| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
and c-Erb A1 proteins
* UMR-866 Différenciation Cellulaire et Croissance (INRA-UMII-ENSAM), Unité d’Endocrinologie Cellulaire, Institut National de la Recherche Agronomique (INRA), 34060 Montpellier Cedex 1, France; and
EMI INSERM 0014, Laboratoire de Microscopie Electronique, Faculté de Médecine, BP 184, 54505 Vandoeuvre-lès-Nancy Cedex, France
1Correspondence: UMR-866 Différenciation Cellulaire et Croissance, Unité d’Endocrinologie Cellulaire, Institut National de la Recherche Agronomique, 2 place Viala, 34060 Montpellier Cedex 1, France. E-mail: cabello{at}ensam.inra.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(mt-RXR), another nuclear receptor, is located in the mitochondrial matrix. We found that mt-RXR is produced after cytosolic or intramitochondrial enzymatic cleavage of the RXR nuclear receptor. After mitochondrial import and binding to specific sequences of the organelle genome, mt-RXR induces a ligand-dependent increase in mitochondrial RNA levels. mt-RXR physically interacts with p43 and acts alone or through a heterodimerical complex activated by 9-cis-retinoic acid and T3 to increase RNA levels. These data indicate that hormonal regulation of mitochondrial transcription occurs through pathways similar to those that take place in the nucleus and open a new way to better understand hormone and vitamin action at the cellular level.—Casas, F., Daury, L., Grandemange, S., Busson, M., Seyer, P., Hatier, R., Carazo, A., Cabello, G., Wrutniak-Cabello, C. Endocrine regulation of mitochondrial activity: involvement of truncated RXR and c-Erb A1 proteins.
Key Words: mitochondrial import • mitochondrial transcription • mt-DNA • triiodothyronine • 9-cis-retinoic acid
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Several studies have established that organelle activity is involved in regulation of cell proliferation and differentiation independently of ATP production (2
3
4
5)
. The physiological relevance of this reevaluation has been emphasized by the observation that disruption of the mt-TFA gene (tfam) encoding a constitutive mitochondrial transcription factor led to embryonic lethality (6)
. It is also well supported by the extreme severity of diseases induced by rearrangements of the mitochondrial genome. This set of data highlights the requirement for tight regulation of mitochondrial activity for proper development.
Given this physiological importance of the organelle, the observation that several hormones regulate mitochondrial activity was not unexpected (7
8
9
10
11
12
13
14
15)
, but requires new studies in order to characterize the pathways involved in this regulation. This knowledge could bring new ways to better understand mitochondrial deficiencies associated with degenerative diseases. Characterization of hormonal receptors in the organelle (16)
has provided new insight into the regulation of mitochondrial genome transcription. p43, a truncated version of the triiodothyronine nuclear receptor c-Erb A 1 (NR1A1 according to the Nuclear Receptors Nomenclature Committee, 1999), acts as a triiodothyronine (T3)-dependent transcription factor of the organelle (17)
. p43 bound to specific mitochondrial DNA sequences and increased RNA steady-state levels in the organelle, with a latency period of <5 min (17)
. We found that p43 bound to a DR2 sequence located in the mitochondrial D-loop by forming a complex with a 45 kDa truncated form of another member of the nuclear receptor superfamily, PPAR
2 (18)
. These data enabled us to clarify the direct regulation of mitochondrial transcription by thyroid hormone (10
, 13
, 19)
and provided an additional explanation of the thyromimetic influence of fibrates reported in other studies (8
, 20)
. This observation, which already suggested that mitochondrial genome transcription could be regulated in a similar way to nuclear transcription, also raised the question of the identity of proteins interacting with p43 on four other response elements recorded in the mitochondrial genome sequence (21)
.
In the nucleus, RXR receptors display two kinds of activity. They act as 9-cis-retinoic acid-dependent transcription factors that bind DR1 response elements found in the promoter of target genes. However, they are also major partners for other nuclear receptors such as VDR, RAR, PPAR, and T3R by forming heterodimers and supporting the transcriptional activity of each specific ligand (22
, 23)
. A striking observation is that disruption of the RXR gene induces embryonic lethality associated with cardiac failure and alterations in mitochondrial gene expression (24)
and that an adipocyte targeted RXR suppression in 4-wk-old mice induces severe hypothermia during fasting (25)
. Conversely, retinoic acid up-regulates mitochondrial gene expression (26
, 27)
; vitamin A, the precursor of naturally occurring retinoids, influences mitochondrial activity in rat liver and heart (7
, 11)
. All these data implying a retinoid influence at mitochondrial level led us to search for the presence of RXR-related proteins in the organelle.
