Unusual sesquiterpene glucosides from Amaranthus retroflexus

Article abstract of DOI:10.1016/j.tet.2006.07.017

Implementing the phytochemical study of the weed Amaranthus retroflexus, four new sesquiterpene glucosides were isolated from the methanolic extract of the plant. The structures of these metabolites are determined on the basis of the mass spectrometry, and 1D and 2D NMR spectroscopies (DQ-COSY, TOCSY, HSQC, HSQC-TOCSY, HMBC, and NOESY). Two compounds are characterized by a new aglycone and differed from the site of glucosylation. The other two compounds are dimeric diastereoisomers. All the glucoside sesquiterpenes were tested on the wild species Taraxacum officinale to evaluate the role of this weed in the habitat and on the seed of A. retroflexus to verify the potential autotoxic effect of the plant.

Full text of DOI:10.1016/j.tet.2006.07.017

Tetrahedron 62 (2006) 8952–8958
Unusual sesquiterpene glucosides from Amaranthus retroflexus
b
a
a
Antonio Fiorentino,a, Marina DellaGreca, Brigida D’Abrosca, Annunziata Golino,
*
Severina Pacifico,a Angelina Izzoa and Pietro Monacoa
aDipartimento di Scienze della Vita, Seconda Universitaꢀ di Napoli, via Vivaldi 43, I-81100 Caserta, Italy
bDipartimento di Chimica Organica e Biochimica, Universitaꢀ Federico II, via Cintia 4, I-80126 Napoli, Italy
Received 10 April 2006; revised 15 June 2006; accepted 6 July 2006
Available online 28 July 2006
Abstract—Implementing the phytochemical study of the weed Amaranthus retroflexus, four new sesquiterpene glucosides were isolated
from the methanolic extract of the plant. The structures of these metabolites are determined on the basis of the mass spectrometry, and 1D
and 2D NMR spectroscopies (DQ-COSY, TOCSY, HSQC, HSQC–TOCSY, HMBC, and NOESY). Two compounds are characterized by
a new aglycone and differed from the site of glucosylation. The other two compounds are dimeric diastereoisomers.
All the glucoside sesquiterpenes were tested on the wild species Taraxacum officinale to evaluate the role of this weed in the habitat and on the
seed of A. retroflexus to verify the potential autotoxic effect of the plant.
Ó 2006 Elsevier Ltd. All rights reserved.
1. Introduction
The amarantholidosides I–VI have been tested on the wild
species Taraxacum officinale to verify their impact on other
weeds and on the seed of A. retroflexus to test their potential
autotoxic effects.
One of the most useful aspects of allelopathy in manipulated
ecosystems is the role in agriculture.1 Many plants have been
studied and a number of new phytotoxic secondary metabo-
lites have been isolated and characterized.2,3 Recent investi-
gations are focused on the effects of weeds on crops,4 crops
on weeds,5 and crops on crops.6 The goal of these studies is
the use of allelochemicals as growth regulators and natural
pesticides to promote sustainable agriculture.7,8
2. Results and discussion
The EtOAc fraction of the MeOH fresh leaf extract of A. ret-
roflexus L. was chromatographed on asilicagel column (flash
chromatography) using CHCl3–EtOAc and CHCl3–MeOH
mixtures as eluents in increasing polarity. The most polar
fractions yielded seven compounds: four amarantholidols
A–D and three amarantholidosides I–III (1–3).9 Continuing
the phytochemical study of the fraction we isolated four
new glucosides 4–7, of which two of them are dimeric
(Figs. 1 and 2).
InthesearchfornewallelochemicalsfromplantsoftheMedi-
terranean area, we recently studied the weed Amaranthus
retroflexus, known as redroot pigweed, a summer annual
invasive plant, widely distributed in Italy and commonly
found in cultivated lands. From the methanolic extract of the
plant, we isolated and characterized new free and glucosyl-
ated nerolidol sesquiterpenes, named amarantholidols and
amarantholidosides, respectively.9 These compounds, when
tested on the cultivated species Lactuca sativa, showed
a moderate phytotoxic activity down to 10ꢀ9 M.
Compound 4, named amarantholidoside IV, has been iden-
tified as (3S,6E,10R)-10-b-D-glucopyranosyloxy-3,11-di-
hydroxy-3,7,11-trimethyldodeca-1,6-diene.
Its molecular formula C21H38O8 was deduced on the basis of
its ESIMS spectrum, which showed the pseudomolecular
peak at m/z 441 [M+Na]+, and the elemental analysis.
In this work we completed the study of the organic extract
and described the isolation and characterization of four new
nerolidol glucosides, named amarantholidosides IV–VII.
Two of them, in fact, showed an unusual dimeric structure.
The mass spectrum showed fragment ion at m/z 279 and 203,
both diagnostic for the presence of a hexose moiety in the
molecule. The H NMR spectrum (Table 1) showed, in the
1
Keywords: Amaranthus retroflexus; Amaranthaceae; Nerolidol glucosides;
Amarantholidosides; NMR analysis; Phytotoxic effect; Autotoxic effect;
Taraxacum officinale.
* Corresponding author. Tel.: +39 0823 274576; fax: +39 0823 274571;
e-mail: antonio.fiorentino@unina2.it
downfield region, an ABX system as three double doublets
centered at d 5.91, 5.19, and 5.02, a partially overlapped sig-
nal at d 5.21, and a double doublet at d 3.40. The doublet at
d 4.44 (J¼7.5 Hz) and two double doublets at d 3.72 and
0040–4020/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2006.07.017

A. Fiorentino et al. / Tetrahedron 62 (2006) 8952–8958
8953
14
13
OGlc
15
OH
OH
HO
HO
12
O
1
10
O
5
3
HO
HO
1
amarantholidoside I
OH
OH
OGlc
OH
4
amarantholidoside IV
OH
2
14
15
13
12
OH
amarantholidoside II
OH
1
10
O
O
5
3
OH
OGlc
OH
HO
OH
OH
3
amarantholidoside III
5
amarantholidoside V
Figure 1. Structures of amarantholidosides I–III (1–3).
