The biogenesis of the mitochondrial inner membrane is dependent on two distinct 70 kDa protein complexes. TIMM8a partners with TIMM13 in the mitochondrial intermembrane space to form a 70 kDa complex and facilitates the import of the inner membrane substrate TIMM23. We have identified a new class of substrates, citrin and aralar1, which are Ca 2+ -binding aspartate/glutamate carriers (AGCs) of the mitochondrial inner membrane, using cross-linking and immunoprecipitation assays in isolated mitochondria. The AGCs function in the aspartate–malate NADH shuttle that moves reducing equivalents from the cytosol to the mitochondrial matrix. Mohr-Tranebjaerg syndrome (MTS/DFN-1, deafness/dystonia syndrome) results from a mutation in deafness/dystonia protein 1/translocase of mitochondrial inner membrane 8a ( DDP1 / TIMM8a ) and loss of the 70 kDa complex. A lymphoblast cell line derived from an MTS patient had decreased NADH levels and defects in mitochondrial protein import. Protein expression studies indicate that DDP1 and TIMM13 show non-uniform expression in mammals, and expression is prominent in the large neurons in the brain, which is in agreement with the expression pattern of aralar1. Thus, insufficient NADH shuttling, linked with changes in Ca 2+ concentration, in sensitive cells of the central nervous system might contribute to the pathologic process associated with MTS.
MTS is a recessive X-linked neurodegenerative disorder characterized by progressive sensorineural deafness, dystonia, cortical blindness, dysphagia and paranoia resulting from mutations in the DDP1 / TIMM8a locus ( 1 – 3 ). Neuronal cell loss was shown in the optic nerve, retina, striate cortex, basal ganglia and dorsal roots of the spinal cord in two unrelated families with different mutations in the DDP1 / TIMM8a locus ( 3 ). In contrast to most mitochondrial diseases that affect both muscular and neural tissues ( 4 ), MTS defects are located in neural tissue and do not seem to affect assembly of the respiratory complexes ( 5 , 6 ). Rather, specific defects in mitochondrial function that contribute to MTS are not well understood.
Yeast contain five small Tim proteins, Tim8p, Tim9p, Tim10p, Tim12p and Tim13p, that are 25% identical and share a conserved ‘twin CX 3 C’ motif ( 7 ). Previous studies have suggested that the cysteine residues of the twin CX 3 C motif coordinate zinc, but a set of recent studies suggest that the cysteine residues form intramolecular disulfide bonds ( 7 – 12 ). Tim8p partners with Tim13p to form a soluble 70 kDa complex in the intermembrane space ( 13 – 15 ). Tim9p partners with Tim10p to form a similar 70 kDa complex, but a fraction of Tim9p and Tim10p associate with Tim12p and Tim22p at the inner membrane ( 16 , 17 ). TIM8 , the DDP1 homolog, and TIM13 are not essential for viability, but the remaining small Tim proteins are essential for viability ( 13 ). Genetic studies suggest that the essential function of the small proteins is at the inner membrane ( 18 ). The small Tim proteins have a chaperone-like function, binding to the hydrophobic regions of inner membrane substrates, in the intermembrane space ( 8 , 9 ). The 70 kDa complexes in the intermembrane space thus facilitate import, but are not essential for the process.
In mammals, DDP1/TIMM8a partners with TIMM13 to form a 70 kDa complex in the mitochondrial intermembrane space and is part of the TIM22 translocation apparatus for the import and assembly of inner membrane proteins ( 5 , 19 ). MTS results from defects in assembly of the DDP1/TIMM8a–TIMM13 complex ( 5 ). An additional 70 kDa complex is formed by small Tim proteins TIMM9 and TIMM10 ( 20 ).
Studies in Saccharomyces cerevisiae show that the Tim9p–Tim10p complex binds to most carrier proteins such as the ADP/ATP carrier, phosphate carrier and dicarboxylate carrier and the import components, Tim17p and Tim22p ( 16 , 21 , 22 ). In contrast, the only inner membrane substrate that has been previously identified for Tim8p–Tim13p is the import component Tim23p that contains a large N-terminal domain in the intermembrane space ( 5 , 14 , 15 , 21 ). Likewise, the DDP1–TIMM13 complex mediates the import of Tim23p and TIMM23 ( 5 , 19 ). Two recent studies also show that the small Tim proteins participate in the early steps of Tom40p [the pore of the translocase of the outer membrane (TOM) complex] biogenesis ( 23 , 24 ). However, inactivation of the small Tim complexes had no noticeable effect on the overall assembly of Tom40p. Because Tim23p and Tom40p are essential for viability and TIM8 and TIM13 are not essential in yeast, a decrease in Tim23p and Tom40p biogenesis may contribute to MTS ( 5 , 6 ). However, pleiotropic defects in both muscular and neural tissues would also be predicted because the TOM complex is required for the import of essentially all mitochondrial proteins. Therefore, additional substrates may depend on DDP1–TIMM13 for biogenesis ( 5 , 6 ).
