3α-Hydroxy-3β-(phenylethynyl)-5β-pregnan-20-ones: Synthesis and pharmacological activity of neuroactive steroids with high affinity for GABA(A) receptors

Article abstract of DOI:10.1021/jm9605344

Neuroactive steroids that allosterically modulate GABA(A) receptors have potential uses as anticonvulsants, anxiolytics, and sedative-hypnotic agents. Recently, a series of pregnanes substituted with simple alkyl groups at the 3β-position were synthesized

Full text of DOI:10.1021/jm9605344

J . Med. Chem. 1997, 40, 73-84
73
3r-Hyd r oxy-3â-(p h en yleth yn yl)-5â-p r egn a n -20-on es: Syn th esis a n d
P h a r m a cologica l Activity of Neu r oa ctive Ster oid s w ith High Affin ity for GABAA
Recep tor s
Ravindra B. Upasani,* Kevin C. Yang, Manuel Acosta-Burruel, Chris S. Konkoy, J ames A. McLellan,
Richard M. Woodward, Nancy C. Lan, Richard B. Carter, and J on E. Hawkinson
CoCensys, Inc., 213 Technology Drive, Irvine, California 92618
Received J uly 24, 1996X
Neuroactive steroids that allosterically modulate GABAA receptors have potential uses as
anticonvulsants, anxiolytics, and sedative-hypnotic agents. Recently, a series of pregnanes
substituted with simple alkyl groups at the 3â-position were synthesized and found to be active
in vitro. The present report describes the synthesis of a series of substituted 3R-hydroxy-3â-
(phenylethynyl)pregnan-20-ones and their in vitro structure-activity relationship determined
by their potency for inhibition of [35S]TBPS binding in rat brain membranes. Appropriate
substitution of the phenyl group results in ligands with particularly high affinity for the
neuroactive steroid site on GABAA receptors (e.g., 4-acetyl 28, IC50 10 nM). The potency of
selected steroids was confirmed electrophysiologically in oocytes expressing cloned human
GABAA R1â2γ2L receptors (e.g., compound 28, EC50 6.6 nM). Consistent with their in vitro
activity, some of the 3â-(phenylethynyl)-substituted steroids displayed anticonvulsant activity
in the pentylenetetrazol (PTZ) and maximal electroshock (MES) tests following ip administration
in mice. Notably, the 3â-[(4-acetylphenyl)ethynyl]-19-nor derivative 36 demonstrated an
attractive anticonvulsant profile (PTZ and MES ED50 values of 2.8 and 9.2 mg/kg, respectively).
A new pharmacophore for the neuroactive steroid site of GABAA receptors is proposed based
upon the high affinity of certain substituted 3â-(phenylethynyl) steroids.
In tr od u ction
occurring progesterone metabolite 3R-hydroxy-5R-preg-
nan-20-one (3R,5R-P).
γ-Aminobutyric acidA (GABAA) receptors are ligand-
gated, chloride-permeable ion channels that mediate
most of the fast postsynaptic inhibitory currents in
mammalian brain.1 GABAA receptor modulators, such
as barbiturates and benzodiazepines, have been used
extensively as anxiolytics, anticonvulsants, sedatives,
and hypnotics.1 More recently, additional GABAA re-
ceptor modulators, such as non-benzodiazepines acting
at the central benzodiazepine site and loreclezole, have
demonstrated the continued utility of this drug target.1
In the present report, we describe the synthesis and
activity of a series of novel 3â-substituted steroids with
in vitro potencies significantly greater than that of
3R,5R-P. Consistent with their high in vitro potencies,
certain of these steroids are potent anticonvulsants in
the pentylentetrazol (PTZ) and maximal electroshock
(MES) seizure models. The high potency of certain of
these 3â-substituted steroids necessitated a reevalua-
tion of neuroactive steroid SAR, resulting in the de-
scription of a new pharmacophore for the steroid binding
site on GABAA receptor complexes.
Following the discovery of the modulatory effects of
neuroactive steroids on GABAA receptors,2 considerable
effort has been made toward characterizing their unique
site on the receptor complex, their physiological and
pharmacological effects, and evaluation of the feasibility
of generating GABA-potentiating drugs with improved
side-effect profiles.3 Interactions of neuroactive steroids
with the receptor complex have been explored primarily
using radioligand binding, electrophysiological, and
chloride flux techniques. Previous in vitro4-9 and in
vivo10-12 structure-activity relationship studies indi-
cated that 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,
although certain other 17â-substitutions such as 17â-
CN retain activity.6,9,12 Numerous modifications to the
steroid nucleus have been made in order to study the
SAR of neuroactive steroids, including substitutions at
the 3â-, 11-, 17-, or 21-positions.4-9 None of these
modifications have resulted, however, in a compound
with appreciably higher potency than the naturally
Ch em istr y
3R-Hydroxy-3â-phenyl-5â-pregnan-20-one (1) was syn-
thesized by adding phenylmagnesium bromide to the
pregnan-3,20-dione derivative (2)7 and purifying the
epimeric mixture of 3R-hydroxy and 3â-hydroxy deriva-
tives by column chromatography. Similarly, 3â-phenyl
compounds with a spacer (3, 4, and 6) were prepared
by the addition of benzylmagnesium chloride, lithium
phenyl acetylide, or (phenylpropynyl)lithium, respec-
tively, to 2 as outlined in Scheme 1. The 3â-(phenyl-
ethynyl) (4) and 3â-(phenylpropynyl) (6) derivatives
were then hydrogenated in the presence of Pd/C to
afford the corresponding 3â-(phenylethyl) (5) and 3â-
(phenylpropyl)pregnanes (7).
The synthesis of 3â-substituted phenylethynyl deriva-
tives is depicted in Scheme 2. The Wittig reaction of
the corresponding benzaldehyde (8) with carbon tetra-
bromide in the presence of 2 equiv of triphenylphosphine
afforded the intermediate phenyldibromoethene (9).13
The dibromoethene derivative was then allowed to react
* Author for correspondence.
X Abstract published in Advance ACS Abstracts, December 15, 1996.
S0022-2623(96)00534-1 CCC: $14.00 © 1997 American Chemical Society

74 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 1
Upasani et al.
Sch em e 1a
a
Reagents: (a) (i) PhCH2MgCl, PhCtCLi, or PhCH2CtCLi, THF, -70 °C to room temperature, (ii) 2 N HCl/acetone, room temperature;
(b) Pd/C, H2, EtOAc, room temperature.
Sch em e 2a
a
Reagents: (a) CBr4/PPh3/CH2Cl2, or CBr4/PPh3/Zn dust/CH2Cl2, 0-20 °C; (b) (i) 2 equiv of n-BuLi/THF, (ii) 2/THF, -70 °C to room
temperature; (iii) 2 N HCl/acetone, room temperature; (iv) chromatography; (c) H2O2/K2CO3/DMSO, room temperature.
Sch em e 3a
a
Reagents: (a) (PPh3)2PdCl2/CuI/CH2Cl2/(NEt3 or HNEt2), room temperature.
with 2 equiv of n-butyllithium to generate, in situ,
lithium phenyl acetylide, which was then reacted with
the ketone 2 to give, after hydrolysis of the 20-ketal,
the corresponding phenylethynyl derivatives (10-20).
An alternate synthetic route was employed to prepare
certain 3â-(phenylethynyl) derivatives as outlined in
Scheme 3. Different iodo- or bromobenzenes (21) were
coupled with 3â-ethynyl derivative (22)7 in the presence
of catalytic amounts of (PPh3)2PdCl2 and CuI to get the
corresponding 3â-(phenylethynyl) derivatives (1, 10-
18, 20, 23-33).14
3â-hydroxy epimers. However, the desired more polar
3R-hydroxy derivatives, in which the hydroxy group is
equatorial (lower TLC Rf), were easily separated from
the less polar 3â-hydroxy (axial, higher TLC Rf) epimers
by flash chromatography using toluene:acetone mix-
tures. Although the stereochemistry of the 3-hydroxy
group was established on the basis of the polarity (TLC
Rf) of the epimers, it was further confirmed by synthe-
sizing unambiguous authentic samples following the
coupling method described in the Scheme 3. (See the
Experimental Section).
Similar coupling of iodobenzenes (34) with the 19-nor
derivative of 22 (compound 35) in the presence of Pd
catalyst afforded the 3R-hydroxy-3â-(substituted phen-
ylethynyl)-5â-19-norpregnanes (36 and 37) in 20-30%
yields (Scheme 4).
Some of the 3â-phenyl compounds (1, 3, 4, 6, 10-13,
15-20) obtained by the addition reaction of the ap-
propriate organometallic reagents with 2, as outlined
in Schemes 1 and 2, were found to be contaminated with
Resu lts a n d Discu ssion
Syn th etic Ra tion a le. Naturally occurring neuro-
active steroids have therapeutic limitations because
they are rapidly metabolized, presumably by conjuga-
tion of the 3R-hydroxy group or oxidation to the corre-
sponding ketone, rendering them inactive as modulators
of GABAergic function. The oxidation of the 3R-hydroxy
group can be blocked by the addition of a substituent

