Effects of new ubiquinone-imidazo[2,1-b]thiazoles on mitochondrial complex I (NADH-ubiquinone reductase) and on mitochondrial permeability transition pore

Article abstract of DOI:10.1016/j.bmc.2004.08.012

In this work we describe the synthesis of a series of imidazo[2,1-b] thiazoles and 2,3-dihydroimidazo[2,1-b]thiazoles connected by means of a methylene bridge to CoQ₀. These compounds were tested as specific inhibitors of the NADH:ubiquinone reductase activity in mitochondrial membranes. The imidazothiazole system when bound to the quinone ring in place of the isoprenoid lateral side chain, may increase the inhibitory effect (with an IC₅₀ for NADH-Q₁ activity ranging between 0.25 and 0.96 μM) whereas the benzoquinone moiety seems to lose the capability to accept electrons from complex I as indicated by very low maximal velocity elicited by the compounds tested. Moreover the low rotenone sensitivity for almost all of these compounds suggests that they are only partially able to interact with the physiological ubiquinone-reduction site. The compounds were investigated for the capability of increasing the permeability transition of the inner mitochondrial membrane in isolated mitochondria. Unlike CoQ₀, which is considered a mitochondrial membrane permeability transition inhibitor, the new compounds were inducers.

Full text of DOI:10.1016/j.bmc.2004.08.012

Bioorganic & Medicinal Chemistry 12 (2004) 5525–5532
Effects of new ubiquinone-imidazo[2,1-b]thiazoles on
mitochondrial complex I (NADH-ubiquinone reductase) and on
mitochondrial permeability transition pore
Aldo Andreani,a,* Massimiliano Granaiola,a Alberto Leoni,a Alessandra Locatelli,a
Rita Morigi,a Mirella Rambaldi,a Maurizio Recanatini,a Giorgio Lenaz,b
Romana Fatob and Christian Bergaminib
aDipartimento di Scienze Farmaceutiche, Universita’ di Bologna, Via Belmeloro 6, 40126 Bologna, Italy
bDipartimento di Biochimica ‘G. Moruzzi’, Universita’ di Bologna, Via Irnerio 48, 40126 Bologna, Italy
Received 25 May 2004; accepted 10 August 2004
Available online 15 September 2004
Abstract—In this work we describe the synthesis of a series of imidazo[2,1-b]thiazoles and 2,3-dihydroimidazo[2,1-b]thiazoles con-
nected by means of a methylene bridge to CoQ0. These compounds were tested as specific inhibitors of the NADH:ubiquinone
reductase activity in mitochondrial membranes. The imidazothiazole system when bound to the quinone ring in place of the isopre-
noid lateral side chain, may increase the inhibitory effect (with an IC50 for NADH-Q1 activity ranging between 0.25 and 0.96lM)
whereas the benzoquinone moiety seems to lose the capability to accept electrons from complex I as indicated by very low maximal
velocity elicited by the compounds tested. Moreover the low rotenone sensitivity for almost all of these compounds suggests that
they are only partially able to interact with the physiological ubiquinone-reduction site. The compounds were investigated for
the capability of increasing the permeability transition of the inner mitochondrial membrane in isolated mitochondria. Unlike
CoQ0, which is considered a mitochondrial membrane permeability transition inhibitor, the new compounds were inducers.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
families of compounds are known to inhibit complex I,10
most of which are commonly used as pesticides, includ-
ing insecticides, miticides, piscicides, and ascaricides.
Mitochondrial NADH-Coenzyme Q (CoQ, ubiquinone)
oxidoreductase (EC 1.6.99.3, complex I) is the most
complicated and the least understood of the respiratory
chain complexes. Despite recent progress in structural
studies,1,2 there is no structural information about com-
plex I comparable with the crystallographic data availa-
ble for other respiratory complexes. This lack of
information is principally due to the complexity of
the system: the mammalian enzyme consists of up to
46 different subunits reaching a total mass of almost
1000kDa.3 Nevertheless, there seems to be a general
agreement on the presence of ubisemiquinones as inter-
mediates for the electron transfer to the bulky quinone
anÀd on the proton pumping stoichiometry of 4H+/
2e .4–9 Another typical feature of complex I is related
to the high number of its inhibitors: over sixty different
The discovery that mitochondrial DNA mutations are at
the basis of a number of human pathologies has opened
a new and extremely active chapter about mitochondrial
research.11,12 A lesion in a structural gene, such as that
due to a point mutation, would interfere with the func-
tion of the polypeptide encoded by that gene, leading to
decreased enzyme activity. Indeed this is often found in
such kind of cytopathies, such as a form of LeberÕs
hereditary optic neuropathy (LHON) where in spite of
normal complex I activity, the flux over the entire respi-
ratory chain is significantly decreased.13 It is predictable
that the quinone-binding sites are involved in a defective
energy conservation present in this form of the disease,
since they take part in the proton translocation activity
of complex I.14 Moreover, in the last decades, the inter-
est in this NADH-ubiquinone reductase is increasing,
due to its possible involvement in the pathogenesis
of human neurodegenerative diseases15,16 such as
*
Corresponding author. Tel.: +39 051 2099714; fax: +39 051
2099734; e-mail: aldo.andreani@unibo.it
0968-0896/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2004.08.012

5526
A. Andreani et al. / Bioorg. Med. Chem. 12 (2004) 5525–5532
O
OH
y
Na2S2O4
O
O
O
O
N
x
R
S
N
O
OH
3a-i
1
6
O
O
O
O
POCl3/DMF
O
.
a) (C2H5)2O BF3
b) Ag2O
OH
y
N
x
R
NaBH4
y
y
N
N
S
N
x
x
R
N
R
S
S
N
2a-i
5a-i
4a-i
Scheme 1. x, y, R see Table 1.
