β1- and β2-adrenoceptor antagonist activity of a series of para- substituted N-isopropylphenoxypropanolamines

Article abstract of DOI:10.1016/S0223-5234(99)00114-2

To further explore the structure-activity relationships of β- adrenoceptor (β-AR) antagonists, a series of 25 para-substituted N- isopropylphenoxy-propanolamines were synthesised, nine of which are new compounds. All have been examined for their ability to antagonise β₁-ARs in rat atria and β₂-ARs in rat trachea. Substitution in the para-position of the phenyl ring is thought to confer β₃-specificity and the selectivity of these compounds for the β₁-AR ranges from 1.5-234. None of the compounds tested were selective for the β₂-AR. Of the 25 compounds studied, 22 had reasonable (pA₂ > 7) potencies for the rat β₁-AR. Only compound 1 displayed reasonable (pA₂ > 7) potency for the rat β₂-AR. Twenty two compounds were used as the training set for comparative molecular field analysis (CoMFA) of antagonist potency (pA₂) at the rat β₁- and β₂-ARs. The inclusion of a number of additional physical characteristics improved the QSAR analysis over models derived solely using the CoMFA electrostatic and steric fields. The final models predicted the β₁- and β₂-AR potency of the compounds in the training set with high accuracy (r² = 0.93 and 0.86 respectively). The final β₁-AR model predicted the β₁-potencies of two out of the three test compounds, not included in the training set, with residual PA₂ values < -0.14, whereas the test compounds were not as well predicted by our final β₂-AR model (residual pA₂ values < -0.38).

Full text of DOI:10.1016/S0223-5234(99)00114-2

Eur. J. Med. Chem. 34 (1999) 919−937
© 1999 Éditions scientifiques et médicales Elsevier SAS. All rights reserved
919
Original article
â1- and â2-Adrenoceptor antagonist activity of a series
of para-substituted N-isopropylphenoxypropanolamines
Simon N. Louis*, Tracy L. Nero, Dimitri Iakovidis, Felicia M. Colagrande, Graham P. Jackman,
William J. Louis
Clinical Pharmacology and Therapeutics Unit, The University of Melbourne, Department of Medicine, Austin and Repatriation Medical
Centre Heidelberg 3084, Victoria, Australia
(Received 8 July 1998; revised 14 April 1999; accepted 20 May 1999)
Abstract – To further explore the structure-activity relationships of â-adrenoceptor (â-AR) antagonists, a series of 25 para-substituted
N-isopropylphenoxy-propanolamines were synthesised, nine of which are new compounds. All have been examined for their ability to
antagonise â1-ARs in rat atria and â2-ARs in rat trachea. Substitution in the para-position of the phenyl ring is thought to confer â3-specificity
and the selectivity of these compounds for the â1-AR ranges from 1.5–234. None of the compounds tested were selective for the â2-AR. Of
the 25 compounds studied, 22 had reasonable (pA2 > 7) potencies for the rat â1-AR. Only compound 1 displayed reasonable (pA2 > 7) potency
for the rat â2-AR. Twenty two compounds were used as the training set for comparative molecular field analysis (CoMFA) of antagonist
potency (pA2) at the rat â1- and â2-ARs. The inclusion of a number of additional physical characteristics improved the QSAR analysis over
models derived solely using the CoMFA electrostatic and steric fields. The final models predicted the â1- and â2-AR potency of the compounds
in the training set with high accuracy (r2 = 0.93 and 0.86 respectively). The final â1-AR model predicted the â1-potencies of two out of the
three test compounds, not included in the training set, with residual pA2 values < –0.14, whereas the test compounds were not as well predicted
by our final â2-AR model (residual pA2 values < –0.38). © 1999 Éditions scientifiques et médicales Elsevier SAS
comparative molecular field analysis / rat â-adrenoceptors / â-adrenoceptor antagonists / N-isopropylphenoxypropanolamines /
quantitative structure activity relationships
1. Introduction
â-Adrenoceptors (â-ARs) are members of the large
family of G-protein coupled receptors [1–4]. It has been
established so far that there are at least three â-AR
subtypes, designated the â1-, â2- and â3-ARs [5–7].
Figure 1. General structure of para-substituted N-isopropyl-
N-Isopropylphenoxypropanolamine (figure 1, R = H) is
considered to be a non-selective â-AR antagonist. Many
structure-activity studies in the past have focused upon
the N-isopropylphenoxypropanolamine core structure
with substitutions in the ortho and/or meta positions [8,
9]. However, the effects of para-substituents upon the
â-blocking activity of this core structure have received
phenoxypropanolamines. The torsion angles τ1 and τ2 define
the conformation of the para-substituent and the oxypropano-
lamine side-chains respectively in relation to the phenyl ring,
for example when τ1 = 0 the para-substituent is planar to the
phenyl ring.
less attention [10–12]. Previous work by us [10, 13] and
others [11, 12, 14, 15] has established that high â1-AR
selectivity and potency can be achieved by para-
substitution of the phenyl ring and/or appropriate substi-
tution of the phenoxypropanolamine amino group (e.g.
-ethyl-3-(4-hydroxyphenyl)urea [13]).
*Correspondence and reprints
Abbreviations: Aryloxypropanolamine, AOPA; â-adrenoceptors,
â-ARs; comparative molecular field analysis, CoMFA; ESP phenyl
ring charge, ESP; flexibility of the para-substituent, flexibility;
length of the para-substituent, length; optimal number of compo-
nents, ONC; partial least squares, PLS; standard error, SE.

920
Figure 2. General procedure for preparing N-isopropylphenoxypropanolamines. Reagents: (i) aqueous KOH, ETOH; (ii) isopropy-
lamine, ETOH.
To further examine the effect of para-substitution upon
both â1- and â2-AR selectivity and potency, we have
synthesised a series of para-substituted N-isopropyl-
phenoxypropanolamines. Nine of these compounds are
new (6, 7, 9, 10, 15, 16, 18, 20 and 22), that is to say that
a chemical abstracts structure database search indicated
that these compounds have not been previously reported.
Eleven previously reported compounds (1, 2, 3, 4, 5, 8,
12, 14, 19, 23 and 25) and five commercially available
compounds (11 (metoprolol), 13 (H 87/07), 17 (betax-
olol), 21 (RO 31-1118) and 24 (cicloprolol)) have also
been synthesised and included in the study in order to
obtain comparative pharmacological data in the rat. All of
these compounds have been examined in our laboratory
for their ability to antagonise â1-ARs in rat atria and
â2-ARs in rat trachea, which are well established as
sources of the â1- and â2-AR subtypes respectively [16].
The synthesis and structure-activity relationships, using
comparative molecular field analysis (CoMFA), of these
para-substituted N-isopropylphenoxypropanolamines are
presented here.
been established for simple phenoxypropanolamines that
the S-isomer is the active isomer, with little â-AR activity
residing with the R-isomer [17–19]. Resolution of the
racemate into the individual isomers or stereospecific
synthesis was therefore not carried out. The correspond-
ing phenols A reacted under aqueous alkaline conditions
with epichlorohydrin to produce the epoxides B. After
isolation, the crude epoxides B were allowed to react with
isopropylamine overnight to furnish the desired com-
pounds C.
Compounds 1, 2, 3, 4, 5, 8, 11 (metoprolol), 12, 13 (H
87/07), 14, 17 (betaxolol), 22, 21 (RO 31-1118), 23, 24
(cicloprolol) and 25, have been previously reported and
were prepared following literature procedures [20–33].
All intermediates and final compounds had virtually
identical physical and chemical data with those reported.
The precursor phenols of compounds 6, 10, 15 and 16
were prepared by reacting bromoethane, n-bromo-
propane, 4-cyclohexylethyl bromide and 4-fluorophenyl-
ethyl bromide with 4-benzyloxyphenol, thus producing
the benzyl ethers D which were subsequently hydroge-
nated giving the desired phenols E (4-ethoxyphenol), F
(4-n-propoxyphenol), G (4-cyclohexylethoxyphenol) and
H (4-fluorophenethoxyphenol) as shown in figure 3.
2. Chemistry
All compounds were prepared as their racemic mix-
tures by the general procedure shown in figure 2. It has
The precursor phenols of compounds 18, 20 and 22
were prepared by the sequence of reactions shown in
Figure 3. Reagents: (i) K2CO3/NaI, anhydrous acetone; (ii) H2, 10% Pd/C.

921
Figure 4. Reagents: (i) ClCH2COOH, NaH, DMF; (ii) LiAlH4, THF; (iii) pTsCl, pyridine; (iv) NaH, DMF; (v) H2, 10% Pd/C.
figure 4. The corresponding alcohols 2-methylpropane-1-
ol, 2-hydroxymethyltetrahydrofuran and n-pentane-1-ol
reacted with chloroacetic acid to furnish the correspond-
ing alkoxy acetic acids I. These were reduced to their
corresponding alkoxy-ethanols J whose tosylates K re-
acted with 4-benzyloxyphenol, to produce the benzyl
ethers L which were subsequently hydrogenated giving
the desired phenols M, N and O.
lation, was reduced with sodium amalgam to produce the
desired phenol Q.
The precursor phenol of compound 9 was prepared
according to the reactions shown in figure 6. Ethyl-4-
hydroxyphenyl propionate was benzylated to produce the
ether R, which was subsequently reduced to give the
protected alcohol S. Hydrogenation of S produced the
desired phenol T.
The precursor phenol of compound 7 was prepared by
the reactions shown in figure 5 [34, 35]. Phenyl-n-
butanoate reacted with aluminium chloride to furnish a
mixture of regioisomers P (2- and 4-hydroxypropio-
phenone), of which 4-hydroxypropiophenone, after iso-
The above mentioned phenols were then used to
produce compounds 6, 7, 9, 10, 15, 16, 18, 20 and 22
following the general procedure shown in figure 2. The
general physical data for these compounds are given in
table I.
Figure 5. Reagents: (i) anhydrous AlCl3; (ii) Zn-Hg, MeOH, HCl.