From previous studies, it has been shown that MTS results from a defect in assembly of the DDP1/TIMM8a–TIMM13 complex ( 5 ) that presumably leads to a defect in the import of inner membrane proteins in specific regions of the central nervous system (CNS) ( 3 , 6 ). We therefore investigated the tissue expression pattern of DDP1 and TIMM13 and the import of inner membrane proteins that are unique to mammals. We have identified a new class of substrates for DDP1–TIMM13: aralar1 and citrin, Ca 2+ -responsive aspartate/glutamate carriers (AGC), in addition to the yeast Agc1p ( 25 – 27 ). The AGC carriers function in the malate–aspartate NADH shuttle ( 25 , 26 ). Aralar1, DDP1 and TIMM13 have similar expression patterns in the brain, particularly in large neurons such as Purkinje cells. These results suggest that a decrease in the NADH shuttle might contribute to molecular defects associated with MTS.
DDP1 and TIMM13 are expressed in Purkinje cells and other large neurons
Previous studies have suggested that defects in the function of basal ganglia contribute to MTS symptoms ( 1 , 6 , 28 – 30 ); but in lieu of an animal model, it is difficult to determine the specific regions of the CNS that might be affected. To begin understanding the molecular basis of MTS, we first investigated the expression pattern of DDP1 and TIMM13 in mouse tissues using immunoblot analysis (Fig. 1 A) and in sectioned CNS using immunocytochemistry with antibodies against mammalian import components (Fig. 1 B–P). Increasing amounts of protein lysate from mouse tissues were separated by SDS–PAGE and blotted with antibodies specific for DDP1, TIMM13 and the matrix protein Hsp60 (Fig. 1 A). The abundance of DDP1 and TIMM13 was decreased in muscle and heart but both proteins expressed to higher levels in brain and liver (Fig. 1 A). In CNS sections, both DDP1 and TIMM13 showed similar specific expression patterns, with the most prominent expression in the soma and the dendritic portion of the Purkinje cells of the cerebellum (Fig. 1 B–D,I,J,M and N), but not in the glial cells. Scattered expression also was detected in the brain stem, olfactory bulb, substantia nigra, hippocampus and striatum (data not shown). Localization was confirmed by using polyglutamylated tubulin and glial fibrillary acidic protein as specific markers for neurons and glial cells, respectively (Fig. 1 E and F). The similar expression pattern thus parallels the biochemical studies that show DDP1 and TIMM13 assembling as partner proteins in the mitochondrial intermembrane space ( 5 ).
We also investigated the expression of matrix chaperone Hsp60 and the inner membrane import component TIMM17a in the cerebellar region to characterize the expression of additional mitochondrial biogenesis components. TIMM17 mediates the import of precursors with an N-terminal targeting sequence ( 31 , 32 ). In yeast, Tim17p import is facilitated by Tim9p–Tim10p, whereas Tim23p import is facilitated by Tim8p–Tim13p ( 21 ). Two genes TIMM17a and TIMM17b are present in the mouse genome and are similar, except for a large sequence divergence at the C-terminus ( 32 ); the polyclonal antibody detected the TIMM17a protein specifically. TIMM17a seemed to be expressed in basket cells ( 33 ) (Fig. 1 O and P). In contrast, Hsp60 protein expression was prominent in most regions of the cerebellum (Fig. 1 K and L).
The AGCs are substrates for DDP1–TIMM13
Because DDP1 and TIMM13 showed specific expression patterns in the CNS, we considered whether a subset of mammalian mitochondrial carrier proteins might have a similar expression pattern. Indeed, studies by Satrústegui and colleagues ( 27 , 34 – 36 ) showed that aralar1 is a mitochondrial carrier with an N-terminal domain in the intermembrane space that has specific expression in the large neurons of the CNS, including Purkinje cells. Aralar1 and citrin are two isoforms of the AGC that display tissue-specific expression. Mutations in citrin have been linked to citrullinemia type II ( 37 ). These carriers contain an N-terminal domain in the intermembrane space that contains four Ca 2+ -binding EF hands ( 25 ) (Fig. 2 A). Thus, the AGCs are activated by an increase in Ca 2+ concentration on the intermembrane space side of the inner membrane, bypassing the requirement to transport Ca 2+ into the mitochondrion ( 25 ). Yeast also contains a similar AGC, Agc1p, with an N-terminal domain but lacking the EF hands ( 26 ). The AGCs are similar to Tim23p because of the soluble N-terminal domain in the intermembrane space (Fig. 2 A), which may be a potential characteristic for DDP1–TIMM13 substrates. In contrast, most carriers such as AAC and the import components Tim17p and Tim22p lack a soluble domain and are import substrates for the Tim9p–Tim10p complex ( 21 ) (Fig. 2 A).
