Phosphorylation of Rat Liver Mitochondrial Glycerol-3-phosphate Acyltransferase by Casein Kinase 2*

We have previously shown rat liver mitochondrial glyc-erol-3-phosphate acyltransferase (mtGAT), which catalyzes the first step in de novo glycerolipid biosynthesis, is stimulated by casein kinase 2 (CK2) and that a phosphorylated protein of (cid:1) 85 kDa is present in CK2-treated mitochondria. In this paper, we have identified the 32 P-la-beled 85-kDa protein as mtGAT. We have also investigated whether the phosphorylation of mtGAT is because of CK2. Mitochondria were treated with CK2 and [ (cid:1) - 32 P]GTP as the phosphate donor. Autoradiography, Western blot, and immunoprecipitation results showed mtGAT was phosphorylated by CK2. Next, we incubated mitochondria with CK2 and either ATP or GTP, in the presence of heparin, a known inhibitor of CK2. Heparin inhibited CK2-induced stimulation of mtGAT activity; this inhibition re-sulted in decreased 32 P-labeling of mtGAT. Additionally, mitochondria were treated with CK2 and [ (cid:1) - 32 P]ATP in the presence of staurosporine (a serine/threonine protein kinase inhibitor), genistein (a tyrosine kinase inhibitor), and 5,6-dichloro-1- (cid:2) - D -ribofuranosylbenzimidazole (DRB, a CK2 inhibitor). Only DRB,

Although mitochondria are involved in many cellular events such as ATP generation, ␤-oxidation, lipid biosynthesis, and apoptosis, they also comprise an intricate role in intracellular signaling via phosphorylation/dephosphorylation. The mitochondrial outer membrane (MOM) 1 occupies a principal inter-face in this signaling, both in possessing kinase activity (1)(2)(3)(4)(5) and protein substrates (6).
One enzyme that resides in the MOM is glycerol-3-phosphate acyltransferase (mtGAT), which catalyzes the first step in de novo glycerolipid biosynthesis via the acylation of glycerol-3phosphate at the sn-1 position to produce lysophosphatidic acid. Lysophosphatidic acid, synthesized at the outer surface of the MOM, can be exported to the endoplasmic reticulum and used for the synthesis of complex phospholipids and triacylglycerol (7,8) or be converted to phosphatidic acid by monoacylglycerol-3-phosphate acyltransferase (AGPAT). The phosphatidic acid is transported to the mitochondrial inner membrane for cardiolipin synthesis (9). Unlike the microsomal isoform, which resides in the endoplasmic reticulum, mtGAT is insensitive to sulfhydryl reagents, such as N-ethylmaleimide, and exhibits substrate specificity for saturated acyl-CoA as the acyl donor (10). The enzyme has been purified (11), and the cDNA has been cloned (12,13). The location of mtGAT makes it a key factor in diverting fatty acids from undergoing ␤-oxidation inside the mitochondria to fatty acid utilization for biosynthetic processes (14,15). Its substrate specificity contributes to the asymmetric distribution of fatty acids in phospholipids and can control the quality of phospholipids necessary for the formation of biomembranes (16,17).
Indirect evidence has been compiled for the posttranslational regulation of this MOM enzyme, mtGAT. First, discordancy exists between mtGAT mRNA levels, protein levels, and enzymatic activity. For example, liver and adipose tissue express high mRNA levels and low protein levels yet have high mtGAT activity; whereas heart and adrenal glands have low mRNA levels and high protein levels but low mtGAT activity (18). Secondly, 5Ј-AMP-activated protein kinase (AMPK) inhibits mtGAT activity. The treatment of cultured rat hepatocytes with 5-amino-4-imidazolecarboxamide riboside, an activator of AMPK, inhibited mtGAT activity (29 -43%). Recombinant AMPK1 and AMPK2 inhibited mtGAT activity 30 -50% in isolated rat liver mitochondria (19). Thirdly, we have previously found, by ProSite analysis, that the mtGAT amino acid sequence contains several putative consensus sequences for casein kinase 2 (CK2) phosphorylation. We have shown that CK2 stimulates mtGAT activity (40 -68%), this stimulation is reversed by protein phosphatase-treatment, and that an ϳ85-kDa protein, the molecular mass of purified rat liver mtGAT, is 32 P-labeled in CK2-treated mitochondria (20). Most recently, West et al. (21) showed that CK2 increased GAT activity in mitochondria isolated from unstimulated Jurkat cells and further increased GAT activity in mitochondria isolated from stimulated Jurkat cells.
