.4
38.4
27.1
76.8
38.3
54.6
18.2
35.7
40.6
46.3
36.6
23.1
124.5
142.6
46.6
66.4
72.9
46.8
39.8
46.2
35.8
78.6
71.4
28.2
16.1
15.5
17.0
20.3
62.1
29.1
19.4
38.5
27.1
76.8
38.3
54.6
18.2
35.7
40.5
46.2
36.6
23.2
124.5
142.5
46.8
66.4
71.3
46.4
40.2
45.9
35.2
79.7
69.1
28.3
16.1
15.5
17.0
20.2
63.9
29.3
19.6
38.4
25.7
88.1
38.7
54.7
18.0
35.6
40.5
46.3
36.3
23.1
123.9
143.3
46.5
66.3
71.8
46.6
39.7
46.7
35.3
80.2
69.7
27.5
16.5
15.4
17.0
20.2
62.5
29.4
19.7
38.4
25.7
88.1
38.7
54.7
18.0
35.6
40.6
46.2
36.3
23.1
124.4
142.6
46.6
66.2
72.2
46.6
39.4
46.0
35.4
77.6
71.9
27.5
16.5
15.5
17.0
20.3
62.3
29.0
19.6
38.4
25.7
88.1
38.7
54.7
18.0
35.6
40.6
46.2
36.3
23.1
124.5
142.6
46.7
66.3
72.8
46.6
39.4
46.1
35.4
77.9
71.9
27.5
16.5
15.5
17.0
20.3
62.1
29.0
19.5
38.2
25.7
88.1
38.6
54.9
17.8
32.3
39.5
46.1
36.2
23.0
122.1
142.9
40.8
33.5
67.0
46.4
38.6
47.3
36.2
90.7
69.8
27.5
16.4
15.6
16.5
26.8
63.0
29.2
19.7
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
21-
21-
21-Ang
21-Ang
21-Ang
21-Ang
3-O-Glc A
1′
168.7
59.3
58.8
13.1
19.2
–
174.5
77.4
70.5
21.4
16.8
–
167.3
128.6
135.4
15.4
20.6
–
167.4
128.6
135.3
15.5
20.6
–
166.4
127.6
137.5
15.4
166.5
127.8
136.7
15.3
105.5
73.6
75.9
71.6
75.3
169.7
51.8
2′
3′
4′
5′
20.4
20.3
6′
OMe
–
–
–
–
22-Ang
22-Ang
28-
3-O-Glc A
22-
22-Ang
21-O-Fuco
1″
166.1
127.6
138.4
15.5
20.4
–
166.5
127.9
137.4
15.5
20.5
–
175.6
33.5
18.9
18.7
–
105.5
73.7
75.9
71.7
75.3
169.7
51.9
175.4
33.6
18.7
18.9
–
166.6
128.2
135.7
15.2
20.3
–
104.4
68.8
72.9
70.3
68.1
15.9
–
2″
3″
4″
5″
6″
–
–
OMe
–
–
–
–
–
3-O-Glc A
3-O-Glc A
3″-O-Ang
1‴
105.5
73.7
75.9
71.7
75.3
169.7
51.9
105.5
73.7
75.9
71.7
75.3
169.7
51.8
166.3
127.2
137.9
15.4
19.9
–
2‴
3‴
4‴
5‴
6‴
OMe
–
4″-O-Ang
1‴'
2‴'
3‴'
4‴'
5‴'
166.6
126.9
138.3
15.4
20.3
3. Conclusion
X. sorbifolium. The NO assay revealed that 3 and 6 possessed stronger
inhibitory effects on LPS-induced NO overproduction than the positive
drug minocycline in activated BV2 cells. The anti-neuroinflammtory
effects of 3 and 6 were also confirmed by the Western blot, RT-qPCR,
Three undescribed triterpenes (1–3) and four previously unreported
saponins (4–7), along with two known ones (8–9), were isolated from
5

G. Chen, et al.
Phytochemistry172(2020)112288
Fig. 4. Inhibitory effects of compounds 3 and 6 on NO production in LPS-activated BV2 cells. Values are means
S.E.M. from three separate experiments and are
normalized to the control group. (#P < 0.05 compared with the control group; *P < 0.05 compared with the LPS group).
and docking experiments, and the results of which showed that both
iNOS and IL-1β expressions were repressed by 3 and 6 in LPS-activated
BV2 cells, which might suggested great potentials for 3 and 6 in the
therapy of neurodegenerative diseases such as AD.
