Koumine, Humantenine, and Yohimbane Alkaloids from Gelsemium elegans

Article abstract of DOI:10.1021/np5009619

Nine new alkaloids of the koumine (1-4), humantenine (5-7), and yohimbane (8, 9) types as well as 12 known analogues were isolated from the leaves and vine stems of Gelsemium elegans. Compound 1 is the first N-4-demethyl alkaloid of the koumine type, compound 7 is the first nor-humantenine alkaloid, and compounds 8 and 9 are the first N-1-oxide and the first seco-E-ring alkaloids, respectively, of the yohimbane type. Compounds 1 and 7 exhibited moderate cytotoxicity against five human tumor cell lines with IC₅₀ values in the range 4.6-9.3 μM. (Figure Presented)

Full text of DOI:10.1021/np5009619

Article
pubs.acs.org/jnp
Koumine, Humantenine, and Yohimbane Alkaloids from Gelsemium
elegans
You-Kai Xu,*,† Lin Yang,† Shang-Gao Liao,‡ Pei Cao,§ Bin Wu,§ Hua-Bin Hu,† Juan Guo,†
and Ping Zhang†
†Key Laboratory of Tropical Plant Resources and Sustainable Use, Xishuangbanna Tropical Botanical Garden, Chinese Academy of
Sciences, Menglun, Mengla, Yunnan 666303, People’s Republic of China
‡Engineering Research Center for the Development and Application of Ethnic Medicines and TCM, School of Pharmacy, Guiyang
Medical College, 9 Beijing Road, Guiyang, Guizhuo 550004, People’s Republic of China
§State Key Laboratory of Phytochemistry and Plant Resources in West China, Kunming Institute of Botany, Chinese Academy of
Science, Kunming, Yunnan 650201, People’s Republic of China
S
* Supporting Information
ABSTRACT: Nine new alkaloids of the koumine (1−4),
humantenine (5−7), and yohimbane (8, 9) types as well as 12
known analogues were isolated from the leaves and vine stems
of Gelsemium elegans. Compound 1 is the first N-4-demethyl
alkaloid of the koumine type, compound 7 is the first nor-
humantenine alkaloid, and compounds 8 and 9 are the first N-
1-oxide and the first seco-E-ring alkaloids, respectively, of the
yohimbane type. Compounds 1 and 7 exhibited moderate
cytotoxicity against five human tumor cell lines with IC50
values in the range 4.6−9.3 μM.
he genus Gelsemium (Loganiaceae), which comprises
Tthree species in Asia (G. elegans) and North America (G.
sempervirens and G. rankinii), has been recognized as a toxic
plant and used as a folk medicine to treat migraines, neuralgia,
sciatica, cancer, and various types of sores.1 A large number of
indole, bisindole, and monoterpenoid alkaloids have been
isolated from the Gelsemium genus.1−4 Pharmacological experi-
ments showed that these Gelsemium alkaloids possessed
cytotoxic, analgesic, anxiolytic, anti-inflammatory, and immu-
nomodulating activities.1,5 During an earlier investigation, five
monoterpenoid indole alkaloids and a new monoterpenoid
alkaloid that selectively inhibited the growth of A-549 tumor
cell lines were isolated from G. elegans.5,6 In the current
investigation, nine new alkaloids, including four of the koumine
type (1−4), three of the humantenine type (5−7), and two of
the yohimbane type (8, 9), as well as 12 known alkaloids were
isolated from the leaves and perennial vine stems of G. elegans.
Among the known compounds, 4R-koumine N-oxide, 11-
methoxy-19-(R)-hydroxygelselegine, and epi-koumidine were
isolated as natural products for the first time, and venoterpine
was isolated from the Loganiaceae family for the first time.
Evaluations of cytotoxicity and nitric oxide production
inhibition revealed that compounds 1 and 7 exhibited moderate
cytotoxic activity against five human tumor cell lines with IC50
values in the range 4.6−9.3 μM. This report describes the
extraction, isolation, structure elucidation, cytotoxicity, and
nitric oxide production inhibitory assays of these compounds.
RESULTS AND DISCUSSION
■
Compound 1 was obtained as a white, amorphous powder and
was assigned the molecular formula C19H18N2O on the basis of
its 13C NMR data and an HREIMS ion at m/z 290.1416 [M]+
(calcd 290.1419), showing 16 mass units less than that of
koumine (10),7 a major alkaloid also isolated during the
investigation. The UV (λmax 221.0 and 261.6 nm) and NMR
(Tables 1 and 2) data indicated the presence of an indolenine
1
chromophore. The H and 13C NMR data (Tables 1 and 2)
were similar to those of koumine,7 including signals assignable
to four aromatic protons [δH 7.54, dd, J = 7.5, 1.0 Hz (H-12);
7.38, td, J = 7.5, 1.0 Hz (H-11); 7.28, td, J = 7.5, 1.0 Hz (H-10);
7.09, dd, J = 7.5, 1.0 Hz (H-9)], one vinyl group [δH 5.28, d, J =
17.8 Hz (H-18); 5.09, d, J = 11.2 Hz (H-18); 4.84, dd, J = 17.8,
11.2 Hz (H-19)], and an oxymethine [δH 4.99, br s (H-3)], two
oxymethylene [δH 4.31, dd, J = 12.0, 4.3 Hz (H-17a); 3.80, d, J
= 12.0 Hz (H-17b)], and an aminomethine [δH 4.49, br s, (H-
5)] proton. The exceptions involved the presence of additional
signals for an imino group [δH 8.52 (s), δC 175.0] and the
absence of signals for an N-4-methyl group. These observations
