In vitro acylation of the transferrin receptor

THE.JOURNAL IC OF BIOLOGICAL CHEMISTRY 1984 by The American Society of Biological Chemists, Inc. Vol. 259. No. 24, Issue of December 25, pp. 15460-15463,1964 Printed in U.S.A. In Vitro Acylation of the Transferrin Receptor* (Received for publication, June 4, 1984) Mohammed Adam,Angel Rodriguez, Claire Turbide, James LarrickS, Edward Meighen, and Rose M. Johnstones From the Departmentof Biochemistw, McGill University, Montreal, Quebec, H3G 1 Y6 Canada and the $Cetus Corporation, Palo Alto, California94303 In vitro fatty acylation of the transferrin receptor with [3H]tetradecanoateor [3H]tetradecanoyl-CoAhas been demonstrated for isolated sheep reticulocyte plasma membranes. Although less than 5%of the receptor was labeled in vitro,the acylated protein could be readily observed after sodium dodecyl sulfate-gel electrophoresis. The acylated transferrin receptor in the reticulocyte membrane wasspecifically precipitated with a monoclonal antibody and was absent from mature red cell membranes. Incorporation of fatty acid was dependent on ATP, and fatty acid was 5-10 times less effectiveas an acyldonor than the acyl-CoA derivative, pointing out the strong potential of this reagent for in vitroacylation of membrane proteins. During in vitro maturation of reticulocytes, the receptor is released in vesicles into the incubation medium. Using reticulocytes labeled with [3H]tetradecanoate,it can be shown that the 3H-labeled receptor is transferred from the cells to the vesicles without loss of acyl groups, suggesting that the vesiculation process does not involve deacylation. branes have been implicated as the sitesof acylation (4). The transferrin receptor isa transmembrane protein (6, 7) which is essential for iron uptake from transferrin (8-10). In rapidly growing cells or reticulocytes which are still synthesizing hemoglobin,this receptor is a prominent componentof the cell surface (11-16). During red cell maturation i n vivo and in uitro, the transferrin receptor is lost and the mature erythrocyte is devoid of transferrin-binding capacity (7, 1719). Recently, it hasbeen shown that the transferrinreceptor of sheep reticulocytes is excised from the cell surface during i n uitro maturation and can be recovered in the medium in the form of a vesicle containing an intactreceptor (20). If the reticulocyte is incubated with the antireceptor antibody and/ or transferrin, the isolated vesicles contain either or both of these ligands (20), showing that the receptor retains its specific binding capacities after being released from the cell. The questionaddressed here is whether theacylation of the transferrin receptor occurs in reticulocytes and whether the acylating activity can be measured in isolated membranes. The ease of preparing reticulocyte membranes and the relative lack of intracellular organelles have made this system ideal for addressing this question.Moreover, it provided an opportunity to determine whetherreceptor released to the medium during in uitro maturation of reticulocytes retained the acyl groups incorporated i n vitro. We have recently found that post-translational modification of the transferrin receptorby phosphorylation can occur in isolated plasma membranesof sheep reticulocytes (1).Since Omary andTrowbridge (2) have demonstrated that fattyacid MATERIALS ANDMETHODS may be incorporated into the transferrin receptor in cultured Reticulocytes were prepared from phlebotomized sheep and incucells many hours after the translational event, the question were arose whetherisolated plasma membranesof reticulocytes are bated i n uitro as described (21). Plasmamembraneghosts still capableof acylating the transferrinreceptor. Although it prepared by the method of Dodge et al. (22) using a %o volume for lysis. The