Distinct patterns of cAMP-dependent protein kinase gene expression in mouse brain

Neuron,

Vol. 3, 71-79,

]uly, 1989, Copyright

0 1989 by Cell Press

Distinct Patterns of CAMP-Dependent Gene Expression in Mouse Brain Gary Cadd and G. Stanley McKnight Department of Pharmacology University of Washington Seattle, Washington 98195

Summary In situ hybridization was used to localize CAMPdependent protein kinase (PKA) mRNAs in the adult mouse CNS. The PKA holoenzyme contains two catalytic (C) subunits and a regulatory (R) subunit dimer. Our studies demonstrate expression of two isoforms of C (Ca and Cb) and four isoforms of R (Rla, RIP, Rlla, and Rllfi) in the CNS. mRNAs for Ca, Rla, and RIP preferentially localize in the neocortex, caudate-putamen, hypothalamus, thalamus, and hippocampus. Hybridization with Cg and Rllg probes is clearly distinguished from the Ca-like pattern by a reduced level of hybridization in the thalamus and by a relative increase in expression in the dentate gyrus compared with cell layers CAl-3 in the hippocampus. Rlla transcripts are very specifically localized in the medial habenula. The differential expression of PKA subunit genes suggests that functional differences in CAMP responses within neural tissues may be mediated by the biochemical properties of specific PKA isoforms. Introduction Biogenic amines and neuropeptides interact with cell surface receptors in the CNS to produce metabolic effects that typically involve the activation of specific intracellular second messenger systems. Among the best characterized second messenger systems is the pathway that involves the generation of CAMP and activation of CAMP-dependent protein kinase (PKA). Protein phosphorylation by PKA has been implicated in a variety of cellular events in the brain, including metabolism, gene transcription, ion channel regulation, and neurotransmitter synthesis (reviewed by Krebs and Beavo, 1979; Nairn et al., 1985). PKA is enriched in synaptic structures, suggesting that it plays an important role in synaptic function. In addition, studies of Aplysia and Drosophila have suggested an involvement of the CAMP second messenger pathway in long-term changes in synaptic plasticity that ultimately alter the behavior of the organism (reviewed by Kandel and Schwartz, 1982; Dudai, 1988). The holoenzyme of PKA is an inactive tetramer composed of a regulatory (R) subunit dimer with a catalytic (C) subunit bound to each R subunit. Biological agents that stimulate adenylate cyclase and cause an increase in CAMP production lead to activation of PKA. When two molecules of CAMP have bound to each R subunit, PKA dissociates, releasing active C subunits. PKA activation is reversed by the hydrolysis of CAMP to 5’ AMP by

Protein

Kinase

phosphodiesterases and the subsequent reassociation of the R and C subunits. The phosphorylation events catalyzed by PKA can be reversed by cellular phosphatases. Two major types of PKA have been identified by DEAE ion exchange chromatography (Reimann et al., 1971) and named type I and type II according to their elution order. The two types of holoenzyme differ in the structure of the R subunit incorporated (RI or RII), but the C subunits are either identical or very similar. Recent work has shown that there are multiple forms of RI, RII, and C, which have distinct tissue distributions. Two clones from mouse genomic and cDNA libraries, designated Rla and RIB, have been shown to encode distinct forms of RI (Clegg et al., 1988). Rla appears to be expressed in all tissues, whereas RIB expression has been reported only in brain and in developing sperm cells. Two forms of RII have been distinguished as products of separate genes, and we refer to these as Rlla and RIIB (Scott et al., 1987; Jahnsen et al., 1986). Rlla, first identified in heart, is now known to be present in many tissues and is highly expressed in male germ cells. Originally thought to be a neural-specific form of RII, RIIB is present in reproductive tissues as well and is induced by the combined action of estrogen and follicle-stimulating hormone in rat granulosa cells (Hedin et al., 1987). In bovine and in mouse tissues, two distinct genes for the C subunit have been identified, termed Ca and CB (Uhler et al., 1986a, 1986b; Showers and Maurer, 1986). Northern blot analysis has shown that Ca is expressed in all tissues examined; CB is expressed primarily in the brain, but is also detectable in many other tissues. A third isoform of C (Cy) has recently been observed in rat testis (T. jahnsen, personal communication). Thus in each case the a forms of the PKA subunits are present in virtually all tissues, whereas the B forms show a restricted tissuedistribution, primarily in brain. In an effort to understand the function of these multiple forms, we first localized mRNA transcripts of PKA subunits in mouse brain using in situ hybridization.

