Infection of B Lymphocytes by a Human Herpesvirus, Epstein-Barr Virus, Is Blocked by Calmodulin Antagonists (Endocytosis/Phenothiazines/Trifluoperazine/R24571) GLEN R

Proc. Natl. Acad. Sci. USA Vol. 81, pp. 4955-4959, August 1984 Medical Sciences

Infection of B lymphocytes by a human herpesvirus, Epstein-Barr virus, is blocked by calmodulin antagonists (endocytosis/phenothiazines/trifluoperazine/R24571) GLEN R. NEMEROW AND NEIL R. COOPER Department of Immunology, Scripps Clinic and Research Foundation, 10666 North Torrey Pines Road, La Jolla, CA 92037 Communicated by Hans J. Muller-Eberhard, April 23, 1984

ABSTRACT Epstein-Barr virus (EBV) is a human her- regulated by the intracellular calcium-binding protein calmo- pesvirus that selectively binds to and infects human B lympho- dulin (7, 9, 12-14). In pursuing our studies of the mecha- cytes (B cells). In the studies presented here, we found that nisms involved in the early stages of infection of human B several phenothiazines, including trifluoperazine, chlorpro- lymphocytes by EBV, we have examined the ability of sev- mazine, prochlorpromazine, and promethazine, blocked EBV eral classes of calmodulin antagonists to block EBV infec- infectivity of isolated adult human B cells as measured either tion and determined the stage at which the infection process by outgrowth of transformed cell colonies-or by [3ll~thymidine is altered. incorporation. Trifluoperazine, chlorpromazine, and prochlorpromazine were equally effective with 20 ,uM fully inhib- MATERIALS iting infectivity, whereas 100 ,AM promethazine was required EBV was isolated from the B95-8 cell line as described (15). for a comparable effect. Inhibition by trifluoperazine was par- Radiolabeled EBV was obtained by addition of 25 /iCi (1 Ci tially reversible. Studies with radiolabeled EBV demonstrated = 37 GBq) of [35S]methionine (New England Nuclear) per that the inhibitors did not impair virus binding to B cells. Elec- ml to B95-8 cells prior to virus purification as described by tron microscopic examination of B lymphocytes revealed that Edson and Thorley-Lawson (16). Inhibitors were obtained trifluoperazine reduced the number of large uncoated cell vac- from the following sources: trifluoperazine (Stelazine) and uoles and the number of membrane microvilli, indicating that R24571 (Calmidazolium), from Boehringer Mannheim; this agent interfered with cell pinocytosis. This process was chlorpromazine (Thorazine) and prochlorpromazine (Com- accompanied by inhibition of EBV endoeytosis iQto B cells. pazine), from Smith Kline & French; haloperidol (Haldol), Phenothiazines bind to and inhibit calmodulin, an intracellu- from McNeill Pharmaceutical (Spring House, PA); phenothi- lar calcium-binding protein that regulates several key en- azine, from Sigma; and promethazine (Phenergan), from zymes, some of which directly affect cytoskeletal elements, Wyeth Labs (New York, NY).' [methyl-3H]Thymidine was although they alsQ may interact nonspecifically with other cel- purchased from Amersham. lular constituents. In this regard, haloperidol, a non-phenothi- EBV Infectivity and Drug Inhibition Assays. Adult human azine calmodulin antagonist, and R24571, a derivative of the B lymphocytes 'were isolated from normal laboratory per- antimycotic miconazole, which is a potent and highly specific sonnel as described (1). Viral infectivity was assayed by calmodulin inhibitor, also blocked EBV infection. These stud- stimulation of DNA synthesis in normal B cells as measured ies suggest that calmodulin or a calmodulin-regulated cellular by incorporation of [3H]thymidine 14 days after infection enzyme(s) is involved in normal cellular endocytic processes in (17) and by outgrowth of transformed B-cell colonies, which B lymphocytes and thereby in the early stages