MitoNEET is an iron-containing outer mitochondrial membrane protein that regulates oxidative capacity

Abstract

Members of the thiazolidinedione (TZD) class of insulin-sensitizing drugs are extensively used in the treatment of type 2 diabetes. Pioglitazone, a member of the TZD family, has been shown to bind specifically to a protein named mitoNEET [Colca JR, McDonald WG, Waldon DJ, Leone JW, Lull JM, Bannow CA, Lund ET, Mathews WR (2004) Am J Physiol 286:E252–E260]. Bioinformatic analysis reveals that mitoNEET is a member of a small family of proteins containing a domain annotated as a CDGSH-type zinc finger. Although annotated as a zinc finger protein, mitoNEET contains no zinc, but instead contains 1.6 mol of Fe per mole of protein. The conserved sequence C-X-C-X₂-(S/T)-X₃-P-X-C-D-G-(S/A/T)-H is a defining feature of this unique family of proteins and is likely involved in iron binding. Localization studies demonstrate that mitoNEET is an integral protein present in the outer mitochondrial membrane. An amino-terminal anchor sequence tethers the protein to the outer membrane with the CDGSH domain oriented toward the cytoplasm. Cardiac mitochondria isolated from mitoNEET-null mice demonstrate a reduced oxidative capacity, suggesting that mito- NEET is an important iron-containing protein involved in the control of maximal mitochondrial respiratory rates.

Type 2 diabetes is a complex disease involving insulin resistance, decreased insulin secretion, dyslipidemia, and altered nutrient partitioning (1). Mitochondria play a well documented role in insulin secretion and the control of glucose and fatty acid oxidation (2, 3). Insulin resistance is associated with decreased mitochondrial mass and compromised oxidative capacity, suggesting a primary role for mitochondrial dysfunction in the pathogenesis of insulin resistance (4–8).

One of the drugs of choice to treat type 2 diabetes is pioglitazone, a member of the thiazolidinedione (TZD) class of insulin sensitizers. The pharmacology of the TZDs has traditionally been attributed to their function as agonists of the peroxisome proliferator-activated receptor γ (PPARγ), a transcription factor involved in adipocyte differentiation and maturation (9). There is a substantial amount of data indicating that the TZDs exert only a portion of their activity via genomic effects on PPARγ and that nongenomic mechanisms are significant and potentially relevant to their clinical effects (10). For example, the efficacy of the TZDs does not correlate with the binding affinities for PPARγ, and these drugs can elicit beneficial effects in the absence of PPARγ in selected tissues (10, 11). In addition, TZD effects can occur too rapidly for PPARγ-mediated transcriptional events to be primarily responsible (11).

In an effort to identify cellular targets of pioglitazone, Colca et al. (12) identified a protein that was cross-linked to a radiolabeled photoaffinity derivative of pioglitazone. This protein had a mitochondrial association, contained the amino acid sequence Asn-Glu-Glu-Thr (NEET), and was named mitoNEET (12). The important role of the TZDs in treating type 2 diabetes and the identification of mitoNEET as a target of pioglitazone binding prompted us to examine this protein in greater detail.

Results and Discussion

MitoNEET Family Proteins Possess Unique “CDGSH-Type Zinc Finger” Domains.

Colca et al. (12) identified mitoNEET as a target for pioglitazone binding. Because there were no additional studies on mitoNEET, we began our efforts by analyzing the domain organization of the protein. By using the simple modular architecture research tool (SMART) and the National Center for Biotechnology Information (NCBI) conserved domain search algorithms, our analysis showed that mitoNEET contains a domain of ≈40 aa (amino acids 55–93) annotated as a CDGSH-type zinc finger (Fig. 1). Two additional cysteines flank the CDGSH sequence. In addition to the CDGSH domain, mitoNEET has a predicted transmembrane domain localized between amino acids 14 and 32, suggesting that it may be a membrane-anchored protein (Fig. 1).

Fig. 1.

A mitoNEET protein family. Alignment of the amino acid sequences of human mitoNEET, Miner1, and Miner2. Amino acids conserved in humans are shown in gray. Amino acids invariant across multiple species are indicated in yellow. The CDGSH domains are noted by red bars. The transmembrane domain of mitoNEET is marked by a blue box.

