Communication Vol. 268 No. 7 Issue of March 5 pp. 4550-4583,1993 THE JOURNAL OF BIOLOGICAL CHEMISTRY

0 1993 by The American Societifor Bibchemistry and Mblecular Biology, Inc. Printed in W.S.A.

SAPSO, a Rat Presynaptic Protein Related to the Product of the Drosophila Tumor Suppressor Gene dlg- A*

(Received for publication, November 30, 1992) Ute KistnerS, Bettina M. Wenzel, Riidiger W. Vehs, Claudia Cases-Langhoff, Abigail M. Garner, Ute Appeltauer, Britta Voss, Eckart D. Gundelfinger, and Craig C. Garner7 From the Center for Molecular Neurobiology, University of Hamburg, Martinistrasse 52, 0-2000 Hamburg 20, Germany and the Slnstitute for Anatomy, Ruhr University Bochum, W-4630 Bochum 1, Germany

A novel synapse-associated protein, SAPSO, accu- mulates around the axon hillock of Purkinje cells in rat cerebellum. By immuno-electron microscopy, SAP90 has been localized to the presynaptic termini of basket cells forming inhibitory, 7-aminobutyric acid (GABA)ergic synapses onto Purkinje cell axon hil- locks. The amino acid sequence for SAP90 has been deduced from the nucleotide sequence of a series of overlapping cDNA clones. SAP90 is related to the gene product encoded by the Drosophila tumor suppressor gene dlg-A. SAP90 and the dlg-A product share an overall sequence identity of 54%. Three distinct do- mains can be identified: (i) a potential cytoskeletal region consisting of three repeats of 90 amino acids in length, (ii) a domain with similarity to SH3, a putative regulatory motif found in the 8rc family of non-recep- tor protein tyrosine kinases and several proteins as- sociated with the cortical cytoskeleton, and (iii) a car- boxyl-terminal domain homologous to yeast guanylate kinase. These features suggest a possible role for SAP90 in a guanine nucleotide-mediated signal trans- duction pathway at a subset of GABAergic synapses in the rat cerebellum.

Synapses are dynamic structures assembled from a complex collection of cytoskeletal, membrane receptor, ion channel, and extracellular matrix proteins whose organization and functional states are regulated via synaptic activity and mul- tiple second messenger systems (1, 2). The identification of regulatory and structural constituents of synapses is essential to understand synaptogenesis and synaptic plasticity. Pro-

* This work was supported by the Federal Government of Germany (Bundesministerium fur Forschung und Technologie). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “adver- tisement’’ in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper hos been submitted to the GenBankTM/EMBL Data Bank with accession number($ X664 74.

$ Submitted in partial fulfillment of the requirements for a Ph.D. from the University of Hamburg.

ll To whom correspondence should be addressed Center for Molec- ular Neurobiology, University of Hamburg, Martinistrasse 52, D-2000 Hamburg 20, Germany.

teins associated with synaptic vesicles and their role in exo- cytosis and transmitter release have been well characterized in the last decade (3).

Structural components of synapses have been best studied at the neuromuscular junction, where many proteins involved in the proper localization of acetylcholine receptors have been described (4). In addition to actin and spectrin, at central nervous system synaptic junctions only a limited number of structural and regulatory proteins have been characterized at the molecular level. These include the subunits of the calcium/ calmodulin dependent protein kinase I1 (5), the F1-20 protein (6), Gephyrin (7), N-CAM (8), NP185 (9), and SNAP-25 (10). Utilizing a combined immunological and molecular approach, we have identified a number of novel and known components in synaptic junctional preparations.

In the present study we describe the cloning of cDNA encoding synapse-associated protein (SAP)’ 90, a protein found enriched in synaptic junctional preparations and lo- cated at the presynaptic termini of a subset of inhibitory synapses in the rat cerebellum.

EXPERIMENTAL PROCEDURES

Protein Isolation, SDS-Polyacrylamide Gel Electrophoresis, and Western Blot Analysis-Proteins were isolated from P30 rat brains for total brain homogenates, supernatants, and pellets and for the fractionation of synaptic vesicles and synaptic junctions as described by Gordon-Weeks (11). Tris extraction was as described by Hayes et al. (12). Protein fractionation by SDS-PAGE on 10% gels, transfer to nitrocellulose membrane (Schleicher & Schull) and immunostain- ing of blots with peroxidase-conjugated secondary antibodies (Di- anova) were performed as previously described (13).

