Proc. Natl. Acad. Sci. USAVol. 88, pp. 3837-3841, May 1991Immunology

A subunit common to an IgG Fc receptor and the T-cell receptormediates assembly through different interactions

(Fcy receptor type HI/CD16/T-cell receptor/CD3/C chain/protein assembly)

TOMOHIRO KUROSAKI, IRENE GANDER, AND JEFFREY V. RAVETCHDeWitt Wallace Research Laboratory, Division of Molecular Biology, Sloan-Kettering Institute, 1275 York Avenue, New York, NY 10021

Communicated by Paul A. Marks, February 1, 1991

ABSTRACT The ; subunit is a component of the Fcyreceptor of natural killer cells (FcyRIll or CD16), as well as themultimeric T-cell receptor/CD3 complex, and is required forassembly of both native receptors. The role of the C subunit inhuman FcyRIIIA assembly differs from its role in T-cellreceptor/CD3 complex assembly. The transmembrane domainof the FcyRIIIA a subunit forms noncovalent interactions withthe comparable domain of the ; subunit and is sufficient forsurface expression of the FcyRIIA complex. In the absence ofthese interactions, sequences in the transmembrane domain ofthe FcyRMA a subunit signal its degradation. Leu-46, presentin the transmembrane domain of the human ; subunit, isimportant for assembly with the FcyRIIIA a subunit. Substi-tution of this leucine with an isoleucine, as found in the mouse{subunit, significantly reduces this interaction. In contrast, themouse and human ; subunits interact with the pentamericT-cell receptor/CD3 complex, resulting in surface expressionof this receptor.

Receptors for IgG immune complexes, the Fcy receptors(FcyRs), are a heterogeneous family of membrane receptorson cells of the immune system that mediate a diverse array ofcellular responses to antibody-antigen complexes (for re-view, see refs. 1-3). FcyRIIIA, the FcyR on natural killer(NK) cells, is an oligomeric complex that mediates antibody-dependant cellular cytotoxicity and NK-cell activation uponcross-linking with immune complexes (4). These responsesare mediated through the stimulation of inositol phospholipidhydrolysis and increased intracellular calcium (5). FcyRIIIAis composed of at least three polypeptide chains: a ligand-recognition subunit, termed FcyRIIIA-a, and the associatedy and ; chains (6-11). FcyRIIIA-a from human NK cells hasbeen shown to exist in three molecular forms in associationwith Y2, {2, or y; chains (11). These associated chains arerequired for surface expression of the a subunit (8-10) andare presumed to be involved in signaling pathways, triggeredby cross-linking of the a subunit by ligand. The y and ;subunits were initially identified as necessary components ofother receptors. The y chain is a component ofthe multimerichigh-affinity Fce receptor for IgE (FceRI) (12), and the ;chain is a component of the T-cell receptor (TCR)/CD3complex (13). The 'y and chains are required for surfacereceptor expression of FceRI and TCR/CD3, respectively,and demonstrate significant sequence homology (14, 15). InTCR/CD3 assembly, the ' chain plays a unique role intargeting pentameric complexes (TCR a and P chains/CD3 'y,8, and E chains) to the cell surface and sparing the incom-pletely assembled complex from degradation (16, 17). In thisreport, we describe results relevant to the mechanism bywhich the ; chain mediates FcyRIIIA assembly and thesequence requirements for these assembly processes.

MATERIALS AND METHODSCell Culture and Antibodies. COS-7 cells were cultured in

modified Eagle's medium containing 10% (vol/vol) fetal calfserum. The peptide synthesized from the cytoplasmic domainof FcyRIIIA-a was used to raise the polyclonal antibodyagainst the a chain. The y and ; chains were made byoverexpressing their respective cDNA sequences in thebacterial system described by Studier and Moffatt (18). Thecomplete coding sequences were cloned into the BamHI siteof pET3xb to create a fusion protein with the T7 gene 10protein and used to transform Escherichia coli BL21(DE3).The fusion protein was induced with isopropyl f3-D-thiogalactoside and the overexpressed molecules were puri-fied by SDS/PAGE. The eluted proteins were used to raiserabbit polyclonal antibodies.DNA Constructions and DNA Transfection. Point mutants

and hybrid molecules were constructed by PCR amplificationor oligonucleotide-directed mutagenesis on single-strandedtemplates and cloned into the pCEXV-3 expression vector(19). DNA was transfected into COS-7 cells by using calciumphosphate precipitation in the presence of 100,M chloro-quine followed by a 20% (vol/vol) glycerol shock.Immunoprecipitation and Fluorescence-Activated Cell

