Portal de Periódicos da CAPES

A Conserved Seven Amino Acid Stretch Important for Murine Mitochondrial Glycerol-3-phosphate Acyltransferase Activity

1999; Elsevier BV; Volume: 274; Issue: 49 Linguagem: Inglês

10.1074/jbc.274.49.34728

1083-351X

Lori K. Dircks, Jinshan Ke, Hei Sook Sul,

Pancreatic function and diabetes

Glycerol-3-phosphate acyltransferase (GPAT) catalyzes the initial and committed step in glycerolipid biosynthesis. We previously cloned the cDNA sequence to murine mitochondrial GPAT (Yet, S-F., Lee, S., Hahm, Y. T., and Sul, H.S. (1993) Biochemistry 32, 9486–9491). We expressed the protein in insect cells which was targeted to mitochondria, purified, and reconstituted mitochondrial GPAT activity using phospholipids (Yet, S.-F., Moon, Y., and Sul, H. S. (1995) Biochemistry 34, 7303–7310). Deletion of the seven amino acids from mitochondrial GPAT,312IFLEGTR318, which is highly conserved among acyltransferases in glycerolipid biosynthesis, drastically reduced mitochondrial GPAT activity. Treatment of mitochondrial GPAT with arginine-modifying agents, phenylglyoxal and cyclohexanedione, inactivated the enzyme. Two highly conserved arginine residues, Arg-318, in the seven amino stretch, and Arg-278, were identified. Substitution of Arg-318 with either alanine, histidine, or lysine reduced the mitochondrial GPAT activity by over 90%. On the other hand, although substitution of Arg-278 with alanine and histidine decreased mitochondrial GPAT activity by 90%, replacement with lysine reduced activity by only 25%. A substitution of the nonconserved Arg-279 with either alanine, histidine, or lysine did not alter mitochondrial GPAT activity. Moreover, R278K mitochondrial GPAT still showed sensitivity to arginine-modifying agents, as in the case of wild-type mitochondrial GPAT. These results suggest that Arg-318 may be critical for mitochondrial GPAT activity, whereas Arg-278 can be replaced by a basic amino acid. Examination of the other conserved residues in the seven amino acid stretch revealed that Phe-313 and Glu-315 are also important, but conservative substitutions can partially maintain activity; substitution with alanine reduced activity by 83 and 72%, respectively, whereas substituting Phe-313 with tyrosine and Glu-315 with glutamine had even lesser effect. In addition, there was no change in fatty acyl-CoA selectivity. Kinetic analysis of the R318K and R318A mitochondrial GPAT showed an 89 and 95%, respectively, decrease in catalytic efficiency but no major change in substrate binding as indicated by the K mvalues for palmitoyl-CoA and glycerol 3-phosphate. These studies indicate importance of the conserved seven amino acid stretch for mitochondrial GPAT activity and the significance of Arg-318 for catalysis. Glycerol-3-phosphate acyltransferase (GPAT) catalyzes the initial and committed step in glycerolipid biosynthesis. We previously cloned the cDNA sequence to murine mitochondrial GPAT (Yet, S-F., Lee, S., Hahm, Y. T., and Sul, H.S. (1993) Biochemistry 32, 9486–9491). We expressed the protein in insect cells which was targeted to mitochondria, purified, and reconstituted mitochondrial GPAT activity using phospholipids (Yet, S.-F., Moon, Y., and Sul, H. S. (1995) Biochemistry 34, 7303–7310). Deletion of the seven amino acids from mitochondrial GPAT,312IFLEGTR318, which is highly conserved among acyltransferases in glycerolipid biosynthesis, drastically reduced mitochondrial GPAT activity. Treatment of mitochondrial GPAT with arginine-modifying agents, phenylglyoxal and cyclohexanedione, inactivated the enzyme. Two highly conserved arginine residues, Arg-318, in the seven amino stretch, and Arg-278, were identified. Substitution of Arg-318 with either alanine, histidine, or lysine reduced the mitochondrial GPAT activity by over 90%. On the other hand, although substitution of Arg-278 with alanine and histidine decreased mitochondrial GPAT activity by 90%, replacement with lysine reduced activity by only 25%. A substitution of the nonconserved Arg-279 with either alanine, histidine, or lysine did not alter mitochondrial GPAT activity. Moreover, R278K mitochondrial GPAT still showed sensitivity to arginine-modifying agents, as in the case of wild-type mitochondrial GPAT. These results suggest that Arg-318 may be critical for mitochondrial GPAT activity, whereas Arg-278 can be replaced by a basic amino acid. Examination of the other conserved residues in the seven amino acid stretch revealed that Phe-313 and Glu-315 are also important, but conservative substitutions can partially maintain activity; substitution with alanine reduced activity by 83 and 72%, respectively, whereas substituting Phe-313 with tyrosine and Glu-315 with glutamine had even lesser effect. In addition, there was no change in fatty acyl-CoA selectivity. Kinetic analysis of the R318K and R318A mitochondrial GPAT showed an 89 and 95%, respectively, decrease in catalytic efficiency but no major change in substrate binding as indicated by the K mvalues for palmitoyl-CoA and glycerol 3-phosphate. These studies indicate importance of the conserved seven amino acid stretch for mitochondrial GPAT activity and the significance of Arg-318 for catalysis. glycerol-3-phosphate acyltransferase AGPAT, 1-acylglycerol-3-phosphate acyltransferase N-ethylmaleimide endoplasmic reticulum kilobase pair 4-morpholinepropanesulfonic acid The first committed step in glycerophospholipid biosynthesis is the acylation of glycerol 3-phosphate in the sn-1 position with a fatty acyl-CoA to form 1-acylglycerol-3-phosphate (lysophosphatidic acid) catalyzed by glycerol-3-phosphate acyltransferase (GPAT)1 (EC2.3.1.15). Lysophosphatidic acid is further acylated in thesn-2 position by 1-acylglycerol-3-phosphate acyltransferase (AGPAT) to form 1,2-diacylglycerol-3-phosphate, which is used for synthesis of all phospholipids and triacylglycerol (reviewed in Refs.1Bell R.M. Coleman R.A. Enzymes. 