Báo cáo khoa học: Frataxin deficiency causes upregulation of mitochondrial Lon and ClpP proteases and severe loss of mitochondrial Fe–S proteins pot

We also addressed the fate of several mitochondrial Fe–S proteins in cardiac tissue of MCK mutants, and showed an overall clearance in protein levels of key mitochondrial Fe–S cluster en

Lon and ClpP proteases and severe loss of mitochondrial Fe–S proteins

Blanche Guillon1, Anne-Laure Bulteau2, Marie Wattenhofer-Donze´3,4, Ste´phane Schmucker3,5, Bertrand Friguet2, He´le`ne Puccio3,4,5,6,7, Jean-Claude Drapier1and Ce´cile Bouton1

1 Institut de Chimie des Substances Naturelles, Centre National de la Recherche Scientifique, Gif-sur-Yvette, France

2 Laboratoire de Biologie et Biochimie Cellulaire du Vieillissement, Universite´ Paris 7, France

3 IGBMC (Institut de Ge´ne´tique et de Biologie Mole´culaire et Cellulaire), Illkirch, France

4 Colle`ge de France, Chaire de ge´ne´tique humaine, Illkirch, France

5 Universite´ Louis Pasteur, Strasbourg, France

6 Inserm, U596, Illkirch, France

7 CNRS, UMR7104, Illkirch, France

Friedreich ataxia (FRDA) is an autosomal recessive

neurodegenerative and cardiodegenerative disease

char-acterized by progressive ataxia and cardiomyopathy

that is associated with deficit in Fe–S enzyme activities

and abnormal cellular iron metabolism [1] FRDA

results from a greatly reduced level of the

mitochon-drial protein frataxin, due to a large GAA repeat

expansion in the gene, which inhibits the transcription

of the frataxin gene through heterochromatin silencing

of the locus [2] Although the exact role of frataxin is still controversial, pathophysiological studies from patient autopsies demonstrated a specific loss of Fe–S protein activities, with accumulation of iron being thought to contribute to an increased level of oxidative stress FRDA mouse models with a tissue-targeted frataxin deficiency have been developed to study the pathophysiology of the disease and the function of frataxin, and to test potential therapeutic agents [3]

Keywords

ClpP; frataxin; Friedreich ataxia; iron-sulfur

cluster; Lon protease

Correspondence

C Bouton, ICSN-CNRS, Avenue de la

Terrasse, 91190 Gif-sur-Yvette, France

Fax: +33 1 69 07 72 47

Tel: +33 1 69 82 30 10

E-mail: Cecile.bouton@icsn.cnrs-gif.fr

(Received 21 October 2008, revised 4

December 2008, accepted 9 December

2008)

doi:10.1111/j.1742-4658.2008.06847.x

Friedreich ataxia (FRDA) is a rare hereditary neurodegenerative disease characterized by progressive ataxia and cardiomyopathy The cause of the disease is a defect in mitochondrial frataxin, an iron chaperone involved in the maturation of Fe–S cluster proteins Several human diseases, including cardiomyopathies, have been found to result from deficiencies in the activ-ity of specific proteases, which have important roles in protein turnover and in the removal of damaged or unneeded protein In this study, using the muscle creatine kinase mouse heart model for FRDA, we show a clear progressive increase in protein levels of two important mitochondrial ATP-dependent proteases, Lon and ClpP, in the hearts of muscle creatine kinase mutants These proteases have been shown to degrade unfolded and dam-aged proteins in the matrix of mitochondria Their upregulation, which was triggered at a mid-stage of the disease through separate pathways, was accompanied by an increase in proteolytic activity We also demonstrate a simultaneous and significant progressive loss of mitochondrial Fe–S pro-teins with no substantial change in their mRNA level The correlative effect

of Lon and ClpP upregulation on loss of mitochondrial Fe–S proteins dur-ing the progression of the disease may suggest that Fe–S proteins are potential targets of Lon and ClpP proteases in FRDA

Abbreviations

DNP, 2,4-dinitrophenylhydrazone; ER, endoplasmic reticulum; FRDA, Friedreich ataxia; MCK, muscle creatine kinase; SDHA, succinate dehydrogenase complex subunit A; Yfh1p, yeast frataxin homolog.

The cardiac conditional model, in which frataxin has

been specifically deleted in striated muscles using a

recombinase expressed under control of the muscle

creatine kinase (MCK) promoter, and the animals

in which are hereafter referred to as MCK mutants,

reproduces important pathophysiological features and

biochemical aspects of the human disease [4,5] These

animals show cardiodegeneration, deficiency of

respira-tory chain complexes I–III and aconitases, and

mito-chondrial iron accumulation, without the presence of

major oxidative stress This model was used to

demon-strate that the deficit in mitochondrial Fe–S cluster

enzyme activities is an early event in FRDA disease,

followed by rapid cardiac dysfunction, whereas

abnor-mal iron accumulation within the mitochondria occurs

at a late stage, pointing to a role of frataxin in Fe–S

cluster biogenesis [5] Since then, many studies have

reported the involvement of frataxin in the maturation

of the Fe–S cluster proteins in yeast [6–10], mammals

[11–14], Drosophila [15] and, recently, bacteria [16] and

plants [17] Frataxin, which has been the focus of

extensive research in the yeast system, specifically

interacts with Fe–S scaffold Isu1⁄ 2 proteins [10,18–20],

and is thought to provide iron for the formation of the

cluster

Evidence is accumulating that proteasome

dysfunc-tion might be associated with cardiomyopathies in

which accumulation of abnormal misfolded proteins

may lead to the formation of potentially toxic

aggre-gates [21] In the mitochondrion, the main organelle

affected in FRDA, various proteases also have

important functions in protein quality control [22]