Sorter (FACS) Analysis. Transfected cells were permeabi-lized with digitonin (10 ,g/ml) for 10 min on ice and iodinatedwith 1251 by 1,3,4,6-tetrachloro-3a,6a-diphenylglycouril(Iodo-Gen) (6). Cells were solubilized in lysis buffer contain-ing 1% digitonin and 0.12% Triton X-100 (11). Cell lysateswere sequentially incubated for 2 hr at 4°C with antibodiesand then with protein A-Sepharose. Digestion with endogly-cosaminidase H (endo H) was performed by incubatingimmunoprecipitated proteins for 16 hr at 37°C with 5 x 10 3units of endo H in 0.1 M sodium phosphate (pH 6.1) con-taining 0.1% Triton X-100, 0.03% SDS, and 20 mM EDTA.Sepharose-bound immune complexes were eluted into sam-ple buffer. FACS analysis was carried out using the anti-FcyIII monoclonal antibody (mAb) 3G8 or a nonbindingirrelevant mAb, B77, as described (10).In Vitro Analysis. The cDNA sequences of a, y, and

chains were cloned into pGEM-3Z and transcribed in vitro.These RNAs were translated in an in vitro translation reac-tion mixture containing 1 ,ug ofthe RNA in a nuclease-treatedreticulocyte lysate (Promega) supplemented with dog pan-creas microsomal membranes (kindly provided by ChrisNicchita, Rockefeller University, New York). Translationproducts were labeled with [35S]methionine and [35S]cys-teine. After 1 hr of incubation at 30°C, a 10-fold volume ofdigitonin lysis buffer was added and the reaction mixture wasimmunoprecipitated with antibodies. The immune complexeswere precipitated with protein A-Sepharose. Proteins were

Abbreviations: FcER, FcE receptor; FcyR, Fcy receptor; TCR,T-cell receptor; NK, natural killer; FACS, fluorescence-activatedcell sorter; endo H, endoglycosaminidase H; mAb, monoclonalantibody; ER, endoplasmic reticulum.

3837

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Oct

ober

11,

202

0

3838 Immunology: Kurosaki et al.

eluted from the beads into SDS/PAGE sample buffer con-taining 50 mM dithiothreitol.

AChose fir

EndoH - + - + -

RESULTS

Previous results have demonstrated (8-10) that the y or;subunit enhances surface expression of FcyRIIIA-a by as

much as 50-fold. To determine the mechanism by which the; subunit mediates FcyRIIIA assembly, we transfected sin-gly, and in combination, FcyRIIIA-a and y and ; chaincDNAs into COS cells and assayed for their surface expres-sion. FcyRIIIA-a alone resulted in barely detectable surfaceexpression; coexpression of a subunit with either y or human; chain resulted in a 20-fold stimulation of surface expressionof FcyRIIIA-a. Surface labeling and immunoprecipitation ofthese transfectants demonstrated oligomeric complexes of achain with homodimeric y or ; chain (Fig. 1 A and B). Thaty and ; chains were found as disulfide-linked dimers was

demonstrated by analyzing the protein complexes immuno-precipitated from these transfectants on nonreducing gels(data not shown). COS cells transfected with Fc'yRIIIA-a andyand human chain cDNAs formed three types ofcomplexeson the cell surface: a chain with disulfide-linked y, A, or y;chain (Fig. 1C). These three complexes are similar to thoseobserved on human NK cells (11), demonstrating that theinteractions of a, y, and {chains can be reconstituted by theircoexpression in COS cells and do not appear to require otherNK-cell-specific proteins.The mechanism by which y and ; subunits facilitate surface

expression ofmultimeric complexes containing a subunit wasdetermined by characterizing the intracellular fate ofFcyRIIIA-a in the presence or absence of y and ; chains.Pulse-chase labeling of the COS cells transfected withFc'yRIIIA-a alone, shown in Fig. 2B, indicates that the a

chain is degraded over the course of 6 hr. Endo H treatmentof pulse-chase-labeled immunoprecipitates demonstratesthat the a chain does not acquire resistance to this enzyme,implying the inability of the a subunit to reach the Golgisystem. In contrast, pulse-chase studies of COS cells trans-fected with y and ; chain cDNAs demonstrate that thesemolecules are stable (data not shown) and are capable ofreaching the cell surface in the absence of a chain (Fig. 1 Aand B). Coexpression of a with y or ; subunit resulted instable accumulation of a subunit, and the processed protein

A

DNA

46 -

30 -

21.5 -

14.3 -

6.5-

a y X

Ii_

B

46 -

30 -

21.5

14.3 -

CZ

. 4

}.Ia| --h

C a

Y

46 -

30

21 5 -

14.3

46- W30 _

?- P:

CChose ;r> .6r

Endo H + - + - +

5-21 53

D

a.