membranes were washed twice in lysing buffer followed by has been shown that red cell plasma membrane proteinsmay one or two washes in 5 mM phosphate buffer, pH 7.4, and resuspended be acylated i n vitro (3,4),specific proteins were not identified. in the same buffer to a final concentration of 1 mg of protein/ml With the exception of studies on red cell membranes, acyla- [3H]Tetradecanoic acid (21 Ci/mmol) was prepared by New England Nuclear. [3H]Tetradecan~yl-CoA(0.76-5.0 Ci/mmol) was prepared tion in eukaryoticsystemshas onlybeen examinedwith by the method of Bishop and Hajra (23). cultured, intact cells (2, 4, 5 ) . A recent paper demonstrated Acylation of the total membranes was performed by incubation at the acylation of a viral protein with microsomes from mam22 “C of 250 ~1 (-250 pgof protein) of the membrane preparation malian cells (31). with 14 ~1 of 0.25 mM [3H]tetradecanoic acid in ethanol and 14 pl of With a cultured cell line, Omary andTrowbridge (2) showed 0.1 M ATP, pH 7. Acylation with [3H]tetradecanoyl-CoA was perthat the acyl moiety of the receptor turned over faster than formed under the same conditions, but the fatty acid and ATP were the protein itself, suggesting that there may be a functional substituted by 10 pl of 0.3 mM [3H]tetradecanoyl-CoA in phosphate role, as well as a structural onefor receptor acylation. No role buffer. Total incorporation of radioactive fatty acid into protein was for acylation or deacylation is known, although many roles determined by placing an aliquot of the membrane incubated with 3H-labeled fatty acid and ATP on filter paper and measuring the havebeenproposedincluding facilitation of anchoring of amount of radioactivity remaining after repeated extraction with proteins into the membrane and formation of membrane buds chloroform/methanol/acetic acid (24). during release of certain viruses (4).Most acylated proteins Immunoprecipitation of thetransferrin receptor from acylated appear to be membrane bound, and the intracellular mem- membranes was carried out as follows. After incubation with either * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ To whom correspondence should be addressed. 3H-labeled fatty acid (+ATP) or [3H]acyl-CoA,the membranes were centrifuged at 31,000 X g. The washed pellet was dissolved with stirring for 30 min on ice in 0.3 ml of 1% Triton X-100 containing 0.5 mM EDTA, 2.5 mM phosphate buffer, pH 8, and 0.15 units/ml aprotinin. The solubilized membranes were diluted 1:l with 25 mM phosphate buffer, pH 7.4, in isotonic saline to bring theTriton 15460 This is an Open Access article under the CC BY license. In Vitro Acyhtionof the Transferrin Receptor 15461 2 3 4 5 concentration to 0.5%. The undissolvedresiduewas removed by 1 2 1 2 1 centrifugation at 35,000 X g for 30 min and discarded. An aliquot (34 pg of protein) of a monoclonal antibodyagainstthetransferrin receptor was added to the supernatant (7). In controls for nonspecific 186Kprecipitation of radioactivity, either nonimmune serum or an equiv186K- .Ialent amount (34 pg) of nonimmune y-globulin was used. Mature red cells from sheep were treated identically to reticulocytes and immunoprecipitated with monoclonal antibody. " After overnight incubation a t 4 "C, either protein A-Sepharose or 100 pI of a 10% suspension of formalin-fixed Staphylococcus aureus was added, and the incubationwas continued for -1-2 h a t 4 "C. The precipitates werewashed 4-6 times in 0.1% Triton X-100,1 mM 0 55K" EDTA, 20 mM NaCI, 8 mM Hepes,' pH 7.4, resuspended in 60 pl of sample buffer consisting of 2% SDS, 10% glycerol, 10% fl-mercapto- 43Kethanol, 154 mM NaCI, and 20 mM Tris/HCI, pH 7.4, and then boiled for 5 min. 30KThe supernatants afterSDS extraction were subjected to gel electrophoresis using the Laemmli procedure (25)and 5-15% polyacryl25K" amidegradient