Synthesis and evaluation of analogues of the partial agonist 6- (propyloxy)-4-(methoxymethyl)-β-carboline-3-carboxylic acid ethyl ester (6- PBC) and the full agonist 6-(benzyloxy)-4-(methoxymethyl)-β-carboline-3- carboxylic acid ethyl ester (Zk 93423) at wild type and recombinant GABA(A) receptors

J . Med. Chem. 1998, 41, 2537-2552

2537

Syn th esis a n d Eva lu a tion of An a logu es of th e P a r tia l Agon ist

6-(P r op yloxy)-4-(m eth oxym eth yl)-â-ca r bolin e-3-ca r boxylic Acid Eth yl Ester

(6-P BC) a n d th e F u ll Agon ist 6-(Ben zyloxy)-4-(m eth oxym eth yl)-â-

ca r bolin e-3-ca r boxylic Acid Eth yl Ester (Zk 93423) a t Wild Typ e a n d

Recom bin a n t GABA_ARecep tor s

Eric D. Cox,^†Hernando Diaz-Arauzo,^†Qi Huang,^†Mundla S. Reddy,^†Chunrong Ma,^†Brad Harris,^‡

Ruth McKernan,^§Phil Skolnick,^‡and J ames M. Cook*^,†

Department of Chemistry, University of WisconsinsMilwaukee, Milwaukee, Wisconsin 53201, Merck Sharp & Dohme Research

Laboratories, Harlow, Essex CM20 2QR, UK, and Laboratory of Neuroscience, National Institute of Diabetes and Digestive and

Kidney Diseases, National Institutes of Health, Bethesda, Maryland 20892

Received J uly 15, 1997

A pharmacophore and an alignment rule have previously been reported for BzR agonist ligands.

The design and synthesis of 6-(propyloxy)-4-(methoxymethyl)-â-carboline-3-carboxylic acid ethyl

ester (6-PBC, 24, IC₅₀) 8.1 nM) was based on this pharmacophore. When evaluated in vivo

this ligand exhibited anticonvulsant/anxiolytic activity but was devoid of the muscle relaxant/

ataxic effects of “classical” 1,4-benzodiazepines (i.e., diazepam). Significantly, 6-PBC 24 also

reversed diazepam-induced muscle relaxation in mice. The 3-substituted analogues 40-46

and 48 of 6-PBC 24 and Zk 93423 27 (IC₅₀) 1 nM) were synthesized and evaluated in vitro to

determine what affect these modifications would have on the binding affinity at recombinant

BzR subtypes. With the exception of the 3-amino ligands 40 and 41, all the â-carbolines were

found to exhibit high binding affinity at BzR sites. The 3-propyl ether derivative 45 was also

evaluated in vivo and found to be devoid of any proconvulsant or anticonvulsant activity at

doses up to 40 mg/kg. The 6-(1-naphthylmethyloxy) and 6-octyloxy analogues 25, 26, 28, and

29 of 6-PBC 24 were synthesized to further evaluate the proposed alignment of agonists vs

inverse agonists in the pharmacophore of the BzR. In addition, ligands 26 and 29 were designed

to probe the dimensions of lipophilic pocket L₃at the agonist site. The activity of 29 was

evaluated in vivo; however, this analogue elicited no pharmacological effects at doses up to 80

mg/kg. These and other related â-carbolines were also examined in five recombinant GABA_A

receptor subtypes. Ligands 52-61 all exhibited moderate to high affinity at GABA_Areceptors

containing R₁subunits. These ligands will be useful in further defining the pharmacophore at

R₁â₃γ₂receptors.

In tr od u ction

subunits (R_1-6, â_1-4, γ_1-3, δ, F_1,2),^1,6,9,10and this molec-

ular diversity likely explains the broad spectrum of

pharmacological actions associated with ligands acting

at these ligand-gated ion channels. The identification

and biological evaluation of highly specific ligands at

GABA_Areceptor subtypes is critical in assigning specific

physiological and pharmacological roles to these sub-

units. Studies from many laboratories have centered

on the design and synthesis of biologically important

ligands which act at GABA_A/Bz receptor sites.^12-25

γ-Aminobutyric acid (GABA) is the major inhibitory

neurotransmitter in vertebrate brain.¹It is estimated

that almost 50% of all synapses in the brain employ this

neurotransmitter.²There are at least three distinct

classes of receptors which mediate the effects of GABA.^3,4

GABA_Areceptors are directly associated with a Cl^-ion

channel while GABA_Breceptors appear to be coupled

to Ca²⁺or K⁺channels through second messenger

systems.^2,3In addition, GABA_Creceptors are also

directly associated with a Cl^-ion channel, and recent

evidence suggests this receptor is structurally related

to GABA_Areceptors.⁵

The GABA_Areceptor contains binding sites for a

variety of ligands including barbiturates, steroids, ben-

zodiazepines, and “cage” convulsants.^4,6The GABA_A

receptor complex possesses significant molecular diver-

sity and is thought to be a pentameric structure as-

sembled from several classes of subunits.¹Molecular

cloning experiments have identified at least 15 different

To date, the composition of any single GABA_Areceptor

ion channel is unknown; however, in situ hybridization

and immunohistochemical studies have shown that

(R₁â_2/3γ₂) is the most abundant GABA_Areceptor subtype

in mammalian brain.⁵In addition, Pritchett et al.¹¹

have shown that complexes which contain the R₁

subunit exhibit high affinity for zolpidem 69, CL 218872

70, and some â-carbolines.⁴Evidence to date has

suggested that most if not all of the subunits investi-

gated do exist in the brain. Recently the binding affinity

of a series of ligands at recombinant R₅â₂/â₃γ₂receptor

subtypes has been shown to directly parallel their

affinities at native R₅subtypes isolated from hippoc-

^†University of WisconsinsMilwaukee.

^‡National Institutes of Health.

^§Merck Sharp & Dohme Research Laboratories.

Published on Web 06/16/1998

2538 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 14

Cox et al.

oxy)-4-(methoxymethyl)-â-carboline-3-carboxylic acid eth-

yl ester (6-PBC, 24) was conceived.^18,19This â-carboline

inhibited PTZ-induced seizures in a dose-dependent

fashion and exhibited anxiolytic activity when evaluated

in an elevated plus-maze paradigm. This partial ago-

nist 24 was devoid of muscle-relaxant activity and

completely antagonized the myorelaxant actions of

diazepam.^18,19A recent examination of the ability to

inhibit [³H]Ro15-1788 binding at five recombinant

GABA_Areceptor subtypes [R₁â₃γ₂(K_i) 0.49 nM), R₂â₃γ₂

(K_i) 1.21 nM), R₃â_3γ2(K_i) 2.20 nM), R₅â₃γ₂(K_i) 2.39

nM), R₆â₃γ₂(K_i) 1343 nM)] indicated 6-PBC possessed

high affinity and some selectivity for the R₁â₃γ₂subtype,

whereas the full agonist Zk 93423 27 [R₁â₃γ₂(K_i) 4.1

nM), R₂â₃γ₂(K_i) 4.2 nM), R₃â₃γ₂(K_i) 6.0 nM), R₅â₃γ₂

(K_i) 4.5 nM), R₆â₃γ₂(K_i) >1000 nM)] was less

selective (see Table 2). More recently, the synthesis and

biological activity of several R₅â₃γ₂selective ligands

(50-70-fold more selective at R₅/R₁subtypes) has been

reported from our laboratory.^12,20-23,25These ligands

will be employed to determine what role receptors which

contain the R₅subunit may play in the human brain

since these subtypes are found principally in the hip-

pocampus.^12,21In addition, a number of â-carbolines

have recently been found to exhibit some subtype

selectivity for the R₁â₃γ₂subunit.⁵

Examination of the biological activity of the partial

agonist 6-PBC 24 prompted further study of â-carbolines

via alteration of the substituent at C(3) on the â-car-

boline skeleton (as well as that of Zk 93423 27). This

study would help to establish what effect the C(3)

substituent would exert on the in vitro binding affinities

of these compounds at BzR subtypes. The â-carbolines

in Schemes 1, 2, and 3 were prepared for this study.

Furthermore, several related â-carbolines (25, 26, 28,

29) which contained substituents at position 6 which

varied (electronically and sterically) were synthesized

and evaluated. These modifications were carried out

both in the agonist (i.e., R₄) CH₂OCH₃) and the inverse

agonist (i.e., R₄) CH₂CH₃) series. In addition, the

receptor subtype selectivities of a number of other

â-carbolines were determined as well.

F igu r e 1. The alignment of the partial agonist 6-PBC (in

black) and the inverse agonist âCCE (in medium black) fitted

to a schematic representation of the inclusive pharmacophore

with CGS 9896 at the benzodiazepine binding site. Note, the

partial agonist 6-PBC is aligned in a vertical fashion, while

that of the inverse agonist âCCE is in a horizontal orientation.

The sites H₁and H₂designate hydrogen bond donor sites on

the receptor protein while A₂represents a hydrogen bond

acceptor site. Interaction with the hydrophobic pocket L₁, as

well as with H₁and A₂is required for potent inverse agonist

activity in vivo. Agonist activity requires interaction with H₁,

H₂, L₁, L₂, and/or L₃. Receptor descriptors S₁, S₂, and S₃are

regions of negative steric repulsion.

ampal tissue.¹²Thus, the use of recombinant receptor

subtypes provides an excellent system to investigate

ligands at the GABA_Areceptor complex.¹²

A comprehensive model of the pharmacophore for

agonists and inverse agonists at the benzodiazepine

receptor has been proposed.^{13,14,18,19,24,26}This model was

developed using the techniques of chemical synthesis,

radioligand binding, and receptor mapping. More than

150 different ligands encompassing 15 structurally

different classes of compounds were employed. This

pharmacophore qualitatively accounts for the relative

affinities, efficacies, and functional effects displayed by

various ligands at the BzR.^{12-14,16,18-27}Briefly, the

pharmacophore required two hydrogen bond donating

sites termed H₁and H₂. The two sites were located

about 6.7 Å from each other. The binding sites H₁and

L₁are common to both inverse agonist and agonist

pharmacophores; however, H₂corresponds to the 4-posi-

tion of â-carbolines. With respect to â-carbolines it was

believed the oxygen atom at this position was critical

for high-affinity binding by directing the ligand into the

active site via formation of a hydrogen bond between

the ether oxygen atom and H₂on the receptor protein.

There is also a lipophilic area with two zones important

for agonist activity termed L₂and L₃. Occupation of

areas L₂and/or L₃as well as interaction at H₁, H₂, and

L₁are important for agonist activity.^18,19,24In addition,

there are regions of repulsive steric interaction which

reduce ligand affinity for the receptor termed S₁, S₂, and

S₃(Figure 1).

Ch em istr y

A computer-based alignment of 6-PBC 24 in the

inclusive pharmacophore of the benzodiazepine receptor

site is illustrated in Figure 1 and has been previously

reported.^{12-14,16,18-25,27}Ligand overlap via computer

modeling illustrated the 6-benzyloxy function of Zk

93423 27 and the 5-phenyl ring of diazepam should fully

occupy lipophilic pocket L₃; however, the partial agonists

CGS 9895 and CGS 9896 and the antagonist 6-methoxy-

4-(methoxymethyl)-â-carbolinecarboxylic acid ethyl es-

ter⁶lacked a substituent which would interact in this

region. On the basis of this evidence, it was decided to

synthesize 6-PBC 24. The 6-n-propyloxy substituent of

24 should only partially occupy L₃and thus should

provide a ligand with a partial agonist profile.^18,19,27As

expected, the â-carboline 24 displayed high affinity at

the BzR (IC₅₀) 8.1 nM). More importantly, however,

24 exhibited anxiolytic/anticonvulsant activity when

evaluated in mice but was devoid of muscle relaxant

activity while the ligand antagonized the muscle relax-

ant/ataxic activity of diazepam.^7,8In addition, ligands

26 and 29 were prepared to further probe the dimen-

During the course of these studies the design and

synthesis of the anxiolytic/anticonvulsant 6-(n-propyl-

BzR Agonist Ligands

J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 14 2539

Ta ble 1. Potencies of Various â-Carboline Ligands To Inhibit [³H]Flunitrazepam Binding to Cortical Membranes

compd

R₃

R₄

R₆

R₉

IC₅₀(nM)^a

24

45

42

48

40

30

26

32

25

31

27

46

43

41

44

29

35

28

34

49

51

71

72

CO₂CH₂CH₃

OCH₂CH₂CH₃

NdCdS

CO₂CH(CH₃)₂

NH₂

CO₂CH₂CH₃

OCH₂CH₂CH₃

NdCdS

CH₂OCH₃

CH₂CH₃

CH₂CH₂CH₃

(CH₂)₇CH₃

CH₂Ph

CH₂-1-naphthyl

CH₂Ph

H (6-PBC)

H

CH₃

H

CH₃

H

CH₃

H (Zk-93423)

H

CH₃

H

CH₃

H

CH₃

H

CH₃

8.1 ( 1.5

43.2

24.7

30.8

697

290

330 ( 30

>1000

1

CH₂CH₃

CH₂OCH₃

CH₂CH₃

CH₂OCH₃

H

63.8

40

271.4

176

NH₂

NHC(S)OCH₃

CO₂CH₂CH₃

55.5 ( 4

730 ( 60

13.9 ( 1.9

>1000

22^b

>5000^b

1^b

CH₂Ph

OCH₃

CH₂Ph

>5000^b

a

Values for compounds are listed with statistical limits and represent an average of three or more experiments. Please see the

b

Experimental Section for further details. See ref 27.

sions of L₃. Ligands 25 and 28 lack an oxygen sub-

stituent at the 4-position and should therefore display

no affinity for the agonist site as compared to their

oxygenated counterparts 26 and 29. The N_a-methyl

analogues 30-35 were prepared as negative controls to

compare to their N_a-H analogues since the methyl

function is believed to undergo negative steric interac-

tions with S₁and should be devoid of activity if the

ligand alignment is correct (with respect to agonist

â-carboline ligands). The synthesis of these â-carboline

ligands was accomplished via the method of Neef et al.²⁸

and is described in detail for 6-PBC 24.

