N-[1-Aryl-2-(1-imidazolo)ethyl]-guanidine derivatives as potent inhibitors of the bovine mitochondrial F1F0 ATP hydrolase

Article abstract of DOI:10.1016/j.bmcl.2003.11.077

A series of substituted guanidine derivatives were prepared and evaluated as potent and selective inhibitors of mitochondrial F₁F₀ ATP hydrolase. The initial thiourethane derived lead molecules possessed intriguing in vitro pharmacological profiles, though contained moieties considered non-drug-like. Analogue synthesis efforts led to compounds with maintained potency and superior physical properties. Small molecules in this series which potently and selectivity inhibit ATP hydrolase and not ATP synthase may have utility as cardioprotective agents.

Full text of DOI:10.1016/j.bmcl.2003.11.077

Bioorganic & Medicinal Chemistry Letters 14 (2004) 1027–1030
N-[1-Aryl-2-(1-imidazolo)ethyl]-guanidine derivatives as potent
inhibitors of the bovine mitochondrial F1F0 ATP hydrolase
Karnail S. Atwal,a,* Saleem Ahmad,a Charles Z. Ding,a,y Philip D. Stein,a John Lloyd,a
Lawrence G. Hamann,a David W. Green,b,{ Francis N. Ferrara,a Paulina Wang,b
W. Lynn Rogers,b Lidia M. Doweyko,a Arthur V. Miller,a Sharon N. Bisaha,a
Joan B. Schmidt,a Ling Li,a Kenneth J. Yost,a Hsi-Jung Lanc and Cort S. Madsenb,*
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 Metabolism and Pharmacokinetics, Bristol-Myers Squibb, Pharmaceutical Research Institute,
PO Box 5400, Princeton, NJ 08543-5400, USA
Received 21 July 2003; accepted 14 November 2003
Abstract—A series of substituted guanidine derivatives were prepared and evaluated as potent and selective inhibitors of
mitochondrial F1F0 ATP hydrolase. The initial thiourethane derived lead molecules possessed intriguing in vitro pharmacological
profiles, though contained moieties considered non-drug-like. Analogue synthesis efforts led to compounds with maintained
potency and superior physical properties. Small molecules in this series which potently and selectivity inhibit ATP hydrolase and
not ATP synthase may have utility as cardioprotective agents.
# 2003 Elsevier Ltd. All rights reserved.
The mitochondrial F1F0 ATPase is responsible for the
majority of ATP synthesis in mammalian cells. How-
ever, it can also hydrolyze ATP during cellular anoxia
so the direction of this enzymes catalytic activity can
have a profound effect on cellular metabolism.1 During
anoxia this electrochemical gradient collapses, and
mitochondrial F1F0 ATPase switches to its hydrolytic
state.2 This hydrolysis of ATP serves no useful purpose
and depletes the ischemic tissue of ATP leading to cell
death.3 The mitochondrial F1F0 ATP synthase activities
in vesicles from ischemic muscle are substantially (ꢀ50–
80%) less than those of control muscle. A naturally
occurring inhibitor, IF1 protein (IF1), may be bound to
the F1 unit under ischemic conditions to inhibit
the mitochondrial F1F0 ATP hydrolase activity of the
enzyme. However, the activity of IF1 is highly pH
dependent and in severe conditions it can only provide
modest control.4 The conversion of mitochondrial F1F0
ATP synthase to hydrolase is reversible, since addition
of oxygen to the mitochondria of ischemic muscle can
reactivate the mitochondrial F1F0 ATPase and the ATP
levels can return to control.
Several inhibitors of F1F0 ATPase have been described,
including efrapeptin,5 oligomycin,6 aurovertin B,7 and
azide.8 Oligomycin targets F0 and reportedly postpones
cell injury by preventing ATP loss during ischemia.9
Prior to our recent disclosure of benzopyran derivatives
related to compound 1,10 the known inhibitors of mito-
chondrial F1F0 ATPase have all been large molecules,
and consequently not orally bioavailable. Through sub-
sequent directed screening efforts, a second small mole-
cule template was discovered, as represented by
thiourethanes such as 2. In this paper, we describe
structural modifications within this series which led to
cyano- and acyl-guanidine derivatives which posses
potent and selective inhibitory activity against
* Corresponding authors. Tel.: +1-609-818-3550 (L.G.H.);
tel: +1-609-818-5238 (C.S.M.); e-mail: lawrence.hamann@bms.com;
cort.madsen@bms.com
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.2003.11.077

