Immunological approaches for probing receptor structure and function

7’iPS- September 1991 [Vol. 121

338

artide we highlight immunological approaches for probing membrane receptors. Two major targets will be the focus: the nico-

Immunological approaches for probing receptor structure and function Suleiman W. Bahouth, Craig C. Ma/bon

Hsien-yu Wang and

clonitrgltns renealcd the primary sequence of numerous membrane receptors, and this infornution catolysed two important effDrfs; modeling O/ receptor sfruchrre by hydroparhy analysis and generating sequence-specific immunological probes &th which these models can be tested experimentully. Craig Malbon and his colleagues outline the recent uduances that illustrufe 80~1 anti-pcpfide antibodies raised to synthetic sequences a/ membrane receptor huve generafed new information on the topology, functional domains and celhdor focalizotio~rof transmembrane signaling elements. 73ey focus on two examples, the C protein-linked &adrenoccptor, and the nicotinic acefylcholir~e receptor, afr infrinsic ion channel receptor. These two classes of receptor provide f~wplnfrs for the analysis of topographical models of membrane proteins wit/r immunological probes, especially anti-peptide nntibodies, and drmonstmfe how these rest& complement those obtained from molecular, biochemical nnd bioph&al feckrriques. AMougb this powcrftrl strategy is not without foulfs, it is likely to continue trl be applied successfsrlly fo the analysis of the sfructurt and function of receptors, ion channels and other nremdrairejvofeins.

Moltwlnr

Membrane receptors are typically low-abundance proteins with major functional significance in transducing information from incoming signals to some output that alters the intracellular tenvironment. Purified nicotinic acetylcholine receptors and the photopigment rhodopsin provide invaluable templates for the cvolution of molecular modeling of membrane receptors based upon data derived initially from physical and biochemical anaIys#. The apphraron of molecular cloning techniques has greatly expanded our knowledge of the primary sequence of these protein@, as well as of membrane receptors whose low abundance

precluded chemical sequencing of the entire, purified molecule6. With this expansion of data came the mom formidable task of developing strategies. for discriminating among models of mceptor structure, organization, and perhaps function. Antibodies to hormone and neurotransmitter receptors have been employed profitably to probe molecular structure and functional domains of membrane proteins’. In contrast tc the arduous task of preparing highly purified proteins from large-scale isolations, the use of synthetic peptides as immunogens for antibody production, based on sequence information derived from microsequencing and molecular cloning, now makes analysis of membrane proteins by immunological techniques more practical and widely available. In addibon to polydonal antibody production, hybrfdoma monodonai technology permits the production rf antibodies to stngie, well-defined antigenic sites of proteins and rjjthetic peptides’. In this brief

tinic acetykholine cation-conducting,

receptor,

a

acetylcholineactivated channel, and the B_ adrenoceptor, a G protein-linked receptor that coupies to adenyhd cyclase and Caz+ channeh?. &Adrenoceptorasatarget The g-adrenoqtor Provides an excellent example in which the careful and thorough application of immunochemical strategies has yielded important new knowledge of the structure and biology of G protein-linked membrane recep tors. Antibodies raised to native p-adrenoceptors isolated on a !.:rge scale by affinity chunnatography and HI’LC prnvided several key pieces of inform&on a-bout the structure of the M. bnmunobiots of cell manbmne fractions separated by SDS-PAGE, transferred to nitroceUulose, and then stained with anti-rucqtor antibodies pennttted a&y&s of receptor molecular m,ass and abundance in a variety of tissues andcellQq?cs.Thephranrcdogitally distinct jI1- and &-rubLypes of mammalian adrenocqtom were shown to be glyooPrx&eins with molecular weights of 65-67 kDa when chemic4y

re&ced

by

thiois and alkylated, diqra&ng the notion that major differwxs in size existed in the msmmalian subtypea, as had been suggested from a comparison of the avian @Iand amphibian &-forma9, Likewise, a momxlonal anti-i&type

antibody to the mcqtor Provided evidence for heteqe&ty in the N-glycosylation ppttums in the &-adrenuqtor of human epidermoid carcinoma C&20. Quantification of rece#or in crude_ subcelhrlar fractions was made possible through the availability of those immunologicai reagents. in a&Won, high-msohttion immunocyto&emical locaiiuHon was made Possible in complex tissues like the brain”, as well as other sites”. Antibodies m4ive with a native molecule can be prepared against chemically synthusiad peptides whose primary sequence haa been obtsined by direct chemical sequencing of the native molecule or fragments derived from it, or by molecular cloning.

