Organization and dynamics of microtubules in Torpedo marmorata electrocyte: Selective association with specialized domains of the postsynaptic membrane

Neuro~ctence Vol 43. No 1, pp 151-162, 1991 Printed m Great Britain

0306-4522/91 $3 00 + 0 00 Pergamon Press plc ,c, 1991 IBRO

ORGANIZATION A N D D Y N A M I C S OF MICROTUBULES IN TORPEDO MARMORATA ELECTROCYTE: SELECTIVE ASSOCIATION WITH SPECIALIZED DOMAINS OF THE POSTSYNAPTIC MEMBRANE B. J. JASMIN,*J.-P CHANGEUX'~a n d J CARTAUD*~ *Blologle Cellulaire des Membranes, Instltut Jacques Monod, CNRS, Unlverslt6 Paris 7, 2 Place Jussleu, 75251 Paris C6dex 05, France tNeurobxologle Mol&ulalre, CNRS, D6partement des Blotechnologles, InstltUt Pasteur, 25 rue du Docteur Roux, 75724 Paris C6dex 15, France Abstract--The dlstnbutlon and subcellular organization of two components of the secretory pathway, the Golgl apparatus and mlcrotubules, have been investigated in Torpedo marmorata electrocyte This highly polarized syncytlum, embryologlcally derived from skeletal muscle cells, displays distinct plasma membrane domains on its innervated and non-innervated faces, and it played a critical role in the identification of the acetylchohne receptor By lmmunocytochemlcal analysis, we show that in the electrocyte, numerous focal Golgl bodies are dispersed throughout the cytoplasm In frequent association with nuclei Under expenmental conditions known to stabilize mlcrotubules, we reveal an elaborate network composed of two populations of mlcrotubules exhibiting different dynamic properties as evaluated by cold-stabihty, resistance to nocodazole and post-translational modification This network appears organized from several nucleating centers located in the medial plane of the cell that are devolded of centrloles The network displays an asymmetric distribution with individual mlcrotubules converging towards the troughs of the postsynaptlc membrane folds In these particular regions, we consistently observed clusters of non-coated vesicles in association with the mlcrotubules. The organization of the mlcrotubules in the electrocyte may thus result in a functional polarization of the cytoplasm In other polarized cells, the particular organization of the secretory pathway accounts for the lntracellular routing of membrane proteins. The orgamzatlon that we have observed m the electrocyte may thus lead to the vectorial delivery of synaptlc proteins to the innervated plasma membrane Furthermore, the abundance of synaptlc proteins makes the electrocyte a unique model with which to decipher the mechanisms revolved m the sorting and targeting of these glycoprotelns

The cellular a n d molecular m e c h a n i s m s by which postsynaptlc m e m b r a n e d o m a i n s are generated a n d m a i n t a i n e d are still poorly u n d e r s t o o d . In this context, the electrocyte from the electric fish Electrophorus electricus a n d Torpedo marmorata m a y represent a n attractive model with which to address these f u n d a m e n t a l issues because of its particular architecture The electrocyte is a flattened syncytium embryologlcally derived from muscle stem cells (refs in R e f 10) It displays a highly asymmetric structure characterized by a differentiation o f its p l a s m a m e m b r a n e into two different functional domains, each with a distinct biochemical c o m p o s i t i o n The noni n n e r v a t e d m e m b r a n e is specialized m the active t r a n s p o r t of ions It is characterized by the accumulatlon o f two integral m e m b r a n e proteins, N a +, K ~-ATPase a n d chloride channel, 2'23'3~49 a n d by a m e m b r a n e skeleton consisting of actln polymers, ++To whom correspondence should be addressed Abbrevtatmns AChR, acetylchohne receptor, EGTA, ethyleneglycolbls(aminoethylether)tetra-acetate, FITC, fluorescem lsothlocyanate, MDCK, Madln-Darby Canine Kidney, MES, 2-(N-morphohno)-ethanesulfontc acid, PBS, phosphate-buffered sahne, TMR, tetramethylrhodamme

