Despite a slight increase in mito-chondrial mass in mtDNA-depleted cells, the mitochondrial protein importactivity was reduced as shown by a decrease in the import of radiolabeledmatrix-
Trang 1a Ca2+-dependent pathway and a reduced mitochondrial protein import
Ludovic Mercy, Aure´lia de Pauw, Laetitia Payen, Silvia Tejerina, Andre´e Houbion,
Catherine Demazy, Martine Raes, Patricia Renard and Thierry Arnould
Laboratory of Biochemistry and Cellular Biology, University of Namur (FUNDP), Namur, Belgium
Keywords
biogenesis; calcium ⁄ CaMKIV; gene
expression; mitochondrial dysfunction;
protein import
Correspondence
T Arnould, Laboratory of Biochemistry and
Cellular Biology, University of Namur
(F.U.N.D.P.), 61 rue de Bruxelles,
mitochond-in mtDNA-depleted cells is also dependent on mitochond-intracellular calcium as itschelation reduces mitochondrial mass Despite a slight increase in mito-chondrial mass in mtDNA-depleted cells, the mitochondrial protein importactivity was reduced as shown by a decrease in the import of radiolabeledmatrix-targeted recombinant proteins into isolated mitochondria and bythe reduced mitochondrial localization of ectopically expressed HA-apo-aequorin targeted to the mitochondria Decrease in ATP content, in mito-chondrial membrane potential as well as reduction in mitochondrial Tim44abundance could explain the lower mitochondrial protein import inmtDNA-depleted cells Taken together, these results suggest that mito-chondrial biogenesis is stimulated in mtDNA-depleted cells and involves acalcium-CREB signalling pathway but is associated with a reduced mito-chondrial import for matrix proteins
Abbreviations
ANT2, adenine nucleotide translocase isoform 2; ATF2, activating transcription factor 2; b-ATPase, beta subunit of Fo-F1-ATPase; CaMKIV, calmodulin-dependent kinase IV; COX I, II, IV and VIII, cytochrome c oxidase subunit I, II, IV and VIII; CPT-1, carnitine palmitoyl transferase-1; CREB, cAMP-responsive element binding protein; cyt c, cytochrome c; DHFR, dihydrofolate reductase; FCCP, carbonyl cyanide
p-trifluoromethoxyphenylhydrazone; HA, hemaglutinin; mtDNA, mitochondrial DNA; MEF2, myocyte enhancer factor 2; mtTFA ⁄ Tfam, mitochondrial transcription factor A; NAO, nonyl acridine orange; NFAT, nuclear factor of activated T cells; NFjB, nuclear factor kappaB; NRF-1 and 2, nuclear respiratory factor-1 and 2; OXPHOS, oxidative phosphorylation; PGC-1a and b, PPARc coactivator-1 a and b; PPARc, peroxisome proliferator-activated TATA-box receptor c; PRC, PGC-1a-related coactivator; R123, rhodamine 123; ROS, reactive oxygen species; Sp1, specificity protein 1; TBP, TATA-binding protein; TNFa, tumor necrosis factor a; TIM, translocase of inner membrane; TOM, translocase of outer membrane; USF-2, upstream stimulatory factor-2; YY1, ying-yang 1; Dwm, mitochondrial membrane protential.
Trang 2Mitochondria play crucial functions in health and
diseases and many mitochondrial disorders, that
mainly affect tissues with high energy demands,
result from mutations or deletions in the
mitochond-rial genome that impair the synthesis of one or more
of the mitochondrial encoded respiratory protein
leading to a decrease in oxidative phosphorylation
(OXPHOS) capacity [1–3] Mitochondrial
prolifer-ation and increase in the expression of respiratory
proteins are a common manifestations found in
patients with mitochondrial myopathies or mtDNA
depletion that is responsible for the so-called
‘ragged-red fibers’ phenotype of skeletal muscle [4] In
addition, several studies have now shown that
mito-chondrial dysfunction leads to the stimulation of
mitochondrial biogenesis For example, muscle from
mouse with myopathy and hypertrophic
cardiomyo-pathy resulting from the targeted inactivation of the
gene encoding the heart muscle isoform of the
aden-ine nucleotide translocator (ANT 1) display abnormal
proliferation of mitochondria [5] In a conditional
knockout mice for mitochondrial transcription factor
A (Tfam), a transcription factor involved in the
regulation of the mitochondrial genome replication
and transcription [6], leading to mtDNA-depletion
and prolonged respiratory chain deficiency, Hansson
et al recently reported that the mitochondrial mass
increases in respiratory chain deficient embryos and
differentiated mouse tissues [7]
If each mammalian cell contains several hundreds
to more than a thousand mitochondria, it is thus
now clear that the size, shape, and abundance of
mitochondria vary dramatically in different cell types
and may change under different energy demand [8]
The abundance of mitochondria in a cell is
deter-mined by division and⁄ or biogenesis of the organelle
[9] that can be defined as a complex biological
pro-cess requiring the synthesis of phospholipids and
cooperative interactions between proteins encoded by
both the nuclear and mitochondrial genes [10,11], the
mitochondrial protein import and their assembly [3]
However, mechanisms leading to mitochondrial
bio-genesis in cells deficient for mitochondrial activity
are still poorly understood
As the protein-coding capacity of mammalian
mtDNA is limited to 13 respiratory subunits that are
necessary for mitochondrial function and integrity,
more than 95% of the genes necessary for
mito-chondrial biogenesis are encoded in the nucleus and
their expression is regulated by the activation of a
small set of specific transcription factors and
signal-ling pathways [9,12,13] The first class of nuclear
transcriptional regulators involved in the biogenesis