AlzheimerÕs and ParkinsonÕs diseases, in diabetes and in
aging.17 On the other hand rotenoids acting as NADH-
ubiquinone oxidoreductase inhibitors induce ornithine
decarboxylase activity and are evaluated as candidate
cancer chemopreventive agents.18,19
in a non-competitive way with the ubiquinone substrate
and interacted with a site, which is mutually exclusive
with that of rotenone but non-exclusive with that of
piericidin and several inhibitors of NADH dehydroge-
nase. Founded on those data, we have synthesized a
new series of different imidazothiazoles (2a–i, Scheme
1) linked to a CoQ0 moiety, in order to evaluate the ef-
fect on complex I and on PTP of the hybrid ubiquinone
and to elucidate the ambiguous quinone specificity in the
interaction with the enzyme.
The identification and characterization of specific qui-
nones involved in the inhibition of complex I, represent
a significant advance to understand the functional mech-
anism of the complex, to define the structural and func-
tional features of the ubiquinone-reduction site of
bovine heart mitochondrial complex I20–25 and also con-
tributes to develop drugs specifically acting on mito-
chondria when complex I deficits are implicated in
human diseases. Some quinones are already known as
complex I inhibitors. The inhibitory action of CoQ2
and other short chain isoprenoid homologs (but not
CoQ1) is well documented in beef heart mitochon-
dria.10,21 In bovine heart submitochondrial particles
(SMP) only CoQ1 and 6-decyl-Q0 (DB) were found to
elicit rates comparable to those calculated for reduction
of endogenous CoQ10 and are used as electron accep-
tors substitutes of the physiological ubiquinones. Also
CoQ0 and the tetramethylbenzoquinone analogue, duro-
quinone (DQ), are electron acceptors.10,21,26 All these
substitutes of the ubiquinones interact with the physio-
logical site(s), in place of the endogenous CoQ and
their reduction is almost completely inhibited by rote-
none. On the other hand, the quinone 2,3-dimethoxy-
5-methyl-1,4-benzoquinone (CoQ0 or ubiquinone0, 1,
Scheme 1) is known to be a potent inhibitor of the so
called permeability transition pore (PTP),27,28 a cyclosp-
orin A-sensitive channel of the inner mitochondrial
membrane.29 Other quinones show different behaviors
being inducers or inhibitors.28 The PTP has been studied
as a target for mitochondria dysfunction in vivo and
although the physiological function of the pore is ob-
scure, more recently the PTP has been considered as a
potential mediator in the mechanism of programmed
cell death.30,31 Furthermore a relation between PTP
and complex I activity has been suggested.32
2. Chemistry
Compounds 3–5 were prepared according to the litera-
ture,36–42 except 3i, 4i, and 5h–i, which were synthesized
according to the Scheme 1; details for the synthesis are
described under Section 5. Compound 3i was synthe-
sized by treating 2-amino-5-chlorothiazole hydrochlo-
ride with chloroacetone. The aldehyde 4i was prepared
by means of the Vilsmeier reaction on the imidazothiaz-
ole 3i and the hydroxymethyl derivatives 5h–i were pre-
pared by reduction of the aldehydes with sodium
borohydride.
2,3-Dimethoxy-5-methyl-1,4-benzoqui-
none 1 was converted to 2,3-dimethoxy-5-methylhydro-
quinone 6 by a conventional reductive step employing
sodium dithionite as the reducing agent. In the subse-
quent step, compound 6 was reacted with the appropri-
ate substituted 5-hydroxymethylimidazo[2,1-b]thiazole
5a–i. The condensation was carried out in dioxane at
room temperature and under nitrogen in order to pre-
vent oxidation of the hydroquinone to the correspond-
ing benzoquinone. Boron trifluoride etherate was used
as the acidic condensing agent.43 The hydroquinones
were not characterized but used as such for a mild
oxidation with silver(I)oxide, which afforded the substi-
tuted 6-imidazo[2,1-b]thiazolylmethyl-2,3-dimethoxy-5-
methyl-benzoquinones 2a–i. All the new derivatives
2a–i are reported in Tables 1 and 2.
3. Biological results
In a previous study33–35 it was shown that thienylimid-
azo[2,1-b]thiazoles were able to inhibit complex I acting
The biological activity of compounds 2a–i on mito-
chondrial NADH dehydrogenase activity has been

A. Andreani et al. / Bioorg. Med. Chem. 12 (2004) 5525–5532
5527
Table 1. Compounds 2a–i
Compounds
x–y
R
Starting 4/5 ref.
Formula
15H13ClN2O4S
Mw
Mp, ꢁC
2a
2b
2c
2d
2e
2f
CH@CH
CH@CH
CH@CH
CH2–CH2
CH2–CH2
H3CC@CH
H3CC@CH
ClC@CH
ClC@CH
Cl
36/36
37/42
C
352.8
332.4
394.4
354.8
334.4
366.8
346.4
428.9
366.8
135–138 (a)
145–150(b)
135–140(b)
156–157 (b)
>300 (b)
CH3
C6H5
Cl
C16H16N2O4S
C21H18N2O4S
C15H15ClN2O4S
C16H18N2O4S
C16H15ClN2O4S
C17H18N2O4S
C21H17ClN2O4S
C16H15ClN2O4S
37/42
38/42
CH3
Cl
38/42
39/39
125–130(c)
115–117 (b)
70–72 (d)
2g
2h
2i
CH3
C6H5
CH3
40/39
41/exp.
exp./exp.
154–156 (b)
Purified by: crystallisation with (a) ethyl ether/esane or by chromatography with the following eluents: (b) acetone/petroleum ether 60–80ꢁC; (c) ethyl
ether; (d) ethyl acetate/petroleum ether 60–80ꢁC.