339 models and established the predictive vaIue of hydropathy analy

The use of site-directed antipeptide antibodies verifies the results obtained from molecular &ning, estabiiding that antibodies to p&i&d sequences do indeed recognize the native molec&. Robing the organization and topography of membrane receptors can be achieved once a panel of site-specific, anti-peptide antibodies becomes available. For a moIecuIe like the ~adrenoc+ tor, predicted to have seven hy drophobic, potentially membranespanning domei&, the task of establishing the topology was quite formidablet3. Antibodies were prrpucd against 11 peptide haptens oorresponding to each of the hydrophilic sequences flanking the hydrophobic domains of the hamster &-adrenoceptor. SpecScity of antibodies can be confirmed by several techniques: by WnunobIotting of immobilized peptide, purified receptor and membrane fractions with varying abundance of receptor expre&m; by immunoprecipitations; and by immunocytochemical techniques (Table I). To obtain anti-receptor antibodies that passed these rigorous characterMions of specificity, antisera were prepared against the immunogen in multiple animals. In some cases immunization of nine animals was required to obtain antibalies which fulfilled these criteria. Chinese hamster wary (CHO) c&s were tran&cM with an expmssion vector harboring the EDNA encodin the &adrenoaptor, and su t Ames exSf Z%Z~Yz%l analysis of receptor topography.

sis for developing models of one class of membrane protein with multiple membrane-spanning domains13. Important features of receptor structure have been revealed uniquely by immunological techniques. The existence of disulfide bridges in both mammalian &and &-adrenoceptors was first deduced bf chemical analysis of purified preparations, which displayed molecular masses of 6% 67 kDa under reducing conditions, compared with 55kDa under non-reducing SDS-PAGE analysis9. Thiol treatment generates a receptor capable of a&iThis high level of expression was vating G. in the absence of agonist required to optimize the anaIysis iigand”,‘5. Thiols also destabilize the binding domain; this is reversby~uno+ochemistry.Tlrmugh seiective permeabiiiution of the ibie by oxidation. The central cell membranes of fixed cells, question of whether the receptor W&S in sifu as the 55 kDa disulfollowed by &&II with antipeptide antibodies to hydrophilic fide-bridged, or the 65 kDa, domains and then aualyais by inspecies was defined by immunodirect immunofiuorescence, each blotting of membrane fractions sequence could be localized either prepared horn cells placed under anaerobic conditions and subto the extracelhdar face of the jected to agents that alkylate free membrane (staining in @ermeabilsulfhydryl groups. The receptor ized and non-permeabilized cells) was shown to exist in situ as the or to the cytoplasmic face of the 55 kDa species, the species which membrane (staining only in peractivates G, when treated by meabilized 041s). Protease sensitivity of the epitopes in intact thiols or agonist”. The versus permeabilized cells pro- sites of covalent insertion of tLadnmocep& photoaffinity vided additional information to ligands were defined using sitec&inn the observations derived directed antibodies to immunofrom indi8wt immuna&ormcenee precipitated receptor fragments in (Fig. 1). The results of thir work tandem with microsequencing’b. discrimillrted between different

c-L/

TiPS - Septrmber 2991 [Vol. 121

340

analysis permitted mapping of structural features of the ligand-

This

binding site. The results suggest that p-adrenoceptor ligands bind within a bundle of the seven domains, membrane-spanning localized to the seventh membrane-spanning region, analogous to the binding of ll-cis-retinal in

the photoreceptor rhodopsin”‘. Although polyclonal antibodies raised against puritied p-adrenoceptors and anti-idiotypic monoclonal antibodies to the &adrenoceptor antagonist alprenolol have been shown to reduce or abolish the capacity of the receptor to bind radiolabeled antagonists’, anti-peptide antibodies to the

eleven hydrophilic domains of the molecule failed to alter radioligand binding properties of the receptort3.