of the organelle includes specific DNA-binding scription factors such as nuclear respiratory factors 1and 2 (NRF-1 and NRF-2) that act on the genescoding for constituent subunits of the OXPHOS sys-tem and mtDNA replication [14–17] Other factorssuch as CREB (cyclic-AMP responsive element-bind-ing protein) [18], PPARc (peroxisome proliferatoractivated receptor gamma) [19,20],or the muscle-spe-cific transcription factor MEF2 (myocyte enhancerfactor 2) [21,22] and general factors such as YY-1(ying yang 1) [23], USF-2 (upstream stimulatory fac-tor-2) [24], and Sp1 (specificity protein 1) [25] havebeen described to act as activators or repressors ofnuclear genes encoding mitochondrial proteins andmore particularly proteins involved in the OXPHOScomplexes A second class of regulators containscoactivators that are unable to bind DNA such asPGC-1a (peroxisome proliferator activated receptorgamma coactivator-1alpha) and related family mem-bers (PRC and PGC-1b) [26] These proteins caninteract with DNA-bound transcription factors inorder to coordinate their action in the expression ofgenes essential for cellular energetics and mitochond-rial biogenesis [27] as recently shown in exercise-induced skeletal muscle adaptation [28]
tran-Numerous signalling pathways have been reported toact upstream of these transcriptional regulators involved
in mitochondrial biogenesis by stimulating the sion of nuclear genes encoding respiratory proteins.Firstly, reactive oxygen species (ROS) have been des-cribed to promote expression of cytochromes c1 and bthrough a H2O2-dependent signalling in human cellsthat respond to defective respiratory function [29].Moreover, a treatment of human MRC-5 lung cells withantimycin A that elevated the intracellular ROS produc-tion induced an increase in the mitochondrial mass inthe cells [30] ROS have also been reported to enhancethe expression of nuclear genes involved in mitochon-drial biogenesis such as NRF-1 and Tfam in rho0HeLaS3 cells [31] Secondly, a nitric oxide (NO)-cGMP-dependent pathway has been reported to control mit-ochondrial biogenesis in several mammalian cell types[32] On the other hand, many links do exist between ahigh cytosolic calcium concentration and the increase inmitochondrial biogenesis as a treatment of muscle cellswith A23187 (a calcium ionophore) triggers the expres-sion of cyt c in a PKC-dependent manner [33] Ojuka
expres-et al also demonstrated that intermittent increases incytosolic calcium stimulate mitochondrial biogenesis inmuscle cells and suggested that calcium is the mediatorresponsible for the increase in mitochondrial population
in response to exercise [34,35] Furthermore, rial dysfunction and calcium homeostasis are closely
Trang 3mitochond-interdependent in cell signalling and cell death [36].
Indeed, it was observed that depletion of mtDNA below
a certain level as well as treatment of mammalian cells
with respiratory inhibitors increased steady-state levels
of cytosolic calcium that may change activities of several
Ca2+-dependent transcription factors such as CREB
[37], nuclear factor of activated T-cells (NFAT),
activa-ting transcription factor 2 (ATF2) and nuclear factor
kappa B (NFjB) that increase OXPHOS gene
expres-sion including subunit Vb of cytochrome c oxidase
(COXVb) and thus stimulate mitochondrial biogenesis
[38] A role of calcium and calmodulin-dependent
kinas-es (CaMKs) in the control of mitochondrial biogenkinas-esis
has also been demonstrated in skeletal muscle of
trans-genic mice that over-express a muscle-specific
constitu-tively active form of CaMKIV [39], a kinase we
previously found to be activated in L929 and 143B
mtDNA-depleted cells and responsible for CREB
activation [37]
During the mitochondrial biogenesis process, the
majority of the thousand or more mitochondrial
pro-teins are required to be imported from
nuclear-enco-ded cytosolically synthesized precursors The import of
these proteins is achieved by different mechanisms
known to operate during the import of the two major
classes of mitochondrial proteins such as the
philic proteins with cleavable presequences and
hydro-phobic proteins with multiple internal signals [40] The
mitochondrial protein import involves an important
group of proteins including translocase of the outer
membrane (TOM) such as Tom40, Tom20, Tom70
and translocase of the inner membrane (TIM) such as
Tim22 and Tim23 family members [40] as well as
numerous chaperones such as Hsp70 [41] forming
effectors, adaptors and receptors of the mitochondrial
protein import machinery Several reports mentioned
the significance of the protein import in the rate of
mitochondrial protein import as several stimuli,
inclu-ding contractile activity of skeletal muscle, thyroid
hormone treatment, and muscle differentiation can
alter the expression of the import proteins that
ulti-mately lead to a change in protein import rate and
mitochondrial phenotype [42–46]
Numerous mtDNA-depleted cell lines have been
generated by long-term treatment with ethidium
bro-mide [47,48] or DNA polymerase-c inactivation [49]
to study important mitochondrial defects in
OXPHOS, calcium homeostasis alteration, ROS
pro-duction and more recently resistance to apoptosis
[37,38,50] It is also interesting to emphasize that
mtDNA-depleted cells maintain their ability to
gener-ate mitochondria-like structure and a mitochondrial
membrane potential (Dwm) [51–53] Thus, even if the
mechanisms involved in the mitochondrial biogenesis
of mtDNA-depleted cells are poorly understood, it isnow more evident that mtDNA is not essential forthe biogenesis of mitochondrial-like structure in pro-liferating cells
