Synthesis, biological activity, and three-dimensional quantitative structure-activity relationship model for a series of benzo[c]quinolizin-3-ones, nonsteroidal inhibitors of human steroid 5α-reductase 1

Article abstract of DOI:10.1021/jm031131o

New 5α-reductase 1 (5αR-1) inhibitors were designed to complete a consistent set of analogues suitable for a 3D QSAR study. These compounds were synthesized by a modification of the aza-Robinson annulation, further functionalized by Pd-catalyzed cross-coupling processes, and were tested with human 5αR-1 expressed in Chinese hamster ovary 1827 cells. It turned out that the potency of the resulting inhibitors was strongly dependent on the type of substitution at the 8 position, with the 1C₅₀ values ranging from 8.1 to 1050 nM. The construction of this homogeneous set of molecules allowed a 3D QSAR study. In particular, comparative molecular field analysis (CoMFA) was used to correlate the potency of the inhibitors with their physicochemical features. Highly accurate evaluations of the atomic point charges were carried out by means of quantum chemical calculations at the DFT/B3LYP level of theory followed by the RESP fitting procedure. It turned out that increasing the reliability of electrostatic parameters greatly affected the statistical results of the QSAR analysis. The 3D QSAR model proposed could be very useful in the further development of 5αR-1 inhibitors, which are suitable candidates to be evaluated as drugs in the treatment of 5αR-1 related diseases such as acne and alopecia in men and hirsutism in women.

Full text of DOI:10.1021/jm031131o

3546
J . Med. Chem. 2004, 47, 3546-3560
Syn th esis, Biologica l Activity, 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 Mod el for a Ser ies of
Ben zo[c]qu in olizin -3-on es, Non ster oid a l In h ibitor s of Hu m a n Ster oid
5r-Red u cta se 1
Ernesto G. Occhiato,† Alessandro Ferrali,† Gloria Menchi,† Antonio Guarna,*,† Giovanna Danza,‡
Alessandra Comerci,‡ Rosa Mancina,‡ Mario Serio,‡ Gianni Garotta,§ Andrea Cavalli,*,# Marco De Vivo,# and
Maurizio Recanatini#
Department of Organic Chemistry “U. Schiff”, University of Firenze, Via della Lastruccia 13,
I-50019 Sesto Fiorentino, Italy, Department of Clinical Physiopathology, Endocrinology Unit,
University of Firenze, Viale G. Pieraccini 6, I-50134 Firenze, Italy, Serono International S.P.R.I., 12,
Chemin des Aulx, CH-1228 Plan-les Ouates, Geneve, Switzerland, and Department of Pharmaceutical Sciences,
University of Bologna, Via Belmeloro 6, I-40126 Bologna, Italy
Received December 8, 2003
New 5R-reductase 1 (5RR-1) inhibitors were designed to complete a consistent set of analogues
suitable for a 3D QSAR study. These compounds were synthesized by a modification of the
aza-Robinson annulation, further functionalized by Pd-catalyzed cross-coupling processes, and
were tested with human 5RR-1 expressed in Chinese hamster ovary 1827 cells. It turned out
that the potency of the resulting inhibitors was strongly dependent on the type of substitution
at the 8 position, with the IC50 values ranging from 8.1 to 1050 nM. The construction of this
homogeneous set of molecules allowed a 3D QSAR study. In particular, comparative molecular
field analysis (CoMFA) was used to correlate the potency of the inhibitors with their
physicochemical features. Highly accurate evaluations of the atomic point charges were carried
out by means of quantum chemical calculations at the DFT/B3LYP level of theory followed by
the RESP fitting procedure. It turned out that increasing the reliability of electrostatic
parameters greatly affected the statistical results of the QSAR analysis. The 3D QSAR model
proposed could be very useful in the further development of 5RR-1 inhibitors, which are suitable
candidates to be evaluated as drugs in the treatment of 5RR-1 related diseases such as acne
and alopecia in men and hirsutism in women.
In tr od u ction
and characterized (5RR-1 and 5RR-2), which have dif-
ferent chromosomal localization, tissue expression pat-
terns, and enzyme kinetic parameters. Whereas the type
2 isozyme is found predominantly in the prostate,
genital skin, seminal vesicles, epididymis, hair follicle,
and liver, the type 1 isozyme occurs predominantly in
sebaceous glands of the skin (including the scalp) and
liver.7-8,11 The pathologies associated with abnormal
production of DHT and the relative importance of the
two isoenzymes have been identified. In particular, the
presence of the 5RR-1 isozyme in the scalp and skin has
prompted an active research effort toward highly selec-
tive 5RR-1 inhibitors12-17 for their use as therapeutic
tools for the treatment of the pathologies of endocrine
disorders such as androgenetic alopecia, male pattern
baldness, acne in men, and hirsutism in women. The
use of selective inhibitors of the 5RR-1 isoenzyme is
mandatory for the treatment of women because it has
been already ascertained that the inhibition of 5RR-2
can cause abnormalities in the male fetus.18,19
The control of the biological action of steroids through
the inhibition of specific enzymes involved in their
metabolism, without significant changes in the overall
profile of the other hormones, has represented an
attractive pharmaceutical target during the past 20
years. For example, several androgen-dependent dis-
orders and diseases such as prostate cancer, benign
prostatic hyperplasia (BPH), acne and androgenetic
alopecia in men, and hirsutism (associated with poly-
cystic ovarian syndrome) in women appear to be related
to 5R-dihydrotestosterone (DHT) production.1-6 For this
reason, much interest has been paid to the synthesis of
inhibitors of 5R-reductase (5RR),7,8 an enzyme that
catalyzes the reduction of testosterone (T) to DHT.9-10
Steroid 5R-reductase (EC 1.3.99.5) is a membrane-
bound, NADPH-dependent enzyme that catalyzes the
selective, irreversible reduction of 4-ene-3-oxosteroids
to the corresponding 5R-3-oxosteroids.9,10 Two human
isozymes of 5R-reductase have been cloned, expressed,
We have recently reported on a novel class of non-
steroidal inhibitors based on the benzo[c]quinolizin-3-
ones structure (Figure 1).17 The design of these com-
pounds came from the observation that an increase of
structural planarity in the corresponding 19-nor-10-
azasteroidal 5R-reductase inhibitors20,21 (obtained by
introducing a double bond in the C ring, Figure 1)
* To whom correspondence should be addressed. For A.G.: phone
and fax, +39 055 4573481/531; e-mail: antonio.guarna@unifi.it. For
A.C.: phone and fax, +39 051 2099735/4; e-mail, andrea.cavalli@unibo.it.
† Department of Organic Chemistry “U. Schiff”, University of
Firenze.
‡ Department of Clinical Physiopathology, University of Firenze.
§ Serono International S.P.R.I.
#
University of Bologna.
10.1021/jm031131o CCC: $27.50 © 2004 American Chemical Society
Published on Web 06/08/2004

Benzo[c]quinolizin-3-ones
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 14 3547
Ta ble 1. Observed and Calculated (Using Model No. 6 of Table
2) 5RR-1 Inhibitory Activities of the Training Seta
IC50
pIC50
pIC50
(nM)
(obsd)
(calcd)
∆
log P
|F|
12
13
14
15
16
17
18
19
20
21
22
23
26
34
35
36
37
38
39
40
41
42
44
45b
46b
47b
48b
49b
50b
8.10
8.09
6.74
6.95
6.71
7.03
7.03
7.08
6.83
6.41
8.06
7.85
6.77
6.42
6.89
6.93
6.66
6.14
6.00
7.01
6.03
6.25
6.92
5.98
6.73
8.25
7.70
6.53
7.31
6.42
7.95
6.70
6.85
6.44
7.01
7.15
7.15
6.97
6.64
7.91
7.50
7.08
6.44
7.02
6.57
6.60
6.07
6.06
6.83
5.92
6.32
7.08
5.85
7.11
7.95
7.12
6.87
7.72
6.84
0.14
0.04
0.10
0.27
0.02
-0.12
-0.07
-0.14
-0.23
0.15
3.88
2.87
4.64
5.35
5.35
5.35
5.52
6.41
6.41
5.49
5.54
3.35
5.54
4.74
3.13
3.66
3.97
4.37
4.96
4.64
6.47
4.86
4.05
2.85
3.73
3.35
2.33
3.22
2.83
2.44
5.69
5.60
6.73
5.54
5.07
5.20
5.28
4.61
4.09
4.26
3.47
4.25
4.48
3.09
3.30
3.83
4.02
3.39
3.12
3.35
2.24
3.75
4.59
2.35
5.06
5.06
2.83
5.54
178
111
194
93.0
92.9
83.7
149
393
8.75
14.2
170
380
128
117
221
731
1000
96.8
942
560
121
1050
185
5.80
20.0
298
49.0
376
0.35
-0.31
-0.02
-0.13
0.36
0.06
0.07
-0.06
0.18
0.11
-0.07
-0.16
0.13
-0.38
0.30
F igu r e 1. Structure of benzo[c]quinolizin-3-one and Eli Lilly
inhibitors, with their azasteroid relative compounds.
0.58
-0.34
-0.41
-0.42
Ch a r t 1
a
The magnitude of the dipole moment (|F|) and the log P values
are reported as well. Compounds previously synthesized by us.17
b
larger aromatic and heteroaromatic moieties; benzyl
ether groups, with different degree and type of substitu-
tion on the phenyl ring; and alkyl and phenyl ester
groups, with various substituents on the phenyl ring.
Some of these substituents have already proved to
strongly affect the potency of the benzo[f]quinolinone
inhibitors;13-16 other have never been introduced in
nonsteroidal 5R-reductase inhibitors. The only substitu-
ent that we maintained on the benzo[c]quinolizinone
skeleton of these new series was the 4-methyl group,
since its presence seems always associated with an
increase of activity, irrespective of the group at position
8.17 The construction of this homogeneous set of mol-
ecules would, moreover, allow a 3D QSAR study that,
due to the lack of models for the active site of 5RR-1,
would be very helpful in interpreting the biological data
and in providing insight for the design of new inhibitors.
afforded molecules with higher inhibition potency to-
ward 5RR-1 (Schemes 1-6 and Chart 1). Benzo[c]-
quinolizin-3-ones, while maintaining the A ring en-
aminone moiety as an essential feature of the 19-nor-
10-azasteroids, lacked the D ring and incorporated a
benzene ring in place of the C ring to have a more planar
overall structure. By analogy to the series of Eli Lilly
compounds13-16 (Figure 1), formally derived from 4-aza-
steroids, we found that benzo[c]quinolizin-3-one deriva-
tives are selective and that some of them are very
potent, competitive inhibitors of 5RR-1. The potency of
these compounds depended on the presence and position
of double bonds on the A ring and on the type and
number of substituents introduced at positions 1, 4, 5,
6, and 8. Although only a limited series of compounds
with variation of the group at position 8 were synthe-
sized, we observed that the activity was strongly af-
fected by this substituent (for example, compounds with
a lipophilic group such as a methyl or a chlorine atom
at position 8 were very active against 5RR-1).17 There-
fore, we synthesized and tested against 5RR-1 a series
of molecules (Schemes 1-6 and Table 1) obtained by
varying the group at the 8 position to assess its influence
on the inhibitory activity. In particular, we introduced
lipophilic substituents, from small aliphatic chains to
In this paper, we present the synthesis and the
biological evaluation of some new 5RR-1 inhibitors
designed to complete a consistent set of analogues
suitable for a 3D QSAR study. Then we describe the
development of a CoMFA model that attempts to cor-
relate the biological activities of this series of 5RR-1
inhibitors with their physicochemical features. The
model was obtained by taking into account the magni-
tude of the dipole moment and the log P of the molecules
along with the classical electrostatic and steric CoMFA
fields. Moreover, to enhance the quality of the descrip-
tion of the electrostatic contribution, density functional
theory based computations were carried out to evaluate
both the atomic partial charges and the dipole moment
values of each inhibitor. Finally, the CoMFA model was
validated by predicting the biological activities of a set

3548 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 14
Occhiato et al.
Sch em e 1a
a
(a) ClCH2CH2COCl, acetone, reflux, 1 h; (b) AlCl3, fusion, 24 h; (c) TBDMSCl, Imidazole, DMF, 50° C, 2 h, then room temp, 14 h; (d)
Lawesson’s reagent, toluene, reflux, 15 min; (e) ethyl vinyl ketone, K2CO3, 18-crown-6, THF, room temp, 3 h; (f) Me2SO4, DBU, toluene,
reflux, 40 min.
Sch em e 2a
of molecules purposely synthesized and not used in
deriving the 3D QSAR equations.
In this paper, besides proposing a model that ration-
alizes the QSAR of a series of 5RR-1 inhibitors, we
present one of the first applications of density functional
theory (DFT) in the building of a 3D QSAR model.
Hopefully, this will help to strengthen the connection
between theoretical physics and life sciences, which
might provide enormous advantages in the drug design
process as recently pointed out by a number of au-
thors.22
Ch em istr y
All target inhibitors are 4-methylbenzo[c]quinolizin-
a
(a) NBS, DMF, 0° C, 2 h; (b) Lawesson’s reagent, toluene,
3-ones bearing diverse substituents at the 8 position.
We envisioned that the best strategy for the preparation
of a large number of them was to have in hand a bulk
quantity of a common intermediate to be easily, and
possibly in a single step, transformed into the target
compounds. Benzo[c]quinolizinones had been previously
synthesized through a strategy based on a Mannich-
Michael tandem reaction involving N-Boc iminium ions
from lactams and 2-silyloxybutadiene derivatives,17 but
the methodology proved to be unsatisfactory in final
yields and unsuitable for large-scale synthesis. More
recently we have described a new synthetic route based
on a modification of the aza-Robinson annulation still
involving lactams as starting materials, which allowed
us to prepare 4-methyl-substituted benzo[c]quinolin-3-
ones in quantities up to 100 g.23 For our purposes, we
focused our attention on the synthesis of two derivatives
to be used as common intermediates in the synthesis of
most inhibitors, for instance, 8-hydroxy-4-methylbenzo-
[c]quinolizin-3-one 7 (Scheme 1), whose 8-OH group
could be functionalized in order to obtain ethers, and
8-bromo-4-methylbenzo[c]quinolizin-3-one 12 (Scheme
2), in which the aryl halide moiety could be used in Pd-
reflux, 15 min; (c) ethyl vinyl ketone, K2CO3, 18-crown-6, THF,
room temp, 3 h; (d) Me2SO4, DBU, toluene, reflux, 40 min.
catalyzed cross-coupling reactions for the preparation
of 8-alkenyl- and alkynyl derivatives, and carbopalla-
dation reactions for the introduction of ester moieties.
Our modification of the aza-Robinson annulation proved
to be suitable for the preparation of both compounds 7
and 12 as depicted in Schemes 1 and 2.
Regarding the synthesis of compound 7 (Scheme 1),
4-hydroxyaniline 1 was initially converted by treatment
with â-chloropropionyl chloride to amide 2, which
underwent an intramolecular Friedel-Crafts alkylation
by fusion at high temperature in the presence of AlCl3
to give lactam 3 in 63% yield. After protection of the
hydroxy group as a TBDMS ether, Lawesson’s reagent
was used to obtain thiolactam 5 (85% yield), which was
subsequently N-alkylated with ethyl vinyl ketone in
THF and in the presence of potassium carbonate as a
base to give 6 in 60% yield, with the TBDMS protection
being lost in the latter reaction. The final cyclization
was achieved through generation of the thioiminium ion
by treatment first with Me2SO4 in refluxing toluene and