spectrln a n d a n k y r m 27,28 O n the o t h e r h a n d , the innervated m e m b r a n e is specialized in the genesis of electrical processes: it receives a dense chohnerglc i n n e r v a t l o n a n d contains high levels o f nicotinic acetylchohne receptor ( A C h R ) molecules 4'4°~: (see also refs m Ref 10). Several n o n - r e c e p t o r m e m b r a n e associated peripheral proteins are also localized within the postsynaptxc m e m b r a n e d o m a i n These include the 43,000 mol. wt protein 44 (and refs in Refs 8, 16 a n d 29) a n d 54,000 mol wt lamln B-related protein, 9 a n d proteins with mol wts of 58,000) 7 87,000, 7 270,000-300,00050 a n d d y s t r o p h l n 2~ Thus, the electrocyte exhibits a striking polarity which can be c o m p a r e d to t h a t of epithelial a n d n e u r o n a l cells (see refs in Refs 12, 36 a n d 43) A t variance with the postsynaptic m e m b r a n e s of b o t h muscle a n d n e u r o n s which represent only a small fraction of the cell surface, the electrocyte displays an " h y p e r t r o p h i e d " postsynaptlc m e m b r a n e d o m a i n , the genesis a n d m a i n t e n a n c e of which rely extensively u p o n the cell's secretory actwlty W e have thus initiated, using the electrocyte of Torpedo marmorata, a series of experiments aimed at elucidating the m e c h a n i s m s involved in the assembly a n d renewal of proteins in postsynaptlc m e m b r a n e d o m a i n s Since 151

152

B J JAbMINet at

m polarized cells the secretory pathway, a e the GolgJ a p p a r a t u s a n d mlcrotubules, is k n o w n to be revolved in the vectorial delivery o f m e m b r a n e constituents, we have, in the present work, investigated the subcellular & s t r l b u t l o n a n d dynamics o f these organelles to u n d e r s t a n d the m e c h a n i s m s Involved m polarized protein t r a n s p o r t m the electrocyte

EXPERINIENTAL PROCEDURES Tzssue preparatton

Because of the known depolymenzmg effects t)~ cold on mlcrotubules, Torpedo marmorata were decerebrated w~thout prior anesthesm on ~ce Columns ol electrocytes were then rapidly &ssected and subjected to e~ther one of the following experimental protocols

Fig. 1. Distribution and organization of the Gotgl in the electrocyte of T marmorata (A) Phase contrast picture o f a cryostat section showing the morphology of the electric tissue In the field, transversal secUons of two electrocytes are evidenced: open arrow points to one nucleus Scale bar = 20/~m (A') Immunofluorescence picture of the same field after m&rect lmmunolabelmg with the CTR 433 antl-Golgl antibody Golbn bodies appear as bright spots &spersed throughout the cytoplasm, sometimes m association with nuclei (open arrow). (B),(C) Electron microscopic views of Golgi regions showing the presence of Clsternae (G), vesicles and mlcrotubules (arrows) Scale bar = 0 5 pm

Mlcrotubules m Torpedo electrocyte (1) for ~mmunofluorescence experiments, columns were fixed at room temperature m 3% paraformaldehyde/0.1 M phosphate buffer, pH 7 4 ( = control conditions) Subsequently, they were impregnated m 25% sucrose (w/v) and rapidly frozen m melting Freon R-22 cooled by liquid nitrogen, (2) for electron microscopic experiments, columns were fixed at room temperature in 2 5% glutaraldehyde, 0 I% tannic acid/0 1 M cacodylate buffer, pH 7 4, postfixed with

153

1% o s m m m tetroxlde, dehydrated in a series of ethanol solutions and embedded m epoxy resin; (3) for taxol treatment, columns were lmtlally incubated for 60 m m at room temperature m M T G buffer 37 [100 m M MES (2-(N-morphohno)ethanesulfomc acid), pH 6 4, 1 m M ethyleneglycolbls(fl-ammoethylether)-N,N,N',N'-tetra-acetlc acid (EGTA), 1 m M GTP, 0 5 m M MgCI:] containing 5 # M of taxol (provided by D Gu6nard, Instltut des Substances Na-

