kinase in Drosophila: Identification of a synapsin I-like protein

Proc. Natl. Acad. Sci. USA Vol. 86, pp. 5988-5992, August 1989 Neurobiology Dynamic properties of the Ca2+/calmodulin-dependent protein kinase in Dro...
Author: Geoffrey Burke
0 downloads 1 Views 1MB Size
Proc. Natl. Acad. Sci. USA Vol. 86, pp. 5988-5992, August 1989 Neurobiology

Dynamic properties of the Ca2+/calmodulin-dependent protein kinase in Drosophila: Identification of a synapsin I-like protein H. MITSCHULAT Institut fur Biologie III der UniversitAt Freiburg, Schanzlestrasse 1, D-7800 Freiburg i. Br., Federal Republic of Germany

Communicated by Martin Lindauer, April 7, 1989 (received for review February 3, 1989)

ABSTRACT Visual adaptation with blue light induces a change in a special light/dark choice behavior in Drosophila. On the molecular level adaptation induces long-term modulation of the in vitro autophosphorylation capacity of a Ca2+/calmodulin-dependent protein kinase. Here I describe a Drosophila phosphoprotein that is a substrate of this protein kinase. The molecular mass and phosphopeptide composition of this protein are similar to those of rat synapsin I. Furthermore, the Drosophila protein shows immunological crossreactivity with monoclonal antibodies against rat synapsin I. I conclude that this 86-kDa protein in Drosophila is homologous to the vertebrate synapsin I.

tions that it is involved in the mechanism of synaptic transmission (for review, see ref. 3). Bahler and Greengard (12) described an actin-bundling activity associated with synapsin I. In their working model synapsin serves to anchor the vesicles to a spectrin/F-actin network in the cytoplasm and prevents prefusion of the vesicles with the presynaptic membrane; therefore, no transmitter release occurs. Bahler and Greengard (12) and Petrucci and Morrow (13) could show that the actin-bundling activity is controlled by phosphorylation. Only phosphorylation by the Ca2+/calmodulin-dependent protein kinase II reduces the actin-bundling and binding activity, whereas phosphorylation by cAMP-dependent protein kinase has a minimal effect on synapsin-actin interaction. Here I report evidence that the autophosphorylation capacity of the 50-kDa Ca2+/calmodulin-dependent protein kinase and its subcellular distribution can be modified by sensory stimulation. In addition, a substrate protein with properties similar to those of vertebrate synapsin I is found in Drosophila. A preliminary report of this work has been presented (14).

Drosophila melanogaster shows a light/dark phototactic choice behavior, which is modified by prolonged visual adaptation. This behavioral plasticity can be seen as a simple form of learning. It can be observed for hours, even if the visual stimulus is removed and the adaptational state of the responding retinula cells is restored (1, 2). Protein phosphorylation has been shown to be involved in mediating neural plasticity during many learning processes (for a review, see ref. 3). Changes in the activity of protein kinases and phosphatases, modulated by different second messenger molecules, provide a mechanism for regulating phosphorylation. Two of the enzymes involved in a simple learning process in invertebrates are the cAMP-dependent and the Ca2+/ calmodulin-dependent protein kinases. For example, in Aplysia the following presynaptic mechanism has been described (4). Brief electric stimuli increase synthesis of intracellular cAMP, activating a cAMP-dependent protein kinase, which phosphorylates a protein at the potassium channel, thus causing a decrease in K+ current. The decrease in K+ current prolongs the action potential, and this in turn increases the duration of the inward Ca2+ current. Ca2+ stimulates other enzymes, including the Ca2+/calmodulin-dependent protein kinase II. In Aplysia this enzyme has been shown to be translocated from the membrane after phosphorylation into the cytoplasm, where it is apparently activated, becoming independent of added Ca2+ (5). In vertebrates synapsin I is a substrate for both Ca2+/ calmodulin-dependent protein kinases (6-9) as well as of cAMP-dependent protein kinase (6, 8, 10). cAMP-dependent protein kinase, Ca +/calmodulin-dependent protein kinase I, and Ca2+/calmodulin-dependent protein kinase II phosphorylate synapsin I at different sites (3). Synapsin I is found in the central and peripheral nervous systems. It is composed of two polypeptides of 78 and 74 kDa, originally called proteins Ia and Ib (10). Later, the nomenclature was changed: protein I is called synapsin I (6). The two polypeptides exhibit similar properties, including immunological cross-reactivity. Synapsin I is associated with the membrane of synaptic vesicles. There are many indica-