Antibody Generation and Purification-100 pg of protein from a purified synaptic junctional preparation (11) were used in a 1:l emulsion with Freund’s complete adjuvant in the first immunization and Freund’s incomplete adjuvant in two additional boosts to generate a polyclonal antiserum in rabbits. Antibody RA2d was affinity-puri- fied from this antiserum using the Xgtll-expressed P-galactosidase- clone 2d fusion protein (13). Antibody MA2d was affinity-purified from a mouse polyclonal antiserum generated against glutathione S- transferase-clone 2d fusion protein, constructed by cloning the 2d cDNA into the pGEX-2T vector (Pharmacia). The fusion protein was purified as described by the manufacturer before immunization of mice. Affinity-purified antibodies were isolated with fusion-proteins immobilized on nitrocellulose filters at 2 pg/cm2. Filters blocked in 5% skim milk were incubated with sera diluted 1:250 for 2 h, washed in phosphate-buffered saline, and the antibody subsequently eluted with glycine buffer, pH 2.7 (13).

Construction and Screening of a cDNA Library-The P25 RNA and a cDNA kit (Pharmacia) were used to construct a random hexanucleotide-primed cDNA library in bacteriophage Xgtll. Double- stranded cDNA with NotI-EcoRI linkers was ligated into the EcoRI site of Xgtll. Phage DNA was packaged with Gigapack Gold

dish on Escherichia coli strain Y1090. The library was screened with (Stratagene) and plated at 50,000 plaque-forming units/l5-cm Petri

the polyclonal antiserum against purified synaptic junctional pro- teins. Positive plaques were visualized by immunoperoxidase staining as described by Garner et al. (13). To reduce nonspecific binding of the polyclonal antisera during screening, the sera were first exposed to nitrocellulose filters presoaked in a bacterial extract that was prepared from sonicated and then lyophilized Y1090 bacteria resus- pended to 5% in TBS (50 mM Tris-HC1, pH 7.4, 150 mM NaCI). Antibodies affinity-purified with the expressed fusion protein from

The abbreviations used are: SAP, synapse-associated protein; SH3, src homology domain 3; PAGE, polyacrylamide gel electropho- resis; EM, electron microscopy; GABA, y-aminobutyric acid; kb, kilobase pair(s).

4580

Synapse-associated Protein SAP90 4581

each purified X clone (13) were used to stain Western blots of brain membrane to determine the molecular weight of the encoded proteins and to stain p30 rat cerebellar sections to assess potential association with synaptic rich regions. cDNAs from plaque-purified X clones were subcloned into pBluescript (Stratagene) and sequenced using T7 polymerase (Pharmacia) as described by the manufacturers. Nucleo- tide and deduced protein sequences were compared to GenBank, EMBL, and NBRF-PIR data bases with search programs from DNASTAR (Madison, WI).

RNA Isolation and Northern Analysis-The isolation of RNA from different tissues and Northern analysis were performed as previously described (13).

Immunocytochemistry and Electron Microscopy-Sprague-Dawley rats under deep anesthesia were fixed by vacuolar perfusion with a solution of 4% paraformaldehyde, 0.05% glutaraldehyde, and 0.2% picric acid in 0.1 M phosphate buffer, pH 7.4. Fixation was followed by 0.15 M sucrose in 0.13 M phosphate buffer, pH 7.4. The brains were removed, frozen at -80 "C after a 48-h pretreatment with 0.09 M sucrose in 0.13 M phosphate buffer, pH 7.4, and stored for up to 3 months. For immunofluorescence staining was performed on 16-pm cryostat sections as previously described (14). Immunoperoxidase staining and electron microscopy was performed as previously de- scribed (15). For controls, sections were incubated with either the affinity-purified preimmune serum or the monospecific anti-SAP90 serum after preabsorption with the nitrocellulose bound antigen, or by omitting the primary antiserum.

cDNA Cloning and Sequencing-Clone 2d was used to isolate clones 2d25 and G3 from a rat brain oligo(dT)-primed XZAPII library (kindly provided by Steve Morley, University of Hamburg) and clone KT 3 from a brain XgtlO library (Clonetech). Double-stranded DNA was sequenced on both strands using the T7-DNA polymerase system (Pharmacia).

