Bioscience Reports, Vol. 17, No. 3, 1997
REVIEW
"Mild" Uncoupling of Mitochondria
A. A. Starkov
Received January 10, 1997
Recently, it was proposed that the thyroid hormone-mediated uncoupling in mitochondria is involved
in the cellular defence system against free radicals (Skulachev V.P. (1996) Quart. Rev. Biophys.
29:169-202). This phenomenon was named "mild" uncoupling. It was postulated to be a proteinmediated process controlled by several factors. The data reported during the past 40 years, pointing to
the protein-mediated uncoupling mechanism in mitochondria, are reviewed in a context of
hypothetical properties of "mild" uncoupling. The mechanism of "mild" uncoupling is suggested to be
the following: (a) mitochondria possess protein(s) that regulate the proton permeability of inner
mitochondrial membrane; (b) these proteins are regulated by binding of an unidentified lowmolecular-weight endogenous compound with properties resembling those of the most active artificial
uncouplers like FCCP and SF6847; (c) the interaction of this compound with its target protein(s) is
modulated by a thyroid hormone in a positive (i.e. enhancing the proton permeability) way and by sex
steroid hormones in a negative way; (e) endogenous fatty acids can attenuate the influence of both
thyroid and steroid hormones.
KEY WORDS: Uncoupling; mitochondria; free radicals; thyroid hormones; steroid hormones.
INTRODUCTION
Recently, it was proposed that the thyroid hormone—mediated uncoupling in
mitochondria is involved in the cellular defence system against reactive oxygen
species (ROS). Such an uncoupling can be favorable for maintenance of the
concentrations of both O2 and one electron O2 reductants at low level irrespective
of the cellular energy demands thus preventing "parasitic" reactions of the ROS
formation [1]. This phenomenon was named "mild" uncoupling. It was postulated
to be a protein-mediated process controlled by several factors.
According to the modern paradigm of chemiosmotic theory, uncouplers of
the oxidative phosphorylation operate as protonophores traversing the hydrophobic membrane region in their protonated and deprotonated forms thereby
increasing the proton permeability of coupling membrane and wasting energy by
means of a futile proton cycle. When this mechanism is operating, the value of the
membrane proton permeability depends on the concentration of the uncoupler
and on the particular chemical properties of the uncoupler molecule such as
Department of Bioenergetics, A. N. Belozersky Institute for Physico-Chemical Biology, Moscow State
University, Moscow 117899, Russia.
273
0144-84«3/97/0600-0273$12.50/0 © 1997 Plenum Publishing Corporation
274
Starkov
hydrophobicity, pK and the degree of charge delocalisation in the anionic form of
the uncoupler. Other factors affecting the proton permeability of the membrane
are the physical and chemical properties of the membrane per se, such as fluidity
and phospholipid composition. To operate, this mechanism required only the
uncoupler and the lipid membrane, no proteins being involved. There are no
reasons to doubt that such a mechanism of uncoupling can account for the effects
of high concentrations of most (if not all) artificial protonophorous uncouplers.
However, during recent years data were obtained that showed that the action of
low concentrations of several uncouplers in mitochondria can be suppressed by
specific inhibitors, which seems to be incompatible with the mechanism described
above. This suggests the involvement of protein(s) in the action of uncouplers and
opens the possibility that the apparently useless and even harmful waste of energy
caused by the uncouplers may be a physiologically important process.
During the past 40 years, some data have been reported that support the
possibility of the protein-mediated uncoupling mechanism in mitochondria (other
than brown fat mitochondria, where the existence of such a mechanism has been
proven). Here, some of these data will be briefly reviewed in a context of
hypothetical properties of "mild" uncoupling.
DATA IN FAVOR OF PROTEIN-MEDIATED MECHANISM OF
UNCOUPLING IN MITOCHONDRIA
The protein-mediated uncoupling means that an uncoupler directly interacts
with a protein(s) at a specific binding site.
The existence of the uncoupler-binding sites has been reported in the past.