1983; 56: 87-111Crossref Scopus (69) Google Scholar, 2Brindley D.N. Vance D. Vance J. Biochemistry of Lipids, Lipoproteins, and Membranes. Elsevier Science Publishers B.V., Amsterdam1991: 171-203Google Scholar, 3Dircks L.K. Sul H.S. Biochim. Biophys. Acta. 1997; 1348: 17-26Crossref PubMed Scopus (60) Google Scholar). GPAT is believed to be a rate-limiting step in phospholipid biosynthesis. There are two isoforms of mammalian GPAT; one resides in the endoplasmic reticulum (ER) membrane and the other in the outer mitochondrial membrane. The two isoforms can be distinguished by their differential sensitivity to the sulfhydryl group modifying reagent,N-ethylmaleimide (NEM): mitochondrial GPAT is resistant to NEM, whereas GPAT in the ER is inactivated by NEM (4Schlossman D.M. Bell R.M. J. Biol. Chem. 1976; 251: 5738-5744Abstract Full Text PDF PubMed Google Scholar, 5Haldar D. Tso W.W. Pullman M.E. J. Biol. Chem. 1979; 254: 4502-4509Abstract Full Text PDF PubMed Google Scholar). In most tissues the mitochondrial enzyme comprises about 10% of GPAT activity, whereas in liver it comprises 50% of the total activity. Mitochondrial GPAT is responsive to nutritional and hormonal regulation, whereas the ER isoform is not; mitochondrial GPAT activity is very low in diabetic or fasted animals and is increased upon administration of insulin to diabetic animals or by refeeding fasted animals a high carbohydrate, fat-free diet (6Bates E.J. Topping D.L. Sooranna S.P. Saggerson D. Mayes P.A. FEBS Lett. 1977; 84: 225-228Crossref PubMed Scopus (46) Google Scholar, 7Bates E.J. Saggerson E.D. Biochem. J. 1979; 182: 751-762Crossref PubMed Scopus (74) Google Scholar). The ER isoform uses saturated and unsaturated fatty acyl-CoAs equally well, whereas the mitochondrial isoform has a preference for saturated fatty acyl-CoAs as a substrate (8Kelker H.C. Pullman M.E. J. Biol. Chem. 1979; 254: 5364-5371Abstract Full Text PDF PubMed Google Scholar, 9Stern W. Pullman M.E. J. Biol. Chem. 1978; 253: 8047-8055Abstract Full Text PDF PubMed Google Scholar). This preference is believed to be responsible for the observed predominance of saturated fatty acids in the sn-1 position, in contrast to unsaturated fatty acids in the sn-2 position, in naturally occurring phospholipids (10Montfoort A. Golde L.M.V. Deenen L.L.V. Biochim. Biophys. Acta. 1971; 231: 335-342Crossref PubMed Scopus (120) Google Scholar, 11Mattson F.H. Luton E.S. J. Biol. Chem. 1958; 233: 868-871Abstract Full Text PDF PubMed Google Scholar). Despite the importance of mammalian GPAT in phospholipid and triacylglycerol biosynthesis, little is known about the structural features that determine substrate binding or catalytic activity. Due to its membrane association, mitochondrial GPAT has been difficult to purify and reconstitute (12Vancura A. Haldar D. J. Biol. Chem. 1994; 269: 27209-27215Abstract Full Text PDF PubMed Google Scholar). It is unlikely that the x-ray crystallography data needed to identify the structural information required for substrate binding and catalysis will be obtained in the near future. Our isolation of a cDNA clone to murine mitochondrial GPAT, which shares 30% identity and 72% similarity over a 300 amino acid region with the previously cloned Escherichia coli GPAT (13Paulauskis J.D. Sul H.S. J. Biol. Chem. 1988; 263: 7049-7054Abstract Full Text PDF PubMed Google Scholar), has enabled us to begin characterizing the structural features required for its function. The protein has been overexpressed in baculovirus-infected insect cells. The enzyme was shown to be targeted to mitochondria, and when purified and reconstituted with phospholipids this mitochondrial GPAT preferred saturated fatty acyl-CoAs as a substrate (14Yet S.F. Moon Y.K. Sul H.S. Biochemistry. 1995; 34: 7303-7310Crossref PubMed Scopus (52) Google Scholar). We showed that murine mitochondrial GPAT containing a deletion of 78 of the most conserved amino acids, residues 250–327, had no catalytic activity (14Yet S.F. Moon Y.K. Sul H.S. Biochemistry. 1995; 34: 7303-7310Crossref PubMed Scopus (52) Google Scholar). These studies suggest that the region comprising the deletion is probably required for GPAT activity. Here we have used amino acid sequence comparisons among known GPATs and AGPATs to predict regions of mitochondrial GPAT that may be important for its activity. We report that a deletion of a seven amino acid stretch in the region most conserved among acyltransferases,312IFLEGTR318, drastically reduced mitochondrial GPAT activity. We found by using arginine-modifying agents and kinetic studies on mitochondrial GPAT with mutations a conserved arginine within this seven amino acid stretch, Arg-318, to be important for catalysis. pcDNA-mitochondrial GPAT was constructed by ligating the 2.8-kb EcoRV-FspI fragment from p3513 (15Yet S.F. Lee S. Hahm Y.T. Sul H.S. Biochemistry. 