..--

32] -am-

FIG. 2. Association of FcyRIIIA-a with yand {chains influencesthe kinetics of its degradation. COS cells expressing FcyRIIIB (A) (8,10, 21), FcyRIIIA-a (B), Fc'yRIIIA-a with y chain (C), orFcyRIIIA-a with human ; chain (D) that were pulse-labeled for 30min at 370C with [35S]methionine (0.25 mCi/ml; 1 Ci = 37 GBq) and[35S]cysteine (0.25 mCi/ml) in carrier-free medium were not incu-bated (0) or were incubated for 2 hr (2h) or 6 hr (6h) in regularmedium. Cell lysates were immunoprecipitated with mAb 3G8 (A) ora rabbit polyclonal antibody against a chain (B, C, and D) andimmunoprecipitates were digested with endo H for 16 hr at 370C (+).

Parallel immunoprecipitates were not treated with endo H (-).

Proteins were separated on a Tricine/SDS/10%o polyacrylamide gel.The percentage of FcyRIIIA-a remaining relative to the amount ofprotein labeled before the chase was determined by densitometricscanning of autoradiograms: a chain alone, 9%o (2h) or 0.1% (6h); aand y chains, 22% (2h) or 14% (6h); a and hC chains, 27% (2h) or 12%(6h).

acquired resistance to endo H (Fig. 2 C and D). Fc-yRIIIB,normally expressed as a glycosyl phosphatidylinositol an-chored protein in neutrophils (6), is rapidly converted to anendo H-resistant molecule, indicating that its processing andmovement to the Golgi are more rapid than FcyRIIIA (Fig.2A). These results indicate that y and ; chains can interferewith the degradation of the a chain and allow these com-plexes to reach the Golgi system. It is likely that this occursthrough noncovalent interactions between the a chain and they or ; chain in the endoplasmic reticulum (ER) as shown inFig. 3. mRNAs for lckyRIIIA-a and 'y and ; chains, synthe-sized in vitro, were translated in a cell-free system in thepresence of microsomes and immonoprecipitated with anti-bodies directed against each chain. These microsome prep-arations contain only ER membranes and are free of con-

RNA a he ma he me h

I- hC/hC- ylhC- Yly 46 - .-

30 - X X4 -^21.5 -

143 -Ab a y a

I.--

FIG. 1. Cell surface expression of FcyRIIIA-a and -y and humanC chains on COS cells. The coding sequences of human FcyRIIIA-a(a), mouse y (y), and human ; (hj) chains were inserted intopCEXV-3 (19). COS-7 cells were transfected with the indicated(above lanes) plasmid DNA (15 ,ug ofeach plasmid) and were surfacelabeled with 115I. Immunoprecipitation experiments with the indi-cated antibodies (below lanes) were performed 72 hr after DNAtransfection. Sepharose-bound immune complexes were eluted intosample buffer containing 2% (wt/vol) SDS and 1% 2-mercapto-ethanol and separated on Tricine/SDS/10o polyacrylamide gels (Aand B). For the nonreducing sample (C), 2-mercaptoethanol wasomitted from the sample buffer. Positions of molecular mass markersare shown in kDa.

A b a a a -a1 2 3 4 5 6

FIG. 3. In vitro association of FcyRIIIA-a with the y and;subunits. In vitro-translated proteins, labeled with [35S]methionineand [35S]cysteine and precipitated with the indicated antibodies,were separated by either Laemmli SDS/PAGE (15% gel)(FcyRIIIA-a plus D or Tricine SDS/PAGE (10%1 gel) (FcyRIIIA-aplus y). Glycosylation intermediates of FcyRIIIA-a were detectedwhen an anti-a chain antibody was used. Note that the fully glyco-sylated form preferentially associates with chains. The relativeintensities of 46-kDa a bands in lanes 4-6 were determined bydensitometric scanning (100%1, 18%, and 10%6, respectively).

