B~,Tchimicaet Biot~'sica Acta, 1136(1992)28-34 © 1992EIs~'/erSe/e:mePub~he~ B.V.All fightsrcser,'ed0167-4889/92/$05.00
28
BBAMCR 13207
Monoclonal antibodies against defmed epitopes of the human transferrin receptor ~3'toplasmic tail S u h a i l a W h i t e a, K a r e n M i l l e r b, C o l i n H o p k i n s b a n d l a n S t u a r t T r o w b r i d g e ~' Deparmwat of Cancer BioDD" The Sulk Institutefor Bio4ogicalStudies, San Diego. CA ¢USA~and b iXopartmentof Biochemirl,~:. Iw47enalCoYlege,London eUIO
(Received II Febma~"]992) Key.words: Transferrinreceptor,Cytoplasm/c~
g~noclonalantiS~d~ F_.~qa~sis
Three murine monnelonal antibodies (mAbs) against the 61-residue amino-terminai cytoplasmic tail of the human transfcrrin receptor (TR) have been produced by immunization of mice with recombinant human TR produced in a baculovims expression system. Mutant human TRs expressed in chick embryo fibroblasts (CEFs) with poim mutations or deletions in their cytoplasmic tails have been used to map the epitopes defined by each of the mAbs. One mAb, H68.4, previously shown to bluek receptor /nternalizat/on, biod~,proximal m the carbox,/-tcrmioal side of the YTRF interoai;Tation signal of TR. The second mAb, H73.2, binds near to the carboxy-tcrmioal side of the H68.4 epitope, whereas the third mAb, 160.1, binds closer to the transmembrane region. H68.4 and H73.2 are auto-antibodies congistent with their epitopes mapping to a region of the human TR that has an identical amino acid sequence to the mouse TR. All three mAbs crossreact with the o'toi~asm/c tm'l of Chinese hamster TIL Double labeiiiog of recombinant human TRs on chick embt3,o fibroblast (CEF) cell membrane preparations with B3/35 and H68.4 antibody-gold conjugates established that receptors in clathrin-coated pits were not labeled with H68.4, implying that assne/ated coated pit proteins may block binding of this mAb.
latreductiea Transferrin receptors (TR) bind the serum transport protein transfenin (TO and mediate uptake of iron into the cell (reviewed ha Refs. 1 and 2). The receptor is a homodimeric type !1 membrane protein consisting of two identical approx. 95-kDa subunits covalently linked by two intermoleealar disulfide bonds [3]. The primary structures of human, mouse, and chicken TRs have been deduced from sequencing of their respective eDNAs and are similar [4-6]. The human TR has an extracellular domain of 671 amino acids, a single 28-residue transmembrane region and a 61-residoe amino* terminal cytoplasmic domain. The TR is a member of the class of ligand transport receptors that are constitutively clustered in coated pits and are rapidly internalized [7,8]. A tetrapepfide sequence, YXRF, in the cytoplasmic tail of the TR that is predicted to be a tight turn has been identified as the recoguhion motif for high-efficiency endocytosig [9]. mAbs have been obtained that react with the external domains ef human, m u s e , rat and chick TRs [J0-13] and have been Correspondenceto: I.S.Trowbridge,Departmentof Cancer Biology, The ~alk Institutefor BiologicalStudies, p.o. Box85800. San Diego, CA 92186-5800. USA.
useful in characterization of the receptor and studies of its fm'~ion [14-16]. Some anti-TR mAbs, either singularly or in combinatiou, block Tf-mediated iron uptake and, as a consequence, inhibit ce~ growth [1719]. Such antibodies are potential therapeutic agents iF: the treatment of cancer [19,20]. Here we report tk,e characterization of three novel mAbs against the cytoplasmic tail of the human TIL The epitopes recognized by each of the antibodies have been mapped by. determining their reactivity with mutant human TRs having cytoplasmic tail deletions or single amino aciO .¢ubstitutions, mAbs reactive with the external domain of the TR are ILl,lily highly species-specific. In contrast, antiTR cytoplasmic tail mAbs show broad species crossreactivity and, unusually, two of the mAbs are auto*antibodies. One anti-TR c~oplasmic tail mAb, H68.4, binds selectively to the cytoplasmic tails of receptors found outside clathrin-coated pits in chicken embryo fibroblast (CEF) plasma membranes prepared from cells expressing recombinant human TRs. Materials and Methods Preparation o f m,4bs
Recombinant human TR was produced in a baculovirus expression system [21] and purified on a hu-
29 man "If affinity column [19,21]. Anti-TR mAbs were obtained by immunization of B a l b / c mice with purified receptor glycoprotein and hybridomas generated exactly as described previously [19]. Hybridomas producing anti-TR mAbs were identified by an E L I S A assay for reactivity to recombinant h u m a n T R and cloned by limiting dilution [19]. Isotyping of mAbs was performed using an E L I S A isotyping kit according to the manuf a c t u r e r s instructions (Mono A b - I D l E A Kit; Zymed Laboratories, San Frandsc~, CA). I-t68.4 and 160.1 are l g G I m A b s and H73.2 is a n l g G 3 mAb.
C C R F - C E M [19], were cultured in RPM11640 medium supplemented with 8 % defined calf bovine serum.
CEF expressing mutant human TRs Mutant human T R s were expressed in C E F using the helper-independent retroviral vector, BH-RCAS, as described previously [22]. The construction, expression, and properties of all the m u t a n t h u m a n T R s used in the present study have been described in detail previously [9,22].
Biochemical procedures Cells and cell culture C / E ClEF were obtained from SPAFAS (Nor.rich, CT) and grown in Duibecco's modified Eagle'z medium ( D M E M ) s u p p l e m e n t e d with 1% ( v / v ) chicken serum, 1% ( v / v ) defined calf bovine ~ r u m (Hyclone, Logan, UT), 2 % ( v / v ) ttyptose broth (Difco, Detroit, MI) [22]. Chinese hamster ovary ( C H O ) cells and mouse L cells were cultured in D M E M supplemented with 8 % defined calf bovine serum. H u m a n T leukemic cells,
Cell-surface iodination, immunoprecipitation studies, and SDS-polyacrylamide gel electropboresis were performed as described previously [21], except that C E F were sufface-iodinated as monolayers. Briefly, confluent C E F cultures in 10-cm tissue-culture dishes were iodinated with 1 mCi N a ~ ! (Amersham, Arlington Heights, IL., approx. 16 m g / p . g ) in 2.5 ml of 0.15 M NaCi/0.01 M sodium phosphate buffer (pH 7.2) (PBS) by the lactoperuxidase-glucose oxidase technique
~.~, _.V. ~'1, ~,~~¢oyV.~q, ",. b
|
,
Wt
|
,
~-r~
Fig. i. mAbs against the cytoplasmicdomain of the human transferrin receptor. (a) Flow cytofluorimetricanalysis of the binding of anti-human TR mAbs t~ CCRF-CEM ceils. Human leukemic T cells, CCRF-CEM (2-106) were stained at 4°C with 100/~1 of hybridoma tissue culture supematant followed by saturating amounts of fluoresceinated goat a n t i - ~ immanoglobulinas the second stage antibody [19]. mAb 133/25 was used as a positive couti~ [or ~[~ biudiBg;the uegative corttrol was tissue culture medium without mAb. (b) lmmunupreeipitation of wikltype and "taille~" (d3-59) mutant TRs by anti-TR mAbs. CEF expressing either wildtype or "tailless"(A3-59) mutant TRs were sufface-indinated and cell b'sates were prepP.,~d as ~ in Materials and Methods. Equal volumes of cell lysate were then immanoprecipitated with anti-TR mAbs and analyzed on a 7.5% SDS-polyacrylamide gel under reducing conditions, mAb B3/25 was used as a positive control. The autoradiograph was exposed for 18 h.