OH
HO
HO
OH
3.85 were in accordance with an anomer and a hydroxy-
methylene H-6 of a sugar moiety, respectively, while the
chemical shift of six carbon signals in the 13C NMR was
in good agreement with a glucose unit. Accordingly, the
HSQC–TOCSYexperiment (Table 1) showed heterocorrela-
tion between these glycidic carbons and the H-10 and H-60
protons ranging from 3.40 to 3.24 ppm. In the upfield region
of the spectrum four singlet methyls at d 1.60, 1.28, 1.17, and
1.13 and eight protons as multiplets were present.
O
O
OH
10
O
10'
HO
O
OH
O
HO
OH
A DQ-COSY experiment showed homocorrelations among
the olefinic protons of the ABX system, the proton at
d 5.21 and a multiplet centered at d 2.03, which correlated
with a double doublet at d 1.51. Among the protons bonded
to oxygen atoms, correlations between the doublet at d 4.44
and a signal at d 3.27 as well as interactions among the dou-
ble doublet at d 3.40 and an overlapped signal centered at
d 1.45 were evident.
OH
6
amarantholidoside VI (10R,10'R) or (10S,10'S)
amarantholidoside VII (10S,10'S) or (10R,10'R)
7
Figure 2. Structures of amarantholidosides IV–VII (4–7).
on the aglycone of amarantholidol IV obtained from treat-
ment with b-glucosidase.10,11 The positive and the nega-
tive DdRꢀS values for the H-9, H-12, and H-15 protons
were found, respectively, on the right and the left sides
of the MTPA plane indicating a R configuration for
C-10 carbon.
The 13C NMR spectrum (Table 1) showed 21 signals confirm-
ing the presence of a glycosilated sesquiterpene. All of the
carbons were identified, on the basis of a DEPT experiment,
as four methyls, six methylenes, eight methines, and three
tetrasubstituted carbons. Their chemical shifts indicated the
presence ofaterminal and an internal double bonds, aglucose
moiety, and three carbinol groups (one of them secondary).
All of the carbons were correlated to the respective protons
using an HSQC experiment as reported in Table 1. In the
same table, the results of an HMBC and HSQC–TOCSY
have also been reported. These experiments were crucial
for the structure elucidation. On the basis of these correla-
tions, the carbinol carbon at d 90.3, bonded to the proton at
d 3.40, was attributed to the C-10 carbon. Its downshift value
could be explained by hypothesizing a linkage with the
glucose. The heterocorrelations, in the HMBC experiment,
between the carbon and the anomeric proton at d 4.44 and,
vice versa, between the anomeric carbon at d 106.4 and the
proton at d 3.40 confirmed the hypothesis. The glucose
moiety was confirmed by GC analysis and the coupling con-
stant value of the anomeric proton in the 1H NMR indicated
the presence of a b-glucose.
The S absolute configuration at the C-3 of compound 4 has
been deduced using a bidentate NMR chiral solvent, as
described by Kishi and co-workers.12,13 The absolute con-
figuration of acyclic alcohols was established from analysis
of the chemical shifts’ behaviors of the adjacent carbons
in (R,R)- and (S,S)-bis-a-methylbenzylpropandiammine
(BMBA-p). The positive (+0.121) and negative (ꢀ0.098)
DdRꢀS values of the amarantholidoside IV were found
on the left and right of the tertiary carbinol carbon, respec-
tively, according to an a-orientation for the hydroxyl and a
b-orientation for the methyl.
Compound 5, named amarantholidoside V, showed amolecu-
lar formula C21H38O8 in accordance with its ESI mass spec-
trum, which showed the pseudomolecular peak at m/z 441
[M+Na]+, and the elemental analysis. This data suggested
that it was an isomer of compound 4.
