The anti-inflammatory secoiridoid glycosides from Gentianae Scabrae Radix: The root and rhizome of Gentiana scabra

Article abstract of DOI:10.1007/s11418-015-0894-8

Gentianae Scabrae Radix is a well-known traditional medicine that is used for the treatment of hepatitis, cholecystitis and inflammatory diseases. It consists mainly of secoiridoid glycosides, with representatives of gentiopicroside, sweroside and swertiamarin. In the present study, a chemical investigation of the CHCl₃ extract of Gentianae Scabrae Radix derived from Gentiana scabra Bunge yielded seven new (1-7) and ten known (8-17) secoiridoid glycosides. Their structures were elucidated by extensive spectroscopic analyses and comparison with literature data. All 17 compounds were evaluated for their inhibitory effects against NO, IL-6 and TNF-α productions induced by lipopolysaccharide (LPS) in RAW264 cells. Among them, 8-epi-kingiside derivatives 1-3; kingiside derivatives 4, 5 and 10; and a sweroside derivative 6 showed inhibition activity against IL-6 production with IC₅₀ values of 51.70-61.10 μM, whereas sweroside derivatives 12 and 15-17 and a swertiamarin derivative 13 showed inhibition effects on both NO and IL-6 productions with IC₅₀ values of 64.74-94.95 and 48.91-75.45 μM, respectively. All the compounds exhibited weak inhibitory activity (IC₅₀ > 100 μM) in a TNF-α bioassay. Finally, a primary structure-activity relationship of these secoiridoid glycosides is discussed.

Full text of DOI:10.1007/s11418-015-0894-8

J Nat Med
DOI 10.1007/s11418-015-0894-8
ORIGINAL PAPER
The anti-inflammatory secoiridoid glycosides from Gentianae
Scabrae Radix: the root and rhizome of Gentiana scabra
•
Yu-Min He Shu Zhu Yue-Wei Ge
•
•
•
•
•
Kohei Kazuma Kun Zou Shao-Qing Cai
Katsuko Komatsu
Received: 16 December 2014 / Accepted: 15 February 2015
ꢀ The Japanese Society of Pharmacognosy and Springer Japan 2015
Abstract Gentianae Scabrae Radix is a well-known tra-
ditional medicine that is used for the treatment of hepatitis,
cholecystitis and inflammatory diseases. It consists mainly
of secoiridoid glycosides, with representatives of gen-
tiopicroside, sweroside and swertiamarin. In the present
study, a chemical investigation of the CHCl3 extract of
Gentianae Scabrae Radix derived from Gentiana scabra
Bunge yielded seven new (1–7) and ten known (8–17)
secoiridoid glycosides. Their structures were elucidated by
extensive spectroscopic analyses and comparison with lit-
erature data. All 17 compounds were evaluated for their
inhibitory effects against NO, IL-6 and TNF-a productions
induced by lipopolysaccharide (LPS) in RAW264 cells.
Among them, 8-epi-kingiside derivatives 1–3; kingiside
derivatives 4, 5 and 10; and a sweroside derivative 6
showed inhibition activity against IL-6 production with
IC50 values of 51.70–61.10 lM, whereas sweroside
derivatives 12 and 15–17 and a swertiamarin derivative 13
showed inhibition effects on both NO and IL-6 productions
with IC50 values of 64.74–94.95 and 48.91–75.45 lM, re-
spectively. All the compounds exhibited weak inhibitory
activity (IC50 [ 100 lM) in a TNF-a bioassay. Finally, a
primary structure–activity relationship of these secoiridoid
glycosides is discussed.
Keywords Secoiridoid glycosides ꢀ Gentiana scabra ꢀ
Gentianaceae ꢀ Anti-inflammatory activity ꢀ 8-epi-kingiside
derivative ꢀ Kingiside derivative
Electronic supplementary material The online version of this
article (doi:10.1007/s11418-015-0894-8) contains supplementary
material which is available to authorized users.
Introduction
The genus Gentiana (Gentianaceae) consists of ap-
proximately 360 species that are distributed mainly in
alpine habitats of temperate regions of Asia, Europe and
the Americas [1]. Many species have been used widely in
traditional medicine. In the Chinese Pharmacopoeia, the
roots and rhizomes of Gentiana scabra Bge., G. man-
shurica Kitag., G. triflora Pall. and G. rigescens Franch.
have been prescribed as the crude drug known as Gentianae
Scabrae Radix, which is used widely for the treatment of
hepatitis, rheumatism, cholecystitis, fungal infection and
allergic inflammations [2]. The former three are also
recorded in the Japanese Pharmacopoeia and are used
mainly as a stomachic or stimulant of appetite [3, 4].
Previous phytochemical studies on Gentianae Scabrae
Radix revealed a series of compounds including iridoids
(mainly secoiridoid glycosides) [5–8], triterpenes [9, 10],
lignans [11] and flavonoids [11]. The four medicinal
Y.-M. He ꢀ S. Zhu ꢀ Y.-W. Ge ꢀ K. Komatsu (&)
Division of Pharmacognosy, Department of Medicinal
Resources, Institute of Natural Medicine, University of Toyama,
2630, Sugitani, Toyama 930-0194, Japan
e-mail: katsukok@inm.u-toyama.ac.jp
Y.-M. He ꢀ K. Zou
College of Medical Science, China Three Gorges University,
8 University Avenue, Yichang 443002,
People’s Republic of China
K. Kazuma
Division of Kampo-Pharmaceutics, Institute of Natural
Medicine, University of Toyama, 2630, Sugitani,
Toyama 930-0194, Japan
S.-Q. Cai
School of Pharmaceutical Sciences, Peking University,
38 Xue-yuan Road, Haidian District, Beijing 100191,
People’s Republic of China
123

J Nat Med
species mentioned above are found to have the common
principal constituents gentiopicroside, sweroside and
swertiamarin. However, the sub-major or minor con-
stituents essentially exhibit the diversity of their chemical
profile, which might be species-specific and are not only
used as markers for identification but are also responsible
for some bio-activities. G. scabra is one of the dominant
species available on the crude drug market, therefore it is
necessary to elucidate its chemical constituents more dee-
ply. As the representative constituents in G. scabra, sec-
oiridoids are evidenced to possess multiple effects, such as
anti-inflammatory [12, 13], hepatoprotective [14, 15] and
neuritogenic activities [16] and are also attractive in their
structural diversity. To date, over 20 secoiridoids had been
identified in G. scabra; however, there are still some minor
secoiridoids that have been detected in our research. With
the aim of enriching the chemical information of G. scabra
and to explore bio-active secoiridoids, seven new secoiri-
doid glycosides, together with ten known ones, have been
isolated from the CHCl3 extract of Gentianae Scabrae
Radix derived from G. scabra.
major mediators playing a role in various inflammatory
diseases [19, 20]. A primary structure–activity relationship
of secoiridoid glycosides is discussed accordingly.