suggested that 1 was likely an N-4-demethly-21-dehydro
derivative of koumine. HMBC correlations between H-21 (δH
8.52, s) and C-5 (δC 61.0) and C-19 (δC 136.6) further
validated the deduction (Figure 1). The relative and absolute
configurations of 1 were established, on the basis of ROESY
Received: December 1, 2014
© XXXX American Chemical Society and
American Society of Pharmacognosy
A
DOI: 10.1021/np5009619
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
1
Table 1. H NMR Spectroscopic Data for 1−4 (δ in ppm, J
in Hz)
a
b
b
c
position
1
2
3
4
d
d
3
5
6
4.99, br s
4.49, br s
4.84,
2.97, br s
4.84,
2.98, br s
4.97, br s
4.28, d (5.2)
1.67, dt (13.6, 2.23, dd (14.3, 2.34, dt (13.7, 2.43, d (17.1)
1.9)
3.2)
1.9)
2.63, dd (13.6, 2.40, br d
2.45, dd (13.7, 2.86, dd (17.1,
3.0)
(14.3)
4.3)
3.7)
9
7.09, dd (7.5,
1.0)
7.65, dd (7.5, 7.94, dd (7.5, 7.40, d (7.4)
1.1)
1.1)
10
11
12
14
7.28, td (7.5,
1.0)
7.37, td (7.5,
1.1)
7.25, td (7.5,
1.1)
7.26, t (7.4)
7.36, t (7.4)
7.38, td (7.5,
1.0)
7.42, td (7.5,
1.1)
7.32, td (7.5,
1.1)
7.54, dd (7.5,
1.0)
7.54, dd (7.5, 7.44, dd (7.5, 7.57, d (7.4)
1.1)
1.1)
1.83, dt (15.0, 1.93, br d
2.6) (14.8)
1.93, br d
(14.8)
1.57, d (15.0)
2.72, dt (15.0, 2.68, dt (14.8, 2.63, dt (14.8, 2.65, dt (15.0,
3.9)
3.7)
3.7)
3.7)
15
16
17
2.02, br d
(12.2)
2.81, br d
(12.2)
2.31, br d
(12.1)
2.49, br d
(12.5)
d
d
2.28, br d
(12.2)
2.73,
2.89, br d
(12.1)
3.50,
3.80, d (12.0)
3.58, d (11.9) 3.58, d (11.9) 3.69, d (12.9)
4.31, dd (12.0, 4.25, dd (11.9, 4.28, dd (11.9, 4.25, dd (12.9,
4.3)
4.5)
4.5)
5.2)
d
18
5.09, d (11.2)
4.73, dd (11.0, 4.83,
7.1)
0.92, d (6.4)
correlation analysis (Figure 1) and by comparing its electronic
circular dichroism (ECD) spectrum to that of koumine (10),
respectively (Figure 3). The structure of 1 was thus established
as N-4-demethyl-21-dehydrokoumine. Compound 1 is the first
koumine-type alkaloid lacking an N-4 methyl group.
e
5.28, d (17.8)
4.96, dd (11.0, 4.90,
1.9)
d
d
19
21
4.84, dd (17.8, 4.79,
11.2)
4.84,
2.91, q (6.4)
8.52, s
4.24, s
3.92, s
3.87, d (14.8)
4.16, d (14.8)
3.82, s
Compounds 2 and 3 were obtained as an inseparable 1:2
1
mixture as determined by H NMR peak integration. The
N4-Me
2.78, s
2.75, s
b
compounds shared the same molecular formula of C20H22N2O2
as determined by their 13C NMR data and HRESIMS ions at
m/z 323.1749 ([M + H]+) These structurally similar koumine
a
Recorded at 500 MHz in methanol-d4. Recorded at 600 MHz in
methanol-d4. Recorded at 600 MHz in CDCl3 (80%) + methanol-d4
(20%) (v/v). Overlapped by other signals. Overlapped by solvent
signals.
c
d
e
1
alkaloids were readily identified by two sets of peaks in the H
NMR data (Table 1) and by an HSQC experiment, in which
the 1,2-disubstituted aromatic groups [δH 7.65/7.94 (H-9, dd, J
= 7.5/7.5, 1.1/1.1 Hz), 7.54/7.44 (H-12, dd, J = 7.5/7.5, 1.1/
1.1 Hz), 7.42/7.32 (H-11, td, J = 7.5/7.5, 1.1/1.1 Hz), and δH
7.37/7.25 (H-10, td, J = 7.5/7.5, 1.1/1.1 Hz)], two vinylic
groups [δH 4.96/4.90 (Ha-18), 4.73/4.83(Hb-18), and 4.79/
4.84 (H-19)], and six oxymethine/methylene [δH 4.84/4.84
(H-3), δH 3.58/3.58 (H-17α, d, J = 11.9/11.9 Hz), δH 4.25/
4.28 (H-17β, dd, J = 11.9/11.9, 4.5/4.5 Hz)], two amino-
methine [δH 2.97/2.98 (H-5, br s/br s)], and two methylene
[δH 1.93/1.93 (H-14α, dd, br d, J = 14.8/14.8 Hz) and δH 2.68/
2.63 (H-14β, dt, J = 14.8/14.8, 3.7/3.7 Hz)] protons were
identified. The 1H and 13C NMR data of 2/3 (Tables 1 and 2)
are similar to those of koumine (10),7 with the distinct
difference being the replacement of signals from a methylene
(δH 3.11, br d, J = 14.2 Hz; δH 3.17, br d, J = 14.2 Hz, δC 57.2)
in koumine by those from an oxymethine (δH 4.24/3.92, s; δC
98.6/103.3) group. These observations suggested that 2 and 3
were likely two 21-hydroxy derivatives of koumine. HMBC and
ROESY correlation analyses (Figure 1) as well as ECD
spectroscopic comparison [Figure 3, with koumine (10)]
further confirmed this conclusion. In particular, the presence of
a 21-hydroxy group was confirmed by HMBC correlations
(Figure 1) between H-21 and C-7 and C-15 and between H3-
N-Me and C-21 in both compounds. Moreover, the ROESY
correlation of H-21/H-6β indicated that the 21-OH group was
α-oriented in 2 (Figure 1), whereas the ROESY correlation H-
21/H-19 was consistent with a β-orientation of the 21-OH
group in 3 (Figure 1). Compounds 2 and 3 were therefore
determined to be 21α- and 21β-hydroxykoumine, respectively.
Failure to separate the two compounds is attributed to the
presence of a 21-aza-hemiacetal group in both compounds,
which allows interconversion of the two epimers. Detection of a
relatively weaker [M + H − H2O]+ ion (m/z 305) but a
stronger total ion current of its product ions in the LC-
(+)-ESIMS chromatogram of the ethanolic extract suggested
that compounds 2 and 3 were natural products (Supporting
Information S2/3−9).
