Quercetin and kaempferol 3-O-[α-L-rhamnopyranosyl-(→2)-α- L-arabinopyranoside]-7-O-α-L-rhamnopyranosides from Anthyllis hermanniae: Structure determination and conformational studies

Article abstract of DOI:10.1021/np200444n

The study reports the isolation and structural identification of two new flavonol triglycosides from the methanolic extract of Anthyllis hermanniae, exhibiting the same glycosylation pattern: quercetin 3-O-[α-L- rhamnopyranosyl-(→2)-α-l-arabinopyranoside]

Full text of DOI:10.1021/np200444n

ARTICLE
pubs.acs.org/jnp
Quercetin and Kaempferol 3-O-[R-L-Rhamnopyranosyl-(1f2)-R-L-
arabinopyranoside]-7-O-R-L-rhamnopyranosides from Anthyllis
hermanniae: Structure Determination and Conformational Studies#
Maria Halabalaki,† Aurꢀelie Urbain,†,‡ Aristea Paschali,† Sofia Mitakou,† Franc-ois Tillequin,§ and
Alexios-Leandros Skaltsounis†,*
†Laboratory of Pharmacognosy and Natural Products Chemistry, Department of Pharmacy, University of Athens,
Panepistimioupolis-Zografou, Athens, 15771, Greece
‡Pharmacognosie et Substances Naturelles Bioactives, UMR 7200ꢀLaboratoire d'Innovation Thꢀerapeutique, Facultꢀe de Pharmacie,
Universitꢀe de Strasbourg, 74 Route du Rhin, 67401 Illkirch, France
§Laboratoire de Pharmacognosie de l’Universitꢀe Renꢀe Descartes, UMR CNRS 8638, Facultꢀe de Pharmacie, 4 Avenue de l’Observatoire,
75006 Paris, France
S Supporting Information
b
ABSTRACT:
The study reports the isolation and structural identification of two new flavonol triglycosides from the methanolic extract of Anthyllis
hermanniae, exhibiting the same glycosylation pattern: quercetin 3-O-[R-L-rhamnopyranosyl-(1f2)-R-L-arabinopyranoside]-7-O-
R-L-rhamnopyranoside (1) and kaempferol 3-O-[R-L-rhamnopyranosyl-(1f2)-R-L-arabinopyranoside]-7-O-R-L-rhamnopyrano-
side (2). A conformational study related to the central arabinoside moiety was carried out including the analysis of the contribution
of NOE effects and acetylation to the elucidation of the 2-O-linked arabinoside configuration of the anomeric carbon. We also report
the total synthesis of a model compound, quercetin 3-O-R-L-rhamnopyranosyl-(1f2)-R-L-arabinopyranoside (3), which verifies
the structures of the isolated compounds.
he genus Anthyllis belongs to the Leguminosae family and
comprises more than 170 herbaceous or shrubby species dis-
investigations of A. sericea led to similar flavonoid derivatives.7
Most of these flavonoid glycosides exhibit complex glycosylation
patterns at C-3, consisting of a galactopyranose linked to one or
two sugar moieties, mainly glucose and rhamnose.8 As far as A.
hermanniae is concerned, there is only one phytochemical study
reporting the presence of isoflavones and chalcones from the
nonpolar fraction.9
T
tributed in Europe, the Middle East, and North Africa. Anthyllis
hermanniae L., which is often called yellow kidney vetch due to its
yellow flowers, is a medium-sized shrub distributed in Asian
countries and in Europe, especially in the northeastern Mediter-
ranean area.1
Despite the large number of Anthyllis species, only a few have
been investigated from a chemical point of view, leading mainly
to the identification of flavonoids and their corresponding glyco-
sides. Quercetin, kaempferol, isorhamnetin, and rhamnocitrin
have been reported as the main aglycones in A. vulneraria2 and A.
onobrychioides, where glycosylation occurs mostly at C-3 with
galactose, glucose, and arabinose.3ꢀ5 From A. onobrychioides, di-
and triglycosides were also identified as 3-O-galactoside deriva-
tives glucosylated at C-30 and C-40,6 while the phytochemical
The current study reports the isolation and structural
identification of two new flavonol triglycosides from the
methanolic extract of A. hermanniae. Both compounds
exhibit the same glycosylation pattern and have been identi-
fied as quercetin 3-O-[R-L-rhamnopyranosyl-(1f2)-R-L-
arabinopyranoside]-7-O-R-L-rhamnopyranoside (1) and kaempferol
Received: May 28, 2011
Published: August 23, 2011
r
Copyright
2011 American Chemical Society and
1939
dx.doi.org/10.1021/np200444n J. Nat. Prod. 2011, 74, 1939–1945
American Society of Pharmacognosy
|

Journal of Natural Products
ARTICLE
3-O-[R-L-rhamnopyranosyl-(1f2)-R-L-arabinopyranoside]-7-O-
R-L-rhamnopyranoside (2). Additional to the phytochemical
analysis, a conformational study concerning the central arabino-
side moiety has been performed. The conformational equilibrium
between the two possible chairlike 4C1 and 1C4 conformations of
the arabinopyranoside unit due to steric hindrance is discussed.
This study also comprises the analysis of NOE effect contributions
as well as the impact of acetylation on the elucidation of the
configuration of the anomeric center of 2-O-linked arabinopyrano-
sides. In addition, quercetin 3-O-[R-L-rhamnopyranosyl-(1f2)-R-L-
arabinopyranoside (3), a reported natural compound, was synthesized.
More specifically, the 1H and 13C NMR deshielded values
for anomeric (δH 5.58, δC 101.6) and especially the H-2 and
C-2 signals (δH 4.13, δC 77.7) indicated that the arabinosyl
moiety was linked at C-2 to the second rhamnosyl unit. The
HMBC correlation between the anomeric proton of the rham-
nosyl (δH 5.12) and C-2 of the arabinosyl moiety (δC 77.7)
unambiguously established the (1f2) linkage between the two
glycosyl moieties.