of EBV infec- were enumerated 7-14 days after infection by phase-contrast tion. light microscopy (15). Stock solutions of the drugs, usually at 1-10 mM, were prepared in sterile saline or RPMI medium Recent studies in our laboratory have shown that Epstein- containing 2% dimethyl sulfoxide in plastic tubes immediate- Barr virus (EBV) infects normal human B cells by an endo- ly before use. For most drug inhibition studies, 2 x 105 B cytic pathway that is somewhat different (1) than the clath- cells in 100 ,x1 were preincubated with varying amounts of rin-receptosome-lysosome pathway used by many other the calmodulin.antagonists, generally at 1:500 or 1:1000 of protein ligands (2) as well as several viruses (3) to enter cells. the stock (1-100 .aM) for 60 min at 370C, after which purified One difference in the endocytic pathway of EBV in normal B EBV was added to the cells at a multiplicity of infection of cells is the lack of participation of clathrin-coated pits and -1.0. The cell cultures were incubated for 14 days,' at'which vesicles in the infectious process. The studies presented here time the cultpres were examined for colony formation and were undertaken to further examine the early events in EBV analyzed'for'[3H]thymidine incorporation (15). Controls in- infection and particularly the' mechanisms involved in EBV cluded cells prdtreated with RPMI medium/dimethyl sulfox- endocytosis into normal B lymphocytes. One element that ide (final concentration = 0.02%) alone. Cell viability, deter- provides an early signal for me'mbrane protein redistribution mined by trypan blue dye exclusion, was also assessed fol- (capping) and cell activation processes in lymphocytes is in- lowing incubation with each drug. The presence of the drugs tracellular calcium (4-8).. Caltcium ions markedly influence did not affept cell viability or the rate of [3H]thymidine incor- many biological processes in lymphocytes, including cap- poration in'B lymphocytes previously transformed by EBV. ping but not patching, cytoplasjnic streaming, endocytosis, Binding of EBV to B Cells. Attachment of purified EBV to and exocytosis (4, 9-11). The effects of calcium on several B cells in the presence and absence of calmoduliq antago- important cell enzymes such as mefhbrane-associated Ca2+, nists was assessed by binding of [35S]methionine-labeled Mg2+-ATPase, phosphodiesterase, and myosin light chain EBV (35S-EBV) to cells. B cells (1 x 106) in 100 1.l' of RPMI kinase implicated in the above noted cellular activities are medium containing 0.2% bovine serum albumin were incubated with varying amounts of the inhibitors at 372C prior to the addition of 3000 cpm of 35S-EBV in 10 /d. Nonspecific The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. Abbreviation: EBV, Epstein-Barr virus. 4955 Downloaded by guest on September 29, 2021 4956 Medical Sciences: Nemerow and Cooper Proc. Natl. Acad Sci. USA 81 (1984)

Table 1. Inhibitors of EBV infectivity Cb x Concentration for a) E CL 50% inhibition of CL-0 EBV infection, AuM £- Colony DNA A2 CD Group Drug formation synthesis C-) Ca CD Phenothiazine Prochlorperazine 5.0 2.5 -E E ._ Chlorpromazine 5.0 2.5 To coo -4- Trifluoperazine 6.0 7.0 Promethazine 40.0 35.0 Butyrophenone Haloperidol 90.0 50.0 Miconazole R24571 1.5 2.0 Inhibitor (MM) neously with EBV. Further experiments were carried out to FIG. 1. Inhibition of EBV infectivity by calmodulin antagonists. determine whether inhibition of EBV infectivity was revers- B cells were incubated with various amounts of trifluoperazine (0), ible. In these studies, trifluoperazine was incubated with the R24571 (A), or phenothiazine (v) and then infected with purified cells for 30 min at 370C, after which the cells were washed EBV. Fourteen days after infection, B-cell cultures were assessed