Compound 4 was obtained as a white, amorphous powder
and was assigned the molecular formula C20H24N2O3 on the
basis of 13C NMR data and the HREIMS ion at m/z 340.1792
[M]+ (calcd 340.1787), showing 16 mass units more than that
of (19S)-hydroxydihydrokoumine,7 a known alkaloid that was
also isolated during this investigation. Comparison of the NMR
data (Tables 1 and 2) of 4 with those of (19S)-
hydroxydihydrokoumine7 revealed that their NMR signals
were similar, the major differences being the deshielding of
peaks for N-4-CH3 (δH 3.82, s; δC 56.9), CH-5 (δH 4.28, d, J =
5.2 Hz; δC 70.1), and CH2-21 (δH 4.16, d, J = 14.8 Hz; 3.87, d, J
= 14.8 Hz; δC 59.6) in 4. These observations suggested that the
B
DOI: 10.1021/np5009619
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Table 2. 13C NMR Spectroscopic Data for 1−4 (δ in ppm)
a
b
b
c
position
1
2
3
4
2
186.2, C
188.2, C
188.2, C
183.2, C
3
71.9, CH
61.0, CH
36.1, CH
58.2, C
72.2, CH
59.2, CH
29.4, CH
58.0, C
72.1, CH
57.5, CH
37.6, CH
56.5, CH
145.1, C
70.4, CH
70.1, CH
30.4, CH
55.7, CH
141.9, C
5
6
7
8
144.4, C
145.1, C
9
123.9, CH
128.2, CH
129.9, CH
121.7, CH
154.9, C
124.7, CH
127.9, CH
129.6, CH
121.7, CH
155.2, C
129.0, CH
127.2, CH
129.0, CH
120.7, CH
155.2, C
124.2, CH
126.7, CH
129.2, CH
121.6, CH
153.5, C
10
11
12
13
14
15
16
17
18
19
20
21
N4-Me
25.8, CH2
34.9, CH
36.1, CH
61.7, CH2
119.1, CH2
136.6, CH
51.1, C
25.7, CH2
25.4, CH
41.1, CH
62.4, CH2
120.8, CH2
135.4, CH
51.7, C
24.9, CH2
31.8, CH
33.9, CH
61.9, CH2
116.8, CH2
139.1, CH
51.7, C
23.8, CH2
26.3, CH
32.3, CH
67.9, CH2
17.2, CH3
65.7, CH
48.2, C
175.0, CH
98.6, CH
41.3, CH3
103.3, CH
43.8, CH3
59.6, CH2
56.9, CH3
a
b
Recorded at 100 MHz in methanol-d4. Recorded at 150 MHz in
c
methanol-d4. Recorded at 150 MHz in CDCl3 (80%) + methanol-d4
(20%) (v/v).
Figure 2. COSY (bold bond) and key HMBC and ROESY (5−7)
correlations of compounds 5−9.
Figure 1. COSY (bold bond) and key HMBC and ROESY
correlations of 1 and 2/3.
structural variation of this compound occurs around N-4.
Considering 4 had an additional oxygen atom relative to (19S)-
hydroxydihydrokoumine, compound 4 was proposed to be the
N-4-oxide derivative of (19S)-hydroxydihydrokoumine. The
deshielding of peaks for H-15 and H-16 was also consistent
Figure 3. ECD spectra of compounds 1−7 and 10.
8
1
with the anisotropic effect of the N-4-oxide. Using H−1H
COSY, HSQC, HMBC, ROESY (Supporting Information S4),
and ECD (Figure 3) data, the structure of 4 was confirmed and
the 1H and 13C NMR data were assigned (Supporting
Information S4). Thus, the structure of compound 4 was
assigned as (19S)-hydroxydihydrokoumine N-4-oxide.
66.1), which suggested that 5 was the 14-hydroxy derivative of
20-hydroxydihydrorankinidine (5a). 1H−1H COSY correlations
as well as HMBC correlations between H-3 (δH 3.34) and the
oxymethine carbon (δC 66.1) and between the protons at δH
4.46 and C-16 (δC 39.3) and C-20 (δC 69.3) (Figure 2)
indicated that the hydroxy group was located at C-14. The
ROESY correlations for H-14/H2-19 suggested that H-14 and
the 20-ethyl group occupied the same side of the two fused
rings (Figure 2). H-14 was thus assigned to be α-oriented and
the hydroxy group was β-oriented. H-14 attains a 90° dihedral
Compound 5 was obtained as a white, amorphous powder
and was assigned the molecular formula C20H26N2O5 on the
basis of its 13C NMR data and the HRESIMS ion at m/z
375.1918 [M + H]+ (calcd 375.1920), i.e., 16 mass units higher
than that of the known 20-hydroxydihydrorankinidine (5a).8
The 1H and 13C NMR data of 5 (Tables 3 and 4) were similar
to those reported for 20-hydroxydihydrorankinidine (5a).
