Benzodiazepine-based selective inhibitors of mitochondrial F 1F0 ATP hydrolase

Article abstract of DOI:10.1016/j.bmcl.2003.11.052

A series of benzodiazepine-based inhibitors of mitochondrial F ₁F₀ ATP hydrolase were prepared and evaluated for their ability to selectively inhibit the enzyme in the forward direction. Compounds from this series showed excellent po

Full text of DOI:10.1016/j.bmcl.2003.11.052

Bioorganic & Medicinal Chemistry Letters 14 (2004) 1031–1034
Benzodiazepine-based selective inhibitors of mitochondrial F1F0
ATP hydrolase
Lawrence G. Hamann,a,* Charles Z. Ding,a,y Arthur V. Miller,a Cort S. Madsen,b
Paulina Wang,b Philip D. Stein,a Andrew T. Pudzianowski,c David W. Green,b,{
Hossain Monshizadeganb and Karnail S. Atwala
aDepartment of Discovery Chemistry, Bristol-Myers Squibb, Pharmaceutical Research Institute, PO Box 5400,
Princeton, NJ 08543-5400, USA
bDepartment of Cardiovascular Biology, Bristol-Myers Squibb, Pharmaceutical Research Institute, PO Box 5400,
Princeton, NJ 08543-5400, USA
cDepartment of Computer-Assisted Drug Design, Bristol-Myers Squibb, Pharmaceutical Research Institute, PO Box 5400,
Princeton, NJ 08543-5400, USA
Received 8 September 2003; revised 14 November 2003; accepted 18 November 2003
Abstract—A series of benzodiazepine-based inhibitors of mitochondrial F1F0 ATP hydrolase were prepared and evaluated for their
ability to selectively inhibit the enzyme in the forward direction. Compounds from this series showed excellent potency and selectivity
for ATP hydrolase versus ATP synthase, suggesting a potentially beneficial profile useful for the treatment of ischemic heart disease.
# 2003 Elsevier Ltd. All rights reserved.
1. Introduction
The impairment of contractile function in ischemic
muscle is associated with mitochondrial levels of
adenosine triphosphate (ATP) and adenosine triphos-
phatases (ATPases).1 ATPases typically catalyze the
hydrolysis of ATP, to adenosine monophosphate (AMP)
or adenosine diphosphate (ADP), plus phosphate ions
and energy. The contractile function of the heart is
regulated by the transport of calcium, sodium, and
potassium ions, which in turn is modulated by ATP and
ATPases. To maintain homeostasis the cells supply of
ATP must be replenished as it is consumed (e.g., with
muscle contraction). Thus, during steady state the rate
of ATP synthesis is closely matched to its rate of con-
sumption. However, unlike other ATPases which func-
tion typically to hydrolyze ATP and release energy, the
mitochondrial F1F0-ATPase (mATPase)2 has both
hydrolytic and synthetic states. As ‘ATP synthase’, the
mATPase catalyzes the production of ATP via oxidative
phosphorylation of ADP and Pi.3
Every year, large numbers of patients die from or
experience complications from ischemic heart disease,
which include acute myocardial infarction, congestive
heart failure, cardiac arrhythmias, or other disorders.
Myocardial ischemia exists when heart tissue demands
for oxygen and substrates exceed the supply. These
imbalances span a large range of various syndromes and
biochemical pathways involved in the pathogenesis of
low-grade to severe ischemic conditions. Major con-
sequences of myocardial ischemia include mechanical
and electrical dysfunction, muscle cell damage, and
development of necrosis. Ultimately, if the ischemia is
sufficiently severe there will be an immediate reduction
(or cessation) of contractile function in the heart and a
potential to generate arrhythmias.
Keywords: Mitochondrial ATP hydrolase; Ischemia; Heart disease.
* Corresponding author. Tel.: +1-609-818-5526; fax: +1-609-818-
3550; e-mail: lawrence.hamann@bms.com
Under normoxic conditions, mATPase modulates ATP
production via its two units, the F1 (catalytic domain)
and F0 (inner membrane domain) complexes,4 whereby
electrons from ATPase substrates are transferred
to oxygen and protons are transported out of the
y
Current Address: Cumbre, Inc., 1502 Viceroy Drive, Dallas, TX
75235-2304, USA.
{
Current Address: Abbott Bioresearch Center, 100 Research Drive,
Worcester, MA 01605-5314, USA.
