REPRODUCTION-DEVELOPMENT
Mitochondria-Associated Membrane Formation in
Hormone-Stimulated Leydig Cell Steroidogenesis:
Role of ATAD3
Leeyah Issop, Jinjiang Fan, Sunghoon Lee, Malena B. Rone, Kaustuv Basu,
Jeannie Mui, and Vassilios Papadopoulos
Research Institute of the McGill University Health Centre (L.I., J.F., S.L., M.B.R., V.P.); Departments of
Medicine (L.I., J.F., S.L., M.B.R., V.P.), Pharmacology and Therapeutics (V.P.), and Biochemistry (V.P.); and
Facility for Electron Microscopy Research (K.B., J.M.), Department of Anatomy and Cell Biology, McGill
University, Montréal, Québec H3G 1A4, Canada
Leydig cell steroidogenesis is a multistep process that takes place in the mitochondria and endoplasmic reticulum (ER). The physical association between these 2 organelles could facilitate both
steroidogenesis substrate availability and mitochondrial product passage to steroidogenic enzymes in the ER, thus regulating the rate of steroid formation. Confocal microscopy, using antisera
against organelle-specific antigens, and electron microscopy studies demonstrated that there is an
increase in the number of mitochondria-ER contact sites in response to hormone treatment in
MA-10 mouse tumor Leydig cells. Electron tomography and 3-dimensional reconstruction allowed
for the visualization of mitochondria-associated membranes (MAMs). MAMs were isolated and
found to contain the 67-kDa long isoform of the adenosine triphosphatase (ATPase) family, AAA
domain-containing protein 3 (ATAD3). The 67-kDa ATAD3 is anchored in the inner mitochondrial
membrane and is enriched in outer-inner mitochondrial membrane contact sites. ATAD3-depleted
MA-10 cells showed reduced production of steroids in response to human choriogonadotropin but
not to 22R-hydroxycholesterol treatment, indicating a role of ATAD3 in the delivery of the substrate cholesterol into the mitochondria. The N terminus of ATAD3 contains 50 amino acids that
have been proposed to insert into the outer mitochondrial membrane and associated organelles
such as the ER. Deletion of the ATAD3 N terminus resulted in the reduction of hormone-stimulated
progesterone biosynthesis, suggesting a role of ATAD3 in mitochondria-ER contact site formation.
Taken together, these results demonstrate that the hormone-induced, ATAD3-mediated, MAM
formation participates in the optimal transfer of cholesterol from the ER into mitochondria for
steroidogenesis. (Endocrinology 156: 334 –345, 2015)
teroid synthesis is critical for the regulation of various
vital processes such as development, reproduction,
and behavior. This multistep process (1) has to be wellregulated to respond effectively to physiological needs by
producing tissue-specific steroids. The main site of this
regulation occurs in the mitochondria, where steroid biosynthesis is initiated by the transfer of cholesterol from the
intracellular stores to the outer mitochondrial membrane
S
(OMM) through the transduceosome, a complex where
cytosolic proteins interact with OMM proteins (2). From
there, cholesterol reaches the cytochrome P450
CYP11A1 enzyme in the inner mitochondrial membrane (IMM), which is involved in the conversion of
cholesterol to pregnenolone, the precursor of all steroids. This process has been demonstrated to occur
through the bioactive 800-kDa protein complex, termed
ISSN Print 0013-7227 ISSN Online 1945-7170
Printed in U.S.A.
Copyright © 2015 by the Endocrine Society
Received June 20, 2014. Accepted October 30, 2014.
First Published Online November 6, 2014
Abbreviations: ACSL4, acyl-coenzyme synthetase long-chain family member 4; ATAD3,
adenosine triphosphatase (ATPase) family AAA domain-containing protein 3; COX IV,
cytochrome c oxidase subunit IV; EM, electron microscopy; ER, endoplasmic reticulum;
GRP78, 78-kDa glucose-regulated protein; hCG, human choriogonadotropin; IMM, inner
mitochondrial membrane; MAM, mitochondria-associated membranes; OMM, outer mitochondrial membrane; qRT-PCR, quantitative real-time PCR; 22R-HC, 22R-hydroxycholesterol; siRNA, small interfering RNA; VDAC, voltage-dependent anion channel.
334
endo.endojournals.org
Endocrinology, January 2015, 156(1):334 –345
doi: 10.1210/en.2014-1503
doi: 10.1210/en.2014-1503
the steroidogenic metabolon, identified at the OMMIMM contact sites (3).
Cellular regulation of steroid formation has been
shown to be dependent upon organelle plasticity and organelle-organelle interactions, specifically between the endoplasmic reticulum (ER) and mitochondria (4). Biochemical and recent electron microscopy (EM) studies
confirmed the presence of specific regions of close apposition between ER membranes and OMM, termed mitochondria-associated membranes (MAMs) (5, 6). ER is a
highly dynamic network that plays a role in many cellular
functions, including calcium homeostasis, protein synthesis, folding, and phospholipid metabolism (7).
The various characteristics of MAMs suggest that they
may be important in the regulation of steroid biosynthesis
(4). MAMs are known as the site of transfer of calcium
from ER to mitochondria and as regulators of mitochondrial homeostasis and apoptosis (8). The presence of a high
concentration of cholesterol in lipid raft microdomains of
MAMs suggests that they could serve as a source of free
cholesterol to mitochondria for steroidogenesis (4, 8, 9).
This proposal would physically allow for the transfer of
cholesterol into mitochondria without the need for a
transfer protein while also enabling the transfer of steroidogenic pathway intermediates out of mitochondria to
the steroidogenic enzymes located in the ER (2).
AAA domain-containing protein 3 (ATAD3) is a member of the large family of ATPases associated with diverse
activities. This highly conserved protein is ubiquitous
(10), and knocking down Atad3 in mice resulted in early
embryonic lethality (11). Although many functions have
been associated with ATAD3, especially in development
(11, 12) and mitochondria DNA integrity (13), the exact
role of ATAD3 remains unclear. Studies conducted on
H295R adrenocortical and MA-10 mouse Leydig cell lines
have implicated this protein in steroidogenesis (3, 12). In
MA-10 cells, ATAD3 was identified in the 800-kDa steroidogenic metabolon, and it appeared to play a role in
hormone-induced OMM-IMM contact site formation, allowing for the passage of lipophilic cholesterol through
the aqueous intermembrane space to reach to CYP11A1 in
IMM (3).
The exact structure of ATAD3 remains unknown, but
studies conducted on the topology of the protein showed
that it was anchored in IMM and enriched at mitochondrial contact sites (12–14). ATAD3 through the N-terminal domain was also suggested to be tethering with the
cytosol and adjacent compartments (15). This finding underlined its importance in maintaining the normal morphology of ER and mitochondria in human lung cancer
cells (16). ATAD3 has been also linked to cancer progression (10). Although most of the work has focused on the
endo.endojournals.org
335
67-kDa ATAD3, there are many variants of the protein
(12, 17). We present evidence herein that there is a hormone-dependent increase in MAM formation in MA-10
Leydig cells. In search of the factors mediating the formation of MAMs, we report that in MA-10 cells, there are
both the regular and long forms of ATAD3, the latter
containing additional amino acids in the N terminus to
form an ␣-helix that drives the insertion of the protein into
OMM and ER at MAM regions, shown to be important in
steroid formation.