1993; 32: 9486-9491Crossref PubMed Scopus (78) Google Scholar) containing the GPAT coding sequence to EcoRV-digested pcDNAI/amp (Invitrogen). pAlter-mitochondrial GPAT was constructed by ligating the 2.8-kb EcoRV-NotI fragment from pcDNA-mitochondrial GPAT containing the mitochondrial GPAT coding sequence into SmaI-NotI-digested pAlter MAX (Promega). Polymerase chain reaction-generated mutants (R318A, R318H, R318K, and F313A) were constructed by amplifying the mitochondrial GPAT coding sequence between nucleotides 1674 and 1911 using primer P1 (5′-CAACATCAAGGCGCCGTACA-3′), which contains a NarI site, and the specific primer from Table I, which contains a SmaI site. The polymerase chain reaction product was purified from an agarose gel, digested with NarI and SmaI, and ligated toNarI-SmaI-digested pcDNA-mitochondrial GPAT. Altered SitesTM-generated mutants (ΔIFLEGTR, R278A, R278H, R278K, R279A, R279H, R279K, F313Y, E315A, and E315Q) were constructed according to the manufacturer's instructions (Promega) using a 1:5:25 pAlter-mitochondrial GPAT template:selection oligonucleotide:mutagenic oligonucleotide ratio. The specific mutagenic oligonucleotides used are shown in Table I. Mutations were confirmed by DNA sequencing using Sequenase (U. S. Biochemical Corp.).Table IOligonucleotides used in construction of mitochondrial GPAT mutantsMutantPrimer sequence1-aTwenty one nucleotides deleted in ΔIFLEGTR are shown in lowercase letters in parentheses. Mutated nucleotides are underlined.Position1-bNucleotide position relative to the mGPAT translation initiation site.ΔIFLEGTR5′-GGT C TT GCC ACT GCG GGA(gcg ggt gcc ttc cag gaa gat)CTC CAG GAA CTG CTG CTG-3′972–916R278A5′-ATC GAG CCT CGC TCT TAT GAA AAA-3′843–820R278H5′-ATC GAG CCT GTG TCT TAT GAA AAA-3′843–820R278K5′-ATC GAG CCT CTT TCT TAT GAA AAA-3′843–820R279A5′-T TTC ATC GAG CGC CCG TCT TAT GAA AAA-3′847–820R279H5′-T TTC ATC GAG GTG CCG TCT TAT GAA AAA-3′847–820R279K5′-T TTC ATC GAG CTT CCG TCT TAT GAA AAA-3′847–820F313A5′-C TGC CCG GGC ACA GGA GGT CTT GCC ACT GCG GGA GCG GGT GCC TTC CAG GGC GAT CTC CAG GA-3′988–925F313Y5′-CC TTC CAG GTA GAT CTC CAG GA-3′947–925E315A5′-GGA GCG GGT GCC TCC CAG GAA GAT CTC C-3′957–930E315Q5′-GGA GCG GGT GCC TTG CAG GAA GAT CTC C-3′957–930R318A5′-C TGC CCG GGC ACA GGA GGT CTT GCC ACT GCG GGA GGC GGT GCC TTC C-3′988–944R318H5′-C TGC CCG GGC ACA GGA GGT CTT GCC ACT GCG GGA GTG GGT GCC TT-3′988–944R318K5′-C TGC CCG GGC ACA GGA GGT CTT GCC ACT GCG GGA CTT GGT GCC TT-3′988–9441-a Twenty one nucleotides deleted in ΔIFLEGTR are shown in lowercase letters in parentheses. Mutated nucleotides are underlined.1-b Nucleotide position relative to the mGPAT translation initiation site. Open table in a new tab CMT cells, COS cells that are stably transfected with SV40 T antigen (16Gerard R.D. Gluzman Y. Mol. Cell. Biol. 1985; 5: 3231-3240Crossref PubMed Scopus (94) Google Scholar), were transfected by the DEAE-dextran method as described previously (14Yet S.F. Moon Y.K. Sul H.S. Biochemistry. 1995; 34: 7303-7310Crossref PubMed Scopus (52) Google Scholar, 17Ausubel F.M. Brent R. Kingston R.E. Moore R.E. Seidman D.D. Smith J.A. Struhl K. Current Protocols in Molecular Biology. John Wiley & Sons, Inc., New York1987Google Scholar), and cells were harvested 3 days after transfection. The transiently transfected cells were homogenized in 0.25 msucrose, 10 mm Tris-HCl, pH 7.4, 1 mm EDTA, 1 mm dithiothreitol, and 1 mm Pefabloc (Roche Molecular Biochemicals), with 8 strokes in a motor-driven Teflon-glass homogenizer at moderate speed and centrifuged at 800 ×g for 10 min. The supernatant was centrifuged at 8000 × g for 10 min to pellet the mitochondrial fraction. For kinetic assays, mitochondria were purified on a hybrid Percoll-metrizamide discontinuous density gradient as described (18Storrie B. Madden E.A. Methods Enzymol. 1990; 182: 203-225Crossref PubMed Scopus (497) Google Scholar). Ten μg of mitochondrial protein was separated on a 10% SDS-polyacrylamide gel, transferred to Immobilon-P (Millipore), and incubated with a specific antibody to mitochondrial GPAT and goat anti-rabbit secondary antibody as described previously (14Yet S.F. Moon Y.K. Sul H.S. Biochemistry. 1995; 34: 7303-7310Crossref PubMed Scopus (52) Google Scholar). The protein concentration was determined by the Bradford method using bovine serum albumen as the standard (19Bradford M.M. Anal. Biochem. 1976; 72: 248-254Crossref PubMed Scopus (215589) Google Scholar). To determine mitochondrial GPAT activity, 25 μg of mitochondrial protein was incubated on ice for 15 min with 0.4 mm N-ethylmaleimide. The assay mixture contained 75 mm Tris-HCl, pH 7.4, 4 mm MgCl2, 1 mg/ml fatty acid-free bovine serum albumin, 8 mm NaF, 30 μm palmitoyl-CoA (Sigma), 0.5 mm glycerol 3-phosphate, and 0.5 μCi of [14C]glycerol 3-phosphate (NEN Life Science Products). The reaction was started by addition of mitochondria and incubated for 30 min at 25 °C. For kinetic measurements of wild type and Arg-318-mutated GPAT, different amounts of mitochondrial protein were incubated for 60 min in the presence of varying concentrations of palmitoyl-CoA and 1 mm glycerol 3-phosphate. Labeled lipids were extracted once with water-saturated butanol, and the butanol phase was back-extracted once with butanol-saturated water and once with water, and quantitated with liquid scintillation counting. Total RNA was prepared from CMT cells transiently transfected with either wild-type or mutant mitochondrial GPATs using Trizol Reagent (Life Technologies, Inc.) according to