Synapse Formation by Neuroblastoma Hybrid Cells M. NIRENBERG, S.P. WILSON, H. HIGASHIDA, A. ROTTER, K. KREUGER, N. Buss, R. RAY, J. KENIMER, M. ADLER, AND H. FUKUI Laboratory of Biochemical Genetics, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland 20205 The most widely held hypothesis concerning the mechanism of synapse formation during the last 20 years has been the chemoaffmity hypothesis of Sperry (1963), i.e., that neurons distinguish appropriate from inappropriate synaptic partner cells by interactions be- tween molecules that code for cell recognition. Al- though Sperry avoided detailed models, many in- vestigators have assumed that neurons or other cells that possess different cell-recognition molecules are generated and that the cells then sort out in appropriate sequences by differential adhesiveness before synaptic connections form. Another quite different hypothesis is that neurons may form relatively stable intercellular attachments during differentiation and that environmen- tal factors and/or transynaptic signals may regulate the expression of genes and thereby determine the final pathway for differentiation and the type of neuron and synapse to be expressed. These hypotheses are not mutually exclusive; both may be involved in the assembly of synaptic circuits. We have established clonal lines of neuroblastoma and somatic hybrid cells that express neural properties and form synapses with cultured striated muscle cells, and we have used the cells as model systems for studies on the effects of extracellular molecules on the expres- sion of neural properties and on synaptogenesis. We find that the expression of stimulus-secretion coupling and synaptogenesis can be regulated in neuroblastoma hybrid cells by receptor-mediated events that are cou- pled to the activation of adenylate cyclase. We also find that many ?S-labeled glycoproteins are affected when cells are shifted to a synapse competent state. RESULTS Synapse Competent and Defective Cell Lines Twenty-six neuroblastoma or hybrid cell lines were tested for their ability to synthesize and release acetyl- choline (ACh) and form synapses with cultured rat striated muscle cells (Higashida et al. 1978; Wilson et al. 1978). Synapses were detected by intracellular recording of miniature end-plate potentials and evoked muscle responses. Photomicrographs of cells from four of the five cell lines found that form abundant synapses are shown in Figure 1. NBrlO-A and NBr20-A cells originated by fusion of mouse neuroblastoma N 18TG-2 (Minna et al. 1972) with clonal BRL-30E rat liver cells; NCB-20, by fusion of N18TG- 2 cells with fetal Chinese hamster brain cells; and NGl08-15 (Nelson et al. 1976), by fusion of Nl8TG-2 with C6BU-I rat gli- oma cells (Amano et al. 1974). Muscle responses to ACh secreted by the neural cells were inhibited rever- sibly by 5 pM d-tubocurarine and irreversibly by 10 nM cr-bungarotoxin. Omission of Ca+ + ions in the me- dium reduced the frequency of miniature end-plate po- tentials (Nelson et al. 1978). The expression of some neural properties by the cells is dependent on prolonged elevation of cellular CAMP. For example, untreated cells have few or no neurites and relatively small soma. However, treatment of cells for 7 days with 1 mM dibutyryl CAMP promotes neurite extension, and decreasing the concentration of fetal bovine serum reduces neurite retraction (Seeds et al. 1970). Neurites are produced in profusion under these conditions as. shown in Figure 1. Other cell lines were found that synthesize ACh and adhere well to myotubes, but do not form synapses. The phenotypes of the cell lines are summarized in Table 1. Cells from five lines synthesize ACh, take up Ca+ + ions from the medium and secrete ACh when depolarized by 80 mM K+ ions, possess small clear vesicles 60 nm in diameter and large dense-core vesicles 180 nm in diameter, synthesize and secrete a protein that stimulates the aggregation of nicotinic ACh receptors (AChR) on myotube plasma membranes, and form many synapses with cultured myotubes. Cells from three lines take up Ca+ + ions slowly and to only a small extent, secrete relatively little ACh when depolarized, and form few synapses with muscle cells. Two cell lines lack functional voltage-sensitive Ca' + channels and therefore do not take up 45Ca+ + or secrete more ACh when depolarized and do not form synapses. Cells from five lines take up Ca+ + when depolarized, but do not respond to the increase in cytoplasmic Ca+ + concentration by secreting ACh, and form few or no synapses. These cells release ACh into the medium in the basal unstimulated state but lack a Ca + +-dependent ACh secretion reaction(s). Three cell lines possess small clear vesicles, but lack both large dense-core vesicles and functional AChR ag- gregation protein, and form few or no synapses. Nine additional cell lines have little or no choline acetyltransferase activity and thus synthesize little or no ACh and do not form synapses with striated muscle cells. Regulation of Synaptogenesis Neuroblastoma and hybrid cells have PGEr receptors that mediate