Optimization of substituted N-3-benzylimidazoquinazolinone sulfonamides as potent and selective PDE5 inhibitors

Article abstract of DOI:10.1021/jm000336j

A previous report from these laboratories identified the N-3-benzylimidazoquinazolinone nucleus as a more selective PDE5 inhibitor template compared to the pyrazolopyrimidine of sildenafil. This paper describes in detail the structure-activity relationshi

Full text of DOI:10.1021/jm000336j

J . Med. Chem. 2000, 43, 5037-5043
5037
Op tim iza tion of Su bstitu ted N-3-Ben zylim id a zoqu in a zolin on e Su lfon a m id es a s
P oten t a n d Selective P DE5 In h ibitor s
David P. Rotella,*,† Zhong Sun,† Yeheng Zhu,† J ohn Krupinski,# Ronald Pongrac,# Laurie Seliger,#
Diane Normandin,# and J ohn E. Macor†
Departments of Discovery Chemistry and Metabolic and Cardiovascular Drug Discovery, Bristol-Myers Squibb Pharmaceutical
Research Institute, P.O. Box 5400, Princeton, New J ersey 08543-5400
Received August 4, 2000
A previous report from these laboratories identified the N-3-benzylimidazoquinazolinone nucleus
as a more selective PDE5 inhibitor template compared to the pyrazolopyrimidine of sildenafil.
This paper describes in detail the structure-activity relationships of a set of sulfonamide
analogues, several of which are both more potent and more selective PDE5 inhibitors in vitro
than sildenafil. The synthesis, in vitro enzyme activity and selectivity, and in vitro functional
and preclinical pharmacokinetic assessment of molecules in this series are described.
In tr od u ction
We recently reported the discovery of 2, which was
shown to be as efficacious as 1 in a functional assay of
PDE5 activity (potentiation of relaxation of rabbit
corpus cavernosum tissue). This compound was derived
from the benzylimidazoquinazolinone sulfonamide 7
(Chart 2), which itself was developed following targeted
screening.5 The improved PDE5 selectivity of 2 and 7
was associated with the N-3-benzyl substituent. This
paper describes structure-activity relationships (SAR)
of the N-3-benzyl moiety, as well as SAR at the sul-
fonamide locus, leading to a group of compounds that
are equal to 2 in terms of in vitro potency and selectiv-
ity. Selected compounds in this series demonstrate
functional activity in the rabbit corpus cavernosum
tissue strip model and have been evaluated in preclini-
cal pharmacokinetic animal models.
Sildenafil (1, Viagra) (Chart 1) is the first and only
orally active phosphodiesterase type 5 (PDE5) inhibitor
available for the treatment of erectile dysfunction (ED).1
PDE5 is the major cGMP hydrolyzing enzyme in human
corpus cavernosum, and inhibition of this intracellular
enzyme elevates levels of the cyclic nucleotide. Upon
sexual stimulation, release of nitric oxide from nona-
drenergic, noncholinergic neurons activates guanylyl
cyclase, and PDE5 inhibition enhances cGMP accumu-
lation in the tissue. Increased cGMP levels cause a
decrease in intracellular calcium concentration, and this
leads to relaxation of vascular smooth muscle and
corpus cavernosum tissue, resulting in increased arterial
blood flow to the penis and ultimately erection.
The age-associated incidence of ED and the poten-
tially large patient population, estimated to be as many
as 30 million men in the United States alone,2 have
combined with the success of sildenafil to provide strong
stimuli for discovery and development of additional
PDE5 inhibitors. Chart 1 illustrates selected chemo-
Ch em istr y
The carboxylic acids 10a -c needed for coupling with
benzimidazoles 12a -g (Scheme 2) were prepared as
shown in Scheme 1. The specific amines illustrated here
were selected following library-based evaluation of over
80 different sulfonamides 8 (X ) H).7 Benzimidazoles
12a -g were prepared as described previously.5 Cou-
pling of these two fragments was done either via the
acid chloride (method A) or via a carbodiimide-mediated
approach (method B, Scheme 2). Cyclization of the
intermediate benzamide was accomplished using potas-
sium tert-butoxide/tert-butyl alcohol (Scheme 2).5 The
overall yield for these transformations was good to
excellent in most cases.
types that have been reported as potent (PDE5 IC50
<
5 nM) PDE5 inhibitors.3 One important issue facing a
new agent is selectivity versus other isozymes in the
phosphodiesterase superfamily.4 In the literature, 3-6
have selectivity data reported for PDE1-3.3a-d We have
previously compared 1 and 2 against PDE1-6 and
shown that 2 is significantly more specific in vitro than
sildenafil.5 PDE1 and 6 are cGMP-hydrolyzing phos-
phodiesterases found in the vasculature and retina,
respectively,6 and inhibition of these isozymes may be
associated with some of the adverse side effects of
sildenafil therapy, e.g., headache and facial flushing
(possibly PDE1) and visual disturbances (PDE6). An
improved, second-generation PDE5 inhibitor would be
one with greater potency and specificity for PDE5,
resulting in an agent with potentially fewer PDE-
associated side effects and greater efficacy as a treat-
ment for ED.
To assess the effect of the linker (sulfonamide vs
amide) on in vitro PDE5 potency, a sulfonamide ana-
logue (16) of 2 and amide derivatives (18a ,b) of sul-
fonamides 13g,j, respectively, were prepared (Scheme
3). The sulfonyl chloride intermediate 15, needed for 16,
was prepared by chlorosulfonation of 14. This crude
material was reacted with ammonia to furnish 16.
Carboxylic acid 175 was coupled with 4-ethylpiperazine
and (R)-3-(dimethylamino)pyrrolidine to furnish 18a ,b,
respectively (Scheme 3).
* To whom correspondence should be addressed. Phone: 609-818-
5398. Fax: 609-818-3450. E-mail: david.rotella@bms.com.
† Department of Discovery Chemistry.
#
Department of Metabolic and Cardiovascular Drug Discovery.
10.1021/jm000336j CCC: $19.00 © 2000 American Chemical Society
Published on Web 11/22/2000

5038 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 26
Rotella et al.
Ch a r t 1
Ch a r t 2
Sch em e 1
Resu lts
ing to 13j, which is one of the most potent PDE5
inhibitors reported to date.