September (autumn), 2018. The Shells of X. sorbifolium was prepared by
removing the kernel and the flesh of the fruits. The material was
identified by Professor Yingni Pan (School of Traditional Chinese
Materia Medica, Shenyang Pharmaceutical University). A voucher
specimen (No.20180921) is deposited in School of Traditional Chinese
Materia Medica, Shenyang Pharmaceutical University.
4. Experimental
4.1. Chemistry
4.4. Acid hydrolysis
NMR spectra were recorded on Bruker ARX-400 and 600 M AVIII
spectrometers, using TMS as an internal standard. HRESI-MS was acquired
on a Bruker micro TOF-Q mass spectrometer in m/z mode. Silica gel
(200–300 mesh) for chromatography was produced by Qingdao Ocean
Chemical Group Company (Qingdao, China). ODS (50 μm) was afforded
by YMC Company (Ltd) (Japan). HPLC separations were performed on a
YMC-pack Prep-ODS column (250 × 20 mm) equipped with a Shimadzu
RID-20AUV detector and a Shimadzu LC-6AD series pumping system
(Tokyo, Japan). Pyridine-d5 were obtained from Sigma-Aldrich Company
(St. Louis, USA). All the chromatographic and analytical grade reagents
were obtained from Tianjin DaMao Chemical Company (Tianjin, China).
For compounds 4–7 (each 2.0 mg), an alkaline hydrolysis experi-
ment (Brenner, 2013) was performed to remove the methyl group in
sugars (dissolved in 1.5 ml 5% KOH:Dioxane 1.5 ml, stirred at 37 °C for
1 h). The powder afforded by the alkaline hydrolysis was dissolved in
4 M HCl (2 mL) and then heated in a H2O bath at 90 °C for 3 h to
dryness. Mixture was extracted by CH2Cl2 and get CH2Cl2 extract. The
water layer was identified using GC method by treating the residue with
L-cysteine methyl ester hydrochloride (0.01 ml) in pyridine (0.02 ml) at
60 °C for 1 h. After this reaction, the solution was treated with N,O-bis
(trimethyl silyl) trifluoroacetamide (0.01 ml) at 60 °C for 1 h. Column:
Supelco STBTM-1 (30 m × 0.25 mmi.d.). Injetor temp.:230. Detector
temp: 230 °C. Column temp.:230. He flow rate: 15 ml/min (methyl-D-
glucuronic acid: 20.16 min). Additonally, residues were analyzed by
HPLC with an NH2P-50 4 E column (4.6 × 250 mm, Showa Denko K.K.,
Japan) and an optical rotation detec-tor (JASCO ORD-4090, JASCO
International Co, Ltd., Japan) to detect the configuration of the fucose
4.2. Plant material
The shell of Xanthoceras sorbifolium Bunge (Sapindaceae) was
collected in Fuxin city (N42.081°, E121.403°), Liaoning province in
Fig. 5. Effects of compounds 3 and 6 on iNOS and IL-1β expressions in LPS-activated BV2 cells. (A) Representative Western blot result of 3 and 6 attenuating iNOS
expression. (B) RT-qPCR results of 3 and 6 suppressing IL-1β expressions. Values are means
S.E.M. from three separate experiments and are normalized to the
control group. (#P < 0.05 compared with the control group; *P < 0.05 compared with the LPS group).
6

G. Chen, et al.
Phytochemistry172(2020)112288
Fig. 6. Docking results by Glide showing 3 perfectly docked into the active pocket of iNOS protein (PDB:4UX6). (A) The docking result of 3 bound to the active pocket
of iNOS. The iNOS protein was represented as gray solid cartoons of solid surfaces while 3 was represented as cyan solid cartoons of solid surfaces. (B) The result of
(A) with iNOS represented as color ribbons. (C) The result of (A) with good contacts between 3 and iNOS represented as white short lines (3 was shown in cyan color).
using standard samples with the following condition: acetonitrile-water
(75:25, v/v); flow rate, 0.8 mL/min. D-fucose, 9.5 min (positive).