Table 2. IR and 1H NMR of compounds 2a–i
Compounds
IR: mmax, cmÀ1
1H NMR:a d, ppm in DMSO-d6
2a
1650, 1605, 1265, 715
2.06 (3H, s, CH3), 3.86 (3H, s, OCH3), 3.89 (3H, s, OCH3), 4.01 (2H, s,
CH2), 7.39 (1H, d, th, J = 4.5), 7.91 (1H, d, th, J = 4.5)
3), 2.13 (3H, s, CH3), 3.85 (3H, s, OCH3), 3.87 (3H, s,
2b
2c
2d
2e
2f
1655, 1635,11(236H5,, 7s7, 0C2H.0
1640, 1605, 1260, 770
1660, 1635, 1600, 1265
1675, 1645, 1600, 1100
1660, 1640, 1600, 1260
1690, 1655, 1610, 1265
1640, 1605, 1260, 765
1640, 1610, 1265, 1025
OCH3), 3.95 (2H, s, CH2), 7.16 (1H, d, th, J = 4.5), 7.66 (1H, d, th, J = 4.5)
1.67 (3H, s, CH3), 3.81 (6H, s, OCH3), 4.20(2H, s, CH 2), 7.29 (1H, d, th,
J = 4.6), 7.36 (3H, m, ar), 7.49 (2H, m, ar), 7.81 (1H, d, th, J = 4.6)
2.00 (3H, s, CH3), 3.72 (2H, s, CH2), 3.84 (2H, t, thn, J = 7.3), 3.88 (3H, s,
OCH3), 3.89 (3H, s, OCH3), 4.12 (2H, t, thn, J = 7.3)
1.97 (3H, s, CH3), 2.01 (3H, s, CH3), 3.67 (2H, s, CH2), 3.79 (2H, t, thn, J = 7.2),
3.87 (3H, s, OCH3), 3.88 (3H, s, OCH3), 3.97 (2H, t, thn, J = 7.2)
2.03 (3H, s, CH3), 2.42 (3H, d, CH3, J = 1.4), 3.85 (3H, s, OCH3), 3.88 (3H, s,
OCH3), 3.94 (2H, s, CH2), 7.67 (1H, q, th, J = 1.4)
2g
2h
2i
2.04 (3H, s, CH3), 2.19 (3H, s, CH3), 2.41 (3H, s, CH3), 3.85 (3H, s, OCH3),
3.88 (3H, s, OCH3), 3.94 (2H, s, CH2), 7.56 (1H, s, th)
1.70 (3H, s, CH3), 3.78 (3H, s, OCH3), 3.79 (3H, s, OCH3), 4.17 (2H, s, CH2),
7.33 (3H, m, ar), 7.43 (2H, m, ar), 8.26 (1H, s, th)
2.03 (3H, s, CH3), 2.11 (3H, s, CH3), 3.86 (3H, s, OCH3), 3.89 (3H, s, OCH3),
3.95 (2H, s, CH2), 8.11 (1H, s, th)
a Abbreviations: th = thiazole, thn = thiazoline, ar = aromatic.
initially tested to evaluate if they were electron accep-
tors. The electron acceptors most commonly employed
as substitutes of the physiological ubiquinones are short
chain CoQ homologues (the quinone most used as
acceptor is CoQ1) and analogues (such as decyl-ubiqui-
none). The experiments were carried out using submit-
ochondrial particles from beef heart. By the kinetic
analysis of the redox reactions, two parameters were
determined, Km and Vmax. As shown in Table 3 it is pos-
sible to divide the compounds in two groups.
that all compounds tested show a very low rotenone sen-
sitivity, suggesting that they are not able to interact with
the physiological CoQ reduction site. In Table 3 is also
reported the values calculated for the ratio kcat/Km that
is taken as index for the complex I specificity versus the
2a–i derivatives. kcat was calculated dividing the Vmax
values by the complex I concentration calculated by
the ferricyanide assay as indicated in Fato et al.26 From
the results listed in Table 3 it appears that complex I has
a specificity for almost all the derivatives tested that is
more than one order of magnitude lower with respect
to that one elicited for CoQ1. Compound 2h is capable
to induce the highest kcat/Km value suggesting that even
if it is a very poor substrate it is able to interact with the
physiological CoQ reduction site as confirmed also by
the high rotenone sensitivity of the NADH-2h reductase
activity. It should be remembered that in the compound
2h the positions 2 and 6 of the imidazothiazole group are
substituted by a chlorine and a phenyl group, respec-
tively. These substitutions make the compound very
hydrophobic indicating that the hydrophobicity is an
important parameter for ensuring the interaction with
complex I. In general, the benzoquinone moiety when
condensed with the imidazothiazole system seems to lose
The first group consists of 2a, 2b, 2d, 2f, 2i, which can-
not be considered electron acceptors, because they have
a low affinity for complex I (Km range from 220to
417lM, whereas Q1 displays Km = 24lM), even though
they have the highest values of Vmax (from 0.134 to
0.255lmolmgÀ1 minÀ1): these values are however low
compared with that elicited by the best acceptor CoQ1
(Vmax = 0.66lmolmgÀ1 minÀ1). The second group con-
sists of 2c, 2e, 2g, and 2h, which are weak electron accep-
tors as shown by the low values of Vmax (from 0.026 to
0.0805lmolmgÀ1 minÀ1), but they have a greater affinity
for the complex, indeed they show a low Km (ranging
from 16.8 to 43.4lM). Moreover it could be considered

5528
A. Andreani et al. / Bioorg. Med. Chem. 12 (2004) 5525–5532
Table 3. Summary of the kinetic parameters of 2a–i in bovine complex Ia
Compounds
Km (lM)
Vmax (lmolmgÀ1 minÀ1
)
kcat (sÀ1
)
kcat/Km (MÀ1 sÀ1
)
Rotenone sensitivity (%)b
logPc
2a
2b
2c
2d
2e
2f
.129764.100
359.81
41.72
62.82
0.255
0.074
2.29
· 105
2.27 · 105
5.72 · 105
1.53 · 105
1.92 · 105
1.83 · 105
3.35 · 105
15.30 · 105
1.95 · 105
.1381.88 · 105
88.10 · 105
501.973
14
43
81.73
23.88
63.46
8.33
1.691
3.040
1.210
0.683
414.35
43.38
0.198
0.026
22
45
417.01
33.48
16.83
0.238
0.035
76.28
11.22
25.80
42.95
57.69
211.54
02.742
25
2g
2h
2i
2.190
2.405
0.080
0.134
803.754
5.6
900
220.59
65.00
24.00
Q0
Q1
0.180
0.660
902.152
a The Km and Vmax values were obtained from at least three independent experiments. The concentration of SMP protein used ranged between 20and
40lg/mL. kcat values were obtained by division of the Vmax by the concentration of complex I of 0.052nmolmgÀ1 calculated by ferricyanide assay.26
b % of NADH-2a–i activity inhibited by 2lM rotenone.
c CLOGP; 4.3 ed. BioByte Corp. Claremont, CA.
the capability of electron acceptor from complex I as
indicated by the very low maximal velocity elicited by
the compounds tested.