3O [23]. Ceils were lysed in 1 ml of i% Nonidet P-40 in PBS and 50-100 /tl of cell lysate used for immunoprecipitation with 50 /zl of hyb~idoma tissue culture supernatant. Antibody-antigen complexes were then adsorbed to fixed S. aureus precoated with rabbit anti-w~J~,e immunoglobulin as described [21]. Samples Were subjected to electrophoresis on 7.5% SDS-polyacrylamide gels under reducing condi_~-:ns.
wm-~
IOeO,
o--
wcn-~qln~
,~
Immuno gold labeling ~ttuties And-human TR n~.bs, 133/25 and H68.4, were complexed to colloidal gold and used to decorate CEF plasma membranes prepared by a surface repi;ca technique as previously described [24]. Cells were incubated with 6 nm 133/25 gold complexes and then the upper plasma membrane was removed to a coverslip exactly as described previously [24]. The stripped membranes were then fixed in 2% paraformaldehyde, quenched and incubated with H68A 3-nm gold complexes as if they were cryosections [25]. Control sections were pre-incubated with free antibody and nonspecific labeling under these conditions was negligible. Finally, still adherent to the coverslip, the stripped membranes were critical point dried, rotary shadowed and whole mounts were prepared for electron mi~oseopy as desen'bed previously [24].
A3-~
w~'ype
~3-3s
Fig. ~ Re~:livi~" of anti-human TR cytoplasmic tail mAbs with mutant human TRs containing cytoplasmic tail deletions. CEF expressing,Aildq.~ or mutant human TRs with either residues 3-28 (A3-28) or residues 3-35 ~3-35) deleted were surface-iodinated. cell II.~ates ¢,ere prepared, and immunoprecipitationsperformed ~ith anti-humanTR mAbs. as described in the legend to Fig. I. B3/25 mAbwas used as a positivecontro!and tissueculture medium v,ithout mAb was used as a negative control. The autoradiograph wasexposedfor 3 da],~s. man TRs. In contrast, the other three mAbs immunoprecipitated wildtype TRs but not the "tailless" mutant receptors, clearly establishing that they were directed against epitopes located in the cytoplasmic tail of the human TR.
Flow cytometry Binding of anti-human TR mAbs to the human T leukemic cell line CCRF-CEM was d e w ,~d by fluorescence-activated cell sorting anal,_ on a Los Alamos design flow cytometer as pre~ ~sly described [191. Results
ldentificati~ of mAbs against the human TR cytoplasmic tail Fusions of mouse spleen cells immunized with recombinant human TR yielded 35 stable, cloned bybridomas producing mAbs against the human TR t.hat were detected by ELISA. Of these antibodies, 32 mAbs were specific for the external domain of the receptor as they bound to TRs expressed on the surface of viable CCRF-CEM cells [19]. As shown in Fig. la, three other anti-TR mAbs, H68A, H73.2, and 160.1, did not bind to viable CCRF-CEM ceils when assayed by flow cytofluorimetry. These antibodies were then tested to determine whether they immunoprecipitated wildtype human TRs or "tailless' (A3-59) mutant human TRs lacking residues 3 to 59 of the 61-residue amino-terminal cytopla~mtc domain [22]. As shown in Fig. lb, 133/25 mAb, an anu'body against the external domain of the human TR [10,14], immunoprecipitatcd both the wildtype and 'tailless" receptors from detergent lysates of surfaee-iodinated CEF-expressing recombinant hu-
Localization of tlw epitopes recognized by anti-human TR cytoplasmic tail nadbs Mutant human TRs with either deletions or single amino acid substitutions in their cytoplasmic domains [9,22] were used to define the epitopes recognized by each of the mAbs. Immunoprecipitation studies with two deletion mutants, A3-28 and A3-35 established that each mAb recognized a different epitope (Fig. 2). In contrast to the other two mAbs, H68A n~,b did not react with the A3-28 mutant TR, localizing its epitope to residues 3-3_8 of the human TR tail. H73.2 mAb immunoprecipitated ~he A3-28 mutant TR, albeit less efficiently than wild .Wpe receptors, but did not react with the A3-35 mutant TR, indicating that the epitope recognized by this mAb involves residues 29-35. mAb HTR-160.1 reacts with both these deletion mutants, indicating that the epitope it recognizes is located between residues 36 and 61. The YXRF tetrapeptide internalization signal of the human TR encompasses residues 20-23 of the cytoplasmic tail, so that mAb H68.4 might bind directly to the receptor internalization signal. However, mAb H68.4 immunoprecipitated the mutan'. TRy, Y20A, T21A, R22A, and F23A efficiently, indicating that the epitope recognized by H68.4 mAb was distinct from the internalization signal (data not shown). In contrast, the mutant TRs A26F and A25-28 were immunoprecipitated poorly by mAb H68.4 (data net shown). We conclude, therefore, that the
Human Mouse Hamster
MAb MAb MAb 1 1468.4 H73.2 160.1 §1 MMDQARSAFSNLFGGEPLSY'r RFSLARQVDGDNSHVEMKLAVDEEENADNNTKANVTKPKR A M--.-S--R~ 'C,
Fig. 3. Schematicdiagramsummarizingthe localization of the epitopesdetected by anti-humanTR cytoplasmictail mAhsto specificregionsof the humanTR cytoplasmicdomain.The completeaminoacid sequenceof the humancytoplasmictail is shownand the positionsat whichthere are aminoacid substitutionsin the mouseand ChinesehamsterTR cytoplasmictails are indicated. The sequencedata are fromRefs.4-6 and 26.
epitope recognized by H68A is proximal to the carboxy-terminal side of the internalization recognition signal and includes residues 25-28 of me human TR cytoplasmic tail. The results of the epitope mapping experiments are summarized schematically in Fig. 3.
of mouse TR and those of human and Chinese hamster TRs lie within residues 52-61 (Fig. 3). Thus, the epitope recognized by mAb !60.1 is likely to be localized to this 10-residue caeboxy-terminal region of the cytoplasmic tail.