HPLC with MeOH/H2O (75: 25) as eluent to afford compound 9
(4.1 mg).
21-O-epoxyangeloyl-22-O-angeloyl-R1-barrigenol (1): [α] + 31.0
(c = 0.33, MeOH). UV (MeOH) λmax 210 and 225 nm. IR(KBr) νmax
3420, 2920, 1722, 1687, 1088 cm−1. HR-ESI-MS m/z: 685.4316 [M-H]-
(calcd. 685.4340 for C40H61O9). 1H and 13C NMR data: see Tables 1
and 2
4.3. Extraction and isolation
The shell of X. sorbifolium (200 kg) were extracted with 70% ethanol
(2000 L each) under reflux for 3 × 2 h. The filtrate was concentrated
under reduced pressure, and afforded the crude extract (20 kg). Ethanol
extract (20 kg) was fractioned on macroporous absorptive resin (D101)
column, eluting with H2O, 20% ethanol/H2O (Fr. A, 1255 g), 50%
ethanol/H2O (Fr. B, 2055 g), 70% ethanol/H2O (Fr. C, 1750 g), 90%
ethanol/H2O (Fr. D, 500 g), successively. Fr. C (1750 g) was chroma-
tographed on silica gel, eluting with CH2Cl2–MeOH (0:100–100:0), to
give six fractions (Fr. D1-D7). Fr. D3 (20.5 g) was subjected to a ODS
column washing with MeOH/H2O (0: 100-100: 0) to get five fractions.
The sub-fraction D3-2 (510 mg) was isolated by HPLC with MeOH/H2O
(80: 20) as eluent to afford compounds 1 (5.8 mg), 2 (13.5 mg), 3
(3.1 mg); sub-fraction D3-3 was isolated by HPLC with MeOH/H2O (75:
25) as eluent to afford compounds 4 (6.5 mg), 5 (8.7 mg), 6 (11.2 mg),
7 (12.4 mg), 8 (16.1 mg); sub-fraction D3-4 (205 mg) was isolated by
21-O-(2,
3-dihydroxy-2-methylbutyryl)-22-O-angeloyl-R1-barri-
genol (2): [α] + 42.0 (c = 0.43, MeOH). UV (MeOH) λmax 210 and
225 nm. IR(KBr) νmax 3425, 2921, 1725, 1680, 1058 cm−1. HR-ESI-MS
m/z: 727.4417 [M+Na]+ (calcd. 727.4397 for C40H64NaO10). 1H and
13C NMR data: see Tables 1 and 2
28-O-isobutyryl-21-O-angeloyl-R1-barrigenol (3): [α]
+
22.0
(c = 0.30, MeOH). UV (MeOH) λmax 210 and 225 nm. IR(KBr) νmax
3418, 2920, 1735, 1685, 1018 cm−1. HR-ESI-MS m/z: 681.4322 [M
+Na]+ (calcd. 681.4342 for C39H62NaO8). 1H and 13C NMR data: see
Tables 1 and 2
3-O-β-D-6-O-methylglucuronopyranosyl-21-O-angeloyl-R1-barri-
genol (4): [α] + 52.0 (c = 0.55, MeOH). UV (MeOH) λmax 210 and
225 nm. IR(KBr) νmax 3423, 2950, 1725, 1675, 1008 cm−1. HR-ESI-MS
7

G. Chen, et al.
Phytochemistry172(2020)112288
m/z: 777.4445 [M-H]- (calcd. 777.4425 for C42H65O13). 1H and 13C
NMR data: see Tables 1 and 2
was retrieved from the Protein Databank Bank (http://www.rcsb.org/).
Before docking, the preparations of proteins and ligands (1–3) were
performed following the standard protocol of the Protein Preparation
and Ligprep Wizards, respectively, of the Schrodinger 2013 Suite.