The 5-(benzyloxy)indole 1 was treated with am-

monium formate in the presence of a palladium catalyst

in ethanol¹⁰to provide 5-hydroxyindole 2 in 92% yield

(Scheme 1). Alkylation of the indole was accomplished

by treating hydroxyindole 2 with propyl iodide and

potassium carbonate in acetone at reflux to afford

5-(propyloxy)indole 3 in excellent yield. A Michael

reaction between indole 3 and 3-hydroxy-2-nitro-5-

oxahexanoic acid ethyl ester 37 in the presence of 10%

acetic acid and toluene at reflux provided indole 6 in

good yield. Raney nickel mediated reduction of the nitro

function provided the amino ethyl ester 12. Cyclization

of 12 was achieved via a modification of the Pictet-

Spengler reaction¹¹to provide the tetrahydro-â-carboline

18. Subsequent decarboxylation of 18 in xylenes at

reflux followed by oxidation using sulfur in DMSO at

140 °C resulted in the fully aromatic â-carboline 24.

Studies on â-carbolines with inverse agonist activity

had demonstrated that derivatives with 3-carboxylic

acid ethyl ester functions exhibit limited water solubility

and are short-lived in vivo due to esterase hydroly-

sis.^13,14On the basis of this, the increased water

solubility (as compared to âCCE) and the inverse

agonist activity of 3-ethoxy-â-carboline,²⁶the 3-carbox-

ylic acid ethyl ester function was converted into the 3-n-

propyloxy substituent.^13,14Consequently, the 6-propyl-

oxy and 6-benzyloxy analogues 45 and 46 (Scheme 2)

were prepared to evaluate water solubility²⁶and half-

life in vivo in comparison to the parent ligands 24 and

27. Furthermore, the binding affinity of these ligands

would provide further evidence to determine if the

interaction of the carbonyl group at C(3) with H₁was

necessary for high-affinity binding at BzR subtypes. The

3-alkoxy ether group has been proposed to release

electron density to the pyridine nitrogen N₍₂₎and would

presumably enhance interaction with hydrogen bond

donor descriptor H₁. A similar result was observed by

Allen et al.¹³in the case of inverse agonists.¹²Thus

6-PBC 24 was treated with 98% hydrazine in the

presence of ethanol at reflux to provide the carbohy-

drazide 38 via the method of Dodd et al.³¹The carbo-

hydrazide 38 was dissolved in concentrated aqueous

HCl at 0 °C followed by the addition of sodium nitrite.

Curtius rearrangement of the intermediate azido com-

pound (not shown) was accomplished in acetic acid/

water at reflux to yield 6-(propyloxy)-4-(methoxymethyl)-

3-amino-â-carboline 40 as the diacetate salt. Conversion

of the amino intermediate 40 into the 3-propyloxy

analogue 45 was attained via diazonium chemistry

using the procedure of Allen et al.¹²Treatment of amine

40 with anhydrous propanol in the presence of isoamyl

nitrite, CuSCN, and KSCN provided 3,6-bis(propyloxy)-

4-(methoxymethyl)-â-carboline 45.

The isothiocyanato derivatives 42 and 43 (Scheme 2)

were proposed as potential irreversible inhibitors. These

ligands could be employed with the 3-substituted de-

2540 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 14

Ta ble 2. R_xâ₃γ₂(K_iValues in nM)^a

Cox et al.

a

K_ivalues represent the mean of two determinations which differed by less than 10%. Data were generated using Ltk^-cell membranes

expressing human R_xâ₃γ₂receptors. [³H]Ro15-1788 and [³H]Ro15-4513 (for cells expressing R₆â₃γ₂receptors) were used as radioligands

at a final concentration of 1-2 nM.

rivatives from the inverse agonist series to help char-

acterize and understand the interaction of ligands at

the GABA_A/Bz receptor at the molecular level.^13,14The

3-isothiocyanato ligands 42 and 43 were prepared by

treatment of the appropriate amino derivative (40 or

41) in chloroform with thiophosgene to provide the

isothiocyanates 42 and 43 in excellent yield.

When the 3-carboxylic acid ethyl ester function of Zk

93423 27 was replaced by a 3-carboxylic acid isopropyl

ester (abecarnil 50, Table 2), the pharmacological profile

of this ligand was altered from a full agonist (see 27) to

a partial agonist (see 30).^32,33To investigate the role

of the isopropyl ester on the agonist activity of â-car-

bolines related to 6-PBC 24, the 3-ethyl ester of 6-PBC

24 was replaced by a 3-isopropyl ester function. Hence,

6-PBC 24 was hydrolyzed in the presence of 10%

aqueous NaOH at reflux followed by treatment with

dilute aqueous HCl to afford the carboxylic acid 47.

Subsequent esterification of acid 47 with 2-propanol in

the presence of anhydrous HCl provided 6-(propyloxy)-

4-(methoxymethyl)-â-carboline 3-carboxylic acid isopro-

pyl ester 48 in 85% yield (Scheme 3).

Resu lts a n d Discu ssion

The potencies of various 3-substituted analogues of

the parent ligands 6-PBC 24 and Zk 93423 27 to inhibit

[³H]flunitrazepam to rat cortical membranes are sum-

marized in Table 1. Replacement of the ethyl ester

function of 24 or 27 with a propyl ether [45 (IC₅₀) 43.2

nM) and 46 (IC₅₀) 63.8 nM)], respectively, significantly

reduced affinity compared to the parent ligands. How-

ever, when evaluated in vivo the more potent analogue

45 exhibited no anticonvulsant activity at doses up to

40 mg/kg against a standard challenge of pentylenetet-

razole. Furthermore, this ligand exhibited no procon-

vulsant activity at doses of 20 and 40 mg/kg in the same

paradigm. The affinities of these two ligands at Bz

binding sites are in agreement with previous studies on

BzR Agonist Ligands

J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 14 2541

Sch em e 1

the effects of substitution at position 3 of â-carbo-

lines.^13,14Initially, it was believed that an ester moiety

at the 3-position was required for a compound to exhibit

high-affinity binding at Bz binding sites.^34-37However,

the high affinity of the partial inverse agonist 3-ethoxy-

â-carboline 54 (Table 3) demonstrated this was not the

case. Subsequently, further studies have suggested at

least two factors affect high-affinity binding with respect

to 3-alkoxy-substituted â-carbolines, one of which is the

lipophilicity of the substituent which interacts at L₁.^13,14

The second factor is the ability of the 3-substituent to

release electron density to the pyridine ring, enhancing

the basicity of the nitrogen atom which results in

greater ligand-receptor interaction at H₁. The lack of

agonist and inverse agonist activity of 3-propyloxy

ligands 45 and 46 in vivo may be a direct result of the

length of the alkyl substituent at position 3. For

example, Allen et al.^13,14demonstrated that 3-ethoxy-

â-carboline 54 displayed inverse agonist properties,

while 3-(propyloxy)-â-carboline 55 elicited an antagonist

response when evaluated in the pentylenetetrazole

paradigm in mice.^12,13Furthermore, the syn conforma-

tion of the carbonyl group and nitrogen function N(2)

in â-carboline-3-carboxylate alkyl esters would enhance

interaction at H₁in vivo, and this interaction is absent

in 45 and 46 which would contribute to the loss of

activity in vivo as well. A direct comparison cannot be

made to the work of Allen et al.¹³since the substitution

pattern is complicated by the presence of the groups at

position 4.

The synthesis of 3-isothiocyanato-â-carboline (IC₅₀)

8 nM) by Allen et al.¹³resulted in a ligand which bound

irreversibly to the Bz binding site. As discussed above,

the ability of the 3-substituent to stabilize the ligand-

receptor hydrogen bond H₁at N(2) is important for high

affinity binding at the Bz binding site. This irreversible

inhibitor provided strong evidence that such an interac-

tion was indeed necessary for high-binding affinity. As

expected, the isothiocyanate derivatives 42 (IC₅₀) 24.7

nM) and 43 (IC₅₀) 40 nM) exhibited moderate affinity

for the Bz binding site. This result further supports the

work of Allen et al. in regard to interactions of ligands

at H₁.^13,14

Examination by molecular modeling combined with

SAR studies have indicated at least three important

hydrophobic receptor regions (L₁, L₂, and L₃) are con-

tained within the Bz binding sites.²⁴The interactions

between ligands and these regions also contribute to the

efficacy of these compounds. For example, studies on

3-substituted â-carbolines have shown a transition from

a methyl ester to a n-propyl ester resulted in an inverse

agonist profile in the former to an antagonist profile in

the latter compound.¹³In addition, the presence of a

bulky substituent such as a tert-butyl ester (âCCt)

provided an antagonist with good Bz₁selectivity.¹⁶

Furthermore, when the 3-ethyl ester moiety of Zk 93423

27 was replaced with an isopropyl ester (abecarnil 50)

the biological profile was altered from that of a full

agonist to a partial agonist. In a complementary

manner, replacement of the 3-ethyl ester of 6-PBC 24

2542 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 14

Cox et al.

Sch em e 2

Sch em e 3

with an isopropyl ester moiety 48 (IC₅₀) 30.8 nM)

provided a ligand with moderate affinity in vitro;

however, agonist activity was no longer observed in vivo

in this series.

The inverse agonist and agonist inclusive pharma-

cophore illustrated in Figure 1 has been developed to

correlate essential pharmacophoric descriptors to bind-

ing affinity in vitro as well as biological activity in vivo.

Evidence from molecular modeling suggests that inverse

agonists (A₂, H₁, and L₁) and agonists (H₁, H₂, L₁, L₂,

and/or L₃) bind to the same domain of the BzR.^24,38

Although H₁and L₁are common to both models, the

remaining points of receptor interaction are clearly

different. With respect to the agonist pharmacophore,

it is believed that an oxygen atom corresponding to the

4-position of the â-carboline nucleus (i.e., the 4-meth-

oxymethyl substituent) will align the ligand into the

active site on the protein complex via generation of a

hydrogen bond between the molecule and a donor site

on the protein (H₂). As stated earlier, the ligand must

form hydrogen bonds with H₁(the pyridine nitrogen of

the â-carboline) and H₂as well as interact with L₁, L₂,

and/or L₃in order to elicit an agonist response with

respect to â-carbolines. The phenyl group of the ben-

zyloxy substituent of Zk 93423 27 is believed to lie in

the lipophilic pocket L₃. It is important to note that

substituents which occupy L₃cannot lie in the same

plane as H₁, H₂, and L₁since they would interfere with

the hydrogen-bonding protein-ligand interactions nec-

As expected, the 3-amino analogues 40 (IC₅₀) 697

nM) and 41 (IC₅₀) 271 nM) exhibited reduced affinities

when compared to their parent ligands (24 and 27). This

reduction is believed to be a result of two unfavorable

interactions.^12,13First, the lack of sufficient lipophilicity

at C-3 may lead to unfavorable ligand-receptor interac-

tions. Second, Allen et al.^13,14have shown the amino

function to exist as the iminopyridine tautomer which

results in diminution of the necessary interaction

between the pyridine nitrogen atom N(2) on the ligand

with the hydrogen bond acceptor portion (H₁) of the

receptor site.

The N_a-methyl congeners 30, 32, 33, and 35 (IC₅₀

values of 290, >1000, 985, 730 nM, respectively) were

synthesized and evaluated as negative controls. As

expected, these ligands exhibited a significant decrease

in affinity to the BzR as compared to their N_a-H

counterparts. It is believed the methyl substituents of

compounds 30 and 35 are interacting with S₁(see

Figure 1), leading to an unfavorable steric interaction.

Ligands 32 and 33 also interact with S₁.^19,24

BzR Agonist Ligands

J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 14 2543

Ta ble 3. R_xâ₃γ₂(K_iValues in nM)^a

a

K_ivalues represent the mean of two determinations which differed by less than 10%. Data were generated using Ltk^-cell membranes

expressing human R_xâ₃γ₂receptors. [³H]Ro15-1788 and [³H]Ro15-4513 (for cells expressing R₆â₃γ₂receptors) were used as radioligands

at a final concentration of 1-2 nM.

essary for high-affinity binding. Thus, the alignment

of the â-carboline nucleus ring A (containing the 6-ben-

zyloxy substituent) is directed toward lipophilic regions

L₁and L₂(see Figure 1).

ligands 26 and 29 were used to further probe the

dimensions of lipophilic pocket L₃.

The 4-ethyl congener 49 of Zk 93423 27 inhibits [³H]-

Flu binding to rat cortical membranes with an IC₅₀

value of 22 nM,²⁷an approximately 20-fold reduction

Examination of Figure 1 illustrates an alignment in

the pharmacophore for inverse agonist â-carboline

ligands that differs from that of the agonist pharma-

cophore. These â-carbolines are devoid of an oxygen

atom at position 4 to interact with the H₂donor site.

Rather, an alignment for inverse agonists, as reported

earlier,^13,14,24is obtained in which the indole N(9)-H

function interacts with A₂on the receptor complex.^13,14,24

This alignment directs the â-carboline ring system

parallel to H₁and A₂rather than perpendicular to this

line. This orientation no longer permits an interaction

with lipophilic pocket L₃, resulting in an inverse agonist

profile. To further study this alignment, several â-car-

bolines 25, 26, 28-35 were synthesized in both the

agonist (R₄) CH₂OCH₃) and inverse agonist series (R₄

) CH₂CH₃) and evaluated in vitro. These alterations

permitted a change in electronic character without

changing the ligand topology at C(4). In addition,

in affinity as compared to the parent ligand 27 (IC₅₀

)

1 nM). This ligand lacked the anticonvulsant activity

associated with agonist ligands. The analogous substi-

tution carried out in the 6-octyloxy series resulted in a

similar outcome. The â-carboline ligand 26 bound with

an IC₅₀of 330 nM, while the 4-ethyl analogue 25

exhibited very low affinity (IC₅₀) >1000 nM). While

further studies are required, the high-moderate affini-

ties of Zk 93423 27, 6-PBC 24, and 29 combined with

the poor affinity of 26 suggests the lipophilic pocket L₃

will not tolerate long-chain alkyl substituents.

The 6-(1-naphthylmethyloxy)-4-methoxymethyl de-

rivative 29 of Zk 93423 27 inhibited [³H]flunitrazepam

binding with an IC₅₀value of 55.5 nM. Although

considerably less potent than the parent ligand 27, this

2544 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 14

Cox et al.

â-carboline exhibited a positive GABA shift of 1.5 which

is consistent with a partial agonist profile of activity.