In this study, to address the question of rial biogenesis in cells depleted of mtDNA, we usedrho-L929 and rho0143 B cells (partially and totallydepleted of mtDNA, respectively) to evaluate the ret-rograde signalling that controls the expression ofnuclear genes encoding mitochondrial proteins andthe activity of mitochondrial matrix-tageted proteinimport We first showed that cells depleted of mtDNAnot only maintain but up-regulate the biogenesis ofmitochondria, as the mitochondrial staining with spe-cific fluroescent dyes and the expression of cyt c areboth increased in these cells We next studied theactivity status of several key transcriptional regulatorsknown to control the biogenesis of mitochondria such
mitochond-as NRF-1⁄ 2, PPARc, MEF2, CREB, Sp1 and YY-1,
as well as the abundance of the coactivator PGC1aand found that CREB is the only overactivated factor
in mtDNA-depleted cells We also showed that CREBregulates cyt c expression and could play a role inmitochondrial biogenesis in mtDNA-depleted cells asthe over-expression of dominant negative mutants(K1-CREB and M1-CREB) decreases both cyt cexpression and nonyl acridine orange (NAO) accumu-lation used to monitor mitochondrial mass in cells.The dependence of mitochondrial biogenesis on intra-cellular calcium in mtDNA-depleted cells was also evi-denced as chelation of intracellular calcium reducesthe abundance of mitochondrial population However,despite a slight increase in mitochondrial populationand cyt c abundance in mtDNA-depleted cells, themitochondrial import activity for matrix proteins isreduced in these cells as we observed a decrease in theimport of radiolabeled matrix-targeted recombinantproteins into isolated mitochondria and a lowermitochondrial localization of ectopically expressedHA-apoaequorin addressed to the mitochondria Wealso clearly showed that lower mitochondrial importfor matrix proteins in mtDNA-depleted cells is associ-ated with a decrease in ATP content, in mitochondrialmembrane potential as well as with a reduction in mit-ochondrial Tim44 abundance, an important effector
of mitochondrial import apparatus Taken together,these results suggest that mitochondrial biogenesisleading to the accumulation of ‘abnormal’ mitochon-dria in mtDNA-depleted cells could be mediated, atleast partly, by a calcium-CREB signalling pathwaybut is associated with a reduced mitochondrial importfor matrix proteins
Trang 4Maintenance of mitochondrial structure
in mtDNA-depleted cell lines
Rho-L929 cells [54] as well as rho0143B cells were
pre-viously characterized by our group for
mtDNA-deple-tion and impaired mitochondrial funcmtDNA-deple-tion [37,51] To
compare mtDNA-depletion in rho-L929 and rho0143B
cells used in this study, COXI expression was
deter-mined by western blot analysis in parental and
mtDNA-depleted cells (Fig 1A) As expected, this
mtDNA-encoded subunit of cytochrome c oxidase is
not expressed in rho0143B and its expression is barely
detectable in rho-L929 cells To investigate whether or
not mtDNA depletion leads to modifications in
mito-chondrial content, the abundance and the morphology
of mitochondria were compared in rho-L929 and L929
cells using transmission electron microscopy (TEM)
(Fig 1B) In rho-929 cells, the morphology of
mito-chondria is clearly different as they appear rounder,
swollen and less dense to electrons as already reported
for several other mtDNA-depleted cell lines [53,55]
When mitochondrial population abundance was
assessed with Mitotracker Red, a specific
mitochond-rial fluorescent probe used for mitochondmitochond-rial mass
detection [56,57], we found a punctuated pattern of
staining that is compatible with a mitochondrial
reten-tion and localizareten-tion of the fluorescent probe in both
mtDNA-depleted and parental cells (Fig 1C)
Quanti-tative analysis of Mitotracker Red accumulation using
spectrofluorimetry also revealed that staining is
dependent on loading time and suggests a slight
increase in rho-L929 cells when assessed after 30 min
in the presence of the dye (supplementary Fig S1)
Similar results were also found for rho0143B cells
stained with Mitotracker Red or NAO, a lipophilic
cation that has a high affinity for mitochondrial
cardi-olipin rich membranes [58] (Figure 7A) We are aware
that in several cell types NAO staining has been
des-cribed recently to be also dependent on the
mitochond-rial membrane potential [59] As the NAO staining is
not reduced, and is even slightly increased in
mtDNA-depleted cells while the mitochondrial membrane
potential is lower in these cells [51,52], these data
sug-gest a higher mitochondrial mass in mtDNA-depleted
cells This statement is also supported by the analysis
of mitochondrial population abundance perfomed by
the quantification of the cell area occupied by
mito-chondria on transmission electron microscopy (TEM)
micrographs from L929 and rho-L929 cells Indeed,
using the nih image software free online (http://rsb
info.nih.gov/nih-image/Default.html) we analysed three
section images (10.7 square inches; magnification
15 600·) taken from random observations and foundthat the surface corresponding to mitochondria repre-sents 11.2 ± 2.3% and 18.8* ± 2.5% (*P < 0.05) forL929 and rho-L929 cells, respectively
mtDNA-deple-by the immunodetection of TBP (B) Electron micrographs of L929 and parental L929 cells showing the presence of rounder shaped mitochondria (arrows) (magnification: 25 200 X) (C) Staining for mitochondrial population with Mitotracker Red in L929, rho- L929, 143B and rho 0 143B incubated with 250 n M of the cationic dye for 30 min and processed for confocal microscopy observation Scale bars ¼ 10 lm and arrows indicate punctuated mitochondrial staining.
