Novel 17β-substituted conformationally constrained neurosteroids that modulate GABAA receptors

Article abstract of DOI:10.1021/jm050271q

The goal of this study was to develop a series of allopregnanolone analogues substituted by conformationally constrained 17β side chains to obtain additional information about the structure-activity relationship of Set-reduced steroids to modulate GABA

Full text of DOI:10.1021/jm050271q

J. Med. Chem. 2005, 48, 5203-5214
5203
Novel 17â-Substituted Conformationally Constrained Neurosteroids that
Modulate GABAA Receptors
Charikleia Souli,† Nicolaos Avlonitis,† Theodora Calogeropoulou,*,† Andrew Tsotinis,‡ Ga´bor Maksay,§
T´ımea B´ıro´,§ Aggeliki Politi,† Thomas Mavromoustakos,† Alexandros Makriyannis, Heribert Reis,† and
Manthos Papadopoulos†
Institute of Organic and Pharmaceutical Chemistry, National Hellenic Research Foundation, 48 Vassileos Constantinou
Avenue, 11635 Athens, Greece, Department of Pharmaceutical Chemistry, Faculty of Pharmacy, University of Athens,
Panepistimioupoli-Zografou, 15771 Athens, Greece, Department of Molecular Pharmacology, Institute of Biomolecular
Chemistry, Chemical Research Center, Hungarian Academy of Sciences, H-1525, Budapest, POB 17, Hungary, and
Center for Drug Discovery, Northeastern University, 360 Huntington Avenue, 116 Mugar Hall, Boston, Massachusetts 02115
Received March 24, 2005
The goal of this study was to develop a series of allopregnanolone analogues substituted by
conformationally constrained 17â side chains to obtain additional information about the
structure-activity relationship of 5R-reduced steroids to modulate GABAA receptors. Specif-
ically, we introduced alkynyl-substituted 17â side chains in which the triple bond is either
directly attached to the 17â-position or to the 21-position of the steroid skeleton. Furthermore,
we investigated the effects of C22 and C20 modification. The in vitro binding affinity for the
GABAA receptor of the new analogues was measured by allosteric displacement of the specific
binding of [3H]4′-ethynyl-4-n-propyl-bicycloorthobenzoate (EBOB) to GABAA receptors on
synaptosomal membranes of rat cerebellum. An allosteric binding model that has been
successfully applied to ionotropic glycine receptors was employed. The most active derivative
is (20R)-17â-(1-hydroxy-2,3-butadienyl)-5R-androstane-3-ol (20), which possesses low nanomolar
potency to modulate cerebellar GABAA receptors and is 71 times more active than the control
compound allopregnanolone. Theoretical conformational analysis was employed in an attempt
to correlate the in vitro results with the active conformations of the most potent of the new
analogues.
Introduction
with this receptor is enantioselective5 and is dependent
on subunit composition.6 Potentiation by the neuroactive
steroids alphaxalone and allotetrahydrodeoxycorti-
costerone (alloTHDOC) requires regions before the
second and beyond the fourth transmembrane segments
of the R1 subunit, respectively.7 The presence of a γ
subunit is not required, unlike the benzodiazepine site
where this is a prerequisite.8 Recent evidence has
implicated the enhanced potentiation of δ subunit-
containing GABAA receptors.8c,9 However, a study on the
neurosteroid effects on recombinant human R4â3δ
GABAA receptors concluded that there is not a single
structural motif within steroid molecules that would
account for the different effects on GABAA receptor
subtypes, a finding that may be consistent with steroid
interactions at multiple sites on the receptor.9a The
physiological and pharmacological actions of neuroster-
oids are topics of widespread interest. Neurosteroids
have been implicated in premenstrual syndrome,10
cognitive and psychiatric dysfunctions, neuroprotec-
tion,11 and GABAA receptor plasticity.12 Withdrawal
from chronic exogenous neurosteroids may be associated
with increased seizure susceptibility due to alterations
in the expression of GABAA receptor subunits.13 The
neurosteroids allopregnanolone and alloTHDOC act as
efficacious allosteric agents of GABAA receptors and
potentiate their Cl- ionophore function.14 They have
been shown to be potent anticonvulsants, anxiolytics,
and antistress agents and have been shown to possess
sedative, hypnotic, and anesthetic activities.15 Preg-
Neurosteroids are synthesized in the central and
peripheral nervous system, particularly in myelinating
glial cells, but also in astrocytes and neurons, and act
on the nervous system.1 Research over the past decade
has elucidated their multiple effects on various neu-
rotransmitter systems.2 Many studies have demon-
strated that they positively or negatively modulate the
function of members of the ligand-gated ion channel
superfamily,3 and most of these studies have focused
on the positive allosteric actions on GABAA receptors.4
GABA is the major inhibitory neurotransmitter in the
mammalian central nervous system, and rapid synaptic
inhibition is mediated through the opening of GABAA
receptor-ionophores. A number of therapeutically im-
portant drugs, including benzodiazepines and sedative
and anesthetic barbiturates, act to enhance the interac-
tion of GABA with this receptor. The GABAA receptor
complex can exist in multiple isoforms and can demon-
strate a variety of pharmacological profiles that arise
from their pentameric structure and the diversity of
their subunits. It has been suggested that, as for other
allosteric modulators, there may be specific neurosteroid
binding sites on GABAA receptors, a theory supported
by previous reports that the interaction of neurosteroids
* To whom correspondence should be addressed: Tel: +3010
7273833. Fax: +3010 7273818. E-mail: tcalog@eie.gr.
† National Hellenic Research Foundation.
‡ University of Athens.
§ Hungarian Academy of Sciences.
Northeastern University.
10.1021/jm050271q CCC: $30.25 © 2005 American Chemical Society
Published on Web 07/20/2005

5204 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16
Souli et al.
nenolone sulfate and dehydroepiandrosterone sulfate,
however, have been shown to inhibit GABAA receptor-
ionophore function4b,16 and enhance memory perfor-
mance in rodents.17
for the synthesis of the 17â-alkynyl-substituted deriva-
tives is described in Scheme 1. Thus, regioselective and
stereoselective reduction of the carbonyl group at C3 of
5R-3,20-pregnanedione (1) with K-Selectride in THF
at -78 °C gave 3R-hydroxy-5R-pregnane-20-one. Protec-
tion of the hydroxyl group as the tert-butyldiphenylsilyl
ether and treatment with NaOBr afforded the corre-
sponding acid 2. The reduction of the acid 2 to the
respective alcohol using LiAlH4 and the subsequent
oxidation with pyridinium chlorochromate afforded the
17â-formyl derivative 3, which in turn was transformed
to the dibromo compound 4 upon treatment with CBr4
and PPh3 using the Corey-Fuchs methodology.28 The
addition of BuLi in THF at -78 °C to gem-dibromide 4
afforded alkyne 5. Sonogashira coupling of alkyne 5 with
4-iodotoluene, using (PPh3)2PdCl2 as a catalyst in the
presence of CuI and pyrrolidine,29 followed by the
deprotection of the C3-hydroxyl group using TBAF in
THF, yielded analogue 6. Furthermore, the treatment
of 4 with TBAF in THF resulted in alkynyl bromide 7.
The addition of n-BuLi to gem-dibromide 4 at -78 °C
and the quenching of the resulting alkynyllithium
intermediate with acetaldehyde afforded the corre-
sponding alkynyl alcohol 8, which, upon deprotection
of the C3-hydroxyl, yielded derivative 9.
The synthesis of the 21-alkynyl-substituted deriva-
tives is described in Schemes 2-4. The addition of
lithium trimethylsilylacetylide to the aldehyde 3 af-
forded a mixture of the epimeric alcohols, 10a (20R) and
10b (20S) in a ratio of 3 to 1, respectively, which were
easily separated by flash column chromatography
(Scheme 2). The less polar diastereomer 10a was
attributed the 20R absolute configuration on the basis
of Cram’s rules30 and previous analogous reactions with
3â-hydroxy-17â-formylandrost-5-ene.31
To obtain compounds 10a and 10b stereoselectively,
we employed another route starting from the acid 2
(Scheme 2). Thus, the formation of the corresponding
Weinreb amide32 and the subsequent treatment with
lithium triisopropylsilylacetylide afforded the alkynyl
ketone 11. The asymmetric reduction of ketone 11 with
catecholborane in the presence of (S)-3,3-diphenyl-1-
ethyltetrahydro-3H-pyrrolo[1,2-c][1,3,2]oxazaborol as a
catalyst33 afforded the 20S epimeric alcohol 12b as the
only product. Alternatively, the stereoselective re-
duction of ketone 11 with BH3-Me2S in the presence of
(S)-3,3-diphenyl-1-ethyltetrahydro-3H-pyrrolo[1,2-c][1,3,2]-
oxazaborol as a catalyst34 afforded the 20R epimeric
alcohol 12a as the only product, and reduction with
NaBH4 afforded a 1:1 mixture of epimeric alcohols 12a
and 12b. The treatment of the alkynyl ketone 11 with
a hydrogen fluoride-pyridine complex in methylene
chloride resulted in the removal of the tert-butyldi-
phenylsilyl protecting group as well as the C22-triiso-
propylsilyl group to afford the ynone 13. The synthesis
of derivatives 14-16 and 18 is depicted in Scheme 3.
Thus, the removal of both silyl groups in compound 12a
with TBAF in THF yielded diol 14. The formation of
the allene derivative 15 was effected by the reaction of
14 with paraformaldehyde, CuI, and diisopropylamine
in refluxing dioxane35 (compound 15 has a 20S config-
uration because of a change in the priorities of the C20
substituents). Furthermore, alcohol 12a was treated
with one equivalent of sodium hydride, followed by the
Previous in vitro and in vivo structure-activity
relationship (SAR) studies indicated that although
steroidal positive allosteric GABAA receptor modulators
must possess a 3R-hydroxylated A ring and a keto group
at C20 of the 17â-acetyl side chain, certain other 17â
substituents that are hydrogen-bond acceptors, such as
17â-CN, retain activity.5,18 A number of modifications
to the steroid nucleus have been made to study the SAR
of neuroactive steroids, including substitutions at the
3â-, 11-, 17-, or 21-positions19 as well as 17a-aza-D-
homosteroid analogues20 and benz[e]indene deriva-
tives.21 In an attempt to identify steroids with improved
activity against GABAA receptors, we introduced al-
kynyl-substituted 17â side chains. Thus, in the present
report we describe the synthesis and activity of two
series of 3R-hydroxy-substituted steroids in which the
triple bond is directly attached to either the 17â-position
or the 21-position of the steroid skeleton. Furthermore,
we investigated the effects of C22 and C20 modification.
In vitro binding studies of the SAR of neuroactive
steroids on GABAA receptor-ionophores have frequently
used [35S]tert-butylbicyclophosphorothionate (TBPS),
which is a cage convulsant.14,18,19 The channel-blocking
radioligands [35S]TBPS, [3H]tert-butylbicycloorthoben-
zoate (TBOB), and [3H]4′-ethynyl-4-n-propyl-bicyclo-
orthobenzoate (EBOB)22 bind within the ion channel.23
Consequently, these binding studies have revealed
several allosteric binding interactions of pharmacologi-
cal relevance.14,18,24,25 Here, we use an allosteric binding
model26 that has been successfully applied to glycine
receptors belonging to the same superfamily of ionotro-
pic receptors.27 GABAA receptors of rat cerebellum were
used for binding because (1) their subunit composition
has been clarified: R1âγ2, R1R6âγ2, R6âγ2, R6âδ, and
R1R6âδ6d and (2) the presence of δ subunits possibly
confers high affinity to neuroactive steroids.8c,9a This
report provides a description of the SAR for the interac-
tion of the GABAA receptor with new pregnane steroids.
Chemistry
The synthetic strategy followed for the preparation
of the neurosteroid analogues 6, 7, 9, 13-16, 18-21,
and 23 is depicted in Schemes 1-4. The synthetic route

