Azaflavones compared to flavones as ligands to the benzodiazepine binding site of brain GABAA receptors

Article abstract of DOI:10.1016/j.bmcl.2008.09.092

A series of azaflavone derivatives and analogues were prepared and evaluated for their affinity to the benzodiazepine binding site of the GABA_A receptor, and compared to their flavone counterparts. Three of the compounds, the azaflavones 9 and 12 as well as the new flavone 13, were also assayed on GABA_A receptor subtypes (α₁β₃γ_2s, α₂β₃γ_2s, α₄β₃γ_2s and α₅β₃γ_2s), displaying nanomolar affinities as well as selectivity for α1- versus α2- and α3-containing receptors by a factor of between 14 and 26.

Full text of DOI:10.1016/j.bmcl.2008.09.092

Bioorganic & Medicinal Chemistry Letters 18 (2008) 5713–5716
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Azaflavones compared to flavones as ligands to the benzodiazepine binding
site of brain GABAA receptors
Jakob Nilsson a, Elsebet Østergaard Nielsen b, Tommy Liljefors c, Mogens Nielsen c, Olov Sterner a,
*
a Department of Organic Chemistry, Lund University, P.O.B. 124, SE-221 00 Lund, Sweden
b NeuroSearch A/S, DK-2750 Ballerup, Denmark
c Department of Medicinal Chemistry, The Danish University of Pharmaceutical Sciences, Universitetsparken 2, DK-2100 Copenhagen, Denmark
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of azaflavone derivatives and analogues were prepared and evaluated for their affinity to the ben-
zodiazepine binding site of the GABAA receptor, and compared to their flavone counterparts. Three of the
compounds, the azaflavones 9 and 12 as well as the new flavone 13, were also assayed on GABAA receptor
Received 11 September 2008
Revised 24 September 2008
Accepted 26 September 2008
Available online 30 September 2008
subtypes (
a
1b3c2s
,
a
2b3c2s
3-containing receptors by a factor of between 14 and 26.
Ó 2008 Elsevier Ltd. All rights reserved.
, a4b3c2s and a5b3c2s), displaying nanomolar affinities as well as selectivity for
a1- versus
a
2- and a
Keywords:
Flavone
Azaflavone GABAA benzodiazepine binding
site
GABA (
c
-aminobutyric acid) is the major inhibitory neurotrans-
types of ligands has improved the potency considerably,9,10 com-
pounds 12 and 13 were also prepared and included in this study.
The azaflavones 9, 10 and 12 as well as the 2-arylquinolines 16
and 17 were prepared according to the procedure shown in
Scheme 1. 2-Arylquinolones 9 and 10 were synthesized in a four-
step procedure starting with an acylation of toluidine and N-meth-
yltoluidine under Sugasawa conditions to give 3 and 4.11 The
amides 5 and 6 were generated by reaction with 5-bromo-2-
methoxybenzoyl chloride prepared from 5-bromo-2-methoxy-
benzaldehyde following a previously published protocol.8
Cyclization of 5 and 6 in presence of potassium tert-butoxide
followed by a demethylation with boron tribromide gave quino-
lones 9 and 10.12 Successive treatment of 7 with phosphorus
oxychloride, boron tribromide and morpholine gave 16. In a
Suzuki-Miyaura cross-coupling reaction with B-benzyl-9-borabicy-
clo[3.3.1]nonane, the bromo atom of 9, 16 and 11 was substituted
by a benzyl group.13 Condensation of ethyl benzoyl acetate and
4-ethylaniline was carried out neat in the presence of a catalytic
amount of Sc(OTf)3 to give b-enamine 19 (see Scheme 2),14 and
subsequent cyclization of the b-enamine in diphenyl ether at
250 °C gave quinolone 20.
mitter in the central nervous system.1 The GABAA receptor is a
chloride ion channel complex, consisting of five subunits from
eight different classes with multiple isoforms (
a1-6, b1-4
,
c
1-4, d,
e,
p
b,
, h and
q
1-3).2,3 The most abundant GABAA receptor contains
a,
c subunits in a 2:2:1 stoichiometry, and receptors with different
subtype composition are associated with different physiological ef-
fects. While 1-containing receptors are implicated in sedation and
anterograde amnesia, 2- and/or 3-containing receptors appear to
a
a
a
be important for anxiolytic activity.4,5 The GABAA receptor has sev-
eral allosteric modulatory sites, of which the one for benzodiaze-
pines (BZDs) has attracted most attention. A pharmacophore
model comprising agonists, inverse agonists, and antagonists for
this binding site was proposed in 19956 and further developed in
a recent study with synthetic flavones,7,8 resulting in the potent
50-bromo-20-hydroxy-6-methylflavone 11 (Ki = 0.9 nM). In Figure
1, 11 is displayed in the pharmacophore model, and it should be
noted that flavones only can interact as a hydrogen bond acceptor
with H2 of the H2/A3 hydrogen bond donating/accepting site. Aza
analogues of flavones, for example, compounds 9 and 12, could
shed light on the difference between a hydrogen bond acceptor
and donor by interacting with A3, and thereby provide valuable
information for the pharmacophore model. Compounds 16 and
17 were also considered to be of interest, as they, just as the flav-
ones, can interact with H2. In addition, as the introduction of a
benzyl group in the position facing the interface region for other
The affinity for the benzodiazepine binding site of the GABAA
receptor was determined by displacement of 3H-flumazenil in
rat cortical tissue as previously described.7 The results are shown
in Table 1. The subtype selectivity of some of the most potent
compounds was tested by assaying their ability to displace 3H-
flumazenil in membranes from HEK293 cells expressing rat
a1b3c2, a2b3c2, a3b3c2 and a5b3c2 GABAA receptor subtypes
* Corresponding author.
(Table 2).
E-mail address: Olov.Sterner@organic.lu.se (O. Sterner).
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.09.092