Neuroactive Steroids with High Affinity for GABAA Receptors
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 1 75
Sch em e 4a
a
Reagents: (a) (PPh3)2PdCl2/CuI/CH2Cl2/NEt3, room temperature.
Ch a r t 1
Ta ble 1. Inhibitory Potency in the TBPS Binding Assay and
Anticonvulsant Activity of 3â-Substituted Pregnane Derivatives
Sch em e 5
at the 3â-position.3 Recently, a series of pregnanes
substituted with simple alkyl groups at the 3â-position
were synthesized and found to be active in vitro, as
indicated by their activity as allosteric inhibitors of the
binding of the chloride channel-associated radioligand
[35S]-tert-butylbicyclophosphorothionate (TBPS) in rat
brain cortical membranes.7 During the synthesis of the
potent 3â-ethynyl-3R-hydroxy-5â-pregnan-20-one (22,
TBPS IC50 39 nM),7 a byproduct with a dimeric struc-
ture (38, Chart 1) was isolated. Although this steroid
derivative had relatively low potency in the TBPS assay
(IC50 2.9 µM),17 even this level of activity was surprising
considering the bulk of the 3â-substituent. Encouraged
by these results it was decided to add a comparatively
smaller phenyl group at the 3â-position. Although more
potent than the dimeric compound 38, substitution at
the 3â-position with a phenyl group (1, IC50 380 nM)
resulted in a 10-fold loss in affinity compared to the
corresponding 3â-methyl compound (39, IC50 37 nM).7
Syn th esis a n d Eva lu a tion of Sp a cer Un its. The
relatively low activities of 1 and 38 suggested that bulky
substituents adjacent to the A ring make contact with
a forbidden volume in the binding site. Therefore, a
further modification of the 3â-position was carried out
by adding a spacer between the phenyl ring and the
steroid nucleus to evaluate whether a bridging unit
could provide enhanced potency (Scheme 5). The spac-
ers used were flexible or rigid chains of 1-3 carbon
atoms. The resulting phenyl derivatives with a spacer
were more active than the phenyl compound 1 without
the spacer (Table 1). Although large differences in
potency were not observed with the different spacers
examined, the affinity was generally higher with spacers
containing unsaturation and was lower as the phenyl
ring was moved further away from the steroid nucleus.
The optimal spacer was found to be the two-carbon rigid
ethynyl group (compound 4, IC50 100 nM). These
PTZ
compd
Z
IC50 (nM)a
Imax (%)
protection (%)b
1
3
4
5
6
7
380 ( 50
110 ( 10
100 ( 10
150 ( 20
131 ( 4
89 ( 2
101 ( 3
100 ( 1
102 ( 7
92 ( 5
25
38
25
0
13
13
CH2
CtC
(CH2)2
CtCCH2
(CH2)3
190 ( 40
91 ( 2
a
Compounds were incubated with rat brain cortical P2 mem-
branes and 2 nM [35S]TBPS in the presence of 5 µM GABA as
described in the Experimental Section. Values are means and
SEMs of three or four independent experiments. Hill values were
1.0 for all compounds. The PTZ protection is the percent of
animals (n ) 8) not exhibiting PTZ-induced seizures 10 min
b
following a 10 mg/kg ip dose of steroid.
results suggested the presence of an auxiliary binding
pocket for hydrophobic groups in the neuroactive steroid
binding site located close to the position occupied by the
steroid A ring.
St r u ct u r e-Act ivit y R ela t ion sh ip s of P h en yl-
eth yn yls. In order to probe the nature of the binding
pocket, a set of substituents on the phenyl ring of 3â-
substituted steroids containing the ethynyl spacer was
selected to provide systematic variation in lipophilic,
electronic, steric, and hydrogen-bonding properties (Table
2). The activity of these 3â-(phenylethynyl) compounds
in the TBPS binding assay was found to be highly
dependent on the substitution on the phenyl ring.
Substitution with an electron-withdrawing group at the
4-position of phenyl ring increased the affinity, e.g.,
4-CN (13, IC50 60 nM), 4-CF3 (27, IC50 56 nM), and
4-NO2 (17, IC50 42 nM). Furthermore, substitution at
the 4-position on the phenyl ring with electron-with-
drawing groups capable of accepting hydrogens to form
hydrogen bonds gave the most potent compounds with
the IC50 values ∼10 nM, e.g., 4-COMe (28, 10 nM),
4-CO2Et (29, 12 nM). Substitution with electron-
donating groups capable of donating hydrogen to form

76 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 1
Upasani et al.
Ta ble 2. Inhibitory Potency in the TBPS Binding Assay and
Anticonvulsant Activity of 3â-(Phenylethynyl)pregnane
Derivatives
PTZ
protection
compd
R
IC50 (nM)a
Imax (%)
(%)
4
10
11
12
13
14
15
16
17
18
19
20
23
H
100 ( 10
42 ( 4
100 ( 1
84 ( 5
88 ( 5
78 ( 5
85 ( 10
96 ( 1
90 ( 3
100 ( 4
98 ( 7
89 ( 2
102 ( 2
93 ( 3
30 ( 6
57 ( 11
100 ( 1
84 ( 4
17 ( 2
84 ( 1
88 ( 4
106 ( 2
97 ( 4
44 ( 14
54 ( 13
34 ( 3
65 ( 3
91 ( 1
96 ( 4
97 ( 2
94 ( 1
25
25
0
50
44
13
0
0
13
0
4-OMe
4-Cl
4-NMe2
4-CN
4-CONH2
4-Ph
4-OH
70 ( 15
40 ( 5
60 ( 9
1800 ( 400
54 ( 12
2300 ( 200
42 ( 9
4-NO2
2-OMe
3,4-(OMe)2
3,4-OCH2O
2-OH
170 ( 20
290 ( 10
68 ( 2
82 ( 31
5100 ( 500
420 ( 100
83 ( 16
96 ( 48
11800 ( 1800
56 ( 7
10 ( 2
12 ( 3
67 ( 44
1600 ( 200
45 ( 9
13
13
0
F igu r e 1. Functional modulation of human GABAA R1â2γ2L
receptors by 3â-(phenylethynyl)pregnane derivatives. Top:
Potentiation of GABA-activated currents by steroids. Data are
plotted as mean values ( SEM, expressed as a fraction of the
maximum current elicited by GABA. FR, fractional response.
Bottom: Sample recordings comparing potentiation of GABA-
activated currents by 28, 3R,5R-P, diazepam, and sodium
pentobarbital. Records are from a single oocyte using a holding
potential of -70 mV. Drugs were applied as indicated by bars.
The control GABA concentration was 8 µM. A train of depo-
larizing pulses (+10 mV, upward deflections) was used to time
drug applications and monitor membrane conductance. Scale
bars represent 100 nA and 1 min.
24
25
26
3-OH
4-Me
4-NH2
0
50
13
27
28
29
30
4-CF3
0
75
19
0
4-COMe
4-CO2Et
4-CHO
31
3-COMe
0
1500 ( 300
93 ( 13
55 ( 11
12 ( 3
32
33
36
37
4-CONEt2
4-COPh
4-COMe, 19-nor
4-CO2Et, 19-nor
13
38
100
13
26 ( 6
a
See Table 1 and Experimental Section for TBPS binding and
PTZ protection assays. Values listed for compounds 23, 26, 30,
and 31 are the high- and low-affinity components of the two-
component inhibition curves. Hill values for one component
inhibitors were 1.0 except for compounds 10 (1.21 ( 0.08), 12 (1.26
( 0.17), 18 (1.25 ( 0.06), 24 (0.83 ( 0.06), and 32 (1.38 ( 0.12).
Values are means and SEMs of three to five independent experi-
ments.
hydrogen bonds rendered these compounds almost inac-
tive, e.g., 4-OH (16, IC50 2300 nM), 4-NH2 (26, IC50
11 800 nM, predominant low-affinity component). Al-
though, the 4-OH substitution decreased the activity, a
gradual increase in activity was observed in moving the
hydroxy group closer to the steroid nucleus. Hence, the
3-hydroxy derivative 24 was about 5-fold more potent
than the 4-hydroxy-substituted compound 16. Although
complicated by a complex two component inhibition, the
2-hydroxy-substituted compound 23, with the hydroxy
group adjacent to the steroid nucleus, inhibited TBPS
binding at lower concentrations, with a high-affinity
component 28-fold more potent than compound 16
(Table 2).
Thus, addition of a hydrogen bond accepting group
at the 4-position of the phenyl ring generally increased
the affinity and addition of a hydrogen bond donating
group decreased the affinity. As predicted, replacing
the hydrogens of the hydrogen bond donor hydroxy and
amino groups of the relatively inactive 16 and 26 with
alkyl groups to give compounds with hydrogen bond
accepting methoxy (compound 10) and dimethylamino
F igu r e 2. Inhibition of [35S]TBPS binding in rat brain
membranes by 3â-(phenylethynyl)pregnane derivatives: Com-
parison of the potent inhibitor 28 with compounds displaying
complex binding curves. Concentrations of 28 > 1 µM resulted
in an apparent decrease in inhibition of TBPS binding,
presumably due to insolubility of this steroid at these concen-
trations; these data points were not included in the curve fit.
Steroids were incubated with 2 nM [35S]TBPS for 90 min at
room temperature in the presence of 5 µM GABA.
(compound 12) groups, respectively, resulted in several-
fold increases in potency (Scheme 6).
Further evidence for the importance of the hydrogen
bond accepting capability of the 4-substituent was
obtained with the low-affinity compound 14, R )
4-CONH2. Since, the CONH2 group can act as a
hydrogen bond acceptor or donor, it was predicted that
eliminating the hydrogen bond donating capability of
the group should result in an active compound. To test
this theory, the N,N-diethylcarboxamido derivative 32,
which does not have hydrogen bond donating capability,
was synthesized and found to be 19 times more active
than its parent (Scheme 7), providing additional support

Neuroactive Steroids with High Affinity for GABAA Receptors
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 1 77
Sch em e 6
Sch em e 7
Ta ble 3. Functional Modulation of GABAA Receptors by
3â-(Phenylethynyl)pregnane Derivatives
for the importance of hydrogen bond accepting ability
of 4-substituents for receptor potency.
Electr oph ysiology. Functional modulation of GABAA
receptors by selected steroids was evaluated electro-
physiologically in Xenopus oocytes expressing cloned
human R1â2γ2L GABAA receptor complexes. Oocytes
injected with a mixture of R1, â2, and γ2L cRNAs (∼1
ng each) produced strong expression of functional
GABAA receptors. The range of maximal GABAA-
activated currents (GABAMAX) was between 2200 and
4470 nA, and the mean maximal response was 2820 (
130 nA (n ) 20). GABA concentration-response curves
were measured in a sample of oocytes. The EC50 value
was 22 ( 2.2 µM, and the slope was 1.0 ( 0.01 (n )
13), consistent with a pharmacologically homogeneous
population of receptors.
efficacy
compd
R
EC50 (nM)a
slope
(FR)
n
12
16
23
26
28
30
36
4-NMe2
4-OH
2-OH
18 ( 1.5
>1000
340 ( 34
> 1000
6.6 ( 0.69
31 ( 5.8
8.2 ( 1.4
1.1 ( 0.53
0.76 ( 0.002
4
4
3
3
3
4
3
1.6 ( 0.11
0.51 ( 0.029
4-NH2
4-COMe
4-CHO
4-COMe
1.2 ( 0.07
1.3 ( 0.20
1.4 ( 0.20
0.94 ( 0.010
0.88 ( 0.023
0.91 ( 0.06
(19-nor)
Modulation of GABA responses by a selected group
of steroids (compounds 12, 16, 23, 26, 28, 30, 36) is
shown in Figure 1, top. Modulation was determined by
adjusting the GABA concentration to produce a current
approximately 5% of the GABAMAX. The mean GABA
concentration required to elicit the 5% response was 4.1
( 1.9 µM, and the mean response was 0.048 ( 0.001,
expressed as a fraction of GABAMAX (n ) 20). Modula-
tion was measured as the potentiation of the control
GABA response combined with any current activated
directly by the steroids. Such steroid-activated currents
were small, ranging from 0 to 10% of GABAMAX and
a
Steroid potentiation of 5% GABA responses in Xenopus oocytes
expressing R1â2γ2L subunits of human recombinant GABAA
receptors as described in the Experimental Section.
were only appreciable when applying micromolar con-
centrations of steroids.
The 3â-(4-acetylphenyl)ethynyl derivatives 28 and 36
were potent, full efficacy modulators of GABA-activated
responses (Table 3). The 4-(dimethylamino) and 4-al-
dehydo analogues 12 and 30, respectively, were several-
fold less potent than 28 and 36. The 4-(dimethylamino)
analogue 12 might also have been less efficacious.

