Rapsyn Variants in Ciliary Ganglia and Their ... - Wiley Online Library

2 downloads 0 Views 262KB Size Report
(Burns et al., 1997). ... 436 and 1,089–1,069, respectively, of chick rapsyn (Burns et ...... Yang S.-H., Armson P. F., Cha J., and Phillips W. D. (1997) Clustering.
Journal of Neurochemistry Lippincott Williams & Wilkins, Inc., Philadelphia © 1999 International Society for Neurochemistry

Rapsyn Variants in Ciliary Ganglia and Their Possible Effects on Clustering of Nicotinic Receptors William G. Conroy and Darwin K. Berg Department of Biology, University of California, San Diego, La Jolla, California, U.S.A.

a7 subunits (Couturier et al., 1990; Schoepfer et al., 1990; Anand et al., 1993; Conroy and Berg, 1998). Such receptors (a7-nAChRs) have a high relative permeability to calcium (Bertrand et al., 1993; Seguela et al., 1993; Vernino et al., 1994) and are capable of influencing a variety of calcium-dependent events. Examples include neurotransmitter release (McGehee et al., 1995; Gray et al., 1996; Coggan et al., 1997; Aramakis and Metherate, 1998; Guo et al., 1998; Li et al., 1998; Radcliffe and Dani, 1998), second messenger cascades (Vijayaraghavan et al., 1995), and neurite extension (Chan and Quik, 1993; Pugh and Berg, 1994; Fu and Liu, 1997; Fu et al., 1998). The receptors can also participate in postsynaptic signaling (Zhang et al., 1996; Ullian et al., 1997; Frazier et al., 1998; Chang and Berg, 1999; Hefft et al., 1999). Which effect a7-nAChR activation produces in a given situation is likely to depend heavily on the exact positioning of the receptor. A preparation in which the subcellular distribution of a7-nAChRs has been studied in some detail is the chick ciliary ganglion. The ganglion contains ciliary and choroid neurons in about equal numbers (Landmesser and Pilar, 1974), and at the end of embryogenesis, the neurons average 106 a7-nAChRs/cell (Chiappinelli and Giacobini, 1978; Corriveau and Berg, 1994). The receptors appear to be excluded from postsynaptic densities on the neurons (Jacob and Berg, 1983; Loring et al., 1985) and instead are configured in clusters on the cell surface (Wilson Horch and Sargent, 1995). The clusters have been shown on ciliary neurons to represent mats of folded somatic spines heavily endowed with a7-nAChRs (Shoop et al., 1999). If the major cytoplasmic loop of the a7 gene product is replaced with that of a3, the resulting receptors can be redirected in vivo to postsynaptic den-

Abstract: Nicotinic acetylcholine receptors (nAChRs) containing the a7 gene product can influence a range of cellular events in neurons, depending on receptor location. On chick ciliary neurons, the receptors are concentrated on somatic spines, but little is known about mechanisms responsible for sequestering them there. Rapsyn is a 43-kDa protein essential for clustering nicotinic receptors at the vertebrate neuromuscular junction. RTPCR confirmed previous studies showing that the chick ciliary ganglion expresses rapsyn transcripts, including several splice variants lacking part or all of exon 2. Heterologous expression of rapsyn constructs, together with nicotinic receptor constructs, shows that chicken fulllength rapsyn can induce clustering of both muscle and neuronal nicotinic receptors. Splice variants lacking one or both leucine zipper motifs of exon 2 are unable to cluster the receptors, though, like full-length rapsyn, they cluster themselves. Immunological analysis demonstrates the presence of full-length rapsyn in chick muscle extracts but fails to detect either full-length or splicevariant versions of rapsyn at significant levels in ganglion extracts. The results suggest that rapsyn does not cluster a7-nAChRs on ciliary neurons in any way similar to that of receptors at the neuromuscular junction where rapsyn and the receptors are present in approximately equimolar amounts. Key Words: Rapsyn—Nicotinic receptor— a7—Ciliary ganglia—Neurons. J. Neurochem. 73, 1399 –1408 (1999).

Tethering neurotransmitter receptors at appropriate sites on the cell surface is a critical feature of synapse formation. For most neurons, this represents a complex challenge because of the multiple synaptic contacts they receive and the diversity of receptor subtypes they express. Recently, considerable progress has been made in identifying some of the submembrane components likely to influence the distribution of several types of neurotransmitter receptors at synapses, including glutamate, GABAA, and glycine receptors (for reviews, see Kirsh et al., 1996; Craig, 1998; Craven and Bredt, 1998). In contrast, almost nothing is known about components controlling the distribution of nicotinic acetylcholine receptors (nAChRs) on neurons. One of the most abundant neuronal nAChRs is a species that binds a-bungarotoxin (aBgt) and contains

Received April 19, 1999; revised manuscript received May 28, 1999; accepted May 31, 1999. Address correspondence and reprint requests to Dr. D. K. Berg at Department of Biology, 0357, University of California, San Diego, 9500 Gilman Dr., La Jolla, CA 92093, U.S.A. Abbreviations used: aBgt, a-bungarotoxin; GFP, green fluorescent protein; mAb, monoclonal antibody; nAChR, nicotinic acetylcholine receptor; PBS, phosphate-buffered saline; PBS-TX, PBS plus 0.5% Triton X-100; SDS-PAGE, sodium dodecyl sulfate–polyacrylamide gel electrophoresis.

1399

1400

W. G. CONROY AND D. K. BERG

sities (Williams et al., 1998). How native a7-nAChRs are normally concentrated on somatic spines is unknown. Rapsyn is a peripheral membrane protein known to be essential for clustering muscle nAChRs at the vertebrate neuromuscular junction (for review, see Sanes and Lichtman, 1999). It is thought to interact directly with the receptor and to be present in approximately equimolar amounts with it. Rapsyn transcripts, together with splice variants, are expressed in the chick ciliary ganglion (Burns et al., 1997). Co-transfection experiments with a mammalian rapsyn construct and neuronal nAChR genes show that the full-length protein can cluster the receptors (Feng et al., 1998; Kassner et al., 1998). Disruption of the mammalian rapsyn gene, however, caused no obvious alteration in the distribution of at least one neuronal nAChR subtype on a population of ganglionic neurons (Feng et al., 1998). The experiments reported here were undertaken to assess the likelihood that rapsyn or a splice variant thereof is responsible for the sequestering of a7-nAChRs on somatic spines, as it is for the concentration of nAChRs at the neuromuscular junction. MATERIALS AND METHODS RT-PCR Total RNA was obtained from ciliary ganglia and pectoral muscle dissected from embryonic day 17 chicks using QIAshredder and RNeasy spin columns according to the manufacturer (Qiagen, Santa Clarita, CA, U.S.A.). RNA was reverse transcribed using a ThermoScript RT-PCR System (Life Technologies, Gaithersburg, MD, U.S.A.) with 0.4 mg of RNA, random hexamers for priming, and a reaction temperature of 50°C. The PCR reactions included 5 ml of cDNA as template and primers CACTGCTCATGCAGAGATGG and GCTCATGGCAGAATCGTACCG corresponding to nucleotides 417– 436 and 1,089 –1,069, respectively, of chick rapsyn (Burns et al., 1997). The PCR was performed with a hot start modification and touchdown cycling parameters with an initial annealing temperature of 65°C and a 1°C decrease in annealing temperature after every two cycles until 51°C was reached; this was followed by 10 cycles at 51°C (Dieffenbach and Dveksler, 1995). PCR products were separated on agarose gels and visualized by staining with ethidium bromide. Products from the RT-PCR were subcloned into pGEM-T Easy (TA cloning vector; Promega, Madison, WI, U.S.A.) and sequences confirmed by digestion with the restriction endonucleases EcoRI, PstI, HindIII, KpnI, and NcoI and by sequencing (Retrogen, San Diego, CA, U.S.A.).