Species crossreacticity o f anti-HTR cytoplasmic tail mAbs
The cytoplasmic tails o f TRs clustered In clathrin-coated pits are inaccessible to mAb H68.4
The primary structure of the TR cytoplasmic tail is highly conserved in the three mammalian species for which cDNA nucleotide sequences encoding their cytoplasmic tails have been determined (Fig. 3, Refs. 4-6 and 26). Strikingly, the deduced amino acid sequences of mouse and Chinese hamster TRs are identical to that of the human TR in the regloa of the cytoplasmic tail where the epitopes for H68A and H73.2 map are located, implying that both mAbs should crossreact with mouse and Chinese hamster TRs. To test this hypothesis, immunoprecipitatie:~ studies were performed using mouse L cells and CHO cells. Consistent with the results of epitope mapping, both H68A and H73.2 mAbs reacted with mouse and Chinese hamster TRs (Fig. 4). In contrast, the third anti-TR cytoplasmic tail mAb, 160.1, immunoprecipitated the Chinese hamster TR but not the mouse TR. With one exception, the amino acid differences between the cytoplasmic tail
c°•r
The distribution of recombinant human TRs in the plasma membranes of transfected CEF has been described in detail previously [24]. These studies used 133/25 gold complexes and demonstrated that at expression levels in the range 2- l0 s to 8- l0 s receptors per cell, 6% of the human TR were located in cIathrin coated domains. Some of these lattices covered invagihating membrane and were typical coated pits while others were fiat lattices, unrelated to invaginated membrane. Conjugates of the anti-TR cytoplasmic tail mAb H68A did not label lattice-coated areas. To investigate this question further, quantitative, double label experiments were carried out in which the external domain of the wildtype human TRs were labelled with B3/25, complexed with 6-nm gold particles and their cytoplasmic tails subsequently labelled with H68A complexed to 3-nm gold. As shown in Fig. 5, the 6-nm B3/25
i
-" ~ |
H73.2 160.1 None
L
[
C')
e
"1"
0
RI7217 ~HTR
H68.4
o
H68.4 H73.2
41
t60.I None
Fig. 4. Crossreactivityof anti-humanTR cytoplasmictail mAbswith mouseand ChinesehamsterTRs. MouseL cells and Chinesehamsterovaw cells were suffacc-iodinatcd, cell lysatcs were prepared, and immunopreci~'itations with human TR mAbs were then performed as described in the legend to Fig. 1. mAb R]7 217 is a rat anti-mouse TR mAb [38]. The immuanprecipitate analyzed in the lane marked aHTR was prepared using a rabbit anti-human TR antiserum generated by immunization with purified recombinant human TR produced in a baculovirus system [2][.
Fig. 5. Stripped membranes labelled on the external surface with B3/25 6-nm gold complexes atKI on the cytoplasmic surface with H68A 3-rim gold complexes. Circles identify colocalization of both particle sizes. (a) Two invaginated pits are shown: (b) the larger aggregates of 6-rim gold particles identify flattened clathrin lattices. Also shown are fibrous cytoskeletal elements and ill-defined, election-opaque masses which are probably invaginated clathrin-coated pits. (a) ×98000:. (b) ×98000.
33 TABLE I Distribution o f TR ~thin duthrin lattices and labelled on their external domains relatire to TR labelled on their ¢3"toplnsmic domains
No. 6-rimgold pat-lickscounted Withinlatticecoated domains Outside latticecoated domains
% of 6-nmparlicles with 3-nmparticles within 25-nmradius
180
I
470
86
immunogold complexes can be seen through the full thickness of the plasma membrane and, as described previously, they are frequently distributed within lattice-coated areas. H68.4 immunogold complexes are not found in lattice-coated areas. However, outside these areas the majority of H68.4 gold particles (86%) are distributed within 25 nm of B3/25 6-rim gold complexes (Table D. We interpret these observations to mean that gold complexes can label aggregates of human T R from both sides of the membrane, but when these receptors become incorporated into a lattice, their cytoplasmic domains are no longer accessible to the H68A gold particles. Discussion The results we have reported clearly establish that mAbs H68A, H73.2 and 160.1 are directed against the human T R cytoplasmic tail. Although many mAbs have been obtained against the external domain of the human T R [10,19,26], none against the cytoplasmic domain of the receptor have previously been repo~ted. One reason that mAbs against the cytoplasmic domain of human T R have been difficult to generate is that the primary structures of the cytoplasmic tails of mammalian T R s are highly conserved. There are four amino acid differences between mouse and human [6,28], and apparently only one between Chinese hamster and human [26]. In contrast, the large external domains of human and mouse T R s are only 77% identical [6]. Two of the three mAbs, H68.4 and H73.2, react with epitopes within a region (residues 24-35) of the human "IR cytoplasmic tail that is completely conserved in monse and Chinese hamster and, therefore, it is not surprising that the mAbs recognize T R s from these species. The fact that apparently high-affinity autoantibodies can be generated against the T R cytoplasmic tail presumably indicates that its intracellular location on the cytoplasmic fa~e of the cell membrane is an innnunologically privileged site from which it is not released in an antigenic form. All three mAbs also crossreact with Chinese hamster TR, and, given the highly conserved nature of the amino-terminal region of the T R cytoplasmic domain, it is likely that many
