Synthesis, biochemical evaluation, and classical and three-dimensional quantitative structure-activity relationship studies of 7-substituted- 1,2,3,4-tetrahydroisoquinolines and their relative affinities toward phenylethanolamine N-methyltransferase and t

118

J . Med. Chem. 1999, 42, 118-134

Syn th esis, Bioch em ica l Eva lu a tion , a n d Cla ssica l a n d Th r ee-Dim en sion a l

Qu a n tita tive Str u ctu r e-Activity Rela tion sh ip Stu d ies of

7-Su bstitu ted -1,2,3,4-tetr a h yd r oisoqu in olin es a n d Th eir Rela tive Affin ities

tow a r d P h en yleth a n ola m in e N-Meth yltr a n sfer a se a n d th e r₂-Ad r en ocep tor ^†,1

Gary L. Grunewald,* Vilas H. Dahanukar, Ravi K. J alluri, and Kevin R. Criscione

Department of Medicinal Chemistry, School of Pharmacy, University of Kansas, Lawrence, Kansas 66045

Received J uly 24, 1998

7-Substituted-1,2,3,4-tetrahydroisoquinolines (7-substituted-THIQs) are potent inhibitors of

phenylethanolamine N-methyltransferase (PNMT, EC 2.1.1.28), the enzyme involved in the

biosynthesis of epinephrine. Unfortunately, most of these compounds also exhibit strong affinity

for the R₂-adrenoceptor. To design a selective (PNMT vs R₂-adrenoceptor affinity) inhibitor of

PNMT, the steric and electrostatic factors responsible for PNMT inhibitory activity and R₂-

adrenoceptor affinity were investigated by evaluating a number of 7-substituted-THIQs. A

classical quantitative structure-activity relationship (QSAR) study resulted in a three-

parameter equation for PNMT (PNMT pK_i) 0.599π - 0.0725MR + 1.55σ_m+ 5.80; n ) 27, r

) 0.885, s ) 0.573) and a three-parameter equation for the R₂-adrenoceptor (R₂pK_i) 0.599π

- 0.0542MR - 0.951σ_m+ 6.45; n ) 27, r ) 0.917, s ) 0.397). These equations indicated that

steric effects and lipophilicity play a similar role at either active site but that electronic effects

play opposite roles at either active site. Two binding orientations for the THIQs were postulated

such that lipophilic and hydrophilic 7-substituents would not occupy the same region of space

at either binding site. Using these two binding orientations, based on the lipophilicity of the

7-substituent, comparative molecular field analysis (CoMFA) models were developed that

showed that the steric and electrostatic interactions at both sites were similar to those

previously elaborated in the QSAR analyses. Both the QSAR and the CoMFA analyses showed

that the steric interactions are similar at the PNMT active site and at the R₂-adrenoceptor

and that the electrostatic interactions were different at the two sites. This difference in

electrostatic interactions might be responsible for the selectivity of THIQs bearing a nonlipo-

philic electron-withdrawing group at the 7-position. These QSAR and CoMFA results will be

useful in the design of potent and selective (PNMT vs R₂-adrenoceptor affinity) inhibitors of

PNMT.

In tr od u ction

only central epinephrine in regulating baroreceptor

reflex activity.^5,6This observation provided an impetus

to develop a potent PNMT inhibitor which can be

utilized as a novel antihypertensive agent. Unfortu-

nately, most of the well-studied PNMT inhibitors, such

as 1, exhibit strong affinity for the R₂-adrenoceptor.⁷

Thus, a potent PNMT inhibitor, which exhibits minimal

R₂-adrenoceptor affinity, would be an important phar-

macological tool to delineate the role of epinephrine in

the CNS.

In the biosynthesis of epinephrine, phenylethanola-

mine N-methyltransferase (PNMT, EC 2.1.1.28) cata-

lyzes the conversion of norepinephrine to epinephrine

using S-adenosyl-L-methionine as the cofactor.²Epi-

nephrine accounts for 5-10% of the total catecholamine

content of the central nervous system (CNS),³and it has

been proposed, based on the anatomical localization of

epinephrine neurons in the CNS, to control vital func-

tions such as blood pressure, respiration, and secretion

of pituitary hormones.⁴Studiessusing the desoxycor-

ticosterone salt-type hypertension model in ratsswith

the PNMT inhibitor SK&F 64139 (1, 7,8-dichloro-

1,2,3,4-tetrahydroisoquinoline) showed a decrease in

blood pressure and heart rate.⁵However, no acute blood

pressure changes were noted with the peripherally

active PNMT inhibitor SK&F 29661 (2, 7-aminosulfonyl-

1,2,3,4-tetrahydroisoquinoline),⁵implying the role of

The PNMT inhibitory activity and the R₂-adrenocep-

tor affinity of 1,2,3,4-tetrahydroisoquinolines (THIQs)

are largely a function of the aromatic substitution. For

example, SK&F 72223 (3, 5,8-dimethoxy-1,2,3,4-tet-

rahydroisoquinoline) possesses high affinity for the R₂-

adrenoceptor but is virtually devoid of PNMT inhibitory

activity,⁸while 2 is a potent PNMT inhibitor that

possesses low R₂-adrenoceptor affinity.^8,9Thus, depend-

ing on the aromatic substituent and the substitution

pattern, the THIQ (4, 1,2,3,4-tetrahydroisoquinoline)

nucleus can exhibit high PNMT inhibitory activity, high

R₂-adrenoceptor affinity, or both.

^†This paper is dedicated to the memory of Ray W. Fuller (deceased

1996), noted neuropharmacologist of Eli Lilly and one of the leading

scientists in the initial studies on the importance of PNMT in the

central nervous system. Ray was a strong supporter of our initial

studies in this area, as well as a friend. He is deeply missed.

* To whom correspondence should be addressed. Phone: (785) 864-

4497. Fax: (785) 864-5328. E-mail: ggrunewald@ukans.edu.

Fuller et al.¹⁰have studied the effect of aromatic

substituents on PNMT inhibition in a series of benzyl-

amines. An excellent correlation of the PNMT inhibitory

Published on Web 12/30/1998

PNMT Inhibitors: Classical QSAR and CoMFA

J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 1 119

to contribute equally. A number of 7- and/or 8-sulfur-

substituted THIQs were reported in the patent litera-

ture, but no detailed PNMT inhibitory or R₂-adrenocep-

tor affinities were reported.¹⁷A preliminary comparative

molecular field analysis (CoMFA) study on PNMT-active

ligands has been reported by our laboratory.¹⁸On the

basis of an active site model for PNMT, 30 benzylamine-

type PNMT inhibitors were aligned and subjected to the

CoMFA technique. A reasonably predictive model was

established by the statistics of the analysis. Several

examples^19,20have been reported in which a combined

approach of traditional QSAR and CoMFA techniques

has led to the development of a meaningful QSAR.

Based on these literature SAR results, it appeared

that optimal potency and selectivity for the PNMT

active site, as opposed to the R₂-adrenoceptor, is associ-

ated, at least in part, with the 7-substituent on the

THIQ ring. Hence, an investigation of the dependence

of PNMT inhibitory activity and R₂-adrenoceptor affinity

on the 7-substituent on THIQ (4) was undertaken. Some

7-substituted-THIQs (such as 5-7) had been previously

evaluated,²¹and additional 7-substituents were chosen

for the variation in their electronic and steric properties

(8-33). Results from previous studies²¹indicated that

the acidity of the aminosulfonyl group in 2 is not

responsible for the observed selectivity for PNMT vs the

R₂-adrenoceptor of this compound but that some other

factor, possibly having to do with the presence of the

sulfonyl group when attached to the aromatic ring,

might be responsible. The possibility of the dependence

of potency and selectivity on the sulfonyl moiety in the

aminosulfonyl group²¹prompted the inclusion of 17-

19 and 21-23.

activity with the Hammett σ and the hydrophobic

constant π was observed. Benzylamines bearing a meta-

substituent were generally more potent than the cor-

responding ortho- or para-substituted benzylamines. In

this quantitative structure-activity relationship (QSAR)

study, only a narrow range of lipophilic substituents,

such as chloro, bromo, iodo, fluoro, trifluoromethyl, and

methyl, were used. The effect of the aromatic substitu-

ent on the pI₅₀indicated that lowering the pK_aincreased

the potency of benzylamines, and it was proposed that

benzylamines bind to the active site of PNMT in the

nonprotonated form.¹⁰Using quantum chemical indices

derived from semiempirical molecular orbital calcula-

tions, Otto et al.¹¹examined the same data set of

benzylamines studied earlier by Fuller. This study

indicated a good correlation between the PNMT inhibi-

tory activity of benzylamines and the composite variable

that accounts for the electrostatic interaction between

the enzyme and the inhibitor. The pI₅₀values for nine

new mono- and disubstituted-benzylamines were pre-

dicted with the derived QSAR equation. Subsequently,

Lukovits¹²applied the method of principal component

analysis to the QSAR of benzylamines using the quan-

tum chemical indices derived by Otto. Results from

Lukovits’ analysis suggested that the PNMT inhibitory

activity of benzylamines was dependent on charge-

transfer phenomena, as well as on electrostatic forces.

Studies by Mercier et al.¹³utilized molecular topology

and molecular electronic structure-based QSAR meth-

ods to correlate a set of 22 of the benzylamines which

had been previously evaluated by Fuller. Excellent

correlations were obtained with these methods, and the

PNMT inhibitory activity of putatively more active

benzylamines was predicted using their QSAR equation.

THIQs contain a constrained benzylamine moiety,

and such constraint has been shown to increase affinity

for the PNMT active site.¹⁴We therefore decided to

extend these studies on benzylamines with a series of

carefully designed THIQ analogues.

A classical regression analysis and a three-dimen-

sional QSAR using the CoMFA technique were em-

ployed to sort out the differences in binding require-

ments for 7-substituted-THIQs at the PNMT active site

and the R₂-adrenoceptor. The CoMFA technique can

accommodate variation in the basic skeleton and is

independent of steric, electronic, and hydrophobic pa-

rameters for substituents. Therefore, in contrast to a

traditional QSAR, there are no restrictions in the choice

of substituents in a CoMFA.²²Such a study may lead

to the identification and characterization of the steric

and electronic factors on which biological activity and

selectivity (PNMT vs R₂-adrenoceptor affinity) are based.

Ch em istr y

The chemistry used in the synthesis of 7-substituted-

THIQs was developed with the aim of synthesizing most

of the desired analogues from a key intermediate.

Electrophilic substitution reactions on THIQ occur

mostly at the 7-position. Nitration,^23,24chlorosulfona-

tion,^9,25and Friedel-Crafts acetylation²⁶are some of the

known literature examples of electrophilic substitution

reactions on THIQ. In acidic media, the protonated

amine of THIQ directs the electrophile to the 5- or

7-positionsmeta with respect to the protonated benzyl-

amine moiety.²⁷More of the 7-substituted product is

obtained, probably due to peri hydrogen interaction at

the 5-position which disfavors electrophilic substitution

at that position.

For aromatic chloro-substituted THIQs, it was estab-

lished by Kaiser et al.¹⁵that a chlorine at the 7-position

is optimal as compared to the 5-, 6-, or 8-position and

that the 7,8-dichloro substitution pattern (as in 1)

resulted in maximum inhibition of PNMT. The effect of

THIQ with aromatic chloro substituents on PNMT

inhibitory activity was quantified with the Fujita-Ban

approach by Singh,¹⁶using data published by the SK&F

group.¹⁵The QSAR analysis indicated a negative con-

tribution from a chloro substituent at the 6-position on

the THIQ nucleus. The 7-chloro substituent had the

highest positive contribution to potency in comparison

with chlorine atoms at the 5- or 8-position, which appear

A key intermediate, 7-nitro-THIQ (8),²³was obtained

by nitration of 4 and was converted to another central

120 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 1

Grunewald et al.

Sch em e 1

Sch em e 3

Sch em e 2

Sch em e 4

species. Synthesis of 17 from 16 required that the THIQ

nitrogen be protected with a triphenylmethyl (trityl)

group to yield 34. Metalation of 34 followed by overnight

treatment with methyl disulfide and then with 3 N HCl

resulted in 17. Oxidation of 17 to sulfoxide 18 and

sulfone 19 was accomplished with trifluoroperacetic acid

in trifluoroacetic acid.³²Acidic conditions avoided N-

oxidation of the secondary aliphatic amine to the N-

oxide. Similarly, using a halogen-metal exchange re-

action, followed by trapping the lithiated species with

ethyl trifluoroacetate,³³resulted in the formation of the

trifluoroacetyl derivative 20. This compound was char-

acterized as its hydrochloride salt because the free base

underwent rapid polymerization. The infrared spectrum

and elemental analysis data of 20 indicated the presence

of the hydrated form of the ketone.

The syntheses of other sulfones were based on the

addition of a Grignard reagent to a sulfonyl fluoride.³⁴

Chlorosulfonation of N-acetyl-THIQ (35)⁹gave the

7-chlorosulfonyl derivative 36⁹which was converted to

fluorosulfonyl 37 under phase-transfer conditions³⁵in

the presence of potassium fluoride and 18-crown-6 ether

(Scheme 3). Reaction of the fluorosulfonyl with an excess

of the desired Grignard reagent gave the corresponding

sulfone (21 or 22) with concomitant removal of the acetyl

group.

intermediate, 7-cyano-THIQ (10), by reduction followed

by diazotization.²¹Nitrile 10 was readily converted to

methoxycarbonyl 11, hydroxymethyl 12, aminomethyl

13, carboxylic acid 14, and carboxamide 15 (Scheme 1).

The conversion of nitrile 10 to the carboxamide deriva-

tive 15 without protection of the THIQ nitrogen neces-

sitated the use of potassium fluoride supported on

alumina.²⁸

Halogen-metal exchange on 7-bromo-THIQ (16) pro-

vided an easy access to a variety of aromatic substitu-

ents which could not be prepared by direct electrophilic

substitution. The Sandmeyer bromination reaction was

used to synthesize 16 from the aromatic amine 9.

Because the Sandmeyer bromination is performed under

acidic conditions, the secondary aliphatic amine in 9

does not require protection as it will be protonated and

thus be unreactive.²⁹Using a modified procedure for the

Sandmeyer bromination of 7-amino-3,4-dihydroisoquin-

oline,³⁰a 73% yield of 16 was obtained after Kugelrohr

distillation (Scheme 2).

In the attempted halogen-metal exchange reaction

on 16 to produce 17, an additional equivalent of n-BuLi

was used to remove the acidic hydrogen from the

secondary amine and the resulting 7-lithiated species

was trapped with methyl disulfide.³¹The N-butyl-7-

thiomethyl-THIQ isolated after the workup resulted

from the reaction of butyl bromide formed in the

halogen-metal exchange reaction with the N-lithiated

Trichloromethylsulfonyl-THIQ (23) was prepared from

38 (Scheme 4). Deprotonation of the acidic methane-

sulfonyl group with 1 equiv of n-BuLi resulted in a

monoanion which was treated with an excess of carbon

tetrachloride.³⁶The resulting monochlorosulfonesbeing

more acidic than the trichloromethyl anionswas depro-

tonated and chlorinated. This sequence was repeated

until all the acidic hydrogen atoms were replaced by

chlorine atoms.

7-Iodo-THIQ (24) was required to synthesize 7-tri-

fluoromethyl-THIQ (25). Iodination of THIQ with I₂/Ag₂-

PNMT Inhibitors: Classical QSAR and CoMFA

J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 1 121

Sch em e 5

Benzoylation under identical conditions provided a

single regioisomer (45), the regiochemistry of which was

confirmed by a one-dimensional difference nuclear

Overhauser effect (NOE) experiment. Compound 45 was

deprotected under acidic conditions to obtain benzoyl

derivative 27, and 45 was also hydrogenated at 50 psi

under acidic conditions, followed by the removal of the

trifluoroacetyl group, to afford benzyl derivative 28.

Three additional derivativess7-acetyl-THIQ²⁶(29),

7-methyl-THIQ⁴⁰(30), and 7-N-(p-chlorophenyl)-7-ami-

nosulfonyl-THIQ⁴¹(31)swere prepared according to

literature procedures. 7-Hydroxy-THIQ (32) and 7-meth-

oxy-THIQ (33) were available from an earlier study

conducted in our laboratory.⁴²

Bioch em istr y

All compounds were evaluated as their hydrochloride

or oxalate salts for their activity as PNMT inhibitors.

Bovine adrenal PNMT,⁴³which had been purified ac-

cording to the method of Connett and Kirshner through

the isoelectric precipitation step,⁴⁴was used. In vitro

activity was assessed by use of a standard radiochemical

assay that has been previously described.⁴⁵Inhibition

constants in this investigation were determined by using

three different concentrations of the inhibitor with

phenylethanolamine as the substrate.

Sch em e 6

Sch em e 7

R₂-Adrenoceptor binding assays were performed using

cortex obtained from male SpraguesDawley rats.⁴⁶[³H]-

Clonidine was used as the radioligand to define the

specific binding and phentolamine was used to define

the nonspecific binding. Clonidine was used as the

ligand to define R-adrenergic binding affinity to simplify

the comparison with previous results.