Results Distribution of PKA lsoforms and Nonneuronal Tissues

in Neuronal

As a preliminary step, we examined the overall distribution of PKA subunit mRNAs in a 14 day mouse fetus using the complementary j5S-labeled RNA probes described in Figure 1. mRNAs for all Rand C subunits were found in high amounts in the developing brain and spinal cord (Figure 2). In general, the a subunit genes were expressed ubiquitously throughout the fetus; the B subunit genes showed a more tissue-restricted expression. The mRNAs encoding Ca, Rla, and Rlla were prominent in brain, heart, and liver, whereas CB and RIB mRNAs were restricted primarily to the brain and spinal cord (Figure 2). RIIB mRNA was detected in brain and spinal cord as well as in the developing liver (Figure 2F).

Neuron 72

Figure 1. Map of PKA cDNA Construct RNA Probes

Clones

Used to

Regions from which complementary RNA probes for PKA subunits were synthesized are indicated by filled bars. Complete open reading frames are defined by start and stop codons. Restriction endonuclease sites used in construction of the probes are indicated: 8, E&Ill; E, EcoRI; SC, Seal; 5, Sall; H, Hindlll.

a forms

0 forms

Figure 2. Expression of PKA Subunits Neuronal and Nonneuronal Tissues of Day Mouse Fetus

in 14

Sagittal sections of mouse fetus were hybridized with %labeled probes complementary to Ca (A), CB (D), Rla (B), RIB (El, Rlla (C), and RllB (F) mRNA, as detailed in Experimental Procedures, and apposed to film for 13 days. mRNAs encoding a forms of PKA were observed in most tissues, including brain, heart, and liver (A-C). Expression of the B forms, however, was confined primarily to brain and spinal cord (D-F). Fetal liver also showed large amounts of RIIB mRNA (F). Abbreviations: 8, brain; S, spinal cord; H, heart; L, liver.

a

a

PKA Gene 73

Expression

in Mouse

Brain

Ca

ca

Controls

Transferrin Figure

3. Ca and CB Describe

Two General

Categories

of PKA Expression

in Mouse

Brain;

Transferrin

Expression

Defines

Oligodendrocytes

Frontal sections of adult mouse brain were hybridized with 3sS-labeled probes complementary to Ca (A-C), CB (D-F), and transferrin (I) %labeled probe with a SOB-fold molar mRNA as described. Specificity of hybridization was demonstrated by incubation of the respective excess of unlabeled probe for Ca (C) and CB (Ht. Exposure times: 6 days (A-C); 8 days (D-H); 5 days (I). Abbreviations: cc, corpus callosum; ci, internal capsule; CP caudate-putamen area; G, granular layer, cerebellum; Hm, medial habenular nucleus; Hp, hippocampus; LC, locus coeruleus; N, neocortex; 0, subfornical organ; ot, optic tract; Pr, piriform cortex: Pv paraventricular thalamic nucleus,

Studies of later developmental stages have shown that RllB mRNA in liver decreases to near background levels by birth (data not shown). These in situ results extend previous Northern blot studies of PKA expression (Uhler et al., 1986a, 198613; Jahnsen et al., 1986; Showers and Maurer, 1986; Scott et al., 1987; Clegg et al., 1988) and serve as positive and negative tissue controls for brain sections. Ca and Cp mRNA Localization Typifies Two General Patterns of PKA Expression in Brain In situ hybridization to adult mouse brain revealed distinctive patterns of PKA isoform expression. The expression of PKA genes formed two general patterns in mouse brain, within which individual subunits showed specific differences. The two general categories are typified by the patterns of Ca and CB expression. mRNA for both Ca and CB was observed in large amounts in the neocortex, piriform cortex, hypothalamus, medial habenular nuclei, and reticular thalamic nuclei (Figures 38 and 3E). Ca mRNA, but not CB mRNA, was concentrated in thalamic nuclei. The Ca pattern of expression also varied from the CB pattern in the hippocampus. mRNA encoding Ca was equally distributed between the dentate gy-