MATERIALS AND METHODS Electrophysiology. Electroretinogram (ERG) recording techniques used were identical to those described by Broda and Wright (15). Phosphorylation Assay. Heads of 15 instantly frozen flies were cut off in the cold (liquid nitrogen, - 1960C) and quickly homogenized (10 sec) in 50 p1L of potassium phosphate buffer (0.1 M, pH 6.8) containing 2 M glycerol. For some experiments this crude head extract was separated into a membrane fraction and a cytosolic fraction by centrifugation either at 40,000 x g for 20 min or at 400,000 x g for 10 min at 4(C. Protein phosphorylation was assayed in potassium phosphate buffer (50 mM, pH 6.8) containing 5 mM MgCl2, 2 p.M ATP, and [a-32P]ATP (300-600 Ci/mmol; 1 Ci = 37 GBq). Other additions (1 mM CaCl2, 50 p.M cAMP, or 10 mM EGTA) are as indicated in each case. The reaction was carried out at room temperature (18-220C) for 2.5 min and stopped by addition of a half volume of Tris-HCl buffer (0.5 M, pH 6.8) containing 10% 2-mercaptoethanol and 6% SDS. Release of the 50-kDa Protein Kinase from the Membrane Fraction. Extensively washed membrane fractions were phosphorylated in potassium phosphate buffer (50 mM, pH 6.8) containing 5 mM MgCI2, 1 mM CaC12, and ATP, as indicated. After subsequent centrifugation, a half volume of Tris HCl buffer (see above) was added to the supernatant and the remaining pellet was resuspended in the same buffer. Proteins were separated by gel electrophoresis (see below). In some experiments synapsin I purified from rat (a gift of L. J. DeGennaro, Max Planck Institute for Biochemistry, Martinsried) was added to the phosphorylation assay.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Abbreviation: ERG, electroretinogram.

5988

Neurobiology: Mitschulat SDS/PAGE. By using the method of Laemmli (16), proteins were separated on 12.5% gels. The gels were stained with Coomassie blue R-250, destained, dried, and exposed to Kodak X-Omat AR film. Autoradiograms were scanned with a Kontron Uvikon 810 spectrophotometer. Immunoblots. Immunoblotting was carried out as described by Towbin et al. (17). Samples were run on 12.5% gels and electroblotted onto nitrocellulose filters presaturated with TBS buffer (10 mM Tris, pH 7.4/0.9% NaCl/0.02% NaN3) containing 2% bovine serum albumin overnight. Incubation with antibodies against rat synapsin I (generous gifts of L. J. DeGennaro) was carried out at room temperature for 1 hr. After washing with TBS containing an additional 0.05% SDS and 0.1% Triton X-100 for 1.5 hr, the filters were incubated with the second antibody (anti-mouse IgG) for 1 hr and washed again. Filters were stained by using second antibody-linked alkaline phosphatase; the reaction was carried out in a Tris HCl buffer (0.1 M, pH 8.8) containing 0.1 M NaCl, 0.005 M MgCl2, 0.05% (wt/vol) 5-bromo4chloro-3-indolyl phosphate, and 0.017% (wt/vol) nitroblue tetrazolium. Peptide Mapping. By using the method of Cleveland et al. (18) the proteins were digested partially after in vitro phosphorylation of the proteins of the different fractions and SDS/PAGE. Protein bands from dried SDS gels were cut out, rehydrated in a Tris HCl buffer (0.125 M, pH 6.8) containing 0.001 M EDTA and 0.1% SDS. The sections were placed in the sample wells of a second SDS gel, overlayered with Staphylococcus aureus V8 protease (0.004 unit per sample; purchased from Sigma), and digestion was carried out throughout the run in the stacking gel. After staining with Coomassie blue R-250 and destaining, the gels were dried and autoradiographed by exposing to Kodak X-Omat AR film. Cleavage with Cyanogen Bromide. Sections of dried gels were cut out and each gel section was incubated for 24 hr at room temperature in 0.2 ml of 1.3% cyanogen bromide (purchased from Sigma) in 70% aqueous formic acid; controls were incubated in acid alone. Acid was removed from the extracts by evaporation to dryness from water, and pellets were washed with 0.2 ml of H20 and dried again for another three times. The residues were resuspended in 50 ,ul of 2% SDS, 2% 2-mercaptoethanol, and 1o sucrose at 100°C. The solutions were neutralized with 2.5 M Tris HCI (pH 8.0) and the peptides were separated on a second SDS gel [method described by Eppler et al. (19)].