We report here the existence of a 44 kDa protein related to RXR in the mitochondrial matrix. This protein, produced by enzymatic cleavage of the RXR nuclear receptor, is imported into mitochondria, binds to specific sequences of the organelle genome, and induces alone or through a heterodimerical complex a ligand-dependent increase in mitochondrial RNA levels.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Mitoplast, outer membrane, intermembrane, and matrix fractions were obtained as described previously (29)
. The purity of submitochondrial fractions (data not shown) was tested by measuring the specific activities of monoamine oxidase (30)
(outer membrane), malate dehydrogenase (30)
(matrix), and cytochrome oxidase (inner membrane).
Western blot analysis and antibodies
Mitochondrial extracts (50 µg) were electrophoresed onto 10% SDS-PAGE gels and blotted onto PDVF membranes. Mouse RXR protein used as control was synthesized in vitro using reticulocyte lysate (Promega, Madison, WI, USA). The presence of RXR polypeptides was assessed with either the polyclonal antibody RPRX(A) raised against an RXR amino terminus epitope or two monoclonal antibodies (4RX-3A2, 1RX-6G12) raised against epitopes in the RXR D-E domains (31)
. Signals were revealed using a chemiluminescent Western blot procedure (ECF kit, Amersham, Arlington Heights, IL, USA) and analyzed with a PhosphorImager (Molecular Dynamics, Sunnyvale, CA, USA).
Electron microscopy
Electron microscopy was performed on Wistar liver preparations (ultramicrotomy and immunolabeling). Specimens were fixed with 4% paraformaldehyde/0.1% glutaraldehyde in 0.1 M sodium cacodylate, pH 7.4, for 3 h at 4°C and rinsed in the same buffer. After dehydration in a graded series of ethanol, liver samples were embedded in Bioacryl (Tebu, France) overnight at room temperature. Polymerization was performed under UV light for 72 h at 4°C. Ultrathin frozen sections were collected on formvar-coated nickel grids. Sections were rehydrated in 0.1 M sodium cacodylate, pH 7.4, then incubated overnight at 4°C with 4RX-3A2 antibody (diluted 1/5000 in 0.1 M sodium cacodylate, pH 7.4). After several washes, sections were incubated with gold-conjugated protein A (diluted 1/100) for 1 h. They were then washed and fixed in 1% glutaraldehyde in 0.1 M sodium cacodylate, pH 7.4. Sections used as negative controls were incubated with preabsorbed 4RX-3A2 antibody (diluted 1/5000 in 0.1 M sodium cacodylate, pH 7.4). Ultrathin sections were slightly contrasted with uranyl acetate and examined with a Philips CM12 electron microscope.
RXR cleavage experiments
For calpain cleavage experiments, 3 µL of in vitro translated protein product was mixed with m-calpain (Sigma, St. Louis, MO, USA) and 5 µL of 400 mM Tris-HCL pH 7.5/24 mM CaCl2/2 mM DTT at 30°C as described (32)
. After 30 min, reactions were stopped by the addition of 2x Laemmli buffer. The resulting product was run on SDS gels and analyzed using a PhosphorImager (Molecular Dynamics).
Mitochondrial import
Import experiments were performed as described previously (17)
using highly purified isolated rat liver mitochondria. After import experiments, mitochondria were treated with proteinase K for 10 min at 4°C; 10% of the amount of reticulocyte lysate added to mitochondria for import experiments was loaded in the control lane. Mitoplasts were prepared by digitonin treatment, mitochondria and rabbit reticulocyte lysate were depleted in ATP and ADP by apyrase, and the organelle membrane potential was decreased by addition of 1 µM fluoryl cyanide m-chlorophenylhydrazone (FCCP) in the incubation medium.
EMSA experiments
EMSA were performed as described previously (16)
, with [32P]-labeled oligonucleotide probes corresponding to the mitochondrial T3RE-like sequences identified on the Rattus norvegicus mitochondrial genome. These sequences were 12S (acgttAGGTCAAGGTGTaggc); 16S (agcgCGACCTatttaagAGTTCAtatc); DR2 (gtcaAGGCATgaAGGTCAgcac); RSV-TRE (ttgaTGCCTTcctcaacatagccgtcAAGGCAtgaag); IP6 (ccaaTGACCTaaaaccAGGTGActtc). Binding specificity was assessed by the parallel use of an unrelated probe (agcttctctgtgatttaatgccagcgcg). Highly purified mitochondrial protein extracts were prepared on heparin agarose columns, as described (16)
. The presence of mt-RXR and p43 in the binding complex was assessed respectively with monoclonal antibodies 4RX-3A2 and LA038 (Quality Biotech, Camden, NJ, USA) raised against a specific sequence of the DNA binding domain of c-Erb A. An antibody raised against GST was used as a control.