Resu lts a n d Discu ssion

The results of in vitro biochemical evaluation of the

compounds in this study are presented in Table 1. For

comparison purposes, those THIQ derivatives which had

been reported earlier^21,42are also included. All THIQs

competitively inhibited the binding of the substrate

(phenylethanolamine), which was indicative of binding

at the PNMT active site. For an inhibitor to be highly

selective, it should have high PNMT inhibitory affinity

but low R₂-adrenoceptor affinity.

SO₄/H₂SO₄was reported to give mainly a mixture of 6-

and 7-iodo-THIQ and a minor amount of 8-iodo-THIQ.³⁷

Diazotization of 7-amino-THIQ (9), followed by addition

to aqueous KI, gave rise to multiple products and black

tar. Use of biphasic conditions³⁸to reduce tar formation

failed to provide reasonable yields of 24. Hence, 7-amino-

N-acetyl-THIQ (39) was prepared in an analogous

manner to the corresponding N-trifluoroacetyl deriva-

tive.²¹Diazotization and subsequent treatment with KI

and a catalytic amount of CuI under biphasic conditions

(CHCl₃/H₂O) gave 40 in 74% yield (Scheme 5). Trifluo-

romethylation of 40 was accomplished by a novel

procedure³⁹which used methyl chlorodifluoroacetate in

the presence of CuI and KF as the trifluoromethylating

reagent.

7-Acetamido-THIQ (26) was made in a straightfor-

ward manner from 42, which was synthesized as

described previously.²¹The greater hydrolytic suscep-

tibility of the aliphatic trifluoroacetamide as compared

to the aromatic acetamide led to the formation of 26 on

treatment of 43 with K₂CO₃in MeOH (Scheme 6).

The synthesis of 7-benzoyl-THIQ (27) (Scheme 7) was

based on the literature precedence of Friedel-Crafts

acetylation of N-trifluoroacetyl-THIQ (44) to afford a

single regioisomer, 7-acetyl-THIQ (29).²⁶A complex is

probably formed between AlCl₃and the trifluoroacetyl

group which directs the acylation at the 7-position.

A. Qu a lita tive Con sid er a tion s. Examination of

Tables 1 and 2 revealed the following trends in SAR in

7-substituted-THIQs:

(1) There is no obvious correlation between substitu-

ent lipophilicity and PNMT affinity (compare 24 and

2). However, as the electron-withdrawing ability of the

substituent increased, the PNMT potency also in-

creased. This was evident from the trifluoromethyl

analogue 25 which was about 15-fold more active than

the methyl analogue 30, and this trend is further

supported by other examples in the table (e.g., 27 as

compared to 28).

(2) A precipitous drop in the PNMT inhibitory activity

of the aminomethyl derivative 13 as compared to the

hydroxymethyl analogue 12 is noted. This molecule

actually contains two benzylamine moieties: the con-

formationally restricted benzylamine of the THIQ nucleus

and the conformationally flexible benzylamine compris-

ing the 7-aminomethyl substituent. Thus, it is not clear

how this molecule would bind at the active site of

122 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 1

Grunewald et al.

Ta ble 1. In Vitro Activities of

Ta ble 2. Parameters for 7-Substituted-THIQs Used in the

QSAR Study

7-Substituted-1,2,3,4-tetrahydroisoquinolines as Inhibitors of

PNMT and Binding of [³H]Clonidine at the R₂-Adrenoceptor

PNMT R₂

(A) PNMT K_i

(B) R₂K_i

B/A

compd

R

pK_i

4.00 -1.82

6.45 0.00

π

σ_m

0.53 12.28

0.00 1.03

0.20 18.17

0.47 31.20

MR

compd

R

( SEM (µM) ( SEM (µM) selectivity

2

4

5

6

7

SO₂NH₂

H

NHSO₂Me

N(SO₂Me)₂

CH₂NHSO₂Me

NO₂

6.26

5.01

4.32

2.35

3.48

6.38

4.57

5.87

5.17

4.97

2.72

3.33

4.19

6.53

6.21

4.35

5.87

5.24

4.85

6.43

6.74

3.73

5.54

4.21

5.31

5.57

5.58

4.70

2

4

5

6

7

SO₂NH₂

H

0.55 ( 0.04^a100 ( 20^a

180

0.036

0.23

0.042

0.055

10

0.11

5.2

0.68

0.21

0.0063

1.0

0.36

0.79

0.67

1.1

120

5.6

1.5

9.7

26

0.59

7.2

0.48

9.3

0.077

1.2

0.13

2.4

0.85

0.022

9.7 ( 0.4^b

48 ( 2^a

0.35 ( 0.11^b

11 ( 0.1^a

190 ( 10^a

18 ( 1^a

4.3 ( 0.3

3.1 ( 0.1

7.3 ( 0.2

4.6 ( 0.2

2.3 ( 0.3

12 ( 1

470 ( 10

23 ( 1

0.23 ( 0.13

0.41 ( 0.05

51 ( 1

4.96 -1.18

NHSO₂Me^a

3.72 -1.51

a

N(SO₂Me)₂

∼4500^a

4.74 -1.23^a-0.10^a22.60^a

CH₂NHSO₂Me^a

330 ( 10^a

0.41 ( 0.05

27 ( 3

c

8

9

5.37 -0.28

0.71

7.36

5.42

6.33

8

9

NO₂

c

NH₂

CN

5.51 -1.23 -0.16

NH₂

10

11

12

13

14

15

16

17

18

19

20

21

24

25

26

27

28

29

30

32

33

5.14 -0.57

5.34 -0.01

5.64 -1.03

4.92 -4.09

0.56

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

CN

CO₂Me

CH₂OH

1.4 ( 0.1

6.7 ( 0.3

11 ( 1

1900 ( 100

470 ( 30

64 ( 3

0.29 ( 0.03

0.61 ( 0.05

45 ( 2

1.3 ( 0.1

5.7 ( 0.3

14 ( 1

9.8 ( 1.0

1.3 ( 0.1

0.37 ( 0.04

0.18 ( 0.03

190 ( 10

2.9 ( 0.1

61 ( 3

CO₂Me

CH₂OH

CH₂_-NH₃

0.36 12.87

0.00 7.19

0.32 10.10^a

+ d

+

CH₂NH₃

- d

CO₂

3.33 -4.36 -0.10

6.05

9.81

8.88

CONH₂

Br

SMe

CONH₂

Br

4.64 -1.49

0.28

0.39

6.64

6.39

0.86

0.61

SMe

0.15 13.82

0.52 13.70

0.60 13.49

0.63 11.20

0.62 33.2

0.35 13.94

SOMe

SO₂Me

COCF₃

SO₂Ph

SO₂CH₂CHdCH₂

SO₂CCl₃

I

SOMe

SO₂Me

COCF₃

SO₂Ph

I

4.29 -1.58

3.79 -1.63

4.49

4.67

6.66

5.89

5.04 -0.97

4.57

5.33

5.21 -0.55

6.44

160 ( 10

32 ( 1

0.02

0.27

1.12

0.88

21 ( 1

95 ( 2

34 ( 1

CF₃

0.43

5.02

0.22 ( 0.04

1.3 ( 0.1

9.1 ( 0.3

27 ( 2

NHCOMe

COPh

CH₂Ph

COCH₃

CH₃

0.21 14.93

0.34 30.33

CF₃

1.05

NHCOMe

COPh

2.01 -0.08 30.01

0.38 11.18

5.65

CH₂Ph

COMe^e

Me^f

4.7 ( 0.4

6.1 ( 0.2

0.36 ( 0.06

6.3 ( 0.5

2.2 ( 0.2

0.47 ( 0.04

0.56 -0.07

4.9 ( 0.2

2.7 ( 0.1

OH

OMe

5.65 -0.67

6.33 -0.02

0.12

2.85

7.87

SO₂NH-C₆H₄-p-Cl^g2.6 ( 0.2

OH^h

2.6 ( 0.1^h

21 ( 2^h

a

Values obtained from the Pomona College Medicinal Chem-

OMe^h

istry database.

Reference 21. Reference 47. ^cReference 23. The moiety is

a

b

d

ionized at the pH of the assay [pH ) 8.0 (PNMT) or 7.7 (R₂)].

ceptor affinity of 19 was lower than that of 2. This might

be a reflection of the better electron-withdrawing capa-

bility of the methylsulfonyl group in comparison with

the aminosulfonyl group (see Table 2 for σ_mvalues). The

good selectivity (PNMT vs the R₂-adrenoceptor) of 19

prompted us to synthesize and evaluate more THIQ

7-sulfones. Changing the substituent from methyl to

phenyl, allyl, or trichloromethyl, as in 21, 22, or 23,

resulted in a drop in selectivity. For phenylsulfonyl 21,

the PNMT K_iincreased and the R₂-adrenoceptor K_i

decreased. In the case of allylsulfonyl 22, the low

selectivity resulted from a reduction in PNMT inhibitory

potency. The PNMT inhibitory activity of 19 was

retained by trichloromethylsulfonyl 23, but the R₂-

adrenoceptor affinity increased by about 5-fold. Anili-

nosulfonyl 31 was reported to be as potent as 2 in

inhibiting PNMT, but no R₂-adrenoceptor affinity was

reported.⁴¹In our assays, 31 proved to be 5-fold less

potent at PNMT and had enhanced R₂-adrenoceptor

affinity as compared to 2.

B. Developm en t of an Active Site Bin din g Model.

A binding model for 7-substituted-THIQs at the PNMT

active site and the R₂-adrenoceptor was developed based

on the SAR data discussed above and on previous SAR

studies on benzylamines.^18,48As both lipophilic and

hydrophilic substituents are acceptable at both the

PNMT and the R₂-adrenoceptor binding sites, it is

unlikely that they bind in the same orientation. If the

aromatic ring and the lone pair on the nitrogen form

the essential interacting points of the THIQ pharma-

cophore at the two binding sites, then the THIQ nucleus

g

h

^eReference 26. ^fReference 40. Reference 41. Reference 42.

PNMT. On the other hand, the binding requirements

at the R₂-adrenoceptor seemed to be less restrictive than

those at the PNMT active site, since 12 and 13 have

comparable R₂-adrenoceptor affinities.

(3) The affinity toward the R₂-adrenoceptor increased

as the lipophilicity of the 7-substituent increased. Thus,

7-halo-substituted-THIQs 16 and 24 had high affinity

toward the R₂-adrenoceptor. A 13-fold loss of affinity of

the benzyl analogue 28 with respect to the methyl

analogue 30 suggested the presence of a region of

limited steric bulk tolerance around the 7-position of

lipophilic THIQs at the R₂-adrenoceptor. Introduction

of electron-withdrawing substituents such as NO₂, CN,

and CF₃resulted in lowered affinity toward the R₂-

adrenoceptor and therefore increased selectivity toward

PNMT vs the R₂-adrenoceptor.

(4) Enhancement of the electron-withdrawing capabil-

ity of the 7-sulfur substituent on THIQ diminished the

affinity toward the R₂-adrenoceptor. Changing the oxi-

dation state of the thiomethyl group in 17 to the methyl

sulfoxide 18 led to a 125-fold loss in binding affinity.

Because compound 18 was tested as a racemate, the

stereoselectivity in binding of the methylsulfinyl group

in 18 was not probed. The sulfonyl group in the

aminosulfonyl moiety of 2 seemed to be responsible for

its lowered affinity toward the R₂-adrenoceptor.²¹In

contrast to the methylsulfonyl 19, the increased selec-

tivity of 2 was the consequence of its higher potency at

PNMT. It was interesting to note that the R₂-adreno-

PNMT Inhibitors: Classical QSAR and CoMFA

J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 1 123

F igu r e 1. (a) Two possible orientations of 7-substituted-THIQs in which optimal interaction of the aromatic ring and lone pair

on the nitrogen can occur. Orientation A is proposed for lipophilic (+π) substituents, while hydrophilic (-π) substituents are

proposed to bind in orientation B. (b) SYBYL-generated stereoview of 7-bromo-THIQ (16) in orientation A, superimposed on 2 in

orientation B, showing that the THIQ ring nitrogen lone pairs can reach the same area. Lipophilic and hydrophilic substituents

occupy different regions of space when fitted according to this hypothesis. The optimized geometries (Tripos force field) of 2 and

16 were fitted using three points: both ends of a normal (2 Å long) passing through the centroid of the THIQ aromatic ring and

the end of the axial lone pair (2.4 Å long) on the THIQ nitrogen. Hydrogens were omitted for clarity.

Ta ble 3. One-, Two-, and Three-Parameter Equations for the QSAR of PNMT (Eq 1)

equations^a

r

s

F_calc

F_tab

one parameter

0.480((0.266)π + 5.27((0.41)^b

F_(1,26)

13.78

3.34

F_(1,26,0.01)

7.72

F_(2,25,0.01)

5.57

F_(3,24,0.01)

4.72

0.589

0.337

0.341

0.957

1.11

-0.0444((0.0499)MR + 5.56((0.78)

1.54((1.71)σ_m+ 4.55((0.64)

3.41

two parameter

F_(2,25)

16.87

9.93

4.77

F_(3,24)

21.43

0.567((0.224)π - 0.0643((0.0362)MR + 6.16((0.60)

0.466((0.251)π + 1.40((1.39)σ_m+ 4.87((0.55)

1.84((1.60)σ_m- 0.0.534((0.0468)MR + 5.17((0.80)

0.758

0.665

0.526

0.787

0.901

1.03

three parameter

0.561((0.183)π - 0.0729((0.0300)MR + 1.79((1.01)σ_m+ 5.77((0.54)

0.853

0.642

a

b

n ) 28. Numbers in parentheses are 95% confidence intervals.

can be oriented in two possible ways as shown in Figure

in Table 2 were used in the initial regression analysis

with a number of QSAR parameters.^50aRegression

analyses and statistical calculations were done using

the computer program QSAR-PC.^50bTable 2 lists the

PNMT pK_i[experimental (-log K_i) value for PNMT

inhibition] and the R₂pK_i[experimental (-log K_i) for

the R₂-adrenoceptor affinity] values along with the

substituent parameters⁵¹for the 7-substituted-THIQs

used in the development of eqs 1 and 3.

QSAR of P NMT Liga n d s. Of the nine variables

surveyed,^50athe most significant single-variable equa-

tions for hydrophobic, steric, or electronic effects were

those involving π, MR, or σ_m, respectively. As shown in

Table 3, the most significant single-variable equation

contained the π term. Stepwise analysis showed that

the best (statistically significant) two-parameter equa-

tion was obtained with π and MR (F-test for the

difference between equations: F_(1,25)) 13.45 vs F_(1,25,0.01)

) 7.77) and the best three-parameter equation with π,

MR, and σ_m(F_(1,24)) 13.57 vs F_(1,24,0.01)) 7.82), with

the statistical significance after addition of a new

parameter assessed by the F-test at the 99% confidence

level. Thus, eq 1 was derived:

1. This model would allow the binding of both lipophilic

and hydrophilic substituents in different regions of

space at either active site, and as shown in Figure 1b,

the axial lone pairs from the THIQ nitrogen in either

orientation could reach a common region of space. [See

Development of the CoMFA Models (in section D)

below.]

The qualitative examination of data presented on the

R₂-adrenoceptor affinity of 7-substituted-THIQs, as well

as limited SAR studies on THIQ-type PNMT inhibitors,

has indicated that an identical binding model may hold

true for these ligands at the R₂-adrenoceptor. As shown

in the sections below, such a simplistic unified model

can be used to explain the literature QSAR results and

can provide a guideline for the alignment of these

molecules for a CoMFA study.

C. Develop m en t of th e QSAR Equ a tion s. A clas-

sical Hansch multivariate regression analysis was used

to derive QSAR equations for our data set. The following

substituents from Table 1 were not included in the

Hansch analysis as their physicochemical parameters

were unavailable: SO₂CH₂CHdCH₂(22), SO₂CCl₃(23),

and SO₂NH-C₆H₄-p-Cl (31). Parameter values for

compound 20 (COCF₃) were included in both analyses,

although the experimental evidence indicated the pres-

ence of the hydrated form of the ketone. Under the assay

conditions (phosphate buffer at pH 8 or Tris buffer at

pH 7.7), the carboxylic acid 14 is expected to be ionized

and the amine of 13 protonated. Thus, the parameters

for the ionized states of those groups were used in the

QSAR analyses.⁴⁹Methyl ester 11 was determined to

be stable under the assay conditions by NMR analysis

(see Experimental Section). All of the compounds listed

pK_i) 0.561((0.183)π - 0.0729((0.0300)MR + 1.79

((1.01)σ_m+ 5.77((0.54) (1)

n ) 28, r ) 0.853, s ) 0.642,

F_(3,24)) 21.43, F_(3,24,0.01)) 4.72

In Table 3, all single-parameter (π, MR, and σ_m) and

all combinations of two-parameter equations used in the

development of the three-parameter equation (eq 1) are

provided.

124 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 1

Grunewald et al.