In Vitr o P DE5 Activity. Compounds 13a -k , 16,
and 18a ,b were screened against PDE1-6 as previously
described.5 The data in Table 1 indicate that there is a
measurable preference for an N-ethylpiperazine moiety
(e.g. 13b) over the N-methyl derivative (7). The pyrro-
lidine sulfonamides 13h ,i illustrate a stereospecific
effect of the amine substituent on PDE5 potency, with
the R enantiomer preferred approximately 20-fold over
the S isomer. Comparison of the 4-fluorobenzyl deriva-
tive 13a with 7 suggests that in this series of sulfona-
mides, a substituted benzyl moiety has a potency-
enhancing effect. In the 2-methoxy analogue 13d , the
substituent exerts a more significant effect, resulting
in a compound equipotent with 2. A similar tendency
was observed in the (R)-pyrrolidine sulfonamides, lead-
Selectivity for PDE5 over other PDEs is uniformly
enhanced by the N-3-benzyl moiety, resulting in com-
pounds with little, if any, affinity for PDE1-4. In the
case of PDE6, the effect of the benzyl substituent and
sulfonamide was more complex. N-methyl sulfonamide
13a , with a 4-fluorobenzyl substituent, was 60-fold
selective for PDE5 compared to the unsubstituted
analogue 7, which was 20-fold selective. The 50-fold
selectivity shown by N-ethyl sulfonamide 13b was due
primarily to improved PDE5 potency, rather than
decreased affinity for PDE6 when compared to 13a .
Comparison of the PDE6 IC50 values for N-ethylpipera-
zine (13b) and pyrrolidine (13h ) sulfonamides clearly
indicated that the latter had the undesired effect of
improving PDE6 potency. Interestingly, in this series,

N-3-Benzylimidazoquinazolinone Sulfonamides
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 26 5039
Sch em e 2
13j/18b. We hypothesize that the geometry of the
sulfonamide linkage allows the more complex amines
to interact favorably with the enzyme. In contrast, the
planar primary carboxamide is preferred by the enzyme,
compared to the corresponding sulfonamide.
In Vitr o F u n ction a l Eva lu a tion . We had previ-
ously noted that addition of an N-3-benzyl moiety had
variable effects on the efficacy of PDE5 inhibitors in the
rabbit corpus cavernosum strip assay, where amide 2
was as active as sildenafil, while sulfonamide 7 was only
weakly efficacious.5 In this test, the compounds are
delivered in solution to a strip of rabbit corpus caver-
nosum tissue suspended in an aqueous tissue bath. To
show activity in this model, molecules must efficiently
diffuse into cells where the enzyme is located. The data
in Table 2 indicate that only a few of these sulfonamides
are active in this functional test. There is no obvious
association between in vitro PDE5 potency and efficacy,
as 13b with an IC50 of 1.1 nM was inactive, while 13f
with the same IC50 had moderate activity. Similarly, 2
and 13j are equipotent in vitro, but 2 was more
efficacious in this functional assay. Of the three mol-
ecules that showed measurable efficacy at 30 nM
(13d ,f,k ), only one, 13d , showed a dose response when
tested at 300 nM (data not shown). From these results,
we conclude that activity in the tissue strip assay is not
a simple consequence of in vitro inhibition of PDE5.
These results cannot be explained by significant differ-
ences in lipophilicity because the cLogP of sulfonamides
13a -k as well as that of 7 and amide 2 are all within
0.6 units of each other.9 Because of the activity of
carboxamide 2 in the rabbit corpus cavernosum strip
assay, a pharmacokinetic evaluation was carried out.
The drug was administered as a solution to fasted dogs
at a dose of 8 mg/kg. The compound showed poor oral
bioavailability (F e 10%) and as a consequence was not
considered for further study in animal models of ED.
Ta ble 1. PDE5 IC50 and Selectivity Ratios for Other PDEsa
IC50 (nM)
PDE5
IC50 ratio
compd
PDE1/5 PDE2/5 PDE3/5 PDE4/5 PDE6/5
1
2
7
1.6 ( 0.5
0.48 ( 0.1
5.3 ( 1.1
3.3 ( 0.5
1.1 ( 0.6
1.6 ( 0.2
0.62 ( 0.09
1.4 ( 0.3
1.1 ( 0.3
1.7 ( 0.9
1.1 ( 0.05
27 ( 9
140
>104
>105
>104
>104
>104
>104
>104
>104
7000
>104
>104
NT
3500
>105
8800
>104
>104
>104
>104
>104
>104
>104
>104
NT
2600
4200
600
8
>105
3400
>104
>104
>104
>104
>104
>104
>104
>104
NTc
60
20
60
50
40
90
80
50
80
30
3
70
50
20
NT
NT
13a
13b
13c
13d
13e
13f
13g
13h
13i
13j
13k
16
>104
>104
>104
>104
>104
3500
>104
4000
NT
Con clu sion s
In this paper, we have demonstrated a general
strategy for the preparation of highly PDE5-selective
sulfonamide-based N-3-benzylimidazoquinazolinones,
some of which are more potent than sildenafil against
the isolated enzyme. Substitution of the benzyl group
improves in vitro potency relative to the parent com-
pounds. Two different amines were identified which
gave potent, selective PDE5 inhibitors, and stereoselec-
tive effects on PDE5 potency were demonstrated using
chiral amines. All of these compounds are more selective
than sildenafil in vitro against other phosphodiesterase
isozymes. In particular, substantially improved specific-
ity against other cGMP-hydrolyzing phosphodiesterases
(PDE1,6) has been achieved. The SAR results in this
work indicate that the N-3-benzyl moiety is the primary
feature associated with this increase in specificity,
especially relative to PDE6. In the sulfonamide series,
this can be amplified or attenuated depending on the
amine used. There appears to be a divergence between
the SAR data obtained for sulfonamides and carboxa-
mides using the same amine, and the details of this
phenomenon remain to be elucidated. Our data suggests
that the amine moiety interacts within the active site
of PDE5 and does not project into bulk solvent. In vitro
functional efficacy in this series can be obtained, but
0.20 ( 0.06
0.52 ( 0.1
3.2 ( 0.5
35b
>104
>104
>104
NT
>104
>104
>104
NT
>104
>104
>104
NT
9600
2500
700
NT
NT
18a
18b
40b
NT
NT
NT
a
Enzyme sources: PDE1, bovine heart; PDE2, rat kidney;
PDE3, human platelet; PDE4, rat kidney; PDE5, human platelet;
PDE6, bovine retina. All entries are averages based on at least 3
determinations, except where noted. b Average of 2 determinations.
c NT ) not tested.
a 4-fluoro substituent (13j: PDE6 IC50 ) 14 nM) seemed
to improve PDE6 affinity (13h : PDE6 IC50 ) 33 nM),
unlike the effect observed in the piperazine sulfona-
mides (e.g. 13a ).