Since not all the compounds investigated are electron
acceptors, we tested if they inhibited the NADH-ubiqui-
none reductase activity of complex I using submito-
chondrial particles. We assayed the NADH-O2 activity
for 2a–i by measuring their inhibition of the overall elec-
tronic flux determined by the rates of NADH oxidation
(NADH-O2 oxidation activity) and the NADH-CoQ1
activity (NADH-Coenzyme Q reductase) using Q1 as
electron acceptor while blocking complex III (ubiquinol
cytochrome c reductase) with antimycin and complex IV
with KCN (see Table 4). Compounds 2c, 2e, and 2h were
not inhibitors. The presence of an aromatic ring in posi-
tion 6 of the imidazothiazole moiety (2c and 2h) or the
saturation in 2,3 position (2e) brings to lack of the inhib-
itory effect. On the contrary 2a, 2b, 2d, 2f, 2g, and 2i in-
hibit the NADH dehydrogenase but with different
features (see Table 4).
Natural CoQ homologues are hydrophobic molecules
and their membrane/water partitioning has to be taken
into account in any consideration concerning their spe-
cificity and kinetics of interaction. For this reason, we
calculated the octanol/water partition coefficient of the
derivatives 2a–i (see Table 3). The calculated logP val-
ues were estimated with the CLOGP program (CLOGP;
4.3 ed. BioByte Corp. Claremont, CA). There is almost
no relationship between the affinity of these compounds
for complex I (Km) and their partition coefficients (cor-
relation coefficient r2 = 0.18). This result brings to con-
clude that even if hydrophobicity is a property
essential to influence the affinity of 2a–i for the complex,
it may not be the only one involved. By contrast if we
calculate the correlation coefficient excluding the most
deviating compounds (2e and 2g) the r2 value is higher:
0.79. This observation strengthens the idea that both
hydrophobicity and some other structural parameters
have to be considered to explain the interaction of the
complete series of products with complex I. Moreover
the low rotenone sensitivity of complex I activity for al-
most all these compounds (see Table 3) suggests that
they are only partially able to interact with the physio-
logical CoQ-reduction site.
Compounds 2b, 2d, 2f, and 2g appear to be inhibitors of
complex I only, since they exert similar extent of inhibi-
tion on NADH-CoQ1 and NADH-O2 activities. Com-
pound 2a has an IC50 for NADH-O2 (0.24 0.07lM)
lower with respect to IC50 for NADH-CoQ1
(0.96 0.17lM); for this derivative a titration of the
inhibitory effects on the reductase activity showed that
it acts prevalently as competitive inhibitor of CoQ1.
Compound 2i inhibits only NADH-O2 activity suggest-
ing that its preferential target is one of the other redox
Table 4. NADH-CoQ1 and NADH-O2 activity in submitochondrial particles (SMP) of beef heart mitochondria in the presence of compounds 2a–i
Compounds
IC50 (lM)a for NADH-CoQ1 activity
IC50 (lM)a for NADH-O2 activity
Concentration inducing 50% swelling
(PTP-activating)
2a
2b
2c
2d
2e
2f
0.960 0.170
0.280 0.200
No inhibition
0.650 0.210
No inhibition
0.250 0.040
0.770 0.090
No inhibition
No inhibition
––
0.240 0.070
0.570 0.260
No inhibition
0.570 0.040
No inhibition
0.210 0.001
1.420 0.530
No inhibition
0.680 0.220
––
2.83lM
4.40lM
18.7lM
4.50lM
42.5lM
5.70lM
4.60lM
3.60lM
2.80lM
PTP-inhibitingb
2g
2h
2i
Q0
a The IC50 values were averaged from three to six independent experiments.
b See Ref. 28.

A. Andreani et al. / Bioorg. Med. Chem. 12 (2004) 5525–5532
5529
reasons we tested our derivatives, which have an imida-
zothiazole system in place of an isoprenoid side chain,
on permeability transition of isolated rat liver mitochon-
dria. They turned out to be PTP openers since they were
able to induce mitochondrial swelling independently
from their inhibitory effect on complex I. As shown in
Table 4 the tested derivatives induced 50% mitochond-
rial swelling at 2.83–5.70lM with the exception of 2c
and 2e, which present lower capacity to induce swelling
(50% swelling at 18.7 and 42.5lM, respectively).
In order to explain the remarkably different biological
behaviors displayed by compounds 2a–i, we undertook
some studies on the geometrical conformations and elec-
trostatic properties of the molecules. In particular, a
conformational analysis was carried out on all com-
pounds and, for each of the low energy conformers,
we calculated both the energy of their frontier orbitals
(HOMO and LUMO) and the two dihedral angles
formed by the planes containing the two cyclic moieties
(the quinone ring and the imidazothiazole system).
From the results obtained (data not shown) no apparent
relationship emerged between the calculated stereoelec-
tronic parameters and the biological profile of com-
pounds 2a–i as determined by the above reported
assays. This suggests that compounds interact with dif-
ferent binding subsites on the complex target system
or, alternatively, that a seemingly unspecific effect (dri-
ven at least in part by the hydrophobicity of the com-
pounds) is at the basis of the biological response.
Figure 1. Dixon plot of NADH-CoQ1 activity in SMP at varying
concentrations. Up: compound 2a: the straight line refers to the
NADH-CoQ1 activity in presence of a large excess of CoQ1. This
behavior refers to a competitive mechanism of inhibition. Down:
compound 2b: the dotted line refers to the NADH-CoQ1 activity in
presence of a large excess of CoQ1. This behavior refers to a
uncompetitive mechanism of inhibition.