However, signals from the CH2-14 (δH 2.13, 2.20; δC 24.2)
have been replaced by those from an oxymethine (δH 4.46; δC
1
angle with both H-3 and H-15, which is consistent with its H
NMR singlet resonance. The configuration of the C-7 spiro
center was deduced to be S as in other Gelsemium alkaloids
from the negative Cotton effect at ca. 260 nm9 in its ECD data
C
DOI: 10.1021/np5009619
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
1
Table 3. H NMR Spectroscopic Data for 5−9 and 11 (δ in ppm, J in Hz)
a
a
a
b
a
a
position
5
6
7
8
9
11
c
d
d
d
d
3
5
6
3.34,
3.57,
3.56,
2.22,
2.32,
3.62, br d (6.6)
4.12, m
3.55, br d (9.4, 2.9)
1.96, br d (15.7, 2.3)
8.04, d (4.0)
8.09, d (4.0)
8.98, d (6.9)
8.81, d (6.9)
8.57, d (6.1)
8.35, d (6.6)
2.03, dd (15.1, 6.9)
2.18, dd (15.1, 4.3)
7.38, d (8.3)
d
2.34,
9
7.47, d (7.6)
7.15, t (7.6)
7.35, t (7.6)
7.05, d (7.6)
4.46, s
7.40, d (8.3)
8.06, d (8.2)
7.29, br t (8.2)
7.62, br t (8.26)
7.91, d (8.2)
9.75, s
8.43, d (8.0)
7.53, br t (8.0)
7.78, br t (8.0)
7.85, d (8.0)
9.23, s
8.17, d (7.7)
7.38, t (7.7)
7.62, t (7.7)
7.67, d (7.7)
8.38, s
10
11
12
14
6.68, dd (8.3, 2.4)
6.67, dd (8.3, 2.4)
6.65, d (2.4)
6.62, d (2.4)
1.87, dd (15.9, 8.3)
2.33, dd (15.9, 7.9)
2.17, dd (15.1, 7.2)
2.50, ddd (15.1, 11.8, 6.7)
3.01, m
d
15
16
17
2.18, br d (5.9)
2.22,
d
2.34,
2.55, m
2.23, dd, (8.6, 4.1)
4.20, dd (11.0, 4.1)
4.35, d (11.0)
3.11 (2H), m
1.93 (2H), m
3.54 (2H), t (7.5)
2.88 (2H), t (7.5)
3.18 (2H), br s
1.98 (2H), br s
4.39 (2H), br s
3.92, dd (10.8, 5.5)
4.28, d (10.8)
1.20, d (6.5)
3.77, q (6.5)
18
19
1.12, t (7.4)
1.90 (2H), m
2.83 (2H), m
1.98 (2H) br s
3.01 (2H), br s
2.00, dq (14.2, 7.4)
2.05, dq (14.2, 7.4)
3.44, d (11.5)
8.78, s
7.36, s
21
2.50, d (14.7)
3.22, d (14.7)
8.55, s
9.42, s
8.86, s
3.50, d (11.5)
1′
2′
4.14 (2H), q (7.1)
1.24 (3H), t (7.1)
N1-OMe
4.01 (3H), s
3.96 (3H), s
3.93 (3H), s
C-11-OMe
3.83 (3H), s
3.83 (3H), s
a
b
c
d
Recorded at 600 MHz in methanol-d4. Recorded at 500 MHz in methanol-d4. Overlapped by solvent signals. Overlapped by other signals.
Table 4. 13C NMR Spectroscopic Data for 5−9 and 11 (in Methanol-d4, δ in ppm)
a
a
a
b
a
a
position
5
6
7
8
9
11
2
176.4, C
83.1, CH
59.5, CH
35.0, CH
56.9, C
177.0, C
74.2, CH
55.0, CH
25.4, CH
56.5, C
175.1, C
74.6, CH
55.1, CH
36.1, CH
56.3, C
126.2, C
133.1, C
131.1, C
3
132.7, C
132.2, C
132.3, C
5
122.5, CH
116.4, CH
114.3, C
128.2, CH
119.3, CH
126.1, C
127.3, CH
117.1, CH
123.2, C
6
7
8
132.7, C
126.9, CH
125.1, CH
129.8, CH
108.7, CH
139.3, C
66.1, CH
47.6, CH
39.3, CH
64.1, CH2
9.5, CH3
24.6, CH2
69.3, C
122.5, C
128.2, CH
109.1, CH
162.0, C
95.8, CH
141.3, C
31.8, CH2
35.0, CH
35.5, CH
68.5, CH2
17.0, CH3
69.2, CH
74.7, C
122.6, C
127.8, CH
109.1, CH
162.1, C
95.7, CH
141.1, C
30.5, CH2
24.7, CH
35.7, CH
66.8, CH2
116.6, C
122.3, C
122.4, C
9
121.6, CH
121.4, CH
128.2, CH
111.8, CH
141.5, C
123.2, CH
123.5, CH
131.5, CH
113.9, CH
143.7, C
122.6, CH
123.1, CH
130.4, CH
113.8, CH
142.6, C
10
11
12
13
14
122.0, CH
147.7, C
138.4, CH
135.3, C
120.7, CH
151.5, C
15
16
30.4, CH2
23.0, CH2
23.0, CH2
27.3, CH2
133.6, C
27.8, CH2
35.3, CH2
173.9, C
30.6, CH2
22.9, CH2
22.9, CH2
27.5, CH2
135.2, C
17
18
19
188.7, CH
118.7, C
167.1, C
20
135.3, C
21
64.1, CH2
43.3, CH2
155.9, CH
134.6, CH
124.8, CH
61.8, CH2
14.5, CH3
136.0, CH
1′
2′
N1-OMe
64.3, CH3
64.2, CH3
56.2, CH3
64.2, CH3
56.2, CH3
C-11OMe
a
b
Recorded at 150 MHz. Recorded at 100 MHz.
(Figure 3). Compound 5 was therefore defined as 14β,20α-
dihydroxydihydrorankinidine.
of 6 bore a resemblance to those of 5. The spin pattern of the
aromatic protons indicated the presence of the aromatic
methoxy group (δH 7.40, d, J = 8.3 Hz; δH 6.68, dd, J = 8.3, 2.4
Hz; δH 6.65, d, J = 2.4 Hz) at either C-10 or C-11. Moreover,
distinctive coupling patterns and chemical shifts for the
oxymethine proton (δH 4.46, s, in 5 and δH 3.77, q, in 6)
suggested that the two compounds had different oxygenation
Compound 6 was obtained as an amorphous powder and
was assigned the molecular formula C21H28N2O6 on the basis of
its 13C NMR data and the HRESIMS ion at m/z 405.2025 [M
+ H]+ (calcd 405.2026), i.e., 30 mass units higher than that of
1
compound 5. The H and 13C NMR spectra (Tables 3 and 4)
D
DOI: 10.1021/np5009619
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
sites. The coupling patterns of the oxymethine proton (δH 3.77,
q, J = 6.5 Hz) and the methyl group (δH 1.20, d, J = 6.5 Hz)
indicated that the hydroxy group was at C-19, and HMBC
correlations between H3-18/C-19, H3-18/C-20, H-19/C-20,
and H-19/C-15 verified the deduction. The relative config-
uration of 6 was assigned by a ROESY experiment (Figure 2),
in which the ROESY correlations of H-5/H-6α, H-5/H-16, and
H-16/H-17α showed their α-orientations. Compound 6 had an
ECD curve similar to that of compound 5 (Figure 3), indicating
that they have the same absolute configuration. Compound 6
was therefore defined as 11-methoxy-19,20α-dihydroxydihy-
drorankinidin. However, the configuration of C-19 remains
undetermined.