B

-- - + - -4-

Proc. Natl. Acad Sci. USA 88 (1991)

a - 4 9-

21! tiI . .-.4 3

x E

4w

a is -Y

Dow

nloa

ded

by g

uest

on

Oct

ober

11,

202

0

Proc. Natl. Acad. Sci. USA 88 (1991) 3839

taminating Golgi (22). The same complexes observed to formamong a, y, and ; chains in vivo are also formed in vitro,indicating that only ER membranes are required for theseinteractions.FcyRIIIA-a contains a signal that results in its degradation.

We have shown (10) that deletion of the cytoplasmic domainof FcyRIIIA-a did not affect its degradation. A series ofchimeric molecules were constructed with FcyRIIIA-a and yand ; chains to identify this degradation signal (Fig. 4A).When the transmembrane and cytoplasmic domains of the achain were replaced with the comparable domains of the y or; chain, stable accumulation and cell surface expression ofthese chimeric molecules were observed (ahZ or ay, respec-tively). The same result was obtained when the mouseFcyRII transmembrane and cytoplasmic domains were sub-stituted for mouse Fc'yRIII-a sequences (R. L. Weinshank,

AStructure of gene

hZ -COLAb

mZ -AvAc VAIY,

hmZ JVACOL AF

mhZ -Vc VA I /Iy

hZ(Y) --VAC0 L

hZ(I) -V CVc~PyF

mZ(F) -act\IvN0

mZ(L) --cV

ahZ A

'aQ Q

B

human -y

mouse

human S

mouse S

and J.V.R., unpublished results). These experiments suggestthat the 21-amino acid transmembrane sequence of the achain contains a signal that determines its degradation. Theinteraction ofFcyRIIIA-a with y or {chain in the ER protectsit from degradation and probably masks this signal. y and ;chains may also provide the appropriate signal for directingthe a-chain-containing complex from the ER to the Golgi.The sites of interaction between a and ; chains could be

mapped as a consequence of our previous observation thatcoexpression of the human ; chain with a chain in COS cellsresults in efficient surface expression, whereas the mouse ;chain is 5-7% as efficient in its ability to function in theassembly of either human or mouse FcyRIIIA-a (10). Chi-meric molecules constructed between human and mouse gchains were tested for their ability to enhance FcyRIIIA-asurface expression (Fig. 4A). The sequences of the ; chain

Expressione efficiency (%)

100

7

75

5

107

35

5

25

77

140

leader peptide |extra- |cellular

transmembrane I cytoplasmic

FIG. 4. Mapping the domains of the ; chain that are required for surface expression of FcyRIIIA-a. (A) Chimeric molecules and pointmutations of the chain. Point mutants and hybrid molecules were transfected singly or cotransfected with FcyRIIIA-a into COS-7 cells. Surfaceexpression was determined by FACS analysis using mAb 3G8 (anti-FcyRlII) and expressed as a fraction of wild type (FcyRIIIA-a plus human; chain). Data were calculated as mean values from three or four experiments. The schematic structure of each molecule is shown, with theimmunoglobulin-like extracellular domains indicated by semi-circles and the transmembrane domains asjagged lines. h, Human; Z, A; m, mouse.Letters in parentheses are substituted amino acids (single-letter code). (B) Sequence comparison of y and ; chains is shown for the regionresponsible for association with FcyRIIIA-a (NH2-terminal extracellular domain, transmembrane domain, and first 21 residues of theintracytoplasmic domain). Arrow indicates the Leu-46 in human, which is conserved among human y and mouse y chains but differs in mouseC chain and which influences the ability of ; chain to interact with FcyRIIIA-a.

MIPAVVLLLILLVEQAAA *LGo*EP4LCYILDAILFLYGIVLTLLYdRLKIQVRKAAITSYEKSDGVY

MISAVILFLLLLVEQMAA *LG *EPQLCYILDAVLFLYGIVLTLLYCRLKCIQVRKAAIASREKADAVYa a 0000

MKALFTAAILIAAQLPITEAQSFGLLDPKLGYLIDGILFIYGVILTALFlFVKYSRSAEPPAYQQGQNQLY*a@0 0 00 e0*.*a0*00a.. ' 0*

MKWKVSVIACILHVRFPGAEAQSFGLLDPI&CYLIDGILFIYGVIITAL FSRSAETANLQDPNQLY

Immunology: Kurosaki et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

11,

202

0

3840 Immunology: Kurosaki et al.

that confer this difference on the surface expression of a

chain were mapped to the 9-amino acid extracellular domain,the transmembrane domain, and the first 21 amino acids ofthe cytoplasmic domain. The amino acid sequences of humanand mouse{chains were compared to that of y chain (Fig. 4B)to identify sequences that are conserved between human;and y chains and divergent in mouse{ chain. Such sequences