1028
K. S. Atwal et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1027–1030
mitochondrial F1F0 ATP hydrolase without inhibition
of the synthase. A select potent compound was found to
be orally active in rats.
The guanidine derivatives were generally prepared
according to Scheme 1. The commercially available
substituted chloroacetophenones 3 were reacted with
three molar equivalents of imidazole in acetonitrile to
give 1-imidazoloacetophenones 4. Reduction with
sodium borohydride in ethanol provided alcohol 5,
which readily reacted with commercially available N-
arylisothiocyanates to give thiocarbamates (6a–e and 7).
The carbamate analogues 8 were prepared from
benzylic alcohols 5 and 2,4-dichlorophenylisocyanate.
Conversion of alcohols 5 to amines 9 was carried out in
two steps: reaction of 5 with diphenylphosphorylazide
(DPPA) in presence of diazabicycloundecane (DBU) in
THF,11 and subsequent reduction of the resultant azide
with triphenylphosphine in aqueous THF. Thioureas 10
were prepared by reacting amines 9 with appropriately
substituted phenylthioisocyanates. Coupling of amines
9 with the sodium salts of N-cyano-N0-arylthioureas,
prepared by addition of sodium cyanamide to aryl-
thioisocyanates, in the presence of N,N-dimethyamino-
ethyl-N0-propylcarbodiimide in DMF gave the desired
cyanoguanidines (11a–f).12
Scheme 1. General synthesis of thiocarbamates and cyanoguanidines: (a)
imidazole, CH3CN; (b) NaBH4, EtOH; (c) NaH, DMF, 0ꢁC, R-NCS or
RNCO; (d) DPPA, DBU, THF; then Ph3P, THF, H2O; (e) aryliso-
thiocyanate; (f) N,N-dimethylaminoethyl-N0-propylcarbodiimide, DMF.
For the construction of 2,5-bis-trifluorophenyl sub-
stituted analogues, the required bromoacetophenone 14
was prepared by a two-step sequence (Scheme 2). The
commercially available benzoyl chloride 12 was converted
to acetophenone 13 by treatment with methylmagnesium
bromide in the presence of tributylphosphine,13 followed
by the bromination of the resulting acetophenone 13 to
give bromoacetophenone 14. Treatment of 14 with
imidazole gave the imidazolo-acetophenone 15, which
could then be carried on to the final product 11f using
the sequence described in Scheme 1.
Scheme 2. (a) MeMgBr, Bu3P; (b) Br2, HOAc; (c) imidazole, CH3CN.
The acylguanidine derivatives 19 were prepared in a
single step from unsubstituted guanidines 17 by
acylation with carboxylic acids 18 in the presence of
carbonyl diimidazole (Scheme 3). Guanidines 17 were
generated from the thioureas 16 by treatment with
methanolic ammonia (6 N) in the presence of HgO.
Since the mitochondrial F1F0 ATPase is ubiquitous,
several sources (rat, porcine, bovine, human) of the
enzyme were evaluated. Because of its low cost and
availability, the bovine enzyme was chosen for screen-
ing. Based on minimum structural requirements for
inhibitors in our previously disclosed series based on
analogues of compound 1,10 we undertook targeted
screening of compounds with structural similarity.
Upon doing so, we discovered that thiourethane
Scheme 3. (a) NH3, MeOH, HgO; (b) RCO2H (18), CDI.

