THE JOURNAL OF COMPARATIVE NEUROLOGY 416:475–495 (2000) Differential Expression of ␥-Aminobutyric Acid Type B Receptor-1a and -1b mRNA Variants in GABA and non-GABAergic Neurons of the Rat Brain FENGYI LIANG,1* YUMIKO HATANAKA,2 HARUMI SAITO,1 TETSUO YAMAMORI,3 AND TSUTOMU HASHIKAWA1 1Laboratory for Neural Architecture, Brain Science Institute, Riken, Wako, Saitama 351–0198, Japan 2Laboratory for Neuronal Recognition Molecules, Brain Science Institute, Riken, Wako, Saitama 351–0198, Japan 3Division of Speciation Mechanisms 1, National Institute for Basic Biology, Okazaki, Aichi 444–8585, Japan ABSTRACT To understand the heterogeneity of ␥-aminobutyric acid type B receptor (GABABR)mediated events, we investigated expression of GABABR1a and 1b mRNA variants in GABA and non-GABAergic neurons of the rat central nervous system (CNS), by using nonradioactive in situ hybridization histochemistry and, in combination with GABA immunocytochemistry, double labeling. In situ hybridization with a pan probe, which recognizes a common sequence of both GABABR1a and GABABR1b mRNA variants, demonstrated widespread expression of GABABR1 mRNA at various levels in the CNS. Both GABABR1a and GABABR1b were expressed in the neocortex, hippocampus, dorsal thalamus, habenula, and septum, but only GABABR1a was detected in cerebellar granule cells, in caudate putamen, and most hindbrain structures. A majority of GABA neurons in cerebral cortex showed hybridization signals for both GABABR1a and GABABR1b, whereas those in most subcortical structures expressed either or neither of the two. GABA neurons in thalamic reticular nucleus and caudate putamen hybridized primarily for GABABR1a. Purkinje cells in the cerebellar cortex expressed predominantly GABABR1b. GABA neurons in dorsal lateral geniculate nucleus did not display significant levels of either GABABR1a or GABABR1b mRNAs. These data suggested widespread availability of GABABR-mediated inhibition in the CNS. The differential but overlapping expression of GABABR1 mRNA variants in different neurons and brain structures may contribute to the heterogeneity of GABABR-mediated inhibition. Some GABA neurons possessed, but others might lack the molecular machinery for GABABR-mediated disinhibition, autoinhibition, or both. J. Comp. Neurol. 416:475–495, 2000. r 2000 Wiley-Liss, Inc. Indexing terms: GABAB receptor; double labeling; immunocytochemistry; in situ hybridization; forebrain; cerebellum As the main inhibitory neurotransmitter in the central nervous system (CNS), ␥-aminobutyric acid (GABA) plays important roles in regulating neuronal activity, plasticity, and pathogenesis. Its action is mediated through distinct type A, B, or C (GABAA, GABAB, or GABAC) receptors. In contrast to the ionotropic GABAA and GABAC receptors that mediate fast inhibitory postsynaptic potentials (IPSP), the metabotropic GABAB receptor (GABABR) mediates slow, long-lasting inhibitory neuronal responses that have been implicated in many important physiological functions and pathologic alterations in the brain (Bowery et al., 1990; Soltesz and Crunelli, 1992; Bonanno and Raiteri, 1993b; Mott and Lewis, 1994; Misgeld et al., 1995; Bowery, r 2000 WILEY-LISS, INC. 1997; Deisz, 1997; Bettler et al., 1998). GABABRs have been demonstrated in both pre- and postsynaptic components of both excitatory and inhibitory neurons in the Grant sponsor: Riken BSI; Grant number: 57911; Grant sponsor: NIBB Cooperative Research Program; Grant number: 98–173. Yumiko Hatanaka’s current address is: Division of Behavior and Neurobiology, National Institute for Basic Biology, Myodaiji, Okazaki, Aichi 444– 8585, Japan. *Correspondence to: Dr. Fengyi Liang, Laboratory for Neural Architecture, Brain Science Institute, Riken, 2–1 Hirosawa, Wako, Saitama 351– 0198, Japan. E-mail: [email protected] Received 26 March 1999; Revised 4 October 1999; Accepted 5 October 1999 476 central nervous system. Their existence in glia cells has also been reported (Hosli and Hosli, 1990; Hosli et al., 1990; Fraser et al., 1994). Presynaptic GABABR-mediated auto-/heteroinhibition represents a classic, well-known model for regulation of neurotransmitter release and has been subjected to extensive pharmacologic and electrophysiological studies in a variety of brain structures. Postsynaptic GABABRs mediate a late IPSP that may results in long-term changes in neurotransmission and dynamic response behavior of neurons (Misgeld et al., 1995; Davies and Collingridge, 1996; Mouginot and Gahwiler, 1996; Bowery, 1997; Deisz, 1997; Isaacson and Hille, 1997; Bettler et al., 1998). Among other findings, previous pharmacologic and electrophysiological studies have suggested the existence of different subtypes of GABABRs at pre- or postsynaptic sites and in different cell types and brain structures (Dutar and Nicoll, 1988; Bonanno and Raiteri, 1993b; Fukuda et al., 1993; Pitler and Alger, 1994; Bowery, 1997; Deisz et al., 1997; Bettler et al., 1998). The distribution of GABABR in the CNS has been demonstrated with receptor binding autoradiography (Bowery et al., 1987; Chu et al., 1990; Knott et al., 1993; Turgeon and Albin, 1993). Expression of GABABR1 mRNA in the rat brain and in the monkey thalamus has been shown by using radioisotope labeled riboprobes that recognize both of the two major GABABR1 mRNA variants, GABABR1a and GABABR1b (Kaupmann et al., 1997; Munoz et al., 1998; Lu et al., 1999). In addition, GABABR1 immunoreactivity in the rat brain and retina and GABABR1a and GABABR1b mRNA variant distribution in the cerebellar cortex and mouse retina have been reported (Kaupmann et al., 1998b; Koulen et al., 1998; Zhang et al., 1998; Fritschy et al., 1999; MargetaMitrovic et al., 1999). However, it is still unclear how the two major GABABR1 mRNA variants are distributed in different neuronal cell types in the CNS and how the possible differences in their distribution may contribute to the heterogeneity of GABABR-mediated late IPSP and hetero-/autoinhibition of neurotransmitter release in various CNS structures. To address these issues, we mapped, at cellular level, the expression of GABABR1 gene and its two major mRNA variants, GABABR1a and GABABR1b, in the rat brain by using nonradioactive in situ hybridization