Novel Neurosteroids that Modulate GABAA Receptors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16 5205
Scheme 1a
a
(i) K-Selectride, THF, -78 °C; (ii) TBDPSiCl, DMF, imidazole, 55 °C; (iii) NaOBr, dioxane, H2O, rt; (iv) LiAlH4, THF, reflux; (v)
PCC, CH2Cl2, rt; (vi) CBr4, PPh3, CH2Cl2, 0 °C; (vii) n-BuLi, THF, -78 °C; (viii) (PPh3)2PdCl2, CuI, pyrrolidine, 4-iodotoluene, rt; (ix)
TBAF, THF, rt; (x) n-BuLi, CH3CHO, THF, -78 °C.
addition of methyl iodide to yield the respective 20R-
methoxy derivative, which, upon the addition of TBAF
in THF, yielded the 3R-hydroxy-(20R)-methoxy alkynyl
derivative 16. The selective removal of the C22-triiso-
propylsilyl group in compound 12a, using 1.5 equiv of
TBAF in THF, yielded alkyne 17. The Sonogashira
coupling of alkyne 17 with 4-iodoanisole in the presence
of (Ph3P)4Pd, CuI, and pyrrolidine, followed by treat-
ment with TBAF in THF, yielded diol 18. Reactions
analogous to those described in Scheme 3 using the
(20S)-alcohol 12b afforded the steroid analogues 19, 20,
21, and 23 (Scheme 4).
pounds 6, 7, and 9. Among these analogues, only the
22-bromoalkynyl derivative 7 had a potency similar to
the control compound allopregnanolone as a displacer
(Ks ) 350 ( 72 and 198 ( 47 nM, respectively).
Substitution of the triple bond at C22 by a tolyl group
is deleterious to GABAA receptor activity, rendering the
resulting compound 6 20 times less active (Ks ) 4100
( 265 nM). Analogue 9, substituted by a C22-hydroxy-
ethyl group, is ∼8 times less active (Ks ) 946 ( 404
nM) than the prototype, 5R-reduced steroid allopreg-
nanolone. We then set out to explore the activity of
compounds containing a triple bond at position C21 of
the pregnane system in conjunction with either a keto
function, a hydroxyl function, or an ether function at
C20. The replacement of the C21 methyl group in
allopregnanolone by an alkynyl functionality, analogue
13, resulted in a 4-fold reduction of activity (844 ( 205
nM). As mentioned previously, GABAAergic steroids
contain a hydrogen-bond-acceptor group located at C17.
Even though adding a triple bond in conjugation with
the carbonyl group should improve its hydrogen-bond-
acceptor properties to a small extent, we observe a
Results and Discussion
The results of the allosteric displacement of [3H]-
EBOB binding from GABAA receptors on synaptosomal
membranes of rat cerebellum for the compounds under
study are reported in Table 1.
Because 3R-hydroxy-5R-androstane-17â-carbonitrile
was reported to produce great potentiation of GABAA
receptor function,5 we prepared 17â-alkynyl derivatives
of 3R-hydroxy-5R-androstane and, specifically, com-

5206 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16
Souli et al.
Scheme 2a
a
(i) n-BuLi, trimethylsilylacetylene, THF, -78 °C; (ii) CH3ONH(CH3)‚HCl, DMAP, EDCI, CH2Cl2, rt; (iii) n-BuLi, (triisopropylsilyl)-
acetylene, THF, -78 °C; (iv) (S)-3,3-diphenyl-1-ethyltetrahydro-3H-pyrrolo[1,2-c][1,3,2]oxazaborol, catecholborane, CH2Cl2, -78 °C; (v)
(S)-3,3-diphenyl-1-ethyltetrahydro-3H-pyrrolo[1,2-c][1,3,2]oxazaborol, BMS, CH2Cl2, -78 °C; (vi) C5H5N‚(HF)x, CH2Cl2, rt.
Scheme 3a
Scheme 4a
a
(i) TBAF, THF, rt; (ii) formaldehyde, CuI, diisopropylamine,
a
(i) TBAF, THF, rt; (ii) formaldehyde, CuI, diisopropylamine,
dioxane, reflux; (iii) NaH, MeI, THF, rt; (iv) 1.5 equiv of TBAF,
THF, rt; (v) 4-iodoanisole, (PPh3)4Pd, CuI, pyrrolidine, rt.
dioxane, reflux; (iii) NaH, MeI, THF, rt; (iv) 1.5 equiv of TBAF,
THF, rt; (v) 4-iodoanisole, (PPh3)4Pd, CuI, pyrrolidine, rt.
decrease in displacing activity. Therefore, we speculate
that the decrease in activity in analogue 13 is not
related to the hydrogen-bond-accepting capacity of the
C20 keto function and that the alkynyl group is un-
favorable for interaction with the steroid site of the
receptor. Additionally, we examined the activity of the
C22-alkynyl-substituted pregnanediols. Because we were
interested in investigating the impact of the orientation
(R or â) of the reduced substituent at the C20 position
on the modulating activity of the GABAA receptor, we
stereoselectively synthesized both C20 epimers. Inter-
estingly, alkynyl pregnanediols possess higher displac-
ing potency than ynone 13. Furthermore, the activity
was also dependent on the C20 stereochemistry, with
the 20S isomer 19 being 3.4 times more active (Ks )
145 ( 49 nM) than the corresponding 20R epimer 14
(Ks ) 494 ( 158 nM) and slightly more potent than
allopregnanolone. Methylation of the free hydroxyl
group in compounds 14 and 19 adversely affects the
potency of the corresponding analogues 16 and 21,
which possess Ks ) 924 ( 490 and 802 ( 227 nM,
respectively. The hydroxyl functionality in analogues 14
and 19 can interact with the receptor, both as a
hydrogen-bond acceptor and as a hydrogen-bond donor,
whereas the methyl ethers of 16 and 21 are only
hydrogen-bond-accepting substituents. It has been re-
ported previously that 5R-androstane-3R,17â-diol is a
relatively poor GABAA receptor modulator, whereas
17â-methoxy-5R-androstane-3R-ol, which possesses a
hydrogen-bond-accepting substituent at C17, is an ac-
tive, positive allosteric modulator.19f Therefore, hydro-
gen bonding is not the only parameter that controls the
affinity of the receptor for C22-alkynyl pregnanediols.
Furthermore, the substitution of the alkynyl functional-
ity at C22 in alcohols 18 and 23 by a 4-methoxyphenyl
substituent results in a more pronounced decrease in

Novel Neurosteroids that Modulate GABAA Receptors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16 5207
Table 1. Allosteric Displacement of [3H]EBOB Binding to
nanolone. Furthermore, these pregnanediols were shown
to be behaviorally active in the rat Vogel test,36 which
is a paradigm that is predictive of anxiolytic activity,
and the 20R-ol shows good selectivity for anxiolytic
activity versus motor impairment.
Additionally, 3R-hydroxy-pregnan-20-ols have partial
agonist activity at GABAA receptors.37 Functional data
reported by Belelli et al.25a confirm previous bio-
chemical,25b neurochemical,37a and behavioral studies36b
that demonstrate that certain endogenous pregnanediols
produce a more subtle modulation of GABAA receptor
function compared with that of anesthetic steroids.
Synaptosomal Membranes of Rat Cerebelluma
compound
KS (nM)
R
6 (n ) 3)
4100 ( 265
350 ( 72
2.4 ( 0.1
5.8 ( 1.1
4.3 ( 0.4
5.3 ( 1.4
3.3 ( 0.7
2.6 ( 0.1
3.0 ( 0.3
.10b
7 (n ) 3)
9 (n ) 3)
946 ( 404
844 ( 205
494 ( 158
810 ( 230
924 ( 490
1640 ( 234
145 ( 49
13 (n ) 5)
14 (n ) 5)
15 (n ) 6)
16 (n ) 6)
18 (n ) 3)
19 (n ) 5)
20c
2.6 ( 0.3
1.7 ( 0.1
4.0 ( 0.5
>10b
2.8 ( 1.2
21 (n ) 5)
23 (n ) 5)
allopregnanolone (n ) 4)
802 ( 227
870 ( 292
198 ( 47
As an initial step to correlate the in vitro results with
the active conformations of pregnanediols 14, 19, 15,
and 20, a combination of minimization algorithms with
stochastic and systematic conformational analysis was
employed assuming a normal chair conformation for
ring A.5 As shown in Figure 2, only one lowest-energy
conformer was found for each compound. The two
diastereomers, 15 and 20, differ in the orientation of
both the allene and the hydroxyl groups, whereas
compounds 14 and 19 differ only in the orientation of
the hydroxyl group. These differences are better ob-
served in Figure 3 in which superimposition of the
lowest-energy conformers of the two diastereomeric
pairs is shown. According to previous reports,19f,20
steroids containing hydrogen-bond-accepting groups at
C17 above the plane defined by the D-ring on the â side
of the steroid exhibit improved GABAA receptor activity.
For the four pregnanediols under study, the C20 hy-
droxyl group, which could serve as a hydrogen-bond
acceptor, lies above the plane of the D-ring, in the global
minimum-energy conformers. This is also the case for
the majority of the conformers with energies up to 6
kcal/mol higher than the global minimum. This suggests
that the direction adopted by the hydroxyl group is not
the determining factor for optimizing the interactions
of pregnanediols with their receptor binding site.
To explore the conformational space, which might
influence productive binding to the receptor, the contour
plots for alkynyl diols 14 and 19 and allenyl diols 15
and 20 were obtained, and they are depicted in Figure
4. Each grad scan analysis for the molecules under study
was performed around the τ1 [C(17)-C(20)-C(21)-
C(22)] and τ2 [C(17)-C(20)-O(20)-H] dihedral angles
by rotating each angle in increments of 10°. As can be
seen in Figure 4, for the two active compounds 19 and
20, τ1 and τ2 occupy common low-energy conformational
space. The same is observed for the other pair of
compounds, 14 and 15, which possess lower activity.
>10b
a Dissociation constants (KS) of the neuroactive steroids and
cooperativity factors (R) with [3H]EBOB binding were determined
via eq 1 and the ternary allosteric model (Scheme 5). Data are
the mean ((SEM) of n experiments. b Cannot be determined
accurately because full displacement approached nonspecific
[3H]EBOB binding. c Determined from fitting to the pooled data
of five experiments. Some experiments could not be fitted sepa-
rately.
Figure 1. Displacement of [3H]EBOB binding to rat cerebellar
GABAA receptors by the epimeric neurosteroids 20 (O) and 15
(b). Specific binding in the presence of the steroids (BES) over
the control (BE). Points are mean ( SEM of 3-5 experiments.
Curve fitting parameters are summarized in Table 1.
activity (3.3- and 6-fold, respectively; compound 18: Ks
) 1640 ( 234 nM and compound 23: Ks ) 870 ( 292
nM). The replacement of the alkynyl group in alcohols
18 and 23 by a propadienyl functionality affected the
displacing potency of the respective compounds 15 and
20 very dramatically. In analogue 15, there is a 1.6-
fold decrease in affinity (Ks ) 810 ( 230 nM) with
respect to alcohol 14, whereas in steroid 20, to our
delight, there is a 52-fold increase in affinity (Ks ) 2.8
( 1.3 nM) (Figure 1). Analogue 20 is the most potent
compound of the derivatives described in the present
study, being ∼71 times more active than allopreg-
nanolone. As we have seen from the pregnanediols
described, the 20S isomers are more active than the 20R
congeners. This observation corroborates with previous
studies concerning the modulation of human recombi-
nant GABAA receptors by the endogenous 5R- or 5â-
pregnan-3,20-diol.25b In particular, for the 5R series, the
introduction of the 20-hydroxyl group in the R orien-
tiation produced only a modest (approximately 2-fold)
decrease in potency, whereas the 20â-hydroxyl steroid
was ∼24 times less potent than the parent allopreg-
Previous studies5,20 have noted the importance of
O-O distance (3R-OH H-donor and side chain H-
acceptor) for several anesthetic steroids that modulate
GABAA receptors. The O3-O20 distance in the active
compounds 19 and 20 is 9.97 and 10.04 Å, respectively,
whereas, for the less-potent analogues 14 and 15, the
distance is 10.64 and 10.66 Å, respectively. However,
only the O3-O20 distance of the less-active analogues
is within the range of the corresponding distance values
(10.19-11.22 Å) found in active allopregnanolone ana-
logues.5,20 Hence, we propose that for pregnanediols, the
favorable distances between the hydrogen-bonding groups
do not fall within the range of the anesthetic steroid
allopregnanolone but are shorter. However, additional