The data in Table 1 indicate that amide and sulfona-
mide linker SARs appear to be different with the same
set of amines. Using ammonia, an amide (2) is preferred
over a sulfonamide (16). However, with piperazine and
(dimethylamino)pyrrolidine, the sulfonamide moiety is
clearly preferred: compare for example 13g/18a and

5040 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 26
Rotella et al.
Sch em e 3
Ta ble 2. Rabbit Corpus Cavernosum Functional Assaya
1-1.5 h, then washed twice with water and twice with brine.
The organic solution was dried and concentrated to furnish
the product sulfonamide esters. These materials were used
without further purification and were found to be single spots
by TLC analysis (1:1 acetone/methylene chloride).
The appropriate methyl ester was dissolved in 40-50 mL
of THF with 10-15 mL of water and 1 equiv of LiOH hydrate.
The reaction was heated to reflux overnight and solvents were
removed by rotary evaporation to furnish a quantitative yield
of the lithium salt of the carboxylic acid as a white solid.
% control relaxation
integral (30 nM)
% control relaxation
integral (30 nM)
compd
compd
1
2
7
13a
13b
13c
13d
150 ( 20
140 ( 10
120 ( 10
90 ( 10
90 ( 6
90 ( 10
140 ( 20
13e
13f
13g
13h
13j
13k
16
100 ( 10
130 ( 10
100 ( 10
110 ( 10
120 ( 5
130 ( 30
100 ( 5
5-[(4-Eth yl-4-p ip er a zin yl)su lfon yl]-2-p r op oxyben zoic
a cid lith iu m sa lt (10a ): LRMS [MH+] 357; 1H NMR (400
MHz, D2O) δ 7.73 (d, 1Η, J ) 2.4 Ηz), 7.65 (1H, dd, J ) 2.5,
9.1 Hz), 7.57 (1H, d, J ) 2.5 Hz), 7.14 (1H, d, J ) 9.1 Hz),
4.05 (t, 2H, J ) 6.5 Hz), 3.03 (br s, 4H), 2.52 (br. s, 4H), 2.40
(apparent q, 2H, J ) 7.2, 14.4 Hz), 1.79-1.87 (m, 2H), 1.03 (t,
3H, J ) 7.2 Hz).
(R)-5-[(3-(Dim eth yla m in o)p yr r olid in yl)su lfon yl]-2-p r o-
p oxyben zoic a cid lith iu m sa lt (10b): HPLC retention time
2.14 min; LRMS [MH+] 357.
a
Control (untreated) response ) 100%. Values are based on the
response of at least 4 independent tissue strips.
activity in the corpus cavernosum model requires ef-
ficient cell penetration of the inhibitor. This result
highlights the challenges associated with identifying an
efficacious PDE5 inhibitor that may be useful in vivo.
Exp er im en ta l Section
General details have been previously described.5
(S)-5-[(3-(Dim eth yla m in o)p yr r olid in yl)su lfon yl]-2-p r o-
p oxyben zoic a cid lith iu m sa lt (10c): HPLC retention time
2.14 min; LRMS [MH+] 357.
The preparation of N-3-benzylbenzimidazoles 12a -g was
carried out as reported previously.5 The yield reported for each
compound is the net yield for the three-step transformation
from the 2,4-dinitro-5-chlorobenzamide precursor.
5-Am in o-1-[[(2-ch lor op h en yl)m eth yl]a m in o]-1H-ben z-
im id a zole-6-ca r boxa m id e (12a ): yield 100%; LRMS [MH+]
301; NMR (400 MHz, CD3OD) δ 9.54 (s, 1H), 8.36 (s, 1H), 7.68
(s, 1H), 7.35-7.42 (m, 4H), 5.75 (s, 2H).
5-Am in o-1-[[(2-m eth oxyph en yl)m eth yl]am in o]-1H-ben z-
im id a zole-6-ca r boxa m id e (12b): yield 87%; LRMS [MH+]
297; NMR (400 MHz, CD3OD) δ 9.67 (s, 1H), 8.68 (s, 1H), 8.00
(s, 1H), 7.67 (dd, 1H, J ) 1.5, 7.4 Hz), 7.38-7.42 (m, 1H), 7.03-
7.12 (m, 2H), 5.77 (s, 2H), 3.86 (s, 3H).
5-Am in o-1-[[(3-flu or op h en yl)m eth yl]a m in o]-1H-ben z-
im id a zole-6-ca r boxa m id e (12c): yield 100%; LRMS [MH+]
285; NMR (400 MHz, CD3OD) δ 9.74 (s, 1H), 8.57 (s, 1H), 8.00
(s, 1H), 7.42-7.50 (m, 1H), 7.29-7.39 (m, 2H), 7.10-7.20 (m,
1H), 5.86 (s, 2H).
5-Am in o-1-[[(3-m eth oxyph en yl)m eth yl]am in o]-1H-ben z-
im id a zole-6-ca r boxa m id e (12d ): yield 94%; LRMS [MH+]
297; NMR (400 MHz, CD3OD) δ 9.51 (s, 1H), 8.40 (s, 1H), 7.67
(s, 1H), 7.36 (t, 1H, J ) 7.8 Hz), 6.96-7.07 (m, 3H), 5.70 (s,
2H), 3.81 (s, 3H).
Meth yl 4-ch lor osu lfon yl-2-p r op oxyben zoic Acid (9). To
a solution of methyl salicylate (25 g, 0.16 mol) in 300 mL of
acetone were added 34 g of potassium carbonate (0.25 mol)
and 84 g (0.49 mol) of 1-iodopropane. The mixture was heated
to reflux for 24 h, then cooled and filtered. The filtrate was
concentrated and the residue taken up in 500 mL of ether.