rus and pyramidal cell layers CAl-3 (Figure 3B). In contrast, although CB expression was high in the dentate gyrus, little CB mRNA was observed in the CAl-3 cell layers (Figure 3E). In more rostra1 sections, Ca and CB mRNAs were also observed in the subfornical organ, the paraventricular thalamic nuclei, and the bed nucleus of the stria terminalis (Figures 3A and 3D). Caudally, Ca and CB mRNAs colocalized in the locus coeruleus and other hindbrain structures (Figures 3C and 3F). Within the Ca and CB were obcerebellum, mRNAs encoding served in the granular layer, but little hybridization was observed in the white matter of the folia (Figures 3C and 3F). Binding specificity was examined by competition with a SOO-fold molar excess of unlabeled probe (Figures 3G and 3H). The unlabeled probe eliminated specific binding of Ca and CB, leaving only a background signal. Competition between the labeled Ca probe and unlabeled CB probe was not observed in parallel experiments, indicating that these conditions allowed specific hybridization (data not shown). Areas of white matter, such as the corpus callosum, the internal capsule, and the optic tract showed much

Figure 4. Differential Expression of Ca and CB in the Hippocampus and Cerebellum Dark-field photomicrographs ofCa (A and B) and CB (C and D) expression were made from emulsion-coated sections exposed for 15 days as described. Ca expression was observed in both the dentate gyrus and the CA3 cell layer in the hippocampus (A). CB expression, however, was confined primarily to the dentate gyrus (0. Differential expression was also observed in the cerebellum. Both C subunit mRNAs were found in the granular layer (B and D). Purkinje cells, however, expressed large amounts of Ca (B), but not of CB (D). Abbreviations: CA3, pyramidal cell layer CA3, hippocampus; DC, granule cell layer of the dentate gyrus, hippocampus; G, granular layer, cerebellum; M, molecular layer, cerebellum; P Purkinjecell layer, cerebellum.

less hybridization to PKA probes compared with neuronal layers. We reasoned that the low amount of PKA signal was due to the low abundance of neuronal perikarya and the relatively high abundance of oligodendrocyte perikarya. As a control, we localized oligodendrocyte perikarya using a probe for transferrin mRNA (Bloch et al., 1985). We observed intense hybridization of the transferrin probe in the areas of low PKA expression as shown in Figure 31, confirming that these areas are rich in oligodendrocytes. Microscopic examination of hybridized sections confirmed specific quantitative differences between the expression patterns of Ca and Cg mRNAs in the hippocampus. Figure 4C illustrates that we observed relatively less hybridization to C8 probes in the pyramidal cell layer CA3 compared with the level of expression in the granule cell layer of the dentate gyrus. mRNA en-

coding Ca, however, was readily as shown in Figure 4A. Another was observed in the cerebellum.

apparent in both areas, interesting difference Although both C

subunits were expressed in the granular and molecular layers, Purkinje cells preferentially expressed Ca, compared with CD, mRNA, as depicted in Figures 48 and 4D.

Rla and RIO Expression Pattern of Expression

Is Similar

to the Ca

Rla and RI8 mRNAs colocalized in most brain sections observed, differing primarily in the relative amounts of mRNA present. As shown in Figure 5, both transcripts were detected in large amounts in cells of the neocortex, piriform cortex, and medial habenular nuclei. Both RI mRNAs were observed in the reticular thalamic nuclei and dorsomedial hypothalamic nuclei. RI mRNA was also observed in various other thalamic and hypotha-

PKA Gene

Expression

in Mouse

Brain

75

Figure 5. Patterns pression Resemble

of Rla and RIB Gene That of Ca Expression

Ex-

Frontal sections of mouse brain were hybridized with probes complementary to either Rla (A-C) or RIB (D-F) mRNA as described in Experimental Procedures. Dark-field photomicrographs of Rla (C) and RIB (F) expression in the hippocampus were made from emulsion-coated sections as described. Abbreviations: CA3, pyramidal cell layer CA3, hippocampus; DC, granule cell layer of the dentate gyrus, hippocampus; C, granular layer, cerebellum; Hm. medial habenular nucleus; Hp, hippocampus; LC, locus coeruleus; N, neocortex; Pr, piriform cortex. Exposure times: 6 days (A, B,D, and E): 15 days (C and F); 4 days (El.