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.11.052

1032
L. G. Hamann et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1031–1034
mitochondrial inner matrix, creating an electrochemical
proton gradient. The energy resulting from proton flow
back across the mitochondrial membrane is captured by
the F0 domain which drives the F1 domain to synthesize
ATP. Under ischemic conditions, however, this electro-
chemical gradient collapses, and mATPase switches to
its hydrolytic state, with a concomitant down-regulation
of F1F0-ATP synthase (up to 50–80% in ischemic
muscle). This leads to a futile cycling and wasting of
ATP which may increase cardiac contractility, vasocon-
striction, sensitivity to vasoactive agents, and arterial
blood pressure. The conversion of F1F0-ATP synthase
to F1F0-ATP hydrolase is reversible, as addition of
substrate and oxygen to the mitochondria of ischemic
muscle can reactivate F1F0-ATPase and return ATP
levels to control levels.5 A native peptide, IF1 inhibitor
protein (IF1)6 may bind to the F1 unit under ischemic
conditions to inhibit the ATP hydrolase activity of the
enzyme; however, IF1 binding is highly pH dependent.
Concentrations of its active form decrease at high pH,
requiring saturating levels of active IF1 to fully inhibit
the enzyme.
analogue synthesis, while selectivity for hydrolase versus
synthase inhibition was maintained throughout.
2. Chemistry
All compounds were prepared by appending a variety of
peripheral aryl-containing moieties to an efficiently
assembled 2-phenethylbenzodiazepine core (Scheme 1).13
2-Aminobenzylamine (4) was reacted with ethyl 2-oxo-4-
phenylbutyrate in refluxing toluene using a Dean-Stark
apparatus for removal of water to give tetrahydro-
quinazoline 5. Reduction of 5 with triethylsilane in TFA
and 1,2-dichloroethane gave tetrahydrobenzodiazepin-3-
one 7 through the intermediacy of amino ester 6.
Reduction of 7 to give diamine 8 was effected with
lithium aluminum hydride in THF. The diamine 8 was
then either sulfonylated with a sulfonyl chloride or alkyl-
ated with an alkyl bromide under standard conditions to
give N-sulfonyl or N-alkyl-substituted tetrahydro-
benzodiazepines 9 respectively. Reductive amination with
the appropriate imidazolyl aldehyde under standard
conditions afforded inhibitors 10–31.
Several inhibitors of mATPase have been described,
including efrapeptin,7 oligomycin,8 aurovertin B,9 and
azide.10 Oligomycin targets F0 and reportedly postpones
cell injury by preserving ATP during ischemia. How-
ever, until recent disclosures from these laboratories (111
and 212), the only inhibitors of mATPase known in the
literature were large proteins or peptides which are not
orally bioavailable.
3. Biological results and discussion
Compounds 10–31 were assayed for their ability to
inhibit the activity of bovine mitochondrial F1F0-ATP
hydrolase in submitochondrial particles in vitro
(Table 1).14 The initial lead for this series (3) selectively
inhibited the hydrolase with an IC50 of 221 nM. A cur-
sory evaluation of electronic variation of the peripheral
aryl moiety ranging from electron-donating to neutral to
electron-withdrawing (compounds 10–17) revealed a
modest impact on activity, with little evidence of a strong
electronic component. Increasing the surface area of the
We report herein the discovery and preliminary struc-
ture–activity relationships (SAR) of a series of small-
molecule, selective inhibitors of mitochondrial F1F0-
ATP hydrolase. These small-molecule inhibitors are
devoid of inhibitory activity against the corresponding
ATP synthase and are of interest for the development of
oral agents for the treatment of ischemic conditions
mediated through this important pathway. An initial
lead against this target (3) was discovered through
directed screening of small molecule libraries using
pharmacophore searching based on an earlier sub-
stituted guanidine lead in this program.12 Substantial
improvements in potency were rapidly achieved through
Scheme 1. (a) Ph(CH2)2C(=O)CO2Et, toluene, reflux, Dean–Stark
trap; (b) 1,2-DCE, TFA, Et3SiH; (c) NaOH/MeOH; (d) LAH, THF;
(e) RS(=O)2Cl, NEt3, CH2Cl2; (f) NaH, DMF, RCH2Br; (g)
imidazolylcarboxaldehyde, Na(OAc)3BH, DCE, HOAc.

L. G. Hamann et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1031–1034
1033
Table 1. Inhibition of bovine mitochondrial F1F0 ATP hydrolase
assay results for compounds 3 and 10–31
The present series of mATPase inhibitors includes a
number of potent and selective inhibitors of the hydro-
lase activity of this enzyme. Lead optimization efforts
successfully achieved a 40-fold improvement in mAT-
Pase potency while maintaining selectivity for inhibition
of the hydrolase versus the synthase activity of this
enzyme.15 Modifications to the imidazole portion of the
initial lead compound helped reduce a potential CYP
liability, with only modest impact on target potency.
