Covalent binding of fatty acid to the transferrin receptor in cultured human cells.

The human transferrin receptor could be fluorographically detected after immunoprecipitation from a leukemic T-cell line labeled with [3H]palmitic acid. The label was found ony in association with the human transferrin receptor and not in association with two other major plasma membrane glycoproteins, demonstrating that the incorporation of radioactivity was not due to metabolism of the palmitate. Treatment of sodium dodecyl sulfate-polyacrylamide gels containing the [3H]palmitate-labeled transferrin receptor with hydroxylamine, prior to fluorography, resulted in release of a substantial fraction of the label from the molecule. In addition, at least part of the label released from immunoprecipitates of the transferrin receptor by treatment with hydroxylamine was identified as palmitohydroxamate, providing further evidence that the labeled fatty acid is covalently bound to the receptor. A proteolytic fragment (Mr = 70,000) derived from the portion of the transferrin receptor exposed on the cell surface can be obtained by trypsin digestion of intact or Nonidet P-40-solubilized cells. When cells were labeled with [3H]palmitic acid, none of the radioactivity could be detected in the tryptic fragment. Thus, the bound palmitate appears to be associated with the region of the molecule that is in close proximity to the plasma membrane.


Institute, San Diego, California 92138
The human transferrin receptor could be fluorographically detected after immunoprecipitation from a leukemic T-cell line labeled with [3Hlpalmitic acid. The label was found only in association with the human transferrin receptor and not in association with two other major plasma membrane glycoproteins, demonstrating that the incorporation of radioactivity was not due to metabolism of the palmitate. Treatment of sodium dodecyl sulfate-polyacrylamide gels containing the [3Hlpalmitate-labeled transferrin receptor with hydroxylamine, prior to fluorography, resulted in release of a substantial fraction of the label from the molecule. In addition, at least part of the label released from immunoprecipitates of the transferrin receptor by treatment with hydroxylamine was identified as palmitohydroxamate, providing further evidence that the labeled fatty acid is covalently bound to the receptor.
A proteolytic fragment (Mr = 70,000) derived from the portion of the transferrin receptor exposed on the cell surface can be obtained by trypsin digestion of intact or Nonidet P-40-solubilized cells. When cells were labeled with [3Hlpalmitic acid, none of the radioactivity could be detected in the tryptic fragment. Thus, the bound palmitate appears to be associated with the region of the molecule that is in close proximity to the plasma membrane.
Covalent attachment of fatty acids to proteins has been demonstrated in only a few cases. In bacteria, for example, Braun's lipoprotein of Escherichia coli cell wall ((1) and references therein) is a case that has been extensively studied. In the case of animal viruses, Schmidt et al. have shown vesicular stomatitis virus G glycoprotein (2) and Sindbis virus El and E2 glycoproteins (3) to have covalently attached palmitate. Proteolipids isolated from the sarcoplasmic reticulum of rabbit skeletal muscle (4) or from brain white matter (5) were also found to have fatty acids covalently bound to protein moieties. More recently, Schlesinger et al. (6) presented evidence for fatty acid acylation of proteins from a cultured human cell line. Their results suggest that covalent linkage of fatty acids to mammalian proteins may not be uncommon. However, none of the labeled proteins in these studies have been identified.
Using a monoclonal antibody, we have identified a human cell surface antigen that is selectively expressed on proliferating cells (7,8). This antigen is a glycoprotein with an apparent molecular weight of 95,000 under reducing conditions on SDS-polyacrylamide gels. Under nonreducing conditions, the apparent molecular weight is approximately 190,000 suggesting that the molecule exists in the membrane as a disulfide-bonded dimer. More recently, we have shown that this glycoprotein is the human transferrin receptor (9). In this report we describe another feature of the glycoprotein: its covalent association with fatty acid.

MATERIALS AND METHODS
Cell Lines-Human leukemic T-cell lines CCRF-CEM and RPMI 8402 were used (10 (Amersham, 1190 Ci/mmol) and cell surface lactoperoxidase-catalyzed iodination were carried out as previously described (11). Immunoprecipitation and SDS-Polyacrylamide Gel Electrophoresis-Immunoprecipitates were prepared from 2 X 106 cells using monoclonal antibodies that recognize, respectively, antigenic determinants on the human transferrin receptor, HLA, or the human homologue of murine T200 glycoprotein (7,9,12). Antibody-antigen complexes were collected with 0.15 ml of a 1% fixed Staphylococcus aureus suspension (v/v) and were then washed 3 times with 0.4 ml of 0.5% deoxycholate, 0.5% Nonidet P-40, 0.05% SDS in phosphatebuffered saline. The antigens were released in 40-70 gl of either reducing or nonreducing electrophoresis sample buffer (7, 13) by boiling for 2 min. After pelleting, 10-30 gl of each sample were subjected to SDS-polyacrylamide gel electrophoresis (7.5% or 10% gels) as previously described (11). After electrophoresis, gels were processed for fluorography (14) either directly or after fixing and staining. Both treatments gave similar results in terms of intensity of [3H]palmitate bands.

DISCUSSION
In this report, we have presented evidence for the covalent association of pahnitic acid with the cell surface transferrin receptor of cultured human T leukemic cell lines. Thus, we have extended earlier observations suggesting that certain proteins in cultured human cells contain fatty acid residues (6), by identifying a specific plasma membrane glycoprotein that is modified in this manner.
The functional significance of the covalent attachment of fatty acid to the transferrin receptor is not clear. One possibility is that the covalently bound fatty acid anchors the glycoprotein in the plasma membrane. This would be consistent with the observation that [3H]palmitate labels only that portion of the transferrin receptor containing the region that interacts with the lipid bilayer. It is noteworthy that neither of the transmembrane glycoproteins HLA (17) nor human T200 (11,12 and Footnote 2) glycoproteins labeled with [3Hjpalmitate. Consequently, covalent fatty acid attachment may not be a general feature of transmembrane glycoproteins and may indicate that the transferrin receptor represents a different class of membrane-bound glycoproteins, particularly since we have not been able to clearly demonstrate that the receptor is exposed on the cytoplasmic side of the plasma membrane.2 Another possible role for the addition of fatty acid residues to membrane glycoproteins is in their transport to the cell surface. Studies of the vesicular stomatitis virus G glycoprotein suggest that addition of fatty acid to this molecule may occur as the glycoprotein moves to the plasma membrane (19,20). Covalent binding of fatty acid to the transferrin receptor offers a new model for studying fatty acid biosynthetic modffication of glycoproteins.