Proc. Natl. Acad. Sci. USA 86 (1989)

5989

Light

offI

on

mV

a)

cpm Tr I a

600j

I

b)

400+

200+

.I

cpr n

c

2001

loot

C)

0

1'

2' 5'

6'

FIG. 1. Comparison of the electrophysiological and biochemical effects of visual stimulation. (a) Redrawing of a typical ERG record of a wild-type fly during stimulation. The previously dark-adapted fly is stimulated by white light and after 5 min the light is switched off. (b) In vitro phosphorylation of the 50-kDa protein kinase from head homogenates during the same visual adaptation (two different experiments are shown). (c) In vitro phosphorylation of the 50-kDa kinase in membrane fractions during visual adaptation.

autophosphorylation capacity (back-phosphorylation) can be measured in the membrane after the light is switched on (Fig. lc). Release of the Protein Kinase from the Membrane Fraction.

RESULTS Effects of Visual Adaptation on ERG Activity and Ca2l Dependency of the Ca2+/Calmodulin-Dependent Protein Kinase. The extent to which phototactic behavior of Drosophila is controlled by the two different visual subsystems (peripheral receptor cells R1-R6 and the central receptor cells R7 and R8) has been the subject of many studies (20-26). If the living fly is exposed to white light after prolonged adaptation in the dark, ERG monitoring shows that elicited response of the photoreceptor cells is much quicker than modulation of both the cAMP-dependent protein kinase (27) and the Ca2+/ calmodulin-dependent protein kinase (Fig. 1 a and b), which can be identified as phosphoproteins from crude extracts of the fly brain. The in vitro autophosphorylation capacity of the Ca2+/ calmodulin-dependent protein kinase increases quickly if the light is switched on, and after some seconds it decreases to a lower level than measured initially in the dark-adapted flies. If the light is switched off, it slowly increases again. The autophosphorylation of dark-adapted flies can be stimulated in vitro with added Ca2 , whereas it seems to be nearly Ca2+-independent after light-adaptation of the flies (28). Parallel to the dynamics of autophosphorylation of the 50-kDa kinase in crude extracts, a decrease of 50-kDa kinase

Fractionating crude head extracts by centrifugation yields autophosphorylation activity of the enzyme predominantly in the membrane fraction. As has been reported (28), a certain amount ofthe kinase can be solubilized after incubation of the membrane with Ca2+, Mg2+, and ATP. Fig. 2 shows that the release only occurs in the presence of ATP (four right-hand lanes). Furthermore, Fig. 2 shows that the release of the enzyme does not depend on Ca2+: the amounts of label in the two lanes on the right are not significantly different. In contrast, the enzyme activity in phosphorylating another substrate does depend on Ca2+. For example, in vitro phosphorylation of rat synapsin I by the released enzyme is stimulated by Ca2' (14).

Identification of Ca2+/Calmodulin-Dependent Protein Kinase in Different Fractions. Comparison of the patterns of

labeled phosphopeptides of the 50-kDa protein kinase of different fractions after digestion with protease of Staphylococcus shows that the protein kinase in crude head extracts, in membrane and cytoplasmic fractions, and in the supernatant containing the released enzyme are all identical. The molecular masses of the labeled fragments are 14, 12, 11, and 9 kDa (Fig. 3B). Furthermore, after chemical cleavage with cyanogen bromide, a phosphorylated fragment with a molecular mass of 11 kDa was found in all fractions (Fig. 3A).