However, for complete structural elucidation, the configura-
tion of the anomeric center of the arabinopyranosyl moiety
needs to be determined. In general, this is a complicated task for
2-substituted arabinosides, mainly due to the known equili-
4
1
brium between the two C1 and C4 chair conformations of
pyranosides.12 This equilibrium with significant alteration of
3JH1,H2 values has been observed in various 2-O-glycosylated
pentopyranosides of bulky aglycones.13,14 For compound 1 the
3JH1,H2 (5.4 Hz) value favors an R-configuration of the anomeric
carbon. This configuration was also described for calabricoside, a
flavonol triglycoside with a similar diglycoside moiety isolated
from Putoria calabrica.15
In order to obtain further structural information, the acetyla-
tion of 1 was carried out, yielding the peracetylated deriv-
ative 1ac. The NMR data of 1ac provided further support to
the assignment of the rhamnose (1f2) arabinose linkage, since
the COSY spectrum of 1ac showed a cross-peak between the
anomeric proton of arabinose (δH 5.58) and the upfield proton
H-2 at δH 4.13.16 Moreover, the presence of L-arabinopyranosyl
and L-rhamnopyranosyl substituents was confirmed by acidic
hydrolysis of 1 followed by TLC analysis and measurement of the
optical rotation values of the sugars obtained from the water-
soluble fraction and comparison with reference standards.
Finally, alkaline hydrolysis of 1 was also carried out in order
to selectively remove the rhamnoside at C-7, leading to quercetin
3-O-R-L-rhamnopyranosyl-(1f2)-R-L-arabinopyranoside (3).
Compound 3 was previously isolated from Brassica nigra; how-
ever its description is incomplete since only 13C NMR data are
given.17 Thus, compound 1 was identified as quercetin 3-O-[R-L-
rhamnopyranosyl-(1f2)-R-L-arabinopyranoside]-7-O-R-L-rham-
nopyranoside. We propose the trivial name hermannioside A for
new compound 1.
Compound 2 was obtained as an amorphous, yellow powder.
Its HR-MS data indicated a molecular formula of C32H37O18 on
the basis of its pseudomolecular ion at m/z 709.1992 (err. 0.96
ppm). The use of UV shift reagents combined with 1H NMR data
indicated that 2 was a kaempferol derivative. As for compound 1,
H-6 and H-8 gave two meta-coupled doublets at δH 6.48 and 6.77
(J = 2.1 Hz). The B-ring protons appeared as two doublets of
doublets at δH 8.10 (dd, J = 8.8, 2.0 Hz, H-20, H-60) and 6.94 (dd,
J = 8.8, 2.0 Hz H-20, H-60), characteristic of an AA0BB0 spin
system. 1H and 13C NMR data in methanol-d4 revealed the same
glycosidic signals as for 1, i.e., three anomeric signals at δH 5.59
(d, J = 1.4 Hz), 5.58 (d, J = 5.1 Hz), and 5.12 (d, J = 1.1 Hz). The
methyl groups of the rhamnopyranosyl moieties were observed
as doublets at δ 1.28 and 1.10. Compound 2 was then assigned as
a kaempferol analogue of 1. HMBC correlations confirmed the
link between the anomeric proton (δH 5.59) of the rhamnosyl
moiety and C-7 (δC 163.6) of kaempferol. In the same way, there
was a correlation between the anomeric proton of the arabinosyl
moiety (δH 5.58) and C-3 (δC 135.9), revealing glycosylation at
this position. Additionally, both C-2 and H-2 of the arabinosyl
unit were deshielded (δC 77.1, δH 4.13), and the carbon
correlated with the rhamnosyl anomeric proton at δH 5.12,
’ RESULTS AND DISCUSSION
Compound 1 was obtained as an amorphous, yellow powder.
The molecular formula was deduced from the high-resolution
MS data to be C32H37O19 ([M ꢀ H]ꢀ at m/z 725.1943, err. 1.13
ppm). Its UV spectrum was characteristic of a flavonol, while the
addition of usual shift reagents as well as the 1H NMR spectrum
in methanol-d4 indicated that 1 was a quercetin derivative. The
characteristic ABX spin system of the B-ring was denoted by
resonances at δH 6.92 (d, J = 8.4 Hz, H-50), 7.62 (dd, J = 8.4, 2.1
Hz, H-60), and 7.68 (d, J = 2.1 Hz, H-20). The 1H NMR spectrum
also exhibited two meta-coupled doublets at δH 6.48 and 6.76 (J =
2.1 Hz) corresponding to H-6 and H-8, respectively. In addition
to the quercetin moiety, three sugar residues were revealed by the
corresponding anomeric protons at δH 5.58 (d, 5.4 Hz), 5.57
(br s), and 5.12 (d, J = 1.3 Hz). The two latter signals correspond to
two R-rhamnopyranosyl moieties on the basis of their coupling
constants and the presence of characteristic doublets of rham-
nosyl methyl groups observed at δH 1.28 and 1.08, respectively.
The attribution of the methyl groups to the respective rhamno-
pyranosides was confirmed by COSY and HMBC correlations.
One of these rhamnose moieties was linked to the flavonol
nucleus at C-7, as proved by the HMBC correlation between the
anomeric proton at δH 5.57 and C-7 (δC 163.9). Furthermore,
the chemical shifts of C-2 and C-3 (δC 159.7 and 135.9,
respectively) indicated that another glycosylation occurred at
C-3 of the quercetin nucleus. Obvious was the HMBC correla-
tion of the most deshielded anomeric proton at δH 5.58 with C-3.
The 1D and 2D (HSQC, COSY, and HMBC) spectra showed
this sugar residue to be an arabinopyranosyl moiety.10,11
The second rhamnosyl unit was found to be linked at the
glycosidic part of the molecule and not directly to the aglycone.