Benzo[c]quinolizin-3-ones
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 14 3549
Sch em e 3a
a
(a) MeI, K2CO3, acetone, reflux, 8 h; (b) K2CO3, acetone, reflux, 6 h.
Sch em e 4a
a
(a) PhCCH, PdCl2(PPh3)2, CuI, Et3N, reflux, 6 h; (b) H2, Pd/BaSO4, pyridine, room temp, 16 h; (c) R2BX, PdCl2(PPh3)2, Na2CO3(aq) 2
M, THF, 80° C, 4-24 h; (d) Bu3SnCHdCH2, Pd(OAc)2, PPh3, Et3N, 95° C, 24 h; (e) H2, (PPh3)3RhCl, C6H6, 40° C, 6 h.
then with DBU as a base to give 7 in 31% yield after
chromatography.
the same procedure using substituted benzyl bromides
in yields ranging from 32% to 75% after chromato-
graphic purification.
The synthesis of 12 was carried out (Scheme 2)
starting from commercially available lactam 8. This was
easily converted to bromo derivative 9 by using N-
bromosuccinimide in DMF. By application of the same
strategy as above, after formation of the thiolactam 10,
in this case a low yield (22%) was obtained in the
N-alkylation step to give 11. This result was essentially
due to competitive S-alkylation of the thiolactam.
Despite this, after the final cyclization compound 12 was
obtained in sufficient amount for all subsequent func-
tionalizations.
As mentioned above, compound 7 was used as the
starting material for the preparation of ethers (Scheme
3); 8-methoxyquinolizinone 13 was obtained in 87%
yield by treatment with MeI in refluxing acetone and
in the presence of K2CO3. Also, a wide variety of
substituted benzyl ethers (14-20) were prepared with
8-Bromo-4-methylbenzo[c]quinolizin-3-one 12 was a
more versatile precursor for the introduction of new
substituents on position 8 because of the possibility of
performing a series of different Pd-catalyzed coupling
reactions, including Sonogashira, Stille, Suzuki-Miyaura,
and carbopalladation reactions. The first three processes
were used for the introduction of small- to medium-sized
lipophilic groups and heterocyclic rings on position 8.
Under the conditions of the Sonogashira reactions, 12
reacted with phenylacetylene (Scheme 4) to give alkynyl
derivative 21 in 31% yield. In this case, conversion to
21 was not complete, but we were lucky to witness
precipitation of the coupling product in Et3N (used as a
base and solvent) mixed with triethylammonium salts.
Derivative 21 was afterward hydrogenated to 22 over
Lindlar catalyst to obtain the Z alkenyl derivative 22

3550 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 14
Occhiato et al.
Sch em e 5a
a
(a) ClCH2CH2COCl, acetone, reflux, 1 h; (b) AlCl3, fusion, 6 h; (c) Lawesson’s reagent, toluene, reflux, 15 min; (d) ethyl vinyl ketone,
K2CO3, 18-crown-6, THF, room temp, 3 h; (e) Me2SO4, DBU, toluene, reflux, 40 min.
Sch em e 6a
a
(a) R1OH, PPh3CO, PdCl2, Et3N, C6H6, 115 °C, 24 h; (b) R2C6H4OH, PPh3CO, PdCl2, Et3N, C6H6, 115 °C, 24 h.
(89% yield after chromatography). These conditions
ensured regioselectivity, since only partial reduction
of the triple bond occurred. Suzuki-Miyaura cross-
couplings were employed in order to insert heterocyclic
substituents on the tricyclic structure (Scheme 4);
3-pyridyl, 2-furanyl, and 2-thiofenyl derivatives 23, 24,
and 25, respectively, were easily prepared performing
couplings with the suitably substituted boronic acids at
80 °C under (Ph3P)2PdCl2 (5%) catalysis in THF with 2
M Na2CO3(aq) as a base. The couplings were complete
in 4-24 h and afforded products in 67-77% yield after
chromatography. Synthesis of E styryl derivative 26 was
achieved in a similar way; the Suzuki reaction was
carried out under the same conditions as above with
2-styrylbenzo[1,3,2]dioxaborole prepared by addition of
catecholborane to phenyl acetylene and gave 26 in 33%
yield after chromatographic purification. A Stille cross-
coupling was finally exploited to obtain vinyl derivative
27 (Scheme 4), employing tributylvinylstannane as the
nucleophilic reagent. Selective hydrogenation of the
vinyl moiety, leaving the enaminone moiety unaltered,
was achieved by using the Wilkinson catalyst. This last
reaction was carried out in benzene at 40 °C and was
complete in 6 h, affording 28 in 75% yield.
thiolactam-based strategy already described for 7 and
12, as depicted in Scheme 5.
Benzo[c]quinolizin-3-ones, bearing ester groups on
position 8, were obtained by carbopalladation of 8-Br
derivative 12 in the presence of a nucleophile (alcohols
and phenols were used) under high pressure (50 bar) of
CO (Scheme 6). We employed the same reaction condi-
tions in all cases, only increasing the amount of the
catalyst in the case of poorer nucleophiles such as
phenols bearing an electron-withdrawing group on the
ring. The reactions were conducted in a steel autoclave,
using 5% PdCl2 as a catalyst in the presence of 10%
Ph3P with Et3N as a base and in benzene as a solvent
for 24 h at 120 °C. Different equivalents of nucleophile
were used. Yields of carbopalladation ranged from 46%
to 88% after chromatography, with the exception of the
tert-butyl ester derivative 38, which underwent acid-
catalyzed hydrolysis during purification on silica gel,
and the p-tolyl ester derivative 44, which required an
additional chromatographic purification. By this meth-
odology, alkyl esters 35-39 with increasing bulkiness
in the alcoholic moiety R1 were prepared, as well as aryl
esters 40-44 with polar and lipophilic substituents R4
on the aromatic moiety.
The synthesis of 8-phenyl-substituted compound 34
(Scheme 5), which is structurally related to the hetero-
cyclic derivatives 23-26, could be realized through a
Suzuki reaction using phenylboronic acid. However, we
did not evaluate this possibility, since this compound
had previously been obtained starting from commer-
cially available biphenyl-4-ylamine 29 by employing the
In h ibition Tests
The synthesized compounds were tested with the
human recombinant isozyme of 5RR-1 to evaluate the
inhibitory potency using finasteride as a control.17a,24
The assays were performed using stably transfected
Chinese hamster ovary (CHO) 1827 cells25 incubated for

Benzo[c]quinolizin-3-ones
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 14 3551
Ta ble 2. Summary of the PLS Runs with Different Sets of Independent Variables and Quantum Chemical Levels of Theory for the
Electrostatics Calculationsa
contribution
model
no.
optimal no.
of components
independent variables
steric only
q2
scross
F
r2
s
steric electrostatic log P
|F|
1
2
3
4
5
4
2
3
3
4
0.480 0.484 3.381 0.887 0.225 1.000
0.260 0.554 1.289 0.672 0.369
0.542 0.445 1.290 0.836 0.266 0.332
0.526 0.453 4.061 0.784 0.305 0.319
0.591 0.429 5.290 0.892 0.220 0.233
electrostatic only
1.000
0.668
0.607
0.625
steric and electrostatic
steric, electrostatic, and log p
steric, electrostatic, and
dipole moment
0.075
0.143
6
ster ic, electr osta tic, log p, a n d
d ip ole m om en t (DF T)b
steric, electrostatic, log p, and
dipole moment (PM3)c
3
0.628 0.401 6.197 0.841 0.262 0.233
0.548
0.060 0.159
7
3
0.229 0.577 1.087 0.628 0.401 0.479
0.384
0.037 0.099
a
b
The analysis illustrated by the contour maps of Figure 2 is bold. Both partial charges and dipole moment are calculated at the DFT
level. c Both partial charges and dipole moment are calculated at the PM3 level.
Ta ble 3. Observed and Calculated 5RR-1 Inhibitory Potency of
3
30 min with [H]-testosterone at the Km concentration
(2 µM for 5RR-1) and each inhibitor in the 10-9-10-5
M concentration range. Data were processed with the
program ALLFIT26 using the four-parameter logistic
equation to calculate the IC50 values. The interassay
reproducibility of the method was good as assessed by
calculating the mean IC50 of finasteride, which resulted
in 911 ( 85 nM (CV ) 9.4%, n ) 15) and 21 ( 1.8 nM
(CV ) 8.6%, n ) 12) for 5RR-1 and 5RR-2, respectively,
consistent with the established selectivity reported for
the inhibitor.19
the Test Seta
IC50
pIC50
pIC50
(nM)
(obsd)
(calcd)
∆
log P
|F|
24
25
27
28
43
133
79
21.6
14
965
6.88
7.10
7.67
7.85
6.02
7.30
7.19
7.52
7.12
6.14
-0.42
-0.09
0.15
0.73
-0.12
4.12
4.62
3.57
3.88
4.86
4.81
4.35
4.15
5.03
3.60
a
The predictions were obtained by using model no. 6 of Table
2.
For the mechanism of action of the control inhibitor
finasteride toward human recombinant 5RR-1, it was
demonstrated that finasteride behaves as an irreversible
or psudoirreversible inhibitor when tested in the intact
cell system.19 Finastride was described as an active site
directed slow time dependent inhibitor of human 5RR-
1.27
pounds, a 3D QSAR model was developed by employing
the CoMFA procedure. To this aim, several statistical
analyses (PLS) were carried out to establish the best
set of independent variables to be used for the calcula-
tion of the CoMFA model. The DFT-calculated dipole
moment and the log P values were taken into account
along with the classical CoMFA independent variables
(steric and electrostatic fields). In Table 2, the statistics
of the PLS runs carried out with the steric and electro-
static fields only, those with the two fields plus either
dipole moment or log P values, and finally, the PLS
analysis with all of the computed independent variables
are reported. Moreover, to assess the relevance of the
DFT-based electrostatic calculations, CoMFA analyses
with geometries, charges, and dipole moments calcu-
lated at the PM3 semiempirical level were also carried
out (model no. 7, Table 2). Clearly, the computations at
the DFT level greatly improved the statistics of the PLS
analyses even though, concerning the geometry, the
semiempirical PM3 and the ab initio DFT calculations
provided fairly similar results (the mean rmsd of non-
hydrogen atoms between PM3- and DFT-optimized
molecules was equal to 0.26 Å).
Molecu la r Mod elin g
The molecular models were built by properly modify-
ing the crystallographic 19-nor-10-aza-androstenedione
skeleton retrieved from the Cambridge Structural
Database.21 After minimization, the inhibitors were
thoroughly studied by means of Monte Carlo conforma-
tional searches, and the conformers were classified
using cluster analysis.
The conformations chosen for the 3D QSAR analyses
were further optimized at the DFT level, and the partial
charges of the compounds were evaluated by means of
the restrained electrostatic potential (RESP) fitting
procedure.
The 3D QSAR analyses were carried out using the
CoMFA method. The alignment of the molecules within
the Cartesian space was accomplished by superimposing
atom by atom the benzo[c]quinolizin-3-one skeleton of
the set of compounds. The molecules bearing a flexible
group at position 8 of the benzo[c]quinolizin-3-one
moiety were aligned by using the Z and E styryl
derivatives 22 and 26 as templates.
The best CoMFA analysis for the training set of 5RR-1
inhibitors (Table 1) was that obtained by taking into
account both the dipole moment and log P values along
with steric and electrostatic fields (model no. 6, Table
2). The use of these variables afforded the best statistics
for both cross-validated and non-cross-validated PLS.
The selected 3D QSAR model has an optimal number
of components equal to 3 and descriptive and predictive
Resu lts
abilities evaluated by the statistical parameters r2
)
The inhibitory potency of the present series of 5RR-1
inhibitors is reported in Table 1. The newly synthesized
compounds (12-44) are all, with a few exceptions, good
5RR-1 inhibitors, with IC50 values ranging from 8 nM
to roughly 1 µM.23 To quantitatively rationalize the
structure-activity relationships of the series of com-
0.841 and s ) 0.262, and q2 ) 0.628 and scross ) 0.401.
In addition, the predictive properties of the CoMFA
model were more rigorously tested by calculating the
5RR-1 inhibitory potency of a set of molecules purposely
synthesized and tested (Table 3). The resulting predic-
tive correlation coefficient r2
was equal to 0.676.40
pred