Coimmunoprecipitation experiments
Immunoprecipitations of mitochondrial proteins were performed with antisera raised against RXR (4RX-3A2) (31)
, c-Erb A (RHTII) (16)
, and ADP/ATP translocator. Purified mitochondrial matrix extracts (200 µg) were mixed with 1 µL of antiserum and 300 µL of binding buffer (HEPES 50 mM, NaCl 400 mM, Nonidet P40 1%, Aprotinin 1 µg/mL, phenylmethylsulfonyl fluoride 100 µg/mL) for 3 h at 4°C. Samples were incubated with 30 µL of protein G-Sepharose for 1 h at 4°C and washed four times with binding buffer. Antibodies specificities were assessed using [35S]-p43, [35S]-RXR/Cp, or [125I]-ANT.
In organello mitochondrial transcription assays and Northern blot analysis
As described previously (33)
, transcription assays were performed using isolated rat liver mitochondria at a final concentration of 2 mg of protein/mL in 0.5 mL of incubation buffer (sucrose 25 mM, sorbitol 75 mM, KCl 100 mM, K2HPO4 10 mM, EDTA 0.5 mM, MgCl2 5 mM, ATP 1 mM, glutamate 10 mM, malate 2.5 mM, Tris pH7.4 10 mM, BSA 1 mg/mL). Incubation was at 37°C for 60 min in the presence of 5 µL of rabbit reticulocyte lysate (containing RXR/Cp, p43 or unprogrammed lysate for control). Mitochondrial RNA was extracted twice at room temperature (33)
, then analyzed by electrophoresis through 1.4% formaldehyde-agarose gels and transferred onto a nylon membrane. Membranes were prehybridized at 65°C for 30 min (0.5 M phosphate buffer, 1 mM EDTA, 7% SDS, 1% BSA), and [32P]dCTP-labeled DNA probes were added. Specific DNA probes for the mouse mitochondrial cytochrome c oxydase subunit III (COX III) and NADH subunit I (ND I) were generated by PCR (17)
. Hybridization was performed at 65°C for 24 h. After hybridization, membranes were washed according to the following protocol: 2x SSC buffer, 5 min at 65°C (25 mM phosphate buffer, 0.25% SDS and where 1x SSC is 150 mM NaCl, 15 mM sodium citrate); 0.6x SSC buffer, 2x 20 min at 65°C; 0.1x SSC buffer, 5 min at 65°C. Membranes were analyzed using a PhosphorImager (Molecular Dynamics).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-related protein is located in the matrix of rat liver mitochondria
A–C). A RXR-related protein with an apparent molecular mass in SDS-PAGE of 
44 kDa was observed by immunoblotting analysis using two specific RXR antibodies raised respectively against sequences overlapping the hinge domain and the ligand binding domain of the receptor (4RX-3A2), and located in the carboxyl-terminal part of the receptor (1RX-6G12) (Fig. 1A
). Western blot performed with an antibody raised against the NH2 terminus of RXR (RPRX) failed to detect any signal (Fig. 1A
). After mitochondrial subfractionation, we showed that the 44 kDa protein was detected in the matrix fraction (Fig. 1C
). We have thus provided the first evidence of the occurrence of a 44kDa protein immunologically related to RXR in rat liver mitochondrial matrix (mt-RXR). We failed to detect this truncated protein in rat liver nuclear extracts, which indicates it displays a specific mitochondrial localization (data not shown).
|
In agreement with this set of data, using 4RX-3A2 antibody in electron microscopy studies we directly observed a specific liver mitochondrial labeling that does not appear when using the RXR preabsorbed antibody (Fig. 2
A, C). Labeling associated with the organelle was at least as important as that recorded in nuclei (Fig. 2A, B
).