We therefore investigated the import pathway of the AGCs using yeast and mammalian mitochondria because the import pathway of TIMM23 is conserved among organisms ( 5 , 38 ). To test if the small Tim complexes bind directly to the inner membrane substrates, we synthesized the precursors in a reticulocyte lysate system for in organello import assays. Large integral membrane proteins often do not translate well in an in vitro transcription–translation assay (data not shown) and full-length citrin, aralar1 and Agc1p did not translate robustly. However, our previous studies have shown that the carriers are translocated through the TOM as a loop ( 8 , 9 ); and the targeting information is located in the C-terminal membrane spanning domains ( 22 , 39 ). In addition, Palmieri et al. ( 25 ) have shown that the mitochondrial targeting information in the AGCs is located in the C-terminal carrier domain and the short calcium-binding mitochondrial carriers are alternatively spliced at the N-terminus ( 40 ). Thus, modifications to the N-terminus do not affect the import pathway of the carriers.
Citrin contains an N-terminal soluble domain from 1 to 323 followed by the C-terminal carrier domain (324–676). A citrin construct (designated citrin 267–676 ) lacking the N-terminal 266 amino acids, but including 57 residues of the soluble domain, translated well and was used for import studies. Citrin 267–676 was imported into mouse liver mitochondria, followed by cross-linking and immunoprecipitation with antibodies against DDP1 and TIMM13 (Fig. 2 B). During the import assays, the mitochondrial membrane potential was dissipated because these conditions trap the import intermediate in the intermembrane space and have been shown to provide optimal binding with the small Tim complexes ( 16 , 41 ). Antibodies against DDP1 and TIMM13 immunoprecipitated cross-linked citrin, indicating a direct interaction between DDP1–TIMM13 and the substrate (denoted by asterisks, Fig. 2 B).
We analyzed the substrate specificity of the small Tim proteins in greater detail using yeast mitochondria (Fig. 2 C) because protein import is highly conserved between yeast and mammals. The import pathways of citrin 267–676 , Agc1p (yeast AGC), Tim23p (Tim8–Tim13p substrate) and AAC (Tim9p–Tim10p substrate) were investigated in detail using cross-linking and immunoprecipitation with antibodies against the small Tim proteins and a control Tim22p. For optimal translation, the N-terminus of Agc1p (Agc1 508–902 ) was removed at amino acid 507, leaving the mitochondrial carrier domain for translation ( 26 ). Substrates were imported into yeast mitochondria in the absence of a membrane potential followed by cross-linking and immunoprecipitation with antibodies against the small Tim proteins. As shown previously, the Tim9p–Tim10p complex, but not the Tim8p–Tim13p complex, bound specifically to AAC (Fig. 2 C). In contrast, Tim23p import was mediated by the Tim8p–Tim13p complex, although cross-linking has been shown for Tim9p and Tim10p ( 14 ). The Tim8p–Tim13p complex as well as Tim10p cross-linked strongly to citrin 267–676 , whereas all four small Tim proteins cross-linked to the Agc1 508–902 precursor. Because Tim9p and Tim10p are present both in the 70 kDa complex and the inner membrane 300 kDa Tim22p complex, it is possible that binding might occur at the inner membrane. As a specificity control, Tim22p has been included because, in the absence of a membrane potential, Tim22p does not bind to the carriers ( 42 ).
We investigated whether the soluble N-terminal domain might be recognized specifically by the Tim8p–Tim13p complex by investigating the import of a set of aralar1 constructs with increasing deletions at the N-terminus (Fig. 2 D). The N-terminal soluble domain is from 1 to 330 and the C-terminal 331–680 residues contain the mitochondrial carrier domain. The following aralar1 constructs were generated—aralar1 189–680 , aralar1 356–680 and aralar1 335–680 —and used for import studies. Both the Tim8p–Tim13p complex and the Tim9p–Tim10p complex were cross-linked to aralar1 (Fig. 2 D). As the N-terminal domain of aralar1 was shortened, the cross-links to Tim10p became more prominent, but Tim8p, Tim9p and Tim13p still bound to aralar1. The Tim8p–Tim13p and the Tim9p–Tim10p complexes thus bound to multiple regions of aralar1. However, because previous studies have shown that DDP1–TIMM13 does not bind to the carriers and import components that reside almost entirely in the inner membrane, citrin, aralar1 and Agc1p are proposed to be new substrates for the DDP1–TIMM13 complex.