4. Conclusion
Most complex I inhibitors are hydrophobic or amphi-
pathic compounds and it seems that they generally act
as ubiquinone antagonists. Kinetic studies suggested
that these inhibitors can be grouped into three classes
represented by piericidin (class I/A-type), rotenone
(class II/B-type), and capsaicin (C-type).10 Direct com-
petition experiments with inhibitors from different
classes revealed that they all share one common binding
domain with overlapping sites.24 Radioligand binding
and photoaffinity labeling studies confirmed that a wide
variety of structurally different inhibitors act at a com-
mon binding domain in complex I.25,46 Since imidazo-
thiazoles also inhibit complex I,33 we are prompted to
imagine that also compounds 2a–i might bind in the
large amphipathic pocket, which is thought to accept a
plethora of chemically different compounds and to con-
tain the ubiquinone binding site(s) of complex I. The
interactions of the different functional groups in the
imidazothiazole system with the binding environment
of several subunits outlining a cavity-like structure,
determine the variation of inhibitory profile from a
prevalently competitive inhibition to an almost uncom-
petitive behavior with respect to the acceptor substrate,
CoQ1.
complexes of the respiratory chain located downstream
the complex I. With respect to the mechanism of inhibi-
tion, it appears that 2a behaves as a competitive inhibi-
tor of CoQ1, whereas 2b has an uncompetitive behavior
as shown by titration (see Fig. 1); the other show an
intermediate mechanism of inhibition.
In the literature there are many reports dealing with the
PTP of isolated mitochondria,27–29,44,45 a field expanded
by the observation that the mitochondria soon after iso-
lation begin to swell, lose their ability to phosphorylate
and release matrix contents unless Ca2+ ions are ex-
cluded from the suspension. More recently the PTP
has been considered a calcium dependent cyclosporine
A-sensitive channel.29 CoQ0 is a PTP-inhibitor as dem-
onstrated by an analysis of quinone with structural var-
iants useful to identify the features required for the
regulation of the PTP, whereas CoQ1 is a PTP-binding
analogue that has no effects per se, but counteracts the
effects of both inhibitors and inducers of PTP. More-
over, the structure–activity correlations indicate that
minor modifications in isoprenoid side chain can turn
a PTP-inhibitor into an activator.27 For the above
This study shows that when the substituent in position 6
is a phenyl group (2c, 2h) the compounds are no inhib-
itors and seem to be very poor electron acceptors, but
they appear to be able to reach the physiological site(s)

5530
A. Andreani et al. / Bioorg. Med. Chem. 12 (2004) 5525–5532
even if the amount of these compounds that can interact
with this site is very low as indicated by the low Vmax in-
duced. Chlorine has the opposite effect since compounds
2a, 2d, and 2f are the best inhibitors of the series and
they are also quite weak electron acceptors. In com-
pounds 2b, 2e, 2g, and 2i, bearing the same substituent
in position 6, the behavior is quite different and deter-
mined by the substituents at the 2–3 positions.
sults seem to support the hypothesis of a direct interac-
tion with the PTP: in fact all the compounds tested are
PTP inducer irrespective of their effect on complex I.
5. Experimental
5.1. Chemistry
The melting points are uncorrected. Analyses (C, H, N)
were within 0.4% of the theoretical values. Bakerflex
plates (silica gel IB2-F) were used for TLC. Kieselgel
60(Merck) was used for column chromatography. The
IR spectra were recorded in Nujol on a Nicolet Avatar
The polar 1,4-benzoquinone ring of ubiquinones (CoQs)
is directly involved in redox reactions. Molecular orbital
calculation47,48 demonstrated that the conformations of
the methoxy groups in the 2- and 3-positions affect the
electrical potential of the oxidized form of CoQs or sem-
iquinone radicals through conformational interconver-
sion. The presence of this polar ring in compounds
2a–i is responsible for their electron acceptor activity.
kcat/Km ratio shows that complex I has the same ÔaffinityÕ
for all compounds 2a–i indicating that the higher Vmax
values elicited by some of them are due to their interac-
tion with a non-physiological site. Only compound 2h
seems to be more specific for complex I, in fact its
kcat/Km value is the highest, moreover it is the only com-
pounds showing an 80% of sensitivity to rotenone. From
the study of their electrostatic properties, it is evident
that the benzoquinone part of the new molecules could
be directly involved in redox reactions. In any case,
the substitution of the hydrophobic isoprenyl tail of
ubiquinones with the hydrophobic imidazothiazole sys-
tem causes loss of capability to be electron acceptors
from complex I, so the tail in the 6 position of the
ubiquinone series seems not only to increase the
hydrophobicity and to facilitate lateral diffusion in
biomembranes49 but it is also indispensable to ensure a
correct interaction with the active site of the enzyme.
Our results are in good agreement with the concept of
a large and rather unspecific inhibitor binding pocket
that could provide an amphipathic ÔrampÕ guiding ubiq-
uinone from the membrane domain into the catalytic
site located in the hydrophilic domain of complex I. In
this kind of picture it is possible to guess that the imid-
azothiazole system interacts with initial part of that
amphipathic ÔrampÕ preventing the interaction of the
benzoquinone moiety with the catalytic site (perhaps
the methylene bridge is too short). The benzoquinone
ring guides the imidazothiazole group to the complex I
active site as demonstrated by the lower IC50 of
these compounds in comparison with that induced by
the imidazothiazole system without the quinone
(IC50 = 0.2–1.5lM vs 75–80lM).34 On the other hand
the imidazothiazole system may prevent the reduction
of the benzoquinone ring. The molecular skeleton allows
for structural modifications that are presently under
study to find the distance between the imidazothiazole
group and the benzoquinone ring, which reverses the
inhibitory effect. This should give us information about
the distance between the inhibition and the electron
acceptor sites.
1
320E.S.P.; mmax is expressed in cmÀ1. The H NMR
spectra were recorded in (CD3)2SO on a Varian Gemini
(300MHz); the chemical shift (referenced to solvent sig-
nal) is expressed in d (ppm) and J in Hz. 2,3-Dimethoxy-
5-methyl-1,4-benzoquinone and chloroacetone were
purchased from Aldrich. 2-Amino-5-chlorothiazole
hydrochloride was purchased from Lancaster. Other
chemicals were commercial products of analytical grade.