Additional structure–activity studies focusing on unex-
plored regions of this scaffold have the potential of
leading to the identification of suitable drug candidates
acting through this novel pathway. Such agents are
expected to restore energy balance to ischemic cells, and
may thereby find utility in the clinical treatment of
ischemic heart disease and its associated complications.
Compda
R
Z
imidazole
F1F0 ATP
hydrolase inhibition
IC50, mMb
3
4-F-Ph
Ph
4-OH-Ph
4-OMe-Ph
2,5-di-Cl-Ph
4-(AcNH)-Ph
4-CN-Ph
2-Cl-4-CN-Ph
3-NO2-Ph
Naphth-1-yl
Thiophen-2-yl
SO2
SO2
SO2
SO2
SO2
SO2
SO2
SO2
SO2
SO2
SO2
5-yl
5-yl
5-yl
5-yl
5-yl
5-yl
5-yl
5-yl
5-yl
5-yl
5-yl
5-yl
5-yl
5-yl
5-yl
5-yl
5-yl
0.221
0.282
0.667
0.077
0.158
2.981
0.255
0.939
0.423
0.338
0.636
1.777
2.935
2.405
0.077
0.008
2.138
2.352
>10.0
9.623
0.151
0.077
0.022
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Benzofurazan-7-yl SO2
Quinolin-8-yl
Bn
CF3
4-t-Bu-Ph
4-t-Bu-Ph
4-t-Bu-Ph
4-t-Bu-Ph
4-F-Ph
SO2
SO2
SO2
SO2
CH2
CH2 4-Me-5-yl
SO2 2-yl
SO2 2-Me-5-yl
SO2 4-Me-5-yl
SO2 4-Me-5-yl
SO2 4-Me-5-yl
References and notes
1. Harris, D. A.; Das, A. M. Biochem. J. 1991, 280, 561.
2. Jennings, R. B.; Murry, C. E.; Steenbergen, C. J.; Reimer,
K. A. Circulation, 1990, 82 (Suppl. 2), II 2.
3. (a) Rouslin, W.; Erickson, J. L. E.; Solaro, R. J. Am. J.
Physiol. 1986, 250, H503. (b) Jennings, R. B.; Reimer,
K. A.; Steenbergen, C. J. J. Mol. Cell. Cardiol. 1991, 23,
1449.
4. (a) Abrahams, J. P.; Leslie, A. G. W.; Lutter, R.; Walker,
J. E. Nature 1994, 370, 621. (b) Tsunoda, S. P.; Rodgers,
A. J.; Aggeler, R.; Wilce, M. C.; Yoshida, M.; Capaldi,
R. A. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 6560. (c)
Stock, D.; Leslie, A. G. W.; Walker, J. E. Science 1999,
286, 1700.
5. Bosetti, F.; Yu, G.; Zucchi, R.; Ronca-Testoni, S.;
Solaini, G. Mol. Cell. Biochem. 2000, 215, 31.
6. Pullmann, M. E.; Monroy, G. C. J. Biol. Chem. 1963, 238,
3762.
7. Abrahams, J. P.; Buchanan, S. K.; van Raaij, M. J.;
Fearnley, I. M.; Leslie, A. G.; Walker, J. E. Proc. Natl.
Acad. Sci. U.S.A. 1996, 93, 9420.
8. (a) Cunningham, C. C.; George, D. T. J. Biol. Chem.
1975, 250, 2036. (b) Walker, J. E.; Lutter, R.; Dupuis, A.;
Runswick, M. J. Biochemistry 1991, 30, 5369.
9. (a) Moyle, J.; Mitchell, P. FEBS Lett. 1975, 56, 55. (b)
Satre, M.; Bof, M.; Vignais, P. V. J. Bacteriol. 1980, 142,
768. (c) van Raaij, M. J.; Abrahams, J. P.; Leslie, A. G.;
Walker, J. E. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 6913.
10. (a) Vigers, G. A.; Ziegler, F. D. Biochem. Biophys. Res.
Commun. 1968, 30, 83. (b) Daggett, S. G.; Tomaszek,
T. A., Jr.; Schuster, S. M. Arch. Biochem. Biophys. 1985,
236, 815. (c) Noumi, T.; Maeda, M.; Futai, M. FEBS
Lett. 1987, 213, 381.