1940
dx.doi.org/10.1021/np200444n |J. Nat. Prod. 2011, 74, 1939–1945

Journal of Natural Products
ARTICLE
confirming the (1f2) interglycosidic linkage between the two
sugar moieties. Acetylation of 2 yielded the peracetylated
derivative 2ac, while acidic hydrolysis led to kaempferol, L-arabino-
pyranose, and L-rhamnopyranose. Compound 2 was thus identi-
fied as kaempferol 3-O-[R-L-rhamnopyranosyl-(1f2)-R-L-
arabinopyranoside]-7-O-R-L-rhamnopyranoside, a new compound
for which we propose the trivial name hermannioside B.
Recently, a compound with similar NMR data to 2 was iso-
lated from Matthiola longipetala and identified as kaempferol
3-O-(200-R-L-rhamnopyranosyl)-β-L-arabinopyranoside-7-O-R-
L-rhamnopyranoside.18 On the basis of the low coupling constant
for the anomeric proton of the arabinoside (δ 5.46, J = 4.6 Hz,
DMSO-d6), the authors claimed the β-configuration of the
arabinosyl unit. The same authors also identified a 40-O-gluco-
pyranoside derivative for which the arabinosyl moiety had similar
values (δH 5.48, J = 4.8 Hz, DMSO-d6). Measuring the coupl-
ing constants of the anomeric protons of the arabinosyl units
of hermannioside A (δH 5.48, J = 4.9 Hz) and hermannioside B
(δ 5.47, J = 4.8 Hz) in DMSO-d6, it was found that the observed
values were similar to those of the compounds from M. long-
ipetala. On the basis of these data it became obvious that the
complete elucidation of the configuration of the anomeric center
of arabinose is problematic. Indeed for the compounds isolated
from M. longipetala an R-configuration of arabinoside with an
increased population of 1C4 conformers could not be excluded.
The structural elucidation of flavonoid glycosides with two,
three, or even more sugar moieties is not easy to achieve due to
the complexity of NMR spectra in the sugar region. The difficulty
in complete assignment concerns both the position and config-
uration of the interglycosidic linkage.19 Previous studies have
reported that 2-O-glycosyl-L-arabinopyranosides exhibit unusual
Figure 1. Possible NOESY correlations between H-1 and H-3 of the
arabinopyranosyl moiety.
Furthermore, it is important to note that almost all the natural
3-O-arabinosylated flavonoids are reported with an R-L-arabino-
syl moiety. There are a few publications referring to flavonoids O-
linked to β-L-arabinosyl moieties, and on the basis of the above
discussion many of them might be reappraised. Indeed, some
authors have confused L-arabinose with D-arabinose, and so they
assigned a β-configuration when they observe high 3JH1,H2 values
(around 7 Hz). In some other cases the β-configuration is
assigned on the basis of quite low to intermediate J values
(<∼5 Hz). Most of the time, authors evade this problem and
do not mention the configuration of the arabinosyl unit, or even
1
the H NMR data for the sugar moiety.26 The most in-depth
study is the one of Abdel-Shafeek et al.,14 reporting the isolation
and characterization of a quercetin derivative linked at C-3 to a
2-O-substituted arabinosyl moiety. They observed a JH1,H2 value
of 1.8 Hz, which could be characteristic of a β-configuration;
however, an R-configuration of arabinose with a strong predo-
3
glycosylation shifts.20,21 Significant alteration of JH1,H2 values
has been observed in various 2-O-glycosylated arabinopyrano-
sides of bulky aglycones. These atypical values could be explained
by a specific conformational equilibrium between the two
4
1
possible chairlike conformations C1 and C4. Indeed, when a
bulky substituent is linked to C-2 of an arabinopyranosyl unit, the
equilibrium between 4C1 and 1C4 conformers is altered in favor of
the latter, since the 1- and 2-substituents adopt the stereoche-
mically and thermodynamically more stable anti orientation.12
Durette and Horton analyzed this phenomenon using low-
temperature NMR to freeze this equilibrium so that the inter-
conversion is slow enough and the anomeric signals due to both
1C4 and 4C1 conformers could be distinguished.22 They proposed
that the coupling constant JH1,H2 observed at room temperature
results from the weighted average between the coupling value
Je and Ja, when the proton is exclusively equatorial (theoretical
J = 1 Hz) or axial (theoretical J = 8 Hz), respectively, and could
be calculated according to the following equation: Jobs = NeJe +
NaJa, where Ne and Na represent the mole fractions of both
conformers.
1
minance of C4 conformers population (>90%) could not be
excluded.
Surprisingly, no one has attempted to clarify these issues for
flavonoid glycosides using NOE effects. NOESY could be a use-
ful spectrometric experiment to solve the problem of anome-
ric proton configuration of 2-O-linked arabinose. Indeed, for
R-arabinopyranosides, a strong NOE correlation should be ob-
served only between the anomeric proton (H-1) and H-3 in the
4C1 conformer. In contrast, this correlation is weak or impossible
1
in the C4 conformer of R-arabinose as well as in the case of
β-arabinose whatever its conformation (Figure 1). In the case of
compounds 1 and 2, the NOESY spectrum exhibited correlation
between the anomeric protons of the arabinosyl moiety (δH 5.58
and 5.48) and protons 3 (δH 3.91 and 3.67), respectively, an
observation in agreement with an R-configuration for the
anomeric center of arabinoside, in both compounds.
3
The JH1,H2 value of the anomeric proton of methyl-R-L-
arabinoside decreased from 6.9 to 4.9 Hz12 after a 2-O-glucosyl
linkage was added. Interestingly, when the steric hindrance is
particularly strong, due to a bulky 1-O-aglycone, the observed
3JH1,H2 value could reach low values similar to those observed
with a β-configuration (around 2ꢀ3 Hz).12,23,24 Therefore, it
appears that in the case of 2-O-linked arabinopyranosides an
intermediate or low value of the coupling constant does not
permit the accurate assignment of the configuration of the
anomeric proton.
In order to unequivocally confirm the structures of hermanniosides
A and B compared to the compounds isolated from Matthiola
longipetala,18 but mainly to investigate further the config-
uration of the arabinosyl moiety, the synthesis of a model
compound was performed. Quercetin 3-O-R-L-rhamnopyranosyl-
(1f2)-R-L-arabinopyranoside (3), previously isolated from Brassica
nigra17 and also obtained by alkaline hydrolysis of 1, was selected to be
synthesized in a regio- and stereoselective manner.