Materials and Methods
Cell culture
MA-10 mouse (18) and R2C rat tumor Leydig cells were
maintained in DMEM/Ham’s F12 (50:50) supplemented with
5% fetal bovine serum and 2.5% horse serum. H295R-S2 human adrenocortical cells were cultured in the same medium with
2.5% Ultroser G.
Establishment of the stable cell line Mito-H by
fluorescence-activated cell sorting
MA-10 cells were used to establish a cell line that stably expressed the mitochondrial reduction-oxidation sensitive green
fluorescent protein 1 (19). In brief, cells transfected with 4 ␮g of
plasmid pRA306 were grown, as described above. Cells were
initially supplemented with 500 ␮g/mL G418 and maintained in
medium with 400 ␮g/mL G418. Single colonies of cells expressing roGFP were sorted using the high-speed cell sorter, FACSAria
II (BD Biosciences) and used for expansion. Cells were then
grown in media supplemented with 1% penicillin/streptomycin
for 2 passages. A highly expressing line, designated Mito-roGFP
MA-10 cells, was selected, tested for its ability to form progesterone in response to human choriogonadotropin (hCG; National Hormone and Peptide Program) in a manner similar to its
parent MA-10 cells (data not shown), and maintained in
DMEM/F-12 medium containing 400 ␮g/mL G418.
Isolation of organelles and MAMs
Cells were grown in 150-mm dishes to 70% confluence,
washed with PBS, and harvested. The preparation of crude membranes, MAM regions, and ER membranes was performed
following a well-established methodology (20) with minor modifications. In brief, cells were homogenized with a glass potter
followed by 3 cycles of freezing and thawing. Homogenates were
centrifuged at 600g for 5 minutes at 4°C. The collected supernatants were centrifuged twice (7000g for 10 minutes followed
by 10 000g for 10 minutes at 4°C) to obtain crude mitochondria.
These preparations were resuspended in isolation medium
(250mM mannitol, 5mM HEPES [pH 7.4], and 0.5mM EGTA).
To confirm the enrichment of the crude mitochondrial fraction,
we performed immunoblot analysis using anti-cytochrome c oxidase subunit IV (COX IV) and anti-voltage-dependent anion
channel (VDAC). To isolate ER membranes, the supernatant
obtained after the centrifugation at 7000g for 10 minutes at 4°C
was centrifuged at 20 000g for 30 minutes at 4°C. Further centrifugation of the supernatant (100 000g for 1 hour) resulted in
336
Issop et al
MAMs in Leydig Cell Steroidogenesis
the isolation of the ER (pellet) and cytosolic fraction (supernatant). ER membranes were resuspended in isolation buffer
(225mM mannitol, 75mM sucrose, 0.1mM EDTA, and 30mM
Tris-HCl [pH 7.4]). To confirm the enrichment of the isolated
fractions we looked for calreticulin (Abcam) and Bip proteins
using specific antibodies. To yield MAMs, crude mitochondria
were layered on a Percoll solution (225mM mannitol, 25mM
HEPES [pH 7.4], 1mM EGTA, and 30% Percoll [vol/vol]) followed by centrifugation at 95 000g for 30 minutes. MAMs were
collected and purified by ultracentrifugation at 100 000g for 1
hour at 4°C. MAMs were resuspended in the same isolation
medium as crude mitochondria, and their purity was tested using
anti-acyl-coenzyme synthetase long-chain family member 4
(ACSL4) and anticalnexin antibodies. Protein concentration was
measured using the Bradford assay (Bio-Rad).
RNA extraction and quantitative real-time PCR
analysis
Cells were plated in 6-well plates at a density of 30 ⫻ 104 cells
per well. After 48 hours, cells were treated with hCG (50 ng/mL).
At the indicated time points, cells were harvested and total RNA
was extracted using the RNeasy PLUS Mini Kit (QIAGEN).
RNA concentration was determined by measuring absorbance at
260 nm using the NanoDrop ND-1000 (Thermo Scientific).
Samples were normalized to total RNA content and reversetranscribed using the Transcriptor First Strand cDNA Synthesis
Kit (Roche Applied Science). The resulting cDNA samples were
diluted with nuclease-free water and subjected to quantitative
real-time PCR (qRT-PCR) using the SYBR green dye technique
on a LightCycler 480 System (Roche Applied Science). The results reported for each RNA product were normalized to Gapdh
mRNA to correct for differences in the amounts of template
cDNA. The following oligonucleotide sequences of the sense and
antisense primers were used: Atad3 (NM_179203.3), forward
5⬘-CGG ATG CTG AAG AGC GAA CAG ATC CG-3⬘ and reverse 5⬘-CTA GCC TGG TGC TGT CGT GTC-3⬘; Star
(NM_019779.2), forward 5⬘-CGG TCT CTA TGA AGA ACT
TGT GGA CCG3⬘ and reverse 5⬘-ATC TCC TTG ACA TTT
GGG TTC CA-3⬘; Gapdh (NM_008084), forward 5⬘-CGG
TGA CTC CAC GAC ATA CTC AGC ACC G-3⬘ and reverse
5⬘-CAC CAT CTT CCA GGA GCG AGA C-3⬘ was used as an
endogenous control. The qRT-PCRs were performed in triplicate, and the comparative cycle threshold (⌬⌬Ct) method was
used to demonstrate the relative transcription of the target gene
compared with the reference gene.
Small interfering RNA transfection studies
Cells were plated onto 6-well plates at an initial concentration
of 3 ⫻ 105 cells per well, and immediately transfected using the
TriFECTa kit DsiRNA duplex (Integrated DNA Technologies)
using Lipofectamine RNAiMAX (Life Technologies). The following small interfering RNA (siRNA) duplexes (150nM) were
used for Atad3 (NM_179203): duplex 1, 5⬘-CCAUCGCAA
CAAGAAAUACCAAGAA-3⬘; duplex 2, 5⬘-CCAGUUUGAC
UAUGGAAAGAAAUGC-3⬘; and duplex 3, 5⬘-AGGACAAAU
GGAGCAACUUCGACCC-3⬘. Gene expression and target gene
knockdown were evaluated by qRT-PCR. After 48 hours, transfected cells were plated in 96-well plates at a density of 2.5 ⫻ 104
per well to determine their ability to form progesterone or to be
Endocrinology, January 2015, 156(1):334 –345
directly lysed, as previously described, for EM and immunoblot
analyses.