3552 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 14
Occhiato et al.
F igu r e 3. Dipole moment vector of compound 39 calculated
at DFT/B3LYP (left) and PM3 semiempirical (right) levels.
This physical observable is quite different between the two
quantum chemical levels of theory in terms of both magnitude
and direction.
contours (Figure 2b, red), two main regions were identi-
fied. One (bottom) was generated around both the
carbonyl oxygen of the esters 35, 36, 39, and 42, as well
as the sp3 oxygen of the ethers and esters 20, 37, 38,
40, and 41. The other (top) corresponded to the carbonyl
oxygen of 37, 38, 40, and 41, as well as to the phenyl
ring of the inhibitors aligned onto 22 (14-19, 35, 36,
39, 42), the latter being clearly embedded in the
electrostatic positive region. All these molecules are of
relatively lower potency than 12, 21, and 46, and for
this reason, the graphical model shows some apparently
contradictory features. However, our interpretation is
that an increase of the positive electrostatic potential
in that region is required in order to obtain more potent
compounds.
The contributions to the CoMFA model of both the
magnitude of the dipole moment (|F|) and log P are
shown in Table 2. These values indicate a higher
contribution of |F| than log P to the calculated biological
activity. Moreover, from the data of Table 1, it appears
that a decrease of |F| is associated with an increase of
the inhibitory potency, whereas it seems that log P has
to be increased to obtain more potent compounds.
Concerning the dipole moment, these were also
estimated at the PM3 semiempirical level. In Figure 3,
the dipole moment of compound 39, calculated at DFT
(left) and PM3 (right) levels, is reported. The dipole
moment is shown in green, and the vector is reported
in terms of magnitude and X, Y, and Z components. It
clearly appears that the dipole moment was poorly
estimated by PM3 with respect to that evaluated
ab initio, in terms of both vector magnitude and direc-
tion (i.e., the orientation of the vector in Cartesian
space).
F igu r e 2. (a) View of the steric CoMFA STDEV*COEFF
contour maps. The regions where increasing the molecular
volume increases 5RR-1 inhibitory activity are green (0.0035
level), and the regions where increasing the volume decreases
the activity are yellow (-0.012 level). The inhibitors shown
are the Z-22 (violet) and E-26 (orange) diastereoisomers of the
phenylvinyl derivative. (b) View of the electrostatic CoMFA
STDEV*COEFF contour maps. The regions where increasing
the positive charges increases 5RR-1 inhibitory activity are red
(0.03 level), and the regions where increasing the negative
charges increases the activity are blue (-0.02 level). The
inhibitors shown are our most potent 5RR-1 inhibitors 46
(violet) and 39 (orange).
Thus, r2
turned out to be of the same order of
pred
magnitude as q2 and it can therefore be considered
satisfactory.
The CoMFA model no. 6 accounting for the 3D QSAR
of compounds of Table 1 is illustrated by the contour
maps shown in Figure 2. Sterically favorable regions
(Figure 2a, green contours) are quite split with a main
appearance around the 4-methyl group, which was
already pointed out as a fundamental substituent for
an optimal 5RR-1 inhibitory profile.17a Sterically un-
favorable regions (Figure 2a, yellow contours) are
located around the substituent at position 8. In par-
ticular, the phenyl group of the styryl-substituted
derivative 26 (diastereoisomer E, orange in Figure 2a),
along with all inhibitors aligned onto it, protrudes inside
the sterically unfavorable region, whereas the dia-
stereoisomer Z (22, violet in Figure 2a) does not make
any contact with that region. Electrostatic positive and
negative regions (Figure 2b) are located around the
substituent at position 8 as well. In particular, the
negative CoMFA contours (blue) are close to the C8
atom of the benzo[c]quinolizinone moiety, where the
chlorine (compound 46, violet in Figure 2b) and the
bromine (compound 12) atoms, as well as the triple bond
of 21, and the oxygen atoms of 14-19, 35, 36, 39 (orange
in Figure 2b), and 42 are located. This means that in
such a region, where relatively more potent compounds
orient their 8-substituent, an increase of negative
electrostatic potential should provide more potent 5RR-1
inhibitors. Concerning the positive electrostatic CoMFA
Discu ssion
In this study, we present a CoMFA model for a series
of 5RR-1 inhibitors currently under investigation for the
treatment of several related endocrine disorders. Our
CoMFA model describes the 3D QSAR for a series of
benzo[c]quinolizin-3-one derivatives, of which a few have
already been reported.17 So far, interpretation of the
inhibition experimental results has been based on

Benzo[c]quinolizin-3-ones
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 14 3553
qualitative structure-activity relationships.17 We have
already shown that the overall planarity of the benzo-
[c]quinolizinones could be a fundamental feature for
good binding with 5RR-1,17a although favorable interac-
tions between the C-ring unsaturations and the enzyme
active site should also be taken into account.28 Besides
this important feature, ring substitution plays a fun-
damental role in modulating the inhibitory activity. We
have already shown, for example, that changing the
position of a single methyl group on the benzo[c]-
quinolizin-3-one structure can dramatically affect the
inhibition potency.17 In particular, substitution at the
8 position had a major effect on the activity, and for this
reason, we undertook this study aimed at obtaining,
through a quantitative study, a deeper insight into these
structure-activity relationships. To do this, an initial
accurate evaluation of the atomic partial charges was
carried out, and to our knowledge, for the first time a
CoMFA model was developed using electrostatic com-
putations at the DFT/B3LYP level. This was required
because both the atomic partial charges and the dipole
moment at the semiempirical quantum chemical level
did not allow development of a significant 3D QSAR
model (see Table 2). For the set of molecules under
investigation in the present study, the electronic cor-
relation effects seem to be of paramount importance to
correctly describe the charge distribution. In fact, the
high conjugation within the tricyclic system and the
electron delocalization make the electronic correlation
effects crucial for a correct description of the charge
distribution as well as of chemical and geometrical
features of the molecules. Therefore, an accurate treat-
ment of the electrostatic properties of the molecules was
required in order to achieve a good CoMFA model.
the statistics of the model (Table 2). Again, the dipole
moment calculated at the semiempirical level was fairly
different from those estimated by means of DFT (Figure
3), in terms of both magnitude and direction. This once
more shows that the exchange and correlation electronic
effects, which are well accounted for by means of DFT/
B3LYP-based computations,29 can be fundamental when
evaluating physicochemical features for this series of
molecules.
The importance of the dipole moment in both enzyme-
inhibitor recognition and interaction has been recently
pointed out by Sulpizi et al., who have performed a
structure-activity relationship study on a series of
inhibitors of herpes simplex virus type 1 thymidine
kinase.30 The dipole moment is a physicochemical
parameter that accounts well for long-range interaction
during the enzyme-inhibitor recognition phase, whereas
the electrostatic field of the ligands may be useful to
identify the interactions at the active site of the biologi-
cal counterparts. In the present case, this might be
taken to mean that our QSAR study not only reveals
the role of physicochemical properties involved in the
enzyme-inhibitor interaction (classical CoMFA fields)
but also takes into account an aspect that can be related
to the recognition phase between the ligands and their
biological counterpart. However, Sulpizi et al. pointed
out that the dipole moment could be important for the
enzyme-inhibitor stabilization energy as well. There-
fore, since so far we do not know anything about the
three-dimensional features of the biological target
(5RR-1), the present CoMFA model suggests a role
for the dipole moment in driving both enzyme recogni-
tion and binding by the ligands. The determination of
the 5RR-1 three-dimensional structure will hopefully
verify such a hypothesis, allowing the study of these
inhibitors in the light of the active site features of the
enzyme.
Concerning log P, despite the low contribution of this
parameter to the CoMFA model, it appears that an
increase of the log P value might lead to more potent
inhibitors. This is in good agreement with our previous
observations that lipophilic groups at the 8 position
increase the inhibitory potency,17a a feature also re-
ported for some Eli Lilly compounds.13-16 However, it
should be noted that the experimental assays to assess
the inhibitory potency are usually carried out in a cell
system, which implies that the molecules have to pass
the cell membrane to interact with the enzyme. There-
fore, more hydrophobic inhibitors can penetrate through
this barrier better than hydrophilic ones.
Model no. 6 of Table 2 shows that the electrostatic
contribution to the final 3D QSAR equation is remark-
ably higher than the steric one. For example, in Figure
2a the steric CoMFA contour maps indicate that the
positive steric contribution is almost irrelevant and
mainly located around the 4-methyl group. This is in
line with the fact that the most potent inhibitors of the
series (12 and 46) do not bear a bulky substituent,
indicating that their very high potency is more likely
related to electrostatic effects. On the other hand, the
negative steric contours (yellow) seem to be much more
descriptive, accounting for the biological activities of the
low potency inhibitors superimposed onto 26.
Concerning the electrostatic CoMFA maps, they span
very well-defined regions of the Cartesian space. This
feature does not seem to be very often observed in
CoMFA studies. In particular, the negative fragments
at position 8 of the most and least potent inhibitors point
inside the negative (blue in Figure 2b) and the positive
(red in Figure 2b) electrostatic regions, respectively.
Such an accurate description of the electrostatic 3D
features of the series of 5RR-1 inhibitors was possible
only for the 3D QSAR model built by using the DFT-
based atomic partial charges. Thus, we can observe that
ab initio quantum chemical calculations of geometry and
electrostatics of the molecules can play a crucial role
for the “standard” CoMFA procedures.
Con clu sion s
A 3D QSAR model was developed for a series of non-
steroidal 5RR-1 inhibitors having the benzo[c]quinolizin-
3-one structure. These were synthesized by a modifica-
tion of the aza-Robinson annulation and further func-
tionalized by Pd-catalyzed cross-coupling processes and
tested with human 5RR-1 expressed in CHO 1827 cells.
The potency of the resulting inhibitors was strongly
dependent on the type of substitution at the 8 position,
with the IC50 values ranging from 8.1 to 1050 nM. To
obtain a robust statistical model in building the 3D
QSAR model, two “nonstandard” variables had to be
added to the classical CoMFA fields, i.e., dipole moment
Moreover, the dipole moment and the log P values,
included as independent variables in the computation
of the 3D QSAR equation, indeed contributed to improve

3554 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 14
Occhiato et al.
2.87 (t, J ) 7.8 Hz, 2H), 2.56 (t, J ) 7.8 Hz, 2H), 0.94 (s, 9H),
0.15 (s, 6H); 13C NMR (CDCl3) δ 171.9 (s), 151.3 (s), 131.3 (s),
124.8 (s), 119.6 (d), 118.6 (d), 116.1 (d), 30.6 (t), 25.6 (t), 25.5
(q), -4.5 (s), -4.5 (q); MS (m/z) 277 (M+, 73), 220 (100). Anal.
(C15H23NO2Si) C, H, N.
6-(ter t-Bu tyld im eth ylsila n yloxy)-3,4-d ih yd r oqu in olin -
2(1H)-th ion e (5). A mixture of 4 (390 mg, 1.4 mmol) and
Lawesson’s reagent (284 mg, 0.7 mmol) was suspended in
anhydrous toluene (8 mL) and then refluxed under nitrogen
atmosphere for 15 min. The solution was then cooled to room
temperature and evaporated. The residue was purified by
chromatography (EtOAc/petroleum ether 1:15, Rf ) 0.30) to
give 4 (350 mg, 85%) as a yellow solid: 1H NMR (CDCl3) δ
9.54 (s, 1H), 6.71-6.61 (m, 3H), 3.04 (t, J ) 7.8 Hz, 2H), 2.78
(t, J ) 7.8 Hz, 2H), 0.94 (s, 9H), 0.16 (s, 6H).
and log P. Moreover, the statistics of the 3D QSAR
equation were greatly affected by the point charges and
dipole moment accuracy. In particular, DFT calculations
using the exchange and correlation functional B3LYP
were required to get charges and dipole moment to be
employed in the PLS analyses. In this respect, this
paper represents one of the first examples in medicinal
chemistry in which the QSAR was obtained by calculat-
ing parameters at the DFT quantum chemical level. We
believe that the rapid increase in CPU speed will make
such calculations routine in future computational me-
dicinal chemistry investigations and provide advantages
in the drug design process.22 Finally, the 3D QSAR
model proposed could be very useful in the further
development of these inhibitors, which are compounds
amenable to evaluation as drugs in the treatment of
5RR-1 related diseases such as acne and alopecia in men
and hirsutism in women.
6-(ter t-Bu t yld im et h ylsila n yloxy)-1-(3-oxop en t yl)-3,4-
d ih yd r oqu in olin -2(1H)-th ion e (6). Compound 5 (350 mg,
1.2 mmol), anhydrous K2CO3 (378 mg, 2.7 mmol) (kept at 140
°C for 12 h before use), and 18-crown-6 (41 mg, 0.15 mmol)
were suspended in anhydrous THF (20 mL). After the mixture
was cooled to 0 °C, the first portion of ethyl vinyl ketone (160
µL, 1.6 mmol) was added dropwise under stirring and nitrogen
atmosphere. After the addition was complete, the solution was
left at 0 °C for 5 min and then at room temperature for 1 h.
The solution was cooled again to 0 °C, and a second portion of
ethyl vinyl ketone (160 µL, 1.6 mmol) was added. The solution
was left under stirring for 1.5 h at room temperature, the
reaction being monitored by TLC. Then Na2SO4 was added,
and the solution was filtered and concentrated, providing crude
6, which was purified by chromatography (EtOAc/petroleum
ether 1:2, Rf ) 0.13) affording pure 6 (190 mg, 60%) as a yellow
solid: mp 157-159 °C; 1H NMR (CDCl3) δ 6.98 (d, J ) 8.8
Hz, 1H), 6.70-6.64 (m, 2H), 4.70 (t, J ) 7.7 Hz, 2H), 3.10 (t,
J ) 7.7 Hz, 2H), 2.97 (t, J ) 7.7 Hz, 2H), 2.70 (t, J ) 7.7 Hz,
2H), 2.47 (q, J ) 7.3 Hz, 2H), 1.05 (t, J ) 7.7 Hz, 3H).
8-Hyd r oxy-4-m eth yl-2,3,5,6,-tetr a h yd r o-(1H)-ben zo[c]-
qu in olizin -3-on e (7). Compound 6 (150 mg, 0.57 mmol) was
suspended under a nitrogen atmosphere in anhydrous toluene
(10 mL), and then freshly distilled Me2SO4 (150 µL, 1.5 mmol)
was added dropwise at room temperature and under stirring.
After the addition was complete, the flask was placed in an
oil bath at 140-150 °C. After the mixture was refluxed for 15
min, DBU (230 µL, 1.5 mmol) was added dropwise and the
resulting solution was refluxed for a further 20 min. Then it
was cooled to room temperature, diluted with CH2Cl2, and
washed with water. The organic layer was dried over Na2SO4
and evaporated, affording an orange solid that was washed
with MeOH. After this treatment, pure compound 7 (41 mg,
31%) was obtained as a yellow solid: 1H NMR (DMSO-d6) δ
6.67-6.52 (m, 3H), 3.72 (t, J ) 7.7 Hz, 2H), 2.61-2.42 (m, 6H),
1.64 (s, 3H); 13C NMR (DMSO-d6) δ 188.7 (s), 156.1 (s), 152.0
(s), 132.5 (s), 127.8 (s), 114.5 (d), 114.3 (d), 113.3 (d), 103.0 (s),
44.6 (t), 35.1 (t), 25.9 (t), 24.3 (t), 9.9 (q); MS (m/z) 203 (6), 160
(6), 84 (100), 66 (84). Anal. (C14H15NO2) C, H, N.
6-Br om o-3,4-d ih yd r oqu in olin -2(1H)-on e (9). To a solu-
tion of 3,4-dihydroquinolin-2(1H)-one 8 (10.19 g, 69.3 mmol)
in DMF (220 mL) cooled at 0 °C, a solution of N-bromo-
succinimide (13.1 g, 73.4 mmol) in DMF (220 mL) was added
dropwise. The mixture was stirred at 0 °C for 2 h, then water
(440 mL) was added and the solution was extracted with
EtOAc/toluene 1:1 (400 mL + 2 × 200 mL). The organic phase
was washed with water (2 × 200 mL), then dried over Na2SO4
and evaporated, affording a white solid, which was purified
by crystallization (EtOH/H2O 8:2), providing pure 9 (13.63 g,
87%) as white crystals: 1H NMR (CDCl3) δ 8.87 (s, 1H), 7.28
(s, 1H), 7.26 (d, J ) 8.8 Hz, 1H), 6.68 (d, J ) 8.8 Hz, 1H), 2.93
(t, J ) 8.1 Hz, 2H), 2.60 (t, J ) 8.1 Hz, 2H).
6-Br om o-3,4-d ih yd r oqu in olin -2(1H)-th ion e (10). 10 was
prepared as reported for 5. Starting from 9 (6.01 g, 26.6 mmol)
and refluxing the reaction mixture for 45 min, pure 10 (5.07
g, 79%) was obtained as a yellow solid after chromatography
(EtOAc/petroleum ether 1:5, Rf ) 0.30): 1H NMR (CDCl3) δ
9.53 (s, 1H), 7.30 (m, 2H), 6.70 (d, J ) 8.8 Hz, 1H), 3.07 (t, J
) 8.0 Hz, 2H), 2.85 (t, J ) 8.0 Hz, 2H).
Exp er im en ta l Section
1. Ch em istr y. All the reactions were performed under
nitrogen unless otherwise stated. Chromatographic separa-
tions were performed under pressure on silica gel using flash
column techniques. Rf values refer to TLC carried out on 25
mm silica gel plates (Merck F254), with the same eluant
indicated for the column chromatography. IR spectra were
recorded on a Perkin-Elmer 881 spectrophotometer in CDCl3
solution. 1H NMR (200 MHz) and 13C NMR (50.33 MHz)
spectra were recorded on a Varian XL 200 instrument in CDCl3
solution. Mass spectra were carried out via EI at 70 eV on
5790A-5970A Hewlett-Packard and QMD 1000 Carlo Erba
instruments. Microanalyses were carried out with a Perkin-
Elmer 240C elemental analyzer.
6-Hyd r oxy-3,4-d ih yd r oqu in olin -2(1H)-on e (3). To a re-
fluxing solution of p-aminophenol 1 (10.4 g, 95.3 mmol) in
acetone (25 mL) was slowly added a solution of 3-chloro-
propanoyl chloride (5 mL, 52.4 mmol) in acetone (10 mL). Then
the solution was refluxed for 1 h and finally cooled to room
temperature. The resulting suspension was transferred into
a flask containing a solution of 6 N HCl in water (150 mL),
and the solution was extracted with 3 × 150 mL of diethyl
ether. The organic phase was washed with a saturated solution
of NaHCO3 and brine, then dried over Na2SO4 and evaporated
under reduced pressure, affording amide 2 (7.81 g, 75%) as a
brown solid: mp 85-87 °C; 1H NMR (CDCl3) δ 7.33 (AB
system, 2H), 7.11 (s, 1H), 6.77 (AB system, 2H), 3.85 (t, J )
6.4 Hz, 2H), 2.75 (t, J ) 6.4 Hz, 2H). A flask equipped with a
mechanical stirrer and containing compound 2 (7.7 g, 38.5
mmol) was put in an oil bath heated at 160 °C, and after
complete melting of 2, AlCl3 (18.0 g, 193 mmol) was added in
small portions and the mixture was left under stirring at 160
°C for 24 h. After the mixture was cooled to room temperature
and then to 0 °C, 10% HCl (100 mL) was added and the
mixture was extracted with EtOAc. The organic phase was
dried over Na2SO4, and after concentration lactam 3 (3.96 g,
63%) was obtained as a dark-red solid sufficiently pure for the
next step: 1H NMR (DMSO-d6) δ 6.91-6.76 (m, 3H), 3.06 (t,
J ) 7.5 Hz, 2H), 2.70 (t, J ) 7.5 Hz, 2H); MS m/z 163 (M+,
19), 134 (15), 109 (19), 91 (100).
6-(ter t-Bu tyld im eth ylsila n yloxy)-3,4-d ih yd r oqu in olin -
2(1H)-on e (4). To a solution of hydroxylactam 3 (300 mg, 1.8
mmol) in dry DMF (5 mL) under a N2 atmosphere, imidazole
(310 mg, 4.5 mmol) and TBDMSCl (329 mg, 2.2 mmol) were
added. The solution was heated to 50 °C for 2 h, then cooled
and maintained at room temperature overnight. After this
period the solution was diluted with 5% NaHCO3 (25 mL) and
extracted with petroleum ether and CH2Cl2. The organic phase
was dried and concentrated, and the crude brown solid was
purified by chromatography (EtOAc/petroleum ether 1:2, Rf
) 0.28) to give 4 (490 mg, 96%) as white crystals: mp 151-
1
153 °C; H NMR (CDCl3) δ 7.72 (s, 1H), 6.63-6.53 (m, 3H),