Among these, the ATP-stimulated Lon protease,

which forms homo-oligomeric complexes, degrades

misfolded and damaged proteins in the matrix space,

similarly to the proteasome function in the cytoplasm

[23] By mediating complete proteolysis, Lon thereby

prevents aggregation and deleterious effects on

mito-chondrial functions A second ATP-stimulated

prote-ase, named ClpP, has also been identified in the

matrix of mammalian cells [24] and associates with

the ATPase ClpX subunits in vitro to effect its

ATP-dependent proteolytic activity [25,26] Although

sub-strate specificity has not been defined yet, several

studies have demonstrated that Fe–S cluster proteins

can be preferential substrates for Lon and⁄ or ClpXP

proteases in different systems [27–30] Indeed, it has

recently been demonstrated in yeast that Fe–S cluster

integrity in proteins is one of the major determinants

of susceptibility to degradation by Pim1, the yeast

homolog of human Lon protease [28] By means of a

proteomic approach using wild-type and Pim-1

mutant strains, the authors identified five Pim-1

substrate proteins, including two Fe–S proteins (the homoaconitase Lys4 and Yjl200c, a putative aconi-tase isozyme) Using an in organello degradation assay, they also demonstrated that improper assembly

of Fe–S clusters on Yjl200c and aconitase (aco1) led

to their increased susceptibility to degradation Inter-estingly, in mammals, mitochondrial aconitase has been identified as a good proteolytic substrate for Lon under mild oxidative conditions [29] Finally, one mutational study of ClpP performed in plants showed that the Rieske Fe–S protein can be a sub-strate for this protease [30]

In this study, we investigated whether mitochondrial Lon and ClpP proteases are regulated in the heart of conditional MCK mice during the progression of the FRDA cardiac disease We found a progressive increase in mitochondrial Lon and ClpP protease expression and activity in cardiac tissues of the MCK mutant over the course of the disease Moreover, the proteases are upregulated through two distinct mecha-nisms, as Lon upregulation is transcriptional, whereas that of ClpP is post-transcriptional, acting either by increasing its protein translation or by decreasing its rate of turnover We also addressed the fate of several mitochondrial Fe–S proteins in cardiac tissue of MCK mutants, and showed an overall clearance in protein levels of key mitochondrial Fe–S cluster enzymes that followed the elevated mitochondrial ATP-stimulated proteolytic activity in this FRDA model

Results

Modulation of Lon and ClpP expression in cardiac muscles of MCK mutants

Using the MCK mutants, we investigated the effect of frataxin deficiency on the possible regulation of the two mitochondrial matrix ATP-dependent proteases

We first evaluated Lon and ClpP mRNA levels by quantitative real-time PCR from total RNA heart extracts of different age groups of control and MCK mutants A significant and progressive increase in Lon mRNA levels was observed in mutant mice between 5 and 10 weeks of age, whereas ClpP mRNA levels were not affected (Fig 1A) Indeed, the Lon protease mRNA level increased  2.5-fold at 5 weeks of age in the hearts of MCK mutants as compared with control mice, and continued to rise, increasing  4-fold at

10 weeks of age We next investigated the possible regulation of both proteases in MCK mutants at the protein level Immunoblot analysis using specific anti-bodies against peptides clearly showed a progressive

increase in Lon protein content, starting at the

interme-diate stage (5 weeks) of cardiomyopathy progression

[5] in the MCK mutants (Fig 1B) Quantitative

immu-noblot analysis revealed  2.6-fold, 5.0-fold and

6.7-fold increases in Lon protein expression at 5, 7 and

10 weeks of age, respectively, in MCK mutants as

compared with control mice Interestingly, the ClpP

protein level was also progressively enhanced, with

 3-fold, 3.5-fold and 4.5-fold increases at 5, 7 and

10 weeks of age in MCK mutants, respectively, despite

no change in mRNA levels (Fig 1A,B)

ATP-stimulated proteolytic activity of ClpP and Lon proteases in heart mitochondria of MCK mutants

To investigate whether the increase in Lon and ClpP protein levels is accompanied by increased protein functionality, ATP-stimulated proteolytic activity, reflecting both ClpP and Lon activities [31], was measured in heart mitochondria from frataxin-deficient mice The cytosolic contamination of mito-chondrial fractions was checked by performing

A

B

Fig 1 Regulation of mitochondrial Lon and ClpP protease expression in the hearts of control and MCK mutant mice (A) Total RNA was isolated from the hearts of control and MCK mutant mice at 3, 5 and 10 weeks of age and used to measure Lon protease and ClpP mRNA levels by quantitative real-time PCR The mRNA expression levels were expressed as fold change between MCK and control samples (value 1) and normalized to 18S ribosomal RNA The experiment was repeated at least three times with independent RNA samples, and the average ± standard deviation of the three replicates is depicted in the bar graphs Statistical analysis was performed using Student’s t-test:

***P < 0.0001 (B) Mitochondrial extracts from the hearts of control (C) and MCK (M) mice at 3, 5, 7 and 10 weeks of age were analyzed

by immunoblotting with antibodies against Lon and ClpP Immunolabeled protein bands of interest were then quantified using a ChemiDoc imaging system and QUANTITY ONE software (BioRad, Marne-La-Coquette, France), and were normalized using antibodies against prohibitin and ATP2 as mitochondrial loading controls *P < 0.01.