Ta ble 4. Correlation Matrix for the Parameters of the 28

Compounds Used in Developing QSAR Eqs 1 and 3

R₂pK_i) 0.482((0.154)π - 0.0573((0.0248)MR +

6.22((0.41) (3)

PNMT

π

σ_m

MR

1

n ) 28, r ) 0.826, s ) 0.539,

F_(2,25)) 26.77, F_(2,25,0.01)) 5.57

π

σ_m

MR

1

0.038

0.217

1

0.177

In Table 7, all single-parameter (π, MR, and σ_m) and

all combinations of two-parameter equations, along with

the equation containing all three parameters, are

provided. The single-parameter equation in π was found

to be most significant, explaining 39% of the variance

in the data. The stepwise development of eq 3 showed

that only the two-parameter equation derived by the

addition of the MR parameter was statistically justified

A correlation matrix (Table 4) showed that no strong

correlations exist between the parameters. While π is

orthogonal with respect to σ_m, some correlation between

π and MR for the substituents chosen was observed.

Most of the bulk parameters (including MR) show some

correlation with hydrophobicity.⁵²Analysis of the cal-

culated pK_ivalues indicated that aminosulfonyl 2 had

a residual greater than 2 standard deviations (residual

) 1.45). Thus, 2 was dropped from the regression, and

eq 2 was derived:

(F-test for the difference between equations: F_(1,25)

)

22.69 vs F_(1,25,0.01)) 7.77), and it was surprising to note

that σ_mwas not statistically significant in eq 3.

The residual for the aminomethyl compound 13 was

greater than 2 standard deviations (residual ) 1.24).

Thus, 13 was dropped from the regression, and eq 4 was

derived:

pK_i) 0.599((0.167)π - 0.0725((0.0268)MR +

1.55((0.917)σ_m+ 5.80((0.48) (2)

n ) 27, r ) 0.885, s ) 0.573,

F_(3,23)) 27.61, F_(3,23,0.01)) 4.76

R₂pK_i) 0.599((0.129)π - 0.0542((0.0186)MR -

0.951((0.623)σ_m+ 6.45((0.34) (4)

In Table 5, all single-parameter (π, MR, and σ_m) and

all combinations of two-parameter equations, along with

the three-parameter equation (eq 2), are provided.^53a

The stepwise development of eq 2 showed that the best

two-parameter equation contained π and MR (F-test for

the difference between equations: F_(1,24)) 17.69 vs

F_(1,24,0.01)) 7.82) and the best three-parameter equation

n ) 27, r ) 0.917, s ) 0.397,

F_(3,23)) 40.53, F_(3,23,0.01)) 4.76

In Table 8, all single-parameter (π, MR, and σ_m) and

all combinations of two-parameter equations, along with

the three-parameter equation (eq 4), are provided.^53b

The hydrophobic constant (π) was the most important,

as its single-parameter equation explained 47% of the

variance in the data. The stepwise development of eq 4

showed that the best two-parameter equation contained

π and MR (F-test for the difference between equations:

F_(1,24)) 31.85 vs F_(1,24,0.01)) 7.82) and the best three-

contained π, MR, and σ_m(F_(1,23)) 12.26 vs F_(1,23,0.01)

)

7.88). Values of PNMT pK_iwere calculated according

to eq 2, and the residuals are presented in Table 6.

Equation 2 was in general agreement with the

hydrophobic and electronic effects inferred from qualita-

tive considerations and accounted for about 75% of the

variance in the data. It is surprising that the equation

underpredicted the more active aminosulfonyl (2) and

methylsulfonyl (19) compounds, although the latter was

predicted within 2 standard deviations. However, previ-

ous studies²¹had indicated that there was some type

of interaction that might be responsible for the enhanced

potency of these compounds.

parameter equation contained π, MR, and σ_m(F_(1,23)

)

9.93 vs F_(1,23,0.01)) 7.88). It is interesting to note that

with the removal of compound 13 from the regression,

σ_mbecame a statistically significant term in the new

equation. The R₂-adrenoceptor pK_ivalues were calcu-

lated according to eq 4, and the residuals are presented

in Table 6.

Equation 2 was quite similar to that obtained by

Fuller et al.¹⁰for meta-substituted benzylamines in

terms of the positive contributions of π and σ to PNMT

inhibitory activity. The large positive coefficient associ-

ated with σ_mindicates that electron withdrawal by the

substituent leads to a favorable interaction, probably

via formation of a charge-transfer complex as suggested

by Lukovits,¹²between an electron-rich region at the

PNMT active site and the electron-deficient benzene

ring in THIQ. A negative correlation with MR suggests

that steric hindrance was experienced by the 7-sub-

stituent in binding at the PNMT active site or that

repulsive van der Waals forces dominate⁵²in the region

around the 7-substituent at the PNMT active site.

QSAR of th e r₂-Ad r en ocep tor Liga n d s. An ap-

proach identical to that used in the development of the

QSAR for PNMT inhibitors was applied to the develop-

ment of a QSAR for the same ligands at the R₂-

adrenoceptor. However, among π, MR, and σ_m, only π

and MR were significant in the initial equation (Table

7):

Com p a r ison of QSAR Equ a tion s. The QSAR equa-

tions of PNMT (eq 2) and the R₂-adrenoceptor (eq 4) for

7-substituted-THIQs differed mainly in terms of the sign

of the coefficient of the electronic parameter σ_m(+1.55

for PNMT vs -0.951 for the R₂-adrenoceptor). The

coefficient of π in both eqs 2 and 4 (0.599 vs 0.599) was

the same and suggested that hydrophobic interactions

play a similar role in increasing affinity at both sites.

The coefficient of MR in the R₂-adrenoceptor equation

(0.0542, eq 4) was less than that in the PNMT equation

(0.0725), which would indicate a somewhat greater

amount of steric bulk tolerance at the R₂-adrenoceptor

site. The opposite signs on the coefficients of σ_m(+1.55

for PNMT, -0.951 for R₂) indicate the principal differ-

ence between the binding requirements at the PNMT

active site and at the R₂-adrenoceptor. Thus, THIQs

which are more selective for PNMT vs the R₂-adreno-

ceptor would require a 7-substituent with a large

positive σ_mvalue, of which aminosulfonyl 2 is one such

PNMT Inhibitors: Classical QSAR and CoMFA

J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 1 125

Ta ble 5. One-, Two-, and Three-Parameter Equations for the QSAR of PNMT (Eq 2)

equations^a

r

s

F_calc

F_tab

one parameter

F_(1,25)

18.19

3.31

F_(1,25,0.01)

7.77

F_(2,24,0.01)

5.61

F_(3,23,0.01)

4.76

0.524((0.253)π + 5.23((0.38)^b

0.649

0.342

0.312

0.897

1.11

1.12

-0.0439((0.0497)MR + 5.51((0.79)

1.40((1.76)σ_m+ 4.56((0.65)

two parameter

0.613((0.201)π - 0.0652((0.0320)MR + 6.13 ((0.53)

0.506((0.244)π + 1.16((1.36)σ_m+ 4.91((0.53)

1.72((1.65)σ_m- 0.0525((0.0473)MR + 5.16((0.81)

three parameter

2.70

F_(2,24)

24.01

11.40

4.21

F_(3,23)

27.61

0.817

0.698

0.509

0.694

0.861

1.03

0.599((0.167)π - 0.0725((0.0268)MR + 1.55((0.917)σ_m+ 5.80((0.48)

0.885

0.573

a

b

n ) 27. Numbers in parentheses are 95% confidence intervals.

Ta ble 6. Experimental and Calculated PNMT pK_i(Eq 2) and

R₂pK_i(Eq 4) Values of 7-Substituted-THIQs

7-substituents in the same region of space), significantly

lower cross-validated r²values were obtained for either

the PNMT or the R₂-adrenoceptor models than when the

molecules were aligned in orientations A or B, as

described in the legend for Figure 1a. The orientations

of the 30 compounds used in the PNMT and the

R₂-adrenoceptor CoMFA are listed in Table 9. Individual

consideration of either steric or electrostatic fields in

either the PNMT or R₂CoMFA resulted in almost 50%

lower cross-validated r²values as compared to the

analyses when both steric and electrostatic fields were

used.

As sulfoxide 18 was synthesized and evaluated as a

racemate, it was excluded from the data set to avoid

any ambiguities in the input of data and the interpreta-

tion of the model. As was mentioned previously, experi-

mental data indicated the presence of the hydrated form

of the ketone of trifluoroacetyl analogue 20, and so this

structure was used in the CoMFA.

CoMF A of 7-Su bstitu ted -THIQs for P NMT. As

was done in the PNMT QSAR (eq 1), 13 and 14 (charged

forms) were included in the analysis; however, this

resulted in a negative cross-validated r²for the CoMFA.

Surprisingly, when the neutral forms of analogues 13

and 14 were used in the analysis, the cross-validated

r²improved to 0.477. As analogue 13 contains two

benzylamine moieties, either of which could theoreti-

cally bind at the active site, a check was made with the

compound “flipped” (Figure 2) so that the 7-position

benzylamine was in the area of space occupied by the

THIQ nitrogen. While this only improved the cross-

validated r²slightly to 0.546, the nonvalidated r²was

improved from 0.851 to 0.904 (Table 10) with three

components, with an increase in the F-test value from

49.64 to 81.84. Thus, the flipped orientation of analogue

13 was used in the development of the PNMT models.⁵⁵

The residuals of the nonvalidated analysis are presented

in Table 9, and all activity predictions were within the

95% confidence interval. In this model, the relative

contributions of electrostatic and steric potential to the

CoMFA regression equation indicated that electrostatic

forces dominate (Table 11). The CoMFA steric and

electrostatic contour maps for PNMT are shown in

Figure 3a,b, respectively.

PNMT pK_i

exptl calcd ∆pK_iexptl calcd ∆pK_i

6.26 4.64^a

1.62 4.00 4.19 -0.19

5.01 5.73 -0.72 6.45 6.40

4.32 4.09 0.23 4.96 4.57

R₂pKi

compd

R

2

4

5

6

7

SO₂NH₂

H

NHSO₂Me

N(SO₂Me)₂

0.05

0.39

0.31

0.15

0.16

2.35 3.36 -1.01 3.72 3.41

CH₂NHSO₂Me 3.48 3.27

0.21 4.74 4.59

0.18 5.37 5.21

0.15 5.51 5.58 -0.07

0.00 5.14 5.24 -0.10

8

9

NO₂

NH₂

CN

6.38 6.20

4.57 4.42

5.87 5.87

10

11

12

13

14

15

16

17

18

19

20

21

24

25

26

27

28

29

30

32

33

CO₂Me

CH₂OH

CH₂_-NH₃

CO₂

CONH₂

Br

5.17 5.42 -0.25 5.34 5.41 -0.07

4.97 4.66

2.72 3.12 -0.40 4.92 3.15^a

3.33 2.60

0.31 5.64 5.45

0.19

1.77

+

0.73 3.33 3.61 -0.28

4.19 4.63 -0.44 4.64 4.76 -0.12

6.53 6.28

6.21 5.40

4.35 4.67 -0.32 4.29 4.27

5.87 4.78

0.25 6.64 6.12

0.81 6.39 5.93

0.52

0.46

0.02

SMe

SOMe

SO₂Me

COCF₃

SO₂Ph

I

1.09 3.79 4.18 -0.39

5.24 5.98 -0.74 4.49 5.26 -0.77

4.85 4.52

6.43 6.01

6.74 6.63

0.33 4.67 4.23

0.42 6.66 6.04

0.44

0.62

CF₃

0.11 5.89 6.30 -0.41

0.17

0.78 4.57 5.12 -0.55

4.21 4.71 -0.50 5.33 6.11 -0.78

5.31 5.25 0.06 5.21 5.16 0.05

5.57 5.62 -0.05 6.44 6.55 -0.11

5.58 5.38 0.20 5.65 5.79 -0.14

4.70 5.40 -0.70 6.33 5.90 0.43

NHCOMe

COPh

CH₂Ph

COMe

Me

3.73 4.47 -0.74 5.04 4.87

5.54 4.76

OH

OMe

a

Compounds not included in the regression.

example.⁵⁴Increasing selectivity by manipulating π or

MR would have much less impact, given the similarity

of lipophilic effects in both QSAR equations and the

increased steric bulk tolerance at the R₂-adrenoceptor.

D. Com p a r a tive Molecu la r F ield An a lysis (CoM-

F A). CoMFA²²is a molecular modeling based three-

dimensional QSAR technique that can calculate nonco-

valent molecular (steric and electronic) interactions

around a set of aligned ligands and correlate variations

in these interactions with observed properties.

Develop m en t of th e CoMF A Mod els. Geometry

optimization and charge calculations for all of the

compounds used in the development of these CoMFA

models were done as detailed in the Experimental

Section.

CoMF A of 7-Su bstitu ted -THIQs for th e R₂-A-

d r en ocep tor . The alignment rule for compounds used

in the development of the R₂CoMFA model was the

same as that used in the PNMT model. It is a rather

simplistic hypothesis that 7-substituted-THIQs bind in

a similar fashion at both the PNMT active site and the

R₂-adrenoceptor. Nevertheless, a higher cross-validated

r²was obtained when the molecules were aligned in

The alignment of ligands is extremely important in a

CoMFA. In the absence of an alignment rule (i.e., when

all 7-substituted-THIQs were aligned in the same

fashion, with the THIQ nucleus superimposed and all

126 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 1

Grunewald et al.

Ta ble 7. One-, Two-, and Three-Parameter Equations for the QSAR of the R₂-Adrenoceptor (Eq 3)

equations^a

r

s

F_calc

F_tab

one parameter

0.404((0.203)π + 5.43((0.31)^b

F_(1,26)

16.84

4.59

F_(1,26,0.01)

7.72

F_(2,25,0.01)

5.57

F_(3,24,0.01)

4.72

0.627

0.387

0.294

0.730

0.864

0.896

-0.0404((0.0387)MR + 5.71((0.61)

-1.05((1.38)σ_m+ 5.48((0.52)

2.45

two parameter

F_(2,25)

26.77

12.50

3.20

F_(3,24)

22.89

0.482((0.154)π - 0.0573((0.0248)MR + 6.22((0.41)

0.416((0.188)π - 1.17((1.04)σ_m+ 5.76((0.41)

-0.840((1.33)σ_m- 0.0362((0.0388)MR + 5.89((0.67)

0.826

0.707

0.452

0.539

0.676

0.853

three parameter

0.485((0.142)π - 0.0531((0.0232)MR - 0.883((0.777)σ_m+ 6.42((0.42)

0.861

0.496

a

b

n ) 28. Numbers in parentheses are 95% confidence intervals.

Ta ble 8. One-, Two-, and Three-Parameter Equations for the QSAR of the R₂-Adrenoceptor (Eq 4)

equations^a

r

s

F_calc

F_tab

one parameter

0.503((0.220)π + 5.43((0.30)^b

F_(1,25)

22.09

4.56

F_(1,25,0.01)

7.77

F_(2,24,0.01)

5.61

F_(3,23,0.01)

4.76

0.685

0.393

0.293

0.695

0.878

0.913

-0.0410((0.0395)MR + 5.73((0.63)

-1.04((1.41)σ_m+ 5.48((0.53)

2.34

two parameter

F_(2,24)

40.65

17.20

3.13

F_(3,23)

40.53

0.591((0.151)π - 0.0587((0.0215)MR + 6.25((0.36)

0.523((0.199)π - 1.24((0.966)σ_m+ 5.78((0.38)

-0.828((1.36)σ_m- 0.0368((0.0397)MR + 5.91((0.68)

0.879

0.767

0.455

0.465

0.624

0.868

three parameter

0.599((0.129)π - 0.0542((0.0186)MR - 0.951((0.623)σ_m+ 6.45((0.34)

0.917

0.397

a

b

n ) 27. Numbers in parentheses are 95% confidence intervals.

Ta ble 9. CoMFA Alignment Rule and Residual Values for

7-Substituted-1,2,3,4-tetrahydroisoquinolines Calculated

According to the CoMFA Model at the Active Site of PNMT and

at the R₂-Adrenoceptor

residuals^a

CoMFA

compd

7-substituent R

SO₂NH₂

orientation

PNMT

R₂

F igu r e 2. Dual orientations of 7-aminomethyl-THIQ (13)

showing the molecule in the standard and “flipped” orienta-

tions.