5714
J. Nilsson et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5713–5716
Figure 1. Binding mode of (a) the high affinity flavonoid 50-bromo-20-hydroxy-6-methyl flavone (11) and (b) the 2-aryl-quinolone 12, in the pharmacophore model discussed
in the text. H1 is a H-bond donor site, A2 is a H-bond acceptor site, H2/A3 is a site that both accepts and donates H-bonds, L1, L2 and L3 represents lipophilic pockets while S1,
S2, S3, S4 and S5 denote regions of steric repulsive ligand–receptor interaction (or receptor essential volume).
In general, the azaflavones appear to be less potent than the
than that of their corresponding flavone analogues 11,8 13 and 21.8
flavones. The affinities of 9, 12 and 20 are 24, 3 and 7 times lower
It has been postulated that BZDR ligands require the ability to
R
O
O
NH
a
b
R
N
R
N
H
Br
O
1: R = H
2: R = Me
3: R = H
4: R = Me
5: R = H
6: R = Me
MeO
c
O
X
O
O
OH
OMe
Br
OH
Br
e
d
X
X
9: X = NH
11: X = O
12: X = NH
13: X = O
9: X = NH
10: X = NMe
7: X = NH
8: X = NMe
Ph
O
Cl
Cl
N
OMe
Br
f
OMe
Br
OH
Br
d
N
H
N
14
7
15
g
O
N
O
N
OH
OH
Br
e
N
N
Ph
17
16
Scheme 1. Reagents and conditions: (a) AlCl3, BCl3, MeCN, toluene, reflux, 2 h, then HCl (1 M), 80 °C, 30 min (yield 48% for 3 and 30% for 4); (b) 5-bromo-2-methylbenzoyl
chloride, Et3N, THF, rt, 18 h (yield 82% for 5 and 84% for 6); (c) KOtBu, tBuOH, 70 °C, 16 h (yield 86% for 7 and 100% for 8); (d) BBr3, CH2Cl2, 40 °C, 8 h (yield 79% for 9, 95% for 10
and 89% for 15); (e) K3PO4, Pd(OAc)2, S-Phos, B-Bn-9-BBN, DMF, 60 °C, 16 h (yield 67% for 12, 57% for 13 and 85% for 17); (f) POCl3, 90 °C, 1 h (yield 95%); (g) DIPEA,
morpholine, DMF, 90 °C, 1 h (yield 81%).