5.1.1. Synthesis of 2-chloro-6-methylimidazo[2,1-b]thi-
azole 3i. 2-Amino-5-chlorothiazole hydrochloride
(2.56g, 15mmol) was dissolved in water (80mL) and
the solution was cautiously basified by dropwise addi-
tion of 15% NH4OH. The obtained precipitate was col-
lected by filtration and dissolved in THF (30mL). The
solution was added of chloroacetone (1.4mL, 15mmol)
and NaHCO3 (1.2g, 15mmol). The reaction mixture
was stirred and kept under reflux for 20h. After cooling
THF was evaporated under reduced pressure, the crude
product was suspended in water (80mL) and extracted
with CHCl3 (3 · 30mL). The organic phase was dried
over anhydrous sodium sulfate, filtered, and the filtrate
was concentrated under reduced pressure to afford an
oil (1.4g) that was used without further purification in
1
the next step. C6H5ClN2S (172.6). H NMR: 2.21 (3H,
d, CH3, J = 1.2), 7.48 (1H, q, H-5, J = 1.2), 8.15 (1H,
s, H-3).
5.1.2. Synthesis of 2-chloro-6-methylimidazo[2,1-b]thi-
azole-5-carboxaldeyde 4i. The Vilsmeier reagent was pre-
pared at 0–5ꢁC by dropping POCl3 (1.8mL, 20mmol)
into a stirred solution of DMF (1.9mL, 25mmol) in
CHCl3 (5mL). Compound 3i (2g, 10mmol) in CHCl3
(60mL) was added dropwise to the Vilsmeier reagent
while maintaining stirring and cooling. The reaction
mixture was kept for 3h at room temperature and under
reflux for 8h. Chloroform was removed under reduced
pressure and the resulting oil was poured onto ice. The
crude aldehyde 4i thus obtained was extracted with chlo-
roform and crystallized from ethanol with a yield of
65%. C7H5ClN2OS (200.6) mp 120–122ꢁC. IR: 1645,
1305, 1260, 885. 1H NMR: 2.55 (3H, s, CH3), 8.53
(1H, s, H-3), 9.83 (1H, s, CHO).
5.1.3. General procedure for the synthesis of the substi-
tuted 5-hydroxymethylimidazo[2,1-b]thiazoles 5h–i. The
appropriate aldehyde 4h–i (3mmol) was dissolved in
methanol (250mL) and treated portionwise, under cool-
ing and stirring, with sodium borohydride (226.9mg,
From this study it is difficult to say if these new quinone
inhibitors 2a–i are able to mediate the observed swelling
of mitochondria through a binding to a specific site or to
an alteration of the membrane potential. Preliminary re-

A. Andreani et al. / Bioorg. Med. Chem. 12 (2004) 5525–5532
5531
6mmol). Stirring at room temperature was maintained
for 3h, then the reaction mixture was refluxed for 4h.
Methanol was evaporated under reduced pressure and
by addition of water the crude hydroxymethyl derivative
was obtained (by filtration for compound 5h and by
extraction with chloroform for compound 5i). These
compounds were crystallized from methanol with a yield
of 65% for 5h and 74% for 5i.
NADH (75lM) and varying ubiquinone concentrations
following the decrease in absorbance in a double wave-
length spectrophotometer at 340 À 380nm and using an
extinction coefficient of 3.5mMÀ1 cmÀ1. NADH-O2
reductase activity was assayed essentially in the same con-
ditions avoiding only KCN and antimycin A in the assay
mixture.
5.3. Measurement of mitochondrial swelling
Compound 5h C12H9ClN2OS (264.7) mp 168–171ꢁC;
IR: 3170, 1010, 1000, 765; 1H NMR: 4.76 (2H, d,
CH2, J = 5.5), 5.46 (1H, t, OH, J = 5.5), 7.33 (1H, t,
ar, J = 7.3), 7.45 (2H, t, ar, J = 7.3), 7.74 (2H, d, ar,
J = 7.3), 8.31 (1H, s, H-3).
Mitochondria were prepared from the livers of male al-
bino Wistar rats (Charles-River, Calco, LC, Italy),
weighing 150–175g essentially as previously described.55
Rat liver coupled mitochondria permeabilization to su-
crose (swelling) was determined by monitoring ab-
sorbance changes at 540nm with a Jasco V-550
spectrophotometer equipped with magnetic stirring
and thermostatic control. The incubation medium con-
tained 0.2M sucrose, 1mM KH2PO4, 10mM Tris–
Mops, pH7.4; 5mM succinate–Tris, 10lM EGTA–Tris,
1lgmLÀ1 rotenone, 1lgmLÀ1 oligomycin; final volume
2mL at 25ꢁC. The assay was performed in presence of
0.5mg mitochondrial protein pre-loaded with 10lM
Ca2+ pulse followed by the addition of 0.5mM EGTA
to prevent the redistribution of Ca2+ released after par-
tial mitochondrial swelling, then the pore opening was
tested monitoring the decrease in absorbance at
540nm due to the addition of FCCP (carbonylcyanide
Compound 5i C7H7ClN2OS (202.7) mp 167–168ꢁC dec;
1
IR: 3100, 1305, 995, 790; H NMR: 2.20(3H, s, CH ),
3
4.58 (2H, d, CH2, J = 5.4), 5.12 (1H, t, OH, J = 5.4),
8.13 (1H, s, H-3).
