Structure-activity relationship study on α1 adrenergic receptor antagonists from beer

Article abstract of DOI:10.1016/j.bmcl.2009.08.068

Hordatine A and aperidine have been previously isolated from beer as active ingredients, which bind to muscarinic M₃ receptor. In addition, these compounds have exhibited antagonist activity against the α_1A adrenoceptor. Although the relative structures of these two molecules have previously been determined, the absolute stereochemistry was unclear. Hence, to elucidate the absolute stereochemistry of natural hordatine A, we synthesized each enantiomer of hordatine A and aperidine from optically pure dehydrodi-p-coumaric acid. Several additional related compounds were also synthesized for structure-activity relationship studies. Chiral column HPLC analysis demonstrated that the absolute stereochemistry of natural hordatine A is (2S,3S), while based on the isomerization mechanism, the stereochemistry of aperidine is (2R,3S). The α_1A adrenoceptor binding activity of (2R,3R)-hordatine A is the most potent among the enantiomeric pairs of hordatines and aperidines. Furthermore, the related, synthetic compound, (2R,3R)-methyl benzofurancarboxylate exhibits antagonist activity against the α_1A adrenoceptor at a lower concentration than that of hordatine A.

Full text of DOI:10.1016/j.bmcl.2009.08.068

Bioorganic & Medicinal Chemistry Letters 19 (2009) 5905–5908
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Structure–activity relationship study on a1 adrenergic receptor antagonists
from beer
a
a
a
a
a
Toshiyuki Wakimoto a, , Makoto Nitta , Kana Kasahara , Taketo Chiba , Ye Yiping , Kuniro Tsuji ,
*
Toshiyuki Kan a, Haruo Nukaya a, Masaji Ishiguro b, Minako Koike c, Yoshiaki Yokoo c, Yoshihide Suwa c
a School of Pharmaceutical Sciences, University of Shizuoka and Global COE Program, 52-1 Yada, Suruga-ku, Shizuoka-shi, Shizuoka 422-8526, Japan
b Suntory Institute for Bioorganic Research, 1-1-1 Wakayamadai, Shimamoto-cho, Mishima-gun, Osaka 618-8503, Japan
c Technical Development Department, Suntory Holdings Limited, 5-2-5 Yamazaki, Shimamoto-cho, Mishima-gun, Osaka 618-0001, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Hordatine A and aperidine have been previously isolated from beer as active ingredients, which bind to
muscarinic M3 receptor. In addition, these compounds have exhibited antagonist activity against the a1A
adrenoceptor. Although the relative structures of these two molecules have previously been determined,
the absolute stereochemistry was unclear. Hence, to elucidate the absolute stereochemistry of natural
hordatine A, we synthesized each enantiomer of hordatine A and aperidine from optically pure dehydro-
di-p-coumaric acid. Several additional related compounds were also synthesized for structure–activity
relationship studies. Chiral column HPLC analysis demonstrated that the absolute stereochemistry of nat-
ural hordatine A is (2S,3S), while based on the isomerization mechanism, the stereochemistry of aperi-
dine is (2R,3S). The a1A adrenoceptor binding activity of (2R,3R)-hordatine A is the most potent among
the enantiomeric pairs of hordatines and aperidines. Furthermore, the related, synthetic compound,
(2R,3R)-methyl benzofurancarboxylate exhibits antagonist activity against the a1A adrenoceptor at a
lower concentration than that of hordatine A.
Received 13 June 2009
Revised 18 August 2009
Accepted 19 August 2009
Available online 21 August 2009
Keywords:
Hordatine A
Aperidine
Absolute stereochemistry
a1A adrenoceptor
Ó 2009 Elsevier Ltd. All rights reserved.