However, when evaluated in vivo, 29 elicited no con-

vulsant, proconvulsant, or anticonvulsant activity at

doses up to 80 mg/kg. The decrease in potency in vitro

of 29 when compared to Zk 93423 27 has resulted in a

ligand with no detectable in vivo activity at the doses

studied. Examination of the affinities at recombinant

receptor subtypes illustrated in Table 2 indicated ligand

29 is nearly 20-fold more selective for the R₁â₃γ₃site

compared to the other subtypes evaluated. The full

agonist Zk 93423 27 [R₁â₃γ₂(K_i) 4.1 nM), R₂â₃γ₂(K_i)

4.2 nM), R₃â₃γ₂(K_i) 6.0 nM), R₅â₃γ₂(K_i) 4.5 nM),

R₆â₃γ₂(K_i) >1000 nM)] exhibited no remarkable

subtype selectivity for diazepam sensitive (DS) GABA_A

receptor isoforms while the partial agonist 6-PBC 24

[R₁â₃γ₂(K_i) 0.49 nM), R₂â₃γ₂(K_i) 1.21 nM), R₃â₃γ₂

(K_i) 2.20 nM), R₅â₃γ₂(K_i) 2.39 nM), R₆â₃γ₂(K_i) 1343

nM)] displayed moderate R₁â₃γ₃selectivity. These data,

coupled with a GABA shift of 1.5, suggest 29 may

exhibit a partial agonist profile. The lack of biological

activity may therefore be due to the inability of this

ligand to penetrate the blood/brain barrier. Further

studies are needed to fully address this issue.

4-ethyl congener (horizontal/inverse agonist alignment)

is devoid of an oxygen atom at this position, thus

negating a vertical alignment. Furthermore, ligand 49

exhibited a very similar selectivity for R₁-containing

receptors as the partial agonist 6-PBC 24 when com-

pared to affinities at the R₂â₃γ₂, R₃â₃γ₂, and R₅â₃γ₂

receptor subtypes. If these compounds bound in the

same orientation with very similar receptor subtype

selectivity, one would have expected similar biological

activity.

Further support for the proposed alignment of â-car-

boline ligands at the Bz binding site is illustrated in

Table 3. The 3-substituted â-carboline ligands âCCE

53 and 3-EBC 54 as well as 3-PBC 55 all exhibit

moderate to good selectivity for the R₁subtype. The

partial agonist 6-PBC 24 exhibited moderate R₁â₃γ₂

selectivity compared to the three other DS subtypes

evaluated. The selectivity of these ligands for GABA_A

receptors containing the R₁subunit is similar in vitro;

however, there is a large difference among them when

evaluated in vivo. The 3-substituted ligands âCCE 50

and 3-EBC 52 elicit either proconvulsant or convulsant

activity while 6-PBC 24 exhibits anticonvulsant/anxi-

olytic actions in rodents. Furthermore 24 antagonized

the ataxic/muscle relaxant activity of diazepam.^18,19The

similar affinities for the R₁subunit of these ligands

when compared to a very different biological profile for

these compounds (see 24, 50, and 52) further supports

an alignment for inverse agonists â-carbolines different

from that of agonist â-carbolines. Recently, 12 different

3,4,6-trisubstituted â-carboline ligands with a 4-meth-

oxymethyl substituent were evaluated using computer

modeling techniques.³⁹When these ligands were evalu-

ated using a vertical alignment, a cross-validated r²

value of 0.5 (in a CoMFA analysis, see 6-PBC) was

obtained. These results, when taken together with the

affinities at recombinant GABA_Areceptors, further

support the proposed alignment for 4-methoxymethyl-

substituted â-carboline ligands at the BzR; nonetheless,

further studies are required to address this issue.

Several studies using GABA_Areceptor subtypes have

been carried out with 3-substituted â-carboline lig-

ands,^13-15some of which are illustrated in Table 3 (see

below). The potent inverse agonist âCCE 53 bound to

rat synaptosomal membranes with an IC₅₀value of 5

nM. Examination of data from these studies have

shown that in the 3-alkoxy series there is a steady

increase in potency in vitro as chain length is increased

from a 3-methoxy moiety (not shown) (IC₅₀) 124 nM),

to 3-ethoxy 54 (IC₅₀) 24 nM), to a 3-n-propyloxy group

55 (IC₅₀) 11 nM). When the chain length is further

increased to provide the n-butyloxy analogue 56 (IC₅₀

) 98 nM), the potency is decreased by almost an order

of magnitude. Furthermore, side chains with â or δ

branching (ligands 58, 59, and 60 with IC₅₀values of

93, 500, and 471 nM, respectively) also resulted in a

significant decrease in affinity. As expected, the sub-

type binding affinities at the R₁â₃γ₂subtype mimics the

in vitro binding affinities to rat cortical membranes

since the R₁â_2/3γ₂subtype is the most abundant receptor

subtype in rodent cortex.¹For example, 3-EBC 54 and

3-PBC 55, as well as ethers 56, 58, and 60, exhibit some

R₁subtype selectivity. In addition, âCCt 52 is more

selective for R₁subtypes than âCCE 53, indicating that

interactions in the pharmacophore at region L₁have

Replacement of the 4-methoxymethyl moiety of 29

with an ethyl substituent 28 resulted in a 4-fold increase

in affinity (IC₅₀) 13.9 nM). This is in agreement with

the proposed alignment rule at the Bz site;²⁴thus if the

6-substituted benzyloxy moiety of Zk 93423 27 fully

occupies lipophilic pocket L₃as has been previously

proposed,²⁴then the naphthyl analogue 29 should

exhibit a significant reduction in affinity at the Bz

binding site. Furthermore, the 4-ethyl analogue 28

should exhibit potent binding affinity. Ligands which

possess a 4-ethyl substituent are devoid of agonist

activity; therefore the ligand was not screened in vivo.

As expected, the N_a-methyl analogues 31, 32, 33, and

34 exhibited drastically decreased affinities in vitro in

agreement with the previous alignment of agonist and

inverse agonist â-carbolines.^{13,14,18,19,24,27}

The affinities of selected â-carboline ligands at re-

combinant GABA_Areceptors are illustrated in Tables

2-5. The proposed alignment for agonist â-carboline

ligands is thought to be in a vertical orientation whereas

that for inverse agonists is believed to be in a horizontal

fashion with respect to A₂and H₁(see Figure 1).²⁴The

data from receptor subtype selectivity studies further

supports this hypothesis. Several issues must be ad-

dressed when comparing the data for Zk 93423 27 to

that for the 4-ethyl analogue 49, the first of which stems

from the difference between the activity of these two

ligands in vivo. The â-carboline Zk 93423 27 elicited

anticonvulsant and anxiolytic as well as ataxic and

muscle relaxant activity, while ligand 49 was devoid of

any anticonvulsant activity when evaluated against a

standard challenge of pentylenetetrazole.²⁷The only

difference between these two ligands is the substituents

at the 4-position. The 4-methoxymethyl moiety of 27

is thought to be interacting at the pharmacophoric

descriptor H₂, resulting in the formation of a hydrogen

bond interaction between the protein complex and the

ligand. This interaction is believed to direct the agonist

ligand into the agonist binding site, resulting in a

vertical orientation of the â-carboline ligand. The

BzR Agonist Ligands

J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 14 2545

Ta ble 4. R_xâ₃γ₂(K_iValues in nM)^a

a

K_ivalues represent the mean of two determinations which differed by less than 10%. Data were generated using Ltk^-cell membranes

expressing human R_xâ₃γ₂receptors. [³H]Ro15-1788 and [³H]Ro15-4513 (for cells expressing R₆â₃γ₂receptors) were used as radioligands

at a final concentration of 1-2 nM.

played a role in R₁selectivity. This property may be

exploited in the future to design â-carbolines even more

selective for R₁isoforms.

subtypes. These imidazobenzodiazepines may be useful

in discerning the underlying mechanisms associated

with memory and learning since the R₅subtype is

located primarily in the hippocampus.¹²These results

combined with computer modeling of these congeners

may provide a detailed analysis of this receptor subtype.

As illustrated in Tables 2-5, many of the â-carboline

ligands shown here exhibit moderate to high selectivity

for the R₁â₃γ₂receptor subtype. The receptor subtype

selectivity of these ligands, when combined with SAR

data as well as computer modeling techniques, may

provide detailed information regarding the topography

of the R₁-containing receptor subtypes.

In conclusion, the in vitro evaluation of ligands 24,

26-29, 40-46, and 48 has shown these â-carbolines all

bind with high affinity to the Bz binding sites on GABA_A

receptors. The receptor subtype selectivity data il-

lustrated in Tables 2-5 has provided further evidence

that the proposed alignment for inverse agonist â-car-

boline ligands is different than that of agonist â-carbo-

lines. Analysis of the recombinant receptor subtype

selectivity data for ligands 52-61 demonstrated that

these ligands all exhibit moderate to high affinity for

the R₁subtype. The data illustrated in Tables 2-5,

combined with SAR studies and computer modeling

analysis, may provide detailed information describing

the pharmacophore for receptor subtypes containing the

R₁subunit. In addition, the receptor subtype selectivity

data has provided further evidence that the diazepam

insensitive (DI) site R₆â₃γ₂is devoid of the lipophilic

The potent inverse agonist DMCM 62 and its 3-car-

bethoxy analogue 63 are illustrated in Table 4. Both

ligands exhibit high affinity for DS receptor subtypes.

In addition, both â-carbolines display moderate affinity

for the R₆â₃γ₂(DI) subtype. In an effort to create a

longer-lived, more water soluble analogue of 62, the

3-methoxy 64 and 3-ethoxy 65 DMCM analogues were

synthesized and evaluated. These modifications re-

sulted in significant reductions in ligand affinity. The

orientation of the carbonyl group in ligands 60 and 61

has been proposed^12-15to align in a syn conformation

which results in the formation of a three-centered

hydrogen bond with pharmacophoric descriptor H₁.

This interaction leads to increased binding affinity. The

3-alkoxy analogues are unable to generate this interac-

tion which may account for the decreased affinity at the

recombinant receptor subtypes.

The negative controls 32 and 65 are shown in Table

4. Methylation of the indole N_a-H moiety rendered both

ligands inactive at the five major receptor subtypes

studied. These data are in agreement with previously

published results.^13-15,24

Recently, Liu et al. have synthesized and evaluated

imidazobenzodiazepines at the five recombinant GABA_A

receptor subtypes illustrated here.^12,20-23Several of the

ligands examined exhibited potent R₅â₃γ₂subtype se-

lectivity (∼65-70-fold) compared to the other receptor

2546 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 14

Ta ble 5. R_xâ₃γ₂(K_iValues in nM)^a

Cox et al.

a

K_ivalues represent the mean of two determinations which differed by less than 10%. Data were generated using Ltk^-cell membranes

expressing human R_xâ₃γ₂receptors. [³H]Ro15-1788 and [³H]Ro15-4513 (for cells expressing R₆â₃γ₂receptors) were used as radioligands

at a final concentration of 1-2 nM.

liquid scintillation counting. K_ivalues were calculated using

the least-squares iterative fitting routine of RS/l analysis

software (BBN Research System, Cambridge, MA) and are the

means of two determinations which differed by less than 10%.

pocket L₃found in the diazepam sensitive (DS) sites

(R₁â₃γ₂, R₂â₃γ₂, R₃â₃γ₂, R₅â₃γ₂).²⁴

Exp er im en ta l Section

Ra d ioliga n d Bin d in g in Ra t Br a in Mem br a n es. The

potencies of test compounds to displace [³H]flunitrazepam from

benzodiazepine binding sites were determined through a

modification of previously described procedures.²⁷In brief, rats

were killed by decapitation, and the cerebral cortex was

removed. Tissues were disrupted in 100 volumes of Tris-

citrate buffer (50 mM, pH 7.4) using a Brinkman polytron (15

s, setting 6). Tissues were centrifuged for 20 min (4 °C) at

20000g. The supernatant was discarded and the tissue pellet

resuspended in an equal volume of buffer. This “washing”

procedure was repeated three times. Tissues were either used

fresh or stored at -70 °C until used. Incubations (0.5 mL)

consisted of tissue suspension (0.1 mL, ∼0.1 mg of protein),

0.05 mL of NaCl solution (2.5M), 0.05 mL of [³H]flunitrazepam

(final concentration, ∼1 nM), and drugs and/or buffer to

volume. Nonspecific binding was determined using Ro15-1788

(final concentration 10 µM). Incubations (0-4 °C) were

initiated by addition of radioligand and terminated after 60

min by rapid filtration under vacuum through GF/B filters

Ra d ioliga n d Bin d in g to Recom bin a n t GABA_ARecep -

tor s. In brief, the affinity of compounds at GABA_Asubtypes

was measured by competition with [³H]Ro15-1788 (83 Ci/mmol;

NEN) to Ltk^-cells expressing human GABA_Areceptors

composed of R₁â₃γ₂, R₂â₃γ₂, R₃â₃γ₂, R₅â₃γ₂, and R₆â₃γ₂.^41,42Cells

were removed from culture by scraping into phosphate-

buffered saline, centrifuged at 3000g and resuspended in 10

mL of phosphate buffer (10 mM KH₂PO₄, 100 mM KCl, pH

7.4 at 4 °C) for each tray (25 cm²) of cells. Radioligand binding

assays were carried out in a volume of 500 µL which contained

100 mL of cells, [³H]Ro15-1788 at a concentration of 1-2 nM

and test compound in the range 10^-9-10^-5M. Nonspecific

binding was defined by 10^-5M diazepam and typically

represented less than 5% of the total binding. For cells

expressing R₆â₃γ₂, [³H]Ro15-4513 was used as radioligand.

Assays were incubated to equilibrium for 1 h at 4 °C and

harvested onto GF/B filters (Brandel) by filtration using a

Tomtec cell harvester and washing with ice-cold assay buffer.

After drying, filter-retained radioactivity was detected by

BzR Agonist Ligands

J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 14 2547

with two 5-mL washes of ice-cold buffer. IC₅₀values were

estimated using InPlot 4.0 (GraphPAD, San Diego, CA) with

at least six concentrations of inhibitor. Values represent X (

SEM of at least three determinations. Compounds with

potencies >1000 nM were generally only tested twice.

5-(Octyloxy)in d ole (4). 5-Hydroxyindole 2 (5.1 g, 0.038

mol) was added to a solution containing acetone (100 mL) and

1-iodooctane (37.2 g, 0.155 mol). To the stirred solution was

added potassium carbonate (26.9 g, 0.195 mol), at which time

the mixture was brought to reflux under an atmosphere of N₂.