Benzo[c]quinolizin-3-ones
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 14 3555
6-Br om o-1-(3-oxop en t yl)-3,4-d ih yd r oq u in olin -2(1H )-
th ion e (11). 11 was prepared as reported for 6. Starting from
10 (1.80 g, 7.43 mmol), pure 11 (528 mg, 22%) was obtained
as a yellow solid after chromatography (EtOAc/petroleum ether
1:9, Rf ) 0.24): 1H NMR (CDCl3) δ 7.37-7.29 (m, 2H), 6.97
(d, J ) 8 Hz, 1H), 4.70 (t, J ) 8.4 Hz, 2H), 3.12 (t, J ) 7.3 Hz,
2H), 2.96 (t, J ) 7.7 Hz, 2H), 2.73 (t, J ) 7.3 Hz, 2H), 2.43 (q,
J ) 7.3 Hz, 2H), 1.05 (t, J ) 7.3 Hz, 3H).
8-Br om o-4-m et h yl-2,3,5,6,-t et r a h yd r o-(1H )-b en zo[c]-
qu in olizin -3-on e (12). 12 was prepared as reported for 7.
Starting from 11 (528 mg, 1.62 mmol), pure 12 (199 mg, 42%)
was obtained as an orange solid after chromatography (EtOAc/
petroleum ether 1:1, Rf ) 0.32): 1H NMR (CDCl3) δ 7.32 (dd,
J 1 ) 8.8 Hz, J 2 ) 2.6 Hz, 1H), 7.24 (s, 1H), 6.79 (d, J ) 8.4 Hz,
1H), 3.87 (t, J ) 7.7 Hz, 2H), 2.77-2.61 (m, 6H), 1.79 (s, 3H);
13C NMR (CDCl3) δ 190.7 (s), 155.3 (s), 139.7 (s), 130.4 (d),
128.8 (s), 125.9 (s), 114.4 (d), 114.0 (d), 106.6 (s), 45.1 (t), 35.4
(t), 26.1 (t), 24.7 8t), 10.1 (q); MS (m/z) 293 (72), 292 (M+, 72),
291 (77), 262 (16), 248 (14), 235 (15), 212 (17), 154 (20), 149
(23), 128 (19), 89 (18), 84 (100), 77 (25). Anal. (C14H14BrNO)
C, H, N.
353 (M+, 23), 228 (100), 200 (53), 172 (15), 125 (20), 89 (14),
57 (22). Anal. (C21H20ClNO2) C, H, N.
8-(4-Ch lor ob en zyloxy)-4-m et h yl-2,3,5,6,-t et r a h yd r o-
(1H)-ben zo[c]qu in olizin -3-on e (17). 17 was prepared as
reported for 14, starting from 12 (25 mg, 0.11 mmol) and using
4-chlorobenzyl bromide (45 mg, 0.22 mmol). After chromatog-
raphy (EtOAc/petroleum ether 3:2, Rf ) 0.25) pure 15 (15 mg,
38%) was obtained as a pale-yellow solid: mp 145-149 °C;
1H NMR (CDCl3) δ 7.36 (ps, 4H), 6.85-6.79 (m, 3H), 5.01 (s,
2H), 3.89 (t, J ) 7.4 Hz, 2H), 2.73-2.61 (m, 6H), 1.81 (s, 3H);
13C NMR (CDCl3) δ 190.2 (s), 156.3 (s), 153.4 (s), 143.2 (s),
134.6 (s), 133.7 (s), 128.7 (d), 128.6 (d), 128.3 (s), 114.9 (d),
113.7 (d), 113.1 (d), 105.2 (s), 69.7 (t), 45.3 (t), 35.4 (t), 26.4
(t), 25.2 (t), 10.2 (q); MS (m/z) 353 (M+, 31), 230 (47), 228 (100),
200 (82), 172 (24), 127 (29), 125 (58), 89 (25), 57 (18). Anal.
(C21H20ClNO2) C, H, N.
8-(4-Tr iflu or om et h ylb en zyloxy)-4-m et h yl-2,3,5,6,-t et -
r a h yd r o-(1H)-ben zo[c]qu in olizin -3-on e (18). 18 was pre-
pared as reported for 14, starting from 12 (40 mg, 0.17 mmol)
and using 4-trifluoromethylbenzyl bromide (55 µL, 0.35 mmol).
After chromatography (EtOAc/petroleum ether 1:1, Rf ) 0.29)
pure 18 (47 mg, 72%) was obtained as a pale-yellow solid: mp
107-111 °C; 1H NMR (CDCl3) δ 7.67-7.48 (m, 4H), 6.90-6.78
(m, 3H), 5.10 (s, 2H), 3.88 (t, J ) 8.2 Hz, 2H), 2.73-2.61 (m,
6H), 1.81 (s, 3H); 13C NMR (CDCl3) δ 190.2 (s), 156.2 (s), 153.2
(s), 141.0 (s), 134.8 (s), 128.3 (s), 127.2 (2d, 2C), 125.4 (s), 114.8
(d), 113.7 (d), 113.1 (d), 105.2 (s), 69.5 (t), 45.2 (t), 35.4 (t),
26.4 (t), 25.1 (t), 10.2 (q) (CF3 nd); MS (m/z) 387 (M+, 19), 228
(100), 159 (21), 149 (56). Anal. (C22H20F3NO2) C, H, N.
8-(2,4-Bis-tr iflu or om eth ylben zyloxy)-4-m eth yl-2,3,5,6-
tetr a h yd r o-(1H)-ben zo[c]qu in olizin -3-on e (19). 19 was
prepared as reported for 14, starting from 12 (25 mg, 0.11
mmol) and using 2,4-bis-trifluoromethylbenzyl bromide (40 µL,
0.22 mmol). After chromatography (EtOAc/petroleum ether
2:3, Rf ) 0.21) pure 19 (16 mg, 32%) was obtained as a pale-
yellow oil: 1H NMR (CDCl3) δ 7.96-7.82 (m, 3H), 6.94-6.80
(m, 3H), 5.30 (s, 2H), 3.91 (t, J ) 7.7 Hz, 2H), 2.76-2.63 (m,
6H), 1.83 (s, 3H); 13C NMR (CDCl3) δ 190.3 (s), 156.2 (s), 152.7
(s), 135.1 (s), 128.9 (2d, 2C), 128.4 (s), 126.1 (d), 123.1 (s), 120.7
(s), 115.6 (s), 114.7 (d), 113.8 (d), 113.1 (d), 105.4 (s), 65.8 (t),
45.2 (t), 35.3 (t), 26.3 (t), 25.1 (t), 10.1 (q); MS (m/z) 455 (M+,
8), 234 (17), 228 (100), 200 (33), 177 (19), 160 (15) (2 CF3 nd).
Anal. (C23H19F3NO2) C, H, N.
8-Meth oxy-4-m eth yl-2,3,5,6,-tetr a h yd r o-(1H)-ben zo[c]-
qu in olizin -3-on e (13). To a solution of 12 (40 mg, 0.17 mmol)
in acetone (3 mL) under a nitrogen atmosphere, K2CO3 (120
mg, 0.87 mmol) and MeI (110 µL, 1.7 mmol) were added. The
mixture was refluxed for 8 h, then cooled to room temperature,
filtered, and concentrated, affording 13 (37 mg, 87%) as a
yellow oil: 1H NMR (CDCl3) δ 6.86 (d, J ) 8.4 Hz, 1H), 6.76-
6.69 (m, 2H), 3.87 (t, J ) 7.7 Hz, 2H), 3.77 (s, 3H), 2.76-2.58
(m, 6H), 1.79 (s, 3H); 13C NMR (CDCl3) δ 190.3 (s), 156.4 (s),
154.6 (s), 134.3 (s), 128.2 (d), 113.8 (d), 113.7 (s), 112.0 (d),
105.0 (s), 55.6 (q), 45.2 (t), 35.4 (t), 26.4 (t), 25.1 (t), 10.1 (q);
MS (m/z) 243 (M+, 85), 212 (56). Anal. (C15H17NO2) C, H, N.
8-Ben zyloxy-4-m eth yl-2,3,5,6,-tetr a h yd r o-(1H)-ben zo-
[c]qu in olizin -3-on e (14). To a solution of 12 (40 mg, 0.17
mmol) in acetone (3 mL) under a nitrogen atmosphere, K2CO3
(120 mg, 0.87 mmol) and benzyl bromide (42 µL, 0.35 mmol)
were added. The mixture was refluxed for 6 h, then cooled to
room temperature, filtered, and concentrated, affording pure
14 (39 mg, 70%): 1H NMR (CDCl3) δ 7.39-7.33 (m, 5H), 6.84-
6.78 (m, 3H), 5.02 (s, 2H), 3.87 (t, J ) 7.7 Hz, 2H), 2.72-2.58
(m, 6H), 1.79 (s, 3H); 13C NMR (CDCl3) δ 190.2 (s), 156.3 (s),
153.6 (s), 136.9 (s), 134.4 (s), 128.5 (d), 128.1 (s), 127.9 (d),
127.3 (d), 114.8 (d), 113.7 (d), 113.1 (d), 105.0 (s), 70.4 (t), 45.2
(t), 35.4 (t), 26.4 (t), 25.1 (t), 10.1 (q); MS (m/z) 319 (M+, 19),
228 (100), 200 (20), 91 (37), 86 (17). Anal. (C21H21NO2) C, H,
N.
8-(2,5-Bis-tr iflu or om eth ylben zyloxy)-4-m eth yl-2,3,5,6-
tetr a h yd r o-(1H)-ben zo[c]qu in olizin -3-on e (20). 20 was
prepared as reported for 14, starting from 12 (40 mg, 0.17
mmol) and using 2,5-bis-trifluoromethylbenzyl bromide (65 µL,
0.34 mmol). After chromatography (EtOAc/petroleum ether
1:1, Rf ) 0.21) pure 20 (60 mg, 75%) was obtained as a yellow
8-(2-Ch lor ob en zyloxy)-4-m et h yl-2,3,5,6,-t et r a h yd r o-
(1H)-ben zo[c]qu in olizin -3-on e (15). 15 was prepared as
reported for 14, starting from 12 (25 mg, 0.11 mmol) and using
2-chlorobenzyl bromide (45 mg, 0.22 mmol). After chromatog-
raphy (EtOAc/petroleum ether 3:2, Rf ) 0.29) pure 15 (25 mg,
65%) was obtained as a pale-yellow solid: mp 105-109 °C;
1H NMR (CDCl3) δ 7.51 (m, 1H), 7.37 (m, 1H), 7.30-7.24 (m,
2H), 6.86-6.80 (m, 3H), 5.14 (s, 2H), 3.89 (t, J ) 7.4 Hz, 2H),
2.73-2.60 (m, 6H), 1.81 (s, 3H); 13C NMR (CDCl3) δ 190.3 (s),
156.3 (s), 153.4 (s), 134.7 (s), 134.6 (s), 132.5 (s), 129.3 (d), 129.0
(d), 128.7 (d), 128.5 (s), 126.9 (d), 114.8 (d), 113.8 (d), 113.2
(d), 105.2 (s), 67.6 (t), 45.3 (t), 35.5 (t), 26.4 (t), 25.2 (t), 10.2
(q); MS (m/z) 353 (M+, 12), 228 (100), 200 (28), 125 (25), 83
(39), 71 (42), 57 (78). Anal. (C21H20ClNO2) C, H, N.
1
solid: mp 149-154 °C; H NMR (CDCl3) δ 8.06 (s, 1H), 7.82
(d, J ) 7.7 Hz, 1H), 7.68 (d, J ) 7.7 Hz, 1H), 6.90-6.81 (m,
3H), 5.24 (s, 2H), 3.88 (t, J ) 7.7 Hz, 2H), 2.75-2.60 (m, 6H),
1.80 (s, 3H); 13C NMR (CDCl3) δ 190.2 (s), 156.2 (s), 152.9 (s),
137.3 (s), 135.2 (s), 134.0 (s), 128.5 (2s, 2C), 126.6 (d), 125.5
(d), 125.0 (d), 115.0 (d), 113.9 (d), 113.3 (d), 105.4 (s), 66.1 (t),
45.3 (t), 35.5 (t), 26.4 (t), 25.2 (t), 10.2 (q) (2 CF3 nd); MS (m/z)
455 (M+, 12), 228 (100), 200 (19). Anal. (C23H19F3NO2) C, H,
N.
8-(P h en ylet h yn yl)-4-m et h yl-2,3,5,6,-t et r a h yd r o-(1H )-
ben zo[c]qu in olizin -3-on e (21). Under a N2 atmosphere 12
(101 mg, 0.35 mmol), PdCl2(PPh3)2 (12 mg, 0.017 mmol), and
CuI (7 mg, 0.034 mmol) were suspended in 2 mL of anhydrous
NEt3, and then phenylacetylene (45 µL, 0.41 mmol) was added.
The mixture was heated to reflux for 6 h. After this period,
the mixture was cooled to room temperature and the yellow
solid precipitated was separated from the solution, dissolved
in Et2O (20 mL), and washed with 3 × 10 mL of H2O. The
organic layer was dried over Na2SO4 and evaporated under
reduced pressure, obtaining a yellow residue. Chromatography
(CH2Cl2/MeOH 40:1, Rf ) 0.25) afforded 21 with 31% yield:
1H NMR (CDCl3) δ 7.52-7.30 (m, 7H), 6.90 (d, J ) 8.5 Hz,
1H), 3.92 (t, J ) 8.3 Hz, 2H), 2.78-2.62 (m, 6H), 1.81 (s, 3H);
13C NMR (CDCl3) δ 190.5 (s), 155.2 (s), 140.3 (s), 131.3 (d),
131.1 (d), 130.7 (d), 128.2 (d), 128.0 (d), 126.5 (s), 123.3 (s),
8-(3-Ch lor ob en zyloxy)-4-m et h yl-2,3,5,6,-t et r a h yd r o-
(1H)-ben zo[c]qu in olizin -3-on e (16). 16 was prepared as
reported for 14, starting from 12 (25 mg, 0.11 mmol) and using
3-chlorobenzyl bromide (45 mg, 0.22 mmol). After chromatog-
raphy (EtOAc/petroleum ether 3:2, Rf ) 0.25) pure 15 (12 mg,
32%) was obtained as a pale-yellow solid: mp 119-123 °C;
1H NMR (CDCl3) δ 7.43 (s, 1H), 7.27 (m, 1H), 6.85-6.79 (m,
3H), 5.02 (s, 2H), 3.89 (t, J ) 7.4 Hz, 2H), 2.74-2.61 (m, 6H),
1.82 (s, 3H); 13C NMR (CDCl3) δ 190.2 (s), 156.2 (s), 153.3 (s),
139.0 (s), 134.7 (s), 134.5 (s), 129.7 (d), 128.1 (s), 128.0 (d),
127.2 (d), 125.1 (d), 114.9 (d), 113.7 (d), 113.1 (d), 105.2 (s),
69.6 (t), 45.2 (t), 35.4 (t), 26.4 (t), 25.1 (t), 10.1 (q); MS (m/z)