2

4

5

6

7

B

A

B

A

B

A

B

A

B

A

B

0.32

-0.63

0.44

0.28

-0.28

0.57

-0.44

-0.28

0.07

-0.12

-0.73

-0.18

-0.62

0.38

0.03

0.25

-0.12

-0.01

-0.05

0.09

0.27

-0.45

0.51

0.04

-0.29

0.28

0.01

0.26

0.37

0.04

-0.11

0.25

-0.05

0.25

H

NHSO₂Me

N(SO₂Me)₂

CH₂NHSO₂Me

NO₂

NH₂

CN

CO₂Me

CH₂OH

CH₂NH₂

CO₂H

CONH₂

Br

Ta ble 10. Selected Statistical Results of the CoMFAs

(Cross-Validated Runs)

0.17

8

9

-0.26

-0.44

-0.07

0.21

-0.24

-0.35

-0.07

-0.21

0.23

-0.13

-0.31

-0.09

0.28

0.16

-0.34

0.29

-0.23

-0.08

-0.13

-0.05

0.12

analysis

n

no. of components

r²

press s

10

11

12

13

14

15

16

17

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

PNMT

R₂

30

1

2

3

1

2

3

0.260

0.413

0.546

0.598

0.682

0.694

0.989

0.897

0.804

0.585

0.530

R₂

SMe

Ta ble 11. Statistical Results of the CoMFAs (Nonvalidated

Runs, Three Components)

SO₂Me

COCF₃

b

SO₂Ph

SO₂CH₂CHdCH₂

SO₂CCl₃

I

CF₃

NHCOMe

COPh

CH₂Ph

COMe

Me

SO₂NH-C₆H₄-p-Cl

OH

conventional

contributions

analysis

n

r²

s

F value^asteric electrostatic

PNMT

R₂

30 0.904 0.369

30 0.915 0.279

81.84

93.70

0.396

0.428

0.604

0.572

a

The table F-value for F_(3,26,0.01)) 4.64.

forms of analogues 13 and 14 were used in the analysis,

the cross-validated r²improved to 0.686. The “flipped”

orientation (Figure 2) of compound 13 was also used,

which improved the cross-validated r²slightly to 0.694

and improved the nonvalidated r²from 0.852 to 0.915

(Table 10) with three components, with an increase in

the F-test value from 77.73 to 93.70.

0.13

0.41

-0.06

0.74

OMe

a

Negative logarithm of the difference between actual and

b

predicted values. Hydrated form of the ketone was used.

orientation A or B (Table 9), as opposed to a single

orientation in which all of the 7-substituents were

superimposed. As in the PNMT model, when the neutral

The predictive power of this CoMFA model was better

(cross-validated r²) 0.694, “press s” ) 0.530, optimal

number of components ) 3) than that of the PNMT

PNMT Inhibitors: Classical QSAR and CoMFA

J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 1 127

F igu r e 3. Contour maps for the PNMT and the R₂-adrenoceptor CoMFAs of 7-substituted-THIQs. SK&F 29661 (2) and compound

16 are superimposed in their proposed alignments as the reference molecules. For the steric maps, the contours where steric bulk

is positively correlated with activity are shown in green, while the contours where steric bulk is negatively correlated with activity

are shown in yellow. For the electrostatic maps, the contours where electron deficiency is correlated with activity are shown in

cyan, while the contours where electron density is correlated with activity are shown in magenta. (a) PNMT steric: the contours

are drawn at the contribution levels (positive:negative) of 85:30. (b) PNMT electrostatic: the contours are drawn at the contribution

levels (positive:negative) of 85:25. (c) R₂steric: the contours are drawn at the contribution levels (positive:negative) of 90:30. (d)

R₂electrostatic: the contours are drawn at the contribution levels (positive:negative) of 70:5.

CoMFA model (cross-validated r²) 0.546, “press s” )

0.804, optimal number of components ) 3). The CoMFA

contour maps, generated after a nonvalidated analysis

using three components, are presented in Figure 3c,d.

In contrast to the PNMT CoMFA, in which all activities

were predicted within the 95% confidence interval, the

predicted activity of methoxy analogue 33 was outside

the 95% confidence interval (residual ) 0.74).

philic substituents were projected. There appeared to

be more steric bulk intolerance in the region of the

nonlipophilic 7-substituents (orientation B).

The sign of the coefficient of the electronic parameter

(σ_m) was the only significant difference between QSAR

eqs 2 and 4. This difference in the electronic require-

ments at the PNMT active site and at the R₂-adreno-

ceptor can be readily seen in the electrostatic contour

maps for the ligands in the PNMT and the R₂-adreno-

ceptor CoMFAs (Figure 3b,d). Large contours represent-

ing an area where electron density is favored were

associated with nonlipophilic substituents in the CoM-

FA for PNMT, and indicated that an increase in nega-

tive charge in this region was positively correlated with

PNMT inhibitory potency. Electron-deficient substitu-

ents will pullsvia induction or by resonancesthe aro-

matic π electron cloud to increase their negative charge.

This was consistent with the QSAR result for PNMT

ligands, which indicated that an increase in the electron-

withdrawing ability of the 7-substituent should increase

the PNMT inhibitory potency. On the other hand, large

contours representing an area where electron deficiency

is favored were associated with nonlipophilic substitu-

ents in the CoMFA electrostatic map for the R₂-adreno-

ceptor. As implied above, the electron-withdrawing

substituents will have more electron density around

them and will experience electrostatic repulsion in this

area. Hence, THIQs bearing an electron-withdrawing

group were predicted to have lower affinity toward the

R₂-adrenoceptor, as would be expected given the sign

E. Com p a r ison of Cla ssica l QSAR a n d CoMF A

Resu lts. The traditional QSAR analysis of 7-substituted-

THIQs for the active site of PNMT and for the R₂-

adrenoceptor has established the importance of lipophi-

licity (π) and size (MR) of the 7-substituent. From

comparison of QSAR eqs 2 and 4, it appears that at both

the PNMT active site and at the R₂-adrenoceptor, the

hydrophobic and steric interactions of 7-substituted-

THIQs were similar. When the steric field contours for

PNMT and R₂-adrenoceptor ligands from the CoMFA

were compared, they seemed to overlap with each other

quite well (Figure 3a,c). MR is a rough steric index of

the substituent, and though it lacks directionality, it can

be used to account for steric interactions at a given

position. In a CoMFA, the steric contours represent

critical van der Waals interactions of the molecule with

probe atoms, and these interactions are correlated with

biological activity. The outcome of the CoMFA not only

confirmed the QSAR analysis but also confirmed the

three-dimensional location of the steric bulk tolerance

regions in space with respect to the THIQ skeleton.

Examination of Figure 3a,c revealed a few positive

contours associated with the region where bulky lipo-

128 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 1

Grunewald et al.

from magnesium. Hexanes refer to the mixture of hexane

isomers (bp 40-70 °C). Solvents were routinely distilled prior

to use; anhydrous tetrahydrofuran (THF), ether (Et₂O), and

benzene were obtained by distillation from sodium-benzophe-

none ketyl; dry methylene chloride (CH₂Cl₂) was obtained by

distillation from phosphorus pentoxide. n-Butyllithium was

standardized by titration against 2,5-dimethoxybenzyl alcohol.

All reactions requiring anhydrous conditions and/or an inert

atmosphere were performed under a positive nitrogen (N₂) or

argon (Ar) flow, and all glassware was oven-dried and/or flame-

dried.

Amine hydrochloride salts were prepared by the addition

of a solution of methanolic HCl to a methanolic solution of the

amine, followed by crystallization of the resulting hydrochlo-

ride from MeOH-Et₂O.

The compounds SK&F 64139 (1), SK&F 29661 (2), and

SK&F 72223 (3) were kindly provided by Smith Kline and

French Laboratories, Smith Kline Corp., Philadelphia, PA.

S-Adenosyl-^L-methionine was obtained from Sigma Chemical

Co. [methyl-³H]-S-Adenosyl-L-methionine that was used in the

radiochemical assay was purchased from New England Nuclear

Corp. (Boston, MA). Bovine adrenal glands were obtained from

Davis Meat Processing (Overbrook, KS). [³H]Clonidine used

in the R₂-adrenoceptor binding assay was purchased from

Amersham Corp. (Arlington Heights, IL).

of the coefficient of σ_min the QSAR equation (eq 4) for

the R₂-adrenoceptor.

Su m m a r y a n d Con clu sion

A series of 7-substituted-THIQs were synthesized and

evaluated for PNMT inhibitory activity and R₂-adreno-

ceptor affinity. Both qualitative and quantitative treat-

ments of the data (see section B above) required a

binding model in which two different binding modes

were assigned to 7-substituted-THIQs based on the

lipophilicity of the 7-substituent. A classical QSAR and

CoMFA were successfully developed to determine the

influence of the steric and electrostatic interactions

involved in binding of 7-substituted-THIQs at the

PNMT active site and at the R₂-adrenoceptor. The

results from these two different techniques were in

general agreement with each other. The CoMFA model

provided a hypothetical “receptor map” which indicated

a degree of similarity in the steric interactions at the

PNMT active site and the R₂-adrenoceptor. However,

major differences were found in the electrostatic inter-

actions, and a strong nonlipophilic, electron-withdraw-

ing substituent was determined to be optimal for

imparting PNMT over R₂-adrenoceptor selectivity to the

THIQ nucleus. Since these models were calculated from

a limited set of data on 7-substituted-THIQs, they

should not be overinterpreted. However, they can

provide an important tool for future drug design of

selective inhibitors of PNMT.

7-Meth oxyca r bon yl-1,2,3,4-tetr a h ydr oisoqu in olin e Hy-

d r och lor id e (11‚HCl). Nitrile 10²¹(0.205 g, 1.29 mmol) was

added to a saturated methanolic HCl solution (20 mL, pre-

pared by bubbling HCl in dry MeOH), and to the resulting

suspension was added water (0.03 mL). The mixture was

heated to reflux for 18 h. The solution was cooled, and solvent

was removed on a rotary evaporator to yield a colorless solid

which was treated with 5% NaHCO₃. The solution was

extracted with CH₂Cl₂(thrice), dried over anhydrous Na₂SO₄,

and evaporated to give a colorless semisolid (0.243 g, 98.0%).

Recrystallization of the semisolid from CH₂Cl₂-hexanes pro-

duced white fluffy needles: mp 216-218 °C dec; IR (KBr) 3390

(NH), 2949, 2730, 1700 (CO), 1585, 1430, 1290, 1270, 1200,

Exp er im en ta l Section

All reagents and solvents were reagent grade or were

purified by standard methods before use. Melting points were

determined in open capillaries on a Thomas-Hoover melting

point apparatus calibrated with known compounds but are

otherwise uncorrected. Proton nuclear magnetic resonance

spectra (¹H NMR) were recorded on a Varian XL-300 or a GE

QE Plus spectrometer with CDCl₃as the solvent, and chemical

shifts are reported in parts per million (ppm) relative to

tetramethylsilane (TMS, 0.00 ppm). Carbon nuclear magnetic

resonance spectra (¹³C NMR) were recorded on a Varian XL-

300 spectrometer with CDCl₃as the solvent, and the chemical

shifts are reported in ppm relative to CDCl₃(77.00 ppm). For

the hydrobromide salts of the phenolic amines, NMR spectra

were recorded in deuterated dimethyl sulfoxide (DMSO-d₆) and

the chemical shifts are reported relative to DMSO (2.49 ppm

for ¹H and 39.50 ppm for ¹³C). Multiplicity abbreviations: s,

singlet; d, doublet; t, triplet; q, quartet; m, multiplet; bs, broad

singlet; e, exchangeable. Infrared spectra were obtained on a

Perkin-Elmer 1420 infrared spectrophotometer. Electron-

impact mass spectra (EIMS) were obtained on Ribermag R10-

10 mass spectrometer. The relative intensities of the mass

spectrum peaks are listed in parentheses. Flash chromatog-

raphy⁵⁶was performed using silica gel 60 (230-400 mesh)

supplied by Universal Adsorbents, Atlanta, GA. Preparative

centrifugal thin-layer chromatography (PCTLC) was per-

formed on a Harrison model 7924 Chromatotron (Harrison

Research, Palo Alto, CA) using Merck silica gel 60 PF254/

CaSO₄‚0.5H₂O binder on 1-, 2-, or 4-mm thickness plates.

Analytical TLC was performed by using silica gel with a

fluorescent indicator coated on 1- × 3-in. glass plates in 0.2-

mm thickness (Whatman MKGF silica gel 200µ). Bulb-to-bulb

distillations were carried out on a Kugelrohr distillation

apparatus (Aldrich Chemical Co., Milwaukee, WI), and oven

temperatures were recorded. Combustion analyses were per-

formed on a Hewlett-Packard model 185B CHN analyzer at

the University of Kansas by Dr. Tho Ngoc Nguyen.

1100, 805 cm^-^{1 1}H NMR (CDCl₃) δ 7.78-7.68 (m, 2 H, H-6

;

and H-8), 7.12 (d, 1 H, J ) 8 Hz, H-5), 4.02 (s, 2 H, H-1), 3.88

(s, 3 H, CO₂CH₃), 3.13 (t, 2 H, J ) 5.9 Hz, H-3), 2.83 (t, 2 H,

J ) 5.9 Hz), 2.46 (bs, e, 1 H, NH); ¹³C NMR (CDCl₃) δ 166.9

(CO₂Me), 140.1, 135.6, 129.2, 127.4, 126.9, 51.7 (OCH₃), 47.8

(C-1), 43.2 (C-3), 29.1 (C-4).

The hydrochloride salt was recrystallized from MeOH-Et₂O

as colorless needles: mp 222-223 °C dec; EIMS m/z 192 (M

+ 1, 6), 191 (M⁺, 33), 190 (M⁺- 1, 100), 176 (15), 162 (46),

131 (73), 103 (48), 77 (50), 51 (35). Anal. (C₁₁H₁₃NO₂‚HCl) C,

H, N.

The stability of the methyl ester to hydrolysis in the assay

conditions was determined by an NMR experiment, using a

procedure described previously.⁵⁷Compound 11‚HCl in pD 9.03

0.5 M phosphate buffer was monitored by NMR and showed

no evidence of hydrolysis over a period of 60 min (twice the

standard PNMT or R₂-adrenoceptor assay incubation time).

7-Hyd r oxym eth yl-1,2,3,4-tetr a h yd r oisoqu in olin e Ox-

a la te (12). To dry ether (20 mL) under N₂was added LiAlH₄

(0.30 g, 7.8 mmol), and the suspension was heated to reflux

for 1 h. After cooling to room temperature, compound 11 (0.50

g, 2.6 mmol) was added as a solution in dry THF (5 mL), and

the mixture was allowed to stir at room temperature for 12 h.

The reaction mixture was cooled in an ice bath and quenched

carefully with water (0.2 mL) and 10% KOH (0.2 mL). The

resulting suspension was filtered and extracted with CH₂Cl₂

(four times), dried over anhydrous Na₂SO₄, and evaporated to

give a colorless solid (0.38 g, 88%) which was crystallized from

CH₂Cl₂-hexanes: mp 132-134 °C; IR (KBr) 3260, 3050, 1450,

1360, 1340, 1120, 1060, 970, 905, 805 cm^-1; ¹H NMR (CDCl₃)

δ 7.10-6.97 (m, 3 H, ArH), 4.53 (s, 2 H, ArCH₂O), 3.93 (s, 2

H, H-1), 3.45-3.15 (m, b, e, 2 H, NH and OH), 3.07 (t, 2 H, J

) 6 Hz, H-3), 2.75 (t, 2 H, J ) 5.9 Hz, H-4); ¹³C NMR (CDCl₃)

δ 138.7, 134.9, 132.8, 128.5, 124.0, 63.3, 47.4, 43.0, 28.1.

As the hydrochloride salt was difficult to crystallize, the

oxalate salt of the amine was prepared. A saturated solution

Unless otherwise stated, all methanol (MeOH) and ethanol

(EtOH) used were anhydrous and were prepared by distillation

PNMT Inhibitors: Classical QSAR and CoMFA

J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 1 129

of anhydrous oxalic acid in dry Et₂O was added to a methanolic

solution of the amine 12, and the oxalate salt obtained was

crystallized from MeOH-Et₂O: mp 155-156 °C; EIMS m/z

164 (M + 1, 5), 163 (M⁺, 46), 162 (M⁺- 1, 100), 134 (70), 132

(60), 130 (20), 117 (24), 105 (94), 91 (85), 77 (68), 51 (36), 44

(28). Anal. (C₁₀H₁₃NO‚C₂H₂O₄) C, H, N.

cold 48% HBr (2.4 mL) in water 8.1 mL). To this stirred

solution was added sodium nitrite (0.38 g, 5.6 mmol) dropwise

as a solution in water (4.5 mL). The reddish-colored solution

turned orange, and the presence of excess HNO₂was checked

by starch-iodide paper. Excess HNO₂was destroyed by adding

urea (0.20 g) to the reaction mixture (a negative starch-iodide

test was obtained at this point). This diazotized solution was

added to a well-stirred, warm (35 °C) mixture of CuBr (2.17

g, 15.1 mmol), 48% HBr (5.1 mL), and water (12.6 mL). After

the completion of the addition of the diazotized solution, the

reaction mixture was warmed to 75-80 °C and stirred at that

temperature for 1.5 h. The reaction mixture was allowed to

stand overnight and then cautiously made alkaline with 50%

KOH solution. The blue copper salts were filtered, and the

filtrate was extracted with CH₂Cl₂(four times), dried over

anhydrous Na₂SO₄, and evaporated to get a dark-red oil (0.88

g). Bulb-to-bulb distillation (110-115 °C, 0.15 mmHg) yielded

compound 16 as a clean colorless oil (0.75 g, 73%): ¹H NMR

(CDCl₃) δ 7.21 (dd, 1 H, J ) 1.6, 8.1 Hz, ArH), 7.13 (m, 1 H,

ArH), 6.92 (d, 1 H, J ) 8.1 Hz, ArH), 3.94 (s, 2 H, H-1), 3.09

(t, 2 H, J ) 6 Hz, H-3), 2.70 (t, 2 H, J ) 5.9 Hz, H-4), 2.33 (bs,

e, 1 H, NH); ¹³C NMR (CDCl₃) δ 137.8, 133.5, 130.8, 128.8,

119.0, 119.0, 47.7 (C-1), 43.4 (C-3), 28.4 (C-4).