Docking studies were implemented via Glide method in XP (Extra-
Precision) mode.
3-O-β-D-6-O-methylglucuronopyranosyl-21-O-angeloyl-22-O-iso-
butyryl-R1-barrigenol (5): [α] + 35.0 (c = 0.23, MeOH). UV (MeOH)
λmax 210 and 225 nm. IR(KBr) νmax 3425, 2910, 1725, 1680,
1015 cm−1. HR-ESI-MS m/z: 871.4838 [M+Na]+ (calcd. 871.4820 for
C
46H72NaO14). 1H and 13C NMR data: see Tables 1 and 2
3-O-β-D-6-O-methylglucuronopyranosyl-21,22-di-O-angeloyl-R1-
4.9. Statistics
barrigenol (6): [α] + 41.0 (c = 0.43, MeOH). UV (MeOH) λmax 210 and
225 nm. IR(KBr) νmax 3421, 2915, 1721, 1685, 1019 cm−1. HR-ESI-MS
m/z: 883.4843 [M+Na]+ (calcd. 883.4820 for C47H72NaO14). 1H and
13C NMR data: see Tables 1 and 2
IC50, inhibitory rate, and RT-qPCR values are means
S.E.M. from
five samples. The significance of the difference between means was
determined by ANOVA. The level of significance was determined by
using Duncan's multiple-range test.
3-O-β-D-6-O-methylglucuronopyranosyl-21-O-(3,4-di-O-angeloyl-β-
D-fucopyranosyl) barrigenol C (7): [α] + 45.0 (c = 0.21, MeOH). UV
(MeOH) λmax 210 and 225 nm. IR(KBr) νmax 3425, 2910, 1725, 1681,
1010 cm−1. HR-ESI-MS m/z: 1013.5473 [M+Na]+ (calcd. 1013.5450
for C53H82NaO17). 1H and 13C NMR data: see Tables 1 and 2
Declaration of competing interest
The authors declare no competing financial interests.
Acknowledgements
4.5. Cell culture and in vitro evaluation of NO production
BV2 cells were cultured in Dulbecco's modified Eagle's medium
(DMEM) supplemented with 10% heat-inactivated FBS in a humidified
environment with 5% CO2 at 37 °C. The measurement of NO production
was conducted as previously reported (Li et al., 2015). Briefly, accu-
mulation of nitrite (NO2−), an indicator of NO synthase activity, in
culture medium was measured using the Griess reaction. Cells
(5 × 104 cells/well) were plated on 96-well microliter plates and
treated with each compound (0.1, 1, 10, 50 μM) in presence of bacterial
lipopolysacchride (LPS; 100 ng/ml) for 24 h. Fifty microliter culture
medium supernatants were mixed with 50 μl Griess reagent (part I: 1%
sulfanilamide; part II: 0.1% naphthylethylene diamide dihydrochloride
and 2% phosphoric acid) at room temperature. Fifteen minutes later,
the absorbance was measured at 540 nm.
This work was supported partially by the National Natural Science
Foundation of China (81872768, 81673323, U1903122, U1603125),
LiaoNing Revitalization Talents Program (XLYC1807118), LiaoNing
BaiQianWan Talents Program (2018), Program for Liaoning Innovative
Talents in University (LR2017043) and the Fundamental Research
Funds for the Central Universities of China (N182008004,
N182006001). Overseas Training Project of Liaoning Colleges and
Universities (2018LNGXGJWPY-YB024).
References
Brenner, D.A., 2013. Reversibility of liver fibrosis. Gastroenterol. Hepatol. 9, 737–739.
Buckwalter, M.S., Wyss-Coray, T., 2004. Modelling neuroinflammatory phenotypes in
vivo. J. Neuroinflammation 1, 10.
Chen, G., Han, Y., Feng, Y., Wang, A., Li, X., Deng, S., Zhang, L., Xiao, J., Li, Y., Li, N.,
2019. Extract of Ilex rotunda Thunb alleviates experimental colitis-associated cancer
via suppressing inflammation-induced miR-31-5p/YAP overexpression.
Phytomedicine 62, 152941.