We further investigated the import of citrin 267–676 , aralar1 189–680 and control AAC into isolated yeast mitochondria expressing the human DDP1/TIMM8a–TIMM13 complex instead of the yeast Tim8p–Tim13p complex (Fig. 3 ). Import assays were completed with protease treatment to remove non-imported precursor and carbonate extraction to confirm that the imported citrin 267–676 , aralar1 189–680 and AAC were inserted into the inner membrane. AAC import was identical in mitochondria lacking the yeast Tim8p–Tim13p complex and expressing the human DDP1–TIMM13 complex, confirming that AAC import is not dependent on the Tim8p–Tim13p complex. The import of the citrin 267–676 and aralar1 189–680 into mitochondria that lacked the Tim8p–Tim13p complex was decreased 80%. Import, however, in mitochondria with the DDP1/TIMM8a–TIMM13 complex was restored to wild-type levels. Similar import results also were obtained with Tim23p ( 5 ).
A lymphoblast cell line derived from a new MTS family is defective in mitochondrial import
Loss of function of the AGCs leads to citrullinemia type II, a disease that affects liver function ( 37 ). At least one common metabolic denominator seems to be a decreased activity in the aspartate–malate NADH shuttle that is important for moving reducing equivalents from the cytoplasm to the mitochondrial matrix. Recently, a lymphoblast cell line was provided from a new family with MTS containing the Q34X mutation in DDP1, which results from a point mutation in the ddp1 gene (100C→T). This mutation yields a truncated DDP1 protein lacking the twin CX 3 C motif, which is important for assembly of the small Tim complexes ( 5 , 7 ). Previously, fibroblast cell lines only were available for metabolic studies and failed to show notable metabolic differences ( 5 ). We therefore characterized the lymphoblast cell line (Fig. 4 ); the control and Q34X lymphoblast line were identical with respect to growth and general function (data not shown).
We investigated whether the DDP1–TIMM13 complex assembled in lymphoblast cell lines. Lysates were subjected to co-immunoprecipitation assays using antibodies against DDP1 and Timm13 (Fig. 4 A) followed by immunoblot analysis with the DDP1 antibody. In the control lymphoblast cell line, DDP1 immunoprecipitated with its cognate antibody and co-immunoprecipitated with TIMM13 antibody. However, DDP1 was not detectable in the Q34X cell line, because it failed to immunoprecipitate; a smaller-sized DDP1 protein and TIMM13 also were not detected (data not shown). We tested the import of matrix-targeted Su9-DHFR, Tim23p and citrin 267–676 into semi-permeabilized lymphoblast cell lines (Fig. 4 B). In the presence of a membrane potential, Su9-DHFR, Tim23p and citrin 267–676 were imported into the control lymphoblasts. Insertion of citrin and Tim23p into the inner membrane was confirmed by carbonate extraction. In the Q34X lymphoblast cell line, import of all substrates was severely decreased. Because Su9-DHFR import depends on TIMM23 biogenesis, a decreased import of TIMM23 may cause a subsequent decrease in the import rate of substrates such as Su9-DHFR that depend on the TIM23 complex for import ( 21 ). The Q34X cell line thus seems to have a general import defect.
Previous experiments have shown that loss of the DDP1–TIMM13 complex does not affect the OXPHOS system significantly ( 5 , 6 ). Because the AGCs are important for transferring reducing equivalents from the cytosol to the mitochondria, we investigated whether loss of DDP1–TIMM13 affected mitochondrial NADH levels (Fig. 5 ). In previous experiments, yeast cells lacking Agc1p were unable to grow on acetate as a carbon source, but were able to grow on ethanol ( 26 ). Yeast cells lacking the Tim8p–Tim13 complex (Δtim8Δtim13) were serially diluted on minimal dextrose, acetate and ethanol media (Fig. 5 A). Growth was similar between the parental and the Δtim8Δtim13 strain on dextrose medium, but the Δtim8Δtim13 strain showed decreased growth on both acetate and ethanol media. The decreased growth on acetate media may be attributed to a defect in Agc1p biogenesis, whereas the decreased growth on ethanol may be ascribed to a defect in Tim23p biogenesis.