RESULTS AND DISCUSSION

Identification and Characterization of a Novel SAP-cDNA clones encoding potential synapse-associated proteins (SAPS) were isolated by screening a rat brain Xgtll expression library with a rabbit polyclonal antiserum generated against a puri- fied synaptic junctional preparation. A rabbit antibody, affin- ity-purified using the fusion protein encoded by the X clone 2d (RABd), was immunoreactive with a 90-kDa protein pres- ent only in brain membrane preparations (Fig. la) and spe- cifically localized in the cerebellum around the Purkinje cell axon hillock (data not shown). The molecular weight and the subcellular distribution of the encoded protein was confirmed by staining Western blots and rat brain cerebellar sections with a mouse affinity-purified antibody (MA2d) generated against the coding region of clone 2d fused with glutathione S-transferase (Figs. 1 and 2). The antibodies RA2d and MA2d recognize a protein designated SAP90 with the same mobility in brain membrane preparations (Fig. la). By fluorescence microscopy, MA2d in cerebellar sections reacted intensely in a patchy pattern around the Purkinje cell axon hillock (Fig. 2b). Immunoperoxidase staining of sections, for immuno- electron microscopy, at the light microscopic level revealed that stellate and basket cells as well as the molecular layer are also stained (Fig. 2d). The peroxidase reaction product a t the EM level was found in the soma of stellate and basket cells and in the presynaptic boutons of symmetric synapses in the molecular layer, on the Purkinje cell soma (Fig. 2e), and around the Purkinje cell axon hillock. Synapses on the Purkinje cell soma and around the axon hillock (diagrammed in Fig. 2a) are GABAergic (16) and originate from basket cells, the axons of which can be visualized with a neurofila- ment antibody (Fig. 2c). The data demonstrate that SAP90 is a presynaptic protein located in part at GABAergic synapses. Interestingly, the 6 isoform of protein kinase C exhibits a similar distribution to SAPSO, although it is not known if the kinase is located in synapses (17).

Is SAP90 in presynaptic termini associated with synaptic vesicles or the membrane cytoskeleton? This was examined

a

kDa

100

70

1 2 3

f

k b

6.2

3.8

b

C I-

1 2 3 4

9

d

c e

5 1 2 3 4

1 2 3 4 5 1 2 3 4 5 6 1 8

FIG. 1. Distribution of SAP90 and its encoding RNA. Panel a, immunoblots of postnatal day 30 (P30) rat brain membrane protein preparations. Strips of nitrocellulose were incubated with affinity- purified antibodies against the polypeptide encoded by clone 2d. Lane I , rabbit antibody RA2d; lane 2, mouse antibody MA2d diluted 1:20; lane 3, MA2d used neat. Panels b and c, P30 brain membranes stained with MA2d ( b ) or a monoclonal against spectrin (c). Lane 1, 30,000 X g supernatant of brain membranes extracted with 1% Triton X- 100,0.3 M NaCl (pH 8.0), 1 mM EDTA, 0.5 mM dithiothreitol, 0.02% phenylmethylsulfonyl fluoride; lane 2, 30,000 X g pellet of membrane extract in lane 1. This membrane pellet was then sequentially ex- tracted and pelleted at 150,000 X g with 1.0 M Tris, pH 8.0 (lane 3), 1.0 M Tris, and 1% Triton X-100 (lane 4 ) , 1% SDS (lane 5). Panels d and e, immunoblots of P30 rat brain subcellular fractions (16 pg of each) stained with the MA2d antibody ( d ) or with a synaptophysin monoclonal antibody (e). Lane 1 , synaptosomes; lune 2, osmotically lysed synaptosomes; lane 3, synaptic vesicle preparation; lane 4, synaptic junction membranes. Panel f, Northern blot of total rat brain RNA (10 pg) from postnatal days 1, 7, 15, 30, and 75 (lanes 1-5, respectively) probed with the cDNA from clone 2d. Panelg, Northern blot of total RNA (10 pg) from P30 brain, heart, muscle, liver, lung, kidney, thymus, and testes (lanes 1-8, respectively).

by fractionating brain membranes over sucrose density gra- dients (11). SAP90 was not only enriched in the synaptosome fractions, but when fractionated further to separate synaptic vesicles from synaptic junctions, it was found highly enriched in the latter (Fig. Id) in contrast to synaptophysin (Fig. le). SAP90 was also found to be tightly bound to brain mem- branes. It could not be released with alkali, 0.5 M MgC12, or 1% Triton X-100 (not shown); however, about 30% could be extracted along with spectrin with 1.0 M Tris, pH 8.0, or completely with 1% SDS (Fig. 1, b and c). These data dem- onstrate that SAP90 is tightly bound perhaps with the cortical cytoskeleton in presynaptic termini, but not directly via actin, ankyrin, or spectrin known to be released by Tris extraction (12).