1941
dx.doi.org/10.1021/np200444n |J. Nat. Prod. 2011, 74, 1939–1945

Journal of Natural Products
ARTICLE
Figure 2. Scheme for the synthesis of 3 and 3ac.
(a) Hg(CN)2, benzeneꢀnitromethane; 68%; (b) acetone, H2SO4, reflux; 53%; (c) pyridine, Ac2O; 92%; (d) H2ꢀPdꢀC; 100%; (e) HBr/OHAc,
HOAcꢀCHCl3; 0 °C, 43%; (f) PhCH2N+(Et)3Brꢀ, KOHꢀCHCl3 anh., 30%; (g) MeONaꢀMeOH; 80%.
The starting material chosen for the synthesis of 3 was benzyl
3,4-O-isopropylidene-β-L-arabinopyranoside (4), readily avail-
able from L-arabinose in two steps by a classical procedure.27,28
Condensation with 2,3,4-tri-O-acetyl-β-L-rhamnopyranosyl bro-
mide (5)29 in the presence of mercuric cyanide in benzene-
nitromethane gave benzyl 2,3,4-tri-O-acetyl-R-L-rhamnopyranosyl-
(lf2)-3,4-O-isopropylidene-β-L-arabinopyranoside (6) (Figure 2).30
Removal of the isopropylidene protecting group of 6 by treatment
with acetone and H2SO4 afforded 7, which after acetylation gave
compound 8. Debenzylation of 8 through hydrogenolysis, fol-
lowed by treatment with HBr in HOAc, furnished 2,3,4-tri-O-
acetyl-(R-L-rhamnopyranosyl-(1f2)-3,4-di-O-acetyl-β-L-arabi-
nopyranosyl bromide (9),13,31,32 which was immediately used for
the glycosidation of 7,40-di-O-benzylquercetin (10),33 which
under stereospecific phase-transfer-catalyzed conditions led
exclusively to the R-arabinoside (11).34 The resulting 7,40-di-
O-benzylquercetin 3-O-[2,3,4-tri-O-acetyl-R-L-rhamnopyranosyl-
(1f2)-3,4-di-O-acetyl-R-L-arabinopyranoside] (11) was hydro-
genolyzed to remove the benzyl group to afford the corresponding
glycoside 12. A portion of 12 via Zemplen deacetylation yielded
the targeted quercetin 3-O-R-L-rhamnopyranosyl-(1f2)-
R-L-arabinopyranoside (3), identical to the natural compound
previously isolated from Brassica nigra,17 while another por-
tion was acetylated to yield peracetylated quercetin 3-O-R-L-
rhamnopyranosyl-(1f2)-R-L-arabinopyranoside (3ac) (Figure 2).
3 at δH 5.43 with a coupling constant of 5.0 Hz (DMSO-d6) and
at δH5.59 with a coupling constant of 7.0 Hz (CDCl3) in 3ac.
According to these data, the proposed structures for herman-
niosides A and B were confirmed since the JH1,H2 for the
arabinose unit observed in DMSO-d6 of 3 (5.0 Hz) was similar
to that of 1 (J = 4.9 Hz) and 2 (J = 4.8 Hz). These observations
confirm the initial hypothesis of enhancement in the population
l
of the C4 conformers due to steric hindrance. Regarding the
acetylated derivative 3ac, the coupling constant of the anomeric
proton was significantly increased (J = 7.0 Hz, CDCl3), a
phenomenon that is also observed in the acetylated products
of hermanniosides 1ac (J = 6.6 Hz, CDCl3) and 2ac (J = 6.8 Hz,
CDCl3), suggesting that O-acetylation stabilizes the 4C1 conformation.35
Overall, it is evident that the observed coupling constants of
4ꢀ5 Hz for the anomeric proton of an arabinopyranosyl moiety
linked to a rhamnopyranosyl unit at C-2 correspond to the
R-configuration of the anomeric carbon of arabinoside and is the
result of the alternation of the equilibrium between the 4C1 and
lC4 conformations of arabinose. Consequently, on the basis of
these data the proposed β-configuration for the anomeric carbon
of arabinose in the flavonol glycosides isolated from Matthiola
longipetala18 should be revised to R. Finally, taking into con-
sideration the observed coupling constants as well as the
equation of Durette and Horton we conclude that the population
of the 4C1 and 1C4 conformers is respectively 67% and 33% for
hermannioside A (1) and 60% and 40% for hermannioside B (2).
Likewise, for the corresponding acetylated derivatives 1ac and
1
The H NMR study of 3 and its peracetylated derivative 3ac
revealed the anomeric proton signal of the arabinosyl moiety in
1942
dx.doi.org/10.1021/np200444n |J. Nat. Prod. 2011, 74, 1939–1945

Journal of Natural Products
ARTICLE
1
Table 1. H and 13C NMR Assignment for 1 and 2 (600 MHz, in methanol-d4) and 1ac and 2ac (600 MHz, in CDCl3)
hermannioside A (1)
δH (J in Hz)
1ac
hermannioside B (2)
δH (J in Hz)
2ac
position
δC
δH (J in Hz)
δC
δH (J in Hz)
2
159.7
135.9
180.1
164.4
100.9
163.9
95.9
159.7
135.9
180.1
163.4
100.6
163.6
95.6
3
4
5
6
6.48 (d, 2.1)
6.74 (d, 2.5)
7.04 (d, 2.5)
6.48 (d, 2.1)
6.77 (d, 2.1)
6.74 (d, 2.5)
7.07 (d, 2.5)
7
8
6.76 (d, 2.1)
9
158.4
107.9
123.4
117.7
146.6
150.5
116.8
123.8
158.0
107.9
123.1
132.1
116.5
150.1
116.5
132.1
10
10
20
30
40
50
60
7.68 (d, 2.1)
7.94 (d, 2.1)
8.10 (dd, 8.8, 2.0)
6.94 (dd, 8.8, 2.0)
8.10 (dd, 8.8, 2.0)
7.18
6.92 (d, 8.4)
7.28 (d, 8.5)
6.94 (dd, 8.8, 2.0)
8.10 (dd, 8.8, 2.0)
7.18
8.10
7.62 (dd, 8.4, 2.1)
6.74 (dd, 8.5, 2.1)
3-O-L-Ara
1
5.58 (d, 5.4)
4.13 (dd, 6.6, 5.4)
3.84 (m)
101.6
77.7
72.9
69.0
66.0
5.60 (d, 6.6)
5.58 (d, 5.1)
4.13 (dd, 6.9, 5.2)
3.83 (dd, 9.5, 3.5)
3.81 (m)
100.8
77.1
70.2
68.8
65.4
5.60 (d, 6.8)
2
4.01 (dd, 9.9, 6.6)
5.04 (dd, 9.9, 3.3)
5.15 (m)
4.02 (dd, 9.2, 6.8)
5.05 (dd, 9.2, 3.4)
5.15 (m)
3
4
3.82 (m)
5a
3.81 (m)
3.70 (dd, 12.8 2.4)
3.47 (dd, 12.8 1.7)
3.82 (m)
3.69 (dd, 12.6, 2.7)
3.46 (dd, 12.8 1.7)
5b
3.42 (dd, 13.5, 4.7)
3.40 (m)
3-O-L-Rha
1
5.12 (d, 1.3)
3.94 (dd, 3.2, 1.3)
3.73 (dd, 9.5, 3.2)
3.37 (t, 9.5)
102.7
73.4.