To identify other proteins with a CDGSH domain, we performed BLAST searches by using the human mitoNEET CDGSH sequence (residues 55–93). These searches yielded two related human proteins, which we refer to as Miner1 and Miner2 for MitoNEET-related 1 and 2 (Fig. 1). Both mitoNEET and Miner1 have a single CDGSH-type zinc finger, whereas Miner2 contains two of these domains (Fig. 1). Multiple species alignments of the CDGSH-type zinc finger domains from mitoNEET, Miner1, and Miner2 revealed that this family contains the consensus sequence: C-X-C-X₂-(S/T)-X₃-P-X-C-D-G-(S/A/T)-H. In addition to the invariant proline, aspartic acid, and glycine residues, this motif also contains three invariant cysteines and an invariant histidine. Although zinc finger proteins are abundant in the genome, this CDGSH-type zinc finger domain is unique.

The Annotated CDGSH-Type Zinc Finger Protein Contains Iron.

To our surprise, purified recombinant mitoNEET was strikingly red in color, as were recombinant proteins of Miner1 and Miner2. Because zinc finger proteins are not reported to be red in color, this suggested that mitoNEET likely had an unanticipated “cofactor.” To explore this further, mitoNEET_27–108 (lacking amino acids 1–26 of the amino terminus) was expressed in bacteria as a fusion protein with an amino-terminal maltose binding protein (MBP)-His tag. The recombinant MBP-His-mitoNEET_27–108 was found to contain only trace amounts of zinc (0.01 mole of Zn per mole of protein) when analyzed by inductively coupled plasma–high-resolution mass spectrometry (ICP-HRMS) (Table 1). This value was equivalent to the amount of zinc bound to the MBP-His protein alone (when expressed without mitoNEET) and did not change significantly when the bacterial culture media was supplemented with 5 μM ZnCl₂ (data not shown). The ICP-HRMS metal analysis detected 1.6 mole of Fe per mole of MBP-His-mitoNEET_27–108 protein (Table 1), with virtually no iron bound to the MBP-His tag when it was expressed alone (0.008 mole of Fe per mole of protein). Similar iron stoichiometries were seen with GST fusion proteins of both full-length mitoNEET and mitoNEET_27–108 (data not shown). Furthermore, ICP-HRMS analysis showed that recombinant mitoNEET contained no significant amounts of other metals, including nickel, copper, magnesium, calcium, chromium, or manganese (Table 1). These data reveal that mitoNEET binds iron and that the CDGSH domain is not a zinc finger. The fact that all three proteins in the mitoNEET family share this domain, and all three proteins are also red, suggests that it is highly likely that all members of the family are iron-containing proteins.

Table 1.

Metal content (moles of metal per mole of protein) of MBP-His-mitoNEET_27–108 protein determined by ICP analysis

Metal	Concentration
Iron	1.6 ± 0.60
Zinc	0.01 ± 0.01
Nickel	0.09 ± 0.07
Copper	0.04 ± 0.03
Magnesium	<0.001
Calcium	<0.001
Chromium	<0.001
Manganese	<0.001

Data are the average of four independent protein preparations. Values for MBP-His protein were below the limits of detection.

Subcellular Localization of mitoNEET, Miner1, and Miner2.

To address the cellular localization of the mitoNEET family members, mitoNEET, Miner1, and Miner2 were tagged with the carboxyl-terminal V5 epitope and transiently expressed in COS-7 cells. Immunocytochemistry using an anti-V5 antibody showed colocalization of mitoNEET and Miner2 with the mitochondrial marker, MitoTracker Red (Fig. 2 A and C). In contrast, Miner1 appeared to be localized to the endoplasmic reticulum, with strong perinuclear staining and a lacy network fanning out across the cell (Fig. 2B). Expression of the mitoNEET family of proteins as carboxyl-terminal fusion proteins with EGFP gave similar expression patterns (data not shown).

Fig. 2.

