Neuronal cAMP-dependent protein kinase type II is concentrated in mushroom bodies of Drosophila...

Neuronal cAMP-Dependent Protein Kinase Type II IsConcentrated in Mushroom Bodies of Drosophilamelanogaster and the Honeybee Apis mellifera

Uli Muller

Institut fur Neurobiologie der Freien Universitat Berlin, Konigin-Luise-Str. 28-30,14195 Berlin, Germany

Received 29 January 1997; accepted 28 February 1997

ing mediated via cAMP-dependent signaling, the mod-ABSTRACT: In both Drosophila melanogasterulatory functions of transmitters on PKA activity inand the honeybee Apis mellifera, cyclic adenosineKenyon cells from the honeybee were tested. Agentsmonophosphate (cAMP)-dependent processes havewhich elevate cytoplasmic Ca2/ levels have no effectsbeen implicated in mechanisms of learning. This studyon PKA activity in cultured Kenyon cells. Dopamine,characterizes the type II cAMP-dependent protein ki-serotonin, and octopamine, however, cause an in-nase (PKAII), the major target of cAMP in adultcrease in PKA activity in Kenyon cells. The modula-animals. In both species, PKAII is restricted to neu-tion of PKA activity by octopamine, the putativeronal tissue, in which it accounts for more than 90%transmitter of the unconditioned stimulus in associa-of total PKA activity. Although the intensity of PKAIItive olfactory learning in the honeybee, together withimmunoreactivity differs between distinct brain re-the findings on the central role of the cAMP cascadegions, labeling is detectable in all neuropiles and mostin Drosophila mushroom bodies, suggests a major im-somata. While the visual neuropiles, the antennalplication of PKAII-mediated phosphorylation inlobes, and structures of the central brain exhibit inter-learning and memory in both Drosophila and Apis.mediate immunostaining, the mushroom bodies showq 1997 John Wiley & Sons, Inc. J Neurobiol 33: 33–44, 1997high labeling and contain a three- to fourfold higherKeywords: Kenyon cells; learning; memory; mush-PKA activity compared to other neuropiles. Since theroom body; PKAmushroom bodies are central sites of olfactory learn-

INTRODUCTION (Frank and Greenberg, 1994; DeZazzo and Tully,1995). Investigations in Aplysia, the most detailedanalyzed model system, indicated that the cAMPThe major role of the second-messenger cycliccascade and phosphorylation via cAMP-dependentadenosine monophosphate (cAMP) in the modula-protein kinase (PKA) are fundamental for changestion of neuronal activity and its implication in learn-in synaptic transmission (Kandel and Schwartz,ing and memory have been well documented (Wa-1982; Byrne, 1987). In the honeybee, the modula-laas and Greengard, 1991). Recent studies havetion of the PKA activity by the unconditioned stim-demonstrated that cAMP-dependent phosphoryla-ulus in olfactory conditioning suggests an implica-tion of transcription factors seem to be a basic re-tion of the cAMP cascade in mechanisms of associa-quirement for the formation of distinct forms oftive learning (Hildebrandt and Muller, 1995a,b) .long-term memory in vertebrates and invertebratesCharacterization of the behavioral Drosophila mu-tants dunce and rutabaga confirms the general andbasic function of the cAMP cascade in learning andContract grant sponsor: Deutsche Forschungsgemeinschaft;memory. While rutabaga encodes a Ca2/ /calmodu-contract grant number: SFB 515/C3

q 1997 John Wiley & Sons, Inc. CCC 0022-3034/97/010033-12 lin-dependent adenylyl cyclase, an enzyme respon-

33

1839/ 8P1F$$1839 05-27-97 13:07:54 nbioa W: Neurobio

34 Muller

mutants dunce and rutabaga were used. (g-32P) adeno-sible for the formation of cAMP, dunce encodes asine triphosphate (ATP) was obtained from NEN/Du-phosphodiesterase, an enzyme for cAMP degrada-Pont. Phosphatase inhibitor I was purified from bovinetion (reviewed in Davis, 1993; Davis et al., 1995).brain according to the procedure described for skeletalRecent work presents evidence that associative ol-muscle (Huang and Glinsmann, 1976). For the purifica-factory learning in Drosophila is impaired by thetion, DEAE-Sephacel (Pharmacia) , CM-Sephadex C-50

expression of constitutively activated Gas in the (Sigma), Sephacryl S-300 (Pharmacia) , and 8-(2-mushroom bodies (Connolly et al., 1996). Although aminohexyl)amino cAMP-Agarose (Sigma) were used.the exact targets of cAMP are unknown in Drosoph- Molecular weight kits were obtained from Sigma. Tissueila, several investigations suggest a major role of Tek II was obtained from Lab-Tek Products. All chemi-the PKA in learning and memory. Both the expres- cals were of analytical grade.sion of a PKA inhibitor in wild-type flies (Drain etal., 1991) and a decrease of the PKA activity in

Preparation of Tissuesmutants displaying a defect in the catalytic subunit(DCO) (Skoulakis et al., 1993) impair learning. The preparation of tissues was carried out under perma-However, only RI and C subunits have been cloned nent liquid nitrogen cooling. Freeze-drying of Drosophilain Drosophila (Foster et al., 1988, Kalderon and and honeybee heads facilitated the dissection of the tis-Rubin, 1988), although there exist different types sues (Muller and Spatz, 1989; Hildebrandt and Muller,

1995a). After separation, equal amounts of protein fromof PKA characterized by distinct R subunits (RIeach tissue were homogenized in 100-mL extractionand RII) (Taylor et al., 1990).buffer (50 mM Tris-HCl, pH 7.5; 0.1 M NaCl; 1 mMPKA are ubiquitous proteins, and their highlyEDTA; 1 mM EGTA; and 10 mM 2-mercaptoethanol)conserved structure and properties have been dem-at 07C in a microhomogenizer. After homogenization,onstrated in Drosophila melanogaster and the hon-samples were used immediately.eybee Apis mellifera (Foster et al., 1984, Muller

and Spatz, 1989, Taylor et al., 1990; Altfelder andMuller, 1991). Binding of cAMP to the regulatory Determination of PKA Activity(R) subunits of the holoenzyme (R2C2) causes the