J. Nilsson et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5713–5716
5715
O
OEt
O
O
O
H
a
b
N
OEt
Ph
N
H
18
19
20
Scheme 2. Reagents and conditions: (a) Sc(OTf)3, 4-ethylaniline (yield 77%); (b) diphenyl ether, reflux, 30 min (yield 63%).
H2/A3 hydrogen acceptor/donor does not occupy the exactly same
space or bind ligands with the same binding angle. In particular,
the H2 interacting ligands could be somewhat tilted compared to
the A3 interacting ligands. This would imply that a SAR study can-
not in a straight forward fashion be translated between the H2 and
A3 interactive benzodiazepine analogues. However, it is interesting
to note that this general trend is not followed by 7, which is more
potent than its flavone analogue 22,8 indicating that a hydrogen
bond from the ligand to A3 is as important as one from H2 to the
ligand in compounds of comparable planarity. An interesting
observation is that the shift from 50-bromo- to 50-benzyl substitu-
tion does not significantly affect the affinity in the flavone series
(11 and 13), whereas a 10-fold increase in affinity is observed in
the azaflavone series (9 and 12). For the two arylquinolines 16
and 17, it is obvious that the potency is considerably lower com-
pared to the corresponding flavones and azaflavones. In other
types of BZDR ligands it has been shown that the interaction of ring
systems similar to the morpholine group with H1 is acceptable,17
excluding steric interactions as the cause for the low affinity. In-
stead, the strong hydrogen bond between the 20-hydroxyl group
and the quinoline nitrogen (considerably stronger than that pres-
ent in the 20-hydroxylflavones, ꢀ22 kJ/mol for 16 compared to
1.4 kJ/mol for 11) will hamper compounds 16 and 17 from adopt-
ing the active conformation.
Table 1
Ki values of flavone analogues tested on 3H-flumazenil binding in vitro to rat cortical
membranes
O
N
N
O
X
R1
R2
R3
OH
R3
16, 17
7, 9, 10, 11, 12, 13, 20, 21, 22
Compound
R1
R2
R3
X
Ki valuea (nM)
7
9
–CH3
–CH3
–CH3
–CH3
–CH3
–CH3
—
–OMe
–OH
–OH
–OH
–OH
–OH
—
—
–H
–H
–OMe
–Br
–Br
–Br
–Br
–CH2C6H5
–CH2C6H5
–Br
–CH2C6H5
–H
NH
NH
NMe
O
NH
O
—
—
NH
O
O
730 130
22
4
10
118
12
13
16
17
20
218
228
1100 310
0.9 0.2
2.0 0.3
0.6 0.3
840 55
400 68
1200 260
180 40
>1500
—
–CH2CH3
–CH2CH3
–CH3
–H
–Br
Subtype affinity testing was performed with compounds 9, 12
and 13 on recombinant
receptor subtypes (Table 2). All compound investigated in this
study display selectivity for 1b3c2s over the other receptor sub-
a1b3c2s, a2b3c2s, a4b3c2s and a5b3c2s
a
Each Ki value is mean SD of three determinations.
a
types. Interestingly, the substitution of a bromo to a benzyl group
in the 6-position of the aryl quinolone (9–12) resulted in 5 times
higher a2/a1 Ki ratio and 8 times higher a3/a1 Ki ratio as well as
an increased affinity, making the highly subtype selective deriva-
tives 12 and 13 valuable for the development of a subtype specific
pharmacophore model.
Table 2
The affinity of selected flavone analogues tested on 3H-flumazenil binding to
2b3c2s 3b3c2s and 5b3c2s GABAA receptor subtypes
a
1b3c2s
,
a
,
a
a
a
a
a
a
Compound
Ki a1 (nM)
Ki a2 (nM)
Ki a3 (nM)
Ki a5 (nM)
9
12
13
39
1.2 0.4
0.80 0.25
8
120 34
18 2.9
99 20
31 7.8
nd
7.3 0.8
2.4 1.1
19
9
11
6
a
Acknowledgements
Each Ki value is mean SD of three determinations. nd, not determined.
Financial support from the Swedish Board for Scientific Re-
search (VR), the KAW foundation, the Research School for Pharma-
ceutical Sciences at Lund University, and the NeuroScience
PharmaBiotec Research Center, Denmark, is gratefully acknowl-
edged. M.N. was supported by the Carlsberg Foundation.
adopt a planar or close to planar arrangement of the ring systems
for an efficient binding,6 and the 20-hydroxyl group constitutes a
sterical hindrance for the adoption of a coplanar conformation
among all the potent analogues tested in this study. The energy dif-
ference between the coplanar conformation and the lowest energy
conformation, with a twisted conformation, was calculated to
13 kJ/mol for flavone 13, implying that 20-hydroxyl substituted
flavones are unlikely to adopt a planar conformation upon binding.
Conformational analyses were performed by Macromodel (version
9.5),15 and force field calculation were undertaken using MMFFs in
gas phase.16 Instead, the biologically active conformation is proba-
bly somewhat twisted. For the azaflavones, the additional sterical
hindrance between the N–H and the aryl group makes a planar
conformation even less probable (35.7 kJ/mol difference between
planar and most stable conformation for 12, and 23 kJ/mol for
compound 20). The N-methylated 10 is consequently considerably
less potent. A reasonable interpretation of the SAR data is that the
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2008.09.092.
References and notes
1. Sieghart, W. Pharmacology 2006, 54, 231.
2. Chebib, M.; Johnston, G. A. R. J. Med. Chem. 2000, 43, 1427.
3. Johnston, G. A. R. Curr. Pharm. Des. 2005, 11, 1867.
4. Rudolph, U.; Crestani, F.; Benke, D.; Brunig, I.; Benson, J. A.; Fritschy, J. M.;
Martin, J. R.; Bluethmann, H.; Mohler, H. Nature 1999, 401, 796.
5. Rudolph, U.; Crestani, F.; Möhler, H. Trends Pharmacol. Sci. 2001, 22, 188.
6. Zhang, W.; Koehler, K.; Zhang, P.; Cook, J. Drug Des. Discov. 1995, 12, 193.