the activation of adenylate cyclase (Sharma 707 708 NIRENBERG ET AL. Figure 1. Neuroblastoma hybrid cells from lines that form many synapses with cultured myotubes were treatea for 7 days with I mM dibutyryl CAMP, and the concentration of fetal bovine serum was reduced from 5% to I % between the fifth and seventh days (H. Higashida et al., in prep.). Bar, 50 Gm. et al. 1975b, 1977). Cells were cultured in the presence dibutyryl CAMP or theophylline, on the percentage of of PGEi to increase the endogenous rate of synthesis myotubes tested that were innervated and the frequency of CAMP and elevate intracellular levels of CAMP of miniature end-plate potentials of myotubes. The (Matsuzawa and Nirenberg 1975; Sharma et al. 1975a). cells were cocultured and treated with the compounds Treatment of cells with PGEi has no immediate effect shown in Table 2 for 5-7 days before myotubes were on cell membrane potential or rate of ACh secretion assayed for synapses by intracellular microelectrode (McGee et al. 1978). recording. Treatment of cells with 1 mM dibutyryl In Table 2 are shown the effects of treating CAMP. 1 inM theophylline, or 10 PM PGE, increased NG108-15 cells and rat myotubes with PGEi, or a the percentage of muscle cells tested that were inner- cyclic nucleotide phosphodiesterase inhibitor such as vated from 15% to approximately 60% and also Table 1. Cell Line Phenotypes K+ -dependent Vesicles AChR ACh large aggrega- Cell forma- 45ca+ + 13H]ACh small dense don lines tion uptake release Cl.%% core protein Synapse : + +++ +++ + + + -I-++ + + + + + + + 2 + + + + 5 + ++ + + + - or + 3 + ++ f + - or + 9 - Data from Higashida et al. (1978). Rotter et al. (1978). Wilson et al. (1978). and N.A. Busis, M.P. Daniels. H.C. Bauer, P.A. Pudimat, P. Sonderegger. A.E. Schaffner. and M. Nirenherg. SYNAPTOGENESIS BY NEUROBLASTOMA CELLS 709 Table 2. Effect of Culture Conditions on Synaptogenesis and ACh Secretion by NG108-15 Cells Culture conditions % Synaptic myotubes responses/ with mini synapses myotube Control Dibutyryl CAMP (1 mM) Theophylline (1 mM) PGE, (10 PM) PGEl (10 FM) + theophylline (1 mM) I5 0.7 55 14 64 10 63 11 98 32 Each value is the mean of values obtained from more rhan 75 myotubes (Higashida er al. 1978). resulted in 14- to 20-fold increases in the mean miniature end-plate-potential frequencies (Higashida et al. 1978). Presumably, each miniature end-plate poten- tial reflects the response of a myotube to the spon- taneous release of ACh from a single NG108-15 vesi- cle. The effects of PGE, and theophylline were ad- ditive or synergistic since 98% of the myotubes tested were innervated and a 45fold increase in the miniature end-plate-potential frequency was observed. As shown in Figure 2, the effects of PGE,, theophylline, or dibutyryl CAMP on synaptogenesis are expressed slowly. Half-maximal increases in myotube innervation required l-2 days of treatment. In other ex- periments not shown here, cells were incubated with PGEl, theophylline, dibutyryl CAMP, or PGEl and theophylline for 5-7 days; the cells were then incubated for an additional 4-14 days in the absence of the com- ti I ' ' I ' a_ @ REGULATION OF SYNAPSE FORMATION @ 2 & IOO- PGEI tTpEOPHYLLlNE F zj ao- Y = 60- 0 DIWTYRYL CAMP 1 J) -I Figure 2. Effects of PGEl, theophylline, or dibutyryl CAMP on the percentage of myotubes innervated by NGl08-15 cells are shown as a function of time of coculture. At zero time, 2 x lo5 mechanically dissociated NG108-15 cells were added to each 35-mm petri dish, which contained well-differentiated myotubes that had formed from myoblasts dissociated from newborn rat hind-limbs during 9 days of culture. The cells were cocultured in Dulbecco's modified Eagle's medium, 5 % horse serum, 100 pM hypoxanthine, 16 pM thymidine. and the following when indicated: Control, no addition; 1 mM theo- phylline; 10 tiM PGEI; 10 PM PGE, and 1 mM theophylline; or I mM dibutyryl CAMP. Synapses were detected by in- tracellular microelectrode recording of miniature end-plate potentials. Each point is the mean of values obtained from IO-20 myotubes. (Data from Higashida et al. 1978.) pounds to determine whether the effects on synapses were reversible. On cessation of treatment, synapses and ACh secretion rates slowly returned to control values over a period of 7-10 days. Thus, the effects of the compounds on synaptogenesis were expressed slow- ly, and on withdrawal of the compounds, the effects were reversed slowly. Treatment of NG108-15 cells with PGEl and R020-1724, a CAMP phosphodiesterase inhibitor, also increased the number of synapses. How- ever, no reversal was detected due to withdrawal of these compounds. Effects of Culture Conditions on ACh Storage and Secretion As shown in Figure 3A, intracellular ACh levels of NG108-15 cells increased eightfold and threefold when cells were treated for 3 or more days with 10 pM PGE, and 1 mM theophylline or with 1 mM dibutyryl CAMP, respectively. The cells were incubated with 30 pM [3H]choline for 17 hours to label intracellular ACh. Cells were then washed, and intracellular [3H]ACh was extracted and