prior to the addition of EBV. As shown in Fig. 2, inhibition for colony formation (closed symbols) and for ['Hithymidine incor- of viral transformation by trifluoperazine was partially re- poration (open symbols). Each point represents the mean ± SEM of triplicate cultures. versible, with 50-60% of infectivity being restored after several wash steps. Trifluoperazine failed to inhibit infectivity if added after the cells and EBV had been in contact for 5 min binding obtained heat-inactivated aver- levels, with virus, at 370C, suggesting that the drug was at a stage levels. acting early in aged 5-10% of the specific binding After further incu- the internalization process. bation at 40C for 60 min, the time of maximal binding at 40C Effect of Phenothiazines and Other Calmodulin Antagonists the cell were (or 370C), samples microcentrifuged (Beckman) on EBV Attachment. To determine whether the calmodulin a cushion of 5% the cell and through dextran T-10, pellet inhibitors interfered with EBV attachment, B cells were supernate, after solubilization in scintillation fluid (Cyto- treated with varying doses of the drugs and subsequently in- scint, San Diego, CA), were assayed for radioac- separately cubated with purified 35S-EBV. As shown in Fig. 3, none of in an LS-8000 p-scintillation counter and tivity (Beckman), the phenothiazines or non-phenothiazine calmodulin antago- the amount of bound labeled virus associated with the B cells nists (Table 2) altered the binding of radiolabeled virus at was determined. doses that significantly restricted viral infectivity. These Electron Studies. B 5 x 105 in 100 of Microscopic cells, A.l studies suggested that the drugs were acting at a stage subse- medium, were preincubated with medium alone or with 20 quent to viral attachment. ,.M trifluoperazine prior to incubation with purified EBV for Electron Microscopic Studies of EBV Internalization. Elec- 60 at 40C. The cells were then warmed to min rapidly 370C tron microscopic studies were performed to analyze the ef- for and 60 to initiate viral and 5, 10, 15, 30, min endocytosis fect of calmodulin antagonists on viral internalization. In then fixed in modified Karnovsky fixative and processed for these studies, B cells were pretreated with 20 tLM trifluoper- transmission electron microscopy as described (1). At least azine for 30 min at 40C, next incubated with EBV at 40C to 50 randomly selected equatorial thin sections of EBV-infect- allow attachment, and then warmed to 370C for 5-, 10-, 15-, ed cells were examined and the location of virus on the outer 30-, and 60-min intervals to initiate endocytosis. As depicted plasma membrane (smooth surface or or in- invaginations) in Fig. 4, relatively few (-10%) of the EBV particles were side uncoated vacuoles (receptosomes) was noted. An aver- observed in endocytic vacuoles of drug-treated cells, where- age of 100 virions for each time in point both the drug-treated as 80% of the virions were inside uncoated vacuoles in the and untreated cells was counted. untreated B cells. No difference in the lack of EBV endocy- RESULTS tosis was noted in the drug-treated B cells between 5 and 60 Inhibition of Infectivity Calmodulin Initial by Antagonists. X 40- -40 studies showed that trifluoperazine, a phenothiazine calmodulin antagonist, blocked EBV infection of B cells in a dose- 35- -35 U, dependent manner as measured by outgrowth of transformed CA 30- -30 =. cell colonies or by stimulation of DNA synthesis (Fig. 1). = 25- -25< percent was Fifty inhibition achieved with -5 ,AM trifluoper- _ 2020- -20 m x azine, whereas inhibition was nearly complete at 11 AtM. = Several other phenothiazines, including prochlorperazine, = 15- -15 = = chlorpromazine, and promethazine, also blocked EBV infec- 0 10- -10,; tion (Table 1), whereas