histochemistry. Of particular interests and obvious functional roles are GABABRs on GABA neurons. GABABRs on somata/ dendrites of GABA neurons may serve as a mechanism to reduce GABA neuronal activity (disinhibition) by recurrent collaterals or by GABAergic input from other GABA neurons. GABABRs on GABAergic axon terminals (autoreceptors) play important roles in regulating GABA release in a frequency-dependent manner. Both processes may act as gating mechanisms for the generation of long-term potentiation (LTP) and other types of use-dependent modification of neuronal excitability (Floran et al., 1988; Deisz and Prince, 1989; Giralt et al., 1990; Davies et al., 1991; Mott and Lewis, 1991; Metherate and Ashe, 1994; Doze et al., 1995; Hosford et al., 1995; Morishita and Sastry, 1995; Brucato et al., 1996; Caddick and Hosford, 1996; Bonanno et al., 1997). Although GABAB receptors on GABAergic axon terminals or somata have been demonstrated pharmacologically and electrophysiologically in a variety of brain structures, further clarification of their distribution and heterogeneity at molecular level would not only advance F. LIANG ET AL. our understanding of the function of this important class of inhibitory receptor, but also facilitate the search for more selective GABABR agonists/antagonists (Floran et al., 1988; Harrison et al., 1988; Giralt et al., 1990; Seabrook et al., 1991; Davies and Collingridge, 1993; Lambert and Wilson, 1993; Doze et al., 1995; Ulrich and Huguenard, 1996; Jarolimek and Misgeld, 1997; Kim et al., 1997; Le Feuvre et al., 1997; Vigot and Batini, 1997; Mouginot et al., 1998; Koulen et al., 1998). Therefore, the second aim of the present study was to investigate the expression of GABABR1a and 1b mRNA variants in different types of GABA neurons of the rat brain by using immunocytochemistry in situ hybridization histochemistry double-labeling techniques. Our results indicated that the distribution of GABABR1a in the CNS markedly differed from that of GABABR1b in different GABA and non-GABAergic neurons. Preliminary results have been reported in abstract form (Liang et al., 1998). MATERIALS AND METHODS cDNA cloning and in vitro transcription Rat brain mRNA was isolated from a 12-day-old rat pup (Wistar strain). First strand cDNA was synthesized by reverse transcription by using SuperScript II kit (GibcoBRL, Grand Island, NY). GABABR1 cDNA fragments were amplified by polymerase chain reaction and isolated by electrophoresis on agarose gel. They were then subcloned into the EcoRV site of pBluescript II SK vector (Stratagene, La Jolla, CA). JM109 competent cells (Stratagene) were transformed by the respective recombinant plasmids and cultured on LB plates containing ampicillin, isopropyl1-thio-␤-D-galactoside (IPTG) and 5-bromo-4-chloro-3indolyl-␤-D-galactoside (Xgal). White colonies were picked up from the plate and cultured in 2⫻ TY medium containing ampicillin. Recombinant plasmids were purified from the culture and the inserts were verified by sequence analysis by using an ABI 310 or ABI 377 DNA sequencer (Perkin-Elmer, Foster, CA). Three cDNA fragments were cloned. GABABR1a fragment is specific for GABABR1a variant, has a C⫹G ratio of 61.78% and corresponds to nucleotides 1 to 437 of rat GABABR1a gene (EMBL accession number Y10369). GABABR1b is specific for GABABR1b variant, has a C⫹G ratio of 72.73% and corresponds to nucleotides 21 to 141 of rat GABABR1b gene (EMBL accession number Y10370). GABABR1p recognizes both GABABR1a and GABABR1b variants, has a C⫹G ratio of 57.60% and corresponds to nucleotides 2408–2841 (2060–2493) of rat GABABR1a (GABABR1b) gene. The purified plasmids were linearized and then used as templates for in vitro transcription of digoxigenin-labeled antisense (or sense control) RNA probes with T3 (or T7) RNA polymerase. The nucleoside triphosphate mix for in vitro transcription consists of 10 mM ATP, 10 mM CTP, 10 mM GTP, 6.5 mM UTP, and 3.5 mM digoxigenin-11-UTP. According to the supplier (Boehringer Mannheim, Mannheim, Germany) average digoxigenin-11-UTP incorporation into transcribed RNA is about 4–5% of all nucleotides. In situ hybridization Sixteen adult rats (Wistar strain) of either sex, ages 10 to 14 weeks and weighing 160–240 grams, were deeply anesthetized with Nembutal (60 mg/kg, i.p.) and transcardially perfused with physiological saline (about 50 ml) EXPRESSION OF GABAB RECEPTOR-1 mRNA VARIANTS IN RAT CNS followed by 4% paraformaldehyde in 0.1 M phosphate buffer (800 ml, pH 7.4). The brain and spinal cord were removed and postfixed in the same fixative for 4–6 hours at 4°C. They were then cryoprotected in 30% sucrose at 4°C for 24–48 hours. Thirty-five-micrometer frozen sections were cut on a freezing microtome. All procedures involving experimental animals were approved by Riken Animal Care Committee. In situ hybridization histochemistry followed a protocol modified from previous descriptions (Liang and Jones, 1997; Hatanaka and Jones, 1998). Free-floating sections were rinsed twice in phosphate buffered saline (pH 7.4) and then treated sequentially with 0.75% glycine (15 minutes ⫻ 2), 0.3% Triton X-100 (10 minutes), 1 µg/ml proteinase K (15 minutes) and 0.25% acetic anhydride in 0.1 M triethanolamine (15 minutes). After two washes in 2⫻ saline sodium citrate (SSC; 1⫻ SSC ⫽ 0.15 M NaCl, 0.015 M Na3C6H5O7 ⭈ 2H2O) and 1 hour prehybridization at 60°C in a hybridization buffer (pH 7.0) without probe, sections were hybridized at 60°C for approximately 14–16 hours in a new hybridization buffer (pH 7.0) that contained one of the digoxigenin-labeled RNA probes. The hybridization buffer consisted of 2–5⫻ SSC (see below), 2% blocking reagent (Boehringer Mannheim), 50% deionized formamide, and 0.1% N-lauroylsarcosine. SSC concentration in the hybridization buffer was varied between 2⫻ and 5⫻ to offset differences in melting temperature derived from variations in the length and G⫹C content of each probes. Probe concentrations of approximately 0.07–0.18 µg/ml for GABABR1a, GABABR1p and 0.25 µg/ml for GABABR1b were used. At the end of hybridization, sections were sequentially treated in the following solutions (except for ribonuclease A, all containing 0.1% N-lauroylsarcosine): 2⫻ SSC plus 50% deionized formamide, 60°C, twice, 