Anti-peptide antibodies have been employed to compare the structure of fl-adrenoceptor sub-

types, as well as the structures of other G protein-linked receptors. All but four anti-peptide antibodies generated to sequences of hamster &-adrenoceptor also displayed immunoreactivity toward human placental and rat fat cell fit-

adrenoceptors, reflecting the level of sequence identity that exists between the two subtypes”. Anti-

bodies prepared against synthetic peptides that correspond to the cytoplasmic loop between membrane-spanning domains 1 and 2, as well as those directed against the serinelthreonine-rich region of the C-terminus of the bovine photoreceptor rhodopsin, also recognized purified &-adrenoceptor”. Edmundson helical pm-

jections of the primary sequence of the fi-adrenoceptor identified the sequence found in rhodopsin. These data suggest conservation of structural domains of rhodopsin in the padrenoceptor. Among the possible functions for these domains are protein k&se recogni tiim sites or sequences involved in the binding and activation of C proteins. The binding domains of rhodopsin for the retinal C protein transducin were mapped using site-directed anti-peptih antibodies in tandem with reconstitution of rhodopsin and tram+ ducin into phospholipid vesides’s. Photobleached rhodopsin binding of transducin wan shown to reduce the binding of anti‘de antibodies to loop 3-4 i!3@ the C-terminal sequences of rhodop-

TiPS- September1992/b’oI.121

sin. By contrast, binding of the antibody raised against a 14 amino acid peptide corresponding to a sequence within loop 5-6 of rhodopsin was unaffected by the presence of transducin. These results suggest a preferential involvement of regions in or near loop 3-4 and the C-terminus in the binding of transducin by photobleached rhodopsin. Receptor localization and organization can also be addressed with immunocytochemical techniques. For example, mapping of B-adrenoceptor expression in the rat brain and a variety of other tissues has been achieved’**‘2. Application of this strategy to questions concerning receptor expression in differentiation and development will yield new insights. Perhaps more intriguing is the organization of the B-adrenoceptor in situ as probed by indirect immtmofluomzcence 1920.The receptors do not display a random, diffuse pattern in stained cells, but rather a punctiform organizreceptor comolecular order. Figure 2 displays the glucocorticoid-induced up~gulation of &admnoc+om. During agonist, induced desensitization, when receptor binding and activity declines, the receptorshave been shown to retain the ability to be

stained by anti-peptide antibodies directed against extracelluku domains, i.e. the receptors may disaggregate by moving latere!ly but do not apparently leave the cell surf.lce or ‘internalti in response to agonist20. Further study will be required to define whether or not lateral scquest&ion within the plane of the lipid bilayer plays a functional role in desensitization nAChRazatarget The nicotinic acetylcholine receptor is a pentameric membrane protein composed of four homologous glycqrotcin subunits with a stoichiometry of or, B, y and b2’. These fundamentally similar, but distinct, gene products are arranged symmetrically around a central ion channel, to which all subunits contributeza. Establishing the topography of this ligand-gated ion channel has been a major oal of neurobiologista. The fok! ing pattern of the receptor has been approached by

341

chemical, structural and immunological strategies, as we!! ..J by hydropathy analysis. The results, often seemingly controversiai, have contributed to an evolution in the model and a re-assessment of the strategies employed to test the existing models. Hydropathy analysis of the primary sequences of the nicotinic receptor subunits revealed via molecular cloning provided the early models for membrane folding of each subunit. Four hydrophobic domains, each composed of -20 amino acyl residues, were predicted to form membranespanning ar-helices in each subuniP. These four putative membrane-spanning domains were designated Ml, M2, M3 and M4 (Fig. 3a-d). Based solely upon this organization, both the N- and C-termini would be predicted to be extracellular (Fig. 3a). The region N-terminal to Ml, constituting the acetylcholine binding site, the putative N-glycosylation site (further N-terminal), and the N-terminus itself (generated by the loss of a signal peptide) must al&be extracellular. An additional putative membrme-spanning region termed MA, which displays a periodic alteration of polar and nonpolar residues with a periodicity of an a-helix, was pmposed subsequently between M3 and MI%-. The assumption that a fifth membrme-spanning region exists, hypothesized to provide a hydrophil:: iining to the ion channel, necessitates that the C-terminus is intracelular rather than extracellular (Fig. .3b). The use of sue-directed antipeptide antibodies to sequences of the nicotinic receytor subunits, together with immunoelectron and immunofluorescence microscopic techniques, provide an alternative strategy with -which to d&riminate among the various models for receptor folding advanced by hydropathy analysis (Fig. 3a-c). This approach confirmed that the N-terminus of each subunit was extracellular. The region between the N-terminus and Ml in each subunit accounts for -40-45% of the protein. This region on the a subunit contains a site of potential Nglycosylation (Asnlll? and the sites reactive to the affinity label MBT’A(vie. Cysl92 and Cys193p3’ (Fig, 3). Other indirect immuno-