mammalian T R s will be recognized by one or more of these anti-human T R cytoplasmic tail mAbs. For example, H68.4 has been used to detect T R s from the rat adrenal pheochromocytoma PC-12 [.29]. The broad species crnssreactivity of the mAbs against the T R cytoplasmic tail contrasts to the relatively high species-specificity of mAbs against the external domain cf the human TR. From a panel of 32 murine mAbs against the human TR, only 6 crnssreacted with baboon or macaque T R s (Ref. 19; unpublished results). An interesting feature of mAb H68.4 is its selective reactivity with T R s outside clathrin-coated pits when the cytoplasmic faces of plasma membrane preparations are stained with antibody-gold complexes. The mAb H68A epitope maps close to the T R internalization signal, which is thought to bind to plasma membrane coated pit adaptor complexes [30-32]. A likely explanation ~'or the selective reactivity of H68.4 is, therefore, that its binding is sterically blocked by adaptor proteins associated with the cytoplasmic tails of receptors in clathrin-coated pits. This interpretation is consistent with the ability of n•Ab H68A to inhibit internalization of T R s [33], as would be expected if the mAb could reciprocally block binding of adaptor proteins to the T R cytoplasmic tail. mAbs against the cytoplasmic tail of the human T R are likely to have several potentially important uses in elucidating the mechanism of receptor-mediated endocytosis. T R s are abundantly expressed on the surface of many cultured cell lines and have been extensively used as markers for early endosomes [34-37]. Because of their broad species crossreactivity, the anti-human T R cytoplasmic tail mAbs can be used to detect T R s of species for which anti-TR r n A ~ arc not available. For example, mAb H68.4 has recently been used to demonstrate the colocalizatinn in endocytic vesicles of syaaptopbysin, a synaptic vesicle protein, with T R s in transfected CHO cells [29]. As H68.4 mAb has been shown in this study to discriminate between T R s in coated pits and receptors in uncoated regions of the plasma membrane and the mAb is known to block T R endocytosis [33], it is likely to be valuable in studying the interaction of the T R cytoplasmic tail with coated pit adaptor proteins. Finally, mAbs against the cytoplasmic tail of the numan T R may also prove useful in the purification of early endnsomes from cell homogenates by immunoaffinity chromatography. ~ m e n t s This work was supported by grants CA34787 and CA37641 from the National Cancer Institute and by NATO Collaborative Research Grant No. 880393. We thank Carol Black for technical assistance in generating anti-TR bybridomas, Jim Coilawn for generous advice throughout this work, and Joan Stewart for
34 word processing. C.R.H. and ILM. were supported by a n M R C programme g r a n t a w a r d e d t o C . I L H . RiP.fences 1 Tro~brk~e, I S . ~ L JL~g. S . White, S . Esekog~. V. a~ ~1, M. O991) in B/m~:h~m~o~ of Plasm~ H ~ m c ~ s i s , q['~m~mbc~isand Iron Proteins (Hcmket, i-LC., edJ. Karger Ba~L Sw/tzerD~d 58.139-147. 2 Htmtmrs, H A . and F-inch. C A . [1987) Pl~'siol. R~,'. 67. 520-582. 3 Ji~g.. S. and T ~ !.S. (1987) EMBO J. 6. 327-331. 4 Schneider. C . O~sen. M J . [ ~ . D. and Williams. J.G. (1984) Nalure 311, 6T5-678. 5 M~,e~l, A.. Ku~n. LC. and RuddY, F.H. (1984) CeLl 39, 267-274. 6 Gerhardt. E.M. Chart. L-N.L, Jmg, S . Qi. M. arid Trowln/dg¢, I.S. (1991) Gene 102. 249-254. 7 GoMstein. .LL. Brown. M-S. Anderson, ILG.W. Russe~ D.W. amt Sch~ider, W.J. ( i ~ 5 ) Ann. Re,,-. Cell. B/ol. !. !-39. 8 T ~ , IS. {1991) Current Opin~om in Cell Bioh~y 3, 633-642, enatmn ~ d 1062. 9 Cetlawn. $.F~ Stangel. M-. Kelm. LA.. Esekoms~ V~ Jing. S. Tro~,~lg~. i.S. and Ta/n~. J-~L (19g01 Cell 63,1061-1072. 10 Tro~d~r~ge, I.S. and O ~ a ~ . M.B. (1981) Proc. Nail. Ac~L SoL USA 78. 3O39-3O43. I I Trowbrid~e, I S . ~ , J. and Schuhe, R. (19821J. Cell. Phil, of. !12, 4~~~10. 12 Je~rl'~es, W.A., Bt-d~l~, MR., W~ianl~ A.F. and Hunt. S.V. (1985) I ~ .~, ~ 3 3 - ~ I , 13 ~ J-A., Marsha~ J. and Hayman. MJ. (1985) Bh~chem. J. 232, 735-741. 14 Oread, M.B. and T ~ . I.S. (1981) J. Biol. Chem. 256. 12888-12892. 15 Schneider. C . Sul~erland. 1~ Newman. R.A. and Greaves. M.F. (1982) J. B/oL ~ 257, 8516--8522. 16 ~ , C.R. and Tro,'~oridge. I.S. (1983) J. Cell Biol. 97. 508-52L 17 Trowbr~ge. I-S. and i.~q~z. F. ( 19821Proc. Natl. Acad. Sci. USA 79,1175-1179.
18 ~ . 19 ~
J.F. and Schelt¢. ILl. (19841MoL Cell Biol. 4.1675-1681. $ . T~tte. L S e t ~ u ~ P.A.. l~lherford. M. and TrowIS. {1990) C.qmfer ~ 50. 6295-6301. 20 Tro~ln/dge. l.S. {1988) Progressin Allergy 45.121-146. 21 ll)¢~mgo. D . L aad T ~ i c l g ¢ . I-S. (1988) J~ Biol. Chem. 263. 13586-! 339-TM 22 Jin$. S_ Spewer. 1". Miler. K.. Hopkins. C. and Tro~bridge, i-S. 11990} J. Cell ~ 110, 283-.-m~. 23 Tro~m'~ge, LS, Ralph, P. and Bevan, M J . (1975) P r ~ . Nail. SoL USA 72,157-161. 24 l~k~r, g-. Slip.am M . Tro~bridge. I-S. and Hoplfin~ C.IL (1991) Cell 6.5. 621-632. 25 M./I~, K.. ~ J . K a ~ , EL. S~alessinger. J. and HopC-R. {19S61J. Ceil ~ 102, .50g-509. 26 A~'xmcz. E-. ~ N. and D ~ R J . 11990) Biochcm. L 267. 31-35. 27 ~afl~*rtam~ I L D e l l . D_ S d ~ r , C . Newman. P~ K - ~ h c a d . J. ~ G r e ~ s , ~,tL( i $ 1 ) Proc. l,:dlL ~ ~ USA 78. 4515-
28 Ro~enberger. S_ ~ace~tta. BJ. and Kuhn. LC. (19871 Cell 49, 423-431. 29 Camero~ P - L .~.~H~f, T.C. J a ~ R. and I ~ CamHIL P. (19911 J. Cell Biol. ll5. f51-1fi4. 30 Pearse. B.M.F. (1~8) EMBO J. 7, 3331-3336. 31 G ~ m ~ , $.N, ~ , E. and Pem~e0 B.M.F. (1989) EMBO J. 8, 1041-1047. 32 Pear~. B.~LF. and R o b s o n . MS. (1990) Ann. Rev. Cell. B~L 6.151-171. 33 ~ S.L and S a ~ , E (19911 !. Cell B~L !14, 869-880. 34 Fuller. CA~ and Hop~fim, C R . (1989)I. Cell Science94, 685-694. 35 Beamme~e.. BID. 2,m:l ~ C.R. (1989) B/ocherr~ J. 264, 137-149. 36 Bca~,,me~e, B.D~ Gibson, A. and Hopkins, C.R. (19901 J. Cell B ~ . I l L 1811-1823. 37 Woodman. P.G. and Warren. G. (1991) J. Cell Biol. 112, 1133i 141. 38 ~ . J.F. a~d SchuRe. ILl. (1985) ~ C¢~. B/oL 5,1814-1821.
