Synthesis and in vitro activity of 3β-substituted-3α-hydroxypregnan- 20-ones: Allosteric modulators of the GABA(A) receptor

Article abstract of DOI:10.1021/jm960021x

Two naturally occurring metabolites of progesterone, 3α-hydroxy-5α- and 5β-pregnan-20-one (1 and 2), are potent allosteric modulators of the GABA(A) receptor. Their therapeutic potential as anxiolytics, anticonvulsants, and sedative/hypnotics is limited b

Full text of DOI:10.1021/jm960021x

J . Med. Chem. 1997, 40, 61-72
61
Syn th esis a n d in Vitr o Activity of 3â-Su bstitu ted -3r-h yd r oxyp r egn a n -20-on es:
Alloster ic Mod u la tor s of th e GABAA Recep tor †
Derk J . Hogenkamp,*,‡ S. Hasan Tahir,‡ J on E. Hawkinson,‡ Ravi B. Upasani,‡ Mian Alauddin,§
Catherine L. Kimbrough,‡ Manuel Acosta-Burruel,‡ E. R. Whittemore,‡ R. M. Woodward,‡ Nancy C. Lan,‡
Kelvin W. Gee,∇ and Michael B. Bolger|
CoCensys, Inc., 213 Technology Drive, Irvine, California 92618, Department of Radiology, School of Medicine, and Department
of Pharmaceutical Sciences, School of Pharmacy, University of Southern California, Los Angeles, California 90033, and
Department of Pharmacology, College of Medicine, University of California, Irvine, California 92717
Received J anuary 9, 1996X
Two naturally occurring metabolites of progesterone, 3R-hydroxy-5R- and 5â-pregnan-20-one
(1 and 2), are potent allosteric modulators of the GABAA receptor. Their therapeutic potential
as anxiolytics, anticonvulsants, and sedative/hypnotics is limited by rapid metabolism. To avoid
these shortcomings, a series of 3â-substituted derivatives of 1 and 2 was prepared. Small
lipophilic groups generally maintain potency in both the 5R- and 5â-series as determined by
inhibition of [35S]TBPS binding. In the 5R-series, 3â-ethyl, -propyl, -trifluoromethyl and
-(benzyloxy)methyl, as well as substituents of the form 3â-XCH2, where X is Cl, Br, or I or
contains unsaturation, show limited efficacy in inhibiting [35S]TBPS binding. In the 5â-series,
the unsubstituted parent 2 is a two-component inhibitor, whereas all of the 3â-substituted
derivatives of 2 inhibit TBPS via a single class of binding sites. In addition, all of the
3-substituted 5â-sterols tested are full inhibitors of [35S]TBPS binding. Electrophysiological
measurements using R1â2γ2L receptors expressed in oocytes show that 3â-methyl- and 3â-
(azidomethyl)-3R-hydroxy-5R-pregnan-20-one (6 and 22, respectively) are potent full efficacy
modulators and that 3R-hydroxy-3â-(trifluoromethyl)-5R-pregnan-20-one (24) is a low-efficacy
modulator, confirming the results obtained from [35S]TBPS binding. These results indicate
that modification of the 3â-position in 1 and 2 maintains activity at the neuroactive steroid
site on the GABAA receptor. In animal studies, compound 6 (CCD 1042) is an orally active
anticonvulsant, while the naturally occurring progesterone metabolites 1 and 2 are inactive
when administered orally, suggesting that 3â-substitution slows metabolism of the 3-hydroxyl,
resulting in orally bioavailable steroid modulators of the GABAA receptor.
Ch a r t 1. Naturally Occurring Neuroactive Steroids
In tr od u ction
Therapeutically useful anticonvulsants, anxiolytics,
and sedative-hypnotics such as benzodiazepines (BZ)
and barbiturates mediate their action by binding to
distinct allosteric modulatory sites on the GABAA
receptor-Cl- channel complex. There is now a large
body of evidence for an additional modulatory site on
GABAA receptors that binds neuroactive steroids.1 The
prototypical ligand for this binding site is epiallopreg-
nanolone (3R-hydroxy-5R-pregnan-20-one, 1; Chart 1)
an endogenous metabolite of progesterone that demon-
strates potent modulatory effects at the GABAA recep-
tor.
Most neuroactive steroids studied for their interaction
with the GABAA receptor are endogenously occurring
metabolites of progesterone and deoxycorticosterone.1
Anesthetic properties of a large number of steroids have
previously been described, but whether the observed
anesthetic properties are wholly mediated by GABAA
receptors, or by charges in membrane fluidity, is still
uncertain.2
one (5RTHDOC) have been shown to allosterically
inhibit the binding of [35S]-tert-butylbicyclophospho-
rothionate ([35S]TBPS), a cage convulsant which binds
in close proximity to the chloride channel portion of the
GABAA receptor.3 We have recently demonstrated that
the rank order potency of 15 neuroactive steroids in
inhibiting [35S]TBPS binding parallels their ability to
potentiate GABA-induced currents in electrophysiologi-
cal studies.4 This finding suggests that the TBPS
inhibition assay is generally a good measurement of the
relative potencies of neuroactive steroids at the GABAA
receptor. In addition, two other metabolites of 1 and 2,
5R-pregnane-3R,20R-diol and 5â-pregnane-3R,20â-diol,
are representative of neuroactive steroids that show
limited inhibition of [35S]TBPS binding even at high
concentrations.5,6 The limited efficacy of these neuro-
active steroids has also been demonstrated electrophysi-
ologically,7 indicating that the [35S]TBPS binding assay
can also provide an estimate of the relative functional
efficacies of neuroactive steroids.
Endogenously occurring neuroactive steroids such as
1, its 5â-epimer 2, or 3R,21-dihydroxy-5R-pregnan-20-
† Presented in part at the 209th National Meeting of the American
Chemical Society, Anaheim, CA, April 2-6, 1995.
* Author for correspondence.
‡ CoCensys, Inc.
§ Department of Radiology.
∇ University of California.
Given their mechanism of action, endogenous neuro-
| Department of Pharmaceutical Sciences.
X Abstract published in Advance ACS Abstracts, December 15, 1996.
active steroids are expected to have anticonvulsant,8
S0022-2623(96)00021-0 CCC: $14.00 © 1997 American Chemical Society

62 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 1
Hogenkamp et al.
Sch em e 1
Ch a r t 2
Sch em e 2
anxiolytic,9 and sedative-hypnotic10 activities. Indeed,
these actions have been demonstrated in a variety of
animal models.8-10 Nevertheless, naturally occurring
neuroactive steroids have therapeutic limitations be-
cause they are rapidly metabolized, presumably by
conjugation of the 3R-hydroxyl or oxidation to the
corresponding ketones.11 Phillips et al.12 have described
the preparation of simple 3â-alkyl-substituted deriva-
tives of 1 that increase the in vivo half-life compared to
1 as anesthetics. Expanding on this work, we have
synthesized and characterized the in vitro activity of a
series of 3â-substituted derivatives of 1 and 2.
12
ylide15 regioselectively and stereoselectively gave the 3R-
epoxide16 4a . Only under forcing conditions, refluxing
in THF, was any reaction with excess ylide observed.
Ethers 15-18 were prepared by addition of epoxide 4a
to a solution of sodium in the desired alcohol. Some 17-
epimerization was observed under these reaction condi-
tions, but the 17-epimer was easily removed by recrys-
tallization or column chromatography.
Reaction of 4a with tetramethylammonium chloride
in DMF in the presence of an excess of acetic acid
afforded the â-chloro alcohol 26 in 68% yield. In the
absence of acetic acid, the reaction failed to go to
completion even with a large excess of chloride. Pre-
sumably, an equilibrium between the epoxide and the
alkoxide formed by chloride addition is driven to comple-
tion by protonation of the alkoxide with acid. Addition
of tetramethylammonium bromide under similar condi-
tions gave a mixture of bromide 27, 3-(hydroxymethyl)-
5R-pregn-2-en-20-one, and 3â-(acetoxymethyl)-3R-hy-
droxy-5R-pregnan-20-one, identified by 1H NMR. Column
chromatography gave 27 in 22% yield. Reaction of 4a
with NaI in THF/methanol/HOAc afforded iodide 28 in
67% yield. The addition of sodium thiomethoxide and
sodium azide13 to 4a gave the expected opening products
20 and 22, respectively. Hydrogenolysis of 18 gave the
diol 19 in 91% yield.
Opening of epoxide 4a was generally effective with
nucleophiles that were compatible with the 20-ketone.
However, reaction of 4a with tetrabutylammonium
fluoride gave a mixture of the desired fluoride 25 and
its 3- and 17-epimers. To avoid 17-epimerization, the
reaction was repeated with the protected epoxide 5a
(Scheme 2). Opening of the epoxide was only observed
in very concentrated reaction mixtures with carefully
dried (azeotropic removal of water) tetrabutylammo-
nium fluoride. After deprotection, 25 was isolated in
59% yield. Reaction of 5a with potassium cyanide in
EtOH gave the desired nitrile 21 in 52% yield after
deprotection. Hydrolysis of the cyanide 21 gave primary
Ch em istr y
The preparation of the 3â-substituted-3R-hydroxy-5R-
and -5â-pregnan-20-ones listed in Tables 1 and 2 was
accomplished using two methods depending on the
nature of the 3â-substituent and whether the steroid
was a 5R- or 5â-pregnane: (1) Reaction of organome-
tallic reagents with 20,20-ethylenedioxypregnan-3-one
(3) or (2) addition of nucleophiles to (3R)-spiro[oxirane-
2′,5R-pregnan]-20-one (4) or to a 20-ketal-protected
analog (5; Scheme 1).13 Compounds 6-8 in the 5R-
series and compounds 30 and 31 in the 5â-series are
covered by Glaxo patents.12 Of these compounds, only
the preparation and characterization of 6 was de-
scribed.12a
For comparison with the 3â-substituted analogs, 3R-
hydroxy-5R-pregnan-20-one (1) was prepared from 3â-
hydroxypregn-5-en-20-one (hydrogenation followed by
Mitsunobu reaction with trifluoroacetate and hydrolysis
to the 3R-ol). In the 5R-series, the alkyne 10 was
prepared from 20,20-(ethylenedioxy)-5R-pregnan-3-one
(3a ; prepared in three steps from 3â-hydroxypregn-5-
en-20-one) and lithium acetylide-ethylenediamine com-
plex followed by deprotection and separation of the 3R-
and 3â-ols. Reaction of 3a with allylmagnesium chlo-
ride and deprotection afforded 11 again accompanied
by its 3-epimer. The less polar 3R-ols (axial hydroxyl)
were easily separated from the more polar 3â-ols
(equatorial hydroxyl) by flash chromatography using
hexane/acetone mixtures. The 3â-trifluoromethyl de-
rivative 24 was isolated as the minor isomer (ratio of
3â- to 3R-ol 10:1) from the fluoride-initiated addition of
14
TMSCF3 to 3a .
The synthesis of 15-18, 20, 22, and 26-28 by the
oxirane method is outlined in Table 3. Reaction of 5R-
pregnane-3,20-dione with dimethylsulfoxonium meth-

