Expression of uncoupling protein-3 in subsarcolemmal
and intermyofibrillar mitochondria of various mouse muscle
types and its modulation by fasting
Maria Jimenez, Cedric Yvon, Lorenz Lehr, Bertrand Le
´
ger, Patrick Keller, Aaron Russell, Franc¸oise Kuhne,
Pierre Flandin, Jean-Paul Giacobino and Patrick Muzzin
Department of Medical Biochemistry, Faculty of Medicine, University of Geneva, Switzerland
Uncoupling protein-3 (UCP3) is a mitochondrial inner-
membrane protein abundantly expressed in rodent and
human skeletal muscle which may be involved in energy
dissipation. Many studies have been performed on the
metabolic regulation of UCP3 mRNA level, but little is
known about UCP3 expression at the protein level. Two
populations of mitochondria have been described in skeletal
muscle, subsarcolemmal (SS) and intermyofibrillar (IMF),
which differ in their intracellular localization and possibly
also their metabolic role. To examine if UCP3 is differen-
tially expressed in these two populations and in different
mouse muscle types, we developed a new protocol for
isolation of SS and IMF mitochondria and carefully valid-
ated a new UCP3 antibody. The data show that the density
of UCP3 is higher in the mitochondria of glycolytic muscles
(tibialis anterior and gastrocnemius) than in those of oxi-
dative muscle (soleus). They also show that SS mitochondria
contain more UCP3 per mg of protein than IMF mito-
chondria. Taken together, these results suggest that oxida-
tive muscle and the mitochondria most closely associated
with myofibrils are most efficient at producing ATP. We
then determined the effect of a 24-h fast, which greatly
increases UCP3 mRNA (16.4-fold) in muscle, on UCP3

+
transporter sensitive to nucleotides and fatty
acids [9].
Many studies have been performed on the metabolic
regulation of brown adipose tissue and muscle UCP3
mRNA expression in both rodents and humans (for review
see [3–5]). Very few of the control mechanisms of UCP3
observed in muscle at the mRNA level have so far been
studied at the protein level. The reasons for this are the
difficulty in obtaining specific antibodies and validation
tests and also good Western blot conditions.
Skeletal muscle mitochondria consist of two distinct
subfractions, the subsarcolemmal (SS) and intermyofibrillar
(IMF) mitochondria, located beneath the sarcolemma and
between the myofibrils, respectively. These two mitochond-
rial populations possess different characteristics, such as
higher cardiolipin content and more elevated state-3 respir-
ation rate in IMF mitochondria [10].
In this study, we validated an antibody to UCP3 using a
fully controlled Western blot technique and present a
comparative study in mice of UCP3 protein expression in
SS and IMF mitochondria of various muscle types. We also
examined UCP3 protein expression in the two mitochon-
dria populations of fed and fasted mouse gastrocnemius
muscle.
Correspondence to P. Muzzin, De
´
partement de Biochimie Me
´
dicale,

Three-month-old female C57BL/6J mice fed ad libitum a
standard laboratory chow and maintained under a 12-h
light/dark cycle at 23 °C were used. All animals were caged
individually during the experimental periods. Mice were
either fed ad libitum orfastedforaperiodof24hwithfree
access to water. The animals were killed by cervical
dislocation, and tibialis anterior, gastrocnemius and soleus
muscles were carefully dissected and kept on ice. In fasted
animals and the respective controls, one gastrocnemius
muscle was used for the mitochondria preparation and the
other for RNA isolation. All experiments were performed in
accordance with the Office Ve
´
te
´
rinaire de Gene
`
ve author-
ization covering animal experiments.
Preparation of muscle mitochondria
SS and IMF mitochondria were prepared from skeletal
muscle by the following procedure. Muscle (50–250 mg)
was minced with scissors in 5 mL ice-cold homogenization
buffer containing 100 m
M
sucrose, 180 m
M
KCl, 10 m
M
EDTA, 5 m