br s, δC 27.5 (CH2-19) ] in 11 (Tables 3 and 4, S11) by those
from two carboxylic/ester carbons (δC 167.1, C-19; δC 173.9,
C-18), an ethoxy group [δH 1.24 (3H), t, J = 7.1 Hz, δC 14.5,
(CH3-2′); δH 4.14 (2H), q, J = 7.1 Hz, δC 61.8, (CH2-1′)], and
two contiguous methylene groups [δH 3.54 (2H), t, J = 7.5 Hz,
δC 27.8 (CH2-16); δH 2.88 (2H), t, J = 7.5 Hz, δC 35.3 (CH2-
17)]. These observations suggested that ring E in 11 was
opened in 9/9a. 2D NMR spectroscopic analysis (Figure 2) not
only confirmed the above inference but also established the
1
remaining substructure for 9. In particular, the H−1H COSY
correlations between H-16/H-17 and H-1′/H-2′ and HMBC
correlations between H-16 and both C-18 and the carbon at δC
173.9 suggested the connection of −CH2(16)−CH2(17)− and
−O−CH2(1′)−CH3(2′) via ester carbonyl carbon C-18.
HMBC correlations between H-16 and C-14 and C-20 located
the C-16−C-17−C-18−O−C-1′−C-2′ substructure at C-15.
Moreover, HMBC correlation between H-21 and C-19
indicated the remaining carbon (δC 167.1) as C-20 and
suggested this carbon to be part of a carboxylic group on the
basis of elemental constitution analysis. Thus, the structure of 9
was elucidated as depicted, and this compound was named seco-
semperviroic acid. Compound 9 is the first known yohimbane-
type E-seco alkaloid.
Although the structure of seco-sempervirinic acid (9) is of
great interest, the presence of an ethoxy group in 9 also
emphasizes whether 9 is a natural product or an artifact. Thus,
the methanol crude extract was chemically analyzed by LC-
ESIMS, and the absence of 9 in the extract suggested that it was
an artifact.
Eleven known indole monoterpenoid alkaloids, koumine
(10),7 (19R)-hydroxydihydrokoumine,7 (19S)-hydroxydihydro-
koumine,7 (4R)-koumine N-oxide,11 (4S)-koumine N-oxide,11
N-demethoxyhumantenine,8 11-methoxy-(19R)-hydroxygelse-
legine,12 sempervirine (11),10 koumidine,13 epi-koumidine,14
and gelsemine,15 and a monoterpenoid alkaloid, venoterpine,16
were also isolated. Their structures were determined by
comparison of their experimental and reported NMR and MS
data. Among the known compounds, (4R)-koumine N-oxide,
11-methoxy-(19R)-hydroxygelselegine, and epi-koumidine were
obtained as natural products for the first time, and venoterpine
was isolated from the Loganiaceae family for the first time.
The new compounds 1−7 and 9 (8 was not tested due to its
poor solubility) were evaluated for their cytotoxic activity
against human myeloid leukemia HL-60, hepatocellular
carcinoma SMMC-7721, lung cancer A-549, breast cancer
MCF-7, colon cancer SW480, and human bronchial epithelial
BEAS-2B cell lines. Compounds 1 and 7 exhibited moderate
cytotoxicity against the five human tumor cell lines, with IC50
values in the range 4.6−9.3 μM (Table 5). Compounds 1−7
and 9 were also tested against nitric oxide production in LPS-
activated RAW264.7 macrophages, but none of the tested
compounds exhibited significant inhibitory activity (IC50 >25
μM).
The molecular formula of compound 7 was established by its
13C NMR data and the HREIMS ion at m/z 370.1542 [M]+
(calcd for C20H22N2O5, 370.1529). The 1H and 13C NMR data
(Tables 3 and 4) of 7 resembled those of 6. The distinct
difference was the replacement of signals from the C-18−C-21
subunits in 6 by those from an α,β-unsaturated formyl
functionality (CHO: δH 18.78, s, δC 188.7; C: δC 118.7; CH:
δH 7.36, s, δC 155.9). H−1H COSY, HSQC, and HMBC
(Figure 2) correlation analyses facilitated the structure
elucidation of compound 7. In particular, HMBC correlations
between H-19 and C-20 and C-21 were consistent with an α,β-
unsaturated formyl moiety, and HMBC correlations between
H-5/C-21 and H-15/C-20 suggested the connection of this
group to C-5 (via N) and C-15, respectively. The ROESY
correlations (Figure 2) and similar ECD spectra in the 195−
240 nm range indicated that 7 and 6 shared the same
configuration in their common structural units. The structure of
7 was thus established as depicted and named norhumantenine
A. This is the first report of a 19-nor-humantenine-type alkaloid.
Compound 8 was obtained as a yellow powder and was
assigned the molecular formula C19H16N2O on the basis of its
13C NMR data and the HREIMS ion at m/z 288.1280 [M]+
(calcd for 288.1263), showing an extra oxygen atom relative to
sempervirine (11),10 which was also isolated during this
investigation (Tables 3 and 4 and Supporting Information
1
S11). The major differences in the H NMR data of 8 and 11
were the shielded signals for H-5 (ΔδH −0.53), H-6 (ΔδH
−0.26), H-9 (ΔδH −0.11), H-10 (ΔδH −0.09), H2-19 (ΔδH
−0.18), and H-21 (ΔδH −0.31) and deshielded signals for H-12
(ΔδH 0.24) and H-14 (ΔδH 1.37), with the most significant
shifts occurring at H-5, H-6, H-21, and H-14. These
observations suggested that the shielding-related structural
variation occurs around N-4 and the deshielding-related one
occurs near C-12 and C-14. Considering the presence of a
conjugated π-electron system between N4 and N1 in 11, it is
likely that such a resonance system disappears in 8 and that the
additional O is located at N1 to form an N→O dative covalent
bond, the presence of which disfavors formation of a
conjugated system between N4 and N1 and is consistent with
the above NMR data. Reduction of compound 8 to 11
suggested it to be the N-1-oxide of sempervirine (11), and it
was named sempervirinoxide. This is the first N-1-oxide
alkaloid of the yohimbane type.