After 48 h, analysis by TLC indicated the absence of starting

material. The potassium carbonate was removed by filtration

and the acetone removed under reduced pressure. The residue

was taken up in CH₂Cl₂(350 mL) and washed with water. The

aqueous layer was extracted (3 × 150 mL) with CH₂Cl₂. The

combined organic layers were washed with brine and dried

(K₂CO₃). Flash chromatography (silica gel, CH₂Cl₂/hexanes)

provided 5-(octyloxy)indole 4 (7.43 g) in 80% yield. 4: mp 95

In Vivo. Adult male NIH/Swiss mice (∼30 g) were injected

intraperitoneally (ip) with graded doses of the compounds (0.1

mL, diluted Emulphor/saline, 1:9) or an equal volume of vehicle

(0.1 mL, diluted Emulphor/saline, 1:9). Groups of 5-10 mice

were injected in graded doses and 10 min later were suspended

by their forepaws on a 1.5-mm-thick wire 60 cm above the

benchtop to assess muscle relaxation; three falls in <1 min

was considered positive for muscle relaxation. At 15 min

postinjection, consistent with previous work in this paradigm,¹²

mice were injected with PTZ (80 mg/kg) to assess anticonvul-

sant activity, or 40 mg/kg to assess proconvulsant activity.

Tonic and clonic convulsions with loss of righting reflex were

considered positive for seizure activity; mild myoclonic jerks

and straub tail were not counted as seizures.^12,27

Ma ter ia ls. Melting points were taken on a Thomas-Hoover

melting point apparatus or an Electrothermal Model IA8100

digital melting point apparatus and are reported uncorrected.

Proton NMR spectra were recorded on a Bruker 250-MHz

multiple-probe instrument or a GE 500-MHz spectrometer.

Infrared spectra were recorded on a Nicolet DX FTIR BX V5.07

spectrometer or a Mattson Polaris IR-10400 instrument. Low-

resolution mass spectral data (EI/CI) were obtained on a

Hewlett-Packard 5985 B GC-mass spectrometer, while high-

resolution mass spectral data were obtained on a Finnigan HR

1

°C; IR (KBr) 3256 cm^-1; H NMR (CDCl₃) δ 0.89 (t, 3H, J )

6.5 Hz), 1.32 (m, 12 H), 1.81 (m, 2H), 4.00 (t, 2H, J ) 6.5 Hz),

6.47 (t, 1H, J ) 2.0 Hz), 6.66 (dd, 1H, J ) 2.4, 8.7 Hz), 7.13

(m, 2H), 7.29 (m, 1H), 8.03 (s, 1H, br); MS (CI) m/e 245 (M⁺

1, 100), 162 (12), 133 (48). Anal. (C₁₆H₂₃NO) C, H, N.

+

5-(1-Na p h th ylm eth yloxy)in d ole (5). Into a 500 mL

round-bottom flask were added 5-hydroxyindole 2 (9.6 g, 0.072

mol), 1-(chloromethyl)naphthalene (60 mL, 70.8 g, 0.4 mol),

potassium carbonate (45 g, 4.5 equiv), and acetone (75 mL).

The reaction mixture was stirred at reflux under an atmo-

sphere of N₂for 18 h at which time analysis by TLC indicated

the absence of starting material. The potassium carbonate

was filtered off and the acetone removed under reduced

pressure. The excess 1-(chloromethyl)naphthalene was re-

moved using a wash column (silica gel, hexanes). Flash

chromatography (silica gel, CH₂Cl₂, hexanes, 1:1) provided the

5-methoxynaphthylindole 5 (15.5 g) in 78% yield. 5: mp 104-

mass spectrometer. Microanalyses were performed on

a

1

105 °C; IR (KBr) 3416 cm^-1; H NMR (CDCl₃) δ 5.54 (s, 2H),

Perkin-Elmer 240C carbon, hydrogen, and nitrogen analyzer.

Analytical TLC plates employed were E. Merck Brinkman UV

active silica gel (Kieselgel 60 F254) on plastic, and silica gel

60b for flash chromatography was purchased from E. M.

Laboratories. All chemicals were purchased from Aldrich

Chemical Co. unless otherwise stated. All reactions were

carried out under an atmosphere of nitrogen, and all solvents

were dried according to literature procedures.

The synthesis of ligands 49,²⁷50,²⁷51,¹⁶52,¹⁶53,¹³54,¹³55,¹³

56,¹⁴57,¹⁵58,¹⁵59,¹⁵60,¹⁴61,¹⁵62,¹⁷63,¹⁷64,¹⁷65,¹⁷66,²⁷67,²⁷

and 71²⁷have been previously reported.

6.51 (s, 1H), 7.02 (dd, 1H, J ) 2.3, 8.9 Hz), 7.20 (m, 1H), 7.31

(m, 2H), 7.53 (m, 3H), 7.87 (m, 2H), 8.15 (m, 2H); MS (CI) m/e

274 (M⁺+ 1, 35), 141 (100). Anal. (C₁₉H₁₅NO) C, H, N.

3-[5-(1-Na p h th ylm eth yloxy)in d ol-3-yl]-2-n itr o-4-(m eth -

oxym eth yl)bu ta n oic Acid Eth yl Ester (11) (P r oced u r e

a ). Into a 250 mL round-bottom flask equipped with a reflux

condenser was added indole 5 (5.1 g, 18.6 mmol), hydroxy nitro

ethyl ester 37 (10.75 g), toluene (125 mL), and glacial acetic

acid (12 mL). The solution was heated to reflux under an

atmosphere of N₂. After 1.25 h, analysis by TLC (silica gel,

EtOAc/hexanes, 1:1) indicated the absence of starting material.

The toluene was removed under reduced pressure, and the

excess acetic acid was removed azeotropically through several

distillations using toluene under reduced pressure. The brown

oil was chromatographed (silica gel, EtOAc/hexanes, 1:1) to

provide 11 as a mixture of diastereomers (8.4 g, 97%). 11: IR

5-Hyd r oxyin d ole (2). To a solution of 5-benzyloxyindole

1 (10.0 g, 0.044 mol) in dry EtOH (600 mL) were added

ammonium formate (5.9 g, 0.094 mol) and 10% Pd/C (1.0 g),

and the solution was stirred at room temperature under an

atmosphere of N₂until analysis by TLC indicated the complete

consumption of starting material. The catalyst was removed

by vacuum filtration over a bed of Celite. The Celite was

rinsed with hot EtOH (500 mL). The organics were combined

and removed under reduced pressure. The residue was taken

up in CH₂Cl₂and washed with water. The aqueous layer was

then washed with CH₂Cl₂(3 × 100 mL). The combined organic

layers were then washed with brine and dried (K₂CO₃), and

the solvent was removed under reduced pressure to provide

5-hydroxyindole 2 (5.4 g) in 90% yield. 2: mp 107 °C (lit.⁴⁰

1

(neat) 3423, 2987, 2930, 1750 (CdO) cm^-1; H NMR (DMSO-

d₆) δ 1.28 (t, 1H, J ) 7.0 Hz), 1.71 (m, 2H), 3.72 (m, 2H), 3.86

(s, 3H), 6.06 (s, 2H), 6.48 (t, 1H, J ) 10.2 Hz), 7.35 (d, 1H, J

) 8.5 Hz), 7.74 (s, 1H), 7.76 (s, 1H), 7.86 (s, 1H), 8.06 (m, 3H),

8.23 (d, 1H, J ) 7.0 Hz), 8.46 (m, 3H), 8.66 (d, 1H, J ) 6.7

Hz), 11.45 (br s, 1H); MS (EI) m/e 462 (M⁺, 1.2), 141 (100).

This material was employed without further purification in a

later step.

3-[5-(1-Na p h th ylm eth yloxy)in d ol-3-yl]-2-n itr op en ta n o-

ic Acid Eth yl Ester (10). A solution of indole 5 (7.39 g, 0.027

mol), hydroxy nitro ester 36 (10.35 g) in toluene (180 mL), and

glacial acetic acid (18 mL) was treated as described in

procedure a. Flash chromatography (silica gel, CH₂Cl₂/hex-

anes, 1:1) provided 10 as a mixture of diastereomers (7.25 g)

in 60% yield. 10: IR (neat) 3420, 2982, 1748 (CdO) cm^-1; ¹H

NMR (CDCl₃) δ 1.26 (t, 1H, J ) 7 Hz), 1.75 (m, 2H), 3.72 (m,

2H), 5.96 (s, 2H), 6.48 (m, 1H), 7.32 (d, 1H, J ) 8.1 Hz), 7.71

(s, 1H), 7.77 (s, 1H), 7.88 (s, 1H), 8.06 (m, 3H), 8.20 (d, 1H, J

) 6.9 Hz), 8.46 (m, 3H), 8.63 (d, 1H, J ) 6.7 Hz); MS (EI) m/e

446 (M⁺, 3), 141 (100). This material was employed without

further purification in a later step.

1

mp 107-108 °C); IR (KBr) 3402, 3226, 1623 cm^-1; H NMR

(CDCl₃) δ 6.22 (d, 1H), 6.61 (dd, 1H, J ) 3.7 and 10 Hz), 6.86

(d, 1H J ) 2 Hz), 7.19 (dd, 1H, J ) 3.7 and 10 Hz), 7.20 (s,

1H), 8.61 (s, 1H), 10.79 (s, 1H); MS (CI) m/e 134 (M⁺+ 1, 100).

5-(P r op yloxy)in d ole (3). To a slurry of 5-hydroxyindole

2 (29 g, 0.218 mol) and solid K₂CO₃(5 equiv, 1.1 mol, 150 g)

in acetone (600 mL) was added 1-iodopropane (5 equiv, 1.1

mol, 185.3 g, 106.3 mL). The resulting solution was heated

to reflux under an atmosphere of N₂and monitored by TLC

(silica gel, 9:1 CHCl₃:CH₃OH). The reaction was cooled to room

temperature, and the solids were removed by filtration. The

solvent was removed under reduced pressure to provide

5-(propyloxy)indole 3 as a dark brown oil (38 g) in 98% yield.

1

3: IR (neat) 3411, 1474, 1454 cm^-1; H NMR (CDCl₃) δ 1.04

3-(5-(Octyloxy)in d ol-3-yl)-2-n itr o-4-(m eth oxym eth yl)-

bu ta n oic Acid Eth yl Ester (8). Into a solution of toluene

(175 mL) and glacial acetic acid (17 mL) were added 5-(octy-

loxy)indole 4 (6.4 g, 0.026 mol) and hydroxy nitro ester 37 (16.3

g). The solution was treated as described in procedure a.

Flash chromatography (silica gel, EtOAc/hexanes, 25:75)

(t, 3H, J ) 7.4 Hz), 1.82 (sextet, 2H, J ) 6.6, 7.4 Hz), 3.95 (t,

2H, J ) 6.6 Hz), 6.45 (s, 1H), 6.86 (dd, 1H, J ) 2, 10 Hz), 7.12

(d, 4H, J ) 8.3 Hz), 7.25 (d, 1H, J ) 8.3 Hz), 8.5 (s, 1H); MS

(CI) m/e 176 (M⁺+ 1, 100), 134 (35); HRMS m/e 175.0997

(C₁₁H₁₃NO requires 175.0991).

2548 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 14

Cox et al.

provided 8 as a mixture of diastereomers (4.85 g) in 43% yield.

8: IR (film) 3374, 2931, 1736 (CdO) cm^-1; MS (CI) m/e 435

(M⁺+ 1, 35), 302 (100). This material was employed without

further purification in a later step.

(s, 1H), 7.23 (m, 1H), 7.97 (s, 1H, br); MS (EI) m/e 404 (M⁺, 3),

302 (100), 190 (36), 146 (38) 102 (30). This material was

employed in a later step.

3-(5-(Octyloxy)in d ol-3-yl)-2-a m in op en ta n oic Acid Eth -

yl Ester (13). The nitro ethyl ester 7 (6.25 g, 15 mmol) was

treated as described in procedure b to provide 13 as a mixture

of diastereomers (4.52 g) in 80% yield. 13: IR (neat) 3381,

2931, 2860, 1736 (CdO) cm^-1; ¹H NMR (CDCl₃) δ 0.89 (m, 8H),

1.15 (t, 3H, J ) 7.1 Hz), 1.29 (m, 8H), 1.81 (m, 5H), 3.24 (m,

1H), 3.76 (t, 1H, J ) 6.6 Hz), 4.04 (m, 5H), 6.85 (m, 1H), 7.06

3-(5-(Octyloxy)in d ol-3-yl)-2-n itr op en ta n oic Acid Eth yl

Ester (7). 5-(Octyloxy)indole 4 (7.12 g, 29 mmol) was treated

as described in procedure a to provide 7 as a mixture of

diastereomers (6.56 g) in 55% yield. 7: IR (neat) 3430, 2938,

2860, 1757 (CdO) cm^-1; ¹H NMR (CDCl₃) δ 0.86 (m, 8H), 1.30

(m, 11H), 1.50 (m, 2H), 1.82 (m, 2H), 3.87 (t, 2H, J ) 7.0 Hz),

4.01 (t, 2H, J ) 6.5 Hz), 4.29 (q, 1H, J ) 7.1 Hz), 5.45 (m,

1H), 6.85 (m, 1H), 7.05 (m, 1H), 7.24 (m, 1H), 7.99 (s, 1H, br);

MS (EI) m/e 418 (M⁺, 80), 298 (60), 286 (51), 186 (100). This

material was employed without further purification in a later

step.

3-(5-(P r op yloxy)in d ol-3-yl)-2-n itr o-4-(m eth oxym eth yl)-

bu ta n oic Acid Eth yl Ester 6. A solution composed of

5-(propyloxy)indole 3 (38.0 g, 0.22 mol), hydroxy nitro ester

36 (45.6 g), toluene (400 mL), and glacial acetic acid (38 mL)

was treated as described in procedure a. A wash column (silica

gel, EtOAc) furnished the ester 6 as a mixture of diastereomers

(73.1 g) in 95% yield. HRMS (Finnigan HR) m/e 364.1634

(C₁₈H₂₆N₂O₆requires 364.1625). This material was employed

without further purification in a later step.

3-(5-(Ben zyloxy)in d ol-3-yl)-2-n itr o-4-(m eth oxym eth yl)-

bu ta n oic Acid Eth yl Ester 9. A solution composed of

5-(benzyloxy)indole 1 (200 g, 0.22 mol), hydroxy nitro ester 37

(40.0 g), toluene (300 mL), and glacial acetic acid (50 mL) was

treated as described in procedure a. A wash column (silica

gel, EtOAc) furnished the ester 9 as a mixture of diastereomers

(30 g) in 82% yield. 9: IR (KBr) 1748 (CdO) cm^-1; MS (CI)

m/e 413 (M⁺+ 1, 100). This material was employed without

further purification in a later step.