Presently, we have investigated whether the stimulation of mtGAT by CK2 was because of the phosphorylation of the acyltransferase itself or the phosphorylation of another protein such as a stimulant protein associated with the MOM (11). Furthermore, we also wanted to determine whether this phosphorylation is result of CK2. Here we have identified the 32 P-labeled 85-kDa protein as mtGAT by Western blot and immunoprecipitation. Preparation of Mitochondria-Rat liver mitochondria were isolated from 175-200-g male Sprague-Dawley rats by differential centrifugation as described previously (10). Protein concentration was maintained at 18 mg/ml as determined by the Bradford assay (22). The purity of the mitochondria was assessed by performing the GAT assay in the presence of 4 mM N-ethylmaleimide; microsomal contamination was always less then 5%.

Materials-Male
Preparation of ATP and GTP-Mg 2ϩ -ATP and Mg 2ϩ -GTP stocks were prepared according to Ref. 23, with a slight modification allowing for the addition of magnesium. ATP or GTP (20 mM) was prepared and diluted 1:1 with 20 mM MgCl 2 to obtain 10 mM Mg 2ϩ -ATP or 10 mM Mg 2ϩ -GTP stocks which were stored at Ϫ20°C.
Preparation of Heparin Stock-A heparin stock was prepared by dissolving heparin in deionized water to a final concentration of 50 mg/ml (24). Before experiments, the heparin stock was further diluted with deionized water to a final concentration of 0.5 mg/ml, which was added to the reaction mixtures to obtain the indicated heparin concentrations.
Isolation of MOM-Rat liver MOM was isolated by the osmotic swelling method according to Ref. 25. The crude MOM pellet was resuspended in 20 mM phosphate buffer, pH 7.2, layered on top of a discontinuous sucrose density gradient consisting of 3.6 ml each of 25.2, 37.7, and 51.3% sucrose in 20 mM phosphate buffer, pH 7.2, and centrifuged at 35,000 rpm for 3 h in a SW-41Ti rotor. The 25.2/37.7% interface was collected using a 3-ml syringe fitted with a 22.5-gauge needle, diluted 4 times with deionized water, centrifuged at 35,000 ϫ g for 60 min in a JA25.50 rotor, suspended in S.E. (0.3 M sucrose, 1 mM EDTA, pH 7.5), and stored at Ϫ80°C. The cytochrome c oxidase assay was performed on the isolated MOM, mitochondrial inner membrane, mitoplasts, and whole mitochondria to determine the purity of the mitochondrial outer membrane preparation. Cytochrome c oxidase activity, as determined using kit purchased from Sigma Chemicals, indicated less than 15% contamination of MOM with inner membrane (data not shown).
The GAT Assay-The GAT assay was performed as described previously by measuring the amount of sn-[2-3 H]-glycerol-3-phosphate incorporated into 1-butanol-extractable phospholipids, lysophosphatidic acid, and phosphatidic acid (26). Asolectin was omitted from the incubation medium. Each assay was initiated by the addition of mitochondria (160 g).
Incubation of Mitochondria with CK2-Mitochondria were incubated with ATP and CK2 as described previously (20). For CK2 inhibition assays, mitochondria were incubated in the presence of 0, 2, 10, 15, 25, and 50 g/ml heparin (27)(28)(29). Alternatively, mitochondria were incubated with rat liver cytosol (53 or 106 g) instead of CK2. The same conditions were followed for assays in which GTP was substituted for ATP.