gels. For radioautography,the gels were stained, destained, treated with ENHANCE, dried, and radioautographed a t FIG.2. Acylation of thetransferrinreceptor inisolated -70 "C for 2 weeks. T o examine the lability of the 3H incorporated sheep reticulocyte and erythrocyte membranes. Incorporation to hydroxylamine, the procedure described in Ref. 26 was followed. of 3H label was carriedout for 60 min a t room temperatureas described under "Materials andMethods." Left panel, after incubation with [3H]tetradecanoic acid (+ATP) (21 Ci/mmol) or with ['HI RESULTS tetradecanoyl-CoA (0.76 Ci/mmol), reticulocyte membranes were disIncubation in uitro of sheep reticulocyte plasma membranes solved in SDS electrophoresis sample buffer and electrophoresed on with [''Hltetradecanoic acid (+ATP) leads to the incorpora- slab gels. Lune I shows a radioautogram of ['Hltetradecanoic acid lane 2 showsa tion of fatty acid into the isolated plasma membrane. Based incorporationintototalmembraneproteins,and radioautogram of [3H]tetradecanoyl-CoA incorporation in total memon itsinsolubility in chloroform/methanol/acetic acid and the brane proteins. Center panel, Coomassie Blue stains of immunopredependence of the incorporation of 'H-labeled fatty acid on cipitates from reticulocyte and erythrocyte membranes (lanes 1 and ATP, a covalent linkage to the protein issuggested (Fig. 1). 2, respectively). Right panel; radioautographs of "H incorporation into immunoprecipitates from reticulocyte membranes (lanes 1-3) and Radioautographs of SDS gels of the total plasma membranes of reticulocytes labeled with 3H-fattyacid or ['Hlacyl- mature cell membranes (lanes 4 and *5).Membranes were incubated CoA are shown in Fig. 2 (left panel). Only a few polypeptides with [3H]tetradecanoic acid minus ATP (lane I), plus ATP (lanes 2 and 4 , or with ['H]tetradecanoyl-CoA (lanes 3 and 5 ) . After incubaare acylated, the major peptide labeled being one with the tion, the membranes were dissolved and immunoprecipitated (see molecular size of thetransferrinreceptor(93kDa).This "Materials and Methods"). polypeptide can be specifically precipitated by a monoclonal antibody to the transferrin receptor and is present in reticu-radioactivity between ['Hltetradecanoyl-CoA and ["]tetralocytes butnot in mature cell membranes (Fig. 2, center decanoic acid. This result suggests that thelevel of acylation panel). in uitro of the transferrin receptor is even higher with acylThe radioautographs of the immunoprecipitates show that CoA than with free fatty acid. The incorporationof fatty acid thetransferrinreceptorisacylated with eitherfatty acid is absolutely dependentonthe presence of ATP (Fig. 2, (+ATP) or acyl-CoA (Fig. 2, right panel). Although a lower compare lanes 1 and 2, right panel). As expected, an acylated intensity on the autoradiogram is observed for the transferrin band a t t h eposition of the transferrin receptor is not evident receptor after acylation with ['Hltetradecanoyl-CoA (lane 3 ) , with mature erythrocyte membrane (Fig. 2, lanes 4 and 5). this change is much less than the20-fold difference in specific Although not shown in this study, labeled palmitate is also incorporatedintothe 93-kDapeptide. However, because higher specific activity of tetradecanoate was available to us, the latterwas used routinely. The rates of incorporation of acyl groups from free fatty acid and acyl-CoA into the transferrinreceptor are compared in Fig. 3. Within 20 min, a near-plateau value for acyl transfer from acyl-CoA to the transferrinreceptor is reached, whereas with "-labeled fatty acid (+ATP) as substrate, only one fifth of the amount of acyl-CoA is incorporated after 2 h (-0.05 pmol/pg of protein). Although some dilution of the labeled compound by the presence of endogenous cold fatty acid in the