Neurobiology: Mitschulat

5990

+

- ATP +-

Proc. Natl. Acad. Sci. USA 86 (1989)

+-

+ -

A

ATP +-

a

b

d

c

Ca2+

kDa

9 kDa =

57 -50

kDa

t

-"57

-50 11

-+

B

-+ a _

FIG. 2. ATP-dependent release of the Ca2+/calmodulindependent protein kinase from the membrane. Extensively washed membrane fractions were incubated with potassium phosphate buffer containing MgC12 and CaCl2 or MgCl2 and EGTA in the absence (-) (four left-hand lanes) or presence (+) (four right-hand lanes) of [32P]ATP and subsequently centrifuged. The two left lanes of each block show the phosphoproteins remaining in the membrane fraction; the two lanes at the right side show those that are released from the membrane. Without ATP no Ca2+/calmodulin-dependent protein kinase can be labeled by further phosphorylation following centrifugation.

The 86-kDa Phosphoprotein. We detected a phosphoprotein with the molecular mass of 86 kDa in crude head extracts and membrane fractions, which acted as an endogenous substrate of the Ca2+/calmodulin-dependent protein kinase. This protein shows phosphorylation states strongly correlated to the autophosphorylation of the 50-kDa protein kinase

(Fig. 4).

Differences in the labeling of both Ca2+/calmodulindependent protein kinase and the 86-kDa protein could be detected in extracts of both dark and light preadapted flies (data not shown). In the presence of Ca2+ the labeling of both was increased (Fig. 5). The effect of cAMP on the phosphorylation of the 86-kDa protein was also tested. Since the cAMP-dependent kinase is found predominantly in the cytoplasmic fraction (27), the phosphorylation in this fraction was tested. Phosphorylation in the presence of cAMP (Fig. 6, lane a) results in a higher level of labeling of the 86-kDa protein than in the absence of the second messenger (Fig. 6, lane b). In parallel, the amount of label in the 57-kDa regulatory subunit of the cAMPdependent protein kinase phosphorylated in the presence of cAMP (0.01 mM) (Fig. 6, lane a) is higher than in the assay without cAMP (Fig. 6, lane b). Fig. 7, lane a, shows a Western blot with monoclonal antibodies against rat synapsin I bound to Drosophila 86-kDa protein of a membrane fraction (lane b shows the corresponding autoradiogram of the 32P-labeled phosphoproteins). Antibody specificity is demonstrated in Fig. 8a. There is a strong antibody binding to 86- and 80-kDa proteins of rat brain (crude extracts); binding to other proteins cannot be detected (Fig. 8b shows the corresponding autoradiogram). To identify the 86-kDa protein in Drosophila, both the 86-kDa protein and rat synapsin I were digested with V8 protease. The proteins are very similar with respect to their phosphopeptide composition (Table 1). Only a few differences can be found: in the pattern of the fly the 16.1-kDa peptide is missing, and the 14.5-kDa peptide in synapsin I is

b VW

± c

_

+

d .

9

kDa 14.0

9.0 IL

FIG. 3. Identification of the Ca2+/calmodulin-dependent protein kinase in different fractions by chemical cleavage with cyanogen bromide (B) and digestion with Staphylococcus aureus V8 protease (A). Shown is the peptide pattern of the kinase of crude head extracts (a), membrane fraction (b), cytoplasmic fraction (c), and the supernatant containing the released kinase (d). Cyanogen bromide treatment (+), controls (-), as well as the digestion were carried out with protein of eight flies each.

at the position of 14.3 kDa. The other four peptides in the low molecular mass range and the two peptides of 34 and 38 kDa are identical. The apparent differences in the high molecular mass range can be explained by the fact that too much protein from rat compared to Drosophila was digested by the same amount of protease in a limited time, leading to only partially digested high molecular mass fragments.