Results and discussion
The CHCl3 extract of Gentianae Scabrae Radix derived
from G. scabra, which was found to possess potential anti-
inflammatory activity (IC50 112.80 lg/mL NO inhibition
and 176.81 lg/mL IL-6 inhibition, respectively) in the
preliminary screening, was subjected to a successive col-
umn chromatography over normal and reverse phase silica
gel, and further purified by preparative HPLC to afford
seven new secoiridoid glycosides (1–7), together with ten
known ones (Fig. 1).
Compounds 1–7 were all obtained as white, amorphous
solids. Hydrolysis of 1–7 with 1.0 N HCl each yielded D-
glucose.
The HR-ESI-MS of 1 displayed a pseudomolecular ion
peak at m/z 525.1596 [M ? H]?, indicating the molecular
1
formula of C24H28O13. The H-NMR spectrum (Table 1)
The inhibitory effects of the compounds isolated from
Gentianae Scabrae Radix on LPS-stimulated NO, IL-6 and
TNF-a productions in macrophages were evaluated be-
cause the induction of nitric oxide (NO) synthesis has been
suggested as one of the major responses to inflammatory
stimuli in macrophages [17, 18], and some pro-inflamma-
tory cytokines, such as IL-6 and TNF-a, are considered as
showed doublets at dH 7.51 (d, J = 0.8) and 5.23 (d,
J = 7.6) that were attributed to H-3 and H-1, being
indicative of an iridoid nucleus. A set of aromatic proton
signals at dH 7.46 (ddd, J = 8.0, 1.6, 1.6 Hz), 7.40 (dd,
J = 2.8, 1.6 Hz), 7.26 (t, J = 8.0 Hz) and 7.01 (ddd,
J = 8.0, 2.8, 1.6 Hz) suggested the presence of a
Fig. 1 Chemical structures of compounds 1–17 isolated from Gentianae Scabrae Radix
123

J Nat Med
123

J Nat Med
Fig. 2 Key H–1H COSY,
HMBC and NOE correlations
for compounds 1 (a, b) and 5
(c, d)
1
(b)
(a)
(c)
(d)
m-disubstituted aromatic ring. Moreover, a methoxy proton
signal at dH 3.70 (s) and an anomeric proton signal at dH
4.71 (d, J = 8.4) were also recognized. The coupling
constant of the anomeric proton signal indicated that the
linkage is of b-configuration. The 13C-NMR spectrum
(Table 1) revealed 24 carbon signals, including signals at
dC 154.4 (C-3) and 96.5 (C-1) assignable to the iridoid
nucleus, as well as signals from three carbonyl groups (dC
174.8, 168.2, 167.8), a b-D-glucopyranosyl group (dC
100.9, 77.7, 75.8, 74.6, 71.6 and 64.2) and a methoxy group
(dC 51.9). The 2D-NMR (1H-1H COSY, HMQC and
HMBC) data, as well as the characteristic signal of C-10
(dC 21.5) in the 13C-NMR spectrum clearly indicated an
8-epi-kingiside structure in 1. As reported in the literature,
chemical shifts of C-10 are indicatively different in
kingiside (ca. dC 18.0) and 8-epi-kingiside (ca. dC 21.5)
types of secoiridoids [21, 22]. Furthermore, NOE correla-
tions of H-10 with H-9 and of H-8 with H-1 (Fig. 2) re-
vealed a b-orientation of methyl group (C-10) at C-8,
which was in agreement with the previously described
configuration of an 8-epi-kingiside skeleton [23]. In addi-
tion to the carbon signals belonging to 8-epi-kingiside
moiety, seven of the remaining carbon signals were as-
signed to m-hydroxybenzoyl group by analyses of the 2D-
NMR data. The linkage of m-hydroxybenzoyl group to C-60
of the sugar unit in 8-epi-kingiside moiety was deduced
from the proton signals of Ha-60 (dH 4.52) and Hb-60 (dH
4.71), which were downshifted by 0.90 and 0.80 ppm
compared with 8-epi-kingiside [22] and confirmed by the
observation of HMBC correlations from H-60 to C-700. The
(1S,5S,8R,9S) absolute configuration of 1 could be assigned
on the basis of the relative configuration deduced from the
typical 13C-NMR chemical shifts and NOE data, as well as
the specific rotation ([a]D -28.72ꢁ) which was close to that
of 8-epi-kingiside tetraacetate with unequivocally absolute
structure [24, 25]. Thus, the structure of 1, including ab-
solute stereochemistry, was established and characterized
as 60-O-(300-hydroxybenzoyl)-8-epi-kingiside.
The HR-ESI-MS data for 2 and 3 indicated molecular
formula of C24H28O13 and C24H28O14, respectively. The
1H- and 13C-NMR data of 2 and 3 (Table 1) revealed
considerable similarity to those of 1. However, the ap-
pearance of p-disubstituted aromatic signals at dH 6.81 (2H,
dd, J = 7.2, 2.0 Hz, H-300, 500), 7.86 (2H, dd, J = 7.2,
2.0 Hz, H-200, 600) in the 1H-NMR data of 2 and o,m-
trisubstituted aromatic signals at dH 7.01 (dd, J = 8.4,
1.6 Hz, H-400), 6.75 (t, J = 8.0 Hz, H-500), 7.35 (dd,
1
J = 8.0, 1.6 Hz, H-600) in the H-NMR data of 3, com-
pletely differed from m-disubstituted aromatic signals de-
tected in 1. The 13C- and 2D-NMR data also supported the
above observation. The location of the p-hydroxybenzoyl
group in 2 and o,m-dihydroxybenzoyl group in 3 were both
determined at the C-60 of the sugar unit in the 8-epi-
kingiside moiety by the proton signal values of Ha-60 and
Hb-60 (dH 4.50, 4.68 for 2 and dH 4.55, 4.77 for 3). The
HMBC correlations of H-60 with C-700 in either 2 or 3
completely supported the deduced connections. Further-
more, the specific rotations of 2 ([a]D -28.20ꢁ) and 3 ([a]D
-21.32ꢁ) were similar to that of 1, suggesting that 2 and 3
had the same stereochemistry as 1. Thus, the structure of 2
123

J Nat Med
was characterized as 60-O-(400-hydroxybenzoyl)-8-epi-
kingiside, and 3 was characterized as 60-O-(200,300-dihy-
droxybenzoyl)-8-epi-kingiside.