At much higher concentrations of the MA2d antibody ad- ditional proteins of a similar molecular weight to SAP90 are present in brain membranes (Fig. la). Whether all are related is not clear; however, we have identified a second set of cDNA clones encoding a 97-kDa protein related to SAP90 that is expressed in the cerebellar molecular layer: suggesting that a

* B. M. Wenzel, U. Kistner, R. W. Veh, C. Cases-Langhoff, A. M. Garner, B. Voss, B. D. Gundelfinger, and C. C. Garner, manuscript in preparation.

4582 Synapse-associated Protein SAP90

sb.

FIG. 2. Immunostaining of rat brain cerebellar sagittal sec- tions. a, diagram of cerebellum immunofluorescence photo of section stained with MA2d ( b ) or neurofilament monoclonal antibody (c; Sigma). d, MA2d immunoperoxidase staining. MA2d reacts specifi- cally with basket cell axons and synapses around the Purkinje cell axon hillock (ah). With immunoperoxidase, the molecular layer (ml) as well as stellate ( s t ) and basket (bc) cells stained. The neurofilament antibody stained axon fibers in the molecular and granule cell layer ( g l ) as well as basket cell axons (ba) surrounding the Purkinje cell axon hillock. pcl, Purkinje cell layer; mf, mossy fibers; cf, climbing fibers. e, electron micrograph of a MA2d-stained symmetric synapse of a basket cell axon on a Purkinje cell body. The upper half of the terminus contains dark reaction product, while the lower half is filled with densely packed, oval-shaped synaptic vesicles. Below the syn- aptic cleft (arrows) there are three subsynaptic cisterns (cis), which may be artificially widened due to the immunocytochemical treat- ment. Ng means glial cell process; bar represents 0.1 pm.

family of these proteins exists, perhaps with each at a different type of synapse.

SAP90 is expressed primarily in the brain. Based on North- ern analysis of different rat tissues, the cDNA from clone 2d hybridizes specifically to a prominent 3.8-kb mRNA and to a lesser extent with a 6.2-kb mRNA both found only in brain (Fig. lg). The 6.2-kb mRNA disappears during development, whereas the 3.8-kb mRNA increases with a peak at postnatal day 15 (Fig. l f ) correlating with the late appearance of the protein (not shown).

Primary Structure of SAP90 Deduced from cDNA Se- quence-The structure of SAP90 was characterized by using clone 2d to isolate a series of overlapping cDNAs (Fig. 3a). The determined nucleotide sequence for these cDNAs con- tains an open reading frame for a protein with 724 amino acid residues (Fig. 3b) and a calculated M, of 80,396. Analysis of the deduced amino acid sequence indicates that SAP90 is a hydrophilic protein lacking membrane-spanning domains or hydrophobic sequences that could insert it into the plasma membrane.

A protein data base search (NBRF-PIR) revealed that SAP90 is structurally similar to the Drosophila dlg-A gene product (dlg-Ap) (18) (Fig. 3b). The two proteins share 54% identical amino acid residues that cover the entire length of SAPSO, but are interrupted by two additional domains only

A ATG H P H H A P TGA A A

I " I I

24 - c"---( 200bp

G3 I 4 2425 t I

I I KT3

FIG. 3. Restriction map of cDNA clones encoding SAP90 and comparison of the deduced amino acid sequences (in sin- gle-letter notation) to the dlg-A protein. Panel a, restriction map and organization of cDNA clones used to establish the nucleotide sequence of SAPSO. Labeled restriction sites: H, HincII; P, PstI; A, ApaI. The filled box indicates the 2172-base pair open reading frame. An in-frame stop codon was found at nucleotide -45 upstream of the proposed ATG coding for the initiator methionine. Panel b, the deduced amino acid sequence of SAP90 is aligned with dlg-Ap (3), v- crk (30), and the yeast guanylate kinase (5) protein sequences. Iden- tical amino acids are bored; the three 90-amino acid residue repeats, the SH3 domain, and the guanylate kinase domain are double-under- lined and marked. The underlined amino acids in the guanylate kinase domain represent residues that are involved in nucleotide binding. Repeats 1 and 2 show a stronger similarity to each other than to repeat 3. The SH3 domain shows the strongest similarity to the SH3 domain of v-crk, although there are some additional amino acids in the sequences of SAP90 and dlg-Ap.