quantitated. More than 90% of the in- tracellular [3H]ACh was found to be particulate. Treat- ment of cells for 5 or more days with dibutyryl CAMP (Daniels and Hamprecht 1974) or with PGEI and theo- phylline (Wilson et al. 1978) resulted in marked in- creases in the abundance of small clear vesicles and large dense-core vesicles. These results suggest that the increase in intracellular ACh is due, at least in part, to NG,08-,5 CELLS B CELLULAR CAM C PROTEIN/FLASK ~~~~~~--:? Figure 3. Effects of dibutyryl CAMP or PGEl and theophyl- line on intracellular [3H]ACh levels of NG108-15 cells (A) and intracellular CAMP (8); (C) cell proteins per flask are shown as a function of time (Wilson et al. 1978). (A) NG 1 OS- 15 cells were incubated in culture medium containing approximately 30 FM (merItyl-3H]choline chloride for 17 hr to enable cells to synthesize [`H]ACh. Cells were washed 3 times with isotonic buffer, and then intracellular r3H]ACh was extracted (Toru and Aprison 1966) and separated from other radioactive compounds by high-voltage paper electro- phoresis (Potter and Murphy 1967). Where indicated, the medium was supplemented with the following: Control, no addition; 1 mM dibutyryl CAMP; or 10 CM PGEl and 1 mM theophylline. (B) CAMP was extracted from cells and purified as described by Matsuzawa and Nirenberg (1975); the amount of CAMP in each sample was determined by inhibition of the binding of [3H]cAMP to CAMP-dependent protein kinase. (C) Protein was determined by a modification of the method of Lowry et al. (1951). 710 NIRENBERG ET AL. an increase in the number of ACh storage vesicles in ceils. In the presence of 10 PM PGEI and 1 mM theophylline, CAMP levels of NG108-15 cells increased markedly and then decreased somewhat during the re- mainder of the incubation period due to partial desen- sitization of PGEl receptors (Fig. 3B). However, CAMP levels of treated cells were higher than those of control cells throughout the `I-day period examined. The rate of cell division also decreased after the first day of treatment with PGE, and theophylline; thus, the rate of accumulation of protein per flask decreased in the presence of PGE, and theophylline (Fig. 3C). The effect of treating NG108-15 cells for 0, 1, 3, or 5 days with 1 mM dibutyryl CAMP on the ability of cells to secrete ACh in response to a depolarizing stimulus is shown in Figure 4A. The ceils were in- cubated with 27 pM [3H]choline for 1 hour, washed by perfusion, and then depolarized with 80 mM K+ (replacing 80 mM Na+) during the period shown. [`H]ACh secreted into the medium was separated from other 3H-labeled compounds and quantitated as described by McGee et al. (1980). Logarithmically `dividing, untreated control cells did not respond to depolarization by secreting .ACh. However, cells treated 1, 3, or 5 days with 1 mM dibutyryl CAMP became increasingly responsive to depolarization with respect to the amount of ACh secreted (McGee et al. 1978). Cells treated with PGE, and theophylline for 7 days secreted twice as much ACh when depolarized as did cells that had been treated with dibutyryl CAMP (Fig. 4B). REC+ILATm OF K'-CEFENCENT ~]ACH RELEASE FFCt.4 CELLS LLS TREATED WITH MBUTYRIL- 8' ' CAMP FCft 0-5@AYS CELLS TREATED VJTH Ffx, MINUTES Figure 4. [3H]ACh secretion by untreated control NG108-15 cells or by cells treated for 1, 3, or 5 days with I mM dibutyryl CAMP (A) or by cells cultured for 7 days with or without IO pM PGEl and I mM theophylline (B) is shown in response to depolarization of cells with 80 mM K+ (replacing 80 rnr+i Na+ in the medium). The cells were incubated with [merhyl-3H]choline chloride to enable them to synthesize [3H]ACh. and then washed and depolarized. (Data from McGee et al. [1978]; the figure was redrawn.) t3H]ACh was separated from other radioactive compounds and quantitated as described by McGee et al. (1980). The basal rate of [3H]ACh secretion by unstimulated cells (5.4 mM K+) was subtracted from each value shown. Conditional Expression of Stimulus-Secretion Coupling in NBrlO-A Ceils Depolarization of neuronal terminals is known to ac- tivate voltage-sensitive Ca+ + channels; Ca + + ions then flow into the neuronal terminals and increase the rate of secretion of neurotransmitter. Thus, the effects of treating cells for 7 days with PGEl and theophylline or dibutyryl CAMP on Ca+ + uptake via voltage- sensitive Ca+ + channels were determined. 45Ca+ + ions rapidly bound to NBrlO-A cells initially and/or entered cells incubated in medium containing 5.4 mM K+ ions, but uptake soon plateaued. Depolarization of cells with 80 mM K+ resulted in a marked increase in 45Ca+ + up- take. Half-maximal inhibition of depolarization- dependent uptake of 45Ca+ + was obtained with 9x lo-' M D600; La3+, Co++, or Ni++ ions also in- hibited K + dependent 45Ca+ + by cells. Electrophysio- logical studies revealed Ca+ + spikes; action potentials also were obtained when strontium or barium ions were substituted for Ca+ + ions (not shown). As shown in Figure 5, logarithmically dividing; un- treated control NBrlO-A cells lacked functional voltage-sensitive Ca + + channels and did not take up 4SCa+ + ions when cells were depolarized. However, cells treated for 7 days with 10 pM PGEl and 1 mM theophylline or with 1 mM dibutyryl CAMP took up 45Ca+ + via voltage-sensitive Ca+ + channels when cells were depolarized with 80 mM K+ (Rotter et al. 1979). <+-DEPENDENT SPECIFIC 45C02+ JPTAKE BY CELLS EXPOSED TO DIFFERENT CONDITIONS PGEI +THEOPHYLLINE Lm CONTROLCELLS 11-1,1 1 I 2 4M1N:TES8 lo Figure 5. Effect of culture conditions on the expression of func- tional voltage-sensitive Ca+ + channels of NBrlO-A cells. 