9-Aminomethyl-9,10-dihydroanthracene (AMDA) analogs as structural probes for steric tolerance in 5-HT2A and H1 receptor binding sites

Article abstract of DOI:10.1016/j.bmcl.2009.12.064

Synthesis, radioligand binding and molecular modeling studies of several 9-aminomethyl-9,10-dihydroanthracene (AMDA) analogs were carried out to determine the extent of the steric tolerance associated with expansion of the tricyclic ring system and amine

Full text of DOI:10.1016/j.bmcl.2009.12.064

Bioorganic & Medicinal Chemistry Letters 20 (2010) 935–938
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
9-Aminomethyl-9,10-dihydroanthracene (AMDA) analogs as structural probes
for steric tolerance in 5-HT2A and H1 receptor binding sites
Jitesh R. Shah a, Philip D. Mosier a, Srinivas Peddi a, Bryan L. Roth b, Richard B. Westkaemper a,
*
a Department of Medicinal Chemistry, School of Pharmacy, Virginia Commonwealth University, Richmond, VA 23298, USA
b Department of Pharmacology, University of North Carolina School of Medicine, Chapel Hill, NC 27599, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Synthesis, radioligand binding and molecular modeling studies of several 9-aminomethyl-9,10-dihydro-
anthracene (AMDA) analogs were carried out to determine the extent of the steric tolerance associated
with expansion of the tricyclic ring system and amine substitution at 5-HT2A and H1 receptors. A mixture
of (7,12-dihydrotetraphene-12-yl)methanamine and (6,11-dihydrotetracene-11-yl)methanamine in a
75–25% ratio was found to have an apparent Ki of 10 nM at the 5-HT2A receptor. A substantial binding
affinity for (7,12-dihydrotetraphene-3-methoxy-12-yl)methanamine at the 5-HT2A receptor (Ki = 21 nM)
was also observed. Interestingly, this compound was found to have 100-fold selectivity for 5-HT2A over
the H1 receptor (Ki = 2500 nM). N-Phenylalkyl-AMDA derivatives, in which the length of the alkyl chain
varied from methylene to n-butylene, were found to have only weak affinity for both 5-HT2A and H1
receptors (Ki = 223 to 964 nM). Our results show that large rigid annulated AMDA analogs can be steri-
cally accommodated within the proposed 5-HT2A binding site.
Received 17 November 2009
Revised 14 December 2009
Accepted 15 December 2009
Available online 23 December 2009
Keywords:
5-HT2A receptor
H1 receptor
G protein-coupled receptor (GPCR)
GPCR modeling
AMDA
Phenylethylamines
Ó 2009 Elsevier Ltd. All rights reserved.
Over the past few years, much information has been generated
about 5-HT2 receptor subtypes (5-HT2A, 5-HT2B, 5-HT2C) and their
ligands.1,2 However, it is still not clear exactly how serotonergic
ligands (agonists/partial agonists/antagonists/inverse agonists)
including 5-HT itself interact with the receptors. In the absence
of a crystal structure for any member of the 5-HT2 receptor family,
homology models generated from a representative GPCR such as
rhodopsin or a b-adrenoceptor are used to study possible drug–
receptor interactions.3–5 Homology modeling and ligand docking,
supported by experimental data, have allowed us to propose pos-
sible ligand-receptor interactions responsible for the observed
affinity of the lead compound 9-aminomethyl-9,10-dihydroanthra-
cene (AMDA) and its analogues at the human 5-HT2A and H1 recep-
tors.6–10 This work expands upon these earlier structure–affinity
relationship studies via new AMDA analogs designed to probe
the stereoelectronic accessibility of the receptor binding sites. In
particular, annulated analogs of AMDA were synthesized and
tested to determine the extent of bulk tolerance within the binding
site of both 5-HT2A and H1 receptors. Further, to explore the poten-
tial differences between the 5-HT2A and H1 receptors in the vicinity
of the ammonium ion binding site (known to be D3.32 based on
experimental and modeling data),11,12 N-phenylalkyl analogues of
AMDA were synthesized and tested.