The PKA activity was determined using phosphatase in-dissociation of active catalytic (C) subunits fromhibitor 1 (I1) as substrate. I1 was specifically phosphory-the R subunit. The free C subunits can phosphory-lated by insect PKA, while phosphorylation by other ki-late various substrate proteins and thus modulatenases was negligible (Hildebrandt and Muller, 1995b).the function of the latter. After the decay of cAMP,The kinase activity was measured by adding 10 mL of

the free C subunits are inactivated by reassociation the sample to 20 mL of phosphorylation buffer [50 mMwith the R subunits. Tris-buffer, pH 7.5, containing 0.1 M NaCl, 10 mM

The fact that type II PKA is the predominant MgCl2 , 10 mM 2-mercaptoethanol, 2 mg I1, 30 mM ATP,form in adult D. melanogaster as well as in A. mel- 0.1 mCi (g-32P)ATP (5000 Ci/mmol)] with or withoutlifera (Foster et al., 1984; Muller and Spatz, 1989; cAMP (Hildebrandt and Muller, 1995a,b) . The reaction

was terminated by the addition of 6 mL of sodium dodecylAltfelder and Muller, 1991), and the major role ofsulfate (SDS) sample buffer (0.5 M Tris-buffer, pH 6.8,PKA in learning and memory demand for a furthercontaining 5% SDS, 5% 2-mercaptoethanol, and 20%characterization of type II PKA. The present studyglycerol) after 60 s at 207C. Samples were subjected toinvestigated the distribution of the regulatory sub-SDS-polyacrylamide gel electrophoresis (PAGE) and 32Punit RII of the type II PKA in Apis and Drosophilalabeling was visualized by exposing the gel to Kodak X-using biochemical and immunological techniques.Omat AR films. Autoradiograms were scanned with aThe results demonstrate that in both species, typeKontron Uvikon 810 spectrophotometer, and 32P incorpo-

II PKA accounts for the major part of PKA and is ration into I1 was quantified using NIH Image. Conditionshighly concentrated in the mushroom bodies. This, were chosen to keep the exposure of the films in lineartogether with the supposed role of the cAMP cas- range. For calibration of exposure of the film, 32P incorpo-cade in mediating mechanisms of olfactory learning ration into I1 was determined using a scintillation counter.in the mushroom bodies, makes the type II PKA aprimary candidate in mediating the cAMP-depen-

Determination of Transmitter-Evokeddent processes implicated in learning and memory.Modulation in PKA Activity in Kenyon

MATERIALS AND METHODS Cells from the HoneybeeMaterials Kenyon cells were dissected from honeybee brains 3 days

before adult eclosion, as described by Kreissl and BickerApis mellifera were taken from a regular hive. Drosophilamelanogaster, wild-type strains Canton-S, Berlin, and the (1992). To allow neurite outgrowth, cells were cultured


PKA II in Drosophila and Honeybee 35

for 4 days in eight-well strips (Greiner; 1F8 strip) . An row of wells, each sample was diluted 11 times one byone. This dilution range (from 1 to 1/2048) covers theeight-well strip contains Kenyon cells from two animals

(É80,000 Kenyon cells /well) . One hour prior to stimula- linear range for the quantitative determination of the anti-gen. After binding of the antigen to the wells (2 h attion with drugs or neurotransmitters, the medium was

exchanged by 100 mL Ringer’s solution (10 mM phos- 47C), the remaining binding sites were blocked by incu-bation with 1% bovine serum albumin (BSA) in PBS (2phate buffer, pH 7.2, containing 137 mM NaCl, 3 mM

KCl, 5 mM MgCl2, 2 mM CaCl2 , and 50 mM sucrose) h at 47C). To test for differences in the binding of theantibodies to fixated and nonfixated antigen, proteinper well. Prior to stimulation, 80 mL of Ringer’s solution

was removed from the wells and 5 mL of Ringer’s solu- bound to the wells was treated with 4% formaldehyde/PBS for 15 min, followed by washing with PBS (3tion containing the indicated drugs was applied to the

cells using an eight-channel pipette. After 30 s, the incu- 1 10 min). Subsequently, the wells were incubated withthe primary antibodies (1/2000 in 1% BSA/PBS) for 3bation was terminated by transferring the eight-well strip

into liquid nitrogen. PKA activity was determined as de- h at room temperature, or at 47C overnight. After washing(3 1 10 min in PBS), the wells were incubated withscribed by Hildebrandt and Muller (1995a,b) with the

following modifications. An eight-well strip containing biotinylated secondary antibody (1/2000 in 1% BSA/PBS) for 2 h at room temperature. Plates were washedthe frozen Kenyon cells was thawed and 5 mL of phos-

phorylation buffer each [50 mM Tris-buffer, pH 7.5, con- (3 1 10 min in PBS) and streptavidin phosphatase wasapplied for 1 h. After the final washes (3 1 10 min intaining 60 mM MgCl2, 60 mM 2-mercaptoethanol, 12

mM EGTA, 2 mg phosphatase inhibitor 1, 30 mM ATP, PBS), 200 mL substrate solution (0.1 M Tris-HCl, pH8.8, containing 1 mM MgCl2 and 1 mM o-nitrophenyl-0.1 mCi (g-32P)ATP] was immediately applied to the

wells using an eight channel pipette. After 30 s, the reac- phosphate) was added to each well. For quantification ofthe antigen, the substrate conversion by alkaline phospha-tion was terminated by the addition of 6 mL of SDS

sample buffer to each well. Samples were subjected to tase was measured with an ELISA reader (SLT 400ATX)at 405 nm (vs. 620-nm background) in distinct time inter-SDS-PAGE and exposed to Kodak X-Omat AR films as

described above. vals after starting the reaction.Using the method of immobilization of the antigen to

nitrocellulose, protein samples at various concentrationsAntibodies against the Regulatory were bound to nitrocellulose filters (5 1 5 mm). AllSubunit RII procedures with the exception of the staining reaction