Citrin and aralar1 overexpression in cultured cells stimulate the malate–aspartate shuttle, leading to increased intramitochondrial NADH and increased 3-(4,5-dimethyl-thiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) reduction by mitochondria ( 25 ). We adopted this assay, replacing MTT with 2,3-bis(2-methoxy-4-nitro-5-sulfophenyl)-5-[(phenylamino) carbonyl]-2H-tetrazolium hydroxide (XTT) and measured XTT reduction in the Q34X lymphoblast cell line. Cells were permeabilized with digitonin and XTT reduction was measured under basal conditions, in the presence of 1 m M glutamate, 5 m M malate and 10 m M lactate (G+M+L) or 2 m M succinate. The Q34X lymphoblasts displayed a trend for a decreased XTT reduction under basal conditions and in the presence of metabolic substrates (Fig. 5 B), which reflect a decrease in NADH levels. We also investigated the abundance of aralar1 using a commercial antibody under similar conditions (Fig. 5 C); aralar1 amounts were normalized against TOMM40 of the outer membrane. In the control lymphoblast cells, the abundance of aralar1 increased in the presence of glutamate, malate and lactate, conditions that should increase NADH shuttle activity. In the Q34X cell line, the aralar1 levels remained similar in all carbon sources. Loss of the DDP1–TIMM13 complex in lymphoblasts thus affects mitochondrial function.
MTS is an X-linked disease characterized by a variety of symptoms including deafness, dystonia, paranoia, blindness and mental deterioration ( 1 , 2 , 6 , 30 ). The symptoms can vary between and within families and female carriers have developed symptoms later in life ( 1 , 29 ). As opposed to many mitochondrial diseases that affect muscular and neural tissues, MTS seems to predominantly affect neural tissues. Neuronal cell loss was shown in the CNS in two unrelated families with different mutations in the DDP1 / TIMM8a locus ( 3 ). Understanding the molecular basis has been difficult, especially because a suitable model has not been developed. As a result, we have continued biochemical characterization of DDP1 and its partner protein TIMM13 in mammalian systems and yeast.
We have obtained a new lymphoblast cell line from an MTS patient in which a point mutation in the ddp1 gene (100→T) codes for a truncated DDP1 protein (Q34X) that lacks the twin CX3C motif. The shortened DDP1 protein did not assemble with TIMM13 into a stable complex, which has been shown previously with mutations in the DDP1 gene ( 5 , 6 ). The Q34X cell line showed general defects in import, but obvious differences in growth rate were not observed. Previous studies with fibroblast cell lines have failed to show defects in the oxidative phosphorylation system ( 5 , 6 ). In the Q34X lymphoblast cell lines, a general decrease in NADH and aralar1 levels was observed. Although possible metabolic changes can occur from establishing lymphoblast cell lines ( 43 ), the Q34X cell line seems to be a reliable model for assessing the function of the NADH shuttle and protein import system.
To better understand the molecular basis of MTS, we investigated the expression pattern of DDP1 and TIMM13 in the CNS with immunocytochemistry. DDP1 and TIMM13 were prominently expressed in the soma and dendritic portion of Purkinje cells, and scattered expression also was detected in the brain stem, olfactory bulb, substantia nigra, hippocampus, and striatum. In contrast, Hsp60 expression seemed to be present in both neurons and glial cells and TIMM17a expression was localized to basket cells. These studies support the concept that expression of mitochondrial proteins throughout the CNS is not uniform ( 44 ). As an example, cytochrome oxidase expression is highest in the dendritic portion of neurons of the basal ganglia, thalamus, brain stem and spinal cord, where metabolic activity is the highest ( 44 ). The calcium-regulated carrier aralar1 also was localized preferentially in neurons (including Purkinje cells) with a high cytochrome oxidase content ( 36 ), suggesting that aralar1 may function in energy-dependent processes. Our investigations suggest that DDP1–TIMM13 function may be most important in assembly of mitochondria in the neuronal cells and suggest that misassembly of mitochondria in a specific subset of neurons might contribute to MTS.
Most members of the mitochondrial carrier family have a conserved membrane topology with six transmembrane domains and the N- and C-termini facing into the intermembrane space ( 45 ). Yeast have three dozen members and mammals have additional ones such as citrin, aralar1 and the uncoupling proteins that are not present in yeast. The import pathway of the carriers has been characterized and the targeting information is repeated among the membrane spanning domains ( 8 , 9 , 46 ). The small Tim proteins act as chaperones to maintain the import competency of the hydrophobic inner membrane proteins in the aqueous intermembrane space. The Tim9p–Tim10p complex has been shown to bind to most substrates tested including the ADP/ATP carrier, dicarboxylate carrier and phosphate carrier ( 21 ).
In contrast, only one inner membrane substrate, Tim23p, has been identified for Tim8p–Tim13p ( 5 , 14 , 15 , 19 , 21 ). The Tim8p–Tim13p complex binds to both the transmembrane and N-terminal domains of Tim23p ( 9 ), and the Tim9p–Tim10p complex also was shown to bind to Tim23p ( 14 , 15 ), but because Tim9p and Tim10p function both in the intermembrane space and inner membrane, binding could occur at a later stage associated with insertion. However, specific interactions between the Tim8p–Tim13p complex and other inner membrane substrates have not been documented. We therefore characterized the import pathway of the mammalian specific carriers, citrin and aralar1, because aralar1 showed a similar expression pattern in neural tissue ( 36 ). Indeed, the DDP1–Timm13 complex binds to the AGCs and the mammalian DDP1–TIMM13 complex restored citrin and aralar1 import into mitochondria lacking the homologous yeast complex.