Trang 5rho-These data show that the abundance of
mitochond-rial population in cells without mtDNA is maintained
and even slightly increased when compared to parental
cells
Effect of mtDNA depletion on some
mitochondrial markers
As mitochondrial biogenesis is dependent on the
expres-sion of numerous nuclear-encoded genes, we next
deter-mined the expression level of several key mitochondrial
markers that cover energetic pathway or mitochondrial
protein import machinery such as the b-subunit of the
Fo-F1-ATPase (b-ATPase), the adenine nucleotide
translocator isoform 2 (ANT2), COXVb or Tom40
and Tim44 The relative mRNA abundance of
b-ATPase, COXVb and Tim44, determined by
real-time PCR, is significantly up-regulated (> twofold
increase) in rho0143B cells compared to parental cells
(Fig 2A) However, we found that some of these
genes might be variously expressed at the protein level
when assessed on cleared cell lysates Indeed, the
Fo-F1-ATPase b-subunit is similarly expressed in both
143B and rho0143B cells (Fig 2B), a data in agreement
with a previous report showing that the expression of
b-ATPase is unchanged in rho0HeLa S3 and rho0143B
[52] The reason for this discrepancy between mRNA
and protein abundance is unknown but could involve
a post-transcriptional regulation as it has been
pro-posed before for the over-expression of Tfam and
NRF-1 at the transcriptional level that was not
reflec-ted at the protein level in mtDNA-deplereflec-ted cells
[31,60] However, this regulation might also be
rela-tively specific as Tim44 was found to be over-expressed
at the protein level in rho0143 B cells, a data in
accordance with the increase in the messenger RNA
for this marker (Fig 2A)
To discriminate between a transcriptional regulation
and mRNA stabilization in the accumulation of these
transcripts, we next transfected cells with plasmids
encoding chloramphenicol acetyl transferase (CAT)
reporter gene driven by the authentic promoter of the
cyt cor the b-ATPase gene (Fig 2C) CAT activity was
significantly up-regulated (respectively three and sixfold
increase) in rho0143B, a result that is consistent with a
positive transactivation of these genes In order to
make sure that the activation of the cyt c promoter is
really the result of a mitochondrial inhibition and not a
consequence of an indirect long-term cell adjustment
to mtDNA depletion, we tested the effect of
mito-chondrial metabolic inhibitors on the promoter activity
143B cells were first transiently transfected with the cyt
c-CAT plasmid and then incubated for 24 h with 1 lm
antimycin A (a complex III inhibitor) or 10 lmcarbonyl cyanide p-trifluoromethoxyphenylhydrazone(FCCP), a mitochondrial uncoupler that both impairthe OXPHOS We found that cyt c promoter was alsoactivated in response to both treatments (Fig 2D).These results show that mitochondrial activity impair-ment per se is responsible for the up-regulation of cyt cgene expression and several other mitochondrialmarkers while ANT2 does not seem to be over-expressed in mtDNA-depleted cells (Fig 2C)
As cyt c is a common marker used to characterizemitochondrial biogenesis [32] and in order to directlyaddress both the expression and the distribution of theprotein, the endogenous expression of cyt c was firstanalysed by western blotting performed on proteinsextracted from enriched-mitochondrial fractions ofmtDNA-depleted cell lines (Fig 3A) The protein ismore abundant (two- to threefold increase) in themitochondria of both mtDNA-depleted cell lines sug-gesting that not only the protein is over-expressed but
is also imported into mitochondria These data havebeen confirmed by immunostaining of cyt c in thedifferent cell lines (Fig 3B) Quantification of fluores-cence signals in cell sections indicates both over-expres-sion and a wider distribution of the protein inmtDNA-depleted cells (Fig 3C) Taken together, thesedata strongly suggest an over-expression of cyt c thatmight be associated with a more abundant mitochond-rial population in mtDNA-depleted cells The role andthe functional significance of cyt c over-expression inmitochondria of mtDNA-depleted cells is an intriguingobservation that should be addressed in the future.The regulation of mitochondrial marker expression
in mtDNA-depleted cells is a process that mightinvolve the activation of transcription factors described
to control the biogenesis of mitochondria