3â-Substituted-3R-hydroxypregnan-20-ones
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 1 63
Sch em e 3
Ch a r t 3
F igu r e 1. Crystal structure of 6.
Sch em e 4
amide 29. A DMSO suspension of 5a reacted smoothly
with an excess of lithium acetylide-ethylenediamine
complex to give 12 in 93% yield after liberation of the
20-ketone. No reaction was observed between 5a and
lithium acetylide-ethylenediamine complex or lithium
trimethylsilylacetylide in THF.
Hydrogenation of the azide (H2, Pd/C) and alkylation
(formalin, formic acid) afforded the tertiary amine 23
in 36% yield (Scheme 3). Hydrogenation of 12 gave the
3â-propyl derivative 8. A procedure for the synthesis
of allylic alcohols from epoxides recently reported by
Alcaraz et al.17 was employed for the preparation of the
allylic alcohol 9 as a single epimer. Thus, addition of
an excess of dimethylsulfonium methylide to epoxide 5a
followed by hydrolysis of the 20-ketal gave 9 in 70%
yield. Hydrogenation of 9 (H2, 5% Pd/C) gave the
saturated analog 7.
Sch em e 5
The syntheses of compounds 6, 13, and 14 from 3R-
hydroxy-3â-(iodomethyl)-5R-pregnan-20-one (28) is given
in Scheme 4. Hydrogenolysis of 28 over 5% Pd/C gave
6 in quantitative yield. Reaction of 28 under free radical
conditions with allyltributylstannane or propargyltriph-
enylstannane gave 13 and 14 in 35% and 11% yields,
respectively.18
A crystal structure of 6 (Figure 1) confirmed the
stereochemistry of the 3â-methyl substituent. The 3â-
methyl in 6 then confirms the identity of the 3R-epoxide
4a and the 3-stereochemistry of its derivatives (com-
pounds 7-9, 12-23, and 25-29, Table 1). The stere-
ochemistry of the allyl group in 11 was confirmed by
hydrogenation and comparison with authentic 8 pre-
pared from 3â-(3-propyne) 12.
In the 5â-series, 3R-hydroxy-5â-pregnan-20-one (2)
was prepared from 20,20-(ethylenedioxy)-5â-pregnan-
3-one (3b) by reduction with lithium tri-tert-butoxyalu-
minum hydride and hydrolysis to unmask the 20-
ketone. The 5â-pregnanes 30 and 32-34 (Table 2) were
prepared from 3b (Scheme 5). Addition of MeMgBr,
vinylmagnesium bromide, lithium acetylide-ethylene-
diamine complex, or 2-propynylmagnesium bromide to
3b gave a mixture of 3R- and 3â-ols which were
separated by column chromatography (3R-ols identified
as the more polar isomers). Hydrogenation of 33 (H2,
Pd/C) afforded the 3â-ethylpregnane 31 in 66% yield

64 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 1
Hogenkamp et al.
Ta ble 1. Potency and Efficacy of 3â-Substituted-5R-pregnanes
in Inhibiting [35S]TBPS Binding in Rat Brain Cortical
Membranesa
Sch em e 6
compd
R3
IC50 (nM)
Imax (%)
diazepam
Ro 16-6028
91
6.1
49
27
pentobarbital
1700
100
1
6
7
8
H
51 ( 5
95 ( 1
94 ( 1
69 ( 1
42 ( 3
94 ( 1
93 ( 1
59 ( 3
43 ( 1
59 ( 6
57 ( 4
104 ( 1
104 ( 4
94 ( 3
57 ( 3
101 ( 5
46 at 10 µM
38 ( 4
99 ( 7
0 at 10 µM
44 ( 4
79 ( 4
33 ( 3
44 ( 2
50 ( 6
28 at 10 µM
Me
Et
Pr
H2CdCH
HCtC
H2CdCHCH2
HCtCCH2
H2CdCHCH2CH2
H2CdC)CHCH2
MeOCH2
EtOCH2
n-PrOCH2
BnOCH2
HOCH2
MeSCH2
NCCH2
N3CH2
Me2NCH2
F3C
FCH2
80 ( 18
257 ( 24
173 ( 33
120 ( 7
64 ( 8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
231 ( 25
50 ( 7
325 ( 65
365 ( 47
76 ( 9
230 ( 51
655 ( 77
376 ( 50
2134 ( 79
>10000
614 ( 54
27 ( 2
>10000
266 ( 43
228 ( 51
71 ( 11
224 ( 45
702 ( 143
>10000
ClCH2
BrCH2
ICH2
H2NCOCH2
after recrystallization. The yield of 3â-methylpregnane
30 in the Grignard addition was found to be variable,
and a more reproducible procedure was found using the
epoxide 5b (Scheme 6). Reaction of 5b with an excess
of lithium aluminum hydride and deketalization gave
30, identical with the more polar isomer formed in the
Grignard reaction.
a
Compounds (1 nM-10 µM) were incubated with rat brain
cortical P2 membranes and 2 nM [35S]TBPS in the presence of 5
µM GABA as described in the Experimental Section. Values are
means and SEMs of 3-9 independent experiments except for the
inactive compounds 23 and 29 (n ) 2). Hill values were 1.0 for
all active compounds except 16 (1.2 ( 0.2) and 18 (1.6 ( 0.1). The
concentration of steroid inhibiting 50% specific binding (IC50) and
the maximal extent of inhibition (Imax) were calculated by fitting
the data to the sigmodial function. Data for pentobarbital are
taken from ref 5, and those for diazepam and Ro 16-6028
(bretazenil) are from ref 20. Data for compounds 1 and 6 are taken
from ref 26b, and data for 24 are from ref 24.
(Chloromethyl)- and (methoxymethyl)pregnanes 35
and 36 were obtained from epoxide 5b as described for
the preparation of the 5R-analogs 26 and 15. The
synthesis of epoxide 5b is given in Scheme 6. Reaction
of 20,20-(ethylenedioxy)-5â-pregnan-3-one (3b) with
dimethylsulfoxonium methylide gave exclusively the
undesired 3â-epoxide. A 2:1 mixture of the 3R- and 3â-
epoxides was isolated from the reaction of 3b and
dimethylsulfonium methylide. In an attempt to improve
on this ratio, 3b was converted to the exocyclic alkene
37 and subjected to epoxidation. Epoxidation of similar
exocyclic alkenes with hydrogen peroxide in benzonitrile
results in a 2:1 ratio of equatorial to axial attack.19
Unfortunately, no reaction with 37 was observed under
these conditions. Epoxidation of 37 with m-chloroper-
oxybenzoic acid gave a 1:2 ratio of 3R- and 3â-epoxides
which could be separated by flash chromatography. The
more polar 3R-epoxide (TLC with hexane/acetone mix-
tures) was identified by comparison with the less polar
3â-epoxide formed from dimethylsulfoxonium methylide
and 3b. More conveniently, the mixture of epoxides was
separated after reaction with the desired nucleophile.
Tables 1 and 2, respectively. The concentration of
steroid inhibiting 50% specific binding (IC50) and the
maximal extent of inhibition (Imax) were calculated by
fitting [35S]TBPS inhibition data to a sigmodial function.
Compounds with Imax > 90% are considered full ago-
nists, while compounds exhibiting incomplete displace-
ment of [35S]TBPS binding at high concentration are
partial agonists (Imax < 90% at > 1 µM). For compari-
son, the IC50 and Imax values for the full agonist
benzodiazepine diazepam20 and partial agonist benzo-
diazepine Ro 16-6028 (bretazenil)20 are included along
with data for pentobarbital.5 The potencies of the
synthetic steroids vary from more potent than 1 and 2
to completely inactive (IC50 > 10 000 nM). Lipophilic
substituents are generally accommodated at the 3â-
position, while the polar, hydrogen bond-donating groups
tested (3â-hydroxymethyl 19 and 3â-carboxamidomethyl
29) and the tertiary amine 23 reduce or abolish activity.
For simple alkyl substituents at the 3â-position, there
is no simple relationship between the size of the group
and the ability to inhibit [35S]TBPS binding. In the
P h a r m a cology
[
35S]TBP S Bin d in g Stu d ies. The ability of 3â-
substituted-3R-hydroxy-5R- and -5â-pregnan-20-ones to
allosterically modulate the binding of [35S]TBPS in rat
brain cortical membranes is compared to the endog-
enously occurring progesterone metabolites 1 and 2 in

3â-Substituted-3R-hydroxypregnan-20-ones
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 1 65
Ta ble 2. Potency and Efficacy of 3â-Substituted-5â-pregnanes
in Inhibiting [35S]TBPS Binding in Rat Brain Cortical
Membranesa
Ta ble 3. Reaction of (3R)-Spiro[oxirane-2′,5R-pregnan]-20-one
(4a ) with Nucleophiles
compd
R3
IC50 (nM)
Imax (%)
compd
reagents
yield (%)
2
H
44 ( 11
12380 ( 5040
37 ( 10
65 ( 3
35 ( 3
104 ( 1
95 ( 4
103 ( 1
101 ( 1
101 ( 2
103 ( 2
103 ( 1
15
16
17
18
20
22
26
27
28
MeONa
EtONa
n-PrONa
BnONa
MeSNa
NaN3
Me4NCl
Me4NBr
NaI
58
57
40
52
57
73
68
22
67
30
31
32
33
34
35
36
Me
Et
135 ( 7
H2CdCH
HCtC
HCtCCH2
ClCH2
43 ( 4
39 ( 5
214 ( 20
243 ( 52
40 ( 5
MeOCH2
a
See footnote to Table 1 for methods and data analysis.
Compound 2 (0.2 nM-100 µM) inhibits TBPS binding with two-
components. Hill values were 1.0 for all active compounds except
31 (1.3 ( 0.1), 33 (1.12 ( 0.03), 34 (1.2 ( 0.1), and 35 (1.3 ( 0.1).
Data for compound 30 are taken from ref 26a.
loss in activity in contrast to the same substitution in
the 5R-series which results in compounds which are
approximately equipotent with the parent 1. Unlike
their 5R counterparts, 31, 34, and 35 are full inhibitors
of [35S]TBPS binding.
The 3â-hydroxy isomers of compounds 1, 6, 24, 30,
and 32 were tested in the [35S]TBPS binding assay and
found to be from 50- to 1000-fold less active than the
3R-ols. The strict stereochemical requirement of 3R-ol
for activity has previously been noted.22
series 3â-H, -Me, -Et, and -Pr (1 and 6-8, Table 1), 1
and 6 are equipotent while a 5-fold loss is observed for
the 3â-ethylpregnane 7. The 3â-propyl homolog 8,
however, is more potent than 7. The addition of a
methylene to alkyne 10 gives the homolog 12 with equal
potency. The ethers 15-17 show a 3-fold loss of activity
for each methylene added. The corresponding thioether
(20) of 15 is inactive. The 3â-trifluoromethyl derivative
24 and the 3â-fluoromethyl analog 25 show substantial
(4.5-5.2-fold) reduction in activity compared to 1. The
other halogens (26-28) show a decrease in activity with
increasing size. The most potent compound in this
series appears to be the azide 22, which is 2-fold more
potent than 1.
In addition to exhibiting a range of potencies, the
steroids examined show varied levels of efficacy in
inhibiting [35S]TBPS binding. In general, active 5R-
compounds incorporating a 3â-group XCH2 show limited
inhibition if X contains unsaturation (compounds 11-
14 and 21) or if X is Cl, Br, or I (halomethyls 26-28).
Simple alkyl groups larger than methyl (3â-Et, 7, and
3â-Pr, 8) also appear to be limited efficacy inhibitors in
the assay, as are the benzyloxymethyl (18) and trifluo-
romethyl (24) compounds. Compounds with unsatura-
tion in the 3â-substituent directly adjacent to the
3-position of the steroid A-ring are full inhibitors, even
when the corresponding saturated analog is a limited
inhibitor (e.g., compounds 9 and 10 compared with
compound 7).
The in vitro binding results for the 5â-pregnanes
examined are given in Table 2. The parent 3â-unsub-
stituted compound 2 is known to be a two-component
modulator of [35S]TBPS binding.21 The addition of a 3â-
substituent results in derivatives which inhibit [35S]-
TBPS binding with a single-component as do all the 5R-
compounds listed in Table 1. While the 3â-methyl-
pregnane 30 is equipotent with 2, the addition of a
methylene spacer in 31 results in a 3.6-fold loss in
activity. Unsaturated groups directly attached to the
steroid nucleus at the 3â-position (32 and 33) and the
methoxymethyl derivative 36 exhibit no loss in potency
compared to 2. The 3â-(2-propyne) derivative 34 and
the 3â-chloromethyl alcohol 35 result in about a 5-fold
Electr op h ysiology. Coinjection of R1, â2, and γ2L
cRNAs into oocytes resulted in strong expression of
functional GABAA receptors. The mean maximum
GABA-activated current was 2700 ( 100 nA (current
elicited by 3 mM GABA) (n ) 22), and the range of
maximum responses was between 1875 and 3630 nA.
To further characterize the receptors, GABA concentra-
tion-response curves were measured in a sample of
cells. The EC50 was 40 ( 3 µM, and the slope value
was 1.3 ( 0.1 (n ) 13), giving no indication of a
pharmacologically heterogeneous population of recep-
tors.
Functional modulation of GABAA receptors by ste-
roids was assayed as shown in Figure 2. A maximum
response was elicited by applying 3 mM GABA. All
other currents were scaled to this value. The GABA
concentration was then adjusted to elicit a fractional
current of approximately 0.05 (5%); the mean GABA
concentration was 5.9 ( 0.6 µM, and the mean response
was 0.051 ( 0.001 nA (n ) 22). Potency and efficacy of
modulation were assayed by measuring effects of in-
creasing concentrations of steroid on the fixed 0.05
control response. Steroid-activated currents, i.e., cur-
rents observed in the absence of GABA, were measured
as the peak response elicited during preincubations
(Figure 2, arrow). Five steroids were selected for
comparison with the [35S]TBPS inhibition experi-
ments: 1, 6, 19, and 22-24. Modulation was measured
in terms of total current (Figure 3, upper panel) and
steroid-activated currents (Figure 3, lower panel).
The endogenous steroid 1 was a potent, full efficacy
modulator of R1â2γ2L receptors expressed in oocytes
(Figure 3, Table 4). At concentrations > 30 nM, 1, in
the nominal absence of GABA, also activated inward
membrane currents. These responses were sensitive to
inhibition by bicuculline and picrotoxin, indicating that