7-Am in om eth yl-1,2,3,4-tetr a h yd r oisoqu in olin e Dih y-

d r och lor id e (13‚2HCl). Nitrile 10²¹(0.30 g, 1.9 mmol) was

dissolved in dry THF (10 mL), and BH₃‚Me₂S complex (2.0 M

in THF, 2 mL, 4 mmol) was added dropwise. The solution was

heated to reflux for 14 h under N₂followed by removal of the

THF by distillation. The semisolid obtained was treated with

saturated methanolic HCl (10 mL) and heated to reflux for 5

h. After cooling the solvent was removed on a rotary evaporator

to give a colorless solid which was made alkaline with 10%

KOH solution and extracted with CH₂Cl₂(four times). The

combined CH₂Cl₂extracts were washed with brine (once), dried

(anhydrous K₂CO₃), and evaporated to produce a colorless thick

oil (0.30 g) which was distilled bulb-to-bulb (120-130 °C, 0.1

mmHg) to give a colorless oil (0.14 g, 44%): ¹H NMR (CDCl₃)

δ 7.10-7.04 (m, 2 H, ArH), 6.95 (m, 1 H, ArH), 3.97 (s, 2 H,

H-1), 3.78 (s, 2 H, ArCH₂NH₂), 3.09 (t, 2 H, J ) 5.9 Hz, H-3),

2.74 (t, 2 H, J ) 5.9 Hz, H-4), 1.51 (s, e, 3 H, NH₂and NH);

¹³C NMR (CDCl₃) δ 140.5, 135.8, 133.0, 129.1, 124.5, 124.5,

48.1 (C-1), 45.9 (ArCH₂NH₂), 43.7 (C-3), 28.7 (C-4).

The colorless needles of the hydrochloride salt were obtained

by crystallization from MeOH-Et₂O: mp 245-247 °C dec; IR

(KBr) 2940, 2710, 1580, 1400, 1195, 840, 810 cm^-1; EIMS m/z

214 (M + 3, 17), 213 (M + 2, 59), 212 (M + 1, 100), 211 (M⁺,

59), 184 (72), 182 (72), 132 (39), 130 (33), 103 (84), 77 (64), 51

(61). Anal. (C₉H₁₀BrN‚HCl) C, H, N.

The dihydrochloride salt was crystallized from MeOH-Et₂O

as fine colorless crystals: mp 324-325 °C dec; IR (KBr) 3180,

2900, 2800, 1585, 1500, 1365, 1060, 820 cm^-1; EIMS m/z 162

(M⁺, 47), 161 (M⁺- 1, 93), 145 (57), 133 (54), 132 (100), 130

(23), 118 (29), 115 (29), 105 (29), 103 (24), 91 (61), 77 (49), 51

(21). Anal. (C₁₀H₁₄N₂‚2HCl) C, H, N.

7-Br om o-N-tr iph en ylm eth yl-1,2,3,4-tetr ah ydr oisoqu in -

olin e (34). To a stirred solution of 16 (0.564 g, 2.66 mmol) in

CH₂Cl₂(10 mL) was added triethylamine (0.296 g, 0.410 mL,

2.92 mmol), and the solution was stirred in an ice bath.

4-(Dimethylamino)pyridine (0.033 g, 0.27 mmol) and triphen-

ylmethyl chloride (1.115 g, 4.00 mmol) were added in one

portion, and the reaction mixture was stirred vigorously. After

a few minutes a copious precipitate of triethylammonium

hydrochloride was observed in the reaction mixture. The

reaction was stirred at room temperature for 18 h, quenched

with 1 N NaOH, and extracted with CH₂Cl₂(thrice), and the

combined organic extracts were dried over anhydrous Na₂SO₄.

Evaporation gave a yellow-orange-colored solid (1.705 g). The

solid was passed through a silica plug using hexanes/EtOAc

(3:1) as eluent. The pale-yellow solid thus obtained was

subjected to PCTLC (silica, 4 mm; CH₂Cl₂/hexanes/MeOH/NH₄-

OH, 150:450:10:1, as the eluent) to yield a colorless foamy solid

(1.147 g, 94.90%): mp 170-177 °C (smears on melting); IR

(KBr) 1590, 1480, 1440, 1030, 740, 700 cm^-1; ¹H NMR (CDCl₃)

δ 7.54-7.52 (m, 6 H, ArH), 7.28-6.94 (m, 12 H, ArH), 3.45-

3.35 (m, 2 H, H-1), 2.94-2.90 (m, 2 H, H-3), 2.55-2.45 (m, 2

H, H-4); ¹³C NMR (CDCl₃) δ 142.1, 138.2, 133.7, 130.1, 129.2,

129.0, 128.8, 127.4, 126.0, 118.6, 76.4 (CPh₃), 50.38 (C-1), 45.98

(C-3), 29.24 (C-4); MS m/z (CI, NH₃) 454 (M + 1, 0.1), 378 (0.2),

376 (0.2), 243 (Ph₃C⁺, 100), 165 (10). Anal. (C₂₈H₂₄BrN) C, H,

N.

7-Met h ylt h io-1,2,3,4-t et r a h yd r oisoq u in olin e H yd r o-

ch lor id e (17‚HCl). In a flame-dried three-neck round-bottom

flask equipped with a thermometer, magnetic stir bar, and a

N₂inlet was placed 34 (1.08 g, 2.37 mmol) in dry THF (15

mL). The solution was chilled to -78 °C, and n-BuLi (1.95 M

in hexanes, 1.34 mL, 2.62 mmol) was added dropwise. A dark-

red-colored solution was obtained and was allowed to warm

to -10 °C, over a period of 1 h. The reaction mixture was again

chilled to -78 °C and a solution of methyl disulfide (0.27 g,

0.26 mL, 2.8 mmol) in dry THF (3 mL) was added. The reaction

mixture was stirred overnight, then cooled in an ice bath, and

treated with acetone (4 mL) and concentrated HCl (2 mL). The

pale-yellow colorless mixture was stirred overnight, and then

the organic solvents were removed on a rotary evaporator. The

residue obtained was diluted with water (15 mL) and washed

with CH₂Cl₂(twice). The aqueous layer was made alkaline

with 6 N NaOH and extracted with CH₂Cl₂(four times). The

combined CH₂Cl₂extracts were dried over anhydrous Na₂SO₄

and evaporated to yield a colorless oil (0.27 g). Bulb-to-bulb

1,2,3,4-Tetr a h yd r oisoqu in olin e-7-ca r boxylic Acid Hy-

d r och lor id e (14‚HCl). In a thick-walled tube with a sidearm

was placed nitrile 10²¹(0.25 g, 1.6 mmol), and to it concen-

trated HCl (1.9 mL) was added. The reaction mixture was

cooled in a dry ice-acetone bath, and vacuum was applied to

the sidearm to evacuate the tube. The reaction mixture was

evacuated thrice and then heated in the sealed tube at 150 °C

for 9 h. After cooling to room temperature, the colorless solid

which formed was collected by filtration and crystallized from

water (0.32 g, 93%): mp 338-340 °C dec; IR (KBr) 3140, 1690,

1590, 1570, 1400, 1270, 1235, 1130, 870 cm^-^{1 1}H NMR

;

(DMSO-d₆) δ 9.88 (bs, e, 2 H, NH₂⁺), 7.82-7.78 (m, 2 H, H-6

and H-8), 7.33 (d, 1 H, J ) 7.8 Hz, H-5), 4.3 (s, 2 H, H-1), 3.34

(t, 2 H, J ) 5.6 Hz, H-3), 3.08 (t, 2 H, J ) 5.6 Hz, H-4); ¹³C

NMR (DMSO-d₆) δ 166.9 (CO₂H), 137.3, 129.5, 129.0, 128.0,

43.2 (C-1), 40.1 (C-3), 24.8 (C-4); EIMS m/z 177 (M⁺, 30), 176

(M⁺- 1, 100), 149 (40), 148 (64), 132 (21), 131 (33), 103 (26),

91 (14), 77 (44), 51 (26), 45 (39) 44 (17), 43 (30), 42 (19). Anal.

(C₁₀H₁₁NO₂‚HCl) C, H, N.

7-Am in oca r bon yl-1,2,3,4-tetr a h yd r oisoqu in olin e Hy-

28

d r och lor id e (15‚HCl). To a suspension of KF-Al₂O₃(1.0

g) and t-BuOH (10 mL) was added nitrile 10²¹(0.300 g, 1.89

mmol), and the mixture was heated to reflux under N₂for 24

h. The suspension was filtered to remove the alumina, and

the solid was washed with CHCl₃and water. The filtrate was

extracted with CHCl₃(thrice), dried (anhydrous Na₂SO₄), and

evaporated to yield a pale-yellow-colored oil (0.095 g, 28.4%).

(The low yield could be avoided by just filtering and evaporat-

ing the reaction mixture. The amide had very low solubility

in organic solvents.) The pale-yellow solid was crystallized

from CH₂Cl₂-hexanes to afford a colorless solid: mp 173-

175 °C dec; ¹H NMR (DMSO-d₆) δ 7.97 (s, 2 H, ArCONH₂),

7.64-7.56 (m, 2 H, H-6 and H-8), 7.11 (d, 1 H, J ) 7.9 Hz,

H-5), 3.94 (s, 3 H, H-1 and NH, e), 3.03 (t, 2 H, J ) 5.9 Hz,

H-3), 2.78 (t, 2 H, J ) 5.9 Hz, H-4); ¹³C NMR (DMSO-d₆) δ

168.3, 138.3, 135.7, 131.0, 128.5, 125.2, 124.5, 47.6, 42.9, 28.6.

The hydrochloride salt was crystallized from MeOH-Et₂O:

mp 260-262 °C; IR (KBr) 3460 (NH₂), 3340 (NH₂), 3100, 1650

(CONH₂), 1600, 1400, 1380, 770 cm^-1; MS (CI, NH₃) m/z 177

(M + 1, 100), 175 (M⁺- 1, 11). Anal. (C₁₀H₁₂N₂O‚HCl) C, H,

N.

7-Br om o-1,2,3,4-t et r a h yd r oisoq u in olin e H yd r och lo-

r id e (16‚HCl). Amine 9 (0.75 g, 5.06 mmol) was added to ice-

130 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 1

Grunewald et al.

distillation (95-100 °C, 0.05 mmHg) gave a colorless oil (0.24

NH₄OH, 150:450:10:1, as the eluent) gave a yellow solid (1.5

g, 56%): IR (film) 3300 (NH), 2920, 1595, 1425, 1255, 1190,

g, 50%) which was identified as 7-trifluoroacetyl-N-triphenyl-

1

950, 800 cm^-1; H NMR (CDCl₃) δ 7.04-6.92 (m, 3 H, ArH),

methyl-THIQ: IR (KBr) 1705 (CO) cm^-1; H NMR (CDCl₃) δ

3.96 (s, 2 H, H-1), 3.10 (t, 2 H, J ) 6 Hz, H-3), 2.72 (t, 2 H, J

) 5.9 Hz, H-4), 2.44 (s, 3 H, SCH₃), 1.69 (bs, e, 1 H, NH); ¹³C

NMR (CDCl₃) δ 136.6, 135.0, 132.0, 129.7, 124.9, 124.7, 48.2

(C-1), 43.8 (C-3), 28.7 (C-4), 16.3 (SCH₃).

The hydrochloride salt was crystallized from MeOH-Et₂O:

mp 249-251 °C dec (lit.¹⁷mp 247-248 °C); EIMS m/z 181 (M

+ 2, 4), 180 (M + 1, 10), 179 (M⁺, 66), 178 (27), 150 (100), 135

(16), 117 (13), 91 (15). Anal. (C₁₀H₁₃NS‚HCl) C, H, N.

7.82 (d, 1 H, J ) 7.8 Hz, H-6), 7.65 (bs, 1 H-8), 7.56-7.50 (m,

6 H), 7.30-7.15 (m, 10 H), 3.52 (bs, 2 H, H-1), 3.08-3.06 (m,

2 H, H-3), 2.57 (bs, 2 H, H-4). This compound (0.68 g, 1.4 mmol)

was dissolved in a mixture of acetone (10 mL) and 3 N HCl (5

mL). After stirring for 1 h at room temperature, the colorless

reaction mixture was concentrated (4 mL) on a rotary evapora-

tor and washed with CH₂Cl₂(thrice) to remove triphenyl-

methanol. Evaporation of the aqueous layer gave a colorless

solid (0.36 g) which was crystallized from MeOH-Et₂O to give

20‚HCl as a white crystalline solid (0.34 g, 89% or 44% from

34): mp 240 °C dec; IR (KBr) 3290, 2940, 2790, 1580, 1170,

1050, 940, 840 cm^-1; ¹H NMR (CDCl₃) δ 9.84 (bs, 2 H, NH₂⁺),

7.40-7.30 (m, 2 H), 7.27 (m, 1 H), 4.27 (s, 2 H, H-1), 3.36-

3.30 (m, 2 H, H-3), 3.05-3.01 (m, 2 H, H-4); EIMS m/z 229

(M⁺, 31), 228 (M⁺- 1, 75), 200 (13), 132 (25), 131 (100), 130

(20), 103 (39), 77 (60), 51 (32). Anal. (C₁₁H₁₀F₃NO‚HCl‚0.5H₂O)

C, H, N.

7-F lu or osu lfon yl-N-a cetyl-1,2,3,4-tetr a h yd r oisoqu in o-

lin e (37). 7-Chlorosulfonyl-N-acetyl-1,2,3,4-tetrahydroisoquin-

oline⁹(36; 6.63 g, 24.2 mmol) was dissolved in dry acetonitrile

(20 mL), and to it were added anhydrous KF (2.72 g, 48.4

mmol) and 18-crown-6 ether (0.06 g, 0.2 mmol). The suspension

was stirred under N₂for 14 h, and a thick slurry resulted

which was diluted with water (100 mL). The emulsion was

extracted with EtOAc (thrice); the combined EtOAc extracts

were dried over anhydrous Na₂SO₄and evaporated. The yellow

oil (4.8 g) thus obtained was passed through a silica plug

(hexanes/EtOAc, 1:1, as the initial eluent to remove nonpolar

impurities and then acetone as the final eluent) to afford a

pale-yellow syrup (3.83 g). Bulb-to-bulb distillation (165 °C,

0.05 mmHg) yielded a pale-yellow oil (3.60 g, 57.2%) that

solidified on standing: mp 80-90 °C; IR (film) 1650 (CO), 1450,

1390 (SO₂), 1210 (SO₂), 1090, 860, 770 cm^-1; ¹H NMR (CDCl₃)

δ 7.84-7.79 (m, 2 H, H-6 and H-8), 7.46-7.42 (m, 1 H, H-5),

4.85-4.75 (m, 2 H, H-1), 3.87 and 3.76 (t, 2 H, J ) 5.9 Hz,

H-3, conformers), 2.07-2.96 (m, 2 H, H-4), 2.22-2.21 (m, 3

H, COCH₃); EIMS m/z 258 (M + 1, 5), 257 (M⁺, 31), 214 (M⁺

- COCH₃, 34), 200 (12), 186 (10), 130 (8), 116 (13), 103 (10),

77 (11), 43 (100). Anal. (C₁₁H₁₂FNO₃S) C, H, N.

7-P h en ylsu lfon yl-1,2,3,4-tetr a h yd r oisoqu in olin e Hy-

d r och lor id e (21‚HCl). Sulfonyl fluoride 37 (0.26 g, 1.0 mmol)

was dissolved in dry THF (20 mL) and cooled to -78 °C.