Phosphorylation of mtGAT with [␥-32 P]ATP, [␥-32 P]GTP, and CK2-Mitochondria (200 g) were incubated with [␥-32 P]ATP (5Ci, 200 M) and CK2 as described previously (20). The samples were washed three times in ice-cold phosphate-buffered saline and centrifuged at 12,000 ϫ g for 5 min (Eppendorf 5410). The mitochondrial pellet was suspended in 1ϫ Laemmli sample buffer and boiled for 5 min; 150 g was loaded onto a 1.5-mm mini SDS-polyacrylamide gel (7.5%), run at a constant current of 30 mA for 1 h at 4°C, transferred to Sequiblot PVDF membrane, and immunoblotted as described below. Another gel was stained with Simply Blue according to the manufacturer's protocol and dried for 2 h in a Bio-Rad gel dryer (model 583). After immunoblotting, the membrane was rinsed 2 times with deionized water and exposed at Ϫ80°C to Kodak BioMax MS film in a cassette with an intensifying screen for 48 h. Kodak GBX developer and fixer solutions were used for developing both immunoblot and autoradiography films. The same conditions were used when [␥-32 P]GTP was substituted for [␥-32 P]ATP.
Immunodetection of mtGAT Using IM1GAT Antibody-After SDS-PAGE, proteins were transferred to a Sequiblot PVDF membrane at a constant voltage of 80 V for 90 min at 4°C. Immunodetection was performed using the protein detector Western blot kit LumiGLO system, according to manufacturer's protocol, with the following modifications: the PVDF membrane was rocked overnight at 4°C with primary antibody (1:1000 dilution) and a 1:2000 dilution of secondary antirabbit antibody was used. The primary antibody used was IM1GAT, an affinity-purified rabbit polyclonal anti-peptide antibody synthesized against N-terminal amino acid sequence CGHYNGEQLGKPKKNES by GeneMed Synthesis, Inc. The membrane was placed in a cassette with intensifying screens and exposed to Kodak BioMax MS film.
Immunoprecipitation of 32 P-Labeled mtGAT with IM1GAT Antibody-Whole mitochondria (700 g) or MOM (200 g) were treated with CK2 and [␥-32 P]ATP as described above. The mitochondria were washed with ice-cold phosphate-buffered saline and centrifuged at top speed for 5 min at 4°C. Mitochondria and MOM were solubilized in 1 ml of ice-cold radioimmune precipitation assay buffer (1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 1ϫ phosphate-buffered saline, pH 7.2), incubated on ice for 15 min, rotated overnight at 4°C with IM1GAT antibody (10 g). Protein A/G plus-agarose (20 l) was added, and the mixtures were rotated at 4°C for 60 min. The agarose pellet was washed 2 times with ice-cold radioimmune precipitation assay buffer (1 ml), suspended in 1.5ϫ Laemmli sample buffer, boiled for 5 min, centrifuged for 5 min, and the supernatant was loaded onto a 1.5-mm thick SDS-PAGE (7.5%) minigel. Electrophoresis and immunoblotting were performed as described above. For heparin experiments, mitochondria were treated as above in the presence of 0, 2, 15, or 50 g/ml heparin. The above experiments were repeated with [␥- 32  buffer, and cytosol (53 g) in a 50-l reaction volume. In addition, cytosol (150 g) was incubated on ice for 60 min with [␥-32 P]GTP (5 Ci, 200 M) and CK2 buffer in a 50-l reaction volume. The reaction mixture was transferred to a Microcon-10 microconcentrator and centrifuged at 12,500 rpm at 4°C for 20 min. Ice-cold phosphate-buffered saline was added to the Microcon-10 column and centrifuged for an additional 2 h. The Microcon-10 column was inverted and pulse-centrifuged at 13,000 ϫ g for 20 s to collect the retentate. The MOM pellets were suspended in 1ϫ Laemmli sample buffer, and the cytosol was suspended in 1.5ϫ Laemmli sample buffer (1:2), boiled for 5 min, and loaded onto a 1.5-mm thick mini SDS-polyacrylamide gel (7.5%).
NetPhos 2.0 Analysis-NetPhos (version 2.0) is an artificial neural network that predicts sites for serine, threonine, or tyrosine phosphorylation in eukaryotic proteins (30) available on the internet at www.expasy.org (31). The amino acid sequence deduced from rat liver mtGAT cDNA was analyzed for potential phosphorylation sites containing serine or threonine residues irrespective of consensus sequences for kinase phosphorylation sites. Results are returned indicating the position of the serine or threonine, the ϩ4 and Ϫ4 flanking amino acids, the score (0.000 -1.000), and the predicted serines or threonines for phosphorylation. Threshold value is automatically set at 0.500.