Fig 2 0 r g a m z a t l o n of the mlcrotubule network in the electrocyte of T marmorata. ~ - T u b u h n antigens were detected by redirect lmmunofluorescence experiments using the B-5-1-2 antibody m cryostat sections of electric tissue (A) Distribution of mlcrotubules m an electrocyte fixed at room temperature The arrows point to axons running along one surface of the cell (B), (B') Double fluorescence experiment showing a part of electrocyte labeled with fluorescem lsothlocyanate (FITC)-conjugated ct-bungarotoxm m order to reveal the A C h R - n c h innervated m e m b r a n e domain of the cell (B) and antltubuhn antibody (B') Note the accumulation of microtubules in the innervated side of the cell leading to an asymmetric distribution of the mlcrotubule network. Microtubules also accumulate m the permuclear area (open arrow) and around focl (arrows) The non-innervated m e m b r a n e m B is outlined to vlsuahze the limit of the cell Scale bar = 20 t~m

B J J~SMIN et al

154

turelles, Centre Nauonal de la Recherche Scaentlfique, Glfsur-Yvette, France) and 5% sucrose, (4) for nocodazole treatment, columns were incubated for 30 min at room temperature an MTG buffer (w~thout GTP) containing 1 or 10 #M of nocodazole (Sigma, St. Lows, MO), (5) for the regrowth of the macrotubule network, columns were lmtially incubated m MTG buffer (without GTP) at 4°C for 45 rain, and subsequently allowed to recover at room temperature for various periods of Ume (0, 1, 2, 5, 10, 20 and 60man). Following treatments (3), (4) and (5), the columns were eather fixed as described m (1) or (2) For comparison, we have also fixed columns of electrocytes at 4°C for both immunofluorescenee and electron macrocop~c experiments.

Antibodws The following monoclonal antabodies were used (1) the B-5-1-2 anta-~-tubuiin antibody; 35 (2) the 6-11B-I antibody specific for acetylated ct-tubulin, 34 and (3) the CTR433 ant~-Golgl antibody directed against an mtralumenal epitope of the medial compartment of the Golg~ z0

lmmunofluorescence Frozen sectmns (4gin) were obtained by cuttmg the columns an a cryostat (SLEE, London) at -20°C. The sections were recovered onto ovatbumm-coated glass slides, mr dried and stored at -70°C until further analyses The presence of the vanous antigens was detected by redirect ammunofluorescence expenments. After an lnmal wash in phosphate-buffered saline (PBS), pH 7.4, nonspecific binding was blocked by premcubating the sections an 5% decomplemented goat serum and 1% bovine serum albumin in PBS for 15 min. Sections were then incubated with one of the monoclonal antabodies for 60-120min Tetramethylrhodamine (TMR)-conjugated goat anti-mouse IgG (Kirkegaard & Perry Laboratories, Gaithersburg, MD) was used as a second antibody. For the detection of the Golg~ anUgens, the seetaons were first permeabflized with 0.1% Tnton X-100 in PBS for 5ram. In all of these experiments, the AChR-nch membrane of the electrocytes was identified by labeling wath fluorescein isothiocyanate (FITC)-conjugated ~-bungarotoran (1/ag/ml, Sigma). Photomicrographs were obtained wath a Leitz Aristoplan photomicroscope equipped with epittuorescence illummauon (filters for FITC and TMR), using Plan x 63 (N.A 1 40) and x 100 (N.A. 1 32) ammersaon optacs. Kodak T-Max films were used and set at 800 ASA, and developed accordingly

Electron rnwroscopy For the ultrastruetural observatton of the Golg~ apparatus and of microtubules, columns of electrocytes fixed and embedded as described above were sectioned, and the ultrat~n sections were stained w~th lead citrate and uranyl acetate. Phflips EM 410 and EM 400 electron microscopes were used to observe the sections, and micrographs were obtained on Kodak EM 4489 films RESULTS