DISCUSSION In agreement with findings in Aplysia (5) and Hermissenda (29), a Ca2+/calmodulin-dependent protein kinase appears to be involved in long-term adaptational processes in Drosophila neural tissues. This involvement is suggested by the increased phosphorylation of the 50-kDa Ca2+/calmodulindependent protein kinase measured in vitro in preparations from blue-light-adapted flies compared with those adapted to white light (2, 28). Adaptation of the flies to darkness leads to a high amount of autophosphorylation capacity (i.e., after in vitro back-phosphorylation the amount of label is, in-

creased) of the membrane-bound 50-kDa protein. If the flies are exposed to light, the amount of autophosphorylation capacity decreases within seconds. In contrast, the capacity of the 50-kDa kinase of previously blue-light-adapted flies cannot be restored quickly, instead lasting for hours (28). This paper describes the dynamics of biochemical changes

induced by visual stimulation of the fly with white light and some of the properties of the enzymes and one involved substrate. Comparison of ERG (Fig. la) and autophosphorylation of the 50-kDa protein kinase (Fig. lb) points out that the response of the photoreceptor cells is much quicker than

.

Neurobiology: Mitschulat 86 kDa 57 kDa

Proc. Natl. Acad. Sci. USA 86 (1989)

5991

a b 0

1.0

kDa

86 > 80 go

."-

0-

FIG. 6. The 86-kDa protein in the cytoplasmic fraction. Crude extracts were centrifuged and the cytoplasmic fraction was phosphorylated in the presence (lane a) or absence (lane b) of added cAMP (0.01 mM). The 86kDa protein as well as 57-kDa cAMPdependent protein kinase show a higher amount of label in the presence of cAMP than in its absence. In contrast, addition of cAMP to the membrane fraction has no effect on phosphorylation of the 86-kDa protein (data not shown).

-a57 -a50

00

0.5

0

0 0

0

0

*

*S.

.

0

* *-

0

0

r = 0.78 P l" 0.01 e

0

0.5

1.0

1.5 50 kDa

57 kDa

FIG. 4. Correlation of phosphorylation of the 54 0-kDa protein kinase to 86-kDa phosphoprotein. Each point represe nts a separate experiment. After adaptation of the flies to different lij ght conditions (blue, red, yellow, and white light, as described in rref. 28), phosphorylation was carried out in crude head extracts of six flies each.

the modulation of the enzyme. After illuminat ion of darkadapted flies autophosphorylation of50-kDa kina tse increases quickly for about 5 sec and drops within the ne xtminute to) a level lower than that observed in the dark (Fi Since labeled phosphate added to the in vitro be bound only to those sites of the kinase th at were not ation cn can be phosphorylated in vivo, only back-phosphoryh ition . observed. Therefore, during the first seconds of visualistimulation many more sites of the kinase might be occupied by labeled ATP than in the later course of illumii suggests in vivo dephosphorylation of the ki dark-adaptation. . to simlar Fractionating of crude extracts showed thai t, simir toud Aplysia (5), the dephospho form of the enzyn predominantly to the membrane fraction, and lesis is found in

ga mbgh

b

a w

OR~~~m

the cytoplasm (28). In both organisms after phosphorylation and subsequent centrifugation the protein kinase is partially released from the membrane and can be detected in the supernatant (5, 28). For Drosophila this release does not depend on stimulation with Ca2' but is caused by phosphorylation. The disappearance of label on the kinase monitored in the membrane fraction during illumination after darkadaptation (Fig. ic) leads me to the assumption that the active form of the enzyme is the phosphorylated one, which might be found in the cytoplasm after in vivo labeling. No differences in the amount of label in the released protein can be detected after autophosphorylation of the membrane-bound kinase dependent on the presence of in vitro added Ca2+ (Fig. 2). This is in agreement with the finding that the Ca2+-dependent component of the autophosphorylation of the released 50-kDa protein kinase in Aplysia is reduced (5). Ca2 -dependent enzyme activity in whitelight-adapted flies is different from the activity in flies adapted

to darkness

(28). From these observations

the fol-

lowing hypothesis has been developed. In vivo the protein kinase of dark-adapted flies is dephosphorylated and bound to the membrane. It is assumed that sensory stimulation and subsequent increase in neural activity in the optical lobes cause Ca2 -dependent autophosphorylation and the ATPdependent release of the enzyme. The phospho form may be active in the cytoplasm to phosphorylate substrate proteins. Peptide mapping by digestion with a protease and chemical

cleavage with cyanogen bromide of the protein kinase derived from the different fractions strongly suggest that the enzyme is identical wherever it is found. Furthermore, no a

b

-.

kDa 8 6

50

FIG. 5. Ca2+/calmodulin-dependency of the 86-kDa protein and 50-kDa protein kinase. A membrane fraction washed with buffer containing 3 mM EDTA was phosphorylated in the presence (a) or absence (b) of added CaC-12 (0.1 mM) and calmodulin (1 ELM). Labeling of the 86-kDa protein and 50-kDa protein kinase is increased in the presence of Ca2' and calmodulin.