Statistical Analysis-GraphPad Prism 3.0 software (San Diego, CA) was used to perform one way analysis of variance with Tukey's multiple comparison test and unpaired, two-tailed t test. Error bars (Figs. 3 and 8) indicate standard error.
Densitometry Analysis-Autoradiogram and Western blot x-ray films were scanned into the computer, and the signals were analyzed using the UN-SCAN-IT gel automated digitizing system (version 5.1) from the Silk Scientific, Inc. (Orem, UT).

NetPhos 2.0 Analysis of Rat Liver mtGAT Amino Acid
Sequence-Rat liver mtGAT has been shown to contain 14 putative consensus sequences for CK2 phosphorylation (20). We further tested, by using another bioinformatics program, Net-Phos 2.0 (30), these consensus sequences with respect to all potential S/T residues that can be phosphorylated within the mtGAT amino acid sequence. Eleven of the 14 putative CK2 consensus sequences found previously were within the Net-Phos 2.0-predicted S/T phosphorylation sites ( Table I). Each of the NetPhos 2.0-predicted S/T phosphorylation sites was subjected to a protein BLAST search for short exact matches. Only one site showed homology with a protein kinase C and CK2 substrate protein 3 in Mus musculus (accession number Q99JB8) and a similar protein in Rattus norvegicus (accession number XP_230279). The result of the BLAST search suggests that the NetPhos 2.0-predicted site (ITHTSRKDE), within rat mtGAT, is a likely target for CK2 phosphorylation, because it shares homology to another CK2 substrate, CK2 substrate protein 3.
Specificity of the Polyclonal IM1GAT Anti-peptide Antibody-Before the hypothesis that the 85-kDa 32 P-labeled protein is mtGAT could be tested experimentally, it was necessary to establish that the polyclonal anti-peptide IM1GAT antibody is suitable for Western blot analysis of mtGAT. The IM1GAT antibody was raised against an oligopeptide corresponding to amino acids 362-377 of the N-terminal region of mtGAT. IM1GAT IgG was shown previously (32) to immunoreact with GAT in mitochondria isolated from rat liver and COS-1 cells expressing the recombinant protein. We wanted to determine the suitability of IM1GAT for Western analysis. Rat liver mitochondria (150 g), MOM (150 g), and cytosol (150 g) fractions were analyzed for the presence of mtGAT by Western blot with the IM1GAT antibody. An immunoreactive band of ϳ85 kDa was detected in the mitochondrial fraction (Fig. 1, lane 4). A band of ϳ10 times greater intensity appeared in the MOM fraction (Fig. 1, lane 5), and, as expected, no band was detected in the cytosolic fraction (lane 6). Lanes 1-3 of Fig. 1 represent the Simply Blue-stained gel for the mitochondrial, MOM, and cytosolic sample, respectively. When Yet et al. (33) immunoblotted mouse liver mitochondria with an antibody raised against a trpE-p90 fusion protein, two bands close in molecular weight were detected. They attributed the lower band to a proteolytic degradation product resulting from storage of the mitochondria. Also, the lower bands seen in isolated MOM fractions could arise from the isolation procedure. Isolation of MOM involves osmotic swelling and rupture. During this process proteases inside the mitochondria, not normally in contact with cleavage sites of mtGAT, may interact with mtGAT resulting in its cleavage. For example, an ATP-dependent, vanadate-sensitive endoprotease is present in rat liver mitochondria and localized primarily to the matrix followed by the intermembrane space (34). Also, Smac/DIABLO, a protein redundant to mammalian serine protease Omi/HtrA2, is located in the intermembrane space and abundantly expressed in the heart, liver, kidney, and testes (35). Another possibility for the lower molecular weight bands in the MOM is that the IM1GAT antibody nonspecifically recognizes AGPAT, the acyltransferase that catalyzes the formation of phosphatidic acid from lysophosphatidic acid. GAT and AGPAT isoforms share a high degree of amino acid similarity; however, AGPAT isoforms have low molecular masses (30 -40 kDa) (36). The IM1GAT antibody is polyclonal and may recognize AGPAT with less affinity and irreproducibility. This would explain why in some circumstances a lower molecular mass band around 40 kDa is present. As with GAT, AGPAT is present in low abundance in the MOM; therefore, the likelihood that AGPAT would be detected in isolated MOM versus whole mitochondria is enhanced. However, these results indicate that the IM1GAT antibody recognizes mtGAT with relatively high specificity and is suitable for Western analysis.