3556 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 14
Occhiato et al.
116.1 (s), 112.7 (d), 106.8 (s), 89.1 (s), 88.9 (s), 45.0 (t), 35.5
(t), 26.1 (t), 24.7 (t), 10.2 (q); MS m/z 313 (M+, 100), 284 (14),
256 (15). Anal. (C22H19NO) C, H, N.
Na2CO3(aq) 2 M (0.7 mL) were added. The mixture was
vigorously stirred and heated to reflux for 24 h, then cooled
to room temperature. H2O (10 mL) was added, and the mixture
was extracted with 3 × 10 mL of CH2Cl2. The organic layer
was washed with H2O, dried over Na2SO4, and evaporated
under reduced pressure, affording the crude product, which
was chromatographed using EtOAc/light petroleum 1:1 as
eluant (Rf ) 0.27). The pure product was obtained with 33%
yield: mp 180-182 °C; 1H NMR (CDCl3) δ 7.50-7.23 (m, 7H),
7.02 (s, 2H), 6.92 (d, J ) 8.4 Hz, 1H), 3.93 (t, J ) 7.6 Hz, 2H),
2.82-2.62 (m, 6H), 1.82 (s, 3H); 13C NMR (CDCl3) δ 190.5 (s),
155.6 (s), 139.8 (s), 137.3 (s), 131.0 (s), 128.6 (d), 127.6 (d),
127.3 (d), 127.2 (d), 126.8 (s), 126.2 (d), 126.1 (d), 125.4 (d),
113.0 (d), 106.5 (s), 45.1 (t), 35.5 (t), 26.3 (t), 25.0 (t), 10.2 (q);
MS m/z 315 (M+, 100), 286 (36), 258 (38). Anal. (C22H21NO) C,
H, N.
8-Z-Styr yl-4-m eth yl-2,3,5,6,-tetr a h yd r o-(1H)-ben zo[c]-
qu in olizin -3-on e (22). In a Schlenk container 21 (20 mg, 0.06
mmol) was dissolved in pyridine (1 mL), then Pd/BaSO4 (Pd
10%, 25 mg) was added. The reaction vessel was filled with
H2, and the mixture was stirred at room temperature for 16
h. The mixture was then diluted with 15 mL of Et2O and
filtered over a Celite layer, and then the solution was washed
with 2 × 10 mL of H2O, dried over Na2SO4, and evaporated
under reduced pressure. The crude product was chromato-
graphed (EtOAc/petroleum ether 1:1, Rf ) 0.31), yielding pure
22 (17 mg, 89%) as a yellow solid: 1H NMR (CDCl3) δ 7.50-
6.95 (m, 7H), 6.76 (d, J ) 8.6 Hz, 1H), 6.50 (AB, 1H), 3.87 (t,
J ) 7.6 Hz, 2H), 2.69-2.59 (m, 6H), 1.79 (s, 3H); 13C NMR
(CDCl3) δ 190.5 (s), 155.6 (s), 139.9 (s), 137.4 (s), 131.1 (s), 128.6
(d), 127.7 (d), 127.4 (d), 127.3 (d), 126.9 (s), 126.3 (d), 126.1
(d), 125.5 (d), 113.0 (d), 106.4 (s), 45.2 (t), 35.6 (t), 26.4 (t),
25.0 (t), 10.2 (q); MS m/z 315 (M+, 45), 286 (67), 258 (41). Anal.
(C22H21NO) C, H, N.
8-P ir id -3-yl-4-m eth yl-2,3,5,6,-tetr a h yd r o-(1H)-ben zo[c]-
qu in olizin -3-on e (23). Under a N2 atmosphere 12 (57 mg,
0.20 mmol) was dissolved in THF (3 mL), and then PdCl2-
(PPh3)2 (14 mg, 0.02 mmol), the boronic acid (HO)2BPy (49 mg,
0.30 mmol), and Na2CO3(aq) 2 M (0.7 mL) were added. The
mixture was vigorously stirred and heated to 80 °C for 4 h,
then cooled to room temperature. H2O (6 mL) was added, and
the mixture was extracted with 3 × 10 mL of Et2O. The organic
layer was washed with H2O, dried over Na2SO4, and evapo-
rated under reduced pressure, affording the crude product,
which was chromatographed (CH2Cl2/MeOH 10:1, Rf ) 0.31)
to give pure 23 (45 mg, 77%) as a yellow solid: 1H NMR
(CDCl3) δ 8.80 (d, J ) 1.9 Hz, 1H), 8.53 (d, J ) 3.3 Hz, 1H),
7.83 (d, J ) 8.0 Hz, 1H), 7.44 (dd, J 1 ) 8.5 Hz, J 2 ) 1.9 Hz,
1H), 7.34 (d, J ) 2.6 Hz, 1H), 7.31 (d, J ) 5.2 Hz, 1H), 7.03 (d,
J ) 8.4 Hz, 1H), 3.95 (t, J ) 7.5 Hz, 2H), 2.86-2.63 (m, 6H),
1.82 (s, 3H); 13C NMR (CDCl3) δ 190.6 (s), 163.2 (s), 155.5 (s),
148.0 (d), 147.8 (d), 140.5 (s), 135.7 (s), 133.7 (d), 132.1 (s),
130.8 (s), 127.4 (d), 126.2 (d), 123.5 (d), 113.4 (d), 45.1 (t), 35.5
(t), 26.3 (t), 25.0 (t), 10.2 (q); MS m/z 290 (M+, 100), 261 (25),
247 (21), 233 (28), 84 (92). Anal. (C19H18N2O) C, H, N.
8-Vin yl-4-m eth yl-2,3,5,6,-tetr ah ydr o-(1H)-ben zo[c]qu in -
olizin -3-on e (27). Under a N2 atmosphere 12 (100 mg, 0.34
mmol), Pd(OAc)2 (4 mg, 0.017 mmol), and PPh3 (18 mg, 0.068
mmol) were dissolved in anhydrous NEt3, then tributylvinyl-
stannane (150 mL, 0.51 mmol) was added. The solution was
heated to 95 °C for 24 h. After this period, the solvent was
evaporated, the residue was dissolved in 3 mL of EtOAc, 5
mL of KF(aq) 1 M were added, and the mixture was vigorously
stirred for 1 h. The biphasic mixture was then filtered over
Celite, and the filtrating agent was washed with EtOAc. The
organic layer (25 mL) was separated from the aqueous layer,
washed with brine (2 × 15 mL), dried over Na2SO4, and
evaporated under reduced pressure, affording the crude prod-
uct, which was chromatographed using EtOAc/light petroleum
1:1 as eluant (Rf ) 0.27). The pure product was obtained with
1
43% yield. H NMR (CDCl3) δ 7.24 (d, J ) 8.5 Hz, 1H), 7.19
(s, 1H), 6.88 (d, J ) 8.5 Hz, 1H), 6.63 (dd, J 1 )17.6 Hz, J 2
)10.8 Hz, 1H), 5.64 (d, J ) 17.6 Hz, 1H), 5.15 (d, J ) 10.8 Hz,
1H), 3.91 (t, J ) 7.7 Hz, 2H), 2.78-2.60 (m, 6H), 1.81 (s, 3H);
13C NMR (CDCl3) δ 190.5 (s), 155.7 (s), 139.9 (s), 135.8 (d),
131.2 (s), 126.6 (s), 125.7 (d), 125.2 (d), 112.8 (t), 112.3 (d),
106.2 (s), 45.1 (t), 35.5 (t), 26.3 (t), 24.9 (t), 10.2 (q); MS m/z
239 (M+, 62), 238 (100), 210 (21), 182 (24). Anal. (C16H17NO)
C, H, N.
8-Eth yl-4-m eth yl-2,3,5,6,-tetr ah ydr o-(1H)-ben zo[c]qu in -
olizin -3-on e (28). Into an autoclave 8-vinyl-Q1-4M (15 mg,
0.063 mmol) was dissolved in C6H6 (1 mL), and then the
catalyst (PPh3)3RhCl (3 mg, 3.1 µmol) was added. The auto-
clave was filled with H2 (100 psi) and was heated to 40 °C for
6 h. After this period the solvent was evaporated and the
residue was dissolved in 10 mL of Et2O and washed with H2O
(2 × 10 mL), then dried over Na2SO4 and evaporated under
reduced pressure, affording the crude product, which was
chromatographed using EtOAc/light petroleum ether 2:1 as
eluant (Rf ) 0.33). The pure product was obtained with 75%
yield: 1H NMR (CDCl3) δ 7.04 (dd, J 1 ) 8.5 Hz, 1H), 6.96 (s,
1H), 6.86 (d, J ) 8.5 Hz,1H), 3.89 (t, J ) 7.7 Hz, 2H), 2.78-
2.52 (m, 8H), 1.80 (s, 3H), 1.20 (t, J ) 8 Hz, 3H); 13C NMR
(CDCl3) δ 190.4 (s), 156.3 (s), 138.0 (s), 137.4 (s), 127.2 (d),
126.7 (d), 126.5 (s), 112.7 (d), 105.5 (s), 45.1 (t), 35.5 (t), 28.0
(t), 26.4 (t), 24.9 (t), 15.7 (q), 10.1 (q); MS m/z 241 (M+, 74),
226 (64), 198 (35), 83 (29). Anal. (C16H19NO) C, H, N.
6-P h en yl-3,4-d ih yd r oqu in olin -2(1H)-on e (31). 31 was
prepared as reported for 3. Starting from 29 (5.00 g, 29.5
mmol), amide 30 was obtained in 98% yield: mp 177-179 °C;
1H NMR (CDCl3) δ 7.55 (ps, 5H), 7.52 (s, 1H), 7.43-7.30 (m,
4H), 3.88 (t, J ) 6.6 Hz, 2H), 2.81 (t, J ) 6.6 Hz, 2H); MS m/z
259 (M+, 15), 169 (100). The cyclization of 30 (2.00 g, 7.7 mmol)
to the corresponding lactam was carried out by heating at
180-190 °C for 6 h, obtaining 31 (961 mg, 56%) after
purification by chromatography (EtOAc/petroleum ether 3:7,
Rf ) 0.29): 1H NMR (CDCl3) δ 7.77 (s, 1H), 7.53-7.20 (m, 7H),
6.80 (d, J ) 8.5 Hz, 1H), 3.01 (t, J ) 7.6 Hz, 2H), 2.65 (t, J )
7.6 Hz, 2H).
8-F u r a n -2-yl-4-m eth yl-2,3,5,6,-tetr a h yd r o-(1H)-ben zo-
[c]qu in olizin -3-on e (24). 24 was prepared as reported for 23,
starting from 12 (88 mg, 0.30 mmol) and using (HO)2BFur (50
mg, 0.45 mmol). After chromatography (EtOAc/petroleum
ether 1:1, Rf ) 0.42) pure 24 (57 mg, 67%) was obtained as a
yellow solid: 1H NMR (CDCl3) δ 7.52 (dd, J 1 ) 8.8 Hz, J 2
)
1.8 Hz, 1H), 7.43 (ps, 2H), 6.94 (d, J ) 8.8 Hz, 1H), 6.54 (d, J
) 3.3 Hz, 1H), 6.44 (dd, J 1 ) 1.8 Hz, J 2 ) 1.5 Hz, 1H), 3.93 (t,
J ) 7.5 Hz, 2H), 2.82-2.62 (m, 6H), 1.82 (s, 3H); 13C NMR
(CDCl3) δ 190.5 (s), 155.5 (s), 153.5 (s), 141.6 (d), 139.5 (s),
126.8 (s), 124.6 (s), 123.1 (2C, 2d), 113.0 (d), 111.6 (d), 106.3
(s), 103.9 (d), 45.1 (t), 35.5 (t), 26.3 (t), 24.9 (t), 10.2 (q); MS
m/z 279 (M+, 100), 250 (39), 222 (55), 140 (29), 83 (31). Anal.
(C18H17NO2) C, H, N.
8-Th iop h en -2-yl-4-m eth yl-2,3,5,6,-tetr a h yd r o-(1H)-ben -
zo[c]qu in olizin -3-on e (25). 25 was prepared as reported for
23, starting from 12 (88 mg, 0.30 mmol) and using (HO)2BTfn
(58 mg, 0.45 mmol). After chromatography (EtOAc/petroleum
ether 3:2, Rf ) 0.45) pure 24 (61 mg, 68%) was obtained as a
yellow solid: 1H NMR (CDCl3) δ 7.46 (dd, J 1 ) 8.5 Hz, J 2
)
2.2 Hz, 1H), 7.44 (d, J ) 2.2 Hz, 1H), 7.21 (ps, 2H), 7.05 (dd,
J 1 ) 4.4 Hz, J 2)4.4 Hz, 1H), 6.93 (d, J ) 8.5 Hz, 1H), 3.93 (t,
J ) 7.7 Hz, 2H), 2.82-2.62 (m, 6H), 1.82 (s, 3H); 13C NMR
(CDCl3) δ 190.4 (s), 155.3 (s), 143.5 (s), 139.6 (s), 127.9 (s), 127.8
(d), 126.9 (s), 125.0 (d), 124.9 (d), 123.9 (d), 122.1 (d), 112.9
(d), 106.2 (s), 44.8 (t), 35.3 (t), 26.0 (t), 24.7 (t), 9.9 (q); MS m/z
295 (M+, 100), 266 (23), 252 (19), 238 (24), 149 (25), 57 (37).
Anal. (C18H17NOS) C, H, N.
8-E-Styr yl-4-m eth yl-2,3,5,6,-tetr a h yd r o-(1H)-ben zo[c]-
qu in olizin -3-on e (26). Under a N2 atmosphere 12 (60 mg,
0.20 mmol) was dissolved in THF (3 mL), and then PdCl2-
(PPh3)2 (14 mg, 0.02 mmol), the boronic ester (0.45 mmol), and
6-P h en yl-3,4-d ih yd r oqu in olin -2(1H)-th ion e (32). 32 was
prepared as reported for 5. Starting from 31 (230 mg, 1.03
mmol), 32 (243 mg) was obtained in 99% yield as an orange
solid after purification by chromatography (CH2Cl2/petroleum