FIG. 7. Western blot of antibodies against rat synapsin I bound to Drosophila protein. Lane a, binding of the antibodies against vertebrate synapsin I to the 86-kDa protein (arrowhead) of Drosophila; lane b, the corresponding autoradiogram of the in vitro labeled membrane fraction of the fly.

5992

_m

Proc. Natl. Acad. Sci. USA 86 (1989)

Neurobiology: Mitschulat b

a

rSyn I

b

*-Syn I a

FIG. 8. Western blot of antibodies against rat synapsin I. (a) Specificity of the antibodies. Strong antibody binding of the two phosphopeptides of synapsin I (Syn I) with molecular masses of 80 and 86 kDa is shown. (b) The corresponding autoradiogram of the phosphoproteins in crude brain extracts of rat.

differences in enzyme phosphopeptide composition can be found between white-light-, blue-light-, or dark-adapted flies (data not shown). Phosphorylation of an 86-kDa phosphoprotein in the brain of Drosophila is correlated to the autophosphorylation of the Ca2+/calmodulin-dependent protein kinase (Fig. 4). This protein therefore might be one of the substrates of Ca2+/ calmodulin-dependent protein kinase in the nerve cell. Because of the similar molecular mass to vertebrate synapsin Ia, I tried to find out whether both synapsin I and the 86-kDa protein share further common properties. Synapsin I is reported to be not only a substrate of Ca2+/calmodulin-dependent protein kinase but also of the cAMP-dependent protein kinase. Since cAMP-dependent protein kinase in Drosophila was found predominantly in the cytoplasmic fraction, the cAMP-dependent phosphorylation of the 86-kDa protein in that fraction was tested. Phosphorylation of the 86-kDa protein is shown to be altered in Fig. 6; the amount of label is higher in the presence of added cAMP than when no cAMP is added. Monoclonal antibodies against rat synapsin I bind to the 86-kDa protein in the membrane fraction of Drosophila (Fig. 7). The immunological cross-reactivity indicates that rat synapsin I and Drosophila 86-kDa protein share common epitopes. Furthermore, the proteins are very similar with respect to their phosphopeptide composition: molecular mass of four of six peptides in the low molecular mass range are identical, and one (14.3 kDa) is slightly shifted (14.5 kDa). Table 1. Peptide composition of rat synapsin I and Drosphila 86-kDa protein Drosophila Rat 86-kDa protein synapsin I Peptide, kDa High molecular mass 44.0 42.0 38.0 36.0 34.0 Low molecular mass 17.0 16.0 + (14.3 kDa) 14.5 13.3 12.1 +

+

11.5 +, Existing;

-,

missing.