Identification of the 85-kDa 32 P-Labeled Band as mtGAT-We further investigated whether the 85-kDa 32 P-labeled band previously observed (20) is mtGAT. Although the 32 P-labeled 85-kDa band appeared only in the CK2-treated mitochondria ( Fig. 2A, left panel), immunoreactive bands of equivalent intensities were present in both untreated and CK2-treated mitochondria ( Fig. 2A, right panel). We correlated Western blot data with immunoprecipitation to determine whether the 32 Plabeled 85-kDa protein is mtGAT. Because mtGAT protein is not abundantly expressed in the mitochondria, we also immunoprecipitated GAT from purified MOM to increase the amount of mtGAT in the immunoprecipitation reaction. Mitochondria and MOM were incubated with [␥-32 P]ATP and kinase buffer in the presence and absence of CK2. The 32 P-labeled 85-kDa band was present in immunoprecipitates from whole mitochondria treated with CK2 (Fig. 2B, MITO, CK2) and MOM treated with CK2 (MOM, CK2) but not untreated mitochondria (MITO, AB)  1. Immunodetection of GAT from isolated rat liver mitochondria with IM1GAT antibody. Rat liver mitochondria (ϳ150 g), MOM (ϳ150 g), and cytosol (ϳ150 g) were subjected to 7.5% SDS-PAGE and Western blotted with IM1GAT antibody (see "Experimental Procedures"). Lanes 1-3 are of the gel stained with Simply Blue and lanes 4 -6 are of the immunoblotted PVDF membrane; 1 and 4, mitochondria; 2 and 5, MOM; 3 and 6, cytosol; arrows, kaleidoscope molecular weight marker. and untreated MOM (MOM, AB). The two minor 32 P-labeled bands, because they did not appear in the immunoprecipitates from whole mitochondria, most likely arose from the MOM isolation procedure (refer to previous section).
Phosphorylation of mtGAT by CK2-Several approaches were taken to determine whether the phosphorylation of mt-GAT is because of direct phosphorylation by CK2. CK2 is one of the few, if not the only, protein kinases that can utilize GTP as efficiently as ATP, as the phosphate donor (37). Therefore, we used GTP in lieu of ATP as the phosphate donor and incubated mitochondria in the absence or presence of CK2. CK2 stimulated mtGAT activity ϳ1.92-fold over mitochondria alone and 1.42-fold over the stimulation by GTP and the CK2 buffer (Fig.  3). This stimulation by CK2 when GTP was used was comparable to that when ATP was used.
The use of GTP as the phosphate donor in the enzyme activity study does not provide direct evidence that CK2 itself is phosphorylating mtGAT. To address this issue, we repeated the above experiments with [␥-32 P]GTP as the phosphate donor. The 32 P-labeled 85-kDa band was present only in mitochondria treated with CK2 (Fig. 4A, left panel), whereas mt-GAT was present in both untreated and CK2-treated mitochondria in equivalent quantities (right panel). To further support these data, mtGAT was immunoprecipitated from both mitochondria and MOM after treating with [␥-32 P]GTP and CK2 buffer in the presence or absence of CK2. As in the immunoprecipitation experiments with [␥-32 P]ATP, 32 P-labeled mtGAT was immunoprecipitated from mitochondria and MOM treated with CK2 (Fig. 4B).
The CK2 inhibitor, heparin, was used to further show the phosphorylation of mtGAT by CK2. Mitochondria were incubated with various concentrations of heparin (2, 10, 15, 25, and 50 g/ml) in the presence or absence of CK2, along with ATP.