Transfections QT-6 cells were transiently transfected using the calcium phosphate method as described previously (Kassner et al., 1998). Expression constructs encoding the chick a3-, b4-, and a7-nAChR subunits and a green fluorescent protein (GFP) construct (Kassner and Berg, 1997; Kassner et al., 1998) and the mouse rapsyn and muscle nAChR subunits a, b, g, and d (Phillips et al., 1991a; Blount and Merlie, 1989) have been previously described. The chick rapsyn expression construct (Rapsyn) was made by engineering an NcoI site directly after the coding sequence of chick rapsyn by PCR using the primers C43P1 (AAGTTCCATGAATGCGTGGAG) and C43P2 (CTAGCCATGGTCACACGTATCCGGGT) and as template the plasmid Ch43K.1 (Burns et al., 1997). A BsmI–NcoI frag-

J. Neurochem., Vol. 73, No. 4, 1999

ment of the resulting PCR product was combined with an NsiI–BsmI fragment of Ch43K.1 after blunting the NsiI site; it was cloned into an SmaI–NcoI fragment of the vector RSVLTR-43K, thus removing the mouse rapsyn sequences and replacing them with the chick. The expression constructs encoding the splice variants of rapsyn lacking exon 2 (RapEx2) and lacking the first leucine zipper of exon 1 (RapZ1) were made by using an overlap extension PCR method (Dieffenbach and Dveksler, 1995). In the first PCR reaction, the primers cg4.p1 (CATGGTAACGATGAG) and cg4.p3 (GATATGCTGAAGGACTATGAGAAAGC) were used for RapEx2, and the primers cg4.p1 and cg11.p3 (AGATATGCTGAAGGGTACCACTGTGAG) were used for RapZ1 with the template Rapsyn. In the second PCR reaction, the products of the first PCR were used as templates with the primers cg4.p1 and C43P2. SphI–BstEII fragments of the resulting PCR products were cloned into an SphI–BstEII fragment of the plasmid Rapsyn, giving the constructs RapEx2 and RapZ1. The expression construct for the variant of chick rapsyn lacking the second leucine zipper (RapZ2) was made by taking an EcoNI–BstEII fragment of the RT-PCR product cloned into pGEM-T Easy and cloning it into an EcoNI–BstEII fragment of Rapsyn. Combinations of plasmids for co-transfection used equal amounts of each plasmid.

Immunofluorescent staining Immunofluorescent localization of nAChRs and rapsyn in transfected cells was conducted as previously described (Kassner et al., 1998). Cells grown on coverslips were fixed with 2% paraformaldehyde in 150 mM sodium phosphate (pH 7.4) for 20 min at room temperature. When rapsyn expression was examined, the fixative contained 2% paraformaldehyde in phosphate-buffered saline (PBS; 150 mM sodium chloride plus 10 mM sodium phosphate, pH 7.4) containing 10 mM sodium m-periodate, 100 mM lysine, and 0.1% saponin. After rinsing in PBS, the cells were incubated overnight at 4°C with primary antibodies in PBS containing 0.5% Triton X-100 (PBS-TX) and 10% donkey serum, rinsed again with PBS, incubated with secondary antibodies for 1 h at room temperature, and rinsed with PBS. Coverslips were mounted on glass slides with Vectashield (Vector Laboratories, Burlingame, CA, U.S.A.), and images were analyzed with a Noran Odyssey confocal laser scanning microscope. The following primary antibodies were used: monoclonal antibody (mAb) 35 (which recognizes a1, a3, and a5) for a3b4- and a1b1gd-nAChRs; goat anti-a7 antiserum (Santa Cruz Biotechnology, Santa Cruz, CA, U.S.A.) for a7-nAChRs; and mAbs 1234A and 1579A (Bloch and Froehner, 1987) for rapsyn. Secondary antibodies included Cy3-donkey anti-mouse, rat, and goat antisera (Jackson ImmunoResearch, West Grove, PA, U.S.A.) for their respective primary antibodies. In some experiments, surface nAChRs were labeled by incubating live cells with mAb 35 for 1 h at 4°C in culture medium containing 20 mM HEPES (pH 7.4), rinsed with PBS, fixed in paraformaldehyde for 20 min, and then processed with secondary antibodies as described above. Hundreds of cells were examined for each labeling condition from multiple (4 – 12) staining experiments. Cells were excluded from the analysis if they were rounded up (poorly attached to the substratum) because of questions about their viability.

Detergent solubility assays The effects of rapsyn on the detergent solubility of nAChRs heterologously expressed in QT-6 cells were investigated as previously described using 125I-aBgt (Kassner et al., 1998).

RAPSYN VARIANTS IN CILIARY GANGLIA Anti-chick rapsyn mAbs Two new anti-rapsyn mAbs, Rap-1 and Rap-2, were made against a fusion protein (C43Kprot1B) containing the sequence of chick rapsyn from amino acid 38 to 412. The fusion protein was expressed in bacteria using a pRSET vector (Invitrogen, Carlsbad, CA, U.S.A.), purified from inclusion bodies, and used to immunize mice for mAb production as previously described (Conroy et al., 1992). Hybridomas were screened for reactivity with the immunizing peptide as well as staining of cells transfected with chick rapsyn. The mAbs Rap-1 and Rap-2 were used as dilutions of ascites fluid or were purified on protein G columns and coupled to Actigel (Sterogene Bioseparations, Carlsbad, CA, U.S.A.) at 2– 4 mg/ml. The epitopes of all the mAbs used were mapped to regions of chick rapsyn by testing for precipitation of [35S]Metlabeled fusion proteins translated in vitro and containing the amino acid sequences 38 – 412 (C43Kprot1B), 38 –224 (C43Kprot2), 38 –147 (C43Kprot3), 38 –114 (C43Kprot4), and 38 –97 (c43Kprot5). The mAbs Rap-1 and Rap-2 were able to immunoprecipitate all of the fusion proteins, indicating that their epitopes lay between amino acids 38 and 97 of rapsyn. The mAbs 1201C, 1579A, and 19F4A reacted with only the fusion protein containing the C-terminus of rapsyn, C43Kprot1B, indicating that their epitopes lay between amino acids 224 and 412. The mAb 22F10A did not react with any of the fusion proteins but did react with the full-length protein, indicating that its epitope is probably in the N-terminal portion of rapsyn and therefore not present in any of the fusion proteins. The expression of the fusion proteins as well as the specific activity of the [35S]Met-labeled products were normalized by precipitating with an mAb to the fusion protein tag found in all constructs; the mAb should have recognized all of the fusion proteins equally well. The immunoprecipitations were also confirmed by examining immunoblots of the expressed fusion proteins.