Hordatine A (1) was initially discovered in 1996 from barley as
located in both the central and peripheral nervous systems, are of
great therapeutic relevance due to their important roles in the con-
traction of smooth muscle tissue, especially in the cardiovascular
an antifungal compound, which was predominantly distributed in
the shoots of seedlings.1–3 Subsequent biochemical and genetic
studies have revealed that 1 can act as a phytoalexin for pathogen
defense.4–6 Stoessl has originally reported that 1 (diacetate) can be
system and lower urinary tract.11,12 Moreover, the
a1 adrenoceptor
subtype can be further divided into a1A, a1B, and a1D. Selective
obtained in the optically active form (½a D23
ꢀ
¼ þ69), and the relative
ligands acting at these subtypes have therapeutic potential in the
treatment of hypertension, stress urinary incontinence, benign
prostatic hyperplasia, and lower urinary tract symptoms.
Thus, determining the absolute stereochemistry of 1 should be
important not only for predicting the binding mode to a1 adreno-
ceptors, but also for developing subtype-selective antagonists.13–15
Herein, we report the complete structure of natural 1 as well as the
stereochemistry was tentatively proposed to be cis.1,2 Later, Yoshi-
hara et al. reported that the relative stereochemistry of 1 was trans
as well as grossamide based on the 1H NMR data.7 However, in
both cases, the absolute chemistry of 1 has yet to be reported.
In 2004, we reported the isolation of two active ingredients
from beer in the guidance of muscarinic M3 receptor binding activ-
ity,8 and the relative structures of these two active compounds, 1
and aperidine (2), were disclosed (Fig. 1).9 Furthermore, we found
that these compounds exhibited antagonist activity against the a1
adrenoceptor.
a1A adrenoceptor antagonist activities of all stereoisomers of 1 and
related synthetic compounds.
First, the enantiomeric purity of 1 from beer has been analyzed
by chiral column HPLC, which revealed natural 1 is a single enan-
tiomer.16 In order to determine the absolute stereochemistry of 1,
we attempted the following methodology: (1) optical resolution
and absolute stereochemical assignment of dehydrodi-p-coumaric
acid (4), (2) synthesis of 1 from optically pure 4, (3) comparison of
retention time using chiral column HPLC and the optical rotation
between natural and synthetic 1. Racemic 4 was synthesized by
single electron oxidation with horseradish peroxidase and H2O2.
The optical resolution of 4 was performed by chiral column HPLC
Adrenergic receptors, which belong to the seven transmem-
brane spanning, G-protein-coupled receptor super family, mediate
the actions of the endogenous catecholamines, adrenaline, and
noradrenaline. Adrenergic receptors are broadly categorized into
a1, a
2, and b subtypes.10 Of these, a1 adrenoceptors, which are
* Corresponding author. Fax: +81 54 264 5745.
E-mail address: wakimoto@u-shizuoka-ken.ac.jp (T. Wakimoto).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.08.068

5906
T. Wakimoto et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5905–5908
With optically active 4 in hand, each enantiomer of 1 was syn-
NH2
NH
thesized according to a modified Stossel’s method (Fig. 3). Pro-
tected agmatine was prepared by condensation between N,N0-
diBoc-imidazoyl guanidine and N-Cbz-diaminobutane in good
yield. Subsequent hydrogenolysis of the Cbz group afforded N,N0-
diBoc-agmatine. Coupling 4 with N,N0-diBoc-agmatine using PyBOP
in the presence of HOBt gave the N,N0,N00,N000-tetra-Boc-hordatine A
(10). Finally, deprotection of the Boc group with TFA yielded 1.
Chromatographic comparison using chiral column HPLC demon-
strated that the retention time of natural 1 was identical to that
of synthetic (2S,3S)-1. We therefore concluded that the absolute
stereochemistry of natural 1 is (2S,3S). However it was unexpected
that this configuration was the opposite to that of (ꢁ)-ephedradine
A, which was a macrocyclic spermine alkaloid having similar dihy-
drobenzofuran skeleton.18
HN
HN
4
O
O
H
N
2'
3S
H2N
4'
2S
N
H
3'
OH
NH
O
Hordatine A (1)
NH2
HN
NH
To confirm our assignment using the modified Mosher’s meth-
od, chiral column HPLC analysis was applied to synthetic (2R,3R)-
dimethyl-dehydrodi-p-coumarate (12), which was prepared from
the intermediate (11) for total synthesis of (ꢁ)-ephedradine A.19
Consequently, 1 had antipodal stereochemistry to (ꢁ)-ephedradine
A (Fig. 4), which is consistent with the result from the modified
Mosher’s method. It has been known that some natural lignans
have been isolated as non-uniform stereochemical isomers, includ-
ing the enantiomeric form or a racemic mixture6 and lignans with
a dihydrobenzofuran skeleton, such as 1 and ephedradine as
well.20
HN
O
O
H
N
3S
H2N
2R
N
H
OH
NH
O
Aperidine (2)
Figure 1. Structures of hordatine A (1) and aperidine (2).