Distributton o f the Golgt apparatus and microtubules m T. m a r m o r a t a electrocyte The distribution of the Golgi was followed m the electrocyte with the CTR433 antibody previously

used to localize the Golgi in several tissues and species from chicken to human This antibody has been reported to label the medial Golga cisternae an muscle cells? ° In the electrocyte, It gave a punctate staining scattered throughout the medial plane of the cells (Fig. 1A). Also, labeled spots were seen close to the nuclei. At the electron microscope level, typical Golgi bo&es (Fig IB) were observed at regularly spaced intervals. They always corresponded to nncrotubule-nch areas (foci) (Fig. 1B, C). A possible reason for the fact that mlcrotubules were not formerly observed in the electrocyte (see, for example, Ref. 48) may be related to the use of fixation procedures in the cold. To curtail an eventual coldlabllity o f m l c r o t u b u l e s i n T marmorataelectrocytes, columns o f electrocytes were fixed at r o o m temperature ( = control conditions) with or without taxol, an agent known to promote the stability of microtubules. 37'38 Immunofluorescence experiments using the B-5-1-2 antibody, specific for all ct-tubulins, 35 strongly decorated the mIcrotubule network m the cytoplasm of electrocytes subjected to either one of these experimental treatments (Fig. 2). Throughout the cytoplasm, strongly labeled foci were consistently observed which, together with individual macrotubules, formed a meshline an the medial plane of the cell (Fig. 2A). Some of the foci were located in the vicinity of nuclei around which mlcrotubules were present (Fig. 2B). The microtubules tended to accumulate more closely to the innervated membrane domain than to the non-innervated one (Fig 2B). At the electron microscope level, m both control and taxol-treated electrocytes, we observed rmcrotubules surrounding nuclei (not shown) and numerous asters (Fig. 1B, C), which most likely correspond to the foci observed in ~mmunofluorescence experiments. In the asters, Golgi cisternae and vesicles were consistently observed in addition to the mlcrotubules Yet, despite extensive electron microscope observations, centrloles were never found an these regions.

Interactions between mtcrotubules, veswles and the postsynaptic membrane A striking feature revealed by our electron microscope studies is the observation that mlcrotubules converged towards the bottom of the folds of the postsynaptic membrane (Fig. 3). In some cases, microtubules appear to terminate at the membrane level. In ad&tlon, at the bottom of the folds, we consistently observed an accumulation of vesicles (mostly non-coated) for which, an most instances,

Fig 3. Electron microscopic visuahzataon of the macrotubules at the innervated stde of the electrocyte Most of the microtubules accumulated in the postsynapuc area of the cell (arrows in A), where they often converge to the bottom of the folds (D-F) Some microtubules appear to terminate at the membrane level (D and E) whereas others (open arrows an A and F), observed in cross-section, contacted the membrane tangentaally Clusters of non-coated vesicles (60-100 nm m diameter) were consistently observed at the bottom of the folds (D-F), often m direct association with microtubules (B,D,F). In C, vesicles were aligned along a membranous casternae at the bottom of a fold (A, C and D are from Ussue fixed at room temperature and B, E and F are from taxol-treated tassue ) NT. nerve terminal Scale bars = 0 5/~m

Mlcrotubules m Torpedo electrocyte

157

an apparent assocxatlon wxth the mlcrotubules was evident (Fig 3). The vesicles appeared as clusters tightly packed along the mlcrotubules which contact the postsynaptlc membrane. These images were never observed outside the bottom of the folds nor were they seen m conditions which disrupt orgamzatlon of the mlcrotubules (cold and nocodazole treatments) mdicatxng (1) the specific locahzatlon of these interactions and (11) the strict dependence of these vesicle accumulations upon

Some coated vesicles were observed m these experlments. At variance with the &strlbutlon of the non-coated vesicles at the level of the postsynaptlc membrane (see above), the coated ones were found scattered along the entire postsynapt~c membrane domain without any apparent pnvdeged dlstnbut~on (not shown)

m~crotubules Moreover, m these expenments, we never observed a similar association of m~crotubules with the non-innervated membrane domain Thus, m agreement with the lmmunofluorescence experiments, the distribution of mlcrotubules m the electrocyte is clearly asymmetric. Non-coated veslcles were also observed within the non-innervated membrane domain, particularly m regions located close to nuclei (not shown),

The stabihty properties of the mlcrotubules were first investigated by fixing columns of electrocytes at 4 C and by subsequent processing for lmmunofluorescence and electron microscope experiments Under these condmons, at both the optical (F~g 4A) and ultrastructural levels (not shown), mlcrotubules could not be identified m the cytoplasm of the electrocyte However, ~mmunostalnmg was observed m axons.