Both Tim23p and AGCs share the common property of a soluble N-terminal domain that might be an ‘earmark’ for DDP1–TIMM13 substrates. Our previous studies have shown that the Tim8p–Tim13p complex binds to both the soluble and the membrane spanning domains of Tim23p, whereas the Tim9p–Tim10p complex mediates the import of typical carriers and import components Tim17p and Tim23p ( 9 ). With the shortened aralar1 constructs, the DDP1–TIMM13 complex was cross-linked, indicating that the DDP1–TIMM13 complex also binds to the carrier domain of aralar1. Tim9p and Tim10p also bound to aralar1, and Tim10p cross-links were more prominent as the N-terminal domain was shortened. When studying the biogenesis of carriers that lack a soluble domain, we have not previously identified substrates for the DDP1–TIMM13 complex; instead the Tim9p–Tim10p complex only showed specificity for the carriers embedded in the membrane. This study shows that both the DDP1–TIMM13 and Tim9p–Tim10p complexes facilitate the biogenesis of the AGCs, presumably by binding to multiple locations in the substrate. Because of the shared expression pattern with the DDP1–TIMM13 complex, the DDP1–TIMM13 interaction and the decreased NADH levels of the Q34X lymphoblast cell line, the AGC biogenesis requires the DDP1–TIMM13 complex. Future studies will be aimed at determining the specific motifs to which the small Tim proteins bind.
General defects in OXPHOS have not been associated with MTS ( 5 , 6 ). However, our studies suggest that a specific defect might be localized to the large neurons in the cerebellum and basal ganglia, which are metabolically active and are enriched in expression of DDP1, TIMM13 and aralar1. A molecular basis for MTS may lie in decreased NADH levels in the mitochondrion caused by a decrease in the aspartate–malate NADH shuttle and a decrease in Tim23 biogenesis (leading to defects in matrix-targeted import). Calcium signaling is particularly important in the CNS ( 47 ) including the optic nerve and hair cells of the ear. Expression of the AGCs in these tissues has the advantage of coordinating increases in cytosolic [Ca 2+ ] with transfer of reducing equivalents from the cytosol to the mitochondrion, bypassing the requirement for Ca 2+ transport into the mitochondrion ( 25 , 48 ). The crucial role of calcium in the CNS suggests that this mechanism may be an important one for mitochondrial energy production and may be important in contributing to the pathogenesis of MTS. Curiously, mutations in citrin have been linked to citrullinemia type II ( 37 ), so citrullinemia type II might be associated with MTS. However, a citrin knock-out mouse model failed to develop citrullinemia type II ( 49 ), suggesting that additional defects contribute to the disease. These studies lay the foundation for specific investigations aimed at understanding the molecular basis of MTS in patients and animals models and should provide additional insights into the complex functions of mitochondria in neurons.
MATERIALS AND METHODS
Plasmids, strains and lymphoblast cell lines
A shortened version of mouse citrin (citrin 267–676 ) was purchased from Open Biosystems, Huntsville, AL (catalog number MMM1013). This protein contains only 409 amino acids and lacks the N-terminus. Agc1 508–902 was prepared by PCR using genomic DNA from the WT GA74 yeast strain with primers 5′-ACGCGTCGACATGGAATTGCAAAAAAATCAAAATG-3′ and 5′-CCGGAATTCTCACCCGTTAATGCTTCTTAG-3′) and cloned into the transcription vector pSP64. Aralar1 189–680 was prepared by RT–PCR (Stratagene) using primers 5′-TCCCCGCGGATGGTCACCATCCGGTCCCA-3′ and 5′-CGGGATCCTCATTGGGCTGCTGCCGCC-3′. AAC, Su9-DHFR and TIM23 transcription constructs have been described previously ( 16 , 21 ).
Standard conditions were used for growth, manipulation and transformation of yeast ( 50 , 51 ). Growth studies with the Δtim8Δtim13 and parental strain GA74 were initiated with mid-log phase pre-cultures grown in YPD ( 5 ). Cultures were diluted to an OD 600 of 0.1 and by a factor of 3 and spotted onto synthetic minimal medium plates (0.67% yeast nitrogen base, 0.12% ammonium sulfate, 0.1% KH 2 PO 4 , pH 4.5, amino acid supplements) containing different carbon sources (3% Na acetate, 2% ETOH or 2% dextrose). Plates were incubated for 3 days at 30°C.