Using the photoaffinity label 2-azido-4-nitrophenol (which is a structural analog of
2,4-dinitrophenol) Hanstein and Hatefi found the specific uncoupler-binding
site(s) in the inner mitochondrial membrane. Photoaffinity labeling experiments
revealed that uncouplers bind predominantly to a 30 kDa polypeptide (see ref. [2]
and ref. therein). The uncoupler-binding sites were demonstrated in mitochondria
isolated from beef heart, rat liver, yeast, and Euglena gracilus, but not in
erythrocyte ghosts (see ref. [3] for review). Covalent binding of 2-azido-4nitrophenol to the specific site induced no uncoupling but rather appeared to
freeze mitochondria to a certain degree in a permanent state 4 not amenable to
stimulation by ADP or uncouplers. Further studies revealed that 2-azido-4nitrophenol also binds to a 56 kDa polipeptide, which was identified as the
a-subunit of FrATPase. The authors suggested that the mitochondrial
uncoupler-binding site is formed by the 30 kDa uncoupler-binding protein and
a-subunit of F, [3].
Using another uncoupler capable of photoaffinity labeling, 2-nitro-4azidocarbonylcyanidephenylhydrazone (N3CCP), Katre and Wilson demonstrated
the existence of a high-affinity binding site for uncouplers in rat heart mitochondria [4], rat liver mitochondria, pigeon heart mitochondria [5] and later in
membranes of Paracoccus denitrifications and Tetrahymena pyriformis [6].
Further studies by Katre and Wilson revealed that the N3CCP-binding site
Mitochondrial uncoupling
275
differed from the 2-azido-4-nitrophenol-labeled site. First, the N3CCP-labeIed
polypeptide was of 10-12 kDa, and second, covalent binding of the N3CCP to
nitochondria resulted in the irreversible uncoupling.
Working with another class of uncouplers, namely substituted 3,5dichlorosalicylanilides, Storey et al. demonstrated that the difference in uncoupling activity of these uncouplers was related to their steric properties. The
authors proposed that uncouplers bind to a specific site [7].
An additional evidence in favor of the protein-mediated mechanism of the
uncoupler action was obtained in studies on the long-chain fatty acid-induced
uncoupling in mitochondria of various tissues. Besides a well known effect of fatty
acids in brown fat mitochondria containing a specific fatty-acid binding uncoupling protein, uncoupling by fatty acids was shown to be mediated by ATP/ADP
antiporter in heart, muscle, and liver mitochondria [8,9]. Studies by Skulachev's
group revealed also that uncoupling induced by 2,4-dinitrophenol [8] and some
other uncouplers [10] may be partially mediated by ATP/ADP-antiporter.
Recently, we showed that uncoupling induced by FCCP, CCCP, or SF6847 can be
prevented or suppressed by 6-ketocholestanol which is a 6-keto cholesterol
derivative. The effect of this compound was shown in mitochondria and
cytochrome oxidase proteoliposomes. On the other hand, 6-ketocholestanol did
not decrease the bilayer membrane conductance increased by these protonophores [11]. It has been suggested that cytochrome oxidase and perhaps some
other inner mitochondrial membrane proteins mediate the uncoupling action of
FCCP, CCCP, and SF6847. We hypothesized (a) that the SF6847-like uncouplers
are active at very low concentrations because they can cross the mitochondria
membrane with the help of membrane proteins and (b) that 6-ketocholestanol
changes the environment (e.g., dipole potential) of cytochrome oxidase or other
protein(s) thereby decreasing the transport rate of these uncouplers [11].
All these studies, although they cannot be considered as direct evidence,
provide a basis for the suggestion that the uncoupling activity of most (if not all)
well-known artificial uncouplers in mitochondria is, at least partially, mediated by
a specific interaction between an uncoupler and proteins in the inner mitochondrial membrane. These proteins may differ for different classes of uncouplers,
which may explain the apparent discrepancy in the results of Hanstein and Hatefi
(two uncoupler-binding proteins of 30 and 59 kDa) and of Katre and Wilson (a
single protein of 10-12 kDa). In this respect, it is interesting that in the above
mentioned studies [9,10], 6-ketocholestanol strongly suppressed the uncoupling
induced by low concentrations of FCCP and CCCP which are chemically related
to N3CCP, but was almost without effect on the uncoupling induced by
2,4-dinitrophenol which is related to 2-azido-4-nitrophenol.
The assumption of protein-mediated uncoupling logically points to one more
interesting possibility, namely the existence of endogeneous compound(s) controlling the degree of coupling in mitochondria. These compounds may have no
protonophorous activity per se, i.e. in lipid bilayer, but rather they should control
the proton permeability of mitochondrial membrane by interacting with membrane proteins.