lamic nuclei and to a lesser extent in the caudate-putamen area. The hybridization pattern of RIB observed in the hippocampus was unlike that of other PKA genes. Rla, Ca, and Rlla all were expressed at comparable levels in the CAl-3 pyramidal cell layers and the granule cell layer of the dentate gyrus (Figure SC). RIB hybridized preferentially in CAl-3 cells layers compared with the dentate gyrus (Figure SF), a pattern opposite that of the other PKA B forms. In more rostra1 sections, a large amount of Rla and RIB mRNA was observed in the subfornical organ, paraventricular thalamic nuclei, and suprachiasmatic nuclei (data not shown). Caudally, these mRNAs were also observed in the locus coeruleus (Figures 5B and 5E). Within the cerebellum, Rla and RIB mRNA was detected in Purkinje cells and the granular and molecular layers (data not shown). Myelinated fiber tracts such as the corpus callosum, the internal capsule, and the optic tract showed little Rla and RIB hybridization compared with neuronal layers.

Rlla and RI@ Are Discreetly

Expressed

in Brain

In contrast to the type I R subunits, the type II R subunits were expressed in very discrete areas in mouse brain. A5 shown in Figure 6, transcripts for Rlla were found almost exclusively in the medial habenular nuclei (Figure 6A). Much lower levels of Rlla mRNA were also detected, after long exposures, in the neocortex, hippocampus, piriform cortex, reticular thalamic nuclei, and some hypothalamic areas. mRNAfor RIIB was present in the medial

habenular nuclei, but high levels were also seen in the neocortex, caudate-putamen area, piriform cortex, reticular thalamic nuclei, and hippocampus (Figure 6D). Sagittal sections also showed the intense localization of Rlla mRNA in the medial habenula (Figure 6B). Much lower levels of Rlla mRNA were observed in other parts of the fore-, mid-, and hindbrain. The RI@ probe, in contrast, hybridized strongly in the caudate-putamen, piriform cortex, and supraoptic nuclei, as well as in the locus coeruleus, as shown in Figure 6E. The B form of RII was observed primarily in the dentate gyrus of the hippocampus and, to a lessor extent, in the CAl-3 cell layers, a pattern similar to that observed for CB. Although present in low amounts, Rlla mRNA conformed to the a pattern of expression, present in approximately equal amounts in both the pyramidal cell layer CA3 and the granule cells of the dentate gyrus (Figure 6C). Purkinje cells and the molecular layer of the cerebellum showed the presence of both Rlla and RIIB transcripts, whereas less RI@ than Rlla mRNA was observed within the granular layer (data not shown). As in the case of the other PKA mRNAs, little hybridization of Rlla and RIIB probes was observed in areas through which myelinated fiber tracts pass.

Discussion In situ hybridization with specific probes for the R and C subunits of PKA demonstrated that both isoforms of the C subunit (Ca and C(3) and all four types of R subunit

Figure 6. Rlla and RllB Show creet Patterns of Expression

Highly

Dis-

Frontal sections (A, C, D, and F) and sagittal sections (B and E) of mouse brain were hybridized with probes complementary to either Rlla (A-C) or RllB (D-F) mRNA as described in Experimental Procedures. Dark-field photomicrographs of Rlla (C) and RI@ (F) expression showed distinctive patterns in the hippocampus. Abbreviations: cerebellum; CA3, pyramidal cell layer CA3, hippocampus; DC, granule cell layer of the dentate gyrus, hippocampus; Hm, medial habenular nucleus; Hp, hippocampus; LC, locus coeruleus; N, neocortex: Pr, piriform cortex; SO, suraoptlc nucleus. Exposure times: 10 days (A, 6, D, and E); 15 days iC and F).