but not widely employed. Examples where this concept has been
successfully applied include analogs of adenosine,13 ATP,14
cyclophosphamide,15 isotryptamine16 and allocolchicinoids.17 By
carrying out a systematic annulation and substitution, one can
infer information about the dimensions as well as the interactions
within the binding site.18,19 AMDA and its analogues contain a rigid
tricyclic scaffold. Previous SAR studies by Westkaemper and co-
workers5,7 using the 5-HT2A receptor have indicated that there
may be a substantial amount of steric tolerance around the rigid
tricyclic scaffold of AMDA, as determined by substituent effects
at the C3-position. However, the largest substituents examined
to date (3-phenylpropyl-AMDA, Ki = 3.2 nM; 3-n-hexyl-AMDA,
Ki = 7 nM; 3-n-pentyloxy-AMDA, Ki = 23 nM)7 have a high degree
of rotational freedom due to alkyl chain flexibility. There are cur-
rently no known 5-HT2A ligands whose rigid core exceeds the
dimensions of the annulated AMDA analogs reported here. For
example, in the case of tricyclic antidepressants, ergot alkaloids
and apomorphine the longest dimension is equivalent to three lin-
early fused rings. As a more stringent test of bulk tolerance, rigid
and elongated structures were constructed by addition of fused
rings to the tricyclic structure of AMDA. If accommodated, the
AMDA pharmacophore expanded via benzfusion would also pro-
vide a new rigid scaffold for the introduction of substituents that
may have access to additional binding site residues more distant
from the tricyclic core than those accessible with previously-em-
ployed scaffolds.
The concept of using annulated analogues of established ligands
as dimensional probes for ligand binding sites has been reported
The initial compounds for which syntheses were attempted
were dihydrotetracenes (1a and 1b, Table 1). However, the
* Corresponding author. Tel.: +1 804 828 6449; fax: +1 804 828 7625.
E-mail address: rbwestka@vcu.edu (R.B. Westkaemper).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.12.064

936
J. R. Shah et al. / Bioorg. Med. Chem. Lett. 20 (2010) 935–938
Table 1
vored
a-substituted product. The binding affinity data obtained
Observed binding affinity data for 6,11-dihydrotetracene (1a and 1b), 7,12-dihydro-
tetraphene (2a and 2b) and 9,10-dihydroanthracene (3a–d) structural probes at
5-HT2A and H1 receptors
for this mixture produced an apparent Ki value of 10 nM (5-
HT2A). Thus it was concluded that at least one of the isomers would
be predicted to bind with Ki 6 10 nM. The annulated methoxy-ana-
logue 1b was also synthesized using a similar procedure. The at-
tempted synthesis also resulted in a mixture of two isomers (1b
and 2b). Due to the challenges faced in the final cyclization step
and difficulty in the separation of products, the reaction resulted
in a very low yield of the final products. While a mixture was ob-
tained, it was possible to separate the methoxy-substituted iso-
mers, although only dihydrotetraphene isomer 2b was obtained
preparatively. Dihydrotetraphene 2b was determined to have sub-
stantial affinity for the 5-HT2A receptor (Ki = 21 nM) while possess-
ing very low affinity for the H1 receptor (Ki = 2500 nM).
In order to explore potential differences between the 5-HT2A
and H1 receptor binding sites in the vicinity of the ligand ammo-
nium ion, various phenylalkylamine analogs of AMDA (3a–d) were
synthesized and tested. It has previously been shown10 that
increasing either the length of the alkyl linker connecting the dihy-
droanthracene core to the amine or the degree of N-methylation
progressively increases the affinity of AMDA analogs for H1, but
not for 5-HT2A. The binding affinity of the larger N-benzyl-AMDA
(3a) has also been reported9 and shown to have a Ki value of
721 nM for the 5-HT2A receptor, dramatically lower than for the
parent compound AMDA (Ki = 20 nM). In order to further investi-
gate the possibility that the ligand N-substituent may interact with
more remote hydrophobic regions of the receptor in a manner sim-
ilar to what has been proposed7 for 3-phenylpropyl AMDA, analogs
3b–d containing longer flexible linkers between the amine and the
phenyl ring were synthesized and tested. The observed affinity
data showed that each of the compounds 3b–d had relatively
low affinity (Ki = 223–964 nM) for both 5-HT2A and H1 receptors
with practically no selectivity, indicating that the larger phen-
ylalkylamine substituents are not sterically well-tolerated in either
receptor.