28
BBAMCR 13207
Monoclonal antibodies against defmed epitopes of the human transferrin receptor ~3'toplasmic tail S u h a i l a W h i t e a, K a r e n M i l l e r b, C o l i n H o p k i n s b a n d l a n S t u a r t T r o w b r i d g e ~' Deparmwat of Cancer BioDD" The Sulk Institutefor Bio4ogicalStudies, San Diego. CA ¢USA~and b iXopartmentof Biochemirl,~:. Iw47enalCoYlege,London eUIO
(Received II Febma~"]992) Key.words: Transferrinreceptor,Cytoplasm/c~
g~noclonalantiS~d~ F_.~qa~sis
Three murine monnelonal antibodies (mAbs) against the 61-residue amino-terminai cytoplasmic tail of the human transfcrrin receptor (TR) have been produced by immunization of mice with recombinant human TR produced in a baculovims expression system. Mutant human TRs expressed in chick embryo fibroblasts (CEFs) with poim mutations or deletions in their cytoplasmic tails have been used to map the epitopes defined by each of the mAbs. One mAb, H68.4, previously shown to bluek receptor /nternalizat/on, biod~,proximal m the carbox,/-tcrmioal side of the YTRF interoai;Tation signal of TR. The second mAb, H73.2, binds near to the carboxy-tcrmioal side of the H68.4 epitope, whereas the third mAb, 160.1, binds closer to the transmembrane region. H68.4 and H73.2 are auto-antibodies congistent with their epitopes mapping to a region of the human TR that has an identical amino acid sequence to the mouse TR. All three mAbs crossreact with the o'toi~asm/c tm'l of Chinese hamster TIL Double labeiiiog of recombinant human TRs on chick embt3,o fibroblast (CEF) cell membrane preparations with B3/35 and H68.4 antibody-gold conjugates established that receptors in clathrin-coated pits were not labeled with H68.4, implying that assne/ated coated pit proteins may block binding of this mAb.
latreductiea Transferrin receptors (TR) bind the serum transport protein transfenin (TO and mediate uptake of iron into the cell (reviewed ha Refs. 1 and 2). The receptor is a homodimeric type !1 membrane protein consisting of two identical approx. 95-kDa subunits covalently linked by two intermoleealar disulfide bonds [3]. The primary structures of human, mouse, and chicken TRs have been deduced from sequencing of their respective eDNAs and are similar [4-6]. The human TR has an extracellular domain of 671 amino acids, a single 28-residue transmembrane region and a 61-residoe amino* terminal cytoplasmic domain. The TR is a member of the class of ligand transport receptors that are constitutively clustered in coated pits and are rapidly internalized [7,8]. A tetrapepfide sequence, YXRF, in the cytoplasmic tail of the TR that is predicted to be a tight turn has been identified as the recoguhion motif for high-efficiency endocytosig [9]. mAbs have been obtained that react with the external domains ef human, m u s e , rat and chick TRs [J0-13] and have been Correspondenceto: I.S.Trowbridge,Departmentof Cancer Biology, The ~alk Institutefor BiologicalStudies, p.o. Box85800. San Diego, CA 92186-5800. USA.
useful in characterization of the receptor and studies of its fm'~ion [14-16]. Some anti-TR mAbs, either singularly or in combinatiou, block Tf-mediated iron uptake and, as a consequence, inhibit ce~ growth [1719]. Such antibodies are potential therapeutic agents iF: the treatment of cancer [19,20]. Here we report tk,e characterization of three novel mAbs against the cytoplasmic tail of the human TIL The epitopes recognized by each of the antibodies have been mapped by. determining their reactivity with mutant human TRs having cytoplasmic tail deletions or single amino aciO .¢ubstitutions, mAbs reactive with the external domain of the TR are ILl,lily highly species-specific. In contrast, antiTR cytoplasmic tail mAbs show broad species crossreactivity and, unusually, two of the mAbs are auto*antibodies. One anti-TR c~oplasmic tail mAb, H68.4, binds selectively to the cytoplasmic tails of receptors found outside clathrin-coated pits in chicken embryo fibroblast (CEF) plasma membranes prepared from cells expressing recombinant human TRs. Materials and Methods Preparation o f m,4bs
Recombinant human TR was produced in a baculovirus expression system [21] and purified on a hu-
29 man "If affinity column [19,21]. Anti-TR mAbs were obtained by immunization of B a l b / c mice with purified receptor glycoprotein and hybridomas generated exactly as described previously [19]. Hybridomas producing anti-TR mAbs were identified by an E L I S A assay for reactivity to recombinant h u m a n T R and cloned by limiting dilution [19]. Isotyping of mAbs was performed using an E L I S A isotyping kit according to the manuf a c t u r e r s instructions (Mono A b - I D l E A Kit; Zymed Laboratories, San Frandsc~, CA). I-t68.4 and 160.1 are l g G I m A b s and H73.2 is a n l g G 3 mAb.
C C R F - C E M [19], were cultured in RPM11640 medium supplemented with 8 % defined calf bovine serum.
CEF expressing mutant human TRs Mutant human T R s were expressed in C E F using the helper-independent retroviral vector, BH-RCAS, as described previously [22]. The construction, expression, and properties of all the m u t a n t h u m a n T R s used in the present study have been described in detail previously [9,22].
Biochemical procedures Cells and cell culture C / E ClEF were obtained from SPAFAS (Nor.rich, CT) and grown in Duibecco's modified Eagle'z medium ( D M E M ) s u p p l e m e n t e d with 1% ( v / v ) chicken serum, 1% ( v / v ) defined calf bovine ~ r u m (Hyclone, Logan, UT), 2 % ( v / v ) ttyptose broth (Difco, Detroit, MI) [22]. Chinese hamster ovary ( C H O ) cells and mouse L cells were cultured in D M E M supplemented with 8 % defined calf bovine serum. H u m a n T leukemic cells,
Cell-surface iodination, immunoprecipitation studies, and SDS-polyacrylamide gel electropboresis were performed as described previously [21], except that C E F were sufface-iodinated as monolayers. Briefly, confluent C E F cultures in 10-cm tissue-culture dishes were iodinated with 1 mCi N a ~ ! (Amersham, Arlington Heights, IL., approx. 16 m g / p . g ) in 2.5 ml of 0.15 M NaCi/0.01 M sodium phosphate buffer (pH 7.2) (PBS) by the lactoperuxidase-glucose oxidase technique
~.~, _.V. ~'1, ~,~~¢oyV.~q, ",. b
|
,
Wt
|
,
~-r~
Fig. i. mAbs against the cytoplasmicdomain of the human transferrin receptor. (a) Flow cytofluorimetricanalysis of the binding of anti-human TR mAbs t~ CCRF-CEM ceils. Human leukemic T cells, CCRF-CEM (2-106) were stained at 4°C with 100/~1 of hybridoma tissue culture supematant followed by saturating amounts of fluoresceinated goat a n t i - ~ immanoglobulinas the second stage antibody [19]. mAb 133/25 was used as a positive couti~ [or ~[~ biudiBg;the uegative corttrol was tissue culture medium without mAb. (b) lmmunupreeipitation of wikltype and "taille~" (d3-59) mutant TRs by anti-TR mAbs. CEF expressing either wildtype or "tailless"(A3-59) mutant TRs were sufface-indinated and cell b'sates were prepP.,~d as ~ in Materials and Methods. Equal volumes of cell lysate were then immanoprecipitated with anti-TR mAbs and analyzed on a 7.5% SDS-polyacrylamide gel under reducing conditions, mAb B3/25 was used as a positive control. The autoradiograph was exposed for 18 h.