The ether solution was washed twice with water and twice
with brine, then dried and concentrated to give a faintly yellow
oil. This material was a single spot by TLC (1:1 methylene
chloride/hexane, Rf 0.4) and was used without further purifica-
tion.
The crude propoxy ether was added dropwise at 0 °C to a
mixture of 35 mL of chlorosulfonic acid and 9.6 mL of thionyl
chloride over 30 min. The dark red reaction was stirred
overnight and was allowed to warm slowly to room tempera-
ture. The reaction was cautiously poured over 500 g of ice and
stirred to deposit a yellow solid which was recrystallized from
cyclohexane to furnish 13 g (0.045 mol, 28% yield) of fine white
needles: mp 58-59 °C; 13C NMR (100 MHz, CDCl3) δ 164.7,
163.9, 135.6, 132.7, 131.8, 121.5, 113.6, 71.6, 52.9, 22.6, 10.7.
Gen er a l P r oced u r e for Syn th esis of Ca r boxylic Acid s
10a -c. The appropriate amine (1.1 equiv) was dissolved in
methylene chloride (5-20 mL) along with 1.3 equiv of triethy-
lamine and cooled to 0 °C. Sulfonyl chloride 9 was dissolved
in methylene chloride (10-15 mL) and added dropwise over
30 min to this solution. The reaction was stirred at 0 °C for

N-3-Benzylimidazoquinazolinone Sulfonamides
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 26 5041
5-Am in o-1-[[(4-Ch lor op h en yl)m eth yl]a m in o]-1H-ben z-
im id a zole-6-ca r boxa m id e (12e): yield 72%; LRMS (M + H)
301; NMR (400 MHz, CD3OD) δ 9.55 (s, 1H), 8.44 (s, 1H), 7.76
(s, 1H), 7.55-7.59 (m, 2H), 7.16-7.21 (m, 2H), 7.03-7.12 (m,
2H), 5.75 (s, 2H).
Physical data for 12f,g have been previously reported.5
Coupling of benzimidazoles 12a -g with carboxylic acids
10a -c using methods A and B (Scheme 2) have been previ-
ously described along with the cyclization of the resulting
benzamides to afford the target imidazoquinazolinones.5 The
specific coupling method used and net yield for these two steps
are indicated with each compound. Physical data for target
compounds 13a -k are as follows.
1-[[3-[1-[(4-F lu or op h en yl)m et h yl]-7,8-d ih yd r o-8-oxo-
1H-im id a zo[4,5-g]qu in a zolin -6-yl]-4-p r op oxyp h en yl]su l-
fon yl]-4-m eth ylp ip er a zin e (13a ): method B; yield 80%; mp
227-228 °C; NMR (400 MHz, CDCl3) δ 10.78 (s, 1H), 8.97 (d,
1H, J ) 2.4 Hz), 8.29 (apparent d, 2H, J ) 7.5 Hz), 8.17 (s,
1H), 7.86 (dd, 1H, J ) 2.4, 8.7 Hz), 7.04-7.26 (partially
obscured by solvent, m, 7H), 5.44 (s, 2H), 4.26 (t, 2H, J ) 6.5,
13 H), 3.13, (br s, 4H), 2.51 (br s, 4H), 2.27 (s, 3H), 2.02-2.07
(m, 2H), 1.18 (t, 3H, J ) 7.5 Hz). Anal. Calcd for C30H31N6O4-
SF‚1.3H2O: C, H, N.
1-[[3-[7,8-Dih yd r o-8-oxo-1-(p h en ylm eth yl)-1H-im id a zo-
[4,5-g]qu in a zolin -6-yl]-4-p r op oxyp h en yl]su lfon yl]-4-eth -
ylp ip er a zin e (13b): method A; yield 64%; mp 135-137 °C;
NMR (400 MHz, CDCl3) δ 10.8 (s, 1H), 8.93 (s, 1H), 8.31 (s,
1H), 8.25 (s, 1H), 8.18 (s, 1H), 7.84 (dd, 1H, J ) 2.5, 8.7 Hz)
7.30-7.40 (m, 2H), 7.18-7.28 (m, 2H) 7.16 (d, 1H, J ) 8.9 Hz),
5.46 (s, 2H), 4.24 (t, 2H, J ) 6.5 Hz), 3.13 (br s, 4H), 2.56 (br
s, 4H), 2.40 (apparent q, 2H, J ) 7.2, 14.4 Hz), 1.98-2.07 (m,
2H), 1.16 (t, 3H, J ) 7.4 Hz), 1.02 (t, 3H, J ) 7.2 Hz). Anal.
Calcd for C31H34N6O4S‚H2O: C, H, N.
1-[[3-[1-[(2-Ch lor op h en yl)m et h yl]-7,8-d ih yd r o-8-oxo-
1H-im id a zo[4,5-g]qu in a zolin -6-yl]-4-p r op oxyp h en yl]su l-
fon yl]-4-eth ylp ip er a zin e (13c): method A; yield 49%; mp
225-226 °C; NMR (400 MHz, CDCl3) δ 10.8 (s, 1H), 8.97 (d,-
1H, J ) 2.1 Hz), 8.28 (d, 1H, J ) 3 Hz), 8.2 (s, 1H), 7.86 (dd,
1H, J ) 2.1, 8.6 Hz), 7.1-7.4 (m, 6H), 5.4 (s, 2H), 4.3 (t, 2H,
J ) 6.4 Hz), 3.1 (br s, 4H), 2.5 (br s, 4H), 2.4 (dd, 2H, J ) 7.1,
14.3 Hz), 2.02-2.07 (m, 2H), 1.17 (t, 3H, J ) 6.4 Hz), 1.02 (t,
3H, J ) 7.1 Hz). Anal. Calcd for C31H33N6O4SCl‚1.2H2O: C,
H, N, Cl.