N-deletion studies
The cDNA coding sequence used to generated mouse Atad3a
was amplified from mammalian gene collection clone ID 65362
(IMAGE:6530155; Thermo Scientific). The Atad3 cDNA coding
sequence was subcloned into pEGFP (Clontech) at the EcoR1
and BamHI sites (ATAD3-pEGFP). To generate an N-truncated
ATAD3 construct, the first 50 amino acids were deleted from the
coding sequence, and the appropriate regions were amplified by
PCR using designed primers (forward 5⬘-TCG AAT TCA TGC
CGA CGG GCC TGG AACG-3⬘ and reverse 5⬘TGG ATC CCA
GCA GCT GAG GAG TGA AGG ATG-3⬘), which were gel
purified, ligated, and reinserted into the pEGFP vector. The constructs were verified by sequencing. Cells plated in 6-well plates
were transfected using Lipofectamine 2000 (Life Technologies)
and 6 ␮g of the N-deletion construct. After 48 hours, cells were
plated in 96-well plates at a density of 2.5 ⫻ 104 to determine
progesterone formation by RIA.
Immunoblot analysis
Cells plated at a density of 3 ⫻ 105 were grown in 6-well plates
for 72 hours. Proteins were extracted using ice-cold radioimmunoprecipitation assay (RIPA) buffer (Cell Signaling Technology).
Proteins (10 ␮g for MAM, 15 ␮g for mitochondrial membranes,
and 25 ␮g for whole-cell and ER membranes) were electrophoretically separated on a 4% to 20% Tris-glycine SDS-PAGE, and
transferred to a polyvinylidene fluoride membrane and blocked
for 90 minutes at room temperature in blocking buffer (20nM
Trizma base, 100mM NaCl, 1% Tween 20, 5% skim milk).
Membranes were incubated overnight at 4°C with primary antisera against ATAD3, VDAC, COX IV, calreticulin, Bip protein,
ACSL4, and calnexin. Membranes were washed and incubated
for 1 hour at room temperature with secondary antirabbit IgG
horseradish peroxidase-linked antibody (Cell Signaling Technology). Proteins of interest were visualized using the Amersham
chemiluminescence kit and a FUJI image reader LAS4000
(FUJIFILM) for capturing images. The intensity of each band was
measured using Multigauge version 3.0 software (FUJIFILM). Image analysis was performed using IMAGE J version 1.43u
software.
Immunocytochemistry and confocal microscopy
MA-10 cells (2 ⫻ 104 per well) were plated in 24-well glassbottom dishes (FluoroDish; World Precision Instruments) in
triplicate and incubated until 60% confluent. Cells were treated
for 2 hours with and without 50 ng/mL hCG. At the end of the
treatment, cells were fixed with 4% paraformaldehyde for 10
minutes, permeabilized with 10% Triton X-100 for 3 minutes,
and blocked with 10% goat serum for 1 hour. Cells were incubated with anti-ATAD3 or anti-ACSL4 or 3␤-hydroxysteroid
dehydrogenase or CYP11A1 antibodies overnight at 4°C. The
next day, the wells were washed with PBS and stained with a
secondary antirabbit IgG F(ab⬘)2 fragment raised in goat (Alexa
Fluor 488 conjugate) and an antigoat IgG F(ab⬘)2 fragment
raised in rabbit (Alexa Fluor 546 conjugate; Cell Signaling Technology) at room temperature for 1 hour. Cells were washed with
0.5M Tris-HCl (pH 7.6), and an appropriate concentration of
4⬘,6-diamidino-2-phenylindole was used for nuclear staining for
doi: 10.1210/en.2014-1503
3 minutes at room temperature. Confocal microscopy was performed using an Olympus Fluoview FV1000 laser confocal microscope (UPLSAP, ⫻60).
Live cell imaging studies
The MA-10 stable cell line Mito-roGFP was cultured in 35-mm
FluoroDish culture dishes to 75% confluence. The following day,
cells were transfected with pDsRed2-ER vector (Clontech Laboratories) using Lipofectamine 2000. pDsRed2-ER is a mammalian
expression vector designed to label the ER in living cells using the
ER-targeting sequence of calreticulin and the ER retention sequence KDEL fused to the 5⬘ and 3⬘ ends of DsRed2, respectively.
After 24 hours, cells were incubated with Hoechst 33342 (Life
Technologies) to counterstain nuclear DNA. Cells were exam-
endo.endojournals.org
337
ined using a scanning laser confocal microscope (Olympus FluoView FV1000) with an oil immersion objective (UPLSAP,
⫻100), and the following excitation/emission wavelengths (nm):
blue (Hoechst), 350/461; green (roGFP), 498/516; and red
(DsRed), 558/583. Images were captured with FluoView software (version 3.1), and quantitative colocalization analysis was
performed using Image-Pro Plus (version 6.3) and ImageJ version
1.47 (http://rsbweb.nih.gov/ij). Statistical analysis of the quantitative data was performed using Mann-Whitney U tests (Prism
version 5.0; GraphPad).
Electron microscopy
Control and ATAD3-depleted MA-10 cells were incubated
with or without hCG for 2 hours, washed in a sodium cacodylate
buffer (0.1M, pH7.4) and scraped. After
centrifugation (1500g for 5 minutes), the
pellet was fixed with 2.5% glutaraldehyde overnight at 4°C. After washing 3
times with 0.1M sodium cacodylate
washing buffer, cells were poststained
with 1% OsO4 for 2 hours at 4°C. Cells
were dehydrated by incubation in increasing acetone concentrations from
30% to 100%. After step infiltration
with different mixes of acetone-Epon
(1:1, 1:2, and 1:3 vol/vol), samples were
embedded in Epon. Polymerization was
performed by incubation at 60°C for 48
hours. Blocks were trimmed and cut at
90 to 100 nm thick with an UltraCut E
ultramicrotome (Reichert-Jung), transferred onto a 200-mesh Cu grid, and
poststained with 4% uranyl acetate for 8
minutes and then with Reynold’s lead for
5 minutes. Cells on the grids were observed
with a transmission EM (FEI Tecnai 12;
FEI) operated at 120 kV, and images were
collected with a CCD camera (AMT XR
80 C).
Electron tomography studies
and 3-dimensional
reconstruction
Figure 1. The ER-mitochondria physical interaction in control and hormone-stimulated MA-10
cells. A, Confocal microscopy imaging studies performed in the MA-10 Mito-roGFP cell line (b
and e) transfected with DsRed-ER (a and d). Merged images are shown in c and f. Scale bars, 2
␮m. B, Higher magnification of the merged images from hormone-treated cells. White dots
represent the colocalization of DsRed-ER and mito-roGFP. C, Intensity of colocalization between
the ER and mitochondria in cells treated with or without hCG for 2 hours. ER-mitochondria
interactions were quantified by image analysis. Results shown are from 20 control and 15 hCGtreated cells. Statistical analysis was performed using the Mann-Whitney U test; *, P ⫽ .0343. D,
Progesterone levels by hCG-treated MA-10 Mito-roGFP cells. Progesterone levels were
normalized to untreated control cells. E, Contact site formation between the ER and
mitochondria upon hCG treatment visualized by EM (magnification, ⫻9300; scale bar, 500 nm)
The white circles represent MAM regions. F, Quantification of MAMs based on image analysis of
the EM images. Mitochondria from 30 control and hCG-treated cells were analyzed.