The 1H NMR spectrum showed the vinyl protons at d 5.91,
5.18, and 5.02, the H-6 olefin proton at d 5.21, the methine
H-10 at d 3.43, two multiplets at d 2.25 and 2.03, a double
doublet at d 1.50, three singlet methyls at d 1.24, 1.22, and
The absolute configuration to the tertiary C-3 and second-
ary C-10 carbinol carbons has been assigned using two
different methods. The R configuration at the C-10 carbon
has been determined using a modified Mosher’s method

8954
A. Fiorentino et al. / Tetrahedron 62 (2006) 8952–8958
Table 1. NMR data of amarantholidoside IV (4) in CD3OD
1H (d)
J (Hz)
DQ-COSY
13C (d)
DEPT
CH2
HMBC (H/C)
HSQC–TOCSY (H/C)
1t
1c
2
3
4
5
6
7
8
5.19 dd
5.02 dd
5.91 dd
—
1.51 dd
2.03 m
5.21 ov
—
2.30 m
2.25 m
1.45 m
3.40 dd
—
1.17 s
1.13 s
1.60 s
1.28 s
4.44 d
3.27 ov
3.27 ov
3.37 ov
3.37 ov
3.72 dd
3.85 dd
17.4, 1.5
10.8, 1.5
17.4, 10.8
—
9.3, 7.5
—
—
1c, 2
1t, 2
1c, 1t
—
5
4, 6
5
—
112.0
2
2, 3
1, 3, 4, 15
—
2, 3, 5, 6, 15
3, 4, 6, 7
4, 5, 8, 14
—
6, 7, 9, 10, 14
6, 7, 9, 10, 14
7, 8
glu1, 8, 9, 11, 12, 13
—
10, 11, 13
10, 11, 12
6, 7, 8
1, 2, 3, 4, 5
10, glu2, glu3
glu1, glu3, glu4
glu1, glu2, glu4
glu5, glu6
glu6
1, 2
1, 2
1, 2
—
4, 5, 6, 14
4, 5, 6, 14
4, 5, 6
—
8, 9, 10
8, 9, 10
8, 9, 10
8, 9, 10
—
—
—
—
146.3
74.8
43.5
CH
C
CH2
CH2
CH
C
23.7
126.1
136.0
37.0
—
—
9
CH2
9
8, 10
9
—
—
—
6
—
9
10
11
12
13
14
15
glu1
glu2
glu3
glu4
glu5
glu6
—
10.5, 1.5
—
—
—
—
—
7.5
—
—
—
—
11.7, 2.1
11.7, 5.7
30.7
90.3
73.8
23.7
26.4
16.1
27.6
106.4
76.0
77.9
71.4
78.4
62.6
CH2
CH
C
CH3
CH3
CH3
CH3
CH
CH
CH
CH
CH
CH2
—
glu2
glu1, glu2, glu3, glu4, glu6
glu1, glu2, glu3, glu4, glu6
glu1, glu2, glu3, glu4, glu6
glu2, glu3, glu4, glu6
glu2, glu3, glu4, glu6
glu3, glu4, glu6
glu3, glu4, glu6
glu1, glu3
glu2, glu4
glu3, glu5
glu4, glu6
glu5, glu6
glu5, glu6
glu4, glu5
glu4, glu5
d¼doublet; dd¼double doublet; m¼multiplet; ov¼overlapped; s¼singlet.
1.21, and a doublet methyl at d 1.60. In the same spectrum,
the sugar signals were present as a doublet at d 4.49 (glu1),
two double doublets at d 3.81 and 3.64 (glu6), and a double
doublet at d 3.15 (glu2) besides other signals obscured by
the solvent signal.
H-glu1 proton suggested the presence of a b-anomer. The
2D NMR experiments confirmed the same aglycone of com-
pound 4. In the HMBC experiment, the signal at d 81.8,
assigned to the C-11 carbon, showed correlations with the
anomeric proton at d 4.49, the H-12 and H-13 methyls and
with the H-10 double doublet, which itself correlates, in
the HSQC experiment, to the carbon at d 78.1. This latter
carbon correlated with the H-9 protons at d 1.35 and 2.20
and with the H-12 and H-13 methyls. This data led to the hy-
pothesize of the presence of a linkage between the C-glu1 of
The main differences observed in the 13C NMR spectrum,
when compared to amarantholidoside IV, are for C-10 and
C-11 carbons (Table 2). The sugar moiety was in good accor-
dance with the presence of a glucose and the J value of the
Table 2. NMR data of amarantholidoside V (5) in CD3OD
1H (d)
J (Hz)
DQ-COSY
13C (d)
DEPT
CH2
HMBC (H/C)
HSQC–TOCSY (H/C)
1t
1c
2
3
4
5
6
7
8
5.18 dd
5.02 dd
5.91 dd
—
1.50 dd
2.03 m
5.21 ov
—
2.30 m
2.25 m
2.20 m
1.35 m
3.43 dd
—
1.21 s
1.22 s
1.60 d
1.24 s
4.49 d
3.15 dd
3.36 ov
3.37 ov
3.26 dt
3.64 dd
3.81 dd
17.4, 1.5
10.5, 1.5
17.4, 10.5
—
1c, 2
1t, 2
1c, 1t
—
5
4, 6
5
—
9
9
8, 10
112.0
2, 3
2, 3
1, 3, 4, 15
—
2, 3, 5, 6, 15
4, 6, 7
3, 5, 7, 8, 14
—
1, 2
1, 2
1, 2
—
146.3
73.8
43.5
CH
C
CH2
CH2
CH
C
9.3, 7.5
4, 5, 6, 14
4, 5, 6, 14
4, 5, 6
—
8, 9, 10
8, 9, 10
8, 9, 10
8, 9, 10
8, 9, 10
—
23.7
126.0
135.9
37.8
—
—
CH2
6, 7, 9, 10, 14
6, 7, 9, 10, 14
7, 8
9
—
30.7
CH2
10
11
12
13
14
15
glu1
glu2
glu3
glu4
glu5
glu6
10.5, 1.5
—
—
—
0.9
—
8.1
9.0, 7.5
9
—
78.1
81.8
21.3
23.8
16.0
27.6
98.6
75.1
77.7
71.6
77.7
62.6
CH
C
8, 9, 11, 12, 13
glu1, 10, 12, 13
10, 11, 13
10, 11, 12
6, 7, 8
1, 2, 3, 4, 5
11, glu2, glu3
glu1, glu3, glu4
glu1, glu2, glu4
glu5, glu6
glu6
glu4, glu5
glu4, glu5
CH3
CH3
CH3
CH3
CH
CH
CH
CH
CH
CH2
6
glu2
glu1, glu2, glu3, glu4, glu6
glu1, glu2, glu3, glu4, glu6
glu1, glu2, glu3, glu4, glu6
glu2, glu3, glu4, glu6
glu2, glu3, glu4, glu6
glu3, glu4, glu6
glu1, glu3
glu2, glu4
glu3, glu5
glu4, glu6
glu5, glu6
glu5, glu6
9.0, 8.4
11.7, 5.1
11.7, 2.1
glu3, glu4, glu6
d¼doublet; dd¼double doublet; m¼multiplet; ov¼overlapped; s¼singlet.