Reduced expression and activity of NRF-1, NRF-2and Tfam in mtDNA-depleted cells
It has been reported previously that mtDNA-depletedHeLa cells display increased mRNA levels of NRF-1and Tfam genes [61] We thus evaluated NRF-1 andNRF-2, two major transcription factors that controlthe expression of several nuclear genes encoding mito-chondrial proteins [6,14,62] and Tfam expression[60,63,64] Interestingly, while NRF-1 expression isonly slightly reduced in both mtDNA-depleted cells(10–20%), NRF-2 expression is strongly decreased inrho0143B and rho-L929 cells by 60 and 80%, respect-ively (Fig 4A) As both factors have been implicated
in the control of Tfam expression [6] which is known
to regulate mtDNA transcription and replication [65],
Trang 6we thus monitored the expression of Tfam in the
murine cell line [66] and show that this factor is
down-regulated in rho-L929 cells (Fig 4B) These results are
in agreement with data obtained in rho-C2C12 cells[60] and suggest that the activity of NRFs is decreased
in mtDNA-depleted cells To test this hypothesis, cells
of clear lysate proteins (35 lg) prepared from 143B and rho0143B cells using specific antibody to Tim44 and to the F1-ATPase b subunit Equal protein loading between lanes was determined by the immunodetection of a-tubulin (C) Promoter activity of ANT2, cyt c and b subunit
of F1-ATPase determined by CAT activity in transiently cotransfected 143B (white) and rho 0 143B (black) cells with CAT reporter constructs driven by the authentic promoter of these genes and a plasmid encoding b-galactosidase CAT activity (cpm: count per minute) was deter- mined 48 h post-transfection and normalized for b-galactosidase activity Results are expressed in percentages of control cells (n ¼ 4) (*,***): significantly different from control cells with, respectively, P < 0.05, and P < 0.001 (D) Effect of antimycin A and FCCP on the promoter activity of cyt c determined by CAT activity 143 B cells were transiently transfected with a CAT reporter construct driven by the cyt c promoter and were incubated or not (control, CTL) for 6 h with 1 l M antimycin A or 10 l M FCCP CAT activity (cpm) was determined
48 h post-transfection and normalized for b-galactosidase activity Results are expressed in percentages of control cells as means for n ¼ 2.
Trang 7were transiently transfected with luciferase reporter
constructs driven by a minimal TK promoter linked to
either four copies of the binding site for NRF-1 (4X
NRF-1) or the Tfam authentic promoter responsive to
NRF-1 [57] Under these conditions, a dramatic and
highly significant decrease in luciferase activity was
obtained for both constructs in rho0143 B (Fig 4C)
These results suggest that the transactivation mediated
by NRF-1 is reduced in mtDNA-depleted cells As
NRF-1 DNA-binding and activity have been shown to
be positively regulated after phosphorylation by casein
kinase II [67], the activity of this enzyme was assessed
in vitro after immunoprecipitation of the kinase from
L929 and rho-L929 Results show an importantdecrease in the activity of casein kinase II in mtDNA-depleted cells (Fig 4D) Furthermore, in human fibro-blasts, ROS production has also been reported tomediate a retrograde signalling pathway that canenhance the expression of NRF-1 and Tfam mRNA inrho0HeLa or antimycin A-treated cells [29,31] ROSproduction was thus determined in L929 and rho-L929cells using the dichlorofluorescein (DCF) probe Inthese conditions, while antimycin A, used as a positivecontrol, triggers a significant increase in ROS produc-tion in L929, we found that ROS generation wasreduced by almost 40% in rho-L929 cells (Fig 4E) as
A
B
C
Fig 3 Expression of cytochrome c protein
is enhanced in mtDNA-depleted cells (A)
Western blot analysis of mitochondrial cyt c
(mtcyt c) abundance performed on proteins
extracted from mitochondrial-enriched
frac-tions of 143B, rho 0 143B, L929, and
rho-L929 cells Equal protein loading between
mtDNA-depleted and corresponding parental
cell lines was determined by the
immuno-detection of the nuclear-encoded COXIV.
(B) Immunostaining of cyt c and confocal
microscopy analysis performed on
para-formaldehyde-fixed and Triton permeabilized
143B, rho0143B, L929 and rho-L929 cells.
(C) Analysis of fluorescence intensity
per-fomed on cell sections presented in B using
the QUANTIFY software from Leica
Fluores-cence intensity profiles are plotted from A
to B direction for the different cell lines.