Control and hCG-treated cells were
prepared as described above. Cells were
stained with 1% tannic acid in 0.1M sodium cacodylate buffer for 1 hour at 4°C
after the osmium step but before the dehydration. The 200-nm-thick sections
were placed on carbon-coated 200-mesh
copper grids and poststained with 4%
uranyl acetate for 15 minutes, then with
Reynold’s lead for 12 minutes. Images
were acquired using a FEI G2 F20 CryoSTEM operated at 200 kV (FEI). Images
were recorded with a Gatan Ultrascan
4k⫻4k Digital CCD Camera System at a
nominal magnification of ⫻9600 corresponding to a pixel size of 1.17 nm using
a ⫺62° to 62° tilt. For electron tomog-
338
Issop et al
MAMs in Leydig Cell Steroidogenesis
raphy, data collection was done at an electron dose of approximately 1500 electrons/Å2 per tomogram. Focusing was done on
an adjacent area to minimize electron dose exposure. Tilt series
were taken using the FEI software in the angular range between
⫺62° and ⫹62° with 2° increments in low tilts (up to tilt angle
30°) and 1° increments at the high tilts (from tilt angles 31°– 62°).
For the estimated sample thickness, this would be sufficient for
a resolution of 2 nm, following the Crowther formula (21). They
were then aligned, cropped, and binned using the ETomo program from the IMOD suite. The 3-dimensional visualization and
volume analysis was done using Amira Resolve RT 5.2.2 (FEI) by
guided segmentation.
Radioimmunoassay
MA-10 cells were plated into 96-well plates at a density of
2.5 ⫻ 104 cells per well. Then, 24 hours upon seeding, cells were
treated with 50 ng/mg of hCG or 20␮M 22R-hydroxycholesterol
(22R-HC) for 2 hours in serum-free media. Mito-roGFP MA-10
cells were treated at different time points. Media were collected,
and progesterone and pregnenolone production were determined by RIA using specific antiprogesterone and antipregnenolone antisera (MP Biomedicals), as previously described
(22). RIA data were analyzed using Prism version 5.0).
Statistical analysis
Data are expressed as means, SEM, and number and analyzed
using Student’s t test or the Mann-Whitney U test using Prism
version 5.0. P ⬍ .05, P ⬍ .01, and P ⬍ .001 were used as indicators of the level of significance. The results shown are means ⫾
SEM from at least 3 independent experiments performed in
triplicate.
Results
ER-mitochondria communication is increased upon
hormone treatment of MA-10 cells
To examine the communication between the ER and
mitochondria, we performed confocal imaging studies using the MA-10 Mito-roGFP cell line, where the mitochondria transmit green fluorescence, transfected with the
pDsRed2-ER expression vector designed to label ER. Cells
were treated with and without hCG for 30, 60, and 120
minutes. The results obtained (Figure 1A and Supplemental Figure 1) demonstrated that upon hormone stimulation, both the mitochondrial network and the ER were
remodeled, and increased colocalization could be seen in
the merged images (in yellow); colocalization, highlighted
in white, was more evident with higher magnification (Figure 1B). Measurement of the intensity of colocalization
before and after hCG treatment was performed. Image
analysis demonstrated that the overlapping signal between DsRed-ER and mito-roGFP was significantly increased after 2 hours hCG treatment (Figure 1C) starting
as early as 30 minutes (Supplemental Figure 1). Steroid
formation in these time points is shown in Figure 1D. To
Endocrinology, January 2015, 156(1):334 –345
Figure 2. Ultrastructural interaction between mitochondria and ER. A,
ER-mitochondria interactions in control MA-10 cells (a) and after
reconstruction (b). The 3-dimensional visualization and volume
reconstruction on a cropped section were obtained using the AMIRA
program (c and d). B, ER-mitochondria interactions in hCG-treated
MA-10 cells. The 3-dimensional reconstructed mitochondria are shown
in green and ER in yellow. Scale bar, 500 nm.
confirm the hormone-induced remodeling of ER-mitochondria interaction, we carried out EM studies under the
same conditions and quantified MAM formation by image
analysis. The white circles on the EM image (Figure 1E)
represent MAM formation. We observed a 3.5-fold significant increase of the percentage of MAMs formed upon
hCG treatment (Figure 1F). A similar effect was observed
using the cAMP analog, dibutyryl-cAMP (1mM) instead
of hCG (data not shown). We then investigated the ultrastructural interaction between mitochondria and ER in 3
dimensions by electron tomography studies (Figure 2). In
hCG-treated cells, the number of interactions occurring
between the 2 organelles appeared to be increased compared with control cells as shown in the tomography movies (Supplemental Videos 1 and 3 for control and 2 and 4
for hCG-treated cells). Through the reconstructed 3-dimensional models, we further confirmed the structural
difference of this interaction in hCG-treated cells (Figure
2B, a– d) compared with control (Figure 2A, a– d). Indeed,
the surface of contact site around the mitochondria with
ER is increased upon hormone stimulation compared with
control.
ATAD3 is present in mitochondria and ER in
MA-10 cells
To determine the subcellular localization of ATAD3 in
MA-10 cells, we performed immunocytochemistry studies
followed by confocal microscopy. Figure 3A shows that
ATAD3 colocalizes (overlapping signals of immunostaining) with the IMM protein CYP11A1 and with 3␤-hydroxysteroid dehydrogenase, the enzyme involved in the
conversion of pregnenolone to progesterone, which is
present in both the ER and mitochondria.
These data were further confirmed by immunoblot
analysis of isolated ER and mitochondrial membrane fractions (Figure 3B). Indeed, ATAD3-immunoreactive pro-
doi: 10.1210/en.2014-1503
Figure 3. Different forms of ATAD3 are present in ER and
mitochondria. A, Immunocytochemistry followed by confocal imaging
performed using antibodies against ATAD3 (a and d), CYP11A1 (b),
and 3␤-HSD (e). Panels c and f represent the merged images and
show the colocalization of ATAD3 with the mitochondrial and ER
markers CYP11A1 and 3␤-HSD, respectively. Confocal images are
representatives of 3 independent experiments. B, Immunoblots
showing the presence of ATAD3 in different fractions (25 ␮g of cell
extracts, 25 ␮g ER, and 15 ␮g mitochondrial proteins) of MA-10, R2C,
and H295R-S2 cells. In MA-10 cells, the 67-kDa long isoform of
ATAD3 is present primarily in the mitochondria, whereas the 57-kDa
short form of ATAD3 is present in the ER. The same profile is observed
in R2C, whereas only the long form is predominant in H295R-S2 cells.