found in dlg-Ap (Figs. 3b and 4). The identities of the two proteins are clustered in three domains (Figs. 3b and 4). The first domain (amino acid residues 64-397 of SAPSO) contains three novel 90-amino acid repeats and shows similarities to a variety of cytoskeletal proteins (19, 20). The second domain (amino acid residues 435-493 of SAPSO) is a SH3 (src homol- ogy 3) region (21). The third domain (amino acid residues

Synapse-associated Protein SAP90 4583

SAPBO I 1 . ' I I

rl R r3 SH3 guanylate kinase . . . . 1 , .,.:..:.:.:.. , r \ \ \ \ \ \ \ \ \ \ y 5&a

dlg-A t I I I 1 Baau, PEST

.... /i , . , I - \\\\\\\\\1 r l R opa r3 SH3 guanyiate kinase

FIG. 4. Schematic comparison. of the primary structural similarities between SAP90 and dlg-A. The 90-amino acid resi- due repeats ( r l , r2, and r3), SH3, and guanylate kinase domains are conserved in both proteins. Found only in dlg-Ap is an opa repeat and a PEST sequence (3). The latter is thought to be involved in rapid degradation of proteins.

534-724 of SAPSO) is similar to the yeast guanylate kinase (22), with all amino acid residues essential for nucleotide binding conserved (23). SAPSO, therefore, appears to be a mosaic molecule. In the carboxyl-terminal end, it possesses a putative guanylate kinase involved in guanine nucleotide me- tabolism, whereas the amino-terminal half contains a poten- tial cytoskeletal domain that could attach SAP90 to the synapse-associated cytoskeleton, where it may also perform a structural role. Interestingly, SAP90 has a SH3 domain that is found in the src family of non-receptor tyrosine kinases, in several membrane cytoskeletal proteins and membrane-asso- ciated signal transduction proteins (21,24). SH3 domains are thought to mediate the clustering of cytoplasmic signal-trans- ducing proteins at the plasma membrane (25) and to regulate their activity (26). In a similar fashion, the presence of a SH3 domain in SAP90 could function to regulate the guanylate kinase or to cluster signal transduction proteins in the pre- synaptic terminus.

The overall structural similarity of SAP90 to dlg-Ap (Fig. 4) provides clues to the function of SAP90 at inhibitory synapses. The dlg-A gene is a tumor suppressor gene with its protein product located at the septate junction of the em- bryonal and larval imaginal disc (18). Septate junctions are thought to be the invertebrate equivalent of the vertebrate tight junctions (27). Although synapses and tight junctions are not anatomically very similar, both are sites with a highly organized membrane cytoskeleton (28, 29). In homozygous dlg mutants, the imaginal disc epithelial cells proliferate be- yond their normal boundaries (30). In addition the septate junctions do not form (31) and exhibit a disorganized actin membrane cytoskeleton (18). It was therefore proposed that dlg-Ap is probably involved in organizing the membrane cytoskeleton (18); SAP90 may have a similar role at synapses.

The potential guanylate kinase at the carboxyl end of SAP90 (Fig. 36) implies that this protein could either be involved in the cGMP cycle (32) and/or in regulating GTP synthesis. In the first case, SAP90 could be part of a guanine nucleotide signal transduction process at inhibitory synapses. Such a pathway involving an increase in cGMP concentration, caused by nitric oxide activation of guanylate cyclase, is correlated to the formation of long term depression at excit- atory synapses (2, 33). Similarly, cGMP second messenger

systems are involved in signal transduction at the photore- ceptor cells of the retina (34). As an alternative, SAP90 could influence GTP levels altering the activity of receptor-coupled G-proteins or small Ras-like GTP-binding proteins (35). One interesting candidate in presynaptic termini is ra63, a ~21'""- like protein involved in neurotransmitter vesicle fusion and release of neurotransmitter from nerve termini (3, 36).

Acknowledgments-We thank Klaus M. Herrmann for critical reading of the manuscript.

REFERENCES 1. Lynch, G., Bodsch, W., and Baudry, M. (1988) in Intrinsic Determinants

of Form and Function: Neurology and Neurobiology (Lasek, R.J., and Black, M.M., eds) Vol. 37, pp. 217-243, Alan R. Liss, New York

2. Fazeli, M. S. (1992) Trends Neurosci. 16 , 115-117 3. Siidhof, T. C., and Jahn, R. (1991) Neuron 6,665-677 4. Froehner, S. C. (1991) J. CellBiol. 1 1 4 , 1-7 5. Wu, K., Sachs, L., Carlin, R. K., and Siekevitz, P. (1986) Mol. Brain Res.