A. Fiorentino et al. / Tetrahedron 62 (2006) 8952–8958
8955
the glucose and the C-11 of the aglycone. The NOEs ob-
served, in an NOESY experiment, among the H-glu1 and
H-12 protons confirmed this hypothesis. The enzymatic hy-
drolysis with b-glucosidase afforded the same aglycone 4a
of the previous glucoside.
sesquiterpene bonded to a glucose unit at C-5. The presence
of 21 carbon signals in the 13C NMR and ESI mass spectrum
indicated a high degree of symmetry in the molecule. The
mass spectrum also showed fragments at m/z 438
[MꢀC21H35O7+Na]+ and 404 [MꢀC21H35O8ꢀH2O+Na]+
due to the cleavage of the ether bridge, and at m/z 488
[Mꢀ2ꢁC6H11O5]+ and 472 [MꢀC6H11O5ꢀC6H11O6]+ due
to the loss of both the sugar unities. In accordance with the
hypothesis of a dimeric structure, the C-10 value (d 89.8)
was downshifted with respect to the known amarantholidol
II.9 This value suggested ethereal bridge among the C-10
carbons.
Compounds 6 and 7 have been identified as two diastereo-
mers and named amarantholidosides VI and VII, respec-
tively.
Compounds 6 and 7 had a molecular formula C42H70O15, as
resulted by the elemental analysis and the ESI mass spec-
trum, which showed the pseudomolecular peak at m/z 837
[M+Na]+.
Amarantholidoside VII NMR data overlapped with those of
the previously described compound, except slight differ-
ences for C-6–C-14 carbon values (Table 3). Unfortunately,
these differences could be due to a different configuration for
the C-10 and C-100 carbons. The isolated quantities of these
compounds did not allow further spectroscopic investiga-
tions and we were not able to define the configurations at
C-10 and C-100 carbons.
In the 1H NMR spectrum of amarantholidoside VI (Table 3),
the ABX system of H-1 cis, H-1 trans, and H-2 protons and
the H-6 and H-12 olefinic protons was evident. In the region
of the protons geminal to oxygen, the signals of the anomeric
and of the hydroxymethyl glu6 of a sugar moiety were pres-
ent, and the H-10 triplet at d 4.22 and further four protons
ranging from 3.10 to 3.40 ppm were also present. In the
upfield region of the spectrum three methyls at d 1.72,
1.67, and 1.28, a methylene as triplet at d 2.06, and a double
doublet centered at d 1.96 were identifiable. This data
together with the information from the 13C NMR spectrum
led to the hypothesize of the presence of a D1,6,11 nerolidol
The amarantholidosides I–VI have been tested on the wild
species T. officinale to define their phytotoxic role in the hab-
itat and on the seed of A. retroflexus to evaluate the potential
autotoxic activity; the results are reported in Figures 3 and 4,
respectively.
Table 3. NMR data of amarantholidosides VI (6) and VII (7) in CD3OD
1H (d)
J (Hz)
DQ-COSY
13C (d)
DEPT
CH2
HMBC (H/C)
HSQC–TOCSY
(H/C)
6
7
6
7
6
7
1t/10t
1c/10c
2/20
3/30
4/40
5.23 dd
5.01 dd
5.96 dd
—
5.23 dd
5.01 dd
5.96 dd
—
17.1, 1.5
9.0, 1.5
11.1, 1.2
—
17.4, 1.2
10.8, 1.2
17.4, 10.8
—
1c/10c, 2/20,
1t/10t, 2/20
1c, 1t
—
5/50
112.0
112.0
2/20, 3/30
1/10, 2/20
1/10, 2/20
1/10, 2/20
—
2/20, 3/30
146.4
74.0
48.1
146.4
74.0
48.1
CH
C
CH2
1/10, 3/30
—
1.96 dd
1.96 dd
14.1, 7.8
14.7, 8.1
2/20, 3/30, 5/50,
6/60, 15/150
4/40, 5/50, 6/60
1.65 ov
1.65 ov
2/20, 3/30, 5/50,
6/60, 15/150
5/50
6/60
4.90 ov
5.11 dd
4.90 ov
5.13 dd
—
9.6, 1.2
—
9.0, 0.9
4/40, 6/60
5/50
71.3
126.6
71.3
126.8
CH
CH
glu1/10, 3/30, 4/40, 6/60
4/40, 5/50, 7/70,
8/80, 14/140
4/40, 5/50, 6/60
4/40, 5/50, 6/60
7/70
8/80
—
2.06 t
—
2.06 t
—
7.5
—
7.5
—
9/90
141.1
36.7
141.8
36.5
C
CH2
—
—
8/80, 9/90, 10/100
6/60, 7/70, 9/90,
10/100, 14/140
7/70, 8/80, 10/100, 11/110
7/70, 8/80, 10/100, 11/110
8/80, 9/90, 11/110,
12/120, 13/130
—
9/90
1.57 dd
1.52 dd
4.22 t
1.61 dd
1.58 dd
4.22 t
6.6, 1.8
6.6, 1.2
6.6
6.6, 1.7
6.6, 1.2
6.3
8/80, 10/100
8/80, 10/100
9/90
30.1
89.8
30.1
89.6
CH2
CH
8/80, 9/90, 10/100
8/80, 9/90, 10/100
10/100
11/110
12/120
13/130
14/140
15/150
glu1/10
—
—
—
—
—
—
—
7.8
—
—
—
—
—
7.5
—
—
145.5
114.2
17.1
16.7
28.6
145.9
114.1
17.1
16.7
28.6
C
—
—
—
—
4.94 s
1.72 s
1.67 s
1.28 s
4.23 d
4.94 s
1.73 s
1.68 s
1.28 s
4.24 d
CH2
CH3
CH3
CH3
CH
10/100, 11/110, 13/130
10/100, 11/110, 12/120
6/60, 7/70, 8/80
2/20, 3/30, 4/40
5/50, 10/100, glu2/20
—
6/60
—
—
glu2/20
100.0
100.1
glu1, glu2, glu3,
glu4, glu6
glu1, glu2, glu3,
glu4, glu6
glu1, glu2, glu3,
glu4, glu6
glu2, glu3,
glu4, glu6
glu2, glu3,
glu4, glu6
glu3, glu4, glu6
glu3, glu4, glu6
glu2/20
glu3/30
glu4/40
glu5/50
glu6/60
3.15 dd
3.30 m
3.29 ov
3.26 m
3.15 dd
3.30 m
3.29 ov
3.26 m
8.9, 7.8
—
8.9, 7.8
—
glu1/10, glu3/30
glu2/20, glu4/40
glu3/30, glu5/50
glu4/40, glu6/60
78.1
75.1
71.9
78.1
62.9
78.1
75.1
71.9
78.1
62.9
CH
CH
CH
CH
CH2
5/50, glu1/10, glu3/30,
glu4/40
glu1/10, glu2/20,
glu4/40
—
—
glu5/50, glu6/60
—
—
glu6/60
3.85 dd
3.66 dd
3.86 dd
3.66 dd
11.7, 2.1
11.7, 5.7
12.3, 2.1
12.3, 6.3
glu5/50, glu6/60
glu5/50, glu6/60
glu4/40, glu5/50
glu4/40, glu5/50
d¼doublet; dd¼double doublet; m¼multiplet; ov¼overlapped; s¼singlet; t¼triplet.