logical evidence suggests, however, that a portion of the Q subunit of the nicotinic receptoramino acids 156-179 - is exposed at the cytoplasmic surfaceas,~~.To account for these observations, a short non-a-helical transmembrane domain between residue 156 and the site of potential Nglycosylation (M6), and a second between residue 179 and the site of agonist binding (M7), have been proposed (Fig. 3d). The existence of two additional membranespanning domains, M6 and M7, would, however, be inconsistent with the extended distribution of mass indicated by the radius of gyration and the volume of t!le receptolJ0. The existence of transmembrane domains Ml, M2 and M3, has been established by several criteria and there is strong evidence for the participation of .M2 in the formation of the cation channelZp-J’. To probe the disposition of MA, several antibodies recognizing peptide sequences within and flanking MA in the u subunit were shown to localize to the cytoplasmic surface of the nicotinic receptopjzu (Fig. 3c). In addition, proteolytic cleavage of the nicotinic receptor and microsequencing of the released peptides, revealed that the majority of the proteolytic fragments or@nate from the region between M3 and M4%, where MA was purported to cross the membrane2”~u. Deletion mutagenesis studies have also indicated that MA is not necessary for channel functionas. The amphipathic o-helical character of this region strongly suggests that MA lies in an interphase between polar and nonpolar environments, such as the surface of the membrat#. These data provide strong evidence against a disposition for MA across the lipid bilayer (Fig. 3b,c). A cytoplasmic disposition for the region between M3 and M4 may be critical for receptors such as the glutamate-operated ion channel (tht NMDA receptor) in the CNS whose membrane organization is anabous to the nicotinic mceptar. Within this domain of the glutamate recephw an important exchange in a small number Of amino acids (termed ‘Bip and flop’) imparts different pharmacological and kinetic properties to the currents evoked by Bgand in

TPS - September 1991/Vol. 221 1

b CHO

b__

d

the various glutamate receptors3’. The sequence extending from around residues W-426 in the o subunit of the nicotinic receptor and a corresponding region extending from around residues 456174 in the fr subunit define the boundary of M4. A region on the immedfate N-terminal side of M4 was immunolocnlized to the cytoplasmic surface ot the receptor’s (Fig. 3c,d). The region between MI and the C-terminus, in both the a and B subunits of the nicotinic receptor, was also immunolocalized to the cytoplasmic surface of the receptoP3. These studies, suggesting a cytoplasmic localization of M4 and the Cterminus, are inconsistent with results obtained by biochemical analyses”. For the Torpedo receptor, compelling evidence has accumulated establishing the

C-terminus of h subunit to be extracellular, occurring as a dimer with a disulfide cross-link between CysSOO of two 6 subunitsJe,4°,4i (Fig. 3e). Cl