20 minutes each; ribonuclease A buffer (0.01 M Tris-HCl buffer, pH 8.0, plus 1 mM ethylenediaminetetraacetic acid and 0.5 M NaCl), room temperature, 10 minutes; ribonuclease A (0.02 mg/ml in ribonuclease A buffer), 37°C, 30 minutes; 2⫻ SSC, 60°C, twice, 20 minutes each; 0.2⫻ SSC, room temperature and then 60°C, 20 minutes each. To visualize hybridization signals by immunoalkaline phosphatase histochemistry, sections were rinsed in 0.1 M Tris-HCl buffered saline (TBS, pH 7.5, 10 minutes), incubated in 1% blocking reagent (in TBS, 60 minutes), and then in alkaline phosphatase-conjugated sheep antidigoxigenin antibody (Fab fragments, Boehringer Mannheim, 1:1,500 dilution in TBS, pH7.5, plus 1% blocking reagent, 60 minutes). After three washes in TBS plus 0.1% Tween 20, alkaline phosphatase (AP) activity was visualized either by nitroblue tetrazolium (NBT) and 5-bromo-4chloro-3-indolyl phosphate (BCIP) or by 2-hydroxy-3naphtoic acid-28-phenylanilide phosphate and fast red TR. The former reaction gave rise to a blue precipitate in hybridization-positive cells, whereas the latter produced a red fluorescent deposit with a maximum absorbance at 550 nm and maximum emission at 562 nm. The NBT-BCIP reaction lasted for 6–18 hours at room temperature in a substrate solution consisting of 0.033% NBT, 0.017% BCIP, 0.001 M levamisole, 0.1 M NaCl, and 0.05 M MgCl2 ⭈ 6H2O in 0.1 M Tris-HCl buffer, pH 9.5. The fluorescent AP reaction lasted for 0.5–2 hours at room temperature in a substrate solution consisting of 0.01% 2-hydroxy-3-naphtoic acid-28-phenylanilide phosphate, 0.025% fast red TR, 0.001 M levamisole, 0.1 M NaCl, and 0.01 M MgCl2 in 0.1 477 M Tris-HCl buffer (pH 8.0). Sections were then either processed further for GABA immunocytochemistry for double labeling (see below) or mounted on chrome alumgelatin–coated glass slides, air-dried, and cover-slipped by using an aqueous mounting medium. Negative control experiments included hybridization that used sense riboprobes and ribonuclease A treatment after hybridization in 0.01 M Tris-HCl buffer, pH 8.0, and 1 mM ethylenediaminetetraacetic acid without NaCl. Positive controls consisted of in situ hybridization with antisense riboprobes for glutamic acid decarboxylase-67 (GAD67) and brain-derived neurotrophic factor (BDNF). For each animal, a series of sections was also stained with thionin for cytoarchitecture. GABA immunocytochemistry For GABA immunocytochemistry or double labeling of GABA immunoreactivity and GABABR1 mRNA, sections were rinsed twice in 0.01 M phosphate buffered saline (pH 7.4) before being placed in 5% normal goat serum plus 0.3% Triton X-100 in 0.01 M phosphate buffered saline (pH 7.4) for 1 hour at room temperature. They were then incubated for 12 hours at room temperature in an affinitypurified, rabbit anti-GABA serum (Sigma, St. Louis, MO; 1:2,000 dilution) and washed in 0.01 M phosphate buffered saline (pH 7.4). For GABA immunoperoxidase, subsequent processes followed instructions supplied with the avidin biotinylated-peroxidase complex (ABC) kit from Vector Laboratories (Burlingame, CA). Peroxidase activity was visualized by H2O2 (0.01%) and diaminobenzidine tetrahydrochloride (DAB·4HCl, 0.05%) that gave rise to a brownish deposit at the site of GABA immunoreactivity. For GABA immunofluorescence, bound antibody was revealed by goat anti-rabbit IgG secondary antibody conjugated to fluorescein (Vector Laboratories; 1:100) that yielded green fluorescence in GABA positive cells. For control reactions of GABA immunocytochemistry, the primary antibody was replaced by equal volume of normal rabbit serum. Double labeling For studying the distribution of GABABR1 mRNA signal in GABA neurons, a sequential double-labeling procedure was carried out on the same histologic sections to visualize first GABABR1a, GABABR1b or GABABR1p mRNA signals by in situ hybridization histochemistry and then GABA immunoreactivity by immunocytochemistry. As with any double- or multiple-labeling techniques, there exists the possibility of masking of one reaction by the other(s) and there were those cells that were marginally (ambiguously) positive for one marker or the other; therefore, their identity as single- or double-labeled cells could not be definitely resolved. These issues were dealt with in the present study as follows: First, in addition to a nonfluorescent double-labeling technique that used NBT-BCIP for revelation of GABABR1 mRNAs and DAB/H2O2 for GABA immunoreactivity, a fluorescent double-labeling procedure was carried out by using the red fluorescent AP reaction for visualization of GABABR1 mRNAs and then immunofluorescence for GABA. Preparations from fluorescent doublelabeling experiments were compared with those from nonfluorescent double labeling to take the advantages of both methods and to cross-verify the results. Second, in addition to conventional microscope, Bio-Rad MRC600 laser scanning confocal microscope (Bio-Rad, Hercules, CA) was used to analyze fluorescent double-labeled materi- 478 F. LIANG ET AL. als to ascertain the identity of individual cells. Finally, only those cells with discernible nuclei were considered for drawing conclusions on the coexistence of GABA and GABABR1. Data analysis Both fluorescent and nonfluorescent histologic preparations were observed under a Nikon Eclipse 800 microscope. Fluorescent histochemical materials were further analyzed and imaged by a Bio-Rad MRC-600 laser scanning confocal microscope. Levels of hybridization signals for GABABR1 mRNAs in various brain structures were estimated by intensity of NBT-BCIP reaction products or fluorescence. Hybridization signals with antisense probes in medial habenular nucleus or Purkinje cells (for GABABR1b and GABABR1p) were denoted as very strong (⫹⫹⫹⫹, very high levels) and background hybridization similar to that with sense control probes as negative (-). Intermediate intensities were designated successively as strong (⫹⫹⫹, high levels), medium (⫹⫹, medium levels), weak (⫹, low levels), and very weak or marginal (⫾, very low levels). Initial ratings of relative signal levels in various structures were carried out independently by two investigators and wherever disagreements arose, a third investigator was called upon to