Dropwise addition of phenylmagnesium bromide (3 M solution

in Et₂O) to the stirred solution under N₂produced a yellow-

red reaction mixture. The mixture was allowed to warm to

room temperature and stirred for 14 h. The red solution was

cooled in an ice bath and quenched with saturated NH₄Cl

solution followed by 3 N HCl (10 mL). The resulting yellow

solution was washed with CH₂Cl₂(thrice) to remove nonbasic

impurities. The aqueous layer was made alkaline with solid

KOH and filtered through Celite to remove Mg(OH)₂, and the

Celite layer was washed with CH₂Cl₂(thrice). The filtrate was

extracted with CH₂Cl₂(thrice), the organic washes and extracts

were combined and dried (anhydrous K₂CO₃), and the solvent

was removed by rotary evaporation to yield a yellow oil (0.14

g, 51%): ¹H NMR (CDCl₃) δ 7.91 (d, 2 H, J ) 7.2 Hz, ArH),

7.66 (d, 1 H, J ) 7.8 Hz, ArH), 7.60-7.46 (m, 4 H, ArH), 7.20

(d, 1 H, J ) 7.9 Hz, ArH), 4.02 (s, 2 H, H-1), 3.10 (t, 2 H, J )

5.9 Hz, H-3), 2.80 (m, 3 H, H-4 and NH).

(()-7-Meth ylsu lfin yl-1,2,3,4-tetr ah ydr oisoqu in olin e Hy-

d r och lor id e (18‚HCl). To compound 17 (0.255 g, 1.42 mmol)

was added trifluoroacetic acid (3 mL), and the mixture was

stirred in an ice bath. Trifluoroperacetic acid (6.37 N, 0.250

mL, 1.59 mequiv) was added in one portion, and the mixture

was stirred for 14 h at room temperature. At this stage TLC

showed little starting material and a slight positive starch-

iodide test. The trifluoroacetic acid was carefully removed on

a rotary evaporator. Traces of trifluoroacetic acid were re-

moved by adding benzene to the residue followed by evapora-

tion. The colorless oil was made alkaline with 3 N NaOH and

extracted with CHCl₃(thrice). After washing the CHCl₃

extracts with brine and drying over anhydrous Na₂SO₄,

evaporation of the combined CHCl₃extracts yielded a pale-

orange-red oil (0.211 g). The oil was purified by PCTLC (silica

gel, 4 mm) using CH₂Cl₂/MeOH/NH₄OH (250:20:1) as the

eluent to give 18 as a viscous colorless liquid (0.117 g, 44.8%):

IR (KBr) 3460 (NH), 2900, 2780, 1420, 1015 (SO), 950, 850

cm^-1; ¹H NMR (CDCl₃) δ 7.37-7.34 (m, 2 H), 7.25-7.22 (m, 1

H), 4.08 (s, 2 H, H-1), 3.17 (t, 2 H, J ) 5.8 Hz, H-3), 3.12 (b, e,

1 H, NH), 2.87 (t, 2 H, J ) 5.8 Hz, H-4), 2.70 (s, 3 H, SOCH₃);

¹³C NMR (CDCl₃) δ 142.6, 138.0, 136.7, 130.1, 121.1, 120.9,

47.7, 43.8, 43.1, 28.7.

The hydrochloride salt was crystallized from MeOH-Et₂O

as a colorless solid: mp 225 °C dec (lit.¹⁷mp 200-202 °C);

EIMS m/z 196 (M + 1, 11), 195 (M⁺, 54), 194 (M⁺- 1, 65),

180 (38), 179 (57), 178 (46), 166 (10), 151 (100), 150 (36), 130

(38), 117 (20), 102 (27), 91 (69), 77 (75), 51 (49), 45 (45). Anal.

(C₁₀H₁₃NOS‚HCl) C, H, N.

7-Meth ylsu lfon yl-1,2,3,4-tetr a h yd r oisoqu in olin e Hy-

d r och lor id e (19‚HCl). Amine 17 (0.281 g, 1.57 mmol) was

dissolved in trifluoroacetic acid (3 mL), the pale-yellow-colored

solution was chilled in an ice bath, and trifluoroperacetic acid

(4 M, 0.81 mL, 3.2 mmol) was added. The reaction mixture

was allowed to stir for 24 h at room temperature. After workup

as for compound 18, followed by evaporation of the organic

layer, a pale-yellow-colored solid was obtained (single spot on

TLC, 0.281 g, 84.9%): mp 124-126 °C; IR (KBr) 3320 (NH),

2900, 1410, 1290 (SO₂), 1130 (SO₂), 960, 810, 760 cm^-^{1 1}H

;

NMR (CDCl₃) δ 7.68-7.60 (m, 2 H, H-6 and H-8), 7.27 (d, 1

H, J ) 8.0 Hz, H-5), 4.07 (s, 2 H, H-1), 3.16 (t, 2 H, J ) 5.9

Hz, H-3), 3.03 (s, 3 H, SO₂CH₃), 2.88 (t, 2 H, J ) 5.9 Hz, H-4),

1.88 (bs, e, 1 H, NH); ¹³C NMR (CDCl₃) δ 141.6, 137.7, 137.4,

130.3, 125.2, 124.6, 48.1, 44.5, 43.2, 29.3.

The hydrochloride salt was crystallized from MeOH-Et₂O

as small colorless needles: mp 262-264 °C dec (lit.¹⁷mp 258-

260 °C); EIMS m/z 212 (M + 1, 9), 211 (M⁺, 29), 210 (M⁺- 1,

100), 182 (22), 131 (59), 103 (34), 91 (73), 77 (62), 51 (33). Anal.

(C₁₀H₁₃NO₂S‚HCl) C, H, N.

7-Tr iflu or oa cetyl-1,2,3,4-tetr a h yd r oisoqu in olin e Hy-

d r och lor id e (20‚HCl). Compound 34 (2.9 g, 6.4 mmol) was

dissolved in dry THF (20 mL) and treated with n-BuLi (1.4

M, 5 mL, 7 mmol) at -78 °C under a positive pressure of N₂.

After stirring for 1 h at -78 °C, the reaction mixture was

transferred via a double-tipped needle to a three-neck flask

containing a solution of trifluoroethyl acetate (1.0 g, 0.83 mL,

7.0 mmol) in dry THF (10 mL) at -78 °C. The reaction mixture

was allowed to warm to room temperature over a period of 12

h and then quenched with saturated Na₂CO₃. The reaction

mixture was extracted with EtOAc (thrice); the combined

extracts were dried over anhydrous Na₂SO₄and evaporated

to afford a yellow solid (3.0 g). Purification of the crude product

by column chromatography (silica gel; CH₂Cl₂/hexanes/MeOH/

The hydrochloride salt was crystallized as a pale-yellow solid

from MeOH-Et₂O: mp 135-137 °C; IR (KBr) 3400, 1440, 1300

(SO₂), 1150 (SO₂), 1130, 1090, 880, 720, 680 cm^-1; EIMS m/z

274 (M + 1, 10), 273 (M⁺, 33), 272 (M⁺- 1, 100), 244 (21), 132

(M⁺- SO₂Ph, 29), 131 (67), 130 (26), 125 (60), 103 (17), 91

(44), 77 (94), 65 (18), 51 (36). Anal. (C₁₅H₁₅NO₂S‚HCl‚H₂O) C,

H, N.

7-Allylsu lfon yl-1,2,3,4-tetr a h yd r oisoqu in olin e Hyd r o-

ch lor id e (22‚HCl). Treatment of allylmagnesium bromide (1

M solution in THF, 18 mL, 18 mmol) with sulfonyl fluoride

37 (0.78 g, 3.0 mmol) in dry THF (50 mL) in a similar manner

as described above gave a pale-yellow oil (0.71 g). Purification

PNMT Inhibitors: Classical QSAR and CoMFA

J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 1 131

by PCTLC (silica, 2 mm) using CH₂Cl₂/MeOH /NH₄OH (250:

17:1) as the eluent furnished a pale-yellow oil (0.46 g, 65%):

IR (film) 3310, 1420, 1300 (SO₂), 1130 (SO₂), 1080, 800, 630

(49), 130 (11), 115 (29), 103 (38), 77 (51), 43 (100). Anal. (C₁₁H₁₂

INO) C, H, N.

-

7-Iod o-1,2,3,4-tetr a h yd r oisoqu in olin e Hyd r och lor id e

(24‚HCl). Iodoamide 40 (0.60 g, 2.0 mmol) was dissolved in

EtOH (5 mL) and heated to reflux with 3 N HCl (5 mL) for 4

h. A thick suspension resulted on cooling, and on removal of

the solvent a solid was obtained. The solid was treated with

20% KOH and extracted with CH₂Cl₂(four times). The organic

layers were combined, dried (anhydrous Na₂SO₄), and evapo-

rated to yield a buff-colored solid (0.50 g). Purification by

PCTLC (silica, 2 mm) using CH₂Cl₂/MeOH/NH₄OH (250:17:

1) as the eluent gave a buff-colored solid (0.45 g, 87%), mp

83-85 °C. The hydrochloride salt was crystallized from

MeOH-Et₂O: mp 243-244 °C dec; IR (KBr, HCl salt) 2920,

2760, 1575, 1420, 1345 (SO₂), 1155 (SO₂), 820 cm^-1; IR (KBr)

3300 (NH), 2940, 1475, 1125, 800, 780 cm^-1; ¹H NMR (CDCl₃)

δ 7.43-7.35 (m, 2 H, H-6 and H-8), 6.82 (d, 1 H, J ) 8.2 Hz,

H-5), 3.94 (s, 2 H, H-1), 3.10 (t, 2 H, J ) 5.9 Hz, H-3), 2.71 (t,

2 H, J ) 5.9 Hz, H-4), 1.85 (s, e, 1 H, NH); ¹³C (CDCl₃) δ 138.3,

134.9, 134.8, 134.4, 131.1, 90.6 (C-7), 47.6 (C-1), 43.5 (C-3),

28.6 (C-4).

cm^-^{1 1}H NMR (CDCl₃) δ 7.60-7.53 (m, 2 H, H-6 and H-8),

;

7.27-7.25 (m, 1 H, H-5), 5.83-5.72 (m, 1 H, CHdCH₂), 5.36-

5.30 (m, 1 H), 5.17 (d, 1 H, J ) 17 Hz), 4.07 (s, 2 H, H-1), 3.79

(d, 2 H, J ) 7.3 Hz, CH₂CHdCH₂), 3.16 (t, 2 H, J ) 5.9 Hz,

H-3), 2.88 (t, 2 H, J ) 5.9 Hz, H-4), 1.92 (bs, e, 1 H, NH).

The hydrochloride salt was obtained as a pale-yellow

crystalline solid: mp 188-190 °C; EIMS m/z 238 (M + 1, 10),

237 (M⁺, 31), 236 (M⁺- 1, 100), 196 (20), 132 (40), 131 (M⁺

-

SO₂C₃H₅, 59), 130 (21), 129 (19), 103 (34), 91 (10), 77 (40). Anal.

(C₁₂H₁₅NO₂S‚HCl) C, H, N.

7-Met h ylsu lfon yl-N-t r ip h en ylm et h yl-1,2,3,4-t et r a h y-

d r oisoqu in olin e (38). Compound 19 (0.72 g, 3.4 mmol) was

subjected to the same procedure as described for compound

34, to afford a pale-yellow solid (1.0 g, 66%): mp 145-150 °C;

IR (KBr) 2920, 1480, 1440, 1305 (SO₂), 1140 (SO₂), 960, 760,

1

740, 700 cm^-1; H NMR (CDCl₃) δ 7.55-7.52 (m, 6 H), 7.31-

7.15 (m, 12 H), 3.52 (bs, 2 H, H-1), 3.08-3.06 (m, 2 H, H-3),

2.98 (s, 3 H, SO₂CH₃), 2.57 (bs, 2 H, H-4); ¹³C NMR (CDCl₃) δ

141.75, 137.76, 129.68, 129.11, 127.86, 127.68, 127.19, 126.27,

125.60, 124.67, 77.18 (CPh₃), 50.74 (C-1), 45.77, 44.50, 30.13

(C-4); MS (CI NH₃) m/z 376 (M⁺- Ph, 2), 244 (21), 243 (Ph₃C⁺,

100), 165 (28). Anal. (C₂₉H₂₇NO₂S) C, H, N.

The hydrochloride salt was crystallized from MeOH-Et₂O:

mp 243-244 °C dec; IR (KBr, HCl salt) 2920, 2760, 1575, 1420,

1345 (SO₂), 1155 (SO₂), 820 cm^-1; EIMS m/z 260 (M + 1, 6),

259 (M⁺, 49), 258 (M⁺- 1, 75), 231 (10), 230 (76), 132 (M⁺

I, 27), 103 (77), 77 (100). Anal. (C₉H₁₀IN‚HCl) C, H, N.

-

7-Tr ich lor om eth ylsu lfon yl-1,2,3,4-tetr a h yd r oisoqu in o-

lin e Hyd r och lor id e (23‚HCl). Compound 38 (0.48 g, 1.1

mmol) was dissolved in dry THF (10 mL) and treated with

n-BuLi (2.0 M, 0.6 mL, 1.2 mmol) at -78 °C under a positive

pressure of N₂. The resulting dark-yellow-orange reaction

mixture was stirred at -78 °C for 45 min, CCl₄(1.0 mL) was

added, and the reaction mixture was allowed to warm to room

temperature. After stirring for 14 h, the dark-red reaction

mixture was quenched with a mixture of acetone and HCl (2:

1, 6 mL). After an identical workup as described for compound

17, a red-orange oil was isolated. Purification by PCTLC (silica,

2 mm) using CH₂Cl₂/MeOH/NH₄OH (250:20:1) as the eluent

gave a pale-orange oil (0.13 g, 39%): ¹H NMR (CDCl₃) δ 7.89-

7.80 (m, 2 H, H-6 and H-8), 7.35 (d, 1 H, J ) 8.2 Hz, H-5),

4.14 (s, 2 H, H-1), 3.21 (t, 2 H, J ) 6.0 Hz, H-3), 2.96 (t, 2 H,

J ) 6.0 Hz, H-4).

7-Tr iflu or om eth yl-N-a cetyl-1,2,3,4-tetr a h yd r oisoqu in -

olin e (41). To a solution of iodoacetamide 40 (1.5 g, 5.0 mmol)

in anhydrous DMF were added copper(I) iodide (0.95 g, 5.0

mmol), dry potassium fluoride (0.29 g, 5.0 mmol), and methyl

chlorodifluoroacetate (1.4 g, 1.1 mL, 10 mmol). Effervescence

was observed after the reaction mixture was heated to reflux,

and a dark-brown solution resulted. After heating to reflux

for 8 h under N₂, the dark-brown solution was diluted with

water (20 mL) and filtered through Celite to remove precipi-

tated solids. The Celite bed was washed with CH₂Cl₂(thrice),

and the combined filtrate was extracted with CH₂Cl₂(thrice).

The organic washes and extracts were combined and evapo-

rated to yield a dark-brown oil (1.26 g) which was passed

through a silica plug (CHCl₃/THF, 20:1 as the eluent) to afford

a brown oil (1.1 g). Further purification by PCTLC (silica, 4

mm; CHCl₃/THF, 20:1, as the eluent) gave a brown oil (0.94

g, 78%). A small amount of this product was distilled bulb-to-

bulb (120 °C, 0.1 mmHg) to get an analytically pure sample:

IR (film) 3460, 1640 (CO), 1325, 1230, 1160, 1070, 825, 735

cm^-1; ¹H NMR (CDCl₃) δ 7.44-7.12 (m, 3 H, ArH), 4.77-4.62

(m, 2 H, H-1), 3.84-3.60 (m, 2 H, H-3), 2.98-2.85 (m, 2 H,

H-4), 2.20-2.11 (m, 3 H, COCH₃); MS EIMS m/z 243 (M⁺, 43),

200 (M⁺- COCH₃, 24), 186 (12), 185 (18), 172 (26), 103 (17),

91 (18), 77 (13), 43 (100). Anal. (C₁₂H₁₂F₃NO) C, H, N.

7-Tr iflu or om eth yl-1,2,3,4-tetr a h yd r oisoqu in olin e Hy-

d r och lor id e (25‚HCl). Acidic hydrolysis of the trifluoroac-

etamide 41 (0.89 g, 3.7 mmol) as described for the iodo

derivative yielded a colorless oil (0.49 g, 67%): ¹H NMR

(CDCl₃) δ 7.38-7.17 (m, 3 H, ArH), 4.04 (s, 2 H, H-1), 3.15 (t,

2 H, J ) 5.9 Hz, H-3), 2.84 (t, 2 H, J ) 5.9 Hz, H-4), 1.88 (bs,

e, 1 H, NH).

The hydrochloride salt was obtained as colorless flakes: mp

242-244 °C dec; IR (KBr, HCl salt) 2775, 1580, 1340, 1170,

1110, 1075, 820 cm^-1; EIMS m/z 202 (M + 1, 6), 201 (M⁺, 36),

200 (M⁺- 1, 100), 182 (10), 172 (56), 151 (16), 132 (M⁺- CF₃,

15), 103 (34). Anal. (C₁₀H₁₀F₃N‚HCl) C, H, N.

The hydrochloride salt was crystallized from MeOH-Et₂O:

mp 243-244 °C dec; IR (KBr) 2920, 2760, 1575, 1420, 1345

(SO₂), 1155 (SO₂), 820 cm^-1; EIMS m/z 317 (M + 4, 2), 315 (M

+ 2, 4), 313 (M⁺, 6), 312 (M⁺- 1, 16), 196 (32), 167 (15), 132

(100), 131 (58), 130 (43), 117 (42), 103 (50), 84 (18), 77 (69), 47

(29). Anal. (C₁₀H₁₀Cl₃NO₂S‚HCl) C, H, N.

7-Iod o-N-a cet yl-1,2,3,4-t et r a h yd r oisoq u in olin e (40).