922
Figure 6. Reagents: (i) BzBr, K2CO3, NaI, anhydrous acetone; (ii) LiALH4, THF; (iii) H2, 10% Pd/C.
3. Pharmacology
SE from 4–9 individual experiments. The â-AR antago-
nist potencies and the subtype selectivity are given in
table II.
All compounds were evaluated for their ability to
antagonise â1- and â2-ARs in rat atria and tracheal rings
respectively. Cumulative concentration-response curves
were obtained in each preparation as described by Van
Rossum [36] and curves were fitted by computer analysis
according to the method of Zabrowsky et al. [37] using
the sigmoidal fit function of the Origin graphics pack-
age [38]. The antagonist potencies, or pA2 values, were
calculated using equation 1 [39] and represent the mean ±
antagonist
dose ratio − 1
͓
ͩ ͪ
͔
pA2 = –log
(1)
where the dose ratio =
(-)-isoprenaline at EC50 in the presence of antagonist
͓
͔
(-)-isoprenaline at EC50 in the absence of antagonist
͓
͔
Table I. Physical data for compounds 6, 7, 9, 10, 15, 16, 18, 20 and 22.
Compound
Yield
(from phenol)
M.p. (°C)
HCl salta
Formulab
6
3.66 g
(63.4 %)
4.35 g
(72.1 %)
2.41 g
(39.7 %)
3.56 g
(58.6 %)
4.91 g
(66.3 %)
11.71 g
(61.4 %)
4.00 g
(55.3 %)
3.59 g
(48.4 %)
3.90g
87.5–88.5
70.5–73.0
C14H23NO3.HCl
C16H27NO2.HCl
C15H25NO3.HCl
C15H25NO3.HCl
C20H33NO3.HCl
C20H26FNO3.HCl
C18H31NO4.HCl
C11H31NO5.HCl
C19H33NO4.HCl
7
9
68.5–71.5
10
15
16
18
20
22
116.5–119.0
131.0–133.0
136.0–138.0
89.5–91.3
165.5–167.0
77.0–79.0
(52.1 %)
aEthanol-ether was the solvent system used for recrystallisation. bEmpirical formula; elemental analyses were carried out for the nine
compounds shown and were within 0.4% per element.

923
Table II. Antagonist activity, subtype selectivity, flexibility of para-substituent, π, ESP phenyl ring charge at 90° and 0° and length of the
para-substituent for the compounds examined in this study.
Compound R Group
â1-ARa pA2
â2-ARb pA2
â1-AR
Flexi-
πe
ESP phenyl ESP phenyl Lengthf
ring charge ring charge (Å )
selectivityc bilityd
at 90o (e)
at 0o (e)
Training Compounds
1
8.09 ± 0.14
6.60 ± 0.18
7.13 ± 0.04
6.58 ± 0.19
7.41 ± 0.13
7.08 ± 0.27
7.49 ± 0.26
7.38 ± 0.19
7.13 ± 0.10
7.16 ± 0.14
7.60 ± 0.17
7.47 ± 0.14
6.43 ± 0.17
6.47 ± 0.11
6.34 ± 0.09
6.56 ± 0.18
5.94 ± 0.16
6.23 ± 0.23
5.39 ± 0.16
6.58 ± 0.34
5.72 ± 0.38
6.43 ± 0.13
4.2
1.5
4.6
1.7
7.1
14
0
0
0
1
2
2
3
3
3
3
3
0.00
–0.86
0.47
–0.04
1.37
0.36
1.82
–0.09
0.06
0.84
0.11
–0.83
–0.83
0.00
1.41
1.53
2.30
3.84
3.61
5.00
5.06
4.86
4.79
4.86
2
–0.67
–0.71
–0.68
–0.76
–0.71
–0.72
–0.75
–0.72
–0.68
–0.72
–0.67
–0.71
–0.72
–0.72
–0.74
–0.68
–0.69
–0.69
–0.72
–0.69
3
4
5
6
7
18
8
98
9
3.5
28
10
11
15
12
13
14
7.98 ± 0.21
7.37 ± 0.09
6.67 ± 0.09
6.77 ± 0.28
5.97 ± 0.19
5.36 ± 0.18
16
25
20
4
4
2
0.57
–0.13
2.06
–0.74
–0.68
–0.68
–0.70
–0.72
–0.69
6.14
5.91
5.08
15
16
7.65 ± 0.12
8.01 ± 0.12
5.93 ± 0.23
5.64 ± 0.12
52
4
4
2.57
2.13
–0.66
–0.70
–0.70
–0.73
6.45
8.72
234
17
8.06 ± 0.15
6.38 ± 0.13
48
5
0.96
–0.73
–0.70
7.39

924
Table II. Continued
Compound R Group
â1-ARa pA2
â2-ARb pA2
â1-AR
Flexi-
πe
ESP phenyl ESP phenyl Lengthf
ring charge ring charge (Å )
selectivityc bilityd
at 90o (e)
at 0o (e)
Training Compounds
18
7.75 ± 0.22
7.61 ± 0.15
5.69 ± 0.20
5.40 ± 0.18
115
162
6
5
1.13
1.76
–0.73
–0.70
8.54
9.51
19
20
–0.72
–0.69
–0.69
–0.73
7.74 ± 0.09
5.49 ± 0.27
178
6
0.32
9.36
21
22
7.98 ± 0.21
8.04 ± 0.23
5.68 ± 0.10
5.80 ± 0.24
200
174
7
8
2.05
1.66
–0.72
–0.72
–0.73
–0.72
12.4
10.8
Test Compounds
23
7.22 ± 0.08
7.73 ± 0.07
6.47 ± 0.17
5.44 ± 0.10
5.6
1
6
0.92
0.72
–0.77
–0.73
–0.75
–0.70
2.50
8.59
24
25
195
8.04 ± 0.30
5.81 ± 0.25
170
5
1.31
–0.74
–0.70
8.36
aâ1-AR antagonist pA2 value ± SEM determined in isolated spontaneously beating rat atria. bâ2-AR antagonist pA2 value ± SEM determined
c
d
in rat tracheal chain preparation. Selectivity = antilog (pA2 â1-AR – pA2 â2-AR). Flexibility was defined as the number of fully rotatable
non-H bonds in the para-substituent. eπ was logP of the compound in question minus logP of compound 1, logP was calculated using PrologP.
fLength was defined as the distance between the para-carbon atom of the phenyl ring and the most distant non-H atom.
4. Results and discussion
6.77) was the most potent of the para-substituted com-
pounds at inhibiting rat â2-ARs (c.f. reference compound
1, pA2 = 7.47).
4.1. Isolated tissue preparations
Overall, the compounds had higher potencies for the
â1-AR than the â2-AR, with the â1/â2 selectivities
ranging from 1.5–234. Compounds 1, 2, 3, 4, 9 and 23,
however, were at the lower end of the selectivity scale
and are considered to be non-selective in the rat.
Functional potencies of compounds 1–25 for inhibiting
(-)-isoprenaline-induced: (i) â1-AR mediated chronotro-
pic effects in spontaneously beating rat atria; and (ii)
â2-AR mediated relaxation of rat tracheal chain previ-
ously contracted with 1 µM carbachol are listed in
table II. The unsubstituted reference compound 1 had a
high potency at both the â1-AR (pA2 = 8.09) and â2-AR
(pA2 = 7.47). Para-substitution reduced the potency of
the compounds for both â1- and â2-ARs (compounds
2–11, 13–15, 18, 19, 20, 23 and 24; table II), except for
compounds 12, 16, 17, 21, 22 and 25 (â1-AR pA2s =
7.98–8.06) which had similar â1-AR potencies to the
unsubstituted compound 1. Compound 12 (â2-AR pA2 =
The animal species used in this study to determine both
the â1- and â2-AR functional potency of the compounds
was the rat. Previously published â1-AR antagonist
functional potency, or activity data has been obtained
from both the rat and guinea-pig [30, 40], whereas the
guinea-pig is the commonly used animal species for
published â2-AR antagonist activity [30, 40]. By using
the rat as the source for both â1- and â2-ARs, any

925
Figure 7. (a) Structural alignment of the training set of compounds (compounds 1–22) when τ1 = 90°. (b) Structural alignment of the
training set of compounds when τ1 = 0°.
interspecies differences between the receptor subtypes
was removed.
the superimposition of the compounds when τ1 = 90° and
figure 7b when τ1 = 0°.
4.2. Alignment of molecules
4.3. CoMFA analysis for antagonist potency at the rat
â1-AR
The alignment of the compounds is the most important
feature of CoMFA analysis [41]. It has been proposed that
AOPA type compounds interact with â-ARs via a three
point pharmacophore consisting of the â-hydroxyl group,
the amino group and an electron rich moiety which is
usually a phenyl ring [42, 43] and more recent site-
directed mutagenesis studies support this hypoth-
esis [44–47]. If we assume the compounds all act via the
same mechanisms, then these common points of interac-
tion must superimpose and hence we have assumed that
the AOPA core structure remains fixed for all compounds
in our conformational analysis.
The CoMFA (SYBYL version 6.40) results for antago-
nist potency at rat â1-ARs for the twenty-two compounds
in the training set are given in table III. A range of
column filtering values were used (1.0–16.0 kcal/mol).
Column filtering omits from the analysis columns, lattice
points whose variance is less than the specified value
(SYBYL). Partial least squares analysis (PLS) of the data
identified that when s1 was set to 90° (figure 1) and the
default CoMFA settings were used, which included both
steric and electrostatic fields, the highest cross validated
r2 (q2) value was obtained when column filtering was set
to 4.0 kcal/mol (q2 = 0.24; optimum number of compo-
nents (ONC) = 4; table III). When s1 = 0°, q2 was
maximised when column filtering was set at 5.0 kcal/mol
(q2 = 0.38 and ONC = 2; table III).
Previous studies within our laboratory [10] examined
two conformations of each molecule ie. τ1 = 90° or 0°. In
the present study we have examined the effect of these
values of τ1 on our CoMFA analysis. Figure 7a displays