(Rla, RIB, Rlla, and RI@) areexpressed within the mouse CNS. Table 1 summarizes the data from these and other studies. In the brain, the highest levels of expression localize to neuronal cell layers, although PKA subunits are certainly expressed in nonneuronal ceils as well. A recent comparison of astrocytes, oligodendrocytes, and neuronal perikarya isolated from rat neocortex showed similar levels of CAMP binding activity in all three preparations (Stein et al., 1987). We used the pattern of transferrin gene expression to identify areas of myelinproducing oligodendrocytes (Bloch et al., 1985). These regions of white matter within the brain contain relatively few neuronal perikarya. Most PKA subunit probes produce a pattern of hybridization that is opposite that seen with the transferrin probe, further supporting our conclusion that PKA expression is preferentially neuronal. The expression of PKA in adult mouse brain can be divided into two general categories based upon the hybridization patterns of probes to Ca and CB. Within these general categories, however, the patterns of expression of each subunit show individual differences. The Ca pattern of expression, shared by Rla and RIB, shows preferential hybridization in the neocortex, caudate-putamen, hypothalamus, thalamus, and hippocampus. The CB pattern, shared by RIIB, is distinguished from the Ca pattern by a reduced level of hybridization in the thalamus and the pyramidal cell layers CAl-3 of the hippocampus. Rlla is highly expressed only in the medial habenula. With longer exposure times, autoradi-

ograms of Rlla hybridization expression, albeit at low

levels.

resemble the Ca pattern Although these data

of do

not necessarily reflect protein levels in these regions, they suggest that isoform-specific holoenzymes may be formed in the brain. The wide distribution of PKA gene expression in the brain, reflects, in part, its importance as a second messenger system for a great variety of neurotransmitters and neuromodulators. The axonal projections of neurons using many of these transmitters have been mapped in some detail. None of the projection maps, however, coincide with patterns of expression of individual PKA subunits. Evidently no particular isoform of PKA is dedicated solely to the transduction of a specific neurotransmitter. It appears, rather, that a particular complement of PKA subunits may serve to transduce many neurotransmitters. The differential expression of PKA subunits within specific regions of the brain such as the hippocampus and the medial habenula suggests that discrete functions may be associated with a particular subtype of PKA. This premise is supported by a growing body of evidence that documents biochemical differences between subunits. For example, different forms of RII have been shown to bind specifically to cellular macromolecules. RI@ binds the calmodulin binding protein P75 with high affinity (Sarkar et al., 1984). Rlla, which binds less well to P75, shows high affinity for the microtubule-associated protein MAP2 (Leiser et al., 1986). This suggests that a portion of type II kinase may be tethered to specific regions

PKA Gene 77

Table

Expression

1. Summary

in Mouse

Brain

of PKA Gene

Expression Relative

in Mouse Amount

Brain of Signal

Ca

CP

Rla

+++ +++ ++ +++ +++ +++ ++

+++ +++ ++ +++ +++ +++ +

+++ +++ t+ ++t +++ +++ ++

+++ +++ ++ +++ t+ ++ ++

++ +++ ++

++ +++ +

+++ ++ ++

++ ++ ++

ND ++ ++

ND + ++

ND ++ +t+

ND t+ +t

+++ +++

++t +

+++ +++

++ +t+

Hindbrarn __Locus coeruleus

+++

++

++

++

+

t

Cerebellum ___ .-~ Purkinje cells Molecular layer Granular layer

+++ ++ +++

+ ++ +++

+++ ++ +++

+++ ++ +++

++ + ++

++ + +

Corpus

callosum

+

+

+

+

+

+

Internal

Capsule

+

+

+

+

t

+

+

+

+

+

+

t

Nucleus/Area

Rlla

Forebrain Neocortex Piriform cortex Caudate-putamen Medial habenula Bed n stria terminalis Subfornical organ Clobus pallidus Thalamus Paraventricular n Reticular n Ventrolateral n Hypothalamus Supraoptic n Paraventricular n Suprachiasmatic n Hippocampal formation Dentate gyrus CA1 -3

Optic __--

tract

+, weak;

+ +, moderate;

+ + +, strong

signal,

n, nucleus;