7-Amino-N-acetyl-1,2,3,4-tetrahydroisoquinoline (39) was pre-

pared from 8 in the same manner as described for 42.²¹To

the aromatic amine 39 (7.00 g, 36.8 mmol) was added

concentrated HCl (13.6 mL) in water (45 mL), and the solution

was cooled in an ice bath. Sodium nitrite (2.64 g, 40.6 mmol)

was added dropwise as a solution in water (15 mL). The

diazotized amine was stirred in the ice bath for 30 min and

then transferred in small portions to a vigorously stirred

biphasic mixture consisting of CHCl₃(150 mL), potassium

iodide solution (12.2 g, 73.6 mmol of KI in 45 mL water), and

copper(I) iodide (0.50 g). After stirring for 7 h at room

temperature, the brown suspension was diluted with CHCl₃

(100 mL), transferred to a separatory funnel, and washed with

10% Na₂S₂O₃solution (thrice). The organic layer was dried over

anhydrous Na₂SO₄followed by evaporation to yield a yellow

oil (11.4 g) which was passed through a silica plug (CHCl₃/

THF, 20:1, as the eluent) to afford a relatively pure pale-yellow

oil (8.21 g, 74.1%). Analytically pure pale-yellow oil was

obtained by bulb-to-bulb (150-160 °C, 0.01 mmHg) distilla-

tion: IR (film) 3450, 1630 (CO), 1430, 1240, 1225, 1025, 750,

730 cm^-1; ¹H NMR (CDCl₃) δ 7.50-7.46 (m, 2 H, ArH), 6.92-

6.87 (m, 1 H, ArH), 4.67 and 4.57 (s, 2 H, H-1), 3.79 and 3.66

(t, 2 H, J ) 5.9 Hz, H-3, conformers), 2.86-2.78 (m, 2 H, H-4),

2.17 (s, 3 H, COCH₃); EIMS m/z 302 (M + 1, 21), 301 (M⁺,

46), 286 (M⁺- CH₃, 10), 258 (M⁺- COCH₃, 25), 244 (16), 230

7-Acet a m id o-N-t r iflu or oa cet yl-1,2,3,4-t et r a h yd r oiso-

qu in olin e (43). To a solution of the amine 42²¹(1.38 g, 5.65

mmol) in CH₂Cl₂(10 mL) were added triethylamine (1.14 g,

1.57 mL, 11.3 mmol) and acetyl chloride (0.66 g, 0.60 mL, 8.5

mmol). The reaction mixture was stirred at room temperature

for 6 h and then quenched with ice water. After extraction of

the reaction mixture with CH₂Cl₂(thrice), the combined

extracts were washed with 1 N HCl (thrice). Drying (anhy-

drous Na₂SO₄), followed by the removal of the solvent, gave a

yellow solid which was crystallized from EtOAc-hexanes to

afford small pale-yellow crystals (1.19 g, 73.5%): mp 167-

168 °C; IR (KBr) 3320 (NH), 1690-1670 (CO), 1595, 1535,

132 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 1

Grunewald et al.

1

1500, 1410, 1180, 1145, 820 cm^-1; H NMR (CDCl₃) δ 8.96-

The resultant pale-yellow semisolid was dissolved in EtOH (5

mL) and heated to reflux with 3 N KOH (3 mL) under N₂for

3 h. The solvent was removed on a rotary evaporator to afford

a solid which was suspended in water (15 mL). Extraction with

CH₂Cl₂(thrice), drying (K₂CO₃), and evaporation yielded a

pale-yellow oil (0.15 g). Purification by PCTLC (silica, 1 mm)

using CH₂Cl₂/MeOH/NH₄OH (250:17:1) as the eluent yielded

28 as a pale-yellow oil (0.12 g, 80%): ¹H NMR (CDCl₃) δ 7.30-

7.10 (m, 5 H, ArH), 6.99-6.91 (m, 2 H, ArH), 6.82-6.80 (m, 1

H, ArH), 4.00-3.88 (m, 4 H, H-1, PhCH₂Ar), 3.08 (t, 2 H, J )

5.9 Hz, H-3), 2.74 (t, 2 H, J ) 5.8 Hz, H-4), 2.11 (bs, e, 1 H,

NH).

The free base was converted to the hydrochloride salt and

crystallized as shiny flakes from MeOH-Et₂O: mp (HCl) 215-

216 °C dec; IR (KBr) 2940, 2780, 1580, 1490, 1450, 1420, 1060,

960, 790, 710, 690 cm^-1; EIMS m/z 224 (M + 1, 12), 223 (M⁺,

60), 222 (M⁺- 1, 100), 194 (70), 179 (32), 178 (31), 165 (19),

132 (45), 115 (21), 91 (PhCH₂⁺, 48), 77 (14), 65 (14), 51 (18).

Anal. (C₁₆H₁₇N‚HCl‚0.3H₂O) C, H, N.

Ra d ioch em ica l Assa y for P NMT Activity. The assay

used in this study has been described elsewhere.⁴⁵Briefly, a

typical assay mixture consisted of 50 µL of 0.5 M phosphate

buffer (pH 8.0), 25 µL of 10 mM unlabeled AdoMet, 5 µL of

[methyl-³H]AdoMet, containing approximately 3 × 10⁵dpm

(specific activity approximately 15 mCi/mmol), 25 µL of

substrate solution (phenylethanolamine), 25 µL of inhibitor

solution, 25 µL of the enzyme preparation, and sufficient water

to achieve a final volume of 250 µL. After incubation for 30

min at 37 °C, the reaction mixture was quenched by addition

of 250 µL of 0.5 M borate buffer (pH 10.0) and was extracted

with 2 mL of toluene/isoamyl alcohol (7:3). A 1-mL portion of

the organic layer was removed, transferred to a scintillation

vial, and diluted with cocktail for counting. The mode of

inhibition was ascertained to be competitive in all cases

reported in Table 1 by inspection of the 1/V vs 1/S plots of the

data. All assays were run in duplicate with three inhibitor

concentrations over a 5-fold range. K_ivalues were determined

by a hyperbolic fit of the data.

r₂-Ad r en ocep tor Ra d ioliga n d Bin d in g Assa y. The ra-

dioligand receptor binding was performed according to the

method of U′Prichard et al.⁴⁶Male Sprague-Dawley rats were

decapitated, and the cortexes were dissected out and homog-

enized in 20 volumes (w/v) of ice-cold 50 mM Tris/HCl buffer

(pH 7.7 at 25 °C). Homogenates were centrifuged thrice for

10 min at 50 000g with resuspension of the pellet in fresh

buffer between spins. The final pellet was homogenized in 200

volumes (w/v) of ice-cold 50 mM Tris/HCl buffer (pH 7.7 at 25

°C). Incubation tubes containing [³H]clonidine (specific activity

ca. 19.2 mCi/mmol, final concentration 4.0 nM), various

concentrations of drugs, and an aliquot of freshly resuspended

tissue (800 µL) in a final volume of 1 mL were used. Tubes

were incubated at 25 °C for 30 min, and the incubation was

terminated by rapid filtration under vacuum through GF/B

glass fiber filters. The filters were rinsed with three 5-mL

washes of ice-cold 50 mM Tris buffer (pH 7.7 at 25 °C). The

filters were counted in vials containing premixed scintillation

cocktail. Nonspecific binding was defined as the concentration

of bound ligand in the presence of 2 µM phentolamine. All

assays were run in quadruplicate with five inhibitor concen-

trations over a 16-fold range. IC₅₀values were determined by

a log-probit analysis of the data, and K_ivalues were deter-

mined by the equation K_i) IC₅₀/(1 + [clonidine]/K_D), as all

Hill coefficients were approximately equal to 1.

8.93 (m, e, 1 H, NH), 7.61-7.25 (m, 2 H, H-6 and H-8), 7.11-

7.07 (m, 1 H, H-5), 4.76 and 4.72 (s, 2 H, H-1, conformers),

3.89-3.91 (m, 2 H, H-3), 2.93-2.87 (m, 2 H, H-4), 2.15 (s, 3

H, COCH₃); EIMS m/z 287 (M+1, 9), 286 (M⁺, 50), 245 (14),

244 (100), 243 (M⁺- COCH₃, 13), 229 (69), 217 (11), 147 (15),

131 (21), 119 (42), 91 (20), 43 (56). Anal. (C₁₃H₁₃F₃N₂O₂) C, H,

N.

7-Acet a m id o-1,2,3,4-t et r a h yd r oisoq u in olin e H yd r o-

ch lor id e (26‚HCl). Acetamide 43 (0.55 g, 1.9 mmol) was

dissolved in 20% aqueous MeOH (12 mL) and treated with

K₂CO₃(0.34 g, 2.4 mmol). After stirring at room temperature

for 14 h, the reaction mixture was evaporated to dryness, and

the residue was suspended in water (15 mL) and extracted

with EtOAc (thrice). The combined EtOAc extracts were dried

(anhydrous Na₂SO₄) and evaporated to yield a yellow solid

(0.33 g). PCTLC (silica, 2 mm) was performed on the crude

product (CH₂Cl₂/MeOH/NH₄OH, 250:17:1, as the initial eluent

and then CH₂Cl₂/MeOH/NH₄OH, 250:34:1, as the final eluent)

to afford a pale-yellow solid (0.22 g, 61%): mp 130-132 °C;

IR (KBr) 3280 (NH), 3220 (NHCOCH₃), 1665 (CO), 1590, 1540,

1410, 1250, 820 cm^-^{1 1}H NMR (CDCl₃) δ 9.15 (s, e, 1 H,

;

NHCOCH₃), 7.30-7.26 (m, 2 H, H-6 and H-8), 6.98 (d, 1 H, J

) 7.9 Hz, H-5), 3.94 (s, 2 H, H-1), 3.32 (bs, e, NH), 3.11 (t, 2

H, J ) 6.0 Hz, H-3), 2.75 (t, 2 H, J ) 5.8 Hz, H-4), 2.11 (s, 3

H, COCH₃); ¹³C NMR (CDCl₃) δ 168.3 (CO), 136.1, 134.6, 129.2,

128.9, 117.9, 117.2, 47.2 (C-1), 43.0 (C-3), 27.6 (C-4), 23.8

(COCH₃).

The hydrochloride salt was obtained as a yellow solid: mp

235 °C dec; EIMS m/z 190 (M⁺, 17), 189 (M⁺- 1, 12), 161 (19),

148 (32), 120 (17), 119 (100), 118 (23), 91 (18), 57 (11). Anal.

(C₁₁H₁₄N₂O‚HCl‚0.3H₂O) C, H, N.

7-Ben zoyl-N-tr iflu or oa cetyl-1,2,3,4-tetr a h yd r oisoqu in -

olin e (45). The compound was prepared from 44 according to

the procedure described²⁶for the synthesis of 7-acetyl-1,2,3,4-

tetrahydroisoquinoline in which benzoyl chloride was used

instead of acetyl chloride. The yellow-orange crude product was

subjected to flash chromatography (silica; CHCl₃/THF, 20:1,

as the eluent) to yield an orange solid (70%) which was

crystallized to afford a colorless crystalline light flakes from

hexanes-EtOAc (17%): mp 116-117 °C; IR (KBr) 1685

(NCOCF₃), 1650 (COPh), 1410, 1280, 1190, 1170, 1150, 1130,

910, 700 cm^-^{1 1}H NMR (CDCl₃) 7.82-7.74 (m, 2 H, ArH),

;

7.68-7.58 (m, 3 H, ArH), 7.52-7.45 (m, 2 H, ArH), 7.32-7.22

(m, 1 H, ArH), 4.78-4.89 (m, 2 H, H-1), 3.96-3.84 (m, 2 H,

H-3), 3.10-3.00 (m, 2 H, H-4); EIMS m/z 334 (M + 1, 13), 333

(M⁺, 53), 256 (16), 208 (16), 131 (15), 115 (11), 105 (PhCO⁺,

100), 77 (78), 69 (11), 51 (22). Anal. (C₁₈H₁₄F₃NO₂) C, H, N.

The regiochemistry of the product was confirmed by a one-

dimensional difference nuclear Overhauser effect (NOE) ex-

periment. When the protons at approximately 4.83 ppm were

irradiated (corresponding to the 1-position on the THIQ

nucleus), a positive NOE response was detected at 7.62 ppm

(corresponding to the 8-position on the THIQ nucleus). This

singlet was the only NOE enhancement observed.

7-Ben zoyl-1,2,3,4-tetr a h yd r oisoqu in olin e Hyd r och lo-

r id e (27‚HCl). The hydrolysis procedure as described²⁶for the

preparation of 7-acetyl-1,2,3,4-tetrahydroisoquinoline was per-

formed on 45 (0.35 g, 1.0 mmol) to yield 27 (0.25 g, 86%) as a

colorless solid: mp 244-246 °C dec; IR (KBr) 1640 (CO), 1600,

1580, 1270, 1250, 720 cm^-1; ¹H NMR (DMSO-d₆) δ 9.90 (bs, e,

2 H, NH₂⁺), 7.75-7.55 (m, 7 H, ArH), 7.41 (d, 1 H, J ) 8.1 Hz,

ArH), 4.34 (s, 2 H, H-1), 3.39-3.37 (m, 2 H, H-3), 3.14-3.12

(m, 2 H, H-4); ¹³C NMR (DMSO-d₆) δ 195.2 (CO), 137.3, 136.9,

135.2, 132.8, 129.6, 129.0, 128.6, 128.5, 128.2, 43.3 (C-1), 40.1

(C-3), 24.9 (C-4); EIMS m/z 238 (M + 1, 8), 237 (M⁺, 44), 236

(M⁺- 1, 78), 208 (26), 132 (20), 131 (36), 015 (PhCO⁺, 34), 77

(100), 51 (30). Anal. (C₁₆H₁₅NO‚HCl) C, H, N.

Com p a r a tive Molecu la r F ield An a lysis (CoMF A) Mod -

elin g. For all of the compounds in this study, with the

exception of those containing SO₂NHR substructures, the

minimum energy conformations were calculated without elec-

trostatics using the Tripos force field, followed by reminimi-

zation with charges calculated by the AM1 method in MOPAC

(SYBYL 6.2 implementation). The semiempirical AM1 option

in MOPAC was used for the geometry optimization and charge

calculation of compounds with SO₂NHR substructures, as

torsion parameters for such were not available in the Tripos

force field. The conformations of compounds bearing side

7-Ben zyl-1,2,3,4-t et r a h yd r oisoq u in olin e H yd r och lo-

r id e (28‚HCl). Amide 45 (0.22 g, 0.66 mmol) was dissolved

in EtOH (10 mL), and to this solution were added Pd/C (10%,

0.5 g) and HClO₄(70%, 0.2 mL). The suspension was hydro-

genated at 50 psi for 8 h at room temperature. The reaction

mixture was filtered through Celite and evaporated to dryness.

PNMT Inhibitors: Classical QSAR and CoMFA

J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 1 133

(13) Mekeyan, O.; Mercier, C.; Bonchev, D.; Dubois, J . E. Compara-

tive Study of DARC/PELCO Methods. II. Modelling PNMT

Inhibitory Potency of Benzylamines and Amphetamines. Eur.

J . Med. Chem. 1993, 28, 811-819.

(14) Fuller, R. W.; Molloy, B. B.; Hemrick, S. K. Inhibition in vitro

of Rabbit Adrenal Norepinephrine N-Methyltransferase by

2,3,4,5-Tetrahydro-1H-2-benzazepines. Biochem. Pharmacol. 1979,

28, 528-530.

chains were calculated by the “systematic search” option in

SYBYL to locate the global minimum energy conformation. If

possible, side chains were oriented such that they occupied

the same region of space, even though this necessitated the

use of a local minimum energy conformation (within 2 kcal/

mol of the corresponding global minimum.)

Molecules were aligned in orientations A (lipophilic) or B

(hydrophilic) and fit using three points: both ends of a normal

(2 Å long) passing through the centroid of the THIQ aromatic

ring and the end of the axial lone pair (2.4 Å long) on the THIQ

nitrogen. The CoMFA procedure was then implemented by

defining a three-dimensional lattice around the superimposed

molecules which had a grid spacing of 2 Å and extended beyond

the shape of all the molecules by a distance of at least 4 Å.

The electrostatic (Coulombic) and steric (van der Waals)

interactions were then sampledsusing an sp³-hybridized

carbon containing a unit positive charge as a probe atomsat

each intersection point of the lattice, except at those points

where the steric interaction energy indicated that the probe

atom was “inside” the aligned molecules (i.e., energy greater

than 30 kcal/mol). The CoMFA model was developed using a

partial least-squares (PLS) analysis, as implemented in SYBYL

6.2. Cross-validated (leave-one-out) runs were used to refine

the model and evaluate the number of components that yielded

the best r².

(15) Bondinell, W. E.; Chapin, R. W.; Girard, G. R.; Kaiser, C.; Krog,

A. J .; Pavloff, A. M.; Schwartz, M. S.; Silvestri, J . S.; Vaidya, P.

D.; Lam, B. L.; Wellman, G. R.; Pendleton, R. G. Inhibitors of

Phenylethanolamine N-Methyltransferase and Epinephrine Bio-

synthesis. 1. Chloro-Substituted 1,2,3,4-Tetrahydroisoquinolines.