The lymphocytes were obtained from MTS patient and a non-affected control subject. The MTS patient contains the 100C→T mutation in ddp1 , resulting in a Q34X mutation in the DDP1 protein. Individuals who cannot be identified provided informed consent prior to their inclusion in this study. Cells were immortalized using standard procedures with Epstein–Barr virus supernatant transformation ( 43 ). Immortalized cells were cultured at 37°C in 5% CO 2 on RMPI 1640 medium supplemented with 2 m ML -glutamine, 10% fetal calf serum and antibiotics (100 U/ml penicillin, 100 µg/ml streptomycin).
Mitochondrial import and co-immunoprecipitation studies
Proteins were synthesized in a rabbit reticulocyte lysate in the presence of [ 35 S]methionine after in vitro transcription of the corresponding gene (AAC, citrin, Agc1p and Tim23p) ( 52 ). The radiolabeled precursor was incubated with isolated yeast mitochondria at 25°C in import buffer [1 mg/ml bovine serum albumin, 0.6 M sorbitol, 150 m M KCl, 10 m M MgCl 2 , 2.5 m M EDTA, 2 m M ATP, 2 m M NADH, 20 m M K + -HEPES (pH 7.4)]. Where indicated, the potential across the mitochondrial inner membrane was dissipated with 1 µ M valinomycin. Non-imported radiolabeled proteins were removed by treatment with 20 µg/ml of trypsin for 15–30 min on ice; then trypsin was inhibited with 200 µg/ml of soybean trypsin inhibitor.
The translocation intermediates of AAC, citrin, Agc1p and Tim23p were cross-linked to adjacent proteins with 0.1 m Mm -maleimidobenzoyl- N -hydroxysuccinimide ester (MBS). After import, protease was omitted and mitochondria were washed, resuspended at 1 mg/ml in import buffer and incubated with the cross-linker on ice for 30 min followed by quenching with 100 m M Tris–HCl (pH 8.0). For immunoprecipitation, solubilized mitochondria were incubated in co-immunoprecipitation buffer [20 m M HEPES–KOH (pH 7.4), 0.2 M sucrose, 50 m M NaCl, 1 m M PMSF] with the corresponding polyclonal antibodies coupled to protein A–Sepharose.
Imports into lymphoblast mitochondria were performed in semi-permeabilized cells ( 53 ), and 0.5×10 6 cells/import reaction were resuspended in import buffer (250 m M sucrose, 10 m M KCl, 1 m M EGTA, 10 m M HEPES–KOH, pH 7.2, 0.1 m M GTP, 1 m M DDT, 10 m M succinate, 2 m M ATP, 1.5 m M creatine phosphate and 0.015 mg/ml creatine phosphokinase) and permeablized with digitonin (0.4 µg/ml) for 2 min at 30°C. The precursor was then added and imports were performed as in yeast mitochondria.
Rat liver mitochondria were purified as described ( 54 ). Import was performed as described earlier in import buffer (220 m M mannitol, 70 m M sucrose, 10 m M HEPES–KOH, 150 m M KCl, 1 m M MgCl 2 and 1 m M EGTA, pH 7.6), supplemented with 1 mg/ml bovine serum albumin, 2 m M ATP, 0.1 m M GTP, 1.5 m M creatine phosphate, 15 µg/ml creatine kinase, 1 m M dithiothreitol and 10 m M succinate ( 55 ). The translocation intermediates were cross-linked to adjacent proteins with 0.1 m M MBS.
For analysis of DDP1/Timm8a in lymphoblast cultures, cells from a control and the patient cell line Q34X were grown in RPMI medium. Cells were lysed in 0.2 M sucrose, 100 m M NaCl, 1 m M EDTA, 1 m M PMSF, 20 m M HEPES, pH 7.4 and 0.16% n -dodecyl maltoside. Immunoprecipitation was done as previously described ( 5 ). Bound proteins were separated by SDS–PAGE and analyzed by fluorography.
To measure the NADH reduction by mitochondria in lymphocyte cells, an XTT-based kit (Sigma) was used. The assay was performed as described previously ( 25 ). In total, 10 6 cells/reaction were switched to glucose free media (HCSS: 5.4 m M KCl, 0.12 m M NaCl, 0.8 m M MgCl 2 , 1 m M CaCl 2 and 25 m M HEPES, pH 7.4). Cells were incubated for 4 h; then washed with Ca 2+ -free HCSS media. After a pre-incubation in Ca 2+ -free HCSS media containing 20 µ M digitonin, the XTT reagent was added, either alone (basal condition) or together with 1 m M glutamate, 5 m M malate and 10 m M lactate or 2 m M succinate and incubated for 30 min at 37°C in the dark. The absorbance was measured at OD 540 . For immunoblot studies, cells were incubated for 12 h in glucose free media as for the XTT reduction assays and then switched to basal media supplemented as in the XTT reduction assay. Cells were collected after 8 h and equal numbers of cells were assayed for aralar1 (monoclonal, BD Biosciences) and TOMM40 (polyclonal) by immunoblot analysis. Aralar1 and TOMM40 levels were quantitated by scanning laser densitometry. Aralar1 levels were normalized to TOMM40, with 100% being set as the amount of aralar1 under basal conditions in control lymphoblasts.