Trang 8already reported for other cell lines depleted of
mtDNA [68,69] These cells are also less-responsive to
an antimycin A treatment Taken together, these
results support a lower expression and activity of
NRF-1 and Tfam in mtDNA-depleted cells
Activity of YY1, Sp1, PPARc, MyoD, MEF2 and
CREB in mtDNA-depleted cellsBeside the crucial role of NRF-1 and NRF-2 in theregulation of OXPHOS genes [15], the transcriptionalcontrol of numerous nuclear genes encoding mito-chondrial proteins also involves other transcriptionfactors such as YY1, Sp1, PPARc, MyoD, MEF2 andCREB [13,70–74] Using a sensitive colorimetric assaysystem as previously described for NF-jB [75], we thusmeasured the DNA-binding activity of these transcrip-tion factors to specific synthetic DNA consensussequence in nuclear protein extracts prepared from143B L (L929) and rho0143B cells (Fig 5A) or fromL929 and rho-929 cells (Fig 5B) The amount of Sp1,PPARc and MyoD that binds to DNA is reduced inboth mtDNA-depleted cell lines while MEF2 DNA-binding activity is unchanged in these cells Theseresults suggest that a chronic inhibition of mitochond-rial activity impairs the DNA-binding activity of
mtDNA-of proteins from clear lysates mtDNA-of L929 and rho-L929 cells Equal loading between lanes was determined by the immunodetection of a-tubulin (C) mtDNA-depletion decreases the activity of a NRF-1- responsive synthetic promoter as well as the activity of the authen- tic Tfam promoter 143B and rho 0 143 B cells were transiently cotransfected with 0.25 lg of a CMV ⁄ b-gal expression plasmid and 0.5 lg of the 4X NRF-1-Luc construct or 0.5 lg of the Tfam promo- ter-Luc construct Luciferase activity was determined 24 h post- transfection and normalized for b-galactosidase activity Results are expressed in percentages of control 143B cells as means ± 1 SD for n ¼ 3 (**: significantly different from control cells with
P < 0.01) (D) Casein kinase II activity is reduced in rho-L929 cells The enzyme was immunoprecipitated from cleared lysates of L929 and rho–L929 cells In vitro activity was then determined in the presence of a synthetic peptide and [c- 32 P]ATP as described in the
‘experimental procedures’ Results represent the radioactivity ciated with the substrate and are expressed in cpm as means for two samples The amount of immunoprecipitated kinase in the dif- ferent conditions is shown on the western blot below (E) ROS pro- duction is reduced in rho-L929 cells Cells were incubated for
asso-30 min at 37 C with 5 l M DCF and then incubated or not with
1 l M antimycin A for 60 min Cells were then lysed before associated fluorescence was measured with a spectrofluorimeter Results are expressed in arbitrary fluorescence units as means ± S.D for n ¼ 4 **, ***: significantly different from L929 cells as determined by an ANOVA I and Sheffe´’s contrasts with, respectively,
cell-P < 0.01 and cell-P < 0.001.
Trang 9several key transcription factors involved in the control
of genes encoding mitochondrial proteins
In order to study the activity of YY1 and PPARc
further, cells were transiently transfected with
con-structs encoding a luciferase reporter gene driven bypromoters responsive to these factors (Fig 5C) Thetranscriptional activity of these factors was eitherunchanged (YY1) or slightly decreased (PPARc) inrho0143B cells Using luciferase constructs that are specifically activated by Sp1 or NRF-2, we also observed adecrease in the transcriptional activity of these factors
in both mtDNA-depleted cells (data not shown) On thecontrary, CREB, a transcription factor we previouslyidentified as specifically activated in cells with impairedmitochondrial activity [37], is activated in rho0143Bcells as shown by a 2.5-fold increase in the luciferaseactivity encoded by a reporter construct driven by thea-inhibin promoter that contains several CRE sites [76].Several of the nuclear transcription factors that con-trol the expression of genes encoding mitochondrialproteins are coordinated by PGC-1a, which inducesNRF-1 and NRF-2 expression and coactivates severalmitochondrial regulatory factors such as NRF-1,MEF2 or PPARc [76a] In skeletal muscle, it has beenshown that a p38 MAPK signalling stimulates PGC1aexpression and promotes mitochondrial biogenesis [28].Indeed, consequently to its activation, PGC1a causes
an increase in mRNA for several genes encoding ochondrial proteins such as cyt c, COXII and COXIV,the b-ATPase, CPT-1 and uncoupling proteins (UCPs)
mit-in a cell type-selective manner [19,57,77] In both L929 and rho0143B cell lines, PGC1a expression isdecreased, as shown by a strong reduction of PGC-1apromoter activity and protein abundance analysed bywestern blotting and immunostaining (supplementaryFig S2)
rho-Role of a CREB⁄ CaMKIV pathway in themitochondrial biogenesis of mtDNA-depletedcells
We showed previously that a CaMKIV-CREB ling pathway is specifically activated in cells withimpaired mitochondrial activity [37] and more recentlyseveral studies reported the importance of this pathway
signal-in the regulation of mitochondrial biogenesis signal-in etal muscle [39,78] To test the potential role of CREB
skel-in the biogenesis of mitochondria skel-in mtDNA-depletedcells, we over-expressed K-CREB and M1-CREB, twodominant negative mutants of CREB [79] and meas-ured their effect on the CAT reporter gene driven bythe cyt c promoter, containing two functional CREsites [18] While inhibition efficiency is rather differentfor both dominant negative forms that could beexplained by either their different mechanism of action
or their respective level of expression, the sion of both dominant mutants significantly reduces
over-expres-Fig 5 Effect of mtDNA-depletion on Sp1, PPARc, MyoD and MEF2
DNA-binding activity and transactivation Microwells containing the
DNA probes were incubated with 10 lg of nuclear proteins prepared
either (A) from 143B (white) and rho 0 143B (black) or (B) from L929
(white) and rho-L929 (black) After the colorimetric reaction,
absorb-ance was measured at 490 nm and the results were expressed in
percentages of corresponding controls as means ± 1 SD for n ¼ 3.