3O [23]. Ceils were lysed in 1 ml of i% Nonidet P-40 in PBS and 50-100 /tl of cell lysate used for immunoprecipitation with 50 /zl of hyb~idoma tissue culture supernatant. Antibody-antigen complexes were then adsorbed to fixed S. aureus precoated with rabbit anti-w~J~,e immunoglobulin as described [21]. Samples Were subjected to electrophoresis on 7.5% SDS-polyacrylamide gels under reducing condi_~-:ns.
wm-~
IOeO,
o--
wcn-~qln~
,~
Immuno gold labeling ~ttuties And-human TR n~.bs, 133/25 and H68.4, were complexed to colloidal gold and used to decorate CEF plasma membranes prepared by a surface repi;ca technique as previously described [24]. Cells were incubated with 6 nm 133/25 gold complexes and then the upper plasma membrane was removed to a coverslip exactly as described previously [24]. The stripped membranes were then fixed in 2% paraformaldehyde, quenched and incubated with H68A 3-nm gold complexes as if they were cryosections [25]. Control sections were pre-incubated with free antibody and nonspecific labeling under these conditions was negligible. Finally, still adherent to the coverslip, the stripped membranes were critical point dried, rotary shadowed and whole mounts were prepared for electron mi~oseopy as desen'bed previously [24].
A3-~
w~'ype
~3-3s
Fig. ~ Re~:livi~" of anti-human TR cytoplasmic tail mAbs with mutant human TRs containing cytoplasmic tail deletions. CEF expressing,Aildq.~ or mutant human TRs with either residues 3-28 (A3-28) or residues 3-35 ~3-35) deleted were surface-iodinated. cell II.~ates ¢,ere prepared, and immunoprecipitationsperformed ~ith anti-humanTR mAbs. as described in the legend to Fig. I. B3/25 mAbwas used as a positivecontro!and tissueculture medium v,ithout mAb was used as a negative control. The autoradiograph wasexposedfor 3 da],~s. man TRs. In contrast, the other three mAbs immunoprecipitated wildtype TRs but not the "tailless" mutant receptors, clearly establishing that they were directed against epitopes located in the cytoplasmic tail of the human TR.
Flow cytometry Binding of anti-human TR mAbs to the human T leukemic cell line CCRF-CEM was d e w ,~d by fluorescence-activated cell sorting anal,_ on a Los Alamos design flow cytometer as pre~ ~sly described [191. Results
ldentificati~ of mAbs against the human TR cytoplasmic tail Fusions of mouse spleen cells immunized with recombinant human TR yielded 35 stable, cloned bybridomas producing mAbs against the human TR t.hat were detected by ELISA. Of these antibodies, 32 mAbs were specific for the external domain of the receptor as they bound to TRs expressed on the surface of viable CCRF-CEM cells [19]. As shown in Fig. la, three other anti-TR mAbs, H68A, H73.2, and 160.1, did not bind to viable CCRF-CEM ceils when assayed by flow cytofluorimetry. These antibodies were then tested to determine whether they immunoprecipitated wildtype human TRs or "tailless' (A3-59) mutant human TRs lacking residues 3 to 59 of the 61-residue amino-terminal cytopla~mtc domain [22]. As shown in Fig. lb, 133/25 mAb, an anu'body against the external domain of the human TR [10,14], immunoprecipitatcd both the wildtype and 'tailless" receptors from detergent lysates of surfaee-iodinated CEF-expressing recombinant hu-
Localization of tlw epitopes recognized by anti-human TR cytoplasmic tail nadbs Mutant human TRs with either deletions or single amino acid substitutions in their cytoplasmic domains [9,22] were used to define the epitopes recognized by each of the mAbs. Immunoprecipitation studies with two deletion mutants, A3-28 and A3-35 established that each mAb recognized a different epitope (Fig. 2). In contrast to the other two mAbs, H68A n~,b did not react with the A3-28 mutant TR, localizing its epitope to residues 3-3_8 of the human TR tail. H73.2 mAb immunoprecipitated ~he A3-28 mutant TR, albeit less efficiently than wild .Wpe receptors, but did not react with the A3-35 mutant TR, indicating that the epitope recognized by this mAb involves residues 29-35. mAb HTR-160.1 reacts with both these deletion mutants, indicating that the epitope it recognizes is located between residues 36 and 61. The YXRF tetrapeptide internalization signal of the human TR encompasses residues 20-23 of the cytoplasmic tail, so that mAb H68.4 might bind directly to the receptor internalization signal. However, mAb H68.4 immunoprecipitated the mutan'. TRy, Y20A, T21A, R22A, and F23A efficiently, indicating that the epitope recognized by H68.4 mAb was distinct from the internalization signal (data not shown). In contrast, the mutant TRs A26F and A25-28 were immunoprecipitated poorly by mAb H68.4 (data net shown). We conclude, therefore, that the
Human Mouse Hamster
MAb MAb MAb 1 1468.4 H73.2 160.1 §1 MMDQARSAFSNLFGGEPLSY'r RFSLARQVDGDNSHVEMKLAVDEEENADNNTKANVTKPKR A M--.-S--R~ 'C,
Fig. 3. Schematicdiagramsummarizingthe localization of the epitopesdetected by anti-humanTR cytoplasmictail mAhsto specificregionsof the humanTR cytoplasmicdomain.The completeaminoacid sequenceof the humancytoplasmictail is shownand the positionsat whichthere are aminoacid substitutionsin the mouseand ChinesehamsterTR cytoplasmictails are indicated. The sequencedata are fromRefs.4-6 and 26.
epitope recognized by H68A is proximal to the carboxy-terminal side of the internalization recognition signal and includes residues 25-28 of me human TR cytoplasmic tail. The results of the epitope mapping experiments are summarized schematically in Fig. 3.
of mouse TR and those of human and Chinese hamster TRs lie within residues 52-61 (Fig. 3). Thus, the epitope recognized by mAb !60.1 is likely to be localized to this 10-residue caeboxy-terminal region of the cytoplasmic tail.