Skulachev proposed that thyroid hormones may be such compounds [1,12].
276
Starkov
Indeed, the effects of hyper- and hypothyroidism on the mitochondrial membrane
proton conductance are well known (see ref. [13] for a review and ref. [14] for
recent data on the interaction of triiodothyronine with mitochondrial ADP/ATP
carrier). It was also shown that diiodothyronine (T2) quickly increased the oxygen
consumption in perfused liver of hypothyroid rats, whereas similar effects of
triiodothyronine and thyroxine were abolished by the deiodinase inhibitor,
thereby suggesting diiodothyronine that is responsible for stimulation of mitocnondrial respiration [15]. Later, an indication on the specific 3,5-T2-binding sites
in rat liver mitochondria was reported [16].
3,5-T2 and 3.3'-T2 were shown to stimulate the activity of rat liver
cytochrome c oxidase in both liver homogenate from hypothyroid animals and in
isolated mitochondria. The maximal effect was observed at 1 fiM of T2. The
authors concluded that T2 can directly stimulate the cytochrome oxidase activity
in hypothyroid mitochondria, probably through the involvement of a cytoplasmic
factor [17].
Additional evidence that thyroid hormones directly control the permeability
of mitochondrial membranes was provided by Kozhomkulov et al. [18]. It was
shown that a protein isolated from rat liver mitochondria can increase the
conductance of a planar bilayer lecithin membrane by about 10-fold, and the
addition of 10"7M thyroxine further increases the conductance [18].
Recently, it was shown that the membrane potential of rat liver mitochondria
decreased by addition of FCCP or SF6847, but not of 2,4-dinitrophenol or
palmitate, can be further decreased by 2-8 fj.M thyroxine, whereas no detectable
effect of thyroxine was found in absence of the uncoupler. Diiodothyronine and
triiodothyronine were less efficient. This effect was observed only in the presence
of bovine serum albumin. Male sex steroid hormones reversed the thyroxine
effect [19].
The antagonism of the thyroid hormones as catabolics and of steroid
hormones as anabolics at the level of the whole organism or a tissue is a
well-known fact. Some of these effects are mediated by protein synthesis
controlled at the genome level. For the thyroid hormones, however, several
non-genomic effects on mitochondria have also been reported (reviewed in ref.
[13]). The non-genomic effects of steroid hormones in mitochondria have also
been reported (some of them will be discussed on page 277).
DIRECT EFFECTS OF STEROID HORMONES ON ISOLATED
MITOCHONDRIA
Post-transcriptional and post translational actions of steroid hormones on
isolated mitochondria were described for high hormone concentrations which
were hardly physiological. For instance, progesterone and some other steroids
were shown to inhibit mitochondrial respiration by direct action on Complex 1 of
respiratory chain (see ref. [20] and ref. therein) at the rotenonebinding site [21].
This effect seems to be nonspecific and can be readily demonstrated with various
hydrophobic steroids applied in micromolar concentrations (40 /nM and higher).
Mitochondrial uncoupling
277
Higher concentrations (400-600 /u.M) of several steroid hormones were shown to
stimulate mitochondrial ATPase activity [22]. This effect did not apparently
correlate with the hydrophobicity of hormones or their ability for binding to
bovine serum albumin. However, an influence of steroid structure on the
stimulatory effect on the ATPase activity was found [22].
It was also shown that 100 /^M progesterone inhibited the succinatesupported State 3 respiration of rat liver mitochondria. The inhibition was
abolished by 2,4-dinitrophenol, Ca 2+ , gramicidin +NH4C1, EDTA. It differed
from the oligomycin effect in that the progesterone inhibited only ATP-synthesis
but not the 2,4-dinitrophenol stimulated ATPase activity. The authors concluded
that progesterone affected mitochondrial respiration by increasing the alkalinization of the mitochondrial matrix. A similar effect was shown also with
pregnanolone, whereas testosterone and a number of other steroids were without
effect [23].
Jung and Brierley have shown that high concentrations (hundreds of
micromoles) of progesterone, testosterone, and corticosterone inhibit passive
swelling and activate (the respiration-dependent) contraction of rat heart mitochondria suspended in nitrate-containing media at pH 8.3. The authors concluded
that these hormones arrested cation influx through the inner mitochondrial
membrane by a nonspecific mechanism [24].