verify the results. Low-magnification images of nonfluorescent preparations were photomicrographed first and then digitized with a Nikon LS-4500AF film scanner from negative films. High magnification photomicrographs were grabbed directly from the microscope to a Macintosh G3 computer by means of a digital color image acquisition system (Sony DXC-950 3CCD color video camera, Scion CG-7 color image grabber and Scion Image 1.62 software). Figures were prepared from digitized images by using Adobe Photoshop 5.0 and printed on a full-color digital printer (Fujix Pictography 3000, Fuji Film, Tokyo, Japan). For best printouts, the original image resolutions were scaled to 320 or 400 pixels per inch with Adobe Photoshop software. RESULTS All GABABR1a, GABABR1b, or GABABR1p mRNApositive stainings were located in the cell bodies. Many cells in a majority of neural structures were GABABR1 mRNA positive, but levels of mRNA signals in different cells varied to a large extent. Neurons in the medial habenular nucleus and cerebellar Purkinje cells showed the strongest hybridization for GABABR1p riboprobe. Marked differences in the distributions of GABABR1a and GABABR1b mRNA variants were observed in many brain structures, with some neuronal cell types expressing neither and others expressing either or both. Different subtypes of GABA neurons were distinguished based on their possession or lack of GABABR1a and/or GABABR1b mRNA signals. Table 1 summarizes relative levels of in situ hybridization signals for GABABR1a, GABABR1b, or GABABR1p mRNA variant in major brain structures/cell types. It should be noted that GABABR1b antisense probe might produce a hybridization signal of lower intensity by one order of magnitude in comparison with GABABR1a and GABABR1p because of its short length and extremely low probability of incorporating digoxigenin-11-UTP (see Discussion section). Any correlation between staining in- TABLE 1. Relative Levels of Hybridization Signals for GABABR1 mRNA Variants in Major CNS Structures1 Brain regions Neocortex Hippocampus Septum Habenula Thalamus Basal ganglia Hypothalamus Cerebellum Brain stem Spinal cord Structure/cells Layer V pyramidal cells Other layers Pyramidal cells Granule cells Other neurons LSD Others Medial Lateral TRN/VLG Other nuclei Caudate putamen Subthalamus SNC SNR Supraoptic nucleus Others Basket/stallete cells Purkinje cells Golgi cells Granule cells Deep cerebellar nuclei Colliculus Periaquiduct gray Pontine nuclei Cranial nerve nuclei Others Motoneurons Other laminae White matter glia cells GABABR1a GABABR1b GABABR1p (⫹⫹)–(⫹⫹⫹) (⫹)–(⫹⫹⫹) (⫹⫹⫹) (⫹⫹) (⫹)–(⫹⫹) (⫹⫹)–(⫹⫹⫹) (⫹⫹) (⫹⫹)–(⫹⫹⫹) (⫹⫹) (⫹)–(⫹⫹) (⫹⫹)–(⫹⫹⫹⫹) (⫹⫹⫹) (⫺)–(⫹⫹) (⫹⫹)–(⫹⫹⫹) (⫹) (⫺)–(⫹⫹) (⫹⫹)–(⫹⫹⫹) (⫹⫹) (⫹⫹⫹⫹) (⫾)–(⫹⫹) (⫺)–(⫾) (⫹⫹)–(⫹⫹⫹⫹) (⫹⫹)–(⫹⫹⫹) (⫹)–(⫹⫹⫹) (⫹⫹)–(⫹⫹⫹) (⫹⫹) (⫹)–(⫹⫹) (⫹⫹)–(⫹⫹⫹) (⫹⫹) (⫹⫹⫹⫹) (⫹⫹) (⫹)–(⫹⫹) (⫹⫹)–(⫹⫹⫹⫹) (⫹)–(⫹⫹) (⫹)–(⫹⫹) (⫹⫹) (⫹)–(⫹⫹) (⫺)–(⫾) (⫺)–(⫾) (⫹) (⫺) (⫹)–(⫹⫹) (⫹)–(⫹⫹) (⫹⫹)–(⫹⫹⫹) (⫹)–(⫹⫹) (⫹⫹)–(⫹⫹⫹) (⫹)–(⫹⫹) (⫺)–(⫾) (⫺)–(⫾) (⫹⫹)–(⫹⫹⫹) (⫹)–(⫹⫹) (⫹) (⫺)–(⫾) (⫹⫹)–(⫹⫹⫹) (⫹)–(⫹⫹) (⫺) (⫹⫹⫹⫹) (⫺) (⫺) (⫾) (⫹⫹⫹⫹) (⫹⫹)–(⫹⫹⫹) (⫹)–(⫹⫹) (⫹)–(⫹⫹) (⫹)–(⫹⫹) (⫺) (⫺)–(⫾) (⫹)–(⫹⫹) (⫹)–(⫹⫹) (⫹)–(⫹⫹⫹) (⫹⫹) (⫺)–(⫾) (⫺)–(⫾) (⫹)–(⫹⫹⫹) (⫹⫹) (⫹⫹)–(⫹⫹⫹) (⫹)–(⫹⫹) (⫹⫹)–(⫹⫹⫹) (⫹)–(⫹⫹) (⫺)–(⫹) (⫺)–(⫾) (⫹) (⫺)–(⫾) (⫹⫹)–(⫹⫹⫹) (⫹)–(⫹⫹) (⫹⫹)–(⫹⫹⫹) (⫹)–(⫹⫹) (⫹) (⫺)–(⫾) (⫹) 1Some but not necessarily all neurons in specified structures were positive for the mRNA. Hybridization signals: undetected (⫺), marginal/very weak (⫾), weak (⫹), medium (⫹⫹), strong (⫹⫹⫹), very strong (⫹⫹⫹⫹). Ratings were not for direct comparison between GABABR1a/1p and GABABR1b signals for reasons stated in the Discussion section. LSD, dorsal lateral septal nucleus; SNC, substantia nigra pars compacta; SNR, substantia nigra pars reticulata; TRN, thalamic reticular nucleus; VLG, ventral lateral geniculate nucleus. tensity and relative mRNA levels for GABABR1a/1p was not directly applicable to that for GABABR1b. Also for this reason, the levels of GABABR1b mRNA expression in the CNS as presented in the present study may represent an underestimation. Although intensity of hybridization signals in sections from different animals slightly varied, its relative ratings among different brain structures in individual animals kept consistent. GABABR1b sense control probe yielded a very weak hybridization signal in the pyramidal cell layer of hippocampal CA1–3 and granule cell layer of the dentate gyrus (Fig. 1B). Other CNS structures did not show any detectable hybridization other than a very faint background staining. Hybridization with sense control probes for GABABR1a and GABABR1p did not give rise to positive signal above background in any CNS structures (Fig. 1A). Posthybridization treatment of sections with ribonuclease A at low salt concentration also resulted in loss of positive hybridization signals. Neocortex Many neurons across all neocortical layers were GABABR1 mRNA positive. Figure 2 compares the expression pattern of GABABR1a (Fig. 2A) and GABABR1p (Fig. 2B) in the primary somatosensory cortex. Left panels in Figure 3 illustrates GABABR1b hybridization in different layers of the motor cortex. EXPRESSION OF GABAB RECEPTOR-1 mRNA VARIANTS IN RAT CNS 479 Fig. 1. Photomicrographs showing hybridization by ␥-aminobutyric acid (GABA)BR1a and GABABR1b sense control probes in the hippocampus. A: Hybridization with GABABR1a sense probe produced no hybridization signal above background. Gr, granule cell layer; Or, stratum oriens; Py, stratum pyramidalis; Rad, stratum radiatum. B: Hybridization with GABABR1b sense probe gave rise to a very weak hybridization signal in pyramidal cells of CA1-CA3 and in granule cells of the dentate gyrus. Scale bar ⫽ 200 µm in B (applies to A,B). In layer I, scattered neurons showed medium levels of GABABR1a/1p mRNA, but only a few neurons in this layer were weakly positive for GABABR1b mRNA. In layers II and III, hybridization signals for GABABR1a/1p variant were medium to high in many neurons of variable sizes, but those for GABABR1b were low to marginal. Among different neocortical areas, layers II-III neurons of the temporal, occipital, and the barrel cortex showed more pronounced hybridization than other areas. Hybridization signals in layer IV neurons were generally