the unsubstituted parent compound, c D 5 phenothiazine, lacked inhibitory activity at the highest con- C-, centration of drug tested (Fig. 1). In other experiments, 20 A B C D E ,M phenothiazine failed to inhibit infectivity (data not shown). Haloperidol, a butyrophenone, also inhibited infec- FIG. 2. Reversibility and time course of EBV inhibition by tri- tivity, although it was approximately 1/5th to 1/10th as fluoperazine. B cells were incubated with medium alone (A) or with effective as trifluoperazine in blocking infection. By con- 20 uM trifluoperazine (B) during infection with EBV and assayed for colony formation or trast, R24571, a highly specific calmodulin inhibitor (18), (hatched bars) ['H]thymidine incorporation (open bars). In parallel studies (C), B cells were washed after treat- was 5 times more potent than trifluoperazine (Fig. 1); com- ment with trifluoperazine prior to infection with EBV. B cells were plete inhibition of EBV infection was achieved with 3 ,uM also incubated with EBV at 4'C for 60 min (D) or at 4'C for 60 min R24571. Inhibition of viral the infectivity by above antago- and then at 37'C for 5 min with EBV (E) prior to addition of 20 /LM nists also occurred when these agents were added simulta- trifluoperazine. Downloaded by guest on September 29, 2021 Medical Sciences: Nemerow and Cooper Proc. NatL. Acad Sci. USA 81 (1984) 4957 280- EXTERNAL CELL MEMBRANE ENOOCYTIC 0% , %s -.0S o 100 VACUOLES I %-4C. . 200- 1-4.cr1-- -.3-- %J '0 b- 80 O. _ 120- '- 60- E Ca 40- .0 .2 40- a 140- co a 20 H co 100- (Smooth Surface) (Invaginations) 60- FIG. 4. Cellular distribution of EBV in B cells treated with trifluoperazine. B lymphocytes were preincubated with medium alone (open bars) or with 20 ,uM trifluoperazine (hatched bars) prior to incubation at 40C for 60 min with EBV. After warming the cells to Ii l0 io io 4o 370C for 10 min, the cells were prepared for transmission electron Phenothiazine microscopy and 50 randomly selected sections of EBV-infected cells (pM) were examined for the location of virus particles. FIG. 3. Effect of phenothiazines on attachment of EBV to B cells. B cells were preincubated with varying amounts oftrifluopera- cles (Figs. 4 and 5); attachment of EBV to cell surface viral zine (o---o), chlorpromazine (o---o), promethazine ( e-), or pro- receptors was unaffected (Fig. 3; Table 2). Previous studies chlorperazine (A -) prior to incubation with 35S-EBV for 60 min at 370C. Nonspecific binding, obtained with heat-inactivated 35S-EBV, have shown that the EBV endocytic pathway in normal B was 18 cpm. Binding of "S-EBV was reduced to 10-20 cpm in the lymphocytes proceeds exclusively via internalization into presence of human immune serum to EBV. such vesicles and does not involve clathrin-coated pits and vesicles or fusion of endocytic vesicles with lysosomes (1), min at 370C. Representative electron microscopic sections of in sharp distinction to the clathrin-receptosome-lysosome EBV-infected B cells at 10 min at 37TC are shown in Fig. 5. pathway utilized by a number of protein ligands (2), includ- EBV was found in membrane invaginations or inside vacu- ing several viruses (3). oles in the absence of trifluoperazine (Upper Left and Lower The drugs capable of retarding EBV infection, although of Left), whereas EBV was exclusively on external smooth different chemical structures, are known to be potent inhibi- membrane surfaces in identical reaction mixtures containing tors of the intracellular calcium-binding protein calmodulin 20 uM trifluoperazine (Upper Right and Lower Right). The (9, 11, 19). Inhibition of calmodulin-mediated intracellular electron micrographs also revealed an overall decrease in the events follows the firm binding of such agents to calcium- number of cell membrane microvilli and