4.6. Western blot analysis
The Western blot experiment was performed with the previously
reported method (Chen et al., 2019). Briefly, Whole-cell lysates were
prepared using 2% SDS, sonicated and centrifuged (15,000 g) at 4 °C for
15 min. The supernatants were boiled for 5 min and size-fractionated by
SDS/PAGE (10% acrylamide). After transferring proteins on to ni-
trocellulose filters, the blots were incubated with primary antibodies
recognizing iNOS and GPADH (1:1000, Cell Signaling Technology,
Beverly, MA) for 2 h at room temperature; following incubation with
secondary antibodies for 1 h at room temperature, immunocomplexes
were developed by using chemiluminescence.
Chen, G., Zhou, D., Li, X.Z., Jiang, Z., Tan, C., Wei, X.Y., Ling, J., Jing, J., Liu, F., Li, N.,
2017. A natural chalcone induces apoptosis in lung cancer cells: 3D-QSAR, docking
and an in vivo/vitro assay. Sci. Rep. 7, 10729.
Chen, Y.J., Takeda, T., Ogihara, Y., Iitaka, Y., 1984. Studies on the constituents of
Xanthoceras sorbifolia Bunge. II. Major sapogenol and a prosapogenin from the fruits
of Xanthoceras sorbifolia Bunge. Chem. Pharm. Bull. (Tokyo) 32, 3378–3383.
Chi, T.Y., Wang, L.H., Qu, C., Yang, B.Z., Ji, X.F., Wang, Y., Okuyama, T., Yoshihito, O.,
Zou, L.B., 2009. Protective effects of xanthoceraside on learning and memory im-
pairment induced by Abeta(25-35) in mice. J. Asian Nat. Prod. Res. 11, 1019–1027.
Deng, S., Wang, A., Chen, X., Du, Q., Wu, Y., Chen, G., Guo, W., Li, Y., 2019. HBD inhibits
the development of colitis-associated cancer in mice via the IL-6R/STAT3 signaling
pathway. Int. J. Mol. Sci. 20.
Li, J., Jiang, Z., Li, X., Hou, Y., Liu, F., Li, N., Liu, X., Yang, L., Chen, G., 2015. Natural
therapeutic agents for neurodegenerative diseases from a traditional herbal medicine
Pongamia pinnata (L.) Pierre. Bioorg. Med. Chem. Lett 25, 53–58.
4.7. Reverse transcription (RT) and quantitative real-time PCR analyses
Li, N., Wang, Y., Li, X., Zhang, H., Zhou, D., Wang, W., Li, W., Zhang, X., Li, X., Hou, Y.,
Meng, D., 2016a. Bioactive phenols as potential neuroinflammation inhibitors from
the leaves of Xanthoceras sorbifolia Bunge. Bioorg. Med. Chem. Lett 26, 5018–5023.
Li, Y., Xu, J., Xu, P., Song, S., Liu, P., Chi, T., Ji, X., Jin, G., Qiu, S., Hou, Y., Zheng, C.,
Wang, L., Meng, D., Zou, L., 2016b. Xanthoceras sorbifolia extracts ameliorate den-
dritic spine deficiency and cognitive decline via upregulation of BDNF expression in a
rat model of Alzheimer's disease. Neurosci. Lett. 629, 208–214.
The RT-qPCR experiment was performed as previously described
(Deng et al., 2019). Total RNA from BV2 cells was extracted Trizol
according to the manufacturer's instructions. Reverse transcription was
carried out to obtain cDNA using the Master Mix kit (Takara, Shiga,
Japan) following standard protocols. The mRNA levels of IL-1β were
Li, Z.L., Li, X., Li, L.H., Li, N., Yu, M., Meng, D.L., 2005. Two new triterpenes from the
husks of Xanthoceras sorbifolia. Planta Med. 71, 1068–1070.
assessed sing
a Step One Plus real-time PCR system (BIO-RAD,
Meng, D.L., Shang, L., Feng, X.H., Huang, X.F., Che, X., 2016. Xanthoceraside hollow gold
nanoparticles, green pharmaceutics preparation for poorly water-soluble natural anti-
AD medicine. Int. J. Pharm. 506, 184–190.