Compound 9 was assigned the molecular formula
C21H18N2O4 on the basis of its 13C NMR data and the
HREIMS ion at m/z 362.1262 [M]+ (calcd 362.1267). The 1H
and 13C NMR data (Tables 3 and 4) resembled those of
sempervirine (11); differences included replacement of the
signals from ring E [δH 3.18 (2H), br s, δC 30.6 (CH2-16); δH
1.98 (4H), br s, δC 22.9 (CH2-17 and CH2-18); δH 3.01 (2H),
Table 5. Cytotoxic Activity (IC50 μM) of Selected
Compounds against Cancer Cell Lines
compound HL-60 SMMC-7721 A-549 MCF-7 SW480 BEAS-2B
1
4.6
8.5
2.0
5.3
7.3
5.2
4.9
9.3
8.6
9.3
>10
>10
7.3
>10
>10
8.6
>10
>10
7
cisplatin
E
DOI: 10.1021/np5009619
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
N-4-Demethyl-21-dehydrokoumine (1): amorphous powder;
D
EXPERIMENTAL SECTION
■
[α]25 −203 (c 0.01, MeOH); UV (MeOH), λmax (log ε) 221
General Experimental Procedures. Optical rotations were
obtained with a JASCO P-1020 polarimeter equipped with a 1 dm
path length cell. UV spectra were measured with a Shimadzu UV-
2401A equipped with a 1 cm path length cell. ECD spectra were
recorded on a JASCO 810 spectrometer. IR spectra (KBr) were
determined on a Bruker Tensor-27 infrared spectrophotometer. 1D
and 2D NMR spectra were recorded on Bruker AM-400, Bruker DRX-
500, and Bruker Avance III 600 spectrometers with TMS as an internal
standard. EIMS and HREIMS spectra were recorded on an AutoSpec
Premier P776 instrument; ESIMS and HRESIMS were measured with
a Finnigan MAT 90 mass spectrometer; LC-ESIMS was done on an
ACQUITY UPLC SYNAPT G2MS (Waters Corp., USA), equipped
with a BEH C18 column (1.7 μm, 2.1 × 50 mm; Waters Corporation).
The mobile phase included (A) pure H2O and (B) MeOH. The
concentration of eluent B was changed from 20% to 95% within 10
min (linear gradient). The flow rate of the eluent was 0.4 mL min−1,
the injection volume of the extract was 1 μL, and the column oven was
set at 30 °C. Identification of compounds 2 and 3 as natural products
in G. elegans was analyzed by LC-PDA-(+)-ESIMS/MS (Supporting
Information S2/3−9). Semipreparative HPLC was carried out using a
Waters system consisting of a 600 pump and a 2996 photodiode array
detector. Silica gel (200−300 mesh, Qingdao Marine Chemical
Factory, Qingdao, China), Sephadex LH-20 gel (40−70 μm,
Amersham Pharmacia Biotech AB, Uppsala, Sweden), and MCI gel
(CHP20/P120, 75−150 μm, Mitsubishi Chemical Corporation,
Japan) were used for column chromatography.
(4.12), 261 (3.59) nm; ECD (c 4.13 × 10−1 mol/L, MeOH, 20 °C)
λmax (Δε) 255 (−18.55), 225 (45.56), 199 (46.5); IR (KBr) νmax 3442,
2923, 2852, 1630, 1384, 1075 cm−1; 1H and 13C NMR data, see Tables
1 and 2; HREIMS m/z 290.1416 [M]+ (calcd for C19H18N2O,
290.1419).
21α-Hydroxylkoumine (2) and 21β-hydroxylkoumine (3): amor-
phous powder; [α]21 −232 (c 0.1, MeOH); UV (MeOH), λmax (log
D
ε) 220.6 (4.27), 263.4 (3.69) nm; ECD (c 5.65 × 10−1 mol/L, MeOH,
20 °C) λmax (Δε) 267 (−24.01), 225 (29.85), 207 (−15.75); IR (KBr)
νmax 3422, 2927, 2866, 1636, 1447, 1085 cm−1; 1H and 13C NMR data,
see Tables 1 and 2; HRESIMS m/z 323.1749 [M + H]+ (calcd for
C20H23N2O2, 323.1760).
(19S)-Hydroxydihydrokoumine N-4-oxide (4): amorphous powder;
[α]25 −73 (c 0.1, MeOH); UV (MeOH), λmax (log ε) 222.2 (3.70),
D
262.2 (3.11) nm; ECD (c 2.45 × 10−1 mol/L, MeOH, 20 °C) λmax
(Δε) 263 (−9.29), 228 (9.92), 205 (−13.56), 195 (−3.53); IR (KBr)
νmax 3441, 2957, 2923, 1630, 1454, 1040 cm−1; 1H and 13C NMR data,
see Tables 1 and 2; HREIMS m/z 340.1792 [M]+ (calcd for
C20H24N2O3, 340.1787).
1β,20α-Dihydroxydihydrorankinidine (5): white, amorphous pow-
der, [α]26.5 −211 (c 0.1, MeOH); UV (MeOH), λmax (log ε) 201.6
D
(2.81), 251.4 (1.69) nm; ECD (c 3.21 × 10−1 mol/L, MeOH, 20 °C)
λmax (Δε) 261 (−12.08), 228 (26.52), 210 (−45.63); IR (KBr) νmax
1
3422, 2926, 1704, 1618, 65, 1042, 750 cm−1; H and 13C NMR data,
see Tables 3 and 4; HRESIMS m/z 375.1918 [M + H]+ (calcd for
C20H27N2O5, 375.1920).