K. S. Atwal et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1027–1030
Table 1. Inhibition of bovine mitochondrial F1F0 ATP hydrolase assay results for 6–20a
1029
Compd
R1
R2
X
Y
F1F0 ATP hydrolase inhibition IC50 (mM)b
6a
6b
6c
6d
6e
7
8
10
11a
11b
11c
11d
11e
11f
19a
19b
19c
19d
20ac
20bc
2,4-Cl2
2,4-Cl2
2,4-Cl2
2,4-Cl2
2,4-Cl2
2,4-Me2
2,4-Cl2
2,4-Cl2
2,4-Cl2
2,4-Cl2
2,4-Cl2
2,4-Cl2
2,4-Cl2
2,5-(CF3)2
2,4-Cl2
2,4-Cl2
2,4-Cl2
2,4-Cl2
2,4-Cl2
2,4-Cl2
4-Cl
H
4-Me
2,4-Cl2
2,4-Me2
4-Cl
2,4-Cl2
2,4-Cl2
2,4-Cl2
4-Cl
2-Cl
2,3-Cl2
3-Cl
O
O
O
O
O
O
O
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
S
S
S
S
0.43ꢂ0.16
3.60ꢂ1.10
0.66ꢂ0.12
0.030ꢂ0.013
0.23
S
S
O
S
NCN
NCN
NCN
NCN
NCN
NCN
31.4
6.77
2.41
0.60ꢂ0.16
8.8
2.23
2.49ꢂ0.72
9.17
2,4-Cl2
4-Cl
4-Cl
4-Cl
4-Cl
0.71ꢂ0.34
0.033ꢂ0.02
0.28
CO(3-CN-Ph)
CO(4-CN-Ph)
CO(4-Cl-Ph)
COEt
CO(3-CN-Ph)
CO(3-CN-Ph)
0.082ꢂ0.03
2.27
0.018ꢂ0.016
>100
4-Cl
4-Cl
a The assays for the mitochondrial F1F0 ATP hydrolase and synthase are described.14
b For all compounds, mitochondrial F1F0 ATP synthase IC50>100 mM.
c Compounds 20a and 20b are the independent resolved enantiomers of compound 19a, though the absolute stereochemistry is unknown.
compound 6a is a potent inhibitor of mitochondrial
F1F0 ATP hydrolase (IC50=0.43 mM), though does not
inhibit the synthase (IC50>300 mM) (Table 1). In the
aniline region of 6a, it was discovered that chloro sub-
stitution is important, as the protio analogue 6b is about
10-fold less active. This loss of activity was not fully
restored in its 4-methyl analogue 6c, possibly indicating
the requirement of an electron withdrawing group. The
dichloro compound 6d was an order of magnitude more
potent (IC50=0.03 mM), while maintaining selectivity
for inhibition of the ATP hydrolase. The corresponding
dimethyl analogue 6e is approximately 10-fold less
active (IC50=0.23 mM) than the corresponding dihalo
compound 6d, and dimethyl analogue 7 is nearly devoid
of inhibitory activity, suggesting that halogen substitu-
tion is favorable in both the 1-aryl-2-imidazoloethyl and
aniline regions of the molecule.
chemotype optimization, as this functionality is known
to exist in more drug-like molecules and can be readily
prepared.16
The structure–activity relationships of cyanoguanidine
derivatives are shown in Table 1 (11a–f). Mono-chloro
analogues 11b, 11c, and 11e all retain a moderate level
of activity, though all are inferior to 2,4-dichloro
analogue 11a. Unlike the thiourethane series of inhibi-
tors, 2,4-dichloro substitution on the 1-aryl-2-imidazo-
loethyl moiety does not appear to be mandatory for
activity, as evidenced by bis-trifluoromethyl analogue
11f (IC50=0.71 mM). The favorable activity of this
compound within this series may also serve to support
the hypothesis that the electron-withdrawing character
of the halogenated aryls contained in earlier analogues
is an important contributing factor for inhibitory activ-
ity. As further modifications within the cyanoguanidine
series failed to achieve additional gains in inhibitory
potency beyond that of 11a,f (IC50=0.6–0.7 mM), the
strategy chosen involved making changes directly to
the guanidine core.
Although thiourethane 6d is a potent inhibitor in vitro,
this compound is not considered to be a drug like
molecule. It was observed to exhibit chemical instability
in aqueous buffer by reversion to isothiocyanate and
aniline, and carries toxic liabilities associated with the
thiourethane functionality.15 We therefore turned our
attention to the exploration of stable functionalities to
replace the thiourethane moiety. Replacement of the
thiourethane group in 6d with a thiourea gave 10, which
is 50-fold less potent than 6d. The carbamate derivative
8 and the cyanoguanidine derivative 11a were also less
potent inhibitors. Despite the reduced in vitro potency
associated with this initial cyanoguanidine containing
analogue 11a, we focused on this moiety for further
The most expedient functionality to investigate was that
of the acylguanidine, which can be readily prepared
from the unsubstituted guanidines. The 3-cyano-
benzoylguanidine containing analogue 19a offered at
least 100-fold improvement in potency relative to the
corresponding cyanoguanidine compound 11b, exhibit-
ing an IC50 value of 0.033 mM. The 4-cyano analogue
19b was about 10-fold less potent than its 3-cyano
derivative 19a. The 4-chloro analogue 19c also retained