8956
A. Fiorentino et al. / Tetrahedron 62 (2006) 8952–8958
germination
germination
0
0
b
b
a
a
b
b
a
b
b
b
a
a
a
b
b
b
b
a
a
a
a
a
a
a
a
a
-25
-50
-25
-50
a
a
b b
a
a
a
a
a
a
a
1
2
3
4
5
6
1
2
3
4
5
6
root elongation
root elongation
0
-25
-50
0
-25
-50
b
b b
b
b
b
b
b
a
bb
b
a
b
bb
b
a
a
b
a
a
b
a
b
a
b
aab
a
a
a
a
a
b
1
2
3
4
5
6
1
2
3
4
5
6
shoot elongation
shoot elongation
a
b
b
b
b
a
a
b
b
b
b
b
b
b
b
b
b
0
-25
-50
0
-25
-50
b
b
b
b
b
b
b
b
b
b
1
2
3
10-6
4
5
6
1
10-4
2
10-5
3
4
5
6
10-4
10-5
10-7
10-8
10-9
10-6
10-7
10-8
10-9
Figure 3. Bioactivity of compounds 1–6 on the germination, root elonga-
tion, and shoot elongation of Taraxacum officinale. Values are presented
as percentage differences from control and they are significantly different
with P>0.05 for Student’s t-test: (a) P<0.01; (b) 0.01<P<0.05. A
positive percentage represents stimulation while negative values represent
inhibitions.
Figure 4. Bioactivity of compounds 1–6 on thegermination, root elongation,
and shoot elongation of Amaranthus retroflexus. Values are presented as per-
centage differences from control and they are significantly different with
P>0.05 for Student’s t-test: (a) P<0.01; (b) 0.01<P<0.05. A positive per-
centage represents stimulation while negative values represent inhibitions.
3. Experimental
On T. officinale, the greatest effects were registered on the
root elongation. Compound 1 was more active: the chemical
inhibited the root development of about 50% at the highest
concentration tested. The effects on the germination were
modest and with exception of compound 1 it seemed
independent from the concentration. Furthermore, the auto-
toxicity, evident in both germination and root elongation
(Fig. 4), was moderate. The results showed slow stimulating
and/or inhibiting effects on shoot elongation of the test
species.
3.1. General procedures
NMR spectra were recorded at 300 MHz for 1H and 75 MHz
for 13C on a Varian 300 spectrometer Fourier transform
NMR in CD3OD at 25 ꢂC. Proton-detected heteronuclear
correlations were measured using a gradient heteronuclear
single-quantum coherence (HSQC), optimized for JHC
140 Hz, a gradient heteronuclear multiple bond coherence
(HMBC), optimized for nJHC¼8 Hz, and a gradient heteronu-
clear single-quantum coherence–total correlation spectros-
copy (HSQC–TOCSY), optimized for nJHC¼8 Hz and for a
mixing time¼0.09. Optical rotations were measured on a
Perkin–Elmer 141 in MeOH solution. Electrospray mass
spectra were recorded using a Waters Z-Q mass spectrometer
equipped with an electrospray ionization (ESI) probe
1
¼
To the best of our knowledge, this is the first report that
describes the isolation of dimeric nerolidol derivatives
from natural sources. The presence of secondary and tertiary
chiral alcohols on aglicone moiety led us to use two different
methods to assign the absolute configuration.

A. Fiorentino et al. / Tetrahedron 62 (2006) 8952–8958
8957
operating in positive or negative ion mode. The scan range
was 80–2000 m/z.
on RP-18 column eluting with H2O–MeOH–MeCN (5:3:2)
to have pure glucoside 4 (30.1 mg) and a mixture, which
was chromatographed by FCC on SiO2, eluting with
CHCl3–MeOH–0.1% TFA–H2O (100:20:3:1). The eluate
collected from 50 to 90 ml was purified on RP HPLC, using
the Polar-RP-80A column and eluting with H2O–MeOH–
MeCN (15:1:4) to have pure 5 (3.0 mg).