5208 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16
Souli et al.
Figure 2. Lowest-energy conformers of compounds 20 (a), 15 (b), 19 (c), and 14 (d).
shown heterogeneous displacement of [35S]TBPS bind-
ing in rat brain regions.25b Three of our findings cannot
be reconciled with receptor heterogeneity. (1) Three
compounds (pregnanediols 18 and 23 and allopreg-
nanolone) out of the thirteen closely related steroids of
this study produced full maximal displacement. Accord-
ingly, allopregnanolone has shown full, homogeneous
displacement of [35S]TBPS binding as well.14 (2) The R
values of the other 10 steroids vary between 1.7 and
5.8 (Table 1), which would correspond to different
receptor subpopulations with a high affinity for the
steroids. (3) Lower-affinity populations showing the
displacement of [3H]EBOB binding could not be ob-
served within the solubility limits of the steroids. KS
values do not significantly depend on the binding models
(allostery or heterogeneity), and they correlate with the
pharmacological potencies of neuroactive steroids to
modulate GABAA receptor-ionophore activity for [35S]-
TBPS binding.14 On the other hand, R values do not
show an apparent correlation with any known GABAer-
gic pharmacological parameters of neurosteroids.
Figure 3. (A) Superimposition of compounds 20 and 15. (B)
Superimposition of compounds 14 and 19.
Conclusions
The goal of this study was to develop a series of
allopregnanolone analogues substituted by conforma-
tionally constrained 17â side chains to obtain additional
information about the SAR of 5R-reduced steroids to
modulate GABAA receptors. Specifically, we investigated
the presence of alkynyl or propadienyl functionalities
at positions C17 and C20 of the steroid skeleton. The
most active derivative is (20R)-17â-(1-hydroxy-2,3-buta-
dienyl)-5R-androstane-3-ol (20), which possesses low
nanomolar potency to modulate cerebellar GABAA re-
ceptors. Theoretical conformational analysis showed
that the two active compounds 19 and 20 adopt similar
defined critical dihedral angle values, which could be
an indication of conformational space preference.
studies with other pregnanediols will be needed to refine
the SAR of this group of compounds.
The ternary allosteric model of Ehlert26 has been
successfully applied to ionotropic glycine receptors27 but
not to [3H]EBOB binding to GABAA receptors. There-
fore, the binding data need to be discussed. Full
maximal displacement of [3H]EBOB binding means
strong negative cooperativity between steroidal and [3H]-
EBOB binding in the allosteric model. The resultant
cooperativity factors cannot be determined exactly (R
> 10). On the other hand, the majority of our neuro-
active steroids can be characterized by partial maximal
displacement and weaker cooperativity with R values
between 1 and 10 (Table 1). However, partial maximal
displacement can also be interpreted with a subpopu-
lation of cerebellar GABAA receptors that are insensitive
for that steroid. Accordingly, allopregnanolone metabo-
lites with 20-hydroxylated pregnanediol structures have
Experimental Section
NMR spectra were recorded on a Bruker AC 300 spectrom-
1
1
eter operating at 300 MHz for H and 75.43 MHz for 13C. H
NMR spectra are reported in units of δ relative to internal

Novel Neurosteroids that Modulate GABAA Receptors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16 5209
Figure 4. Contour plots generated for compounds 14, 15, 19, and 20 by rotating around the τ1 and τ2 dihedral angles in increments
of 10°. The contour levels, which are up to 6 kcal mol-1 higher than the lowest-energy minimum, are shown.
CHCl3 at 7.24 ppm. 13C NMR shifts are expressed in units of
δ relative to CDCl3 at 77.0 ppm. 13C NMR spectra were proton
noise decoupled. All NMR spectra were recorded in CDCl3.
Silica gel plates (Merck F254) were used for thin-layer chro-
matography. Chromatographic purification was performed
with silica gel (200-400 mesh). Analyses, indicated by the
symbols of the elements, were carried out by the micro-
analytical section of the Institute of Organic and Pharmaceuti-
cal Chemistry of the National Hellenic Research Foundation.
1,1-Dibromo-2-[3r-(tert-butyldiphenylsilyloxy)-5r-
androstan-17â-yl]-ethylene (4). Triphenylphosphine (3.14
g, 6 mmol) was added to a solution of tetrabromomethane (1.99
g, 6 mmol) in anhydrous methylene chloride (31 mL) at 0 °C,
and the resulting mixture was stirred for 10 min. Subse-
quently, a solution of 3R-(tert-butyldiphenylsilyloxy)-5R-an-
drostan-17â-carboxaldehyde (3) (540 mg, 1 mmol) in methylene
chloride (6 mL) was added, and the mixture was stirred at 0
°C for an additional 10 min. The reaction mixture was diluted
with ethyl acetate, and the organic layer was washed with a
water/saturated aqueous NaCl solution and was dried with
anhydrous Na2SO4. The solvent was evaporated in vacuo, and
the residue was purified by flash column chromatography
using petroleum ether 40-60 °C/ether 95/5 as eluent to afford
560 mg (93%) of compound 4 as a solid. mp: 69-71 °C; Anal.
(C37H50OSiBr2) C, H.
in anhydrous THF (8 mL) at -78 °C, and the resulting mixture
was stirred at this temperature for 1 h. The mixture was
quenched by the addition of saturated aqueous NH4Cl solution.
Ethyl acetate was added, and the organic layer was extracted
with water and brine and was dried with anhydrous Na2SO4.
Evaporation of the solvent in vacuo followed by purification
of the residue by flash column chromatography [petroleum
ether 40-60 °C/ether (97/3)] afforded 160 mg (89%) of the
desired compound 5 as a solid. mp: 60-62 °C; Anal. (C37H50-
OSi) C, H.
3r-(tert-Butyldiphenylsilyloxy)-17â-[2-(4-tolyl)ethynyl]-
5r-androstane. A solution of 5 (120 mg, 0.23 mmol) in
pyrrolidine (2 mL) was added to a mixture of tetrakis tri-
phenylposphine palladium (0) (13.28 mg, 0.0115 mmol) and
CuI (4.4 mg, 0.023 mmol) in pyrrolidine (1 mL). 4-Iodotoluene
(150 mg, 0.69 mmol) was added to the resulting mixture and
was stirred at rt for 24 h. Subsequently, saturated aqueous
NH4Cl solution was added, followed by the addition of ethyl
acetate. The organic layer was extracted with water and brine
and was dried with anhydrous Na2SO4. The solvent was
evaporated in vacuo, and the residue was purified by flash
column chromatography [petroleum ether 40-60 °C/acetone
(80/20)] to afford 100 mg (72%) of 3R-(tert-butyldiphenylsi-
lyloxy)-17â-[2-(4-tolyl)ethynyl]-5R-androstane as a solid. mp:
79-81 °C; Anal. (C44H56OSi) C, H.
17â-Ethynyl-3r-(tert-butyldiphenylsilyloxy)-5r-andros-
tane (5). A solution of n-BuLi (1.6 M) in hexanes (0.45 mL,
0.72 mmol) was added to a solution of 4 (230 mg, 0.33 mmol)
17â-[2-(4-Tolyl)ethynyl]-5r-androstane-3r-ol (6). A so-
lution of 3R-(tert-butyldiphenylsilyloxy)-17â-[2-(4-tolyl)eth-
ynyl]-5R-androstane (0.074 mmol) in anhydrous THF (3 mL)