3-[5-(1-Naph th ylm eth yloxy)in dol-3-yl]-2-am in o-4-(m eth -

oxym eth yl)bu ta n oic Acid Eth yl Ester (17) (P r oced u r e

b). Into a 500 mL round-bottom flask were added nitro ethyl

ester 11 (8.1 g, 17 mmol), ethanol (250 mL), and activated

Raney nickel (7 g). The air in the flask was removed under

vacuum and an atmosphere of hydrogen placed over the

solution. After 36 h, analysis by TLC indicated the absence

of starting material. The Raney nickel was removed by

filtration over a bed of Celite, and the ethanol was removed

under reduced pressure. A wash column (silica gel, EtOAc/

hexanes, 1:1) provided the amino compound 17 as a mixture

of diastereomers (4.85 g) in 67% yield. 17: IR (neat) 3382,

2984, 1733 (CdO) cm^-1; ¹H NMR (CDCl₃) δ 1.14 (m, 3H), 3.32

(s, 2H), 3.39 (s, 1H), 3.69-3.79 (m, 4H), 3.96 (s, 1H), 4.03-

4.14 (m, 3H), 5.53 (s, 2H), 6.95 (dd, 1H J ) 2.3, 4.7 Hz), 7.12

(d, 1H, J ) 2.5 Hz), 7.30 (m, 2H), 7.49 (m, 3H), 7.63 (d, 1H, J

) 6.9 Hz), 7.66 (m, 2H), 8.10 (d, 1H, J ) 6.9 Hz), 8.12 (m, 1H);

MS (EI) m/e 432 (M⁺, 10), 332 (40) 141 (100). This material

was employed without further purification in a later step.

(m, 1H), 7.25 (m, 1H), 7.94 (s, 1H, br); MS (CI) m/e 388 (M⁺

1, 100), 372 (42), 286 (48). This material was employed in a

later step.

+

3-(5-(P r op yloxy)in d ol-3-yl)-2-a m in o-4-(m eth oxym eth -

yl)bu ta n oic Acid Eth yl Ester 12. The nitro ethyl ester 6

(7.0 g, 0.021 mol) was treated as described in procedure b to

provide 12 as a mixture of diastereomers (5.0 g) in 79% yield.

12: ¹H NMR (CDCl₃) δ 1.04 (t, 3H, J ) 7.5 Hz), 1.79 (sextet,

2H, J ) 7 Hz), 3.15 (s, 3H), 3.90 (t, 2H, J ) 7 Hz), 6.7 (s, 1H),

7.2 (d, 1H, J ) 8 Hz), 7.27 (d, 1H, J ) 8 Hz), 9.48 (s, 1H); MS

(CI) m/e 335 (M⁺+ 1, 100); HRMS (Finnigan HR) m/e 334.1892

(C₁₈H₂₄N₂O₆requires 334.1889). This material was employed

without further purification in a later step.

3-(5-(Ben zyloxy)in d ol-3-yl)-2-a m in o-4-(m eth oxym eth -

yl)bu ta n oic Acid Eth yl Ester 15. The mixture of nitro ethyl

ester 9 (4.0 g, 9.7 mmol) and Raney nickel (8 g) in ethanol

(400 mL) was treated as described in procedure b to provide

15 as a mixture of diastereomers (3.1 g) in 83% yield. 15: IR

(neat) 3339, 2987, 2938, 1742 cm^-1; MS (CI) m/e 413 (M⁺+ 1,

100). This material was employed in a later step without

further characterization.

6-(1-Na p h th ylm eth yloxy)-4-(m eth oxym eth yl)-3-(eth ox-

yca r b on yl)-1,2,3,4-t et r a h yd r o-â-ca r b olin e-1-ca r b oxylic

Acid (23) (P r oced u r e c). Into a 250 mL round-bottom flask

was added aminoindole 17 (5.07 g, 0.012 mol) in ethyl acetate

(29 mL). A solution of glyoxylic acid monohydrate (1.03 g,

0.014 mol) in water (14 mL) was added dropwise to the

solution. The pH was adjusted to 5 with 10% aqueous

potassium carbonate, and the mixture was allowed to stir for

16 h at room temperature. The aqueous and organic phases

were separated, and the aqueous phase was extracted (2 × 25

mL) with ethyl acetate. The organic phases were combined,

and the solvent was removed under reduced pressure to yield

the carboxylic acid 23 as a dark orange oil which solidified

upon standing (5.72 g) in 92% yield. 23: mp >300 °C; IR (KBr)

1

3360-3200 br, 2924, 1736 (CdO) cm^-1; H NMR (DMSO-d₆)

δ 1.11 (t, 3H, J ) 7 Hz), 3.26 (s, 3H), 3.52 (m, 4H), 4.01 (t, 2H,

J ) 6.7 Hz), 4.08 (s, 1H), 4.94 (s, 1H), 5.52 (s, 2H), 6.75 (dd,

1H, J ) 2.3, 7.2 Hz), 7.23 (m, 2H), 7.54 (m, 3H), 7.68 (d, 1H,

J ) 7.2 Hz), 7.92 (m, 2H), 8.15 (d, 1H, J ) 7.2 Hz), 10.6 (s,

1H); MS (EI) m/e 488 (M⁺, 0), 141 (100). This material was

employed in a later step.

3-[5-(1-Naph th ylm eth yloxy)in dol-3-yl]-2-am in open tan o-

ic Acid Eth yl Ester (16). Into a 250 mL round-bottom flask

were placed nitro ethyl ester 10 (7.02 g, 0.015 mol), ethanol

(200 mL), and activated Raney nickel (6.2 g), and the resulting

mixture was treated as described in procedure b. A wash

column (silica gel, EtOAc/hexanes, 1:1) provided the amino

compound 16 (4.1 g) in 66% yield as a mixture of diastereo-

6-(1-Na p h th ylm eth yloxy)-4-eth yl-3-(eth oxyca r bon yl)-

1,2,3,4-tetr a h yd r o-â-ca r bolin e-1-ca r boxylic Acid (22). A

solution of glyoxylic acid monohydrate (1.078 g, 0.117 mol) in

H₂O (12 mL) was added dropwise to a vigorously stirred

solution of amino ethyl ester 16 (3.96 g, 0.0095 mol) in ethyl

acetate (25 mL). The reaction was treated as described in

procedure c to provide the carboxylic acid 22 (3.2 g) as a

mixture of diastereomers in 80% yield. 22: mp >300 °C; IR

1

mers. 16: IR (neat) 3382, 2983, 1740 (CdO) cm^-1; H NMR

1

(CDCl₃) δ 0.93 (m, 3H), 1.20 (m, 5H), 1.85 (m, 1H), 3.72 (t,

2H, J ) 7.1 Hz), 4.07 (m, 2H), 5.51 (s, 2H), 6.96 (dd, 1H, J )

2.3, 7.0 Hz), 7.06 (m, 1H), 7.27 (m, 2H), 7.51 (m, 3H), 7.62 (d,

1H, J ) 6.9 Hz), 7.87 (m, 2H), 8.13 (m, 1H); MS (EI) m/e 416

(M⁺, 2), 314 (100), 141 (78). This material was employed in a

later step.

(KBr) 3402, 2980, 2938, 1743 (CdO) cm^-1; H NMR (DMSO-

d₆) δ 0.87 (m, 3H), 1.05 (t, 3H, J ) 7.2 Hz), 1.71 (m, 2H), 4.01

(m, 3H), 4.22 (q, 2H, J ) 7.0 Hz), 4.89 (s, 1H), 5.51 (s, 2H),

6.78 (dd, 1H, J ) 2.3, 8.8 Hz), 7.22 (m, 2H), 7.53 (m, 3H), 7.66

(d, 1H, J ) 7.0 Hz), 7.93 (m, 2H), 8.14 (d, 1H, J ) 8.7 Hz),

10.6 (s, 1H); MS (EI) m/e 444 (M⁺, 0), 141 (100). This material

was employed in a later step.

3-(5-(Octyloxy)in d ol-3-yl)-2-a m in o-4-(m eth oxym eth yl)-

bu t a n oic Acid E t h yl E st er (14). The nitro ethyl ester 8

(4.85 g, 0.012 mol) was treated as described in procedure b to

provide 14 as a light brown oil (3.52 g) in 75% yield. 14: IR

6-(Oct yloxy)-4-(m et h oxym et h yl)-3-(et h oxyca r b on yl)-

1,2,3,4-tetr a h yd r o-â-ca r bolin e-1-ca r boxylic Acid (20). The

amino ethyl ester 14 (2.4 g, 6.0 mmol) was treated as described

in procedure c to provide 20 as an orange solid (2.53 g) in 92%

yield. 20: mp >300 °C; IR (KBr) 2924, 1743 (CdO), 1651,

1

(neat) 3374, 2917, 1736 (CdO) cm^-1; H NMR (CDCl₃) δ 0.89

(m, 3H), 1.15 (t, 3H, J ) 7.1 Hz) 1.32 (m, 9H), 1.48 (m, 1H),

1.80 (m, 2H), 3.35 (s, 3H), 3.41 (s, 2H), 3.76 (m, 3H), 4.07 (m,

3H), 4.11 (m, 2H), 6.83 (m, 1H), 7.04 (d, 1H, J ) 2.3 Hz), 7.11

1

1384, 1216 cm^-1; H NMR (CDCl₃) δ 0.89 (t, 3H, J ) 6.8 Hz),

1.26 (m, 12 H), 1.46 (m, 1H), 1.77 (m, 2H), 3.15 (s, 3H), 3.33

BzR Agonist Ligands

J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 14 2549

(m, 2H), 3.66 (m, 2H), 3.95 (m, 2H), 4.29 (m, 1H), 6.61 (m,

1H), 6.69 (m, 1H), 7.20 (m, 1H), 9.47 (s, 1H); MS (CI) m/e 460

(M⁺+ 1, 0), 418 (100), 385 (24), 315 (26). This material was

employed in a later step.

3.52 (s, 3H), 4.09 (t, 2H, J ) 6.5 Hz), 4.52 (q, 2H, J ) 7.1 Hz),

5.38 (s, 2H), 7.24 (dd, 1H, J ) 2.4, 8.9 Hz), 7.46 (d, 1H, J )

9.0 Hz), 7.80 (d, 1H, J ) 2.2 Hz), 8.62 (br s, 1H), 8.87 (s, 1H);

MS (CI) m/e 414 (M⁺+ 1, 100), 301 (10). Anal. (C₂₄H₃₂N₂O₄)

C, H, N.

6-(Octyloxy)-4-eth yl-3-(eth oxycar bon yl)-1,2,3,4-tetr ah y-

d r o-â-ca r bolin e-1-ca r boxylic Acid (19). The amino ethyl

ester 13 (1.24 g, 3.2 mmol) was treated as described in

procedure c to provide 19 as a mixture of diastereomers (1.29

g) in 91% yield. 19: mp >300 °C; IR (KBr) 2924, 2854, 1743

(CdO) cm^-1; ¹H NMR (CDCl₃) δ 0.89 (t, 3H, J ) 6.2 Hz), 0.99

(m, 2H), 1.30 (m, 12H), 1.47 (m, 2H), 1.79 (m, 2H), 3.68 (m,

1H), 3.70 (m, 3H), 4.08 (m, 2H), 4.28 (m, 2H), 4.49 (m, 1H),

6.78 (m, 1H), 6.96 (m, 1H), 7.16 (m, 1H), 7.32 (s, 1H); MS (CI)

m/e 444 (M⁺+ 1, 3), 430 (20), 402 (100). This material was

employed in a later step.

6-(P r op yloxy)-4-(m eth oxym eth yl)-3-(eth oxyca r bon yl)-

1,2,3,4-tetr a h yd r o-â-ca r bolin e-1-ca r boxylic Acid (18). The

amino ethyl ester 12 (15 g, 4.5 mmol) was treated as described

in procedure c to provide the carboxylic acid 18 as a mixture

of diastereomers (10.2 g) in 90% yield. 18: mp 160-163 °C;

¹H NMR (CDCl₃) δ 3.27 (s, 3H), 3.73 (m, 2H), 3.90 (m, 2H),

4.0 (m, 2H), 6.77 (d, 1H), 6.92 (s, 1H), 6.96 (s, 1H), 7.04 (s,

1H), 7.07 (d, 1H), 7.23 (s, 1H), 9.12 (s, 1H); MS (CI) m/e 347

(M⁺+ 1, 100). This material was employed in a later step.

6-(Octyloxy)-4-eth yl-â-car bolin e-3-car boxylic Acid Eth -

yl Ester (25). The â-carboline 19 (3.89 g, 0.0085 mol) in

xylenes (50 mL) was heated at reflux, followed by oxidation

in DMSO (21 mL) with sulfur powder (0.55 g) as described in

procedure d. A wash column (silica gel, CH₂Cl₂) was run on

the black residue followed by crystallization from hot ethyl

acetate to provide 25 as light beige crystals (1.01 g, 30%). 25:

1

mp 163-165 °C; IR (KBr) 2931, 1701 (CdO) cm^-1; H NMR

(CDCl₃) δ 0.89 (t, 3H, J ) 6.9 Hz), 1.30-1.56 (m, 16H), 1.86

(m, 2H), 3.56 (q, 2H, J ) 7.3 Hz), 4.08 (t, 2H, J ) 6.5 Hz), 4.5

(q, 2H, J ) 7.1 Hz), 7.25 (dd, 1H, J ) 2.3, 8.9 Hz), 7.49 (d, 1H,

J ) 8.9 Hz), 7.71 (d, 1H, J ) 2.2 Hz), 8.82 (s, 1H), 8.95 (br s,

1H); MS (EI) m/e 396 (M⁺, 52), 367 (35), 350 (17), 255 (19),

238 (38), 211 (100), 181 (36). Anal. (C₂₄H₃₂N₂O₃) C, H, N.