Cl

El

Antibodies, in particular sitedirected anti-peptide antibodies, provide powerful tools with which the localization, organization, structure and function of membrane receptors can be approached. Immunocytochemishy can illuminate receptor expression in cells or in complex collections of cells such as the brain: The organization of membrane mceptom, espectally those with multiple membrane-spanning domains, is a formidable question to address. She-directed, anti-peptide antibodies used in

tandem with immunocytochemical techniques can aid in establiihing the diapo&ton of receptor domains as extmc&k or cytoplasmic, but the rem&a from this approach are best complemented by data from chemical or biochemical strategies. Structural features of receptai* can be rweakd bv w of antibodies capable of _ immunopmcipitattng fragments derived fnnll receptors subjected to chemical modification, phokaffinity labeling, etc. The size of the antibody probe itself must be consfdered when planning immuno&emiul strategies. In some caaea ak-&ected anti-peptide antibodis have been shown to be uoeful probes of functional features (i.e. contact sitea) of the mceptot that are sensitive to shicfdfng. None of these approaches is devoid of the exper-

TiPS - September 1992IVol. 121 imental pitfalls common to the use of antibodies. Not only must the specificities of the antibodies be characterized rigorously, bbt also the immunocytochemistry must be conducted at zn!i’bodv titers within the range of dih&ons at which specitkity was determined. There is little doubt, however, that even with well-defined experimental limitations, this new dimension of creating specific antibodies to defined regions of complex membrane proteins will continue to expand our repertoire of analytical tools. Acknowledgements This work was supported by grants DK30111 and DK2!5410 from the NIH (to C.C.M.), a Grantin-Aid from the Tennessee Affiliate-American Heart Association (to S.W.B.), and a grant from the National Research Council of the Republic of China (to H-y.W.).

1 Barnard. E. A., Miiedi, R. and Sumikavsa, K. (1982) Pwc. R. Sot. Lordon Ser. B 215.241-246 2 Ovchlnnikov, Y. A. (1%2) FEBS L&t. 148.179-191 3 Nuda, M. et PI. (19B2) Noturr 299,

793-797 4 Nob, M. ct al. (1983) Nature 302, 251-M 5 Nathnr, 1. and Hugness, D. S. (19B3) Cell 34,8B7-814 6 Dbam, R. A. F. ct al. (19B6) Notwe 321, 7w9 7 Malbun, C. C., Muxham, C. P. and Bmndwein, H. in The Beh-adrewt$r Receptor (Perkins, J., ed.), Humrna Rau (in pmm) 8 Gilman. A. G. (1987) Anuu. Rerr. Biochmn. S6,61%49 9 G~uWIO, M. P., Muxham, C. P. and Mdt,on, C. C. (1985) J. Biol. Chm. 260, 7665-7674 10 Cervantes-Olivia, P. et al. (1988) Biothem. j.Z50,133-143 11 Wanaka, A. rt rf. (1989) Brain Rrs. 485, 125-140 12 T&no, T. et aI (;989) Rrain Res. 499, 174-179 13 Wang H-y., Lipfe& L., Malbnn, C. C. and Bahouth, S. (1989) J. Biof.Chew 264, 1442~lr*431 14 Pederw, S. E. and Rose, E. M. (1985) /. Bid. Chent.260,14150-14157 15 Moxham, C. P., Ross, E. M., George, S. T. and Malbun, C. C. (19B8) Mol. Phamscof. 33,4&-492 16 Wag, S. K-F., Slaughter, C.. Ruoho, A. and Russ, E. M. (198B)J. Biol. Chem. 263, Rli-1928 17 Weln, E, wk. J., Johnson, G. L. and Malbun, C. C. (19B7) J. Biol. C/rem. 262 43l9-4323 18 Weiss, E., Kc&w, D. J. and Johnson, G. I. (1%) /. Btot.Chem.263,6150-6154 19 Waq, H-y., Rerrioa, M. and M&on, C. C. (1999) Biocbtm.J. 263,519-532 20 Wang, H-y., Berries, M. and Mrlbon,