4sCa+ + uptake was due to activation of voltage-sensitive Ca+ + channels of untreated, IogarithmicaJly dividing control NBrlO-A cells and cells cultured for 6 days with 1 mM dibutyryl CAMP or 10 FM PGEl and I mM theophylline. The cells were depolarized with 80 mM K+ (in place of 80 mM Na+). Values for 45Ca++ binding to cells and/or uptake at 5.4 mM K+, which were not inhibited by 100 CM D600 and were not mediated by voltage-sensitive Ca+ + channels were subtracted. Ca+ + uptake dependent on depolarization was completely inhibited by 100 pM D600. (Data from Rotter et al. 1979) SYNAPTOGENESIS BY NEUROBLASTOMA CELLS 711 Figure 6. Acquisition of functional voltage-sensitive Ca' + channels by NBrlO-A cells cultured with or without 1 mM dibutyryl CAMP is shown as a function of time of treatment. Ca+ + action potentials were detected by intracellular rnicro- electrode recording; each point is the m&n of values obtained from 20 NBrlO-A cells. Logarithmically dividine NBrlO-A cells reached confluency on app;oximately ihe fourth day of culture. The medium contained 5 pM tetrodotoxin to inhibit voltage- sensitive Na+ channels. (Data from Rotter et al. 1979.) Four methods were used to measure Ca+ + uptake de- pendent on cell depolarization: 45Ca+ + flux was measured; net uptake of Ca+ + ions from the medium by cells was determined using a Ca+ +-specific elec- trode; Ca+ + ion concentrations were determined by a spectrophotometric assay with murexide (P. Darvenezia and M. Nirenberg, in prep.); and Ca++ action potentials were demonstrated by electrophysiological methods with intracellular microelectrode recording. Results obtained by each method showed that the expression of voltage-sensitive Ca+ + channels is increased by treat- ment of cells with compounds that elevate cellular CAMP levels. The acquisition of voltage-sensitive Ca+ + channel activity by NBrlO-A cells measured with intracellular microelectrodes is shown in Figure 6 as a function of days of culture of cells with or without 1 mM dibutyryl CAMP. Cells were stimulated electrically with depolarizing pulses of current in the presence of 5 pM tetrodotoxin to inhibit voltage-sensitive Na+ channels. Voltage-sensitive Ca+ + channel activity was deter- mined at the resting membrane potential and at a mem- brane potential of - 90 mV adjusted with steady cur- rent. Only 10% of logarithmically dividing, untreated cells at low density possessed functional voltage- sensitive Ca+ + channels, and the Ca+ + spikes that were detected were relatively weak. The proportion of cells with Ca+ + spikes increased from 10% to 50% as untreated cells multiplied and formed confluent monolayers, whereas 100% of the cells tested that were treated for 3 days with dibutyryl CAMP generated Ca+ + spikes when stimulated electrically. The mean maximum rate of rise of Ca+ + spikes, a measure of Ca+ + chan- nel activity, increased approximately 20-fold when cells were treated with dibutyryl CAMP (M. Adler and M. Nirenberg, iri prep.) (Fig. 6B). A smaller increase was observed when logarithmically dividing, untreated con- trol cells formed confluent monolayers, revealing a small cell concentration or contact-dependent regula- tion of Ca+ + channel expression. These and the previous results show that when CAMP levels are elevated, cells acquire at similar rates voltage-sensitive Ca+ + channels, depolarization-dependent ACh secre- tion, and synapses. Nitrendipine and other dihydropyridines have been shown to reduce voltage-sensitive Ca+ + channel activ- ity in smooth muscle (for review, see Triggle and Swamy 1983), and specific binding sites for [`H]ni- trendipine have been found in smooth muscle (Bolger et al. 1982). cardiac muscle (Ehlert et al. 1982), and brain (Murphy and Snyder 1982). The receptors for nitren- dipine are thought to be either part of the voltage-sen- sitive Ca+ + channel complex or regulators of channel activity. As shown in Table 3, few or no specific bind- ing sites for [3H]nitrendipine were detected in washed membranes prepared from logarithmically dividing control NBrlO-A ceils, whereas specific binding sites were found in membranes prepared from NBrlO-A cells that had been treated for 8 days with 10 pM PGE, and 1 mM theophylline (M. Nirenberg, L. Anderson, and A. Rotter, in prep.). Scatchard analysis revealed a single class of specific binding sites for [3H]nitrendi- pine with a dissociation constant of 2 x lo-lo M, which agrees well with values reported for other tissues (Bolger et al. 1982; Murphy and Snyder 1982). The maximum number of specific