66 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 1
Hogenkamp et al.
Ta ble 4. Potency and Efficacy of 3â-Substituted-5R-pregnanes
in Functional Modulation of Cloned R1â2γ2L GABAA Receptors
Expressed in Xenopus Oocytesa
compd
R3
EC50 (nM)
slope
efficacy (FR)
n
1
6
19
22
23
24
H
CH3
HOCH2
N3CH2
Me2NCH2
F3C
160 ( 50
260 ( 40
∼20000
1.5 ( 0.1
1.2 ( 0.1
0.91 ( 0.01
0.80 ( 0.04
3
5
3
3
2
3
150 ( 50
>50000
1.4 ( 0.2
1.2 ( 0.1
0.80 ( 0.05
0.30 ( 0.02
390 ( 70
a
F igu r e 2. Sample records illustrating potentiation of GABA-
activated currents by 22 in Xenopus oocytes expressing human
R1â2γ2L receptor subunits. For this cell the 0.05 (5%) GABA
response required 6 µM GABA. Responses to 3 mM GABA
(maximum currents) are included for comparison. Steroid-
activated currents were apparent at higher concentrations of
22 (e.g., arrow). Holding potential was -70 mV, periodically
pulsed to -60 mV to help time drug applications. Inward
current is indicated by downward deflection. Individual records
were separated by 2-10 min intervals of wash.
Potentiation of GABA-activated currents by steroids was
measured as described in Results and the Experimental Section.
EC50 and slope values are the optimum logistic fits of data
presented in Figure 3. For 19 an approximate EC50 was estimated
by extrapolation, assuming full efficacy modulation (Table 1).
Efficacy is expressed in terms of a fractional response (FR)
calculated with respect to the maximum GABA current elicited
in the cell. Data are given as the mean ( SEM, quoted to two
significant figures.
The azido analog 22 showed equal potency to 1 in
modulation of GABA responses. The methyl-substituted
analog 6 was 1.6 times less potent. Both steroids were
full efficacy modulators and, like 1, caused some direct
activation of R1â2γ2L receptors. EC50 values for steroid-
activated currents were 0.28 µM for 22 and 0.73 µM for
6. The fractional peak efficacy of responses elicited in
the absence of GABA by 22 and 6 was approximately
0.07.
The (trifluoromethyl)pregnane 24 was 2.4 times less
potent than 1. As seen in the [35S]TBPS binding, 24
showed distinctly low-efficacy modulation with maxi-
mum potentiation only 0.3 of the peak GABA response.
Changing DMSO concentrations over the range 0.1-
1% had little effect on levels of modulation induced by
1 µM 24, suggesting that solubility was not a factor in
limiting efficacy. The 3â-(hydroxymethyl)pregnane 19
potentiated GABA currents with low-potency, and the
tertiary amine 23 was inactive at concentrations up to
10 µM. The latter compound was also tested for
antagonism at steroid sites. At 10 µM, 23 did not inhibit
modulation induced by 30 nM 6, indicating it is not a
steroid site antagonist. The low-efficacy modulator 24
and the low-potency compound 19 elicited only small
steroid-activated currents that were <0.02 of the maxi-
mum GABA response.
F igu r e 3. Functional modulation of human R1â2γ2L recep-
tors by 3â-substituted-5R-pregnanes: (upper panel) potentia-
tion of GABA responses by steroids and (lower panel) steroid-
activated currents in the nominal absence of GABA; note the
expansion of the y-axis. Data are plotted as mean (SEM,
expressed as a fraction of the maximum current elicited by
GABA; FR, fractional response. The number of separate
experiments is given in parentheses. Smooth curves, best fits
of the logistic equation to the data. Optimum EC50 and slope
values for data in the upper panel are given in Table 3. EC50
(µM) and slope values for the steroid-activated currents are,
respectively: 1, 0.98, 1.2; 6, 0.73, 1.2; and 22, 0.27, 1.7.
Con clu sion s
The 3â-modifications made to the naturally occurring
progesterone metabolites 1 and 2 generally maintain
good potency at the neuroactive steroid binding site as
determined by inhibition of [35S]TBPS binding and
confirmed in three cases by electrophysiological assay.
Small lipophilic groups in this position are well accom-
modated, as well as larger groups incorporating unsat-
uration. In the 5R-series, 3â-ethyl, -propyl, -trifluoro-
methyl, and -(benzyloxy)methyl, as well as substituents
of the form XCH2, where X contains unsaturation or is
Cl, Br, or I, show limited efficacy in inhibiting [35S]TBPS
binding. Limited efficacy may be due to partial agonism
at the neuroactive steroid binding site and/or neuroac-
tive steroid receptor heterogenity. The latter possibility
the currents were due to direct steroidal activation of
GABAA receptors. Maximum steroid-activated currents
for 1 were between 0.15 and 0.2 of the peak GABA
response, and the EC50 was 0.98 µM.

3â-Substituted-3R-hydroxypregnan-20-ones
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 1 67
are referenced to polystyrene film at 1602 cm-1
.
Melting
exists since numerous GABAA receptor subunits and
variants are expressed in rat cortex.23 Electrophysi-
ological and binding studies using recombinant recep-
tors have confirmed that the trifluoromethyl compound
24 (Co 2-1970) is a partial agonist for the neuroactive
steroid site of the GABAA receptor.24 It is possible that
partial agonist neuroactive steroids may have superior
in vivo profiles compared to full agonists. Partial
agonist benzodiazepines have been found to have im-
proved therapeutic profiles in preclinical studies com-
pared to benzodiazepine full agonists, including less
sedation, decreased interaction with alcohol, and di-
minished abuse potential.25
In the 5â-series, addition of a 3â-substituent to the
two-component inhibitor 2 gives compounds which
inhibit [35S]TBPS binding via a single class of binding
sites. Although the nature of the two-component inhibi-
tion of TBPS binding in rat cortical membranes has not
been fully explained, GABAA receptors could contain two
nonequivalent steroid binding sites per receptor complex
which may interact with negative cooperativity. All of
the 3â-substituted-5â-pregnanes tested (Table 2) are
one-component full inhibitors of [35S]TBPS binding,
suggesting that these neuroactive steroids are full
agonists which are nonselective and do not display
negative cooperativity.
The apparent potency of steroids as measured by
electrophysiological assay was consistently lower than
that measured by [35S]TBPS binding, with ranges of
1.5-10-fold for the compounds tested. Nonetheless,
there was generally good agreement between the two
assay systems.4 In particular, the electrical assays
confirmed that 1, 6, and 22 are potent full efficacy
agonists, that 24 is a low-efficacy modulator, that 19 is
a low-potency modulator, and that 23 is inactive. The
potency difference found between 1 and 22 in the [35S]-
TBPS binding assay was not apparent in electrophysi-
ological assays, and while 1 and 6 are equipotent in the
[35S]TBPS binding assay, 6 is 1.6-fold less potent than
1 with R1â2γ2L receptors. These modest differences in
relative potencies may simply reflect differences be-
tween allosteric modulation of the TBPS binding site
as opposed to the channel itself or may be a consequence
of differences in subtype selectivity. For example, some
classes of the GABAA receptor found in rat brain
membranes could have different sensitivities for steroids
than the cloned R1â2γ2L receptors expressed in oocytes.
In addition to maintaining in vitro potency, the
presence of 3â-substitution in the steroids synthesized
has been shown to improve oral in vivo activity when
compared with 1 and 2. For example, compound 6 (CCD
1042) is an orally active anticonvulsant in animals,
whereas 1 and 2 are inactive.26 Thus, appropriate 3â-
substitution of neuroactive steroids results in potent
GABAA receptor modulators which have activity follow-
ing oral administration and should prevent conversion
to metabolites with hormonal side effects. These com-
pounds represent a clear step forward in the develop-
ment of neuroactive steroids for treatment of neuro-
psychiatric disorders.
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 . Flash chro-
matography on silica gel (230-400 mesh; Mallinckrodt) was
carried out as described by Still.27 HPLC grade solvents were
obtained from Baxter. THF and ether were distilled from
sodium and benzophenone. Methylene chloride and DMSO
were dried over CaH2 and distilled. Other synthetic reagents
were obtained from Aldrich Chemical Co. (Milwaukee, WI) and
used as received unless otherwise specified. 20,20-(Ethylene-
dioxy)-5R- and -5â-pregnan-3-ones (3a ,b, respectively) and 5R-
pregnane-3,20-dione were obtained from DioSynth (Oss, The
Netherlands) or prepared as described in the literature.28 3R-
Hydroxy-5R-pregnan-20-one (1) was obtained from Steraloids
(Windham, NH) or prepared as described in the literature.29
The 5â-isomer 2 was obtained from DioSynth. 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 Biochemicals International (Natick, MA).
(3R)-Sp ir o[oxir a n e-2′,5r-p r egn a n ]-20-on e (4a ). To a
stirred solution of trimethylsulfoxonium iodide (5.29 g, 24.0
mmol) in 75 mL of DMSO was added NaH (97%; 488 mg, 19.7
mmol). After stirring at room temperature for 1 h, a suspen-
sion of 5R-pregnane-3,20-dione (1.54 g, 4.86 mmol) in 50 mL
of DMSO was added dropwise via addition funnel. After 2.5
h, the reaction mixture was poured into ice-cold water and
extracted with ether (3×). The combined ether layers were
then washed with brine, dried (MgSO4), filtered, and concen-
trated in vacuo. Recrystallization from 50 mL of 1:1 MeOH/
acetone gave 1.18 g (73%) of 4a as a white solid: mp 161 °C;
1H NMR (200 MHz, CDCl3) δ 2.62 (s, 2H), 2.54 (t, 1H), 2.11 (s,
3H), 0.84 (s, 3H), 0.61 (s, 3H). Anal. (C22H34O2) C, H.
20,20-(E t h ylen ed ioxy)-(3R)-sp ir o[oxir a n e-2′,5r-p r eg-
n a n e] (5a ). A mixture of trimethylsulfoxonium iodide (3.80
g, 17.3 mmol) and potassium tert-butoxide (95%; 1.87 g, 15.8
mmol) in 50 mL of dry THF was refluxed for 1.5 h. To the
resulting mixture at room temperature was added solid 3a
(4.81 g, 13.3 mmol). After stirring at room temperature for 3
h, the solvent was removed and the residue was partitioned
between CH2Cl2, and water. The aqueous layer was extracted
with CH2Cl2, and the combined organic layers were washed
with a saturated NaCl solution, dried (MgSO4), filtered, and
evaporated to dryness. Recrystallization from 1:1 MeOH/
acetone afforded 3.15 g (63%) of 5a as a white solid: mp 169-
171 °C; 1H NMR (300 MHz, CDCl3) δ 4.05-3.85 (m, 4H), 2.61
(s, 2H), 1.29 (s, 3H), 0.84 (s, 3H), 0.76 (s, 3H). Anal. (C24H38O3)
C, H
20,20-(E t h ylen ed ioxy)-(3R)-sp ir o[oxir a n e-2′,5â-p r eg-
n a n e] (5b). To a solution of 37 (1.05 g, 2.94 mmol) in 10
mL of dry CH2Cl2 was added anhydrous potassium carbonate
(500 mg, 3.62 mmol). The resulting mixture was cooled in an
ice/water bath, and solid 3-chloroperoxybenzoic acid (50-60%
by weight; 860 mg, 2.99 mmol if 60%) was added in a few
portions. After stirring for 1 h, the mixture was filtered and
the precipitated benzoate salt was washed with CH2Cl2. The
solvent was removed in vacuo, and the residue was chromato-
graphed (10% hexane/CH2Cl2). Along with 679 mg (62%) of
the undesired epoxide (Rf 0.3, 100% CH2Cl2) the desired
epoxide 5b (407 mg) was isolated contaminated with 20% of
the undesired epoxide. The impure desired epoxide was
rechromatographed (10% hexane/CH2Cl2). The desired epoxide
(Rf 0.23 100% CH2Cl2) was isolated as a white solid in 26%
yield: mp 115.5-117.5 °C. 1H NMR (300 MHz, CDCl3) δ 4.00-
3.83 (m, 4H), 2.58 (s, 2H), 2.39 (t, 1H, J ) 13.2 Hz), 1.30 (s,
3H), 0.98 (s, 3H), 0.75 (s, 3H). Anal. (C24H38O3) C, H.
3r-Hyd r oxy-3â-m eth yl-5r-p r egn a n -20-on e (6). A sus-
pension of 4a (10.0 g, 30.3 mmol) in 150 mL of 1:1 THF/MeOH
was treated with sodium iodide (7.1 g, 45 mmol) and glacial
acetic acid (4.5 mL, 45 mmol). After heating at reflux for 1.5
h, an additional 1.0 g of NaI and 0.65 mL of HOAc were added,
and refluxing was continued for an additional 30 min. The
resulting mixture was diluted with 175 mL of 1:1 methanol/
THF, and solid NaOAc (8 g) was added followed by 2 g of 5%
Pd/C. After stirring under an atmosphere of hydrogen gas for
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 CDCl3 with tetramethylsilane
as reference. 19F NMR spectra in CDCl3 are referenced to
FCCl3 at 0.00 ppm. Infrared spectra were obtained as KBr
pellets on a Perkin-Elmer 1600 Series FTIR instrument and