Localization of MitoNEET family proteins. (A–C) Immunofluorescent staining of V5-tagged mitoNEET (A), Miner1 (B), and Miner2 (C) transiently transfected for 24 h in COS-7 cells. MitoNEET family proteins appear green. Colocalization of mitoNEET and Miner2 (Left) with MitoTracker Red (MR) (Center) appears yellow in the merged images (Right). DNA is stained blue with DAPI. (D) Immunofluorescent imaging in COS-7 cells of endogenous mitoNEET, MitoTracker Red, and merged images.

MitoNEET mRNA and Protein Are Widely Expressed.

RT-PCR analysis of mRNA from murine tissues revealed that mitoNEET is widely expressed with the highest level of mRNA seen in the heart [supporting information (SI) Fig. 6A]. To further explore the tissue distribution of the mitoNEET protein, we raised polyclonal antiserum against the recombinant protein. Immunofluorescent staining of COS-7 cells to visualize endogenous mito-NEET gave a reticular, mitochondrial pattern that overlapped with MitoTracker Red (Fig. 2D). The mitochondrial staining was blocked by recombinant mitoNEET protein (data not shown). The antibody did not recognize transiently expressed V5-tagged Miner1 or Miner2 (data not shown). These data suggest that the endogenous mitoNEET protein is also localized to the mitochondria.

We also examined the expression of mitoNEET protein in various mouse tissues by immunohistochemical staining (SI Fig. 6B). The MitoNEET protein was present in many tissue types, including insulin-responsive tissues such as the liver, adipose, skeletal muscle, and heart (images 2, 3, 4, and 5), with cells lining the ducts (image 2) and ventricles (image 5) staining particularly robustly. MitoNEET staining appeared to be absent in both the endocrine (islet) and exocrine cells of the pancreas (image 6). There was also no staining with preimmune serum (image 1).

MitoNEET Is Directed to the Outer Mitochondrial Membrane (OMM) by an Amino-Terminal Signal Sequence.

Many nuclear-encoded mitochondrial proteins possess an amino-terminal signal sequence that is recognized by the translocase of the outer membrane complex (13, 14). After mediating import into the mitochondria, this targeting sequence is often, but not always, cleaved by mitochondrial proteases (15). To determine whether mitoNEET was indeed being directed to the mitochondria by an amino-terminal signal sequence, the protein lacking the first 32 amino acids and containing a carboxyl-terminal V5 epitope tag was expressed in COS-7 cells. As expected, the full-length protein was localized to the mitochondria (Fig. 3A), whereas deletion of the first 32 residues completely abolished the mitochondrial localization of mitoNEET (Fig. 3B). When the first 32 amino acids of mitoNEET were fused to the amino terminus of EGFP, this construct was able to redirect EGFP from a nuclear and cytoplasmic localization (Fig. 3C) to the mitochondria (Fig. 3D). This demonstrates that the first 32 amino acids of mitoNEET are both necessary and sufficient to direct mitochondrial localization.

Fig. 3.

MitoNEET is directed to the OMM by an amino-terminal signal sequence. (A–D) Immunofluorescent imaging of transiently expressed mitoNEET proteins in COS-7 cells: V5-tagged full-length mitoNEET (A), V5-tagged mitoNEET lacking amino acids 1–32 (B), EGFP (C), and EGFP with amino acids 1–32 of mitoNEET fused to the amino terminus (D). (E) Immunogold electron microscopic staining of mitochondria in rat kidney tissue sections with either preimmune serum (PI) or anti-mitoNEET antibody. (F) Rat liver mitochondrial subfractions were separated by SDS/PAGE and analyzed by immunobloting by using antibodies against known mitochondrial marker proteins. Abbreviations: SMP, submitochondrial particles; IMS, intermembrane space; VDAC, voltage-dependent anion channel; SMAC, second mitochondrial activator of caspases; PTPMTI, protein tyrosine phosphatase of the mitochondria-1.