1-[[3-[7,8-Dih yd r o-1-[(2-m eth oxyp h en yl)m eth yl]-8-oxo-
1H-im id a zo[4,5-g]qu in a zolin -6-yl]-4-p r op oxyp h en yl]su l-
fon yl]-4-eth ylp ip er a zin e (13d ): method A; yield 70%; mp
180-182 °C; NMR (400 MHz, CDCl3) δ: 10.7 (s, 1H), 8.97 (dd,
1H, J ) 2.2 Hz), 8.43 (s, 1H), 8.23 (apparent d, 2H, J ) 7.5
Hz), 7.86 (dd, 1H, J ) 2.3, 8.7 Hz), 7.15-7.34 (m, partially
obscured by solvent, 3H), 6.93 (apparent t, 2H, J ) 7.8 Hz),
5.44 (s, 2H), 4.26 (t, 2H, J ) 6.5 Hz), 3.89 (s, 3H), 3.13 (br s,
4H), 2.55 (br s, 4H), 2.40 (apparent q, 2H, J ) 7.1, 14.3 Hz),
2.01-2.09 (m, 2H), 1.18 (t, 3H, J ) 7.5 Hz), 1.02 (t, 3H, J )
7.2 Hz). Anal. Calcd for C32H36N6O5S‚H2O: C, H, N, S.
1-[[3-[1-[(3-F lu or op h en yl)m et h yl]-7,8-d ih yd r o-8-oxo-
1H-im id a zo[4,5-g]qu in a zolin -6-yl]-4-p r op oxyp h en yl]su l-
fon yl]-4-eth ylp ip er a zin e (13e): method B; yield 93%; mp
115-117 °C; NMR (400 MHz, CDCl3) δ 10.7 (s, 1H), 8.97 (d,
1H, J ) 2.3 Hz), 8.61 (d, 1H, J ) 2.4 Hz), 8.28 (apparent d,
2H, J ) 2.1 Hz), 8.20 (s, 1H), 7.86 (dd, 1H, J ) 2.3, 8.7 Hz),
7.33 (apparent q, 1H, J ) 5.9, 7.9 Hz), 7.17 (d, 1H, J ) 8.8
Hz), 7.00-7.06 (m, 2H), 6.90 (d, 1H, J ) 9.1 Hz), 5.47 (s, 2H),
4.26 (t, 2H, J ) 6.5 Hz), 3.13 (br s, 4H), 2.55 (br. S, 4H), 2.41
(apparent q, 2H, J ) 7.1, 14.3 Hz), 2.04 (apparent q, 2H, J )
6.9, 14.1 Hz), 1.17 (t, 3H, J ) 7.5 Hz), 1.02 (t, 3H, J ) 7.2 Hz).
Anal. Calcd for C31H33N6O4SF‚2H2O: C, H, N, F.
5.43 (s, 2H), 4.25 (t, 2H, J ) 6.5 Hz), 3.76 (s, 3H), 3.13 (br s,
4H), 2.55 (br s, 4H), 2.40 (apparent q, 2H, J ) 7.2, 14.4 Hz),
2.01-2.07 (m, 2H), 1.17 (t, 3H, J ) 7.5 Hz), 1.02 (t, 3H, J )
7.2 Hz). Anal. Calcd for C32H36N6O5S‚1.2H2O: C, H, N.
1-[[3-[1-[(4-F lu or op h en yl)m et h yl]-7,8-d ih yd r o-8-oxo-
1H-im id a zo[4,5-g]qu in a zolin -6-yl]-4-p r op oxyp h en yl]su l-
fon yl]-4-eth ylp ip er a zin e (13g): method B; yield 52%; mp
155-158 °C; NMR (400 MHz, CDCl3) δ 10.78 (s, 1H), 8.97 (d,
1H, J ) 2.4 Hz) 8.30 (s, 1H), 8.27 (s, 1H), 8.18 (s, 1H), 7.86
(dd, 1H, J ) 2.4, 8.7 Hz), 7.15-7.26 (m, partially obscured by
solvent, 3H), 7.05 (apparent t, 1H, J ) 6.7 Hz), 5.44 (s, 2H),
4.26 (t, 2H, J ) 6.5 Hz), 3.13 (br s, 4H), 2.55 (br s, 4H), 2.41
(apparent q, 2H, J ) 7.1, 14.2 Hz), 2.00-2.09 (m, 2H), 1.18 (t,
3H, J ) 7.4 Hz), 1.02 (t, 3H, J ) 7.2 Hz). Anal. Calcd for
C
31H33N6O4SF‚0.5H2O: C, H, N, F.
(R)-1-[[3-[7,8-Dih yd r o-8-oxo-1-(p h en ylm eth yl)-1H-im i-
d a zo[4,5-g]q u in a zolin -6-yl]-4-p r op oxyp h en yl]su lfon yl]-
3-(d im eth yla m in o)p yr r olid in e (13h ): method B; yield 71%;
mp >300 °C (HCl salt); NMR (400 MHz, CDCl3) δ 10.79 (s,
1H), 8.97 (d, 1H, J ) 2.4 Hz), 8.35 (s, 1H), 8.22-8.26 (m, 2H),
7.95 (dd, 1H, J ) 2.2, 8.8 Hz), 7.34-7.39 (m, 2H), 7.21-7.24
(m, partially obscured by solvent, 4H), 5.48 (s, 2H), 4.28 (t,
2H, J ) 6.5 Hz), 3.43-3.65 (m, 4H), 3.32-3.39 (m, 1H), 2.74
(br s, 6H), 2.30-2.37 (m, 1H), 2.01-2.06 (m, 2H), 1.17 (t, 3H,
J ) 7.4 Hz). Anal. Calcd for C31H34N6O4S‚H2O: C, H, N.
(S)-1-[[3-[7,8-Dih yd r o-8-oxo-1-(p h en ylm eth yl)-1H-im i-
d a zo[4,5-g]q u in a zolin -6-yl]-4-p r op oxyp h en yl]su lfon yl]-
3-(d im eth yla m in o)p yr r olid in e (13i): method B; yield 67%;
mp 190 °C dec (HCl salt); NMR (400 MHz, CDCl3) δ 10.78 (s,
1H), 9.00 (d, 1H, J ) 2.4 Hz), 8.33 (s, 1H), 8.26 (s, 1H), 8.18,
(s, 1H), 7.94 (dd, 1H, J ) 2.3, 8.7 Hz), 7.34-7.39 (m, 2H), 7.18-
7.24 (m, partially obscured by solvent, 4H), 5.47 (s, 2H), 4.27
(t, 2H, J ) 6.4 Hz), 3.62-3.66 (m, 1H), 3.48-3.52 (m, 1H),
3.31-3.38 (m, 1H), 3,11-3.22 (m, 1H), 2.80-3.00 (m, 1H), 2.35
(br s, 6H), 2.01-2.06 (m, 2H), 1.17 (t, 3H, J ) 7.4 Hz). Anal.