|
RXR nuclear receptor is cleaved by m-calpain or intramitochondrial calpain-like activity
Several arguments led us to test the possibility that the mitochondrial 44 kDa-RXR-related protein could be a truncated version of the nuclear receptor. Previous studies reported the presence in hepatocytes of a 44kDa RXR protein resulting from the amino-terminal cleavage of the nuclear receptor (34
35
36)
. This cleavage significantly reduces the amounts of RXR located in the nucleus, suggesting an extranuclear location for this truncated protein (35)
. Consequently, we examined the possibility that this protein could be addressed to mitochondria. We performed digestion experiments with [35S]-RXR synthesized in rabbit reticulocyte lysate as described by Noguchi et al. (32)
(Fig. 3
A, B). According to Matsushima-Nishiwaki et al. (34)
, we found that m-calpain (Sigma) efficiently cleaved [35S]-RXR to generate a 44 kDa protein (Fig. 3A
). As shown for the mitochondrial RXR-related protein, the absence of a signal in Western blot experiments using an antibody raised against the A/B domain (Fig. 3B
) proved that the truncated RXR protein produced by m-calpain cleavage (RXR/Cp) was deleted from the receptor amino-terminal domain. Mobility of the resulting protein in SDS-PAGE was similar to that recorded for mt-RXR (Fig. 3C
).
|
In addition to cytosolic calpain, several studies have reported the occurrence of a calpain-like activity in mitochondria (37
38
39)
. To test the possibility that such activity in mitochondria could cleave RXR, we performed digestion experiments by incubating [35S]-RXR synthesized in rabbit reticulocyte lysate with 1) whole mitochondrial extracts, 2) proteins belonging to mitoplast, 3) proteins belonging to mitochondrial outer membrane and intermembrane space, 4) mitochondrial outer and inner membranes extracts (Fig. 4
A). We found that protein extracts from whole mitochondria and outer membrane/intermembrane space efficiently cleaved [35S]-RXR to generate a 44 kDa protein (Fig. 4A
), whereas mitoplast and outer membrane/inner membrane extracts failed to do so (Fig. 4A
). Therefore, it appeared that proteolytic activity is restricted to the mitochondrial intermembrane space. To exclude the possibility that the full-length RXR could be cleaved by some cytosolic proteases copurified with mitochondria, we performed digestion experiments by incubating [35S]-RXR with the supernatant collected after mitochondrial removal from the incubation medium used in these assays. That no cleavage was recorded in these conditions strongly argues in favor of the occurrence of an intramitochondrial cleavage (Fig. 4B
).
|
To extend our knowledge of proteolytic activity involved in intramitochondrial RXR cleavage, several protease inhibitors were used in digestion experiments performed by incubating [35S]-RXR with whole mitochondrial extracts. We found that proteolytic cleavage of the nuclear receptor was inhibited by cysteine protease inhibitors, such as antipain and leupeptin, and by a calpain inhibitor such as ALLN (Fig. 4C
). However, it was not inhibited by most serine protease inhibitors (aprotinin, PMSF), metalloprotease inhibitors (EDTA, EGTA), or an aspartate protease inhibitor (pepstatin) (Fig. 4C
). These data imply that the enzymatic activity is related a cysteine protease belonging to the calpain family, in agreement with previous studies describing a calpain-like activity in mitochondria (37
38
39)
. We observed that cytosolic m-calpain or mitochondrial calpain-like activity generated truncated RXR products with a similar apparent molecular weight. All these data raised the possibility that mt-RXR could be a truncated form of RXR resulting from cleavage by cytosolic m-calpain or by mitochondrial calpain-like activity.
After cytosolic or intramitochondrial enzymatic cleavage, RXR is imported into the mitochondrial matrix
To test the possibility that after cleavage, RXR could be addressed to mitochondria, we studied the import of [35S]-RXR or of [35S]-RXR/Cp in isolated rat liver mitochondria as described previously for p43 (17)
(Fig. 5
A–D). At the end of the import experiment, the reaction mixture was treated with proteinase K to rule out contamination by nonimported proteins. In these conditions, we found that both proteins are protected against proteolytic activity inside the organelles (Fig. 5
, lanes 3, 7). The observation that after import, RXR and RXR/Cp protein sensitivity to proteinase K is restored after solubilization of mitochondrial membranes by a detergent treatment (Fig. 5A
, lanes 4, 8), conclusively demonstrated that protection of RXR proteins to peptidase enzyme is provided by mitochondrial membranes. Therefore, these data establish that both RXR proteins are translocated into the organelle. However, whereas RXR/Cp was not cleaved during the import processes, the apparent molecular mass of full-length RXR decreased to 44 kDa after internalization (Fig. 5A
, lane 3).