nitrendipine-binding sites in membranes from NBrlO-A cells treated with PGE, and theophylline was found to be 61 fmoles/mg of membrane protein, which corresponds to approximate- ly 16,000 sites for nitrendipine per cell. SB37-B neuroblastoma x L cell hybrid cells synthesize ACh but have little or no voltage-sensitive Ca+ + channel activ- ity and thus do not secrete ACh when cells are depolar- ized and do not form synapses with muscle cells. Spe- cific binding sites for [3H]nitrendipine were not detected in membranes prepared from SB37-B cells grown with or without PGEl and theophylline. These results show that the CAMP-dependent expression of Table 3. PGEl and Theophylline Regulate the Number of Nitrendipine-binding Sites Expressed by NBrlO-A Cells Treatment of cells fmoles 13H]nitrendipine (5-8 days) specifically boundlmg protein NBr IO-A control <2 10 FM PGEl + I mM theophylline I7 SB37-B control <2 lo PM PGE, + 1 mM theophylline <2 Data from M. Nirenberg. L. Anderson. R. Rotter. and R. Ray (in prep.). The reaction-mixture components and con- ditions were similar to those reported by Murphy and Snyder (1982). Each reaction mixture contained 0.2 nM f3H]nitrendipine and approximately 250 rg of protein (a washed membrane fraction pelleted at 48,COOg). 712 NIRENBERG ET AL. voltage-sensitive Ca+ + channels in NBrlO-A cells is accompanied by an increase in the number of specific binding sites for [3H]nitrendipine and suggest that elevation of cellular CAMP results in an increase in the number of Ca+ + channels in cells. CAMP-dependent Changes in Glycoproteins of NGlOS-15 Cells Elevation of CAMP levels of neuroblastoma or hybrid cells gradually results in the acquisition by cells of functional action potential channels for Na+, K+ (Nelson et al. 1978). and Ca++ (Rotter et al. 1979), which suggests that CAMP regulates the synthesis or catabolism of channel molecules. We therefore deter- mined the effect of PGE, on glycoproteins expressed 2ocl 140 100 60 40 200 m 140 k loo 2- 60 40 200 140 100 60 40 CONTROL PGEl by NC 10% I5 cells. Logarithmically dividing control NG108-15 cells and cells treated with 10 pM PGE, for 7 days were incubated for 18 hours with [`?S]methi- onine. %-labeled glycoproteins soluble in 1 o/c Triton X-100 were fractionated by chromatography on wheat germ agglutinin-. Lens cufinaris agglutinin-, and Ric- inus communis agglutininii-agarose columns. %-la- beled glycoproteins were eluted with 0.5 M N-acetyl- glucosaminq, 0.2 M cr-methylmannoside, and 0.2 M lactose, respectively, desalted, and subjected to two- dimensional polyacrylamide gel electrophoresis and autoradiography (Fig. 7). Autoradiographs of duplicate two-dimensional gels (triplicates in other experiments), exposed for 7 hours and 1, 2, and 4 days, were com- pared with the aid of a computer and programs de- signed for analysis of two-dimensional gels (Vo et al. 1981). Computer-generated plots based on the 48 auto- radiographs analyzed in this experiment were obtained. I WHEAT GERM III 1 I I I IIIJ 5.0 5.5 6.0 6.5 5.0 5.5 6.0 6.5 PH Figure 7. Two-dimensional electrophoresis of ?S-labeled glycoprotein fractions. NGl08-15 cells treated for 6 days with IO pM PGEr and logarithmically dividing, untreated control cells were incubated for 18 hr in Dulbecco's mod- ified Eagle's medium that contained 2% fetal bovine serum and 5 PM L-[3sS]methionine (106 &i/ml of medium). The ceils were harvested and washed in isotonic phosphate-buffered saline. The pelleted cells were then lysed. and the proteins were solubilized by the addition of a solution containing 10 mM HEPES (pH 7.4); I50 mM NaCI; I mM CaClz, 1 mM MgClz, 1 mM MnClz; I mM phenylmethylsulfonyl fluoride, and 1% Triton X-100. Cell lysates were centrifuged at 100.6OOg for 1 hr at 3oC. and glycoproteins in the supernatant portions were fractionated by wheat germ agglutinin-. L. culinaris agglutinin-, and R. communis ag- glutininn-agarose column chromatography. 35S- labeled glycoproteins were eluted with 0.5 M N- acetylglucosamine, 0.2 M a-methylmannoside. or 0.2 M lactose, respectively, dialyzed against 0.01 M NH~HCOJ, and lyophilized. Two por- tions from each sample were subjected to two- dimensional polyacrylamide gel electrophoresis on duplicate 7.5% acrylamide gels according to the method of O'Farrell (1975) with I .O x lo6 cpm (IO-20 rg of protein) applied to each gel. Autoradiographs were exposed for 7 hr and I, 2. and 4 days; those shown were exposed for 2 days. Arrows indicate some of the CAMP-de- pendent differences in 35S-labeled glycopro- teins. i.e., control cells vs. cells treated with PGEr. (Data from K.E. Krueger, M.I. Miller. and M. Nirenberg, in prep.) SYNAPTOGENESIS BY NEUROBLASTOMA CELLS 713 In most cases, different species of 35S-labeled glyco- proteins were eluted from the three lectin columns. Most species of 35S-labeled glycoproteins were syn- thesized by both control cells and PGEi-treated cells. However, numerous CAMP-dependent changes in the 35S-labeled glycoproteins were found as summarized in Table 4. Twelve %-labeled glycoproteins were ex- pressed by PGEi-treated cells that were not