The HPLC apparatus consisted of a pump (Shimadzu LC-
10AD), a refractive index detector (Shimadzu RID-10A),
and a Shimadzu Chromatopac C-R6A recorder. Preparative
HPLC was performed using NH2 (Luna 10 mm, 250ꢁ10 mm
i.d., Phenomenex) column. Analytical HPLC was performed
using RP-8 (Luna 5 mm, 250ꢁ4.6 mm i.d., Phenomenex) or
Polar-RP-80A (Synergi 4 mm, 250ꢁ4.6 mm i.d., Pheno-
menex) columns. Analytical TLC was performed on Merck
Kieselgel 60 F254, RP-18 F254, or RP-8 F254 plates with
0.2 mm layer thickness. Spots were visualized by UV light
or by spraying with H2SO4–AcOH–H2O (1:20:4). The plates
were then heated for 5 min at 110 ꢂC. Preparative TLC was
performed on Merck Kieselgel 60 F254 plates, with 0.5 or
1 mm film thickness. Flash column chromatography (FCC)
was performed on Merck Kieselgel 60 (230–400 mesh) at
medium pressure. Column chromatography (CC) was per-
formed on Merck Kieselgel 60 (70–240 mesh). Filtrations
on solid phase extraction (SPE) cartridge were performed
on Waters NH2 or C8 Sep-Pak Cartridge.
3.3. Characterization of the metabolites 4–7
Amarantholidoside IV [(3S,6E,10R)-10-b-D-glucopyrano-
syloxy-3,11-dihydroxy-3,7,11-trimethyldodeca-1,6-diene] 4.
Colorless oil; IR (CHCl3) nmax cmꢀ1: 3400, 2932, 1577; 1H
NMR (300 MHz, CD3OD): Table 1; 13C NMR (75 MHz,
CD3OD): Table 1; ESIMS m/z 441 [M+Na]+, 279 [Mꢀglc+
Na]+, 203 [glc+Na]+; [a]2D5 ꢀ12.2 (c 0.16, MeOH). Anal.
Calcdfor C21H38O8: C, 60.27;H, 9.15. Found:C, 60.31;H, 9.17.
Amarantholidoside V [(3S,6E,10R)-11-b-D-glucopyrano-
syloxy-3,10-dihydroxy-3,7,11-trimethyldodeca-1,6-diene] 5.
Colorless oil; IR (CHCl3) nmax cmꢀ1: 3389, 2921; 1H
NMR (300 MHz, CD3OD): Table 2; 13C NMR (75 MHz,
CD3OD): Table 2; ESIMS m/z 441 [M+Na]+, 279
[Mꢀglc+Na]+; [a]D25 ꢀ14.6 (c 0.12, MeOH). Anal. Calcd
for C21H38O8: C, 60.27; H, 9.15. Found: C, 60.30; H, 9.18.
3.2. Plant material, extraction, and isolation of the
metabolites
Plants of A. retroflexus L. were collected in May 2003, in the
vegetative state, in Recale, near Caserta (Italy), and identi-
fied by Dr. Assunta Esposito of the Second University of
Naples. A voucher specimen (CE118) has been deposited
at the Herbarium of the Dipartimento di Scienze della Vita
of Second University of Naples.
Amarantholidoside VI [(3S,30S,5S,50S,6E,60E,10x,100x)-
10,100-oxybis(5-b-D-glucopyranosyloxy-3-hydroxy-3,7,11-
trimethyldodeca-1,6,11-triene] 6. Colorless oil; IR (CHCl3)
nmax cmꢀ1: 3397, 2932; 1H NMR (300 MHz, CD3OD): Table
3; 13C NMR (75 MHz, CD3OD): Table 3; ESIMS m/z 837
[M+Na]+, 488 [Mꢀ2ꢁC6H11O5]+, 472 [MꢀC6H11O5ꢀ
C6H11O6]+, 438 [MꢀC21H35O7+Na]+, 404 [MꢀC21H35O8ꢀ
H2O+Na]+, 354 [C16H27O7+Na]+; [a]2D5 ꢀ26.4 (c 0.26,
MeOH). Anal. Calcd for C42H70O15: C, 60.30; H, 8.66.
Found: C, 60.32; H, 8.69.
Fresh leaves of A. retroflexus (42 kg) were extracted with
MeOH–H2O (1:9) for 48 h at 20 ꢂC and then in MeOH for
5 days. The methanolic extract was dissolved in water and
then extracted with EtOAc. The organic fraction was dried
with Na2SO4 and concentrated under vacuum, yielding
18.4 g of crude residue.
Amarantholidoside VII [(3S,30S,5S,50S,6E,60E,10x,100x)-
10,100-oxybis(5-b-D-glucopyranosyloxy-3-hydroxy-3,7,11-
trimethyldodeca-1,6,11-triene] 7. Colorless oil; IR (CHCl3)
nmax cmꢀ1: 3394, 2931; 1H NMR (300 MHz, CD3OD):
Table 3; 13C NMR (75 MHz, CD3OD): Table 3; ESIMS
m/z 837 [M+Na]+, 488 [Mꢀ2ꢁC6H11O5]+, 472
[MꢀC6H11O5ꢀC6H11O6]+, 438 [MꢀC21H35O7+Na]+, 404
[MꢀC21H35O8ꢀH2O+Na]+, 354 [C16H27O7+Na]+; [a]2D5
ꢀ43.3 (c 0.14, MeOH). Anal. Calcd for C42H70O15:
C, 60.30; H, 8.66. Found: C, 60.28; H, 8.29.
The EtOAc fraction was chromatographed on silica gel, with
CHCl3–EtOAc and CHCl3–MeOH solutions, to give four
fractions A–D.