343 C. C. (1989) Biochrm.1.263.533-538 21 R~]noids, j. A. and ‘Karlin. A. (1978) Biochcaistrv17.2035-2038 22 Dani, J. (&Ii) Trrttds Nrrrosci. 12, 125-128 23 Noda, M. et nl. (1983) Nature 302, 528-532 24 Finer-Moore, J. and Struud. R. M. (1964) Proc. Nutl Acad. Sci. USA 81,15!&159 25 Guy, H. R (1984) Biophys. I. 45249-261 26 Kao, P. N. et of. (1984) I. Biot. Chcm. 259, 11662-1166s 27 Changwx, J. P. (1990) TrendsPharmncol. Sri. 11,4B%92 28 Craido, M., Hochschwender, S., Sarin, V., Fox, J. L. and Lindstrom, J. (1985) Pmt. NotI Acod. Sri. USA 82, ZOM-ZCOE 29 Pederson.S. E., Bridgman, P. C,Sharp, S. 0. and Cohen, J. 8. (1990) J. Biol. Cbem.265,569-581 30 Karlin, A. Hamry Lcct~weSer. (in press) 31 Stroud, R. M.. McCarthy. M. P. and Shuster, M. (1990) Bioclwmistry 29, 1100%11023 32 Ratnam, M. et al. (1986) Biochrmistry25, 2621-2632

33 Maelicke. A. et al. (1989) Biochemistry 28,1356-140s 34 Moore. C. R. ct zi. (1989) Biochemistry 28,9164-9191 35 Mishina, M. rt al. (1985) Notwr 313, 364-369 36 Finer-Moore, J.. Bazan, F., Rubin, J. and Stroud. R. M. (1989) in Prediction of Protein Stntctarc and the Prhipirs of Prokin Conform&ion (Fasman, G., cd.), pp. 719-759, Plenum Pma 37 Summer, B. et al. (1990) Science 249, 15&l-158s 38 Ratnam, M., Le Nguyen, D., Rivier, J., Sagcnt, P. B. and Lindsbran, 1. (1986) Biorhrmistry25.2633-2~3 39 Young, E. F. et al. (1985) Proc. Natl Acad. Sci. USA &?,626-630 40 McCrea, P. D., Popor, J-L. and Engdman, D. M. (1987) EMBO J. 6,361~3626 41 DiPaok, M., Czajkowski, C. and K&in, A. (1989) 1. Biol.Ckcm. 264,15457-15463

MBTA: C(Nmakimido) ammonium iodide

benz#rimethyl

Proto-oncogenetranscription factorsand epilepsy James I. Morgan and Tom Curran Chemically and electrici*!!yinduced seizures elicit the rapid transcriptional a&&ion in neurons of ;I,‘&s ofgenes referred to us cellular immedirtte-early genes. Since the products of these genes include transcription factors and cytokines, they ore proposed to bej!:voloed in coupling neuronal excitation to a complex, and poorly understood, programme of dlular responses that involves the regulation of gene expression. Products of two cellular immrdiateearly genes, c-fos and c-jun, are componenk of the kanscription factor AP-I. In this reoiezu,Jim Morgan and Tom Curran discuss how thesegene products have begun to reveal some of the molecular details of sfimufus-transcription coupling in the nervous system following seizures. in addition, fhese genes have provided novel reagents and concepts for investigating the biochemical and cellular sequelae of ~izun in the CNS, and point towards new avenues of research and potenliol therapeutic targets b epilepsy. Neuronal plasticity is a property of most nervous systems whereby the response of a circuit is modifieci a.: a consequence of altered input. Such changes range& their duration from the transient, lasting from seconds to hours, to those that are evident for weeks or, indeed, for the life span of the anima!. This implies that there must be celhdar mechanisms that couple stimuli to permanent alterations in neuronal phenotype. Conventionally, plasticity is asJ. I. Morgan is in thr Dqrrtmrrt 01 Neuw scinxe, aad T. Curran is bt the Dqwtment of Molrnlar Om-ololpyand Virology. at thr Rwhlr Instttrtr of M&calor B!u!ogy. Rochr Rrsrarch Cmter, Nutlry, Nj 07110. USA.

sociated with normal neurophysiological processes such a? adaptation, facilitation and potenHowever, pezmanent tiation. alteration of neuronal behaviour is also a charackristi~~of several nemopathological states, induding epilepsy, suggesting that similar mechanisms may underlie the genesis or maintenance of these conditions. The search for the biochemical links between cell stimulation and changes in cell phenotype has led to investigations of the arpreasion of the c-fos and c-jun protosncossnca in various rodent models of epilepsy. Excitation of neuions has two general consqucnces. The first is a short-lived biochemical and