Calcd for C31H34N6O4S‚2.2H2O: C, H, N.
(R)-1-[[3-[1-[(4-F lu or op h en yl)m et h yl]-7,8-d ih yd r o-8-
oxo-1H-im id a zo[4,5-g]qu in a zolin -6-yl]-4-p r op oxyp h en yl]-
su lfon yl]-3-(d im eth yla m in o)p yr r olid in e (13j): method B;
yield 69%; mp 223-225 °C; NMR (400 MHz, CDCl3) δ 10.80
(s, 1H), 9.02 (d, 1H, J ) 2.3 Hz), 8.30 (s, 1H), 8.28 (s, 1H),
8.18 (s, 1H), 7.93 (dd, 1H, J ) 2.3, 8.7 Hz), 7.22-7.25 (m, 1H),
7.17 (d, 1H, J ) 8.8 Hz), 7.04-7.08 (apparent t, 1H, J ) 8.6
Hz), 5.44 (s, 2 H), 4.27 (t, 2H, J ) 6.5 Hz), 3.63 (apparent q,
1H, J ) 7.2, 9.1 Hz), 3.44-3.49 (m, 1H), 3.35 (apparent t, 1H,
J ) 7.2 Hz), 3.01 (apparent t, 1H, J ) 9.1 Hz), 2.63-2.70 (m,
1H), 2.19 (s, 6H), 2.02.-2.09 (m, 3H), 1.65-1.72 (m, 3H), 1.18
(t, 3H, J ) 7.2 Hz). Anal. Calcd for C31H33N6O4SF‚H2O: C, H,
N.
(R)-1-[[3-[1-[(4-Ch lor op h en yl)m et h yl]-7,8-d ih yd r o-8-
oxo-1H-im id a zo[4,5-g]qu in a zolin -6-yl]-4-p r op oxyp h en yl]-
su lfon yl]-3-(d im eth yla m in o)p yr r olid in e (13k ): method B;
yield 75%; mp 220-222 °C; NMR (400 MHz, CDCl3) δ 10.80
(s, 1H), 9.02 (d, 1H, J ) 2.5 Hz), 8.28 (s, 1H), 8.27 (s, 1H),
8.18 (s, 1H), 7.93 (dd, 1H, J ) 2.4, 8.7 Hz), 7.32-7.35 (m, 2H)
7.17 (apparent d, 1H, J ) 8.6 Hz), 5.44 (s, 2 H), 4.27 (t, 2H, J
) 6.5 Hz), 3.63 (apparent q, 1H, J ) 7.0, 9.3 Hz), 3.45
(apparent q, 1H, J ) 2.8, 9.4 Hz), 3.32-3.37 (m, 1H), 3.01
(apparent t, 1H, J ) 9.1 Hz), 2.63-2.70 (m, 1H), 2.19 (s, 6H),
2.02.-2.07 (m, 3H), 1.65-1.71 (m, 3H), 1.18 (t, 3H, J ) 7.4 Hz).
Anal. Calcd for C31H33N6O4SCl‚H2O: C, H, N, Cl.
1-[3-[7,8-Dih yd r o-8-oxo-1-(p h en ylm eth yl)-1H-im id a zo-
[4,5-g]qu in azolin -6-yl]-4-pr opoxyph en yl]su lfon am ide (16).
Benzimidazole 12g was coupled with 2-propoxybenzoic acid
and cyclized to quinazolinone 14 as described earlier.5 Com-
pound 14 (0.10 g, 0.23 mmol) was added portionwise over 15
min to an ice-cold solution of 5 mL of chlorosulfonic acid. The
reaction was stirred at 0 °C for 5 h, then cautiously poured
over ice to precipiatate a tan solid which was collected by
filtration, washed with water and dried. Without further
purification, this material was dissolved in 8 mL of a 0.5 M
solution of ammonia in dioxane. The reaction was stirred at
room temperature for 4 days, then concentrated to a tan solid
which was purified by flash chromatography eluting with 9:1
1-[[3-[7,8-Dih yd r o-1-[(3-m eth oxyp h en yl)m eth yl]-8-oxo-
1H-im id a zo[4,5-g]qu in a zolin -6-yl]-4-p r op oxyp h en yl]su l-
fon yl]-4-eth ylp ip er a zin e (13f): method A; yield 55%; mp
117-118 °C; NMR (400 MHz, CDCl3) δ 10.7 (s, 1H), 8.97 (d,
1H, J ) 2.4 Hz) 8.33 (s, 1H), 8.27 (s, 1H), 8.18 (s, 1H), 7.86
(dd, 1H, J ) 2.4, 8.7 Hz), 7.24-7.28 (partially obscured by
solvent, m, 1H), 7.16 (d, 1H, J ) 8.8 Hz), 6.86 (dd, 1H, J )
2.4, 8.0 Hz), 6.81 (d, 1H, J ) 7.9 Hz), 6.74 (d, 1H, J ) 2.4 Hz),

5042 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 26
Rotella et al.
methylene chloride:methanol to furnish 0.012 g of the desired
compound as a white solid (0.025 mmol, 12% net yield): NMR
(400 MHz, DMSO) δ 9.3-9.4 (br s, exch. with D2O), 8.81 (s,
1H), 8.32 (s, 1H), 8.13 (s, 1H), 8.00 (s, 1H), 7.92 (d, 1H, J ) 10
Hz), 7.30-7.46 (m, 5H), 7.21 (apparent t, 2H, J ) 8.4 Hz), 5.68
(s, 2H), 4.10 (t, 2H, J ) 6.5 Hz), 1.73 (m, 2H), 0.94 (t, 3H, J )
7.1 Hz); HRMS calcd for 507.1376, obsd (MH+) 508.1454.
calculated as the ratio of the IC50 value for inhibition of one
PDE divided by the IC50 value for inhibition of PDE5.