Dynamtcsofrmcrotubules lr/T marmorata electrocvte

Fig. 5 Detection of acetylated mlcrotubules m the electric tissue In both control (A') and taxol-treated (B') electric tissue, acetylated mlcrotubules were detected only m axons and at the nuclear penphery (open arrow in B') Left panels correspond to ~-bungarotoxm labehng and nght panels to 6-11B-I ant~-acetylated tubuhn labehng Scale bar = 20 pm

158

B J JASMINet al

In a second series of experiments, the stablhty of the mlcrotubules was challenged by exposure to increasing concentrations of the depolymenzmg agent nocodazole At the lowest concentration tested (1/~ M), most cytoplasmic mlcrotubules were depolymenzed (compare Fig. 4B, C to Fig. 2) except around nuclei and at some foci. At the electron rmcroscope level, no mlcrotubules were seen close to the postsynaptlc membrane following this treatment (not shown), thus ruling out a differential resistance of the microtubules associated with the postsynapUc membrane. In adjacent axons, however, the microtubule network persisted. At the highest nocodazole concentration (10/~ M), even the perlnuclear network

was partially &srupted (Fig. 4D), as shown by the punctate labehng observed The 6-1 1B-1 monoclonal antibody &rected against the acetylated isoform of ~t-tubuhn34 was used to determine whether subsets of mlcrotubules were posttranslationally modified. In control and taxol-treated electrocytes, only the pennuclear microtubules were labeled along with the axons, m&cating that they contamed acetylated ~t-tubuhn (Fig. 5). Electron microscope immunogold experiments using the 6-1 1B-I antibody confirmed that only axons were labeled at the level of the innervated membrane (not shown) To determine the nucleation sites of microtubules m the electrocyte, regrowth experiments were carried

Fig 6 Mlcrotubute regrowth expenments After complete depolymenzatlon of rmcrotubules m the cold (see Experimental Procedures and Fig 4), the columns of eleetrocytes were allowed to recover at room temperature for 1 nun (A), 2 mm (B) or 5 mm (C) and processed for lmmunofluoresceneewith the B-5-1-2 antibody. Regrowth proceeded from the focl of the electrocyte towards the cell pertphery Scale bar = 20 #m

Mlcrotubules m Torpedo electrocyte out. After a 45 rain cold exposure, all the microtubules present m the electrocyte were depolymerlzed (see Fig. 4A). The t~me-course study (recovery time ranging from 0 to 60 mm) revealed a timedependent reconstltut~on of the microtubule network (see example in Fig. 6) which was almost complete after 20 mm. The regrowth proceeded radially from the regularly spaced focl towards the periphery of the cell. Electron m~croscopy observations of electrocytes at low recovery times (0, 1 and 2 min) demonstrated that no mlcrotubules were present at the postsynapt~c membrane. Taken together, these observations md~cate that there ~s no nucleation at the level of the postsynaptlc membrane and, therefore, that the mlcrotubules of the electrocyte have their plus ends oriented towards the plasmalemma DISCUSSION In the present work, we have examined the distnbutlon and the subcellular orgamzatlon of the Golgl and mlcrotubules, which compose the secretory pathway, m Torpedo marmorata electrocyte Our results indicate that the electrocyte contains numerous focallzed Golgi bodies that are dispersed throughout the cytoplasm and that it d~splays an asymmetric m~crotubule network which preferenUally projects towards specmhzed domains of the postsynaptlc membrane Furthermore, we demonstrate the presence m these domains of clusters of non-coated vesicles