72.7
74.1
70.6
18.0
4.96 (br s)
5.12 (d, 1.1)
3.91 (dd, 3.2, 1.6)
3.73 (dd, 9.6, 3.3)
3.38 (t, 9.6)
102.1
72.1
72.3
73.9
69.2
17.7
4.96 (br s)
2
5.16 (m)
5.16 (m)
3
5.42 (dd, 10.0, 3.3)
5.00 (t, 10.0)
4.30 (dd, 10.0, 6.2)
0.96 (d, 6.2)
5.42 (dd, 10.2, 3.4)
5.00 (t, 10.2)
4.30 (m)
4
5
3.82 (m)
3.82 (m)
6
1.05 (d, 6.2)
1.10 (d, 6.2)
0.94 (d, 6.3)
7-O-L-Rha
1
2
3
4
5
6
5.57 (br s)
100.4
73.0
71.9
74.5
71.7
18.5
5.54 (br s)
5.59 (d, 1.4)
99.6
71.7
71.4
73.6
71.3
17.9
5.53 (br s)
4.04 (dd, 3.2, 1.7)
3.86 (m)
5.40 (dd, 3.3, 1.6)
5.35 (dd, 9.9, 3.3)
5.12 (t, 9.9)
4.04 (dd, 3.3, 1.7)
3.85 (dd, 9.5, 3.3)
3.50 (t, 9.5)
5.39 (dd, 3.4, 1.7)
5.35 (dd, 10.2, 3.4)
5.11 (m)
3.50 (t, 9.5)
3.61 (m)
3.88 (dd, 9.9, 6.2)
1.27 (d, 6.2)
3.61 (dd, 9.4, 6.2)
1.28 (d, 6.2)
3.87 (m)
1.26 (d, 6.2)
1.27 (d, 6.2)
4
2ac, these populations are altered by stabilization of the C1
conformation and are calculated to be 80% and 20% for the first
and 82% and 18% for the latter.
spectrometer. The 2D experiments (COSY, HSQC, HMBC, and NOESY)
were performed using standard Bruker microprograms. The residual 1Hand
13C signals of methanol-d4 (δH 3.31, δC 49.5, respectively), DMSO-d6 (δH
2.50, δC 39.4, respectively), and chloroform-d (δH 7.26, δC, respectively)
were used as internal standards. FABMS (thioglycerol) was performed on a
VG Micromass 70-70 spectrometer, LR-ESIMS spectra were obtained in
negative mode using a Thermo-Finnigan MSQ mass spectrometer, and
HR-ESIMS spectra were obtained on a hybrid Thermo Scientific LTQ-
Orbitrap Discovery mass spectrometer (negative mode). Analytical TLC
was performed on Merck precoated silica gel 60 F254 plates. Spots were
visualized using UV light and vanillinꢀH2SO4 reagent. Column flash
chromatography was carried out using Si gel, Merck, 0.04ꢀ0.06 mm. A
Thermo Finnigan HPLC system (Thermo Finnigan, San Jose, CA)
connected to a Spectral System UV2000 PDA detector was employed for
the isolation procedure. The reagents used for the synthesis of the
model compound were reagent grade and were used without further
purification, while solvents were distilled from the appropriate drying
In conclusion, caution should be exercised in determin-
ing the identity of pentopyranosides and especially in the
elucidation of their configurations. Nevertheless, the
NOESY correlations and JH1,H2 coupling constants of the
acetylated derivatives of the compounds under investiga-
tion could be useful tools for their complete structure
elucidation.
’ EXPERIMENTAL SECTION
General Experimental Procedures. Optical rotations were
obtained using a Perkin-Elmer 141 polarimeter. UV spectra were
obtained using spectroscopic grade MeOH on a Shimadzu-160A spectro-
photometer. NMR spectra were recorded on a Bruker Avance 600 MHz
1943
dx.doi.org/10.1021/np200444n |J. Nat. Prod. 2011, 74, 1939–1945

Journal of Natural Products
ARTICLE
agents before use when needed. All reactions involving air- or
moisture-sensitive reagents or intermediates were performed under a
nitrogen or argon atmosphere.
Plant Material. Whole plants of Anthyllis hermanniae L. were
collected in May 2007 in the vicinity of Athens, Greece, on Ymittos hill
(250 m alt.), and identified by Dr. Eleftherios Kalpoutzakis. A voucher
specimen (No. NEK 001) was deposited at the Herbarium of the
Division of Pharmacognosy, University of Athens.