The mitochondrion is a highly compartmentalized organelle in which the location of proteins coordinating specific functions is tightly controlled. Determining the exact position of a protein within the mitochondrion is important because the milieu of each compartment will dictate potential interacting proteins and possible biochemical roles. To elucidate mitoNEET's submitochondrial localization, we first performed immunogold electron microscopic labeling of rat kidney tissue. Ultrathin cryosections were probed with a protein A-purified anti-mitoNEET antibody and a 10-nm gold particle-conjugated goat anti-rabbit secondary antibody. Positive staining was detected in most mitochondria with little staining of other subcellular compartments. As shown in Fig. 3E, gold particles could be seen decorating the perimeter of the mitochondria. We counted 1,101 gold particles from numerous fields and scored them according to their proximity to the OMM. Located “on/near” the OMM was defined as being within 30 nm of the OMM, which is the approximate combined length of the primary and secondary antibodies plus the gold particle. The gold particles were partitioned as follows: 74.3% on/near the OMM, 14.1% inside the mitochondria, and 11.6% cytosolic. These results suggest that mitoNEET is likely located at the OMM.

To further define the submitochondrial location of mito-NEET, we separated highly purified rat liver mitochondria into the following fractions: OMM, mitoplasts (consisting of inner membrane and matrix), intermembrane space (soluble material between the inner and outer membranes), submitochondrial particles (vesicles composed of inside–out mitochondrial membranes devoid of most soluble material), and soluble (material from the matrix and intermembrane space). Equal amounts of protein from each fraction were separated by SDS/PAGE and analyzed by immunoblotting with antibodies against mitoNEET and marker proteins (Fig. 3F). MitoNEET was detected only in the membrane-containing fractions (submitochondrial particles, mitoplasts, OMM). Furthermore, it was highly enriched in the OMM fraction, as was the voltage-dependent anion channel protein, VDAC, a known marker for the mitochondrial outer membrane (16). In our experiments, mitoNEET consistently displayed an apparent molecular mass of 13 kDa by immunoblotting, suggesting that the amino-terminal localization sequence is not cleaved from the protein (data not shown). The calculated molecular mass based on amino acid composition is 12.2 kDa. Collectively, these different approaches suggest an OMM localization for mitoNEET.

MitoNEET Is an Integral Outer Membrane Protein Oriented Toward the Cytoplasm.

It is possible for proteins to associate with membranes because of protein–protein or protein–lipid interactions. To determine whether mitoNEET is an integral or peripheral membrane protein, rat liver mitochondria were first swollen by osmotic shock and then treated with either high salt (200 mM KCl) or high pH (0.1 M Na₂CO₃; pH 11.5) and analyzed by SDS/PAGE and immunoblotting. Cytochrome c, which associates with the inner membrane, was stripped efficiently from the mitochondria under these conditions (Fig. 4A) as expected (17, 18). In contrast, mitoNEET remained associated with the mitochondrial pellet as did other known integral membrane proteins such as BCL2 and voltage-dependent anion channel protein (Fig. 4A). Our results suggest that mitoNEET is an integral membrane protein, consistent with the SMART predictions that mitoNEET possesses a single transmembrane region.

Fig. 4.

MitoNEET is an integral membrane protein oriented toward the cytoplasmic face of the OMM. (A) Immunoblot analysis of swollen rat liver mitochondria treated to remove peripheral membrane proteins. Mitochondria were washed with either 200 mM KCl (high salt) or 0.1 M Na₂CO₃, pH 11.5 (high pH), and the proteins remaining in the pellets were separated by SDS/PAGE and immunoblotted. (B) SDS/PAGE and immunoblot analysis of trypsin-treated rat heart mitochondria (osmotically shocked). (C) SDS/PAGE and immunoblot analysis of trypsin-treated rat heart mitochondria (low pH washed). BCL2, B cell lymphoma-2; SMAC, second mitochondrial activator of caspases; VDAC, voltage-dependent anion channel protein.