could be responsible for this functional difference inFcyRIIIA-a interaction. Leu-46 of the human; chain isconserved in y chain at position 39 but is an isoleucine inmouse; chain. To determine the role of this difference in theability of; chain to interact with a chain, single amino acidsubstitutions were made at this position in mouse and humanC chains (Figs. 4A and 5). When Leu46 was mutated toisoleucine in human; chain, the interaction with a chain wasreduced by 65%. Similarly, mutation of Ile-46 to leucine in themouse sequence led to a 5-fold stimulation in its ability toassemble with a chain (Figs. 4A and 5). Other amino aciddifferences between mouse and human; chains in the trans-membrane region and the 21 residues of the cytoplasmicregions were tested for their role in a- chain interactions. Noeffect was observed in the efficiency of the surface expres-sion of a chain when these single and double mutants were

transfected with a chain cDNA (Fig. 4A). The Leu -* Ile

substitution did not affect the in vivo stability of; chain as

determined in pulse-chase experiments; the mouse{ proteinwas able to reach the cell surface as efficiently as its humancounterpart (data not shown). Thus the difference in theefficiency of interaction of FcyRIIIA-a with human or mouse

C chain results from the reduced ability of mouse; chain toform a noncovalent complex with a chain in the ER, as can

be seen from in vitro reconstitution experiments shown inFig. 3. When mouse; mRNA was translated in a cell-freeextract supplemented with microsomal membranes and a

chain, a complex between these chains could not be detected,in contrast to the situation when human; chain was used.y and; chains exist as disultide-linked homo- and het-

erodimers (7, 11, 16, 20, 23) in all the receptor complexes inwhich they have been found. To determine if disulfide-linkeddimerization of these subunits is essential for their ability toassemble with FcyRIIIA-a, we mutated the cysteine atposition 32 to a serine in the human; chain and constructedthe homologous cysteine mutation (Cys-25 -* Ser) in the 'y

chain. These molecules were expressed on the cell surface

0)m0E

z

0

hz mz

mz(L)

Log Fluorescence Intensity

FIG. 5. FACS analysis of COS cells transiently expressing theindicated mutants and FcyRIIIA-a. y* and hC* indicate the y mutant

(Cys-25 -- Ser) and the human ; mutant (Cys-32 -. Ser), respec-

tively. Control fluorescence (dotted line) was determined using a

nonbinding mAb, B77. h, Human; z, C; m, mouse. Letters inparentheses are substituted amino acids (single-letter code).

but were incapable of disulfide-linked dimerization, as dem-onstrated on nonreducing gels (data not shown). There aretwo cysteines (Cys-25 and Cys-44) in a y chain. Mutation ofCys-44 to Ser did not affect disulfide-linked dimerization of'y chains, indicating that Cys-25 is responsible for disulfidebond formation. These mutatedf and human; chains werestill able to associate with FcyRIIIA-a upon coexpression inCOS transfectants, although surface expression of the com-

plex was reduced by 40% for the; mutant and 10%1 for the y

mutant (Fig. 5).

DISCUSSIONThe ligand-binding subunits of low-affinity IgG Fc receptors,FcyRII and FcyRIII, differ in their ability to be efficientlyexpressed on the surface of heterologous cells. MouseFc'yRIII-a was :=10% as efficient as FcyRII in its ability to beexpressed on the surface of transfected cells (7, 10, 24). Thismodest level of surface expression for FcyRIII-a wasachieved only after increasing the concentration of stable a

mRNA in the transfectants by the truncation of its 3' un-translated sequences (thereby deleting degradation signals)and increasing the concentration of transfected DNA. Sur-face expression of the a subunit could then be reproduciblydetected in the absence of associated chains.