Benzo[c]quinolizin-3-ones
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 14 3557
ether 4:3, Rf ) 0.26): 1H NMR (CDCl3) δ 9.40 (s, 1H), 7.53-
7.31 (m, 7H), 6.86 (d, J ) 8.5 Hz, 1H), 3.12 (t, J ) 6.6 Hz,
2H), 2.92 (t, J ) 6.6 Hz, 2H).
6-P h en yl-1-(3-oxop en t yl)-3,4-d ih yd r oq u in olin -2(1H )-
th ion e (33). 33 was prepared as reported for 6. Starting from
32 (243 mg, 1.02 mmol), pure 33 (250 mg, 76%) was obtained
as a yellow solid after chromatography (EtOAc/petroleum ether
1:8, Rf ) 0.23): 1H NMR (CDCl3) δ 7.56-7.29 (m, 7H), 7.17
(d, J ) 8.4 Hz, 1H), 4.78 (t, J ) 9.1 Hz, 2H), 3.18 (t, J ) 7.7
Hz, 2H), 3.02 (t, J ) 7.7 Hz, 2H), 2.83 (t, J ) 7.7 Hz, 2H), 2.49
(q, J ) 7.3 Hz, 2H), 1.07 (t, J ) 7.3 Hz, 3H).
δ 7.84 (dd, J 1 ) 8.5 Hz, J 2 ) 2.2 Hz, 1H), 7.72 (s, 1H), 6.92 (d,
J ) 8.5 Hz, 1H), 3.94 (t, J ) 7.3 Hz, 2H), 2.82-2.63 (m, 6H),
1.81 (s, 3H), 1.57 (s, 9H); 13C NMR (CDCl3) δ 193.5 (s), 168.5
(s), 155.8 (s), 142.4 (s), 129.5 (d), 128.8 (d), 126.0 (s), 123.0 (s),
112.1 (d), 107.5 (s), 63.5 (s), 45.1 (t), 35.5 (t), 28.3 (q), 26.1 (t),
24.8 (t), 10.2 (q); MS m/z 313 (M+, 20), 256 (100), 240 (20), 83
(72). Anal. (C19H23NO3) C, H, N.
8-Car boxyben zyl-4-m eth yl-2,3,5,6,-tetr ah ydr o-(1H)-ben -
zo[c]qu in olizin -3-on e (39). 39 was prepared as reported for
36, starting from 12 (50 mg, 0.17 mmol) and using benzyl
alcohol (1 mL, 10 mmol). After chromatography (EtOAc/
petroleum ether 1:1, Rf ) 0.29) pure 39 (42 mg, 71%) was
8-P h en yl-4-m et h yl-2,3,5,6,-t et r a h yd r o-(1H )-b en zo[c]-
qu in olizin -3-on e (34). 34 was prepared as reported for 7.
Starting from 33 (250 mg, 0.77 mmol), pure 34 (45 mg, 20%)
was obtained as a pale-yellow solid after chromatography
(EtOAc/petroleum ether 45:70, Rf ) 0.28): mp 161-165 °C;
1H NMR (CDCl3) δ 7.56-7.29 (m, 7H), 7.00 (d, J ) 8.8 Hz,
1H), 3.94 (t, J ) 8.1 Hz, 2H), 2.90-2.62 (m, 6H), 1.81 (s, 3H);
13C NMR (CDCl3) δ 190.5 (s), 166.7 (s), 155.9 (s), 140.3 (s),
135.3 (s), 132.7 (s), 128.8 (d), 127.0 (d), 126.6 (d), 126.4 (d),
126.2 (d), 113.2 (d), 106.2 (s), 45.1 (t), 35.5 (t), 26.3 (t), 25.0
(t), 10.1 (q); MS m/z 289 (M+, 100), 260 (47), 246 (36), 232 (51)
165 (29), 57 (30). Anal. (C20H19NO) C, H, N.
8-Ca r b oxym et h yl-4-m et h yl-2,3,5,6,-t et r a h yd r o-(1H )-
ben zo[c]qu in olizin -3-on e (35). Into an autoclave 12 (50 mg,
0.17 mmol), PdCl2 (4 mg, 0.023 mmol), PPh3 (12 mg, 0.046
mmol), NEt3 (30 µL, 0.22 mmol), and MeOH (1 mL, 25 mmol)
were mixed with 3 mL of benzene. The autoclave was filled
with CO (50 bar) and then heated to 115 °C for 24 h. After
this period the mixture was cooled to room temperature and
the solvent and the excess alcohol were evaporated under
reduced pressure, giving a yellowish residue. Chromatography
of the crude product (eluant EtOAc/light petroleum 1:1, Rf )
0.36) afforded pure 35 (37 mg) with 80% yield: mp 106-108
°C; 1H NMR (CDCl3) δ 7.90 (dd, J 1 ) 8.8 Hz, J 2 ) 2.0 Hz, 1H),
7.78 (d, J ) 2.0 Hz, 1H), 6.95 (d, J ) 8.8 Hz, 1H), 3.95 (t, J )
7.4 Hz, 2H), 3.87 (s, 3H), 2.85-2.63 (m, 6H), 1.81 (s, 3H); 13C
NMR (CDCl3) δ 190.6 (s), 166.5 (s), 154.6 (s), 144.1 (s), 130.3
(d), 128.9 (d), 122.9 (s), 112.2 (d), 107.7 (s), 51.9 (q), 45.1 (t),
35.5 (t), 26.0 (t), 24.8 (t), 10.2 (q); MS m/z 271 (M+, 75), 242
(40), 227 (38), 214 (19). Anal. (C16H17NO3) C, H, N.
8-Ca r boxyeth yl-4-m eth yl-2,3,5,6,-tetr a h yd r o-(1H)-ben -
zo[c]qu in olizin -3-on e (36). 36 was prepared as reported for
35, starting from 12 (50 mg, 0.17 mmol) and using EtOH (1
mL, 17 mmol). After chromatography (EtOAc/petroleum ether
1:1, Rf ) 0.34) pure 36 (30 mg, 60%) was obtained as a white
solid: mp 110-112 °C; 1H NMR (CDCl3) δ 7.90 (d, J ) 8.1
Hz, 1H), 7.79 (s, 1H), 6.95 (d, J ) 8.1 Hz, 1H), 4.34 (q, J ) 7.0
Hz, 2H), 3.95 (t, J ) 7.3 Hz, 2H), 2.79-2.63 (m, 6H), 1.81 (s,
3H), 1.37 (t, J ) 7.0 Hz, 3H); 13C NMR (CDCl3) δ 190.7 (s),
166.1 (s), 154.7 (s), 144.0 (s), 129.6 (d), 128.9 (d), 126.1 (s),
123.2 (s), 112.2 (d), 107.6 (s), 60.7 (t), 45.1 (t), 35.5 (t), 26.0 (t),
24.7 (t), 14.4 (q), 10.2 (q); MS m/z 285 (M+, 100), 256 (55), 240
(29), 77 (44). Anal. (C17H19NO3) C, H, N.
1
obtained as a white solid: mp 107-110 °C; H NMR (CDCl3)
δ 7.94 (dd, J 1 ) 8.8 Hz, J 2 ) 2.2 Hz, 1H), 7.81 (s, 1H), 7.44-
7.34 (m, 5H), 6.94 (d, J ) 8.8 Hz, 1H), 5.32 (s, 2H), 3.94 (t, J
) 7.4 Hz, 2H), 2.82-2.63 (m, 6H), 1.81 (s, 3H); 13C NMR
(CDCl3) δ 190.7 (s), 165.9 (s), 154.7 (s), 144.2 (s), 131.9 (s), 129.8
(d), 129.0 (d), 128.5 (d), 128.3 (d), 128.1 (d), 126.1 (s), 122.7
(s), 112.2 (d), 107.7 (s), 66.5 (t), 45.1 (t), 35.5 (t), 26.0 (t), 24.7
(t), 10.2 (q); MS m/z 347 (M+, 90), 277 (54), 240 (32), 183 (27),
91 (100), 77 (38). Anal. (C22H21NO3) C, H, N.
8-Ca r b oxyp h en yl-4-m et h yl-2,3,5,6,-t et r a h yd r o-(1H )-
ben zo[c]qu in olizin -3-on e (40). Into an autoclave 12 (50 mg,
0.17 mmol), PdCl2 (8 mg, 0.046 mmol), PPh3 (24 mg, 0.092
mmol), NEt3 (60 µL, 0.44 mmol), and phenol (500 mg, 5.3
mmol) were mixed with 3.5 mL of benzene. The autoclave was
filled with CO (50 bar) and then heated to 120 °C for 24 h.
After this period the mixture was cooled to room temperature
and the solvent was evaporated under reduced pressure. The
residue was dissolved in CH2Cl2 (10 mL) and washed with H2O
(2 × 10 mL) and KOH 5% (2 × 10 mL). The organic layer was
dried over Na2SO4 and after evaporation of the solvent the
obtained crude product was chromatographed (EtOAc/petro-
leum ether 1:1, Rf ) 0.25), affording pure 40 (50 mg, 88%):
1
mp 183-184 °C; H NMR (CDCl3) δ 8.06 (dd, J 1 ) 8.8 Hz, J 2
) 2.2 Hz, 1H), 7.95 (s, 1H), 7.45-7.17 (m, 5H), 7.02 (d, J )
8.8 Hz, 1H), 3.99 (t, J ) 7.5 Hz, 2H), 2.87-2.66 (m, 6H), 1.83
(s, 3H); 13C NMR (CDCl3) δ 190.7 (s), 164.6 (s), 154.4 (s), 150.9
(s), 144.7 (s), 130.4 (d), 129.5 (d), 129.3 (d), 126.3 (s), 125.7
(d), 122.0 (s), 121.6 (d), 112.4 (d), 108.0 (s), 45.2 (t), 35.5 (t),
26.0 (t), 24.8 (t), 10.3 (q); MS m/z 333 (M+, 36), 240 (100), 211
(51). Anal. (C21H19NO3) C, H, N.
4-Meth yl-2,3,5,6,-tetr a h yd r o-(1H)-ben zo[c]qu in olizin -
3-on e-8-ca r boxylic Acid p-(ter t-Bu tyl)p h en yl Ester (41).
41 was prepared as reported for 40, starting from 12 (50 mg,
0.17 mmol) and using p-tert-butylphenol (510 mg, 3.4 mmol).
After chromatography (EtOAc/petroleum ether 1:1, Rf ) 0.29)
pure 41 (46 mg, 70%) was obtained as a white solid: 1H NMR
(CDCl3) δ 8.07 (d, J ) 8.8 Hz, 1H), 7.94 (s, 1H), 7.42 (AB, 2H),
7.10 (AB, 2H), 7.01 (d, J ) 8.8 Hz, 1H), 3.99 (t, J ) 7.7 Hz,
2H), 2.87-2.66 (m, 6H), 1.83 (s, 3H), 1.32 (s, 9H); 13C NMR
(CDCl3) δ 190.7 (s), 164.7 (s), 154.5 (s), 148.4 (s), 144.5 (s), 131.8
(s), 130.3 (d), 129.4 (d), 126.2 (d), 122.1 (s), 120.8 (d), 112.3
(d), 107.9 (s), 97.8 (s), 45.1 (t), 35.4 (t), 34.4 (s), 31.4 (q), 25.9
(t), 24.7 (t), 10.2 (q); MS m/z 389 (M+, 4), 240 (100), 212 (15).
Anal. (C25H27NO3) C, H, N.
4-Meth yl-2,3,5,6,-tetr a h yd r o-(1H)-ben zo[c]qu in olizin -
3-on e-8-ca r boxylic Acid p-(Ca r boxym eth yl)p h en yl Ester
(42). 42 was prepared as reported for 40, starting from 12 (50
mg, 0.17 mmol) and using p-caboxymethylphenol (530 mg, 3.5
mmol). After chromatography (EtOAc/petroleum ether 1:1, Rf
) 0.37) pure 42 (38 mg, 58%) was obtained as a white solid:
1H NMR (CDCl3) δ 8.10 (AB, 2H), 8.03 (d, J ) 8.8 Hz, 1H),
7.94 (s, 1H), 7.28 (AB, 2H), 7.02 (d, J ) 8.8 Hz, 1H), 3.99 (t, J
) 7.7 Hz, 2H), 3.92 (s, 3H), 2.84-2.67 (m, 6H), 1.83 (s, 3H);
13C NMR (CDCl3) δ 190.8 (s), 166.2 (s), 164.1 (s), 154.5 (s),
154.3 (s), 144.9 (s), 131.1 (d), 130.5 (d), 129.6 (d), 127.5 (s),
126.4 (s), 121.7 (d), 121.4 (s), 112.4 (d), 108.2 (s), 52.2 (q), 45.2
(t), 35.5 (t), 25.9 (t), 24.8 (t), 10.3 (q); MS m/z 391 (M+, 4), 240
(67), 211 (7), 150 (8), 105 (100), 83 (37); MS m/z 391 (M+, 11),
240 (100), 212 (44). Anal. (C23H21NO5) C, H, N.
8-Ca r boxyisop r op yl-4-m eth yl-2,3,5,6,-tetr a h yd r o-(1H)-
ben zo[c]qu in olizin -3-on e (37). was 37 prepared as reported
for 36, starting from 12 (50 mg, 0.17 mmol) and using
2-propanol (1 mL, 13 mmol). After chromatography (EtOAc/
petroleum ether 1:1, Rf ) 0.37) pure 15 (30 mg, 59%) was
1
obtained as a white solid: mp 129-131 °C; H NMR (CDCl3)
δ 7.90 (dd, J 1 ) 8.8 Hz, J 2 ) 1.8 Hz, 1H), 7.77 (d, J ) 1.8 Hz,
1H), 6.94 (d, J ) 8.8 Hz, 1H), 5.21 (hp, J ) 6.2 Hz, 1H), 3.95
(t, J ) 7.7 Hz, 2H), 2.85-2.63 (m, 6H), 1.81 (s, 3H), 1.34 (d, J
) 6.2 Hz, 6H); 13C NMR (CDCl3) δ 190.7 (s), 165.5 (s), 154.7
(s), 143.9 (s), 129.6 (d), 128.8 (d), 126.0 (s), 123.7 (s), 112.1 (d),
107.6 (s), 68.0 (d), 45.1 (t), 35.5 (t), 26.0 (t), 24.8 (t), 22.0 (q),
10.2 (q); MS m/z 299 (M+, 57), 256 (100), 240 (20). Anal.
(C18H21NO3) C, H, N.
8-Ca r boxy-ter t-bu tyl-4-m eth yl-2,3,5,6,-tetr a h yd r o-(1H)-
ben zo[c]qu in olizin -3-on e (38). 38 was prepared as reported
for 36, starting from 12 (50 mg, 0.17 mmol) and using tert-
butyl alcohol (1 mL, 10 mmol). After chromatography (EtOAc/
petroleum ether 1:1, Rf ) 0.29) pure 38 (6 mg, 11%) was
4-Meth yl-2,3,5,6,-tetr a h yd r o-(1H)-ben zo[c]qu in olizin -
3-on e-8-ca r boxylic Acid m -(Ca r boxym eth yl)p h en yl Ester
(43). 43 was prepared as reported for 40, starting from 12 (50
1
obtained as a white solid: mp 161-163 °C; H NMR (CDCl3)