CFX96TM Real-time System, C1000TM Thermal Cycle, USA). For IL-1β
mRNA quantification: the following sense and antisense primers were
used: forward primer 5′-ATGGCAACTGTTCCTGAACTCAACT-3′, re-
verse primer 5′- AGGACAGGTATAG ATTCTTTCCTT-3'.
Nakamura, S., Chen, G., Nakashima, S., Matsuda, H., Pei, Y., Yoshikawa, M., 2010.
Brazilian natural medicines. IV. New noroleanane-type triterpene and ecdysterone-
type sterol glycosides and melanogenesis inhibitors from the roots of pfaffia glo-
merata. Chem. Pharm. Bull. (Tokyo) 58, 90–695.
Rong, W., Ding, K., Guo, S., Xie, F., Li, Q., Bi, K., 2019. Metabolomics analysis of
Xanthoceras sorbifolia husks protection of rats against Alzheimer's disease using li-
quid chromatography mass spectrometry. J. Chromatogr. B Analyt. Technol. Biomed.
Life Sci. 1126–1127, 121739.
4.8. Docking studies
The docking studies were conducted with Grid-based Ligand
Docking with Energetics (Glide) provided in the Schrodinger 2013
package suit as previously reported (Chen et al., 2017; Zhou et al.,
2019). The three-dimensional crystal structure of iNOS (PDB ID: 4UX6)
́
Torres-Valencia, J.M., Cerda-Garcıa-Rojas, C.M., Joseph-Nathan, P., 1998.
Enantioselective preparation of (2R,3R)-(+)- and (2S,3S)-(−)-2,3-epoxy-2-methyl-
butanoic acids and some derivatives. Tetrahedron: Asymmetry 9, 757–764.
8

G. Chen, et al.
Phytochemistry172(2020)112288
Wang, D., Dan, S., Li, X., Dan, L., Xi, R., Gao, H., Wang, X.J.R.A., 2016. Barrigenol tri-
terpenes from the husks of Xanthoceras sorbifolia Bunge and their antitumor activ-
ities. 6, 27434–27446.
Yang, C.Y., Ha, W., Lin, Y., Jiang, K., Yang, J.L., Shi, Y.P., 2016. Polyphenols isolated
from Xanthoceras sorbifolia husks and their anti-tumor and radical-scavenging ac-
tivities. Molecules 21.
Wang, D., Su, D., Yu, B., Chen, C., Cheng, L., Li, X., Xi, R., Gao, H., Wang, X., 2017. Novel
anti-tumour barringenol-like triterpenoids from the husks of Xanthoceras sorbifolia
Bunge and their three dimensional quantitative structure activity relationships ana-
lysis. Fitoterapia 116, 51–60.
Yu, L., Tang, X., Chen, L., Wang, M., Jian, J., Cao, S., Wang, X., Kang, N., Qiu, F., 2012.
Oleanane-type triterpenoid saponins from Xanthoceras sorbifolia Bunge. Fitoterapia
83, 1636–1642.
Zhang, X., Zhang, S., Yang, Y., Wang, D., Gao, H., 2019. Natural barrigenol-like tri-
terpenoids: a comprehensive review of their contributions to medicinal chemistry.
Phytochemistry 161, 41–74.
Wu, C.F., Bi, X.L., Yang, J.Y., Zhan, J.Y., Dong, Y.X., Wang, J.H., Wang, J.M., Zhang, R.,
Li, X., 2007. Differential effects of ginsenosides on NO and TNF-α production by LPS-
activated N9 microglia. Int. Immunopharm. 7, 313–320.
Zhou, D., Li, X., Chang, W., Han, Y., Liu, B., Chen, G., Li, N., 2019. Antiproliferative
steroidal glycosides from rhizomes of Polygonatum sibiricum. Phytochemistry 164,
172–183.
Xiao, W., Wang, Y., Zhang, P., Li, N., Jiang, S., Wang, J.H., Huang, J., Li, X., 2013.
Bioactive barrigenol type triterpenoids from the leaves of Xanthoceras sorbifolia
Bunge. Eur. J. Med. Chem. 60, 263–270.
9