Differences in the high molecular mass range might be explained by only partially digested protein in the case of rat synapsin I, but the two peptides of the Drosophila protein pattern can be found in the pattern of synapsin I too. Since molecular mass and phosphopeptide composition of both proteins are similar, both proteins are substrates of a Ca2+/ calmodulin-dependent kinase, phosphorylation of both is partially dependent on cAMP, and immunological crossreactivity could be shown, I conclude that the 86-kDa protein in Drosophila is homologous to vertebrate synapsin I. Several proteins that cross-react with a monoclonal antibody raised against mammalian synapsin I were identified in insects (J. Buxbaum, Y. Dudai, R. Jahn, and P. Greengard, personal communication). These proteins had a molecular mass of 70-225 kDa. In Drosophila a protein of 130 kDa cross-reacted with four unique monoclonal antibodies raised against mammalian synapsin I. This protein is acid-soluble and is a substrate for cAMP-dependent protein kinase. Because of the dynamic properties of the 50-kDa protein kinase during visual stimulation and long-term adaptational processes, I speculate that it is involved in the plasticity of transmitter release in the optical lobes. This mechanism, which can be analyzed easily in Drosophila, might be involved in neural information storage during adaptational processes or even learning. I thank Dr. L. J. DeGennaro for the generous gift of antibodies against rat synapsin I and him as well as Drs. R. Willmund, J. Buxbaum, and R. Cassada for critical reading of the manuscript. The work was supported by the Deutsche Forschungsgemeinschaft. 1. Willmund, R. & Fischbach, K. F. (1977) J. Comp. Physiol. 118, 261-271. 2. Willmund, R. (1986) J. Insect Physiol. 32, 1-8. 3. Nestler, E. J. & Greengard, P. (1984) Protein Phosphorylation in the Nervous System (Wiley, New York). 4. Schwartz, J. H. & Greenberg, S. M. (1987) Annu. Rev. Neurosci. 10, 459-476. 5. Saitoh, T. & Schwartz, J. H. (1985) J. Cell Biol. 100, 835-842. 6. Huttner, W. B., DeGennaro, L. J. & Greengard, P. (1981) J. Biol. Chem. 256, 1482-1488. 7. Kennedy, M. B., McGuinness, T. & Greengard, P. (1983) J. Neurosci. 3, 818-831. 8. Huttner, W. B. & Greengard, P. (1979) Proc. Natl. Acad. Sci. USA 76, 5402-5406. 9. Nairn, A. C. & Greengard, P. (1987) J. Biol. Chem. 262, 7273-7281. 10. Ueda, T. & Greengard, P. (1977) J. Biol. Chem. 252, 5155-5163. 11. Kilimann, M. & DeGennaro, L. (1985) EMBO J. 4, 1997-2002. 12. Bahler, M. & Greengard, P. (1987) Nature (London) 326, 704-707. 13. Petrucci, T. C. & Morrow, J. S. (1987) J. Cell Biol. 105, 1355-1363. 14. Mitschulat, H. & Willmund, R. (1988) in Modulation of Synaptic Transmission and Plasticity in Nervous Systems, eds. Hertting, G. & Spatz, H.-Ch. (Springer, Berlin), pp. 387-400. 15. Broda, H. & Wright, R. (1978) J. Insect Physiol. 24, 681-684. 16. Laemmli, U. K. (1970) Nature (London) 227, 680-685. 17. Towbin, H., Staehelin, T. & Gordon, J. (1979) Proc. Natl. Acad. Sci. USA 76, 4350-4354. 18. Cleveland, D. W., Fischer, S. G., Kirschner, M. W. & Laemmli, U. K. (1977) J. Biol. Chem. 252, 1102-1106. 19. Eppler, C. M., Bayley, H., Greenberg, S. M. & Schwartz, J. H. (1986) J. Cell Biol. 102, 320-331. 20. Harries, W. A., Stark, W. S. & Walker, J. A. (1976) J. Physiol. (London) 130, 161-171. 21. Stark, W. S. (1975) J. Comp. Physiol. 96, 343-356. 22. Heisenberg, M. & Buchner, E. (1977) J. Comp. Physiol. 117, 127-162. 23. Jacob, K. G., Willmund, R., Folkers, E., Fischbach, K. F. & Spatz, H.-Ch. (1977) J. Comp. Physiol. 116, 209-225. 24. Hu, K. G. & Stark, W. S. (1977) J. Comp. Physiol. 121, 241-252. 25. Fischbach, K. F. (1979) J. Comp. Physiol. 130, 161-171. 26. Miller, G. V., Hansen, K. N. & Stark, W. S. (1981) J. Insect Physiol. 27, 813-819. 27. Mitschulat, H. & Willmund, R. (1988) Insect Biochem. 18, 551-556. 28. Willmund, R., Mitschulat, H. & Schneider, K. (1986)Proc. Natl. Acad. Sci. USA 83, 9789-9793. 29. Acosta-Urquidi, J., Alkon, D. L. & Neary, J. T. (1984) Science 224, 1254-1257.

Suggest Documents