68 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 1
Hogenkamp et al.
18 h, the catalyst was removed by filtration and the solvent
was removed under reduced pressure. The solution of the
residue in CH2Cl2 was washed successively with a saturated
NaHCO3 solution, a 10% aqueous Na2S2O3 solution, and a
saturated NaCl solution. After drying (MgSO4), the solvent
was filtered and evaporated. The crude product was recrystal-
lized twice from MeOH to give 5.1 g (51%) of 6 as a white
solid: mp 190-192 °C (lit.12a mp 189 °C); 1H NMR (300 MHz,
CDCl3) δ 2.53 (t, 1H, J ) 8.8 Hz), 2.11 (s, 3H), 1.20 (s, 3H),
0.75 (s, 3H), 0.60 (s, 3H). Anal. (C22H36O2) C, H.
3â-Eth yl-3r-h yd r oxy-5r-p r egn a n -20-on e (7). A solution
of 9 (312 mg, 0.91 mmol) in 10 mL of EtOH was treated with
40 mg of 5% Pd/C and stirred under a hydrogen atmosphere
for 24 h. Removal of the catalyst by filtration and concentra-
tion in vacuo gave crude 7. Recrystallization from 3:1 hexane/
acetone gave 222 mg (71% yield) of 7 as a white solid: mp
163-165 °C; 1H NMR (300 MHz, CDCl3) δ 2.53 (t, 1H, J ) 8.8
Hz), 2.11 (s, 3H), 0.91 (t, 3H, J ) 7.4 Hz), 0.74 (s, 3H), 0.60 (s,
3H). Anal. (C23H38O2) C, H.
EtOAc (3 × 25 mL), and the pooled organic layers were washed
with brine, dried (Na2SO4), filtered, and concentrated. Column
chromatography (elution with a gradient from 100% CH2Cl2
to 1.5% acetone/CH2Cl2) and recrystallization from hexane
afforded 154 mg (12%) of the alkene as a white solid: mp 163-
164 °C; 1H NMR (300 MHz, CDCl3) δ 5.89 (ddt, 1H, J ) 16.9,
10.3 Hz), 5.15 (dd, 1H, J ) 10.3, 2.1 Hz), 5.11 (dd, 1H, J )
15.9, 2.0 Hz), 2.54 (t, 1H), 2.11 (s, 3H), 0.74 (s, 3H), 0.60 (s,
3H). Anal. (C24H38O2) C, H.
3r-Hyd r oxy-3â-(2-p r op yn yl)-5r-p r egn a n -20-on e (12). A
suspension of lithium acetylide-ethylenediamine complex
(2.48 g, 26.9 mmol) in 12 mL of dry DMSO was cooled in an
ice/water bath, and solid 5a (2.02 g, 5.39 mmol) was added.
After stirring at room temperature for 21 h, the mixture was
cooled to 0 °C and added to 100 mL of an ice-cold saturated
NH4Cl solution. After addition of 100 mL of water, the
mixture was extracted with EtOAc (4 × 50 mL). The combined
EtOAc extracts were washed with a saturated NaCl solution,
dried (Na2SO4), filtered, and concentrated. The 20-ketal
obtained was dissolved in 25 mL of THF and treated with 5
mL of a 1 N HCl solution and 5 mL of acetone. After stirring
for 3 h, the reaction mixture was cooled in an ice/water bath.
The precipitate that formed was isolated and washed with a
total of 35 mL of 4:1:1 THF/water/acetone, affording 1.3 g (68%)
of 12 as a white solid: mp 218-223 °C dec. The mother liquor
was added to 10 mL of a saturated NaHCO3 solution and
extracted with EtOAc (3 × 50 mL). The organic layers were
combined, washed with a saturated NaCl solution, dried, and
concentrated to give an additional 487 mg of 12: total yield
3r-Hyd r oxy-3â-p r op yl-5r-p r egn a n -20-on e (8). As de-
scribed for the synthesis of 7, hydrogenation of 12 afforded 8
in 32% yield after recrystallization from hexane/acetone: 1H
NMR (300 MHz, CDCl3) δ 2.53 (t, 1H), 2.11 (s, 3H), 0.74 (s,
3H), 0.60 (s, 3H). Anal. (C24H40O2‚2/3H2O) C, H.
3â-Eth en yl-3r-h yd r oxy-5r-p r egn a n -20-on e (9). A sus-
pension of trimethylsulfonium iodide (365 mg, 1.79 mmol) in
4 mL of THF at -10 °C was treated with a 2.5 M solution of
n-BuLi in hexanes (0.75 mL, 1.87 mmol). The resulting
mixture was allowed to warm to 0 °C and then recooled to -10
°C, and a solution of 5a in 3 mL of dry THF was added via
syringe. The reaction mixture was allowed to warm to room
temperature and stirred overnight. After partitioning between
water and EtOAc, the reaction mixture was dried (Na2SO4),
filtered, and concentrated. Removal of the 20-ketal with 1 N
HCl/acetone/THF and flash chromatography (15% acetone/
hexane) gave 9 in 70% yield as a white solid: mp 163-165
1
1.79 g (93%); H NMR (300 MHz, CDCl3) δ 2.53 (t, 1H, J )
8.8 Hz), 2.33 (s, 1H), 2.32 (s, 1H), 2.11 (s, 3H), 2.17 (s, 1H),
0.75 (s, 3H), 0.60 (s, 3H). Anal. (C24H36O2) C, H.
3â-(Bu t-3-en yl)-3r-h ydr oxy-5r-pr egn an -20-on e (13). Dry
benzene (3 mL) was degassed for 15 min with argon. Next
were added successively 3R-hydroxy-3â-(iodomethyl)-5R-preg-
nan-20-one (28; 263 mg, 0.574 mmol), allyltributyltin (97%;
0.5 mL, 1.56 mmol), and azobis(isobutyroylnitrile) (20 mg, 0.12
mmol). The resulting mixture was refluxed for 3.5 h and
allowed to cool to room temperature. After removing the
solvent, the residue was subjected to flash chromatography
twice, eluting with CH2Cl2 and 7:1 hexane/acetone. Compound
13 was obtained as a white solid (75 mg, 35%): mp 140-141
°C; 1H NMR (300 MHz, CDCl3) δ 5.86 (ddt, 1H, J ) 17.0, 10.2,
6.6 Hz), 5.04 (d, 1H, J ) 17.1 Hz), 4.96 (d, 1H, J ) 10.1 Hz),
2.53 (t, 1H), 2.12 (s, 3H), 0.75 (s, 3H), 0.61 (s, 3H). Anal.
(C25H40O2) C, H.
3â-(Bu tadi-2,3-en yl)-3r-h ydr oxy-5r-pr egn an -20-on e (14).
Propargyl bromide (80 wt % solution in toluene; 1.7 mL, 11
mmol) was added to a mixture of Mg (471 mg, 19.4 mmol) and
dry Et2O (25 mL). HgCl2 (15 mg) was then added. The
resulting mixture was first stirred at room temperature and
then warmed briefly. The turbid reaction mixture was stirred
at <20 °C (cold water bath) for ∼1 h, after which time most of
the Mg had reacted. Propargyl bromide (80 wt % in toluene;
0.4 mL, 2.69 mmol) was then added, and the resulting mixture
was stirred at <20 °C for ∼1 h. A suspension of triphenyltin
chloride (95%; 5.04 g, 12.4 mmol) in Et2O (50 mL) was added
dropwise, and the reaction mixture was stirred at room
temperature for ∼3 h. A saturated aqueous NH4Cl solution
(10 mL) was added, and the two phases which formed were
separated. The organic phase was dried (MgSO4), filtered, and
evaporated under reduced pressure to give a white solid.
Recrystallization from hexane (50 mL) furnished triphenyl-
prop-2-ynylstannane as white, plate-shaped crystals (3.92 g,
81%).
1
°C; H NMR (CDCl3, 300 MHz) δ 5.92 (dd, 1H, J ) 17.3, 10.7
Hz), 5.22 (d, 1H, J ) 17.3 Hz), 5.00 (d, 1H, J ) 10.7 Hz), 2.53
(t, 1H, J ) 8.9 Hz), 2.20-1.98 (m, 2H), 2.11 (s, 3H), 0.78 (s,
3H), 0.60 (s, 3H); IR 3083 (dCH2), 1700 (CdO), 993 (dCH),
908 cm-1 (dCH). Anal. (C23H36O2) C, H.
3â-Eth yn yl-3r-h yd r oxy-5r-p r egn a n -20-on e (10). Acety-
lene gas was bubbled through a mixture of lithium acetylide-
ethylenediamine complex (2.75 g, 90%, 27.5 mmol) in dry
benzene (60 mL) at a moderate rate. The mixture was then
heated to 50-55 °C and treated in parts with 3a (9 g, 25
mmol). After stirring at this temperature for 5 h and then at
room temperature for another 17 h, the resulting suspension
was cooled to 10 °C and was treated with saturated NaCl
solution (5 mL). The solvent was removed, and the residue
was taken up in water. The water insoluble product was
collected by filtration, washed with water, and dried under
vacuum. Recrystallization from EtOAc gave unreacted 3a
(3.35 g). The mother liquor was evaporated to dryness, and
the residue was purified by column chromatography. Elution
with a toluene/acetone mixture (92:8) gave unreacted 3a (1.3
g) followed by 20-ketal-protected 10. Removal of the ketal
(aqueous HCl/THF/acetone) followed by column chromatogra-
phy (92:8 toluene/acetone mixture) gave 280 mg (3%) of 10:
mp 175-177 °C; 1H NMR (300 MHz, CDCl3) δ 2.52 (t, 1H),
2.42 (s, 1H), 2.11 (s, 3H), 0.81 (s, 3H), 0.60 (s, 3H). Anal.
(C23H34O2) C, H.
3r-Hyd r oxy-3â-(2-p r op en yl)-5r-p r egn a n -20-on e (11). A
2 M solution of allylmagnesium chloride in THF (5 mL, 10
mmol) was diluted with 20 mL of THF and cooled to -75 °C.
A solution of 3a (1.00 g, 2.77 mmol) in 15 mL of dry THF was
added to the reaction mixture via syringe pump over 30 min.
The reaction mixture was stirred cold for 5 h and then allowed
to warm to room temperature. After stirring overnight, the
reaction mixture was added to ice-cold water and extracted
with EtOAc (3 × 50 mL). The pooled organic layers were
washed with a saturated NaCl solution, dried (Na2SO4),
filtered, and concentrated to dryness. The residue was dis-
solved in 7 mL of THF and 3 mL of acetone and treated with
1 mL of a 1 M HCl solution. After stirring overnight, the
reaction mixture was added to 5 mL of a saturated NaHCO3
solution and 15 mL of water. The mixture was extracted with
A mixture containing 28 (234 mg, 0.51 mmol), triphenyl-
prop-2-ynylstannane (795 mg, 2.04 mmol), azobis(isobuty-
roylnitrile) (18 mg, 0.11 mmol), and dry benzene (2 mL) was
degassed for ∼10 min. After 2 h at reflux, the solvent was
removed, and the residue was subjected to flash chromatog-
raphy, affording 14 as a white solid (20 mg, 11%): 1H NMR
(300 MHz, CDCl3) δ 5.14 (m, 1H), 4.78 (m, 2H), 2.53 (m, 1H),
2.11 (s, 3H), 0.74 (s, 3H), 0.60 (s, 3H). Anal. (C25H38O2‚3/8H2O)
C, H.
3r-H yd r oxy-3â-(m et h oxym et h yl)-5r-p r egn a n -20-on e
(15). To a solution of 4a (5.14 g, 15.6 mmol) in CH3OH (600
mL) was added Na metal (∼600 mg), and the resulting solution

