Mitochondrial glycerol phosphate acyltransferase contains two transmembrane domains with the active site in the N-terminal domain facing the cytosol.

Maria R Gonzalez-Baró,Rosalind A Coleman,Deborah A Granger

doi:10.1074/jbc.m107885200

Abstract

The topography of mitochondrial glycerol-3-phosphate acyltransferase (GPAT) was determined using rat liver mitochondria and mutagenized recombinant rat GPAT (828 aa (amino acids)) expressed in CHO cells. Hydrophobicity analysis of GPAT predicts two transmembrane domains (TMDs), residues 472-493 and 576-592. Residues 224-323 correspond to the active site of the enzyme, which is believed to lie on the cytosolic face of the outer mitochondrial membrane. Protease treatment of rat liver mitochondria revealed that GPAT has a membrane-protected segment of 14 kDa that could correspond to the mass of the two predicted TMDs plus a loop between aa 494 and 575. Recombinant GPAT constructs containing tagged epitopes were transiently expressed in Chinese hamster ovary cells and immunolocalized. Both the C and N termini epitope tags could be detected after selective permeabilization of only the plasma membrane, indicating that both termini face the cytosol. A 6-8-fold increase in GPAT-specific activity in the transfected cells confirmed correct protein folding and orientation. When the C terminus and loop-tagged GPAT construct was immunoassayed, the epitope at the C terminus could be detected when the plasma membrane was permeabilized, but loop-epitope accessibility required disruption of the outer mitochondrial membrane. Similar results were observed when GPAT was truncated before the second TMD, again consistent with an orientation in which the loop faces the mitochondrial intermembrane space. Although protease digestion of the HA-tagged loop resulted in preservation of a 14-kDa fragment, consistent with a membrane protected loop domain, neither the truncated nor loop-tagged enzymes conferred GPAT activity when overexpressed, suggesting that the loop plays a critical structural or regulatory role for GPAT function. Based on these data, we propose a GPAT topography model with two transmembrane domains in which both the N (aa 1-471) and C (aa 593-end) termini face the cytosol and a single loop (aa 494-575) faces the intermembrane space.

Highlights

Catalyzes the first and committed step in de novo cellular glycerolipid synthesis, the formation of 1-acyl-sn-glycerol-3-phosphate from glycerol-3-phosphate and long chain fatty acyl-CoA substrates [1, 2]
Protease treatment of rat liver mitochondria revealed that glycerol-3-phosphate acyltransferase (GPAT) has a membraneprotected segment of 14 kDa that could correspond to the mass of the two predicted transmembrane domains (TMDs) plus a loop between aa 494 and 575
With the second loop construct, Tr576Myc, which is truncated just before the second TMD, cells became stained with the Myc antibody only after intracellular membranes had been disrupted with Triton X-100; the Myc epitope was not accessible to the antibody in cells permeabilized with digitonin, indicating that the shortened C terminus was located in the intermembrane space. These results indicate that GPAT C- and N-terminal domains are located in the cytosolic face of the outer mitochondrial membrane (OMM) and that the loop lies in the intermembrane space

Summary

EXPERIMENTAL PROCEDURES

Materials—Rabbit polyclonal anti-GPAT antibody raised against the gel-purified protein expressed in bacteria was commercially produced by Immunodynamics, La Jolla, CA. The polymerase chain reaction product was inserted in the BamHI-XbaI sites of the multicloning site of pcDNA3.1 This construct pcDNA3.1-GPAT-FLAG is referred to as GFLAG. Immunoblotting—The mitochondrial protease digestion products were separated on a 4 –20% gradient (8% for the cellular particulate of CHO cells expressing recombinant GPAT) polyacrylamide gel containing 1% SDS and transferred to a polyvinylidene difluoride membrane (Bio-Rad). GPAT Assays—GPAT was assayed in rat liver mitochondria (20 – 80 ␮g of protein) and in total particulate preparations from CHO cells expressing the recombinant GPAT constructs (40 – 80 ␮g of protein). GPAT activity was assayed with immobilized palmitoylCoA on agarose beads (180 ␮M) (Sigma) using rat liver mitochondria or CHO cell total particulate preparations under isosmotic conditions (intact) or hyposmotic conditions (10 mM Tris-HCl, pH 7.4), which disrupted the OMM. Disruption of the OMM was monitored by loss of adenylate kinase activity

RESULTS

TABLE I The active site of GPAT faces the cytosol

DISCUSSION