This decreased efficiency of Fc'yRIII-a as compared toFcyRII expression results from its requirement for subunitinteractions. Both the y subunit of FceRI and the C subunit ofTCR/CD3 associate with FcyRIIIA-a (7-11). The transmem-brane domain of the C subunit interacts directly with thecomparable domain of Fc'yRIIIA-a in the ER and is sensitiveto changes in specific amino acids in this domain. This rolefor Cchain in assembly of FcyRIIIA-a is different than its rolein TCR/CD3 assembly. Direct interaction ofr chain with theligand-binding subunit TCRa does not occur; rather a Cchaininteracts with the pentameric precursorafyye (16, 17). Bothhuman and mouse C chains interact with the pentamericTCR/CD3 complex sparing it from lysosomal degradation,thus resulting in its surface expression. Despite the extensivesequence homology between mouse and human C chains,only the human C chain is capable of directly interacting withFcyRIIIA-a, indicating that the amino acid difference (LeuIle) between human and mouse; molecules distinguishes one

type of receptor assembly (FcyRIIIA) and another (TCR/CD3). Thus, in mouse NK cells, one type of FcyRIIIAcomplex,a{2, would be predicted to be absent, althoughcomplexes containing a, y2, or ayC would still be capable ofassembling.The fate of the isolated ligand-binding subunits of

FcyRIIIA or TCR/CD3 is similar. Both FcyRIIIA-a andTCRa contain signals in their transmembrane domains thatresult in their degradation in the ER (25, 26). These moleculesdo not acquire endo H-resistant glycosylation, and blockingthe transport of these molecules from the ER to the Golgi withbrefeldin A does not affect their fate (ref. 25 and data notshown). However, noncovalent interactions of an associatedsubunit with these sequences prevent their degradation. Inthis respect, the role of the y or C chain in FcyRIIIA assemblyis similar to that of the CD38 chain in TCR/CD3 assembly(26). The assembly of FcyRIIIA described here is likely to beshared by the tetrameric Fc receptor for IgE, FcERI. It hasa ligand-binding a subunit that shares substantial sequence

homology to Fc-yRIIIA-a, particularly in its transmembranedomain (27, 28) and, for the human receptor, interactsdirectly with the y chain. From the results presented herein,we would expect that the interaction of the transmembranedomain of FceRI-a with 'y chain will be similar to theinteraction of FciyRIIIA-a with y and human f chains.The interactions between these ligand-binding subunits

and their associated chains in the ER result from specific

a, III ..,.

IIIIj h I

\ ~~~hz(DI

Proc. Natl. Acad. Sci. USA 88 (1991)

A

11 11 1111 11111

Dow

nloa

ded

by g

uest

on

Oct

ober

11,

202

0

Proc. Natl. Acad. Sci. USA 88 (1991) 3841

protein interactions, occurring in the lipid bilayer. Both theFcyRIIIA-a and the TCRa have charged amino acids in theirtransmembrane domain. These residues appear to play acritical role in the assembly of these subunits in the ER.Mutation of the Asp-222 residue in the transmembrane do-main of FcyRIIIA-a decreases its susceptibility to degrada-tion and reduces its ability to interact with y and human gchains (10). Similar observations have been made for TCRaassembly with CD38 (26). In contrast to the role of thetransmembrane domains in assembly, truncation of the cy-toplasmic domains of either the a or the human ; chain hasa less dramatic effect on this process (10). In addition to therole of charged residues interrupting the hydrophobic stretchof amino acids spanning the bilayer, a surprising degree ofspecificity was observed in the hydrophobic interactions thatoccur in this domain. The Leu -* Ile substitution in thetransmembrane domain of ; chain resulted in a significantdecrease in its ability to interact with Fc'yRIIIA-a.Thus the results presented herein indicate that a high

degree of specificity is contained within the transmembranedomains of oligomeric receptor complexes. The role of thisdomain is considerably more dynamic and complex thanpreviously thought. Rather than functioning as a passivehydrophobic nonspecific anchor through the lipid bilayer,this domain contains specific recognition signals for degra-dation, receptor assembly, and subcellular targeting.

We are grateful to Dr. Chris Nicchita of the Rockefeller Universityfor his assistance with the in vitro translation experiments and toMari Kurosaki and Jahan Dadgar for technical assistance. Thesestudies were supported by grants to J.V.R. from the NationalInstitutes of Health. T.K. is a Charles A. Dana fellow and recipientof an award from the Foundation for Cancer Research (Tokyo). I.G.is supported by a fellowship from the Deutsche Forschungsgemein-schaft.