3558 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 14
Occhiato et al.
mg, 0.17 mmol) and using m-caboxymethylphenol (530 mg, 3.5
mmol). After chromatography (CH2Cl2/MeOH 40:1, Rf ) 0.32)
pure 43 (47 mg, 71%) was obtained as a white solid: 1H NMR
(CDCl3) δ 8.06 (dd, J 1 ) 8.8 Hz, J 2 ) 2.2 Hz, 1H), 7.95-7.86
(m, 3H), 7.52-7.38 (m, 2H), 7.02 (d, J ) 8.8 Hz, 1H), 3.99 (t,
J ) 7.7 Hz, 2H), 3.90 (s, 3H), 2.91-2.67 (m, 6H), 1.83 (s, 3H);
13C NMR (CDCl3) δ 190.7 (s), 166.0 (s), 164.3 (s), 154.3 (s),
150.7 (s), 144.8 (s), 131.5 (s), 130.4 (d), 129.5 (d), 129.3 (d),
126.8 (d), 126.3 (2C, d + s), 122.8 (d), 121.4 (s), 112.4 (d), 108.1
(s), 52.2 (q), 45.1 (t), 35.4 (t), 25.9 (t), 24.7 (t), 10.2 (q); MS m/z
391 (M+, 8), 240 (100), 211 (17). Anal. (C23H21NO5) C, H, N.
and correlation functional B3LYP29 and with the 6-31G* basis
set. In detail, the chosen conformations for the CoMFA
analysis were first geometrically optimized until a nuclear
convergence of 3 × 10-4 on the gradient was reached, and then
the ESP was calculated. Finally, by means of the RESP
procedure, the atomic partial charges were obtained fitting
them on the ESP. The dipole moment was directly obtained
as a physical observable from the DFT calculations. All
quantum chemical computations were carried out using the
Gaussian 98 suite of program (Gaussian 98, Gaussian, Inc.,
Pittsburgh, PA, 1998).
4-Meth yl-2,3,5,6,-tetr a h yd r o-(1H)-ben zo[c]qu in olizin -
3-on e-8-ca r boxylic Acid p-(Meth yl)p h en yl Ester (44). 44
was prepared as reported for 40, starting from 12 (50 mg, 0.17
mmol) and using p-methylphenol (355 µL, 3.4 mmol). After
chromatography (EtOAc/petroleum ether 1:1, Rf ) 0.40) pure
44 (25 mg, 46%) was obtained as a white solid: 1H NMR
(CDCl3) δ 8.05 (d, J ) 8.4 Hz, 1H), 7.93 (s, 1H), 7.20 (AB, 2H),
7.05 (AB, 2H), 7.01 (d, J ) 8.4 Hz, 1H), 3.98 (t, J ) 7.4 Hz,
2H), 2.87-2.66 (m, 6H), 2.35 (s, 3H), 1.83 (s, 3H); 13C NMR
(CDCl3) δ 190.8 (s), 164.8 (s), 154.5 (s), 148.6 (s), 144.6 (s), 135.4
(s), 130.4 (d), 129.9 (d), 129.5 (d), 126.3 (s), 122.1 (s), 121.3
(d), 112.3 (d), 108.0 (s), 45.2 (t), 35.5 (t), 26.0 (t), 24.8 (t), 20.9
(q), 10.3 (q); MS m/z 347 (M+, 8), 240 (100), 211 (26), 154 (7),
78 (9). Anal. (C22H21NO3) C, H, N.
3.4. 3D QSAR th r ou gh CoMF A. A 3D QSAR model was
obtained by means of the comparative molecular field analysis
(CoMFA).38 The most critical step in the CoMFA procedure is
the alignment of the molecules in Cartesian space. In the
present work, the 5RR-1 inhibitors were aligned by super-
imposing atom by atom the benzo[c]quinolizin-3-one skeleton
of the set of compounds. The molecules bearing a flexible group
at position 8 of the benzo[c]quinolizin-3-one moiety (14-20,
35-39, 40-42, and 44) were aligned by using the Z and E
phenylvinyl derivatives (22, 26) as templates. To this aim, both
diastereoisomers were synthesized, and because they present
quite different inhibitory potencies (see Table 1), the inhibitors
with high biological activity (pIC50 > 6.5; 14-19, 35, 36, 39,
and 42) were aligned onto the Z-22, and those with low potency
(pIC50 < 6.5; 20, 37, 38, 40, 41, and 44) were aligned onto the
E isomer 26. A CoMFA table was built containing the biological
activities of the set of 5RR-1 inhibitors (the dependent vari-
ables, pIC50obsd in Table 1) and the values of steric and
electrostatic fields at discrete points of the Cartesian space
surrounding the molecules (the independent variables). The
fields were generated by using an sp3 carbon atom with a
formal charge of +1 as a probe. The region was generated
automatically around the molecules by fixing a grid spacing
of 2 Å. Moreover, two additional independent variables were
taken into account to compute the 3D QSAR equation, i.e., the
dipole moment calculated at the DFT level and the log P value
estimated with the ClogP program39 (Table 1). The statistical
analyses were carried out by applying the PLS procedure to
the appropriate variables and using the standard scaling
method (COMFA_STD). Furthermore, to reduce the number
of independent variables, an energy cutoff value of 30 kcal/
mol was selected for both electrostatic and steric fields. The
minimum σ for further filtering of the independent variables
was set to 2.0 kcal/mol. Cross-validated PLS runs were carried
out to establish the optimal number of components (the latent
variables) to be used in the final fitting models. The number
of cross-validated groups was always equal to the number of
compounds (leave-one-out procedure), and the optimal number
of latent variables was chosen by considering the lowest
standard error of prediction (scross). All of the independent vari-
ables were checked against each other for possible collinearity;
no significant cross-correlations were found. In particular,
collinearity was not found between the dipole moment and the
electrostatic field or between log P and the steric field. Finally,
scrambling analyses were performed to rule out the possibility
of chance correlations.
2. Biology. All compounds were tested according to the
procedure already reported.17a
3. Molecu la r Mod elin g. 3.1. Con str u ction of th e Mod -
els. The molecular models were built by properly modifying
the 19-nor-10-aza-androstenedione skeleton retrieved from the
Cambridge Structural Database (CSD code NOVDOR21) and
by adding the proper fragments from the standard library of
SYBYL (Tripos Inc., St. Louis, MO). The models were mini-
mized first by using steepest descent and then conju-
gate gradient until a convergence of 0.05 kcal mol-1 Å-1 on
the gradient was reached. When needed, conformational
searches were carried out, and the minimum energy con-
formers were optimized by means of the semiempirical Hamil-
tonian PM331 as implemented in the SYBYL graphic interface
to MOPAC (keyword PRECISE). The geometry of the confor-
mations chosen for the CoMFA analysis was first optimized
at the HF/6-31G* level of theory and eventually at DFT/
B3LYP/6-31G* to properly account for the electron correlation
effects.
3.2. Con for m a tion a l Sea r ch . For the benzo[c]quinolizin-
3-one skeleton, the 1R,2â,6R conformation was kept fixed
because it was previously determined to be a minimum energy
conformation and to better fit onto the well-known 5RR-1
selective inhibitor LY191704 (Figure 1).17a The conformational
analyses were therefore carried out to sample the rotatable
bonds of the substituent at position 8 of the benzo[c]quinolizin-
3-one skeleton. Since some inhibitors bear a fairly flexible
substituent in that position, the conformational space was
sampled by means of Monte Carlo analyses32 using the MMFF
force field33 and the generalized Born/surface area (GB/SA)
continuum solvation model34 for water as implemented in
MacroModel software.35 In a Monte Carlo study, the phase
space of a molecule is sampled by randomly changing dihedral
angle rotations or atom positions. Then the trial conformation
is accepted if its energy has decreased from the previous one.
If the energy is higher, various criteria can be applied to accept
or reject the Monte Carlo trial. In the present simulations,
the number of Monte Carlo steps was set equal to 7000 and
the trial conformation was accepted if the energy was lower
than that of the previous conformation or if its energy was
within an energy window of 100 kJ /mol. Then the conforma-
tions were classified by means of a cluster analysis36 using
geometrical parameters as filtering screens.
Ack n ow led gm en t. This work was supported by
grants from MIUR and Serono International (Geneva,
Switzerland). The quantum chemical DFT calcula-
tions were carried out on an SGI Origin 3800 ma-
chine, taking advantage of a grant provided by
the supercomputer center CINECA, Bologna, Italy.
Mr. Sandro Papaleo, Mrs. Brunella Innocenti, and
Mr. Claudio Piazza are acknowledged for their technical
assistance.
3.3. Electr osta tic Com p u ta tion s. In a CoMFA model, the
electrostatic contribution to the final 3D QSAR equation
strongly depends on the quality of the point charges. For this
reason, we thought to compute the inhibitor charges by means
of the RESP procedure,37 i.e., by fitting them to an electrostatic
potential (ESP) calculated at the DFT level using the exchange
Su p p or tin g In for m a tion Ava ila ble: Elemental analysis
data. This material is available free of charge via the Internet
at http://pubs.acs.org.