78 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 1
Upasani et al.
The 2-hydroxyl analogue 23 was approximately 10-fold
less potent than 12 and 30 and demonstrated interme-
diate efficacy modulation, whereas the 4-hydroxyl and
4-amino analogues 16 and 26, respectively, were weak
potentiators. The measurement of maximal fractional
responses for 16 and 26 was hindered by solubility
limitations.
sentative high-affinity analogs with marked in vivo
activity, compounds 28 and 36, were evaluated for
anticonvulsant potency against PTZ- and MES-induced
seizures following intraperitoneal administration. The
neuroactive steroid 28, administered intraperitoneally
in 50% HP-â-CD 30 min prior to the convulsant stimu-
lus, produced a dose-dependent blockade of seizures
with an ED50 of 8.2 mg/kg (95% confidence interval 5.1-
13.3) against PTZ and an ED50 of 7.4 mg/kg (95% CI
5.2-10.5) against MES. Compound 36 exhibited a
different in vivo potency profile with ED50 values of 2.8
mg/kg (1.7-4.5) against PTZ and 9.2 mg/kg (6.0-14.0)
against MES. The degree of neurobehavioral impair-
ment produced by 28 and 36 was evaluated using the
hanging-wire test. Compound 28 produced a dose-
dependent increase in hanging-wire failure with a TD50
of 10.0 mg/kg (6.3-15.7), whereas 36 displayed a TD50
of 13.2 mg/kg (9.7-17.9). The dose at which 28 pro-
duced ataxia was only 1.2 times greater than the dose
required for anticonvulsant efficacy against PTZ, whereas
for compound 36, the difference was 4.7. Thus, of the
pair, the 19-nor compound 36 congener displays phar-
macological properties better suited to development as
a potential antiepileptic medication. The generality of
this finding remains to be determined.
The anticonvulsant profile of compound 36 is similar
to that reported previously for the endogenous neuro-
active steroids 3R,5R-P and 3R-hydroxy-5â-pregnan-20-
one (3R,5â-P).15 Endogenous neurosteroids, such as
3R,5R-P and 3R,5â-P, are unsuitable as therapeutic
agents, however, as they are readily oxidized and/or
conjugated at the 3R position,12 resulting in compounds
that are inactive at GABAA receptors, but potentially
active at hormonal steroid receptors.4-6 3â-Substitu-
tion, which in part prevents metabolism of the 3R-
hydroxy group, has been suggested to enhance the
bioavailability of pregnane steroids without altering
their primary pharmacologic properties.3,7 Indeed, ga-
naxolone (CCD 1042; 3R-hydroxy-3â-methyl-5R-preg-
nan-20-one), a 3â-methylated synthetic analogue of the
endogenous neuroactive steroid 3R,5R-P, has been shown
to retain the anticonvulsant activity of the endogenous
steroid 3R,5R-P, while acquiring oral bioavailability and
other properties that would be expected to facilitate its
use as an antiepileptic drug.16 Thus, the 3â-[(4-
acetylphenyl)ethynyl]-5â-19-nor derivative 36 would be
predicted to be of similar therapeutic utility.
In these types of assays, 0.5 nM 28 produced a level
of potentiation of GABA responses comparable to a 60-
fold higher concentration of the prototypical neuroster-
oid 3R,5R-P, a 200-fold higher concentration of the
benzodiazepine diazepam, and a 20 000-fold higher
concentration of the barbiturate sodium pentobarbital
(Figure 1, bottom).
Cor r ela tion of in Vitr o Assa ys. The rank order of
potency for potentiation of GABA-evoked currents in
Xenopus oocytes expressing R1â2γ2L receptors was 28
g 36 >12 > 30 > 23 > 26 ∼ 16. Similarly, the rank
order potency for inhibition of TBPS binding in rat brain
membranes for this set of compounds was 28 g 36 >
12 g 30 (high-affinity component) g 23 (high-affinity
component) > 16 > 26 (low-affinity component). The
good correlation between electrophysiological and bind-
ing assays was expected based on previous studies6 and
indicates that allosteric inhibition of TBPS binding is a
useful tool to develop steroid structure-activity rela-
tionships. However, complex binding curves were ob-
served for four compounds in this study (23, 26, 30, and
31) (Figure 2). Three of these were evaluated electro-
physiologically (23, 26, and 30). Indeed, these com-
pounds were selected for functional confirmation of
activity because of their unusual inhibition curves in
the binding assay. For compounds 23 and 30, the high-
affinity component in the binding assay correlates more
closely to functional activity in the electrophysiological
assays, whereas the low-affinity component for com-
pound 26 appears most relevant. These complex bind-
ing curves may be due to receptor subtype selectivity,
negative cooperativity, partial agonism, multiple sites
or mechanism of action, and/or combinations of these
factors. In principle, evaluation of the [35S]TBPS inhi-
bition curves for these compounds in membranes from
cell lines expressing recombinant receptor complexes
may help to resolve this question, although this was not
addressed in the present study. In any case, steroids
displaying complex binding curves that have been
examined electrophysiologically all show unremarkable
concentration-modulation curves. In practical terms,
these results indicate that compounds with unusual
profiles in the binding assay should be evaluated
functionally to confirm activity.
P h a r m a cop h or e Mod el
Previous studies4-9 have demonstrated that a hydro-
gen bond donating a 3R-hydroxyl group is an essential
element for the primary docking of the ligand to the
receptor. Replacing the hydrogen of the hydroxyl with
methyl, thus eliminating the ability of the steroid to
donate a hydrogen to form a hydrogen bond in this
region, results in 250-fold reduction in potency.17 On
the basis of the SAR of several 3â-(phenylethynyl)-
pregnane derivatives, we propose that an additional
hydrogen bond accepting group (e.g., carbonyl group of
an acetylphenyl or carbethoxyphenyl) located at a
certain distance from the A ring of the steroid can take
part in secondary docking giving rise to a tightly bound
ligand-receptor complex. Thus, the 3R-hydroxy group
is an absolute requirement for binding (primary dock-
ing), whereas neuroactive steroid affinity can be aug-
mented via an additional hydrogen bonding interaction
In Vivo P h a r m a cology. Preliminary in vivo phar-
macological testing was conducted using the PTZ-
induced seizure model. Compounds were administered
in 50% hydroxypropyl-â-cyclodextrin (HP-â-CD) at a
dose of 10 mg/kg, ip, 10 min prior to PTZ injection.
These data, which are presented in Tables 1 and 2, belie
a simple relationship between in vitro affinity and in
vivo potency and/or efficacy. For example, although
compounds 29 and 36 exhibit identical IC50 values (12
nM), 36 produces 100% protection against PTZ-induced
seizures at 10 mg/kg, whereas 29 yields only 19%
protection. Poor bioavailability and/or metabolic lability
may contribute to these discrepancies, although these
possibilities were not further investigated. Two repre-