5210 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16
Souli et al.
was treated with a 1 M solution of (n-Bu)4N+F- (1.48 mL, 1.48
mmol), and the resulting solution was stirred at rt for 48 h.
The reaction was cooled to 0 °C and was quenched with a
saturated aqueous NH4Cl solution. Ethyl acetate was added,
and the organic layer was extracted with water and brine and
was dried with anhydrous Na2SO4. Evaporation of the solvent
in vacuo, followed by purification of the residue by flash
column chromatography (methylene chloride), afforded the
desired compound 6 in 50% yield as a solid. mp: 178-180 °C;
1H NMR (δ): 0.77 (s, 3H, CH3), 0.81 (s, 3H, CH3), 0.86-2.08
(m, 23H), 2.30 (s, 3H, CH3 aromatic), 4.02 (bs, 1H, 3â-Η), 7.03-
acetone (92/8)] afforded, in 85% total yield, the two diastere-
omers 10a and 10b in a 3:1 ratio, respectively, each as a solid.
(20R)-1-[3r(tert-Butyldiphenylsilyloxy)-5r-androstan-
17â-yl]-3-trimethylsilyl-2-propyn-1-ol (10a) Less-polar di-
astereomer: 233 mg, mp: 68-71 °C; 1H NMR (δ): 0.15 (s, 9H,
Si(CH3)3), 0.71 (s, 6H, 18,19-CH3), 0.84-2.07 (m, 23H), 1.06
(s, 9H, C(CH3)3), 3.99 (bs, 1H, 3â-H), 4.27 (d, J ) 9.54 Hz, 1H,
20-H), 7.32-7.42 (m, 6H), 7.63-7.66 (m, 4H); 13C NMR (δ):
0.09, 11.39, 12.39, 19.36, 20.62, 24.30, 25.42, 27.06, 28.55,
29.32, 32.23, 32.66, 35.47, 36.05, 36.17, 39.36, 39.56, 42.54,
54.56, 56.00, 56.75, 65.21, 68.08, 91.48, 107.43, 127.42, 129.37,
134.90, 135.77; Anal. (C41H59O2Si2) C, H.
7.06 (d, J ) 7.97 Hz, 2H), 7.25-7.27 (d, J ) 7.97 Hz, 2H); 13
C
NMR (δ): 11.18, 13.77, 20.57, 21.31, 24.61, 28.74, 28.99, 29.28,
32.05, 32.19, 35.84, 36.11, 36.16, 37.48, 39.14, 42.88, 44.44,
54.40, 54.83, 66.51, 82.43, 90.94, 121.18, 128.81, 131.38,
137.17; Anal. (C28H38O) C, H.
(20S)-1-[3r(tert-Butyldiphenylsilyloxy)-5r-androstan-
17â-yl]-3-trimethylsilyl-2-propyn-1-ol (10b) More-polar di-
1
astereomer: 78 mg. mp: 62-65 °C; H NMR (δ): 0.15 (s, 9H,
Si(CH3)3), 0.68 (s, 3H, CH3), 0.71 (s, 3H, CH3), 0.75-2.23 (m,
23H), 1.05 (s, 9H, C(CH3)3), 3.99 (bs, 1H, 3â-H), 4.10-4.15 (m,
1H, 20-H), 7.31-7.39 (m, 6H), 7.62-7.65 (m, 4H); 13C NMR
(δ): 0.17, 11.49, 12.78, 19.39, 20.51, 23.91, 26.62, 27.08, 28.58,
29.37, 32.30, 32.75, 35.25, 36.08, 36.21, 38.61, 39.42, 41.92,
54.62, 56.20, 56.72, 64.92, 68.11, 90.23, 107.31, 127.46, 129.40,
134.83, 135.80; Anal. (C41H59O2Si2) C, H.
17â-(2-Bromoethynyl)-5r-androstane-3r-ol (7). A solu-
tion of 4 (0.074 mmol) in anhydrous THF (3 mL) was treated
with a 1 M solution of (n-Bu)4N+F- (1.48 mL, 1.48 mmol), and
the resulting solution was stirred at rt for 48 h. The mixture
was quenched with a saturated aqueous NH4Cl solution. Ethyl
acetate was added, and the organic layer was extracted with
water and brine and was dried with anhydrous Na2SO4.
Evaporation of the solvent in vacuo, followed by purification
of the residue by flash column chromatography [methylene
chloride/ethyl acetate (95/5)], afforded the desired compound
3r-(tert-Butyldiphenylsilyloxy)-5r-androstan-17â-car-
boxylic acid methoxy methyl amide. A solution of 3R-(tert-
butyldiphenylsilyloxy)-5R-androstan-17â-carboxylic acid (2)
(490 mg, 0.88 mmol) in anhydrous dichloromethane (5 mL)
was treated with N,O-dimethylhydroxylamine hydrochloride
(129 mg, 1.32 mmol), EDCI (161 mg, 1.32 mmol), and DMAP
(253 mg, 1.32 mmol), and the resulting mixture was stirred
at rt for 2h. The reaction was quenched by the addition of
saturated aqueous NaCl, which was followed by an extraction
with ethyl acetate. The organic layer was washed with 5% HCl
and brine, and was then dried over anhydrous Na2SO4.
Evaporation of the solvent in vacuo and purification of the
residue by flash column chromatography [petroleum ether
40°-60 °C/acetone (85:15)] afforded the corresponding Weinreb
amide as a white solid (520 mg, 98% yield).
1-[3r-(tert-Butyldiphenylsilyloxy)-5r-androstan-17â-
yl]-3-triisopropylsilyl-propynone (11) A solution of n-BuLi
(1.6 M) in hexanes (1.3 mL, 2.06 mmol) was added to a solution
of triisopropylsilyl acetylene (0.7 mL, 3.09 mmol) in anhydrous
THF (10 mL) dropwise at -40 °C. The resulting solution was
stirred for 1h at -40 °C, and subsequently, a solution of 3R-
(tert-butyldiphenylsilyloxy)-5R-androstan-17â-carboxylic acid
methoxy methyl amide (620 mg, 1.03 mmol) in anhydrous THF
(1.3 mL) was added dropwise to the acetylide solution at -10
°C under an argon atmosphere. The reaction mixture was
further stirred for 2 h at -10 °C and was quenched by the
addition of saturated aqueous NH4Cl, which was followed by
an extraction with diethyl ether. The organic layer was washed
with brine and was then dried over anhydrous Na2SO4.
Evaporation of the solvent and purification of the residue by
flash column chromatography [petroleum ether 40°-60 °C/
acetone (98:2)] afforded ketone 11 as a viscous oil (520 mg,
70% yield).
1
7 (70%) as a solid. mp: 165-167 °C; H NMR (δ): 0.75 (s, 3H,
CH3), 0.76 (s, 3H, CH3), 0.89-1.79 (m, 22H), 2.15 (t, J ) 9.5
Hz, 1H, 17R-H), 4.01 (bs, 1H, 3â-H); 13C NMR (δ): 11.18, 13.74,
20.52, 24.52, 28.45, 28.73, 29.02, 32.01, 32.19, 35.84, 36.03,
36.17, 37.37, 39.02, 39.14, 43.22, 44.32, 54.34, 54.60, 66.52,
81.89; Anal. (C21H31OBr) C, H.
4-[3r-(tert-Butyldiphenylsilyloxy)-5r-androstan-17â-
yl]-3-butyn-2-ol (8). A solution of n-BuLi (1.6 M) in hexanes
(2.76 mL, 4.4 mmol) was added to a solution of 4 (308 mg,
0.44 mmol) in anhydrous THF (11 mL) at -78 °C, and the
mixture was stirred for 2 h at this temperature. Acetaldehyde
(0.032 mL, 0.57 mmol) was added at -78 °C, and the
temperature was permitted to rise to 25 °C. The addition of a
2 M HCl solution at 0 °C was followed by an extraction with
ethyl acetate. The organic layer was washed with NaHCO3
and brine and was dried with anhydrous Na2SO4, and the
solvent was evaporated in vacuo. Subsequent purification of
the residue by flash column chromatography [using petroleum
ether 40-60 °C/ethyl acetate (90/10) as eluent] afforded 160
mg (62%) of the desired compound 8 as a solid. mp: 68-70 °C;
Anal. (C39H54O2Si) C, H.
17â-(3-Hydroxy-1-butynyl)-5r-androstane-3r-ol (9). Re-
moval of the protecting group from 8 was accomplished
following the same procedure as that used for compound 6 to
afford 83% of the desired product 9 after purification by flash
column chromatography [dichloromethane/ethyl acetate (90/
10)]. mp: 169-171 °C; 1H NMR (δ): 0.72 (s, 3H, CH3), 0.76 (s,
3H, CH3), 0.83-1.99 (m, 22H), 1.39 (s, 3H, CH(OH)CH3), 2.14
(t, J ) 9.36 Hz, 1H, 17R-H), 4.01 (bs, 1H, 3â-H), 4.5 (m, 1H,
CH(OH)CH3); 13C NMR (δ): 11.19, 13.68, 20.53, 24.53, 24.95,
28.46, 29.01, 29.15, 32.04, 32.20, 35.84, 36.06, 36.17, 37.38,
39.15, 42.10, 44.02, 54.38, 54.77, 58.68, 66.52, 84.22, 85.94.
Anal. (C23H36O2) C, H.
3-Trimethylsilyl-1-[3r-(tert-butyldiphenylsilyloxy)-5r-
androstan-17â-yl]-2-propyn-1-ol (10a,b). A solution of 1.6
M n-BuLi in hexanes (1.07 mL, 1.715 mmol) was added to a
solution of trimethylsilylacetylene (0.24 mL, 1.715 mmol) in
anhydrous THF (5.7 mL) at 0 °C, and the resulting mixture
was stirred for 2 h. Subsequently, the mixture was cooled to
-78 °C, a solution of 3 (310 mg, 0.57 mmol) in anhydrous THF
(5.7 mL) was added, and the mixture was stirred at this
temperature for 3 h. The quenching of the reaction by the
addition of NH4Cl was followed by an extraction with ethyl
acetate. The organic layer was washed with water and brine
and dried with anhydrous Na2SO4, and the solvent was
evaporated in vacuo. Subsequent purification of the residue
by flash column chromatography [petroleum ether 40-60 °C/
(20S)-1-[3r(tert-Butyldiphenylsilyloxy)-5r-androstan-
17â-yl]-3-triisopropyllsilyl-2-propyn-1-ol (12b). A solution
of ketone 11 (120 mg, 0.16 mmol) in anhydrous dichlo-
romethane was treated with (S)-3,3-diphenyl-1-methyltetrahy-
dro-3H-pyrrolo[1,2-c][1,3,2]oxazaborole (1 M) in toluene (0.024
mL, 0.024 mmol). The resulting mixture was cooled to -78
°C, and a solution of catecholborane (0.02 mL, 0.192 mmol) in
anhydrous dichloromethane was then added dropwise. After
the mixture had been stirred for 5h at -78 °C, methanol (1
mL) was added, and the solution was warmed to rt and diluted
with diethyl ether. The organic layer was first washed with a
mixture of 1N NaOH-saturated aqueous NaHCO3 (1:2) until
it became colorless, then washed with brine, and then was
dried over anhydrous Na2SO4. After evaporation of the solvent,
the residue was purified by flash column chromatography
[petroleum ether 40°-60 °C/acetone (90:10)] to afford 12b as
1
a white solid (110 mg, 95% yield, and 100% ee). H NMR (δ):
0.72 (s, 3H, CH3), 0.73 (s, 3H, CH3), 1.09 (s, 21H), 1.23 (s, 9H),