Immunoblots and immunoprecipitations Total lysates of ciliary ganglia and muscle were prepared by homogenizing tissue in 2% Triton X-100 in PBS and then adding 33 concentrated sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer. The samples were heated to 100°C for 2 min, centrifuged to remove insoluble material, and then subjected to SDS-PAGE and immunoblotting as previously described (Conroy and Berg, 1995). Nitrocellulose membranes containing the blotted material were probed with primary antibodies overnight at 4°C, washed with PBS containing 0.05% Tween 20, and developed with goat anti-mouse IgG antibodies coupled with horseradish peroxidase. Signals were visualized by enhanced chemiluminescence. Molecular mass markers for the blots included phophorylase B (97.4 kDa), bovine serum albumin (66.2 kDa), ovalbumin (45 kDa), and carbonic anhydrase (31 kDa) (lowrange standards; Bio-Rad, Hercules, CA, U.S.A.). An aliquot of the lysate before adding SDS sample buffer was taken for determining the number of 125I-aBgt binding sites. Aliquots were incubated with 10 nM 125I-aBgt in the absence and presence of 1 mM unlabeled aBgt for 1 h at room temperature; the samples were then filtered over GF/B filters (Whatman) treated with 0.5% polyethylenimine to collect membranes and solubilized nAChRs. All of the available mAbs against rapsyn were tested for their ability to immunoprecipitate rapsyn from chick muscle extracts. Aliquots of the mAbs containing ;1–5 mg of mAb were incubated with protein G–Sepharose (0.05 ml) for 2 h at room temperature, washed with PBS-TX, and then incubated

1401

with 1-ml extracts of muscle. The muscle extracts were prepared by incubating muscle membranes for 2 h with 2% Triton X-100 in 50 mM sodium phosphate buffer (pH 7.4) containing the following protease inhibitors: iodoacetamide (0.4 mM), benzamidine (5 mM), phosphoramidon (5 mg/ml), soybean trypsin inhibitor (10 mg/ml), leupeptin (10 mg/ml), pepstatin A (20 mg/ml), EDTA (5 mM), EGTA (5 mM), aprotinin (2 mg/ml), and phenylmethylsulfonyl fluoride (1 mM). Insoluble material was removed by centrifugation, and the recovered supernatant fraction was incubated overnight with mAb–protein G–Sepharose. The resins were washed with PBS-TX four times, and material was eluted with SDS-PAGE sample buffer and processed for SDS-PAGE and immunoblotting. Analysis of lysate before and after depletion indicated that mAb Rap-1 was best for immunoprecipitation, depleting half or more of the rapsyn in a single pass. Accordingly, mAb Rap-1 was used routinely to immunoprecipitate rapsyn from all tissue extracts prepared as described above with one difference: Muscle samples were prepared from muscle membrane preparations, whereas ganglion samples were prepared from whole ganglia. Muscle membrane preparations were generated by homogenizing muscle tissue in PBS containing phenylmethylsulfonyl fluoride, leupeptin, aprotinin, and pepstatin at concentrations used for the solubilization buffer. Membrane fragments were collected by centrifugation at 20,000 g for 30 min, washing by resuspending in fresh buffer, and recentrifuging and then resuspended at 4 ml/g of tissue in buffer and stored frozen at 280°C until use.

Materials White Leghorn embryonated chick eggs were obtained locally and maintained at 39°C in a humidified incubator (Boyd et al., 1991). aBgt was obtained from Biotoxins (St. Cloud, FL, U.S.A.) and iodinated to a specific activity of 2–5 3 1017 cpm/mol using chloramine T. Unless otherwise indicated, all other reagents were from Sigma (St. Louis, MO, U.S.A.). The chick rapsyn cDNA construct Ch43K.1 was provided by Dr. Joseph Margiotta (Ohio Medical College, Toledo, OH, U.S.A.). The QT6 cells, mammalian rapsyn cDNA construct RSV-LTR43K, and mouse muscle nAChR genes were provided by Drs. Joshua Sanes, John Merlie, and Elizabeth Apel (Washington University, St. Louis, MO, U.S.A.). The GFP cDNA construct was obtained from Dr. Christine Holt (Cambridge University, Cambridge, U.K.). The mAbs 35 and 270 were provided by Dr. Jon Lindstrom (University of Pennsylvania, Philadelphia, PA, U.S.A.); mAbs 1201C, 1579A, and 1234A were provided by Dr. Stanley Froehner (University of North Carolina, Chapel Hill, NC, U.S.A.); and mAbs 19F4A and 22F10A were provided by Dr. Jonathan Cohen (Harvard Medical School, Boston, MA, U.S.A.).

RESULTS Alternatively spliced rapsyn transcripts in ciliary ganglia and muscle Rapsyn transcripts in chick ciliary ganglia were analyzed and compared with those in skeletal muscle by using RTPCR. PCR primers were designed to amplify a region encoding the amino-terminal half of the rapsyn protein. Based on the mammalian gene (Gautam et al., 1994), this region should include all of exons 2–4 and portions of exons 1 and 5 and is alternatively spliced in neurons and muscle (Burns et al., 1997; Feng et al., 1998). The same reproducible pattern of RT-PCR products was obtained J. Neurochem., Vol. 73, No. 4, 1999

1402

W. G. CONROY AND D. K. BERG

FIG. 1. Rapsyn transcripts in chick ciliary ganglia. A: RT-PCR generated a major product of ;670 bp (top arrow) and minor products of ;520 (middle arrow) and 330 bp (bottom arrow) in both ganglion (middle lane) and muscle (right lane) extracts. Omission of reverse transcriptase from the reaction provided a negative control with no PCR products (left lane). B: Sequence analysis of cloned PCR products indicated that the major species corresponded to undeleted rapsyn (Rapsyn). The PCR band of 520 bp contained two species: one lacking the first leucine zipper motif of exon 2 (RapZ1) and the other lacking the second leucine zipper therein (RapZ2), based on the mammalian gene structure (Gautam et al., 1994). The PCR band of 330 bp lacked all of exon 2 (RapEx2). The upper bar indicates the eight exons of the rapsyn gene (Ex 1– 8) and the corresponding amino acid positions demarcating the borders. LZ, leucine zipper motif.

with RNA from both embryonic day 17 ciliary ganglia and embryonic day 17 pectoral muscle (Fig. 1A). Of the three bands seen, the largest and most intense corresponded in size to the predicted 672-base pair sequence encoding the rapsyn segment without deletions. The two smaller bands corresponded in size to alternatively spliced rapsyn transcripts previously described in ciliary ganglion RNA as having deletions of ;150 and 340 base pairs in this region (Burns et al., 1997). To confirm that the RT-PCR products were rapsyn transcripts, the bands were excised, subcloned into a TA cloning vector, and analyzed by restriction digestion and DNA sequencing. For the largest band, six clones from both ciliary ganglia and muscle were analyzed, and all had a restriction digest pattern corresponding to rapsyn. DNA sequencing of two such clones confirmed the presence of an uninterrupted rapsyn sequence. Clones derived from the middle-sized RT-PCR bands of both ciliary ganglion and muscle RNA yielded two different restriction digest patterns. Sequencing several clones from each group identified two species; each had a deletion of ;150 nucleotides within exon 2, but they were nonoverlapping. The first variant lacked the region encoding one leucine zipper motif, whereas the second lacked the sequence for the other leucine zipper (Fig. 1B). Clones derived from the smallest of the three RT-PCR bands yielded restriction digest patterns corresponding to that expected for a rapsyn transcript missing all of exon 2. Sequencing two such clones from ciliary ganglion and muscle RNA confirmed the exact deletion of exon 2 (Fig. 1B). The results corroborate those of Margiotta and colleagues on rapsyn transcripts in chick ciliary ganglia (Burns et al., 1997) and identify an additional splice variant in the exon 2 region as well. J. Neurochem., Vol. 73, No. 4, 1999