Extraction and Isolation. Air-dried powdered plant (1.1 kg) was
extracted at room temperature successively with CH2Cl2 (3 ꢁ 2 L) and
MeOH (3 ꢁ 2 L) for 2 days each time. The MeOH extract was
concentrated to give a residue (58 g), which was applied to a silica gel
column and eluted with a CH2Cl2/MeOH gradient to yield 18 fractions.
Fractions 15ꢀ17 were combined and concentrated (7.8 g). An aliquot
(1 g) was dissolved in 50% aqueous MeCN (20 mg/mL), passed
through nylon acrodisc filters (0.45 μm, Waters), and subjected to
semipreparative HPLC-PDA using a reversed-phase C18, Supelcosil
SPLC-18 column (250 ꢁ 10 mm, 5 μm, Supelco, Sigma-Aldrich).
The gradient conditions were as followed: eluent A: H2O, B: MeCN;
gradient: 4% to 12% B in 30 min, then to 20% in 15 min, followed by a 30
min gradient to 50% B, to finish with 70% B within a total analysis time of
80 min; flow rate: 3 mL/min. Two peaks were collected to afford 1 (34
mg, tR 45.0 min) and 2 (32 mg, tR 47.4 min).
(4.8), 270.2 sh (4.3), 353.6 (4.3) nm; + NaOMe, 265.2, 400.5; + NaOAc,
258.0, 358.4; + NaOAc/H3BO3, 260.0, 374.0;+AlCl3, 274.5, 439.5; + AlCl3/
HCl, 271.5, 402.0; 1H and 13C NMR Table 1; HR-ESIMS m/z 725.1943
[M ꢀ H]ꢀ (calcd for C32H38O19, 725.1935).
Peracetyl quercetin 3-O-[R-L-rhamnopyranosyl-(1f2)-R-L-ara-
binopyranoside]-7-O-R-L-rhamnopyranoside (1ac): amorphous, brown
solid; [R]20D ꢀ23.8 (c 0.1, MeOH); 1H NMR Table 1; HR-ESIMS m/z
1187.3069 [M ꢀ H]ꢀ (calcd for C54H60O30, 1187.3085).
Kaempferol 3-O-[R-L-rhamnopyranosyl-(1f2)-R-L-arabinopyranoside]-
7-O-R-L-rhamnopyranoside, hermannioside B (2): amorphous, yellow pow-
der; [R]20D ꢀ86.9 (c 0.6, MeOH); UV (MeOH) λmax (log ε) 225.0 (4.5),
265.0 sh (4.3), 348.5 (4.2) nm; + NaOMe, 274.0, 387.0; + NaOAc, 269.5 sh,
348.5; + NaOAc/H3BO3, 270.0 sh, 359.5 sh; + AlCl3, 233.0 sh, 273.5, 348.5; +
1
AlCl3/HCl, 230.0 sh, 273.0; H and 13C NMR Table 1; HR-ESIMS m/z
709.1992 [M ꢀ H]ꢀ (calcd for C32H37O18, 709.1985).
Peracetyl kaempferol 3-O-[R-L-rhamnopyranosyl-(1f2)-R-L-ara-
binopyranoside]-7-O-R-L-rhamnopyranoside (2ac): amorphous, brown
1
solid; [R]20 ꢀ20.3 (c 0.1, MeOH); H NMR Table 1; HR-ESIMS
D
m/z 1129.3045 [M ꢀ H]ꢀ (calcd for C52H57O28, 1129.3031).
Quercetin 3-O-[R-L-rhamnopyranosyl-(1f2)-R-L-arabinopyrano-
side (3): [R]20D ꢀ58.8 (c 0.9, MeOH); 1H NMR (600 MHz, DMSO-d6)
7.55 (1H, dd, J = 8.2, 2.2 Hz, H-60Q), 7.56 (1H, d, J = 2.2 Hz, H-20Q),
6.78 (1H, d, J = 8.2 Hz, H-50Q), 6.28 (1H, d, J = 2.0 Hz, H-8Q), 6.05
(1H, d, J = 2.0 Hz, H-6Q), 5.43 (1H, d, J = 5.0 Hz, H-1), 4.95 (1H, br s,
H-10), 4.10 (1H, dd, J = 8.0, 5.3 Hz, H-2), 3.70 (1H, m, H-20), 3.68 (1H,
m, H-50), 3.67 (1H, m, H-3), 3.66 (1H, m, H-4), 3.64 (1H, m, H-5a),
3.42 (1H, dd, J = 9.2, 3.1 Hz, H-30), 3.42 (1H, dd, J = 13.5, 4.0 Hz, H-5β),
3.21 (1H, t, J = 9.2, Hz, H-40), 0.95 (3H, d, J = 6.0 Hz, CH3-60); 13C
NMR (50 MHz, DMSO-d6) 176.6 (C-4Q), 168.5 (C-7Q), 160.8 (C-
5Q), 156.4 (C-2Q), 155.2 (C-9Q), 149.8 (C-40Q), 145.7 (C-30Q),
132.9 (C-3Q), 121.1 (C-60Q), 120.2 (C-10Q), 116.0 (C-50Q), 115.6
(C-20Q), 102.1 (C-10Q), 99.9 (C-1), 99.6 (C-10), 99.4 (C-6Q), 93.9
(C-8Q), 74.5(C-2), 71.9 (C-40), 71.0 (C-20), 70.6 (C-30), 69.9 (C-3),
68.4 (C-50), 66.6 (C-4), 64.1 (C-5), 17.4 (C-6); FABMS (m/z) 581
[M + H]+; anal. calcd for C26H29O15.
Acetylation and Hydrolysis of 1 and 2. Acetylation. Com-
pounds 1 (5 mg) and 2 (5 mg) were treated with Ac2O (0.5 mL) and
pyridine (0.5 mL), at room temperature, overnight and gave the
peracetylated derivative 1ac (91%) and the peracetylated derivative
2ac, respectively (1H NMR, Table 1).