detected in extracts of logarithmically dividing control cells. In addition, the radioactivities of 29 %-labeled glycopro- teins expressed by cells treated with PGEi were 2.5 to IO-fold higher than those of control cells. CAMP- dependent decreases in the 35S-labeled glycoproteins were also found. Forty-eight 35S-labeled glycoproteins were expressed by control cells but not by cells treated with PGEi. In addition, the radioactivities of 25 %- labeled glycoproteins expressed by control cells were 2.5-6.5fold higher than those of cells treated with PGEi. No changes in the 35S-labeled glycoproteins were detected when cells were incubated with 10 pM PGEi for 20 minutes (not shown here). DISCUSSION Neuroblastoma hybrid cells were obtained that syn- thesize ACh and form synapses with cultured striated muscle cells. Other cell lines were found that syn- thesize ACh, adhere to myotubes, but form few or no synapses. Different kinds of presynaptic defects were identified, such as cell lines without voltage-sensitive Ca+ + channel activity, large dense-core vesicles, depolarization-dependent ACh secretion, or active pro- tein that increases the aggregation of nicotinic AChR on myotube plasma membranes. No immediate effect of CAMP was detected on voltage-sensitive Ca+ + channel activity, ACh secre- tion, or the percentage of myotubes that were inner- vated. In contrast, these properties and synaptogenesis were regulated by prolonged elevation of CAMP levels of hybrid cells. Activation of adenylate cyclase with 10 PM PGEl for 5 days, or treatment of cells with I mM dibutyryl CAMP, resulted in lo-lOO-fold increases in voltage-sensitive Ca+ + channel activity, 15-45-fold in- creases in spontaneous secretion of ACh at synapses, and 5-lo-fold increases in the number of muscle cells innervated. Treated cells also contained considerably more large dense-core vesicles and small clear vesicles and had higher levels of intracellular ACh than did un- treated cells. Treatment of NBr'lO-A cells with 10 pM PGEi and 1 mM theophylline resulted in the appearance of speci- fic binding sites for [jH]nitrendipine, a putative antag- onist of voltage-sensitive Ca+ + channels. Specific ni- trendipine-binding sites were not detected in SB37-B cells, which lack voltage-sensitive Ca+ + channel activ- ity. These results suggest that cells acquire more voltage-sensitive Ca + + channels when cellular CAMP is elevated for a number of days. By regulating the expression of Ca+ + channels in hybrid cells, CAMP would be expected to function as a conditional switch that modulates the activities of many Ca+ +-dependent reactions in the neural cells and couples these reactions to the cell membrane potential and to transynaptic or hormonal stimuli that affect adenylate cyclase activity. Glycoproteins of NG 1 OS- 15 hybrid cells grown in the presence of [35S]methionine with or without 10 pM PGEi were solubilized and fractionated by wheat germ agglutinin, lentil lectin, or ricin column chromatog- raphy and by two-dimensional gel electrophoresis. Ele- vation of cellular CAMP levels resulted in the disap- pearance of some glycoproteins, the appearance of new glycoproteins of different relative molecular weights, large increases or decreases in the radioactivities of some "5S-labeled glycoproteins, and changes in the iso- electric points of still other 35S-labeled glycoproteins. These results agree with and extend previous reports of differentiation-specific changes in neuroblastoma pro- teins or glycoproteins (Truding et al. 1974, 1975; Charalampous 1977; Garvican and Brown 1977; Prashad et al. 1977; Akeson and Hsu 1978; Prashad and Rosenberg 1978; Rosenberg et al. 1978; Zisapel and Littauer 1979; Littauer et al. 1980; Atkinson and Bramwell 1981; Sugiyama et al. 1980; Imada and Im- ada 1982). Treatment of neuroblastoma or hybrid cells with di- butyryl CAMP has been shown to alter the levels of some species of polysomal mRNA (Morrison et al. 1980; U.Z. Littauer, pers. comm.). When "undif- ferentiated" neuroblastoma cells were shifted to a more differentiated state by altering various growth condi- tions, some species of polysomal poly(A)+ RNA were no longer expressed (Felsani et al. 1978; Grouse et al. Table 4. CAMP-dependent Changes in 35S-labeled Glycoproteins No. of "S-labeled glycoproteins Effects of CAMP on control PGEI `%-labeled glycoproteins cells cells Increased expression 0 I2 Increased expression 29 29t Decreased expression 48 0 Decreased expression 25 251 The experiment is described in Fig. 7 (K.E. Krueger, M. Miller, and M. Nirenberg in prep.). The total number of 35S-labeled roteins detected in extracts of control cells was 317. Twelve additional b S-labeled proteins were detected in extracts of cells that had been treated with PGEI, 714 NIRENBERG ET AL. 1980) and many new species of poly(A)+ RNA ap- peared (Grouse et al. 1980). Marked effects of CAMP or dibutyryl CAMP on mRNA levels of other eukaryotic cells have been reported (Derda et al. 1980; Maurer 1981; Miles et al. 1981; Lamers et al. 1982; Landfear et al. 1982; Wu and Johnson 1982; Mangiarotti et al. 