Novel Neurosteroids that Modulate GABAA Receptors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16 5211
2.26-1.25 (m, 22H), 4.01 (bs, 1H, 3â-H), 4.23 (bs, 1H, 20H),
7.41-7.33 (m, 6H), 7.68-7.64 (m, 4H).
was added. The slurry was cooled to 0 °C, and a solution of
alcohol 12a (100 mg, 0.14 mmol) in THF (2 mL) was added.
The reaction mixture was stirred at rt, and, after 30 min, MeI
(0.02 mL, 0.28 mmol) was added; stirring was then continued
for 5 h. The reaction was quenched by saturated aqueous NH4-
Cl and was diluted with ethyl acetate, and the organic layer
was washed with brine and was dried over anhydrous Na2-
SO4. After the evaporation of the solvent, the residue was
purified by flash column chromatography [petroleum ether
40-60 °C/dietyl ether (98:2)] to afford (20R)-3-triisopropyl-1-
methoxy-1-[3R-(tert-butyldiphenylsilyloxy)-5R-androstan-17â-
yl]-2-propyne as a white solid (70 mg, 70% yield).
(20R)-1-[3r(tert-Butyldiphenylsilyloxy)-5r-androstan-
17â-yl]-3-triisopropyllsilyl-2-propyn-1-ol (12a). A solution
of (S)-3,3-diphenyl-1-methyltetrahydro-3H-pyrrolo[1,2-c][1,3,2]-
oxazaborole (1 M) in toluene (0.084 mL, 0.084 mmol) was
added to a solution of ketone 11 (30 mg, 0.042 mmol) in 3 mL
of THF. The resulting solution was cooled to -30 °C, and then
0.02 mL (0.21 mmol) of borane methyl sulfide complex was
added. After the reaction was completed, it was quenched by
the slow addition of methanol (1.0 mL). The reaction mixture
was diluted with diethyl ether, and the organic layer was
washed with saturated NH4Cl, 5% NaHCO3, and brine. The
organic layer was dried over anhydrous Na2SO4 and was
concentrated in vacuo. The residue was purified by flash
column chromatography [petroleum ether 40°-60 °C/acetone
(90:10)] to afford 12a as a white solid (28 mg, 93% yield, and
100% ee). 1H NMR (δ): 0.71 (s, 6H, 18,19-CH3), 0.84-2.07 (m,
23H), 1.06 (s, 9H, C(CH3)3), 3.99 (bs, 1H, 3â-H), 4.27 (d, J )
9.54 Hz, 1H, 20-H), 7.32-7.42 (m, 6H), 7.63-7.66 (m, 4H).
17â-(3-Oxo-1-propynyl)-5r-androstan-3r-ol (13). A solu-
tion of 11 (155 mg, 0.215 mmol) in anhydrous dichloromethane
(9 mL) was treated with HF-pyridine (0.86 mmol, 0.024 mL)
at 0 °C, and the resulting mixture was stirred at rt for 2 h.
The addition of water at 0 °C was followed by an extraction
with methylene chloride. The organic layer was washed with
brine and was dried with anhydrous Na2SO4, and the solvent
was evaporated in vacuo. Subsequent purification of the
residue by flash column chromatography [dichloromethane/
ethyl acetate (95/5)] afforded the desired compound 13 in 76%
yield as a solid. mp:148-150 °C; 1H NMR (δ): 0.67 (s, 3H, CH3),
0.77 (s, 3H, CH3), 0.80-2.27 (m, 22H), 2.62 (t, J ) 8.8 Hz, 1H,
17R-H), 3.21 (s, 1H, acetylenic), 4.04 (bs, 1H, 3â-H); 13C NMR
(δ): 11.23, 13.49, 20.64, 22.39, 24.26, 28.46, 29.05, 31.97, 32.22,
35.53, 35.90, 36.17, 38.69, 39.15, 54.29, 56.84, 64.81, 66.52,
73.85, 78.65, 97.77, 197.91; Anal. (C22H32O2) C, H.
(20R)-17â-(1-Hydroxy-2-propynyl)-5r-androstan-3-ol
(14). Compound 12a (200 mg, 0.31 mmol) was treated with a
solution of 1 M (n-Bu)4N+F- (7.8 mL, 7.8 mmol) in hexanes at
rt for 24 h. The reaction was completed using the same
procedure as that used for compound 6 to afford, after
purification by flash column chromatography using dichlo-
romethane/ethyl acetate 90/10 as elution solvent, alcohol 14
in 80.5% yield. mp: 197-199 °C; 1H NMR (δ): 0.73 (s, 3H, 18-
CH3), 0.77 (s, 3H, 19-CH3), 0.80-2.10 (m, 23H), 2.41 (d, J )
1.81 Hz, 1H), 4.03 (bs, 1H, 3â-H), 4.24-4.72 (m, 1H, 20-H).
13C NMR (δ): 11.23, 12.24, 20.59, 24.30, 25.34, 28.53, 29.03,
32.03, 32.20, 35.43, 35.91, 36.15, 39.16, 39.63, 42.56, 54.45,
56.07, 56.67, 64.68, 66.01, 72.50, 85.44; Anal. (C22H34O2) C,
H.
(20S)-17â-(1-Hydroxy-2,3-butadienyl)-5r-androstane-
3-ol (15). Paraformaldehyde (20 mg), copper hypobromide (4
mg, 0.027 mmol), and diisopropylamine (0.07 mL, 0.5 mmol)
were sequentially added to a solution of alcohol 14 (83 mg,
0.25 mmol) in anhydrous dioxane (3 mL). The resulting
mixture was refluxed for 24 h. The addition of HCl 2 M and
extraction with ethyl acetate were followed by an extraction
of the organic layer with NaHCO3, water, and brine and a
drying with anhydrous Na2SO4. The solvent was evaporated
in vacuo, and the residue was purified by flash column
chromatography, using dichloromethane/ethyl acetate 85/15
as elution solvent, to afford allenic alcohol 15 (45 mg, 52%).
mp: 184-187 °C.; 1H NMR (δ): 0.76 (s, 3H, CH3), 0.78 (s, 3H,
CH3), 0.85-2.10 (m, 23H), 4.03 (bs, 1H, 3â-H), 4.07 (m, 1H,
20-H), 4.82 (dd, J ) 1.23 Hz, 2H, 23-H), 5.20 (q, J ) 6.6 Hz,
1H, 21-H).; 13C NMR (δ): 11.18, 12.42, 20.64, 24.43, 25.29,
28.54, 29.03, 32.06, 32.18, 35.39, 35.88, 36.13, 39.15, 39.85,
42.74, 54.41, 56.03, 56.75, 66.59, 72.56, 77.16, 94.89, 207.05.;
IR(cm-1): 1995 (CdCdCH2), 1255 (C-O); Anal. (C23H36O2) C,
H.
(20R)-17â-(1-Methoxy-2-propynyl)-5r-androstan-3r-
ol (16) The deprotection of 3R-OH and the removal of the
triisopropylsilyl group in 3-triisopropylsilyl-1-methoxy-1-[3R-
(tert-butyldiphenylsilyloxy)-5R-androstan-17â-yl]-2-propyne were
accomplished in one step by adding 1 M (n-Bu)4N+F- solution
in THF (1.9 mL, 1.9 mmol) to a solution of the above compound
(50 mg, 0.076 mmol) in anhydrous THF (3 mL). The resulting
solution was stirred at rt for 24 h and was then taken through
the same process used to obtain compound 6 to afford, after
purification by flash column chromatography using petroleum
ether 40-60 °C/acetone 82/18 as the elution solvent, methoxy
1
derivative 16. mp: 161-163 °C; H NMR (δ): 0.69 (s, 3H, 18-
CH3), 0.77 (s, 3H, 19-CH3), 0.79-2.24 (m, 23H), 2.44 (d, J )
1.83 Hz, 1H), 3.37 (s, 3H, OCH3), 3.70 (dd, J ) 9.8 Hz, 1H,
20-H), 4.03 (bs, 1H, 3â-H); 13C NMR (δ): 10.74, 11.94, 20.44,
23.89, 26.62, 28.54, 29.02, 32.03, 32.16, 35.17, 35.89, 36.11,
38.63, 39.14, 41.74, 54.28, 54.37, 55.90, 56.13, 66.58, 73.11,
74.59, 82.95; Anal. (C23H35O2) C, H.
(20R)-1-[3r-(tert-Butyldiphenysilyloxy)-5r-androstan-
17â-yl]-2-propyn-1-ol (17) A solution of 1 M (n-Bu)4N+F- in
THF (0.72 mL, 0.72 mmol) was added to a solution of 12a (310
mg, 0.48 mmol) in THF (10 mL), and the resulting mixture
was stirred at rt for 0.5 h. The reaction mixture was taken
through the same process used to obtain compound 6 to afford,
after purification by flash column chromatography using
petroleum ether 40-60 °C/acetone 82/18 as the elution solvent,
compound 17 (230 mg, 84%). Anal. (C38H52O2Si) C, H.
(20R)-1-{3r-(tert-Butyldiphenysilyloxy)-5r-androstan-
17â-yl-[3-(4-methoxyphenyl)-2-propynyl]}-1-ol. A solution
of 17 (100 mg, 0.176 mL) in pyrrolidine (2.5 mL) was added to
a solution of 4-iodoanisole (82.38 mg, 0.35 mmol), CuI (19 mg,
0.1 mmol), and tetrakis(triphenylphosphine) palladium (0) (58
mg, 0.05 mmol) in pyrrolidine (1.5 mL). The resulting mixture
was stirred for 24 h at rt. The quenching of the reaction by
the addition of NH4Cl was followed by an extraction with ethyl
acetate. The organic layer was washed with water and brine
and dried with (Na2SO4), and the solvent was evaporated in
vacuo. Subsequent purification of the residue with flash
column chromatography using petroleum ether 40-60 °C/
acetone 80/20 as the eluent afforded, in 84% yield, (20R)-1-
{3R-(tert-butyldiphenysilyloxy)-5R-androstan-17â-yl-[3-(4-meth-
oxy phenyl)-2-propynyl]}-1-ol as a solid. mp: 63-65 °C; Anal.
(C45H57O3Si) C, H.
(20R)-17â-[1-Hydroxy-3-(4-methoxyphenyl)-2-propynyl]-
5r-androstan-3r-ol (18). The removal of the protecting group
in (20R)-1-{3R-(tert-butyldiphenysilyloxy)-5R-androstan-17â-
yl-[3-(4-methoxyphenyl)-2-propynyl]}-1-ol was accomplished
following the same procedure as that followed for compound
6 to afford, after flash column purification using dichlo-
romethane/ethyl acetate (90/10) as the elution solvent, com-
1
pound 18 in 77% yield. mp: 185-188 °C.; H NMR (δ): 0.76
(s, 3H, 19-CH3), 0.78 (s, 3H, 18-CH3), 0.81-2.08 (m, 23H), 3.79
(s, 1H, OCH3), 4.03 (bs, 1H, 3â-H), 4.45-4.49 (m, 1H, 20-H),
6.80 (d, J ) 8.8 Hz, 2H), 7.32 (d, J ) 8.8 Hz, 2H).; 13C NMR
(δ): 11.19, 12.38, 20.58, 24.30, 25.45, 28.52, 29.02, 32.02, 32.16,
35.40, 35.87, 36.14, 39.14, 39.61, 42.56, 54.42, 55.26, 56.07,
57.07, 65.28, 66.60, 84.40, 89.22, 113.86, 114.99, 133.03,
159.52; Anal. (C29H39O3) C, H.
(20R)-3-Triisopropyl-1-methoxy-1-[3r-(tert-butyldi-
phenylsilyloxy)-5r-androstan-17â-yl]-2-propyne. Sodium
hydride (60% dispersion in mineral oil, 12 mg, 0.28 mmol) was
washed with dry hexane (2 × 2 mL), and then 1.5 mL of THF
(20S)-17â-(1-Hydroxy-2-propynyl)-5r-androstan-3-ol
(19). Compound 12b (200 mg, 0.31 mmol) was treated with a
solution of 1 M (n-Bu)4N+F- (7.8 mL, 7.8 mmol) in hexanes at
rt for 24 h. The reaction was completed using the same