(C) Effect of mtDNA-depletion on the transcriptional activity of YY1,
PPARc and CREB 143B and rho0143B cells were transiently
transfected with 0.25 lg of a CMV ⁄ b-gal plasmid and 0.5 lg of
responsive luciferase constructs responsive to either YY1
(Msx2SS-Luc), PPARc (3X-PPRE-TK-Luc) or CREB Luciferase activity was
determined 24 h post-transfection and normalized for
b-galacto-sidase activity Results are expressed as percentages of controls as
means ± 1 SD for n ¼ 3 *, **, ***: significantly different from
corresponding controls as determined by an ANOVA I and Sheffe´’s
contrasts with, respectively, P < 0.05, P < 0.01, and P < 0.001.
Trang 10the activity of cyt c promoter in rho0143B cells
(Fig 6A) The over-expression of K-CREB and
M1-CREB also decreases the expression of endogenous
cyt c in rho-929 cells (Fig 6B,C) Furthermore, the
over-expression of a CaMKIV dominant mutant
(CaMKIVT200A) [80] is also able to repress cyt cexpression in rho–L929 cells (Fig 6B,C), suggestingthat a calcium⁄ CaMKIV-CREB pathway might beinvolved in the induction of cyt c expression inmtDNA-depleted cells As a single mitochondrial pro-tein marker is not enough to characterize mitochond-rial biogenesis, we next used NAO dye to monitortotal abundance of the mitochondrial population inrho0143 B cells that over-express either K-CREB orM1-CREB After 48 h, we consistently found a reduc-tion of about 15–20% in the NAO signal in these cells(Fig 7A), suggesting that global mitochondrial abun-dance can be reduced by inhibiting CREB activity.Taken together, these data suggest that the presence ofmitochondria in mtDNA-depleted cells could bedependent on an active CaMKIV-CREB pathway AsmtDNA-depletion causes a sustained increase in cyto-solic calcium that activates cell signalling such asCaMKIV-CREB or JNK pathway [37,38], and becauseintermittent or sustained increase in cytosolic calcium
of skeletal muscle during exercise results in an increase
A
B
C
Fig 6 Cytochrome c up-regulation is dependent on CREB in
mtDNA-depleted cells (A) cyt c promoter activity in transiently
cotransfected 143B (white) and rho 0 143B (black) cells with
plas-mids encoding K-CREB, M1-CREB, a CREB-sensitive CAT reporter
construct driven by the cyt c promoter and an expression plasmid
encoding the b-galactosidase CAT activity was determined in cell
lysates 48 h post-transfection Substrate-associated radioactivity
(cpm) was normalized for b-galactosidase activity and results are
expressed in percentages of 143B control cells as means ± 1 SD.
for n ¼ 3 (***: significantly different from 143B control cells with
P < 0001; + and + + + : significantly different from rho 0 cells with,
respectively, P < 0.05 and P < 0.001) (B) Representative western
blot image of cyt c expression assessed in L929 and rho-L929 cells
transiently transfected with either plasmids encoding K-CREB,
M1-CREB, CaMKIV(T200A) or a pGL2 empty vector (L929 and
rho-L929 control cells) Equal loading was determined by the
immu-nodetection of a-tubulin (C) Quantification of cyt c expression after
optical density determination of the different signals and
normaliza-tion by the abundance of a-tubulin The mean value in L929 cells
was set as a reference for comparison and results are expressed in
fold-increase as means ± 1 SD (n ¼ 3) *: significantly different
from L929 cells with P < 0.05. Fig 7 Mitochondrial mass is dependent on CREB and calcium in
mtDNA-depleted cells (A) Spectrofluorimetric determination of mitochondrial abundance measured by NAO accumulation in 143B (white) and rho0143B (black) cells transiently transfected with a plasmid encoding K-CREB or M1-CREB for 48 h or (B) in cells incu- bated in the presence of BAPTA (10 l M ) for 72 h Results are expressed in arbitrary fluorescence unit normalized for protein con- tent as means ± 1 SD for n ¼ 3 (A) or means for n ¼ 2 (B).