immunoblotting using prohibitin as mitochondrial

marker and vinculin and proteasome 20S subunit

as cytosolic markers (Fig 2A) As shown in Fig 2A,

very little cytosolic contamination was observed in

the mitochondrial fraction Values corresponding to

ATP-dependent proteolytic activities in control and

mutant mice (Table S1) were used to calculate

the fold changes in ClpP and Lon protease activities

found in MCK mutant versus control mice at 5, 7

and 8 weeks of age We showed that ClpP⁄ Lon

protease activity, which was low in the heart

mitochondria of 5-week-old MCK mutants, increased

 2-fold to  2.5-fold between 7 and 8 weeks of age

(Fig 2B)

Level of carbonylated proteins in heart mitochondria of MCK mutants at different ages Stimulation of Lon proteolytic activity that may depend on increases in carbonylated proteins has been reported in an in vivo cardiac ischemia–reperfusion model [32] and in yeast frataxin homolog (Yfh1p)-defi-cient yeast cells [33] In MCK mutants, Seznec et al [34] did not find any evidence of increased cellular oxi-dative stress Rather, a reduction in oxidized proteins

in the hearts of MCK mutants was detected from 7

to 10 weeks In this previous report, carbonylated proteins were measured in total extracts of frataxin-deficient mice, leading to an underestimation of pos-sible oxidative stress in mitochondria We therefore performed subcellular fractionation from cardiac tissue

of wild-type and MCK mutant mice in order to detect carbonylated proteins in the mitochondrial protein fractions In mitochondria of control mice, appreciable amounts of oxidized proteins were detected, probably due to the oxidative metabolism of mitochondria under normal conditions [35] The total amount of oxi-dized proteins did not increase in frataxin-deficient mice as compared with control mice at any age tested (Fig 2C) Carbonylated proteins of mitochondrial frac-tions were also quantified by ELISA using carbonylated standards, and the results indicated that their levels were similar in both control and MCK mutant mice at any age and represented < 0.1 nmolÆmg)1of total proteins Therefore, in contrast to the Yfh1p-deficient yeast model of FRDA, increase in Lon proteolytic activity in

Fig 2 Lon ⁄ ClpP protease activity and level of oxidized proteins in the heart mitochondria of MCK mutant mice (A) Cytosolic contami-nation of the mitochondrial fraction was checked by immunoblot-ting with protein extracts (40 lg) from the first (500 g) and second (10 000 g) pellets using antibodies against vinculin, proteasome 20S and prohibitin A representative result of cell fractionation is shown C, control; M, MCK (B) Mitochondrial fractions were pre-pared from the hearts of control and MCK mutant mice at 5, 7 and

8 weeks of age and assayed for ATP-dependent proteolytic activity Two hearts of MCK mutant or control mice were pooled to mea-sure proteolytic activity per dot Each triangle, diamond or square corresponds to the fold change in mitochondrial ATP-dependent proteolytic activity of MCK versus control mice at the different stages indicated Detailed measures can be found in Table S1 (C) Carbonylated proteins of mitochondrial fractions (10 lg) from control (C) and MCK mutant (M) mice were detected after derivati-zation of their carbonyl groups using a solution of dinitrophenylhydr-azine, SDS ⁄ PAGE and immunoblotting using a primary antibody against DNP as described in Experimental procedures A digitized image of Ponceau staining was used to check equal loading of each lane (not shown) Experiments were performed at least three times, and a representative result is shown.

A

B

C

the MCK⁄ FRDA mouse model is not linked to

accumu-lation of oxidized proteins

Regulation of mitochondria-encoded genes in

MCK mutants at different ages

Beside its protease function, it has been reported that,

in vitro, Lon binds to specific regions of the

light-strand and heavy-light-strand promoters of mitochondrial

DNA [36,37], and this was recently confirmed in living

cells, pointing to another function of this protease in

the regulation of mitochondrial DNA replication and

gene expression [38] We therefore examined whether

mitochondria-encoded gene expression was affected in

MCK mutant hearts The mRNA expression profile of

12 mitochondria-encoded genes involved in

com-plexes I, III, IV and V of oxidative phosphorylation

was determined in the hearts of wild-type and MCK

mutant mice at different stages of the disease Apart

from ND3 at 5 weeks, there was no significant decrease

or increase in gene expression of any of the

mitochon-dria-encoded genes tested at 3 and 5 weeks of age in

the hearts of MCK mutants as compared with control

littermate mice (Table 1) In contrast, at a late stage,

five genes of complex I, cytochrome b of complex III

and two genes of complex IV were slightly but

signi-ficantly downregulated in MCK mutants versus

controls

Decrease in mitochondrial Fe–S protein levels in

MCK mutants

The effect of frataxin deficiency on the abundance of

mitochondrial Fe–S cluster-containing proteins was

also examined in the MCK mutants Three Fe–S sub-units of complex I (NDUFS3), complex II (SDHB) and complex III (Rieske) of the respiratory chain were selected for immunoblot analysis, as well as ferrochela-tase, a [2Fe–2S] enzyme required for the last step in heme biosynthesis [39], and the [4Fe–4S] aconitase of the tricarboxylic acid cycle A significant decrease in the protein level of every mitochondrial Fe–S protein tested was observed at 5 weeks in the heart of MCK mutants, where frataxin is completely deleted, as com-pared with control samples (Fig 3A) Quantitative immunoblot analysis revealed a similar pattern in the time course of mitochondrial Fe–S cluster protein loss, whereas expression of the mitochondrial Atp2 b-sub-unit of the ATP synthase complex, which does not contain an Fe–S cluster, was not significantly affected