Neuroactive Steroids with High Affinity for GABAA Receptors
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 1 79
as reference. Melting points were determined on a Thomas-
Hoover capillary melting point apparatus and are uncorrected.
Elemental analyses were performed by Galbraith Laboratories,
Knoxville, TN, or Robertson Microlit Laboratories, Madison,
NJ . Elemental analyses for the undesired side products, 3â-
hydroxy epimers, were not obtained. Flash chromatography
on silica gel (230-400 mesh, Mallinckrodt) was carried out as
described by Still.18 HPLC grade solvents were obtained from
Baxter. THF and ether were distilled from sodium and
benzophenone. Methylene chloride, Et3N, and Et2NH were
dried over CaH2 and distilled. Other synthetic reagents were
obtained from Aldrich Chemical Co. (Milwaukee, WI) and used
as received unless otherwise specified. 3â-Ethynyl-3R-hy-
droxy-5â-pregnan-20-one and 5â-pregnane-3,20-dione cyclic 20-
(1,2-ethanediyl acetal) were obtained from DioSynth BV (The
Netherlands). γ-Aminobutyric acid (GABA) was purchased
from Sigma Chemical Co. (St. Louis, MO). [35S]TBPS (60-
100 Ci/mmol) was obtained from New England Nuclear
(Boston, MA), and unlabeled TBPS was from Research Bio-
chemicals International (Natick, MA).
F igu r e 3. A schematic pharmacophore model for the neuro-
active steroid binding site of GABAA receptors.
(secondary docking). This secondary docking may be
the basis for the exceptionally high in vitro activity of
some of the phenylethynyl derivatives. For example,
compound 28 is 5- or 24-fold more potent than 3R,5R-P
(IC50 51 nM or EC50 160 nM)7 in TBPS binding or
electrophysiological assays, respectively.
3r-Hyd r oxy-3â-p h en yl-5â-p r egn a n -20-on e (1). A solu-
tion of 5â-pregnane-3,20-dione cyclic 20-(1,2-ethanediyl acetal)7
(2, 720 mg, 2 mmol) in dry THF (20 mL) was treated with
phenylmagnesium bromide (3M in THF, 6 mmol, 2 mL) at -70
°C. After the mixture was stirred at this temperature for 3 h
and then at room temperature for 2 h, it was quenched with
2 N HCl (10 mL). The solvent was removed, and the residue
was dissolved in acetone (20 mL). After 1 N HCl (5 mL) was
added, the solution was stirred at room temperature for 0.5
h. Saturated NaHCO3 solution was added to neutralize the
acid. The solvents were removed, and the residue was
extracted with CH2Cl2. The organic layer was washed with
water, dilute NaHCO3 solution, water, and brine. After being
dried over anhydrous MgSO4, the solution was filtered and
evaporated to yield the crude product (1.3 g). This crude
product was then dissolved in a small amount of CH2Cl2 and
poured on a column of silica gel. On elution with toluene:
acetone mixture (95:5) gave 3â-hydroxy-3R-phenyl-5â-pregnan-
20-one (420 mg) as a first fraction: TLC Rf 0.52 (toluene:
acetone, 95:5); mp 186-191 °C; 1H NMR (CDCl3) δ 7.21-7.61
(m, 5H), 2.58 (m, 1H), 2.13 (s, 3H), 1.04 (s, 3H), 0.63 (s, 3H).
Further elution with the same solvent mixture yielded 3R-
hydroxy-3â-phenyl-5â-pregnan-20-one (1, 185 mg): TLC Rf
0.45 (toluene:acetone, 95:5); mp 182-184 °C; IR 2938, 2868,
From the above observations it is possible to propose
a pharmacophore model for the binding site. The
schematic of the model is shown in Figure 3. The model
entails a critical primary docking through hydrogen-
bonding interactions at A and B and a secondary
docking through a strong hydrogen bonding interaction
at C. Although region B may well involve a hydrogen
bond donor group in the binding pocket, it should be
pointed out that certain other side groups besides acetyl
(in particular CN) also provide for high affinity primary
docking.6,9 Although the pharmacophore model pre-
sented is useful in designing steroid derivatives with
high in vitro potency, it does not predict their in vivo
activities. Future work will focus on use of this model
in conjunction with other approaches to design potent
compounds with improved in vivo profiles.
1687 cm-1 1H NMR (CDCl3) δ 7.26-7.62 (m, 5H), 2.56 (m,
;
1H), 2.13 (s, 3H), 0.84 (s, 3H), 0.60 (s, 3H). Anal. (C27H38O2)
C, H.
Con clu sion s
The present report describes the synthesis of a series
of substituted 3R-hydroxy-3â-(phenylethynyl)pregnan-
20-ones and their in vitro structure-activity relation-
ship determined by their potency for inhibition of
[35S]TBPS binding in rat brain membranes. Appropri-
ate substitution of the phenyl group results in ligands
with particularly high affinity for the neuroactive
steroid site on GABAA receptors (e.g., 4-acetyl 28, IC50
10 nM). The potency of some of these steroids was
confirmed electrophysiologically in oocytes expressing
R1â2γ2L receptors (e.g., compound 28, EC50 6.6 nM).
Consistent with their in vitro activity, some of these 3â-
(phenylethynyl)-substituted steroids displayed anticon-
vulsant activity in the pentylenetetrazol (PTZ) and
maximal electroshock (MES) tests following ip admin-
istration in mice. Notably, the 3â-(4-acetylphenyl)-19-
nor derivative 36 provided an attractive anticonvulsant
profile (PTZ and MES ED50 values of 2.8 and 9.2 mg/
kg, respectively). A new pharmacophore for the neu-
roactive steroid site of GABAA receptors is proposed
based upon the high affinity of certain substituted 3â-
(phenylethynyl) steroids.
3â-Ben zyl-3r-h yd r oxy-5â-p r egn a n -20-on e (3). A solu-
tion of 5â-pregnane-3,20-dione, cyclic 20-(1,2-ethanediyl
acetal)7 (2, 360 mg, 1 mmol) in dry THF (20 mL) was added
dropwise to a solution of benzylmagnesium chloride (2 M in
THF, 3 mmol, 1.5 mL) in dry THF (20 mL) at -70 °C. After
the mixture was stirred at this temperature for 40 min and
then at room temperature for 2.5 h, it was quenched with
saturated NH4Cl (2 mL). The solvent was removed, and the
residue was dissolved in acetone (20 mL). After 2 N HCl (2
mL) was added, the solution was stirred at room temperature
for 20 min. Saturated NaHCO3 solution was added to neutral-
ize the acid. The solvents were removed, and the residue was
extracted with CH2Cl2. The organic layer was washed with
water, dilute NaHCO3 solution, water, and brine. After being
dried over anhydrous MgSO4, the solution was filtered and
evaporated to yield the crude product (400 mg). This crude
product was then dissolved in a small amount of CH2Cl2 and
poured on a column of silica gel. On elution with toluene:
acetone mixture (97:3) gave 3R-benzyl-3â-hydroxy-5â-pregnan-
20-one (250 mg) as a first fraction: TLC Rf 0.57 (toluene:
1
acetone, 9:1); mp 150-152 °C; H NMR (CDCl3) δ 7.15-7.38
(m, 5H), 2.75 (s, 2H), 2.53 (m, 1H), 2.13 (s, 3H), 0.95 (s, 3H),
0.61 (s, 3H). Further elution with the same solvent mixture
yielded 3â-benzyl-3R-hydroxy-5â-pregnan-20-one (3, 30 mg):
TLC Rf 0.50 (toluene:acetone, 9:1); mp 133-141 °C; IR 2929,
2868, 2359, 1687 cm-1; 1H NMR (CDCl3) δ 7.18-7.42 (m, 5H),
2.89 (s, 2H), 2.52 (m, 1H), 2.11 (s, 3H), 1.03 (s, 3H), 0.60 (s,
3H). Anal. (C28H40O2) C, H.
Exp er im en ta l Section
Ch em istr y. 1H NMR spectra were recorded on a Varian
200 or 300 MHz spectrometer in CDC13 with tetramethylsilane

80 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 1
Upasani et al.
3r-Hyd r oxy-3â-(p h en yleth yn yl)-5â-p r egn a n -20-on e (4).
Meth od A. A solution of 5â-pregnane-3,20-dione cyclic 20-
(1,2-ethanediyl acetal) (2, 180 mg, 0.5 mmol) in dry THF (20
mL) was treated with lithium phenyl acetylide (1 M in THF,
1.5 mmol, 1.5 mL) at -70 °C. After being stirred the mixture
at this temperature for 1 h and then at room temperature for
2 h, it was quenched with 2 N HCl (1 mL). The solvent was
removed, and the residue was dissolved in acetone (20 mL).
After 1 N HCl (5 mL) was added, the solution was stirred at
room temperature for 0.5 h. Saturated NaHCO3 solution was
added to neutralize the acid. The solvents were removed, and
the residue was extracted with CH2Cl2. The organic layer was
washed with water, dilute NaHCO3 solution, water, and brine.
After being dried over anhydrous MgSO4, the solution was
filtered and evaporated to yield the crude product (300 mg).
This crude product was then dissolved in a small amount of
CH2Cl2 and poured on a column of silica gel. Elution with
toluene:acetone mixture (95:5) gave 3R-hydroxy-3â-(phenyl-
(m, 1H), 2.11 (s, 3H), 0.92 (s, 3H), 0.59 (s, 3H). Anal.
(C30H44O2) C, H.
Gen er a l Meth od for th e P r ep a r a tion of 2,2-Dibr o-
m op h en yleth en e Der iva tives (9).13 To a well-stirred solu-
tion of CBr4 (2.985 g, 9 mmol) in CH2Cl2 (20 mL) at 0 °C were
added PPh3 (4.75 g, 18.1 mmol) and then the benzaldehyde
derivative (8, 9 mmol). Stirring was continued at 0 °C for 20
min, at which point water (20 mL) was added. The organic
layer was separated, washed with water, dried (anhydrous
MgSO4), filtered, and evaporated to leave the crude product.
Purification of this by column chromatography on silica gel
(hexane:acetone, 9:1) afforded the 2,2-dibromophenylethene
derivative (9).
3r-Hydr oxy-3â-(4-m eth oxyph en yleth yn yl)-5â-pr egn an -
20-on e (10). Meth od A. A solution of 2,2-dibromo-1-(4′-
methoxyphenyl)ethene19 (prepared by the Wittig reaction of
4-methoxybenzaldehyde with carbon tetrabromide in the pres-
ence of triphenylphosphine) (875 mg, 3 mmol) in dry THF (15
mL) was treated under N2 with n-BuLi (2.5M in THF, 6 mmol,
2.4 mL) at -78 °C. The mixture was stirred at this temper-
ature for 0.5 h and then at room temperature for 1 h more. It
was then recooled to -78 °C, and a solution of 5â-pregnan-
3,20-dione cyclic 20-(1,2-ethanediyl acetal) (2, 720 mg, 2 mmol)
in dry THF (10 mL) was added dropwise over a period of 30
min. After the resulting mixture was stirred at -78 °C for 2
h, the cooling bath was removed and the stirring was continued
at room temperature for another hour. It was then quenched
with 2 N HCl solution (1 mL) at 10 °C. The solvent was
removed, and the residue was then dissolved in acetone (25
mL). After 2 N HCl (10 mL) was added, the solution was
stirred at room temperature for 2 h. Saturated NaHCO3
solution was added to neutralize the acid. The mixture was
concentrated until the product started crystallizing out. After
the mixture stood at room temperature for 1 h, the solid was
collected by filtration, washed with water, and dried. The
crude product was then recrystallized from a mixture of
hexane:acetone (4:1) to yield 10 as colorless rods (400 mg):
TLC Rf 0.32 (toluene:acetone, 95:5); mp 190-194 °C; IR 2932,
ethynyl)-5â-pregnan-20-one (4, 160 mg): TLC Rf 0.22 (toluene:
acetone, 95:5); mp 188-190 °C; IR 2930, 2868, 2361, 1685 cm-
;
1
1H NMR (CDCl3) δ 7.22-7.52 (m, 5H), 2.55 (m, 1H), 2.12 (s,
3H), 0.98 (s, 3H), 0.61 (s, 3H). Anal. (C29H38O2) C, H.
Meth od B. Alternatively, this compound was prepared in
25% yield in a manner analogous to the preparation of that
for 23 (see below) by the Pd-catalyzed coupling reaction of
4-iodobenzene and 22. The crude compound was purified by
column chromatography on silica gel. Elution with hexane:
acetone mixture (4:1) gave the product as a colorless solid,
which was found to be identical to 4 (TLC Rf, mp, NMR).
3r-Hyd r oxy-3â-(2′-p h en yleth yl)-5â-p r egn a n -20-on e (5).
A solution of 3R-hydroxy-3â-(phenylethynyl)-5â-pregnan-20-
one (44 mg) was dissolved in EtOAc (12 mL), Pd/C (5%, 12
mg) was added, and the mixture was hydrogenated at 45 psi
of hydrogen for 15 h at room temperature. Filtration of the
catalyst followed by evaporation of the solvent yielded the
crude product, which was purified by chromatography over
silica gel to isolate the pure 5 (33 mg): TLC Rf 0.40 (hexane:
acetone, 7:3); mp 153-184 °C; IR 2929, 2868, 2349, 1687 cm-
;
1
2865, 2361, 1699 cm-1 1H NMR (CDCl3) δ 7.38 (d, 2H, J )
,
1H NMR (CDCl3) δ 7.12-7.39 (m, 5H), 2.61-2.78 (m, 2H), 2.54
(m, 1H), 2.12 (s, 3H), 0.95 (s, 3H), 0.60 (s, 3H). Anal.
(C29H42O2) C, H.
8.8 Hz), 6.82 (d, 2H, J ) 8.8 Hz), 3.81 (s, 3H), 2.53 (m, 1H),
2.12 (s, 3H), 0.98 (s, 3H), 0.60 (s, 3H). Anal. (C30H40O3) C, H.
Meth od B. Alternatively, this compound was prepared in
45% yield in a manner analogous to the preparation of that
for 23 (see below) by the Pd-catalyzed coupling reaction of
4-iodoanisole and 22. The crude compound was purified by
column chromatography on silica gel. Elution with hexane:
acetone mixture (7:3) gave the product as a colorless solid,
which was found to be identical to 10 (TLC Rf, mp, NMR).
3â-(4-Ch lor op h en ylet h yn yl)-3r-h yd r oxy-5â-p r egn a n -
20-on e (11). Meth od A. This compound was prepared in
57% yield in a manner analogous to the preparation of 10
3r-Hyd r oxy-3â-(3′-p h en yl-1′-p r op yn yl)-5â-p r egn a n -20-
on e (6). A solution of 3-phenyl-1-propyne (232 mg, 2 mmol)
in dry THF (7 mL) was treated under N2 with n-BuLi (2.5 M
in hexane, 2 mmol, 0.8 mL) at -65 °C. The mixture was
stirred at this temperature for 10 min. It was then treated
dropwise with a solution of 5â-pregnane-3,20-dione cyclic 20-
(1,2-ethanediyl acetal) (2, 206 mg, 0.6 mmol) in dry THF (10
mL) over a period of 30 min. After the resulting mixture was
stirred at -78 °C for 0.5 h, it was quenched with saturated
NH4Cl solution (5 mL) at -10 °C. The solvent was removed,
and the residue was then dissolved in acetone (15 mL). After
1 N HCl (8 mL) was added, the solution was stirred at room
temperature for 0.5 h. A 2 N NaOH solution was added to
neutralize the acid. The solvents were removed, and the
residue was extracted with CH2Cl2. The organic layer was
washed with water, dilute NaHCO3 solution, water, and brine.
After being dried over anhydrous MgSO4, the solution was
filtered and evaporated to yield the crude product (460 mg).
This crude product was then dissolved in a small amount of
CH2Cl2 and poured on a column of silica gel. Elution with
toluene:acetone mixture (93:7) gave 3R-hydroxy-3â-(3′-phenyl-
1′-propynyl)-5â-pregnan-20-one (6, 175 mg): TLC Rf 0.46
(hexane:acetone, 7:3); mp 124-132 °C; IR 2929, 2868, 2359,
1697 cm-1; 1H NMR (CDCl3) δ 7.18-7.42 (m, 5H), 3.66 (s, 2H),
2.55 (m, 1H), 2.11 (s, 3H), 0.96 (s, 3H), 0.60 (s, 3H). Anal.
(C30H40O2) C, H.
starting from 2,2-dibromo-1-(4′-chlorophenyl)ethene13
(ob-
tained by the Wittig reaction of 4-chorobenzaldehyde with
carbon tetrabromide in the presence of triphenylphosphine)
and 2. The crude compound was purified by column chroma-
tography on silica gel (hexane:acetone, 4:1): TLC Rf 0.24
(toluene:acetone, 95:5); mp 188-190 °C; IR 2931, 2868, 2349,
1695, 1488 cm-1; 1H NMR (CDCl3) δ 7.38 (d, 2H, J ) 8.5 Hz),
7.27 (d, 2H, J ) 8.5 Hz), 2.52 (m, 1H), 2.12 (s, 3H), 0.99 (s,
3H), 0.61 (s, 3H). Anal. (C29
H37ClO2) C, H, Cl.
Meth od B. Alternatively, this compound was prepared in
53% yield in a manner analogous to the preparation of 23 (see
below) by the Pd-catalyzed coupling reaction of 4-chloroiodo-
benzene and 22. The crude compound was purified by column
chromatography on silica gel. Elution with hexane:acetone
mixture (3:1) gave the product as a colorless solid, which was
found to be identical to 11 (TLC Rf, mp, NMR).
3r-Hydr oxy-3â-(3′-ph en ylpr opyl)-5â-pr egn an -20-on e (7).
A solution of 6 (50 mg) was dissolved in EtOAc (12 mL), Pd/C
(5%, 10 mg) was added, and the mixture was hydrogenated at
30 psi of hydrogen for 45 min at room temperature. Filtration
of the catalyst followed by evaporation of the solvent yielded
the crude product, which was purified by chromatography over
silica gel to isolate the pure 7 (30 mg): TLC Rf 0.48 (hexane:
3â-[[4-(N ,N -Dim e t h yla m in o)p h e n yl]e t h yn yl]-3r-h y-
d r oxy-5â-p r egn a n -20-on e (12). Meth od A. This compound
was prepared in 35% yield in a manner analogous to the
preparation of 10 starting from 2,2-dibromo-1-[4′-(N,N-di-
methylamino)phenyl]ethene (obtained by the Wittig reaction
of 4-(N,N-dimethylamino)benzaldehyde with carbon tetrabro-
mide in the presence of triphenylphosphine) and 2. The crude
compound was purified by column chromatography on silica
gel (hexane:acetone, 9:1): TLC Rf 0.28 (hexane:acetone, 7:3);
acetone, 7:3); mp 41-46 °C; IR 2928, 2869, 1698, 1265 cm-1
1H NMR (CDCl3) δ 7.11-7.38 (m, 5H), 2.59-2.71 (m, 2H), 2.54
;