The data obtained in studies of the in vitro effects of steroid hormones on
isolated mitochondria apparently suggest the lack of a specific regulatory
influence of these hormones on mitochondrial energetics. All the effects require
high nonphysiological concentrations of hormones and affect the energyproducing machinery of mitochondria in a "negative", destructive direction:
inhibition of respiration or ATP-synthesis [20], uncoupler-like effects [22],
nonspecific permeability changes, swelling of mitochondria, and loss of matrix
content [25,26]. These data are in line with the generally accepted explanation of
steroid effects by their ability "to fit or pack into membranes" [24] thus modifying
the mitochondrial membrane properties.
However, these data do not negate the possibility of the regulation of
mitochondrial functions by direct interaction of certan steroid hormones with an
unidentified mitochondrial protein or a number of proteins. Taking into account
the high hydrophobicity and structural rigidity of the steroid core, it may be
suggested that high concentrations of hormones required to affect mitochondrial
properties in vitro are due to the necessity of an additional component (which is
probably a protein of cytoplasmic origin) for proper delivery or orientation of a
hormone molecule to the site of it's action.
Recently, we succeeded in demonstrating a "positive" effect of some steroid
hormones on isolated mitochondria [19], It was shown that testosterone,
dihydrotestosterone, and progesterone can "recouple" mitochondria uncoupled
by low concentrations of potent protonophores such as FCCP, CCCP or SF6847.
This effect required relatively high concentrations of hormones (Q/z = 3550 /tM), and it was completely absent in cytochrome oxidase proteoliposomes or
in R.spheroides chromatophores. The recoupling was seen only in the presence of
bovine serum albumin and was abolished by low (5 /xM) concentrations of
Starkov
278
exogenous fatty acids. Moreover, the effect of these hormones exhibited clear
seasonal dependence and was influenced by the thyroid (and presumably by
noradrenaline) status of animals. Other hormones, such as glucocorticoids and
female sex hormones (estrone, estradiol), were without effect. We proposed that
the recoupling effect of male sex steroids may be responsible for their well-known
anabolic action in many tissues [27] which requried high concentrations of
hormones. It was proposed that sex steroids increase the efficiency of energy
production in mitochondria which are, to some degree, uncoupled in vivo by a
natural compound with properties similar to those of SF6847 [19].
There is a possibility that the mechanism of the steroid-induced recoupling
might not be mediated by binding of a steroid molecule to a specific protien. It
may consist in the modification of the membrane environment of protiens directly
involved in uncoupling. The insertion of a hydrophobic steroid hormone
molecule, which possesses one or two keto-groups, may change the dipole
potential of lipid membrane, thereby changing the accessibility of the site of
protein-uncoupler interaction. It was mentioned in the previous section that
6-ketocholestanol significantly suppresses the uncoupling induced by FCCP,
CCCP, or SF6847. This compound is known as a strong modificator of the dipole
potential of lipid membranes, which can affect the binding and the transport rates
of hydrophobic ions in a maembrane [28]. In this respect, it is interesting that
another potential dipole modificator, namely phloretin, whose effect on the dipole
potential is opposite to that of 6-ketocholestanol, was shown to be both an
uncoupler of mitochondria, when high concentraions of phloretin were used
(~500 /j,M), and an inhibitor of Complex I of the respiratory chain, when low
concentrations (—50/^M) are used, these concentrations are similar to those of
steroid hormones that exert the same effects (see above) [29].
HYPOTHETICAL MECHANISM OF "MILD" UNCOUPLING
Taking into account the above mentioned data, a hypothetical mechanism of
the protein-mediated uncoupling of mitochondria can be outlined as follows: (a)
mitochondria possess a protein or a number of proteins that regulate the proton
permeability of inner mitochondrial membrane; (b) these proteins are regulated
by binding of an unidentified low-molecular-weight endogenous compound with
properties resembling those of artificial uncouplers such as FCCP or SF6847, (c)
the interaction of this compound with its target protein(s) is controlled by a
thyroid hormone in a "positive" (enhancing the proton permeability) way and by
sex steroid hormones in a "negative" way; (e) endogenous long chain fatty acids
can attenuate the influence of both thyroid and steroid hormones. Such a
mechanism looks rather complicated but it can provide cells and whole tissue with
a tool to control various energy-dependent functions of mitochondria. It seems to
be fully compatible with the specific requirements of "mild" mitochondrial
uncoupling. The only one point is missing as yet is the nature of the endogenous
compound mentioned above in (b). Indeed, the finding of such a compound has
not been reported. However, it can be supposed that protonated superoxide
Mitochondria! uncoupling
279
(HO^) can perform this function. The participation of HO^ in uncoupling of
mitochondria has already been suggested (see the article by Shu-sen Liu in this
issue). With this compound functioning as an endogeneous factor controlling the
degree of coupling of mitochondria, the mechanism of "mild" uncoupling
becomes truly efficient because it will sense both the mitochondrial protonmotive
force and the level of oxygen, thus switching on and off depending on the
particular circumstances in a single mitochondrion (for more details, see the
article by Skulachev in this issue). Of course, such a mechanism remains highly
speculative due to the lack of direct evidence.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
Skulachev, V. P. (1996) Quart. Rev. Biophys. 29:169-202.