Species crossreacticity o f anti-HTR cytoplasmic tail mAbs
The cytoplasmic tails o f TRs clustered In clathrin-coated pits are inaccessible to mAb H68.4
The primary structure of the TR cytoplasmic tail is highly conserved in the three mammalian species for which cDNA nucleotide sequences encoding their cytoplasmic tails have been determined (Fig. 3, Refs. 4-6 and 26). Strikingly, the deduced amino acid sequences of mouse and Chinese hamster TRs are identical to that of the human TR in the regloa of the cytoplasmic tail where the epitopes for H68A and H73.2 map are located, implying that both mAbs should crossreact with mouse and Chinese hamster TRs. To test this hypothesis, immunoprecipitatie:~ studies were performed using mouse L cells and CHO cells. Consistent with the results of epitope mapping, both H68A and H73.2 mAbs reacted with mouse and Chinese hamster TRs (Fig. 4). In contrast, the third anti-TR cytoplasmic tail mAb, 160.1, immunoprecipitated the Chinese hamster TR but not the mouse TR. With one exception, the amino acid differences between the cytoplasmic tail
c°•r
The distribution of recombinant human TRs in the plasma membranes of transfected CEF has been described in detail previously [24]. These studies used 133/25 gold complexes and demonstrated that at expression levels in the range 2- l0 s to 8- l0 s receptors per cell, 6% of the human TR were located in cIathrin coated domains. Some of these lattices covered invagihating membrane and were typical coated pits while others were fiat lattices, unrelated to invaginated membrane. Conjugates of the anti-TR cytoplasmic tail mAb H68A did not label lattice-coated areas. To investigate this question further, quantitative, double label experiments were carried out in which the external domain of the wildtype human TRs were labelled with B3/25, complexed with 6-nm gold particles and their cytoplasmic tails subsequently labelled with H68A complexed to 3-nm gold. As shown in Fig. 5, the 6-nm B3/25
i
-" ~ |
H73.2 160.1 None
L
[
C')
e
"1"
0
RI7217 ~HTR
H68.4
o
H68.4 H73.2
41
t60.I None
Fig. 4. Crossreactivityof anti-humanTR cytoplasmictail mAbswith mouseand ChinesehamsterTRs. MouseL cells and Chinesehamsterovaw cells were suffacc-iodinatcd, cell lysatcs were prepared, and immunopreci~'itations with human TR mAbs were then performed as described in the legend to Fig. 1. mAb R]7 217 is a rat anti-mouse TR mAb [38]. The immuanprecipitate analyzed in the lane marked aHTR was prepared using a rabbit anti-human TR antiserum generated by immunization with purified recombinant human TR produced in a baculovirus system [2][.
Fig. 5. Stripped membranes labelled on the external surface with B3/25 6-nm gold complexes atKI on the cytoplasmic surface with H68A 3-rim gold complexes. Circles identify colocalization of both particle sizes. (a) Two invaginated pits are shown: (b) the larger aggregates of 6-rim gold particles identify flattened clathrin lattices. Also shown are fibrous cytoskeletal elements and ill-defined, election-opaque masses which are probably invaginated clathrin-coated pits. (a) ×98000:. (b) ×98000.
33 TABLE I Distribution o f TR ~thin duthrin lattices and labelled on their external domains relatire to TR labelled on their ¢3"toplnsmic domains
No. 6-rimgold pat-lickscounted Withinlatticecoated domains Outside latticecoated domains
% of 6-nmparlicles with 3-nmparticles within 25-nmradius
180
I
470
86
immunogold complexes can be seen through the full thickness of the plasma membrane and, as described previously, they are frequently distributed within lattice-coated areas. H68.4 immunogold complexes are not found in lattice-coated areas. However, outside these areas the majority of H68.4 gold particles (86%) are distributed within 25 nm of B3/25 6-rim gold complexes (Table D. We interpret these observations to mean that gold complexes can label aggregates of human T R from both sides of the membrane, but when these receptors become incorporated into a lattice, their cytoplasmic domains are no longer accessible to the H68A gold particles. Discussion The results we have reported clearly establish that mAbs H68A, H73.2 and 160.1 are directed against the human T R cytoplasmic tail. Although many mAbs have been obtained against the external domain of the human T R [10,19,26], none against the cytoplasmic domain of the receptor have previously been repo~ted. One reason that mAbs against the cytoplasmic domain of human T R have been difficult to generate is that the primary structures of the cytoplasmic tails of mammalian T R s are highly conserved. There are four amino acid differences between mouse and human [6,28], and apparently only one between Chinese hamster and human [26]. In contrast, the large external domains of human and mouse T R s are only 77% identical [6]. Two of the three mAbs, H68.4 and H73.2, react with epitopes within a region (residues 24-35) of the human "IR cytoplasmic tail that is completely conserved in monse and Chinese hamster and, therefore, it is not surprising that the mAbs recognize T R s from these species. The fact that apparently high-affinity autoantibodies can be generated against the T R cytoplasmic tail presumably indicates that its intracellular location on the cytoplasmic fa~e of the cell membrane is an innnunologically privileged site from which it is not released in an antigenic form. All three mAbs also crossreact with Chinese hamster TR, and, given the highly conserved nature of the amino-terminal region of the T R cytoplasmic domain, it is likely that many