ND,

within neurons, which could limit the range of preferred substrates. Association with macromolecules may also localize PKA to microenvironments such that the enzyme may be close to adenylate cyclase and thus see very high concentrations of CAMP Increased sensitivity of the holoenzyme to CAMP appears to be another biochemical specialization of at least one PKA subtype. Recently we compared the activation of holoenzymes formed in vitro from bacterially expressed Rla or RIO and purified bovine Cu. Halfmaximal activation of type Is PKA, (formed from RIP and Ca subunits) occurred at a CAMP concentration that was a-to S-fold lower than that measured for type la holoenzymes (formed from Rla and Ca subunits). This suggests that expression of type lb holoenzymes may provide a more sensitive signaling system (Cadd and McKnight, unpublished data). Given these and other biochemical differences, the ability to regulate the expression of a particular subunit may have important effects on the cell’s response to CAMP Studies have demonstrated RI@ gene induction in response to follicle-stimulating hormone in estrogenprimed rat granulosa cells (Hedin et al., 1987). RI@ protein and mRNA were also found to accumulate in response to 8-bromo-CAMP in Friend erythroleukemic

++ + + +++ +

++ ++ +++ ++ ++

ND +

ND +

t

+t ++ +

t

+++

ND ++

ND ++

+++ +

not determiend.

cells (Schwartz and Rubin, 1983). These observations demonstrate that at least one of the PKA subunit genes can be induced and raise the possibility that neurons might also regulate the expression of specific PKA subunit genes. The discrete expression of PKA subunit genes in the brain, together with the known differences in biochemical properties of individual subunits, implies that particular subtypes of PKA subserve specific signaling events within neurons. Expression or induction of particular PKA genes may be one method of modulating neural responses to stimuli.

Experimental

Procedures

PrObeS Radiolabeled RNA probes were generated from cDNA fragments of PKA subunits cloned into plasmids containing SP6 and T7 promoter sequences. Linearized plasmid was then transcribed using bacterial SP6 orT7 polymerase (Bethesda Research Laboratories] in the presence of [s5S]UTP (New England Nuclear). The single-stranded RNA probes were separated from unincorporated label by gel filtration chromatography, using a G-75 (Pharmacia) column run in lx SET (10 mM Tris [pH 7.51, 5 mM EDNA, 1% SDS). The probes were heat-denatured before application to eliminate possible tertiary structures. Figure 1 illustrates the portion of the cDNA clones of Rla, RIB, Ca, Cg, Rlla, and RI@ used to construct the RNA probes.