5.1.4. General procedure for the synthesis of the
substituted 6-imidazo[2,1-b]thiazolylmethyl-2,3-dimethoxy-
5-methyl-benzoquinones 2a–i. 2,3-Dimethoxy-5-methyl-
1,4-benzoquinone 1 (332.5mg, 2.1mmol) was dissolved
in 120mL of diethyl ether and shaken with sodium dithio-
nite (2.78g, 16mmol) in 70mL of water until the solu-
tion becomes colorless. The mixture was extracted with
diethyl ether. The organic phase was dried over anhy-
drous sodium sulfate, filtered, and evaporated. The
hydroquinone 6, obtained as colorless crystals, was dis-
solved in 30mL of dry dioxane. This solution was added
by a syringe in a three necked flask containing a solution
of the appropriate 5-hydroxymethylimidazo[2,1-b]thiaz-
ole 5a–i (2.1mmol) in 20mL of dry dioxane stirred in a
nitrogen atmosphere at room temperature. The resulting
solution was treated with 0.4mL of boron trifluoride
etherate (3.1mmol) in 10mL of dry dioxane over
20min by means of a dropping funnel. The mixture
was stirred for 10h at room temperature, treated with
50mL of water, and extracted with diethyl ether. The or-
ganic phase was dried over anhydrous sodium sulfate,
treated with silver(I)oxide (486.65mg, 2.1mmol), and
stirred for 3h. The solution was filtered and the filtrate
was concentrated under reduced pressure. The residue
was purified by column chromatography or by crystalli-
zation (see Table 1). Compounds 2a–i were obtained as
orange/red solids with a yield of 15–35%.
p-trifluorometoxyphenylhydrazone)56
that
induces
membrane depolarization; the loss of swelling in pres-
ence of 1lgmLÀ1 cyclosporin A, a specific inhibitor of
the pore opening, is indicative of the involvement of the
permeability transition pore in the swelling process.
The experiment allows a calibration of the absorbance
as a function of the fraction of swollen mitochondria,
where zero is the value obtained from mitochondria
incubated in the absence of Ca2+ and 100 is the value
obtained after all of the mitochondria have undergone
the transition induced by 200nM FCCP according to
Petronilli et al.57,58 The fraction of mitochondria with
open pore was indicated as /. The effect of the different
compounds on the pore opening was evaluated compar-
ing the fraction of swollen mitochondria, induced by
addition of 2a–i to the assay mixture, with that obtained
by the addition of 200nM FCCP.
Acknowledgements
5.2. Enzyme assays
This work was supported by the University of Bologna
(Funds for Selected Research Topics).
Submitochondrial particles (SMP) were prepared from
beef heart mitochondria (BHM) by sonic irradiation of
the frozen and thawed BHM.50 The particles were essen-
tially broken membrane fragments.51 Protein was evalu-
ated by the Biuret method of Gornall et al.52 with
addition of 10% sodium deoxycholate and using bovine
serum albumine (BSA) as standard. NADH-CoQ reduc-
tase was assayed essentially as described by Yagi53 and
modified by Degli Esposti et al.54 in presence of 2mM
KCN and 2lM antimycin A to block complex IV and
III, respectively. Determination of the kinetic constants
was accomplished at quasi saturating concentration of
References and notes
1. Zickermann, V.; Bostina, M.; Hunte, C.; Ruiz, T.;
Radermacher, M.; Brandt, U. J. Biol. Chem. 2003, 278,
29072–29078.
2. Bo¨ttcher, B.; Scheide, D.; Hesterberg, M.; Nagel-Steger,
L.; Friedrich, T. J. Biol. Chem. 2002, 277, 17970–17977.
3. Buchanan, S. K.; Walker, J. E. Biochem. J. 1996, 318, 343–
349.

5532
A. Andreani et al. / Bioorg. Med. Chem. 12 (2004) 5525–5532
4. Weiss, H.; Friedrich, T.; Hofhaus, G.; Preis, D. Eur. J.
Biochem. 1991, 197, 563–576.
5. Vinogradov, A. D.; Sled, V. D.; Burbaev, D. S.; Griven-
nikova, V. G.; Moroz, I. A.; Ohnishi, T. FEBS Lett. 1995,
370, 83–87.
31. Susin, S. A.; Lorenzo, H. K.; Zamzami, N.; Marzo, I.;
Snow, B. E.; Brothers, G. M.; Mangion, J.; Jacotot, E.;
Costantini, P.; Loeffler, M.; Larochette, N.; Goodlett, D.
R.; Aebersold, R.; Siderovski, D. P.; Penninger, J. M.;
Kroemer, G. Nature 1999, 397, 441–446.
6. De Jong, A. M. P.; Albracht, S. P. J. Eur. J. Biochem.
1994, 222, 975–982.
32. Fontaine, E.; Bernardi, P. J. Bioenerg. Biomembr. 1999,
31, 335–345.
7. Di Virgilio, F.; Azzone, G. F. J. Biol. Chem. 1982, 257,
4106–4113.
8. Wikstro¨m, M. FEBS Lett. 1984, 169, 300–304.
9. Brown, G. C.; Brand, M. D. Biochem. J. 1988, 252, 473–
479.
10. Degli Esposti, M. Biochim. Biophys. Acta 1998, 1364, 222–
235.
11. Johns, D. R. N. Engl. J. Med. 1995, 333, 638–644.
12. Schapira, A. H. V. J. Bioenerg. Biomembr. 1997, 29, 105–
107.
33. Andreani, A.; Rambaldi, M.; Leoni, A.; Locatelli, A.;
Ghelli, A.; Ratta, M.; Benelli, B.; Degli Esposti, M. J.
Med. Chem. 1995, 38, 1090–1097.
34. Andreani, A.; Rambaldi, M.; Leoni, A.; Morigi, R.;
Locatelli, A.; Giorgi, G.; Lenaz, G.; Ghelli, A.; Degli
Esposti, M. Eur. J. Med. Chem. 1999, 34, 883–889.
35. Andreani, A.; Rambaldi, M.; Locatelli, A.; Leoni, A.;
Ghelli, A.; Degli Esposti, M. Pharm. Acta Helv. 1994, 69,
15–20.
36. Paolini, J. P.; Lendvay, L. J. J. Med. Chem. 1969, 12,
1031–1034.
37. Andreani, A.; Bonazzi, D.; Rambaldi, M.; Greci, L. Boll.
Chim. Farm. 1979, 118, 694–698.
13. Majander, A.; Huoponen, K.; Savontaus, M. L.; Niko-
skelainen, E.; Wikstro¨m, M. FEBS Lett. 1991, 292, 289–
292.
14. Degli Esposti, M.; Ghelli, A. Biochim. Biophys. Acta 1994,
1187, 116–120.
38. Andreani, A.; Rambaldi, M.; Bonazzi, D. Farmaco-Ed.
Sci. 1980, 35, 573–580.