Differences in ability of rapsyn splice variants to cluster nAChRs Previous studies showed that full-length mammalian rapsyn can cluster both muscle (Froehner et al., 1990; Phillips et al., 1991a) and neuronal (Feng et al., 1998; Kassner et al., 1998) nAChRs when heterologously expressed in co-transfected cells. Similar experiments were performed here to determine whether avian rapsyn also clusters nAChRs and to assess whether deletion of the leucine zipper motifs, as found in the naturally occurring rapsyn variants, impairs the clustering ability of rapsyn. For this purpose, full-length cDNA clones of the splice variants were obtained as described in Materials and Methods. These included constructs encoding full-length rapsyn (Rapsyn), the variant with exon 2 deleted (RapEx2), and the variants lacking either the first leucine zipper (RapZ1) or the second (RapZ2). All four species of chicken rapsyn were able to form clusters alone when heterologously expressed by transient transfection in the QT-6 fibroblast cell line. The clusters could be detected by fixing the cells, permeabilizing them, and visualizing the distribution of rapsyn with anti-rapsyn mAbs followed by immunofluorescent secondary antibodies (Fig. 2). The ability of the rapsyn variants to form such clusters mimicked that seen previously with mammalian rapsyn (Phillips et al., 1991a). Apparently, neither the leucine zipper motifs nor any portion of exon 2 is required for the clustering of rapsyn. Co-expression of the muscle nAChR gene constructs a1, b, g, and d together with full-length chicken rapsyn produced receptor clusters (Fig. 3, top row) as previously found with mammalian rapsyn (Froehner et al., 1990; Phillips et al., 1991a). The clustering was readily appar-

RAPSYN VARIANTS IN CILIARY GANGLIA

1403

FIG. 2. Spontaneous clustering of heterologously expressed rapsyn variants. QT-6 cells were transiently transfected with either a full-length rapsyn construct or each of the splice variants lacking part or all of exon 2. When the cells were subsequently fixed, permeabilized, and immunostained with antirapsyn mAbs, each of the encoded rapsyn products was found clustered in a nonuniform arrangement within the cell: Rapsyn (top left), RapZ1 (top right), RapZ2 (bottom left), and RapEx2 (bottom right). The occasional large, brightly labeled intracellular aggregates seen (e.g., with RapEx2) were found with all the constructs and have been reported for mammalian rapsyn as well (Phillips et al., 1991b, 1993). Bar 5 15 mm.

ent when the patchy distribution of receptor labeling on the surface of co-transfected cells was compared with the diffusely distributed pattern of surface labeling when cells were transfected with the receptor gene constructs without rapsyn (Fig. 3, left column). Interestingly, none of the three splice variants lacking part or all of exon 2 was able to induce clustering of the muscle nAChRs in co-transfected cells (Fig. 3, top row). Nor were the splice variants able to form intracellular clusters of receptor, visualized by fixing and permeabilizing the cells prior to labeling with antibody (data not shown). The four rapsyn constructs were also tested for an ability to cluster neuronal nAChRs heterologously expressed by transient co-transfection of QT-6 cells. The neuronal receptors employed were a7-nAChR and the a3/b4-nAChR. In both cases, the receptor distribution was visualized by labeling with anti-receptor mAbs after the cells were fixed and permeabilized so that both surface and intracellular nAChRs could be seen. This was necessitated by the fact that the levels of surface expression for the neuronal nAChRs, particularly for a7nAChRs, were much lower than for muscle receptors. Low levels are typically encountered with heterologous expression of a7-nAChRs in nonneuronal cells (Cooper and Millar, 1997; Kassner and Berg, 1997). Fortunately, intracellular receptors, which are more abundant in these cases, can be clustered by rapsyn in intracellular membrane, thereby providing an opportunity to assess rapsyn–receptor interactions (Feng et al., 1998; Kassner et al., 1998).

In the absence of rapsyn, neuronal nAChRs were found throughout the intracellular space with the greatest concentrations being in perinuclear regions. This presumably reflects their sequestration in the endoplasmic reticulum and Golgi; relatively little receptor label was found overlying nuclei. Co-expression with full-length chicken rapsyn produced a different distribution, with receptor clearly being aggregated or clustered within the cell (Fig. 3, bottom two rows). Most cells had at least some clusters clearly visible near the perimeter. In ;10% of the cells, large pronounced clusters could be seen, comparable with those obtained with muscle receptors. Similar patterns are produced by full-length mammalian rapsyn (Kassner et al., 1998). In contrast, coexpression of the receptors with rapsyn splice variants lacking part or all of exon 2 produced labeling patterns that were indistinguishable from those of control cells transfected only with the receptor constructs (Fig. 3, bottom two rows). Staining for rapsyn protein shows that the rapsyn spice variants are reliably expressed when co-transfected with the receptor constructs and that the rapsyn protein itself clusters, though not with receptor (data not shown). The results indicate that both leucine zipper motifs of exon 2 are required for rapsyn to cluster nAChRs, but they are not needed to permit clustering of rapsyn itself. Effects of rapsyn variants on detergent solubility of nAChRs In addition to clustering nAChRs, mammalian rapsyn reduces the ability of nAChRs to be extracted by deterJ. Neurochem., Vol. 73, No. 4, 1999

1404

W. G. CONROY AND D. K. BERG

FIG. 3. Muscle and neuronal nAChRs clustered by full-length chicken rapsyn but not by rapsyn lacking one or more leucine zippers. QT-6 cells were transiently transfected with muscle nAChR (top row), a7-nAChR (middle row), or a3/b4 (bottom row) constructs and with GFP (left column; negative control), Rapsyn (middle column), or RapZ1 (right column) constructs. Subsequently, cells were labeled with either mAb 35 without permeabilization to detect muscle nAChRs on the cell surface (top row), anti-a7 antibodies after permeabilization to detect a7-nAChRs (middle row), or mAb 35 after permeabilization to detect a3/b4 nAChRs (bottom row) and then followed with Cy3-labeled secondary antibodies. Receptors expressed in the absence of rapsyn displayed a diffuse distribution (left column), whereas receptor co-expressed with full-length rapsyn was distributed in clusters (middle column). Receptor co-expressed with the rapsyn splice-variant RapZ1, which lacks the first leucine zipper, was indistinguishable in distribution from that seen with control cells lacking rapsyn and expressing only receptors (right column). The other rapsyn splice variants lacking part or all of exon 2 also failed to cluster the receptors (data not shown). Bar 5 15 mm.