Next our attention turned to determination of absolute stereo-
chemistry of 2. During the final step of the total synthesis of 1, 2
was obtained as a minor product. We assumed that 2 was formed
by an acid catalyzed isomerization during the deprotection step,
and the most favorable condition to form 2 was a pH of 8.0 in an
aqueous buffer, which yielded a product ratio cis:trans = 1:3. More-
over, based on the chiral column HPLC analysis, natural 2 was so
optically pure that racemization did not occur during the isomeri-
zation step. Hence, we proposed that isomerization mechanism
from 1 to 2 could be via a p-quinone methide intermediate because
this isomerization route retained the C-2 stereochemistry.21,22
Treatment of N,N0,N00,N000-tetra-Boc-hordatine A with TFA-d did
not lead to a H/D exchange in 1H NMR spectrum, which supports
an isomerization mechanism via p-quinone methide. Therefore,
the absolute stereochemistry of natural 2 was determined as
(2R,3S).
PGME
O
O
+0.10 +0.28
-0.04 -0.04
HO
+0.04
OH
-0.02
+0.22
+0.13
O
-0.04 -0.04
+0.03
Figure 2. Distribution of DdS-R values for PGME derivatives (5) in DMSO-d6.
with chiracel OJ-RH. Taking advantage of the carboxylic acid on the
chiral center, a modified Mosher’s method with PGME amide was
applied to each enantiomer.17 The
Dd values were clearly distrib-
uted (Fig. 2), which led to the assignment of the absolute stereo-
chemistry of each enantiomer. Consequently, the elution order of
the two peaks were (S,S)-(+) and then (R,R)-(ꢁ)-4.
HO
HO
O
O
O
O
O
a
b
S
HO
HO
HO
OH
OH
OH
S
O
O
3
racemate-4
(+) - 4
H
N
N
H
N
N
c
d
e
f
NHBoc
N
NHBoc
NBoc
N
CbzHN
H2N
NBoc
HN
NH2
BocN
NHBoc
6
7
8
9
NHBoc
NH
NH2
NH
NH2
NH
BocN
HN
HN
g
+
HN
HN
HN
O
O
O
O
O
O
H
N
H
H
N
BocHN
H2N
N
H2N
N
H
N
H
N
H
OH
OH
OH
NBoc
NH
NH
O
O
O
10
(+) - 1
(+) - 2
Figure 3. Synthesis of hordatine A (1) and aperidine (2). Reagents and conditions: (a) horseradish peroxidase, H2O2, phosphate buffer (pH 7.4) (34%); (b) ChiralPak OJ-RH; (c)
(Boc)2O, LiH, THF (70%); (d) N-Cbz-diaminobutane, Et3N, CH3CN, H2O (86%); (e) H2, Pd-C, AcOEt (85%); (f) (+)-4, EDCI, HOBt, DMF (53%); (g) (1)TFA, (2) Sephadex G-15 column
chromatography with 0.01 N HCl (92%).

T. Wakimoto et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5905–5908
5907
#1
HO
R
MeO
O
O
O
O
a, b, c
Br
MeO
OBn
OH
OH
R
O
O
11
(2R,3R)-12
#2
HO
O
O
MeO
O
S
d
HO
OH
MeO
S
O
O
(+) -4
(2S,3S)-12
Figure 4. Synthesis of optically active 12. Reagents and conditions: (a) CH2N2, Et2O; (b) methyl acrylate, tri(o-tolyl)phosphine, Pd(OAc)2, NEt3, DMF, 100 °C (36% in 2 steps);
(c) BCl3, CH2Cl2, ꢁ78 °C (89%); (d) CH2N2, Et2O (93%).