Neuron 78

The Rla probe was made complementary to bases 45-339 of the mouse RI cDNA sequence (Clegg, Knickerbocker, and McKnight, unpublished data), containing 54 bp of 5’ noncoding sequence, The RIB probe was made complementary to bases t-319 of the mousecDNAsequencepublished bycleggetal. (1988), containing 78 bp of 5’ noncoding sequence. The Ca probe was made complementary to bases l-410 of the mouse cDNA sequence published by Uhler et al. (1986a) and contained 192 bp of 5’noncoding sequence. The CB probe was made complementary to bases T-390 ofthe mousecDNAsequencepublished by Uhleret al. (1986b) and also contained 192 bp of 5’ noncoding sequence. The Rlla probe was made complementary to bases l-386 of the mouse cDNA sequence published by Scott et al. (1987). The RIIB probe was made complementary to bases 302-1500 of the rat cDNA sequence published by Jahnsen et al. (1986) and contained approximately 300 bp of 3’ noncoding sequence. Thus, whereas the RII probes were constructed from different regions, probes for mRNAs encoding Rla and RIB, as well as probes for those encoding Ca and CB, were made from the same region of homologous cDNAs. In Situ Hybridization In situ hybridization was performed as described by Rogers et al. (1987), with small modifications. Briefly, adult male Swiss Webster mice were sacrificed by decapitation under ether/CO2 anesthesia. Brain tissues were removed, immediately frozen in powdered dry ice, and stored at -7OOC. Transverse and sagittal brain sections 10 pm in thickness were cut in a cryostat and thaw-mounted on slides coated with poly-L-lysine (50 us/ml). A mouse embryo was obtained on day 14 of gestation, frozen in powdered dry ice, and stored at -7OOC. Sagittal sections of the embryo were thawmounted as indicated above. Prior to hybridization, tissue sections were warmed to room temperature and fixed in 4% paraformaldehyde in PBS (PH. 7.4) for 10 min. Slides were then rinsed twice in PBS and treated with 0.25% acetic anhydride in 0.1 M triethanolamine (pH 8.0) for 10 min to reduce nonspecific binding. Sections were prehybridized for 3 hr at 50°C in a buffer containing 50% formamide, lx Denhardt’s solution, 10% dextran sulfate, 10 mM Dm, 250 &ml yeast RNA, 100 &ml denatured salmon sperm DNA, 0.6M NaCI, 1 mM EDTA, and 10 mM Tris (pH 7.5). Slides were then washed briefly in 4x SSC (lx SSC is 150 mM NaCI, 15 mM sodium citrate [pH 7.21). The ?labeled probes were dissolved in the same buffer used for prehybridization, with a final concentration of 2.5 ng of probe per ml (specific activity of ~2.40 x 106 cpmlpmol), and 60 ul was applied to each slide. Slides were covered with Parafilm coverslips and incubated overnight in humidified chambers at 50°C. Coverslips were soaked off in 4x SSC. The slides were treated with 20 &ml RNAase A and subjected to a series of washes, including a stringent wash in 0.1x SSC at 50°C. After dehydration through a graded ethanol series, the sections were allowed to air dry. Autoradiography mRNAs of PKA subunits were localized by apposing the sections to X-ray film (Ultrofilm 3H, LKB Instruments) for 4-13 days. Films were then processed with a Drexel’s Unix-based Image Analysis System for the preparation of prints. For higher resolution, slides were coated with photographic emulsion (NTB-3, Kodak, diluted l:l), exposed for 15 days at 4”C, developed, and counterstained with cresyl violet before coverslipping. Silver grains were visualized using dark-field microscopy. Specificity of In Situ Hybridizations with Homologous mRNAs In addition to binding competition with a 500-fold molar excess of unlabeled probe (Figures 3G and 3H), the ability of probes for these closely related mRNAs to cross-react was also tested. We hybridized %labeled probe in the presence of a 500-fold molar excess of unlabeled probe for the related mRNA under otherwise normal conditions. Under these conditions, specific binding was not affected (data not shown). This indicates that, although Ca and C8 and Rla and RI8 are very similar at the nucleotide level in the regions used to make the probes, the hybridization conditions and RNAase treatment used prevented cross-hybridization. We have

found that incubation with a lOOO-fold molar slight loss of specific labeling due to competition

excess results in a (data not shown).

Analysis of In Situ Hybridization Brain in situ studies were performed in duplicate under identical conditions, using two animals for each experiment. Controls were performed as described for each experiment, Incubation with twice as much probe per section produced no significant increase in hybridization (data not shown). Visual inspection ofthe autoradiograms thus allowed a semiquantitative analysis of the density of probe hybridization. Acknowledgments The authors gratefully acknowledge Drs. Robert Steiner and Kimberly Rogers for expert advice, Dr. Christopher Fisher for assistance with dark-field microscopy, Drs. Margret Miller and Christopher Clegg for helpful comments on the manuscript, Drs. Joanne Richards and John Scott for providing cDNA clones, and Ms. Glenda Froelick for cryostat technical assistance. This work was supported by NIH grant GM 32875. Received

February

24, 1989;

revised

May

4, 1989.