926
Table III. CoMFA results for rat â1-ARs (n = 22).
Analysis variables
τ1
Column filtering
q2
ONC
SE
CoMFA
90o
90o
90o
90o
90o
90o
90o
90o
90o
90o
90o
90o
90o
90o
90o
90o
0o
4
4
13
4
4
9
4
4
13
13
4
9
10
4
5
9
5
5
5
5
5
5
5
5
5
5
5
5
2
5
5
5
0.244
0.527
0.747
0.322
0.186
0.686
0.375
0.398
0.672
0.638
0.247
0.655
0.632
0.259
0.636
0.610
0.384
0.592
0.604
0.524
0.553
0.657
0.571
0.515
0.618
0.601
0.407
0.635
0.599
0.482
0.563
0.564
4
3
10
6
5
4
4
4
10
10
5
4
5
4
7
5
2
2
4
2
2
2
2
3
2
4
3
2
6
3
4
6
0.461
0.354
0.332
0.465
0.493
0.331
0.419
0.411
0.358
0.388
0.474
0.311
0.332
0.457
0.352
0.341
0.394
0.320
0.334
0.346
0.335
0.294
0.328
0.359
0.310
0.335
0.397
0.303
0.357
0.371
0.351
0.373
CoMFA, flexibilitya
CoMFA, ESPb,c
CoMFA, lengthd
CoMFA, π
CoMFA, flexibility, ESP
CoMFA, flexibility, length
CoMFA, flexibility, π
CoMFA, ESP, length
CoMFA, ESP, π
CoMFA, length, π
CoMFA, flexibility, ESP, length
CoMFA, flexibility, ESP, π
CoMFA, flexibility, length, π
CoMFA, ESP, length, π
CoMFA, flexibility, ESP, length, π
CoMFA
CoMFA, flexibilitya
CoMFA, ESPb
0o
0o
CoMFA, lengthc
0o
CoMFA, π
0o
CoMFA, flexibility, ESP
CoMFA, flexibility, length
CoMFA, flexibility, π
CoMFA, ESP, length
CoMFA, ESP, π
0o
0o
0o
0o
0o
CoMFA, length, π
0o
CoMFA, flexibility, ESP, length
CoMFA, flexibility, ESP, π
CoMFA, flexibility, length, π
CoMFA, ESP, length, π
CoMFA, flexibility, ESP, length, π
0o
0o
0o
0o
0o
aFlexibility of the para-substituent. bBold indicates the options used for the final model. cESP phenyl ring charge. dLength of the
para-substituent.
In addition to the PLS analysis of the CoMFA data,
several other physical parameters of the compounds were
determined and included in further CoMFA analysis.
These parameters included the flexibility, length and π of
the para-substituent and the ESP phenyl ring charge
(table II). The flexibility of the para-substituent (flexibil-
ity) was determined by assigning an integer value for the
number of torsion angles which affected the conformation
of the substituent. The term ‘ESP phenyl ring charge’
(ESP) was defined in this study as the sum of the six
phenyl ring carbon atom esp charges. The phenyl ring
charges of all the compounds studied (table II) were
calculated using the AM1 Hamiltonian within MOPAC
(version 6.0), specifying the key words ‘ESP’, ‘PRE-
CISE’ and ‘NOMM’. The length of the para-substituent
(length) was defined as the distance between the para-
carbon of the phenyl ring and the most distant non-H
atom, hence the para-substituent length of compound 1 is
zero (table II). The π value of the para-substituent was
defined as the logP of the compound minus the logP of
compound 1, logP having been calculated using Pallas
PrologP (version 1.1). When these additional parameters
were included with CoMFA they generally improved q2.
For instance, when τ1 = 90° a combination of the CoMFA
fields and ESP gave the highest q2 (q2 = 0.769; ONC =
11; SE = 0.332; table III), and when τ1 = 0° a combina-
tion of the CoMFA fields, ESP and flexibility maximised
q2 (q2 = 0.657; ONC = 2; SE = 0.29; table III). The
former model was selected for further analysis as it
possessed the highest q2 of any of the â1-AR models
examined (table III).

927
Figure 8. (a) Plot of the actual â1-AR pA2 verses the predicted â1-AR pA2 compounds in table II (n = 22) using the final model, non
cross validated, and an optimum number of components of 4. (b) Plot of the actual â2-AR pA2 verses the predicted â2-AR pA2
compounds in table II (n = 22) using the final model, non cross validated, and an optimum number of components of 4.
The final model, without cross validation, was obtained
using the default CoMFA options, 10 ONC and 13 kcal/
mol column filtering, τ1 = 90o, including ESP. This model
had an r2 value of 0.95, a standard error (SE) of 0.15 and
an F (11, 10) of 21.47. The regression equation for ESP,
taken from the PLS output file, was:
sured pA2 values for the non-cross validated analysis and
residual values are given in table IV.
4.4. CoMFA analysis for antagonist potency at the rat
â2-AR
â1-pA2 = 0.995–8.49 ESP
Similarly, the default CoMFA settings and a range of
column filtering values (1.0–16.0 kcal/mol) were used
when analysing the â2-AR data. When τ1 = 90o the
highest q2 value was obtained when column filtering was
The relative contributions of the components were:
steric, 0.88; electrostatic, 0.00; and ESP, 0.12. Figure 8a
displays the relationship between calculated and mea-

928
Table IV. CoMFA Experimental, calculated and residual activities
of the training set of compounds for â1-ARs.
column filtering values of 5 kcal/mol when τ1 = 90° and
6 kcal/mol when τ1 = 0°.
Compound
Experimental Calculated
Residual
1
2
3
4
5
6
7
8
8.09
6.60
7.13
6.58
7.41
7.08
7.49
7.38
7.13
7.16
7.60
7.98
7.37
6.67
7.65
8.01
8.06
7.75
7.61
7.74
7.98
8.04
8.04
6.68
7.02
6.73
7.48
6.96
7.31
7.58
7.12
7.19
7.56
7.95
7.43
6.68
7.61
8.03
8.06
7.75
7.80
7.52
8.03
8.01
0.05
–0.08
0.11
–0.15
–0.07
0.12
0.18
–0.20
0.01
–0.03
0.04
0.03
–0.06
–0.01
0.04
–0.02
0.00
4.5. Comparison of contour maps
Figures 9A and 9B show compound 16 binding within
the steric contour maps for the final â1- and â2−AR
models. The experimental data suggests that para-
substitution decreases antagonist potency at both â1- and
â2-ARs. However, the contour maps, of the â1- and
â2-AR pharmacophores display important differences,
the most striking of which appears to be a bulk preferring
pocket in the â1-AR which can be accessed by long
flexible para-substituents. Compounds that access this
pocket have their â1-AR potency restored and the longer
the substituent the greater the restoration of their activity,
compare the activity of compounds 4, 13 and 22. Com-
pounds with shorter bulky substituents, like compound 14
cannot access this pocket and have low â1-AR potencies.
The â2-AR, on the other hand, seems far more sterically
restricted than the â1-AR in the para-position. For
instance, the addition of the short bulky substituent of
compound 14 reduces the â1-AR potency by 1.42 log
units when compared to the base compound, whereas this
substituent reduces â2-AR potency by 2.11 log units.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
0.00
–0.19
0.22
0.05
0.03
set to 4 kcal/mol (q2 = 0.50; ONC = 5; table V). When τ1
= 0°, q2 was maximised when column filtering was set to
5 kcal/mol (q2 = 0.45; ONC = 2; table V). When the
additional parameters were included in the analysis the
highest q2 occurred when τ1 = 90o, column filtering was
set at 6 kcal/mol and both flexibility and ESP were
considered in addition to the steric and electrostatic fields
generated within CoMFA (q2 = 0.60; ONC = 4; SE =
0.38).
The final model, without cross-validation, for the
â2-AR was obtained using 4 ONC, column filtering set at
6 kcal/mol and included flexibility and ESP. This model
had an r2 value of 0.83, SE of 0.25 and an F (4, 17) of
20.71. The regression equation for flexibility and ESP
was:
Using compound 16 as an example the differences
between the two steric maps can be clearly illustrated. In
the â1-AR model (figure 9A), the para-substituent of
compound 16 does not impinge on the sterically restricted
region (ie. yellow region), however, the terminal fluoro-
phenyl ring is embedded in the bulk preferring pocket
(i.e. green region) and hence the â1-potency is almost
restored to that of compound 1 (table II). By contrast, the
para-substituent of compound 16 is clearly embedded in
the sterically restricted region of the â2-AR model
(figure 9B) and hence this compound has a much lower
â2-pA2 than compound 1 (table II).
4.6. Predictive abilities of the CoMFA models
â2-pA2 = 2.05 + 0.07 flexibility–6.46 ESP.
The relative contributions of the components were:
steric, 0.73; electrostatic, 0.00; flexibility, 0.11; and ESP,
0.161. Figure 8b displays the relationship between the
calculated and measured pA2 values for the cross vali-
dated analysis and residual values are given in table VI.
It is important to note that by using the column filtering
values to maximise the q2 for the â1- and â2-AR models
(column filtering = 13 and 6 kcal/mol, respectively) it
effectively removes the electrostatic field parameter from
the CoMFA analysis. For both models this occurs at
The potencies of three test compounds (table VII), not
in the training set, were predicted at rat â1- and â2-ARs
using the final models and the predict property function in
the QSAR package of SYBYL (version 6.40). For the
â1-AR the predicted potencies for two of the compounds
were in close agreement with those obtained experimen-
tally (residuals < 0.15). For the â2-AR, however, all three
compounds were relatively poorly predicted (residuals
–0.25 to –0.38) suggesting that our CoMFA model of the
â2-AR is not sufficiently resolved.