5212 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16
Souli et al.
procedure as that used for compound 6 to afford, after
purification by flash column chromatography using dichlo-
romethane/ethyl acetate 90/10 as the elution solvent, alcohol
19 in 80.5% yield. mp: 187-190 °C. 1H NMR (δ): 0.69 (s, 3H,
18-CH3), 0.77 (s, 3H, 19-CH3), 0.79-2.21 (m, 23H), 2.49 (d, J
) 2.02 Hz, 1H), 4.03 (bs, 1H, 3â-H), 4.15 (d, J ) 8.6 Hz, 1H,
20-H). 13C NMR (δ): 11.17, 12.73, 20.40, 23.84, 26.32, 28.51,
29.02, 32.02, 32.16, 35.13, 35.87, 36.11, 38.51, 39.15, 41.87,
54.38, 56.19, 56.48, 64.15, 66.58, 85.28; Anal. (C22H34O2) C,
H.
(20R)-17â-(1-Hydroxy-2,3-butadienyl)-5r-androstane-
3-ol (20). Paraformaldehyde (20 mg), copper hypobromide (4
mg, 0.027 mmol), and diisopropylamine (0.07 mL, 0.5 mmol)
were sequentially added to a solution of alcohol 19 (83 mg,
0.25 mmol) in anhydrous dioxane (3 mL). The resulting
mixture was refluxed for 24 h. The addition of HCl 2 M and
extraction with ethyl acetate was followed by an extraction of
the organic layer with NaHCO3, water, and brine and a drying
with anhydrous Na2SO4. The solvent was evaporated in vacuo,
and the residue was purified by flash column chromatography
using dichloromethane/ethyl acetate 85/15 as the elution
solvent to afford allenic alcohol 20 (45 mg, 52%). mp: 174-
177 °C. 1H NMR (δ): 0.66 (s, 3H, 18-CH3), 0.77 (s, 3H, 19-
CH3), 0.79-1.91 (m, 23H), 4.03 (bs, 2H, 3â-H, 20-H), 4.79 (dd,
J ) 0.90 Hz, 2H, 23-H), 5.20 (q, J ) 6.4 Hz, 1H, 21-H). 13C
NMR (δ): 11.17, 12.98, 20.48, 24.11, 25.95, 28.53, 29.02, 32.00,
32.19, 35.14, 35.88, 36.21, 38.92, 39.14, 41.96, 54.42, 56.17,
56.91, 66.57, 72.92, 77.19, 94.76, 207.60; IR(cm-1): 1995 (Cd
CdCH2), 1255 (C-O); Anal. (C23H36O2) C, H.
(20S)-1-Methoxy-1-[3r-(tert-butyldiphenylsilyloxy)-5r-
androstan-17â-yl]-2-propyne. Sodium hydride (60% disper-
sion in mineral oil, 12 mg, 0.28 mmol) was washed with dry
hexane (2 × 2 mL), and 1.5 mL THF was added. The slurry
was cooled to 0 °C, and a solution of alcohol 12b (100 mg, 0.14
mmol) in THF (2 mL) was added. The reaction mixture was
stirred at rt, and, after 30 min, MeI (0.02 mL, 0.28 mmol) was
added; stirring was then continued for 5 h. The reaction was
quenched by saturated aqueous NH4Cl and was diluted with
ethyl acetate, and the organic layer was washed with brine
and was dried over anhydrous Na2SO4. After the evaporation
of the solvent, the residue was purified by flash column
chromatography [petroleum ether 40°-60 °C/diethyl ether (98:
2)] to afford (20S)-1-methoxy-1-[3R-(tert-butyldiphenylsilyloxy)-
5R-androstan-17â-yl]-2-propyne as a white solid (70 mg, 70%
yield).
(20S)-1-{3r-(tert-Butyldiphenysilyloxy)-5r-androstan-
17â-yl-[3-(4-methoxyphenyl)-2-propynyl]}-1-ol. A solution
of 22 (100 mg, 0.176 mL) in pyrrolidine (2.5 mL) was added to
a solution of 4-iodoanisole (82.38 mg, 0.35 mmol), CuI (19 mg,
0.1 mmol), and tetrakis(triphenylphosphine) palladium (0) (58
mg, 0.05 mmol) in pyrrolidine (1.5 mL). The resulting mixture
was stirred for 24 h at rt. The quenching of the reaction by
the addition of NH4Cl was followed by an extraction with ethyl
acetate. The organic layer was washed with water and brine
and dried with anhydrous Na2SO4, and the solvent was
evaporated in vacuo. Subsequent purification of the residue
with flash column chromatography using petroleum ether 40-
60 °C/acetone 80/20 as the eluent afforded, in 84% yield, (20S)-
1-{3R-(tert-butyldiphenysilyloxy)-5R-androstan-17â-yl-[3-(4-
methoxyphenyl)-2-propynyl]}-1-ol as a solid. mp: 50-53 °C;
Anal. (C45H57O3Si) C, H.
(20S)-17â-[1-Hydroxy-3-(4-methoxyphenyl)-2-propynyl]-
5r-androstan-3r-ol (23) The removal of the protecting group
of (20S)-1-{3R-(tert-butyldiphenysilyloxy)-5R-androstan-17â-yl-
[3-(4-methoxyphenyl)-2-propynyl]}-1-ol was accomplished us-
ing the same procedure as that used for compound 6 to afford,
after flash column purification using dichloromethane/ethyl
acetate (90/10) as the elution solvent, compound 23 in 77%
yield. mp: 159-161 °C; 1H NMR (δ): 0.73 (s, 3H, 18-CH3),
0.76 (s, 3H, 19-CH3), 0.85-2.25 (m, 23H), 3.79 (s, 1H, OCH3),
4.03 (bs, 1H, 3â-H), 4.33-4.36 (m, 1H, 20-H), 6.81 (d, J ) 8.8
Hz, 2H), 7.33-7.36 (d, J ) 8.8 Hz, 2H).; 13C NMR (δ): 11.20,
12.78, 20.56, 24.02, 26.62, 28.53, 29.02, 32.03, 32.19, 35.16,
35.87, 36.13, 38.57, 39.15, 42.65, 54.39, 55.27, 56.17, 57.03,
65.03, 66.58, 85.41, 89.17, 113.92, 114.98, 132.94, 159.61; Anal.
(C29H39O3) C, H.
Theoretical Calculations. Compounds 14, 15, 19, and 20
were subjected to energy minimization using steepest descents,
conjugate gradients, and Newton-Raphson algorithms embed-
ded in the Quanta/Charmm software of Accelrys and using an
SG O2 workstation. The minimization procedure was ended
using energy convergence criterion between successive steps
(∆E < 0.001 kcal mol-1). The dielectric constant during the
calculations was set to 1. Monte Carlo conformational search
analysis was applied to obtain low-energy conformers for the
compounds under study. These would serve as initial struc-
tures for grid-scan analysis. The application of Monte Carlo
analysis, which allows full angular window specification and
the random change of all flexible dihedral angles of flexibility,
generated 1000 conformers, which were consequently subjected
to energy minimization using 1000 iterations and the steepest
descents algorithm. The four lowest-energy structures were
then used as starting structures for grid-scan analysis. Each
grid-scan analysis for the four compounds was performed
around the τ1 and τ2 dihedral angles by rotating each angle in
increments of 10°. Thus, 1296 (36 × 36) conformers were
generated. None of these conformers had a lower value than
the initial low-energy conformer derived by Monte Carlo
analysis. The energy of the conformers versus the two dihedral
angles τ1 and τ2 were plotted as contours using 12 contour
levels, each one increasing by 0.5 kcal mol-1 from the lowest-
energy minimum. Therefore, contour levels reach up to 6 kcal
mol-1 from the lowest-energy minimum.
Receptor Binding. Crude synaptosomal membranes were
prepared from the cerebella of male Wistar rats. Cerebellar
homogenates in 0.32 M sucrose were centrifuged at 1.000g for
10 min. The supernatant was centrifuged at 45000g for 30 min.
The pellet was homogenized in distilled water, centrifuged at
45000g for 30 min, washed by suspension in Tris HCl buffer
(pH 7.4), and after two centrifugations as above was frozen.
The thawed membranes were centrifuged in 50 mM Tris HCl
buffer containing 0.2 M NaCl at 10000g for 10 min. For
displacement studies, membrane suspensions were incubated
with 1 nM [3H]EBOB (30 Ci/mmol, Dupont-NEN) in the above
buffer for 2 h at 25 °C. For nonspecific binding, 50 µM
picrotoxin was applied. Duplicate 1-mL samples were filtered
on Whatman GF/B filters under vacuum with a Brandel
harvester and washed with 3 × 3 mL of buffer.
(20S)-17â-(1-Methoxy-2-propynyl)-5r-androstan-3r-
ol (21) The deprotection of 3R-OH and the removal of the
triisopropylsilyl group in (20S)-3-triisopropylsilyl-1-methoxy-
1-[3R-(tert-butyldiphenylsilyloxy)-5R-androstan-17â-yl]-2-pro-
pyne was accomplished in one step by adding 1 M (n-Bu)4N+F-
solution in THF (1.9 mL, 1.9 mmol) to a solution of the above
compound (50 mg, 0.076 mmol) in THF (3 mL). The resulting
solution was stirred at rt for 24 h and was then taken through
the same process used to obtain compound 6 to afford, after
purification by flash column chromatography using petroleum
ether 40-60 °C/acetone 82/18 as the elution solvent, methoxy
1
derivative 21. mp: 171-173 °C; H NMR (δ): 0.66 (s, 3H, 18-
CH3), 0.77 (s, 3H, 19-CH3), 0.86-2.34 (m, 23H), 2.37 (d, J )
1.7 Hz, 1H), 3.36 (s, 3H, OCH3), 3.82 (dd, J ) 10.37 Hz, 1H,
20-H), 4.02 (bs, 1H, 3â-H); 13C NMR (δ): 11.20, 12.36, 20.55,
24.27, 25.36, 28.53, 29.03, 32.04, 32.15, 35.43, 35.88, 36.12,
39.13, 39.32, 42.58, 54.09, 54.41, 55.90, 56.13, 66.58, 73.34,
73.46, 82.89; Anal. (C23H35O2) C, H.
(20S)-1-[3r-(tert-Butyldiphenysilyloxy)-5r-androstan-
17â-yl]-2-propyn-1-ol (22). A solution of 1 M (n-Bu)4N+F-
in THF (0.72 mL, 0.72 mmol) was added to a solution of 12b
(310 mg, 0.48 mmol) in THF (10 mL), and the resulting
mixture was stirred at rt for 0.5 h. The reaction mixture was
taken through the same process used to obtain compound 6
to afford, after purification by flash column chromatography
using petroleum ether 40-60 °C/acetone 82/18 as the elution
solvent, compound 22 in 84% yield (230 mg). Anal. (C38H52O2-
Si) C, H.