5716
J. Nilsson et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5713–5716
7. Dekermendijan, K.; Kahnberg, P.; Witt, M.; Sterner, O.; Nielsen, M.; Liljefors, T. J.
Med. Chem. 1999, 42, 4343.
12. Li, L.; Wang, H.-K.; Kuo, S.-C.; Wu, T.-S.; Lednicer, D.; Lin, C. M.; Hamel, E.; Lee,
K.-H. J. Med. Chem. 1994, 37, 1126.
8. Kahnberg, P.; Lager, E.; Rosenberg, C.; Schougaard, J.; Camet, L.; Sterner,
O.; Østergaard Nielsen, E.; Nielsen, M.; Liljefors, T. J. Med. Chem. 2002,
45, 4188.
13. Flaherty, A.; Trunkfield, A.; Barton, W. Org. Lett. 2005, 7, 4975.
14. Yadav, J. S.; Kumar, V. N.; Rao, R. S.; Priyadarshini, A. D.; Rao, P. P.; Reddy, B. V.
S.; Nagaiah, K. J. Mol. Catal. A: Chem. 2006, 256, 234.
9. Albaugh, P.; Marshall, L.; Gregory, J.; White, G.; Hutchison, A.; Ross, P.;
Gallagher, D.; Tallman, J.; Crago, M.; Cassella, J. J. Med. Chem. 2002, 45,
5043.
15. Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Liskamp, R.; Lipton, M.; Caulfield,
C.; Chang, G.; Hendrickson, T.; Still, W. C. J. Comput. Chem. 1990, 11, 440.
16. Halgren, T. A. J. Comput. Chem. 1999, 29, 720.
10. Lager, E.; Andersson, P.; Nilsson, J.; Pettersson, I.; Østergaard Nielsen, E.;
Nielsen, M.; Sterner, O.; Liljefors, T. J. Med. Chem. 2006, 49, 2526.
11. Prasad, K.; Lee, G. T.; Chaudhary, A.; Girgis, M. J.; Streemke, J. W.; Repic, O. Org.
Process Res. Dev. 2003, 7, 723.
17. Andersson, K. E.; Lundt, B. F.; Jorgensen, A. S.; Braestrup, C. Eur. J. Med. Chem.
1996, 31, 417.
18. (Given in Supplementary Material) Anwer, B.; Sged, S. A.; Tanveer, A. F.
Tetrahedron Lett. 1976, 36, 3217.