in the number of calmodulin complexes (11, 19, 20). Although the present large uncoated cell vacuoles in cells treated with trifluopera- studies have not directly shown calmodulin to be involved in zine, irrespective of the presence or absence of EBV. Thus, endocytosis, and thus in EBV internalization, this is proba- an average of 6 large (:300 nm) vacuoles was noted in equa- ble for the following reasons: (i) R24571, a highly specific tonally sectioned B cells in the presence of the inhibitor calmodulin antagonist, which, unlike trifluoperazine, does compared to 20 such vacuoles in its absence. These studies not bind to a-adrenergic receptors (18), also blocked EBV indicate that trifluoperazine blocks EBV infectivity by inhib- infectivity (Table 1; Fig. 1); (ii) the concentrations of the iting endocytosis. phenothiazines trifluoperazine and chlorpromazine required to inhibit EBV infection (Table 1) are similar to those report- DISCUSSION ed to inhibit other calmodulin-dependent activities (11, 19- 22); (iii) promethazine, a phenothiazine, and haloperidol, a In these studies, several phenothiazines and non-phenothi- butyrophenone, were considerably less efficient than trifluo- azines inhibited the infection of B lymphocytes by EBV. perazine in inhibiting both EBV infection (Table 1) and cal- These agents were found to inhibit infection by retarding in- modulin-dependent activation of phosphodiesterase (21, 23); ternalization of EBV into large thin-walled endocytic vesi- and (iv) the limited time frame for expression of drug-in- duced inhibition (Fig. 2) was similar to comparable findings Table 2. Effect of non-phenothiazine calmodulin antagonists on for other calmodulin-dependent reaction. These aspects ren- EBV attachment der it likely that the actions were mediated through calmodu- Concentration, "S-EBV bound, lin and not via one of the reported nonspecific actions of Drug 1M cpm* phenothiazines on cells (23, 24). The mechanisms involved in receptor-mediated endocyto- Haloperidol 200 293 ± 50 sis in nucleated cells are complex and involve several dis- R24571 3 270 ± 28 tinct reaction sequences (4) and at least two endocytic path- None (control) 202 ± 72 ways (25, 26). In addition to the clathrin-receptosome-lyso- Isolated B cells (1 x 106) were treated for 60 min at 370C with an somal endocytic pathway that has been studied extensively amount of each antagonist that resulted in complete inhibition of in cultured cell lines (2, 3), there exists another endocytic EBV infectivity. cpm values are given as mean ± SEM. pathway in lymphocytes and other cells (12, 25-30) that does *Input of "S-EBV was -3000 cpm. Eighty-five percent binding of not involve clathrin-coated vesicles but, rather, is character- the labeled virus could be achieved with 2 x 107 Raji lymphoblas- formation toid cells with this virus preparation. Nonspecific binding, ob- ized by membrane internalization and the random tained with heat-inactivated 35S-EBV, was 18 cpm. Binding of 35S- of large thin-walled vesicles (1, 12, 25). This pathway, which EBV was reduced to 10-20 cpm in the presence of human immune appears to be the one utilized by EBV to infect B lympho- serum to EBV. cytes, probably represents a general pinocytic mechanism Downloaded by guest on September 29, 2021 4958 Medical Sciences: Nemerow and Cooper Proc. NatL Acad. Sci USA 81 (1984)

S j. j"4 ~ ~ ~

1howt s _ t5Z^*x,'\ .<,s,6,t~~~~~~~~A!:*ti..eW?*i,./r Us}byXo. I2s

*: , ., " r ..:f . ..

a~~~~~~~~~~~~~~~~~~~~~bl &

:-: , . ; L~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Z

Al '

FIG. 5. Electron micrographs of EBV-infected B cells incubated in medium alone (Upper Left and Lower Left) or with 20 AtM trifluoperazine (Upper Right and Lower Right). Arrows indicate the location of virions. (Bars = 1 tum.)