929
Table V. CoMFA results for rat â2-ARs (n = 22).
Analysis variables
τ1
Column filtering
q2
ONC
SE
CoMFA
90o
90o
90o
90o
90o
90o
90o
90o
90o
90o
90o
90o
90o
90o
90o
90o
0o
4
4
6
4
4
6
4
4
4
5
4
5
7
4
12
12
5
5
2
5
5
2
5
7
2
1
7
2
1
1
1
1
0.501
0.574
0.586
0.499
0.364
0.602
0.498
0.387
0.575
0.454
0.284
0.587
0.430
0.372
0.473
0.455
0.445
0.462
0.453
0.463
0.351
0.454
0.479
0.538
0.459
0.386
0.494
0.491
0.387
0.440
0.451
0.475
5
4
4
6
4
4
4
5
5
3
3
4
5
3
2
2
2
4
5
4
5
5
4
9
5
6
9
5
6
10
6
6
0.434
0.389
0.383
0.449
0.475
0.376
0.422
0.481
0.400
0.428
0.490
0.383
0.464
0.459
0.409
0.416
0.420
0.437
0.454
0.437
0.495
0.454
0.430
0.482
0.452
0.497
0.505
0.438
0.497
0.555
0.470
0.460
CoMFA, flexibilitya
CoMFA, ESPb
CoMFA, lengthc
CoMFA, π
CoMFA, flexibility, ESPd
CoMFA, flexibility, length
CoMFA, flexibility, π
CoMFA, ESP, length
CoMFA, ESP, π
CoMFA, length, π
CoMFA, flexibility, ESP, length
CoMFA, flexibility, ESP, π
CoMFA, flexibility, length, π
CoMFA, ESP, length, π
CoMFA, flexibility, ESP, length, π
CoMFA
CoMFA, flexibilitya
CoMFA, ESPb
0o
0o
CoMFA, lengthc
0o
CoMFA, π
0o
CoMFA, flexibility, ESP
CoMFA, flexibility, length
CoMFA, flexibility, π
CoMFA, ESP, length
CoMFA, ESP, π
0o
0o
0o
0o
0o
CoMFA, length, π
0o
CoMFA, flexibility, ESP, length
CoMFA, flexibility, ESP, π
CoMFA, flexibility, length, π
CoMFA, ESP, length, π
CoMFA, flexibility, ESP, length, π
0o
0o
0o
0o
0o
b
c
d
aFlexibility of the para-substituent. ESP phenyl ring charge. Length of the para-substituent. Bold indicates the options used for the final
model.
5. Discussion
site existed that was accessible to long para-substituents
containing a ring system. Our data, however, suggests
that although the pocket accommodated large, hydropho-
bic substituents, eg. compounds 15, 16 and 21, it will also
accept large, less hydrophobic substituents such as com-
pounds 17, 24 and 25 and long flexible carbon chains like
compound 22. The important characteristic is that the
substituent is able to access this pocket, for example,
shorter hydrophobic compounds that cannot reach the
pocket, like compound 14, encounter steric hindrance and
have low â1-potency.
The aim of this study was to determine the optimal
structural requirements of para-substituted N-isopropyl-
phenoxypropanolamines to maximise antagonist potency
and selectivity for â1-ARs. The study also provides
information on the extent to which the CoMFA derived
models of the rat â1- and â2-ARs differed for a series of
para-substituted â-AR antagonists. Although para-
substitution generally reduced antagonist potency at both
â1- and â2-ARs, differences did exist between the two
models. For both models, steric factors were the most
important (0.88 for â1- and 0.73 for â2-ARs). The â1-AR
also possesses a pocket which prefers bulky substituents.
Long flexible compounds that accessed this region had
their â1-AR potency restored which is consistent with our
previous work [10] that suggested a hydrophobic binding
In addition, a number of other physical characteristics
of the compounds seemed to influence their potency at
â1- and â2-ARs. For both â1- and â2-ARs, ESP was
negatively correlated with potency suggesting that the
more negative the phenyl ring charge the higher the â1- or

930
Figure 9. (A) CoMFA contour map of the final model of the â1-AR with compound 16 embedded. Green areas show favoured areas
where increasing bulk is associated with higher potency, yellow areas where increasing bulk is associated with lower potency and blue
areas where positive charge is associated with higher potency. (B) CoMFA contour map of the final model of the â2-AR with compound
16 embedded.
â2-AR potency which supports the hypothesis of previous
workers that an electron rich moiety is an essential
pharmacophoric element for binding to â-ARs [42, 43].
In addition, flexibility was positively correlated with
â2-AR potency and this may be related to the ability of
more flexible para-substituents to avoid sterically re-
stricted regions in the â2-AR.
Unfortunately, our synthetic program was designed to
examine â1-AR potency and hence only compound 1 had
high potency (pA2 > 7.0) at the rat â2-AR and higher
affinity compounds are required before we can confi-
dently predict the structure of the â2-AR pharmacophore.
We can say, however, that the â2-AR appears to be very
sterically restricted around the para-position of the phe-
nyl ring of AOPA antagonists.
On the whole, the potency of the training set of
compounds were predicted better by the â1-AR model (r2
= 0.95; residuals < ± 0.22) than the â2-AR model (r2 =
0.83; residuals < ± 0.80). In particular, the potency of
compound 8 was over estimated for the â2-AR. This may
have been due to the existence of a negative charge
preferring region in the region of the positively charged
amine in the para-substituent. This, however, was not
identified by CoMFA perhaps due to the small number of
compounds examined that possess this characteristic. It is
of interest that the OH-group of compound 9, which has
an electronegative nature, has significantly higher â2-AR
potency than compound 8 and slightly higher activity
than the uncharged substituent of compound 7 (table II),
therefore, this region requires further study.

931
Table VI. CoMFA Actual, calculated and residual activities of the
training set of compounds for â2-ARs.
Table VII. Comparison of predicted and experimentally determi-
ned potencies at rat â1- and â2-ARs for the test compounds.
Compound
Actual
Calculated
Residual
Compound Receptor
Actual
Predicted
Residual
1
2
3
4
5
6
7
8
7.47
6.43
6.47
6.34
6.56
5.94
6.23
5.39
6.58
5.72
6.43
6.77
5.97
5.36
5.93
5.64
6.38
5.69
5.40
5.49
5.68
5.80
7.41
6.34
6.53
6.27
6.50
6.17
6.03
6.19
6.32
5.76
6.21
6.49
6.11
5.39
6.07
5.49
6.47
5.69
5.33
5.40
5.67
5.84
0.06
0.09
–0.06
0.07
0.06
–0.23
0.20
–0.80
0.26
–0.04
0.22
23
24
25
â1-
â1-
â1-
7.22
7.73
8.04
7.52
7.88
7.94
–0.30
0.15
0.10
23
24
25
â2-
â2-
â2-
6.47
5.44
5.81
6.85
5.80
6.06
–0.38
–0.36
–0.25
9
10
11
12
13
14
15
16
17
18
19
20
21
22
the symbols of the elements, results obtained were within
± 0.4% of the theoretical values. Descriptions of the
synthetic procedures for preparing compounds 6, 7, 9, 10,
15, 16, 18, 20 and 22 are outlined below. No attempts
were made to maximise yields.
0.28
–0.14
–0.03
–0.14
0.15
–0.09
0.00
0.07
0.09
0.01
–0.04
6.1.2. Materials
Phenol, 4-methylphenol, 4-ethylphenol, 4-methoxy-
phenol, 4-benzyloxyphenol, 4-fluorophenol, p-cresol,
4-aminophenol, 4-fluorophenethyl bromide, 2-cyclo-
hexyl-1-bromoethane, 2-(4-benzyloxyphenyl)-1-ethanol,
benzyl bromide, 4-toluene sulfonylchloride, phenyl-n-
butanoate, ethylchloroformate, bromoethane, iodomethane,
2-hydroxymethyl tetrahydrofuran, 2-methyl-propane-1-
ol, n-pentanol, isobutanol, cyclopentanol, cyclohexanol,
chloroacetic acid, ethyl-4-hydroxyphenyl propionate,
lithium aluminium hydride and dimethyl formamide
(DMF) were purchased from Aldrich Chemicals. Sodium
hydride and isopropylamine were obtained from Fluka
Chemicals. Anhydrous aluminium chloride was obtained
from Merck. Inorganic reagents were supplied by Ajax
Chemicals. Analytical grade solvents, including carbon
disulphide, were purchased from Rhone Poulenc Austra-
lia. (-)-Isoprenaline and carbachol were obtained from
Sigma. Silica plates (5 × 10 cm, Silica F254) were
purchased from Merck. Silica for column chromatogra-
phy (100 Å, 50 mm) was supplied by Amicon Inc., MA
01923, USA.
The two CoMFA models expand the results of earlier
studies [10–15] that high â1-selectivity and potency can
be achieved by para-substitution of the phenyl ring. For
para-substituted N-isopropylphenoxy-propanolamines it
appears at least two factors contribute to â1-potency.
These are the ability of the para-substituent to access a
bulk preferring pocket and the phenyl ring charge. The
latter is negatively correlated with potency.
6. Experimental section
6.1. Chemistry
6.1.1. General
Melting points were determined using a manual Gal-
lenkamp electrothermal apparatus (range 0–200 °C) in
glass capillary tubes and are uncorrected. IR spectra were
recorded on a Perkin-Elmer FT IR 1600. NMR spectra
were recorded on a Varian Associates EM 360 spectrom-
eter and are expressed in δ using TMS (tetramethyl
silane) as reference. Mass spectra were recorded on a
Finnigan 4000 series GC/MS Mass spectrometer or a
Thermo Instruments GCQ Mass spectrometer using
methane gas as the ionising medium for CI (chemical
ionisation spectra). All spectra were consistent with the
assigned structures. Where analyses are indicated only by
6.1.3. General procedure 1 for the synthesis of 1-
phenoxy-2,3-epoxypropanes (B)
The phenol A (20 mmol), KOH (1.23 g, 22 mmol) and
epichlorohydrin (5.55 g, 60 mmol) in EtOH (60 mL)
were stirred overnight at room temperature. The reaction
mixture was evaporated to dryness and the residue was
partitioned between ether (or EtOAc) and water. The
organic layer was washed with aqueous NaOH (5%),
followed by a water wash and then dried over Na2SO4.
After the solvent was evaporated, the residue was freed
from excess epichlorohydrin under vacuum and used