|
To extend our knowledge of mitochondrial import of RXR proteins, we used mitoplasts (mitochondria deleted from the outer membrane) or pharmacological agents such as apyrase (ATP/ADP phosphatase) and FCCP, respectively, to deplete ATP stores in the organelle and dissipate membrane potential. First, RXR/Cp translocation effectively occurred in mitoplast, ruling out the necessity to interact with the translocator outer membrane complex (Fig. 5B
, lane 8). Second, it was not abrogated by an FCCP-induced dissipation of mitochondrial membrane potential and remained unaltered by depletion of ATP stores by apyrase (Fig. 5B
, lanes 9, 10). In parallel experiments, we found that the full-length receptor was not imported into mitoplasts, indicating that RXR amino-terminal cleavage in the intermembrane space by a calpain-like activity is a prerequisite for its translocation through the inner membrane (Fig. 5B
, lane 3). Using integral organelle, this import was, as observed for RXR/Cp, independent of ATP stores or mitochondrial membrane potential (Fig. 5B
, lanes 4, 5). Last, we found that RXR import was not influenced by thyroid hormone or retinoic acid (Fig. 5 C
, lanes 3, 4). In this set of import experiments, we established that in the calpain inhibitor (ALLN) presence, only the full-length RXR form, unable to reach the matrix, is detected in mitochondria, as shown by protection against external proteinase K (Fig. 5 C
, lane 5). This conclusively demonstrates that intramitochondrial cleavage of RXR occurs in the intermembrane space after its internalization in the organelle.
We studied the time course of RXR and RXR/Cp import (Fig. 5D
). RXR/Cp was detected in mitochondria within the first minutes of the experiment; in addition, the maximal mitochondrial RXR/Cp level was recorded after 15 min of incubation and did not change thereafter, suggesting the involvement of a saturable import process (Fig. 5D
). We observed that significant amounts of the truncated protein were only detected in mitochondria after 15 min of incubation when using full-length RXR. This more delayed import probably results from the additional step of RXR cleavage in the intermembrane space.
mt-RXR binds to specific sequences of the mitochondrial genome in complex with p43
The apparent molecular mass of mt-RXR in SDS-PAGE suggests that amino-terminal cleavage occurs some amino acids upstream the DNA binding domain of the nuclear receptor (Fig. 1B
), raising the possibility that like its nuclear counterpart, it could bind to specific sequences of the mitochondrial genome. Several sequences homologous to nuclear hormone response elements have already been described in the mitochondrial genome (21)
. We performed EMSA experiments using these sequences as probes and highly purified mitochondrial extracts devoid of any significant nuclear contamination.
In agreement with a previous report (17)
, we found that a mitochondrial protein complex specifically bound to all specific probes (Fig. 6
). In contrast, no binding was detected using a probe corresponding to another D-loop sequence (NS; Fig. 6
), indicating that this complex only binds to particular sequences of the mitochondrial genome. Binding specificity was demonstrated by efficient removal of the radioactive complex resulting from the addition of an excess of the corresponding cold probe in association with the inability of an unrelated DNA probe to compete for binding (Fig. 6)
. Whatever the probe used in EMSA experiments, this mitochondrial complex displayed the same migration pattern, suggesting that a common complex binds to all probes. Preincubation of mitochondrial extracts with an antibody raised against RXR fully abrogated this binding, but unrelated antibodies (raised against GST) did not, indicating that mt-RXR is a component of this complex (Fig. 6)
. The first indication that mt-RXR could heterodimerize with p43 (mitochondrial truncated form of c-Erb A1) was provided by data demonstrating that preincubation of mitochondrial extracts with an antibody raised against c-Erb A prevented such binding (Fig. 6)
.
|
To test the possibility that mt-RXR could heterodimerize with p43, we performed coimmunoprecipitation experiments using highly purified mitochondrial extracts (Fig. 7
A–E). We found that p43 was coimmunoprecipitated by an antibody raised against RXR and, conversely, that mt-RXR was coimmunoprecipitated by an antibody raised against c-Erb A (Fig. 7B
). Additional experiments ruled out the possibility that this observation could result from antibodies cross-reactivity (Fig. 7C-E
).
|
Finally, we assessed the relative mitochondrial amounts of mt-RXR in several tissues. Mt-RXR was detected in mitochondria collected from all tissues tested but the brain (Fig. 7F
). The highest amounts of the protein were found in liver organelles (Fig. 7F
). These data underscore that mt-RXR levels displays a strong tissue specificity, close to that reported for p43 (18)
.