To examine binding23 and antagonist24 activities of optically ac-
tive hordatines and aperidines, all stereoisomers of 1 and 2 were
synthesized utilizing optically active 4. In regards to binding activ-
ities to the a1A adrenoceptor, (2S,3S)-(+)-1 and (2R,3R)-(ꢁ)-1
tional group is essential for binding to this receptor.25 Hordatine
A also has a guanidino group as a cationic function. However, there
are two identical functionalities on the molecule, one is located on
the C-4 agmatine side chain and another is C-40 agmatine. We
hypothesized that only one moiety is critical for binding, and the
other is unnecessary. In order to clarify this, we prepared the
two types of hordatine-related compounds by a selective reaction
on each carboxylic acid of dehydrodi-p-coumaric acid. Methyl ben-
zofurancarboxylate derivative (13), which lacked the C-4 agmatine
side chain of 1, was synthesized from monomethylester prepared
by selective methylation of the C-4 carboxylic acid. On the other
hand, benzofurancarboxamide derivative (14), which lacked a
C-40 agmatine side chain, was prepared from 4. Moreover, p-cou-
maroylagmatine (15) was synthesized by condensation between
p-coumaric acid and N,N0,N00,N000-tetra-Boc-agmatine (Fig. 5).
Compound 13 exhibited binding activity to the a1A adrenocep-
tor. The IC50 values for (2S,3S)-13 and (2R,3R)-13 were 2.16 and
exhibited IC50 values of 0.31 lg/mL and 0.29 lg/mL, respectively.
Hence, the stereochemistry on the dihydrobenzofuran ring of 1
does not seem to be essential for binding to the receptor. On the
other hand, the relationship between optical isomers of 2 was dif-
ferent. Synthetic (2S,3R)-(ꢁ)-2 and (2R,3S)-(+)-2 exhibited IC50 val-
ues of 2.44 lg/mL and 0.42 lg/mL, respectively. Interestingly,
these findings indicate unnatural isomers of 1 and 2 are more po-
tent than the corresponding natural isomers. Among all four ster-
eoisomers of 1, (2R,3S)-2 exhibits the most potent antagonistic
activity at the a1A adrenoceptor. It is likely that the R configuration
at the C-2 might be suitable for binding to receptors (Table 1).
Past structure–activity relationship studies of ligands for the
a1A adrenoceptor supports the hypothesis that the cationic func-
0.56 lM, respectively. On the other hand, compound 14 was inac-
tive, while 15 showed only weak activity. The antagonistic activity
of each compound was clearly associated with its binding ability to
the a1A adrenoceptor. In particular, (2R,3R)-13 exhibited a more
potent antagonist activity than that of 1 at a lower concentration
Table 1
Activities of hordatine derivatives against a1 receptor (IC50
)
Binding activity (mM)
Antagonistic activity (mM)
(IC50 = 9.72 lM). The assay results for these molecules are clear-
cut as only compound 13 is active for both binding and the antag-
(2S,3S)-1
(2R,3R)-1
(2R,3S)-2
(2S,3R)-2
(2S,3S)-13
(2R,3R)-13
(2S,3S)-14
(2R,3R)-14
15
0.31
0.29
0.42
2.44
2.16
0.56
>100
>100
57.6
19.3
14.3
4.90
17.7
42.2
9.72
>100
>100
>100
onist test, while 14 is inactive. These data indicate that the C-40
agmatine unit is essential for activity against the a1A adrenoceptor.
Moreover, upon comparing optical isomers, (2R,3R)-13 exhibits the
most potent activity. Interestingly, the antagonistic activity of
(2R,3R)-13 is more potent than natural 1. Furthermore, the weak
activity of coumaroylagmatine (15) suggests that the dihydroben-
zofuran ring of 1 is also important Table 1.