Benzo[c]quinolizin-3-ones
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 14 3559
(18) Thiboutot, D.; Harris, G.; Iles, V.; Cimis, G.; Gilliland, K.; Hagari,
Refer en ces
S.; Activity of the Type
Differences in Isolated Sebaceous Glands and Whole Skin. J .
Invest. Dermatol. 1995, 105, 209-214.
1 5R-Reductase Exhibits Regional
(1) Geller, J . Benign Prostatic Hyperplasia: Pathogenesis and
Medical Therapy. J . Am. Geriatr. Soc. 1991, 39, 1208-1266.
(2) (a) Gormley, G. J . Role of 5R-Reductase Inhibitors in the
Treatment of Advanced Prostatic Carcinoma. Urol. Clin. North
Am. 1991, 18, 93-98. (b) Brawley, O. W.; Ford, G. L.; Thompson,
I.; Perlman, J . A.; Kramer, B. S. 5R-Reductase Inhibition and
Prostate Cancer Prevention. Cancer Epidemiol., Biomarkers
Prev. 1994, 3, 177-182.
(19) (a) Kenny, B.; Ballard, S.; Blagg, J .; Fox, D. Pharmacological
Options in the Treatment of Benign Prostatic Hyperplasia. J .
Med. Chem. 1997, 40, 1293-1314. (b) Harris, G. S.; Kozarich,
J . W. Steroid 5R-Reductase Inhibitors in Androgen-Dependent
Disorders. Curr. Opin. Chem. Biol. 1997, 1, 254-259.
(20) Guarna, A.; Belle, C.; Machetti, F.; Occhiato, E. G.; Payne, H.
P.; Cassiani, C.; Comerci, A.; Danza, G.; De Bellis, A.; Dini, S.;
Marrucci, A.; Serio, M. 19-Nor-10-azasteroids: A Novel Class
of Inhibitors for Human Steroid 5R-Reductases 1 and 2. J . Med.
Chem. 1997, 40, 1112-1129.
(21) Guarna, A.; Occhiato, E. G.; Machetti, F.; Marrucci, A.; Danza,
G.; Serio, M.; Paoli, P. 19-Nor-10-azasteroids, a New Class of
Steroid 5R-Reductase Inhibitors. 2. X-ray Structure, Molecular
Modeling, Conformational Analysis of 19-Nor-10-azasteroids and
Comparison with 4-Azasteroids and 6-Azasteroids. J . Med.
Chem. 1997, 40, 3466-3477.
(22) (a) Cavalli, A., Ed. Density functional theory. Quant. Struct.-
Act. Relat. 2002, 21. (b) Carloni, P.; Alber, F. Quantum Medicinal
Chemistry. In Methods and Principles in Medicinal Chemistry;
Mannhold, R., Kubinyi, H., Folkers, G., Eds.; Wiley-VCH:
Weinheim, Germany, 2003.
(23) Guarna, A.; Lombardi, E.; Machetti, F.; Occhiato, E. G.; Scarpi,
D. Modification of the aza-Robinson Annulation for the Synthe-
sis of 4-Methyl-benzo[c]quinolizin-3-ones, Potent Inhibitors of
Steroid 5R-Reductase 1. J . Org. Chem. 2000, 65, 8093-8095
(24) All new compounds were also tested with human recombinant
5RR-2 isozyme expressed in CHO 1829, some of them displaying
activity toward this isozyme.
(25) Thigpen, A. E.; Cala, K. M.; Russel, D. W. Characterization of
Chinese Hamster Ovary Cell Lines Expressing Human Steroid
5R-Reductase Isozymes. J . Biol. Chem. 1993, 268, 17404-17412.
(26) Program kindly furnished by the following: De Lean, A.;
Munson, P. J .; Guardabasso, V.; Rodbard, D.; Laboratory of
Theoretical and Physical Biology, National Institute of Child
Health and Human Development, National Institutes of Health,
Bethesda, MD 20892.
(27) (a) Tian, G.; Stuart, J . D.; Moss, M. L.; Domanico, P. L.; Bramson,
H. N.; Patel, I. R.; Kadwell, S. H.; Overton, L. K.; Kost, T. A.;
Mook, R. A., J r.; Frye, S. V.; Batchelor, W. K.; Wiseman, J . S.
17â-(N-tert-butylcarbamoyl)-4-aza-5R-androstan-1-en-3-one is an
active site-directed slow time-dependent inhibitor of human
steroid 5R-reductase 1. Biochemistry 1994, 33, 2291-2296. (b)
Bull, H. G.; Garcia-Calvo, M.; Anderson, S.; Baginsky, W. S.;
Chan, H. K.; Ellsworth, D. E.; Miller, R. R.; Stearns, R. A.;
Bakshi, R. K.; Rasmusson, G. H.; Tolman, R. L.; Myers, R. W.;
Kozarich, J . W.; Harris, G. S. Mechanism-based inhibition of
human steroid 5R-reductase by finasteride: enzyme-catalyzed
formation of NADP-dihydrofinasteride, a potent bisubstrate
analog inhibitor. J . Am. Chem. Soc. 1996, 118, 2359-2365.
(28) Guarna, A.; Occhiato, E. G.; Machetti, F.; Trabocchi, A.; Scarpi,
D.; Danza, G.; Mancina, R.; Comerci, A.; Serio, M. Effect of
C-Ring Modifications in Benzo[c]quinolizin-3-ones, New Selective
Inhibitors of Human 5R-Reductase 1. Bioorg. Med. Chem. 2001,
9, 1385-1393.
(29) Becke, A. D. A new mixing of Hartree-Fock and local density-
functional theories. J . Chem. Phys. 1993, 98, 1372-1377.
(30) Sulpizi, M.; Schelling, P.; Folkers, G.; Carloni, P.; Scapozza, L.
The rational of catalytic activity of herpes simplex virus thymi-
dine kinase. a combined biochemical and quantum chemical
study. J . Biol. Chem. 2001, 276, 21692-21697.
(31) Stewart, J . P. P. Optimization of parameters for semiempirical
methods I. Method. J . Comput. Chem. 1989, 10, 209-220.
(32) Chang, G.; Guida, W. C.; Still, W. C. An internal coordinate
Monte Carlo method for searching conformational space. J . Am.
Chem. Soc. 1989, 111, 4379-4386.
(3) Dallob, A. L.; Sadick, N. S.; Unger, W.; Lipert, S.; Geissler, L.
A.; Gregoire, S. L.; Nguyen, H. H.; Moore, E. C.; Tanaka, W. K.
The Effect of Finasteride, a 5R-Reductase Inhibitor, on Scalp
Skin Testosterone and Dihydrotestosterone Concentrations in
Patients with Male Pattern Baldness. J . Clin. Endocrinol.
Metab. 1994, 79, 703-706.
(4) Smith, L. S.; Tegler, J . J . Advances in Dermatology. Annu. Rep.
Med. Chem. 1989, 24, 177-186.
(5) (a) Price, V. H. Treatment of Hair Loss. N. Engl. J . Med. 1999,
341, 964-973 and references therein. (b) Gadwood, R. C.;
Fiedler, V. C. Pathogenesis and Treatment of Alopecia. Annu.
Rep. Med. Chem. 1989, 24, 187-196.
(6) Brooks, J . R. Treatment of Hirsutism with 5R-Reductase Inhibi-
tors Clin. Endocrinol. Metab. 1986, 15, 391-405.
(7) Metcalf, B. W.; Levy, M. A.; Holt, D. A. Inhibitors of Steroid 5R-
Reductase in Benign Prostatic Hyperplasia, Male Pattern Bald-
ness and Acne. Trends Pharmacol. Sci. 1989, 10, 491-495.
(8) (a) Holt, D. A.; Levy, M. A.; Metcalf, B. W. Inhibition of Steroid
5R-Reductase. In Advances in Medicinal Chemistry; Maryanoff,
B. E., Maryanoff, C. A., Eds.; J AI Press Inc: Greenwich, CT,
1993; Vol. 2, pp 1-29. (b) Abell, A. D.; Henderson, B. R. Steroidal
and Non-Steroidal Inhibitors of Steroid 5R-Reductase. Curr.
Med. Chem. 1995, 2, 583-597. (c) Li, X.; Chen, C.; Singh, S.;
Labrie, F. The Enzyme and Inhibitors of 4-Ene-3-oxosteroid 5R-
Oxidoreductase. Steroids 1995, 60, 430-441. (d) Frye, S. V.
Inhibitors of 5R-Reductase. Curr. Pharm. Des. 1996, 2, 59-84.
(9) (a) Russell, D. W.; Wilson, J . D. Steroid 5R-Reductases: Two
Genes/Two Enzymes. Annu. Rev. Biochem. 1994, 63, 25-61. (b)
Wilson, J . D.; Griffin, J . E.; Russell, D. W. Steroid 5R-Reductase
2 Deficiency. Endocr. Rev. 1993, 14, 577-593.
(10) Rasmusson, G. H. Chemical Control of Androgen Action. Annu.
Rep. Med. Chem. 1986, 22, 179-188.
(11) (a) Imperato-McGinley, J .; Guerrero, L.; Gautier, T.; Peterson,
R. E. Steroid 5R-reductase deficiency in man: an inherited form
of male pseudohermaphroditism. Science 1974, 186, 1213. (b)
J enkins, E. P.; Andersson, S.; Imperato-McGinley, J .; Wilson,
J . D.; Russell, D. W. Genetic and pharmacological evidence for
more than one human steroid 5 alpha-reductase J . Clin. Invest.
1992, 89, 293.
(12) J ones, C. D.; Audia, J . E.; Lawhorn, D. E.; McQuaid, L. A.;
Neubauer, B. L.; Pike, A. J .; Pennington, P. A.; Stamm, N. A.;
Toomey, R. E.; Hirsch, K. R. Nonsteroidal Inhibitors of Human
Type I Steroid 5R-Reductase. J . Med. Chem. 1993, 36, 421-
423
(13) Wikel, J . H.; Bemis, K. G.; Audia, J . E.; McQuaid, L. A.; J ones,
C. D.; Pennington, P. A.; Lawhorn, D. E.; Hirsch, K. R.; Stamm,
N. B. QSAR Study of Benzoquinolinones as Inhibitors of Human
Type 5R-Reductase. Bioorg. Med. Chem. Lett. 1993, 3, 1157-
1162.
(14) Abell, A. D.; Erhard, K. F.; Yen, H.-K.; Yamashita, D. S.; Brandt,
M.; Mohammed, H.; Levy, M. A.; Holt, D. A. Preparative Chiral
HPLC Separation of all Possible Stereoisomers of LY191704 and
LY266111 and Their in Vitro Inhibition of Human Types 1 and
2 Steroid 5R-Reductase. Bioorg. Med. Chem. Lett. 1994, 4, 1365.
(15) Smith, E. C. R.; McQuaid, L. A.; Goode, R. L.; McNulty, A. M.;
Neubauer, B. L.; Rocco, V. P.; Audia, J . E. Synthesis and 5R-
Reductase Inhibitory Activity of 8-Substituted Benzo[f]quinoli-
nones Derived from Palladium Mediated Coupling Reactions.
Bioorg. Med. Chem. Lett. 1998, 8, 395-398.
(16) Abell, A. D.; Brandt, M.; Levy, M. A.; Holt, D. A. A Comparison
of Steroidal and Nonsteroidal Inhibitors of Human Steroid 5a-
Reductase: New Tricyclic Aryl Acid Inhibitors of the Type-1
Isozyme. Bioorg. Med. Chem. Lett. 1996, 6, 481-484.
(33) Halgren, T. A. Merck molecular force field. I. Basis, form, scope,
parameterization, and performance of MMFF94. J . Comput.
Chem. 1996, 17, 490-519.
(34) Still, W. C.; Tempzyk, A.; Hawley, R.; Hendrickson, T. Semi-
analytical treatment of solvation for molecular mechanics and
dynamics. J . Am. Chem. Soc. 1990, 112, 6127-6129.
(35) Mohamadi, F.; Richards, N. G. J .; Guida, W. C.; Liskamp, R. M.
J .; Lipton, M. A.; Caulfield, C. E.; Chang, G.; Hendrickson, T.
F.; Still, W. C. MacroModelsan integrated software system for
modeling organic and bioorganic molecules using molecular
mechanics. J . Comput. Chem. 1990, 1, 440-467.
(36) Shenkin, P. S.; McDonald, D. Q. Cluster analysis of molecular
conformations. J . Comput. Chem. 1994, 15, 899-916.
(17) (a) Guarna, A.; Machetti, F.; Occhiato, E. G.; Scarpi, D.; Comerci,
A.; Danza, G.; Mancina, R.; Serio, M.; Hardy, K. Benzo[c]-
quinolizin-3-ones: A Novel Class of Potent and Selective Non-
steroidal Inhibitors of Human Steroid 5R-Reductase 1. J . Med.
Chem. 2000, 43, 3718-3735. (b) Guarna, A.; Occhiato, E. G.;
Scarpi, D.; Tsai, R.; Danza, G.; Comerci, A.; Mancina, R.; Serio,
M. Synthesis of Benzo[c]quinolizin-3-ones: Selective Non-
Steroidal Inhibitors of Steroid 5R-Reductase 1. Bioorg. Med.
Chem. Lett. 1998, 8, 2871-2876. (c) Guarna, A.; Occhiato, E.
G.; Scarpi, D.; Zorn, C.; Danza, G.; Comerci, A.; Mancina, R.;
Serio, M. Synthesis of 8-Chloro-benzo[c]quinolizin-3-ones as
Potent and Selective Inhibitors of Human Steroid 5R-Reductase
1. Bioorg. Med. Chem. Lett. 2000, 10, 353.
(37) Bayly, C. I.; Cieplak, P.; Cornell, W. D.; Kollman, P. A. A well-
behaved electrostatic potential based method using charge
restraints for determining atom-centered charges: the RESP
model. J . Phys. Chem. 1993, 97, 10269-10280.

3560 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 14
Occhiato et al.
(38) Cramer, R. D.; 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.
each observed activity value for each molecule of the test set
from the mean of the observed activity values of the training
set (pIC50obsd values of Table 1) and where PRESS is the sum of
the squared deviations between predicted and observed values
(∆ of Table 3).
(39) ClogP, version 4.3; BioByte Corp.: Claremont, CA.
(40) The r2
was evaluated by solving the equation r2
) (SD -
pred
pred
PRESS)/SD, where SD is the sum of the squared deviations of
J M031131O