A protein on the OMM may be oriented toward the intermembrane space or toward the outside of the mitochondrion where it is free to interact with cytoplasmic proteins. It is possible to distinguish between these two orientations by determining proteolytic susceptibility by using intact versus detergent-solubilized mitochondria. During the course of this study, it became apparent that the mitoNEET protein was resistant to trypsin cleavage (Fig. 4B) despite the fact that it contains many potential cleavage sites. We had learned from in vitro studies with recombinant mitoNEET that the stability of the protein is sensitive to pH (data not shown). Therefore, rat heart mitochondria were briefly incubated in an isosmotic buffer at pH 4.6 to partially unfold the mitoNEET protein and increase susceptibility to trypsinolysis. Subsequently, intact and 1% Triton X-100 solubilized mitochondria were treated with trypsin and analyzed by SDS/PAGE and immunoblotting. The intermembrane space protein second mitochondrial activator of caspases, which is susceptible to digestion by trypsin, was examined to ascertain the integrity of the OMM. With intact mitochondria, trypsin treatment resulted in complete degradation of mito-NEET, whereas second mitochondrial activator of caspases was left intact (Fig. 4C), confirming that the OMM was not compromised by the brief exposure to low pH. Thus, mitoNEET is oriented on the OMM such that its CDGSH domain faces the cytoplasm with its amino terminus serving as a membrane anchor.

MitoNEET Regulates Mitochondrial Oxidative Capacity.

Mice with a targeted disruption of the mitoNEET gene, developed by using a proprietary gene trap technology (19), were acquired from Lexicon Genetics (The Woodlands, TX). The phenotypic characterization of the mice will be described in a subsequent publication. By using mitochondria purified from the hearts of mitoNEET wild-type (+/+) and null littermates, we measured complex I-dependent oxygen consumption to evaluate the state 3 (phosphorylating) and the maximal uncoupler-stimulated (state 3_u) respiration rates. These rates were 36 and 31% lower, respectively, in cardiac mitochondria from mitoNEET-deficient mice when compared with mitochondria from wild-type littermates (Fig. 5A). In contrast, the lack of mitoNEET expression had no effect on rates of resting (state 4_o) respiration (Fig. 5A). These data suggest that the loss of mitoNEET expression results in a decrease in the maximal capacity of heart mitochondria to carry out electron transport and oxidative phosphorylation, yet has no effect on the basal rate of proton leak across the inner mitochondrial membrane. When examined by electron microscopy, the morphology and numbers of mitochondria from the hearts of mitoNEET-null mice were indistinguishable from those of wild-type littermates (data not shown). Immunoblot analysis confirmed the lack of mitoNEET expression in the null mice and revealed similar levels of mitochondrial proteins as control mice when probing either cardiac tissue homogenate or isolated mitochondria (Fig. 5B and SI Fig. 7). Thus, the observed differences in respiration rates are probably not because of a compromise in the quantity or structural integrity of mitochondria in the mitoNEET-null mice. This is in contrast to the mechanism of other known regulators of maximal oxidative capacity, such as PPARγ coactivator-1α and 5′-AMP kinase, which appear to operate through control of mitochondrial biogenesis (20).

Fig. 5.

MitoNEET protein levels affect maximal mitochondrial oxidative capacity. (A) Respiration rates of cardiac mitochondria from mitoNEET (+/+) (filled bars) and (−/−) (open bars) mice. Error bars represent the standard deviation from n = 7 (+/+) and n = 6 (−/−) from three separate preparations (P < 0.01). (B) SDS/PAGE and immunoblot analysis of purified cardiac mitochondria from mitoNEET (+/+) and (−/−) mice. NDUFB6 is a subunit of complex I.

Conclusions and Summary.

Our studies demonstrate that mito-NEET is an integral protein of the OMM with the CDGSH domain facing the cytoplasm. It is directed to the OMM by an amino-terminal mitochondrial targeting sequence that possesses all of the hallmarks of a mitochondrial signal anchor sequence. Signal anchor sequence proteins are anchored to the OMM by a transmembrane domain and present a hydrophilic domain to the cytoplasm (13, 21). They typically have a mitochondrial targeting sequence with a single helical hydrophobic transmembrane segment and a net positive charge following the transmembrane segment to act as a stop transfer signal. In addition, signal anchor sequence proteins possess at least one positively charged amino acid and two or more amino acids containing hydroxyl groups amino-terminal to the transmembrane segment (21). MitoNEET meets all of these criteria (Fig. 1A). Positioned with the CDGSH domain oriented toward the cytoplasm, mito-NEET is ideally suited to communicate signals between the mitochondria and the rest of the cell.