the manufacturer's instructions. RNA was separated on a 1% agarose gel in 2.2 m formaldehyde, 20 mm MOPS, pH 7.0, and 1 mm EDTA and transferred to a nitrocellulose filter. The filter was hybridized with the 2.8-kbSalI-NotI mitochondrial GPAT fragment from pAlter-mitochondrial GPAT, which was labeled using [α-32P]dCTP and the Prime It labeling kit (Stratagene), in 50% formamide, 5× SSC at 42 °C and washed in 0.2× SSC, 0.1% SDS at 50 °C. The reaction contained 10 mm Tris-HCl, pH 7.4, 125 mm sucrose, and 10 μg/ml trypsin, as indicated. Fifteen μg of mitochondrial protein was incubated at 25 °C for the indicated time. The digestion was stopped by addition of 2 mg/ml trypsin-chymotrypsin inhibitor (Sigma). Mitochondrial proteins were then subjected to SDS-polyacrylamide gel electrophoresis and Western blot analysis. Murine mitochondrial GPAT shares 30% identity and overall 72% similarity with E. coli GPAT in a 300 amino acid region. We previously showed that deletion of 78 of the most conserved amino acids in this 300 amino acid region, residues 250–327, abolished mitochondrial GPAT activity, suggesting that this region was important for GPAT catalysis or substrate binding (14Yet S.F. Moon Y.K. Sul H.S. Biochemistry. 1995; 34: 7303-7310Crossref PubMed Scopus (52) Google Scholar). In Fig. 1, we compared known GPAT and various AGPAT sequences and found a 134 amino acid conserved region within the 300 amino acids that we previously found to be conserved between GPATs. The regions that are conserved between the GPATs and AGPATs are likely to be involved in functions that are shared between the two enzymes, such as fatty acyl-CoA binding or the esterification reaction. Within this region, amino acids312IFLEGTR318 of mitochondrial GPAT are the most conserved among these acyltransferases. We therefore constructed a mutant of mitochondrial GPAT in which these seven amino acids,312IFLEGTR318, were deleted. Wild-type and ΔIFLEGTR mitochondrial GPAT were transiently transfected into CMT cells, and total RNA and mitochondria were isolated. As shown in Fig.2, panel A, cells transfected with wild-type and ΔIFLEGTR mitochondrial GPAT contained similar levels of the 3.6-kb mitochondrial GPAT mRNA expressed from the transfected plasmids. The mitochondrial GPAT mRNA expressed from the transfected plasmids is smaller than the endogenous 6.8-kb mRNA due to the truncation of the noncoding regions. Because of its very low levels in CMT cells, the 6.8-kb mitochondrial GPAT mRNA was not observed in this Northern blot. Panel B shows the Western blot of mitochondria isolated from the transfected cells. The Coomassie staining shows that equal amounts of mitochondrial protein were loaded. The ΔIFLEGTR mitochondrial GPAT was correctly localized to mitochondria, as was wild-type mitochondrial GPAT, and that mitochondria isolated from cells expressing ΔIFLEGTR mitochondrial GPAT contained a similar level of the 90-kDa GPAT protein relative to that in mitochondria isolated from cells expressing wild-type mitochondrial GPAT. Due to deletion of only seven amino acids, there was no detectable change in apparent molecular weight of mutant mitochondrial GPATs. Nontransfected cells contained an undetectable level of endogenous mitochondrial GPAT protein. Mitochondria from cells transfected with wild-type mitochondrial GPAT showed a mitochondrial GPAT activity of 24 nmol/min/mg, whereas mitochondria from nontransfected cells showed an activity of 1.0 nmol/min/mg. Mitochondria from cells transfected with ΔIFLEGTR mitochondrial GPAT showed a GPAT activity of only 3.3 nmol/min/mg (Fig. 2, panel C). Subtraction of the endogenous mitochondrial GPAT activity from the mitochondrial GPAT activity of transfected cells indicated that wild-type mitochondrial GPAT activity from the transfected plasmid was approximately 23 nmol/min/mg, whereas ΔIFLEGTR mitochondrial GPAT activity was 2.3 nmol/min/mg. These data clearly indicate that deletion of amino acids 312IFLEGTR318 caused a drastic decrease in mitochondrial GPAT specific activity, demonstrating that this seven amino acid stretch is important for GPAT activity.Figure 2Effect of deletion of seven amino acid conserved region on mitochondrial GPAT activity. Panel A, Northern blot of RNA isolated from CMT cells transiently transfected with either wild-type or ΔIFLEGTR mitochondrial GPAT. CMT cells were transfected, and total RNA was isolated as described under “Experimental Procedures.” Lane 1, untransfected cells;lane 2, wild-type mitochondrial GPAT; and lane 3, ΔIFLEGTR mitochondrial GPAT. Five μg of total RNA were applied to each lane. Panel B, right, Western blot of mitochondria isolated from CMT cells transiently transfected with wild-type or ΔIFLEGTR mitochondrial GPAT; left, Coomassie staining. Ten μg mitochondrial protein was subjected to Western blot analysis as described under “Experimental Procedures.” Mitochondria from nontransfected cells (lane 1) or cells transfected with wild-type mitochondrial GPAT (lane 2) or ΔIFLEGTR mitochondrial GPAT (lane 3). The minor band migrating below the 90-kDa mitochondrial GPAT protein is a degradation product of mitochondrial GPAT that we observe in some of our preparations.Panel C, mitochondrial GPAT activity of mitochondria isolated from CMT cells transfected with wild-type or ΔIFLEGTR mitochondrial GPAT. Mitochondrial GPAT activity was measured as described under “Experimental Procedures,” and the contribution