3â-Substituted-3R-hydroxypregnan-20-ones
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 1 69
was heated at reflux for 16 h. After the reaction mixture had
come to room temperature, glacial acetic acid (3 mL) was added
dropwise. The solvent was then removed under reduced
pressure, and CH2Cl2 and water were added. The aqueous
layer was back-extracted with CH2Cl2. The combined organics
were washed with a saturated NaHCO3 solution, dried (Mg-
SO4), filtered, and evaporated under reduced pressure to give
a white solid (5.55 g), which was found to be a 84:16 mixture
was observed after 4 h at room temperature, so the reaction
mixture was heated at 35-50 °C for 3 days. CH2Cl2 and water
were added. The aqueous layer was back-extracted with CH2-
Cl2. The combined organic layers were washed with a satu-
rated NaHCO3 solution, dried (MgSO4), filtered, and concen-
trated to give a solid. Purification by flash column chromato-
graphy (100% CH2Cl2) afforded the azide 22 (834 mg, 73%) as
1
a white solid: mp 147-148 °C; H NMR (300 MHz, CDCl3) δ
1
of 15 and its 17-epimer on the basis of its H NMR spectrum.
3.23 (s, 2H), 2.53 (t, 1H, J ) 8.7 Hz), 2.12 (s, 3H), 0.76 (s, 3H),
Recrystallization from 1:1 hexane/acetone furnished 15 (3.27
g, 58%) as white crystals: mp 163-164 °C; 1H NMR (300 MHz,
CDCl3) δ 3.40 (s, 3H), 3.19 (s, 2H), 2.54 (t, 1H), 2.12 (s, 3H),
0.76 (s, 3H), 0.61 (s, 3H). Anal. (C23H38O3) C, H.
0.61 (s, 3H); IR 2095 (N3), 1689 cm-1 (CdO). Anal. (C22H35
-
N3O2) C, H, N.
3â-[(Dim eth yla m in o)m eth yl]-3r-h yd r oxy-5r-p r egn a n -
20-on e (23). A solution of 22 (330 mg, 0.883 mmol) in absolute
EtOH was flushed with Ar; 5% Pd/C (94 mg) was added, and
the mixture was flushed successively with Ar and hydrogen.
The reaction mixture was then stirred under an atmosphere
of hydrogen at room temperature for 5 h. The catalyst was
removed by filtration. The resulting filtrate was concentrated
under reduced pressure to an oil to which were added form-
aldeyde (37% in water; 2.0 mL, 72.1 mmol) and formic acid
(95-97%; 0.5 mL, 13.3 mmol). The above mixture, which
became a solution on refluxing, was heated with stirring
between 60 and 110 °C overnight. After the mixture had come
to room temperature, 50% NaOH (∼5 mL) was added drop-
wise. Subsequently ether and water were added, and the
resulting mixture was stirred until a solution formed. The
aqueous layer was back-extracted with ether. The combined
organics were washed successively with water and brine, dried
(MgSO4), filtered, and concentrated in vacuo to give a solid,
which was purified by flash column chromatography. Elution
with acetone gave 23 (120 mg, 36%) as a white solid: mp 126-
128 °C; 1H NMR (300 MHz, CDCl3) δ 2.53 (t, 1H), 2.38 (s, 6H),
2.25 (s, 2H), 2.12 (s, 3H), 0.75 (s, 3H), 0.61 (s, 3H). Anal.
(C22H35N3O2) C, H, N.
A similar procedure was used for the preparation of 3â-
(ethoxymethyl)-3R-hydroxy-5R-pregnan-20-one (16): mp 101-
102 °C; 1H NMR (300 MHz, CDCl3) δ 3.54 (q, 2H, J ) 7.0 Hz),
3.22 (s, 2H), 2.55 (t, 1H, J ) 8.9 Hz), 2.13 (s, 3H), 1.22 (t, 3H,
J ) 7.0 Hz), 0.77 (s, 3H), 0.62 (s, 3H). Anal. (C24H40O3) C, H.
3R-Hydroxy-3â-propoxymethyl-5R-pregnan-20-one (17): mp
1
96-98 °C; H NMR (300 MHz, CDCl3) δ 3.42 (t, 2H, J ) 6.5
Hz), 3.20 (s, 2H), 2.53 (t, 1H, J ) 8.7 Hz), 2.11 (s, 3H), 0.93 (t,
3H, J ) 6.5 Hz), 0.75 (s, 3H), 0.60 (s, 3H). Anal. (C25H42O3)
C, H. 3â-[(Benzyloxy)methyl]-3R-hydroxy-5R-pregnan-20-one
1
(18): mp 125-126 °C; H NMR (300 MHz, CDCl3) δ 7.33 (m,
5H), 4.57 (s, 2H), 3.28 (s, 2H), 2.53 (t, 1H), 2.12 (s, 3H), 0.76
(s, 3H), 0.62 (s, 3H). Anal. (C28H40O3) C, H.
3r-H yd r oxy-3â-(h yd r oxym et h yl)-5r-p r egn a n -20-on e
(19). Argon was passed over a solution of 18 (606 mg, 1.3
mmol) in absolute ethanol (100 mL). Palladium on activated
carbon (5%; 352 mg) was added in one portion, and hydrogen
gas (2-3 L) was maintained over the reaction mixture. After
60 h, the hydrogen was replaced with argon and the reaction
mixture was filtered. The filtrate was evaporated, and the
residue was purified by flash column chromatography (8:1 CH2-
Cl2/acetone) to give 19 as a white solid (440 mg, 91%): mp
183.5-185 °C; 1H NMR (300 MHz, CDCl3) δ 3.41 (s, 2H), 2.55
(t, 1H), 2.13 (s, 3H), 0.78 (s, 3H), 0.62 (s, 3H). Anal.
(C22H36O3‚0.25H2O) C, H.
3r-H yd r oxy-3â-(t r iflu or om et h yl)-5r-p r egn a n -20-on e
(24). To a solution of 3a (20.5 g, 56.9 mmol) in 170 mL of dry
THF was added 10 mL (9.62 g, 67.6 mmol) of neat TMSCF3.
The resulting solution was cooled in an ice/water bath, and
solid n-Bu4NF‚xH2O (115 mg, 0.44 mmol) was added in one
portion. The cold bath was removed, and the reaction mixture
was stirred at room temperature for 5 h. An additional 1 mL
(6.77 mmol) of TMSCF3 was added along with 36 mg (0.14
mmol) of solid n-Bu4NF‚xH2O. After stirring overnight, 35 mL
of a 3 N HCl solution was added. After 5 h, the two-phase
reaction mixture was separated and the aqueous layer was
extracted with EtOAc (3 × 50 mL). The pooled organic layers
were washed with a saturated NaHCO3 solution, dried over
MgSO4, filtered, and concentrated. Recrystallization from 20%
acetone/hexane gave 9.28 g of the undesired 3â-ol. A second
crop gave an additional 2.56 g of the 3â-ol. The mother liquor
from the second crop was concentrated to dryness and recrys-
tallized from hexane. The crystals that formed where sub-
jected to two flash columns (100% CH2Cl2). A total of 1.08 g
3r-H yd r oxy-3â-(t h iom et h oxym et h yl)-5R-p r egn a n -20-
on e (20). To a mixture of 4a (348 mg, 1.05 mmol) and dry
DMF (10 mL), which became a clear solution on heating, was
added NaSCH3 (165 mg, 2.33 mmol). The reaction mixture
was stirred at room temperature for 30 min, and 1 mL of CH3-
OH was added. Subsequently CH2Cl2 and water were added.
Brine was also added to overcome the emulsion formation. The
aqueous layer was back-extracted with CH2Cl2. The combined
organics were washed with water, dried (MgSO4), filtered, and
evaporated under reduced pressure to give a white solid (387
mg), which was found to be a mixture of 20 and its 17-epimer
1
on the basis of its H NMR spectrum. Flash column chroma-
tography (100:1 CH2Cl2/acetone) furnished 225 mg (57% yield)
1
of 20: mp 157-159 °C; H NMR (300 MHz, CDCl3) δ 2.61 (s,
2H), 2.54 (t, 1H), 2.20 (s, 3H), 2.13 (s, 3H), 0.76 (s, 3H), 0.62
(s, 3H). Anal. (C23H38SO2) C, H, S.
1
(5% yield) of the 3R-ol 24 was isolated: mp 187-189 °C; H
NMR (300 MHz, CDCl3) δ 2.53 (t, 1H), 2.12 (s, 3H), 0.79 (s,
3H), 0.61 (s, 3H); 19F NMR (CDCl3) -14.63 ppm (s). Anal.
(C22H33F3O2) C, H.
3â-(F lu or om eth yl)-3r-h yd r oxy-5r-p r egn a n -20-on e (25).
A mixture of n-Bu4NF‚xH2O (7.87 g) and benzene (50 mL) was
refluxed with azeotropic removal of water overnight. The
mixture, which was not a clear solution, was then concentrated
3â-(Cya n om eth yl)-3r-h yd r oxy-5r-p r egn a n -20-on e (21).
Potassium cyanide (210 mg, 3.2 mmol) and 5a (1.0 g, 2.7 mmol)
were stirred in ethanol (175 mL) for 48 h, refluxed for 5 h,
and then allowed to stir at room temperature overnight. Ether
(100 mL) and brine (100 mL) were added to the reaction
mixture. The separated aqueous phase was washed with
ether, and the combined organic phases were washed with
brine. Aqueous HCl (1 M, 6 mL) and THF (15 mL) were added
to the organic phases, and the solution was stirred overnight.
An aqueous NaHCO3 solution (20 mL) was added and then
water (50 mL). The separated aqueous phase was washed
with ether, and the combined organic phases were dried over
MgSO4, filtered, and evaporated. The white solid residue was
purified by flash column chromatography (40:20:1 CH2Cl2/
hexane/acetone) to give 21 as a white solid (493 mg, 52%): mp
218.5-220 °C; 1H NMR (300 MHz, CDCl3) δ 2.52 (t, 1H), 2.49
(s, 2H), 2.12 (s, 3H), 0.79 (s, 3H), 0.61 (s, 3H); IR 2213 (CN),
1693 cm-1 (CdO). Anal. (C23H35NO2) C, H, N.
to ∼10 mL and allowed to cool to room temperature.
A
solution of 5a (2.55 g, 6.81 mmol) in 20 mL of dry benzene
was then added. The reaction was refluxed and made more
concentrated. After the mixture had come to room tempera-
ture (giving a light-yellow solid), ether and water were added.
Since all the solid did not dissolve, CH2Cl2 was added. The
aqueous layer was back-extracted with CH2Cl2. The combined
organic extracts were washed twice with water, dried (Na2-
SO4), filtered, and evaporated under reduced pressure. The
residue was dissolved with heating in 105 mL of 20:1 acetone/
water containing 1 N HCl and allowed to stand at room
temperature for 30 min. The mixture became turbid, so CH2-
Cl2 was added to obtain a clear solution. The solvent was
removed under reduced pressure, giving a white solid to which
were added CH2Cl2 and water. The aqueous layer was back-
extracted with CH2Cl2. The combined organics were washed
3â-(Azid om eth yl)-3r-h yd r oxy-5r-p r egn a n -20-on e (22).
To a solution of 4a (1.01 g, 3.06 mmol) in dry DMF (50 mL)
was added solid NaN3 (226 mg, 3.48 mmol). Glacial acetic acid
(1.05 mL, 17.5 mmol) was added, and the resulting mixture
was warmed in order to give a clear solution. Little reaction