Acidic Hydrolysis. Compounds 1 (5 mg) and 2 (5 mg) were dissolved
in 2 N HCl (2.0 mL) and heated at 100 °C for 3 h. After evaporation of
the solvent under vacuum, the residue was dissolved in H2O (10 mL)
and extracted with EtOAc (3 ꢁ 10 mL). The residue of the aqueous
phase was redissolved after evaporation in 50% MeOH (5 mg/mL). This
fraction and standards of the sugars L-arabinopyranose and L-rhamno-
pyranose (Sigma-Aldrich) were applied on normal-phase TLC, and
CHCl3ꢀMeOHꢀH2O (128:80:16) was used as the mobile phase.
Using preparative TLC with the same solvent system the sugars were
obtained, and their specific rotation values were compared with those of
the standards. Therefore, the identity of both of L-arabinopyranosyl and
L-rhamnopyranosyl moieties was confirmed for 1. In a similar manner,
the presence of the same sugars was also verified for 2.
Alkaline Hydrolysis. Compound 1 (10 mg) was dissolved in 0.5%
KOH (10 mL) and heated for 30 min under an N2 atmosphere. After
neutralization with 2 M HCl and filtration, the mixture was evaporated
to dryness and 3 was obtained (2 mg) by preparative TLC using
MeOHꢀEtOAc (80:20) as solvent.
Peracetyl quercetin 3-O-R-L-rhamnopyranosyl-(1f2)-R-L-arabi-
1
nopyranoside (3ac): [R]20 ꢀ48.6 (c 0.3, MeOH); H NMR (600
D
MHz, CDCl3) 8.05 (1H, dd, J = 8.5, 2.2 Hz, H-60Q), 7.95 (1H, d, J = 2.2
Hz, H-20Q), 7.30 (1H, d, J = 8.5 Hz, H-50Q), 7.22 (1H, d, J = 2.0 Hz,
H-8Q), 6.80 (1H, d, J = 2.0 Hz, H-6Q), 5.59 (1H, d, J = 7.0 Hz, H-1),
5.38 (1H, dd, J = 10.1, 4.0 Hz, H-30), 5.15 (1H, dt, J = 3.5, 2 Hz, H-4),
5.12 (1H, dd, J = 4.0, 2.0 Hz, H-20), 5.06 (1H, dd, J = 10.0, 3.5 Hz, H-3),
5.05 (1H, t, J = 10.0 Hz, H-40), 5.00 (1H, d, J = 2.1 Hz, H-10), 4.31 (1H,
dq, J = 10.0, 6.0 Hz, H-50), 4.03 (1H, dd, J = 10.0, 7.0 Hz, H-2), 3.73 (1H,
dd, J = 13.0, 3.5 Hz, H-5eq), 3.49 (1H, dd, J = 13.0, 2.0 Hz, H-5ax),
1.94ꢀ2.55 (27H, 9 ꢁ s, 9 ꢁ ꢀOCOCH3), 0.97 (3H, d, J = 6.0 Hz, CH3-
60); 13C NMR (50 MHz, CDCl3) 177.3 (C-4Q), 164.2 (C-7Q), 161.2
(C-5Q), 156.1 (C-2Q), 154.2 (C-9Q), 148.5 (C-40Q), 144.9
(C-30Q), 133.6 (C-3Q), 122.1 (C-60Q), 120.8 (C-10Q), 115.2 (C-
20Q), 115.5 (C-50Q), 103.8 (C-10Q), 100.0 (C-1), 99.9 (C-10), 98.6
(C-6Q), 93.9 (C-6Q), 76.5 (C-2), 71.9 (C-40), 70.5 (C-3), 70.5 (C-20),
70.5 (C-30), 68.5 (C-50), 66.1 (C-4), 63.5 (C-5), 20.6ꢀ22.7 (9 ꢁ
CH3COꢀ), 17.4 (C-60); FABMS (m/z) 959 [M + H]+; anal. calcd for
C44H47O24.
Synthesis of 3 and 3ac. A solution of 7,40-di-O-benzylquercetin 3-
O-[2,3,4-tri-O-acetyl-R-L-rhamnopyranosyl-(1f2)-3,4-di-O-acetyl-R-
L-arabinopyranoside (11) (15.0 mg, 0.015 mmol) in MeOH (5 mL)
containing 10% PdꢀC (0.04 g) was submitted to hydrogenolysis (H2, 1
atm), at 20 °C, for 4 h. The catalyst was removed by filtration over Celite,
and the solvent was evaporated under reduced pressure to give crude
quercetin 3-O-[2,3,4-tri-O-acetyl-R-L-rhamnopyranosyl-(1f2)-3,4-di-
O-acetyl-R-L-arabinopyranoside (12). A portion of 12 (5 mg) was
dissolved in 1 N NaOMe in MeOH (3 mL) and was stirred for 3 h at
20 °C. After neutralization by addition of Amberlite IRC 50 H+ ion-
exchange resin and filtration, the solvent was removed by evaporation to
afford 3 as an amorphous solid (3.0 mg). The other part of 12 was
dissolved in pyridine (1 mL), Ac2O (1 mL) was added under stirring,
and the stirring was continued for 72 h at room temperature. Evaporation of
the solvent under reduced pressure afforded 3ac as a foam (5.0 mg).
Quercetin 3-O-[R-L-rhamnopyranosyl-(1f2)-R-L-arabinopyranoside]-
7-O-R-L-rhamnopyranoside, hermannioside A (1): amorphous, yellow
’ ASSOCIATED CONTENT
S
Supporting Information. Detailed description of the
b
synthesis of compounds 3 and 3ac, 1H NMR and 13C NMR
data for the intermediate synthetic products, as well as NMR,
MS, and HRMS spectra for the isolated compounds 1
and 2 are available free of charge via the Internet at http://
pubs.acs.org.
solid; [R]20 ꢀ101.7 (c 0.8, MeOH); UV (MeOH) λmax (log ε) 256.2
D
1944
dx.doi.org/10.1021/np200444n |J. Nat. Prod. 2011, 74, 1939–1945

Journal of Natural Products
ARTICLE
’ AUTHOR INFORMATION
P.; Neszmꢀelyi, A.; Riessmaurer, I.; Wagner, H. Carbohydr. Res. 1981,
93, 43–52.