932
without further purification. All epoxides were homog-
enous by TLC (silica, eluent dichloromethane).
phenyl propiophenone P was obtained in overall 33.4%
yield (27.4 g) and had the following physical data: b.p. =
148–150 °C (11.5 mm Hg); MS m/e 165 (M + 1);
1H-NMR (CD3OD) δ 1.14 (3H, t), 2.39–2.82 (2H, m),
3.62–3.98 (2H, m), 6.97 (5H, dd, the phenolic hydrogen
is hiding under the aromatic protons).
6.1.4. General procedure 2 for preparing N-isopropyl-
phenoxypropanolamines (C)
The crude 1-phenoxy-2,3-epoxypropane B (20 mmol)
and isopropylamine (100 mmol, 5 molar excess) in EtOH
(60 mL) were stirred overnight at room temperature. The
reaction mixture was evaporated to dryness and the
residue was dissolved in ethanol and then treated with
excess ethereal HCl. The crude precipitate was recrystal-
lised from a suitable solvent or purified by column
chromatography (silica, eluent: CH2Cl2/MeOH/NH4OH
28%, 90:9:1).
Reduction of P (27.0 g, 166 mmol) was achieved by its
portion wise addition into freshly prepared sodium amal-
gam (68.0 g of powdered Zn and 5.03 g of HgCl2 in
650 mL of methanol and 320 mL of concentrated HCl),
evaporation of methanol, isolation and chromatography
on silica (eluent CH2Cl2:hexane, 1:1). Q was produced in
1
62.5% yield (15.56 g); MS m/e 151 (M + 1); H-NMR
(CD3OD) δ 1.08 (3H, t), 1.25–2.05 (4H, m), 2.65 (2H, t),
3.87 (5H, m), 6.89 (5H, dd, the phenolic hydrogen is
hiding under the aromatic protons).
6.1.5. 4-Ethoxyphenol (E)
Ethyl bromide (6.0 g, 55 mmol), 4-benzyloxyphenol
(10.11 g, 50 mmol), anhydrous K2CO3 (8.29 g, 60 mmol)
and NaI (0.5 g) were stirred and refluxed in anhydrous
acetone (250 mL) for 48 h. The solvent was evaporated
and the residue was partitioned between dichloromethane
and water. The organic layer was washed with water and
dried over anhydrous MgSO4. The solvent was evapo-
rated and the residual 4-ethoxyphenylbenzylether D (R =
H) was recrystallised from EtOH. Yield = 8.2 g, 81.5%;
m.p. = 64–66 °C; MS m/e 229 (M + 1); 1H-NMR
(CDCl3) δ 1.58 (3H, t), 4.15 (2H, q), 5.17 (2H, s), 7.06
(5H, s), 7.58 (4H, s).
The ether D (R = H) (7.2 g, 36 mmol) in EtOH
(200 mL) was hydrogenated over 10% Pd/C (0.1 g) at
room temperature and atmospheric pressure. When hy-
drogen absorption ceased, the reaction mixture was
filtered and evaporated to dryness. The product E was
recrystallised from EtOH. Yield = 4.26 g, 87.0%; m.p. =
64.5–65.5 °C; MS m/e 139 (M + 1); 1H-NMR (CDCl3) δ
1.61 (3H, t), 4.19 (2H, q), 5.63 (1H, s), 6.93 (4H, s).
6.1.8.
panol (7)
1-(4-n-Butylphenoxy)-3-isopropylamino-2-pro-
Prepared from Q according to the general procedures 1
and 2, chromatographed and isolated as the hydrochlo-
ride. Yield = 4.35 g, 72.1%; m.p. = 70.5–73.0 °C; MS m/e
266 (M + 1); 1H-NMR (CD3OD) δ 1.13 (3H, t), 1.75 (6H,
d), 2.51–2.93 (3H, m), 3.23–3.83 (5H, bm), 4.06–4.31
(2H, m), 4.63–5.12 (2H, m), 6.77–7.28 (4H, dd). Anal.
(C16H27NO2.HCl) C, H, Cl, N.
6.1.9. 4-(n-3≠-Hydroxypropyl)phenol (T)
Ethyl-4-hydroxyphenylpropionate (54.5 g, 250 mmol)
was refluxed with benzyl bromide (44.5 g, 30.9 mL,
260 mmol), anhydrous potassium carbonate (38.7 g,
270 mmol) and sodium iodide (2.0 g) in dried acetone for
48 h. Filtration of the cooled mixture and evaporation of
the solvent produced a solid residue which was taken up
in ethyl acetate, washed with diluted sodium hydroxide
solution and water, dried, filtered and evaporated. The
product, ethyl-4-benzyloxyphenylpropionate, R, was pu-
rified by chromatography (silica, eluent CH2Cl2) and
isolated as an oil. Yield = 68.6 g, 89.1%; MS m/e 309 (M
+ 1); 1H-NMR (CDCl3) δ 1.16 (3H, t), 2.76 (2H, t), 3.82
(2H, t), 4.64 (2H, q), 5.06 (2H, s), 6.88 (4H, s), 7.26 (5H,
s).
6.1.6.
1-(4-Ethoxyphenoxy)-3-isopropylamino-2-pro-
panol (6)
Prepared from E according to the general procedures 1
and 2, chromatographed and isolated as the hydrochlo-
ride. Yield = 3.66 g, 63.4%; m.p. = 87.5–88.5 °C; MS m/e
1
254 (M + 1); H-NMR (CD3OD) δ 1.33 (6H, d), 1.62
Ethyl-4-benzyloxyphenylpropionate (61.6 g, 200 mmol)
in anhydrous THF (200 mL) was added dropwise into a
suspension of lithium aluminium hydride (8.0 g,
210 mmol) in THF (500 mL). The resultant mixture was
refluxed for 2 h, cooled, the excess reagent decomposed
by the dropwise addition of water and filtered. The filtrate
was evaporated to dryness and the residue was purified by
chromatography (silica, eluent CH2Cl2). The isolated
product, 1-(4≠-benzyloxyphenyl)-3-propanol, S, was re-
crystallised from ethanol-ether. Yield = 32.1 g, 66.2%;
(3H, t), 2.16–3.23 (6H, bm), 4.23 (2H, q), 7.10 (4H, s).
Anal. (C14H23NO3.HCl) C, H, Cl, N.
6.1.7. 4-n-Butylphenol (Q)
Phenyl-n-butanoate (82.0 g, 500 mmol) was added into
a stirred suspension of aluminium chloride (75 g) in
carbon disulphide (80 mL) to furnish, after work-up, a
mixture of 2- and 4-hydroxyphenylpropiophenones (bu-
tyrophenones) (64.5 g, 78.6%) which were separated by
repeated fractional distillation. The desired 4-hydroxy-