70 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 1
Hogenkamp et al.
with a saturated NaHCO3 solution, dried (MgSO4), filtered,
and evaporated under reduced pressure. Flash column chro-
matography (CH2Cl2) provided 25 (1.41 g, 59%): mp 201-203
160-161 °C; 1H NMR (300 MHz, CDCl3) δ 2.53 (t, 1H, J ) 8.7
Hz), 2.11 (s, 3H), 1.27 (s, 3H), 0.95 (s, 3H), 0.60 (s, 3H).
Syn th esis fr om Ep oxid e 5b. A solution of 5b (110 mg,
0.294 mmol) in 4 mL of THF at 0 °C was treated with solid
lithium aluminum hydride (28 mg, 0.74 mmol). The resulting
mixture was stirred at room temperature for 1 h and then
heated at reflux for 5 min. Once at room temperature, the
reaction mixture was cooled in an ice/water bath and diluted
with ether. Ice-cold water was added, and the organic layer
was decanted from the two-phase mixture. The aqueous phase
remaining was triturated with ether (3 × 15 mL). The
combined ether layers were washed with water and brine,
dried with Na2SO4, filtered, and concentrated. Column chro-
matography with 20% ethyl acetate/hexane gave 67 mg of
ketal-protected 30. Deketalization in 5 mL of THF and 1 mL
of a 1 M HCl solution and partitioning between a saturated
NaHCO3 solution and ethyl acetate gave 51 mg of crude
product. Column chromatography (30% ethyl acetate/hexane)
afforded 25 mg (25%) of 30 as a white solid: mp 157-158 °C.
1H NMR identical with 30 prepared above. Anal. (C22H36O2)
C, H.
3â-Eth yl-3r-h yd r oxy-5â-p r egn a n -20-on e (31). A solu-
tion of 33 (155 mg) was dissolved in MeOH (12 mL), Pd/C (5%,
12 mg) was added, and the mixture was hydrogenated at 200
kPa pressure overnight at room temperature. Filtration of the
catalyst followed by evaporation of the solvent yielded the
crude product, which was recrystallized from a hexane-
acetone mixture to give 31 (103 mg, 66%): mp 115-117 °C;
1H NMR (300 MHz, CDCl3) δ 2.51-2.72 (m, 1H), 2.12 (s, 3H),
0.74 (s, 3H), 0.60 (s, 3H). Anal. (C23H38O2) C, H.
1
°C; H NMR (300 MHz, CDCl3) δ 4.17 (d, 2H, J H,F ) 48 Hz),
2.54 (t, 1H, J ) 8.8 Hz), 2.13 (s, 3H), 0.77 (s, 3H), 0.61 (s, 3H).
Anal. (C22H35FO2) C, H.
3â-(Ch lor om eth yl)-3r-h yd r oxy-5r-p r egn a n -20-on e (26).
A mixture of 4a (1.00 g, 3.02 mmol), dry tetramethylammo-
nium chloride (696 mg, 6.35 mmol), glacial acetic acid (0.2 mL,
210 mg, 3.49 mmol), and 50 mL of DMF was heated at 90 °C.
After 2h, an additional 60 mL of DMF and 390 mg (3.56 mmol)
of tetramethylammonium chloride were added. After 75 min
at 100 °C, the reaction mixture was allowed to cool and added
to ice-cold water containing 25 mL of a saturated NaHCO3
solution. The resulting mixture was extracted with EtOAc (4
× 100 mL). The pooled organic layers were extracted with 15
mL of a saturated NaHCO3 solution and a saturated NaCl
solution. The solvent was dried (Na2SO4), filtered, and
concentrated. Column chromatography (100% CH2Cl2) and
recrystallization from EtOH afforded 495 mg (44%) of 26 as a
white solid: mp 214-215 °C; 1H NMR (300 MHz, CDCl3) δ
3.47 (s, 2H), 2.52 (t, 1H, J ) 8.9 Hz), 2.11 (s, 3H), 0.75 (s, 3H),
0.60 (s, 3H). Anal. (C22H35ClO2) C, H, Cl.
3â-(Br om om eth yl)-3r-h yd r oxy-5r-p r egn a n -20-on e (27).
A mixture of 4a (362 mg, 1.10 mmol), tetramethylammonium
bromide (322 mg, 2.16 mmol), acetic acid (1.0 mL, 17.5 mmol),
and dry DMF (30 mL) was heated at ∼110 °C. After 3 h the
solvent was removed by distillation under vacuum. To the
resulting solid were added CH2Cl2, water, and brine. The
organic layer was washed with saturated NaHCO3 solution,
dried (MgSO4), filtered, and evaporated under reduced pres-
sure to give a solid (380 mg). Flash chromatography (9:1
hexane/acetone) provided the bromide 27 (98 mg, 22%) as a
white solid: mp 196-197 °C dec; 1H NMR (300 MHz, CDCl3)
δ 3.42 (s, 2H), 2.52 (t, 1H), 2.12 (s, 3H), 0.76 (s, 3H), 0.63 (s,
3H). Anal. (C22H35BrO2) C, H, Br.
3â-Eth en yl-3r-h yd r oxy-5â-p r egn a n -20-on e (32). A solu-
tion of 3b (1.18 g, 3.3 mmol) in dry THF (20 mL) was treated
with a 1 M solution of vinyl magnesium bromide in THF (3.7
mmol, 3.7 mL) at -70 °C. After the mixture was stirred for 5
min and then at room temperature for 2.5 h, the reaction was
quenched with a saturated NH4Cl solution (10 mL). The
solvent was removed, and the residue was extracted with
EtOAc. The organic layer was washed with water, a dilute
NaHCO3 solution, water, and brine. After drying over MgSO4,
the solution was filtered and evaporated to yield the crude
product. This crude product was then dissolved in acetone (20
mL), and a 1 N HCl solution (10 mL) was added. After stirring
for 15 h, the solvents were removed and the residue was
extracted with CH2Cl2. The organic layer was washed with
water, a dilute NaHCO3 solution, water, and brine. After
drying (MgSO4), the solution was filtered and concentrated in
vacuo. Column chromatography (94:6 toluene/acetone) gave
189 mg (17%) of 32: mp 113-116 °C; 1H NMR (300 MHz,
CDCl3) δ 6.13 (dd, 1H, J ) 18.4, 11.0 Hz), 5.31 (d, 1H, J )
18.4 Hz), 5.17 (d, 1H, J ) 11.0 Hz), 2.52 (m, 1H), 2.11 (s, 3H),
0.93 (s, 3H), 0.59 (s, 3H). Anal. (C23H36O2‚0.25H2O) C, H.
3â-Eth yn yl-3r-h yd r oxy-5â-p r egn a n -20-on e (33). Using
the procedure given for the synthesis of the 5R-isomer 10, the
5â-isomer 33 was prepared in 45% yield: mp 196-197 °C; 1H
NMR (300 MHz, CDCl3) δ 2.54 (t, 1H), 2.51 (s, 1H), 2.12 (s,
3H), 0.97 (s, 3H), 0.60 (s, 3H). Anal. (C23H34O2) C, H.
3r-Hyd r oxy-3â-(p r op -2-yn yl)-5â-p r egn a n -20-on e (34).
To a mixture of Mg (268 mg, 11.0 mmol), propargyl bromide
(80 wt %; 1.5 mL, 10.1 mmol), and dry Et2O (15 mL) was added
HgCl2 (10 mg). The resulting mixture was warmed in a water
bath in order to initiate the reaction. The mixture became
turbid and was stirred at <20 °C until all Mg reacted (∼1 h).
A solution of 3b (725 mg, 2.01 mmol) in dry Et2O (20 mL) and
dry THF (10 mL) was cooled to -78 °C. Subsequently, the
propargylmagnesium bromide solution (∼0.6 M; 7 mL), pre-
pared above, was added dropwise at -78 °C. The resulting
mixture was stirred at -78 °C for ∼1 h. The mixture was
then allowed to warm to room temperature where it was
stirred for 1 h; 1 N HCl, p-toluenesulfonic acid monohydrate,
and acetone were added. The two-phase mixture was stirred
at room temperature overnight. Et2O and water were added.
The aqueous layer was back-extracted with Et2O. The com-
bined organics were washed with a saturated aqueous NaH-
CO3 solution and brine, dried (MgSO4), filtered, and concen-
trated. Flash-chromatography twice (10:1 hexane/acetone and
7:1 hexane/ethyl acetate) gave 34 as a white solid (208 mg,
3r-Hyd r oxy-3â-(iod om et h yl)-5r-p r egn a n -20-on e (28).
To a suspension of 4a (3.06 g, 9.26 mmol) and NaI (1.80 g,
12.0 mmol) in 50 mL of 1:1 MeOH/THF was added glacial
acetic acid (1 mL), and the reaction mixture was heated to
reflux. After 4h at reflux, an additional 0.5 mL of glacial acetic
acid was added. The reaction mixture was maintained at
reflux for an additional 2 h and then allowed to stir overnight
at room temperature. The reaction mixture was then concen-
trated in vacuo, and the residue was partitioned between CH2-
Cl2 and water. The aqueous layer was washed twice with
CH2Cl2, and the combined organic layers were extracted with
a saturated NaHCO3 solution, dried (MgSO4), filtered, and
concentrated. Recrystallization from 1:1 acetone/hexane gave
1
2.84 g (67%) of 28 as a white solid: mp 132-133 °C dec; H
NMR (300 MHz, CDCl3) δ 3.30 (s, 2H), 2.53 (t, 1H, J ) 8.8
Hz), 2.12 (s, 3H), 0.74 (s, 3H), 0.60 (s, 3H). Anal. (C22H35IO2)
C, H.
3â-(Ca r boxa m id om eth yl)-3r-h yd r oxy-5r-p r egn a n -20-
on e (29). Potassium carbonate (40 mg, 0.29 mmol) was added
to a solution of 21 (61 mg, 0.17 mmol) in H2O2 (1.5 mL, 30%)
and DMSO (6 mL) at 0-5 °C. The mixture was stirred at room
temperature for 6 h and poured into ice/water. The separated
solid was collected by filtration, washed with water, and dried.
Column chromatography (1:1 acetone/hexane) gave 29 as a
1
colorless solid (20 mg): mp 198-202 °C; H NMR (300 MHz,
CDCl3) δ 5.71-5.91 (br s, 1H), 5.31-5.51 (br s, 1H), 3.61-
3.77 (m, 1H), 2.41-2.62 (m, 1H), 0.74 (s, 3H), 0.70 (s, 3H).
Anal. (C23H37NO3) C, H, N.
3r-Hyd r oxy-3â-m eth yl-5â-p r egn a n -20-on e (30). Syn -
th esis fr om 3b. To a solution of 3b (1.00 g, 2.77 mmol) in 60
mL of dry THF at 0 °C was added 1.72 mL (5.16 mmol) of a 3
M solution of MeMgBr in ether. The reaction mixture was
allowed to warm to room temperature, and an additional 0.86
mL (2.58 mmol) of the Grignard reagent was added. After 1
h, the reaction was quenched with a saturated NH4Cl solution
and the mixture diluted with 50 mL of water and extracted
with EtOAc (3 × 50 mL). The pooled organic layers were dried
(MgSO4), filtered, and concentrated. Column chromatography
(5% acetone/hexane) gave 490 mg of a white solid. Removal
of the 20-ketal (acetone/p-TsOH/water) and recrystallization
from 3:1 hexane/acetone gave 350 mg (38% yield) of 30: mp
1
29%): mp 144-145.5 °C; H NMR (300 MHz, CDCl3) δ 2.53