Neuroactive Steroids with High Affinity for GABAA Receptors
mp 212-216 °C; IR 2920, 2855, 2361, 1690 cm-1 1H NMR
(CDCl3) δ 7.32 (d, 2H, J ) 8.5 Hz), 6.63 (d, 2H, J ) 8.5 Hz),
2.97 (s, 6H), 2.52 (m, 1H), 2.12 (s, 3H), 0.97 (s, 3H), 0.60 (s,
3H). Anal. (C31H43NO2) C, H, N.
Meth od B. Alternatively, this compound was prepared in
45% yield in a manner analogous to the preparation of 23 (see
below) by the Pd-catalyzed coupling reaction of 4-bromo-N,N-
dimethylaniline and 22. The crude compound was purified
by column chromatography on silica gel. Elution with hexane:
acetone mixture (3:1) gave the product as a colorless solid,
which was found to be identical to 12 (TLC Rf, mp, NMR).
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 1 81
;
Meth od B. Alternatively, this compound was prepared in
15% yield in a manner analogous to the preparation of 23 (see
below) by the Pd-catalyzed coupling reaction of 4-bromobiphen-
yl and 22. The crude compound was purified by column
chromatography on silica gel. Elution with hexane:acetone
mixture (4:1) gave the product as a colorless solid, which was
found to be identical to 15 (TLC Rf, mp, NMR).
3â-(4-Hyd r oxyp h en yleth yn yl)-3r-h yd r oxy-5â-p r egn a n -
20-on e (16). Meth od A. This compound was prepared in
46% yield in a manner analogous to the preparation of 10
starting from 2,2-dibromo-1-(4-hydroxyphenyl)ethene21 (ob-
tained by the Wittig reaction of 4-hydroxybenzaldehyde with
carbon tetrabromide in the presence of triphenylphosphine)
and 2. The crude compound was purified by column chroma-
tography on silica gel (hexane:acetone, 9:1): TLC Rf 0.17
(toluene:acetone, 9:1); mp 252-256 °C; IR 3404, 2927, 2873,
3â-(4-Cya n op h en ylet h yn yl)-3r-h yd r oxy-5â-p r egn a n -
20-on e (13). Meth od A. A solution of 2,2-dibromo-1-(4-
cyanophenyl)ethene20 (prepared by the Wittig reaction of
4-cyanobenzaldehyde with carbon tetrabromide in the presence
of 1 equiv of triphenylphosphine and 1 equiv of zinc dust) (824
mg, 3.33 mmol) in dry THF (15 mL) was treated under N2 with
n-BuLi (2.5M in THF, 6 mmol, 2.4 mL) at -78 °C. The
mixture was stirred at -75 °C for 0.5 h, and then a solution
of 5â-pregnane-3,20-dione cyclic 20-(1,2-ethanediyl acetal) (725
mg, 2.014 mmol) in dry THF (10 mL) was added dropwise over
a period of 10 min. After the resulting mixture was stirred at
-70 °C for 2 h, it was quenched with NH4Cl solution (3 mL).
The solvent was removed, and the residue was then dissolved
in acetone (25 mL). After 2 N HCl (10 mL) was added, the
solution was stirred at room temperature for 1 h. Saturated
NaHCO3 solution was added to neutralize the acid. The
solvents were removed, and the residue was extracted with
CH2Cl2. The organic layer was washed with water and brine.
After being dried over anhydrous MgSO4, the solution was
filtered and evaporated to yield the crude product (600 mg).
This crude product was then dissolved in a small amount of
CH2Cl2 and poured on a column of silica gel. Elution with
toluene:acetone mixture (95:5) gave 13 as a brown solid (150
mg): TLC Rf 0.29 (toluene:acetone, 95:5); mp 200-203 °C; IR
1
1683 cm-1; H NMR (CDCl3) δ 7.37 (d, 2H, J ) 8.5 Hz), 6.78
(d, 2H, J ) 8.5 Hz), 5.00 (bs, 1H), 2.52 (m, 1H), 2.12 (s, 3H),
0.98 (s, 3H), 0.60 (s, 3H). Anal. (C29H38O3) C, H.
Meth od B. Alternatively, this compound was prepared in
35% yield in a manner analogous to the preparation of 23 (see
below) by the Pd-catalyzed coupling reaction of 4-iodophenol
and 22. The crude compound was purified by column chro-
matography on silica gel. Elution with hexane:acetone mix-
ture (7:3) gave the product as a colorless solid, which was found
to be identical to 16 (TLC Rf, mp, NMR).
3r-Hyd r oxy-3â-(4-n itr op h en yleth yn yl)-5â-p r egn a n -20-
on e (17). Meth od A. A solution of 2,2-dibromo-1-(4-nitro-
phenyl)ethene22 (prepared by the Wittig reaction of 4-nitroben-
zaldehyde with carbon tetrabromide in the presence of
triphenylphosphine) (296 mg, 1 mmol) in dry THF (20 mL)
was treated under N2 with n-BuLi (2.5M in THF, 2 mmol, 0.8
mL) at -95 °C. The mixture was stirred at -80 to -100 °C
for 0.5 h, and then a solution of 5â-pregnane-3,20-dione cyclic
20-(1,2-ethanediyl acetal) (720 mg, 2 mmol) in dry THF (10
mL) was added dropwise over a period of 10 min. After the
resulting mixture was stirred <-80 °C for 1 h and then at 0
°C for an additional 1 h, it was quenched with NH4Cl solution
(3 mL). The solvent was removed, and the residue was then
dissolved in acetone (25 mL). After 2 N HCl (10 mL) was
added the solution was stirred at room temperature for 1 h.
Saturated NaHCO3 solution was added to neutralize the acid.
The solvents were removed, and the residue was extracted
with CH2Cl2. The organic layer was washed with water and
brine. After being dried over anhydrous MgSO4 the solution
was filtered and evaporated to yield the crude product (400
mg). This crude product was then dissolved in a small amount
of CH2Cl2 and poured on a column of silica gel. Elution with
toluene:acetone mixture (96:4) gave 17 as a brown solid (70
mg): TLC Rf 0.18 (toluene:acetone, 95:5); mp 241-244 °C; IR
2922, 2862, 2346, 1697, 1516, 1336, 855 cm-1; 1H NMR (CDCl3)
δ 8.21 (d, 2H, J ) 8 Hz), 7.59 (d, 2H, J ) 8 Hz), 2.52 (m, 1H),
2.12 (s, 3H), 1.00 (s, 3H), 0.61 (s, 3H). Anal. (C29H37NO4) C,
H, N.
1
2926, 2869, 2222, 1678 cm-1; H NMR (CDCl3) δ 7.60 (d, 2H,
J ) 8.4 Hz), 7.51 (d, 2H, J ) 8.4 Hz), 2.53 (m, 1H), 2.12 (s,
3H), 0.99 (s, 3H), 0.60 (s, 3H). Anal. (C30H37NO2) C, H, N.
Meth od B. Alternatively, this compound was prepared in
25% yield in a manner analogous to the preparation of 23 (see
below) by the Pd-catalyzed coupling reaction of 4-bromoben-
zonitrile and 22. The crude compound was purified by column
chromatography on silica gel. Elution with hexane:acetone
mixture (3:1) gave the product as a colorless solid, which was
found to be identical to 13 (TLC Rf, mp, NMR).
3â-[(4-For m am idoph en yl)eth yn yl]-3r-h ydr oxy-5â-pr eg-
n a n -20-on e (14). Meth od A. A solution of 13 (79 mg, 0.178
mmol) in DMSO (1.5 mL) was treated with a H2O2 (30%, 7
drops) and K2CO3 (12 mg) at 0-5 °C. The mixture was stirred
at room temperature for 1 h. Water (3 mL) was added, and
the precipitated solid was collected by filtration. The crude
product was crystallized from MeOH as colorless prisms (48
mg): TLC Rf 0.15 (hexane:acetone, 7:3); mp 255-262 °C dec;
IR 3346, 3053, 2926, 2854, 2305, 1692, 1663, 1259, 730 cm-1
;
Meth od B. Alternatively, this compound was prepared in
45% yield in a manner analogous to the preparation of 23 (see
below) by the Pd-catalyzed coupling reaction of 4-iodonitroben-
zene and 22. The crude compound was purified by column
chromatography on silica gel. Elution with hexane:acetone
mixture (7:3) gave the product as a colorless solid, which was
found to be identical to 17 (TLC Rf, mp, NMR).
1H NMR (CDCl3) 7.76 (d, 2H, J ) 8.3 Hz), 7.50 (d, 2H, J ) 8.3
Hz), 6.05 (bs, 1H), 6.61 (bs, 1H), 2.53 (m, 1H), 2.12 (s, 3H),
0.99 (s, 3H), 0.61 (s, 3H). Anal. (C30H39NO3) C, H, N.
Meth od B. Alternatively, this compound was prepared in
15% yield in a manner analogous to the preparation of 23 (see
below) by the Pd-catalyzed coupling reaction of 4-bromoben-
zamide and 22. The crude compound was purified by column
chromatography on silica gel. Elution with hexane:acetone
mixture (7:3) gave the product as a colorless solid, which was
found to be identical to 14 (TLC Rf, mp, NMR).
3r-Hydr oxy-3â-[(2-m eth oxyph en yl)eth yn yl]-5â-pr egn an -
20-on e (18). Meth od A. This compound was prepared in
54% yield in a manner analogous to the preparation of 10
starting from 2,2-dibromo-1-(2-methoxyphenyl)ethene (ob-
tained by the Wittig reaction of 2-methoxybenzaldehyde with
carbon tetrabromide in the presence of triphenylphosphine)
and 2. The crude compound was purified by column chroma-
tography on silica gel (hexane:acetone, 4:1): TLC Rf 0.27
(hexane:acetone, 4:1); mp 183-185 °C; IR 2928, 2868, 2361,
3â-(4-Bip h e n yle t h yn yl)-3r-h yd r oxy-5â-p r e gn a n -20-
on e (15). Meth od A. This compound was prepared in 53%
yield in a manner analogous to the preparation of 10 starting
from 2,2-dibromo-1-(4-biphenylyl)ethene (obtained by the Wit-
tig reaction of 4-phenylbenzaldehyde with carbon tetrabromide
in the presence of triphenylphosphine) and 2. The crude
compound was purified by column chromatography on silica
gel (hexane:acetone, 4:1): TLC Rf 0.22 (hexane:acetone, 4:1);
1
1697 cm-1; H NMR (CDCl3) δ 7.38 (d, 1H, J ) 7.5 Hz), 7.28
(m, 1H), 6.89 (m, 2H), 3.87 (s, 3H), 2.52 (m, 1H), 2.12 (s, 3H),
0.99 (s, 3H), 0.61 (s, 3H). Anal. (C30H40O3) C, H.
1
mp 182-184 °C; H NMR (CDCl3) δ 7.32-7.66 (m, 9H), 2.52
Meth od B. Alternatively, this compound was prepared in
45% yield in a manner analogous to the preparation of 23 (see
below) by the Pd-catalyzed coupling reaction of 2-iodoanisole
(m, 1H), 2.12 (s, 3H), 1.00 (s, 3H), 0.61 (s, 3H). Anal.
(C35H42O2) C, H.