Table 2 1H- (400 MHz) and 13C-NMR (100 MHz) spectral data of
compound 6 (d in ppm, J in Hz)
Position
dH
dC
98.6
The HR-ESI-MS data for 4 and 5 indicated a molecular
formula of C24H28O12 and C24H28O14, respectively. The
1H-, 13C- and 2D-NMR data of 4 and 5 (Table 1) revealed
a common kingiside moiety. The carbon signals of C-10 at
dC 17.4 in 4 and dC 17.2 in 5, similar to that of kingiside
(dC 18.4) [21], confirmed the presence of this moiety. The
NOE correlation between H-10 and H-1 was also indicative
of a-orientation of the methyl group (C-10) at C-8 in
kingiside moiety (Fig. 2) [23]. Moreover, the differences
1
5.31, d (2.0)
7.51, d (2.8)
3
153.4
106.7
28.7
4
5
2.82, m
6a
1.61, m
25.8
6b
1.72, m
7a
4.26, td (11.6, 2.4)
4.37, dq (11.6, 2.0)
5.45, ddd (17.2, 10.0, 8.0)
2.65, ddd (8.0, 6.0, 2.0)
5.21, dd (10.0, 2.0)
5.27, dd (17.2, 2.0)
70.0
7b
1
8
132.8
43.4
observed in their H-NMR data, that is, the appearance of
9
five benzoyl signals at dH 7.99 (2H, dd, J = 8.0, 1.2), 7.47
(2H, t, J = 8.0) and 7.61 (t, J = 8.0) in 4, and three o,m-
dihydroxybenzoyl signals at dH 7.03 (dd, J = 8.0, 1.6 Hz),
6.76 (t, J = 8.0 Hz) and 7.26 (dd, J = 8.0, 1.6 Hz) in 5
indicated the presence of a benzoyl group in 4 and o,m-
dihydroxybenzoyl group in 5. These findings were sup-
ported by the corresponding carbon signals and the 2D-
NMR data. Compared with 1, the proton signal of H-20 was
obviously deshielded from dH 3.23 to 4.94 in 4 and to 4.98
in 5, as well as the upfield shift of H-60 in 4 and 5, sug-
gesting that the benzoyl moiety in 4 and o,m-dihy-
droxybenzoyl moiety in 5 were both located at C-20 of
sugar unit of the kingiside moiety. It was further confirmed
by the HMBC correlations of H-20 with C-700 in either 4 or
5. The absolute configurations of 4 and 5 were deduced to
be (1S,5S,8S,9S) based on the above-mentioned relative
configurations and the specific rotation of 4 ([a]D -88.27ꢁ)
and 5 ([a]D -87.76ꢁ) which were close to that of kingiside
tetraacetate with unequivocally absolute structure [24, 25].
Accordingly, the structure of 4 and 5 were established and
characterized as 20-O-benzoyl kingiside and 20-O-(200,300-
dihydroxybenzoyl) kingiside, respectively.
10a
121.3
10b
11
168.0
1-O-glucosyl
10
20
30
40
50
60a
60b
5.11, d (8.4)
97.6
72.2
73.2
70.6
73.4
63.2
4.90, dd (8.4, 10.0)
5.45, dd (10.0, 8.0)
5.20, m
4.00, m
4.11, dd (12.4, 2.8)
4.18, dd (12.4, 3.2)
O-benzoyl
100
200
300
400
500
600
700
131.5
118.4
159.2
123.5
130.8
124.7
166.3
7.62, t (1.6)
7.25, ddd (8.0, 2.4, 1.6)
7.32, t (8.0)
7.55, ddd (8.0, 2.4, 1.6)
300-O-glucosyl
1¢¢¢
2¢¢¢
3¢¢¢
4¢¢¢
5¢¢¢
6¢¢¢a
6¢¢¢b
OCOCH3
4.87, d (7.6)
3.38, m
102.2
74.8
78.3
71.5
78.0
62.7
The HR-ESI-MS data for 6 indicated the molecular
formula of C35H42O19. The 1H-NMR spectrum of 6
(Table 2) showed a set of characteristic signals at dH 5.31
(d, J = 2.0 Hz, H-1), 7.51 (d, J = 2.8 Hz, H-3), 2.82 (m,
H-5), 5.45 (ddd, J = 17.2, 10.0, 8.0 Hz, H-8), 5.21 (dd,
J = 10.0, 2.0 Hz, Ha-10) and 5.27 (dd, J = 17.2, 2.0 Hz,
Hb-10), assignable to a secoiridoid moiety. Two anomeric
proton signals at dH 5.11 (d, J = 8.4 Hz, H-10) and 4.87 (d,
J = 7.6 Hz, H-1¢¢¢), as well as three acetyl methyl proton
signals at dH 1.79, 1.88 and 1.93 (each 3H and s) were also
recognized. The above proton signal pattern was quite
similar to that of trifloroside (14) [4]. However, three
aromatic protons (dH 7.47, 7.45 and 6.87) of the benzoyl
group in 14 (Table S1) were replaced by a set of signals at
dH 7.25 (ddd, J = 8.0, 2.4, 1.6 Hz), 7.32 (t, J = 8.0 Hz),
7.55 (ddd, J = 8.0, 2.4, 1.6 Hz) and 7.62 (t, J = 1.6 Hz) in
6, which indicated the presence of a m-hydroxy benzoyl
group. The 13C-NMR data of 6 (Table 2) were in good
3.40, m
3.28, m
3.51, m
3.61, dd (12.0, 6.0)
3.86, dd (12.0, 2.4)
171.0
171.4
172.3
20.5
OCOCH3
1.79, s
1.88, s
1.93, s
20.6
20.7
agreement with those of 14, except for the aromatic carbon
signals from the m-hydroxy benzoyl group at dC (131.5
C-100), 118.4 (C-200), 159.2 (C-300), 123.5 (C-400), 130.8 (C-
500), 124.7 (C-600) and 166.3 (C-700) in 6. The HMBC
123

J Nat Med
Table 3 1H- (400 MHz) and 13C-NMR (100 MHz) spectral data of
compound 7 (d in ppm, J in Hz)
H-6 (Fig. 3) observed in the NOE experiments indicated
that H-1, H-3 and H-8 were all cofacial (a-orientation),
thus allowed to establish both of the hydroxy group at C-6
and the methoxy group at C-3 to be b-configuration in 7.