in frataxin-deficient mice (Fig 3B) At 3 weeks, the ini-tial stage of the disease, in which the only phenotype observed is the specific deficit in Fe–S enzyme activity, very little change in Fe–S protein levels was observed, suggesting a defect in Fe–S cluster assembly The decrease in mitochondrial Fe–S cluster proteins was clearly apparent at 5 weeks ( 40–70% decrease) in the hearts of MCK mutants, a stage corresponding to the beginning of the cardiac dysfunction Downregula-tion further decreased at 7 weeks, residual Fe–S protein levels reaching  30–20% in MCK mutants as compared with controls, and stabilizing at a plateau of

 20% at 10 weeks These results are in agreement with the strong enzymatic deficiency of aconitase and the respiratory chain previously observed Protein expression levels of other subunits of the respiratory chain were also investigated (Fig 4) The level of the hydrophilic succinate dehydrogenase complex sub-unit A (SDHA), the flavoprotein subsub-unit of com-plex II, which is not an Fe–S cluster protein but which requires the Fe–S proteins to be properly folded into the complex [40], was also visibly reduced at 5 weeks

by 40% as compared with control mice The expres-sion of the ND6, which is a partner subunit of Fe–S complex I, was also significantly affected in frataxin-deficient mice (Fig 4A,B)

To determine whether the decrease in mitochondrial Fe–S proteins and partners was due to a transcrip-tional regulation, reverse transcription followed by real-time qPCR was performed in control and MCK mutants at 3, 5 and 10 weeks of age As shown in Fig 5, no significant change in the mRNA expression

of mitochondrial aconitase and several subunits of the respiratory chain (Rieske, ND6 and SDHA) was observed in the hearts of either control or MCK mutant mice, despite a marked reduction in their pro-tein levels in the mutants The ferrochelatase, Ndufs3

Table 1 mRNA expression of mitochondria-encoded genes in the

heart of control versus MCK mutant mice at 3, 5 and 10 weeks of

age Cytb, cytochrome b; COX, cyclo-oxygenase.

Mitochondrial

subunits

mRNA fold change difference (WT ⁄ MCK)

*P < 0.05; **P < 0.001.

and Sdhb transcripts were slightly but significantly

decreased at a late stage in frataxin-deficient mice

(Fig 5) [41], but this cannot explain the drastic change

seen earlier at the protein level in the hearts of MCK

mutants (Fig 3)

Discussion

Disturbances of proteolytic systems have been

associ-ated with various human diseases [42] A defect in these

enzymatic systems usually causes protein aggregation and subsequent cellular damage [43] In the matrix of mitochondria, two ATP-stimulated serine proteases have been identified, namely ClpP and Lon proteases, which participate in the degradation of improperly folded and damaged proteins [23,44] Whereas Lon is a homo-oligomeric complex, human ClpP forms a het-erocomplex in vitro with ClpX, an ATP-dependent AAA+ chaperone [45] High expression levels of Lon and ClpP have been reported in energy-hungry tissues such as skeletal muscle and heart, suggesting an impor-tant mitochondrial function of these proteases in these tissues [46,47] Little is known about Lon and ClpP in mammals, and to date, regulation of these proteases has never been studied in mitochondrial diseases In the present article, we report the first demonstration that frataxin deficiency causes significant upregulation of both mitochondrial Lon and ClpP proteases in the car-diac mouse model for Friedreich ataxia The increase in protease levels started at the mid-stage of the disease, and was rapidly followed by a boost of their proteolytic activity We also show that Lon upregulation and ClpP upregulation in the hearts of MCK mutants operate

A

B

Fig 3 Levels of mitochondrial Fe–S cluster-containing proteins in

cardiac muscle of control and MCK mutant mice (A) Total protein

extracts (20 lg) from control (C) and MCK mutant (M) mice at 3, 5,

7 and 10 weeks of age were analyzed by SDS ⁄ PAGE and

immuno-blot with specific primary antibodies against frataxin, mitochondrial

aconitase, ferrochelatase, three Fe–S subunits of complex I

(NDUFS3), complex II (SDHB) and complex III (Rieske) and the

ATP2 b-subunit of mitochondrial ATP synthase (B) Immunolabeled

protein bands of interest were quantified using a ChemiDoc

imag-ing system and QUANTITY ONE software (BioRad), and were

normal-ized using glyceraldehyde-3-phosphate dehydrogenase (GAPDH),

vinculin or b-tubulin as loading control These experiments were

performed at least three times independently, and representative

data are shown Statistical analysis was performed using Student’s

t-test: **P < 0.001; ***P < 0.0001.