gent (Phillips et al., 1993; Kassner et al., 1998), presumably because of links to the cytoskeleton. Accordingly, experiments were performed to evaluate the ability of the four chicken rapsyn constructs to alter the detergent solubility of nAChRs. QT-6 cells were transiently co-transfected with either the muscle nAChR a1, b, g, and d constructs or the neuronal a7-nAChR construct, together with one of the four chicken rapsyn constructs. Detergent solubility of nAChRs on the cell surface was assessed by labeling intact cells with 125I-aBgt, extracting with buffer containing Triton X-100, and centrifuging to separate the soluble and insoluble fractions. Unexpectedly, none of the chicken rapsyn constructs, including the full-length species, increased the proportion of muscle nAChRs on the cell surface that scored as insoluble (Fig. 4). Nor did the full-length species or the variant lacking exon 2 have any effect on the solubility of a7-nAChRs expressed on the cell surface. The chicken rapsyn constructs also had no effect on the total number of receptors measured on the cell surface. As a positive control, co-expression of J. Neurochem., Vol. 73, No. 4, 1999

muscle nAChRs with a mammalian rapsyn construct produced a large increase in the proportion of surface receptors that were detergent insoluble (Fig. 4), as previously reported (Kassner et al., 1998). Immunoblot analysis shows that mammalian rapsyn is largely recovered in the insoluble fraction when expressed alone in transfected cells (Kassner et al., 1998). This presumably accounts for how it renders co-clustered nAChRs insoluble. When similar experiments were performed here with chicken full-length rapsyn expressed alone in transfected cells, a different result was obtained. About 80% of the chicken rapsyn was recovered in the soluble fraction (n 5 5 experiments). It is not clear whether this reflects a difference between chicken and mammalian rapsyn in their abilities to form cytoskeletal links or whether it arises from a difference in levels of expressed rapsyn: Mammalian rapsyn was so abundant that it could be detected by protein staining of transfected cell extracts on blots, whereas this was never true for chicken rapsyn (data not shown). In any case, the results clearly indicate that effects on receptor solubility cannot

RAPSYN VARIANTS IN CILIARY GANGLIA

1405

FIG. 4. Effects of rapsyn variants on nAChR solubility in detergent. QT-6 cells were transiently transfected with either muscle nAChR (black columns) or a7-nAChR (gray columns) constructs, together with the indicated rapsyn construct, and subsequently incubated with 125I-aBgt to label receptors on the cell surface. The cells were then solubilized in buffer containing Triton X-100, centrifuged to separate soluble and insoluble material, and quantified for bound radioactivity. Mammalian rapsyn (m Rapsyn), used as a positive control, converted a substantial fraction of the muscle nAChRs to an insoluble form. None of the chick rapsyn constructs had any effect as seen by comparing the results with those obtained when the muscle receptors were expressed in the absence of rapsyn constructs (GFP as a negative control). Chick rapsyn also had no effect on the solubility of a7-nAChRs compared with the negative control with GFP. Values shown for insoluble receptors represent the means 6 SEM of three to nine separate experiments and are expressed as a percent of the total labeled receptors calculated by adding the soluble and insoluble fractions.

be taken as a reliable indicator of rapsyn interactions because the chicken full-length species does promote receptor clustering and yet has no detectable effect on receptor solubility in detergent. Levels of rapsyn protein in ciliary ganglia Though rapsyn transcripts can be identified in ganglionic neurons (Burns et al., 1997; Feng et al., 1998), it has not yet been possible to demonstrate the presence of rapsyn protein in the cells. To optimize our chances of detecting chicken rapsyn in ciliary ganglia, we screened existing anti-rapsyn mAbs as well as generating new ones for recognizing the protein. Anti-rapsyn mAbs raised against nonavian immunogens were tested on immunoblots for cross-reaction with the chick species. Best was mAb 1201C, which readily detected components of the appropriate size for each of the rapsyn variants expressed via transient transfection of QT-6 cells (Fig. 5A). As mAb 1201C recognizes an epitope at the C-terminus, it should bind all four rapsyn species equally well (see Materials and Methods). Also effective were mAbs 22F10A and 1579A (data not shown), which recognize epitopes at the N- and C-termini, respectively. None was very effective at immunoprecipitating avian rapsyn. Using a large fragment of chicken rapsyn as an immunogen, we prepared new anti-rapsyn mAb isolates and tested

FIG. 5. Immunoblots of rapsyn in cell and tissue extracts. A: Extracts were prepared from QT-6 cells transiently transfected with GFP as a negative control, Rapsyn, RapZ1, RapZ2, or RapEx2 and were analyzed on immunoblots probed with mAb 1201C. Components of the expected sizes were obtained in each case; no crossreacting species was obtained in extracts of cells transfected with GFP. Similar results were obtained in two additional experiments. B: Extracts prepared from skeletal muscle (Muscle) and ciliary ganglia (CG) were analyzed on immunoblots probed with the indicated anti-rapsyn mAbs. In muscle extracts, a species of 43 kDa was detected by both mAbs (arrow) and aligned exactly with the full-length rapsyn product (not shown) detected in transfected cells; in ciliary ganglion extracts, multiple components were labeled by both mAbs, but none exactly co-aligned with any of the rapsyn products (not shown) expressed in transfected cells. Numbers of aBgt-binding sites loaded per lane: 3 fmol for muscle, 250 fmol for ciliary ganglia. The experiment was performed seven times with mAb 1201C, three times with mAb 22F10A, and two times each (not shown) with mAbs 1579A and 9E7. C: Material collected from tissue extracts by immunoprecipitation with mAb Rap-1 was eluted and analyzed on immunoblots probed with mAb 1201C. A prominent band of 43 kDa was apparent in the muscle sample (arrow). Very faint bands were sometimes seen at the same position in the ciliary ganglion (cg) and brain samples, but it was not clear that they represented significant labeling over that found in the negative control lane. The intensely labeled bands at ;55 kDa and 27 kDa represent immunoglobulin heavy and light chains eluted from protein G–Sepharose beads that had been coated with mAb for the immunoprecipitations. Numbers of aBgt-binding sites loaded per lane: 0.5 pmol for muscle, 1.5 pmol for ciliary ganglia, and 0.6 pmol for brain. In each panel, the positions of the molecular mass markers are indicated on the left. Similar results were obtained in two additional experiments.