Conversely, the NOESY spectrum of 9 showed cross-peaks
of H-1 with the methoxy group at C-3 and H-8 with both
H-1 and H-6, indicating the hydroxy group at C-6 to be b-
configuration and the methoxy group at C-3 to be a-con-
figuration in 9, which is in agreement with the previously
report [27]. Thus, the structure of 7 was determined to be
6b-hydroxy-3-epi-swertiajaposide A.
Position
dH
dC
96.5
1
5.47, s
3
5.23, d (2.4)
94.4
121.8
156.2
61.9
4
5
6
4.45, ddd (9.2, 5.6, 2.4)
4.13, dd (10.4, 9.2)
4.39, dd (10.4, 5.6)
5.72, ddd (16.8, 10.4, 8.4)
3.54, d (8.4)
7a
70.8
7b
8
134.1
43.8
Notably, 7 was once reported as ‘‘qinjiaoside A’’ [28],
with the methoxy group at C-3 incorrectly determined as in
b-configuration by the evidence of predicting that the
splitting of H-3 was caused by a ‘‘W or M’’ coupling with
H-1(a-orientation). In the present research, the 1H–1H
COSY spectrum of 7 showed a significant cross-peak be-
tween H-3 and H-6 (Fig. 3), rather than between H-3 and
H-1. This fact indicated that the splitting of H-3 was caused
by a homoallylic coupling with H-6. The homoallylic
coupling occurs over five bonds (5J) in allylic systems and
is generally observed when both C–H r bonds are aligned
in parallel with p-orbital of an intervening double bond.
This was supported by an inspection of a Dreiding model
of 7 (Fig. 3), in which both H-3 and H-6 are positioned
axially with the intervening double bond of C-4 and C-5.
Therefore, it is misleading to determine the configuration
of C-3 relying on the coupling constant of H-3 and the
configuration of H-1. The relative configuration of 7 was
finally determined by NOE analysis in the present study.
Accordingly, the ‘‘qinjiaoside A’’ was suggested to be 6b-
hydroxyswertiajaposide A (9) but not the structure reported
by a comprehensive comparison of their NMR data .
The known compounds were identified as gentiopi-
croside (8) [29], 6b-hydroxyswertiajaposide A (9) [27],
20-O-(300-hydroxybenzoyl)-kingiside (10) [23], 4¢¢¢-O-b-D-
glucopyranosyltrifloroside (11) [30], 4¢¢¢-O-b-D-glucopyra-
nosylscabraside (12) [30], gelidoside (13) [31], trifloroside
(14) [4], macrophylloside A (15) [32], scabraside (16) [30]
and deglucosyltrifloroside (17) [33], by comparison of their
spectroscopic data with the reported values.
9
10a
5.28, dd (10.4, 0.8)
5.29, dd (16.8, 0.8)
120.3
10b
11
163.2
1-O-glucosyl
10
20
30
40
50
4.72, d (7.6)
3.22, dd (9.2, 7.6)
3.42, t (9.2)
99.4
74.9
78.1
71.4
78.5
62.6
3.28, t, (9.2)
3.32, m
60a
60b
3.67, dd (12.0, 6.0)
3.89, dd (12.0, 2.4)
3.58, s
3-OCH3
58.2
correlations of H-10 with C-1, H-20 (dH 4.90) with OCO-20
(ester carbonyl carbon, dC 171.0), H-30 (dH 5.45) with
OCO-30 (dC 171.4), H-60 (dH 4.11, 4.18) with OCO-60 (dC
172.3), H-40 (dH 5.20) with C-700, and H-1¢¢¢ with C-300
completely supported the deduced planar structure as
shown in Fig. 1. The absolute configuration of 6 was de-
duced to be (1S,5S,9R) from the same chiral center as
sweroside with unequivocally absolute structure [26].
Thus, the structure of 6 was established and characterized
as 200-dehydroxytrifloroside.
The HR-ESI-MS of 7 displayed a pseudomolecular ion
peak at m/z 405.1396 [M ? H]? corresponding to the
molecular formula of C17H24O11. Analyses of the 1H-, 13C-
and 2D-NMR data (Table 3) revealed that the planar
structure of 7 was identical to that of 6b-hydroxyswer-
tiajaposide A (9) [27]. However, the differences in the
proton signals of H-3 (7: dH 5.23, d, J = 2.4, vs. 9: dH 5.54,
d, J = 1.2), H-6 (7: dH 4.45, ddd, J = 9.2, 5.6, 2.4, vs. 9:
dH 4.15, ddd, J = 2.4, 1.6, 1.2), Ha-7 (7: dH 4.13, dd,
J = 10.4, 9.2, vs. 9: dH 4.41, dd, J = 10.8, 2.4) and Hb-7
(7: dH 4.39, dd, J = 10.4, 5.6, vs. 9: dH 4.43, dd, J = 10.8,
1.6), as well as the carbon signals of C-4 (7: dC 121.8, vs.
9: dC 125.3), C-5 (7: dC 156.2, vs. 9: dC 152.9), C-7 (7: dC
70.8, vs. 9: dC 73.4) and C-9 (7: dC 43.8, vs. 9: dC 46.8)
(Table S1) indicated that 7 was a diastereomer of 9. The
correlation of H-1 with H-3, H-1 with H-8 and H-8 with
All of the compounds were tested for their anti-inflam-
matory activity in vitro through the assay of inhibitory
effect against NO, IL-6 and TNF-a productions in LPS-
stimulated RAW264 macrophage cells (Table 4). Cell
viability under stimulation was evaluated by the CCK-8
assay and all of the test compounds at a dose level up to
100.00 lM or the positive control parthenolide at a dose
level up to 10.00 lM did not exhibit obvious cytotoxicity.