11. Atwal, K. S.; Wang, P.; Rogers, W. L.; Sleph, P.;
Monshizadegan, H.; Ferrara, F. N.; Traeger, S. C.; Green,
D. W.; Grover, G. J. J. Med. Chem. 2004, in press.
12. Atwal, K. S.; Ahmad, S.; Ding, C. Z.; Stein, P. D.; Lloyd,
J.; Hamann, L. G.; Green, D. W.; Ferrara, F. N.; Wang,
P.; Rogers, W. L.; Doweyko, L. M.; Miller, A. V.; Bisaha,
S. N.; Schmidt, J. B.; Li, L.; Yost, K. J.; Lan, H.-J.;
Madsen, C. S. Bioorg. Med. Chem. Lett. 2004, 14,
4-F-Ph
4-t-Bu-Ph
3,4-di-Cl-Ph
a All compounds tested were of greater than 98% purity as determined
by LC/MS and 1H NMR analysis.
b For all compounds, mitochondrial F1F0 ATP synthase IC50>30 mM.
aryl ring at this same position, or changing to hetero-
cyclic aryls (compounds 18–21) likewise failed to show
any improvement in activity. From these initial efforts,
however, it was noted that H-bond donors (11,14) were
particularly ill-tolerated. In contrast, more lipophilic
susbtituents were somewhat favorable (12,13), and
additional lipophilically-substituted analogues were
explored. Direct replacement of the aryl moiety with a
benzyl homologue (22) eroded activity ten-fold, whereas
substitution with a trifluoromethyl group, as in com-
pound 23, improved activity to 77 nM. Potency could
be further improved by an additional order of magni-
tude to 8 nM with a 4-tert-butyl-substituted phenyl
group (24).
Along with the aforementioned improvements in
potency, however, potent inhibition of cytochrome P450
(CYP) isoform CYP2C9 was observed (IC50=38 nM
for compound 24), and efforts were made to identify key
portions of the molecule that might be altered to
diminish or remove this unwanted property while
maintaining mATPase activity. Several changes that
were investigated greatly diminished target potency, and
so were not pursued further, including replacement of
the sulfonyl moiety with CH2 (25), and changing the
point of attachment of the imidazole ring (27). Methyl-
ation of the imidazole ring gave mixed results: methyl-
ation at the 2-position as in compound 28 abolished
mATPase activity, however, methylation at the 4-posi-
tion could be tolerated with modest impact on target
potency while significantly reducing CYP2C9 inhibition
(IC50=3 mM and 2 mM for compounds 30 and 31,
respectively).
preceding paper
j.bmcl.2003.11.077.
in
this
issue.
doi:10.1016/
13. Representative experimental procedures for Scheme 1 are
illustrated for the preparation of compound 29. (a) 2-
Phenethyl-1,2,3,4-tetrahydroquinazoline-2-carboxylic acid
ethyl ester (5). A solution of 2-aminobenzylamine (2.4 g,
20 mmol), and ethyl 2-oxo-4-phenylbutyrate (3.8 mL, 20

1034
L. G. Hamann et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1031–1034
mmol, 1.0 equiv) in 100 mL of toluene was heated at
sequentially. The mixture was allowed to stir at rt for 3 h,
and then saturated K2CO3 solution was added, followed
by solid K2CO3. The suspension was filtered, and the fil-
trate was concentrated under reduced pressure to give the
sulfonamide 9 (ZR=4-F-PhSO2) as a yellow oil (300 mg,
68%). 1H NMR (CDCl3) d 1.79 (m, 2H), 2.72 (m, 2H),
3.00 (m, 1H), 3.30 (dd, 1H), 3.58 (dd, 1H), 4.28 (d, J=15,
1H), 4.45 (d, J=15, 1H), 6.58 (d, J=8.0, 1H), 6.80–7.40
(m, 10H), 7.60 (m, 2H); 13C NMR (CDCl3) 32.32, 34.68,
52.39, 55.61, 115.73, 116.06, 119.70, 121.20, 126.24,
127.15, 128.27, 128.61, 129.69, 129.81, 135.79, 140.65,
147.11. (e) 4-(4-Fluorophenylsulfonyl)-1-(4-methylimidazol-
5-ylmethyl)-2-phenylethyl-2,3,4,5-tetrahydrobenzodiazepine
(29). To a stirred solution of sulfonamide 9 (ZR=4-F-
PhSO2) (80 mg, 0.20 mmol) in 1,2-DCE (1 mL) and acetic
acid (0.5 mL) was added 4-methyl-5-imidazolecarbox-
aldehyde (26 mg, 0.24 mmol, 1.2 equiv), and the mixture
was allowed to stir at rt for 30 min. Sodium triacetoxy-
borohydride (130 mg, 0.60 mmol, 3.0 equiv) was then
added, and additional quantities of the reagents were
added to ensure completion of the reaction. The reaction
mixture was then poured into an EtOAc and NH4OH
solution, and the organic layer was separated, dried, and
concentrated under reduced pressure to give the title
compound 29 as a solid (95 mg, 90%). 1H NMR (CDCl3)
d 1.43 (m, 2H), 2.03 (m, 3H), 2.44 (m, 1H), 2.63 (m, 1H),
3.13 (m, 1H), 3.71 (m, 2H), 4.12 (m, 3H), 4.60 (d, J=10.8,
1H), 6.90–7.40 (m, 12H), 7.81 (m, 2H); MS m/z 505
[M+H]+.