were carried out as described above. For the color reac-Antibodies against type II regulatory subunit (RII) puri-tion, the NC filters were incubated in 20 mL of 0.1 M Tris-fied from Drosophila heads (Muller and Spatz, 1989) orbuffer, pH 8.8, containing 0.1 M NaCl, 1 mM MgCl2, 1against RII subunit purified from honeybee brains weremg 5-bromo-4-chloro-3-indolyl phosphate, and 0.5 mgused. Antibodies against the Drosophila RII subunit shownitroblue tetrazolium. After termination of the reaction,very weak cross reactivity with the honeybee RII subunit,the filters with the insoluble staining product wereand vice versa. Antisera show selective reactions onlyscanned with color scanner UMAX UC840, and the inten-with the regulatory subunits RII and RIIp of the corre-sity of the staining was determined using IMAGE (NIH,sponding species. No cross reactivity with other proteinsBethesda, MD).from the head, thorax, or abdomen was found at 1/10

dilutions. For immunohistochemistry and for Westernblotting dilutions of 1/1000, enzyme-linked immunosor-

Immunohistochemistrybent assay (ELISA) dilutions of 1/5000 were used. Pre-absorption of the antisera with purified RII subunit of the Tissues of Drosophila or Apis were fixed in PBS con-corresponding species leads to a total disappearance of taining 4% formaldehyde for 2 h at 47C. For cryosections,reaction in Western blotting, immunohistochemistry, and the tissue was incubated in PBS with 20% sucrose for 20ELISA. h, soaked in Tissue Tec II for 30 min, frozen, and cut in

a cryostat (7 mm for Drosophila and 12 mM for Apis) . Forparaffin sections, the tissue was dehydrated in increasingQuantitative Determinationgrades of ethanol, terminating in xylol, subsequently in-of the RII Subunitcubated for infiltration in paraplast (Sigma), and embed-ded in the latter.Determination of the amount of antigen in membrane and

soluble fraction was achieved by different techniques. The sections were mounted on poly-D-lysine–coatedslides. In the case of the paraplast-embedded tissue, theWhile in the first method the antigen was bound to the

surface of high-bond polyvinyl chloride plates, it was sections were rehydrated. Then the slides were washedtwo times (10 min) with PBS-Tx (PBS containing 0.1%immobilized onto nitrocellulose filters in the second

method. For quantitative determination using ELISA, the Triton X-100). Prior to antibody application, the slideswere blocked against nonspecific binding with blockingantigen (sample) dissolved in phosphate-buffered saline

(PBS) containing 2 mM EGTA and 2 mM EDTA was solution (PBS-Tx containing 1% BSA and 1% goat se-rum) for 1 h at room temperature. The antibodies, dilutedbound to high-bond plates (Falcon; Pro Bind). Within a


36 Muller

1/1000 to 1/2000 in blocking solution, were applied tothe sections and incubated overnight at 47C. The sectionswere washed (3 1 15 min) in PBS-Tx and incubated for2 h at room temperature with biotinylated anti-mouseimmunoglobulin G (IgG) diluted 1:2000 in blocking so-lution. After three 10-min washes, the slides were incu-bated with streptavidin–alkaline phosphatase complex di-luted 1:2000 in blocking solution for 1 h at room tempera-ture. The color reaction was developed in 20 mL 0.1 MTris-buffer, pH 8.8, containing 0.1 M NaCl, 1 mMMgCl2, 1 mg 5-bromo-4-chloro-3-indolyl phosphate, and0.5 mg nitroblue tetrazolium. After termination of thestaining process, sections were dehydrated and mountedwith Entelan.

RESULTSFigure 1 Immunoblot showing the regulatory subunitsRII from neuronal tissues of Apis mellifera and Drosoph-ila melanogaster. Soluble (s) and particulate (p) frac-Characterization of Type II PKAs intions from neuronal tissue were separated on SDS-gels,Adult Apis and Drosophilatransferred to nitrocellulose filters, and subjected to im-munolabeling using antibodies against RII subunit puri-The insect type II PKA is a tetrameric kinase (RII2fied from Apis or Drosophila, respectively. The upper-C2) consisting of two regulatory (RII) and two cata-band doublets (arrows) represent the RII subunits (RII)lytic subunits (C). It has properties similar to thatfrom the tetrameric kinase; the lower-band doubletsof vertebrate type II PKA. Unlike the vertebrate(arrows) represent the RIIp subunits (RIIp) from thetype II PKA, however, the insect PKAII can bedimeric kinase, which is only found in the soluble frac-

activated by cGMP at physiologic concentrations tion. In each doublet, the upper band represents the phos-(Altfelder and Muller, 1991). After activation, the phoform, and the lower band the dephospoform of thefree C subunit is detected only in the soluble frac- regulatory subunits.tion, while the RII subunit is localized in both thesoluble and the particulate fraction (Fig. 1) . Asdescribed for the vertebrate enzyme, the RII subunit subunit is hardly visible (Mr Å 36,000 and Mr

Å 37,000) (Fig. 1) .exists in phospho- and dephosphoform, which showdifferent electrophoretic mobilities in SDS-PAGE With regard to the RII subunit, only in dunce

could a difference be observed compared to the(phosphoform, Mr Å 57,000; dephosphoform, Mr

Å 53,000). This is also true for Apis RII subunits, wild-type strains. The Drosophila mutant dunceshows a higher level of dephosphoforms of RII andbut here the difference between phosphoform (Mr

Å 50,000) and dephosphoform (Mr Å 49,000) is RIIp subunits and an increased amount of RIIp