Among them, 1–6 and 10 showed inhibitory activity on IL-
6 production with IC50 values of 51.70–63.80 lM, whereas
12, 13, 15, 16, 17 showed inhibition effects on both of NO
and IL-6 productions with IC50 values of 64.74–94.95 and
123

J Nat Med
Fig. 3 Key NOE correlations
for compound 7 (i) and the
homoallylic coupling between
H-3 and H-6 based on Dreiding
model for the aglycone unit of
compound 7 (ii)
Table 4 Inhibition of compounds 1–17 on LPS-stimulated NO and IL-6 productions in RAW264 cells incubated with 12.50–100.00 lM of the
test compounds or 1.25–10.00 lM of parthenolide in the presence of LPS (0.1 lg/mL) for 24 h
Compound
Inhibition of NO
IC50 (lM)b
Inhibition of IL-6
IC50 (lM)
Compound
Inhibition of NO
IC50 (lM)
Inhibition of IL-6
IC50 (lM)
1
2
3
4
5
6
10
a
NAa
NA
NA
NA
NA
NA
NA
51.70 1.32
58.11 1.34
57.55 3.09
58.63 2.53
61.10 1.86
58.19 4.90
63.80 2.38
12
94.95 3.04
90.49 2.99
64.74 6.74
60.55 2.09
79.10 4.65
NA
51.36 0.28
67.91 0.61
48.91 3.59
75.45 2.84
62.08 1.54
NA
13
15
16
17
7–9, 11, 14
Parthenolide
3.10 0.14
1.93 0.20
Test compounds having IC50 values more than 100 lM
b
The IC50 values were calculated using SigmaPlot (version 11.5) with a four parameter logistic nonlinear regression model with the sample
absorbencies over the log concentrations of the test samples. The results are expressed as the mean SD of three independent experiments
48.91–75.45 lM, respectively. All of the test compounds
exhibited weak inhibition (IC50 [ 100 lM) in a TNF-a
bioassay. The results showed that gentiopicroside (8), the
predominant component in Gentianae scabrae Radix, and
other secoiridoid glycosides exhibited relatively weak anti-
inflammatory activity in vitro when compared with the
positive control parthenolide. However, gentiopicroside
has already been proven to be active in anti-inflammatory
studies in vivo, and some in vitro metabolites of gen-
tiopicroside exhibited promising activity [12], which sug-
gested that the secoiridoid glycosides are likely precursors
of a series of anti-inflammatory metabolites such as dihy-
droisocoumarin [34].
decrease. Additionally, 8-epi-kingiside derivatives 1–3
showed more potent inhibitory activity than kingiside
derivatives 4, 5 and 10 in the assay of inhibitory effect
against IL-6 production, which might be an important
factor for optimum anti-inflammation through inhibition of
IL-6 production.
Experimental
Experimental instrumentation and chemicals
1H-, 13C- and 2D-NMR spectra were recorded on a JEOL
ECX400 Delta spectrometer with TMS as internal stan-
dard. Chemical shifts (d) are expressed in ppm and J values
are given in Hz. HR-ESI-MS measurements were carried
out on a Accella HPLC system equipped with an Orbitrap-
EX mass spectrometer (Thermofisher Scientific, Bremen,
Germany). IR spectra were recorded on a JASCO FT/IR-
460 Plus spectrophotometer (JASCO International Co.,
Ltd., Tokyo, Japan). Optical rotations were measured on a
JASCO P2100 digital polarimeter. Column chromatogra-
phy (CC) was performed with normal-phase silica gel (C-
300HG, 40–50 lm, Wako, Tokyo, Japan) and reverse-
phase silica gel (ODS-A, 75 lm, YMC, Kyoto, Japan).
By comparing the IC50 value of test compounds
(Table 4), the substitution of the benzoyl moiety at C-20 in
12 (vs. 11) and 16 (vs. 14), as well as the presence of
hydroxy group at C-5 in 13 (vs. 14), enhanced the in-
hibitory effect against both NO and IL-6 productions,
which agreed well with previously reported results [8, 35].
The absence of the glucose unit at C-300 in 17 (vs. 11 and
14) resulted in obvious increase of the inhibitory effect;
however, it had the opposite effect on inhibitory activity
against IL-6 production when the C-20 substituted with a
benzoyl moiety in 12 (vs. 16). Furthermore, the hydroxy
group at C-200 in 14 (vs. 6) triggered obvious activity
123

J Nat Med
Preparative HPLC was performed on a Waters HPLC
system (Waters Ltd., Milford, MA, USA) equipped with a
Delta 600 pump and 2489 UV/Visible detector on YMC-
pack R&D ODS-A column (250 9 20 mm i.d., 5 lm).
TLC was performed to monitor fractions on silica gel 60
F254 plates (250 lm thickness; Merck, Darmstadt, Ger-
many). All solvents were of analytical grade and purchased
from Wako. For the hydrolysis and sugar analysis, hy-
drochloric acid, pyridine, phenyl isothiocyanate, L-cysteine
methyl ester hydrochloride and D-(?)-glucose (purity
C98 %) were obtained from Wako. Amberlite IRA-400
resin and L-(-)-glucose (purity C99 %) were obtained
from Sigma-Aldrich (St. Louis, MO). For work with cell
cultures, Minimum Essential Medium (MEM), PBS and
0.05 % trypsin–EDTA were obtained from Gibco (Grand
Island, NY, USA). FBS was obtained from HyCloneꢂ
(Waltham, MA, USA). LPS, NaNO2, N-1-naphthylethylene
diamine dihydrochloride, sulfanilamide and DMSO were
obtained from Wako. Parthenolide (purity C90 %) was
obtained from Sigma-Aldrich. Cell counting kit-8 (CCK-8)
was obtained from Dojindo Lab (Kumamoto, Japan) and
commercially available ELISA kits were obtained from
Pepro Tech (Rocky Hill, NJ, USA).