A

B

Fig 4 Levels of SDHA and ND6 respiratory chain subunits in the hearts of control and MCK mutant mice (A) Total protein extracts (20 lg) from control (C) and MCK mutant (M) mice at 3, 5, 7 and

10 weeks of age were analyzed by immunoblot using specific pri-mary antibodies against the ND6 subunit of complex I and the flavoprotein (SDHA) of complex II A representative result of three independent experiments is shown (B) Immunolabeled protein bands of interest were quantified and normalized using vinculin as loading control Statistical analysis was performed using Student’s t-test: **P < 0.001; ***P < 0.0001.

through two distinct mechanisms The increase in Lon

protein level was due to transcriptional regulation, as

the protein increase was mirrored at the mRNA level

The same does not hold true for ClpP, which did not

exhibit a change in transcript level, suggesting

transla-tional or post-translatransla-tional regulation Regarding Lon,

although signals that trigger its upregulation in FRDA

are unknown, some proposals can be put forward The

MCK mutants present signs of endoplasmic reticulum

(ER) stress (H Puccio, unpublished data)

simulta-neously with the increase in Lon expression As it has

been reported that cells subjected to hypoxia or ER

stress exhibit higher Lon mRNA levels [48], it is

tempt-ing to suggest that ER stress leads to Lon protease

upregulation in MCK mutants Besides, depletion of

ATP, which has been shown in tissues of FRDA

patients [49], can lead to regulation of gene expression

in some stressful situations [50,51] As Lon proteolytic

activity is stimulated up to nine-fold by ATP [44], it

can be hypothesized that lack of ATP may compensate for low Lon activity by initiating upregulation of the Lon gene and protein expression Regarding the ClpP protease, very little is known about its physiological function and regulation in mammals One study reported ClpP gene upregulation after accumulation of unfolded proteins within the mitochondrial matrix, which appears to depend on CHOP and C⁄ EBPb ele-ments identified in its promoter [52] However, as the change in ClpP expression in MCK mutants occurred

at a post-transcriptional level, protein accumulation in mitochondria, described by Zhao et al., is not the sig-nal that triggers ClpP protein upregulation in the MCK mutants

Lon protease has been described as a multifunctional enzyme, which behaves like an ATP-stimulated prote-ase, a chaperone or a regulator of mitochondrial DNA replication and gene expression [38,47,53] We have shown that the major activity displayed by Lon in the heart of frataxin-deficient mice was its proteolytic activity, which contrasts with the small change in mito-chondria-encoded gene expression In addition, the prominent accumulation of mitochondria observed in MCK mutants at 6–7 weeks, which becomes excessive

at the final stage of the disease [34], may be related to high Lon expression Indeed, two reports showed that the expression and activity of Lon were increased in cells with enhanced mitochondrial biogenesis [54] and that a population of Lon-deficient cells exhibited fewer mito-chondria [55] Therefore, Lon could, at least in part, be responsible for the prominent accumulation of mito-chondria observed in the hearts of MCK mutants

In the bacterial and yeast systems, it has recently been shown that integrity of Fe–S clusters is a main determinant of susceptibility to Lon and⁄ or ClpP deg-radation [27,28] Interestingly, an important biochemi-cal feature associated with frataxin deficiency in Friedreich ataxia is the specific defect in Fe–S cluster enzyme activities, a very early step in the disease pro-cess [5,14,34] This phenomenon was in part attributed

to imperfect Fe–S protein maturation, as frataxin has been identified as an important component of the Fe–S cluster assembly machinery in mammals and other organisms [6,7,12,15–17] In this study, we have shown that frataxin deficiency causes a severe protein loss for several mitochondrial Fe–S enzymes, contrib-uting to the overall Fe–S deficit in FRDA The decrease in protein level at 5, 7 and 10 weeks was not due to transcriptional regulation, as mRNA levels showed no substantial difference between controls and mutants According to structural studies, Fe–S sub-units and other proteins of complexes I and II actually

fit into each other [40,56] The complex I hydrophobic

A

B

Fig 5 Gene expression profile of mitochondrial Fe–S proteins and

SDHA and ND6 subunits of the respiratory chain in cardiac muscle

of wild-type and MCK mutant mice Total RNA, which was

extracted from the hearts of control and MCK mutant mice at 3, 5

and 10 weeks of age, was used to assess mRNA expression of

genes encoding mitochondrial aconitase, ferrochelatase, NDUFS3,

SDHB and Rieske proteins (A), and ND6 and SDHA proteins (B) by

quantitative real-time PCR The mRNA expression levels in MCK

samples were expressed as fold change over controls (assigned

the value of 1) and were normalized to 18S ribosomal RNA

Experi-ments were performed at least three times, and data are presented

as mean ± standard deviation for three separate experiments

Sta-tistical analysis was performed using Student’s t-test *P < 0.05;

**P < 0.001.

subunits ND1–ND6 form a shell around the Fe–S

protein fragments [57], and SDHA flavoprotein of

complex II dimerizes with the Fe–S protein domain

(Ip), forming a soluble catalytic heterodimer [40]