6-(P r op yloxy)-4-(m et h oxym et h yl)-â-ca r b olin e-3-ca r -

boxylic Acid Eth yl Ester (24). The carboxylic acid 18 (1.2

g, 2.98 mmol) was suspended in xylenes, and the mixture was

heated to reflux under an atmosphere of N₂for 1 h. Then 10%

Pd/C (1.33 g) was added to the yellow solution and the mixture

heated to reflux for 3 h. The mixture was filtered and the

catalyst washed with hot xylenes (100 mL). The xylenes were

combined and removed under reduced pressure to provide a

crude solid. Flash chromatography provided the â-carboline

24 (512 mg) in 54% yield. 24: mp 250-251 °C; (IR (KBr) 2936,

6-(Ben zyloxy)-4-(m eth oxym eth yl)-3-(eth oxyca r bon yl)-

1,2,3,4-tetr a h yd r o-â-ca r bolin e-1-ca r boxylic Acid (21). The

amino ethyl ester 15 (2.9 g, 7.6 mmol) was treated as described

in procedure c to provide the carboxylic acid 21 as a mixture

of diastereomers (2.7 g) in 87% yield. 21: mp 156-157 °C;

MS (CI) m/e 395 (M⁺- 44, 100). This material was employed

in a later step without further characterization.

1

1706 (CdO) cm^-1; H NMR (G.E. 500 MHz, CDCl₃) δ 1.06 (t,

3H, J ) 7 Hz), 1.32 (t, 3H, J ) 7.4 Hz), 1.85 (sextet, 2H, J )

6.6, 7.4 Hz), 3.52 (s, 3H), 4.02 (t, 2H, J ) 6.6 Hz), 4.43 (q, 2H,

J ) 7 Hz), 5.37 (s, 2H), 7.19 (dd, 1H, J ) 1.5, 9 Hz), 7.41 (d,

1H, J ) 9 Hz), 7.75 (s, 1H), 8.74 (s, 1H), 10.2 (s, 1H); HRMS

m/e 342.1581 (C₁₉H₂₂N₂O₄requires 342.1579). Anal.

(C₁₉H₂₂N₂O₄) C, H, N.

6-(1-Na p h th ylm eth yloxy)-4-(m eth oxym eth yl)-â-ca r bo-

lin e-3-ca r boxylic Acid Eth yl Ester (29) (P r oced u r e d ).

Into a 100 mL round-bottom flask were added carboxylic acid

23 (3.23 g, 6.61 mmol) and a mixture of xylenes (65 mL). The

solution was heated at reflux for 1 h and the solvent removed

under reduced pressure. The residue was dissolved in DMSO

(18 mL), followed by the addition of sulfur powder (0.47 g),

and this mixture was stirred at a bath temperature of 144 °C

for 1.5 h. The DMSO was removed by high vacuum distilla-

tion, followed by a wash column (silica gel, EtOAc), and then

purification by flash chromatography (silica gel, EtOAc/hex-

anes, 80:20) provided â-carboline 29 (642 mg) in 22% yield.

6-(Ben zyloxy)-4-(m et h oxym et h yl)-â-ca r b olin e-3-ca r -

boxylic Acid Eth yl Ester (27). The â-carboline 21 (3.05 g,

6.97 mmol) in xylenes (200 mL) at reflux was subjected to

decarboxylation, followed by oxidation in DMSO (100 mL) with

sulfur powder (0.56 g) as described in procedure d. Flash

chromatography of the black residue (silica gel, 9:1 CHCl₃:

MeOH) followed by crystallization from ethyl acetate/ether

furnished the â-carboline 27 (1.45 g) in 31% yield whose

spectral properties were identical to those previously pub-

lished.^27,2827: mp 187 °C (lit.^27,28mp 187 °C).

29: mp 174-176 °C; IR (KBr) 2931, 1694 (CdO) cm^-^{1 1}H

;

NMR (DMSO-d₆) δ 1.33 (t, 3H, J ) 7.1 Hz), 3.27 (s, 3H), 4.34

(q, 2H, J ) 7.1 Hz), 5.10 (s, 2H), 5.65 (s, 2H), 7.39 (dd, 1H, J

) 2.3, 8.9 Hz), 7.48-7.62 (m, 4H), 7.71 (d, 1H, J ) 6.7 Hz),

7.85 (s, 1H), 7.95 (m, 2H), 8.17 (d, 1H, J ) 8.2 Hz), 8.85 (s,

1H); MS (CI) m/e 440 (M⁺+ 1, 100), 299 (33), 143 (30) 111

(16). Anal. (C₂₇H₂₄N₂O₄₎C, H, N.

9-Meth yl-6-(1-n aph th ylm eth yloxy)-4-(m eth oxym eth yl)-

â-ca r bolin e-3-ca r boxylic Acid Eth yl Ester (35) (P r oce-

d u r e e). Into a 50 mL flame-dried round-bottom flask was

added â-carboline 29 (126 mg, 0.285 mmol) in dry DMF (12

mL). To this solution were added 2.5 equiv of iodomethane

and sodium hydride (16 mg, 60% dispersion in mineral oil, 0.39

mmol), and the solution was stirred at room temperature for

12 h under an atmosphere of N₂. The DMF was removed by

heating in vacuo and the residue taken up in CHCl₃and

washed with water. The aqueous layer was extracted with

CHCl₃(3 × 25 mL). The combined organic layers were washed

with brine and dried (K₂CO₃), and the solvent was removed

under reduced pressure. A wash column (silica gel, ethyl

acetate) provided 35 as an off-white solid (77 mg, 60%). 35:

6-(1-Naph th ylm eth yloxy)-4-eth yl-â-car bolin e-3-car box-

ylic Acid Eth yl Ester (28). The â-carboline 22 (3.12 g, 6.61

mmol) in xylenes at reflux was subjected to decarboxylation,

followed by oxidation in DMSO (7 mL) with sulfur powder

(0.156 g) as described in procedure d. The black residue was

purified using flash chromatography (silica gel, EtOAc/hex-

anes, 80:20) to provide 28 (456 mg) in 16% yield. 28: mp

195.3-196.5 °C; IR (KBr) 2973, 1694 (CdO) cm^-^{1 1}H NMR

;

(DMSO-d₆) δ 1.18 (t, 3H, J ) 7.3 Hz), 1.29 (t, 3H, J ) 7.1 Hz),

3.24 (q, 2H, J ) 7.3 Hz), 4.29 (q, 2H, J ) 7.1 Hz), 5.64 (s, 2H),

7.36 (dd, 1H, J ) 2.2, 9.0 Hz), 7.43-7.69 (m, 6H), 7.92 (m,

2H), 8.18 (d, 1H, J ) 7.6 Hz), 8.69 (s, 1H); MS (EI) m/e 424

(M⁺, 0.9), 141 (100). Anal. (C₂₇H₂₄N₂O₃) C, H, N.

1

mp 138-140 °C; IR (KBr) 2931, 1708 (CdO) cm^-1; H NMR

(DMSO-d₆) δ 1.32 (t, 3H, J ) 7.1 Hz), 3.25 (s, 3H), 3.98 (s,

3H), 4.33 (q, 2H, J ) 7.1 Hz), 5.09 (s, 2H), 5.66 (s, 2H), 7.43-

7.58 (m, 5H), 7.64 (m, 2H), 7.92 (m, 2H), 7,15 (d, 1H, J ) 7.7

Hz), 8.97 (s, 1H); MS (EI) m/e 454 (M⁺, 21), 141 (100), 115

(17). Anal. (C₂₈H₂₆N₂O₄) C, H, N.

6-(Octyloxy)-4-(m eth oxym eth yl)-â-ca r bolin e-3-ca r box-

ylic Acid Eth yl Ester (26). The â-carboline 20 (2.40 g,

0.0052 mol) in xylenes (28 mL) at reflux was subjected to

decarboxylation, followed by oxidation in DMSO (13 mL) with

sulfur powder (0.33 g) as described in procedure d. The DMSO

was removed by high vacuum distillation, and a wash column

(silica gel, CH₂Cl₂) was run on the black residue. Crystalliza-

tion from hot ethyl acetate provided the fully aromatic â-car-

boline 26 (820 mg) in 38% yield. 26: mp 147.3-149 °C; IR

(KBr) 3198 (NH), 2931, 1701 (CdO) cm^-1; ¹H NMR (CDCl₃) δ

0.89 (t, 2H, J ) 6.9 Hz), 1.29-1.57 (m, 16H), 1.84 (m, 2H),

9-Meth yl-6-(1-n a p h th ylm eth yloxy)-4-eth yl-â-ca r bolin e-

3-ca r boxylic Acid Eth yl Ester (34). A solution of â-carbo-

line 28 (350 mg, 0.824 mmol) in dry DMF (18 mL) was

methylated as described in procedure e to provide 34 (306 mg,

84%) as a white solid. 34: mp 153-156 °C; IR (KBr) 2980,

1

1715 (CdO) cm^-1; H NMR (DMSO-d₆) δ 1.24 (t, 3H, J ) 7.0

Hz), 1.33 (t, 3H, J ) 7.1 Hz), 3.23 (q, 2H, J ) 6.5 Hz), 3.97 (s,

3H), 4.33 (q, 2H, J ) 7.6 Hz), 5.71 (s, 2H), 7.46-7.62 (m, 4H),

2550 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 14

Cox et al.

7.73 (m, 3H), 7.95 (m, 2H), 8.22 (d, 1H, J ) 7.5 Hz), 8.87 (s,

1H); MS (EI) m/e 438 (M⁺, 2), 141 (100), 115 (29). Anal.

(C₂₈H₂₆N₂O₃) C, H, N.

yellow solution which resulted was cooled to 0 °C followed by

the addition of NaNO₂(87 mg, 1.26 mmol, 1.1 equiv) in H₂O

(2 mL). After being stirred at 0 °C for 35 min, the solution

was brought to alkaline pH with saturated aqueous NaHCO₃.

The precipitate which formed was collected by filtration and

brought up in acetic acid/water (1:1, 60 mL). The suspension

was then heated at reflux during which time N₂was evolved.

After 45 min a new component was observed by TLC which

turned brown some time later under UV light. The solvent

was removed under reduced pressure to furnish an oil which

was taken up in ethanol to precipitate 40 as a yellow solid as

the diacetate salt. After recrystallization from ether the

â-carboline 40 (275 mg, 88% yield) was obtained as the

diacetate salt. 40: mp 267-268 °C; IR (KBr) 3448, 1642, 1562,

9-Meth yl-6-(octyloxy)-4-(m eth oxym eth yl)-â-ca r bolin e-

3-ca r boxylic Acid Eth yl Ester (32). A solution of â-carbo-

line 26 (50 mg, 0.121 mmol) in dry DMF (8 mL) was

methylated as described in procedure e to provide 32 as an

off-white solid (41 mg, 80%). 32: mp 96.2-97.4 °C; IR (KBr)

1

2925, 1708 (CdO) cm^-1; H NMR (CDCl₃) δ 0.89 (t, 3H, J )

6.8 Hz), 1.26-1.56 (m, 16H), 1.84 (m, 2H), 3.51 (s, 3H), 3.96

(s, 3H), 4.10 (t, 2H, J ) 6.5 Hz), 4.52 (q, 2H, J ) 7.1 Hz), 5.37

(s, 2H), 7.31 (dd, 1H, J ) 2.4, 8.9 Hz), 7.42 (d, 1H, J ) 8.9

Hz), 7.84 (d, 1H, J ) 2.3 Hz), 8.86 (s, 1H); MS (CI) m/e 428

(M⁺+ 1, 100), 396 (12), 381 (7), 315 (11). Anal. (C₂₅H₃₄N₂O₄)

C, H, N.

1

1410 cm^-1; H NMR (CDCl₃) δ 1.07 (t, 3H, J ) 7.4 Hz), 1.85

(m, 2H), 3.45 (s, 3H), 3.99 (t, 2H, J ) 6.6 Hz), 4.73 (s, 1H), 5.0

(s, 2H), 7.15 (s, 1H), 7.28 (m, 2H), 7.63 (s, 1H), 7.97 (s, 1H),

8.27 (s, 1H); ¹³C NMR (CDCl₃) δ 10.61, 22.76, 57.76, 67.80,

70.72, 76.60, 108.26, 109.21, 112.15, 118.75, 121.16, 128.48,

128.56, 131.70, 137.69, 153.06; MS (CI) m/e 286 (M⁺+ 1, 100),

154 (65); high-resolution mass spectrum, m/e 285.1481

(C₁₆H₁₉N₃O₂requires 285.1477).

9-Me t h yl-6-(oct yloxy)-4-e t h yl-â-ca r b olin e -3-ca r b ox-

ylic Acid E t h yl E st er (31). A solution of â-carboline 25

(0.119 g, 0.299 mmol) in dry DMF (12 mL) was methylated

using 2.5 equiv of iodomethane and sodium hydride (14 mg,

60% dispersion in mineral oil, 0.35 mmol) as described in

procedure e. A wash column (silica gel, ethyl acetate) on the

residue provided 31 (103 mg, 84%) as an off-white solid. 31:

mp 94-95 °C; IR (KBr) 2924, 1708 (CdO) cm^-1

;

¹H NMR

6-(Be n zyloxy)-4-(m e t h oxym e t h yl)-3-a m in o-â-ca r b o-

lin e (41). A suspension of the carbohydrazide 39 (700 mg,

1.86 mmol) in H₂O (50 mL) and THF (3 mL) was dissolved by

the dropwise addition of aqueous concentrated HCl. The

yellow solution which resulted was cooled to 0 °C followed by

the addition of NaNO₂(141 mg, 2.04 mmol) in H₂O (2 mL).

After 35 min of stirring at 0 °C, the solution was brought to

alkaline pH with saturated aqueous NaHCO₃. The precipitate

which formed was collected by filtration and brought up in

acetic acid/water (1:1, 100 mL). The suspension was then

heated at reflux during which time nitrogen was evolved.

After 45 min a new component was observed by TLC which

turned brown some time later under UV light. The solvent

was removed under reduced pressure and the crude solid

recrystallized from EtOH to furnish the amine 41 as the

diacetate salt (492 mg) in 60% overall yield. 41: mp 175-

176 °C; IR (KBr) 1634, 1559 cm^-1; ¹H NMR (CDCl₃) δ 3.39 (s,

3H), 4.95 (s, 2H), 5.14 (s, 2H), 7.30 (m, 6H), 7.50 (m, 3H), 7.68

(d, 1H, J ) 8.7 Hz), 8.07 (s, 1H), 8.28 (s, 1H); MS (EI 15 eV)

m/e 333 (M⁺, 67), 242 (100). Anal. (C₂₀H₁₉N₃O₂) C, H, N.