by preparative HPLC eluting with CH3CN–H2O (55:45,
v/v), yielded 17 (131.6 mg). Fraction 6 (3.9 g) was sub-
jected to ODS-A CC (50 9 5 cm i.d.) eluting with ace-
tone–H2O (1:9 ? 9:1, v/v) to afford 14 (800.1 mg) and
nine other subfractions. Subfraction 6–3 (158 mg) was
purified by preparative HPLC (CH3CN–H2O, 23:77, v/v) to
yield 2 (9.4 mg) and 4 (3.6 mg). Preparative HPLC al-
lowed for the purification of subfractions 6–4 (48 mg), 6–5
(108 mg) and 6–9 (167 mg) yielded 1 (6.9 mg), 13
(97.6 mg) and 16 (122.5 mg), respectively. Moreover,
purification of subfraction 6–7 (67 mg) yielded 6 (6.1 mg)
and 15 (23.5 mg). Fraction 7 (7 g) was subjected to the
ODS-A CC and eluted with MeOH–H2O (15:85 ? 95:5,
v/v) to afford 16 subfractions. Subfraction 7–5 (2.1 g) was
further separated over silica gel CC (60 9 4 cm i.d.) with a
gradient elution of CHCl3–MeOH (9:1 ? 3:1, v/v), fol-
lowed by purification through the preparative HPLC with
CH3CN–H2O (15:85, v/v) to yield 7 (18.9 mg), 8
(245.2 mg) and 9 (8.2 mg). With a similar process, sub-
fraction 7–8 (55 mg) was purified to yield 5 (5.1 mg) and
10 (3.7 mg), subfraction 7–10 (11 mg), 7–13 (39 mg),
7–14 (16 mg) were purified to yield 3 (2.1 mg), 11
(57.6 mg) and 12 (9.2 mg), respectively.
60-O-(300-hydroxybenzoyl)-8-epi-kingiside (1): white,
25
amorphous solid; ½aꢁD -28.72ꢁ (c 0.05, MeOH); IR (KBr)
Materials
mmax 3393, 1715, 1636, 1602, 1588, 1506, 1457, 1389,
1290, 1094, 1078, 758 cm-1
;
1H-NMR (CD3OD,
Gentianae Scabrae Radix (LOT no. 1201C010901, pro-
duced in Liaoning Province, China) was purchased from
Tochimoto Tenkaido Co. (Osaka, Japan) on 21 June 2012.
The botanical source of the material was identified as G.
scabra by genetic analysis of nucleotide sequence of the
ITS region, which was completely identical to those reg-
istered in GenBank (JQ890595, AB564723, GQ864015,
GQ864016). A voucher sample (TMPW No. 27499) has
been deposited in the Museum of Materia Medica, Institute
of Natural Medicine, University of Toyama, Japan.
400 MHz) and 13C-NMR (CD3OD, 100 MHz) spectral
data are given in Table 1; HR-ESI-MS m/z 525.1596
[M ? H]? (calcd for C24H29O13 525.1603).
60-O-(400-hydroxybenzoyl)-8-epi-kingiside (2): white,
25
amorphous solid; ½aꢁD -28.20ꢁ (c 0.03, MeOH); IR
(KBr) mmax 3394, 1704, 1637, 1608, 1591, 1557, 1517,
1440, 1389, 1281, 1166, 1097, 1074, 854, 770 cm-1
;
1H-NMR (CD3OD,400 MHz) and 13C-NMR (CD3OD,
100 MHz) spectral data are given in Table 1; HR-ESI-MS
m/z 525.1603 [M ? H]? (calcd for C24H29O13 525.1603).
Extraction and isolation
60-O-(200,300-dihydroxybenzoyl)-8-epi-kingiside
(3):
25
white, amorphous solid; ½aꢁD -21.32ꢁ (c 0.05, MeOH); IR
Gentianae Scabrae Radix (1.5 kg) was extracted succes-
sively with CHCl3 (9 L 9 4, 1.5 h each time) and MeOH
(9 L 9 4, 1.5 h each time) under sonication at room tem-
perature. The solvents were removed in vacuo to afford
CHCl3 extract (49 g) and MeOH extract (527 g). CHCl3
extract (44 g used), which had relatively strong anti-in-
flammatory activity, was then separated over a silica gel
column (75 9 6 cm i.d.) eluting with a gradient solvent
condition of CHCl3 and MeOH (0:1 ? 1:1, v/v) to give
seven fractions. Fraction 3 (10.6 g) was subjected to silica
gel CC (60 9 4 cm i.d.) eluting with hexane–acetone
(9:1 ? 1:1, v/v) to afford nine subfractions. Further
separation of subfraction 3–8 (2.2 g) over ODS-A CC
(40 9 3 cm i.d.; acetone–H2O, 1:1 ? 1:0, v/v) and then
(KBr) mmax 3395, 1704, 1637, 1590, 1471, 1440, 1389,
1
1306, 1278, 1157, 1077, 757 cm-1; H-NMR (400 MHz,
CD3OD) and 13C-NMR (100 MHz, CD3OD) spectral data
are given in Table 1; HR-ESI-MS m/z 541.1548 [M ? H]?
(calcd for C24H29O14 514.1552).
25
20-O-benzoyl kingiside (4): white, amorphous solid; ½aꢁD
-88.27ꢁ (c 0.03, MeOH); IR (KBr) mmax 3411, 1716, 1645,
1557, 1540, 1506, 1454, 1386, 1302, 1271, 1070, 715 cm-1
;
1H-NMR (CD3OD, 400 MHz) and 13C-NMR (CD3OD,
100 MHz) spectral data are given in Table 1; HR-ESI-MS
m/z 509.1650 [M ? H]? (calcd for C24H29O12 509.1654).
20-O-(200,300-dihydroxybenzoyl) kingiside (5): white,
25
amorphous solid; ½aꢁD -87.76ꢁ (c 0.05, MeOH); IR (KBr)
123

J Nat Med
mmax 3435, 1691, 1631, 1598, 1471, 1440, 1386, 1352,
a vehicle control. Parthenolide was used as a positive
control. Each sample was tested in triplicate of three in-
dependent experiments.
1
1079, 758 cm-1; H-NMR (CD3OD, 400 MHz) and 13C-
NMR (CD3OD, 100 MHz) spectral data are given in
Table 1; HR-ESI-MS m/z 541.1549 [M ? H]? (calcd for
C24H29O14 541.1552).
Cell viability assay
200-dehydroxytrifloroside (6): white, amorphous solid;
25
½aꢁD -160.45ꢁ (c 0.02, MeOH); IR (KBr) mmax 3423, 1753,
Cell viability was assessed by CCK-8 assay, according to
the manufacturer’s instructions. Briefly, after treatment
with test compound (maximum concentration of
100.00 lM), parthenolide (maximum concentration of
10.00 lM) or vehicle control, 10 lL WST-8 solution was
added to each well and incubated for 2 h at 37 ꢁC. The
absorbance at 405 nm was measured on a microplate
reader. The percentage of cell viability was calculated ac-
cording to the following formula: Cell viability ð%Þ ¼
ðMean absorbance of test samples = mean absorbance of
vehicle controlÞ ꢂ 100 %.