3â-Substituted-3R-hydroxypregnan-20-ones
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 1 71
(t, 1H), 2.52 (s, 1H), 2.51 (s, 1H), 2.15 (s, 1H), 2.11 (s, 3H),
0.94 (s, 3H), 0.59 (s, 3H). Anal. (C24H36O2) C, H.
cRNAs encoding the human R1, â2, and γ2L subunits, ∼1 ng
of cRNA for each subunit per cell. Electrical recordings were
made using a two-electrode voltage clamp in frog Ringer
solution containing (in mM): NaCl, 115; KCl, 2; CaCl2, 1.8;
HEPES, 5; pH 7.4. Drug and wash solutions were applied
directly to the oocyte using a microcapillary “linear array”.24
Steroids were made up in DMSO over the range 0.001-10 mM.
Stocks were diluted 1000-3000-fold in Ringer, and unless
otherwise stated, DMSO was fixed at 0.3% by vol, a concentra-
tion at which the vehicle had little effect on GABA responses.24
Concentration-response data for GABA and steroids, and
concentration-modulation data for steroids, were fit to a four-
parameter logistic equation (Origin, Microcal Inc.) Data are
given as mean ( SEM, all values quoted to two significant
figures.
20,20-(Eth ylen ed ioxy)-3-m eth ylen e-5â-p r egn a n e (37).
A suspension of methyltriphenylphosphonium bromide (3.0 g,
8.40 mmol) in 10 mL of dry THF was treated with a 1 M
solution of KO-t-Bu in THF (8.4 mL, 8.4 mmol). The resulting
red-orange mixture was heated at 45 °C for 30 min. The
reaction mixture was allowed to cool to room temperature, and
3b (2.00 g, 5.55 mmol) was added as a solid in one portion.
After refluxing for 3 h, the reaction mixture was allowed to
cool to room temperature and then placed in an ice/water bath.
After 10 min, the reaction mixture was poured into an ether/
water mixture. The aqueous layer was extracted with ether
(2 × 25 mL), and the combined organic layers were washed
with a saturated NaCl solution, dried (Na2SO4), filtered, and
concentrated. This crude material was triturated with hexane
(2 × 100 mL). The hexane was evaporated in vacuo and the
residue (2.2 g) was subjected to column chromatography (1000
mL of 1.25% acetone/hexane and 500 mL of 1.75% acetone/
hexane) affording the alkene 37 as a white solid: mp 127-
128 °C, weight 1.83 g (91%); 1H NMR (300 MHz, CDCl3) δ 4.57
(s, 2H), 4.03-3.82 (m, 4H), 2.51 (t, 1H, J ) 13.4 Hz), 2.10-
0.97 (m, 22H), 1.29 (s, 3H), 0.93 (s, 3H), 0.75 (s, 3H).
3â-(Ch lor om eth yl)-3r-h yd r oxy-5â-p r egn a n -20-on e (35).
Reaction of 5b with tetramethylammonium chloride as de-
scribed for the synthesis of the 5R-isomer 26 afforded 35 in
90% yield after column chromatography (100% CH2Cl2): mp
183-185 °C. Recrystallization from 15% acetone/hexane gave
an analytically pure sample: 1H NMR (300 MHz, CDCl3) δ
3.73 (dd, 2H, J ) 11.3, 19.5 Hz), 2.53 (t, 1H, J ) 8.8 Hz), 2.11
(s, 3H), 0.95 (s, 3H), 0.59 (s, 3H). Anal. (C22H35ClO2) C, H,
Cl.
Ack n ow led gm en t. We thank Dr. Delia Belelli, Tom
Prins, David Fick, and J ie-Sheng Cheng for technical
assistance, Drs. J ohn Drewe and Wu Yang for the
preparation of GABAA receptor cRNA, and Dr. George
Field for invaluable discussions. The crystal structure
of 6 was determined by Dr. Marilyn Olmstead, Univer-
sity of California, Davis, CA.
Su p p or tin g In for m a tion Ava ila ble: Crystal structure
data for 6 (10 pages); tables of observed and calculated
structure factors (14 pages). Ordering information is given
on any current masthead page.
Refer en ces
(1) Gee, K. W.; McCauley, L.; 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, 207-277.
3r-H yd r oxy-3â-(m et h oxym et h yl)-5â-p r egn a n -20-on e
(36). Using the procedure described for the synthesis of the
5R-isomer, the 5â-isomer 36 was isolated as a white solid in
60% yield: mp 95.5-97 °C; 1H NMR (300 MHz, CDCl3) δ 3.39
(s, 3H), 3.38 (dd, 2H), 2.52 (t, 1H, J ) 8.7 Hz), 2.11 (s, 3H),
0.93 (s, 3H), 0.59 (s, 3H). Anal. (C23H38O3) C, H.
(2) (a) 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. J . Med. Chem. 1965, 8, 426-243. (b)
Gyermek, L.; Iriarte, J .; Crabbe, P. Steroids. CCCX. Structure-
activity relationship of some steroidal hypnotic agents. J . Med.
Chem. 1968, 11, 117-125. (c) Phillips, G. H. Structure-activity
relationships in steroidal anesthetics. J . Steroid Biochem. 1975,
6, 606-613. (d) Makriyannis, A.; DiMeglio, C. M.; Fesik, S. W.
Anesthetic Steroid Mobility in Model Membrane Preparations
As Examined by High-Resolution 1H and 2H NMR Spectroscopy.
J . Med. Chem. 1991, 34, 1700-1703 and references cited therein.
(3) Squires, R. F.; Casida, J . E.; Richardson, M.; Saederup, E. [35S]-
t-butylbicyclophosphorothionate binds with high affinity to brain
specific sites coupled to a γ-aminobutyric acid-A and ion recogni-
tion sites. Mol. Pharmacol. 1983, 23, 326-336.
(4) 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-Butylbicyclophosphorothionate and [3H]-
Flunitrazepam Binding and γ-Aminobutyric AcidA Receptor
Function. Mol. Pharmacol. 1994, 46, 977-985.
(5) 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.
(6) McCauley, L. D.; Liu, V.; Hawkinson, J . E.; Lan, N. C.; Gee, K.
W. Selective actions of certain neuroactive pregnanediols at the
GABAA receptor complex in rat brain. Mol. Pharmacol. 1995,
47, 354-362.
Ster oid In h ibition of TBP S Bin d in g. Rat brain cortical
membranes were prepared as previously described.4,5 Briefly,
cortices were rapidly removed following decapitation of carbon
dioxide-anesthetized Sprague-Dawley rats (200-250 g). The
cortices were homogenized in 10 vols 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 vol of buffer. Aliquots (100
µL) of the membrane suspensions were incubated with 2 nM
[35S]TBPS and 5 µL aliquots of test drug (nine concentrations
ranging from 1 nM to 10 µM final) 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 spectrom-
etry. Nonlinear curve fitting of the overall data for each drug
averaged for each concentration was done using Prism (Graph-
Pad). The data were fit to a partial instead of a full inhibition
model if the sum of squares was significantly lower by F-test.
Hill coefficients were allowed to vary from unity if the sum of
squares was significantly lower by F-test. The concentration
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 ( SEMs
of the individual experiments were calculated.
(7) Belelli, D.; Lambert, J . J .; Gee, K. W. Unpublished results.
(8) (a) Hogskilde, S.; Wagner, J .; Carl, P.; Anker, N.; Angelo, H. R.;
Sorensen, M. B. Anticonvulsant properties of pregnanolone
emulsion compared with althesin and thiopentone. Br. J . An-
aesth. 1988, 61, 462-467. (b) Belelli, D.; Bolger, M. B.; Gee, K.
W. Anticonvulsant profile of the progesterone metabolite 5R-
pregnan-3R-ol-20-one. Eur. J . Pharmacol. 1989, 166, 325-329.
(9) (a) Crawley, J . N.; Glowa, J . R.; Majewska, M. D.; Paul, S. M.
Anxiolytic activity of an endogenous adrenal steroid. Brain Res.
1986, 398, 382-385. (b) Bitran, D.; Hilvers, R. J .; Kellogg, C.
K. Anxiolytic effects of 3R-hydroxy-5R[â]-pregnan-20-one: en-
dogenous metabolites of progesterone that are active at the
GABAA receptor. Brain Res. 1991, 561, 157-161. (c) Wieland,
S.; Lan, N. C.; Mirasadeghi, S.; Gee, K. W. Anxiolytic activity of
the progesterone metabolite 5R-pregnan-3R-ol-20-one. Brain Res.
1991, 565, 263-268.
(10) (a) Gyermek, L. Pregnanolone:
a highly potent, naturally
occurring hypnotic-anesthetic agent. Proc. Soc. Exp. Biol. Med.
1967, 125, 1058-1062. (b) Mendelson, W. B.; Martin, J . V.;
Perlis, M.; Wagner, R.; Majewska, M. D.; Paul, S. M. Sleep
induction by an adrenal steroid in the rat. Psychopharmacology
1987, 93, 226-229.
GABAA Recep tor Exp r ession a n d Electr op h ysiology
in Xen opu s Oocytes. Preparation of oocytes and cRNAs
encoding GABAA receptor subunits were as described else-
where.30 Oocytes were microinjected with a 1:1:1 mixture of