Fractions A and B, eluted with EtOAC–CHCl3 (1:9), give the
free nerolidols previously described. Fraction C eluted with
MeOH–CHCl3 (1:9) was chromatographed on Sephadex
LH-20 eluting with hexane–CHCl3–MeOH (2:1:1) to obtain
three fractions: the first, rechromatographed on RP-18 col-
umn, eluting with MeOH–MeCN–H2O (2:1:2), furnished
pure 3 (9.0 mg); the second fraction was chromatographed
by NH2 HPLC eluting with MeCN–H2O (9:1), which was
purified by RP-8 HPLC, eluting with H2O–MeOH–MeCN
(13:3:4) to give pure 1 (3.0 mg) and 2 (2.0 mg); the third frac-
tion was chromatographed on RP-18 column eluting with
H2O–MeOH–MeCN (2:1:2) to have a fraction that was first
filtered on NH2 Sep-Pak with H2O–MeCN (2:23) and then
purified by analytical RP-8 HPLC, eluting with H2O–
MeOH–MeCN (13:3:4) to give pure glucosides 6 (6.0 mg)
and 7 (2.0 mg). Fraction D eluted with MeOH–CHCl3
(1:1) was chromatographed on Sephadex LH-20 eluting
with hexane–CHCl3–MeOH (1:1:1), collecting fractions of
25 ml volumes. The second fraction was chromatographed
3.3.1. Enzymatic hydrolysis of amarantholidol IV. Toa so-
lution of pure amarantholidoside IV (10 mg) in acetate buffer
(0.5 M, pH 5.0, 5 ml), 40 mg of b-glucosidase (Sigma) was
added. After 24 h at 37 ꢂC in stirring, the mixture was ex-
tracted with EtOAc (5 mlꢁ2), dried over Na2SO4, and evapo-
rated in vacuo. The crude extract was chromatographed on
SiO2 [hexane–CHCl3–MeOH (3:6:1)] to have pure aglycone
(3S,6E,10R)-3,10,11-trihydroxy-3,7,11-trimethyldodeca-1,6-
diene (4a): 1H NMR spectral data (300 MHz, CD3OD):
d 5.91 (1H, dd, J¼17.1, 10.5 Hz, H-2), 5.18 (1H, dd, J¼
17.1, 1.5 Hz, H-1 cis), 5.19 (1H, dd, J¼9.3, 1.2 Hz, H-6),
5.02 (1H, dd, J¼10.5, 1.5 Hz, H-1 trans), 3.22 (1H, dd, J¼
10.8, 1.8 Hz, H-10), 1.61 (3H, s, H-14), 1.24 (3H, s, H-15),
1.15 (3H, s, H-12), 1.12 (3H, s, H-13); 13C NMR (75 MHz,
CD3OD) d 145.0 (CH, C-2), 136.0 (C, C-7), 125.8 (CH, C-6),

8958
A. Fiorentino et al. / Tetrahedron 62 (2006) 8952–8958
112.0 (CH2, C-1), 78.9 (CH, C-10), 73.8 (C, C-11), 74.8 (C,
C-3), 43.4 (CH2, C-4), 37.8 (CH2, C-8), 30.7 (CH2, C-9), 27.5
(CH3, C-15), 26.2 (CH3, C-13), 23.9 (CH2, C-5), 23.7 (CH2,
C-12), 16.5 (CH3, C-14); EIMS m/z 256 [M]+, 238
[MꢀH2O]+, 223 [MꢀCH3ꢀH2O]+; [a]2D5 ꢀ14.6 (c 0.12,
MeOH); [a]2D5 +7.8 (c 0.10, MeOH). Anal. Calcd for
C15H28O3: C, 70.27; H, 11.01. Found: C, 70.61; H, 10.96.
3.4. Phytotoxicity test
Seeds of T. officinale and A. retroflexus, collected during
2005, were obtained from Herbiseed (Twyford, UK).
All undersized or damaged seeds were discarded and the
assay seeds were selected for uniformity.
3.3.2. Preparation of (S)- and (R)-MTPA esters of 4a. (R)-
(ꢀ)-MTPA chloride (5 ml, 26 mmol) was added to a solution
of pure compound (1.5 mg) in dry pyridine (50 ml). After 6 h
under magnetic stirring at room temperature, EtOAc (5 ml)
and H2O (5 ml) were added to the reaction mixture. The
organic layer, separated by centrifugation at 4000 rpm for
10 min, gave a crude extract, which was purified by prepar-
ative TLC eluting with hexane–CHCl3–MeOH (5:4:1).
The test solution (10ꢀ4 M) was prepared using 2-[N-mor-
pholino]ethanesulfonic acid (MES; 10 mM, pH 6) and the
rest (10ꢀ5–10ꢀ9 M) was obtained by dilution. Parallel con-
trols were performed. After adding 10 seeds and 1.0 ml of
test solutions, Petri dishes were sealed with ParafilmÒ to en-
sure closed-system models. Seeds were placed in a growth
chamber KBW Binder 240 at 27 ꢂC in the dark. Germination
percentage was determined daily for 7 days (no more germi-
nation occurred after this time).
1
The (S)-MTPA ester of 4a had the H NMR spectral data
(300 MHz, CD3OD): d 5.90 (1H, dd, J¼17.3, 10.5 Hz,
H-2), 5.12 (1H, dd, J¼17.3, 1.8 Hz, H-1 trans), 4.99 (H-10,
detected by DQ-COSY spectrum), 4.92 (H-1 cis, detected
by DQ-COSY spectrum), 1.59 (3H, s, H-14), 1.51 (H-9,
detected by DQ-COSY spectrum), 1.24 (3H, s, H-15), 1.19
(3H, s, H-12), 1.15 (3H, s, H-13). The (R)-MTPA ester of 4a
After growth, the plants were frozen at ꢀ20 ꢂC to avoid
subsequent growth until the measurement process. Data
are reported as percentage differences from control. Thus,
zero represents the control, positive values represent the
stimulation of the parameter studied, and negative values
represent inhibition.