82 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 1
Upasani et al.
and 22. The crude compound was purified by column chro-
matography on silica gel. Elution with hexane:acetone mix-
ture (4:1) gave the product as a colorless solid, which was found
to be identical to 18 (TLC Rf, mp, NMR).
(toluene:acetone, 95:5); mp 208-210 °C; 1H NMR (CDCl3) δ
7.35 (d, 2H, J ) 7.4 Hz), 7.11 (d, 2H, J ) 7.4 Hz), 2.52 (m,
1H), 2.35 (s, 3H), 2.12 (s, 3H), 0.98 (s, 3H), 0.60 (s, 3H). Anal.
(C30H40O2) C, H.
3â-[(3,4-Dim eth oxyph en yl)eth yn yl]-3r-h ydr oxy-5â-pr eg-
n a n -20-on e (19). This compound was prepared in 45% yield
in a manner analogous to the preparation of 10 starting from
2,2-dibromo-1-(3,4-dimethoxyphenyl)ethene23 (obtained by the
Wittig reaction of 3,4-dimethoxybenzaldehyde with carbon
tetrabromide in the presence of triphenylphosphine) and 2.
The crude compound was purified by column chromatography
on silica gel. Elution with toluene:acetone mixture (96:4) gave
a phenylacetylene compound, which was not characterized.
Further elution with the same solvent yielded 3R-[(3′,4′-
dimethoxyphenyl)ethynyl]-3â-hydroxy-5â-pregnan-20-one (11%)
as a first fraction and 3â-[(3′,4′-dimethoxyphenyl)ethynyl]-3R-
hydroxy-5â-pregnan-20-one as a second fraction (45%): TLC
Rf 0.18 (hexane:acetone, 4:1); mp 82-88 °C; IR 2929, 2867,
2361, 1697 cm-1; 1H NMR (CDCl3) δ 7.05 (dd, 1H, J ) 1.8 and
8.3 Hz), 6.91 (d, 1H, J ) 1.8 Hz), 6.78 (d, 1H, J ) 8.3 Hz),
3.89 (s, 6H), 2.52 (m, 1H), 2.12 (s, 3H), 0.99 (s, 3H), 0.61 (s,
3H). Anal. (C31H42O4) C, H.
3â-[(4-Am in op h en yl)eth yn yl]-3r-h yd r oxy-5â-p r egn a n -
20-on e (26). This compound was prepared in 60% yield in a
manner analogous to the preparation of 23 starting from
4-iodoaniline and 22. The crude compound was purified by
column chromatography on silica gel. Elution with hexane:
acetone mixture (4:1) gave 26 as a pale yellow solid: TLC Rf
0.22 (hexane:acetone, 95:5); mp 203-207 °C dec; IR 3441, 3352,
1
2923, 2859, 2345, 1686 cm-1; H NMR (CDCl3) δ 7.25 (d, 2H,
J ) 8.4 Hz), 6.58 (d, 2H, J ) 8.4 Hz), 3.80 (bs, 2H), 2.52 (m,
1H), 2.11 (s, 3H), 0.97 (s, 3H), 0.60 (s, 3H). Anal. (C29H39
NO2) C, H, N.
-
3r-Hyd r oxy-3â-[[4-(t r iflu or om et h yl)p h en yl]et h yn yl]-
5â-p r egn a n -20-on e (27). This compound was prepared in
31% yield in a manner analogous to the preparation of 23
starting from 4-(trifluoromethyl)iodobenzene and 22. The
crude compound was purified by column chromatography on
silica gel. Elution with toluene:acetone mixture (95:5) gave
27 as a colorless solid: TLC Rf 0.23 (toluene:acetone, 95:5);
3r-Hydr oxy-3â-[[3,4-(m eth ylen edioxy)ph en yl]eth yn yl]-
5â-p r egn a n -20-on e (20). Meth od A. This compound was
prepared in 65% yield in a manner analogous to the prepara-
tion of 10 starting from 2,2-dibromo-1-[3′,4′-(methylenedioxy)-
phenyl]ethene (obtained by the Wittig reaction of 3,4-(meth-
ylenedioxy)benzaldehyde with carbon tetrabromide in the
presence of triphenylphosphine) and 2. The crude compound
was purified by column chromatography on silica gel (toluene:
acetone, 95:5): TLC Rf 0.43 (hexane:acetone, 7:3); mp 188-
mp 201-203 °C; IR 2922, 2863, 2358, 1688, 1312 cm-1 1H
;
NMR (CDCl3) δ 7.58 (d, 2H, J ) 7 Hz), 7.52 (d, 2H, J ) 7 Hz),
2.52 (m, 1H), 2.12 (s, 3H), 0.99 (s, 3H), 0.61 (s, 3H). Anal.
(C30H37F3O2) C, H, F.
3â-[(4-Acetylp h en yl)eth yn yl]-3r-h yd r oxy-5â-p r egn a n -
20-on e (28). This compound was prepared in 51% yield in a
manner analogous to the preparation of 23 starting from
4-iodoacetophenone and 22. The crude compound was purified
by column chromatography on silica gel. Elution with hexane:
acetone mixture (7:3) gave 28 as a colorless solid: TLC Rf 0.40
(toluene:acetone, 9:1); mp 192-194 °C; IR 2935, 2858, 2360,
1700, 1683 cm-1; 1H NMR (CDCl3) δ 7.91 (d, 2H, J ) 8.1 Hz),
7.51 (d, 2H, J ) 8.1 Hz), 2.60 (s, 3H), 2.52 (m, 1H), 2.12 (s,
3H), 0.99 (s, 3H), 0.61 (s, 3H). Anal. (C31H40O3) C, H.
3â-[(4-Car beth oxyph en yl)eth yn yl]-3r-h ydr oxy-5â-pr eg-
n a n -20-on e (29). This compound was prepared in 68% yield
in a manner analogous to the preparation of 23 starting from
4-iodobenzoic acid ethyl ester and 22. The crude compound
was purified by column chromatography on silica gel. Elution
with hexane:acetone mixture (4:1) gave 29 as a colorless
solid: TLC Rf 0.31 (hexane:acetone, 4:1); mp 164-166 °C; IR
1
191 °C; IR 2929, 2868, 2361, 1697 cm-1; H NMR (CDCl3) δ
6.98 (d, 1H, J ) 7.95 Hz), 6.88 (s, 1H), 6.73 (d, 1H, J ) 7.95
Hz), 5.97 (s, 2H), 2.52 (m, 1H), 2.12 (s, 3H), 0.98 (s, 3H), 0.60
(s, 3H). Anal. (C30H38O4) C, H.
Meth od B. Alternatively, this compound was prepared in
30% yield in a manner analogous to the preparation of 23 (see
below) by the Pd-catalyzed coupling reaction of 4-bromo-1,2-
(methylenedioxy)benzene and 22. The crude compound was
purified by column chromatography on silica gel. Elution with
hexane:acetone mixture (7:3) gave the product as a colorless
solid, which was found to be identical to 20 (TLC Rf, mp,
NMR).
1
2922, 2871, 2347, 1717, 1700 cm-1; H NMR (CDCl3) δ 8.01
3â-(2-Hyd r oxyp h en yleth yn yl)-3r-h yd r oxy-5â-p r egn a n -
20-on e (23).A solution of 2-iodophenol (96 mg, 0.44 mmol) and
3â-ethynyl-3R-hydroxy-5â-pregnan-20-one (227, 150 mg, 0.44
mmol) in dry degassed triethylamine (0.5 mL) was stirred
under argon at room temperature. Bis(triphenylphosphine)-
palladium chloride (5 mg) and CuI (5 mg) were added, and
the mixture was stirred at this temperature for 45 min. CH2-
Cl2 (5 mL) was added, and the stirring was continued for
another hour. The TLC showed 100% conversion of the
starting material; hence, the solvent was removed and the
residue was purified by chromatography on silica gel. Elution
with hexane:EtOAc (4:1) gave 23 (40 mg) as a colorless solid;
TLC Rf 0.34 (toluene:acetone, 95:5); mp 210-212 °C; IR 3427,
(d, 2H, J ) 8.1 Hz), 7.49 (d, 2H, J ) 8.1 Hz), 4.39 (m, 2H),
2.52 (m, 1H), 2.12 (s, 3H), 1.42 (m, 3H), 0.99 (s, 3H), 0.61 (s,
3H). Anal. (C32H42O4) C, H.
3â-[(4-F or m ylp h en yl)eth yn yl]-3r-h yd r oxy-5â-p r egn a n -
20-on e (30). This compound was prepared in 25% yield in a
manner analogous to the preparation of 23 starting from
4-bromobenzaldehyde and 22. Triethylamine was replaced
with diethylamine. The crude compound was purified by
column chromatography on silica gel. Elution with hexane:
acetone mixture (7:3) gave 30 as a colorless solid: TLC Rf 0.18
(hexane:acetone, 4:1); mp 212-215 °C; IR 2922, 2858, 2347,
1694, 1684 cm-1 1H NMR (CDCl3) δ 10.01 (s, 1H), 7.91 (d,
;
1
2929, 2868, 2359, 1697 cm-1; H NMR (CDCl3) δ 7.58 (d, 1H,
2H, J ) 8.1 Hz), 7.58 (d, 2H, J ) 8.1 Hz), 2.52 (m, 1H), 2.12
(s, 3H), 1.00 (s, 3H), 0.62 (s, 3H). Anal. (C30H38O3) C, H.
3â-[(3-Acetylp h en yl)eth yn yl]-3r-h yd r oxy-5â-p r egn a n -
20-on e (31). This compound was prepared in 58% yield in a
manner analogous to the preparation of 23 starting from
3-iodoacetophenone and 22. The crude compound was purified
by column chromatography on silica gel. Elution with hexane:
acetone mixture (4:1) gave 31 as a colorless solid: TLC Rf 0.22
(hexane:acetone, 4:1); mp 195-197 °C; IR 2927, 2859, 2361,
1699, 1684 cm-1; 1H NMR (CDCl3) δ 8.01 (s, 1H), 7.91 (d, 1H,
J ) 7.7 Hz), 7.62 (d, 1H, J ) 7.9 Hz), 7.42 (dd, 1H, J ) 7.7
and 7.9 Hz), 2.62 (s, 3H), 2.54 (m, 1H), 2.12 (s, 3H), 0.99 (s,
3H), 0.61 (s, 3H). Anal. (C31H40O3) C, H.
J ) 7.5 Hz), 7.59 (d, 1H, J ) 8 Hz), 7.26 (m, 2H), 6.68 (s, 1H),
2.57 (m, 1H), 2.13 (s, 3H), 0.87 (s, 3H), 0.61 (s, 3H). Anal.
(C29H38O3) C, H.
3â-[(3-Hydr oxyph en yl)eth yn yl]-3r-h ydr oxy-5â-pr egn an -
20-on e (24). This compound was prepared in 25% yield in a
manner analogous to the preparation of 23 starting from
3-iodophenol and 22. The crude compound was purified by
column chromatography on silica gel. Elution with hexane:
acetone mixture (4:1) gave 24 as a colorless solid: TLC Rf 0.31
(toluene:acetone, 85:15); mp 208-213 °C; IR 3500, 2928, 2358,
1677 cm-1; 1H NMR (MeOH-d4) δ 7.15 (m, 1H), 6.70-6.95 (m,
3H), 2.68 (m, 1H), 2.16 (s, 3H), 1.05 (s, 3H), 0.65 (s, 3H). Anal.
(C29H38O3) C, H.
3r-Hyd r oxy-3â-[(4-m eth ylp h en yl)eth yn yl]-5â-p r egn a n -
20-on e (25). This compound was prepared in 46% yield in a
manner analogous to the preparation of 23 starting from
4-iodotoluene and 22. The crude compound was purified by
column chromatography on silica gel. Elution with hexane:
acetone mixture (7:3) gave 25 as a colorless solid: TLC Rf 0.25
3â-[[4-(N,N-Dieth ylca r ba m oyl)p h en yl]eth yn yl]-3r-h y-
d r oxy-5â-p r egn a n -20-on e (32). This compound was pre-
pared in 12% yield in a manner analogous to the preparation
of 23 starting from 4-iodo-N,N-diethylbenzamide and 22. The
crude compound was purified by column chromatography on
silica gel. Elution with hexane:acetone mixture (3:1) gave 32
as a colorless solid: TLC Rf 0.22 (hexane:acetone, 3:1); mp