933
1
m.p. = 59.5–61.5 °C; MS m/e 243 (M + 1); H-NMR
(6H, d), 1.79–2.18 (2H, m), 3.27–3.78 (4H, m), 3.87–4.29
(4H, m), 4.55–6.07 (1H, m), 5.48–5.74 (1H, m), 6.92
(4H, s). Anal. (C15H25NO3.HCl) C, H, Cl, N.
(CDCl3) δ 0.72 (1H, bs), 0.87–1.23 (2H, m), 2.81 (2H, t),
3.87 (2H, t), 5.13 (2H, s), 7.11 (4H, dd), 7.52 (5H, s).
A solution of 1-(4≠-benzyloxyphenyl)-3-propanol
(24.2 g, 100 mmol) in ethanol (600 mL) was hydroge-
nated over 10% Pd/C at ambient temperature and pres-
sure. Filtration of the catalyst and evaporation of the
solvent furnished the desired phenol T which was puri-
fied by chromatography (silica, eluent CH2Cl2) and
isolated as an oil. Yield = 12.81 g, 84.3%; MS m/e 153
(M + 1); 1H-NMR (CDCl3/CD3OD 1:1) δ 1.77–2.23 (2H,
m), 2.77 (2H, m), 3.81 (2H, m), 7.02 (4H, dd).
6.1.13. 4-Cyclohexylethoxyphenol (G)
4-Cyclohexylethyl bromide (10.5 g, 55 mmol), 4-ben-
zyloxyphenol (10.11 g, 50 mmol), anhydrous K2CO3
(8.29 g, 60 mmol) and NaI (0.5 g) were stirred and
refluxed in anhydrous acetone (250 mL) for 48 h. The
solvent was evaporated and the residue was partitioned
between dichloromethane and water. The organic layer
was washed with water and dried over anhydrous
MgSO4. The solvent was evaporated and the residual
4-(4≠-cyclohexylethoxy)phenylbenzylether D (R≠ = cy-
lohexyl) was recrystallised from EtOH. Yield = 13.6 g,
81.4%; m.p. = 68–70 °C; MS m/e 311 (M + 1); 1H-NMR
(CDCl3) δ 0.89–1.96 (13H, m), 3.65–4.02 (2H, m), 4.95
(2H, s), 6.72 (4H, s), 7.24 (5H, s).
6.1.10. 1-(4-(n-3≠-Hydroxypropyl)-phenoxy)-3-isopropyl-
amino-2-propanol (9)
Prepared from T according to the general procedures 1
and 2, chromatographed and isolated as the hydrochlo-
ride. Yield = 2.41 g, 39.7%; m.p. = 68.5–71.5 °C; MS m/e
268 (M + 1); 1H-NMR (CD3OD) δ 1.62 (6H, d),
1.83–2.29 (2H, m), 2.61–3.12 (2H, m), 3.15–3.94 (6H,
m), 4.07–4.36 (2H, m), 6.90–7.48 (4H, dd). Anal.
(C15H25NO3.HCl) C, H, Cl, N.
The ether D (R≠ = cyclohexyl) (12.0 g, 36 mmol) in
EtOH (200 mL) was hydrogenated over 10% Pd/C (0.1 g)
at room temperature and atmospheric pressure. When
hydrogen absorption ceased, the reaction mixture was
filtered and evaporated to dryness. The product G was
recrystallised from EtOH. Yield = 7.78 g, 85.6%; m.p. =
6.1.11. 4-n-Propoxyphenol (F)
1
n-Propyl bromide (6.8 g, 55 mmol), 4-benzyloxy-
phenol (10.11 g, 50 mmol), anhydrous K2CO3 (8.29 g,
60 mmol) and NaI (0.5 g) were stirred and heated to
reflux in anhydrous acetone (250 mL) for 48 h. The
solvent was evaporated and the residue was partitioned
between dichloromethane and water. The organic layer
was washed with water and dried over anhydrous
MgSO4. The solvent was evaporated and the residual
4-n-propoxyphenylbenzylether D (R≠ = Me) was recrys-
tallised from EtOH. Yield = 11.07 g, 83.2%; m.p. =
67–69 °C; MS m/e 243 (M + 1); H-NMR (CDCl3) δ
1.22 (3H, t), 1.70–2.23 (2H, m), 4.03 (2H, t), 5.14 (2H, s),
7.03 (5H, s), 7.52 (4H, s).
The ether D (R≠ = Me) (10.1 g, 42 mmol) in EtOH
(200 mL) was hydrogenated over 10% Pd/C (0.1 g) at
room temperature and atmospheric pressure. When hy-
drogen absorption ceased, the reaction mixture was
filtered and evaporated to dryness. The product F was
recrystallised from EtOH. Yield = 5.12 g, 80.2%; m.p. =
91–93 °C; MS m/e 221 (M + 1); H-NMR (CDCl3) δ
0.77–2.01 (13H, m), 3.65–4.21 (2H, m), 4.47 (1H, s),
6.73 (4H, s).
6.1.14. 1-(4-Cyclohexylethoxyphenoxy)-3-isopropylamino
-
2-propanol (15)
Prepared from G according to the general procedures 1
and 2, chromatographed and isolated as the hydrochlo-
ride. Yield = 4.91 g, 66.3%; m.p. = 131.0–133.0 °C; MS
1
m/e 336 (M + 1); H-NMR (CD3OD) δ 0.91 (6H, d),
1
0.64–1.53 (11H, bm), 2.80 (4H, m), 3.87 (5H, m), 6.43
(4H, s). Anal. (C20H33NO3.HCl) C, H, Cl, N.
6.1.15. 4-Fluorophenethoxyphenol (H)
4-Fluorophenethyl bromide (11.16 g, 55 mmol),
4-benzyloxyphenol (10.11 g, 50 mmol), anhydrous
K2CO3 (8.29 g, 60 mmol) and NaI (0.5 g) were stirred
and refluxed in anhydrous acetone (250 mL) for 48 h. The
solvent was evaporated and the residue was partitioned
between dichloromethane and water. The organic layer
was washed with water and dried over anhydrous
MgSO4. After evaporation of the solvent, the residual
1
67–69 °C; MS m/e 153 (M + 1); H-NMR (CDCl3) δ
1.19 (3H, t), 1.77–2.26 (2H, m), 4.06 (2H, t), 6.73 (1H, s),
6.99 (4H, s).
4-(4≠-fluorophenethoxy)phenylbenzylether
D (R≠ =
6.1.12.
propanol (10)
1-(4-n-Propoxyphenoxy)-3-isopropylamino-2-
4-fluorophenyl) was recrystallised from EtOH. Yield =
15.2 g, 87.9%; m.p. = 113–114 °C; MS m/e 323 (M + 1);
1H-NMR (CDCl3) δ 3.05 (2H, m), 4.12 (2H, m), 5.11
(2H, s), 6.88 (5H, s), 7.10 (4H, dd), 7.55 (5H, dd).
The ether D (R≠ = 4-fluorophenyl) (13.84 g, 40 mmol)
in EtOH (250 mL) was hydrogenated over 10% Pd/C
Prepared from F according to the general procedures 1
and 2, chromatographed and isolated as the hydrochlo-
ride. Yield = 3.56 g, 58.6%; m.p. = 116.5–119.0 °C; MS
m/e 268 (M + 1); 1H-NMR (CD3OD) δ 1.22 (3H, t), 1.72