Heparin reduced CK2 stimulation of mtGAT activity in a dosedependent manner (Fig. 5A) and caused a dose-dependent decrease in the amount of 32 P incorporated into mtGAT by CK2 (Fig. 5B, even numbered lanes). When the same experiments were performed using GTP, heparin reduced the stimulation of mtGAT activity by CK2 (Fig. 6A) and decreased the amount of 32 P incorporated into mtGAT by CK2 (Fig. 6B, even numbered lanes) in a dose-dependent manner.
To further substantiate these heparin data, we incubated mitochondria with various protein kinase inhibitors, including another CK2 inhibitor. We incubated mitochondria with [␥- 32  Western blot signal for the amount of mtGAT present using UN-SCAN-IT gel automated digitizing system (version 5.1) (see Fig. 7). DRB inhibited 32 P incorporation into mtGAT by ϳ38% compared with the control for the vector (Me 2 SO). Genistein failed to inhibit 32 P incorporation into mtGAT, and staurosporine slightly inhibited (ϳ18%) 32 P incorporation into mtGAT (Fig. 7).
Phosphorylation of mtGAT with Rat Liver Cytosol-To establish physiological relevance of the phosphorylation of mtGAT by CK2, we incubated mitochondria with increasing amounts of rat liver cytosol in the presence of either ATP or GTP and kinase buffer. Cytosol stimulated mtGAT activity (Fig. 8), as expected, because liver fatty acid-binding protein is present in the cytosol and known to stimulate mtGAT activity (38,39). However, when mitochondria were incubated with CK2 buffer, cytosol (53 g), and ATP or GTP, mtGAT activity increased ϳ1.56-and 1.24-fold, respectively, over the increase because of cytosol (53 g) alone. When the amount of cytosol was increased to 106 g, stimulation of mtGAT activity by ATP was about the same, but the stimulation of mtGAT activity due to GTP increased to ϳ1.33-fold over that of cytosol (106 g) alone (Fig. 8). Stimulation of mtGAT activity above that of the cytosol by both ATP and GTP suggests that a kinase present in the cytosol can phosphorylate mtGAT. The ability of GTP to stimulate mtGAT activity suggests that CK2 is one of those kinases. To test this possibility, we incubated mitochondria and MOM with [␥-32 P]GTP, CK2 buffer, and cytosol or CK2. A 32 P-labeled 85-kDa band was detected in MOM treated with [␥-32 P]GTP and exogenous CK2. Moreover, a band was also detected in MOM treated with cytosol ( Fig. 9, left panel). No 32 P-labeled 85-kDa band was detected in mitochondria and MOM samples where neither CK2 nor cytosol were added (Fig. 9, left panel). This result corroborated our observation that an ϳ85-kDa protein is 32 P-labeled in the MOM by CK2 when [␥-32 P]ATP is used as the phosphate donor (data not shown). Western blot analysis indicated that the 32 P-labeled 85-kDa band corre- sponded to mtGAT (Fig. 9, right panel). Moreover, no immunoreactive bands were present in the cytosol incubated with [␥-32 P]GTP and CK2 buffer (Fig. 9, right panel), confirming that the 32 P-labeled 85-kDa protein is mtGAT. These data suggest that mtGAT is phosphorylated by CK2-like activity present in rat liver cytosol.

DISCUSSION
Western blot and immunoprecipitation results indicate that the 32 P-labeled 85-kDa protein corresponds to mtGAT (Fig. 2). We further determined if CK2 was involved in the phosphorylation of mtGAT. Because CK2 is known to utilize GTP as efficiently as ATP as the phosphate donor, all experiments were performed with GTP and [␥-32 P]GTP. Mitochondrial GAT activity was significantly stimulated, and the enzyme was phosphorylated by CK2 in the presence of GTP (Figs. 3 and 4). The stimulation and phosphorylation of mtGAT by CK2 was confirmed using the CK2 inhibitor, heparin. Heparin was chosen because it is a known, well characterized, potent, specific, and physiological inhibitor of CK2. Moreover, heparin is a noncompetitive inhibitor with respect to ATP (40 -42). Heparin (2,10,15,25, and 50 g/ml) attenuated both the stimulation of mtGAT activity (Figs. 5A and 6A) and the 32 P labeling of the enzyme (Figs. 5B and 6B) by CK2, in a dose-dependent manner. To corroborate these heparin data, mitochondria were incubated with [␥-32 P]ATP and CK2 in the presence of protein kinase inhibitors staurosporine (a general S/T protein kinase inhibitor), genistein (a tyrosine kinase inhibitor), and DRB (a CK2 inhibitor). Only DRB greatly reduced (ϳ40%) the amount of 32 P incorporated into mtGAT (Fig. 7). Collectively, these data indicate that CK2 catalyzes the phosphorylation of mtGAT.