1 , 167-184 6. Sousa, R., Tannery, N. H., Zhou, S., and Lafer, E. M. (1992) J. Neurosci.

7. Prior, P., Schmitt, B., Grenningloh, G., Pribilla, I., Multhaup, G., 12,2130-2143

Beyreuther, K., Maulet, Y., Werner, P., Langosch, D., Kirsch, J., and

8. Persohn, E., Pollenberg, G. E., and Schachner, M. (1989) J. Comp. Neurol. Betz, H. (1992) Neuron 8, 1161-1170

288,92-100 9. Su, B., Hanson, V., Perry, D., and Puszkin, S. (1991) J. Neurosci. Res. 2 9 ,

10. Oyler, G. A., Higgins, G. A., Hart, H. A., Battenberg, E., Billingsley, M.,

11. Gordon-Weeks P. R. (1987) in Neurochemistry (Turner, A. J., and Bach-

12. Hayes, N. V. L., Rayner, D. A., and Baines, A. J. (1991) J. Neurochem. 67,

13. Garner, C. C., Farmer, L., and Matus, A. (1988) Gene (Amst.) 7 1 , 483-

14. Lichte, B., Veh, R. W., Meyer, H. E., and Kilimann, M. W. (1992) EMBO

15. Mayer, B., Hamaguchi, M., and Hanafusa, H. (1988) Nature 332,272-274 16. Somogyi, P., Hodgson, A. J., Chubb, I. W., Penke, B., and Erdei, A. (1985)

J. Histochem. Cytochem. 33,240-248 17. Wetsel, W. C., Khan, W. A,, Merchenthaler, I., Rivera, H., Halpern, A. E.,

Phung, H. M., Negro-Vllar, A., and Hannun, Y. A. (1992) J. Cell Biol.

18. Woods, D. F., and Bryant, P. J. (1991) Cell 66,451-464 117,121-133

19. Domenighini, M., Relman, D. Ca iau, C., Falkow, S., Prugnola, A., Scarlato, V., and Ra puoli R (1990) dl. Microbiol. 4,787-800

20. Ramirez, F., iosat, 'S.,'D'Alessio, M., Lee, B., Prince, J., Su, M. W., Vissing, H., and Yoshioka, H. (1990) Ann. N . Y. Acad. Sci. 680,74430

21. Musacchio, A,, Gibson, T., Lehto, V.-P., and Saraste, M. (1992) FEBS Lett. 307,55-61

22. Berger, A,, Schiltz, E., and Schulz, G. E. (1989) Eur. J. Biuchem. 184,433-

461-473

Bloom, F. E., and Wilson, M. C. (1989) J. Cell Biol. 109,3039-3052

elard, H. S., kds) pp. 1-26, IRL Press, Oxford

397-405

49013

J. 11,2521-2530

23. Stehle, T., and Schulz, G. E. (1990) J. Mol. Biol. 2 1 1 , 249-254 24. Drubin, D. G., Miller, K. G., and Botstein, D. (1988) J. Cell Biol. 107 ,

443

3551 -35G1 25. Koch, C. A., Anderson, D., Moran, M. F., Ellis, C., and Pawson, T. (1991)

26. Pawson, T. (1988) Oncogene 3,491-495 27. Noirot-Timothee, C., and Noirot, C. (1980) Int. Reu. Cytol. 63,97-140 28. Landis, D. M. D. (1988) in Intrinsic Determinants of Form and Function:

Neurolozv and Neurobiologv (Lasek. R. J.. and Black. M. M.. eds) Vol.

"" " _ _ Science 262,668-674

37, pp. 501-216, Alan R. I&, New pork '

29. Da Silva. P. P.. and Kachar. B. (1982) Cell 28.441-450 30. Stewart,"., M u hy, C., and Fristrom, J. (1972) Deu. Biol. 2 7 , 71-83 31. Woods, D. F., anTBryant, P. J. (1989) Deu. Bzol. 134,222-235 32. Hall, S. W., and Kuhn, H. (1986) Eur. J. Biochem. 161,551-556 33. Garthwaite. J. (1991) Trends Neurosci. 14. 60-67 34. Stryer, L. (1986)Annu. Reu. Neurosci. 9 , 87-119 35. Bourne, H. R., Sanders, D. A., and McCormick, F. (1990) Nature 3 4 8 ,

36. Fischer von Mollard, G., Siidhof, T. C., and Jahn, R. (1990) Nature 349 , 125- 132

79-81