common to many cells. These actions require movement of of intracellular calcium levels is followed by receptor-ligand cellular structures and organelles and most likely involve en- clustering, which, in the case of membrane immunoglobulin, zymatic reactions and other energy-dependent processes. is accompanied by redistribution of cytoplasmic myosin be- Phenothiazines have been observed previously to inhibit this neath the membrane-receptor ligand complexes (32-34). pathway (12, 25), as also observed in the present study, indi- The recruitment of myosin and perhaps other cytoskeletal cating that it also is apparently calmodulin dependent. elements, including tubulin (22) and a-actinin (35), to the site Endocytosis, in certain situations, is preceded by mem- of receptor aggregates probably brings about the contractile brane protein redistribution, patching, and capping. Cross- force necessary for endocytosis. The redistribution of myo- linking of membrane receptors is associated with mobiliza- sin during capping has been shown to be blocked by calmo- tion of intracellular calcium in murine B lymphocytes. The dulin antagonists (10) and calmodulin has been found to re- increase of calcium levels has been postulated to provide an distribute with caps of various types in lymphocytes (5, 7, early stimulus for contractile activity (4, 31). Thus, the rise 13, 14). Endocytosis of immunoglobulin complexes, conca- Downloaded by guest on September 29, 2021 Medical Sciences: Nemerow and Cooper Proc. NatL. Acad Sci USA 81 (1984) 4959 navalin A, and a2-macroglobulin are also inhibited by pheno- 7. Nelson, G. A., Andrews, M. L. & Karnovsky, M. J. (1982) J. thiazines (12, 13, 36). Taken together, these observations Cell Biol. 95, 771-780. suggest that one pathway of endocytosis involves a series of 8. Klausner, R. D., Bhalla, D. K., Dragsten, P., Hoover, R. L. events initiated by a rapid increase in intracellular calcium & Karnovsky, M. J. (1980) Proc. Natl. Acad. Sci. USA 77, 437-441. brought about by receptor cross-linking, which leads to acti- 9. Cheung, W. Y. (1980) Science 207, 19-27. vation of calmodulin at the cytoplasmic cell surface: activa- 10. Braun, J., Fujiwara, K., Pollard, T. D. & Unanue, E. R. (1978) tion of this protein allows activation of one of several poten- J. Cell Biol. 79, 419-426. tial cellular enzymes, such as myosin light chain kinase, 11. Weiss, B., Prozialeck, W., Cimino, M., Barnette, M. S. & whose activity is regulated by calmodulin (5, 37, 38). The Wallace, T. (1980) Ann. N. Y. Acad. Sci. 356, 319-345. activation of this enzyme then generates functionally active 12. Salisbury, J. L., Condeelis, J. S. & Satir, P. (1980) Ann. N. Y. myosin molecules capable of interacting with actin microfil- Acad. Sci. 356, 429-432. aments. Thus, calmodulin activation might be the key step 13. Salisbury, J. L., Condeelis, J. S., Maihle, N. J. & Satir, P. for physically linking cell membrane receptor-ligand com- (1981) Nature (London) 294, 163-166. 14. Bourguignon, L. Y. W. & Balazovich, K. (1980) Cell Biol. Int. plexes to the cytoskeletal system or alternatively by activat- Rep. 4, 947-952. ing the contractile systems of the cell. 15. Nemerow, G. R. & Cooper, N. R. (1981) J. Immunol. 127, The present studies indicate that calmodulin or calmodu- 272-278. lin-dependent processes are involved in the earliest stages of 16. Edson, C. M. & Thorley-Lawson, D. A. (1981) J. Virol. 