Plant Material. The leaves and perennial vine stems of G. elegans
were collected from Xishuangbanna Tropical Botanical Garden
(XTBG), Chinese Academy of Science (CAS), Mengla County,
Yunnan Province, People’s Republic of China, in October 2009 and
were identified by one of the authors (Y.-K.X.). A voucher specimen
(No. GE-2009-1012) was deposited in the herbarium of XTBG.
Extraction and Isolation. The air-dried, powdered plant material
(6.0 kg) was extracted three times (5 days each) with EtOH/H2O
(90/10, v/v, 50 L) at room temperature. Removal of the solvent from
the combined extracts in vacuum afforded a crude residue (600 g).
The aqueous EtOH extract was dissolved in H2O (3.0 L) to form a
suspension, which was acidified with 10% H2SO4 to a pH ca. 3. The
acidic suspension was partitioned with EtOAc to remove the neutral
compounds, and the aqueous phase was basified with Na2CO3 to a pH
ca. 10 and extracted with CHCl3 to yield a crude alkaloid extract (105
g). The crude alkaloid mixture was applied to an MCI gel column
(MeOH/H2O from 2/8 to 5/1, v/v) to yield five major fractions (Frs.
1−5). Fr. 1 (15 g) was subjected to column chromatography (CC)
over silica gel (CHCl3/MeOH, 5/1 to 1/1, v/v) to yield four major
fractions (1a−1d). Frs. 1a−1d were purified first by Sephadex LH-20
CC (MeOH/H2O, 2/3 to 4/1, v/v) and then by semipreparative
HPLC (MeOH/H2O from 3/7 to 6/1, v/v) to yield 2/3 (10 mg), 11-
methoxy-(19R)-hydroxygelselegine (5 mg), 5 (5 mg), and (19S)-
hydroxydihydrokoumine (50 mg), respectively. Fr. 2 (10 g) was
enriched by Sephadex LH-20 CC (MeOH/H2O, from 1/1 to 4/1, v/
v) and separated on silica gel CC (CHCl3/MeOH, from 10/1 to 3/1,
v/v) to give four subfractions, purification of which by semipreparative
HPLC (MeOH/H2O, from 1/1 to 5/1, v/v) yielded (4S)-koumine N-
oxide (15 mg), (4R)-koumine N-oxide (3 mg), 7 (5 mg), and (19R)-
hydroxydihydrokoumine (8 mg). Fr. 3 (12 g) was applied to a silica gel
column eluted with petroleum ether/EtOAC (from 3/1 to 1/5, v/v)
to give seven fractions (3a−3g). Purification of Frs. 3a−3g by HPLC
(MeOH/H2OH, from 3/2 to 5/1, v/v) yielded N-demethoxyhuman-
tenine (5 mg), sempervirine (11) (100 mg), gelsemine (50 mg), 8 (10
mg), 9 (5 mg), koumidine (50 mg), and venoterpine (10 mg). Fr. 4
(17 g) was applied to a silica gel column eluted with petroleum ether/
EtOAC (from 4/1 to 1/5, v/v) to yield three major fractions (4a−4c).
Fr. 4a (6 g) was subjected to silica gel CC (CHCl3/MeOH, from 1/0
to 4/2, v/v) to yield koumine (10) (1500 mg), 4 (4 mg), and epi-
koumidine (15 mg). Fr. 4b (3 g) was enriched by MCI CC (MeOH/
H2O, from 3/2 to 5/1, v/v) and separated by semipreparative HPLC
(MeOH/H2O, 3/2 to 5/1, v/v) to afford 1 (12 mg) and 6 (15 mg).
11-Methoxy-19,20α-dihydroxydihydrorankinidin (6): white, amor-
phous powder; [α]24 −93 (c 0.1, MeOH); UV (MeOH), λmax nm
D
(log ε) 258 (3.44), 213 (4.31) nm; ECD (c 5.98 × 10−1 mol/L,
MeOH, 20 °C) λmax (Δε) 266 (−5.89), 236 (22.02), 215 (−40.93),
196 (16.22); IR (KBr) νmax 3427, 2934, 1719, 1630, 1499, 1456, 1384,
1217, 1119, 1063, 963, 804 cm−1; 1H and 13C NMR data, see Tables 3
and 4; HRESIMS m/z 405.2025 [M + H]+ (calcd for C21H29N2O6,
405.2026).
Norhumantenine A (7): white, amorphous powder; [α]21D −307 (c
0.03, MeOH); UV (MeOH), λmax (log ε) 212 (3.31), and 291 (3.02)
nm; ECD (c 3.53 × 10−1 mol/L, MeOH, 20 °C) λmax (Δε) 290
(−25.6), 237 (10.5), 217 (−28.3), 199 (14.3); IR (KBr) νmax 3448,
2921, 1716, 1631, 1583, 1499, 1384 cm−1; 1H and 13C NMR data, see
Tables 3 and 4; HREIMS m/z 370.1542 [M]+ (calcd for C20H22N2O5,
370.1529).
Sempervirinoxide (8): yellow powder; UV (MeOH) λmax nm (log
ε) 471 (3.35), 355 (4.25), 300 (4.10), 245.5 (4.33), 197.5 (4.11) nm;
IR (KBr) νmax 3440, 2924, 2853, 1640, 1469, 1403, 1347, 1037, 745
1
cm−1; H and 13C NMR data, see Tables 3 and 4; HREIMS m/z
288.1280 [M]+ (calcd for C19H16N2O, 288.1263).
seco-Semperviroic acid (9): white, amorphous powder; UV
(MeOH), λmax nm (log ε) 390 (3.52), 353 (3.55), 282 (3.66), 248
(3.92), 218 (4.12), 204 (4.12) nm; IR (KBr) νmax 3440, 2920, 1725,
1633, 1114 cm−1; 1H and 13C NMR data, see Tables 3 and 4;
HREIMS m/z 362.1262 [M]+ (calcd for C21H18N2O4, 362.1267).
Chemical Transformation of Compound 8 to Compound 11.