72 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 1
Hogenkamp et al.
(11) (a) Craig, C. R. Anticonvulsant activity of steroids: separability
of anticonvulsant from hormonal effects. J . Pharmacol. Exp.
Ther. 1966, 153, 337-343. (b) Craig, C. R.; Deason, J . R.
Anticonvulsant activity of steroids, specificity of structure. Arch.
Int. Pharmacodyn. 1968, 172, 366-372. (c) Rupprecht, R.; Reul,
J . M. H. M.; Trapp, T.; van Steensel, B.; Wetzel, C.; Damm, K.;
Ziegelgaensberger, W.; Holsboer, F. Progesterone Receptor-
Mediated Effects of Neuroactive Steroids. Neuron 1993, 11, 523-
530.
(12) (a) Cook, M. C.; Lawrence, R.; Phillipps, G. H.; Hunter, A. C.;
Newall, C. E.; Stephenson, L.; Weir, N. G. Anaesthetic Steroids
of the Androstane and Pregnane Series. U.S. Patent 3,953,429,
April 27, 1976. (b) Phillipps, G. H.; Lawrence, R.; Newall, C. E.
Steroidal Anaesthetics of the Pregnane and 19-Norpregnane
Series. U.S. Patent 3,983,111, Sept. 28, 1976 (3â-substitution is
H and C1 to C6 alkyl) and reference 2c).
for Steroid Interaction with the γ-Aminobutyric AcidA Receptor
Complex. J . Pharmacol. Exp. Ther. 1987, 241, 346-353. Electro-
physiology: Cottrell, G. A.; Lambert, J . J .; Peters, J . A. Modula-
tion of GABAA receptor activity by alphaxalone. Br. J . Pharma-
col. 1987, 90, 491-500. Peters, J . A.; Kirkness, E. F.; Callachan,
H.; Lambert, J . J .; Turner, A. J . Modulation of the GABAA
receptor by depressant barbiturates and pregnane steroids. Br.
J . Pharmacol. 1988, 94, 1257-1269. Kokate, T. G.; Svensson,
B. E.; Rogawski, M. A. Anticonvulsant Activity of Neuroster-
oids: Correlation with γ-Aminobutyric Acid-Evoked Chloride
Current Potentiation. J . Pharmacol. Exp. Ther. 1994, 270, 1223-
1229.
(23) Wisden, W.; Laurie, D. J .; Monyer, H.; Seeburg, P. H. The
Distribution of 13 GABAA Receptor Subunit mRNAs in the Rat
Brain. I. Telencephalon, Diencephalon, Mesencephalon. J . Neu-
rosci. 1992, 12, 1040-1062.
(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.; Whitte-
more, R. M.; Woodward, R. M. 3R-Hydroxy-3â-trifluoromethyl-
5R-pregnan-20-one (Co 2-1970): A partial agonist at the neu-
roactive steroid site of the γ-aminobutyric acidA receptor. Mol.
Pharmacol. 1996, 49, 897-906.
(13) (a) Tahir, H.; Bolger, M.; Buswell, R.; Gabriel, R.; Stearns, J .
Method for Making 3R-Hydroxy-3â-substituted-pregnanes. U.S.
Patent 5,319,115, J une 7, 1994. (b) Cambie, R. C.; J urlina, J .
L.; Rutledge, P. S.; Woodgate, P. D. Iodo-azide Adducts of
Exocyclic Alkenes; Structure and Solvolysis. J . Chem. Soc.,
Perkin Trans. I 1982, 315-325.
(14) (a) Prakash, G. K. S.; Krishnamurti, R.; Olah, G. A. Fluoride-
Induced Trifluoromethylation of Carbonyl Compounds with
Trifluoromethyltrimethylsilane (TMS-CF3). A Trifluoromethide
Equivalent. J . Am. Chem. Soc. 1991, 111, 393-395. (b) Krish-
namurti, R.; Bellew, D. R.; Prakash, G. K. S. Preparation of
Trifluoromethyl and Other Perfluoroalkyl Compounds with
(Perfluoroalkyl)trimethylsilanes. J . Org. Chem. 1991, 56, 984-
989. (c) Ramaiah, P.; Krishnamurti, R.; Surya Prakarsh, G. K.
1-Trifluoromethyl-1-cyclohexanol. Org. Synth. 1993, 72, 232-
240. (d) Kruse, A.; Siegemund, G.; Schumann, A. Process of the
Preparation of Perfluoroalkyl Compounds, and Pentafluoroeth-
yltrimethylsilane. U.S. Patent 4,968,848, Nov. 6, 1990.
(15) Corey, E. J .; Chaykovsky, M. Dimethylsulfoxonium Methylide
and Dimethylsulfonium Methylide Formation and Application
to Organic Synthesis. J . Am. Chem. Soc. 1965, 87, 1353-1364.
(16) (a) Steroidal 5R-3-ketones give exclusively the R-epoxide: Leh-
mann, H. G.; Engelfried, O.; Wiechert, R. Spiro-3-oxiranyl-5R-
androstan-17â-ols. J . Med. Chem. 1965, 8, 383-384. (b) Cook,
C. E.; Corley, R. C.; Wall, M. E. Steroids. LXXIX. Synthesis and
Reactions of Oxiranes Obtained from 3- and 17-Keto Steroids.
J . Org. Chem. 1968, 33, 2789-2793.
(17) Alcaraz, L.; Harnett, J . J .; Mioskowski, C.; Martel, J . P.; Le Gall,
T.; Shin, D.-S.; Falck, J . R. Novel Conversion of Epoxides to One
Carbon Homologated Allylic Alcohols by Dimethylsulfonium
Methylide. Tetrahedron Lett. 1994, 35, 5449-5452.
(18) (a) Keck, G. E.; Yates, J . B. Carbon-Carbon Bond Formation via
the Reaction of Trialkylallylstannes with Organic Halides. J .
Am. Chem. Soc. 1982, 104, 5829-5831. (b) Pine, S. H.; Sanchez,
B. L. The Formic Acid-Formaldehyde Methylation of Amines.
J . Org. Chem. 1971, 36, 829-832. (c) Baldwin, J . E.; Adlington,
R. M.; Basak, A. Allene Transfer Reactions. A New Synthesis of
Terminal Allenes. J . Chem. Soc., Chem. Commun. 1984, 1284-
1285.
(19) (a) Carlson, R. G.; Behn, N. S. The Stereochemistry of Epoxi-
dation of Rigid Methylenecyclohexane Systems. Determination
of Stereochemistry of Exocyclic Epoxides in Rigid Cyclohexanes
by Nuclear Magnetic Resonance Half Band Widths. J . Org.
Chem. 1967, 32, 1363-1367. (b) Bellucci, G.; Chiappe, C.; Lo
Moro, G.; Ingrosso, G. Facial Stereoselectivity of Two-step
Additions Initiated by Electrophillic Halogens to Methylenecy-
clohexanes. A Comparison with Epoxidation. J . Org. Chem.
1995, 60, 6214-6217.
(20) Finn, D. A.; Gee, K. W. A comparison of Ro 16-6028 with
benzodiazepine receptor ‘full agonists’ on GABAA receptor func-
tion. Eur. J . Pharmacol.-Mol. Pharmcol. 1993, 247, 233-237.
(21) Hawkinson, 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,
169, 157-163.
(25) (a) Scherschlicht, R. The quinolizinone hypnotic Ro 41-3696, a
partial agonist at the flumazenil binding site, induces EEG sleep
in rabbits but lacks sedative side effects in rodents. Presented
at the Biennial IPEG Meeting, Berlin, Germany, Nov. 27-30,
1994. (b) J acobsen, E. J .; TenBrink, R. E.; Stelzer, L. S.; Belonga,
K. L.; Carter, D. B.; Im, H. K.; Im, W. B.; Sethy, V. H.; Tang, A.
H.; VonVoigtlander, P. F.; Petke, J . D. High-Affinity Partial
Agonist Imidazo[1,5-R]quinoxaline Amides, Carbamates, and
Ureas at the γ-Aminobutyric Acid A/Benzodiazepine Receptor
Complex. J . Med. Chem. 1996, 39, 158-175. (c) Tang, A. H.;
Smith, M. W.; Carter, D. B.; Im, W. B.; VonVoigtlander, P. F.
U-90042, A Sedative/Hypnotic Compound that Interacts Differ-
entially with the GABAA Receptor Subtypes. J . Pharmacol. Exp.
Ther. 1995, 275, 761-767.
(26) (a) Gee, K. W.; Bolger, M. B.; Tahir, H.; Wieland, S.; Wood, P.
L.; Lan, N. C. Epalons: The Antiepileptic Activity of Neuros-
teroids Active at the GABAA Receptor. Abstract presented at the
American Epilepsy Society Meeting, Neuropharmacology-Mech-
anisms, December 1995. (b) 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. Characterization of the Anticonvulsant Properties
of Ganaxolone (CCD 1042; 3R-hydroxy-3â-methyl-5R-pregnan-
20-one), a Selective High-Affinity Steroid Modulator of the
GABAA Receptor. J . Pharmacol. Exp. Ther., in press.
(27) Still, W. C.; Kahn, M.; Mitra, A. Rapid Chromatographic
Technique for Preparative Separations with Moderate Resolu-
tion. J . Org. Chem. 1978, 43, 2923-2925.
(28) (a) Nace, H. R.; Nelander, D. H. Norsteroids. V. The Application
of the Benzilic Acid Rearrangement to the Synthesis of A-
Norpregnanes. J . Org. Chem. 1964, 29, 1677-1681. (b) Clinton,
R. O.; Clarke, R. L.; Stonner, F. W.; Manson, A. J .; J ennings, K.
F.; Phillips, D. K. Steroidal Heterocycles. VI. Formylation of A/B-
cis 3-Ketosteroids. Preparation of 5â-Steroidal[3,2-c]pyrazoles.
J . Org. Chem. 1962, 27, 2800-2807.
(29) (a) Browne, P. A.; Kirk, D. N. A Study of the Henbest Reduc-
tion: the Preparation of 3R-Hydroxy-5R- and 3â-Hydroxy-5â-
steroids. J . Chem. Soc. C 1969, 1653-1659. (b) 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. (c) Varasi, M.; Walker, K. A. M.; Maddox, M. L. A
Revised Mechanism for the Mitsunobu Reaction. J . Org. Chem.
1987, 52, 4235-4238. (d) Schubert, K.; Schlegel, J .; Hoerhold,
C. Selektive Hydrierung von Steroidhormonen durch Clostridien
unter anaeroben Bedingungen. (Selective Hydrogenation of
Steroid Hormones by Clostridium under Anaerobic Conditions.)
Z. Naturforsch. B 1963, 18B, 284-286.
(22) Activity of 3â-ols: in vivo, see refs 2b,c and 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. In vitro [35S]-
TBPS, see: ref 5 and Harrison, N. L.; Majewska, M. D.;
Harrington, J . W.; Barker, J . L. Structure-Activity Relationships
(30) Whittemore, E. R.; Yang, W.; Drewe, J . A.; Woodward, R. M.
Functional and pharmacological characterization of the human
GABAA receptor R4-subunit expressed in Xenopus oocytes. Mol.
Pharmacol. 1996, 50, 1364-1375.
J M960021X