39, infection of normal human cells by a member of the herpes- 172-184. virus family, EBV. Few therapeutic agents are available for 17. Robinson, J. & Miller, G. (1975) J. Virol. 15, 1065-1072. the treatment of infection by herpesviruses. Calmodulin an- 18. Van Belle, H. (1981) Cell Calcium 2, 483-494. tagonists, including several phenothiazines and haloperidol, 19. Weiss, B. & Levin, R. M. (1976) Adv. Cyclic Nucleotide Res. are widely used in clinical medicine as tranquilizers and anti- 9, 285-303. 20. Levin, R. M. & Weiss, B. (1977) Mol. Pharmacol. 13, 690- psychotic agents and thus their actions and side effects are 697. known. Since the classes ofdrugs studied here varied in their 21. Levin, R. M. & Weiss, B. (1977) J. Pharmacol. Exp. Ther. effectiveness in blocking EBV infection and simple substitu- 208, 454-459. tions markedly influenced activity in this regard (Table 1), 22. Yahara, I. & Kakimoto-Sameshima, F. (1978) Cell 15, 251-259. other types of agents or various chemical alterations may 23. Roufogalis, B. D. (1981) Biochem. Biophys. Res. Commun. 98, represent or yield even more potent anti-viral agents. Identi- 607-613. fication of the process or enzyme(s) regulated by calmodulin 24. Corps, A. N., Hesketh, T. R. & Metcalfe, J. C. (1982) FEBS and involved in EBV endocytosis may also provide other Lett. 138, 280-284. approaches to blocking the early stages of infection. Such 25. Salisbury, J. L., Condeelis, J. S. & Satir, P. (1980) J. CellBiol. 87, 132-141. investigations may provide useful therapeutic agents for con- 26. Huet, C., Ash, J. F. & Singer, S. J. (1980) Cell 40, 429-438. trolling herpesvirus infections in man. 27. Taylor, R. B., Duffus, N. P. H., Raff, M. C. & de Petris, S. (1971) Nature (London) New Biol. 233, 225-229. We thank Mary Ellen McNaughton for her expert technical assist- 28. Steinman, R. M., Millman, I. S., Muller, W. A. & Cohn, Z. A. ance and Bonnie Weier for her assistance in preparation of the (1983) J. Cell Biol. 96, 1-27. manuscript. This research was supported by National Institutes of 29. Tycko, B. & Maxfield, F. R. (1982) Cell 28, 643-651. Health Grants Al 17354 and CA 14692 and by a Leukemia Society of 30. Griffith, G. R. & Consigli, R. A. (1984) J. Virol. 50, 77-85. America Special Fellowship. 31. Pozzan, T., Arslan, P., Tsien, R. Y. & Rink, T. J. (1982) J. Cell Biol. 94, 335-340. 32. Braun, J., Fujiwara, K., Pollard, T. D. & Unanue, E. R. (1978) 1. Nemerow, G. R. & Cooper, N. R. (1984) Virology 132, 186- J. Cell Biol. 79, 409-418. 198. 33. Bourguignon, L. Y. W. & Singer, S. J. (1977) Proc. Natl. 2. Goldstein, J. L., Anderson, R. G. W. & Brown, M. S. (1979) Acad. Sci. USA 74, 5031-5035. Nature (London) 279, 679-685. 34. Bourguignon, L. Y. W., Tobuyasu, K. T. & Singer, S. J. 3. Helenius, A. & Marsh, M. (1982) Ciba Symp. 92, 59-76. (1978) J. Cell. Physiol. 95, 239-257. 4. Braun, J., Sha'afi, R. & Unanue, E. R. (1979) J. Cell Biol. 82, 35. Geiger, B. & Singer, S. J. (1979) Cell 16, 213-222. 755-766. 36. Cheung, R. K., Grinstein, S. & Gelfand, E. W. (1983) J. Im- 5. Kerrick, W. G. L. & Bourguignon, L. Y. W. (1984) Proc. munol. 131, 2291-2295. Natl. Acad. Sci. USA 81, 165-169. 37. Adelstein, R. S. (1983) J. Clin. Invest. 72, 1863-1866. 6. Bourguignon, L. Y. W. & Kerrick, W. G. L. (1983) J. Membr. 38. Adelstein, R. S. & Klee, C. B. (1981) J. Biol. Chem. 256, Biol. 75, 65-72. 7501-7509. Downloaded by guest on September 29, 2021