COX IV and VDAC are used as markers of mitochondria enrichment,
and calreticulin and GRP78 are markers of the ER. The immunoblots
presented are representative of 3 independent experiments. C,
Representative model of ATAD3 isoform topology in the mitochondria
membranes and the ER. The long isoform of ATAD3 is anchored in the
IMM with the C terminus in the matrix. The exact localization of the N
terminus is not yet known. The short isoform is present in the ER
membranes, but the topology is not known.
teins were found in both ER and mitochondria. Interestingly, 2 different isoforms of the protein were identified,
57 and 67 kDa, present primarily in the ER and mitochondria, respectively. The mitochondrial markers COX
IV and VDAC as well as the ER markers calreticulin and
78-kDa glucose-regulated protein (GRP78) demonstrated
the enrichment of the isolated subcellular fractions. Previous studies of the topology of ATAD3 suggested that it
is anchored in the IMM and its N terminus could interact
with the ER (12, 16). Such localization would facilitate the
formation of MAMs. A schematic diagram of the localization of the 2 isoforms is presented in Figure 3C. Addi-
endo.endojournals.org
339
Figure 4. ATAD3 is present in the MAM regions in MA-10 cells. A,
Immunocytochemistry studies and confocal imaging using ATAD3
antisera (a and d), ER tracker (b), and ACLS4, a marker of MAMs (c).
Merged images show the localization of ATAD3 at localized sites
(shown with a white circle on c). B, Immunoblot analysis of purified
MAMs (10 ␮g protein) showing the presence of the 67-kDa long
isoform of ATAD3. Antibodies against ASCL4 and calnexin confirmed
the presence of MAMs in this fraction. The immunoblots presented are
representative of 3 independent experiments. C, Quantification of
ATAD3 in different membrane fractions (MAM, mitochondria, and ER).
For each band, the OD of the expression levels of ATAD3 protein was
quantified and normalized to the amount of protein loaded.
tional studies in R2C rat Leydig and H295R-S2 adrenocortical cells demonstrated the presence of the 67-kDa
long isoform in all cell lines with the 57-kDa form at higher
levels in rodents (Figure 3B).
ATAD3 is present in MAMs
Using ER Tracker, we confirmed the presence of
ATAD3 in ER membranes by immunocytochemistry followed by confocal imaging (Figure 4A). Surprisingly, the
colocalization was observed mainly at well-defined punctuated sites (Figure 4A, a– c). These specific areas could
correspond to MAM regions, which represent up to 20%
of the mitochondrial surface (23). To examine this possi-
340
Issop et al
MAMs in Leydig Cell Steroidogenesis
Endocrinology, January 2015, 156(1):334 –345
Figure 5. Effect of hCG on ATAD3 levels. A, Time course of the effect of hCG on Atad3 mRNA levels in MA-10 cells. B, Time course of the effect
of hCG on Star mRNA levels. In A and B, mRNA expression was normalized to GAPDH. Results shown are means ⫾ SEM (n ⫽ 3) C, ATAD3 protein
levels in the mitochondria, ER, and MAM fractions isolated from control and cells treated with hCG (50 ng/mL) for 2 hours. ACSL4, COX IV,
calnexin, and GRP78 are used as references for the different fractions. The immunoblots presented are representative of 3 independent
experiments. D, Quantification of ATAD3 in different membrane fractions before and after treatment. The expression levels of ATAD3 were
quantified and normalized to the corresponding protein of reference.
bility, we assessed the localization of ATAD3 in parallel
with ACLS4, a well-known marker of MAMs (24, 25). A
strong colocalization of the signal between the antiATAD3 and anti-ACSL4 used is seen in Figure 3Af. To
confirm these data, we isolated MAMs from MA-10 cells
(20) and performed immunoblot analyses for the presence
of various markers. Figure 4B shows the presence of both
ACSL4 and calnexin, specific markers of MAMs. Moreover, the data obtained showed that ATAD3 is greatly
enriched in MAMs where only the 67-kDa long isoform is
present (Figure 4C).
ATAD3 levels in MA-10 cells are not regulated by
acute treatment with hCG
To understand the role of ATAD3 in hormone-induced
steroidogenesis, we examined ATAD3 expression in response to hCG treatment in MA-10 cells. There was no
effect on Atad3 mRNA levels during the first 2 hours of
hCG treatment (Figure 5A). As predicted, under the same
conditions, the Star gene was induced in a time-dependent
manner (Figure 5B). We analyzed ATAD3 protein levels in
the different membrane fractions before and after stimulation with hCG. There was no difference in ATAD3 levels
in isolated ER, mitochondria, or MAM fractions (Figure
5, C and D). Thus, we hypothesized that ATAD3 might be
involved in MAM formation by regulating the association
between the 2 organelles.
ATAD3 localization regulates hormone-dependent
steroidogenesis
Figure 6, A and B, shows that transfection with specific
siRNAs resulted in the specific reduction of both Atad3
mRNA and protein levels in MA-10 cells. The reduction in
ATAD3 levels was accompanied by reduced progesterone
formation in response to hCG (Figure 6C), but not in response to 22R-HC (Figure 6D), which is in agreement with
recently published data (3). These data suggest that the
effect of ATAD3 is likely due to the mechanism of cho-
doi: 10.1210/en.2014-1503
endo.endojournals.org
341
Figure 6. ATAD3 expression is critical for steroidogenesis. A, Forty-eight hours transfection with siRNAs targeting Atad3 resulted in reduced
Atad3 mRNA levels. B, Forty-eight hours transfection with siRNAs targeting Atad3 resulted in reduced ATAD3 levels. ACSL4 and COX IV were used
as controls. To assess the role of ATAD3 in hormone-stimulated progesterone production, MA-10 cells were treated with siRNAs targeting Atad3,
followed by 2 hours of stimulation either with 50 ng/mL hCG (C and E) or 20␮M 22R-HC (D and F). After 2 hours, progesterone (C and D) or
pregnenolone (E and F) were measured by RIA. Transmission electron micrographs of MA-10 cells were treated with either scrambled siRNAs (G) or
siRNA targeting Atad3 (H). Magnification, ⫻4800; scale bars, 500 nm. Results shown are means ⫾ SEM from 3 independent experiments
performed in triplicate. Statistical analysis was performed using Student’s t test. *, P ⬍ .05; **, P ⬍ .01; ***, P ⬍ .001. Basal levels of
progesterone were 11.7 ⫾ 2.7 ng/mg protein The immunoblots and EM images presented are representative of 3 independent experiments.
lesterol import into the mitochondria, rather than to
CYP11A1 activity or pregnenolone export to the ER. To
further confirm these results, we measured pregnenolone
production by the cells under the same conditions. Indeed,
pregnenolone formation in response to hCG, but not 22RHC, was decreased in ATAD3-depleted MA-10 cells (Fig-
342
Issop et al
MAMs in Leydig Cell Steroidogenesis
Endocrinology, January 2015, 156(1):334 –345
Figure 7. Role of ATAD3 in MAM formation and steroid synthesis. A, Schematic representation of the ATAD3-GFP construct made and its various
domains. Abbreviations: cc, coiled coil domains; TMD, transmembrane domain; WA, Walker A; WB, Walker B. Deletion of the first 50 amino acids
(red rectangle) of ATAD3 was performed. B, The N-deleted ATAD3-GFP construct was transfected in MA-10 Leydig cells: left, phase-contrast
imaging; right, fluorescence imaging. C, At 48 hours after transfection, MA-10 cells were treated with hCG (50 ng/mL) for 2 hours. Progesterone
formation was measured by RIA. Basal levels of progesterone were 21.2 ⫾ 2.1 ng/mg protein. Results shown are the means ⫾ SEM (n ⫽ 5).