Mt-RXR is a ligand-dependent transcription factor of the mitochondrial genome
To elucidate the function of mitochondrial RXR protein, by analogy with p43 (17)
, we examined directly the ability of mt-RXR to activate mitochondrial genome transcription in in organello transcription experiments using isolated rat liver mitochondria, RXR/Cp, and/or p43 proteins synthesized in rabbit reticulocyte lysate (Fig. 8
A, B).
|
Using a COX III probe, two precursor transcripts were detected by Northern blot in mitochondria incubated in the presence of RXR/Cp or p43 (Fig. 8A
). Detection of these transcripts in mitochondria incubated in the presence of unprogrammed rabbit reticulocyte lysate required a much longer film exposure (up to 3 days vs. 20 min, data not shown). However, since both strands of mt-DNA are transcribed, hybridization analysis could not distinguish between H or L strand-derived precursor RNA. As observed for p43 after T3 loading, RXR/Cp transcriptional activity was strongly increased by the addition of 9-cis-retinoic acid in the incubation medium (Fig. 8A
). The size of precursor transcripts induced by RXR/Cp or p43 differed significantly (Fig. 8A
). Using a ND1 probe, we found that the increase in precursor transcript levels was associated with a rise in mature transcript levels (Fig. 8A
).
RXR/Cp-p43 interaction was assessed in additional experiments. In the presence of both receptors, only precursor transcripts corresponding to the influence of p43 alone were detected (Fig. 8B
), suggesting that p43 heterodimerization with mt-RXR induced recognition of p43-specific binding sequences, as shown for nuclear genes. RXR/Cp-p43 transcriptional influence increased after the addition of 9-cis-retinoic acid or T3 to the incubation medium and was even greater in the presence of both ligands (Fig. 8B
).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
.
Therefore, identification of hormonal receptors in the organelle such as p43 has provided new insight into the regulation of mitochondrial genome transcription (17)
and clarified the direct regulation of mitochondrial activity by thyroid hormone (10
, 13
, 19)
. Identification of mitochondrial complexes, including p43 and a truncated form of PPAR2 that binds a specific mitochondrial DNA sequence (18)
, already suggested that mitochondrial genome transcription could be regulated in a similar way to nuclear transcription. Consequently, in addition to the retinoid influence on organelle function, the alterations of mitochondrial activity observed after disruption of the RXR gene led us to search for the presence of RXR-related proteins in the organelle.
In this study we provide convincing evidence that a truncated RXR protein is located in the mitochondrial matrix. First, in all instances, we used highly purified mitochondria devoid of any significant contamination by nuclei or components of other cell compartments. Contaminations by lysosome, microsome, and membrane proteins were always minimal, as verified by measurements of specific markers. Nucleus-specific proteins were not detected in our mitochondrial preparations (ref 16
and data not shown). Second, two antibodies specifically raised against different sequences of RXR recognized the same 44 kDa mitochondrial protein. Third, in situ electron microscopy studies provide direct evidence of the presence of a protein related to RXR in the mitochondrion. Last, in vitro synthesized RXR is indeed translocated into the matrix of isolated organelles and gives rise to a 44 kDa protein during the import process. Subfractionation experiments established that, like in vitro synthesized RXR, the endogenous protein is located in the mitochondrial matrix. All these data strongly argue in favor of the presence of a truncated version of the nuclear receptor RXR in the organelle. This is also well supported by the observation that as in vitro synthesized RXR after mitochondrial import, mt-RXR lacks the amino-terminal domain of the nuclear receptor as shown by the absence of recognition of both proteins by an antibody raised against a part of this sequence.
Previous studies reported that RXR undergoes cytoplasmic amino-terminal cleavage by cathepsin L (35)
or m-calpain (34)
giving rise to a 44 kDa protein probably displaying an extranuclear location (35)
. RXR cleavage was associated with a significant decrease in the amounts of receptor located in the nucleus (35
, 36)
, suggesting that the truncated protein could display an extranuclear location.