MeO
O
O
H
NH2
NH
H2N
N
N
H
O
OH
HN
NH
(2S,3S)-13
HN
O
O
O
H
H2N
N
HO
N
H
OH
NH
O
OH
(2S,3S)-14
15
Figure 5. Synthetic hordatin derivatives.

5908
T. Wakimoto et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5905–5908
References and notes
1. Stoessl, A. Tetrahedron Lett. 1966, 21, 2287.
2. Stoessl, A. Tetrahedron Lett. 1966, 25, 2849.
3. Stoessl, A. Can. J. Chem. 1967, 45, 1745.
4. Davin, L. B.; Wang, H. B.; Crowell, A. L.; Bedgar, D. L.; Martin, D. M.; Sarkanen,
S.; Lewis, N. G. Science 1997, 275, 362.
5. Davin, L. B.; Lewis, N. G. Phytochem. Rev. 2003, 2, 257.
6. Suzuki, S.; Umezawa, T.; Shimada, M. Biosci., Biotechnol., Biochem. 2002, 66,
1262.
7. Yoshihara, T.; Yamaguchi, K.; Sakamura, S. Agric. Biol. Chem. 1983, 47, 217.
8. Yokoo, Y.; Fujii, W.; Hori, H.; Nagao, K.; Suwa, Y.; Taniyama, K.; Tsuji, K.;
Yoshida, T.; Nukaya, H. Alcohol Clin. Exp. Res. 2004, 28, 129S.
9. Yamaji, N.; Yokoo, Y.; Iwashita, T.; Nemoto, A.; Koike, M.; Suwa, Y.; Wakimoto,
T.; Tsuji, K.; Nukaya, H. Alcohol Clin. Exp. Res. 2007, 31, 9S.
10. Civantos, C. B.; Aleixandre, A. A. Pharmacol. Res. 2001, 44, 195.
11. Lowe, F. C. Clin. Ther. 2004, 26, 1701.
12. Muramatsu, I.; Suzuki, F.; Tanaka, T.; Yamamoto, H.; Morishima, S. Yakugaku
Zasshi 2006, 126, 187.
13. Dardonville, C.; Goya, P.; Rozas, I.; Alsasua, A.; Martin, M. I.; Borrego, M. Bioorg.
Med. Chem. 2000, 8, 1567.
14. Ishiguro, M.; Futabayashi, Y.; Ohnuki, T.; Ahmed, M.; Muramatsu, I.; Nagatomo,
T. Life Sci. 2002, 71, 2531.
15. Kinsella, K. G.; Rozas, I.; Watson, W. G. Biochem. Biophys. Res. Commun. 2005,
333, 737.
Figure 6. Complex model of (R,R)-hordatine A at the binding cleft of a1 receptor
models.
16. Racemic hordatine A was synthesized from racemic 4 (Fig. 3).
17. Yabuuchi, T.; Ooi, T.; Kusumi, T. Chirality 1997, 9, 550.
18. Tamada, M.; Endo, K.; Hikino, H.; Kabuto, C. Tetrahedron Lett. 1979, 20, 873.
19. Kurosawa, W.; Kan, T.; Fukuyama, T. J. Am. Chem. Soc. 2003, 125, 8112.
20. Ralph, J.; Peng, J.; Lu, F.; Hatfield, R. D.; Helm, R. F. J. Agric. Food Chem. 1999, 47,
2991.
21. Lundgren, N. L.; Theander, O. Phytochemistry 1988, 27, 829.
22. Wada, H.; Kido, T.; Tanaka, N.; Murakami, T.; Saiki, Y.; Chen, C. M. Chem. Pharm.
Bull. 1992, 40, 2099.
To confirm the above results, hordatine A was docked into a a1A
adrenoceptor model (Fig. 6), which was constructed by homology
modeling using an antagonist-bound form of GPCR as the tem-
plate.26 We assumed the cationic guanidino group interacted with
the carboxylate group of Asp148 in TM3 through an ionic bond.