weak for GABABR1a/1p and very weak to unobservable for GABABR1b mRNA. In the barrel cortex, neurons inside the barrels were visibly weaker in their hybridization signal for GABABR1a and GABABR1p (Fig. 2). Layer V pyramidal neurons, especially those of the motor cortex, showed the strongest hybridization of all neocortical neu- 480 F. LIANG ET AL. Fig. 2. Photomicrographs showing variable expression levels of ␥-aminobutyric acid (GABA)BR1a (A) or GABABR1p (B) mRNA in different layers of the primary somatosensory cortex of the rat brain. Cortical layers are indicated in A. Scale bar ⫽ 200 µm in B (applies to A,B). rons for GABABR1a, 1b, or 1p probe, but hybridization signals in neurons of smaller sizes in layer V were generally medium to low for GABABR1a/1p and low to below detection for GABABR1b mRNA. mRNA levels in most layer VI neurons were low for GABABR1a/1p and low to marginal for GABABR1b (Figs. 2A,B, 3A,C,E). As revealed by fluorescent double labeling and illustrated in Figure 4, a majority of GABA neurons in the neocortex expressed GABABR1a/1p. Many neocortical GABA neurons, especially those in deep layers, also displayed low to medium levels of GABABR1b mRNA (Fig. 3A–F). In layer I, virtually all GABABR1a/1p-positive neurons were GABAergic. Some GABA neurons in layer I also showed low levels of GABABR1b mRNA expression. GABA neurons in layers II–VI were scattered among non-GABAergic GABABR1 mRNA positive neurons. Those in layers II to IV expressed variable levels of GABABR1a/1p and very low to indiscernible GABABR1b mRNA. Most GABA neurons in layers V and VI had weak to medium GABABR1a, 1b, or 1p mRNA expression (Figs. 3A–F, 4A–D). Same results were obtained from nonfluorescent double-labeling experiments (data not shown). Hippocampus and septum Pyramidal neurons of CA1–3 expressed strong GABABR1a and medium GABABR1b mRNA with relatively higher levels of expression for the latter variant in CA1 than in CA2–3. GABABR1p expression represented the summation of GABABR1a and GABABR1b mRNA variants in that pyramidal neurons in CA1 showed more prominent hybridization signals than in CA2–3. Granule cells of the dentate gyrus expressed medium levels of GABABR1a/1p and low levels of GABABR1b mRNA. Scattered neurons in strata oriens, radiatum, and lacunosum moleculare of CA1–3 and molecular, polymorphic layers of the dentate gyrus exhibited medium levels of GABABR1a/1p mRNA. Most of them were also weakly positive for GABABR1b mRNA (Fig. 5A–C). GABA immunocytochemistry revealed a distinct band of dense GABA axon terminals in stratum radiatum at the proximal dendritic segment of CA3 pyramidal cells (Fig. 6A,B). On double-labeling preparations, most GABA neurons stained positive for GABABR1a/1p, but a few of them in strata oriens and pyramidalis displayed little EXPRESSION OF GABAB RECEPTOR-1 mRNA VARIANTS IN RAT CNS 481 Fig. 3. Laser scanning confocal photomicrographs illustrating double labeling of ␥-aminobutyric acid (GABA)BR1b mRNA (left panels) and GABA immunoreactivity (right panels) on the same histologic section of the rat motor cortex. Expression of GABABR1b mRNA in GABA and non-GABAergic neurons in layers I–III (A,B) was very weak or below detection. Many cells in layers V (C,D) and VI (E,F) were moderately positive for GABABR1b. Note the relatively strong hybridization in layer V pyramidal neurons. Arrows in the right panels point to examples of GABA neurons and corresponding arrows at the same locations in the left panels indicate the GABABR1b expression status of the same neurons. Scale bar ⫽ 50 µm in F (applies to A–F). GABABR1a/1p hybridization signal. Both GABA and GABABR1a/1p mRNA levels varied to a certain extent in hippocampal GABA neurons (Fig. 6A–D). GABABR1b signals were also observed in a large proportion of hippocampal GABA neurons (Fig. 7A–D). GABABR1a and GABABR1b coexpressed in the septum. Although there existed considerable variation of GABABR1 mRNA staining intensity among individual neuronal cells, neurons in the dorsal lateral septal nucleus showed stronger GABABR1a, GABABR1b as well as GABABR1p mRNA staining than those in ventral or medial septal nuclei (Fig. 8A). Panels B to G in Figure 8 show fluorescent double labeling of GABABR1 mRNA variants and GABA immunoreactivity in the septum. All GABA neurons in the septum appeared positive for GABABR1a, GABABR1b, and GABABR1p mRNA. 482 F. LIANG ET AL. Fig. 4. Laser scanning confocal photomicrographs showing fluorescent double labeling of ␥-aminobutyric acid (GABA)BR1a (A,C) and GABA (B,D) on the same histologic section of the rat motor cortex. All GABA neurons in layers I, II (A,B), and a majority of GABA neurons in layers V and VI (C,D) were GABABR1a mRNA positive. Arrows in the right panels point to examples of GABA neurons, and corresponding arrows at the same locations in the left panels indicate the GABABR1a expression status of the same neurons. Cortical layers are indicated on the right in A and C. Scale bar ⫽ 100 µm in D (applies to A–D). Thalamus and caudate putamen neurons in this structure seemed GABAergic and expressed low to medium levels of GABABR1a/1p mRNA without detectable GABABR1b (Figs. 9A–C, 10A,B). In the dorsal lateral geniculate nucleus, GABA neurons were smaller in size than the surrounding non-GABAergic neurons, showed strong GABA immunoreactivity, but gave only marginal GABABR1a and no GABABR1b or GABABR1p mRNA signal (Fig. 11A–F). GABA neurons in ventral lateral geniculate nucleus had medium levels of hybridization signals for GABABR1a and GABABR1p (data not shown). Most other thalamic nuclei lacked significant number of GABA neurons. In the caudate putamen and globus pallidus, many neurons expressed low to medium levels of GABABR1a and GABABR1p mRNA, but GABABR1b hybridization signal was mostly not observed. Scattered GABA neurons in the caudate putamen and globus pallidus contained weak to moderate levels of GABABR1a/1p mRNA but expressed no apparent GABABR1b mRNA (Fig. 10C,D). Neurons in a majority of dorsal thalamic nuclei displayed medium to high levels of hybridization signals for both GABABR1a and GABABR1b mRNA. Thalamic reticular nucleus and ventral lateral geniculate nucleus, however, lacked significant hybridization signal for GABABR1b under the