of endogenous mitochondrial GPAT to the mitochondrial GPAT activity was subtracted from the total activity in order to calculate mitochondrial GPAT activity from the transfected plasmid. Values are expressed as nanomoles of glycerol 3-phosphate converted to lysophosphatidic acid per min per mg of mitochondrial protein and represent the mean ± S.E. of four independent experiments.View Large Image Figure ViewerDownload (PPT) Within312IFLEGTR318 is an arginine, Arg-318, that is conserved among all the sequences shown in Fig. 1. We postulated that arginine, with its guanidino group, is well suited to interact with the phosphate groups of glycerol 3-phosphate or of the CoA moiety of fatty acyl-CoA. In fact, Green and Bell (20Green P.R. Bell R.M. Biochim. Biophys. Acta. 1984; 795: 348-355Crossref PubMed Scopus (10) Google Scholar) previously observed that the arginine-modifying agents butanedione and phenylglyoxal inactivatedE. coli GPAT. CoA, however, protected E. coliGPAT from inactivation by arginine-modifying agents, indicating that arginine may be involved in fatty acyl-CoA binding. We tested whether the arginine-modifying agents, phenylglyoxyl and cyclohexanedione, had similar effects on mitochondrial GPAT activity from murine liver. As shown in Fig. 3, both phenylglyoxyl and cyclohexanedione inactivated mitochondrial GPAT in a dose-dependent manner. This result indicates that, similar to E. coli GPAT, one or more arginine residues are important for murine mitochondrial GPAT function. In addition to Arg-318 in the seven amino acid stretch, Arg-278 is also conserved among the sequences shown in Fig. 1. These are the only two conserved arginines in mitochondrial GPAT. To test whether one or both of these arginines are important for mitochondrial GPAT activity, we used site-directed mutagenesis to substitute alanine for each arginine, individually (R278A and R318A). Since alanine has a small nonpolar side chain as compared with arginine with a large basic side chain, we hypothesized that this substitution was likely to affect mitochondrial GPAT activity if either of the arginines were important for GPAT activity. Cells transfected with R278A or R318A GPAT expressed similar amounts of mitochondrial GPAT mRNA as cells transfected with the wild-type GPAT, and mitochondria from cells expressing R278A or R318A GPAT had a similar level of mitochondrial GPAT protein as mitochondria from cells expressing wild-type GPAT (Fig. 4,panels A and B). Because R278A was constructed in the pAlter MAX vector and R318A was constructed in pcDNA-1/amp vector, RNA, protein, and mitochondrial GPAT activity of R278A were compared with those of wild-type mitochondrial GPAT expressed from pAlter-mGPAT and RNA, protein, and mitochondrial GPAT activity of R318A were compared with those of wild-type mitochondrial GPAT expressed from pcDNA-mGPAT. The mitochondrial GPAT mRNA transcribed from pcDNA-mGPAT is larger than that transcribed from pAlter-mGPAT because the pcDNA-1/amp vector contains about 400 nucleotides longer untranslated flanking sequence than the pAlter Max vector. Fig.4, panel C, shows the mGPAT activity of mitochondria isolated from cells expressing wild-type, R278A, or R318A mGPAT. Mitochondria isolated from cells expressing R278A and R318A mGPAT had only 8 and 0.8%, respectively, of the GPAT activity observed from cells expressing wild-type mGPAT. We also substituted both of these arginines with either histidine or lysine (R278H, R278K, R318H, and R318K) to test whether a positive charge at Arg-278 and Arg-318 was sufficient for GPAT activity or if arginine itself was required. Arginine to histidine substitutions for both Arg-278 and Arg-318 reduced mitochondrial GPAT activity to approximately 10 and 3%, respectively, of the wild-type activity (TableII). Substitution of Arg-278 with lysine reduced mitochondrial GPAT activity by only about 25%, whereas substitution of Arg-318 with lysine reduced the enzyme's activity by almost 90%. These data indicate that Arg-318 is important for substrate binding or catalysis of mitochondrial GPAT, because even a conservative substitution of lysine for Arg-318 drastically reduced enzyme activity. On the other hand, a positive charge at position 278 was sufficient for activity since an arginine to lysine substitution did not substantially reduce mitochondrial GPAT activity. As shown in the sequence alignment in Fig. 1, only 2 of 10 arginines (Arg-278 and Arg- within the 134 amino acid region are conserved among the acyltransferases. We also mutated Arg-279, which is not conserved among the sequences shown in Fig. 1, to alanine, histidine, or lysine (R279A, R279H, and R279K). The results in Table II show that, in contrast to Arg-278 and Arg-318, mutation of Arg-279 did not reduce mitochondrial GPAT activity compared with wild-type GPAT. Based on these results, we conclude that Arg-318 is essential for mitochondrial GPAT activity, since amino acid substitutions of Arg-318, including a conservative arginine to lysine substitution, abolished enzyme activity, whereas a lysine substitution for Arg-278 did not significantly affect activity. The Arg-279 to alanine substitution did not affect enzyme activity, demonstrating that a nonconserved arginine residue is not critical for mitochondrial GPAT activity.Figure 4Effect of alanine substitutions for conserved arginine residues on mitochondrial GPAT activity. Panel A, right, Northern blot of RNA isolated