1983). Twelve new 3SS-labeled glycoproteins were expressed by cells treated with PGEI, and 2.5lo-fold increases were found in the radioactivities of 29 additional 3sS- labeled glycoproteins. Concomitantly, neuroblastoma or hybrid cells acquired voltage-sensitive Na+ (Cat- terall et al. 1973). K+ (Nelson et al. 1969), and Ca++ channels; Ca+ +-dependent K + channels; extended long neurites; and acetylcholinesterase specific activity in- creased (Blume et al. 1970). These results suggest that some of the 35S-labeled glycoproteins that are expres- sed to a greater extent in PGEl-treated cells may play a role in some of the newly expressed neuronal mem- brane functions. Synapse-defective cell lines that lack channel activity or the ability to extend neurites and other variant cell lines may provide a means of identify- ing functions of some membrane proteins. In summary, the results show that CAMP regulates synaptogenesis by regulating the expression of voltage- sensitive Ca+ + channels and suggest that CAMP affects posttranslational modifications of some species of glycoproteins and/or regulates gene expression. REFERENCES AKESON. R. and W.C. Hsu. 1978. Identification of a high molecular weight nervous system specific cell surface glycoprotein on murine neuroblastoma cells. Exp. Cell Res. 115: 367, AWANO, T.. B. HAMPRECHT, and W. KEMPER. 1974. High activity of choline acetyltransferase induced in neuroblas- toma x glia hybrid cells. Erp. Cell Res. 85: 399. ATKINSON, M.A.L. and M.E. BRAMWELL. 1981. Studies on the properties of hybrid cells. III. A membrane glycopro- tein found on the surface of a wide range of malignant cells. J. Cell Sci. 48: 147. BLUME: A., F. GILBERT, S. WILSON, J. FARBER, R. ROSEN- BERG, and M. NIRENBERG. 1970. Regulation of acetylcholinesterase in neuroblastoma cells. Proc. Nor/. Acad. Sci. 67: 786. BOLGER, G.T., P.J. GENGO. E.M. LUCHOWSKI, H. SIEGEL, D.J. TRIGGLE. and R.A. JANIS. 1982. High affinity bind- ing of a calcium channel antagonist to smooth and cardiac muscle. Biochem. Biophys. Res. Commun. 104: 1604. CATTERALL, W.A. and M. NIRENBERG. 1973. Sodium uptake associated with activation of action potential ionophores of cultured neuroblastoma and muscle cells. Proc. Narl. Acad Sci. 70: 3759. CHARALAMPOUS. F.C. 1977. Differences in plasma mem- brane organization of neuroblastoma cells grown in the differentiated and undifferentiated states. Arch. Biochem. Biophys. 181: 103. DANIELS, M.P. and B. HAMPRECHT. 1974. The ultrastructure of neuroblastoma glioma somatic cell hybrids. J. Cell Biol. 63: 69 I. DERDA. D.F., M.F. MILES, J.S. SCHWEPPE, and R.A. JUNGMANN. 1980. Cyclic AMP regulation of lactate dehydrogenase: Isoproterenol and N6, 0' `-dibutyryl cyclic AMP increase the levels of lactate dehydrogenase-5 isozyme and its messenger RNA in rat C6 glioma cells. J. Biol. Chem. 255: 11112. EHLERT, F.J., E. IWGA, W.R. ROESKE, and H.I. YAMA- MURA. 1982. The interaction of [3H]nitrendipine with receptors for calcium antagonists in the cerebral cortex and heart of rats. Biochem. Biophys. Res. Commun. 104: 937. FELSANI, A., F. BERTHELOT, F. GROS, and B. CROIZAT. 1978. Complexity of polysomal polyadenylated RNA of undifferentiated and differentiated neuroblastoma cells. Eur. J. Biochem. 92: 569. GARVICAN, J.H. and G.L. BROWN. 1977. A comparative analysis of the protein components of plasma membranes isolated from differentiated and undifferentiated mouse neuroblastoma cells in tissue culture. Eur. J. Biochem. 76: 251. GROUSE, L.D.. B.K. SCHRIER, C.H. LETENDRE, M.Y. ZUBAIRI,. and P.G. NELSON. 1980. Neuroblastoma dif- ferentiation involves both the disappearance of old and the appearance of new poly(A)+ messenger RNA sequences in polyribosomes. J. Biol. Chem. 255: 387 I. HIGASHIDA, H., S.P. WILSON, M. ADLER, and M. NIRENBERG. 1978. Synapse formation by neuroblastoma and hybrid cell lines. Sot. Neurosci. Absrr.4: 591. IMADA, S. and M. IMADA. 1982. Increase of a surface glycoprotein by cyclic AMP in Chinese hamster ovary cells. Dependence on cell-cell interaction. J. Biol. Chem. 257: 9108. LAMERS, W.H., R.W. HANSON, and H.M. MEISNER. 1982. CAMP stimulates transcription of the gene for cytosolic phosphoenolpyruvate carboxykinase in rat liver nuclei. Proc. Natl. Acad. Sci. 79: 5137. LANDFEAR, S.M., P. LEFEBVRE, S. CHUNG, and H.F. LODISH. 1982. Transcriptional control of gene expression during development of Dictyostelium discoideum. Mol. Cell. Biol. 2: 1417. LITTAUER, U.Z., M.Y. GIOVANNI, and M.C. GLICK. 1980. A glycoprotein from neurites of differentiated neuroblastoma cells. J. Biol. Chem. 255: 5448. LOWRY, O.H., N.J. ROSEBROUGH, A.L. FARR, and R.J. RAN- DALL. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193: 265. MANGIAROTTI. G.. A. CECCARELLI, and H.F. LODISH. 1983. Cyclic AMP stabilizes a class of developmentally regulated Dictyostelium discoideum mRNAs. Nature 301: 616. MATSUZAWA, H. and M. NIRENBERG. 1975. Receptor- mediated shifts in cGMP and CAMP levels in neuroblastoma cells. Proc. Nut/. Acud. Sci. 72: 3472. MAURER, R.A. 1981. Transcriptional regulation of the pro- lactin gene by ergocryptine and cyclic AMP. Nurure 294: 94. MCGEE, R., P. SIMPSON, C. CHRISTIAN. M. MATA, P. NELSON, and M. NIRENBERG. 