Statistical analysis was performed using Students’ t test. **, P ⬍ .01.
ure 6, E and F). Interestingly, morphological evaluation of
the cells treated with siRNAs targeting Atad3 showed that
the inner ultrastructure of mitochondria was disturbed
when compared with cells treated with scrambled siRNAs
(Figure 6, G and H). Surprisingly, no decrease in MAM
formation was observed. In contrast, a significant increase
of the contact sites was observed in siRNA-treated cells,
suggesting that the absence of the protein tends to modify
the organelle plasticity and ATAD3 could regulate this
interaction by modulating the distance between these 2
organelles. (means ⫾ SEM of percentage of MAMs
formed were 14.32% ⫾ 2.00% in control vs 25.72% ⫾
2.08% in siRNA-treated cells; 1.79-fold increase).
Role of ATAD3 topology in steroid formation
Considering the putative role of the first 50 amino acids of
ATAD3 in the interaction of the protein with OMM, and the
possibility that this region has the ability to span the OMM
and exit into the cytosol, we deleted this region of the protein
(Figure 7A). The N-terminal–truncated ATAD3 was transfected into MA-10 cells, with the aim of creating a dominantnegative effect competing with the endogenous wild-type
protein. Figure 7B shows that the truncated protein is expressed in the transfected Leydig cells. Figure 7C shows that
the presence of the truncated protein resulted in reduced hormone-induced progesterone formation.
Discussion
Steroidogenic cells are dynamic entities that rapidly respond to hormonal stimuli by metabolizing cholesterol to
tissue-specific steroid products. This is a well-timed and
orchestrated process that involves protein synthesis, lipid
and ion mobilization, and a series of protein-protein interactions (26 –28), all aiming at delivering free cholesterol to CYP11A1 in IMM, which is the rate-limiting step
in steroid biosynthesis. Recently we, and others, proposed
interorganellar communication, mainly between the ER
and the mitochondria, as one of the driving forces that
regulate the access to and availability of the substrate cholesterol into the mitochondria (4, 29, 30).
MA-10 live cell imaging and confocal microscopy using
specific mitochondrial and ER markers, as well as MA-10
cells containing roGFP-labeled mitochondria and vectors
targeting DsRed exclusively to the ER, coupled with EM,
demonstrated that physical interactions between ER and
mitochondria exist in control cells and are increased upon
hCG stimulation. This observation is in agreement with
previous data showing ultrastructural modifications of
testicular interstitial cells in mice treated with gonadotropin (31). Detailed quantitative EM demonstrated a 3.5fold increase in the number of ER-mitochondria contact
sites in response to hCG. Furthermore, electron tomography studies combined with 3-dimensional reconstruction
volume demonstrated an increase of the contact surface
between the 2 organelles.
Associations between ER and mitochondria, leading to
the formation of MAMs, are characterized by their stability and reversibility (8), and they follow the movement
of the microtubules (32) and the state of mitochondrial
fusion (30). Such dynamic changes between organelles de-
doi: 10.1210/en.2014-1503
pend on their own homeostasis and protein and lipid composition and represent a strong parameter in the regulation of many processes.
The formation and protein composition of MAMs have
been studied in other cells, and few proteins have been
reported to be both mitochondria- and ER-bound. MAM
regions are well-organized, and their formation involves
shaping proteins (eg, phosphofurin acidic cluster sorting
protein 2 and mitofusin 2), calcium-dependent chaperones (such as calreticulin), binding Ig protein, and ␴-1,
important for protein folding in the ER (33, 34). MAM
regions are also sites of high concentration and flow of
calcium, important for mitochondria homeostasis, and
they are likely necessary in optimal steroidogenesis as well;
these regions could serve as sites of cholesterol storage.
Indeed, recent studies demonstrated that cholesterol is
highly concentrated in MAMs (35) in lipid raft-like microdomains, which are also rich in ␴-1 receptors involved
in cholesterol distribution (9). Although the exact role of
MAMs in cholesterol storage and delivery in cells throughout the body are unclear (8), hormone-induced rapid free
cholesterol delivery from ER to mitochondria through
MAMs could support the hypothesis that MAM formation might play a role in steroidogenesis. In support of this
hypothesis, ACSL4, one of the markers used to characterize MAMs, is an enzyme previously shown to be involved
in steroid biosynthesis (29, 36).
ATAD3 is present exclusively in multicellular eukaryotes participating in the control of mitochondrial dynamics at the interface between OMM and IMM, controlling various cell functions including mitochondrial
fission, cell growth, and cholesterol transport (3, 10, 12,
17). ATAD3 is anchored in IMM and enriched in OMMIMM contact sites. Although ATAD3 was initially identified as a 67-kDa protein (long form), a number of
ATAD3 isoforms have been reported, including the 57kDa short form (17). In accordance with previous work
(3), we confirmed that ATAD3 plays a role in steroid hormone formation. Indeed, knocking down ATAD3 in
MA-10 cells induced a profound modification of mitochondria cristae organization, which could affect OMMIMM contact site formation, crucial for optimal cholesterol transfer through the aqueous intermembrane space.
Interestingly, we observed that ATAD3 is present in
both the ER and mitochondria in MA-10 cells, in agreement with previous findings in nonsteroidogenic cells (10,
16), representing a potential candidate in the regulation of
MAM formation. However, we observed that the 67-kDa
long form of ATAD3 is present primarily in mitochondria,
whereas the 57-kDa short form is present in ER. This finding raised the possibility of distinct functions for these 2
ATAD3 isoforms, with the mitochondrial long form ex-
endo.endojournals.org
343
Figure 8. Proposed model for the role of ATAD3 in MAM formation
and steroidogenesis. In the control cells, the 67-kDa long ATAD3 form
is anchored in the IMM, creating the bridge between the IMM and
OMM (upper panel). In response to hormone treatment, the 67-kDa
mitochondrial ATAD3 reaches the ER via its N-terminal domain, which
acts as a scaffold protein driving the formation of MAMs (middle
panel). This subcellular organization may facilitate the transfer of free
cholesterol from the ER to the mitochondria, leading to optimal
steroidogenesis. The absence of ATAD3 induced a loss of the integrity
of MAM regions causing steroidogenesis impairment (lower panel).
tending into the ER and participating in or initiating the
formation of MAMs. This hypothesis is supported by the
finding that isolated MAMs contained only the 67-kDa
long form of ATAD3, the one of mitochondrial origin.