from CMT cells transiently transfected with the indicated mGPAT construct; left,ethidium bromide staining. Lane 1, untransfected CMT cells;lane 2, R278A; lane 3, wild type; lane 4, R318A; lane 5, wild type. Twenty μg of total RNA were applied to each lane. Panel B, Western blot of mitochondria isolated from CMT cells transfected with the indicated mGPAT construct. Lane 1, untransfected; lane 2,wild type; lane 3, R278A; lane 4, wild type; andlane 5, R318A. Panel C, mGPAT activity of mitochondria isolated from CMT cells expressing the indicated mGPAT mutant relative to that of mitochondria isolated from CMT cells expressing wild-type mGPAT. Values represent the mean ± S.E. of at least three experiments.View Large Image Figure ViewerDownload (PPT)Table IIEffect of amino acid substitutions on mitochondrial GPAT activityMitochondrial GPATGPAT activity2-amGPAT activity of mitochondria isolated from CMT cells expressing the indicated mutant form of mGPAT relative to that of mitochondria isolated from cells expressing wild-type mGPAT. Values are means ± S.E. of at least three independent experiments.% wild-typeR278H10.2 ± 5.0R278K75.5 ± 13.3R279A101.0 ± 9.5R279H95.0 ± 21.8R279K101.2 ± 11.5F313A16.5 ± 5.2F313Y87.8 ± 21.5E315A28.2 ± 3.6E315Q46.7 ± 3.3R318H2.6 ± 2.4R318K11.1 ± 6.12-a mGPAT activity of mitochondria isolated from CMT cells expressing the indicated mutant form of mGPAT relative to that of mitochondria isolated from cells expressing wild-type mGPAT. Values are means ± S.E. of at least three independent experiments. Open table in a new tab To identify further the arginine residue responsible for inactivation of mitochondrial GPAT by arginine-modifying agents, we treated mitochondria isolated from CMT cells expressing R278K GPAT with phenylglyoxal and cyclohexanedione. As shown in Fig.5, both phenylglyoxal and cyclohexanedione inactivated R278K GPAT, as in the case of wild-type GPAT. These results suggest that modification of an arginine other than Arg-278, probably Arg-318, by phenylglyoxal and cyclohexanedione inactivates mitochondrial GPAT. Due to the difficulty of studying extremely low activity of Arg-318 GPAT, we did not determine the effects of phenylglyoxal and cyclohexanedione on mutants of Arg-318. Nevertheless, the results of our mutagenesis of conserved arginine residues indicate that Arg-318 is the most important for mitochondrial GPAT activity, and mutation of Arg-278 did not change the susceptibility to arginine-modifying agents. To determine whether the overall protein conformation of R318K GPAT was altered by the mutation, we analyzed the R318K mutant by protease digestion. Fig.6 shows a comparison of the proteolytic digestion pattern of wild-type and R318K GPAT. The time course and appearance of digestion products from wild-type and R318K GPAT were similar, indicating that the accessibility of trypsin cleavage sites, and thus the overall protein structure, were probably not different in the wild-type and the mutated proteins. These data indicate that the inactivation of mitochondrial GPAT by the R318K mutation was most likely not caused by altered folding of the protein but was due to effects of the specific mutation on GPAT activity.Figure 6Trypsin digestion of wild-type and R318K mitochondrial GPAT mutants. Mitochondria were isolated from CMT cells expressing either wild-type (lanes 1–4) or R318K (lanes 5–8) mitochondrial GPAT. Fifteen μg of mitochondrial protein were subjected to digestion at 25 °C with 10 μg/ml trypsin for the indicated times. Mitochondrial GPAT digestion products were analyzed by Western blot. Lanes 1 and5, 0 min; lanes 2 and 6, 10 min;lanes 3 and 7, 20 min; lanes 4 and8, 60 min.View Large Image Figure ViewerDownload (PPT) We have determined by the deletion of residues312IFLEGTR318 that these seven conserved amino acids are essential for mitochondrial GPAT activity. Further mutagenesis demonstrated that Arg-318 is critical to enzyme function. Next, we mutated other conserved amino acids adjacent to Arg-318 in the seven amino acid stretch to determine whether they also contributed to mitochondrial GPAT catalytic activity. The phenylalanine at position 313 was substituted with either alanine (F313A) or tyrosine (F313Y), and the glutamic acid at position 315 was substituted with either alanine (E315A) or glutamine (E315Q). Substitution of Phe-313 with alanine reduced mitochondrial GPAT activity by approximately 83%, whereas substitution with tyrosine, another aromatic amino acid, did not significantly reduce activity (Table II). Substitution of Glu-315 with alanine reduced mitochondrial GPAT activity by 72%, whereas the more conservative glutamine substitution decreased mitochondrial GPAT function by 53%. These results indicate that Phe-313 and Glu-315 are important for mitochondrial GPAT activity because substitutions of either with alanine significantly reduced mitochondrial GPAT activity. However, in the case of Phe-313, an aromatic residue is probably sufficient because tyrosine could substitute for phenylalanine with no significant loss of activity. In the case of Glu-315, the polar character and/or chain length of glutamine may also have enabled E315Q mitochondrial GPAT to function, albeit at a reduced efficiency. Fig. 7 summarizes the effects of each substitution of the conserved amino acids in mitochondrial GPAT. As discussed above, substitution of Arg-278 or Arg-318 with alanine or histidine and of Arg-318 