To a stirred suspension of compound 8 (1 mg, 0.0035 mmol) and an
equal weight of 5% Pd/C (1 mg, 0.00047 mmol Pd) in dry MeOH
(0.5 mL) was added anhydrous ammonium formate (31.5 mg, 0.5
mmol) in a single portion under a nitrogen atmosphere. The mixture
was stirred at room temperature, and the progress was monitored by
TLC. After 48 h, the catalyst and the ammonium salt were removed by
filtration through a Celite pad, using a CHCl3/MeOH (15/1) solvent
mixture (ca. 3 mL) as the eluent. The combined organic filtrate upon
evaporation under reduced pressure afforded the target compound 11.
Analysis of Synthetic and Natural Sempervirine (11).
Components 8 and 11 were indistinguishable by Rf values on silica
gel 60 F254 precoated plates. However, when the samples were
dissolved in MeOH in identical concentrations (approximately 1 mg/
mL), 8 appeared dark brown on TLC under visible light, whereas 11
1
was only pale yellow or nearly colorless. The H NMR (methanol-d4,
F
DOI: 10.1021/np5009619
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
500 MHz) spectroscopic data of synthetic 11 are consistent with those
of the natural sample previously isolated (S11-6).
(14) Kitajima, M.; Takayama, H.; Sakai, S. I. J. Chem. Soc., Perkin
Trans. 1 1991, 8, 1773−1779.
Cytotoxicity Assays.17 Inhibition of NO Production Assays. The
assay was performed according to a published method.18 Each
compound was dissolved in DMSO and further diluted in the medium
to produce different concentrations with a maximum concentration of
25 μM. The absorbance was measured at 570 nm with a 2104 Envision
multilabel plate reader (PerkinElmer Life Sciences, Inc., Boston, MA,
USA). Cytotoxicity was determined with the MTT assay. MG-132
(Sigma-Aldrich, Foster City, CA, USA) was used as the positive
control.
(15) Lin, L. Z.; Yeh, S.; Cordell, G. A.; Ni, C. Z.; Clardy, J.
Phytochemstry 1991, 30, 679−683.
(16) Ravao, T.; Richaro, B.; Zeches, M.; Massiot, T.; Le Men-Olivier,
L. Tetrahedron Lett. 1985, 26, 837−838.
(17) Ji, K. L.; Zhang, P.; Hu, H. B.; Hua, S.; Liao, S. G.; Xu, Y. K. J.
Nat. Prod. 2014, 77, 1764−1769.
(18) Ahmad, V. U.; Uddin, G. S.; Bano, S. J. Nat. Prod. 1990, 53,
1193−1197.
None of the eight (1−7, 9) compounds exhibited significant
inhibitory activities against nitric oxide production in LPS-activated
RAW264.7 macrophages.
ASSOCIATED CONTENT
* Supporting Information
■
S
1D and 2D NMR, HRESIMS/HREIMS, IR, and UV spectra of
compounds 1−9 and 11 (1D and 2D), LC-ESIMS/MS analysis
of compounds 2 and 3 in the EtOH crude extracts of G. elegans,
and the 1H NMR spectrum of sempervirine (11). The
Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/np5009619.
AUTHOR INFORMATION
Corresponding Author
*Tel: +86-691-8713169. Fax: +86-691-8713061. E-mail: xyk@
xtbg.ac.cn.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was financially supported by a grant (U1302222)
from the National Natural Science Foundation of China, 135
Program (XTBG-F02) from Xishuangbanna Tropical Botanical
Garden, Chinese Academy of Sciences, and Ethnobotanical
Investigation of Plants for Industrialization in Southwest China
(SQ2012FY4910027).
REFERENCES
■
(1) Jin, G. L.; Su, Y. P.; Liu, M.; Xu, Y.; Yang, J.; Liao, K. J.; Yu, C. X.
J. Ethnopharmacol. 2014, 152, 33−52.
(2) Zhang, Z.; Zhang, Y.; Wang, Y. H.; Zhang, Q.; Yan, X. H.; Di, Y.
T.; He, H. P.; Hao, X. J. Fitoterapia 2012, 83, 704−708.
́
(3) Liang, S.; He, C. Y.; Szabo, L. F.; Feng, Y.; Lin, X.; Wang, Y.
Tetrahedron Lett. 2013, 54, 887−890.
(4) Qu, J.; Fang, L.; Ren, X. D.; Liu, Y. B.; Yu, S. S.; Li, L.; Bao, X. Q.;
Zhang, D.; Li, Y.; Ma, S. G. J. Nat. Prod. 2013, 76, 2203−2209.
(5) Xu, Y. K.; Liao, S. G.; Na, Z.; Hu, H. B.; Li, Y.; Luo, H. R.
Fitoterapia 2012, 83, 1120−1124.
(6) Xu, Y. K.; Yang, S. P.; Liao, S. G.; Zhang, H.; Lin, L. P.; Ding, J.;
Yue, J. M. J. Nat. Prod. 2006, 69, 1347−1350.
(7) Lin, L. Z.; Cordell, G. A.; Ni, C. Z.; Clardy, J. Phytochemistry
1990, 29, 965−968.
(8) Lin, L. Z.; Cordell, G. A.; Ni, C. Z.; Clardy, G. Phytochemistry
1991, 30, 1311−1315.
(9) Yamada, Y.; Kitajima, M.; Kogure, N.; Wongseripipatana, S.;
Takayama, H. Chem.−Asian J. 2011, 6, 166−173.
(10) Gribble, G. W.; Barden, T. C.; Johnson, D. A. Tetrahedron 1988,
44, 3195−3202.
(11) Ponglux, D.; Wongseripipatana, S.; Subhadhirasakul, S.;
Takayama, H.; Yokota, M.; Ogata, K.; Phisalaphong, C.; Aimi, N.;
Sakai, S. Tetrahedron 1988, 44, 5075−5094.
(12) Takayama, H.; Kitajima, M.; Sakai, S. I. Tetrahedron 1994, 50,
11813−11820.
(13) Yeh, S.; Cordell, G. A. Phytochemistry 1987, 26, 2875−2876.
G
DOI: 10.1021/np5009619
J. Nat. Prod. XXXX, XXX, XXX−XXX