1030
K. S. Atwal et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1027–1030
significant potency. While most of the benzoyl derivatives
maintained excellent potency, the aliphatic compounds
such as 19d showed reduced activity. Because all of the
inhibitors evaluated in this series were racemic mixtures,
we then chose to separate the enantiomers of the most
potent compound, 3-cyano derivative 19a, using chiral
HPLC (Chiracel AD1 column; isopropanol/hexanes,
0.1% triethylamine). All of the potency was retained by
single enantiomer 20a (IC50=0.018 mM) of the racemic
19a (the other enantiomer, 20b, was devoid of any
ATPase inhibitory activity), though the absolute
stereochemistry of 20a was not determined.
3. (a) Abrahams, J. P.; Leslie, A. G. W.; Lutter, R.; Walker,
J. 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.
4. Cabezon, E.; Butler, P. J. G.; Runswick, M. J.; Walker,
J. E. J. Biol. Chem. 2001, 275, 25460.
5. 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.
6. (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.
7. (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.
8. (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.
9. Harris, D. A.; Das, A. M. Biochem. J. 1991, 280, 561.
10. Atwal, K. S.; Wang, P.; Lynn Rogers, W.; Sleph, P.;
Monshizadegan, H.; Ferrara, F. N.; Traeger, S. C.; Green
D. J.; Grover, G. J. J. Med. Chem. 2004, in press.
11. Thompson, A. S.; Humphrey, G. R.; DeMarco, A. M.;
Mathre, D. J.; Grabowski, E. J. J. J. Org. Chem. 1993, 58,
5886.
We next evaluated the oral bioavailability of the most
potent compound (20a) of this series. The compound
was dosed to rats (n=2) in polyethyleneglycol:water:
ethanol (1:1:1) and the plasma samples were analyzed
by LC-MS/MS. The compound has an oral bioavail-
ability of 47% and an ivhalf life of 2.1 h. The volume of
distribution (Vss) was 2.37 L/kg and the Cmax was 21
mM. Thus we were able to identify an orally active
compound with good pharmacokinetic properties which
might be suitable for evaluation of pharmacological
activity in animal models of ischemia.
Through modifications to peripheral aryl groups and
the core functional group in a series of guanidine based
analogues, we have identified a potent and selective
inhibitor (20a) of mitochondrial F1F0 hydrolase. The
compound does not inhibit the synthase component of
the enzyme. It has good oral bioavailability and an
excellent half life in rats. This is though to represent a
suitable tool molecule for investigating the role of this
enzyme in disease models. Interestingly, all of the com-
pounds studied in this series showed complete absence
of inhibition of the synthase component of this enzyme,
although at present, the reasons for this selectivity are
not clear. Since the enzymatic reaction of the mito-
chondrial F1F0 ATP hydrolase proceeds in discrete
rotational steps, the forward and the backward
reactions may each follow different steps, and therefore
the determinants of selectivity may be manifested in the
differences in these discreet steps.
12. Atwal, K. S.; Ahmed, S. Z.; O’Reilly, B. C. Tetrahedron
Lett. 1989, 30, 7313.
13. Maeda, H.; Okamoto, J.; Ohmori, H. Tetrahedron Lett.
1996, 37, 5381.
14. The hydrolysis of ATP in submitochondrial particles
(SMP) was measured using a pyruvate kinase:lactate
dehydrogenase coupled enzyme system in microtiter plates
by monitoring the decrease in absorbance at 340 nM. The
assay was done in 50 mM TrisAcetate pH 7.5 containing 2
mM MgCl2, 0.5 mM ATP, 2 mM phosphoenolpyruvate
ATP. The SMP were added and briefly preincubated at
room temperature before measuring NADH oxidation.
The change in absorbance at 340 nm was converted into
mmoles ADPminꢃ1 using the extinction coefficient of
NADH (6220 Mꢃ1 cmꢃ1). The ATP synthase activity was
measured by monitoring the increase in the absorbance at
340 nm using a hexokinase:glucose-6-phosphate dehy-
drogenase coupled assay. Further assay details can be
found in supplementary information for ref 12.
15. Neil, R. A.; Halpert, J. Annu. Rev. Pharmacol. Toxicol.
1982, 22, 321.
References and notes
1. Jennings, R. B.; Murry, C. E.; Steenbergen, C. J.; Reimer,
K. A. Circulation 1990, 82 (Suppl. 2), II2.
2. (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.
16. (a) Atwal, K. S.; Grover, G. J.; Ahmed, S. Z.; Sleph, P. G.;
Dzwonczyk, S.; Baird, A. J.; Normandin, D. E. J. Med.
Chem. 1995, 38, 3236. (b) Ding, C. Z.; Rovnyak, G. C.;
Misra, R. N.; Grover, G. J.; Miller, A. V.; Ahmed, S. Z.;
Kelly, Y.; Normandin, D. E.; Sleph, P. G.; Atwal, K. S. J.
Med. Chem. 1999, 42, 3711.