M
KCl, 1 m
M
EDTA, 5 m
M
MgSO
4
,1m
M
ATP, 50 m
M
Tris/HCl,
pH 7.4, and 0.06% protease inhibitor cocktail using a
Teflon pestle in an ice-cold glass Elvehjem homogenizer
(clearance 0.12 mm for 3 min, 1800 rpm). The resulting
homogenate was centrifuged at 1600 g for 10 min at 4 °C.
After filtration through two layers of surgical gauze, the
supernatant was centrifuged at 15 000 g for 45 min at 4 °C,
and the resulting IMF mitochondrial pellet was resuspended
in an appropriate volume of distilled water. Mitochondrial
protein concentrations were determined as described by
Bradford [12] using the Bio-Rad Protein Assay, with BSA as
a standard. Isolated mitochondria were stored at )20 °Cas
15-lg mitochondrial protein aliquots.
Western blotting
Purified muscle mitochondria (15 lg) were dried under
vacuum and resuspended in 10 lL loading buffer contain-
ing 50% glycerol, 5% SDS, 2.5% bromophenol blue and
0.5
M

peroxidase-labeled secondary antibody. To compare the
UCP3 signals, linear standard curves were constructed using
increasing concentrations of the human and mouse UCP3
recombinant proteins provided by Dr Michele Chiesi at
Novartis (Basel, Switzerland) and Stratagene (La Jolla, CA,
USA), respectively. The specificity of the antibody to UCP3
was tested using UCP3KO muscle mitochondria, which
were described by Gong et al. [7] and generously provided
by Dr Mary Ellen Harper (University of Ottawa, Ottawa,
Ontario, Canada). The replication-defective recombinant
adenoviral vector that contains the human UCP3 cDNA
under the transcriptional control of the cytomegalovirus
promoter was constructed as previously described [13].
Real-time quantitative RT-PCR
Total muscle RNA was isolated using the Trizol reagent
technique according to the manufacturer’s instructions.
Oligo-dT first-strand cDNA was synthesized from 2 lg
total RNA using Superscript II reverse transcriptase.
Real-time PCR was performed using a Lightcycler rapid
thermal cycler system and designated software (Roche
Diagnostics Ltd, Rotkreuz, Switzerland) according to the
Ó FEBS 2002 UCP3 protein quantitation (Eur. J. Biochem. 269) 2879
manufacturer’s instructions. Reactions were performed in a
20-lL reaction mixture containing 50 ngÆlL
)1
of first-
strand cDNA, 0.5 l
M
primers and 2.4 m
M

separated on a 1.2% agarose/formaldehyde gel and trans-
ferred to nylon membrane as described by Boss et al.[16].
To detect UCP3 mRNA, we used a probe derived from a
full-length rat UCP3 cDNA [15]. The probe was labeled
by random priming with [a-
32
P]dCTP (Amersham).
Hybridization and washing were carried out as previously
reported [16]. Blots were exposed to Hyperfilm ECL films
(Amersham) at )80 °C with intensifying screens. The
signals on the autoradiograms were quantified by scanning
photodensitometry using ImageQuant Software version
3.3. Hybridization of the blots with a [c-
32
P]ATP-labeled
synthetic oligonucleotide specific for the 18S rRNA
subunit was used to correct for differences in the amounts
of RNA loaded on to the gel. Student’s unpaired t test
was used to determine statistical significance.
RESULTS
Validation of antibodies to UCP3
Table 1 shows a list of antibodies that have been used for
the Western blot analysis of UCP3 protein expression in
rodents. The studies performed on UCP3KO or transgenic
mice overexpressing UCP3 in their skeletal muscle provide
convincing validation of the antibodies used. The Lilly
antibody to mouse and rat UCP3 [6] and the Chemicon
antibody to human UCP3 (AB3046) [7] were found to react
specifically with mouse UCP3, as the signal observed in the
wild-type animal was found to be abolished in UCP3KO