1978. Regulation of acetylcholine release from neuroblastoma- x glioma hvbrid cells. Proc. Narl. Acud. Sci. 75: 1314. MCKEE, R., C. SMITH, C. CHRISTIAN, M. MATA, P. NELSON, and M. NIRENBERG. 1980. A new capillary tube system for measuring the uptake and release of materials from cultured cells. Anal. Biochem. 101: 320. MILES, M.F., P. HUNG, and R.A. JUNGMANN. 1981. Cyclic AMP regulation of lactate dehydrogenase: Quantitation of lactate dehydrogenase M-subunit messenger RNA in iso- proterenol and N6, 02'-dibutyryl cyclic AMP-stimulated rat C6 glioma cells by hybridization analysis using a cloned cDNA probe. J. Biol. Chem. 256: 12545. MINNA. J., D. GLAZER, and M. NIRENBERG. 1972. Genetic dissection of neural properties using somatic cell hybrids. Nat. New Biol. 235: 225. MORRISON, M.R., S. PARDUE. N. PRASHAD, D.E. CROALL. and R. BRODEUR. 1980. Relative increase in polysomal mRNA for RI CAMP-binding protein in neuroblastoma SYNAF'TOGENESIS BY NEUROBLASTOMA CELLS 715 cells treated with 1, ND-dibutyryl-adenosine 3',-5'-phosphate. Eur. J. Biochem. 106: 463. MURPHY, K.M.M. and S.H. SNYDER. 1982. Calcium an- tagonist receptor binding sites labeled with 13H]nitren- dipine. Eur. J. Pharmacol. 77: 201. NELSON, P.G., C. CHRISTIAN, and M. NIRENBERG. 1970. Synapse formation between clonal neuroblastoma x glioma hybrid cells and striated muscle cells. Proc. Nat/. Acad. Sci. 73: 123. NELSON, P.. W. RUFFNER, and M. NIRENBERG. 1969. Neuronal tumor cells with excitable membranes grown in vitro. Proc. Natl. Acad. Sci. 64: 1004. NELSON, P.G., C.N. CHRISTIAN, M.P. DANIELS, M. HENKART, P. BULLOCK, D. MULLINAX. and M. NIRENBERG. 1978. Formation of synapses between cells of a neuroblastoma x glioma hybrid clone and mouse myotubes. Brain Res. 147: 245. O'FARRELL, P.H. 1975. High resolution two-dimensional electrophoresis of proteins. J. Eiol. Chem. 250: 4007. POTTER, L.T. and W. MURPHY. 1967. Electrophoresis of acetylcholine, choline, and related compounds. Biochem. Pharmacol. 16: 1386. PRASHAD, N. and R.N. ROSENBERG. 1978. Induction of cyclic AMP-binding proteins by dibutyryl cyclic AMP in mouse neuroblastoma cells. Biochim. Biophys. Acra 539: 459. PRASHAD, N., B. WISCHMEYER, C. EVETTS, F. BASKIN, and R. ROSENBERG. 1977. Dibutyryl CAMP-induced protein changes in differentiating mouse neuroblastoma cells. Cell Differ. 6: 147. ROSENBERG, R.N., C.K. VANCE, M. MORRISON, N. PRASHAD, J. MEYNE, and F. BASKIN. 1978. Differentia- tion of neuroblastoma, glioma, and hybrid cells in culture as measured by the synthesis of specific protein species: Evidence for neuroblast glioblast reciprocal genetic regulation. J. Neurochem. 36: 1343. . - ROTTER. A., R. RAY; and M. NIRENBERG. 1979. Regulation of calcium uptake in neuroblastoma or hybrid &Is. A possible mechansim for synapse plasticity. Fed. Proc. 38: 476. SEEDS, N., A.G. GILMAN. T. AMANO, and M. NIRENBERG. 1970. Regulation of axon formation by clonal lines of a neural tumor. Proc. Narl. Acad. Sci. 66: 160. SHARMA, S.K., W.A. KLEE, and M. NIRENBERG. 1975a. Dual regulation of adenylate cyclase accounts for narcotic dependence and tolerance. Proc. Narl. Acad. Sci. 72: 3092. -. 1977. Opiate-dependent modulation of adenylate cyclase. Proc: Narl. A&d. Sci. 74: 3365. SHARMA. S.K.. M. NIRENBERG. and W.A. KLEE. 1975b. Morphine receptors as regulators of adenylate cyclase ac- tivity. Proc. Natl. Acad. Sci. 72: 590. SPERRY, R.W. 1963. Chemoaffinity in the orderly growth of nerve fiber patterns and connections. Proc. Nafl. Acad. Sci. 50: 703. SUGIYAMA, K., M. TOMIDA, Y. HONMA, and M. HOZUMI. 1980. Expression of a cell surface glycoprotein (~180) related to cell-substratum adhesion during differentiation of mouse myeloid leukemia cells. Cancer Res. 40: 3387. TORU, M. and M.H. APRISON. 1966. Brain acetylcholine studies: A new extraction procedure. J. Neurochem. 13: 1533. TRIGGLE, D.J. and V.C. SWAMY. 1983. Calcium antagonists. Some chemical-pharmacologic aspects. Circ. Res. 52: 17 (Suppl. I.). TRUDING, R., M.L. SHELANSKI, and P. MORELL. 1975. Glycoproteins released into the culture medium of dif- ferentiating murine neuroblastoma cells. J. Biol. Chem. 250: 9348. TRUDING, R., M.L. SHELANSKI, M.P. DANIELS. and P. MOR- RELL. 1974. Comparison of surface membranes isolated from cultured murine neuroblastoma cells in the differen- tiated and undifferentiated state. J. Eiol. Chem. 249: 3973. Vo, K.P., M.J. MILLER, E.P. GEIDUSCHEK, C. NIELSON, A. OLSON, and N.H. XUONG. 1981. Computer analysis of two-dimensional gels. Anal. Biochem. li2: 258. - WILSON, S., H. HIGASHIDA. I. MINNA. and M. NIRENBERG. 1978. Defects in synapse format& and acetylcholine release by neuroblastoma and hybrid cell lines. Fed. Proc. 37: 1784. WV, J.R. and L.F. JOHNSON. 1982. Regulation of dihydrofolate reductase gene transcription in methotrexate resistant mouse ftbroblasts. J. Cell. Physiol. 110: 183. ZISAPEL, N. and U.Z. LITTAUER. 1979. Expression of external-surface membrane proteins in differentiated and undifferentiated mouse neuroblastoma cells. Eur. J. Biochem. 95: 51.