membranecould account, in part,for the difference in the degree of acylation by the two precursors, the higher relative rate with acyl-CoA suggests that the latter is the preferred Minute6 (i.e. direct) donor of the acyl groups to the transferrinreceptor. FIG.1. Fatty acid incorporation into total membrane proteins. Isolatedreticulocyte membranes were incubatedin 5 mM The total level of acyl moiety incorporation into the transphosphate buffer, pH.7.4, containing [3H]tetradecanoic acid plus (0) ferrin receptor and other membrane proteins after SDS-gel or minus ( 0 )5 mM ATP at 22 "C. Samples were taken a t intervals, electrophoresis was much lower than that predicted from the and incorporation into acid-insoluble, chloroform/methanol-insoluamount of fatty acid incorporatedintoprotein based on Me material was determined. insolubility in organic solvents (Fig. 1).This discrepancy may ' The abbreviations used are: Hepes, 4-(2-hydroxyethyl)-l-pipera- arise from the instability of the linkages between different proteins and the acyl groups during SDS-gel electrophoresis zineethanesulfonic acid; SDS, sodium dodecyl sulfate. In VitroAcylation of the Transferrin Receptor 15462 1 2 - 0 186K- 3 1 4 186K- 2 3 4 -Q- e r) 0 1 I 1 I 30 60 Minute. 90 120 94K- FIG.3. Kinetics of the specific acylation of the transferrin receptor with [‘Hltetradecanoic acid (+ATP) and [‘Hltetradecanoyl-CoA. The reticulocyte membranes were acylated for different lengths of time with [‘H]tetradecanoic acid (+ATP) (0)(21 Ci/mmol) or (3H]tetradecanoyl-CoA (0)(0.76 Ci/mmol), immunoprecipitated with antibody against the transferrin receptor, and electrophoresed on SDS gels as described under “Materialsand Methods.” After radioautography for 7 days, the area corresponding tothe receptor (93 kDa) was excised, and the gel was swollen and digested in 80% protosol and counted. The protein content of the transferrin receptor (-10 pg) was estimated from the amount of protein on the SDS gel at 93 kDa. or from incomplete removal of phospholipids during extraction with organic solvents. Nonspecific binding of fatty acid to the receptor seems unlikely since incorporation of fatty acid into the receptor was clearly dependent on ATP (Figs. 1 and 2). Moreover, the acyl groupsincorporated into the transferrin receptor are hydroxylamine labile at pH. 7.0 (data not given), suggesting that the acyl groups are bound in ester linkages to the receptor. [3H]Tetradecan~yl-AMPdoes not acylate the receptor. Since the transferrin receptor isexcised and released in vesicular form during the maturation of the reticulocytes in vitro (20), it seemed pertinent to ask whether the released receptor contains the “-labeled acyl groups incorporated in vitro. For these studies, the receptor was labeledby incubation of the cells at 37 “C in a medium containing the ‘H-labeled fatty acid. During the long-term incubation (7), the excised vesiclesformedwerecollected and the extent of receptor labeling in the cellular plasma membranes and in the isolated vesicles was determined (Fig. 4). With time in culture, the amount of protein immunoprecipitated with anti-transferrin receptor antibody decreased in the plasma membrane and increased in the released vesicles (Fig. 4, left panel), a result consistent with the idea that thetransferrin receptor protein is transferred from the cell to the vesicles (20). With time, less radioactivity is recovered in the receptor isolated from cellular membranes, whereas the radioactivity in the released vesicles isincreased (Fig. 4, rightpanel) in a manneranalogous to thechanges in the amountof receptor protein in the plasma membranes and vesicles. Moreover, if reticulocytes are prelabeled with [‘Hltetradecanoic acid, washed with unlabeled tetradecanoic acid, and then cultured for 24 h, the receptor recovered in the