mammalian T R s will be recognized by one or more of these anti-human T R cytoplasmic tail mAbs. For example, H68.4 has been used to detect T R s from the rat adrenal pheochromocytoma PC-12 [.29]. The broad species crnssreactivity of the mAbs against the T R cytoplasmic tail contrasts to the relatively high species-specificity of mAbs against the external domain cf the human TR. From a panel of 32 murine mAbs against the human TR, only 6 crnssreacted with baboon or macaque T R s (Ref. 19; unpublished results). An interesting feature of mAb H68.4 is its selective reactivity with T R s outside clathrin-coated pits when the cytoplasmic faces of plasma membrane preparations are stained with antibody-gold complexes. The mAb H68A epitope maps close to the T R internalization signal, which is thought to bind to plasma membrane coated pit adaptor complexes [30-32]. A likely explanation ~'or the selective reactivity of H68.4 is, therefore, that its binding is sterically blocked by adaptor proteins associated with the cytoplasmic tails of receptors in clathrin-coated pits. This interpretation is consistent with the ability of n•Ab H68A to inhibit internalization of T R s [33], as would be expected if the mAb could reciprocally block binding of adaptor proteins to the T R cytoplasmic tail. mAbs against the cytoplasmic tail of the human T R are likely to have several potentially important uses in elucidating the mechanism of receptor-mediated endocytosis. T R s are abundantly expressed on the surface of many cultured cell lines and have been extensively used as markers for early endosomes [34-37]. Because of their broad species crossreactivity, the anti-human T R cytoplasmic tail mAbs can be used to detect T R s of species for which anti-TR r n A ~ arc not available. For example, mAb H68.4 has recently been used to demonstrate the colocalizatinn in endocytic vesicles of syaaptopbysin, a synaptic vesicle protein, with T R s in transfected CHO cells [29]. As H68.4 mAb has been shown in this study to discriminate between T R s in coated pits and receptors in uncoated regions of the plasma membrane and the mAb is known to block T R endocytosis [33], it is likely to be valuable in studying the interaction of the T R cytoplasmic tail with coated pit adaptor proteins. Finally, mAbs against the cytoplasmic tail of the numan T R may also prove useful in the purification of early endnsomes from cell homogenates by immunoaffinity chromatography. ~ m e n t s This work was supported by grants CA34787 and CA37641 from the National Cancer Institute and by NATO Collaborative Research Grant No. 880393. We thank Carol Black for technical assistance in generating anti-TR bybridomas, Jim Coilawn for generous advice throughout this work, and Joan Stewart for
34 word processing. C.R.H. and ILM. were supported by a n M R C programme g r a n t a w a r d e d t o C . I L H . RiP.fences 1 Tro~brk~e, I S . ~ L JL~g. S . White, S . Esekog~. V. a~ ~1, M. O991) in B/m~:h~m~o~ of Plasm~ H ~ m c ~ s i s , q['~m~mbc~isand Iron Proteins (Hcmket, i-LC., edJ. Karger Ba~L Sw/tzerD~d 58.139-147. 2 Htmtmrs, H A . and F-inch. C A . [1987) Pl~'siol. R~,'. 67. 520-582. 3 Ji~g.. S. and T ~ !.S. (1987) EMBO J. 6. 327-331. 4 Schneider. C . O~sen. M J . [ ~ . D. and Williams. J.G. (1984) Nalure 311, 6T5-678. 5 M~,e~l, A.. Ku~n. LC. and RuddY, F.H. (1984) CeLl 39, 267-274. 6 Gerhardt. E.M. Chart. L-N.L, Jmg, S . Qi. M. arid Trowln/dg¢, I.S. (1991) Gene 102. 249-254. 7 GoMstein. .LL. Brown. M-S. Anderson, ILG.W. Russe~ D.W. amt Sch~ider, W.J. ( i ~ 5 ) Ann. Re,,-. Cell. B/ol. !. !-39. 8 T ~ , IS. {1991) Current Opin~om in Cell Bioh~y 3, 633-642, enatmn ~ d 1062. 9 Cetlawn. $.F~ Stangel. M-. Kelm. LA.. Esekoms~ V~ Jing. S. Tro~,~lg~. i.S. and Ta/n~. J-~L (19g01 Cell 63,1061-1072. 10 Tro~d~r~ge, I.S. and O ~ a ~ . M.B. (1981) Proc. Nail. Ac~L SoL USA 78. 3O39-3O43. I I Trowbrid~e, I S . ~ , J. and Schuhe, R. (19821J. Cell. Phil, of. !12, 4~~~10. 12 Je~rl'~es, W.A., Bt-d~l~, MR., W~ianl~ A.F. and Hunt. S.V. (1985) I ~ .~, ~ 3 3 - ~ I , 13 ~ J-A., Marsha~ J. and Hayman. MJ. (1985) Bh~chem. J. 232, 735-741. 14 Oread, M.B. and T ~ . I.S. (1981) J. Biol. Chem. 256. 12888-12892. 15 Schneider. C . Sul~erland. 1~ Newman. R.A. and Greaves. M.F. (1982) J. B/oL ~ 257, 8516--8522. 16 ~ , C.R. and Tro,'~oridge. I.S. (1983) J. Cell Biol. 97. 508-52L 17 Trowbr~ge. I-S. and i.~q~z. F. ( 19821Proc. Natl. Acad. Sci. USA 79,1175-1179.
18 ~ . 19 ~
J.F. and Schelt¢. ILl. (19841MoL Cell Biol. 4.1675-1681. $ . T~tte. L S e t ~ u ~ P.A.. l~lherford. M. and TrowIS. {1990) C.qmfer ~ 50. 6295-6301. 20 Tro~ln/dge. l.S. {1988) Progressin Allergy 45.121-146. 21 ll)¢~mgo. D . L aad T ~ i c l g ¢ . I-S. (1988) J~ Biol. Chem. 263. 13586-! 339-TM 22 Jin$. S_ Spewer. 1". Miler. K.. Hopkins. C. and Tro~bridge, i-S. 11990} J. Cell ~ 110, 283-.-m~. 23 Tro~m'~ge, LS, Ralph, P. and Bevan, M J . (1975) P r ~ . Nail. SoL USA 72,157-161. 24 l~k~r, g-. Slip.am M . Tro~bridge. I-S. and Hoplfin~ C.IL (1991) Cell 6.5. 621-632. 25 M./I~, K.. ~ J . K a ~ , EL. S~alessinger. J. and HopC-R. {19S61J. Ceil ~ 102, .50g-509. 26 A~'xmcz. E-. ~ N. and D ~ R J . 11990) Biochcm. L 267. 31-35. 27 ~afl~*rtam~ I L D e l l . D_ S d ~ r , C . Newman. P~ K - ~ h c a d . J. ~ G r e ~ s , ~,tL( i $ 1 ) Proc. l,:dlL ~ ~ USA 78. 4515-
28 Ro~enberger. S_ ~ace~tta. BJ. and Kuhn. LC. (19871 Cell 49, 423-431. 29 Camero~ P - L .~.~H~f, T.C. J a ~ R. and I ~ CamHIL P. (19911 J. Cell Biol. ll5. f51-1fi4. 30 Pearse. B.M.F. (1~8) EMBO J. 7, 3331-3336. 31 G ~ m ~ , $.N, ~ , E. and Pem~e0 B.M.F. (1989) EMBO J. 8, 1041-1047. 32 Pear~. B.~LF. and R o b s o n . MS. (1990) Ann. Rev. Cell. B~L 6.151-171. 33 ~ S.L and S a ~ , E (19911 !. Cell B~L !14, 869-880. 34 Fuller. CA~ and Hop~fim, C R . (1989)I. Cell Science94, 685-694. 35 Beamme~e.. BID. 2,m:l ~ C.R. (1989) B/ocherr~ J. 264, 137-149. 36 Bca~,,me~e, B.D~ Gibson, A. and Hopkins, C.R. (19901 J. Cell B ~ . I l L 1811-1823. 37 Woodman. P.G. and Warren. G. (1991) J. Cell Biol. 112, 1133i 141. 38 ~ . J.F. a~d SchuRe. ILl. (1985) ~ C¢~. B/oL 5,1814-1821.