If the phosphorylation of mtGAT by exogenous CK2 has physiological relevance, then rat liver cytosol should possess CK2 activity able to stimulate and phosphorylate mtGAT. Addition of either ATP or GTP significantly stimulated mtGAT activity over that by cytosol alone (Fig. 8). The elevated enzyme activity in the presence of ATP or GTP is probably because of phosphorylation. CK2 is present in rat liver cytosol and appears to be under physiological control (44). An 85-kDa 32 Plabeled band corresponding to mtGAT was detected when isolated MOM was incubated with [␥-32 P]GTP and cytosol (Fig. 9). A band with an intensity obtained with exogenous CK2 would not be expected. Physiological concentrations of CK2 would be far less than the concentration of exogenously added CK2. The lack of the 32 P-labeled 85-kDa band in cytosol-treated mitochondria can be explained by the low protein levels of GAT in rat liver mitochondria. Furthermore, the use of [␥-32 P]GTP as the phosphate donor suggests the kinase activity in the cytosol that phosphorylates mtGAT is CK2. The significance of CK2catalzyed phosphorylation of mtGAT is described below.
The intertransmembrane loop region of mtGAT appears to be important for catalytic activity of the enzyme. Balija et al. (32), using immunoprecipitation and immunoreactivity assays with several different mtGAT anti-sera raised against synthetic peptides, found that the N and C termini of mtGAT are sequestered inside the mitochondria, and the loop region is facing the cytosol. In contrast, Coleman and co-workers (45), using epitope-tagged mtGAT and fluorescence microscopy, found the opposite, that the loop region is sequestered inside the mitochondria, and the N-and C-terminal regions are facing the cytosol. Although the two groups differ regarding topography, data from both laboratories suggest that the loop region is important for catalytic activity. Interestingly, several CK2 sites fall within this region. It is worth noting that the one S/T phosphorylation site with homology to protein kinase C and CK2 substrate protein 3 (Table I) resides within the intertransmembrane loop region. Thus, the effect of CK2 on mtGAT activity could involve the phosphorylation of the protein within this region. Therefore, it is feasible that CK2, the primary cyclic nucleotide-independent protein kinase in rat liver cytosol, could phosphorylate rat liver mtGAT. This would help explain why rat liver, which has low mtGAT protein levels, has the highest mtGAT activity of all tissues (18).
Mitochondrial GAT also appears to be allosterically modulated by ATP, GTP, and citrate. 2 Additionally, GAT from Escherichia coli is similarly regulated by ATP and GTP (46). We postulate that CK2-catalyzed phosphorylation of mtGAT may enhance the allosteric effect of ATP and citrate on mtGAT activity. The phosphorylation of mtGAT by CK2 could sensitize mtGAT to ATP thereby enhancing the stimulation of mtGAT by ATP. Alternatively, the binding of ATP to mtGAT could make it a better substrate for CK2-cataylzed phosphorylation. In either scenario, phosphorylation and allosteric modulation may synergistically affect mtGAT activity. For example, AMPK is FIG. 8. Effect of cytosol on mtGAT activity in the presence of either ATP or GTP. Mitochondria (200 g) were incubated with ATP, GTP, CK2 buffer, or rat liver cytosol (53 or 106 g) as indicated by ϩ below the graph. The GAT assay was initiated by addition of the incubated-mitochondria (160 g). These data represent the -fold change over mitochondria only (specific activity ϭ 1.41 nmol/min/mg of protein). Error bars indicate S.E. *, significant difference between conditions as determined by an unpaired, two-tailed t test, p Ͻ 0.05. For each set of experiments performed, n ϭ 3. not only allosterically activated by 5Ј-AMP but is also phosphorylated on a key threonine residue (Thr 172 ) in its catalytic subunit by an upstream kinase (AMPKK). The phosphorylation of AMPK on Thr 172 may be involved in AMP binding via affecting the sensitivity of AMPK to AMP either by a direct effect of the negative charge or by a conformational change (47).