References Bloch, B., Popovici, T., Levin, M. J., Tuil, D., and Kahn, A. (1985). Transferrin gene expression visualized in oligodendrocytes of the rat brain by using in situ hybridization and immunohistochemistry. Proc. Natl. Acad. Sci. USA 82, 6706-6710. Clegg, C. H., Cadd, C. C., and McKnight, C. S. (1988). Genetic characterization of a brain-specific form of the type I regulatory subunit of CAMP-dependent protein kinase. Proc. Natl. Acad. Sci. USA 85, 3703-3707. Dudai, Y. (1988). Neurogenetic dissection of learning and shortterm memory in Drosophila. Annu. Rev. Neurosci. I I, 537-63. Hedin, L., McKnight, G. S., Lifka, J., Durica, J. M., and Richards, J. S. (1987). Tissue distribution and hormonal regulation of messenger ribonucleic acid for regulatory and catalytic subunits of adenosine 3’,5’-monophosphate-dependent protein kinases during ovarian folliculardevelopment and luteinization in the rat. Endocrinology 120, 1928-1935. Jahnsen, T., Hedin, L., Kidd, V. J., Beattie, W. G., Lohmann, S. M., Walter, U., Durica, J., Schulz, T. Z., Schiltz, E., Browner, M., Lawrence, C. B., Goldman, D., Ratoosh, S. L., and Richards, J. S. (1986). Molecular cloning, cDNA structure and regulation of the regulatory subunit (R-IIs,) of type II CAMP-dependent protein kinase from rat granulosa cells. J. Biol. Chem. 267, 12352-12361. Kandel, learning:

E. R., and modulation

Schwartz, J. H. (1982). of transmitter release.

Molecular biology of Science 278, 433-443.

Krebs, E. G., and Beavo, J. A. (1979). Phosphorylation-dephosphorylation of enzymes. Annu. Rev. Biochem. 48, 923-959. Leiser, M., Rubin, C. S., and Erlichman, J. (1986). Differential binding of the regulatory subunits (RII) of CAMP-dependent protein kinase II from bovine brain and muscle to RII-binding proteins. J. Biol. Chem. 261, 1904-1908. Nairn, A. C., Hemmings, tein kinases in the brain.

H. C., Jr., and Greengard, P (1985). Annu. Rev. Biochem. 54, 931-976.

Pro-

Reimann, E. M., Walsh, D. A., and Krebs, E. C. (1971). Purification and properties of rabbit skeletal muscle adenosine 3’,5’monophosphate-dependent protein kinases. J. Biol. Chem. 246, 1986-1995. Rogers, K. V., Vician, L., Steiner, R. A., and Clifton, D. K. (1987). duced preprosomatostatin messenger ribonucleic acid in periventricular nucleus of hypophysectomized rats determined quantitative in situ hybridization. Endocrinology 12 I, 90-93.

Rethe by

Sarkar, D., Erlichman, I., and Rubin, C. S. (1984). Identification of a calmodulin-binding protein that co-purifies with the regulatory subunit of brain protein kinase II. J. Biol. Chem. 259, 9840-9846. Schwartz,

D.A.,

and

Rubin,

C. S. (1983).

Regulation

of CAMP-

PKA Gene 79

Expressron

in Mouse

Brain

dependent protein kinase subunit levels in Friend erythroleukemic cells. Effects of differentiation and treatment with 8-Br-CAMP and methyl isobutyl xanthine. I. Biol. Chem. 258, 777-784. Scott, j. D., Claccum, M. B., Zoller, M. I., Uhler, M. D., Helfman, D. M., McKnight, G. S., and Krebs, E. G. (1987). The molecular cloning of a type II regulatory subuni! of the CAMP-dependent protein kinase from rat skeletal muscle and mouse brain. Proc. Natl. Acad. SCI. USA 84, 5192-5196. Showers, M. 0, and Maurer, R. A. (1986). A cloned bovine cDNA encodes an alternate form of the catalytic subunit of CAMPdependent protein kinase. J. Biol. Chem. 261, 1628816291. Stern, j. C., Farooq, M., Norton, W. T., and Rubin, C. S. (1987). Differential expression of isoforms of the regulatory subunit of type II CAMP-dependent protein kinase in rat neurons, astrocytes and oligodendrocytes. 1. Biol. Chem. 262, 3002-3006. Uhler, M. D., Carmichael, D. F., Lee, D. C., Chrivia, E. G., and McKnight, C. S. (1986a). Isolation of cDNA ing for the catalytic subunit of mouse CAMP-dependent nase. Proc. Natl. Acad. Sci. USA 83, 1300-1304.

j. C., Krebs, clones codprotein kr-

Uhler, M. D., Chrivia, J. C., and McKnight, G. S. (1986b). Evidence for a second isoform of the catalytic subunit of CAMP-dependent protern kinase. j. Biol. Chem. 261, 15360-15363.