934
(0.1 g) at room temperature and atmospheric pressure.
When hydrogen absorption ceased, the reaction mixture
was filtered and evaporated to dryness. The product H
was recrystallised from toluene. Yield = 7.12 g, 76.1%;
6.1.19. 4-(2≠-Tetrahydrofurylmethoxy)ethoxyphenol (N)
2-Bromomethyl tetrahydrofuran (11.5 g, 55 mmol),
4-benzyloxyphenol (10.11 g, 50 mmol), anhydrous
K2CO3 (8.29 g, 60 mmol) and NaI (0.5 g) were stirred
and heated to reflux in anhydrous acetone (250 mL) for
48 h. The solvent was evaporated and the residue was
partitioned between dichloromethane and water. The
organic layer was washed with water and dried over
anhydrous MgSO4. The solvent was evaporated and the
1
m.p. = 166–168 °C; MS m/e 233 (M + 1); H-NMR
(CDCl3) δ 2.45 (2H, m), 3.56 (2H, m), 4.67 (1H, s), 6.53
(4H, dd), 6.93 (5H, dd, the phenolic hydrogen is hiding
under the aromatic protons).
6.1.16.
1-(4-(4≠-Fluorophenethoxy)-phenoxy)-3-iso-
residual
4-(2≠-tetrahydrofurylmethoxy)ethoxyphenyl-
propylamino-2-propanol (16)
benzyl ether L (R≠≠ = 2≠-tetrahydrofurylmethyl) was
purified by chromatography (silica, eluent CH2Cl2) and
isolated as an oil, homogenous by TLC. Yield = 14.6 g,
80.9%; MS m/e 329 (M + 1); 1H-NMR (CDCl3) δ
1.73–2.29 (4H, m), 3.48–4.39 (9H, m), 5.12 (2H, s), 7.03
(4H, s), 7.55 (5H, s).
Prepared from H according to the general procedure,
chromatographed and isolated as the hydrochloride. Yield
= 11.71 g, 61.4%; m.p. = 136.0–138.0 °C; MS m/e 348
1
(M + 1); H-NMR (CD3OD) δ 1.16 (6H, d), 2.71–3.32
(6H, m), 3.71–4.22 (4H, m), 6.46 (4H, s), 6.73 (4H, dd).
Anal. (C20H26FNO3.HCl) C, H, Cl, F, N.
The ether L (R≠≠ = 2≠-tetrahydrofurylmethyl) (11.8 g,
36 mmol) in EtOH (200 mL) was hydrogenated over 10%
Pd/C (0.1 g) at room temperature and atmospheric pres-
sure. When hydrogen absorption ceased, the reaction
mixture was filtered and evaporated to dryness. The
product N was purified by chromatography (silica, eluent
CH2Cl2) and isolated as an oil, homogenous by TLC.
6.1.17. 4-(2≠-Methylpropoxy)ethoxyphenol (M)
2-Methylpropoxyethyl bromide (9.96 g, 55 mmol),
4-benzyloxyphenol (10.11 g, 50 mmol), anhydrous
K2CO3 (8.29 g, 60 mmol) and NaI (0.5 g) were stirred
and refluxed in anhydrous acetone (250 mL) for 48 h. The
solvent was evaporated and the residue was partitioned
between dichloromethane and water. The organic layer
was washed with water and dried over anhydrous
MgSO4. The solvent was evaporated and the residual
4-(2-methylpropoxyethoxy)phenylbenzylether L (R≠≠ =
2-methylpropyl) was purified by chromatography (silica,
eluent CH2Cl2) and isolated as an oil, homogenous by
TLC. Yield = 13.8 g, 83.6%; MS m/e 301 (M + 1);
1H-NMR (CDCl3) δ 1.16 (6H, d), 1.87–2.35 (1H, m),
3.33–3.60 (2H, m), 3.77–4.35 (4H, m), 5.12 (2H, s), 6.94
(4H, dd), 7.52 (5H, s).
1
Yield = 6.85 g, 80.1%; MS m/e 238 (M + 1); H-NMR
(CDCl3) δ 1.73–2.32 (4H, m), 3.57–4.52 (9H, m), 6.81
(1H, s), 6.91 (4H, s).
6.1.20.
1-4≠-(2≠≠ -Tetrahydrofurylmethoxy)ethoxyphe-
noxy)-3-isopropylamino-2-propanol (20)
Prepared from N according to the general procedures 1
and 2, chromatographed and isolated as the hydrochlo-
ride. Yield = 3.59 g, 48.4%; m.p. = 165.5–167.0 °C; MS
1
m/e 336 (M + 1); H-NMR (CD3OD) δ 0.93 (6H, d),
The ether L (R≠≠ = 2-methylpropyl) (10.8 g, 36 mmol)
in EtOH (200 mL) was hydrogenated over 10% Pd/C
(0.1 g) at room temperature and atmospheric pressure.
When hydrogen absorption ceased, the reaction mixture
was filtered and evaporated to dryness. The product M
was purified by chromatography (silica, eluent CH2Cl2)
and isolated as an oil, homogenous by TLC. Yield =
6.85 g, 86.5%; MS m/e 221 (M + 1); H-NMR (CDCl3)
δ 1.18 (6H, d), 1.97–2.42 (1H, m), 3.38–3.64 (2H, m),
3.80–4.39 (4H, m), 6.96 (1H, s), 7.02 (4H, dd).
1.73–2.32 (4H, m), 2.38–3.23 (6H, bm), 3.57–4.52 (9H,
m), 6.93 (4H, s). Anal. (C19H31NO5.HCl) C, H, Cl, N.
6.1.21. 4-n-Pentyloxyethoxyphenol (O)
n-Pentyloxyethyl bromide (10.72 g, 55 mmol),
4-benzyloxyphenol (10.11 g, 50 mmol), anhydrous
K2CO3 (8.29 g, 60 mmol) and NaI (0.5 g) were stirred
and refluxed in anhydrous acetone (250 mL) for 48 h. The
solvent was evaporated and the residue was partitioned
between dichloromethane and water. The organic layer
was washed with water and dried over anhydrous
MgSO4. The solvent was evaporated and the residual
4-(4≠-n-pentyloxyethoxy)phenylbenzyl ether L (R≠≠ =
n-pentyl) was purified by chromatography (silica, eluent
CH2Cl2) and isolated as an oil, homogenous by TLC.
1
6.1.18. 1-4≠-(2≠≠ -Methylpropoxyethoxyphenoxy)-3-iso-
propylamino-2-propanol (18)
Prepared from M according to the general procedures
1 and 2, chromatographed and isolated as the hydrochlo-
ride. Yield = 4.00 g, 55.3%; m.p. = 89.5–91.3 °C; MS m/e
1
1
326 (M + 1); H-NMR (CD3OD) δ 1.13 (6H, d), 1.75
Yield = 13.7 g, 79.4%; MS m/e 315 (M + 1); H-NMR
(6H, d), 3.37–3.62 (6H, bm), 3.80–4.37 (7H, bm), 6.94
(4H, s). Anal. (C18H31NO4.HCl) C, H, Cl, N.
(CDCl3) δ 1.03 (3H, t), 1.19–2.02 (8H, m), 3.53–4.22
(4H, m), 5.16 (2H, s), 6.87 (4H, s), 7.24 (5H, s).

935
The ether L (R≠≠ = n-pentyl) (11.3 g, 36 mmol) in
EtOH (200 mL) was hydrogenated over 10% Pd/C (0.1 g)
at room temperature and atmospheric pressure. When
hydrogen absorption ceased, the reaction mixture was
filtered and evaporated to dryness. The product O was
purified by chromatography (silica, eluent CH2Cl2) and
isolated as an oil, homogenous by TLC. Yield = 7.52 g,
93.25%; MS m/e 225 (M + 1); 1H-NMR (CDCl3) δ 1.07
(3H, t), 1.22–2.07 (8H, m), 3.57–4.29 (4H, m), 6.87 (4H,
s).
6.2.1.2. Rat isolated tracheal chains
Trachea were excised from adult rats (200–250 g),
dissected free of overlying tissue and cut transversely into
segments about 2 mm wide. Five segments were mounted
in a 20 mL bath maintained at 37 °C at a tension of 1 g.
Relaxation of the segments by (-)-isoprenaline was re-
corded by an isotonic transducer connected to a Grass
Polygraph after tone had been established by administra-
tion of 1 µM carbachol (45 min prior to concentration-
response curves).
6.1.22.
amino-2-propanol (22)
1-(4-n-Pentyloxyethoxyphenoxy)-3-isopropyl-
6.3. Computational methods
Prepared from O according to the general procedures 1
and 2, chromatographed and isolated as the hydrochlo-
ride. Yield = 3.90 g, 52.1%; m.p. = 77.0–79.0 °C; MS m/e
340 (M + 1); 1H-NMR (CD3OD) δ 1.13 (3H, t), 1.75 (6H,
d), 1.36–2.10 (8H, bm), 3.26–4.04 (6H, m), 3.86–4.28
(4H, dt), 7.03 (4H, s). Anal. (C19H33NO4.HCl) C, H, Cl,
N.
6.3.1. Molecular modelling
Modelling studies were performed using the SYBYL
software package (Tripos Inc., version 6.40) on a Silicon
Graphics Indigo 2 UNIX Workstation. The structure of
the active S-isomer of compound 1 was constructed from
standard bond lengths and bond angles using the sketch
routine in SYBYL. The geometry of compound 1 was
fully optimised in vacuo as previous studies [19] have
identified that the propanolamine side-chain of aryloxy-
propanolamine (AOPA) type compounds can adopt com-
mon conformations in solid, theoretical gas and solution
states. Optimisation was conducted using the AM1
Hamiltonian in MOPAC (version 6.0) [48], the keywords
‘PRECISE’, ‘ESP’ and ‘NOMM’ were specified. The
keyword ESP calculates the electrostatic potential derived
atomic charges, these charges have been used since they
are reported to be reliable and highly correlated with ab
initio ESP charges [49, 50].
6.2. Pharmacological methods
6.2.1. Isolated tissue preparations
Studies were carried out on Sprague-Dawley rat iso-
lated atria and tracheal rings according to our method
described previously [16]. All tissues were allowed to
equilibrate for 45 min with Krebs Ringer physiological
salt solution; the composition of which in mmol L–1 was
NaCl, 120; KCl, 5.6; MgSO4, 1.2; CaCl2, 2.5; KH2PO4,
1.4; NaHCO3, 25; glucose 11.2 and EGTA, 0.0025.
Cumulative concentration-response curves were ob-
tained for the non-selective â-AR agonist (-)-isoprenaline
in each preparation [16]. (-)-Isoprenaline was dissolved
and diluted in 1 mg mL–1 ascorbic acid to prevent
oxidation. For the measurement of antagonist activity the
appropriate agent was added to the organ bath at least
30 min after the first control concentration-response
curve was completed and allowed to equilibrate for
10 min before the next concentration-response curve
established. The shift in this curve to the right was
calculated as a pA2 value [39]. At least three concentra-
tions of each antagonist were examined to verify the pA2.
6.3.2. Molecular alignment
Compounds 2–25 were constructed using the mini-
mised structure of compound 1 as a starting point. The
appropriate para-substituent was added using the add
multiple atom function in SYBYL. The structure was
then optimised using the AM1 Hamiltonian in MOPAC,
with the keywords ‘PRECISE’, ‘ESP’ and ‘NOMM’
specified, however, τ1 was fixed at 0° or 90° (figure 1)
and the bond, dihedral and valence angles of the AOPA
core structure were fixed at the optimised values for
compound 1. Therefore, two series of compounds were
created, one where τ1 = 0° and the other where τ1 = 90°.
After optimisation the compounds were superimposed
over the entire N-isopropylphenoxypropanolamine struc-
ture (compound 1) using the linear least squares routine
within SYBYL, as shown in figure 7.
6.2.1.1. Rat isolated spontaneously beating atria
Rat hearts were removed from adult animals
(200–250 g) and placed in Krebs Ringer salt solution (pH
7.4) aerated with 5% CO2 in O2. The atria were dissected
free of the ventricles and overlying tissue and placed in a
20 mL bath maintained at 37 °C and connected to an
isotonic transducer. A tension of 1 g was applied and
chronotropic activity was amplified and recorded on a
Grass Polygraph.
In addition to the optimisation and superimposition of
each compound a number of other physical parameters
were examined:

936
6.3.3. Phenyl ring charge calculations
generated using the PLS algorithm using the ‘leave-one-
out’ cross validation procedure, and the number of
components with the lowest standard error of prediction
value was selected as the ONC. The additional physical
parameters were also included in the cross validated PLS
analysis to determine whether they improved q2. The final
models, non cross-validated, for the â1- and â2-ARs were
generated using CoMFA, the combination of additional
parameters that maximised q2 and the ONC as deter-
mined above.
For each compound in table II (1–25), the AM1 ESP
charges calculated for the six phenyl ring carbon atoms
were added together to obtain the parameter ‘phenyl ring
charge’. Various oxypropanolamine and para-substituent
conformations (ie. different τ1, τ2 values, figure 1) were
examined for their effect upon phenyl ring charge. Both
τ1 and τ2 were varied independently in 90° steps through
a full 360°, giving a total of sixteen conformers examined
for each compound, and the phenyl ring charge was
calculated for each conformer. The phenyl ring charge did
not vary greatly between conformers and therefore the
value obtained when τ1 = 90° and τ2 = 0° was used in this
study. All sixteen phenyl ring charge values calculated for
a particular compound were within ± 0.04e of the value
obtained when τ1 = 90° and τ2 = 0°.
Acknowledgements
The authors would like to thank Ms. Leanne Styan for
her technical assistance and Dr Margaret Wong for her
helpful comments. This work was supported by a grant
from the National Health and Medical Research Council
of Australia.
6.3.4. LogP calculations
The logP values for the twenty five compounds were
calculated using the Pallas PrologP (version 1.1) program
at Swinburne University of Technology, Hawthorn, Vic-
toria, Australia. The π values [51] for the para-
substituents were determined by subtracting the logP of
compound 1 from the compound in question and are
given in table II.
References
[1] Greenidge P.A., Pharm. Acta Helv. 68 (1994) 201–210.
[2] Lafontan M., Belan M., J. Lipid Res. 34 (1993) 1057–1091.
[3] Ostrowski J., Kjelsberg M.A., Caron M.G., Lefkowitz R.J., Ann.
Rev. Pharmacol. Toxicol. 32 (1992) 167–183.
6.3.5. Length of the para-substituent
[4] Strosberg A.D., Prot. Sci. 2 (1993b) 1198–1209.
The length was defined as the distance between the
para-carbon of the phenyl ring and the most distant
non-H atom in the para-substituent, hence the para-
substituent length of compound 1 is zero (table II). The
chain length of the para-substituent (R-group, table II)
was calculated for the relaxed, fully extended low energy
conformation of each molecule.
[5] Emorine L.J., Marullo S., Briend-Sutren M.M., Patey G., Tate K.,
Delavier-Klutchko C., Strosberg A.D., Science 245 (1989)
1118–1121.
[6] Frielle T.J., Collins S., Daniel K.W., Caron M.G., Lefkowitz R.J.,
Kobilka B.K., Proc. Natl. Acad. Sci. USA 84 (1987) 7920–7924.
[7] Kobilka B.K., Dixon R.A.F., Frielle T., Dohlman H.G., Bolanowski
M.A., Sigal I.S. et al., Proc. Natl. Acad. Sci. USA 84 (1987) 46–50.
[8] Donne-Op Den Kelder G.M., Bultsma T., Timmerman H., Rade-
maker B., J. Med. Chem. 31 (1988) 1069–1079.
6.3.6. Flexibility of the para-substituent
[9] Linschoten M.R., Bultsma T., Ijzerman A.P., Timmerman H., J. Med.
Chem. 29 (1986) 278–286.
A measure of the flexibility was incorporated into the
QSAR analysis by assigning an integer for the number of
torsion angles which affect the conformation of the
para-substituent and is given in table II. For example, the
flexibility of compound 4 is assigned the integer 1, as
rotation of terminal methyl, hydroxyl and amine groups
do not affect the conformation of the para-substituent.
[10] Ball J.B., Nero T.L., Iakovidis D., Tung L., Jackman G., Louis W.J.,
J. Med. Chem. 35 (1992) 4676–4682.
[11] Baldwin J.J., Christy M.E., Denny G.H., Habecker C.N., Freedman
M.B., Lyle P.A. et al., J. Med. Chem. 29 (1986) 1065–1080.
[12] Machin P.J., Hurst D.N., Bradshaw R.M., Blaber L.C., Burden D.T.,
J. Med. Chem. 27 (1984) 503–509.
[13] Berthold R., Louis W.J., US Patent, US 4, 425, 362, (1984).
[14] Hoefle M.L., Hastings S.G., Meyer R.F., Corey R.M., Holmes A.,
Stratton C.D., J. Med. Chem. 18 (1975) 148–152.
6.3.7. Development of the CoMFA model
[15] Large M.S., Smith L.H., J. Med. Chem. 25 (1982) 1286–1292.
CoMFA [41] was performed using the QSAR option in
SYBYL. Both steric and electrostatic fields were consid-
ered. The probe atom had the properties of an sp3 carbon
atom and a charge of +1.0. Cut-off values were SYBYL
default values and the grid step size was 2.0 Å. For each
group of compounds (ie. τ1 = 0° or 90°) the effects of
changing column filtering values (≈ 1.0–16.0 kcal/mol)
were examined. The CoMFA QSAR equations were
[16] Louis S.N.S., Nero T.L., Iakovidis D., Jackman G.P., Louis W.J.,
Eur. J. Pharmacol. 367 (1999) 431–435.
[17] Beckett A.H., Fortschr. Arzneim. Forsch. 1 (1959) 455–530.
[18] Easson L.H., Stedman E., Biochem. J. 27 (1933) 1257–1266.
[19] Nero T.L., Louis W.J., Iakovidis D., J. Mol. Struct. (Theochem.) 285
(1993) 251–272.
[20] Brandstrom A.E., Carlsson P.A.E., Carlsson S.A., Corrodi H.R., Ek
L., Ablad B.A.H., US Patent, US 3928601, (1975).

937
[21] Carlsson P.A.E., Brandstrom A.E., Lamm B.R., Ablad B.A.H.,
Carlsson S.A.I., Corrodi H.R., Ek L., DE Patent, DE 2020864,
(1970).
[36] Van Rossum J.M., Hurkmans J.A.T.M., Wolters C.J.J., Arch. Int.
Pharmacodyn. Ther. 143 (1963) 299–330.
[37] Zaborowsky B.R., McMahan W.C., Griffin W.A., Norris F.H.,
Ruffolo Jr. R.R., J. Pharmacol. Methods 4 (1980) 165–178.
[22] Crowther A.F., Gilman D.J., McLoughlin B.J., Smith L.H., Turner
R.W., Wood T.M., J. Med. Chem. 12 (1969) 638–642.
[38] MicroCal Origin, MicroCal Software Inc., USA.
[39] Mackay D.J., Pharm. Pharmacol. 30 (1978) 312–313.
[23] Erez M., Shtacher G., Weinstock M., J. Med. Chem. 21 (1978)
982–984.
[40] Baldwin J.J., Denny G.H., Hirschmann R., Freedman M.B., Ponti-
cello G.S., Gross D.M., Sweet C.S., J. Med. Chem. 26 (1983)
950–957.
[24] Ferrari G., Ferrini R., Casagrande C., Boll. Chim. Farm. 107 (1968)
234–248.
[41] Cramer R.D., Paterson D.E., Bunce J.D., J. Am. Chem. Soc. 110
(1988) 5959–5967.
[25] Jain P.C., Belleau B., Rastogi S.N., Anand N., Tripathi O.N., Srimal
R.C., Indian J. Chem. 12 (1974) 1021–1027.
[42] Triggle D.J., in: Triggle D.J., Triggle C.R. (Eds.), Chemical phar-
macology of the synapse, Academic Press, London, 1976 p. 233.
[26] Leigh T., GB Patent 1185046 (1970).
[27] Loevgren K., Hedberg A., Nilsson J., Lars G., J. Med. Chem. 24
(1981) 451–454.
[43] Belleau B., Ann. NY Acad. Sci. 139 (1967) 580–605.
[44] Dixon R.A.F., Sigal I.S., Strader C.D., Cold Spring Harbor Sympo-
sium Quant. Biology. 3 (1988) 487–497.
[28] Machin P.J., Hurst D.N., Bradshaw R.M., Blaber L.C., Burden D.T.,
Fryer A.D., Melarange R.A., Shivdasani C., J. Med. Chem. 26
(1983) 1570–1576.
[45] Lewell X.Q., Drug Design Discovery 9 (1992) 29–48.
[46] Strader C.D., Candalore M.R., Hill W.S., Dixon R.A.F., Sigal I.S., J.
Biol. Chem. 264 (1989) 13572–13578.
[29] Malguzzi R., Giordano C., EP Patent, EP 50885 (1982).
[30] Manoury P.M., Binet J.L., Rousseau J., Lefevre-Borg F.M., Cavero
I.G., J. Med. Chem. 30 (1987) 1003–1011.
[47] Strader C.D., Sigal I.S., Register R.B., Candelore M.R., Rands E.,
Dixon R.A.F., Proc. Natl. Acad. Sci. USA 84 (1987) 4384–4388.
[31] Rosell G., Mauleon D., Pujol M.D., Granados R., Arch. Farmacol.
Toxicol. (1986) 65–70.
[48] Dewar M.J.S., Zoebisch E.G., Healy E.F., Stewart J.J.P., J. Am.
Chem. Soc. 107 (1985) 3902–3909.
[32] Wick A.E., Binet J., US Patent, US 4515814 (1985).
[49] Besler B.H., Merz Jr. K.M., Kollman P.A., J. Computat. Chem. 11
(1990) 431–439.
[33] Zaagsma J., Nauta W.T., J. Med. Chem. 17 (1974) 507–513.
[34] Vogel A.I., Practical Organic Chemistry 3rd ed. Longmann, London,
1962 pp. 676–962.
[50] Ferenczy G.G., Reynolds C.A., Richards W.G., J. Computat. Chem.
11 (1990) 159–169.
[35] Isagulyants V.I., Tisukova V.N., Favorskaya N.A., Oganesyan P.O.,
Chem. Abstracts 62 (1962) 14545a.
[51] Hansch C., Leo A., Unger S.H., Kim K.H., Nikaitani D., Lien E.J.,
J. Med. Chem. 16 (1973) 1207–1216.