H2N
10
12
5
1
4
R
9
8
11
2
3
7
6
1
2
H2N
R
1
3
4
11
10
9
12
5
8
7
6
2
R
HN
8
1
4
2
3
9
7
6
5
10
3
Compd
R
Ki, nM ( SD)a
H1
5-HT2A
1a + 2a
1b
2b
3a
3b
–H
10 ( 1)b
ND
ND
ND
–OCH3
–OCH3
21 ( 1)
2640 ( 528)
ND
242 ( 27)
725 ( 128)
964 ( 148)
–CH2C6H5
721 ( 223)c
400 ( 34)
223 ( 5)
–(CH2)2C6H5
–(CH2)3C6H5
–(CH2)4C6H5
3c
3d
365 ( 17)
In order to provide potential insights into how the benz-fused
ligands interact with their receptors at an atomic level, automated
ligand docking was carried out using an existing 5-HT2A homology
model, the construction and refinement of which has previously
been described in detail.6,7 This process is briefly described here.
The initial step consisted of sequence alignment of the h5-HT2A
(P28233) sequence with several related class A GPCRs20 using the
CLUSTALX program.21 This produced an unambiguous alignment in
the transmembrane (TM) helical regions of h5-HT2A sequences
with that of both bovine rhodopsin and the b2-adrenoceptor. This
alignment, along with a file containing the atomic coordinates of
the template receptor (PDB ID = 1U19), was used as input to
the MODELLER software package22 to generate a population of 100
a
[3H]Ketanserin-labeled cloned h5-HT2A or [3H]pyrilamine-labeled cloned hH1
sites. Values represent the mean of computer-derived Ki estimates (using GraphPad
PRISM) of multiple determinations. ND = not determined.
b
Represents a mixture (75:25) of 2a and 1a. See text for details.
c
From Runyon et al.9
synthetic methodology chosen for 1a (Scheme 1) resulted in a mix-
ture (75:25) of positional isomers of dihydrotetraphene 2a and
dihydrotetracene 1a, respectively. In retrospect this is consistent
with the greater stability of the intermediate carbocation formed
as a result of electrophilic addition at the
b-carbon of the naphthalene ring, giving rise to the kinetically-fa-
a-carbon versus the
CHO
R
CHO
Br
CN
Br
b
c
a
6a,b
a R = —H
5
4
b
—
R = OCH3
H2N
H2N
H2N
R
OH
R
d
R
1a,b
2a,b
7a,b
Scheme 1. Reagents and conditions: (a) DIBAL-H, CH2Cl2; (b) Pd(PPh3)4, K3PO4, 2-naphthylboronic acid/6-methoxy-2-naphthylboronic acid; (c) TMSCN, ZnI, CH2Cl2, LAH,
THF; (d) CH3SO3H.

J. R. Shah et al. / Bioorg. Med. Chem. Lett. 20 (2010) 935–938
937
different h5-HT2A homology models. Each of these receptors was
subsequently energy-minimized. The automated docking program
GOLD version 3.01 (Cambridge Crystallographic Data Centre, Cam-
bridge, UK)23 was then used to dock selected ligands into the 5-
HT2A populations. Based on the fitness function values, steric and
electronic interactions of the docked poses and reported site-direc-
ted mutagenesis data, one receptor model was selected to repre-
sent the AMDA-antagonist binding site of the h5-HT2A receptor.
This model was subsequently analyzed using PROCHECK and the Pro-
Table facility within SYBYL to assess the geometric integrity of vari-
ous structural elements (bond lengths, torsion angles, etc.) within
the receptor. The target compounds with explicit consideration of
stereoisomers were built and energy-minimized. These structures
were docked into the receptor binding site using GOLD as previously
described.7 Visual inspection of the docked poses in conjunction
with the ChemScore fitness function was used to select the final
solution for each isomer. The GOLD program was able to place the
5-HT2A ligands 2a and 2b in the previously-predicted AMDA-
antagonist binding site (Fig. 1).