(31) Boughandjioua, R. C.; Skaltsounis, A. L.; Seguin, E.; Tillequin,
F.; Koch, M. Can. J. Chem. 1992, 70, 1956–1965.
(32) Schroeder, L. R.; Counts, K. M.; Haigs, F. C. Carbohydr. Res.
1974, 37, 368–372.
Corresponding Author
*Tel: 0030 210 7274598. Fax: 0030 210 7274594. E-mail:
skaltsounis@pharm.uoa.gr.
(33) Jurd, L. J. Org. Chem. 1962, 27, 1294–1297.
(34) Demetzos, C.; Skaltsounis, A. L.; Tillequin, F.; Koch, M.
Carbohydr. Res. 1990, 207, 131–135.
(35) Kitagawa, H. I. I.; Matsushita, K.; Shirakawa, K.; Tori, K.;
Tozyo, T.; Yoshikawa, M.; Yoshimura, Y. Tetrahedron Lett. 1981,
22, 1529–1532.
’ DEDICATION
#This paper is dedicated to the memory of our late colleague
Professor Franc-ois Tillequin.
’ REFERENCES
(1) Brullo, S.; Giusso del Galdo, G. J. Bot. Nomen. 2006, 16, 304–314.
(2) Gonnet, J.-F.; Jay, M. Phytochemistry 1972, 11, 2313–2316.
(3) Gonnet, J.-F. Phytochemistry 1975, 14, 823–828.
(4) Gonnet, J.-F. Phytochemistry 1978, 17, 1319–1323.
(5) Marco, J A.; Barbera, O.; Sanz, J. F.; Sanchez-Parareda, J.
Phytochemistry 1985, 24, 2471–2472.
(6) Barbera, O.; Sanz, J. F.; Sꢀanchez-Parareda, J.; Marco, J. A.
Phytochemistry 1986, 25, 2361–2365.
(7) Adell, J.; Barbera, O.; Marco, J. A. Phytochemistry 1988, 27,
2967–2970.
(8) Marco, J. A.; Adell, J.; Barbera, O.; Strack, D.; Wray, V.
Phytochemistry 1989, 28, 1513–1516.
(9) Pistelli, L.; Spera, K.; Flamini, G.; Mele, S.; Morelli, I. Phyto-
chemistry 1996, 42, 1455–1458.
(10) Beier, C. R.; Mundy, P. B. J. Carbohydr. Chem. 1984, 3,
253–266.
(11) Pawan, K. A. Phytochemistry 1992, 31, 3307–3330.
(12) Mizutani, K.; Hayashi, A.; Rasai, K.; Tanaka, O.; Yoshida, N.;
Nakajima, T. Carbohyd. Res. 1984, 126, 177–189.
(13) Razanamahefa,B.;Demetzos,C.;Skaltsounis,A.L.;Andriantsiferana,
M.; Tillequin, F. Heterocycles 1994, 38, 357–361.
(14) Abdel-Shafeek, A. K.; El-Messiry, M. M.; Shahat, A. A.; Apers,
S.; Pieters, L.; Seif-El Nasr, M. M. J. Nat. Prod. 2000, 63, 845–847.
(15) C-alis, I.; Heilmann, J.; Tasdemir, D.; Linden, A.; Ireland, M. C.;
Sticher, O. J. Nat. Prod. 2001, 64, 961–964.
(16) Stoi, D.; Gorunovi, M.; Skaltsounis, A. L.; Tillequin, F.; Koch,
M. Helv. Chim. Acta 1988, 71, 348–353.
(17) Geiger, H.; Maier, H.; Markham, K. R. Z. Naturforsch. 1983,
38c, 490–491.
(18) Marzouk, M. M.; Kawashty, S. A.; Ibrahium, L. F.; Saleh, N.; A,
M.; Al-Nowaihi, A-S. M. Nat. Prod. Commun. 2008, 3, 1325–1328.
(19) Veit, M.; Pauli, G. F. J. Nat. Prod. 1999, 62, 1301–1303.
(20) Markham, K. R.; Ternai, B.; Stanley, R.; Geiger, H.; Mabry, T. J.
Tetrahedron 1978, 34, 1389–1397.
(21) Liptꢁak, A.; Szurmai, P.; Nꢁanꢁasi, P.; Neszmꢀelyi, A. Tetrahedron
1982, 38, 3489–3497.
(22) Durette, P. L.; Horton, D.; Bhacca, N. S. Carbohydr. Res. 1969,
10, 565–577.
(23) Mizutani, K.; Ohtani, K.; Kasai, R.; Tanaka, O.; Matsura, H.
Chem. Pharm. Bull. 1985, 33, 2266–2272.
(24) Mukhopadhyay, B.; Field, R. A. Carbohydr. Res. 2004, 339,
1285–1291.
(25) Kim, M.-R.; Lee, J.-Y.; Lee, H.-H.; Aryal, D.-Q.; Kim, Y.-G.;
Kim, S.-K.; Woo, E.-R.; Kang, K.-W. Food Chem. Toxicol. 2006,
44, 1299–1307.
(26) Hosoi, S.; Shimizu, E.; Ohno, K.; Yokosawa, R.; Kuninaga, S.;
Coskun, M.; Sakushima, A. Phytochem. Anal. 2006, 17, 20–24.
(27) Sivakumaran, T.; Jones, K. N. Can. J. Chem. 1967, 45, 2493–
2500.
(28) Ballou, C. E. J. Am. Chem. Soc. 1957, 79, 165–166.
(29) Fleet, G. W. J.; Gough, M. J. Tetrahedron Lett. 1982,
23, 4509–4512.
(30) (a) Liptꢁak, A.; Nꢁanꢁasi, P.; Neszmꢀelyi, A.; Riessmaurer, I.;
Wagner, H. Carbohydr. Res. 1981, 93, 43–52. (b) Liptꢁak, A.; Nꢁanꢁasi,
1945
dx.doi.org/10.1021/np200444n |J. Nat. Prod. 2011, 74, 1939–1945