J. Neurochem., Vol. 73, No. 4, 1999

1406

W. G. CONROY AND D. K. BERG

them both on immunoblots and in immunoprecipitations. Two mAbs, Rap-1 and Rap-2, were obtained that were effective on immunoblots; mAb Rap-1 was also efficient at immunoprecipitation. Both mAbs recognize epitopes encoded by exon 2. This was inferred from immunoblot analysis comparing the full-length and splice-variant products and also by immunoblot analysis of in vitrotranslated fusion proteins produced with a series of truncated rapsyn constructs (see Materials and Methods). Total extracts prepared from chick ciliary ganglia and pectoral muscle were analyzed on immunoblots probed with mAbs 1201C and 22F10A. Both mAbs recognized a species of 43 kDa in muscle extracts, as expected for full-length rapsyn (Fig. 5B). The bands co-migrated with a minor nonspecific component of the same size, but it is clear, particularly from the intense labeling with mAb 22F10A, that muscle rapsyn was present at the expected size. Other species were detected as well but apparently represented nonspecific antibody binding because they were not recognized by other anti-rapsyn mAbs. In ciliary ganglion extracts, multiple minor species were detected, but none of the bands was recognized by all of the appropriate mAbs in a reproducible manner. Bands were detected at 43 kDa with the two mAbs shown, but other anti-rapsyn mAbs did not label the bands. Even if all the staining detected with mAbs 1201C and 22F10A at 43 kDa in the ciliary ganglion sample represented rapsyn, it would still constitute only a tiny fraction of that expected from the muscle results because the ciliary ganglion samples had ;80-fold more receptor, judging by the number of aBgt-binding sites, than did the muscle samples. Thus, the ratio of rapsyn to receptor is much lower in the ganglia (if present at all) than in muscle. The prominent band at 38 kDa recognized by mAb 1201C in ganglion extracts was judged nonspecific because it did not exactly co-align with any of the rapsyn splicevariant products expressed in transfected QT-6 cells and because it did not cross-react with any other mAbs specific for either the N- or the C-terminus of rapsyn (data not shown). A second strategy to detect rapsyn in ciliary ganglia exploited the ability of mAb Rap-1 to concentrate the protein by immunoprecipitation. Only species containing at least the first leucine zipper encoded by exon 2 would have been collected in this case, given the specificity of the antibody. Immunoblot analysis of the material probed with mAb 1201C revealed significant amounts of fulllength rapsyn at 43 kDa in samples prepared from muscle tissue but little, if any, in samples prepared either from ciliary ganglia or from brain (Fig. 5C). The number of ganglionic a7-nAChRs exceeded muscle nAChRs by about threefold in the samples. The absence of a significant rapsyn signal in the ganglionic samples indicates that the amount must have been at least an order of magnitude lower than that in muscle. This predicts a ratio of rapsyn to a7-nAChRs in ganglia that is at least 30-fold lower than rapsyn per receptor in muscle. Combining the estimates from the two approaches places an upper limit on the rapsyn-to-receptor ratio that is at least 30- to 80-fold less than in muscle. If rapsyn plays any role in J. Neurochem., Vol. 73, No. 4, 1999

the concentration of a7-nAChRs on somatic spines in the ganglion, it must be doing so by a mechanism that is very different from the one by which it clusters receptors at the vertebrate neuromuscular junction. DISCUSSION The major findings reported here are that chick ciliary ganglia express at least four kinds of rapsyn transcripts and that all four encoded rapsyn products can self-cluster when heterologously expressed in cells; only the full-length version, however, efficiently induces clustering of co-expressed nAChRs. None of the rapsyn forms changes the detergent solubility of co-expressed nAChRs, indicating this is an unreliable marker of rapsyn–receptor interactions. Immunological analysis confirms the presence of fulllength rapsyn in chick muscle extracts but is not sensitive enough to detect significant amounts in ganglion extracts. Apparently, rapsyn does not perform the same kind of receptor-clustering function in neurons that it does in skeletal muscle. The PCR results obtained here identifying rapsyn splice variants in chick ciliary ganglia are consistent with those reported previously (Burns et al., 1997) but differ in requiring only a single round of PCR amplification for detection of product. In addition, another splice variant was identified, having a deletion in exon 2, assuming a chicken gene structure similar to that described for mammalian rapsyn (Gautam et al., 1994). No deletions were detected in exon 3, in contrast to rapsyn splice variants described in mouse and the rat pheochromocytoma cell line PC12 (Feng et al., 1998). The PCR amplification was not quantitative and therefore prohibits a comparison of rapsyn transcript levels in muscle versus ciliary ganglia or even among rapsyn splice variants within a tissue. The immunoblot analysis, however, readily detected full-length rapsyn in muscle and did not reveal smaller species, suggesting that the full-length species dominated. Neither full-length rapsyn nor truncated versions could be reliably detected in ganglion extracts. The inability of splice variants lacking exon 2 to cluster nAChRs clearly implicates the leucine zipper motifs in determining interactions with receptors. The leucine zippers are not necessary for rapsyn clustering itself. Both of these conclusions are consistent with previous studies using mutational analysis of mammalian rapsyn constructs to deduce the functional significance of individual regions (Phillips et al., 1991b). At least one other region of the rapsyn molecule, however, is needed to cluster nAChRs. This is the predicted coiled-coil domain found within amino acids 298–331 of the mammalian sequence, extending from exon 5 into exon 7. Disruptive mutations in this region block rapsyn-induced clustering of receptors without preventing rapsyn self-association (Ramarao and Cohen, 1998). It will be of interest in future studies to determine how the exon 2 and exon 5–7 domains cooperate to achieve receptor clustering. Previous studies showed a correlation between the ability of mammalian rapsyn to induce clustering of

RAPSYN VARIANTS IN CILIARY GANGLIA individual receptor subtypes and the negative effect it had on receptor solubility in detergent (Phillips et al., 1993; Kassner et al., 1998). No such correlation was found in the present studies with chick rapsyn constructs. The explanation may lie in the fact that, unlike mammalian rapsyn, chicken rapsyn is largely soluble under the conditions of expression used here. It is not clear whether the two rapsyn species are fundamentally different in this respect or whether the substantially lower levels of chick rapsyn found in the cells permit most of it to remain soluble. In any case, it is clear that insolubility cannot be considered an obligatory consequence of rapsyn–receptor interactions. Disruption of the rapsyn gene in mice does not prevent clustering of nAChRs recognized by mAb 35 on superior cervical ganglion neurons (Feng et al., 1998). The receptors detected in this case almost certainly correspond to a3*nAChRs described in chick ciliary ganglia (Vernallis et al., 1993; Mandelzys et al., 1994, 1995; Conroy and Berg, 1995; Feng et al., 1998). As a7-nAChRs have a different distribution on neurons than do a3*-nAChRs (Jacob and Berg, 1983; Jacob et al., 1984; Loring et al., 1985; Loring and Zigmond, 1987) and a7-nAChRs appeared to be more readily affected by rapsyn (Kassner et al., 1998), they remained a candidate for rapsyn action. The present experiments indicate, however, that the ratio of rapsyn to a7nAChRs in ciliary ganglion neurons is almost certainly far below the value of ;1 found for rapsyn and nAChRs in muscle where they are co-clustered. For ganglionic rapsyn to have escaped detection here, it would have had to remain insoluble in heated buffer containing 2% SDS. The low levels of rapsyn and related splice variants in the ganglion suggest that they do not cluster a7-nAChRs in any way analogous to the clustering of muscle nAChRs at the neuromuscular junction. These results do not preclude all clustering roles for rapsyn, however, in ciliary ganglia. It is possible, for example, that the protein influences some aspect of a7nAChR distribution by a nonstoichiometric mechanism. Alternatively, rapsyn may cluster other classes or subtypes of receptors on the neurons. Candidates include both glycine and GABAA receptors, which are present on ciliary ganglion neurons and may be much less abundant than a7-nAChRs (McEachern and Berg, 1988; Zhang and Berg, 1995). It is known that rapsyn can induce clustering of heterologously expressed GABAA receptors (Yang et al., 1997). The clustering of a7-nAChRs on somatic spines is likely to be controlled largely by nonrapsyn mechanisms. Acknowledgment: We thank Dr. Joseph Margiotta (Ohio Medical College, Toledo, OH, U.S.A.) for supplying the chick full-length rapsyn cDNA construct; Drs. Joshua Sanes, John Merlie, and Elizabeth Apel (Washington University, St. Louis, MO, U.S.A.) for supplying the QT6 cells, mammalian rapsyn cDNA construct, and mouse muscle nAChR genes; and Dr. Christine Holt (Cambridge University, Cambridge, U.K.) for supplying the GFP cDNA construct. We also thank Dr. Jon Lindstrom (University of Pennsylvania, Philadelphia, PA, U.S.A.) for providing mAbs 35 and 270; Dr. Stanley Froehner (University of North Carolina, Chapel Hill, NC, U.S.A.) for

1407

providing mAbs 1201C, 1579A, and 1234A; and Dr. Jonathan Cohen (Harvard Medical School, Boston, MA, U.S.A.) for providing mAbs 19F4A and 22F10A. Lynn Ogden provided expert technical assistance throughout. Grant support was provided by NIH grants NS35469 and NS12601 and the TobaccoRelated Disease Research Program.