present experimental condition. As a result, GABABR1p mRNA levels in these nuclei were markedly lower than those in other thalamic nuclei, although levels of GABABR1a hybridization signals were comparable to those in the latter. Among different nuclei in the dorsal thalamus, mRNA levels varied in a similar pattern for GABABR1a, 1b, and GABABR1p. Hybridization signals for all three cRNA probes were visibly stronger in neurons of anterodorsal, mediodorsal, laterodorsal, medial geniculate, and dorsal lateral geniculate nuclei than in other thalamic subdivisions. Ventrolateral nucleus exhibited slightly higher signal intensities than the neighboring ventral posteromedial and ventral posterolateral nuclei. Neurons in intralaminar nuclei displayed weaker hybridization signals (Fig. 9A–F, results from anterodorsal and medial geniculate nuclei not illustrated). Upper panels in Figure 10 show fluorescent double labeling of GABABR1p mRNA hybridization (A) and GABA immunoreactivity (B) in thalamic reticular nucleus. All Cerebellum Different cell types in the cerebellar cortex had their unique patterns of expression of GABABR1 mRNA variants. Stellate/basket cells in the molecular layer showed weak GABA immunoreactivity, weak GABABR1a, very Fig. 5. Photomicrographs showing ␥-aminobutyric acid (GABA)BR1a (A), GABABR1b (B), or GABABR1p (C) mRNA expression in the hippocampus. Note the stronger hybridization signal in CA1 pyramidal neurons than in CA2–3 for GABABR1b and 1p. Gr, granule cell layer; Or, stratum oriens; Py, stratum pyramidalis; Rad, stratum radiatum. Scale bar ⫽ 200 µm in C (applies to A–C). 484 F. LIANG ET AL. Fig. 6. Laser scanning confocal photomicrographs showing double labeling of ␥-aminobutyric acid (GABA)BR1 mRNA and GABA immunoreactivity in the hippocampus. A,B: GABABR1p mRNA and GABA immunoreactivity in CA2–3. Note a band of GABAergic axon termi- nals in stratum radiatum in CA3. C,D: GABABR1a mRNA expression in pyramidal neurons and GABAergic interneurons in CA2–3. Scale bars ⫽ 200 µm in B (applies to A,B); 100 µm in D (applies to C,D). weak GABABR1p and undetectable GABABR1b mRNA expression. Purkinje cells were weakly GABA immunoreactive, had strong signals for GABABR1b and GABABR1p, but very weak signals for GABABR1a mRNA. Both Golgi and granule cells showed no signals for GABABR1b mRNA. Golgi cells had medium levels of GABA immunoreactivity and GABABR1a/1p mRNAs. The granule cells were negative for GABA, but weakly positive for GABABR1a/1p mRNA (Fig. 12A–F). Although Purkinje cells showed very weak hybridization signal for GABABR1a at probe concentrations used in the present study, they did show medium to strong hybridization for GABABR1a at higher probe concentrations (⬎0.4 µg/ml) as demonstrated in a series of pilot experiments designed to test the suitable probe concentration and to optimize the specificity and sensitivity of the technique. This positivity for GABABR1a at higher probe concentrations was arbitrarily defined as ‘‘false positive.’’ However, together with the observation that hybridization signal for GABABR1a was stronger than that for GABABR1p in the stellate/basket cells, these phenomena might have resulted from the possible existence of truncated forms of GABABR1a or other unidentified GABABR1 mRNA variants/subunits in the cerebellar cortex. In the deep cerebellar nuclei, two populations of neurons could be distinguished according to their size, GABA immunoreactivity, and GABABR1 mRNA signals. The relatively large neurons, which lacked GABA immunoreactivity, were moderately GABABR1a and GABABR1p mRNA positive. The small GABAergic neurons were weakly GABABR1a and GABABR1p mRNA positive. No GABABR1b mRNA-positive neurons were detected in the deep cerebellar nuclei (data not illustrated). Other CNS structures Low to medium levels of GABABR1a/1p hybridization signals were observed in various hypothalamic nuclei, with relatively strong signal in the supraoptic nucleus EXPRESSION OF GABAB RECEPTOR-1 mRNA VARIANTS IN RAT CNS 485 Fig. 7. Laser scanning confocal photomicrographs showing double labeling of ␥-aminobutyric acid (GABA)BR1b mRNA (A,C) and GABA immunoreactivity (B,D) in CA1 strata oriens, pyramidalis, and radiatum (upper panels) or in stratum radiatum and lacunosum moleculare (lower panels) of the hippocampus. Scale bar ⫽ 50 µm in D (applies to A–D). (Fig. 13A,B). GABABR1b hybridization in most cases could not be ascertained in the hypothalamus because of low signal. GABA neurons in the hypothalamus showed low to medium levels of GABABR1a/1p mRNA. Neurons in both the pars compacta and reticulata of substantia nigra hybridized at variable signal intensities for GABABR1a and GABABR1p (Fig. 13C). Weak expression of GABABR1b mRNA was only detected in substantia nigra pars compacta. Many neurons in various brainstem structures and all laminae of the spinal cord gray matter showed low to medium levels of GABABR1a/GABABR1p mRNA. Some cells in the spinal cord white matter also displayed weak GABABR1a/1p hybridization signal. Figure 13D shows the superior colliculus hybridized for GABABR1a mRNA. Although detected at low levels in the motoneurons of various cranial nerve nuclei and of lamina IX of spinal ventral horn, GABABR1b mRNA expression in most other structures of the brainstem and spinal cord could not be definitely established due to poor hybridization signal. Low to medium levels of GABABR1a/1p hybridization signal were detected in many GABA neurons of the brain stem and spinal cord structures including the superior and inferior colliculi, nucleus of the solitary tract, the superficial laminae of the spinal trigeminal nucleus and laminae II-III of the spinal cord dorsal horn. Figure 13E and F illustrate such an example of GABABR1p expression in GABA neurons in laminae II-III of the spinal cord dorsal horn. DISCUSSION The present study demonstrated widespread expression of GABABR1 receptor gene but differential distributions of the 1a and 1b mRNA variants in a variety of brain structures and neuronal cell types. GABA neurons of 486 F. LIANG ET AL. Fig. 8. A: Photomicrograph showing ␥-aminobutyric acid (GABA)BR1p mRNA expression in the septum. B–G: Laser scanning confocal photomicrographs showing double labeling of GABA (C, E, G) and GABABR1a (B), GABABR1b (D), or GABABR1p (F) mRNA in lateral septal nucleus. Middle column depicts GABABR1 mRNA hybridization and right column illustrates GABA immunoreactivity in the same fields. Arrows in the right panels point to some of the GABA neurons and corresponding arrows at the same locations in the middle panels indicate the expression status of GABABR1 mRNA variants in the same neurons. Scale bars ⫽ 100 µm in A,G (apply to A and B–G, respectively). different areal origins and cell types expressed very differently the GABABR1a and GABABR1b variants. fects have been described in many neural structures and cell types in spinal dorsal horn, cerebellum, substantia nigra, thalamus, hypothalamus, hippocampus, and neocortex, in line with the present findings of widespread expression of GABABR1 mRNA variants in the CNS (Bowery et al., 1980; Newberry and Nicoll, 1984, 1985; Andrade et al., 1986; Howe et al., 1987; Connors et al., 1988; Soltesz et al., 1989a; Hausser and Yung, 1994; Misgeld et al., 1995; Davies and Collingridge, 1996; Mouginot and Gahwiler, 1996; Deisz, 1997; Isaacson and Hille, 1997; Bettler et al., 1998). However, it has been shown that in heterologous mammalian expression systems GABABR1 subunit alone could not fully function as a mature receptor. A GABABR2 receptor subunit has been cloned recently, and heterodimerization of the GABABR1 and GABABR2 subunits appears to be necessary for the assembly of a fully functional GABAB receptor (Jones et al., 1998; Kaupmann et al., 1998a; White et al., 1998; Kuner et al., 1999). Low-level expression of GABABR1 mRNA has also been detected in peripheral organs (Castelli et al., 1999). It remains to be seen whether expression products of the GABABR1 variants serve functions other than mediating late IPSP and auto-/heteroinhibition of transmitter release. Widespread availability of GABABR-mediated inhibition? GABABR1 mRNAs were expressed by surprisingly numerous neuronal cell types throughout the CNS. Although their levels varied considerably in different cell types and CNS structures, these results may still suggest that the GABAB receptor-mediated mechanisms are widely disposable to a variety of neuronal cell types and CNS structures. This corroborated previous data from molecular biological, electrophysiological, pharmacologic, receptor binding autoradiographic as well as recent immunocytochemical works on GABABRs in a variety of brain structures and cell types. The present results of strong GABAB1 receptor gene expression in the medial habenular nucleus, dorsal thalamus, and cerebellar Purkinje cells conform to previous results by receptor binding, in situ hybridization, and immunocytochemistry (Bowery et al., 1987; Chu et al., 1990; Knott et al., 1993; Turgeon and Albin, 1993; Kaupmann et al., 1997, 1998b; Munoz et al., 1998; MargetaMitrovic et al., 1999; Fritschy et al., 1999; Lu et al., 1999). GABABR-mediated physiological and pharmacologic ef- Fig. 9. Photomicrographs showing expression of ␥-aminobutyric acid (GABA)BR1 mRNA variants in the thalamus. A–C: Expression of GABABR1a (A), GABABR1b (B), or GABABR1p mRNA (C), in different nuclei of the thalamus. D–F: Differential expression of GABABR1a (D), GABABR1b (E), or GABABR1p (F) mRNA in the dorsal and ventral lateral geniculate nuclei (DLG and VLG). CL, centrolateral thalamic nucleus; LD, laterodorsal thalamic nucleus; MD, mediodorsal thalamic nucleus; R, thalamic reticular nucleus; VL, ventrolateral nucleus; VPL, ventral posterolateral thalamic nucleus; VPM, ventral posteromedial thalamic nucleus. Scale bars ⫽ 400 µm in C (applies to A–C); 200 µm in F (applies to D–F). 488 F. LIANG ET AL. Fig. 10. Laser scanning confocal photomicrographs showing expression of ␥-aminobutyric acid (GABA)BR1p mRNA in GABA and nonGABAergic neurons in thalamic reticular nucleus and ventral posterolateral nucleus (A,B) and in the caudate putamen (C,D). Left column illustrates expression of GABABR1p mRNA and right column, GABA immunoreactivity. Arrows in the right panels point to some of the GABA neurons, and corresponding arrows at the same locations in the left panels indicate the GABABR1p expression status of the same neurons. For abbreviations, see Figure 9. Scale bar ⫽ 100 µm in D (applies to A–D). In general, the present results on the distribution of GABABR1 mRNAs agree well with that of GABABR1 immunoreactivities (Koulen et al., 1998; Malitschek et al., 1998; Fritschy et al., 1999; Margeta-Mitrovic et al., 1999). The lack of detectable GABABR1b mRNA expression in the caudate putamen, thalamic reticular nucleus, and most brain stem structures, as shown in the present study, is reflected by a negative immunostaining thereof (Fritschy et al., 1999). However, in contrast to the present results of stronger GABABR1a hybridization signals in most CNS structures, Western blotting results suggest GABABR1b as a dominant variant in the adult rat brain (Malitschek et al., 1998; Fritschy et al., 1999; Margeta-Mitrovic et al., 1999). This discrepancy may arise from the fact that mRNA levels do not always reflect the synthesis of protein or the functional relevance of the gene (Tecott et al., 1994; see Lu et al., 1999 for further discussions). More importantly, it should be noted that sequence information specific for GABABR1b is only 140 bp nucleotides. The relative low GABABR1b signal levels were likely to be attributed to the short length and very low U content of GABABR1b antisense probe (U contents of GABABR1a, 1p, and 1b antisense probes were 76, 98, and 8, respectively). Thus, if one digoxigenin-11-UTP is incorporated out of every five probe uracil bases, each copy of GABABR1b antisense probe had on average 1.6 digoxigenin molecules, whereas each copy of GABABR1a and GABABR1p antisense probe might include 15.2 and 19.6 digoxigenin molecules, respectively. Considering the amplifying effect by anti-digoxigenin antibody, the difference could be even bigger in number of bound alkaline phosphatase molecules per mRNA copy between GABABR1a/1p and GABABR1b. Further complicated by a nonlinear correlation between intensity of AP reaction deposits and the number of bound enzyme molecules or the duration of AP reaction, it may be misleading to directly compare the