After confirmation that m-calpain efficiently cleaves in vitro synthesized RXR to generate a 44kDa protein, we found that this truncated protein is imported into isolated organelles without any apparent change in its molecular mass, providing a first molecular origin for mt-RXR. In agreement with previous studies (37
38
39)
, we also observed that a calpain-like proteolytic activity located in the intermembrane space cleaves RXR to generate a 44kDa protein, providing a second molecular origin for mt-RXR. This led us to conclude that, on the one hand, after a cytosolic cleavage of RXR, the resulting protein is directly addressed to the mitochondrial matrix. On the other, the nuclear receptor is imported into the organelle and reaches the mitochondrial matrix after a cleavage occurring in the intermembrane space (Fig. 9
).
|
Typical steps involved in mitochondrial protein import have already been described: 1) binding of the protein to a mitochondrial outer membrane receptor; 2) translocation through the outer and inner mitochondrial membranes, a process generally driven by the mitochondrial membrane potential; 3) unfolding to an active conformation involving ATP-dependent interactions with mitochondrial heat-shock proteins (40)
. The targeting sequence of almost all mitochondrial protein is cleaved by matrix proteases. Using mitoplasts or pharmacological agents such as apyrase and FCCP, we found that after calpain cleavage, the import of a 44 kDa-RXR protein does not involve interaction with the outer membrane receptor and does not depend on mitochondrial membrane potential or ATP stores. These results provide a third example of a protein imported into mitochondria by this unusual process, apart from p43 (17)
and MTF1 (41)
. Work is in progress to characterize this unknown mitochondrial translocation process.
During these experiments, we observed that the full-length nuclear receptor RXR is not imported into the matrix compartment when import assays are performed with mitoplasts and therefore in the absence of outer membrane and components of the intermembrane space, including calpain-like activity. As this import, associated with a cleavage of the receptor to give rise to a 44 kDa protein is easily observed when using integral organelles, it appears that the enzymatic cleavage occurring in the intermembrane space is a decisive event for addressing the full-length receptor to the mitochondrial matrix.
We have previously shown that p43 binds several sequences homologous to nuclear hormone response elements (HRE) located in the mitochondrial genome (17)
. In this study, we have also provided evidence that p43 did not bind to these sequences as a monomer, which suggests the presence of dimerization partners of this receptor in the organelle. The apparent molecular weight of mt-RXR clearly suggests that the amino-terminal cleavage occurs some amino acids upstream the DNA binding domain of the nuclear receptor, raising the possibility that like its nuclear counterpart, it could bind to specific sequences of the mitochondrial genome.
In agreement with this possibility, in EMSA experiments we found that a mitochondrial protein complex including mt-RXR specifically bound to five probes corresponding to the mitochondrial HREs described previously (21)
. The abrogation of DNA binding by antibodies raised against RXR or c-Erb A demonstrated that mt-RXR and p43 are major components of this complex. Coimmunoprecipitation experiments confirmed that mt-RXR and p43 interact directly. These data therefore establish that, as observed at nuclear level, p43 and RXR bind to mitochondrial DNA by forming heterodimerical complexes. In line with this possibility, the assessment of relative mitochondrial amounts of mt-RXR in several tissues demonstrates a strong tissue specificity, very close to that reported for p43 (18)
.
To study the possible transcriptional activity of mt-RXR mediated by binding to the specific mitochondrial DNA sequences, we examined directly the ability of mt-RXR to activate mitochondrial genome transcription in in organello transcription experiments. We found that like p43 and T3 (present data and ref 17
), mt-RXR induced a dramatic increase in the levels of mitochondrial precursor transcripts in the presence of exogenous 9-cis-retinoic acid, paralleled by a rise in the levels of mature transcripts.
The different sizes of precursor transcripts induced by mt-RXR or p43 suggest that each receptor could stimulate transcription through the use of different promoters, as previously reported for the activities of p43 and mt-TFA (17)
. However, characterization of these promoters did not enter into the scope of this study.
Finally, in the presence of both receptors, only precursor transcripts corresponding to the influence of p43 were detected, and transcriptional influence of mt-RXR/p43 increased after the addition of 9-cis-retinoic acid or T3 to the incubation medium and was even greater in the presence of both ligands. As shown for nuclear genes, these data suggest that the mt-RXR/p43 heterodimerical complex induced specific recognition of p43 binding sequences.
Several observations rule out the possibility that this RXR and RXR/p43-induced rise in mitochondrial transcript levels could result from an increase in RNA stability. First, Enriquez et al. (10)
have established that T3 directly influences mitochondrial genome transcription in association with a decrease in mt-RNA half-life. Second, in a previous study (17)
, we found that the dramatic rise in mitochondrial transcripts levels induced by p43 in the presence of T3 is detected in <5 min, too brief a delay to involve only mRNA stabilization.