(Muller and Spatz, 1989).smaller.In addition to this major type II PKA (¢90%)

neural tissues of both Apis and Drosophila contain Distribution of PKA in Adult Apisa minor amount of a dimeric form of a type II PKA and Drosophila(RIIpC) with a smaller regulatory subunit (RIIp)and a single catalytic subunit (C). In both the hon- The distribution of PKA in Apis and Drosophilia

was investigated by immunological detection of RIIeybee and Drosophila, it has been demonstrated thatRIIp derives from specific cleavage of the RII sub- subunit of type II PKA and by biochemical determi-

nation of total PKA activity. Using antibodiesunit (Altfelder and Muller, 1991; Muller and Spatz,1989). In contrast to the RII subunit of the tetra- against the purified RII subunits of each species,

the amounts of RII subunit was quantified by ELISAmeric kinase, the RIIp subunit of the dimeric kinaseis localized only in the soluble fraction. While in technique and by binding of the antigen to nitrocel-

lulose filters. The results derived from the differentDrosophila the phospho- and dephosphoforms ofRIIp show differences in electrophoretic mobility methods of quantification of immobilized antigens

are comparable. Moreover, no difference in binding(Mr Å 39,000 and Mr Å 41,000, respectively) , thechange in electrophoretic mobility of Apis RIIp of the antibodies to formaldehyde-fixated and -non-



Figure 2 Determination of the RII subunit and the total PKA activity in distinct tissues ofApis and Drosophila. The amounts of RII subunit (a,c) and total PKA activity (b,d) of equalamounts of protein of the dissected tissues were compared. Prior to measurement, the tissueswere separated into supernatant and particulate fractions, respectively. The values were normal-ized with respect to the sum of the amounts of RII subunit or PKA activities in the tissuesand set to 1. Each column shows means { S.D. of at least six independent experiments. (al,antennal lobes; ol, optic lobes; cb/mb, central brain/mushroom body; pro, proboscis; th/ab,thorax/abdomen)

fixated protein was observed. Independent of the tained by immunological quantification of RII sub-unit ( type II PKA) matches that of total PKA activ-quantification of RII subunit, and thus, type II PKA,

the total PKA activity was determined by cAMP- ity [Fig. 2; compare (a) with (b) or (c) with (d)] .Comparison of the Drosophila mutants dunce anddependent 32P incorporation into a specific substrate

of PKA. rutabaga to wild-type strains reveals a differencein neither the amount of RII subunit nor the totalIn both Apis and Drosophila, neuronal tissues

contain by far the highest amount of RII subunit as PKA activity (data not shown).As demonstrated by repeated immunoprecipita-well as the highest PKA activity (Fig. 2) . Within

the brain of adult animals, the dissected central tion using the polyclonal antiseras of RII subunit,type II PKA seems to account for the major amountbrain, which contains the mushroom bodies, shows

threefold higher levels of PKA activity per micro- of PKA activity in the neuronal tissues (¢90%).In nonneuronal tissues such as proboscis, thorax,gram of protein than the antennal and visual lobes.

In both species, the ratio between PKA activity in and abdomen, total PKA activity decreased onlyabout 30–50% after immunoprecipitation with RIIthe soluble and particulate fraction is very similar

in distinct brain tissues. Moreover, the ratio ob- subunit antisera. These findings suggest that type II


38 Muller

PKA seems to account for only a part of the mea- nections between sensory neurons, local interneu-rons, and output interneurons which project to thesured cAMP-dependent substrate phosphorylation

in nonneuronal tissue. It is unclear, however, mushroom bodies. In Apis, sensory neurons froman antenna which innervate the rind areas of thewhether these differences are due to yet-unknown

cAMP-dependent kinases in nonneuronal tissue or glomeruli but not the central areas show very weaklabeling. Since the rind areas are labeled at a lowerto distinct inhibitors of PKA activity in neuronal

versus nonneuronal tissues. level and immunostaining is concentrated in thecentral areas of the glomeruli, the staining is mostlikely due to interneurons [Fig. 3(c)] . Somata ofLocalization of the RII Subunit in theantennal lobe interneurons which innervate eitherBrain of Apis and Drosophilathe central area or the whole glomeruli show immu-

The type II PKA in the brain of A. mellifera andnostaining. The glomeruli exhibit a noticeable indi-

D. melanogaster was localized by antibodies againstvidual degree of labeling, with an inhomogeneous

the RII subunit of either species. Immunohisto-distribution within a given glomerulus.

chemistry was performed on both cryosections andIn Drosophila, the antennal nerve which contains

sections of paraffin-embedded tissue. Both proce-the sensory projections into the antennal lobe is

dures led to identical results. Although the intensitydevoid of staining, while the antennal lobe shows

of labeling is very different in distinct regions, im-immunoreactivity [Fig. 3(d)] . The immunostaining

munoreactivity is detectable in all neuropiles andis concentrated in the glomeruli, although the label-

most somata of the brain of both species. The mea-ing within the antennal lobe is more diffuse com-

sured staining intensities in the brain sections arepared to Apis. The inhomogeneously stained glo-

in accordance with the measured amounts of RIImeruli exhibit an individual degree of labeling. In

subunit in neuronal tissue (Fig. 2) . Staining patternsDrosophila, the most striking feature is the very

and intensities of sections of the Drosophila mutantsstrong labeling of clusters of somata [arrows in Fig.

dunce and rutabaga are indistinguishable from sec-3(d)] , which are located at positions described for

tions of the wild-type strains Berlin and Canton-S.antennal lobe interneurons (Stocker, 1994).

Visual Lobes. Using antibodies against RII subunit,Central Brain and Mushroom Bodies. In the cen-all neuropiles in the visual lobes of Apis show im-tral brain, the mushroom bodies are highly orderedmunostaining, while fibers in the outer and innerbilateral symmetric neuronal assemblies which havechiasma were devoid of labeling [Fig. 3(a)] . Com-been implicated in higher control of distinct motorpared to the labeling in the lobula and medulla, theprograms, learning, and memory (deBelle andlamina exhibits a low level of staining. The colum-Heisenberg, 1994; Menzel et al., 1994). They con-nar organization of the neuropile is clearly visible,sist of the calyces, pedunculi, and lobes. The mush-and the different tangential layers within the lobularoom body intrinsic fibers originate from the Ken-and medulla show distinct levels of staining. Somatayon cell somata, which are located around the caly-surrounding the visual neuropiles are labeled, al-ces. The Kenyon cells arborize in the calyces; theirthough at distinct intensities. The fiber tracts con-axons project via the peduncles into the lobes of thenecting the visual lobes with the central brain alsomushroom bodies. The calyces represent the inputexhibit immunoreactivity. In Drosophila, immuno-regions; the lobes are the predominant output areas.staining of the lamina is hardly detectable, while