During these studies, a proteomics paper appeared further
supporting the presence of ATAD3 in MAMs (37). A comparison of the 2 forms indicates that the shorter one is
missing the ATP hydrolysis/binding domain, which is important for both the formation of MAMs and steroidogenesis because mutation studies conducted on this domain impaired the formation of the mitochondrial
network (12). Low expression of the 57-kDa isoform in
344
Issop et al
MAMs in Leydig Cell Steroidogenesis
human H295R-S2 cells suggested a species-specific regulation of expression of this isoform and that it might be
involved in other processes such as development and embryogenesis (17). It is also likely that there is a balance
between the isoforms needed to achieve an equilibrium
important to maintain a certain ER-mitochondria distance, as shown for the mitochondrial optic atrophy 1 in
the regulation of mitochondrial membranes in cristae
proper formation (38, 39).
Biochemical studies combined with topological analysis revealed that the long form of ATAD3 has an N-terminal domain containing 50 amino acids and an additional transmembrane domain proposed to form an
␣-helix that drives the insertion of the protein back into the
inner surface of the OMM (12). Although Fang and colleagues (16) identified the long form of ATAD3 in ER,
mitochondria, and MAMs isolated from lung adenocarcinoma cells, there is no evidence as to how ATAD3 anchors to the ER. Data presented herein indicate that this is
achieved through the N-terminal domain of the protein,
which would span through the OMM into the cytosol to
interact with the ER, thus participating in the formation of
the MAM regions.
Transfection of MA-10 cells containing a construct that
encodes truncated ATAD3 protein, where its N-terminal
domain was removed, resulted in reduced synthesis of hormone-dependent progesterone. This observation suggests
that the expressed truncated protein acted as a dominantnegative, competing with the endogenous wild-type protein in its function in steroid formation. The finding that
MAMs contain the 67-kDa long ATAD3 form suggests
that the N-terminal domain of the protein not only spans
the OMM but also found its way and anchored to the ER.
Oligomerization of this domain, which is important for its
interaction with the OMM, is dependent on the ATP-binding domain of the protein (12), suggesting that ATAD3mediated MAM formation may be an ATP-dependent
process. Thus, mitochondrial homeostasis appears to be
important for the function of ATAD3 in steroidogenesis,
and further investigations appear to be necessary to analyze the exact role and anchoring site of this N-domain in
MAMs. In addition, the association of ATAD3 with other
shaping proteins present at this region, such as dynaminrelated protein 1, phosphofurin acidic cluster sorting protein 2, and mitofusin 2 could represent a new site of regulation for steroidogenesis. This is consistent with the
hypothesis that ATAD3 might be participating in the regulation of steroidogenesis by controlling the optimal distance between the 2 organelles. Indeed, recent studies
demonstrated that the distance between ER and mitochondria can vary and that this is tightly dependent on cell
metabolism (40 – 42).
Endocrinology, January 2015, 156(1):334 –345
In Figure 8, the respective localization of the 2 ATAD3
forms is depicted. The 67-kDa long form of ATAD3 is
present in IMM, with the N terminus extending to OMM
and cytosol. A physical association is present under basal
conditions. Upon hormone stimulation, ATAD3 participates in the reorganization of the mitochondrial membrane, facilitating the association of mitochondria with
ER through its N-terminal domain.
MAM formation could not only allow increased influx
for the steroidogenesis substrate cholesterol but could also
facilitate controlled efflux of the mitochondrial steroid
products pregnenolone and progesterone. This is in contrast to the current hypothesis where progesterone and
pregnenolone synthesized in the mitochondria are believed to freely diffuse to the ER (27). However, 22R-HC–
supported steroidogenesis was not affected in ATAD3depleted cells, suggesting that the role of MAMs is likely
to be limited in cholesterol’s influx into mitochondria.
In summary, we demonstrated herein 1) the hormoneinduced ER-mitochondria interactions and MAM formation, which are functional components of the cellular
pathways responsible for channeling cholesterol into the
mitochondria for steroidogenesis, and 2) identified the 67kDa long form of ATAD3 as a critical component in this
process.
Acknowledgments
We thank Dr M. Ascoli (University of Iowa) for providing the
MA-10 cells, Dr W. Rainey (University of Michigan) for the
H295R-2S cells, and Dr A. F. Parlow (National Hormone and
Peptide Program, Harbor-UCLA Medical Center) for supplying
the hCG.
Address all correspondence and requests for reprints to: Dr V.
Papadopoulos, Research Institute of the McGill University
Health Centre, Montreal General Hospital, 1650 Cedar Avenue,
Room C10-148, Montréal, Québec H3G 1A4, Canada. E-mail:
[email protected].
This work was supported by a grant from the Canadian Institutes of Health Research (MOP-102647) and the Canada Research Chair in Biochemical Pharmacology (to V.P.). M.B.R. was
supported in part by a fellowship from the Research Institute of
the McGill University Health Centre. The Research Institute is
supported in part by a center grant from Fonds de la Recherche
Quebec–Santé.
Disclosure Summary: The authors have nothing to disclose.
References
1. Miller WL, Auchus RJ. The molecular biology, biochemistry, and
physiology of human steroidogenesis and its disorders. Endocr Rev.
2011;32:81–151.
doi: 10.1210/en.2014-1503
2. Papadopoulos V, Miller WL. Role of mitochondria in steroidogenesis. Best Pract Res Clin Endocrinol Metab. 2012;26:771–790.
3. Rone MB, Midzak AS, Issop L, et al. Identification of a dynamic
mitochondrial protein complex driving cholesterol import, trafficking, and metabolism to steroid hormones. Mol Endocrinol. 2012;
26:1868 –1882.
4. Issop L, Rone MB, Papadopoulos V. Organelle plasticity and interactions in cholesterol transport and steroid biosynthesis. Mol Cell
Endocrinol. 2013;371:34 – 46.
5. Csordás G, Várnai P, Golenár T, et al. Imaging interorganelle contacts and local calcium dynamics at the ER-mitochondrial interface.
Mol Cell. 2010;39:121–132.
6. Vance JE. Phospholipid synthesis in a membrane fraction associated
with mitochondria. J Biol Chem. 1990;265:7248 –7256.
7. Giorgi C, De Stefani D, Bononi A, Rizzuto R, Pinton P. Structural
and functional link between the mitochondrial network and the
endoplasmic reticulum. Int J Biochem Cell Biol. 2009;41:1817–
1827.
8. Vance JE. MAM (mitochondria-associated membranes) in mammalian cells: Lipids and beyond. Biochim Biophys Acta. 2014;1841:
595– 609
9. Hayashi T, Su TP. Cholesterol at the endoplasmic reticulum: roles
of the sigma-1 receptor chaperone and implications thereof in human diseases. Subcell Biochem. 2010;51:381–398.