Novel Neurosteroids that Modulate GABAA Receptors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16 5213
(5) Purdy, R. H.; Morrow, A. L.; Blinn, J. R.; Paul, S. M. Synthesis,
metabolism, and pharmacological activity of 3R-hydroxy steroids
which potentiate GABA-receptor-mediated chloride ion uptake
in rat cerebral cortical synaptoneurosomes. J. Med. Chem. 1990,
33, 1572-1581.
Scheme 5
(6) (a) Hill-Venning, C.; Belleli, D.; Peters, J. A.; Lambert, J. J.
Subunit dependent interaction of the general anaesthetic eto-
midate with the gamma-aminobutyric acid type A receptor. Br.
J. Pharmacol. 1997, 120, 749-756. (b) Belleli, D.; Casula, A.;
Ling, A.; Lambert, J. J. The influence of subunit composition
on the interaction of neurosteroids with GABA(A) receptors.
Neuropharmacology 2002, 4, 651-661. (c) Lambert, J. L.; Belelli,
D.; Peden, D. R.; Vardy, A. W.; Peters, J. A. Neurosteroid
modulation of GABAA receptors. Prog. Neurobiol. (Oxford) 2003,
71, 67-80. (d) Po¨ltl, A.; Hauer, B.; Fuchs, K.; Tretter, V.;
Sieghart, W. Subunit composition and quantitative importance
of GABAA receptor subtypes in the cerebellum of mouse and
rat. J. Neurochem. 2003, 87, 1444-1455.
(7) (a) Ueno, S.; Tsutsui, M.; Toyohira, Y.; Minami, K.; Yanagihara,
N. Sites of positive allosteric modulation by neurosteroids on
ionotropic γ-aminobutyric acid receptor subunits. FEBS Lett.
2004, 566, 213-217. (b) Rick, C. E.; Ye, Q.; Finn, S. E.; Harrison,
N. L. Neurosteroids act on the GABAA receptor at sites on the
N-terminal side of the middle of TM2. NeuroReport 1998, 9,
379-383.
(8) (a) Luddens, H.; Korpi, E. R.; Seeburg, P. H. GABAA/benzodi-
azepine receptor heterogeneity: neurophysiological implications.
Neuropharmacology 1995, 34, 245-254. (b) Siegel, E.; Buhr, A.
The benzodiazepine binding site of GABAA receptors. Trends
Pharmacol. Sci. 1997, 18, 425-429. (c) Wohlfarth, K. M.;
Bianchi, M. T.; Macdonald, R. L. Enhanced neurosteroid poten-
tiation of ternary GABAA receptors containing the delta subunit.
J. Neurosci. 2002, 22, 1541-1549.
Binding data were evaluated according to the ternary
allosteric binding model of Ehlert26 shown in Scheme 5. Ratios
of specific binding in the presence (BES) and absence (BE) of
the steroids were fitted to eq 1 (Scheme 5) via nonlinear
regression with GraphPad Prism 4.02 (San Diego, CA).
Saturation analysis of [3H]EBOB binding previously revealed
a homogeneous population of binding sites in rat cerebellum
with KE ) 1.2 ( 0.2 nM. Three steroids (18, 23, and allopreg-
nanolone) resulted in apparently full maximal displacement
of [3H]EBOB binding, and the remaining binding could not be
quantitatively distinguished from nonspecific binding. There-
fore, cooperativity factors could not be exactly determined (R
> 10). On the other hand, steroid 20 resulted in low maximal
displacement (∼30%, Figure 1), and some displacement curves
could not be fitted separately. Therefore, curve fitting was
applied for the pooled experimental data for 20.
(9) (a) Pillai, G. V.; Smith, A. J.; Hunt, P. A.; Simpson, P. B. Multiple
structural features of steroids mediate subtype-selective effects
on human R4â3δ GABAA receptors. Biochem. Pharmacol. 2004,
68, 819-831. (b) Puia, G.; Santi, M. R.; Vicini, S.; Pritchett, D.
B.; Purdy, R. H.; Paul, S. M.; et al. Neurosteroids act on
recombinant human GABAA receptors. Neuron 1990, 4, 759-
765.
Acknowledgment. This work was supported in part
by the Greek General Secretariat for Research and
Technology and ELPEN S.A., as well as Te´T GR-40/2001
and 1/A/005/2004 NKFP MediChem2 of the Hungarian
National Office for Research and Technology.
(10) Britton, K. T.; Koob, G. F. Neuropharmacology: premenstrual
steroids? Nature 1998, 392, 869-870.
(11) (a) Dubal, D. B.; Wilson, M. E.; Wise, P. M. Estradiol:
a
protective and trophic factor in the brain. J. Alzheimer’s Dis.
1999, 1, 265-274. (b) Guarneri, P.; Cascio, C.; Russo, D.;
D’Agostino, S.; Drago, G.; Galizzi, G.; De Leo, G.; Piccoli, F.;
Guarneri, M.; Guarneri, R. Neurosteroids in the retina: neuro-
degenerative and neuroprotective agents in retinal degeneration.
Ann. N.Y. Acad. Sci. 2003, 1007, 117-128.
Supporting Information Available: Experimental pro-
cedures and spectroscopic data for compounds 2 and 3.
Spectroscopic data for compounds 4, 5, 8, 11, 17, 22, 3R-
(tert-butyldiphenylsilyloxy)-17â-[2-(4-tolyl)ethynyl]-5R-andros-
tane, 3R-(tert-butyldiphenylsilyloxy)-5R-androstan-17â-carbox-
ylic acid methoxy methyl amide, (20R)-3-triisopropyl-1-meth-
oxy-1-[3R-(tert-butyldiphenylsilyloxy)-5R-androstan-17â-yl]-2-
propyne, (20R)-1-{3R-(tert-butyldiphenysilyloxy)-5R-androstan-
17â-yl-[3-(4-methoxyphenyl)-2-propynyl]}-1-ol, (20S)-1-meth-
oxy-1-[3R-(tert-butyldiphenylsilyloxy)-5R-androstan-17â-yl]-2-
propyne, (20S)-1-{3R-(tert-butyldiphenysilyloxy)-5R-androstan-
17â-yl-[3-(4-methoxyphenyl)-2-propynyl]}-1-ol. Geometry op-
timization data and low-energy conformers for compounds 14,
15, 19, and 20. Analytical data for compounds 4-23. This
material is available free of charge via the Internet at http://
pubs.acs.org.
(12) (a) Mayo, W.; Lemaire, V.; Malaterre, J.; Rodriguez, J. J.; Cayre,
M.; Stewart, M. G.; Kharouby, M.; Rougon, G.; Le Moal, M.;
Piazza, P. V.; Abrous, D. N. Pregnenolone sulfate enhances
neurogenesis and PSA-NCAM in young and aged hippocampus.
Neurobiol. Aging 2005, 26, 103-114. (b) Fritschy, J. M.; Brunig,
I. Formation and plasticity of GABAergic synapses: physiological
mechanisms and pathophysiological implications. Pharmacol.
Ther. 2003, 98, 299-323.
(13) Smith, S. S.; Gong, Q. H.; Hsu, F. C.; Markowitz, R. S.; Ffrench-
Mullen, J. M.; Li, X. GABAA receptor alpha4 subunit suppression
prevents withdrawal properties of an endogenous steroids.
Nature 1998, 392, 926-930.
(14) (a) Lan, N. C.; Gee, K. W. Neuroactive steroid actions at the
GABAA receptor. Horm. Behav. 1994, 28, 537-544. (b) Hawk-
inson, J. E.; Kimbrough, C. L.; McCauley, L. D.; Bolger, M. B.;
Lan, N. C.; Gee, K. W. The neuroactive steroid 3R-hydroxy-5â-
pregnan-20-one is a two-component modulator of ligand binding
to the GABAA receptor. Eur. J. Pharmacol. 1994, 269, 157-163.
(c) Gee, K. W.; McCauley, L. D.; Lan, N. C. A putative receptor
for neurosteroids on the GABAA receptor complex: the phar-
macological properties and therapeutic potential of epalons. Crit.
Rev. Neurobiol. 1995, 9, 207-227.
(15) (a) Mellon, S. H.; Griffin, L. D. Neurosteroids: biochemistry and
clinical significance. Trends Endocrinol. Metab. 2002, 13, 35-
43. (b) Zorumski, C. F.; Mennerick, S.; Isenberg, K. E.; Covey,
D. F. Potential clinical uses of neuroactive steroids. Curr. Opin.
Invest. Drugs (Thomson Sci.) 2000, 1, 360-369.
References
(1) (a) Paul, S. M.; Purdy, R. H. Neuroactive steroids. FASEB J.
1992, 6, 2311-2322. (b) Robel, R.; Baulieu, E. E. Neurosteroids
biosynthesis and function. Trends Endocrinol. Metab. 1994, 5,
1-8. (c) Kulkarni, S. K.; Reddy, D. S. Neurosteroids: a new class
of neuromodulators. Drugs Today 1995, 31, 433-455. (d) Compag-
none, N. A.; Mellon, S. H. Neurosteroids: biosynthesis and
function of these novel neuromodulators. Front. Neuroendocrinol.
2000, 21, 1-56.
(2) (a) Schumacher, M.; Robel, P.; Baulieu, E. E. Development and
regeneration of the nervous system: a role for neurosteroids.
Dev. Neurosci. (Basel, Switz.) 1996, 18, 16-21. (b) Majewska,
M. D. Neurosteroids: endogenous bimodal modulators of the
GABAA receptor. Mechanism of action and physiological signifi-
cance. Prog. Neurobiol. (Amsterdam, Neth.) 1992, 38, 379-395.
(3) Rupprecht, R.; Holsboer, F. Neuroactive steroids: mechanisms
of action and neuropsychopharmacological perspectives. Trends
Neurosci. 1999, 22, 410-416.
(16) Mienville, J. M.; Vicini, S. Pregnenolone sulfate antagonizes
GABAA receptor-mediated currents via a reduction of channel
opening frequency. Brain Res. 1989, 489, 190-194.
(17) (a) Floods, J. F.; Smith, G. E.; Roberts, E. Dehydroepiandros-
terone and its sulfate enhance memory retention in mice. Brain
Res. 1988, 447, 269-278. (b) Mayo, W.; Dellu, F.; Robel, P.;
Cherkaoui, J.; Le Moal, M.; Baulieu, E. E.; Simon, H. Infusion
of neurosteroids into the nucleus basalis magnocellularis affects
cognitive processes in the rat. Brain Res. 1993, 607, 324-328.
(18) Gee, K. W.; Bolger, M. B.; Brinton, R. E.; Coirini, H.; McEwen,
B. S. Steroid modulation of the chloride ionophore in rat brain:
Structure-activity requirements, regional dependence, and
(4) (a) Lambert, J. J.; Belelli, D.; Hill-Venning, C.; Peters, J. A.
Neurosteroids and the GABAA receptor function. Trends Phar-
macol. Sci. 1995, 16, 95-303. (b) Lambert, J. J.; Peters, J. A.;
Harney, S.; Belelli, D. Steroid modulation of GABAA receptors.
In Handbook of Experimental Pharmacology; Mohler, H., Ed.;
Springer: Berlin, 2001, pp 117-140.