Compared to other brain areas, the mushroomthe medulla, lobula, and lobula plate exhibit abodies in both species show the strongest immuno-clearly visible staining pattern [Fig. 3(b)] . Fibersreactivity with the RII antisera. In Apis, stainingin the optic chiasma are devoid of staining. Theof the mushroom body neuropil displays a clearlystaining in the medulla is clearly stratified in tangen-compartmentalized pattern. In the calyces, the lip,tial layers and exhibits a columnar organization. Al-which receives chemosensory input, and the collar,though the third-order neuropiles, the lobula andwhich receives mainly visual input, display stronglobula plate, are also organized in columns and lay-staining [Fig. 4(a,b)] . Somata of the Kenyon cellsers, the staining is more homogeneous and distinc-show differences in staining intensity. Somata lo-tion between layers is difficult. The somata of thecated in more lateral areas of the calycal cup exhibitvisual neuropiles are stained at different levels.strong labeling, while somata which are located inthe medial area of the calyces or outside the calycesAntennal Lobes. In both species, the labeling in

the antennal lobes is concentrated in the glomeruli display weaker staining [arrows in Fig. 4(b)] . Thisdifferential labeling of the Kenyon cells most likely[Fig. 3(c,d)] , which are the areas of synaptic con-



Figure 3 Immunostaining in the visual lobes and antennal lobes in Apis and Drosophila usingantibodies against RII subunits. While in Apis (a) and Drosophila (b) the retina (r) is devoidof staining, visual neuropiles such as the medulla (me), lobula (lo) , and lobula plate (lp) ( inDrosophila only) show intermediate immunostaining. Staining of the lamina (la) is barelydetectable in Drosophila. The columnar organization and the different tangential layers of theneuropiles show distinct levels of staining. Somata surrounding the visual neuropiles are labeled,although at distinct levels. Immunolabeling with RII subunit antibodies is observed in theantennal lobes of Apis (c) and Drosophila (d) . In Apis, the staining in the antennal lobes isconcentrated in the central area of the glomeruli (g) , which exhibit an individual, inhomoge-neous degree of labeling. The somata of antennal-lobe interneurons (arrow) show weak staining.In Drosophila, the labeling within the antennal lobe is also located in the glomeruli, but ismore diffuse than in Apis. The very strong labeling of a cluster of somata located at positionsof antennal lobe interneurons (arrows) is remarkable. All scales Å 50 mM.

accounts for the characteristic stratified staining pat- duncle most likely presents fibers which project intothe more weakly stained g-lobes [Fig. 4(e)].tern in the peduncle and the a-lobes of the mush-

room body neuropil [Fig. 4(c)] . In addition to the mushroom bodies, structuressuch as the central complex and distinct neuropilIn Drosophila, neither the neuropil of the mush-

room body nor the somata of the Kenyon cells are areas of the protocerebrum are stained to varyingdegrees (Fig. 4) . In both species, all substructureslabeled at the same intensity [Fig. 4(d–g)] . Sub-

groups of Kenyon cell somata located in distinct of the central complex such as the central body,noduli, and protocerebral bridge show a characteris-clusters stain more intensely than surrounding so-

mata [Fig. 4(f )] . Within the calycal neuropil, more tic pattern of immunoreactivity [Fig. 4(a,d)] .intense labeled regions in lateral areas are notice-able. While the calyces, peduncle, and a-1 and b- Stimulation of PKA in Cultured Kenyonlobes exhibit generally strong labeling, the g-lobes Cells from Honeybeesare less intensively labeled [Fig. 4(g)] . Since theintrinsic fibers project via the peduncle into the Since several lines of evidence indicate that the

mushroom bodies are sites mediating learning andlobes, the area of low staining intensity in the pe-


40 Muller

Figure 4 Immunolabeling in the central brain and mushroom bodies of Apis (a–c) andDrosophila (d–g) using antibodies against RII subunits. In both Apis and Drosophila, themushroom body neuropiles such as the calyces (ca) , peduncle (p) , and lobes (a, b, g) exhibitstrong but inhomogeneous labeling. Distinct somata of the mushroom intrinsic Kenyon cells(kc) show differences in their staining intensity [arrows in (b) and (f )] . In Apis, the peduncleand a-lobes exhibit a characteristic stratified staining pattern (a,c) . In Drosophila (g) , the g-lobes are less intensively labeled, while the a- and the b-lobes exhibit strong labeling. A crosssection of the peduncle of Drosophila shows an area with weaker labeling of fibers [(e) ,arrow]. Structures such as the central body (cb) and distinct neuropil areas of the protocerebrumare stained to varying degrees. All scales Å 50 mM.

memory via the cAMP cascade (Davis, 1993; Davis role in learning and memory are capable of modulat-ing the cAMP cascade, and thus PKA activity inet al., 1995), the high concentration of PKA II in

the mushroom body–intrinsic Kenyon cells is of intact Kenyon cells. For this purpose, a techniquewas applied which allowed the application of neuro-special interest. Thus, it is tempting to address the

question of whether neurotransmitters which play a transmitters on intact cells and the subsequent deter-



these cyclic nucleotides directly activate PKA (Alt-felder and Muller, 1991), the stimulatory effect canalso be observed in homogenates of Kenyon cells(data not shown). Forskolin, which directly stimu-lates the adenylyl cyclase, increases PKA activityin intact cells twofold [Fig. 5(a)] . The similar ap-plication of forskolin (30 s) to homogenated cellscauses no significant elevation in PKA activity.