Neuroactive Steroids with High Affinity for GABAA Receptors
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 1 83
1
174-177 °C; IR 2930, 2867, 2361, 1699, 1693 cm-1; H NMR
receptor subunits were prepared as previously described.24
Briefly, oocytes were microinjected with a 1:1:1 mixture of
cRNAs encoding the human R1, â2, and γ2L subunits (ap-
proximately 1 ng of each cRNA per cell). Membrane currents
were recorded with a two-electrode voltage clamp in frog
Ringer’s solution consisting of 115 mM NaCl, 2 mM KCl, 1.8
mM CaCl, and 5 mM HEPES, pH 7.4. Drug and wash
solutions were applied directly to the oocyte using a micro-
capillary “linear array” system. Maximum GABA responses
were elicited by application of 1 mM GABA. Stock solutions
of steroids were dissolved in DMSO over the range of 0.01 µM
to 10 mM. Stocks were diluted 300-1000-fold in frog Ringer’s
solution and DMSO adjusted to 0.3% by volume; 0.3% DMSO
had little effect on GABA responses. Both GABA concentra-
tion-response data and steroid concentration-modulation
data were fit to a four-parameter logistic equation (Origin,
Microcal, Inc.). Unless indicated otherwise, data is expressed
as mean values ( SEM to two significant figures.
In Vivo P h a r m a cology. Male NSA mice weighing be-
tween 15 and 20 g were obtained from Harlan Sprague-
Dawley (San Diego, CA). Upon arrival they were housed in
standard polycarbonate cages (four per cage) containing a
sterilized bedding material in a room of constant temperature
(23.0° ( 2.5 °C) with a 12 h (0700-1900 light) light/dark cycle.
Food (Teklad LM 485) and water were freely available. Mice
were acclimated a minimum of 4 days prior to experimenta-
tion.
(CDCl3) δ 7.48 (d, 2H, J ) 7.6 Hz), 7.31 (d, 2H, J ) 7.6 Hz),
3.52 (bm, 2H), 3.25 (bm, 2H), 2.52 (m, 1H), 2.12 (s, 3H), 0.99
(s, 3H), 0.61 (s, 3H). Anal. (C34H47NO3) C, H, N.
3â-[(4-Ben zoylph en yl)eth yn yl]-3r-h ydr oxy-5â-pr egn an -
20-on e (33). This compound was prepared in 10% yield in a
manner analogous to the preparation of 23 starting from
4-iodobenzophenone and 22. The crude compound was puri-
fied by column chromatography on silica gel. Elution with
hexane:acetone mixture (4:1) gave 33 as a colorless solid: TLC
Rf 0.54 (hexane:acetone, 7:3); mp 78-83 °C; IR 2920, 2837,
2361, 1698, 1695 cm-1; 1H NMR (CDCl3) δ 7.51-7.92 (m, 9H),
2.52 (m, 1H), 2.12 (s, 3H), 1.00 (s, 3H), 0.61 (s, 3H). Anal.
(C36H42O3) C, H.
3â-[(4-Acet ylp h en yl)et h yn yl]-3r-h yd r oxy-5â-19-n or -
p r egn a n -20-on e (36). This compound was prepared in 30%
yield in a manner analogous to the preparation of 23 starting
from 4-iodoacetophenone and 3â-ethynyl-3R-hydroxy-5â-19-
norpregnan-20-one [35, obtained by the addition of lithium
acetylide to 5â-19-norpregnane-3,20-dione cyclic 20-(1,2-
ethanediyl acetal)].7 The crude compound was purified by
column chromatography on silica gel. Elution with hexane:
acetone mixture (4:1) gave 36 as a colorless solid: TLC Rf 0.12
(hexane:acetone, 4:1); mp 64-65 °C; IR 2920, 2868, 2361, 1699,
1684 cm-1; H NMR (CDCl3) δ 7.91 (d, 2H, J ) 8.2 Hz), 7.53
(d, 2H, J ) 8.2 Hz), 2.60 (s, 3H), 2.55 (m, 1H), 2.12 (s, 3H),
1
0.63 (s, 3H). Anal. (C30H38O3) C, H.
3â-[4-Ca r b et h oxyp h en yl)et h yn yl]-3r-h yd r oxy-5â-19-
n or p r egn a n -20-on e (37). This compound was prepared in
20% yield in a manner analogous to the preparation of 23
starting from 4-iodobenzoic acid ethyl ester and 35. The crude
compound was purified by column chromatography on silica
gel. Elution with hexane:acetone mixture (4:1) gave 37 as a
colorless solid: TLC Rf 0.27 (hexane:acetone 4:1); mp 164-
165 °; IR 2922, 2874, 2358, 1713, 1700 cm-1; 1H NMR (CDCl3)
δ 8.01 (d, 2H, J ) 8.1 Hz), 7.49 (d, 2H, J ) 8.1 Hz), 4.39 (m,
2H), 2.55 (m, 1H), 2.13 (s, 3H), 1.40 (m, 3H), 0.63 (s, 3H). Anal.
(C31H40O4) C, H.
P en tylen etetr a zol-In d u ced Seizu r es. Seizures were
induced by administration of 85 mg/kg, sc pentylenetetrazole
(30 min observation period). The dose used was previously
determined to be the CD97. A clonic seizure was defined as
forelimb clonus of g3 s duration. Data were treated quantally.
Maxim al Electr osh ock-In du ced Seizu r es. Seizures were
induced by application of current (50 mA, 60 pulses/s, 0.8 ms
pulse width, 1 s duration, dc) using a Ugo Basile ECT device
(Model 7801). Mice were restrained by gripping the loose skin
on their dorsal surface, and saline-coated corneal electrodes
were held lightly against the two cornea. Current was applied,
and mice were observed for a period of up to 30 s for the
occurrence of a tonic hindlimb extensor response. A tonic
seizure was defined as a hindlimb extension in excess of 90°
from the plane of the body. Results were treated in a quantal
manner.
P h a r m a cology. Ster oid In h ibition of TBP S Bin d in g.
TBPS binding assays using rat brain cortical membranes in
the presence of 5 µM GABA has been described.4,6,7 Briefly,
cortices were rapidly removed following decapitation of carbon
dioxide-anesthetized Sprague-Dawley rats (200-250 g). The
cortices were homogenized in 10 volumes of ice-cold 0.32 M
sucrose using a glass/Teflon homogenizer and centrifuged at
1500g for 10 min at 4 °C. The resultant supernatants were
centrifuged at 10000g for 20 min at 4 °C to obtain the P2
pellets. The P2 pellets were resuspended in 200 mM NaCl/50
mM Na-K phosphate pH 7.4 buffer and centrifuged at 10000g
for 10 min at 4 °C. This washing procedure was repeated
twice, and the pellets were resuspended in 10 volumes of
buffer. Aliquots (100 µL) of the membrane suspensions were
incubated with 2 nM [35S]TBPS and 5 µL aliquots of test drug
dissolved in dimethyl sulfoxide (DMSO) (final 0.5%) in the
presence of 5 µM GABA. The incubation was brought to a final
volume of 1.0 mL with buffer. Nonspecific binding was
determined in the presence of 2 µM unlabeled TBPS and
ranged from 15 to 25%. Following a 90 min incubation at room
temperature, the assays were terminated by filtration through
glass fiber filters (Schleicher and Schuell No. 32) using a cell
harvester (Brandel) and rinsed three times with ice-cold buffer.
Filter bound radioactivity was measured by liquid scintillation
spectrometry. Nonlinear curve fitting of the overall data for
each drug averaged for each concentration was done using
Prism (GraphPad). The data were fit to a partial instead of a
full inhibition model if the sum of squares was significantly
lower by F-test. Similarly, the data were fit to a two
component instead of a one-component inhibition model if the
sum of squares was significantly lower by F-test. The con-
Ha n gin g Wir e. The hanging-wire test used a custom-built
apparatus that consisted of a metal wire (2 mm diameter)
suspended horizontally above a padded surface (25 cm). Mice
were held by the base of the tail, their forepaws placed in
contact with the wire, and then released. Animals were
required to bring both hindpaws in contact with the wire
within 5 s in order to be scored as a pass. Results were treated
quantally.
Ack n ow led gm en t. The authors acknowledge Mike
Suruki and Silvia Robledo for technical assistance. We
thank Dr. Derk J . Hogenkamp and Dr. George Field for
invaluable discussions.
Refer en ces
(1) Sieghart, W. Structure and pharmacology of the γ-aminobutyric
acidA receptor subtypes. Pharmacol. Rev. 1995, 47, 181-233.
(2) Harrison, N. L.; Simmonds, M. A.. Modulation of the GABA
receptor by a steroid anaesthetic. Brain Res. 1984, 323, 287-
292.
(3) Gee, K. W.; McCauley, L. D.; Lan, N. C. A putative receptor for
neurosteroids on the GABAA receptor complex: The pharmaco-
logical properties and therapeutic potential of epalons. Crit. Rev.
Neurobiol. 1995, 9, 207-227.
(4) 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 mech-
anism of action, J . Pharmacol. Exp. Ther. 1988, 246, 803-812.
(5) Harrison, N. L.; Majewska, M. D.; Harrington, J . W.; Barker, J .
L. Structure-activity relationships for steroid interaction with
the γ-aminobutyric acidA receptor complex. J . Pharmacol. Exp.
Ther. 1987, 241, 346-353.
centration of test compound producing 50% inhibition (IC50
of specific binding and the maximal extent of inhibition (Imax
)
)
were determined for the individual experiments with the same
model used for the overall data and then the means ( SEM of
the individual experiments were calculated.
Electr op h ysiology of GABAA Recep tor s Exp r essed in
Xen opu s Oocytes. Oocytes and cRNAs encoding GABAA