Interestingly, protein levels of ND6 and SDHA were

also reduced In contrast, protein expression of both

ATP2, the b-subunit of mitochondrial ATP synthase,

and of mitochondrial prohibitin, neither of which

con-tains Fe–S clusters or is related to Fe–S proteins, was

unaffected in mutant mice, indicating that frataxin

deficiency specifically affects Fe–S proteins and their

protein partners These results are reminiscent of studies

showing that Yfh1p-deficient yeast cells undergo

degra-dation of mitochondrial aconitase [6] and that, in plants,

depletion of chloroplastic NifS, another important

com-ponent for Fe–S cluster biogenesis, decreases the

abun-dance of several Fe–S proteins and partners [58] By

statistical analysis, we found that the time course of

mitochondrial Fe–S protein loss in MCK mutants

sig-nificantly correlates with the progressive increase in the

levels of both mitochondrial Lon and ClpP proteases

(Fig S1) Taking all these data together, it is tempting

to speculate that the frataxin-dependent defect in Fe–S

cluster biogenesis leads to the formation of

mitochon-drial apoenzymes, which are recognized as misfolded or

low-stability proteins, and degraded by Lon and⁄ or

ClpP proteases Although it is unlikely, we cannot

exclude the possibility that aggregation of unfolded

Fe–S protein also participates in the decrease of Fe–S

protein assessed by immunoblot Research on the

devel-opment of specific Lon inhibitors has started very

recently [59] and, when they are available, they will be

very useful in discovering whether mitochondrial

prote-ase activation is a deleterious or protective process in

FRDA To date, reports concerning the understanding

of Lon cellular functions suggest a protective role of

Lon against aggregation and intracellular accumulation

of oxidized proteins in mitochondria [29,32] Therefore,

increased Lon and ClpP activity in MCK mutants, by

preventing the accumulation of carbonylated proteins,

may hide increased oxidative damage in MCK mutants

Specific inhibitors of Lon proteases will help to further

characterize the Lon and ClpP cellular functions and

identify whether Fe–S proteins are specific substrates of

Lon protease in FRDA

Experimental procedures

Animals

MCK mutants were generated by crossing mice

homozy-gous for a conditional allele of Frda (FrdaL3 ⁄ L3) with mice

heterozygous for the deletion of Frda exon 4 (FrdaD⁄ +), which carries a tissue-specific Cre transgene under the con-trol of the MCK promoter [5] In this study, mice carrying the transgene and conditional allele (L3⁄ L; MCK+) are called MCK mutants Control mice were littermates having

at least one normal frataxin-expressing allele All methods employed in this work are in accordance with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication

No 85-23, revised 1996)

Preparation of total protein extracts Using a glass homogenizer (Duall 21), a tissue homogenate was prepared from the heart of wild-type and MCK mutant mice in 100 mm Tris (pH 7.4) and 150 mm NaCl in the pres-ence of protease inhibitors (protease inhibitor cocktail set III; Calbiochem, Darmstadt, Germany) After centrifu-gation at 400 g for 10 min at 4C, red blood cells were lysed

by resuspending the cell pellet in 800 lL of hypotonic solu-tion (Sigma, Saint-Quentin Fallavier, France) The resulting white pellet was then lysed in 100 mm Tris (pH 7.4) contain-ing 0.5% Triton X-100, with the protease inhibitor cocktail After centrifugation at 10 000 g for 10 min at 4C, the supernatant was collected and protein content was deter-mined before storage at)80 C for further measurements

Isolation of cardiac mitochondria Hearts of wild-type and MCK mutant mice were homo-genized in 210 mm mannitol, 5.0 mm Mops, 70 mm sucrose, and 1.0 mm EDTA (pH 7.4) The homogenate was centri-fuged at 500 g for 10 min at 4C to remove tissue frag-ments and cell nuclei, and the supernatant was recentrifuged at 10 000 g for 10 min at 4C to bring down the mitochondrial pellet After two washings in homoge-nized buffer, the mitochondrial pellets were either stored at )80 C for subsequent immunoblot analysis or immediately used for the measurement of ATP-dependent proteolytic activity as well as detection of carbonylated proteins

Immunoblot analysis Total protein extract (20–40 lg) or mitochondrial fraction (60 lg) was loaded on 10% or 12% SDS⁄ PAGE gel (depending on the expected protein size), and proteins were transferred onto a Hybond nitrocellulose membrane Spe-cific proteins were detected by immunoblotting with the indicated primary antibodies and peroxidase-conjugated secondary antibody (DAKO Cytomation, Glostrup, Denmark) Blots were developed with an enhanced chemilu-minescence detection system [Super Signal Pierce (Perbio Science, Berbie`res, France) or Immobilon Western (Milli-pore, Saint-Quentin en Yvelines, France)]

Immunochemical reagents

Rabbit polyclonal anti-peptide Lon protease serum was

kindly provided by L Sweda (Case Western Reserve

Uni-versity, Cleveland, OH, USA) Mouse monoclonal IgG

against complex I subunit NDUFS3 (NADH

dehydroge-nase ubiquinone Fe–S protein), complex II subunit b (one

Fe–S subunit of complex II), Rieske Fe–S protein subunit

of complex III and the flavoprotein subunit of complex II

were from MitoSciences (Eugene, OR, USA) Monoclonal

anti-mitochondrial DNA-encoded ND6 subunit IgG

(Mito-Sciences) was kindly provided by G Dujardin (Centre de

Ge´ne´tique Mole´culaire, Gif-sur-Yvette, France) Rabbit

polyclonal sera were raised against synthetic peptides

corre-sponding to amino acids 45–61 of mouse mitochondrial

aconitase and to amino acids 410–422 of mouse

ferrochela-tase The serum against frataxin used was the 1250

poly-clonal serum [5] Serum against ClpP raised against the

peptide corresponding to amino acids 65–80 of human

ClpP was kindly provided by L Sweda (Oklahoma Medical

Research Foundation, Oklahoma City, OK, USA) Sera

against vinculin (Sigma) and prohibitin (Neomarkers,

Fremont, CA, USA) were used as loading controls Serum

against the ATP2 b1-subunit of yeast ATP synthase was

kindly provided by J Velours (Institut de Biochimie et

Ge´ne´tique Cellulaire, Bordeaux, France)