Treatment of 3-day cultured cells with depolariz-ing agents such as KCl (30 mM) and Ca2/ iono-phore A21187 (5 mM) , as well as acetylcholine (20mM) , which also increases intracellular Ca2/ levelsin honeybee Kenyon cells (Bicker and Kreissl,1994), have no effect on PKA activity [Fig. 5(b)] .The latter is of special interest, since histochemicalstaining suggests a potential cholinergic transmis-sion into the mushroom bodies via a subset of ante-nnoglomerular fibers (Kreissl and Bicker, 1989).The application of inhibitory g-amino-h-butyricacid (GABA) (20 mM) , which evokes a chloridechannel in Kenyon cells from honeybees (Rosen-

Figure 5 Modulation of PKA activity in cultured Ken-boom et al., 1994), is ineffective in changing theyon cells of Apis mellifera. Changes in PKA activityPKA activity in intact cells.evoked by neurotransmitters in situ were tested by

The monoamines dopamine and serotonin (eachapplying drugs to Kenyon cells cultured for 3 days. After5–20 mM) significantly elevate PKA activity in30 s stimulation, the cells were shock-frozen in liquidKenyon cells [Fig. 5(b)] . Octopamine, the putativenitrogen and changes in PKA activity were measured

using a fast in vitro phosphorylation assay. (a) The cells transmitter of the unconditioned stimulus in associa-were stimulated with Br-cAMP, Br-cGMP (5 mM) , or tive olfactory learning in honeybees (Hammer andforskolin (20 mM) . (b) Agents such as KCl (30 mM) , Menzel, 1994, 1995), causes an increase of PKAA21187 (5 mM) , and acetylcholine (20 mM) , which in- activity in Kenyon cells which varies with the con-crease intracellular Ca2/ levels as well as GABA (20 centrations used. While low concentrations of octo-mM) , have no effect on PKA activity. The monoamines pamine (5 mM) induce a strong elevation of PKAdopamine, serotonin, and octopamine significantly ele-

activity, concentrations of 20 mM octopamine leadvate PKA activity in Kenyon cells. Each column showsto a weaker stimulation, with highly variable resultsmeans { S.D. of at least eight independent experiments.[Fig. 5(b)] .(*p õ 0.01, t test) .

Since in all cases the exposure of neurotransmit-ters to homogenized cells failed to induce changein PKA activity, the observed effects of neurotrans-mination of changes in PKA activity by a fast in

vitro phosphorylation assay (Hildebrandt and mitters on PKA seem to require the intact signaltransduction cascade of viable Kenyon cells.Muller, 1995(a,b)] .

Immediately after the dissociation of the Kenyoncells from honeybees (É170,000 cells /mushroombody), only the administration of Br-cAMP (5 mM) DISCUSSIONor Br-cGMP (5 mM) caused an increase in PKAactivity; forskolin (20 mM) and other neurotrans- Although in Drosophila the cAMP cascade has been

implicated in learning and memory, until now theremitters showed no significant effect. One day afterdissociation, forskolin elevated PKA activity in the has been no immunohistological localization of the

type II PKA which is the major target of cAMP incells, while the effects of neurotransmitters on PKAactivity were observable only after 2–3 days, when adult Drosophila and Apis. Thus far, biochemical

studies have failed to identify type I PKA in theoutgrown neurites were clearly visible.In Kenyon cells cultured for 3 days, application nervous tissue of adult Drosophila and Apis (Foster

et al., 1984; Muller and Spatz, 1989; Altfelder andof the membrane permeable Br-cAMP (5 mM) orBr-GMP (5 mM) for 30 s causes a threefold increase Muller, 1991). Since the existence of transcripts of

type I regulatory subunits (RI) have been reportedin PKA activity in the intact cells [Fig. 5(a)] . Since


42 Muller

in adult Drosophila (Kalderon and Rubin, 1998), which effects the cAMP cascade, eliminates asso-ciative learning (Connelly et al., 1996).however, and since only about 90% of the total

neuronal PKA activity is immunoprecipitated by Because of the major role of the RII subunit intargeting and inactivating the C subunit, comparisonantibodies against the RII subunit, we have to as-

sume a small and probably barely detectable frac- of the localization of the Drosophila catalytic sub-unit gene (DCO) (Skoulakis et al., 1993) with thattion of RI in adult animals.

As in mammals, the distinct R subunits, in con- of the RII subunit is of most interest. Although theDCO-encoded C subunit seems to be mainly colo-trast to C subunits, show a high variability in bio-

chemical properties and their amino acid sequences calized with the RII subunit, there are some charac-teristic differences within the mushroom bodies of(Kalderon and Rubin, 1988; Taylor et al., 1990).

While the C subunits from Drosophila and Apis are Drosophila. Unlike the C subunit, immunostainingwith RII antisera exhibit different labeling of dis-very similar to the mammalian enzyme (Foster et

al., 1984, 1988; Muller and Spatz, 1989; Altfelder tinct subsets of Kenyon cell somata, inhomogeneousstaining within the calycal neuropile, a weaker la-and Muller, 1991), the RIIs exhibit different proper-

ties, even between both insect species. The poly- beling of the g-lobes, and probably weaker stainingof axons which project into the g-lobes. Althoughclonal antisera generated against the Apis RII sub-

unit show only very weak cross reactivity to the other explanations are possible, the differences maybe due to the existence of C and R subunits differentDrosophila RII subunit, and vice versa. Also, the

RII subunit from Apis and Drosophila differ in their from the DCO and RII subunit. Feasible candidatesare isoforms of the C and the RI subunits (Fostermolecular weight and electrophoretic mobilities of

the phospho and dephosphoforms (Fig. 1) (Muller et al., 1988; Kalderon and Rubin, 1988). Indepen-dent of the exact colocalization of distinct R and Cand Spatz, 1989; Altfelder and Muller, 1991).