1619, 1489, 1445, 1406, 1371, 1278, 1231, 1069, 991, 903,
1
756 cm-1; H-NMR (CD3OD, 400 MHz) and 13C-NMR
(CD3OD, 100 MHz) spectral data are given in Table 2;
HR-ESI-MS m/z 767.2378 [M ? H]? (calcd for C35H43O19
767.2393).
6b-hydroxy-3-epi-swertiajaposide A (7): white, amor-
25
phous solid; ½aꢁD -79.34ꢁ (c 0.20, MeOH); IR (KBr) mmax
3421, 1715, 1634, 1409, 1314, 1275, 1074, 977, 949 cm-1
;
1H-NMR (CD3OD, 400 MHz) and 13C-NMR (CD3OD,
100 MHz) spectral data are given in Table 3; HR-ESI-MS
m/z 405.1396 [M ? H]? (calcd for C17H25O11 405.1391).
Measurement of NO production
Acid hydrolysis and sugar analyses of compounds 1–7
NO production was determined by measuring the accumu-
lation of nitrite in the cell culture supernatant using Griess
reagent [8]. The absorbance at 560 nm was measured on a
microplate reader. The nitrite concentration was calculated
from the calibration curve prepared using NaNO2 stan-
dards. Inhibition activity was calculated according to the
following formula: Inhibition activity ð%Þ ¼ ½ðAc ꢃ AsÞ=
ðAc ꢃ AbÞꢁ ꢂ 100 %, where Ac, As and Ab represent the
nitrite concentration (lM) of vehicle control, test sample
and blank, respectively.
Acid hydrolysis of 1–7 was carried out using a procedure
similar to that reported in [36]. Briefly, each compound
(0.5 mg) was dissolved in 1.0 N HCl (1.0 mL) and heated
in a water bath (90 ꢁC) for 3 h. The filtrate from the hy-
drolysate was neutralized with Amberlite IRA-400 resin
and concentrated to dryness in vacuo. The residue was then
dissolved in pyridine (0.1 mL) and reacted successively
with L-cysteine methyl ester hydrochloride and phenyl
isothiocyanate at 60 ꢁC for 1 h. Finally, the reaction mix-
ture was subjected to HPLC–PDA system to determine the
D/L configuration of sugar moieties by comparing their
retention times (tR) with those of authentic D- and L-glu-
cose derivatives (tR: D-glucose, 21.6 min; L-glucose,
20.4 min).
Measurement of IL-6 and TNF-a productions
IL-6 and TNF-a levels in the cell-free supernatant were
measured using commercially available ELISA kits, ac-
cording to the manufacturer’s instructions. Absorbance at
405 nm was measured on a microplate reader. Inhibition
activity of the test compounds was calculated according
to the following formula: Inhibition activity ð%Þ ¼
½ðAc ꢃ AsÞ=ðAc ꢃ AbÞꢁ ꢂ 100 %, where Ac, As and Ab are
absorbance of vehicle control, test sample and blank,
respectively.
Cell culture and treatment
The RAW264 mouse macrophage cells (Riken Cell Bank,
Tsukuba, Japan) were cultured in MEM supplemented with
10 % FBS in a humidified 5 % CO2 at 37 ꢁC. For all ex-
periments, the cells were grown to 80–90 % confluence,
then harvested with trypsin–EDTA and diluted to a sus-
pension in fresh media. The suspended cells were seeded
into 96-well plates at density of 1 9 105 cells/well and
allowed to adhere for 6 h, followed by treatment for 24 h
in the presence or absence of test samples with LPS
(0.1 lg/mL). For measurement of NO, IL-6 and TNF-a
productions, cells treated with 0.1 % DMSO in medium
were served as a blank, while cells treated with 0.1 %
DMSO in the presence of LPS (0.1 lg/mL) were served as
Acknowledgments The authors would like to thank Dr. Shengli
Wei from the School of Chinese Pharmacy, Beijing University of
Chinese Medicine, Beijing, China, for helping with sample collection.
They are also grateful to Dr. Feng Li from Joint Usage/Research
Center for Science-Based Natural Medicine Research Promotion
Office, Institute of Natural Medicine, University of Toyama, Toyama,
Japan, for assistance with 2D-NMR experimental acquisition and
troubleshooting. This work was supported by JSPS KAKENHI Grant
Number 24406005.
123

J Nat Med
19. Scheller J, Chalaris A, Schmidt-Arras D, Rose-John S (2011) The
pro- and anti-inflammatory properties of the cytokine interleukin-
6. Biochim Biophys Acta 1813:878–888
20. Bauqh JA, Bucala R (2001) Mechanisms for modulating TNF
alpha in immune and inflammatory disease. Curr Opin Drug
Discov Dev 4:635–650
21. Garcia J, Mpondo EM, Kaouadji M (1990) Kingiside and derivative
from Gentiana pyrenaica. Phytochemistry 29:3353–3355
22. Garcia J, Lavaitte S, Gey C (1989) 8-epikingiside and its vanillate
ester, isolated from Gentiana pyrenaica. Phytochemistry
28:2199–2201
23. Xu M, Wang D, Zhang YJ, Yang CR (2008) Iridoid glucosides
from Gentiana rhodantha. J Asian Nat Prod Res 10:491–498
24. Inouye H, Yoshida T, Tobita S, Tanaka K, Nishioka T (1970)
Absolute struktur des oleuropeins und einiger verwandter glu-
coside. Tetrahedron Lett 28:2459–2464
References
1. Flora of China Editorial Committee (1995) Flora of China
(Gentianaceae), vol 16. Science, Beijing, pp 1–139
2. The State Pharmacopoeia Commission of P. R. China (2010)
Pharmacopoeia of the People’s Republic of China (English ver-
sion), vol 1. China Medical Science, Beijing, pp 205–206
3. The Ministry of Health, Labour and Welfare (2011) The Japanese
Pharmacopoeia (English version), 16th edn. p 1669
4. Ikeshiro Y, Tomita Y (1983) A new bitter secoiridoid glucoside
from Gentiana scabra var. buergeri. Planta Med 48:169–173
5. Tang HQ, Tan RX (1997) Glycosides from Gentiana scabra.
Planta Med 63:388
6. Kakuda R, Iijima T, Yaoita Y, Machida K, Kikuchi M (2001)
Secoiridoid glycosides from Gentiana scabra. J Nat Prod
64:1574–1575
25. Inouye H, Yoshida T, Tobita S, Kanaka K, Nishioka T (1974)
Uber die monoterpenglucoside und verwandte naturstoffe. XXII.