with lysine had the most dramatic effects on mitochondrial GPAT function, reducing activity by at least 90%. Although Arg-278 may play a role, Arg-318 is the most important arginine residue for mitochondrial GPAT activity. A substitution of Phe-313 with alanine and Glu-315 with glutamine also significantly reduced enzyme activity. These results indicate that Phe-313 and Glu-315, which are adjacent to Arg-318, are also important, but substitutions with similar amino acids can restore enzyme activity. Presently we do not know whether Arg-318 plays a critical role in substrate binding or catalysis. Bell and co-workers (20Green P.R. Bell R.M. Biochim. Biophys. Acta. 1984; 795: 348-355Crossref PubMed Scopus (10) Google Scholar) showed that CoA and palmitoyl-CoA protectedE. coli GPAT from inactivation by arginine-modifying agents and speculated that arginines are at or near the active site of the enzyme. In contrast, we observed that CoA did not protect murine mitochondrial GPAT from inactivation by arginine-modifying agents (data not shown). These results suggest that Arg-318 might not be involved in fatty acyl-CoA binding. Therefore, we attempted to examine the potential role of this conserved region in substrate binding by comparing kinetic characteristics. Due to the very low enzymatic activity of the Arg-318 mutants, we further purified the mitochondria by Percoll gradient, eliminating microsomal GPAT contamination. We employed varying concentrations of palmitoyl-CoA and 1 mmglycerol 3-phosphate. As we previously reported, the inhibitory effect of fatty acyl-CoA at higher concentrations makes it difficult to determine the apparent affinity for fatty acyl-CoAs. In calculating the kinetic data, only the palmitoyl-CoA concentrations that do not inhibit GPAT activity were used. The V max for R318K and R318A were 3.6 and 1.8% (0.8 and 0.4 nmol/min/mg, respectively) that of wild-type mitochondrial GPAT (22.2 nmol/min/mg) (TableIII). The K m values for palmitoyl-CoA of R318K and R318A (4.6 and 3.2 μm, respectively) were not increased relative to the K mvalue of wild-type mitochondrial GPAT (11.3 μm). The catalytic efficiency, as estimated byV max/K m, of R318K and R318A were decreased by 89 and 95%, respectively, relative to the wild type. These results suggest that the R318K and R318A mutations did not significantly affect the affinity of the enzyme for palmitoyl-CoA, but catalytic efficiency was reduced. In addition, theK m value for glycerol 3-phosphate of R318A was not significantly changed from the wild-type enzyme (data not shown). We previously reported that murine mitochondrial GPAT was most active with saturated fatty acyl-CoAs of chain length 8–16. Mitochondrial GPAT was less than 30% as active when unsaturated acyl-CoAs were used as compared with palmitoyl-CoA (14Yet S.F. Moon Y.K. Sul H.S. Biochemistry. 1995; 34: 7303-7310Crossref PubMed Scopus (52) Google Scholar). We employed E315Q GPAT, which has approximately 50% of the wild-type mitochondrial GPAT activity, to test whether the preference for saturated fatty acyl-CoA of mitochondrial GPAT was maintained. Fig. 8shows the activity of wild-type (left panel) and E315Q mitochondrial GPAT (right panel), respectively, at increasing concentrations of the indicated fatty acyl-CoAs. Similar to the wild-type enzyme, E315Q mitochondrial GPAT was active with lauroyl-CoA (C12:0) and palmitoyl-CoA (C16:0), was 28% as active with oleoyl-CoA (C18:1), and 43% as active with linoleoyl-CoA (C18:2) as with palmitoyl-CoA when the maximal activity with each fatty acyl-CoA was compared. Overall, E315Q mitochondrial GPAT was about 50% as active as wild-type mitochondrial GPAT with each fatty acyl-CoA. We could not directly compare the apparent K m values of the mutant and wild-type enzymes for the acyl-CoAs because various acyl-CoAs were inhibitory at differing concentrations (14Yet S.F. Moon Y.K. Sul H.S. Biochemistry. 1995; 34: 7303-7310Crossref PubMed Scopus (52) Google Scholar). However, we detected no significant differences in the ability of E315Q mitochondrial GPAT to use the various acyl-CoAs as a substrate. As with the Arg-318 mutations, there was no major change in substrate binding, as indicated by the K m value for glycerol 3-phosphate of E315Q and wild-type GPAT, 0.48 and 0.27 mm, respectively. We previously reported the K m value for mitochondrial GPAT overexpressed in Sf9 insect cells to be 0.67 mm, which is an order of magnitude higher than that of the E. coli enzyme. Regardless, the E315Q mutation does not significantly affect the preference for saturated fatty acyl-CoA as a substrate.Table IIIComparison of kinetic characteristics of Arg-318 mutants with wild-type mitochondrial GPATMitochondrial GPATK m, palmitoyl-CoAV maxV max/K mμmnmol/min/mgWild-type11.3 ± 1.422.2 ± 2.52.0 ± 0.6R318A3.2 ± 0.30.4 ± 0.20.1 ± 0.1R318K3-aAverage of two experiments.4.60.80.23-a Average of two experiments. Open table in a new tab As the first committed, and possibly a rate-limiting, step in phospholipid and triacylglycerol biosynthesis, GPAT is a pivotal enzyme in lipid metabolism. Mitochondrial GPAT exhibits a preference for saturated fatty acyl-CoAs (8Kelker H.C. Pullman M.E. J. Biol. Chem. 1979; 254: 5364-5371Abstract Full Text PDF PubMed Google Scholar, 9Stern W. Pullman M.E. J. Biol. Chem. 1978; 253: 8047-8055Abstract Full Text PDF PubMed Google Scholar) and is regulated by nutritional and hormonal manipulations (6Bates