1. Ravetch, J. V. & Anderson, C. L. (1989) in Fc Receptors andthe Action ofAntibodies, ed. Metzger, H. (Am. Soc. Microbi-ol., Washington), pp. 221-235.

2. Kinet, J.-P. (1989) Cell 57, 351-354.3. Ravetch, J. V. & Kinet, J.-P. (1991) Annu. Rev. Immunol. 9,

457-492.4. Perussia, B., Tutt, M. M., Qiao, W. Q., Kuziel, W. A.,

Tucker, P. W., Trinchieri, G., Bennett, M., Ravetch, J. V. &Kumar, V. (1989) J. Exp. Med. 170, 73-87.

5. Cassatella, M. A., Anegon, l., Cuturi, M. C., Griskey, P.,Trinchieri, G. & Perussia, B. (1989) J. Exp. Med. 169, 549-567.

6. Ravetch, J. V. & Perussia, B. (1989) J. Exp. Med. 170,481-497.

7. Ra, C., Jouvin, M.-H., Blank, U. & Kinet, J.-P. (1989) Nature(London) 341, 752-754.

8. Hibbs, M. L., Selvaraj, P., Carpen, O., Springer, T. M.,Kuster, H., Jouvin, M.-H. & Kinet, J.-P. (1989) Science 246,1608-1611.

9. Lanier, L. L., Yu, G. & Phillips, J. H. (1989) Nature (London)342, 803-805.

10. Kurosaki, T. & Ravetch, J. V. (1989) Nature (London) 342,805-807.

11. Anderson, P., Caligiuri, M., O'Brien, C., Manley, T., Ritz, J.& Schlossman, S. F. (1990) Proc. Nati. Acad. Sci. USA 87,2274-2278.

12. Blank, U., Ra, C., Miller, L., White, K., Metzger, H. & Kinet,J.-P. (1989) Nature (London) 337, 187-189.

13. Samuelson, L. E., Harford, J. B. & Klausner, R. D. (1985) Cell43, 223-231.

14. Weissman, A. M., Baniyash, M., Hou, D., Samelson, L. E.,Burgess, W. H. & Klausner, R. D. (1989) Science 239, 1018-1021.

15. Kuster, H., Thompson, H. & Kinet, J.-P. (1990) J. Biol. Chem.265, 6448-6452.

16. Sussman, J. J., Bonifacino, J. S., Lippincott-Schwartz, J.,Weissman, A. M., Saito, T., Klausner, R. D. & Ashwell, J. D.(1988) Cell 52, 85-95.

17. Weissman, A. M., Frank, S. J., Orloff, D. G., Mercep, M.,Ashwell, J. D. & Klausner, R. D. (1989) EMBO J. 8, 3651-3656.

18. Studier, F. W. & Moffatt, B. A. (1986) J. Mol. Biol. 189,113-130.

19. Ravetch, J. V., Luster, A. D., Weinshank, R. L., Pavlovec,A., Portnoy, D. A., Hulmes, J., Pan, Y.-C. E. & Unkeless,J. C. (1986) Science 234, 718-725.

20. Orloff, D. G., Ra, C., Frank, S. J., Klausner, R. D. & Kinet,J.-P. (1990) Nature (London) 347, 189-191.

21. Lanier, L. L., Cwirla, S., Yu, G., Testi, R. & Phillips, J. H.(1989) Science 246, 1611-1613.

22. Walter, P. & Blobel, G. (1983) Methods Enzymol. 96, 84-93.23. Perez-Montfort, R., Kinet, J.-P. & Metzger, H. (1983) Bio-

chemistry 22, 5722-5728.24. Weinshank, R. L., Luster, A. D. & Ravetch, J. V. (1988) J.

Exp. Med. 167, 1909-1925.25. Bonifacino, J. S., Suzuki, C. K. & Klausner, R. D. (1990)

Science 247, 79-82.26. Bonifacino, J. S., Cosson, P. & Klausner, R. D. (1990) Cell 63,

503-513.27. Kinet, J.-P., Metzger, H., Hakimi, J. & Kochan, J. (1987)

Biochemistry 26, 4605-4610.28. Shimizu, A., Tepler, I., Benfey, P. N., Berenstein, E. H.,

Siraganian, R. P. & Leder, P. (1988) Proc. Natl. Acad. Sci.USA 85, 1907-1911.

Immunology: Kurosaki et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

11,

202

0