Acetyl-CoA carboxylase (ACC) is a rate-limiting enzyme in fatty acid biosynthesis. It catalyzes the synthesis of malonyl-CoA, an intermediate in the synthesis of long-chain fatty acids. ACC and GAT are two enzymes that catalyze the first step in fatty acid and glycerolipid synthesis, respectively. The similarities in regulation between ACC and mtGAT are numerous. First, ACC2 is anchored to the MOM (48), and mtGAT transverses the MOM (49). Secondly, triacylglycerol levels are lowered in both ACC2 Ϫ/Ϫ (50) and mtGAT Ϫ/Ϫ mice (51). Third, allosteric modulators like citrate, ATP, and GTP stimulate ACC and mtGAT activity. Fourth, starvation decreases and refeeding a high carbohydrate diet increases both mtGAT (18,52) and ACC activity (48). Fifth, both ACC (53, 54) and mtGAT activities are stimulated by insulin and epidermal growth factor (55,56). Sixth, both ACC and mtGAT are inhibited as a result of exercise (57). Seventh, AMPK inhibits both ACC and mtGAT activity (58). And like ACC (54), this work shows mt-GAT is phosphorylated by CK2.
These similarities in regulation of ACC and mtGAT may result from a larger purpose within the cell. It is logical that the cell would utilize common regulatory mechanisms to control similar biosynthetic processes. Insulin, a stimulator of lipogenesis, stimulates both mtGAT and ACC. One of the components in the insulin pathway may be CK2. By employing a phosphorylation control mechanism, the cell can feasibly ensure the fatty acids synthesized are used for biosynthesis. Stimulation of both ACC and mtGAT ensures that the fatty acids are diverted from undergoing ␤-oxidation in the mitochondria to synthesis of glycerolipids. A recent report (15) suggests that increased hepatic mtGAT activity contributes to increased lipid biosynthesis and decreased ␤-oxidation of fatty acids.
ACC and GAT are two important enzymes in similar biosynthetic pathways. The cell may have a greater plan for regulating similar anabolic and catabolic pathways. For example, when cellular ATP is high and AMP is low, both the enzymes are stimulated; inversely, when AMP is high and ATP is low, the enzymes are inhibited. This makes much sense, because when cellular ATP is high there is no need to produce energy via ␤-oxidation or glycolysis. Therefore, synthetic processes can utilize the excess ATP. When ATP is low, catabolic processes can proceed at a higher rate to produce more ATP, at the same time anabolic processes can remain inhibited. Increasing fatty acid synthesis (i.e. stimulating ACC) will not ensure that the fatty acids will be diverted toward glycerolipid biosynthesis. Without GAT being similarly regulated, the fatty acids may be used either for biosynthesis or for degradation. It may be that similar modulators (ATP, AMP, and citrate), hormones, and protein kinases (CK2, AMPK, and protein kinase A) are employed to control these processes.
The work indicates that CK2 stimulates mtGAT via the direct involvement of CK2 in the phosphorylation of the acyltransferase. The findings presented here are the first to provide direct evidence that mtGAT is phosphorylated by a protein kinase (phosphorylation of mtGAT by AMPK was shown indirectly). In part, these findings help elucidate the discordancy observed between mtGAT protein levels and enzymatic activity in the rat liver (18). Furthermore, the phosphorylation of mt-GAT by CK2 may serve as a component in the stimulation of the acyltransferase by hormones such as insulin. It would be interesting to see whether CK2 and AMPK coordinately regu-late mtGAT activity (stimulate and inhibit, respectively) in response to various stimuli. Overall, this work contributes to the understanding of both the role of CK2-catalyzed phosphorylation in regulating mitochondrial proteins and in the regulation of lipid metabolism.