for the dihydroanthracene core of 3-methoxy-AMDA. Like the ster-
eoisomers of 3-methoxy-AMDA, the additional methoxy group of
(R)-2b or (S)-2b may be accommodated in the binding pocket
through a slight shift in the position of the polycyclic core, while
the methoxy group may act as an H-bond acceptor for serine resi-
dues S1593.36 or S771.35/S1312.61, respectively (Fig. 1c and d). The
observation that the methoxy-substituted compound 2b showed
substantial affinity (Ki = 21 nM) suggests that the non-linear dihy-
drotetraphene analog 2a is a significant contributor to the ob-
served affinity of the 2a:1a mixture at 5-HT2A. Our homology
modeling and docking studies, however, indicate that the dihyd-
rotetracenes 1a and 1b may also be accommodated in, and have
substantial affinity for, the proposed binding site in a manner anal-
ogous to that of the dihydrotetraphenes 2a and 2b (ChemScores:
(R)-1a, 40.75; (S)-1a, 38.23; (R)-1b, 36.49; (S)-1b, 40.65; (R)-2a,
38.28; (S)-2a, 40.03; (R)-2b, 41.15; (S)-2b, 38.05).
The target compounds were synthesized as shown in Schemes 1
and 2. In the case of the annulated analogs (Scheme 1), commer-
cially available
a-bromo-o-tolunitrile 4, on treatment with DI-
In each solution, the ligand ammonium ion interacts with the
D1553.32 residue in the binding site. Compound 2a, with its
extended aromatic structure, exhibited extensive lipophilic inter-
actions with residues surrounding the dihydrotetraphene core
( Fig. 1a and b). The region of the binding site occupied by the
dihydrotetraphene core is the same as that previously proposed6
BAL-H in anhydrous CH2Cl2 and work up using aqueous HBr,
gave the aldehyde 5 in quantitative yields.24 Palladium-catalyzed
cross-coupling of 2-naphthylboronic acid or 6-methoxy-2-naph-
thylboronic acid with the halide 5 gave the coupled aldehydes 6a
and 6b. Cyanosilylation of the aldehydes 6a and 6b using TMSCN
gave the cyano trimethylsilyl ether as intermediates, which were
Figure 1. Proposed binding mode of isomers of annulated AMDA analogs (a) (R)-2a, (b) (S)-2a, (c) (R)-2b and (d) (S)-2b within the binding site of the 5-HT2A receptor.
Residues within 4 Å of the bound ligands are displayed. Hydrogen bonds are shown in magenta with donor–acceptor distances in Ångströms.

938
J. R. Shah et al. / Bioorg. Med. Chem. Lett. 20 (2010) 935–938
( )n
)
(
n
O
OH
O
NH
NH
a,b
c
a
n = 1
9a-d
3a-d
8
b n = 2
c
n = 3
d n = 4
Scheme 2. Reagents: (a) SOCl2, benzene; (b) phenylethylamine, phenylpropylamine or phenylbutylamine; (c) BH3ÁTHF.
3. Rashid, A.; Manivet, P.; Nishio, H.; Pratuangdejkul, J.; Rajab, M.; Ishiguro, M.;
Launay, J.-M.; Nagatomo, T. Life Sci. 2003, 73, 193.
4. Shapiro, D. A.; Kristiansen, K.; Kroeze, W. K.; Roth, B. L. Mol. Pharmacol. 2000,
58, 877.
5. Westkaemper, R. B.; Glennon, R. A. Curr. Top. Med. Chem. 2002, 2, 575.
6. Dewkar, G. K.; Peddi, S.; Mosier, P. D.; Roth, B. L.; Westkaemper, R. B. Bioorg.
Med. Chem. Lett. 2008, 18, 5268.
7. Runyon, S. P.; Mosier, P. D.; Roth, B. L.; Glennon, R. A.; Westkaemper, R. B. J.
Med. Chem. 2008, 51, 6808.
8. Runyon, S. P.; Peddi, S.; Savage, J. E.; Roth, B. L.; Glennon, R. A.; Westkaemper, R.
B. J. Med. Chem. 2002, 45, 1656.
9. Runyon, S. P.; Savage, J. E.; Taroua, M.; Roth, B. L.; Glennon, R. A.; Westkaemper,
R. B. Bioorg. Med. Chem. Lett. 2001, 11, 655.
10. Shah, J. R.; Mosier, P. D.; Roth, B. L.; Kellogg, G. E.; Westkaemper, R. B. Bioorg.
Med. Chem. 2009, 17, 6496.
11. Individual amino acid residues of the receptor are identified by the traditional
residue identifier indicating the residue’s position in the primary amino acid
sequence, followed by the general Ballesteros–Weinstein GPCR residue
reduced with LAH to give the respective amino alcohols 7a and 7b.