Asymmetrtcal dtstrtbution of the secretory pathway In the electrocyte of T. marmorata, numerous focal Golgl bodies are found scattered throughout the cytoplasm and often colnclde with nuclei. Such polarlzed Golgl orgamzatlon resembles that observed m the postsynapt~c sarcoplasm of ch~ck skeletal muscle fiber2° but differs from that displayed by embryomc myotubes and denervated muscle fibers which appears, m these cells, to be permuclear. 2°'3°'45 Our lmmunofluorescence and electron microscope experiments reveal for the first time an elaborate network of mlcrotubules m T marmorata electrocyte, They also show that this network is detected only under appropriate experimental conditions, 1 e fixed at room temperature or exposed to low concentrataon of taxol (see Experimental Procedures). The network that we observe is composed of m~crotubules which surround nuclei and radmte from asters regularly spaced throughout the cytoplasm. This network occupies the medial plane of the cell from which lndlwdual mlcrotubules project to the postsynapt~c membrane leading to its asymmetrical distribution, Unhke most mterphaslc cells, the m~crotubule network of the electrocyte does not originate from a single m~crotubule-orgamzmg center but, rather, from a multlphc~ty of orgamzmg centers which do not contain centnoles, We observed that the majority of the mlcrotubules m the electrocyte are readily depolymenzed by cold NSC 4~ 1- F

159

and nocodazole exposure, and do not contain acetylated ct-tubuhn. These unstable mlcrotubules may thus represent a dynamic population having a high turnover rateY '26 These properties may be related to the observation that the bulk of the mlcrotubules m the electrocyte are not associated with Mlcrotubule Associated Proteins 1, 2, 5 and tau (Jasmln B J., Changeux J.-P. and Cartaud J., unpubhshed observatlons). Conversely, as in many systems, including the postsynaptlc domain of skeletal muscle fiberfl2 another population of m~crotubules was identified on the basis of its stability properties In the electrocyte, the perlnuclear mlcrotubules, along with those located m Golgl regions, contain acetylated ct-tubuhn and exhiba resistance to low concentrations ofnocodazole, and as such, they represent a stable, less dynamic, population. 39 As for a variety of cells, the stable mlcrotubules in the electrocyte assume a specific location and exhibit regional compartmentallzatlon. Therefore, the asymmetry of the mlcrotubule network m the electrocyte does not rely upon an extensive population of stable and dlfferentmted microtubules but rather depends on the actual dlstrlbutton of the bulk of mlcrotubules Since some of the mlcrotubules which project towards the bottom of the postsynapnc membrane folds appear to terminate at the cytoplasmic face of this membrane domain, possibly wa their plus ends (see Figs 3 and 6), the specific configuration of the m~crotubule network m the electrocyte may result from a mechanism of polymerization from discrete nucleaUon sites (foc0 and selective capture at the postsynaptlc membrane (for discussion, see Ref. 25) As a consequence, the asymmetrical d~stributlon of the m~crotubule network m the electrocyte may confer a functional polarization of the cytoplasm

Identzficatton of speczalized domains wtthm the postsynapttc membrane Consistently, we observed clusters of non-coated vesicles at the level of specmlized domains of the innervated membrane, located at the bottom of the postsynapt~c folds The occurrence of vesicles at this region depended on the integrity of the m~crotubule network, since such images were never observed m electrocytes formerly exposed to cold or nocodazole treatments Their dmmeter (60-100 nm), the absence of an electron-dense coat 0.e clathrm) and the observation that the cytoplasmic domain of the proterns they carried 0 e AChR; Jasmm B. J Changeux J.-P and Cartaud J , unpubhshed observations) is exposed and accessible at the vesicle surface, suggest that they are post-Golg~ vesicles as observed m other cells such as Madm-Darby Canine K~dney (MDCK) cells, 3'~6Baby Hamster Kidney Cells ~l and neurons.3~ The postsynaptlc membrane of the neuromuscular junctxon, and to a lesser extent that of the electrocyte, can be subdwlded into two d~stlnct domains visualxzed by electron m~croscopy, autoradlography and lmmunocytochemlstry Underneath the nerve ending, ,