UCP3KO – Mouse muscle
Cadenas et al. [17] C-Terminus of human UCP3,
AB3046 (Chemicon)
– hUCP3/HEK293 Rat muscle (starvation)
Zhou et al. [19] C-Terminus of human UCP3,
AB3046 (Chemicon)
– – Rat muscle (exercise, hypoxia,
AMPkinase activation)
Clapham et al. [8] C-Terminus of human UCP3,
UCP32-A (a-Diagnostic)
hUCP3Tg – Mouse muscle
Sivitz et al. [18] C-Terminus of human UCP3,
UCP32-A (a-Diagnostic)
– – Rat muscle (fasting, leptin)
Jucker et al. [20,21] C-Terminus of human UCP3,
UCP32-A (a-Diagnostic)
– – Rat muscle (fasting, T3)
Pedraza et al. [22] Peptide sequence between
TM2 and TM3 of human
UCP3, UCP31-A (a-Diagnostic)
– hUCP3/HEK293 Mouse muscle (lactation)
2880 M. Jimenez et al.(Eur. J. Biochem. 269) Ó FEBS 2002
Muscle mitochondria prepared from UCP3KO mice [7]
were compared with wild-type mitochondria. As shown in
Fig. 1A lanes 1 and 2, the strong signal observed in wild-
type mouse mitochondria was absent from those of
UCP3KO mice. Figure 1A (lane 3) shows that the CabrX
antibody reacts with the mouse recombinant protein. The
size of the signal is higher than 34 kDa because of the
presence of a His

was analyzed using mouse gastrocnemius and tibialis muscle
mitochondria. The mean variation between quadruplicates
for four different samples was 27 ± 7% and 20 ± 6% for
UCP3 and COX, respectively. It should be stressed that
larger and unpredictable variations were observed when
values obtained with a given sample on two different gels
were compared. Therefore we only compared values
obtained on the same gel for all our subsequent quantitative
studies.
Preparation of SS and IMF mitochondria
We developed a technique using selective conditions of
mechanical disruption to prepare SS and IMF mitochon-
dria. As shown in Table 2, the quantity of IMF mitochon-
dria recovered from 1 g gastrocnemius muscle was 1.7-fold
higher than SS mitochondria. The specific and total levels of
COX protein were not significantly different in IMF and SS
mitochondria. The yield in mitochondria, which was
determined by comparing the level of COX protein in the
sum of the two mitochondria populations with that in the
homogenate, was 88%.
Expression of UCP3 protein in SS and IMF mitochondria
of various muscles
Figure 3A illustrates the distribution of UCP3 in SS and
IMF mitochondria obtained from different types of mouse
muscle, i.e. tibialis anterior (two-thirds fast oxidative
glycolytic, one-third glycolytic), gastrocnemius (one-third
slow oxidative, one-third fast oxidative glycolytic, one-third
fast glycolytic) and soleus (90% slow oxidative). It can be
seen in Fig. 3A that the UCP3 protein levels in SS
mitochondria (expressed as arbitrary units per mg mitoch-

Table 2. Recovery of SS and IMF mitochondria from gastrocnemius
muscle. The results are expressed as means ± SEM from the number
of experiments in parentheses.
SS IMF
Protein yield
(mg protein per g muscle)
1.9 ± 0.2 (4) 3.2 ± 0.2 (3)
a
COX protein specific level
(arbitrary units per mg protein)
131 ± 12 (4) 86 ± 26 (3)
COX protein total level
(arbitrary units per g muscle)
239 ± 14 (4) 265 ± 69 (3)
a
P < 0.01 in IMF vs. SS mitochondria.
Ó FEBS 2002 UCP3 protein quantitation (Eur. J. Biochem. 269) 2881
for the gastrocnemius, where the level of COX is lower in
IMF than SS mitochondria by 43% (Fig. 3B). As shown in
Fig. 3C, the UCP3/COX ratio in SS mitochondria is higher
in the gastrocnemius than in the soleus (1.5-fold) and in
IMF mitochondria in the tibialis anterior and in the
gastrocnemius than in the soleus (2.0-fold and 1.9-fold,
respectively). In the gastrocnemius and soleus muscles, the
UCP3/COX ratio is 37% and 41% lower, respectively, in
IMF than SS mitochondria. This indicates that the compo-
sitions of the two mitochondrial populations are different.
UCP3 mRNA was determined by quantitative RT-PCR
in the same muscles to allow a comparison between the
respective UCP3 mRNA and protein levels. As shown in