1
had the H NMR spectral data (300 MHz, CD3OD): d 5.89
(1H, dd, J¼17.1, 10.5 Hz, H-2), 5.18 (1H, dd, J¼17.1,
1.2 Hz, H-1trans), 5.05(H-1cis, detectedbyDQ-COSYspec-
trum), 5.00 (H-10, detected by DQ-COSY spectrum), 1.76
(3H, s, H-14), 1.58 (H-9, detected by DQ-COSY spectrum),
1.24 (3H, s, H-15), 1.11 (3H, s, H-12), 1.07 (3H, s, H-13).
Statistical treatment. The statistical significance of differ-
ences between groups was determined by a Student’s
t-test, calculating mean values for every parameter (germi-
nation average, shoot, and root elongation) and their popula-
tion variance within a Petri dish. The level of significance
was set at P<0.05.
3.3.3. 13C NMR of amarantholidol IV in the chiral solvent
R,R- and S,S-BMBA-p. Chiral NMR solvent (R,R)-BMBA-
p [or (S,S)-BMBA-p] was prepared with small modifications
of method reported by Kobayashi et al.13 1,3-Diiodopropane
(10.3 mmol, 1 equiv) was added dropwise over 7 min to
(R)-a-phenetylamine [or (S)-a-phenetylamine] (41.3 mmol,
4 equiv) at 130 ꢂC. After stirring for 30 min, the mixture
was cooled to 80 ꢂC and then poured into aqueous 50%
NaOH solution (300 ml). The resulting free amines were
extracted with EtOAc (300 ml) and the organic layer was
washed with brine, dried over anhydrous K2CO3, filtered,
and evaporated to give the crude oil. The pure BMBA-p
was obtained by distillation as reported by Hulst et al.14
References and notes
1. Duke, S. O.; Romagni, J. G.; Dayan, F. E. Crop Prot. 2000, 19,
583–589.
2. D’Abrosca, B.; Della Greca, M.; Fiorentino, A.; Monaco, P.;
Natale, A.; Oriano, P.; Zarrelli, A. Phytochemistry 2000, 66,
2681–2688.
ꢀ
3. Macias, F. A.; Velasco, R. F.; Castellano, D.; Galindo, J. C. G.
J. Agric. Food Chem. 2005, 53, 3530–3539.
4. Jefferson, L. V.; Pennacchio, M. J. Arid Environ. 2003, 55,
275–285.
5. Turk, M. A.; Tawaha, A. M. Crop Prot. 2003, 22, 673–677.
6. Kato, T.; Saito, N.; Kashimura, K.; Shinohara, M.;
Kurahashi, T.; Taniguchi, K. J. Agric. Food Chem. 2003, 50,
6307–6312.
7. Walter, M. W. Nat. Prod. Rep. 2002, 19, 278–291.
8. Vyvyan, J. R. Tetrahedron 2002, 58, 1631–1646.
9. D’Abrosca, B.; De Maria, P.; Della Greca, M.; Fiorentino,
A.; Golino, A.; Izzo, A.; Monaco, P. Tetrahedron 2006, 62,
640–646.
10. Yokosuka, A.; Mimaki, Y.; Sakagami, H.; Sashida, Y. J. Nat.
Prod. 2002, 65, 283–289.
11. Ohtani, I.; Kusumi, T.; Kashman, Y.; Kakisawa, H. J. Am.
Chem. Soc. 1991, 113, 4092–4096.
12. Kobayashi, Y.; Hayashi, N.; Kishi, Y. Org. Lett. 2002, 4,
411–414.
13. Kobayashi, Y.; Hayashi, N.; Kishi, Y. Tetrahedron Lett. 2003,
44, 7489–7491.
1
(R,R)-BMBA-p (colorless oil) had H NMR spectral data
(300 MHz, CDCl3): d 7.40–7.20 (5H, overlapped), 3.71
(2H, q, J¼6.6 Hz), 2.54 (2H, dt, J¼6.6, 5.1 Hz), 2.47 (2H,
dt, J¼6.6, 5.1 Hz), 1.62 (2H, m), 1.34 (6H d, J¼6.6); 13C
NMR (75 MHz, CDCl3) d 145.3 (C), 128.3 (CH), 126.6 (CH),
126.4 (CH), 58.3 (CH), 46.3 (CH2), 29.9 (CH2), 24.1 (CH3);
[a]2D0 +66.2 (c 0.25, CHCl3). (S,S)-BMBA-p (colorless oil)
had 1H NMR spectral data (300 MHz, CDCl3): d 7.40–7.20
(5H, overlapped), 3.71 (2H, q, J¼6.6 Hz), 2.54 (2H, dt,
J¼6.6, 5.1 Hz), 2.47 (2H, dt, J¼6.6, 5.1 Hz), 1.62 (2H, m),
1.34 (6H, d, J¼6.6); 13C NMR (75 MHz, CDCl3) d 145.3
(C), 128.3 (CH), 126.6 (CH), 126.4 (CH), 58.3 (CH), 46.3
(CH2), 29.9 (CH2), 24.1 (CH3); [a]2D0 ꢀ66.2 (c 0.25, CHCl3).
A Varian Mercury 300 spectrometer (75 MHz) was used to
collect the 13C NMR data of amarantholidoside IV (4,
10 mg) in the chiral solvent BMBA-p (350 ml)–CD3OD
(300 ml), with readout of NMR spectra being adjusted to
0.001 ppm/point (sw¼23980.8, fn¼524288).
14. Hulst, R.; de Vries, N. K.; Feringa, B. L. Tetrahedron:
Asymmetry 1994, 5, 699–708.