160

B J JASMINet al

the crest of the folds appears denser to electrons and is composed primarily of the AChR and 43,000 mol wt protein 5,~3.~8,z7,29.33,4~42Conversely, the troughs of the folds lack both the AChR and the 43,000 mol. wt protein. Despite the recent demonstration of the presence of voltage-gated Na + channels and ankyrin at their level m the neuromuscular junction,~3 their function remains to be elucidated Here, we provide ewdence that these troughs represent a specialized domain of the postsynaptlc membrane where microtubules converge and where post-Golgl vesicles accumulate. Consequently, the troughs of these folds are likely engaged in the local insertion of newly synthesized constituents of the postsynaptlc membrane. Our data, thus, specifically attribute a distinct function to these differentiated areas of the postsynapt~c membrane Conclusion and perspective

The observation that both the innervated and the non-innervated membranes &splay speciahzed areas most likely involved in the reception of post-Golgx vesicles raises the question of how these vesicle accumulations are generated, i.e. via bulk flow vs signal-mediated transport. The peculiar organization of the secretory pathway of the electrocyte is reminiscent of that of MDCK cells (see Ref 1). Since the latter are known to be involved in the intracellular routing of membrane proteins,47 we may assume that the secretory pathway of the electrocyte shares cornmon functional features with that of polarized epithehal cells and, perhaps, even with neurons) z Therefore, the vesicles, parUcularly those observed at the level of the postsynaptlc membrane domain, may be segregated intracellularly m the GolgI and vectonally delivered by way of a microtubule-based organelle transport mechanism to specific domains of the cell surface. This view is coherent with the notion that the secretory pathway is involved in the selec-

tlve delivery of vesicles to distinct area~ of the eel} surface, z4 The identification of the content of the vesicles accumulated both at the innervated and non-innervated membranes will determine whether membrane proteins of the electrocyte are indeed subjected to sorting and targeting mechanisms Recently, we have shown that the postsynaptic sarcoplasm of the neuromuscular junction represents a compartmentalized domain of the muscle fiber that also displays a polar Golgl 2° and a specialized microtubule network 22 These observations, together with the results of Fontame et a l , ~4L' Goldman and Staple ~9 and Brenner et al., 6 on the confined locallzataon of AChR subunlt gene transcripts at the level of the postsynaptlc nuclei, suggest that the postsynaptIc domain of the innervated muscle fiber represents a compartment specialized for the transcription, post-translational processing and transport of synaptlc proteins, in particular of the AChR, at the motor endplate level. The significant similarity displayed by the secretory pathway of both systems, along with the area covered by the postsynaptic membrane domain m the electrocyte, indicates that this cell may provide a useful model to investigate the cellular and molecular basis of sorting and targeting mechanisms of synaptIc proteins Acknowledgements--We are especmlly indebted to Dr

M Bornens for stimulating discussion, helpful suggestions throughout the course of this work and for cntacallyreading the manuscript We are also grateful to Drs E L BenedettI and R Couteaux for discussion, Dr G lhpemo for providIng us with antibodies and M A Ludosky for expert technical assistance The support of the Assocaatlon Francalse contre les Myopathies, the Coll6ge de France, the Centre National de la Recherche Se~entifique,the Umverslt~ Pans VII and the Institut National pour la Santd et la Recherche M6dicale is gratefully acknowledged B J J is supported by a Postdoctoral Fellowship in Sciences from NATO (via the Natural Sciencesand Engmeenng Research Council of Canada)