values; °°°P < 0.005 vs. gastrocnemius values.
Fig. 4. UCP3 mRNA levels in mouse tibialis anterior (TA), gastroc-
nemius (Gn) and soleus (So) muscle. The results, obtained by real-time
quantitative RT-PCR as described in Materials and Methods, are
presented as means ± SEM of values normalized using actin.
*P < 0.05 vs. gastrocnemius values.
2882 M. Jimenez et al.(Eur. J. Biochem. 269) Ó FEBS 2002
determined the UCP3 mRNA levels in gastrocnemius
muscle. We observed that fasting induced a 16.4-fold
(P < 0.001) increase in UCP3 mRNA expression, whereas
it augmented the total amount of UCP3 protein by 1.5-fold
in the gastrocnemius from the same animals.
DISCUSSION
This study first validates a Western blot technique for the
detection and quantitation of UCP3 using a specific and
sensitive antibody raised against 14 amino acids located at
the C-terminus of human UCP3 protein (CabrX). The use
of both mitochondria from knockout mouse gastrocnemius
and homogenates of C
2
C
12
myoblasts infected with an
adenovirus encoding for human UCP3 allowed clear
validation of our Western blot signal. Good cross-reactivity
of the CabrX antibody with mouse recombinant UCP3 was
also demonstrated. The signals can be quantitated and
compared with a reasonable degree of accuracy, but only if
obtained on the same gel.
The density of UCP3 protein (i.e. the intensity of the

the soleus muscle, contains less UCP3 per mg of mitoch-
ondrial protein than glycolytic muscle types and (b) in all
muscle types the mitochondria most closely associated with
the myofibrills, i.e. the IMF mitochondria, have a lower
UCP3 density.
If UCP3 is an uncoupling protein, the oxidative muscle
(soleus) and the IMF mitochondria (which should be most
involved in muscle contraction) would be less prone to
uncoupling and therefore more efficient at producing ATP.
In a recent paper on pig muscle SS and IMF mitochondria,
Lombardi et al. [24] showed that IMF mitochondria had a
higher capacity for ATP production than SS. The possible
role of the lower level of UCP3 in this difference would be
interesting to study.
Most studies on regulation of UCP3 expression have
investigated changes in UCP3 mRNA levels. Fasting has
repeatedly been shown to dramatically increase muscle
UCP3 mRNA in rats and mice [5]. This is surprising
because UCP3, which has been shown to exhibit uncoupling
activity, would be expected to be turned off in muscle under
conditions that dictate energy sparing such as starvation. In
this study, a 24-h fasting period, which was shown to
increase gastrocnemius muscle UCP3 mRNA level 16.4-
fold, was found to induce an increase of 1.5-fold in the
UCP3 protein level of gastrocnemius muscle from the same
animal. Consistent with these observations, Cadenas et al.
[17] and Sivitz et al. [18] reported increases in rat UCP3
protein induced by fasting that were less than half those of
UCP3 mRNA level in experiments carried out in parallel.
These results are in agreement with studies reporting no

muscle. Our data also indicate that SS mitochondria contain
more UCP3 than IMF mitochondria, raising the possibility
that these organelles have different capacities for oxidative
ATP production and a moderate increase in UCP3 protein
content in SS mitochondria of fasted mice.
ACKNOWLEDGEMENTS
This work was supported by grants from the Swiss National Science
Foundation no. 31-53707.98 and 31-54306.98. We are indebted to the
Office Fe
´
de
´
ral du Sport Macolin, the Fonds Euge
`
ne Rapin, the
Fondation du Centenaire de la socie
´
te
´
Suisse d’Assurances ge
´
ne
´
rales sur
la vie humaine pour la sante
´
publique et les recherches me
´
dicales and
the Roche Research Foundation.