This study demonstrates that post-translational events substantially alter RXR location by addressing the receptor to mitochondria and provides important data for the comprehension of hormonal action at the cellular level. First, it throws new light on the influence of retinoids or vitamin A on organelle activity (7
, 11
, 26
, 27)
. It also contributes to explain the metabolic abnormalities recorded after disruption of the RXR gene (24
, 25)
. Second, in association with the identification of truncated c-Erb A1 (16)
and PPAR2 (18)
proteins in the organelle, it opens up a new field in endocrine research as it proves that besides membrane and nuclear receptors, a new class of receptors occurs in mitochondria. Finally, it provides new insight into understanding mitochondrial physiology. Mitochondrial proteins are mostly encoded by nuclear genes; however, 13 enzyme subunits required for organelle activity are encoded by the mitochondrial genome, highlighting the necessity for cooperation between nuclear and mitochondrial gene expression for mitochondriogenesis. Therefore, as put forward in a recent review (19)
, nuclear genes simultaneously encoding a nuclear and a mitochondrial receptor such as c-Erb A (p43) and RXR (mt-RXR) could play a crucial role in these processes. With regard to the origin of mitochondria as a bacterium-like organelle engulfed by a primitive cell not subjected to endocrine regulation, it appears that the organelle and the host cell’s response to hormones probably evolved in parallel. One of the most surprising aspects of this finding was that truncated forms of the nuclear receptor could act in mitochondria in much the same way as their full-length counterparts act at the nuclear level: recognition of responsive elements of the mitochondrial genome similar to those described in nuclear genes, heterodimerization leading to recruitment of RXR on the binding sequences of the other partner, p43 in this study, and ligand-dependent stimulation of transcription (Fig. 10
). A major goal will be to elucidate how this complex could induce activation or recruitment of mitochondrial RNA polymerase on the circular mitochondrial genome. A search for the occurrence of deregulation of mt receptor expression or import and mutations of hormone response elements present in mitochondrial DNA could be a promising way to better understand mitochondrial diseases whose etiology remains unclear.
|
|
ACKNOWLEDGMENTS |
|---|
Received for publication July 30, 2002. Accepted for publication November 18, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. S. Wong, S. Rajagopalan, F. M. Townsley, S. M. Freund, M. Petrovich, D. Loakes, and A. R. Fersht Physical and functional interactions between human mitochondrial single-stranded DNA-binding protein and tumour suppressor p53 Nucleic Acids Res., February 1, 2009; 37(2): 568 - 581. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Pettersson, N. Hanna, M. Lagodich, D. Dupere-Richer, M.-C. Couture, C. Choi, and W. H. Miller Jr. Rexinoids Modulate Steroid and Xenobiotic Receptor Activity by Increasing Its Protein Turnover in a Calpain-dependent Manner J. Biol. Chem., August 8, 2008; 283(32): 21945 - 21952. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. D. Arrington, T. R. Van Vleet, and R. G. Schnellmann Calpain 10: a mitochondrial calpain and its role in calcium-induced mitochondrial dysfunction Am J Physiol Cell Physiol, December 1, 2006; 291(6): C1159 - C1171. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J.-Z. Zeng, D.-F. Sun, L. Wang, X. Cao, J.-B. Qi, T. Yang, C.-Q. Hu, W. Liu, and X.-K. Zhang Hypericum sampsonii induces apoptosis and nuclear export of retinoid X receptor-alpha Carcinogenesis, October 1, 2006; 27(10): 1991 - 2000. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. J. De Blasio, K. L. Gatford, J. S. Robinson, and J. A. Owens Placental restriction alters circulating thyroid hormone in the young lamb postnatally Am J Physiol Regulatory Integrative Comp Physiol, October 1, 2006; 291(4): R1016 - R1024. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. V. Sionov, O. Cohen, S. Kfir, Y. Zilberman, and E. Yefenof Role of mitochondrial glucocorticoid receptor in glucocorticoid-induced apoptosis J. Exp. Med., January 23, 2006; 203(1): 189 - 201. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Takiyama, N. Miyokawa, A. Sugawara, S. Kato, K. Ito, K. Sato, K. Oikawa, H. Kobayashi, S. Kimura, and M. Tateno Decreased Expression of Retinoid X Receptor Isoforms in Human Thyroid Carcinomas J. Clin. Endocrinol. Metab., November 1, 2004; 89(11): 5851 - 5861. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