REFERENCES Anand R., Peng X., and Lindstrom J. (1993) Homomeric and native a7 acetylcholine receptors exhibit remarkably similar but non-identical pharmacological properties, suggesting that the native receptor is a heteromeric protein complex. FEBS Lett. 327, 241–246. Aramakis V. B. and Metherate R. (1998) Nicotine selectively enhances NMDA receptor-mediated synaptic transmission during postnatal development in sensory neocortex. J. Neurosci. 18, 8485– 8495. Bertrand D., Galzi J. L., Devillers-Thiery A., Bertrand S., and Changeux J. P. (1993) Mutations at two distinct sites within the channel domain M2 alter calcium permeability of neuronal a7 nicotinic receptor. Proc. Natl. Acad. Sci. USA 90, 6971– 6975. Bloch R. J. and Froehner S. C. (1987) The relationship of the postsynaptic 43K protein to acetylcholine receptors in receptor clusters isolated from cultured rat myotubes. J. Cell Biol. 104, 645– 654. Blount P. and Merlie J. P. (1989) Molecular basis of the two nonequivalent ligand binding sites of the muscle nicotinic acetylcholine receptor. Neuron 3, 349 –357. Boyd R. T., Jacob M. H., McEachern A. E., Caron S., and Berg D. K. (1991) Nicotinic acetylcholine receptor mRNA in dorsal root ganglion neurons. J. Neurobiol. 22, 1–14. Burns A. L., Benson D., Howard M. J., and Margiotta J. F. (1997) Chick ciliary ganglion neurons contain transcripts coding for acetylcholine receptor-associated protein at synapses (rapsyn). J. Neurosci. 17, 5016 –5026. Chan J. and Quik M. (1993) A role for nicotinic a-bungarotoxin receptor in neurite outgrowth in PC12 cells. Neuroscience 56, 441– 451. Chang K. and Berg D. K. (1999) Dependence of circuit function on nicotinic acetylcholine receptors containing a7 subunits. J. Neurosci. 19, 3701–3710. Chiappinelli V. A. and Giacobini E. (1978) Time course of appearance of a-bungarotoxin binding sites during development of chick CG and iris. Neurochem. Res. 3, 465– 478. Coggan J. S., Paysan J., Conroy W. G., and Berg D. K. (1997) Direct recording of nicotinic responses in presynaptic nerve terminals. J. Neurosci. 17, 5798 –5806. Conroy W. G. and Berg D. K. (1995) Neurons can maintain multiple classes of nicotinic acetylcholine receptors distinguished by different subunit compositions. J. Biol. Chem. 270, 4424 – 4431. Conroy W. G. and Berg D. K. (1998) Nicotinic receptor subtypes in the developing chick brain: appearance of a species containing the a4, b2, and a5 gene products. Mol. Pharmacol. 53, 392– 401. Conroy W. G., Vernallis A. B., and Berg D. K. (1992) The a5 gene product assembles with multiple acetylcholine receptor subunits to form distinctive receptor subtypes in brain. Neuron 9, 679 – 691. Cooper S. T. and Millar N. S. (1997) Host cell-specific folding and assembly of the neuronal nicotinic acetylcholine receptor a7 subunit. J. Neurochem. 68, 2140 –2151. Corriveau R. A. and Berg D. K. (1994) Neurons in culture maintain acetylcholine receptor levels with far fewer transcripts than in vivo. J. Neurobiol. 25, 1579 –1592. Couturier S., Bertrand D., Matter J.-M., Hernandez M.-C., Bertrand S., Millar N., Valera S., Barkas T., and Ballivet M. (1990) A neuronal nicotinic acetylcholine receptor subunit (a7) is developmentally regulated and forms a homo-oligomeric channel blocked by a-Btx. Neuron 5, 847– 856. Craig A. M. (1998) Activity and synaptic receptor targeting: the long view. Neuron 21, 459 – 462. Craven S. E. and Bredt D. S. (1998) PDZ proteins organize synaptic signaling pathways. Cell 93, 495– 498.