E.J. Topping D.L. Sooranna S.P. Saggerson D. Mayes P.A. FEBS Lett. 1977; 84: 225-228Crossref PubMed Scopus (46) Google Scholar, 7Bates E.J. Saggerson E.D. Biochem. J. 1979; 182: 751-762Crossref PubMed Scopus (74) Google Scholar). Therefore, mitochondrial GPAT may be critical for establishing the observed predominance of saturated over unsaturated fatty acids in the sn-1 position of cellular phospholipids and for responding to physiological states that require alterations in the metabolism of glycerophospholipids. Despite the importance of mitochondrial GPAT, very little is known about the primary structural features that mediate substrate binding and catalysis. Understanding the mechanism of mitochondrial GPAT function has been hindered by the difficulty of purifying this membrane-bound enzyme. Our cloning of the mitochondrial GPAT cDNA and determination of the amino acid sequence enabled us to compare the sequences of known acyltransferases involved in phospholipid biosynthesis to predict regions of mitochondrial GPAT that would likely be important for function. We identified a 134-amino acid region that is conserved among known acyltransferases. We showed here that a seven amino acid deletion within this conserved region,312IFLEGTR318, drastically reduced mitochondrial GPAT catalytic activity. We used site-directed mutagenesis to substitute highly conserved amino acids within this region. Our data revealed that Arg-318 is important for mitochondrial GPAT activity since even a conservative substitution from arginine to lysine inactivated the enzyme. On the other hand, a positively charged amino acid can replace the other conserved arginine at position 278 (Table II). Other conserved amino acid residues that are close to Arg-318 also contribute to mitochondrial GPAT function because mutations of those amino acids also resulted in substantial loss of catalytic activity, but conservative replacement could maintain partial activity. In the absence of x-ray crystallographic or other structure data for the glycerolipid acyltransferases, we can only speculate on the role of the essential Arg-318. A possible role for arginines in mitochondrial GPAT is to mediate binding of the substrates glycerol 3-phosphate or palmitoyl-CoA by interacting with the phosphate groups, as has been proposed for other enzymes (21Shanmugasundaram T. Kumar G.K. Shenoy B.C. Wood H.G. Biochemistry. 1989; 28: 7112-7116Crossref PubMed Scopus (10) Google Scholar). Green and Bell (20Green P.R. Bell R.M. Biochim. Biophys. Acta. 1984; 795: 348-355Crossref PubMed Scopus (10) Google Scholar) demonstrated that arginine-modifying agents inactivated E. coli GPAT and CoA or palmitoyl-CoA partly protected GPAT from inactivation, and they proposed that important arginine residues were at or near the active site. Similar to the result with E. coli GPAT, phenylglyoxal and cyclohexanedione also inactivated murine mitochondrial GPAT. However, in contrast to the case with E. coli GPAT, CoA did not protect mitochondrial GPAT from inactivation. Arg-278, Arg-318, or both, which are conserved in acyltransferases involved in glycerolipid biosynthesis, could be the targets for inactivation by arginine-modifying agents. Since R278K mitochondrial GPAT was also inactivated by phenylglyoxal and cyclohexanedione (Fig. 5), it is likely that modification of Arg-318 is responsible for mitochondrial GPAT inactivation. The fact that CoA did not protect mitochondrial GPAT from inactivation might indicate that Arg-318 is not involved in CoA binding. Our observation that a substitution of Arg-318 with either lysine or alanine did not alter affinity for palmitoyl-CoA suggests its potential role in catalysis. In this regard, we did not observe a significant change in K m for glycerol 3-phosphate either. Heath and Rock (22Heath R.J. Rock C.O. J. Bacteriol. 1998; 180: 1425-1430Crossref PubMed Google Scholar) recently reported that the conserved histidine 306 and aspartic acid 311 of E. coli GPAT, which correspond to His-230 and Asp-235 within the 134-amino acid conserved region of murine mitochondrial GPAT, are important for the catalytic activity ofE. coli GPAT because a substitution of alanine for His-306 or glutamic acid for Asp-311 significantly reduces the GPAT activity. These amino acids are part of a HX 4D consensus sequence that is found in glycerolipid acyltransferases from a variety of organisms. Because the K m value of D311E for glycerol 3-phosphate is not significantly different from that of wild-type E. coli GPAT, they hypothesized that the HX 4D consensus sequence was not involved in substrate binding and that, by analogy to chloramphenicol acetyltransferase, the histidine might function as a general base to deprotonate the hydroxyl moiety of the acyl acceptor. Since this sequence is also conserved in murine mitochondrial GPAT, these residues may play a similar role in the mammalian mitochondrial enzyme. Regardless, we have demonstrated that amino acids312IFLEGTR318, Arg-318 in particular, play a significant role in catalysis of mitochondrial GPAT. Further studies are necessary to identify the residues involved in catalysis and to define the regions responsible for determining fatty acyl-CoA substrate specificity. We thank Naxin Yao for constructing some of the mutants, Dr. Keith Allen for preparing the multiple sequence alignment, and Dr. Mary Ann Willaims and members of the Sul laboratory for helpful discussions.