The R subunits, which exist in soluble and mem- subunits, the inhomogeneous staining of the Kenyoncells and mushroom body neuropiles in both speciesbrane-associated forms, are thought to serve as in-

struments to target the catalytic action to specific supports the idea that distinct Kenyon cell subpopu-lations express different levels of RII subunits. Thiscellular destinations (Taylor et al., 1990). In both

insect species, a high ratio of about 35–45% of RII observation is in line with recent findings that Dro-sophila mushroom bodies, and thus Kenyon cellsubunit was located in membrane-associated frac-

tions. Although the exact anchor proteins of the RII subgroups, can be subdivided using enhancer-trapexpression patterns (O’Dell et al., 1995; Yang etsubunit are unknown, it has been demonstrated that

in Drosophila the microtubular proteins responsible al., 1995).While these findings support an important rolefor binding of RII subunit are different from those

in vertebrates (Adam and Friedrich, 1988). of the cAMP-mediated signaling pathway in mush-room bodies, there is increasing evidence that bio-In general, the overall similarities in distribution

and localization of the RII subunit in the brain of genic amines are essential for initiating these bio-chemical cascades. The colocalization of a Dro-Apis and Drosophila are remarkable. The visual

lobe, with its intermediate levels of RII subunits sophila dopamine D1 receptor (DAMB) with therutabaga-encoded adenylyl cyclase, together withand PKA activity, shows immunostaining in the

second- and third-order neuropiles and most somata its ability to mediate dopamine-induced cAMP for-mation, makes DAMB a suitable initiator of the(Fig. 3) . In the antennal lobes of both species, the

labeling is concentrated on the glomeruli and so- cAMP signaling implicated in learning and memory(Han et al., 1996).mata of antennal lobe interneurons. Although char-

acteristic staining was also observed in other areas In both Drosophila and Apis pharmacologic evi-dence indicates an implication of octopamine in learn-of the brain, the high level of immunolabeling in

the mushroom bodies, which have been suggested ing (Mercer and Menzel, 1982; Dudai et al., 1987;Bicker and Menzel, 1989). In Apis, the unconditionedto be sites of olfactory learning (Connolly et al.,

1996; deBelle and Heisenberg, 1994; Menzel et al., stimulus in olfactory learning (sucrose) is mediatedby the VUMmx1 neuron (Hammer, 1993), which is1994), is most remarkable (Fig. 4) . Interestingly,

like the RII subunit, three genes which code for presumably octopaminergic (Kreissl et al., 1994). Therole of octopamine as the transmitter of the uncondi-distinct components of the cAMP cascade and cause

severe defects in olfactory learning in Drosophila tioned pathway is supported by the finding that localinjections of octopamine into the mushroom body ca-are also preferentially expressed in the mushroom

bodies (Davis, 1993). Moreover, in the mushroom lyces of the honeybee substitute for the unconditionedstimulus in olfactory conditioning (Hammer and Men-bodies of Drosophila an impaired Ga signaling,



(1996). Associative learning disrupted by impaired Gszel, 1994). Interestingly, in Kenyon cells, which aresignaling in Drosophila mushroom bodies. Sciencethe potential postsynaptic site of conditioned and un-274:2104–2107.conditioned pathways, biogenic amines are capable of

DAVIS, R. L. (1993). Mushroom bodies and Drosophilachanging the PKA activity. Acetylcholine, the sup-learning. Neuron 11:1–14.posed transmitter of the conditioned pathway in hon-

DAVIS, R. L., CHERRY, J., DAUWALDER, B., HAN, P. L.,eybees (Kreissl and Bicker, 1989), and other agentsand SKOULAKIS, E. (1995). The cyclic AMP system

which increase intracellular Ca2/ levels, have no effect and Drosophila learning. Mol. Cell. Biochem. 149–on PKA activity (Fig. 5). 150:271–278.

Compared to serotonin and dopamine, octopamine DEBELLE, J. S. and HEISENBERG, M. (1994). Associativeshows a conspicuous concentration-dependent effect odor learning in Drosophila abolished by chemical ab-on PKA activity in Kenyon cells. This is probably lation of mushroom bodies. Science 263:692–695.due to different subtypes of octopamine receptors with DEZAZZO, J. and TULLY, T. (1995). Dissection of mem-

ory formation: from behavioral pharmacology to mo-distinct coupling to different second-messenger path-lecular genetics. Trends Neurosci. 18:212–218.ways (Evans and Robb, 1993; Robb et al., 1994) or

DRAIN, P., FOLKERS, E., and QUINN, W. G. (1991).the coexpression of PKAII with different subtypes ofcAMP-dependent protein kinase and the disruption ofoctopamine receptors in distinct Kenyon cell subpopu-learning in transgenic flies. Neuron 6:71–82.lations. Although these experiments demonstrate

DUDAI, Y., BUXBAUM, J., CORFAS, G., and OFARIM, M.transmitter-evoked changes of the PKA activity in(1987). Formamidines interact with Drosophila octo-cultured Kenyon cells in situ, future studies will bepamine receptors, alter the flies’ behavior and reduce

needed to reveal the modulatory function of the trans- their learning ability. J. Comp. Physiol. 161:739–746.mitters on the cAMP cascade in the Kenyon cells EVANS, P. D. and ROBB, S. (1993). Octopamine receptorduring associative olfactory learning. Irrespective of subtypes and their modes of action. Neurochem. Res.the techniques used to unravel the role of the Kenyon 18:869–874.cells in olfactory learning and memory in Apis and FOSTER, J. L., GUTTMAN, J. J., HALL, L. M., and ROSEN,Drosophila, the studies need to take into account the O. M. (1984). Drosophila cAMP-dependent protein

kinase. J. Biol. Chem. 259:13049–13055.indicated differences between Kenyon cell subpopu-FOSTER, J. J., HIGGINS, G. C., and JACKSON, F. R. (1988).lations.

Cloning, sequence, and expression of the DrosophilacAMP-dependent protein kinase catalytic subunit gene.The financial support of the Deutsche Forschungsgem-J. Biol. Chem. 263:1676–1681.einschaft (Grant SFB 515/C3), and support by a Leibniz

FRANK, D. A. and GREENBERG, M. E. (1994). CREB: aaward to R.M. are gratefully acknowledged. The authormediator of long-term memory from mollusks to mam-thanks R. Menzel and S. Meuser for critical reading ofmals. Cell 79:5–8.the manuscript.

HAMMER, M. (1993). An identified neuron mediates theunconditioned stimulus in associative olfactory learn-ing in honeybees. Nature 366:59–63.