Trail, B., Latcham, J., Rastan, S., Harper, A.J., Cadenas, S.,
Buckingham, J.A., Brand, M.D. & Abuin, A. (2000) Mice over-
expressing human uncoupling protein-3 in skeletal muscle are
hyperphagic and lean. Nature (London) 406, 415–418.
9. Echtay,K.S.,Winkler, E.,Frischmuth, K.&Klingenberg, M.(2001)
Uncoupling proteins 2 and 3 are highly active H(+) transporters
and highly nucleotide sensitive when activated by coenzyme Q
(ubiquinone). Proc. Natl Acad. Sci. USA 98, 1416–1421.
10. Cogswell, A.M., Stevens, R.J. & Hood, D.A. (1993) Properties of
skeletal muscle mitochondria isolated from subsarcolemmal and
intermyofibrillar regions. Am. J. Physiol. 264, C383–C389.
11. Krieger, D.A., Tate, C.A., McMillin-Wood, J. & Booth, F.W.
(1980) Populations of rat skeletal muscle mitochondria after
exercise and immobilization. J. Appl. Physiol. 48, 23–28.
12. Bradford, M.M. (1976) A rapid and sensitive method for the
quantitation of microgram quantities of protein utilizing the
principle of protein-dye binding. Anal. Biochem. 72, 248–254.
13. Muzzin, P., Eisensmith, R.C., Copeland, K.C. & Woo, S.L. (1996)
Correction of obesity and diabetes in genetically obese mice by
leptin gene therapy. Proc. Natl Acad. Sci. USA 93, 14804–14808.
14. Higuchi, R., Fockler, C., Dollinger, G. & Watson, R. (1993)
Kinetic PCR analysis: real-time monitoring of DNA amplification
reactions. Biotechnology (NY) 11, 1026–1030.
15. Boss, O., Samec, S., Desplanches, D., Mayet, M.H., Seydoux, J.,
Muzzin, P. & Giacobino, J.P. (1998) Effect of endurance training
on mRNA expression of uncoupling proteins 1, 2, and 3 in the rat.
FASEB J. 12, 335–339.
16. Boss, O., Samec, S., Kuhne, F., Bijlenga, P., Assimacopoulos-
Jeannet, F., Seydoux, J., Giacobino, J.P. & Muzzin, P. (1998)
Uncoupling protein-3 expression in rodent skeletal muscle is

22. Pedraza, N., Solanes, G., Carmona, M.C., Iglesias, R., Vinas, O.,
Mampel, T., Vazquez, M., Giralt, M. & Villarroya, F. (2000)
Impaired expression of the uncoupling protein-3 gene in skeletal
muscle during lactation: fibrates and troglitazone reverse lacta-
tion-induced downregulation of the uncoupling protein-3 gene.
Diabetes 49, 1224–1230.
23. Hesselink, M.K.C., Keizer, H.A., Borghouts, L.B., Schaart, G.,
Kornips, C.F.P., Slieker, L.J., Slopp, K.W., Saris, W.H.M. &
Schrauwen, P. (2001) Protein expression of UCP3 differs between
type 1, type 2a, and type 2b fibers. FASEB J. 15, 1071–1073.
24. Lombardi, A., Damon, M., Vincent, A., Goglia, F. & Herpin, P.
(2000) Characterisation of oxidative phosphorylation in skeletal
muscle mitochondria subpopulations in pig: a study using top-
down elasticity analysis. FEBS Lett. 475, 84–88.
2884 M. Jimenez et al.(Eur. J. Biochem. 269) Ó FEBS 2002