vesicles is radiolabeled. Vesicles incubated with [“Hltetradecanoate (+ATP)do not incorporate the label. These results show that thereleased receptor is not deacylated and is derived fromthe same pool of receptor that has undergone post-translational modification in the reticulocyte. It is not clear whether all the available receptor can still undergo acylation since the level of acyl groupincorporation in isolated membranes is relatively low (less than 5% labeling of the receptor on a molar basis) (Fig. 3). In intact cells, uncertain- FIG.4. Detection of [‘Hltetradecanoic acid in vesicles formed during in vitro incubation of reticulocytes. Incubations were carried out with a 2% intact cell suspension in which the medium contained 1 p~ [3H]tetradecanoicacid. At 3 and 12 h, 25-ml samples were removed and centrifuged to separate the cells and released vesicles (7,700 X g and 100,000 X g pellets, respectively). The vesicles and the plasma membranes prepared from the cells were washed and solubilized in Triton X-100, immunoprecipitated, and gel-electrophoresed as described under “Materials and Methods.” Left panel, Coomassie Blue stains of the proteins in the immunoprecipitates at 3 and 12 h, respectively, from the vesicles (lanes I and 2) and cells (lanes 3 and 4). Note that the amount of the 93-kDa peptide (transferrin receptor) decreases in the cell membranes (lanes 3 and 4 ) and increases in the vesicles (lanes I and 2). The band a t 186 kDa is the dimer of the transferrin receptor. Right panel, radioautographs of lanes corresponding to left panel. ties about the specific activity of the incorporated fatty acid make computations more difficult to interpret. DISCUSSION The present work shows that acylation of the transferrin receptor occurs in isolated membranes under conditions which cannot support de novo protein synthesis. These data are consistent with the observations of Omary and Trowbridge (2) that in uiuo acylation of the receptor is a post-translational event whichmayoccurmany hours afterthe receptor is synthesized. Thus, both in red cells and growing mammalian cells, the ability to modify the receptor persists long after the protein itself has been synthesized. Except for an ATP requirement, the necessary enzymesand cofactors for acylation of the receptor are present in washed, isolated reticulocyte membranes. Since reticulocytes retain residual intracellular structures, such as mitochondria, the possibility couldbe raised that the acylation is catalyzed by enzymes associated with intracellular organelles which adhere to theplasma membrane during isolation. Current studies are underway to assess whether acylation is due to contaminating enzymes or is a function whichresides in the plasma membrane in sheep reticulocytes. It has been postulated that post-translational acylation might be involved in directing, inserting, or anchoring membrane proteins (4). Although in vitro acylation of membranes is not associated with de novo synthesis of receptor nor its insertion into the membrane, these possibilities cannot be Acylation VitroIn Transferrin of the Receptor 15463 4. Magee, A. I., and Schlesinger,M. J. (1982) Biochim. Biophys. eliminated as only a small number of sites could be acylated Acta 694, 279-289 in the present study. Whether the low level of acylation is 5. Schlesinger, M. J., Magee, A. I., and Schmidt, M. F. G. (1980) J. due to the fact that the remaining sites are acylated is not Biol. Chem. 255, 10021-10024 known. Among the putative functionsfor acylated proteins,a 6. Schneider, C., Sutherland, R., Newman, R., and Greaves, M. (1982) J. Biol. Chem. 257, 8516-8522 common feature is membranefusion (27-29). It is interesting 7. Pan, B. T., Blostein, R., and Johnstone, R. M. (1983) Biochem. in this regard that the transferrin receptor is released in J. 210.37-47 vesicular form, a process which is likely to require membrane 8. Jandl, J. H., and Katz, J. H. (1963) J . Clin. Inuest. 