J. Neurochem., Vol. 73, No. 4, 1999

1408

W. G. CONROY AND D. K. BERG

Dieffenbach C. W. and Dveksler G. S. (1995) PCR Primer. A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. Feng G., Steinbach J. H., and Sanes J. R. (1998) Rapsyn clusters neuronal acetylcholine receptors but is inessential for formation of an interneuronal cholinergic synapse. J. Neurosci. 18, 4166 – 4176. Frazier C. J., Buhler A. V., Weiner J. L., and Dunwiddie T. V. (1998) Synaptic potentials mediated via a-bungarotoxin-sensitive nicotinic acetylcholine receptors in rat hippocampal interneurons. J. Neurosci. 18, 8228 – 8235. Froehner S. C., Luetje C. W., Scotland P. B., and Patrick J. (1990) The postsynaptic 43K protein clusters muscle nicotinic acetylcholine receptors in Xenopus oocytes. Neuron 5, 403– 410. Fu W.-M. and Liu J.-J. (1997) Regulation of acetylcholine release by presynaptic nicotinic receptors at developing neuromuscular synapses. Mol. Pharmacol. 51, 390 –398. Fu W.-M., Liou H.-C., and Chen Y.-H. (1998) Nerve terminal currents induced by autoreception of acetylcholine release. J. Neurosci. 18, 9954 –9961. Gautam M., Mudd J., Copeland N. G., Gilbert D. J., Jenkins N. A., and Merlie J. P. (1994) Characterization and mapping of the Rapsyn gene encoding the 43 kDa acetylcholine receptor-associated protein. Genomics 24, 366 –369. Gray R., Rajan A. S., Radcliffe K. A., Yakehiro M., and Dani J. A. (1996) Hippocampal synaptic transmission enhanced by low concentrations of nicotine. Nature 383, 713–716. Guo J.-Z., Tredway T. L., and Chiappinelli V. A. (1998) Glutamate and GABA release are enhanced by different subtypes of presynaptic nicotinic receptors in the lateral geniculate nucleus. J. Neurosci. 18, 1963–1969. Hefft S., Hulo S., Bertrand D., and Muller D. (1999) Synaptic transmission at nicotinic acetylcholine receptors in rat hippocampal organotypic cultures and slices. J. Physiol. (Lond.) 510, 709 –716. Jacob M. H. and Berg D. K. (1983) The ultrastructural localization of a-bungarotoxin binding sites in relation to synapses on chick CG neurons. J. Neurosci. 3, 260 –271. Jacob M. H., Berg D. K., and Lindstrom J. M. (1984) Shared antigenic determinant between the Electrophorus acetylcholine receptor and a synaptic component on chicken ciliary ganglion neurons. Proc. Natl. Acad. Sci. USA 81, 3223–3227. Kassner P. D. and Berg D. K. (1997) Differences in the fate of neuronal acetylcholine receptor protein expressed in neurons and stably transfected cells. J. Neurobiol. 33, 968 –982. Kassner P. D., Conroy W. G., and Berg D. K. (1998) Organizing effects of rapsyn on neuronal nicotinic acetylcholine receptors. Mol. Cell. Neurosci. 10, 258 –270. Kirsch J., Meyer G., and Betz H. (1996) Synaptic targeting of ionotropic neurotransmitter receptors. Mol. Cell. Neurosci. 8, 93–98. Landmesser L. and Pilar G. (1974) Synapse formation during embryogenesis on ganglion cells lacking a periphery. J. Physiol. (Lond.) 241, 715–736. Li X., Rainnie D. G., McCarley R. W., and Greene R. W. (1998) Presynaptic nicotinic receptors facilitate monoaminergic transmission. J. Neurosci. 18, 1904 –1912. Loring R. H. and Zigmond R. E. (1987) Ultrastructural distribution of 125 I-toxin F binding sites on chick ciliary neurons: synaptic localization of a toxin that blocks ganglionic nicotinic receptors. J. Neurosci. 7, 2153–2162. Loring R. H., Dahm R. H., and Zigmond R. E. (1985) Localization of a-bungarotoxin binding sites in the ciliary ganglion of the embryonic chick: an autoradiographic study at the light and electron microscopic level. Neuroscience 14, 645– 660. Mandelzys A., Pie B., Deneris E. S., and Cooper E. (1994) The developmental increase in ACh current densities on rat sympathetic neurons correlates with changes in nicotinic ACh receptor a-subunit gene expression and occurs independent of innervation. J. Neurosci. 14, 2357–2364. Mandelzys A., De Koninck P., and Cooper E. (1995) Agonist and toxin sensitivities of ACh-evoked currents on neurons expressing multiple nicotinic ACh receptor subunits. J. Neurophysiol. 74, 1212– 1221.

J. Neurochem., Vol. 73, No. 4, 1999

McEachern A. E. and Berg D. K. (1988) Benzodiazepine interactions with GABAA receptors on chick ciliary ganglion neurons. Mol. Pharmacol. 34, 129 –135. McGehee D., Heath M., Gelber S., and Role L. W. (1995) Nicotine enhancement of fast excitatory synaptic transmission in CNS by presynaptic receptors. Science 269, 1692–1697. Phillips W. D., Kopta C., Blount P., Gardner P. D., Steinbach J. H., and Merlie J. P. (1991a) ACh receptor-rich membrane domains organized in fibroblasts by recombinant 43-kilodalton protein. Science 251, 568 –570. Phillips W. D., Maimone M., and Merlie J. P. (1991b) Mutagenesis of the 43-kD postsynaptic protein defines domains involved in plasma membrane targeting and AChR clustering. J. Cell Biol. 115, 1713–1723. Phillips W. D., Noakes P. G., Roberds S. L., Campbell K. P., and Merlie J. P. (1993) Clustering and immobilization of acetylcholine receptors by the 43-kD protein: a possible role for dystrophinrelated protein. J. Cell Biol. 123, 729 –740. Pugh P. C. and Berg D. K. (1994) Neuronal acetylcholine receptors that bind a-bungarotoxin mediate neurite retraction in a calciumdependent manner. J. Neurosci. 14, 889 – 896. Radcliffe K. A. and Dani J. A. (1998) Nicotinic stimulation produces multiple forms of increased glutamatergic synaptic transmission. J. Neurosci. 18, 7075–7083. Ramarao M. K. and Cohen J. B. (1998) Mechanism of nicotinic acetylcholine receptor cluster formation by rapsyn. Proc. Natl. Acad. Sci. USA 95, 4007– 4012. Sanes J. R. and Lichtman J. W. (1999) Development of the vertebrate neuromuscular junction. Annu. Rev. Neurosci. 22, 389 – 442. Schoepfer R., Conroy W. G., Whiting P., Gore M., and Lindstrom J. (1990) Brain a-bungarotoxin binding protein cDNAs and mAbs reveal subtypes of this branch of the ligand-gated ion channel gene superfamily. Neuron 5, 35– 48. Seguela P., Wadiche J., Dineley-Miller K., Dani J. A., and Patrick J. W. (1993) Molecular cloning, functional properties, and distribution of rat brain a7: a nicotinic cation channel highly permeable to calcium. J. Neurosci. 13, 596 – 604. Shoop R. D., Martone M. E., Yamada N., Ellisman M. H., and Berg D. K. (1999) Neuronal acetylcholine receptors with a7 subunits are concentrated on somatic spines for synaptic signaling in embryonic chick ciliary ganglia. J. Neurosci. 19, 692–704. Ullian E. M., McIntosh J. M., and Sargent P. B. (1997) Rapid synaptic transmission in the avian CG is mediated by two distinct classes of nicotinic receptors. J. Neurosci. 17, 7210 –7219. Vernallis A. B., Conroy W. G., and Berg D. K. (1993) Neurons assemble acetylcholine receptors with as many as three kinds of subunits while maintaining subunit segregation among receptor subtypes. Neuron 10, 451– 464. Vernino S., Rogers M., Radcliffe K. A., and Dani J. A. (1994) Quantitative measurement of calcium flux through muscle and neuronal nicotinic acetylcholine receptors. J. Neurosci. 14, 5514 –5524. Vijayaraghavan S., Huang B., Blumenthal E. M., and Berg D. K. (1995) Arachidonic acid as a possible negative feedback inhibitor of nicotinic acetylcholine receptors on neurons. J. Neurosci. 15, 3679 –3687. Williams B. M., Tamburni M. K., Schwartz Levey M., Bertrand S., Bertrand D., and Jacob M. H. (1998) The long internal loop of the a3 subunit targets nAChRs to subdomains within individual synapses on neurons in vivo. Nat. Neurosci. 1, 557–562. Wilson Horch H. L. and Sargent P. B. (1995) Perisynaptic surface distribution of multiple classes of nicotinic acetylcholine receptors on neurons in the chicken CG. J. Neurosci. 15, 7778 –7795. Yang S.-H., Armson P. F., Cha J., and Phillips W. D. (1997) Clustering of GABAA receptors by rapsyn/43kD protein in vitro. Mol. Cell. Neurosci. 8, 430 – 438. Zhang Z.-W. and Berg D. K. (1995) Patch-clamp analysis of glycineinduced currents in chick ciliary ganglion neurons. J. Physiol. (Lond.) 487, 395– 405. Zhang Z.-W., Coggan J. S., and Berg D. K. (1996) Synaptic currents generated by neuronal acetylcholine receptors sensitive to a-bungarotoxin. Neuron 17, 1231–1240.