Cyclodehydration of amino alcohols using methanesulfonic acid
gave both the isomers of cyclized products 1a and 1b, and 2a
and 2b, respectively. The synthesis of 9-(N-benzylaminomethyl)-
9,10-dihydroanthracene (3a) was reported previously.9 The
phenylalkylamines 3b–d were prepared by reduction (BH3ÁTHF)
of amides obtained by the treatment of acid chloride with
phenylethyl-, phenylpropyl- and phenylbutylamine, respectively
(Scheme 2).
Binding assays and data analysis were performed through the
NIMH Psychoactive Drug Screening Program (PDSP) using cloned
human receptors. The 5-HT2A competitive binding assay employs
[3H]ketanserin (a 5-HT2A antagonist) as the radioligand, and the
H1 competitive binding assay employs [3H]pyrilamine (an H1
antagonist) as the radioligand. Binding data were analyzed using
PRISM (GraphPad Software, Inc., San Diego, CA). Details of the bind-
ing assay protocol may be found at the PDSP home page, http://
pdsp.med.unc.edu.
identifier as
a superscript. See: Ballesteros, J. A.; Weinstein, H. Methods
Neurosci. 1995, 25, 366.
12. Klabunde, T.; Giegerich, C.; Evers, A. J. Med. Chem. 2009, 52, 2923.
13. Trivedi, B. K.; Bristol, J. A.; Bruns, R. F.; Haleen, S. J.; Steffen, R. P. J. Med. Chem.
1988, 31, 271.
14. Scopes, D. I.; Barrio, J. R.; Leonard, N. J. Science 1977, 195, 296.
15. Ludeman, S. M.; Zon, G. J. Med. Chem. 1975, 18, 1251.
16. Chang-Fong, J.; Addo, J.; Dukat, M.; Smith, C.; Mitchell, N. A.; Herrick-Davis, K.;
Teitler, M.; Glennon, R. A. Bioorg. Med. Chem. Lett. 2002, 12, 155.
17. Büttner, F.; Bergemann, S.; Guénard, D.; Gust, R.; Seitz, G.; Thoret, S. Bioorg.
Med. Chem. 2005, 13, 3497.
Acknowledgments
This work was supported by United States Public Health Service
Grant R01-MH57969 (R.B.W.), R01-MH61887 (B.L.R.) and the
NIMH Psychoactive Drug Screening Program (B.L.R.).
18. Campiani, G.; Ramunno, A.; Fiorini, I.; Nacci, V.; Morelli, E.; Novellino, E.;
Goegan, M.; Mennini, T.; Sulivan, S.; Zisterer, D. M.; Williams, C. D. J. Med.
Chem. 2002, 45, 4276.
19. Chirapu, S. R.; Pachaiyappan, B.; Nural, H. F.; Cheng, X.; Yuan, H.; Lankin, D. C.;
Abdul-Hay, S. O.; Thatcher, G. R. J.; Shen, Y.; Kozikowski, A. P.; Petukhov, P. A.
Bioorg. Med. Chem. Lett. 2009, 19, 264.
Supplementary data
20. Bissantz, C.; Bernard, P.; Hibert, M.; Rognan, D. Proteins 2003, 50, 5.
21. Chenna, R.; Sugawara, H.; Koike, T.; Lopez, R.; Gibson, T. J.; Higgins, D. G.;
Thompson, J. D. Nucleic Acids Res. 2003, 31, 3497.
22. Fiser, A.; Šali, A.. In Methods in Enzymology: Macromolecular Crystallography;
Part, D., Carter, C. W. J., Sweet, R. M., Eds.; Academic Press: San Diego, 2003;
Vol. 374, pp 461–491.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2009.12.064.
References and notes
23. Jones, G.; Willett, P.; Glen, R. C.; Leach, A. R.; Taylor, R. J. Mol. Biol. 1997, 267,
727.
24. Zhang, X.-X.; Lippard, S. J. J. Org. Chem. 2000, 65, 5298.
1. Berger, M.; Gray, J. A.; Roth, B. L. Annu. Rev. Med. 2009, 60, 355.
2. Nichols, D. E.; Nichols, C. D. Chem. Rev. 2008, 108, 1614.