84 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 1
Upasani et al.
(6) Hawkinson, J . E.; Kimbrough, C. L.; Belelli, D.; Lambert, J . J .;
Purdy, R. H.; Lan, N. C. Correlation of neuroactive steroid
modulation of [35S]t-butylbicyclophosporothionate and [3H] fluni-
trazepam binding and γ-aminobutyric acidA receptor function.
Mol. Pharmacol. 1994, 46, 977-985.
(7) 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 that allosterically modulate the GABAA
receptor. J . Med. Chem. 1997, 40, 61-72.
(8) Im, W. B.; Blakeman, D. P.; Davis, J . P.; Ayer, D. E. Studies on
the mechanism of interactions between anesthetic steroids and
γ-aminobutyric acidA receptors. Mol. Pharmacol. 1990, 37, 429-
434.
(9) 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.
(10) Atkinson, R. M.; Davis, B.; Pratt, M. A.; Sharpe, H. M.; Tomich,
E. G. Action of some steroids on the central nervous system of
the mouse. II. Pharmacology. J . Med. Chem. 1965, 8, 426-432.
(11) Gyermek, L.; Iriarte, J .; Crabbe. P. Steroids. CCCX. Structure-
activity relationship of some steroidal hypnotic agents. J . Med.
Chem. 1968, 11, 117-125.
(12) Phillips, G. H. Structure-activity relationships in steroidal
anesthetics. J . Steroid Biochem. 1975, 6, 607-613.
(13) Kostikov, R. R.; Krasikov, A. Ya.; Mandel’shtam, T. V.; Khar-
icheva, E. M. Synthesis and dehydrobromination of 2,2,2-
tribromo-1-arylethanes. Zh. Organich. Khim. 1985, 21 (2), 329-
333 (Russian).
5R-pregnan-20-one and 3R-OH-5â-pregnan-20-one in rodents.
Psychopharmacology 1995, 118, 65-71.
(16) Carter, R. B.; Wood, P. L.; Wieland, S.; Hawkinson, J . E.; Belelli,
D.; Lambert, J . J .; White, H. S.; Wolf, H. H.; Mirsadeghi, S.;
Tahir, S. H.; Bolger, M. B.; Lan, N. C.; Gee, K. W. Characteriza-
tion of the anticonvulsant properties of ganaxolone (CCD 1042;
3R-hydroxy-3â-methyl-5R-pregnan-20-one), a selective high-af-
finity steroid modulator of the GABAA receptor. J . Pharmacol.
Exp. Ther., in press.
(17) Upasani, R. B.; Hawkinson, J . E. Unpublished observations.
(18) Still, W. C.; Kahn, M.; Mitra, A. Rapid chromatographic tech-
nique for preparative separations with moderate resolution. J .
Org. Chem. 1978, 43, 2923-2925.
(19) Morris, J .; Wishka, D. G. Synthesis of 2-amino-6-phenyl-4H-
pyran-4-ones. Synthesis 1994, 43-46.
(20) Milburn, G. H. W.; Campbell, C.; Shand, A. J .; Werninck, A. R.
Thermotropic diacetylenic liquid crystals. Liq. Cryst. 1990, 8 (5),
623-637.
(21) Iguchi, S.; Iwamura, H.; Nishizaki, M.; Hayashi, A.; Senokuchi,
K.; Kobayashi, K.; Sakaki, K.; Hachiya, K.; Ichioka, Y.; Kawa-
mura, M. Development of a highly cardioselective ultra short-
acting â-blocker, ONO-1101. Chem. Pharm. Bull. 1992, 40 (6),
1462-1469.
(22) Bestmann, H. J .; Frey, H. Reactions with phosphinealkylenes.
XXXIX. New methods for the preparation of 1-bromoacetylenes
and aromatic and conjugated enynes. Liebigs Ann. Chem. 1980,
(12), 2061-2071.
(23) Pelter, A.; Ward, R. S.; Little, G. Approaches to 2,6-diaryl-3,7-
dioxabicyclo[3.3.0]octane lignans via asymmetric synthesis of
dihydro- and tetrahydro-furan derivatives. J . Chem. Soc., Perkin
Trans. 1 1990, 2775-2790.
(24) Hawkinson, J . E.; Drewe, J . A.; Kimbrough, C. L.; Chen, J . S.;
Hogenkamp, D. J .; Lan, N. C.; Gee, K. W.; Shen, K. Z.;
Whittemore, E. R.; Woodward, R. M. 3R-Hydroxy-3â-trifluoro-
methyl-5R-pregnan-20-one (Co 2-1970): A partial agonist at the
neuroactive steroid site of the γ-aminobutyric acidA receptor.
Mol. Pharmacol. 1996, 49 (5), 897-906.
(14) Sonogashira, K.; Tohda, Y.; Hagihara, N. A convenient synthesis
of acetylenes: Catalytic substitution of acetylenic hydrogen with
bromoalkenes, iodoarenes, and bromopyridines. Tetrahedron
Lett. 1975, 4467-4470.
(15) Wieland, S.; Belluzi, J . D.; Stein, L.; Lan, N. C. Comparative
behavioral characterization of the neuroactive steroids 3R-OH-
J M9605344