In this study, we show that mitoNEET is the defining member of a small family of proteins that is characterized by the presence of a unique CDGSH-type zinc finger domain. Even though mitoNEET is annotated as a zinc finger protein, it contains no zinc but contains 1.6 moles of Fe per mole of protein. It is highly likely that the conserved sequence C-X-C-X₂-(S/T)-X₃-P-X-C-D-G-(S/A/T)-H is important in binding the iron. This sequence is conserved in all mitoNEET family members, suggesting the possibility that all three proteins are iron containing.

We further demonstrate that mitoNEET is a widely expressed protein that localizes to the OMM. By using mitochondria isolated from the hearts of mitoNEET-null mice, we also demonstrated that complex I-driven state 3 respiration is significantly altered. Our data support the concept that mitoNEET plays a key role in regulating electron transport and oxidative phosphorylation. The role that mitoNEET plays in the etiology of type 2 diabetes and in mediating some of the effects of TZDs remains to be determined. The degree of mitoNEET sequence conservation among vertebrates suggests that mitoNEET most likely serves an important role in mitochondrial function. Indeed, iron-containing proteins frequently function as dynamic redox-sensitive molecules in the mitochondria, participating in critical processes such as electron shuttling through the electron transport chain, regulation of enzymatic activity, and synthesis of heme and iron–sulfur clusters (22, 23). Dysregulation of iron metabolism has also been associated with diabetes (24). We are currently exploring the biochemical properties of mitoNEET and its regulation by pioglitazone.

Materials and Methods

Identification of cDNAs and Construction of Mammalian Expression Plasmids.

Human sequences similar to the mitoNEET protein were identified by tBLASTn and PsiBLAST searches of the NCBI nonredundant database and were then confirmed by BLAST searches of the human expressed sequence tag database. Domain searches were performed by using the SMART algorithms and the NCBI conserved domain search. Human cDNA clones (Invitrogen, San Diego, CA) were obtained corresponding to the published mitoNEET sequence (MGC: 14684, GI:37590611), to an MGC clone for a different but related protein (MGC: 40370, GI:21619026), and a cDNA for a predicted hypothetical protein (GI:42661145). The latter two were named Miner1 and Miner2, respectively. Sequence alignments were performed with the MacVector Program.

The ORFs for human mitoNEET, Miner1, and Miner2 were amplified by PCR and cloned into the pCDNA3.1D-V5-His/Topo vector (Invitrogen). Constructs encoding either mito-NEET lacking amino acids 1–32 (Δ1–32) or only amino acids 1–32 of mitoNEET fused to EGFP were generated by PCR and cloned into the pCDNA3.1D-V5-His/Topo and pEGFP-N1 vectors (Clontech, Palo Alto, CA), respectively.

Metal Quantitation.

The metal content of the purified recombinant MBP-His-mitoNEET_27–108 protein (see SI Materials and Methods for expression details) was determined by using a Finnigan Element ICP-HRMS by Dr. Ted J. Huston (Department of Geological Sciences, University of Michigan, Ann Arbor, MI).

RT-PCR and Immunohistochemistry.

RT-PCR was performed with mouse cDNA samples from Clontech and primers to the entire coding sequence of murine mitoNEET. Digital images of the PCR products after 25 cycles were quantified by densitometry and normalized to the expression of GAPDH in the samples. For immunohistochemical staining, paraffin-embedded mouse tissues were immunostained as described previously (25). Tissue slices were stained with the above-mentioned protein A-purified anti-mitoNEET antibody (1:1,000) and 3-amino-9-ethylcarbazole substrate. Counterstaining was with Mayer's hematoxylin. Images were taken by using a Zeiss (Oberkochen, Germany) microscope with an Axiocam HRc color camera and Axiovision software.

EM.