42, 314-326 fusion, and that the released vesicles contain thenewly incor9. Morgan, E. H. (1981) Mol. Aspects Med. 4, 1-123 porated fatty acid. 10. Aisen, P., and Brown, E. B. (1976) Prog. Hematol. 9, 25-56 The fatty acid composition of the transferrin receptor in 11. Trowbridge, I. S., and Omary, M. B. (1981) Proc. Natl. Acad. Sci. U. S. A. 78,3039-3043 situ has not yet been established. Omary andTrowbridge (2) 12. Sutherland, R., Delia, C., Schneider, R., Newman, R., Kemshead, used palmitate to label the transferrin receptor in growing J., and Greaves, M.(1981) Proc. Natl. Acad. Sci. U. S. A . 78, human cells. In the present experiments, tetradecanoate as 4515-4519 well as palmitate were used successfully to label the receptor. 13. Hamilton, T., Wada,G.H., and Sussman, H. H. (1979) Proc. Natl. Acad. Sci. U. S. A. 76, 6406-6410 It is unknown whether thephysiological effectiveness of the J. W., and Cresswell, P. (1979) J. Supramol. Struct. 11, receptor isinfluenced by the natureof the fattyacid attached. 14. Larrick, 579-586 More significant, however, is the demonstration that tetra15. Galbraith, G. M. P., Goust, J. M., Mercurio, S. M., and Galbraith, decanoyl-CoA is a better acyl donorthantetradecanoate R.M. (1980) J. Zrnrnunol. Immuno@d. 16,387-395 (+ATP) for receptor labeling, whereas tetradecanoyl-AMP is 16. Octave, J . N., Schneider, Y. J., Hoffman, P., Truet, A., and Crichton, R. R. (1979) FEES Lett. 108, 127-130 not used (not shown). Thepreference for acyl-CoA over fatty Bockxmeer, F. M., and Morgan, E.H. (1979) Biochirn. acid is in agreement with the specificity observed in the two 17. Van Biophys. Acta 584, 76-83 other eukaryotic cell-free systems ( 3 , 4,30) in which in vitro 18. Frazier, J . L., Caskey, H. J., Yoffe, M., and Seligman, P. (1982) acylation hasbeen studied. The preferred acylation of proteins J. Clin. Inuest. 69, 853-865 19. Kornfeld, S. (1969) Biochim. Biophys. Acta 194, 25-33 with acyl-CoA over that with fatty acid plus ATP is not universal since in luminescent bacteria (31), some proteins 20. Pan, B. T., and Johnstone,R. M. (1983) Cell 33,967-977 21. Benderoff, S., Johnstone, R. M., and Blostein, R. (1978) Membr. are acylatedby fatty acid plus ATP, but are not acylated with Biochem. 1,89-106 thecorresponding CoA derivative.Although itisperhaps 22. Dodge, J. T., Mitchell, C., andHanahan, D. J . (1963) Arch. Bioclaem. Biophy~.100, 119-130 premature to make any conclusions on the preferred route of acylation since few systems have been studied in vitro, the 23. Bishop, J . E., and Hajra, A. K. (1981) J.Biol. Chern. 256, 95429550 relative efficacy of the CoA derivative over fatty acid plus 24. Rodriguez, A., Wall, L., Riendeau, D., and Meighen, E. (1983) ATP or acyl-AMPsuggests that membrane protein acylation Biochemistry 22,5604-5611 in eukaryotes may follow the acylation route established for 25. Laemmli, U. K. (1970) Nature (Lond.)227, 680-685 26. Omary, M. B., and Trowbridge, I. S. (1981) J. Biol. Chern. 256, the majority of low-molecular-weight compounds. REFERENCES 1. Johnstone, R. M., Adam, A,, Turbide, C., and Larrick, J . (1984) Can. J. Biochern. Cell Biol. 62, 927-934 2. Omary, M. B., and Trowbridge, I. S. (1981) J. Biol. Chem. 256, 12888-12892 3. Marinetti, G. V., and Cattieu, K. (1982) Biochim. Biophys. Acta 685, 109-116 4715-4718 27. Schmidt, M. F. G. (1982) Virology 116, 327-338 28. Scheid, A., and Choppin, P. W. (1974) Virology 57, 475-490 29. Huang, R. T. C., Wahn, K., Klenk, H.-D., and Rott, R. (1980) Virology 104, 294-302 30. Berger, M., and Schmidt, M. F. G. (1984) J . Bid. Chem. 259, 724512.52 _" 31. Wall, L., Rodriguez, A., and Meighen, E. (1984) J . Biol. Chern. 259, 1409-1414 ~~~