Ra bbit Cor pu s Ca ver n osu m Str ip Assa y. P h ysiologica l
sa lt solu tion s: A bicarbonate-buffered salt solution (PSS) was
used, containing (in mM): 118.4 NaCl, 4.7 KCl, 1.2 MgSO4,
1.2 KH2PO4, 1.9 CaCl2, 25.0 NaHCO3, and 10.1 D-glucose. The
solution additionally contained 3 µM indomethacin, 1 µM
atropine, and 5 µM guanethidine. Concentrated stock solutions
(10 mM) of test compounds were prepared and serially diluted
to the appropriate concentrations in DMSO. Appropriate
solvent/time controls were run in parallel.
1-Eth yl-4-[[3-[1-[(4-flu or op h en yl)m eth yl]-7,8-d ih yd r o-
8-oxo-1H-im idazo[4,5-g]qu in azolin -6-yl]-4-pr opoxyph en yl]-
carbonyl]piperazine (18a)and (R)-1-[[3-[1-[(4-Fluorophenyl)-
m e t h y l]-7,8-d ih y d r o -8-o x o -1H -im id a zo [4,5-g ]q u in a -
zolin -6-yl]-4-p r op oxyp h en yl]ca r bon yl]-3-(d im eth yla m i-
n o)p yr r olid in e (18b). Prepared as a part of a library of
analogues of 2 by the following procedure: Carboxylic acid 17
was dissolved in pyridine at room temperature. To this were
added 1.2 equiv of the appropriate amine, HOBt hydrate and
EDAC-HCl, along with 0.15 equiv of DMAP. The reaction was
stirred at room temperature for 4 h. The reaction was poured
into ice water to precipipate the desired products which were
collected by filtration, washed with water and 1.0 N NaOH
then dried to furnish the products which were then analyzed
by HPLC and LC/MS for purity and appropriate mass. Using
the standard HPLC and LC/MS conditions described earlier,5
18a had a retention time of 3.04 min, was at least 95% pure,
and showed an M + H peak of 569. Similarly, 18b had a
retention time of 2.99 min, was at least 90% pure, and showed
an M + H peak of 569. These materials were tested without
further purification as PDE5 inhibitors using the method
described below.
P DE Activity Assa y. Enzymatic activity is assayed using
a commercially available phosphodiesterase scintillation prox-
imity (SPA) assay kit with either [3H]cGMP or [3H]cAMP as
the substrate depending on the PDE of interest (Amersham
product #TRKQ 7090 for cAMP kit; #TRKQ 7100 for cGMP
kit). The manufacturer’s protocol was followed explicitly except
that the reactions are carried out at room temperature and 3
mM nonradioactive cyclic nucleotide was included in the
suspension of SPA beads to stop the synthesis of additional
radioactive products. The radioactive product of the reaction,
[3H]nucleotide monophosphate, preferentially bound to the
SPA beads, excited the scintillant embedded in the beads, and
was quantified on a Packard TopCount liquid scintillation
counter. Nonspecific binding to the beads (the assay blank)
was quantified by adding a 10000-fold excess of nonradioactive
substrate to the reaction mixture prior to the addition of
enzyme to prevent the synthesis of radioactive product. The
assay blank was subtracted from the activity of the enzyme
measured using the SPA kit as described above. The activity
in samples that received test compound was calculated as a
percent of the control activity measured in samples that only
received the vehicle.
En zym e P r ep a r a tion s. For PDE3,5 activity, the enzyme
source was sonicated human platelet homogenates prepared
by the method of Seiler et al.10 PDE5 accounted for ap-
proximately 90% of the [3H]cGMP hydrolytic activity in the
homogenates and PDE3 accounted for over 80% of the [3H]-
cAMP hydrolytic activity. PDE1 from bovine heart was pur-
chased from Sigma and assayed with 1 µM calmodulin and 4
mM calcium present in the reaction buffer and [3H]cGMP as
the substrate. PDE2,4 were purified from rat kidney cytosol
by Mono-Q anion-exchange chromatography as described.11
The substrate for both of these isoforms was [3H]cAMP.
Nonradioactive cGMP (1 µM) was also included in the reaction
buffer for the PDE2 assay because the cAMP hydrolytic
activity of this isoform was stimulated by cGMP. Rod outer
segment membranes were purified from bovine retinas as
described.12 PDE6 was stripped from the rod outer segment
membranes by hypotonic washes and trypsin-activated prior
to the assay as described.13 Trypsin treatment activated PDE6
by proteolytically degrading the inhibitory γ-subunit.
Tissu e p r ep a r a tion : Adult male New Zealand white
rabbits weighing 3.4-3.7 kg were sacrificed by intravenous
Nembutal injection. The entire penis (up to pelvic bone) was
quickly removed and placed in ice-cold aerated PSS. The skin
and connective tissue were carefully removed. The corpus
spongiosum was cut from the groove under the corpus caver-
nosum. The tunica albuginea was dissected from the corpus
cavernosum. The corpus cavernosum was carefully cut along
the midline and connective tissue was cleaned from the
“septal” region. The two corpora were longitudinally cut in half;
each of these strips was then cut in half along the transverse
axis, resulting in eight strips approximately 2 mm × 5 mm.
Each strip was individually mounted for isometric force
recording using silk suture in 10-mL tissue baths. One end
was secured between the two platinum plates of a field
stimulating electrode, which in turn was connected to a
micrometer for control of tissue length. The field stimulating
electrode was attached to a Stimu-splitter and Grass stimula-
tor. The other end of the suture was connected to a Grass
FT.03 force displacement transducer. The strips were bathed
in PSS maintained at 37 °C and bubbled with 95% O2 and 5%
CO2. Data were simultaneously recorded on a Grass recorder
and a Biopac System which additionally allowed measurement
of the integral at each frequency. A preload tension of 1 g was
applied to each strip. Following the equilibration period, the
strips were contracted with 3 µM PE until a steady state level
of force was attained. For the first frequency-response curve,
3 µM DMSO (solvent control) was added to each strip for 10
min. A frequency-response curve of the following parameters
was then performed: 24 V in the form of square wave pulses
of 0.2-ms duration delivered as 10-s train and a 4-min interval
between stimuli. A frequency of 32 Hz was chosen to measure
the effects of test compounds based on the signal-to-noise
response. After a 4-min recovery at 32 Hz, the strips were
washed extensively and allowed at least 50 min to recover
during which time tension adjustments and several washes
were done. The frequency-response curve was repeated in the
presence of test compound. Following another recovery period,
a
third frequency-response curve was performed in the
presence of a higher concentration of test compound.