Total protein extraction and tissue distribution studies
Fresh tissue samples were homogenized in 5 volumes of 50 m M Tris–HCl (pH 7.5), 150 m M NaCl, 1 m M EDTA, 1 m M phenylmethylsulfonyl fluoride (PMSF), 1 m M benzamidine, 10 µg/ml pepstatin A, 2 µg/ml apoprotinin and 50 µg/ml leupeptin. The insoluble material was pelleted by centrifugation at 3000 g for 1 min. The protein concentration was assayed and equal amounts of protein samples were subjected to western blot analysis ( 56 ).
The euthanized animals were perfused through the cardiac ventricle with fixative solution containing 4% paraformaldehyde in 0.1 M phosphate-buffered saline (PBS, pH 7.4). The brain was then removed, post-fixed overnight and cryoprotected by immersion in 30% sucrose (w/v) in 0.1 M PBS at 4°C. After freezing with dry ice, the brain was sliced in 10 µm thick coronal sections on a Reichert-Jung cryostat. Sections were placed on ‘superfrost plus’ slides (Fisher).
Immunohistochemical staining techniques were published previously ( 57 ). Sections were first quenched with 0.3% hydrogen peroxide in 1× PBS. Then sections were incubated for 30 min in 0.1 M PBS containing 6% normal donkey serum and 0.3% Tween to block unspecific binding and then incubated overnight with the primary antibodies: rabbit anti-DDP1 (dilution 1 : 200), rabbit anti-TIMM13 antibody (dilution 1 : 100), mouse anti-polyglutamylated tubulin (dilution 1 : 500, Sigma), mouse anti-glial fibrillary acidic protein (dilution 1 : 400, Sigma), rabbit anti-TIMM17a (dilution 1 : 500, Santa Cruz Biochemical), and rabbit anti-Hsp60 (dilution 1 : 500). The sections were rinsed four times in PBS containing 0.3% Tween. Then biotinylated donkey anti-rabbit IgG (dilution 1 : 300, Vector laboratories, Burlingame, CA, USA) or anti-mouse IgG (1 : 300, Vector laboratories) was applied for 1 h; followed by incubation in the avidin–biotin-peroxidase complex (dilution 1 : 200, Vector Laboratories) for 90 min. The color reaction was carried out by incubating the sections in 50 m M Tris buffer (pH 7.5) containing 0.02% 3,3′-diaminobenzidine (DAB) and 0.015% H 2 O 2 for 15–30 min. After the DAB reaction, the sections were rinsed several times, dehydrated and coverslips were fixed with Permount. Controls included omission of the primary antibody. All of these control sections were devoid of immunoreactivity.
For immunofluorescent techniques, the sections were incubated with a rhodamine-conjugated secondary antibody (1 : 100) for 3 h. The sections were washed and mounted in Vectashield (Vector Laboratory). Sections were viewed and photomicrographs were taken with a Nikon fluorescent microscope equipped with a spot camera.
We are grateful to the MTS family for providing materials to generate the lymphoblast cell line and Barbara Switzer, University of Nebraska Medical Center and Dr William J. Kimberling, Boys Town Research Institute, Omaha, NE, USA for development and contribution of the Q34X cell line. We thank Dr El-Khansa Kaicer for technical assistance, and Dr S. Sampogna, Dr M.-F. Chesselet and Dr M. Hickey for assistance with immunocytochemical studies. This work was supported by funds to C.M.K. from the Beckman Foundation (02035457), the NIH NIDCD (R21DC006663-01), the Muscular Dystrophy Association (03018774), and the Deafness Research Foundation and funds to L.T. from the Oticon Foundation. C.M.K. is a Beckman Scholar. P.J.H. is supported by the Muscular Dystrophy Association (02076534).
Present address: Genome Technology Branch, National Human Genome Research Institute, National Institutes of Health, 50 South Drive, MSC 8004, Bethesda, MD 20892-8004, USA.
- central nervous system
- origin of life
- cell line
- tissue membrane
- mitochondrial proteins
- nicotinamide adenine dinucleotide (nad)
- pathologic processes
- hearing impairment
- mohr-tranebjaerg syndrome
- mitochondrial matrix