5214 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16
Souli et al.
mechanism of action. J. Pharmacol. Exp. Ther. 1988, 246, 803-
812. (b) Harrison, N. L.; Majewska, M. D.; Harrington, J. W.;
Barker, J. L. Structure-activity relationships for steroid inter-
action with the γ-aminobutyric acidA receptor complex. J.
Pharmacol. Exp. Ther. 1987, 241, 346-353.
(27) Maksay, G.; B´ıro´, T. Dual cooperative allosteric modulation of
binding to ionotropic glycine receptors. Neuropharmacology
2002, 43, 1087-1098.
(28) Corey, E. J.; Fuchs, P. L. A synthetic method for formyl f
ethynyl conversion (RCHO f RCCH or RCCR′) Tetrahedron Lett.
1972, 13, 3769-3772.
(19) (a) Han, M.; Zorumski, C. F.; Covey, D. F. Neurosteroid
analogues. 4. The effect of methyl substitution at the C-5 and
C-10 positions of neurosteroids on electrophysiological activity
at GABAA receptors. J. Med. Chem. 1996, 39, 4218-4232. (b)
Hogenkamp, D. J.; Tahir, S. H.; Hawkinson, J. E.; Upasani, R.
B.; Alauddin, M.; Kimbrough, C. L.; Acosta-Burruel, M.; Whit-
temore, E. R.; Woodward, R. M.; Lan, N. C.; Gee, K. W.; Bolger,
M. B. Synthesis and in vitro activity of 3â-substituted-3R-
hydroxypregnan-20-ones: allosteric modulators of the GABAA
receptor. J. Med. Chem. 1997, 40, 61-72. (c) Anderson, A.; Boyd,
A. C.; Byford, A.; Campbell, A. C.; Gemmell, D. K.; Hamilton,
N. M.; Hill, D. R.; Hill-Venning, C.; Lambert, J. J.; Maidment,
M. S.; May, V.; Marshall, R. J.; Peters, J. A.; Rees, D. C.;
Stevenson, D.; Sundaram, H. Anesthetic activity of novel water-
soluble 2â-morpholinyl steroids and their modulatory effects at
GABAA receptors. J. Med. Chem. 1997, 40, 1668-1681. (d)
Upasani, R. B.; Yang, K. C.; Acosta-Burruel, M.; Konkoy, C. S.;
McLellan, J. A.; Woodward, R. M.; Lan, N. C.; Carter, R. B.;
Hawkinson, J. E. 3R-Hydroxy-3â-(phenylethynyl)-5â-pregnan-
20-ones: synthesis and pharmacological activity of neuroactive
steroids with high affinity for GABAA receptors. J. Med. Chem.
1997, 40, 73-84. (e) Nilsson, K. R.; Zorumski, C. F.; Covey, D.
F. Neurosteroid analogues. 6. The synthesis and GABAA receptor
pharmacology of enantiomers of dehydroepiandrosterone sulfate,
pregnenolone sulfate, and (3R,5â)-3-hydroxypregnan-20-one sul-
fate. J. Med. Chem. 1998, 41, 2604-2613. (f) Anderson, A.; Boyd,
A. C.; Clark, J. K.; Fielding, L.; Byford, A.; Gemmell, D. K.;
Hamilton, N. M.; Maidment, M. S.; May, V.; McGuire, R.;
McPhail, P.; Sansbury, F. H.; Sundaram, H.; Taylor, R. Confor-
mationally constrained anesthetic steroids that modulate GABAA
receptors. J. Med. Chem. 2000, 43, 4118-4125. (g) Jiang, X.;
Manion, B. D.; Benz, A.; Rath, N. P.; Evers, A. S.; Zorumski, C.
F.; Mennerick, S.; Covey, D. F. Neurosteroid analogues. 9.
Conformationally constrained pregnanes: structure-activity
studies of 13,24-cyclo-18,21-dinorcholane analogues of the GABA
modulatory and anesthetic steroids (3R,5R)- and (3R,5â)-3-
hydroxypregnan-20-one. J. Med. Chem. 2003, 46, 5334-5348.
(20) Covey, D. F.; Han, M.; Kumar, A. S.; Antonette, M.; de la Cruz,
M.; Meadows, E. S.; Hu, Y.; Tonnies, A.; Nathan, D.; Coleman,
M.; Benz, A.; Evers, A. S.; Zorumski, C. F.; Mennerick, S.
Neurosteroid analogues. 8. Structure-activity studies of N-
acylated 17a-aza-D-homosteroid analogues of the anesthetic
steroids (3R,5R)- and (3R,5â)-3-hydroxypregnan-20-one. J. Med.
Chem. 2000, 43, 3201-3204.
(29) (a) Chemin, D.; Linstrumelle, G. Palladium-catalyzed reaction
of (E) and (Z)-dichloroethenes with 1-alkynes. An efficient
stereospecific synthesis of (E) and (Z)-enediynes. Tetrahedron
1994, 50, 5335-5344. (b) Alami, M.; Ferri, F.; Linstrumelle, G.
An efficient palladium-catalysed reaction of vinyl and aryl
halides or triflates with terminal alkynes. Tetrahedron Lett.
1993, 34, 6403-6406. (c) Sonogashira, K.; Tohda, Y.; Hagihara,
N. A convenient synthesis of acetylenes: catalytic substitutions
of acetylenic hydrogen with bromoalkenes, iodoarenes and
bromopyridines. Tetrahedron Lett. 1975, 16, 4467-4470.
(30) (a) Piatak, D. M.; Wicha, J. Various approaches to the construc-
tion of aliphatic side chains of steroids and related compounds.
Chem. Rev. 1978, 78, 199-241. (b) Chaudhuri, N. K.; Williams,
J. G.; Nickolson, R.; Gut, M. Stereochemistry of the addition
reactions of Grignard reagents to 20-keto steroids. Synthesis of
17R,20R-dihydrocholesterol. J. Org. Chem. 1969, 34, 3759-3766.
(c) Chaudhuri, N. K.; Gut, M. Ethynylation of 17.alpha.-
hydroxypregnenolone 3-acetate. Stereochemistry and acid-
catalyzed rearrangement of the reaction product. J. Org. Chem.
1969, 34, 3754-3758.
(31) Mauvais, A.; Burger, A.; Roussel, J. P.; Hetru, C.; Luu, B.
Acetylenic inhibitors of C-22 hydroxylase of ecsydone biosyn-
thesis. Bioorg. Chem. 1994, 22, 36-50.
(32) Singh, J.; Satyamurthi, N.; Aidhen, I. S. The growing synthetic
utility of Weinreb’s amide. J. Prakt. Chem. (Weinheim, Ger.)
2000, 342, 340-347.
(33) (a) Corey, E. J.; Helal, C. J. Reduction of carbonyl compounds
with chiral oxazaborolidine catalysts: A new paradigm for
enantioselective catalysis and a powerful new synthetic method.
Angew. Chem., Int. Ed. 1998, 37, 1986-2012. (b) Helal, C. J.;
Magriotis, P. A.; Corey, E. J. Direct catalytic enantioselective
reduction of achiral R,â-ynones. Strong remote steric effects
across the C-C triple bond. J. Am. Chem. Soc. 1996, 118,
10938-10939. (c) Corey, E. J.; Helal, C. J. Novel electronic effects
on remote substituents on the oxazaborolidine-catalysed enan-
tioselective reduction of ketones. Tetrahedron Lett. 1995, 36,
9153-9156. (d) Wallbaum, S.; Martens, J. Asymmetric syntheses
with chiral oxazaborolidines. Tetrahedron: Asymmetry 1992, 3,
1475-1504. (e) Corey, E. J.; Bakshi, R. K.; Shibata, S. J. Highly
enantioselective borane reduction of ketones catalyzed by chiral
oxazaborolidines. Mechanism and synthetic implications. J. Am.
Chem. Soc. 1987, 109, 5551-5553.
(34) Parker, K. A.; Lederboer, M. W. Asymmetric reduction. A
convenient method for the reduction of alkynyl ketones. J. Org.
Chem. 1996, 61, 3214-3217.
(21) (a) Hu, Y.; Zorumski, C. F.; Covey, D. F. Neurosteroid ana-
logues: structure-activity studies of benz[e]indene modulators
of GABAA receptor function. 1. The effect of 6-methyl substitu-
tion on the electrophysiological activity of 7-substituted benz-
[e]indene-3-carbonitriles. J. Med. Chem. 1993, 36, 3956-3967.
(b) Han, M.; Hu, Y.; Zorumski, C. F.; Covey, D. F. Neurosteroid
analogues. 3. The synthesis and electrophysiological evaluation
of benz[e]indene congeners of neuroactive steroids having the
5â-configuration. J. Med. Chem. 1995, 38, 4548-4556.
(35) Crabbe, P.; Andre, D.; Fillion, H. Synthesis of homo-dinordrin,
allenyl A-nor and dinor-steroids. Tetrahedron Lett. 1979, 10,
893-896. (b) Fillion, H.; Andre, D.; Luche, J. L. Mecanisme de
l’homologation directe d’acetyleniques en allenes. Tetrahedron
Lett. 1980, 21, 929-930. (c) Grabbe, P.; Nassim, B.; Robert-
Lopes, M. T. One-step homologation of acetylenes to allenes:
4-hydroxynona-1,2-diene. Org. Synth. 1984, 63, 203-205. (d)
Searles, S.; Li, Y.; Nassim, B.; Robert-Lopez, M. T.; Tran, P. T.;
Crabbe, P. Observation on the synthesis of allenes by homolo-
gation of alk-1-ynes. J. Chem. Soc., Perkin Trans. 1 1984, 747-
751.
(36) Carboni, E.; Wieland, S.; Lan, N. C.; Gee, K. W. Anxiolytic
properties of endogenously occurring pregnanediols in two rodent
models of anxiety. Psychopharmacology (Berlin) 1996, 126, 173-
178. (b) Gee, K. W.; McCauley, L. D.; Lan, N. C. A putative
receptor for neurosteroids on the GABAA receptor complex: the
pharmacological properties and therapeutic potential of epalons.
Crit. Rev. Neurobiol. 1995, 9, 207-227.
(22) Maksay, G.; Korpi, E. R.; Uusi-Oukari, M. Bimodal action of
furosemide on convulsant [3H]EBOB binding to cerebellar and
cortical GABAA receptors. Neurochem. Int. 1998, 33, 353-358.
(23) Perret, P.; Sarda, X.; Wolff, M.; Wu, T. T.; Bushey, D.; Goeldner,
M. Interaction of noncompetitive blockers within the γ-ami-
nobutyric acid type A chloride channel using chemically reactive
probes as chemical sensors for cysteine mutants. J. Biol. Chem.
1999, 274, 25350-25354.
(24) Maksay, G. Commentary. From kinetics and thermodynamics
of GABAA receptor binding to ionophore function. Neurochem.
Int. 1996, 29, 361-370.
(25) (a) Belelli, D.; Lambert, J. J.; Peters, J. A.; Gee, K. W.; Lan, N.
C. Modulation of human recombinant GABAA receptors by
pregnanediols. Neuropharmacology 1996, 35, 1223-1231. (b)
McCauley, L. D.; Liu, V.; Chen, J. S.; Hawkinson, J. E.; Lan, N.
C.; Gee, K. W. Selective actions of certain neuroactive preg-
nanediols at the gamma-aminobutyric acid type A receptor
complex in rat brain. Mol. Pharmacol. 1995, 47, 354-362.
(26) Ehlert, F. J. Estimation of the affinities of allosteric ligands using
radioligand binding and pharmacological null methods. Mol.
Pharmacol. 1988, 33, 187-194.
(37) (a) Belelli, D.; Gee, K. W. 5R-Pregnan-3R,20R-diol behaves like
a partial agonist in the modulation of GABA-stimulated chloride
in uptake by synaptoneurosomes. Eur. J. Pharmacol. 1989, 167,
173-176. (b) Bolger, M. C.; Gee, K. W.; Lan, N. T.-Y.; Belelli,
D.; Mirsadeghi, S.; Purdy, R. H.; Tahir, S. H. GABA receptor
modulators. EP0554436, 1993.
JM050271Q