Da ta a n a lysis: All force determinations were made assum-
ing that the lowest level of force following a washout period
defined 100% relaxation and the phenylephrine contraction
prior to DMSO or test compound addition was 0% relaxation.
The direct relaxation induced by test compound, measured
after the 10-min incubation period, was calculated relative to
these values. The peak relaxation responses at 32 Hz were
also measured relative to these values. Values were then
normalized to the maximum peak relaxation attained during
the control frequency-response curve. The integral for the
area under the curve at 32 Hz was determined for a set time
period (3.66 s from immediately prior to stimulation). The
integral for the relaxation response was then calculated as the
difference between the integral for the time period immediately
prior to the start of stimulation and that for the frequency.
All responses were normalized to the maximum attained
during the first frequency-response curve.
Each data set was fit to a sigmoidal dose-response curve
by nonlinear regression using the GraphPad Prism software
package.
Da ta An a lysis. Each data set was fit to a curve for
inhibition at a single site using the nonlinear regression
analysis in the Activity Base/XLFit software package, and IC50
values were calculated by this analysis. Selectivity ratios were
Su p p or tin g In for m a tion Ava ila ble: Elemental analysis
data for compounds 13a -h ,j,k . This material is available free
of charge via the Internet at http://pubs.acs.org.

N-3-Benzylimidazoquinazolinone Sulfonamides
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 26 5043
(5) Rotella, D. P.; Sun, Z.; Zhu, Y.; Krupinski, J .; Pongrac, R.;
Seliger, L.; Normandin, D.; Macor, J . E. N-3-Substituted imi-
dazoquinazolinones: potent and selective PDE5 inhibitors as
potential agents for treatment of erectile dysfunction. J . Med.
Chem. 2000, 43, 1257-1263.
Refer en ces
(1) (a) Brock, G. Sildenafil citrate (Viagra). Drugs Today 2000, 36,
125-134. (b) Lue, T. F. Erectile Dsyfunction. N. Engl. J . Med.
2000, 342, 1802-1818.
(2) Feldman, H. A.; Goldstein, I.; Hatzichristou, D. G.; Krane, R.
J .; McKinlay, J . B. Impotence and its medical and physiological
correlates: results of the Massachusetts male aging study. J .
Urol. 1994, 151, 54-56.
(6) Beavo, J . A. Cyclic nucleotide phosphodiesterases: functional
implications of multiple isoforms. Physiol. Rev. 1995, 75, 725-
748.
(7) Rotella, D. P.; Sun, Z.; Zhu, Y. Unpublished results.
(8) This phenomenon has been observed with other unrelated
structures. Bi, Y. Manuscript in preparation.
(9) These calculations were done using the neutral species of each
molecule.
(10) Seiler, S.; Gillespie, E.; Arnold, A. J .; Brassard, C. L.; Meanwell,
N. A. and Fleming, J . S. Imidazoquinoline derivatives: potent
inhibitors of platelet cAMP phosphodiesterase which elevate
cAMP levels and activate protein kinase in platelets. Thrombosis
Res. 1991, 62, 31-42.
(11) Hoey, M.; Houslay M. D. Identification and selective inhibition
of four distinct soluble forms of cyclic nucleotide phosphodi-
esterase activity from kidney. Biochem. Pharmacol. 1990, 40,
193-202.
(12) Fung, B. K.-K. and Stryer, L. Photolyzed rhodopsin catalyzes
the exchange of GTP for bound GDP in retinal rod outer
sefments. Proc. Natl. Acad. Sci. U.S.A. 1980, 77, 2500-2504.
(13) Berger, A. L.; Cerione, R. A.; Erickson, J . W. Real time confor-
mational changes in the retinal phosphodiesterase γ subunit
monitored by resonance energy transfer. J . Biol. Chem. 1997,
272, 2714-2721.
(3) (a) Lee, S. J .; Konishi, Y.; Yu, D. T.; Miskowski, T. A.; Riviello,
C. M.; Macina, O. T.; Frierson, M. R.; Kondo, K.; Sugitani, M.;
Sircar, J . C.; Blazejewski, K. M. Discovery of potent cyclic GMP
phosphodiesterase inhibitors. 2-Pyridyl- and 2-imidazolylquinazo-
lines possessing cyclic GMP phosphodiesterase and thromboxane
synthesis inhibitory activities. J . Med. Chem. 1995, 38, 3547-
3557. (b) Watanabe, N.; Adachi, H.; Takase, Y.; Ozaki, H.;
Matsukura, M.; Miyazki, K.; Ishibashi, K.; Ishihara, H.; Kodama,
K.; Nishino, M.; Kakiki, M.; Kabasawa, Y. 4-(3-Chloro-4-meth-
oxybenzyl)aminophthalazines: synthesis and inhibitory activity
toward phosphodiesterase 5. J . Med. Chem. 2000, 43, 2523-
2529. (c) Ukita, T.; Nakamura, Y.; Kubo, A.; Yamamato, Y.;
Takahashi, M.; Kotera, J .; Ikeo, T. 1-Arylnaphthalene lignan:
a novel scaffold for type 5 phosphodiesterase inhibitor. J . Med.
Chem. 1999, 42, 1293-1305. (d) Ho, G. D.; Silverman, L.;
Bercovici, A.; Puchalski, C.; Tulshian, D.; Xia, Y.; Czarniecki,
M.; Green, M.; Cleven, R.; Zhang, H.; Fawzi, A. Synthesis and
evaluation of potent and selective c-GMP phosphodiesterase
inhibitors. Bioorg. Med. Chem. Lett. 1999, 9, 7-12.
(4) Corbin, J . D.; Francis, S. H. Cyclic GMP phosphodiesterase-5:
target of sildenafil. J . Biol. Chem. 1999, 274, 13729-13732.
J M000336J