15. Shoffner, J. M.; Brown, M. D.; Torroni, A.; Lott, M. T.;
Cabell, M. F.; Mirra, S. S.; Beal, M. F.; Yang, C. C.;
Gearing, M.; Salvo, R.; Watts, R. L.; Juncos, J. L.;
Hansen, L. A.; Crain, B. J.; Fayad, M.; Reckord, C. L.;
Wallace, D. C. Genomics 1993, 17, 171–184.
16. Flint Beal, M. Biochim. Biophys. Acta 1998, 1366, 211–
223.
17. Lenaz, G.; Bovina, C.; Castelluccio, C.; Fato, R.; Formig-
gini, G.; Genova, M. L.; Marchetti, M.; Merlo Pich, M.;
Pallotti, F.; Parenti Castelli, G.; Biagini, G. Mol. Cell.
Biochem. 1997, 174, 329–333.
18. Luyengi, L.; Lee, I. S.; Mar, W.; Fong, H. H. S.; Pezzuto,
J. M.; Kinghorn, A. D. Phytochemistry 1994, 36, 1523–
1526.
19. Fang, N.; Casida, J. E. Proc. Natl. Acad. Sci. U.S.A. 1998,
95, 3380–3384.
39. Andreani, A.; Rambaldi, M.; Carloni, P.; Greci, L.; Stipa,
P. J. Heterocycl. Chem. 1989, 26, 525–529.
40. Andreani, A.; Rambaldi, M.; Mascellani, G.; Rugarli, P.
Eur. J. Med. Chem. 1987, 22, 19–22.
41. Andreani, A.; Rambaldi, M.; Locatelli, A.; Andreani, F.
Collect. Czech. Chem. Commun. 1991, 56, 2436–2447.
42. Andreani, A.; Bonazzi, D.; Rambaldi, M. Farmaco-Ed.
Sci. 1980, 35, 896–901.
43. Merck & Co., Inc. U.S. Patent 927,531 May 29, 1963;
Chem. Abstr. 1963, 59, 12710F.
44. Fontaine, E.; Eriksson, O.; Ichas, F.; Bernardi, P. J. Biol.
Chem. 1998, 273, 12662–12668.
45. Bernardi, P. J. Biol. Chem. 1992, 267, 8834–8839.
46. Schuler, F.; Yano, T.; Di Bernardo, S.; Yagi, T.; Yanko-
vskaya, V.; Singer, T. P.; Casida, J. E. Proc. Natl. Acad.
Sci. U.S.A. 1999, 96, 4149–4153.
20. Degli Esposti, M.; Ngo, A.; McMullen, G.; Ghelli, A.;
Sparla, F.; Benelli, B.; Ratta, M.; Linnane, A. W.
Biochem. J. 1996, 313, 327–334.
21. Lenaz, G. Biochim. Biophys. Acta 1998, 1364, 207–221.
22. Sakamoto, K.; Miyoshi, H.; Ohshima, M.; Kuwabara, K.;
Kano, K.; Akagi, T.; Mogi, T.; Iwamura, H. Biochemistry
1998, 37, 15106–15113.
47. Robinson, H. H.; Kaha, S. D. J. Am. Chem. Soc. 1990,
112, 4728–4731.
48. Gu, L. Q.; Yu, L.; Yu, C. A. Biochem. Biophys. Acta 1990,
1015, 482–492.
49. Lenaz, G.; Degli Esposti, M.; Fahmy, T.; Fato, R.;
Rugolo, M.; Parenti Castelli, G. Biomed. Clin. Aspects
Coenzyme Q 1984, 4, 33–42.
23. Ohshima, M.; Miyoshi, H.; Sakamoto, K.; Takegami, K.;
Iwata, J.; Kuwabara, K.; Iwamura, H.; Yagi, T. Biochem-
istry 1998, 37, 6436–6445.
50. Beyer, R. E. Methods Enzymol. 1967, 10, 186–194.
51. Fato, R.; Cavazzoni, M.; Castelluccio, C.; Parenti Castelli,
G.; Palmer, G.; Degli Esposti, M.; Lenaz, G. Biochem. J.
1993, 290, 225–236.
24. Okun, J. G.; Lummen, P.; Brandt, U. J. Biol. Chem. 1999,
¨
29, 2625–2630.
52. Gornall, A. G.; Bardawill, C. J.; David, M. M. J. Biol.
Chem. 1949, 177, 751–766.
53. Yagi, T. Arch. Biochem. Biophys. 1990, 281, 305–311.
54. Degli Esposti, M.; Ghelli, A.; Crimi, M.; Estornell, E.;
Fato, R.; Lenaz, G. Biochem. Biophys. Res. Commun.
1993, 190, 1090–1096.
25. Miyoshi, H. J. Bioenerg. Biomembr. 2001, 33, 223–231.
26. Fato, R.; Estornell, E.; Di Bernardo, S.; Pallotti, F.;
Parenti Castelli, G.; Lenaz, G. Biochemistry 1996, 35,
2705–2716.
27. Fontaine, E.; Ichas, F.; Bernardi, P. J. Biol. Chem. 1998,
273, 25734–25740.
28. Ludivine, W.; Nogueira, V.; Leverve, X.; Heitz, M. P.;
Bernardi, P.; Fontaine, E. J. Biol. Chem. 2000, 275, 29521–
29527.
29. Crompton, M.; Ellinger, H.; Costi, A. Biochem. J. 1988,
255, 357–360.
55. Costantini, P.; Petronilli, V.; Colonna, R.; Bernardi, P.
Toxicology 1995, 99, 77–88.
56. Heytler, P. G. Methods Enzymol. 1979, 55, 442–462.
57. Petronilli, V.; Cola, C.; Massari, S.; Colonna, R.; Ber-
nardi, P. J. Biol. Chem. 1993, 268(29), 21939–21945.
58. Petronilli, V.; Nicolli, A.; Costantini, P.; Colonna, R.;
Bernardi, P. Biochim. Biophys. Acta 1994, 1187(2), 255–
259.
30. Liu, X.; Kim, C. N.; Yang, J.; Jemmerson, R.; Wang, X.
Cell 1996, 86, 147–157.