ATP-stimulated protease activity

ATP-stimulated Lon and ClpP protease activity was

deter-mined using casein–fluorescein isothiocyanate (0.5 lgÆlL)1)

as substrate [31] Mitochondrial pellets (50–150 lg) were

incubated in assay buffer containing 50 mm Tris (pH 7.9),

10 mm MgCl2, 1 mm dithiothreitol and 0.1% Triton X-100

in the presence or absence of 8 mm ATP Proteolysis of

fluorescein isothiocyanate-labeled casein was then

per-formed for 90 min at 37C At incubation times of

0–90 min, an aliquot was collected and proteins were

pre-cipitated by adding 8% trichloroacetic acid After

centrifu-gation at 15 000 g for 30 min, supernatant containing

fluorescent peptides was recovered and neutralized by

add-ing sodium borate (0.6 m final concentration, pH 10)

Fluo-rescence was then measured with excitation⁄ emission

wavelengths of 495⁄ 515 nm Activities were expressed as

fluorescence units⁄ min ⁄ mg protein, and a ratio between the

ATP-dependent proteolytic activities found in MCK and

wild-type mice was calculated

Carbonylated protein detection

Carbonylated proteins were detected using the OxyBlot

protein oxidation detection kit according to the

manufac-turer’s protocol (Chemicon International, Temecula, CA,

USA) Briefly, mitochondrial pellets were lysed in 100 mm

Tris (pH 7.5), 0.5% Triton X-100 and 50 mm dithiothreitol After centrifugation, clear lysate was denatured by adding 6% SDS, and the carbonyl groups in proteins were deriva-tized to 2,4-dinitrophenylhydrazone (DNP) using 1· dini-trophenylhydrazine solution for 15 min at 25C After neutralization of the reaction, DNP-derivatized proteins were loaded on a 10% SDS⁄ PAGE gel, transferred to a nitrocellulose membrane, and detected with a primary anti-body specific to the DNP moiety Carbonylated proteins in the mitochondrial fractions from hearts of control and MCK mutant mice were measured using the protein car-bonyl ELISA kit according to the manufacturer’s instruc-tions (BioCell Corp., Auckland, New Zealand)

Quantitative real-time PCR analysis Total RNA was extracted using the SV Total RNA Isolation System kit (Promega, Charbonnie`res-les-bains, France), and

4 lg of total RNA was reverse-transcribed using the High Capacity cDNA Archive Kit (Applied Biosystems, Courta-boeuf, France) according to the manufacturer’s instructions Quantitative real-time PCR was performed using the Roche Light Cycler system and the FastStart DNA master plus SYBR green I kit (Roche Applied Sciences, Meylan, France) Lon protease (forward, 5¢-AGGATCTTGCCTTGTGT GGA-3¢; reverse, 5¢-TGGATGAGGAGCTGAGCAAG-3¢) ClpP (forward, 5¢-CACACCAAGCAGAGCCTACA-3¢; reverse, 5¢-CCCAGCAGAGGAAGTTTCAG-3¢), mitochon-drial aconitase (forward, 5¢-AGGAGTTTGGCCCTGTA CCT-3¢; reverse, 5¢-GCCTTGAATGGTCAGCTTGT-3¢), NDUFS3 (forward, 5¢-CTGTGGCAGCACGTAAGAAG-3¢; reverse, 5¢-ACTCATCAAGGCAGGACACC-3¢), SDHB (forward, 5¢-GGAGGGCAAGCAACAGTATC-3¢; reverse, 5¢-GCGTTCCTCTGTGAAGTCGT-3¢), Rieske (forward, 5¢-TGGTCTCCCAGTTTGTTTCC-3¢; reverse, 5¢-GCAGC TTCCTGGTCAATCTC-3¢) and SDHA (forward, 5¢-CAG AAGTCGATGCAGAACCA-3¢; reverse, 5¢-CGACCCGCA CTTTGTAATCT-3¢) sequence-specific primers were designed to span intron–exon boundaries to generate ampli-cons of approximately 100 bp Values were normalized to the relative amounts of 18S ribosomal cDNA (forward, 5¢-CTGAGAAACGGCTACCACATC-3¢; reverse, 5¢-CGCT CCCAAGATCCAACTAC-3¢) Sequences of mitochondrial gene primers are listed in Table S2

Acknowledgements

We wish to thank Alexandre Diet, Aurelien Bayot, Agnieszka Malinowska and Fabienne Pierre for techni-cal support, and Laurence Reutenauer for the genera-tion and genotyping of all the mice for this project This work was supported by funds from the French National Agency for Research (ANR-05-MRAR-013-01) and the French Medical Research Foundation

(H Puccio) M Wattenhofer-Donze´ is an ATER at

the Colle`ge de France

References

1 Puccio H & Koenig M (2002) Friedreich ataxia: a

para-digm for mitochondrial diseases Curr Opin Genet Dev

12, 272–277

2 Campuzano V, Montermini L, Molto MD, Pianese L,

Cossee M, Cavalcanti F, Monros E, Rodius F, Duclos

F, Monticelli A et al (1996) Friedreich’s ataxia:

autoso-mal recessive disease caused by an intronic GAA triplet

repeat expansion Science 271, 1423–1427

3 Puccio H (2007) Conditional mouse models for

Friedreich ataxia, a neurodegenerative disorder

associ-ating cardiomyopathy Handb Exp Pharmacol 178,

365–375

4 Rotig A, de Lonlay P, Chretien D, Foury F, Koenig

M, Sidi D, Munnich A & Rustin P (1997) Aconitase

and mitochondrial iron–sulphur protein deficiency in

Friedreich ataxia Nat Genet 17, 215–217

5 Puccio H, Simon D, Cossee M, Criqui-Filipe P, Tiziano

F, Melki J, Hindelang C, Matyas R, Rustin P &

Koenig M (2001) Mouse models for Friedreich ataxia

exhibit cardiomyopathy, sensory nerve defect and Fe–S

enzyme deficiency followed by intramitochondrial iron

deposits Nat Genet 27, 181–186

6 Chen OS, Hemenway S & Kaplan J (2002) Inhibition of

Fe–S cluster biosynthesis decreases mitochondrial iron

export: evidence that Yfh1p affects Fe–S cluster

synthe-sis Proc Natl Acad Sci USA 99, 12321–12326

7 Muhlenhoff U, Richhardt N, Ristow M, Kispal G &

Lill R (2002) The yeast frataxin homolog Yfh1p plays a

specific role in the maturation of cellular Fe⁄ S proteins

Hum Mol Genet 11, 2025–2036

8 Zhang Y, Lyver ER, Knight SA, Pain D, Lesuisse E &

Dancis A (2006) Mrs3p, Mrs4p, and frataxin provide

iron for Fe–S cluster synthesis in mitochondria J Biol

Chem 281, 22493–22502

9 Muhlenhoff U, Gerber J, Richhardt N & Lill R (2003)

Components involved in assembly and dislocation of

iron–sulfur clusters on the scaffold protein Isu1p

EMBO J 22, 4815–4825

10 Ramazzotti A, Vanmansart V & Foury F (2004)

Mitochondrial functional interactions between frataxin

and Isu1p, the iron–sulfur cluster scaffold protein, in

Saccharomyces cerevisiae FEBS Lett 557, 215–220

11 Yoon T & Cowan JA (2003) Iron–sulfur cluster

biosyn-thesis Characterization of frataxin as an iron donor for

assembly of [2Fe–2S] clusters in ISU-type proteins

J Am Chem Soc 125, 6078–6084

12 Stehling O, Elsasser HP, Bruckel B, Muhlenhoff U &

Lill R (2004) Iron–sulfur protein maturation in human

cells: evidence for a function of frataxin Hum Mol

Genet 13, 3007–3015

13 Seznec H, Simon D, Bouton C, Reutenauer L, Hertzog

A, Golik P, Procaccio V, Patel M, Drapier JC, Koenig

M et al (2005) Friedreich ataxia: the oxidative stress paradox Hum Mol Genet 14, 463–474

14 Martelli A, Wattenhofer-Donze M, Schmucker S, Bouvet S, Reutenauer L & Puccio H (2007) Frataxin

is essential for extramitochondrial Fe–S cluster proteins in mammalian tissues Hum Mol Genet 16, 2651–2658

15 Anderson PR, Kirby K, Hilliker AJ & Phillips JP (2005) RNAi-mediated suppression of the mitochondrial iron chaperone, frataxin, in Drosophila Hum Mol Genet

14, 3397–3405

16 Layer G, Ollagnier-de-Choudens S, Sanakis Y & Fonte-cave M (2006) Iron–sulfur cluster biosynthesis: charac-terization of Escherichia coli CYaY as an iron donor for the assembly of [2Fe–2S] clusters in the scaffold IscU J Biol Chem 281, 16256–16263

17 Busi MV, Maliandi MV, Valdez H, Clemente M, Zabal-eta EJ, Araya A & Gomez-Casati DF (2006) Deficiency

of Arabidopsis thaliana frataxin alters activity of mito-chondrial Fe–S proteins and induces oxidative stress Plant J 48, 873–882

18 Gerber J, Muhlenhoff U & Lill R (2003) An interaction between frataxin and Isu1⁄ Nfs1 that is crucial for Fe⁄ S cluster synthesis on Isu1 EMBO Rep

4, 906–911

19 Foury F, Pastore A & Trincal M (2007) Acidic residues

of yeast frataxin have an essential role in Fe–S cluster assembly EMBO Rep 8, 194–199

20 Wang T & Craig EA (2008) Binding of yeast frataxin to the scaffold for Fe–S cluster biogenesis, Isu J Biol Chem 283, 12674–12679

21 Powell SR (2006) The ubiquitin–proteasome system in cardiac physiology and pathology Am J Physiol Heart Circ Physiol 291, H1–H19

22 Koppen M & Langer T (2007) Protein degradation within mitochondria: versatile activities of AAA prote-ases and other peptidprote-ases Crit Rev Biochem Mol Biol

42, 221–242

23 Ngo JK & Davies KJ (2007) Importance of the lon pro-tease in mitochondrial maintenance and the significance

of declining lon in aging Ann NY Acad Sci 1119, 78– 87

24 de Sagarra MR, Mayo I, Marco S, Rodriguez-Vilarino

S, Oliva J, Carrascosa JL & Castan JG (1999) Mito-chondrial localization and oligomeric structure of HClpP, the human homologue of E coli ClpP J Mol Biol 292, 819–825

25 Santagata S, Bhattacharyya D, Wang FH, Singha N, Hodtsev A & Spanopoulou E (1999) Molecular cloning and characterization of a mouse homolog of bacterial ClpX, a novel mammalian class II member of the Hsp100⁄ Clp chaperone family J Biol Chem 274, 16311–16319