Hence, the conformation of 1 was minimized and subjected to
manual docking in order to yield reasonable interactions with
the cavity of the receptor binding site calculated and visualized
by the Binding-Site module installed in Insight II. The docking
experiment assumed a salt bridge between the carboxylate group
of Asp148 and C-4 agmatine side chain was not a reasonable inter-
action due to the results of structural optimization with molecular
dynamics and energy minimization using Discover 3. On the other
hand, the docking model considering the C-40 agmatine side chain
as a cationic center displayed a reasonable conformation of 1 in the
receptor model. In this model, the other agmatine side chain did
not significantly interact with any residues of the receptor. These
results are in good agreement with our SAR data.
23. Assay of a1A adrenoceptor binding activity: Test samples were added to
a
suspension of
a homogenate of rat salivary gland containing the a1
adrenoceptor in the presence of 0.06 nM [3H]-prazosin, and the mixture was
incubated at 22 °C for 60 min. The reaction was terminated by vacuum filtering
through glass-fiber filters (GF/B; Perkin Elmer, Wellesley, MA), which were
individually washed several times with 50 mM Tris–HCl. The radioactivity
remaining on the filter was counted in
(MicroScint-O, Perkin Elmer) using TopCount (Perkin Elmer) liquid
scintillation counter. The specific binding of [3H]-prazosin was calculated by
a liquid scintillation cocktail
a
subtracting the nonspecific binding in the presence of 10 lM of phentolamine
from the total binding of [3H]-prazosin. IC50 values were obtained as the
concentration that competitively inhibited the binding of [3H]-prazosin to the
a1 adrenoceptor at 50% of maximal binding.
24. Assay of a1A adrenoceptor antagonist activity: Rings of rat caudal artery denuded
of endothelium were suspended in 20 mL organ baths containing an
oxygenated (95% O2 and 5% of CO2) and pre-warmed (37 °C) physiological
salt solution with the following composition: 118.0 mM NaCl, 4.7 mM KCl,
1.2 mM MgSO4, 2.5 mM CaCl2, 1.2 mM KH2PO4, 25.0 mM NaHCO3, and
11.0 mM glucose (pH 7.4). Yohimbine (1
lM), propranolol (1
lM), pyrilamine
In summary, we disclosed the absolute structures of 1 and 2 uti-
lizing the synthesis of all possible stereoisomers of 1. Moreover,
five additional related compounds were also prepared for the
structure–activity relationship study. The resulting assay data
demonstrated that the guanidino group on C-40 agmatine side
chain is critical for the affinity to the receptor, which was also sup-
ported by the docking study. The asymmetric total synthesis and
development of a subtype-selective antagonist based on these
molecules are currently under investigation in our laboratory.
(1 M), atropine (1 M), and methysergide (1 l
l
l
M) were also present in assay
incubates to block the a2 adrenergic, b-adrenergic, histamine H1, muscarinic
and 5-HT2 receptors, respectively. The tissues were connected to force
transducers for isometric tension recordings, stretched to a resting tension of
1 g, and then allowed to equilibrate for 30 to 60 min. During the equilibration,
the tissues were washed repeatedly and the tension readjusted. The tissues
were exposed to
phenylephrine (10
a
l
submaximal concentration of the reference agonist
M) to obtain control contractile response. After
a
stabilization of the phenylephrine-induced response, the concentration of the
test sample was increased or the reference antagonist prazosin was added
cumulatively. At each concentration, the tissues remained in contact with the
solution until a stable response was obtained or for a maximum of 15 min.
Inhibition of the phenylephrine-induced response by a test sample indicated
antagonist activity at the a1A adrenergic receptors.
Supplementary data
25. Waugh, J. J. D.; Gaivin, J. R.; Zuscik, J. M.; Gonzalez, C. P.; Ross, A. S.; Yun, J.;
Perez, M. D. J. Biol. Chem. 2001, 276, 25366.
26. Ishiguro, M. ChemBioChem 2004, 5, 1210.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2009.08.068.