reflux for 18 h using a Dean–Stark apparatus for removal
of water. The solvent then was removed under reduced
pressure to give ester 5 as a brown oil (6 g, 95%). 1H
NMR (CDCl3) d 1.20 (m, 3H), 2.10 (m, 2H), 2.33 (m,
1H), 2.40 (m, 1H), 4.00 (dd, 2H), 4.70 (m, 2H), 6.60 (d,
1H), 6.75 (t, 1H), 6.90 (m, 1H), 7.05 (t, 1H), 7.15–7.35 (m,
5H). (b) 2-Phenylethyl-2,3,4,5-tetrahydrobenzodiazepin-3-
one (7). To a stirred solution of tetrahydroquinazoline 5
(4.5 g, 14.5 mmol) in 1,2-DCE (40 mL) was added TFA
(10 mL) at 0 ꢀC, followed by triethylsilane (4.0 mL, 25
mmol, 1.7 equiv), and the mixture was allowed to warm to
rt over 2 h. The solvent was then removed under reduced
pressure, and the residue was dissolved in MeOH (50 mL).
To this stirred solution at 0 ꢀC was added 1 N NaOH
to adjust the pH to 13. After 18 h, the precipitate was
collected to give benzodiazepinone 7 as a tan solid (2.5 g,
1
65%). H NMR (CDCl3) d 1.95 (m, 1H), 2.30 (m, 1H),
2.79 (m, 2H), 3.55 (d, J=5, 1H), 3.86 (dd, J=7, 16, 1H),
4.32 (dd, J=12, 6.5, 1H), 4.86 (dd, J=16, 7, 1H), 6.50 (d,
J=8, 1H), 6.65 (m, 2H), 6.90 (d, J=7.4, 1H), 7.07 (m,
1H), 7.20 (5H); 13C NMR (CDCl3) 31.34, 31.71, 44.49,
52.94, 116.76, 117.51, 120.45, 125.28, 127.66, 127.68,
127.93, 128.35, 140.32, 144.54. (c) 2-Phenylethyl-2,3,4,5-
tetrahydrobenzodiazepine (8). To a stirred suspension of
lithium aluminum hydride (0.33 g, 8.6 mmol, 1.0 equiv) in
THF (20 mL) at rt under argon was added a solution of
benzodiazepinone 7 (2.3 g, 8.6 mmol) in 20 mL THF
through an addition funnel, and the mixture was allowed
to stir at rt for 18 h before quenching with the addition of
water (1 mL). The resultant suspension was filtered, and
the filtrate was concentrated under reduced pressure to
give benzodiazepine 8 as a yellow oil (2.1 g, 98%), which
was used directly in the next step. MS m/z 253 [M+H]+.
(d) 4-(4-Fluorophenylsulfonyl)-2-phenylethyl-2,3,4,5-tetra-
hydrobenzodiazepine (9). To a stirred solution of benzo-
diazepine 8 (280 mg, 1.1 mmol) in CH2Cl2 was added
NEt3 (0.23 mL, 1.7 mmol, 1.5 equiv) and 4-fluoro-
benzenesulfonyl chloride (220 mg, 1.1 mmol, 1.0 equiv)
14. Assays using the bovine enzyme in submitochondrial
particles were performed as described in ref 11.
15. None of the compounds in the present series exhibited any
measurable inhibition of F1F0 ATP synthase at
concentrations up to 30 mM. While the molecular basis for
the ability of these agents to inhibit this enzyme in the
forward direction only is presently unknown, current
hypotheses center on the potential for the discreet
rotational steps unique to this enzyme to be different in
forward and reverse directions.