REFERENCES

1 Bacallao R., Antony C, Dottl C, Karsentl E., SteLzer E. H. K. and Stmons K. (1989) The subcellular organization of Madin-Darby Camne Kidney ceils during the formation of a polarized epithelium J Cell Biol. 109, 2817-2832. 2. Bauman A., Changeux J P. and Benda P. (1970) Purification of membrane fragments derived from the non-excitable surface of the eel electroplax Fedn Fur. blochem. Socs Lett. 8, 145-158 3. Bennett M K., Wandinger-Ness A. and Simons K. (1988) Release of putative exocytic vesicles from perforated MDCK cells. Eur. molec Biol. Org. J 7, 4075~1085. 4. Bourgeois J. P., Popot J. L., Ryter A. and Changeux J P. (1973) Consequences of dcnervatlon on the dlstnbutlon of the cholinergic (mcotmic) receptor sites from Electrophorus electr~cus revealed by high resohiuon autoradlography Brain Res. 62, 557-563. 5. BourgaOlSJ. P., Popot J. L., Ryter A and Changeux J. P. (1978) Quantitauve stu&es on the locahzation of chohnerglc receptor protein in the normal and denervated electroplaque from Electrophorus electricus. J. Cell B~ol. 79, 200-216 6. Brenner H. R., Witzemann V. and Sakmann B. (1990) Imprinting of acetylcholIne receptor mRNA accumulation m mammalian neuromuscular synapses Nature, Lond. 344, 544-547 7. Carr C, Fischbach G D. and Cohen J. B. (1989) A novel 87,000-Mr protein assocmted with acetylchohne receptor in Torpedo electric organ and vertebrate skeletal muscle. J Cell Biol. 109, 1753-1764 8 Carr C, McCourt D and Cohen J B. (1987) The 43-kllodalton of Torpedo mcotm~c postsynaptlc membrane purification and primary structure Bzochemtstry 26, 7090-7102. 9. Cartaud A., Courvalin J. C., Ludosky M. A. and Cartaud J (1989) Presence of a protein mamunologlcallyrelated to lamin B in the postsynaptic membrane of Torpedo marmorata electrocyte J Cell B, ol 109, 1745-1752 10. Changeux J. P. (1981) The acetylchohne receptor: an "allostenc" membrane protein Harvey Leet. 75, 85-254 ii Curtis I. de and Slmons K. (1989) Isolation of exocytic carrier vesicles from BHK cells. Cell 58, 719-727

Mlcrotubules m Torpedo electrocyte

161

12 Dottl C G and Slmons K (1990) Polarized sorting of viral glycoprotems to the axons and dendrites of hlppocampal neurons m culture Cell 62, 63-72 13 Flucher B E and Darnels M P (1989) Distribution of Na ÷ channels and ankyrm m neuromuscular junctions is complementary to that of acetylchohne receptors and the 43 kd protein Neuron 3, 163-175 14 Fontame B and Changeux J P (1989) Locahzatlon of mcotmlc acetylchohne receptor ct-subumt transcripts during myogenesls and motor endplate development m the chick J Cell Bzol. 108, 1025-1037 15 Fontame B, Sassoon D , Buckingham M and Changeux J P (1988) Detection of the mcotmlc acetylchohne receptor • -subumt mRNA by m sltu hybn&zatlon at neuromuscular junctions of 15-day-old chick striated muscles Eur molec Btol Org J 7, 603~509 16 Frad D E , Mudd J , Shah V , Carr C , Cohen J B and Merhe J P (1987) cDNA's for the postsynaptlc 43kDa protein of Torpedo electric organ encode 2 proteins with &fferent carboxyl termini Proc natn Acad S
162

B J JASMIN et al

47 Van Zeij1 M J A H and Matlm K. S (1990) Mlcrotubule perturbation inhibits mtracellular transport oI an apical membrane glycoprotem m a substrate-dependent manner an polarized Madm-Darby canane kadney epithelial cells Cell Regulatton I, 921-936 48 Walker J H , Boustead C M., Watzemann V., Shaw G , Weber K. and Osborn M (1985) Cytoskeletal protems at the chohnerglc synapse' dastnbutlon of desmin, actin, fodnn, neurofilaments, and tubulm m Torpedo eleclnc organ Eur J Cell Bwl 38, 123-133 49. Wtute M. M. and Mdler C (1979) A voltage gated anmn channel from electric organ J Cell Baol 254, 10161-10166 50 Woodruff M L , Theriot J and Burden S J (1987) 300kD subsynaptic protean copurlfies wath acetylchohne receptor-rich membrane and as concentrated at neuromuscular synapses J Cell Biol 104, 939-946 (Accepted 6 February 1991 )