J . Med. Chem. 1980, 23, 506-511.

(16) Singh, P. Quantitative Structure-Activity Relationship Studies

of 1,2,3,4-Tetrahydroisoquinoline Derivatives as Inhibitors of

Phenylethanolamine N-Methyltransferase. Indian J . Biochem.

Biophys. 1983, 20, 397-399.

(17) Bondinell, W. E.; Pendleton, R. G. 7- and/or 8-Sulfur Substituted

1,2,3,4-Tetrahydroisoquinoline Compounds. U.S. Patent 4,228,170,

1980.

(18) Grunewald, G. L.; Skjærbæk, N.; Monn, J . A. An Active Site

Model of Phenylethanolamine N-Methyltransferase Using CoM-

FA. In Trends in QSAR & Molecular Modelling 92; Wermuth,

C. G., Ed.; ESCOM Science: Leiden, 1993; pp 513-516.

(19) Carroll, F. I.; Gao, Y.; Rahman, M. A.; Abraham, P.; Parham,

K.; Lewin, A. H.; Boja, J . W.; Kuhar, M. J . Synthesis, Ligand

Binding, QSAR and CoMFA study of 3â-(p-Substituted phenyl)-

tropane-2â-carboxylic Acid Methyl Esters. J . Med. Chem. 1991,

34, 2719-2725.

(20) Debnath, A. K.; Hansch, C.; Kim, K. H.; Martin, Y. C. Mecha-

nistic Interpretation of the Genotoxicity of Nitrofurans (Anti-

bacterial Agents) Using Quantitative Structure-Activity Rela-

tionships and Comparative Molecular Field Analysis. J . Med.

Chem. 1993, 36, 1007-1016.

Molecu la r Gr a p h ics Stu d ies. Molecular modeling was

performed using the SYBYL software package (version 6.2,

Tripos Associates, Inc., St. Louis, MO) on a Silicon Graphics

Indigo²workstation.

(21) Grunewald, G. L.; Dahanukar, V. H.; Caldwell, T. M.; Criscione,

K. R. Examination of the Role of the Acidic Hydrogen in

Imparting Selectivity of 7-Aminosulfonyl-1,2,3,4-tetrahydroiso-

quinoline (SK&F 29661) Toward Inhibition of Phenylethanola-

mine N-Methyltransferase vs the R₂-Adrenoceptor. J . Med.

Chem. 1997, 40, 3997-4005.

(22) Cramer, R. D., III; Patterson, D. E.; Bunce, J . D. Comparative

Molecular Field Analysis (CoMFA). 1. Effect of Shape on Binding

of Steroids to Carrier Proteins. J . Am. Chem. Soc. 1988, 110,

5959-5967.

Ack n ow led gm en t. We thank Dr. Ki Kim (Abbott

Laboratories), Dr. Corwin Hansch (Pomona College),

and our University of Kansas colleagues (Dr. Karl-Heinz

Czaplinski, Mr. Christopher Gunn, and Mr. Timothy M.

Caldwell) for the suggestions and help with the QSAR

and molecular modeling studies. This research was

supported by USPHS Grant HL 34193.

(23) McCoubrey, A.; Mathieson, D. W. isoQuinolines. Part III. The

Nitration of 3:4-Dihydro- and 1:2:3:4-Tetrahydro-isoquinolines.

J . Chem. Soc. 1951, 2851-2853.

Refer en ces

(24) Ajao, A. F.; Bird, C. W. The Preparation and Oxidative Dimer-

ization of 2-Acetyl-7-hydroxy-1,2,3,4-tetrahydroisoquinoline. A

New Approach to Tetrahydroisoquinoline Synthesis. J . Hetero-

cycl. Chem. 1985, 22, 329-331.

(25) Grell, W.; Griss, E.; Kleeman, M.; Kutter, E. N-Acyl-7-(N"-

cyclohexylureido-N′-sulfonyl)-isoquinolines and -Benzazepines

and Alkali Metals Salts Thereof. U.S. Patent 3 725 388, 1973;

Chem. Abstr. 1971, 74, 99903j.

(26) Ali, F. E.; Gleason, J . G.; Hill, D. T.; Krell, R. D.; Kruse, C. H.;

Lavanchy, P. G.; Volpe, B. W. Imidodisulfamides. 2. Substituted

1,2,3,4-Tetrahydroisoquinolinylsulfonic Imides as Antagonists

of Slow-Reacting Substance of Anaphylaxis. J . Med. Chem. 1982,

25, 1235-1240.

(27) Bunting, J . W. Isoquinolines Containing Oxidized Nitrogen

Functions and their Hydrogenated Derivatives. In Isoquinolines;

Kathawala, F. G., Coppola, G. M., Shuster, H. F., Eds.; Wiley:

New York, 1990; Part 2, pp 479-532.

(28) Rao, G. C. Facile Hydration of Nitriles to Amides Using

Potassium Fluoride on Alumina. Synth. Commun. 1982, 12,

177-181.

(29) Kornblum, N.; Iffland, D. C. The Selective Replacement of the

Aromatic Primary Amino Group by Hydrogen in Aromatic-

Aliphatic Diamines. J . Am. Chem. Soc. 1949, 71, 2137-2143.

(30) Koch, R. C. Azabenzocycloalkanes-N-carboxamidines. Belg. Patent

635 308, 1964; Chem. Abstr. 1964, 61, 11978f.

(31) Gronowitz, S.; Kada, R. Sulfur Derivatives of 1-Methylpyrrole.

J . Heterocycl. Chem. 1984, 21, 1041-1043.

(32) Venier, C. G.; Squires, T. G.; Chen, Y.-Y.; Hussmann, G. P.; Shei,

J . C.; Smith, B. F. Peroxytrifluoroacetic Acid. A Convenient

Reagent for the Preparation of Sulfoxides and Sulfones. J . Org.

Chem. 1982, 47, 3773-3774.

(1) Taken in large part from the Ph.D. dissertation of Vilas H.

Dahanukar, University of Kansas, 1994.

(2) Blaschko, H. Catecholamine Biosynthesis. Br. Med. Bull. 1973,

29, 105-109.

(3) Gunne, L.-M. Relative Adrenaline Content in Brain Tissue. Acta

Physiol. Scand. 1962, 56, 324-333.

(4) Goldstein, M.; Lew, J . Y.; Matsumoto, Y.; Ho¨kfelt, T.; Fuxe, K.

Localization and Function of PNMT in the Central Nervous

System. In Pyschopharmacology: A Generation of Progress;

Lipton, M. A., DiMascio, A., Killam, K. F., Eds.; Raven: New

York, 1978; pp 261-269.

(5) Black, J .; Waeber, B.; Bresnahan, M. R.; Gavras, I.; Gavras, H.

Blood Pressure Response to Central and/or Peripheral Inhibition

of Phenylethanolamine N-Methyltransferase in Normotensive

and Hypertensive Rats. Circ. Res. 1981, 49, 518-524.

(6) Hieble, J . P.; McCafferty, J . P.; Roesler, J . M.; Pendleton, R. G.;

Gessner, G.; Carey, R.; Sarau, H. M.; Goldstein, M. Cardiovas-

cular Activity of SK&F 64139 in the Hypertensive Rat. J .

Cardiovasc. Pharmacol. 1983, 5, 889-897.

(7) Toomey, R. E.; Horng, J . S.; Hemrick-Leucke, S. K.; Fuller, R.

W. R₂-Adrenoceptor Affinity of Some Inhibitors of Norepineph-

rine N-Methyltransferase. Life Sci. 1981, 29, 2467-2472.

(8) Goldstein, M.; Saito, M.; Lew, J . Y.; Hieble, J . P.; Pendleton, R.

G. The Blockade of R₂-Adrenoceptors by the PNMT Inhibitor

SK&F 64139. Eur. J . Pharmacol. 1980, 67, 305-308.

(9) Pendleton, R. G.; Gessner, G.; Weiner, G.; J enkins, B.; Sawyer,

J .; Bondinell, W.; Intoccia, A. Studies on SK&F 29661, an Organ

Specific Inhibitor of Phenylethanolamine N-Methyltransferase.

J . Pharmacol. Exp. Ther. 1979, 208, 24-30.

(10) Fuller, R. W.; Molloy, B. B.; Day, W. A.; Roush, B. W.; Marsh,

M. M. Inhibition of Phenylethanolamine N-Methyltransferase

by Benzylamines. 1. Structure-Activity-Relationships. J . Med.

Chem. 1973, 16, 101-106.

(33) Creary, X. Reaction of Organometallic Reagents with Ethyl

Trifluoroacetate and Diethyl Oxalate. Formation of Trifluro-

methyl Ketones and R-Ketoesters via Stable Tetrahedral Ad-

ducts. J . Org. Chem. 1987, 52, 5026-5030.

(34) Frye, L. L.; Sullivan, E. L.; Cusack, K. P.; Funaro, J . M.

Sulfonylation of Organometallic Reagents with Arenesulfonyl

Fluorides: A Simple One-Step Synthesis of Sulfones. J . Org.

Chem. 1992, 57, 697-701.

(11) Otto, P.; Seel, M.; Ladik, J .; Mu¨ller, R. From Retrospective to

Predictive Structure Activity Correlations. J . Theor. Biol. 1979,

78, 197-210.

(12) Lukovits, I. Quantitative Structure-Activity Relationships Em-

ploying Independent Quantum Chemical Indices. J . Med. Chem.

1983, 26, 1104-1109.

134 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 1

Grunewald et al.

(35) Bianchi, T. A.; Cate, L. A. Phase Transfer Catalysis. Preparation

of Aliphatic and Aromatic Sulfonyl Fluorides. J . Org. Chem.

1977, 42, 2031-2032.

(36) Kotake, H.; Inomata, K.; Kinoshita, H.; Sakamoto, Y.; Kaneto,

Y. Halogenation of Alkyl p-Tolyl Sulfones with Carbon Tetra-

halides. Application to the Synthesis of Aromatic Ketones. Bull.

Chem. Soc. J pn. 1980, 53, 3027-3028.

(37) Yu, T.; Brown, L. E.; Beierwaltes, W. H.; Pendleton, R. G.

Synthesis of Inhibitors of Phenylethanolamine N-Methyltrans-

ferase. J . Labelled Compd. Radiopharm. 1979, 16, 173-173.

(38) Chalmers, J . R.; Dickson, G. T.; Elks, J .; Hems, B. A. The

Synthesis of Thyroxine and Related Substances. Part V. A

Synthesis of L-Thyroxine from L-Tyrosine. J . Chem. Soc. 1949,

3424-3433.

(39) Su, D.-B.; Duan, J .-X.; Chen, Q.-Y. Methyl Chlorodifluoroacetate.

A Convenient Trifluoromethylating Agent. Tetrahedron Lett.

1991, 32, 7689-7690.

(40) Bailey, D. M.; DeGrazia, C. G. Hydroxyguanidines. A New Class

of Antihypertensive Agents. J . Med. Chem. 1973, 16, 151-156.

(41) Blank, B.; Krog, A. J .; Weiner, G.; Pendleton, R. G. Inhibitors

of Phenylethanolamine N-Methyltransferase and Epinephrine

Biosynthesis. 2. 1,2,3,4-Tetrahydroisoquinoline-7-sulfonanilides.

J . Med. Chem. 1980, 23, 837-840.

(42) Sall, D. J .; Grunewald, G. L. Inhibition of Phenylethanolamine

N-Methyltransferase (PNMT) by Aromatic Hydroxy-Substituted

1,2,3,4-Tetrahydroisoquinolines: Further Studies on the Hy-

drophilic Pocket of the Aromatic Ring Binding Region of the

Active Site. J . Med. Chem. 1987, 30, 2208-2216.

(43) Adrenal PNMT has been found to be similar to the brain enzyme

in terms of susceptibility to inhibitors; see: Fuller, R. W.

Pharmacology of Brain Epinephrine Neurons. Annu. Rev. Phar-

macol. Toxicol. 1982, 22, 31-55 and references therein.

(44) Connett, R. J .; Kirshner, N. Purification and Properties of Bovine

Phenylethanolamine N-Methyltransferase. J . Biol. Chem. 1970,

245, 329-334.

(45) Grunewald, G. L.; Borchardt, R. T.; Rafferty, M. F.; Krass, P.

Conformationally Defined Adrenergic Agents. 5. Conformational

Preferences of Amphetamine Analogues for Inhibition of Phe-

nylethanolamine N-Methyltransferase. Mol. Pharmacol. 1981,

20, 377-381.

(49) As will be noted later in the CoMFA section, the best CoMFA

models were obtained by using the parameters for the neutral

forms of compounds 13 and 14. However, in the QSAR, the r

values showed significant reduction if the neutral forms were

used (eq 1, r ) 0.853 vs r ) 0.730 if neutral forms used; eq 3, r

) 0.826 vs r ) 0.800 if neutral forms used).

(50) (a) The parameters surveyed for both QSAR analyses were π,

σ_m, MR, hydrogen bond acceptor, hydrogen bond donor, ×c1, L,

B₁, and B₅. (b) Coburn, R. A. Quantitative Structure-Activity

Relationship Studies: Physicochemical-Activity Relationships

Methods (QSAR-PC: PAR); BIOSOFT: Cambridge, U.K., 1987.

(51) Hansch, C.; Leo, A. J .; Hoekman, D. In Exploring QSAR.

Hydrophobic, Electronic and Steric Constants; American Chemi-

cal Society: Salem, MA, 1995.

(52) Franke, R. QSAR Parameters. In QSAR and Strategies in the

Design of Bioactive Compounds; Seydel, J . K., Ed.; VCH:

Weinheim, 1985; pp 59-78.

(53) (a) An increase in the correlation coefficient of eq 2 (PNMT) (r

) 0.912 vs r ) 0.885) could be obtained through the inclusion of

an indicator variable (I) to account for the higher activity of

lipophilic compounds that could bind in orientation A. As shown

in Figure 1, all 7-substituted-THIQs bearing a lipophilic (+π)

substituent which may bind in orientation A were given a value

of I ) 1, while those bearing hydrophilic (-π) substituents which

may bind in orientation B were given a value of I ) 0. However,

the inclusion of I was not justified statistically (F_(1,22)) 6.36 vs

F_(1,22,0.01)) 7.94). (b) Addition of an indicator variable in the R₂-

adrenoceptor equation (eq 4) did not increase the correlation

(0.917 vs 0.917) and was therefore not included in the derivation.

(54) Attempts at using the selectivity (R₂K_i/PNMT K_i) as the

dependent variable with the parameters π, σ_m, and MR showed

that only the single-variable equation with σ_mwas statistically

significant [log(selectivity) ) 2.05(( 1.48)σ_m- 0.686(( 0.552);

n ) 28, r ) 0.487, s ) 0.955, F_(1,26)) 8.10, F_(1,26,0.01)) 7.72].

This equation explains only 24% of the variance in the data and

would indicate that both sites cannot be described by a single

equation.

(46) U′Prichard, D. C.; Greenberg, D. A.; Snyder, S. H. Binding

Characteristics of a Radiolabeled Agonist and Antagonist at

Central Nervous System Alpha Noradrenergic Receptors. Mol.

Pharmacol. 1977, 13, 454-473.

(47) Grunewald, G. L.; Dahanukar, V. H.; Ching, P.; Criscione, K.

R. Effect of Ring Size or an Additional Heteroatom on the

Potency and Selectivity of Bicyclic Benzylamine-type Inhibitors

of Phenylethanolamine N-Methyltransferase. J . Med. Chem.

1996, 39, 3539-3546.

(48) Grunewald, G. L.; Monn, J . A.; Sall, D. J . Approaches to the

Design of Selective Inhibitors of Phenylethanolamine N-Meth-

yltransferase: Development of an Active Site Model. In Epi-

nephrine in the Central Nervous System; Stolk, J . M., U′Prichard,

D. C., Fuxe, K., Eds.; Oxford: New York, 1988; pp 117-140.

(55) This suggests that the use of the flipped orientation might have

been of value in the QSAR analyses. However, there is only one

such compound in the data set, and using an indicator variable

for it would be of limited usefulness in the QSAR analyses.

(56) Still, W. C.; Mitra A.; Kahn, M. Rapid Chromatographic Tech-

nique for Preparative Separations with Moderate Resolution. J .

Org. Chem. 1978, 43, 2923-2924.

(57) Grunewald, G. L.; Sall, D. J .; Monn, J . A. Synthesis and

Evaluation of 3-Substituted Analogues of 1,2,3,4-Tetrahydroiso-

quinoline as Inhibitors of Phenylethanolamine N-Methyltrans-

ferase. J . Med. Chem. 1988, 31, 824-830.

J M980429P

Article Doi

Synthesis, biochemical evaluation, and classical and three-dimensional quantitative structure-activity relationship studies of 7-substituted- 1,2,3,4-tetrahydroisoquinolines and their relative affinities toward phenylethanolamine N-methyltransferase and t

DOI: 10.1021/jm980429p

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jm980429p

Full text of DOI:10.1021/jm980429p

Products guided by the article

R&D Labs maybe for 220247-62-1

Relevant to this article

Hot Product