6-(P r opyloxy)-4-(m eth oxym eth yl)-â-car bolin e-3-isoth io-

cya n a te (42). A suspension of 3-amino â-carboline 40 as the

diacetate salt (50 mg, 0.124 mmol) in a saturated aqueous

solution of NaHCO₃(5 mL) was added to CHCl₃(5 mL). After

10 min of vigorous stirring at room temperature, thiophosgene

(0.011 mL, 0.017 g, 0.148 mmol) was syringed into the

chloroform solution and the mixture was allowed to stir for 1

h. The reaction progress was monitored by TLC (silica gel,

CHCl₃/MeOH, 9.5:0.5). The organic layer was then separated

and the aqueous layer extracted with CHCl₃(3 × 10 mL). The

organic layers were combined, and the solvent was removed

under reduced pressure. The crude solid was then purified

by flash chromatography to furnish the isothiocyanate 42 (14

mg, 34%). 42: mp 166-168 °C; IR (KBr) 3460, 2924 cm^-1; ¹H

NMR (CDCl₃) δ 1.07 (t, 3H, J ) 7.4 Hz), 1.85 (m, 2H), 3.49 (s,

3H), 4.02 (t, 2H, J ) 6.6 Hz), 5.03 (s, 2H), 7.19 (m, 1H), 7.40

(m, 1H), 7.68 (d, 1H, J ) 9 Hz), 8.30 (s, 1H), 8.61 (s, 1H); ¹³C

NMR (CDCl₃) δ 49.72, 55.89, 62.05, 76.99, 108.26, 109.21,

112.15, 118.75, 121.16, 128.48, 128.56, 148.24, 158.32, 175.48,

207.28; MS (CI) m/e 328 (M⁺+ 1, 28), 296 (100); HRMS m/e

327.1043 (C₁₇H₁₇N₃O₂S requires 327.1041).

(CDCl₃) δ 0.89 (t, 2H, J ) 6.7 Hz), 1.32 (m, 8H), 1.51 (m, 8H),

1.81 (m, 2H), 3.58 (t, 2H, J ) 7.5 Hz), 3.96 (s, 3H), 4.09 (t, 2H,

J ) 6.6 Hz), 4.52 (q, 2H, J ) 7.1 Hz), 7.31 (dd, 1H, J ) 2.4,

8.9 Hz), 7.44 (d, 1H, J ) 8.9 Hz), 7.75 (d, 1H, J ) 2.2 Hz),

8.81 (s, 1H); MS (CI) m/e 412 (M⁺+ 1, 100), 299 (12). Anal.

(C₂₅H₃₄N₂O₃) C, H, N.

9-Meth yl-6-(pr opyloxy)-4-(m eth oxym eth yl)-â-car bolin e-

3-ca r boxylic Acid Eth yl Ester (30). A solution of â-carbo-

line 24 (60 mg, 0.175 mmol) in dry DMF (6 mL) was

methylated as described in procedure e. Flash chromatogra-

phy provided the N_a-methyl-â-carboline 30 (35 mg) in 56%

yield. 30: mp 112-113 °C; IR (KBr) 2936, 1706 (CdO) cm^-1

;

¹H NMR (CDCl₃) δ 1.23 (t, 3H, J ) 7 Hz), 1.47 (t, 3H, J ) 7.4

Hz), 1.87 (sextet, 2H, J ) 6.6 Hz), 3.49 (s, 3H), 4.07 (t, 2H, J

) 6.6 Hz), 4.51 (q, 2H, J ) 7 Hz), 5.32 (s, 2H), 7.31 (d, 1H, J

) 9 Hz), 7.41 (d, 1H, J ) 9 Hz), 7.83 (s, 1H), 8.84 (s, 1H); MS

(CI) m/e 357 (M⁺+ 1, 100); HRMS m/e 356.1729 (C₂₀H₂₄N₂O₄

requires 356.1736).

6-(P r op yloxy)-4-(m eth oxym eth yl)-â-ca r bolin e-3-ca r bo-

h yd r a zid e (38). A solution of â-carboline 24 (500 mg, 1.48

mmol) in EtOH (20 mL) which contained 98% anhydrous

hydrazine (5 mL) was heated at reflux under an atmosphere

of N₂for 2.5 h at which time analysis by TLC (silica gel, 9.5:

0.5 EtOAc:MeOH) indicated the disappearance of starting

material. As the solution cooled, white crystals formed. The

crystals were collected by filtration, washed with EtOH, and

dried to furnish the carbohydrazide 38 (300 mg) in 81% yield.

38: mp 230-231 °C; IR (KBr) 3270, 1638 cm^-^{1 1}H NMR

;

(CDCl₃, CF₃COOD, 500 MHz) δ 1.10 (t, 3H, J ) 7.4 Hz), 1.91

(sextet, 2H, J ) 6.6,7.4 Hz), 3.62 (s, 3H), 4.12 (t, 2H, J ) 6.6

Hz), 5.42 (s, 2H), 7.63 (s, 1H), 7.73 (d, 1H, J ) 9.5 Hz), 7.81 (s,

1H), 9.29 (s, 1H), 10.81 (s, 1H); MS (CI) m/e 329 (M⁺+ 1, 100),

297 (21). Anal. (C₁₇H₂₀N₄O₃) C, H, N.

6-(Ben zyloxy)-4-(m eth oxym eth yl)-â-ca r bolin e-3-ca r bo-

h yd r a zid e (39). A solution of â-carboline 27 (800 mg, 2.0

mmol) in EtOH (30 mL) which contained 98% anhydrous

hydrazine (6 mL) was heated at reflux under an atmosphere

of N₂for 7 h at which time analysis by TLC (silica gel, 9.5:0.5

EtOAc:MeOH) indicated the disappearance of starting mate-

rial. The solution was cooled, and the resulting precipitate

which formed was collected by filtration, washed with EtOH,

and dried to provided â-carboline 39 (730 mg) in 95% yield.

6-(Ben zyloxy)-4-(m eth oxym eth yl)-â-car bolin e-3-isoth io-

cya n a te (43). To a stirred solution of the amine 41 as the

diacetate salt (100 mg, 0.220 mmol) in a saturated aqueous

solution of NaHCO₃(10 mL) was added chloroform (10 mL).

After 15 min of vigorous stirring at ambient temperature,

thiophosgene (0.028 g, 0.019 mL, 0.25 mmol) was syringed into

the chloroform solution, and stirring was continued for another

hour. The reaction progress was monitored by TLC (silica gel,

EtOAc/MeOH, 9:1). The chloroform layer was separated and

the water layer extracted with CHCl₃(3 × 60 mL). The

organic layers were combined and dried (K₂CO₃), and the

1

39: mp 243-244 °C; IR (KBr) 3303 cm^-1; H NMR (DMSO-

d₆, CF₃COOD, 250 MHz) δ 3.45 (s, 3H), 5.30 (s, 2H), 7.25 (s,

2H), 7.40 (m, 6H), 7.70 (m, 3H), 9.28 (s, 1H), 11.15 (s, 1H);

MS (CI) m/e 377 (M⁺+ 1, 100), 345 (32). Anal. (C₂₁H₂₀N₄O₃)

C, H, N.

6-(P r op yloxy)-4-(m e t h oxym e t h yl)-3-a m in o-â-ca r b o-

lin e (40). A suspension of the carbohydrazide 38 (380 mg,

1.1 mmol) in H₂O (50 mL) and THF (3 mL) was dissolved by

the dropwise addition of aqueous concentrated HCl. The

BzR Agonist Ligands

J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 14 2551

solvent was removed under reduced pressure. The crude solid

was purified by flash chromatography (silica gel, EtOAc) and

crystallized from benzene to yield the isothiocyanate 43 (69

mg, 83%). 43: mp 189 °C; IR (KBr) 2140 (NCS), 1662 (NdC),

1559, 1498, 1381 (CdS) cm^-1; ¹H NMR (CDCl₃) δ 3.43 (s, 3H),

4.98 (s, 2H), 5.16 (s, 2H), 7.30 (m, 5H), 7.46 (s, 1H), 7.48 (d,

1H, J ) 8.5 Hz), 7.74 (s, 1H), 8.60 (s, 1H); MS (EI) m/e 375

(M⁺, 62), 284 (60) 254 (100). Anal. (C₂₁H₁₇N₃O₂S) C, H, N.

homogeneous (3 h). The reaction was cooled to 5 °C and the

pH adjusted to 4 with aqueous HCl (10%). The yellow solid

which formed was filtered from the medium and washed with

H₂O (5 mL). The solid was air-dried to provide 47 (90 mg) in

98% yield. 47: mp 262 °C dec; IR (KBr) 3408, 3070, 2970,

1616, 1350, 1210, 1098 cm^-1; ¹H NMR (DMSO-d₆, 500 MHz) δ

1.01 (t, 3H, J ) 7.4 Hz), 1.78 (sextet, 2H, J ) 6.6, 7.4 Hz),

2.48 (s, 3H), 3.37 (s, 3H), 4.02 (t, 2H, J ) 6.6 Hz), 5.30 (s, 2H),

7.25 (d, 1H, J ) 8.5 Hz), 7.56 (d, 1H, J ) 8.5 Hz), 7.88 (s, 1H),

8.84 (s, 1H), 12.01 (s, 1H); MS (CI) m/e 315 (M⁺+ 1, 100);

HRMS m/e 314.1254 (C₁₇H₁₈N₂O₄requires 314.1266). This

material was employed directly in the next step.

6-(P r op yloxy)-4-(m et h oxym et h yl)-â-ca r b olin e-3-ca r -

boxylic Acid Isop r op yl Ester (48). The carboxylic acid 47

(90 mg, 0.286 mmol) was dissolved in a solution of 2-propanol

(50 mL) which had been previously saturated with a solution

of anhydrous hydrogen chloride. The reaction mixture was

held at reflux under nitrogen until analysis by TLC (silica gel,

CHCl₃/MeOH, 9.5:0.5) indicated the absence of starting mate-

rial. The solvent was removed under reduced pressure and

the residue dissolved in H₂O (40 mL). The pH of the aqueous

solution was adjusted to 8 with concentrated aqueous NH₄-

OH. The aqueous layer was then extracted with CHCl₃(6 ×

20 mL). The CHCl₃was removed under reduced pressure to

yield a yellow solid. Flash chromatography (silica gel, CHCl₃/

MeOH, 9.5:0.5) on the crude product gave the isopropyl ester

(60 mg, 85%) as a fine yellow powder. 48: mp 243-244 °C;

IR (KBr) 3409, 3156, 2966, 1602 (CdO) cm^-1; ¹H NMR (DMSO-

d₆) δ 1.02 (t, 6H, J ) 7.4 Hz), 1.39 (m, 3H), 1.85 (m, 2H), 3.31

(s, 3H), 4.05 (t, 2H, J ) 6.6 Hz), 5.92 (s, 2H), 7.31 (d, 1H, J )

9.0 Hz), 7.59 (s, 1H), 7.65 (d, 1H, J ) 9 Hz), 9.06 (s, 1H); ¹³C

NMR (CDCl₃) δ 10.58, 22.21, 67.26, 67.82, 71.67, 83.07, 107.12,

107.45, 113.45, 114.02, 115.21, 118.23, 119.74, 121.54, 133.20,

136.98, 137.40, 137.94, 153.73, 169.45; MS (CI) m/e 357 (M⁺

+ 1, 100) 313 (50); HRMS m/e 356.1725 (C₂₀H₂₅N₂O₄requires

356.1736).

6-(Ben zyloxy)-4-(m et h oxym et h yl)-â-ca r b olin e-3-t h io-

ca r ba m a te (44). To a solution of isothiocyanate 43 (30 mg,

0.08 mmol) in dry CH₃OH (10 mL) was added a solution of

dry CH₃OH in which anhydrous HCl (10 mL) had been

dissolved. Examination of the solution by TLC (silica gel,

EtOAc/MeOH, 9.5:0.5, NH₄OH fumes) indicated the formation

of a new component which exhibited an intense purple color

under UV light. The solution was stirred for 2 min and the

solvent removed under reduced pressure at room temperature.

The oil which remained was washed with dry ether (2 × 10

mL) and purified by flash chromatography (silica gel, EtOAc)

to furnish the thiocarbamate 44 (20 mg, 61%). 44: mp 177

°C; IR (KBr) 1655 (NdC), 1559, 1381 (CdS), 1217 cm^-^{1 1}H

;

NMR (CDCl₃) δ 3.39 (s, 3H), 4.05 (s, 3H), 4.88 (s, 2H), 5.16 (s,

2H), 7.23 (m, 4H), 7.35 (m, 2H), 7.48 (m, 2H), 7.68 (s, 1H),

8.40 (s, 1H), 8.60 (s, 1H); MS (EI) m/e 375 (56), 284 (58), 254

(100). Anal. (C₂₂H₂₀N₃O₃S) C, H, N.

6-(P r op yloxy)-4-(m eth oxym eth yl)-â-ca r bolin e 3-n -P r o-

p yl Eth er (45). To a solution of 3-amino-â-carboline 40 (100

mg, 0.248 mmol), dry 1-propanol (25 mL), and isoamyl nitrite

(0.034 g, 0.0399 mL, 0.287 mmol, 1.2 equiv) was added

concentrated H₂SO₄(0.5 mL) dropwise until the â-carboline

40 dissolved. The mixture was stirred at room temperature

for ¹/₂h followed by heating at reflux for ¹/₂h. The reaction

was monitored by TLC (silica gel, CHCl₃/MeOH, 9.5:0.5). After

the mixture was allowed to cool to room temperature, H₂O (250

mL) was added followed by a saturated aqueous solution of

NaHCO₃(25 mL). The mixture was then extracted with CHCl₃

(3 × 25 mL) and the solvent removed under reduced pressure.

Flash chromatography (silica gel, CHCl₃/MeOH, 9.5:0.5) pro-

vided the 3-n-propyl ether 45 (30 mg, 52%). This material was

homogeneous on TLC in two different solvent systems. 45:

Ack n ow led gm en t. We thank NIMH for generous

financial support. We also thank Dr. Eugene DeRose,

Mr. Frank Laib, and Mr. Keith Krumrow for some

spectral and analytical data, respectively.

1

mp 173-174 °C; IR (KBr) 3433, 3221 cm^-1; H NMR (CDCl₃)

δ 1.06 (t, 6H, J ) 7.4 Hz), 1.26 (t, 3H, J ) 7.4 Hz), 1.76 (m,

4H), 3.97 (t, 2H, J ) 6.6 Hz), 5.37 (s, 2H), 7.04 (dd, 1H, J )

2.5, 9 Hz), 7.41 (d, 1H, J ) 2.5 Hz), 7.48 (d, 1H, J ) 9 Hz),

8.29 (s, 1H), 10.20 (s, 1H); ¹³C NMR (CDCl₃) δ 10.58, 14.30,

22.68, 58.17, 61.65, 67.83, 70.40, 107.85 112.60, 119.16, 121.67,

128.74, 129.13, 132.94, 136.14, 137.22, 137.65, 153.99, 167.23;

MS (CI) m/e 329 (M⁺+ 1, 48), 297 (100); HRMS m/e 328.1787

(C₁₉H₂₄N₂O₃requires 328.1786).