Hatefi, Y. (1975) J of Supramol. Structure 3:201-213.
Hanstein, W. G. (1976) Biochim. Biophys. Ada, 456:129-143.
Katre, N. V. and Wilson, D. F. (1977) Arch. Biochem. Biophys. 184:578-585.
Katre, N. V. and Wilson, D. F. (1978) Arch. Biochem. Biophys., 191:647-656.
Katre, N. V. and Wilson, D. F. (1980) Biochim. Biophys. Acta, 593:224-229.
Storey, B. T., Wilson, D. F., Bracey, A., Rosen, S. L. and Stephenson, S. (1975) FEBS Lett.,
49:338-341.
8. Andreyev, A. Yu., et al. (1989) Eur. J. Biochem., 182:585-592.
9. Dedukhova, V. I., Mokhova, E. N., Skulachev, V. P., Starkov, A. A., Arrigoni-Martelli, E. and
Bobyleva, V. A. (1991) FEBS Lett. 259:51-54.
10. Starkov, A. A., Dedukhova, V. I. and Skulachev, V. P. (1994) FEBS Lett. 355:305-308.
11. Starkov, A. A., et al. (1997) Biophys. Acta. (in press).
12. Skulachev, V. P. (1995) Molek. Biol., 29:709-715 (in Russian).
13. Soboll, S. (1993) Biochim. Biophys. Ada, 1144:1-16.
14. Mowbray, J. and Hardy, D. L. (1996) FEBS Lett., 394:61-65.
15. Horst, C, Rokos, H. and Seitz, H. J. (1989) Biochem. J. 261:945-950.
16. Goglia, F., Lanni, A., Horst, C., Moreno, M. and Thoma, R. (1994) J. Mol. Endocrinol.
13:275-282.
17. Lanni, A., Moreno, M., Lombardi, A. and Goglia, F. (1994) Mol. Cell. Endocrinol. 99:89-94.
18. Kozhomkulov, E., Normatov, K., Azimova, Sh. and Marzoev, A. (1984) Problemy Endocrinologii,
30:60-63 (in Russian).
19. Starkov, A. A., Simonian, R. A., Dedukhova, V. I., Mansurova, E. S., Palamarchuk, L. A. and
Skulachev, V. P. (1997) Biochim. Biophys. Acta. (in press).
20. Vallejos, R. H. and Sloppani, A. O. M. (1967) Biochem. Biophys. Acta, 131:295-309.
21. Yelding, K. L., Tomkins, G. M., Munday, J. S. and Cowley, I. J. (1960) J. Biol. Chem.
235:3413-3416.
22. Blecher, M. and White, A. (1960) J. Biol. Chem. 235:3404-3412.
23. Aleksandrowicz, Z., Swierczynski, J. and Zelewski, L. (1972) Eur. J. Biochem. 31:300-307.
24. Jung, D. W. and Brierley, G. P. (1981) Experimentia, 37:237-238.
25. Boveris, A. and Stoppani, A. O. M. (1971) Arch. Biochem. Biophys. 142:150-156.
26. Gallacher, C. H. (1960) Biochem. J. 38:74.
27. Heftmann, E. (1970) Steroid Biochemistry, Academic Press, New-York.
28. Franclin, J. C. and Cafiso, D. S. (1993) Biophys. J., 65:289-299.
29. De Jonge, P. C., Wieringa, T., Van Putten, J. P., Krans, H. M. and Van Dam K. (1983) Biochim.
Biophys. Acta. 722:219-225.