Rat kidney was perfusion fixed with 4% paraformaldehyde in 0.1 M phosphate buffer, and then immersion fixed overnight in 4% paraformaldehyde in phosphate buffer, as described previously (26). The tissue was embedded in 2% gelatin, cryoprotected overnight in 2.3 M sucrose, and sectioned. Immunolabeling was performed by using the anti-mitoNEET antibody at 1:125 dilution and 10-nm gold-conjugated goat anti-rabbit IgG (GE Healthcare Bio-Sciences, Piscataway, NJ) at 1:25 (27). Images were captured at a magnification of ×21,000.

High Salt and Carbonate Treatment of Mitochondria.

Histodenz-purified rat liver mitochondria were treated as described previously (26), with osmotic shock followed by a wash in either high salt (200 mM KCl) or carbonate (0.1 M Na₂CO₃, pH 11.5) (17, 26) and then analyzed by SDS/PAGE.

Proteolysis of Mitochondria.

Mitochondria from rat hearts were purified by differential centrifugation (see SI Materials and Methods), swollen by osmotic shock, and treated with trypsin for 10 min at room temperature in an isotonic buffer [125 mM sucrose, 10 mM Tris (pH 8.0), and 1 mM EGTA] with or without the addition of 1% Triton X-100 as described (26). For the low pH treatment, 200 μg of mitochondria were resuspended in 200 μl of MSHE [a buffer containing 210 mM mannitol, 70 mM sucrose, 5 mM Hepes (pH 7.4), and 1 mM EGTA (pH 4.6)], incubated on ice for 5 min, pelleted, and then trypsin treated for 20 min at 37°C in the buffer described above. Samples were analyzed by SDS/PAGE and immunoblotting.

Isolation of Mouse Heart Mitochondria and Bioenergetic Analysis.

Mice with targeted disruption of the mitoNEET gene on the C57Bl/6 (albino) × 129 SvEv Brd (agouti) background were acquired from Lexicon Genetics through their Omnibank program. The mice were originally developed by using their proprietary gene trap technology (19). For the mitochondrial respiration studies, mice between ages 9–11 weeks of age were used. Cardiac mitochondria were isolated from wild-type and mitoNEET null animals essentially as described by Yen et al. (28) with the following modifications. Hearts were finely minced in Hanks' buffered saline supplemented with 10 mM 2,3-butanedione monoxime, 5% FBS, 30 mM mannitol, 20 mM Hepes, 1 mM EGTA, and 2 mM MgSO₄. Minced tissue was kept on ice in DMEM containing 5% FBS, 10 mM 2,3-butanedione monoxime, 2 mM Gln, 1.8 mM EGTA, 30 mM mannitol, 20 mM Hepes, and 2 mM sodium pyruvate. Mitochondrial isolation was performed in medium containing 210 mM mannitol, 70 mM sucrose, 1 mM EGTA, and 0.1% BSA, pH 7.0. The proteolytic step by using Nargarse was omitted. Mitochondria (0.5 mg/ml) were suspended in 0.25 ml of basal saline medium (125 mM KCl, 5 mM Hepes/KOH, 2 mM phosphate, and 1 mM MgCl₂, pH 7.4; 37°C) supplemented with the complex I-linked substrates glutamate (5 mM) and malate (5 mM) in a Hansatech Oxytherm electrode unit analyzed with Oxygraph Plus software (Hansatech, Pentney King's Lynn, U.K.). State 3 respiration was initiated by the addition of 80 μM ADP. Resting respiration (state 4_o) was then induced by the addition of 5 μg/ml oligomycin, followed by measurement of maximal electron transport chain activity after titration with carbonyl cyanide p-trifluoromethoxyphenylhydrazone (up to 30 nM). Differences in rates were analyzed by two-way ANOVA with a post hoc F test.

Cell Culture, Immunocytochemistry, Preparation of Anti-MitoNEET Antisera, SDS/PAGE, and Immunoblotting.

See SI Materials and Methods.

Abbreviations

TZD

thiazolidinedione

PPARγ

peroxisome proliferator-activated receptor γ

Miner1

MitoNEET-related 1

Miner2

MitoNEET-related 2

MBP

maltose binding protein

ICP-HRMS

inductively coupled plasma–high-resolution mass spectrometry

OMM

outer mitochondrial membrane.