REFERENCES HAMMER, M. and MENZEL, R. (1994). Octopamine localinjections into the mushroom body calyces and theantennal lobe substitute for the unconditioned stimulusADAM, G. and FRIEDRICH, P. (1988). Microtubule associ-(US) in the honeybee olfactory conditioning. Soc. Neu-ated cyclic AMP-dependent protein kinase in Drosoph-rosci. Abstr. 20:582.ila melanogaster. J. Neurochem. 51:1014–1022.

HAMMER, M. and MENZEL, R. (1995). Learning andALTFELDER, K. and MULLER, U. (1991). Cyclic nucleo-memory in the honeybee. J. Neurosci. 15:1617–1630.tide-dependent protein kinases in the neural tissue of

HAN, K. A., MILLAR, N. S., GROTEWIEL, M. S., andthe honeybee Apis mellifera. Insect Biochem. 21:487–DAVIS, R. L. (1996). DAMB, a novel dopamine recep-494.tor expressed specifically in Drosophila mushroomBICKER, G. and KREISSL, S. (1994). Calcium imagingbodies. Neuron 16:1127–1135.reveals nicotinic acetylcholine receptors on cultured

HILDEBRANDT, H. and MULLER, U. (1995a). Octopaminemushroom body neurons. J Neurophysiol. 71:808–mediates rapid stimulation of PKA in the antennal lobe810.of honeybees. J. Neurobiol. 27:44–50.BICKER, G. and MENZEL, R. (1989). Chemical codes for

HILDEBRANDT, H. and MULLER, U. (1995b). PKA activ-the control of behavior in arthropods. Nature 337:33–ity in the antennal lobe of honeybees is regulated by39.chemosensory stimulation in vivo. Brain Res.BYRNE, J. H. (1987). Cellular analysis of associative679:281–288.learning. Physiol. Rev. 67:329–439.

HUANG, F. L. and GLINSMANN, W. H. (1976). SeparationCONNOLLY, J. B., ROBERTS, I. J., ARMSTRONG, D., KAI-

SER, K., FORTE, M., TULLY, T., and O’KANE, C. J. and characterization of two phosphorylase phosphatase


44 Muller

inhibitors from rabbit skeletal muscle. Eur. J. Biochem. tein kinase in Drosophila wild-type and dunce memorymutants. J. Neurogenet. 6:95–114.70:419–426.

O’DELL, K. M. C., ARMSTRONG, J. D., YANG, M. Y., andKALDERON, D. and RUBIN, G. M. (1988). Isolation andKAISER, K. (1995). Functional dissection of the Dro-characterisation of Drosophila cAMP-dependent pro-sophila mushroom bodies by selective feminization oftein kinase genes. Genes Dev. 2:1539–1556.genetically defined subcompartments. Neuron 15:55–KANDEL, E. R. and SCHWARTZ, J. H. (1982). Molecular61.biology of learning: modulation of transmitter release.

ROBB, S., CHEEK, T. R., HANNAN, F. L., HALL, L. M.,Science 218:433–443.MIDGLEY, J. M., and EVANS, P. D. (1994). Agonist-

KREISSL, S. and BICKER, G. (1989). Histochemistry of specific coupling of a cloned Drosophila octopamine/acetylcholinesterase and immunocytochemistry of an tyramine receptor to multiple second messenger sys-acetylcholine receptor-like antigen in the brain of the tems. EMBO J. 13:1325–1330.honeybee. J. Comp. Neurol. 286:71–84. ROSENBOOM, H., GOLDBERG, F., SCHAFER, S., and MEN-

KREISSL, S. and BICKER, G. (1992). Dissociated neurons ZEL, R. (1994). Ionic currents of Kenyon cells possiblyof the pupal honeybee brain in cell culture. J. Neurocy- involved in olfactory learning in insects. Soc. Neurosci.tol. 21:545–556. Abstr. 20:803.

SKOULAKIS, E. M. C., KALDERON, D., and DAVIS, R. L.KREISSL, S., EICHMULLER, S., BICKER, G., RAPUS, J., and(1993). Preferential expression in mushroom bodiesECKERT, M. (1994). Octopamine-like immunoreactiv-of the catalytic subunit of protein kinase A and its roleity in the brain and suboesophageal ganglion of thein learning and memory. Neuron 11:197–208.honeybee. J. Comp. Neurol. 348:583–595.

STOCKER, R. F. (1994). The organization of the chemo-MENZEL, R., DURST, C., ERBER, J., EICHMULLER, S., HAM-sensory system in Drosophila melanogaster: a review.MER, M., HILDEBRANDT, H., MAUELSHAGEN, J.,Cell Tissue Res. 275:3–26.MULLER, U., ROSENBOOM, H., RYBAK, J., SCHAFER, S.,

TAYLOR, S. S., BUECHLER, J. A., and YONEMOTO, W.and SCHEIDLER, A. (1994). The mushroom bodies in(1990). cAMP-dependent protein kinase: frameworkthe honeybee: from molecules to behaviour. In: Neuralfor a diverse family of regulatory enzymes. Annu. Rev.

Basis of Behavioural Adaptations, K. Schildberger andBiochem. 59:971–1005.

N. Elsner, Eds. Fischer Verlag, Stuttgart, pp. 81–102. WALAAS, S. I. and GREENGARD, P. (1991). Protein phos-MERCER, A. R. and MENZEL, R. (1982). The effects of phorylation and neuronal function. Pharmacol. Rev.

biogenic amines on conditioned and unconditioned re- 43:299–349.sponses to olfactory stimuli in the honeybee (Apis mel- YANG, M. Y., ARMSTRONG, J. D., VILINSKY, I., STRAUS-lifera) . J. Comp. Physiol. 145:363–368. FELD, N. J., and KAISER, K. (1995). Subdivision of the

MULLER, U. and SPATZ, H.-C. (1989). Ca2/-dependent Drosophila mushroom bodies by enhancer-trap expres-sion patterns. Neuron 15:45–54.proteolytic modification of the cAMP-dependent pro-


Neuronal cAMP-dependent protein kinase type II is concentrated in mushroom bodies of Drosophila...

Documents

Transcript of Neuronal cAMP-dependent protein kinase type II is concentrated in mushroom bodies of Drosophila...