Trang 11in mitochondria [34,35], we next incubated
mtDNA-depleted cells for 72 h with 10 lm BAPTA, an
intra-cellular calcium chelator, before the mitochondrial
abundance was assessed with NAO staining (Fig 7B)
In these conditions, NAO fluorescence was reduced by
almost 40% in both mtDNA-depleted cell lines,
sug-gesting a decrease in mitochondrial population
Mitochondrial protein import for matrix-targeted
proteins is reduced in mtDNA-depleted cells
As mitochondrial biogenesis in mtDNA-depleted cells
most likely requires mitochondrial protein import, we
next assessed this process by two different
experimen-tal approaches We first adapted an in vitro assay,
that has been used mainly in yeast [81], to
quantita-tively determine the import of radiolabeled
mito-chondrial proteins targeted to the matrix in purified
mitochondria as visualized on the electron micrograph
of enriched mitochondrial fractions (Fig 8A) Two
precursor fusion proteins targeted to the
mitochond-rial matrix, the subunit-9 of the ATPsynthase–
dihidrofolate reductase (DHFR) and a truncated form
of the cytochrome b2 (b2(167)D-DHFR), have been
translated and radiolabeled with [35S]methionine in vitro
(Fig 8B) The mitochondria-associated radioactivity
was then measured on mitochondrial fractions of 143B
and rho0143B cells treated with proteinase K after
the import assay The global mitochondrial importwas reduced by 66% and 85%, respectively, forSu9-ATPase-DHFR and b2(167)D-DHFR proteins(Fig 8C,D) We next wondered if the reduced globalmitochondrial import of matrix proteins in mtDNA-depleted cells could be due to a decrease in theb-barrel Tom40 core of the TOM complex, throughwhich the precursor proteins are passing before beingtransferred to other mitochondrial compartments [82].This is not likely the case as the amount of Tom40 inpurified mitochondria is comparable in rho0 andparental 143B (Fig 8E) despite a strong increase in theexpression of Tom40 in both mtDNA-depleted cell
Fig 8 The mitochondrial protein import is reduced in mitochondria
isolated from mtDNA-depleted cells (A) Electron micrograph of an
enriched mitochondrial fraction prepared for the import assay and
illustrated for mitochondria purified from 143B cells analysed by
transmission electron microscopy Scale bar: 100 nm (B)
Autoradi-ography of the cytochrome b 2 and the ATPase subunit-9 chimeric
proteins translated and radiolabeled in vitro (10% of the output).
b2(167D)-DHFR consists of the first 167 amino acids of the
cyto-chrome b 2 precursor fused to the full-length mouse DHFR by a
lin-ker of two amino acids The cytochrome b2presequence consists
of an amino-terminal matrix-targeting sequence (residues 1–31) and
a sorting sequence (residues 32–80) Su9-DHFR contains the first
66 amino acids of the subunit-9 of Neurospora crassa ATPase
fused to DHFR (C,D) For the mitochondrial protein import assay,
isolated mitochondria (30 lg) from 143B and rho 0 143B cells were
incubated for 10 min at 25 C with reticulocyte lysate containing
35 S-labelled Su9-DHFR (C) or b2(167D)-DHFR (D) The import assay
was then stopped by the addition of 1 l M valinomycin To remove
nonimported preproteins, all samples were treated with proteinase
K (40 lgÆmL)1) for 15 min on ice After mitochondria isolation,
asso-ciated radioactivity was counted Results are presented as
repre-sentative data for three independent experiments and expressed in
c.p.m as means ± 1 SD (E) Western blotting analysis of b-ATPase,
Tom40 and Tim44 abundance performed on mitochondrial purified
fractions of 143B and rho 0 143B cells.
C
D
E
Trang 12lines as shown by western blot analysis of clear cell
lysates and immunostaining (supplementary Fig S3)
In addition, western blot analysis of mitochondrial
fractions prepared from 143B and rho0143B cells
revealed that the abundance of b-ATPase and Tom40
is similar in both cell lines while Tim44, an important
effector of mitochondrial protein import that interacts
with mtHsp70 (mitochondrial heat-shock protein 70)
[41], could only be detected in parental cells (Fig 8E)
While we cannot rule out that the abundance of the
various markers in the mitochondria of
mtDNA-deple-ted cells may result from a different degradation of the
different proteins in the mitochondria of cells depleted
of mtDNA, these results indicate that endogenous
mit-ochondrial proteins might be differentially imported in
the mitochondria of mtDNA-depleted cells
In order to extend our data on the mitochondrialprotein import in mtDNA-depleted cells in situ, theimport activity was also determined in rho-L929 cellstransfected with a cDNA encoding a chimeric proteincontaining HA-tagged apoaequorin and the mito-chondrial presequence of the COXVIII subunit thatspecifically targets the fusion protein to the mitochond-rial matrix [83] Confocal microscopy observations(Fig 9A) and quantitative analysis of fluorescence sig-nals on sections of cells immunostained for HA-taggedapoaequorin and cyt c, used as a mitochondrial mar-ker, revealed a decrease in the colocalization betweenboth proteins in the mitochondria of rho-L929 as evi-denced by the reduced match of fluorescence signalsfound in the overlapping fluorescence profiles for thesecells (Fig 9B)
of HA-apoaequorin in a transfected cell (green), abundance of endogenous cyt c (red), and colocalization (overlay) were then visualized by confocal microscopy after the immunostaining of both proteins (B) Analy- sis of fluorescence intensity perfomed for HA-apoaequorin (green) and cyt c (red) on cell sections using the QUANTIFY software from Leica Fluorescence intensity profiles showing expression level and colocalization are plotted from A to B direction for L929 and rho-L929 cells (representative of about
30 analyses).