10. Li S, Rousseau D. ATAD3, a vital membrane bound mitochondrial
ATPase involved in tumor progression. J Bioenerg Biomembr. 2012;
44:189 –197.
11. Goller T, Seibold UK, Kremmer E, Voos W, Kolanus W. Atad3
function is essential for early post-implantation development in the
mouse. PLoS One. 2013;8:e54799.
12. Gilquin B, Taillebourg E, Cherradi N, et al. The AAA⫹ ATPase
ATAD3A controls mitochondrial dynamics at the interface of the
inner and outer membranes. Mol Cell Biol. 2010;30:1984 –1996.
13. He J, Mao CC, Reyes A, et al. The AAA⫹ protein ATAD3 has
displacement loop binding properties and is involved in mitochondrial nucleoid organization. J Cell Biol. 2007;176:141–146.
14. Da Cruz S, Xenarios I, Langridge J, Vilbois F, Parone PA, Martinou
JC. Proteomic analysis of the mouse liver mitochondrial inner membrane. J Biol Chem. 2003;278:41566 – 41571.
15. Bogenhagen DF, Rousseau D, Burke S. The layered structure of
human mitochondrial DNA nucleoids. J Biol Chem. 2008;283:
3665–3675.
16. Fang HY, Chang CL, Hsu SH, et al. ATPase family AAA domaincontaining 3A is a novel anti-apoptotic factor in lung adenocarcinoma cells. J Cell Sci. 2010;123:1171–1180.
17. Li S, Lamarche F, Charton R, et al. Expression analysis of ATAD3
isoforms in rodent and human cell lines and tissues. Gene. 2014;
535:60 – 69.
18. Ascoli M. Characterization of several clonal lines of cultured Leydig
tumor cells: gonadotropin receptors and steroidogenic responses.
Endocrinology. 1981;108:88 –95.
19. Hanson GT, Aggeler R, Oglesbee D, et al. Investigating mitochondrial redox potential with redox-sensitive green fluorescent protein
indicators. J Biol Chem. 2004;279:13044 –13053.
20. Wieckowski MR, Giorgi C, Lebiedzinska M, Duszynski J, Pinton P.
Isolation of mitochondria-associated membranes and mitochondria
from animal tissues and cells. Nat Protoc. 2009;4:1582–1590.
21. Crowther RA, Amos LA, Finch JT, De Rosier DJ, Klug A. Three
dimensional reconstructions of spherical viruses by fourier synthesis
from electron micrographs. Nature. 1970;226:421– 425.
22. Liu J, Rone MB, Papadopoulos V. Protein-protein interactions me-
endo.endojournals.org
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
345
diate mitochondrial cholesterol transport and steroid biosynthesis.
J Biol Chem. 2006;281:38879 –38893.
Rizzuto R, Pinton P, Carrington W, et al. Close contacts with the
endoplasmic reticulum as determinants of mitochondrial Ca2⫹ responses. Science. 1998;280:1763–1766.
Lewin TM, Van Horn CG, Krisans SK, Coleman RA. Rat liver
acyl-CoA synthetase 4 is a peripheral-membrane protein located in
two distinct subcellular organelles, peroxisomes, and mitochondrial-associated membrane. Arch Biochem Biophys. 2002;404:263–
270.
Lewin TM, Kim JH, Granger DA, Vance JE, Coleman RA. AcylCoA synthetase isoforms 1, 4, and 5 are present in different subcellular membranes in rat liver and can be inhibited independently.
J Biol Chem. 2001;276:24674 –24679.
Kimura T. Transduction of ACTH signal from plasma membrane to
mitochondria in adrenocortical steroidogenesis. Effects of peptide,
phospholipid, and calcium. J Steroid Biochem. 1986;25:711–716.
Miller WL, Bose HS. Early steps in steroidogenesis: intracellular
cholesterol trafficking. J Lipid Res. 2011;52:2111–2135.
Rone MB, Fan J, Papadopoulos V. Cholesterol transport in steroid
biosynthesis: role of protein-protein interactions and implications in
disease states. Biochim Biophys Acta. 2009;1791:646 – 658.
Poderoso C, Duarte A, Cooke M, et al. The spatial and temporal
regulation of the hormonal signal. Role of mitochondria in the formation of a protein complex required for the activation of cholesterol transport and steroids synthesis. Mol Cell Endocrinol. 2013;
371:26 –33.
Duarte A, Poderoso C, Cooke M, Soria G, Cornejo MF, Gottifredi
V, Podesta EJ. Mitochondrial fusion is essential for steroid biosynthesis. PLoS One. 2012;7:e45829.
Aoki A, Massa EM. Early responses of testicular interstitial cells to
stimulation by interstitial-cell-stimulating hormone. Am J Anat.
1972;134:239 –261.
Rowland AA, Voeltz GK. Endoplasmic reticulum-mitochondria
contacts: function of the junction. Nat Rev Mol Cell Biol. 2012;13:
607– 625.
Hayashi T, Rizzuto R, Hajnoczky G, Su TP. MAM: more than just
a housekeeper. Trends Cell Biol. 2009;19:81– 88.
Fujimoto M, Hayashi T. New insights into the role of mitochondriaassociated endoplasmic reticulum membrane. Int Rev Cell Mol Biol.
2011;292:73–117.
Fujimoto M, Hayashi T, Su TP. The role of cholesterol in the association of endoplasmic reticulum membranes with mitochondria.
Biochem Biophys Res Commun. 2012;417:635– 639.
Maloberti P, Cornejo MF, Castillo AF, et al. Enzymes involved in
arachidonic acid release in adrenal and Leydig cells. Mol Cell Endocrinol. 2007;265–266:113–120.
Poston CN, Krishnan SC, Bazemore-Walker CR. In-depth proteomic analysis of mammalian mitochondria-associated membranes
(MAM). J Proteomics. 2013;79:219 –230.
Palmer CS, Osellame LD, Stojanovski D, Ryan MT. The regulation
of mitochondrial morphology: intricate mechanisms and dynamic
machinery. Cell Signal. 2011;23:1534 –1545.
Duvezin-Caubet S, Jagasia R, Wagener J, et al. Proteolytic processing of OPA1 links mitochondrial dysfunction to alterations in mitochondrial morphology. J Biol Chem. 2006;281:37972–37979.
Bravo R, Gutierrez T, Paredes F, et al. Endoplasmic reticulum: ER
stress regulates mitochondrial bioenergetics. Int J Biochem Cell
Biol. 2012;44:16 –20.
Schon EA, Area-Gomez E. Mitochondria-associated ER membranes
in Alzheimer disease. Mol Cell Neurosci. 2013;55:26 –36.
Csordás G, Renken C, Várnai P, et al. Structural and functional
features and significance of the physical linkage between ER and
mitochondria. J Cell Biol. 2006;174:915–921.