Refer en ces

(1) Wisden, W.; Laurie, D. J .; Monyer, H.; Seeburg, P. H. The

distribution of 13 GABA_Areceptor subunit mRNAs in the rat

brain. I. Telencephalon, diencephalon, mesencephalon. J . Neu-

rosci. 1992, 12, 1040-1062.

(2) Young, A. B.; Chu, D. GABA_Aand GABA_Breceptors in mam-

malian brain: potential targets for drug development. Drug Dev.

Res. 1990, 21, 161-167.

6-(Ben zyloxy)-4-(m eth oxym eth yl)-â-ca r bolin e 3-n -P r o-

p yl Eth er (46). To a solution of amine 41 (100 mg, 0.202

mmol) in anhydrous 1-propanol (25 mL) at -20 °C (dry ice/

CCl₄) was added isoamyl nitrite (0.485 mL, 3.62 mmol). The

solution was allowed to stir for 2 min, followed by the addition

of KSCN (0.714 g, 7.35 mmol), and Cu^ISCN (4.473 g, 0.0367

mol) in anhydrous propanol (15 mL) at -20 °C. After 4 h of

stirring, analysis by TLC (silica gel, EtOAc/MeOH, 9.5:0.5)

indicated the absence of starting material. The inorganic salts

were filtered from the medium and the solvent removed under

reduced pressure. The residue was then dissolved in aqueous

NaHCO₃and extracted with EtOAc (3 × 10 mL). The organic

layer was removed under reduced pressure and the crude solid

purified by flash chromatography (silica gel, EtOAc) to yield

the ether 46 (48.6 mg, 58%) which was directly converted into

(3) Bowery, N. G. GABA_Breceptor pharmacology. Annu. Rev.

Pharmacol. Toxicol. 1993, 33, 109-147.

(4) Bormann, J . Electrophysiology of GABA_Aand GABA_Breceptor

subtypes. Trends Neurosci. 1988, 11, 112-116.

(5) Sieghart, W. Structure and Pharmacology of γ-aminobutyric acid

A receptor subtypes. Pharmacol. Rev. 1995, 47, 181-234.

(6) Olsen, R. W.; Tobin, A. J . Molecular biology of GABA_Areceptors.

FASEB J . 1990, 4, 1469-1480.

(7) Wafford, K. A.; Burnett, D. M.; Dunwiddie, T. V.; Harris, R. A.

Genetic differences in the ethanol sensitivity of GABA_Areceptors

expressed in Xenopus oocytes. Science 1990, 249, 291-293.

(8) Wafford, K. A.; Burnett, D. M.; Burt, D. R.; Wang, J . B.; Kofuji,

P.; Dunwiddie, T. V.; Harris, R. A. Ethanol sensitivity of the

GABA_Areceptor expressed in Xenopus oocytes requires 8 amino

acids contained in the γ_2Lsubunit. Neuron 1991, 7, 27-33.

(9) Doble, A.; Martin, I. L. Multiple benzodiazepine receptors: no

reason for anxiety. Trends Pharmacol. Sci. 1992, 13, 76-81.

(10) Levitan, E. S.; Schofield, P. R.; Burt, D. R.; Rhee, L. M.; Wisden,

W.; Kohler, M.; Fujita, N.; Rodriguez, H. F.; Stephenson, A.;

Darlison, M. G.; Barnhard, E. Z.; Seeburg, P. H. Structural and

functional basis for GABA receptor heterogeneity. Nature 1988,

335, 76-79.

(11) Pritchett, D. B.; Seeburg, P. H. γ-Aminobutyric acid A receptor

R5 subunit creates novel type II benzodiazpine receptor phar-

macology. J . Neurochem. 1990, 54, 1802-1804.

(12) Liu, R. Y.; Hu, R.; Zhang, P.; Skolnick, P.; Cook, J . M. Synthesis

and pharmacological properties of novel 8-substituted imidazo-

benzodiazepines: High affinity, selective probes for R5 contain-

ing GABA_Areceptors. J . Med. Chem. 1996, 39, 1928-1934.

the HCl salt. 46: mp 143-144 °C; IR (KBr) 1595 cm^-^{1 1}H

;

NMR (CF₃CO₂D) δ 1.58 (t, 3H, J ) 7.0 Hz), 1.78 (m, 2H), 3.69

(s, 3H), 4.24 (t, 2H, J ) 7.0 Hz), 5.19 (s, 1H), 5.25 (s, 2H), 7.4

(m, 7H), 7.58 (s, 1H), 7.7 (s, 1H), 8.7 (s, 1H); MS (EI) m/e 376

(M⁺, 53), 285 (100) 255 (35) 213 (64). Anal. (C₂₃H₂₄N₂O₃) C,

H, N.

6-(P r op yloxy)-4-(m et h oxym et h yl)-â-ca r b olin e-3-ca r -

boxylic Acid (47). The â-carboline 24 was suspended in 10%

aqueous NaOH (10 mL) and the slurry heated at reflux. The

suspension was held at reflux until the solution became

2552 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 14

Cox et al.

(13) Allen, M. S.; Hagen, T. J .; Trudell, M. L.; Codding, P. W.;

Skolnick, P.; Cook, J . M. Synthesis of novel 3-substituted

â-carbolines as benzodiazepine receptor ligands: probing the

benzodiazepine receptor pharmacophore. J . Med. Chem. 1988,

31, 1854-1860.

(14) Allen, M. S.; Tan, Y.-C.; Trudell, M. L.; Narayanan, K.; Schindler,

L. R.; Martin, M. J .; Schultz, C.; Hagen, T. J .; Koehler, K. F.;

Codding, P. W.; Skolnick, P.; Cook, J. M. Synthetic and computer-

assisted analyses of the pharmacophore for the benzodiazepine

receptor inverse agonist site. J . Med. Chem. 1990, 33, 2343-

2357.

(15) Allen, M. S.; LaLoggia, A. J .; Dorn, L. J .; Martin, M. J .;

Costantino, G.; Hagen, T. J .; Koehler, K. F.; Skolnick, P.; Cook,

J . M. Predictive binding of â-carboline inverse agonists and

antagonists via the CoMFA/GOLPE approach. J . Med. Chem.

1992, 35, 4001-4010.

(16) Cox, E. D.; Hagen, T. J .; McKernan, R. M.; Cook, J . M. Bz1

Receptor subtype specific ligands. Synthesis and biological

properties of âCCt, a Bz1 receptor subtype specific antagonist.

Med. Chem. Res. 1995, 5, 710-718.

(17) Cox, E. D.; Reddy, M. S.; Cook, J . M., manuscript in preparation.

(18) Diaz-Arauzo, H.; Evoniuk, G. E.; Skolnick, P.; Cook, J . M. The

agonist pharmacophore of the benzodiazepine receptor. Synthe-

sis of a selective anticonvulsant/anxiolytic. J . Med. Chem. 1991,

34, 1754-1756.

(19) Diaz-Arauzo, H.; Koehler, K. F.; Hagen, T. J .; Cook, J . M.

Synthetic and computer assisted analysis of the pharmacophore

for agonists at benzodiazepine receptors. Life Sci. 1991, 49, 207-

216.

(20) Liu, R. Y.; Gillard, N. P.; Gu, Z.-Q.; Rice, K.; McKernan, R.; Cook,

J . M. Search for R5 benzodiazepine receptor subtype specific

ligands based on the structure of sarmazenil (Ro15-4513).

Presented at the 56th annual scientific meeting of the college

on problems of drug dependence; The Breakers, West Palm

Beach, J une 18-23, 1994, oral communications VI.

(21) Liu, R.; Zhang, P.; McKernan, R. M.; Wafford, K.; Cook, J . M.

Synthesis of novel imidazobenzodiazepines selective for the

R5â2γ2 (Bz5) GABA_A/benzodiazepine receptor subtype. Med.

Chem. Res. 1995, 5, 700-709.

(27) Hollinshead, S. P.; Trudell, M. L.; Skolnick, P.; Cook, J . M.

Structural requirements for agonist actions at the benzodiaz-

epine receptor: Studies with analogs of 6-(benzyloxy)-(4-meth-

oxymethyl)-â-carboline-3-carboxylic acid ethyl ester. J . Med.

Chem. 1990, 33, 1062-1069.

(28) Neef, L. G.; Eder, V.; Huth, A.; Rahtz, G.; Schmiechen, R.;

Seidelmann, C. Synthesis of 4-substituted â-carbolines. Hetero-

cycles 1988, 20, 1295-1313.

(29) Brieger, G.; Nestrick, T. J . Catalytic transfer hydrogenation.

Chem. Rev. 1974, 74, 567-580.

(30) Cox, E. D.; Cook, J . M. The Pictet-Spengler condensation: A

new direction for an old reaction. Chem. Rev. 1995, 95, 1797-

1842.

(31) Dodd, R. H.; Quannes, C.; Prado de Carvalho, L.; Valin, A.;

Venault, P.; Chapouthier, G.; Rossier, J .; Potier, P. 3-Amino-â-

carboline derivatives and the benzodiazepine receptor. Synthesis

of a selective antagonist of the sedative action of diazepam. J .

Med. Chem. 1985, 28, 824-828.

(32) Stephens, D. N.; Schneider, H.; Kehr, W.; Andrews, J . S.; Rettig,

K. J .; Truski, L.; Schmiechen, R.; Turner, J . D.; J ensen, L. H.;

Petersen, E. N.; Honore, T.; Hansen, J . B. Abecarnil, a metaboli-

cally stable, anxioselective â-carboline acting at benzodiazepine

receptors. J . Pharmacol. Exp. Ther. 1990, 253, 334-343.

(33) Turski, L.; Stephens, D. N.; J ensen, L. H.; Petersen, E. N.;

Meldrum, S. B.; Patel, S.; Hansen, J . B.; Lo¨sher, W.; Schneider,

H. H.; Schmiechen, R. Anticonvulsant action of the â-carboline

abecarnil: Studies in rodents and baboon, Papio papio. J .

Pharmacol. Exp. Ther. 1990, 253, 344-352.

(34) Lawson, J .; Uyeno, E. T.; Nienow, J .; Loew, G. H.; Toll, L.

Structure-activity studies of â-carboline analogs. Life Sci. 1984,

34, 2007-2013.

(35) Borea, P. A.; Gilli, G.; Bertolassi, U.; Ferretti, U. Stereochemical

features controlling binding and intrinsic activity properties of

benzodiazepine-receptor ligands. Mol. Pharmacol. 1987, 31,

334-344.

(36) Fryer, R. I.; Cook, C.; Gilman, N. W.; Walser, A. Conformational

shifts at the benzodiazepine receptor related to the binding of

agonists antagonists and inverse agonists. Life Sci. 1986, 39,

1947-1957.

(37) Corda, M. G.; Blaker, W. D.; Mendelson, W. B.; Guidotti, A.;

Costa, E. â-carbolines enhance shock-induced suppression of

drinking in rats. Proc. Natl. Acad. Sci. U.S.A. 1983, 80, 2072-

2076.

(22) Liu, R. Y.; Zhang, P. W.; Gillard, N. P.; McKernan, R.; Cook, J .

M. Synthesis of ligands selective for the R5 benzodiazepine

receptor (BzR) subtype. Presented at the 209th National ACS

Meeting; Anaheim, CA, April 2-6, 1995.

(23) Liu, R. Y.; Zhang, P. W.; Gillard, N. P.; McKernan, R.; Cook, J .

M. Synthesis and study of ligands selective for the R5â2γ2

subtype. Presented at the 28th Great Lakes Regional Meeting

of the ACS; La Crosse, WI, J une 5-8, 1995; abstract 210.

(24) Zhang, W.; Koehler, K.; Zhang, P.; Cook, J . M. Development of

a comprehensive pharmacophore model for the benzodiazepine

receptor. Drug Des. Discovery 1995, 12, 193-248.

(25) Zhang, P.; Liu, R.; McKernan, R. M.; Wafford, K.; Cook, J . M.

Studies of novel imidazobenzodiazepine ligands at GABA_A/BzR

subtypes: effect of C(3) substituents on receptor site selectivity.

Med. Chem. Res. 1995, 5, 487-495.

(26) Martin, M. J .; Trudell, M. L.; Diaz-Arauzo, H.; Allen, M. S.;

LaLoggia, A. J .; Deng, L.; Schultz, C. A.; Tan, Y.-C.; Bi, Y.;

Narayanan, K.; Dorn, L. J .; Koehler, K. F.; Skolnick, P.; Cook,

J . M. Molecular yardsticks. Rigid probes to define the spatial

dimensions of the benzodiazepine receptor binding site. J . Med.

Chem. 1992, 35, 4105-4117.

(38) Skolnick, P.; Paul, S. New concepts in the neurobiology of

anxiety. J . Clin. Psychiatry 1983, 44, 12-20.

(39) Huang, Q.; Cox, E. D.; Gan, T.; Bennett, D. W.; Cook, J . M.;

McKernan, R. M., manuscript in preparation.

(40) Ek, A.; Witkop, B. The synthesis of labile hydroxytryptophan

metabolites. J . Am. Chem. Soc. 1954, 76, 5579-5583.

(41) Wafford, K. A.; Bain, C. J .; Whiting, P. J .; Kemp, J . A. Functional

Comparison of the role of γ subunits in Recombinant Human

γ-Aminobutyric Acid_A/Benzodizepine Receptors. Mol. Pharmacol.

1993, 44, 437-442.

(42) Hadingham, K. L.; Wingrove, P. B.; Wafford, K. A.; Bain, C. J .;

Kemp, J . A.; Palmer, K. J .; Wilson, A. W.; Wilcox, A. S.; Sikela,

J . M.; Ragan, C. I. Role of the â subunit in determining the

pharmacology of human γ-aminobutyric Acid type A recptors.

Mol. Pharmacol. 1993, 44, 1211-1218.

J M970460B

Article Doi

DOI: 10.1021/jm970460b

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jm970460b

Full text of DOI:10.1021/jm970460b

Products guided by the article

R&D Labs maybe for 210160-59-1

Relevant to this article

Hot Product