Absolutstrukturen des oleuropeins, Kingisids und Morronisids.
Tetrahedron 30:201–209
26. Inouye H, Yoshida T, Nakamura Y, Tobita S (1970) Uber die
monoterpenglucoside. XI. chemische korrelation des asperulosids
mit swerosid. Chem Pharm Bull 18:1889–1899
27. Kikuchi M, Kakuda R, Kikuchi M, Yaoita Y (2005) Secoiridoid
glycosides from Gentiana scabra. J Nat Prod 68:751–753
28. Liu YH, Li XC, Liu YQ, Yang CR (1994) Iridoid glycosides from
Gentiana macrophylla. Acta Botanica Yunnanica 16:85–89
29. Mpondo EM, Chulia AJ (1988) 60-O-b-D-Glucosyl aentiopi-
croside: a new secoiridoid from Gentiana asclepiadea. Planta
Med 54:185–186
30. Kim JA, Son NS, Son JK, Jahng Y, Chang HW, Jang TS, Na MK,
Lee SH (2009) Two new secoiridoid glycosides from the rhi-
zomes of Gentian scabra Bunge. Arch Pharm Res 32:863–867
31. Calis¸ I, Ru¨egger H, Chun Z, Sticher O (1990) Secoiridoid glu-
cosides isolated from Gentiana gelida. Planta Med 56:406–409
32. Tan RX, Wolfender JL, Zhang LX, Ma WG, Fuzzati N, Marston A,
Hostettmann K (1996) Acyl secoiridoids and antifungal constituents
from Gentiana macrophylla. Phytochemistry 42:1305–1313
33. Inouye H, Ueda S, Nakamura Y, Inoue K, Hayano T, Matsumura
7. Jiang RW, Wong KL, Chan YM, Xu HX, Pui-Hay But P, Shaw
PC (2005) Isolation of iridoid and secoiridoid glycosides and
comparative study on Radix Gentianae and related adulterants by
HPLC analysis. Phytochemistry 66:2674–2680
8. Wang SF, Xu YM, Jiang W, Zhang YF (2013) Isolation and
identification of constituents with activity of inhibiting nitric
oxide production in RAW 264.7 macrophages from Gentiana
triflora. Planta Med 79:680–686
9. Kakuda R, Iijima T, Yaoita Y, Machida K, Kikuchi M (2002)
Triterpenoids from Gentiana scabra. Phytochemistry 59:791–794
10. Xu M, Wang D, Zhang YJ, Yang CR (2007) Dammarane triter-
penoids from the roots of Gentiana rigescens. J Nat Prod
70:880–883
11. Xu M, Yang CR, Zhang YJ (2009) Minor antifungal aromatic
glycosides from the roots of Gentiana rigescens (Gentianaceae).
Chin Chem Lett 20:1215–1217
12. Park KS, Kim BH, Chang IM (2010) Inhibitory potencies of
several iridoids on cyclooxygenase-1, cyclooxygnase-2 enzymes
activities, tumor necrosis factor-a and nitric oxide production
in vitro. Evid Based Complement Altern Med 7:41–45
13. Saravanan S, Hairul Islam VI, Prakash Babu N, Pandikumar P,
Thirugnanasambantham K, Chellappandian M, Simon Durai Raj
C, Gabriel Paulraj M, Ignacimuthu S (2014) Swertiamarin at-
tenuates inflammation mediators via modulating NF-jB/I jB and
JAK2/STAT3 transcription factors in adjuvant induced arthritis.
Eur J Pharm Sci 56:70–86
¨
(1974) Uber die monoterpenglucoside und verwandte
H
naturstoffe-XXIV: Triflorosid, ein neues secoiridoidglucosid aus
Gentiana triflora var. japonica. Tetrahedron 30:571–577
34. Wang ZG, Wang SS, Sun YJ, Wang HY, Chen G, Wang XJ,
Hattori M, Zhang HL (2014) New analytical method for the study
of the metabolism of gentiopicroside in rats after oral adminis-
tration by LC–TOF-MS following picolinoyl derivatization. J Sep
Sci 37:237–243
35. Lv T, Xu M, Wang D, Zhu HT, Yang CR, Zhang TT, Zhang YJ
(2012) The chemical constituents from the roots of Gentiana
crassicaulis and their inhibitory effects on inflammatory me-
diators NO and TNF-a. Nat Prod Bioprospect 2:217–221
36. Tanaka T, Nakashima T, Ueda T, Tomii K, Kouno I (2007) Facile
discrimination of aldose enantiomers by reversed-phase HPLC.
Chem Pharm Bull 55:899–901
14. Kondo Y, Takano F, Hojo H (1994) Suppression of chemically
and immunologically induced hepatic injuries by gentiopicroside
in mice. Planta Med 60:414–416
´
´
ˇ ´
´
´
15. Mihailovic V, Katanic J, Misic D, Stankovic V, Mihailovic M,
´
ˇ ´
´
´
´
Uskokovic A, Arambasic J, Solujic S, Mladenovic M, Stankovic
N (2014) Hepatoprotective effects of secoiridoid-rich extracts
from Gentiana cruciata L. against carbon tetrachloride induced
liver damage in rats. Food Funct 5:1795–1803
16. Chiba K, Yamazaki M, Kikuchi M, Kakuda R, Kikuchi M (2011)
New physiological function of secoiridoids: neuritogenic activity
in PC12h cells. J Nat Med 65:186–190
17. Nathan CF, Hibbs JB Jr (1991) Role of nitric oxide synthesis in
macrophage antimicrobial activity. Curr Opin Immunol 3:65–70
18. Kobayashi Y (2010) The regulatory role of nitric oxide in
proinflammatory cytokine expression during the induction and
resolution of inflammation. J Leukoc Biol 88:1157–1162
123