hybridization signal intensity for GABABR1b variant with those for the GABABR1a or 1p probe. Also for the same reason, low-level expression of GABABR1b mRNA in certain structures might have escaped detection. EXPRESSION OF GABAB RECEPTOR-1 mRNA VARIANTS IN RAT CNS 489 Fig. 11. Laser scanning confocal photomicrographs comparing ␥-aminobutyric acid (GABA) immunoreactivity (right column) and GABABR1 mRNA expression (left column) in dorsal lateral geniculate nucleus. Many non-GABAergic neurons stained positive for GABABR1a (A), GABABR1b (C) or GABABR1p (E) mRNA. However, GABA neurons in the same structure, as revealed by simultaneous staining of GABA immunoreactivity on the same histologic sections and illus- trated in B, D, and F, respectively, were only marginally positive for GABABR1a and mostly devoid of GABABR1b or GABABR1p mRNA expression. Arrows in the right panels point to some of the GABA neurons, and corresponding arrows at the same locations in the left panels indicate the expression status of GABABR1 mRNA variants in the same neurons. Scale bar ⫽ 50 µm in F (applies to A–F). In addition to neurons, glia-like cells in the spinal white matter near the pia mater showed weak GABABR1a/1p mRNA expression. This result supports previous findings of GABABR-mediated effects in glial cells of the CNS (Hosli and Hosli, 1990; Hosli et al., 1990; Fraser et al., 1994). Expression of GABABR1 mRNA in glial cells, however, 490 Fig. 12. Photomicrographs showing expression of ␥-aminobutyric acid (GABA)BR1a (A,B), GABABR1b (C), and GABABR1p (E,F) mRNA expression in the cerebellar cortex. A, C, and E depict in situ hybridization results. B and F show non-fluorescent double labeling of GABABR1a or GABABR1p mRNA variants (blue) and GABA immuno- F. LIANG ET AL. reactivity (brown) on the same histologic sections. D: Distribution of GABA-immunoreactive cells in the cerebellar cortex. G, Golgi cell; gr, granule cell layer; P, Purkinje cell; s/b, stellate/basket cell. Scale bar ⫽ 100 µm in F (applies to A–F). EXPRESSION OF GABAB RECEPTOR-1 mRNA VARIANTS IN RAT CNS Fig. 13. Photomicrographs showing expression of ␥-aminobutyric acid (GABA)BR1p mRNA in the hypothalamus (A,B), GABABR1a mRNA in the substantia nigra (C) and superior colliculus (D), and GABABR1p mRNA in spinal cord dorsal horn (E). F: GABAimmunoreactive neurons in the same field as E. GABA neurons were GABABR1p mRNA positive in laminae II-III of the gray matter of the 491 spinal dorsal horn. InG, intermediate gray layer of superior colliculus; LA, lateral anterior hypothalamic nucleus; LH, lateral hypothalamic area; SCh, suprachiasmatic nucleus; SNC, pars compacta of substantia nigra; SNR, pars reticulata of substantia nigra; SO, supraoptic nucleus; SuG, superficial gray layer of superior colliculus. Scale bars ⫽ 200 µm in D (applies to A–D), 100 µm in F (applies to E,F). 492 F. LIANG ET AL. Fig. 14. Schematic diagrams showing the four expression profiles of ␥-aminobutyric acid (GABA)BR1a and/or GABABR1b mRNA variants in different GABA neurons of the rat brain. The possibly differential targeting of GABABR1a and GABABR1b variants into preand postsynaptic GABABR receptors is according to a recent hypothesis, which remains unconfirmed (Bettler et al., 1998; Kaupmann et al., 1998b; Zhang et al., 1998). I: Most GABA neurons in the neocortex, hippocampus, and septum expressed both GABABR1a and GABABR1b. II: GABA neurons in thalamic reticular nucleus, caudate putamen, and in stellate/basket cells of the cerebellum showed mainly GABABR1a mRNA signal. III: As exemplified by Purkinje cells in the cerebellar cortex, this group of GABA neurons expressed predominantly GABABR1b mRNA variant. IV: GABA neurons in the dorsal lateral geniculate nucleus lacked significant expression of both GABABR1a and GABABR1b mRNA variants. GABABR(1a⫹2), GABAB receptor heterodimer formed by GABABR1a and GABABR2 subunits. GABABR(1b⫹2), GABAB receptor heterodimer formed by GABABR1b and GABABR2 subunits. may be restricted to certain regions of the CNS and at a level much lower than in neurons because a previous study did not find significant GABABR1 gene expression in glial cells of hippocampal CA3 region (Kaupmann et al., 1997). The present study, as well, did not detect significant levels of GABABR1a, 1b, or 1p hybridization signal in glia cells in a majority of rat CNS structures other than the spinal white matter. It is unlikely that the widespread distribution of GABABR1a and 1p was an artifact caused by nonspecific hybridization or cross-hybridization of the probes with other types of mRNA. In a series of pilot experiments, probe concentrations ranging from 0.07 to 0.7 µg/ml for GABABR1a and GABABR1p were tested. Although probe concentrations above 0.4 µg/ml gave rise to certain background staining for GABABR1p and strong Purkinje cell positivity for GABABR1a hybridization, probe concentrations ranging from 0.07–0.3 µg/ml produced consistent hybridization patterns for different brain structures and cell types. In addition, hybridization with sense control probes gave no positive signal except for the very weak hybridization for GABABR1b sense probe in the hippocampus. Finally, the present results on the distribution of the two mRNA variants in the cerebellar cortex agree with recently published data in this structure (Kaupmann et al., 1998b). septum expressed both GABABR1a and GABABR1b variants. This finding is in strong contrast to GABA neurons in most subcortical structures. GABA neurons in thalamic reticular nucleus, ventral lateral geniculate nucleus, caudate putamen, and GABAergic stellate/basket cells in the cerebellar cortex hybridized mainly for GABABR1a antisense probe without detectable GABABR1b mRNA signal. Cerebellar Purkinje cells showed mainly GABABR1b mRNA expression. Finally, GABA neurons in the dorsal lateral geniculate nucleus showed little expression of GABABR1 mRNA variants. Presynaptic GABABR-mediated autoinhibition of GABA release has been extensively studied and many important neurophysiological and pathologic processes such as nociception, use-dependent depression of inhibition, LTP, and epileptogenesis have been associated with this selfcontrolling mechanism (Floran et al., 1988; Giralt et al., 1990; Seabrook et al., 1991; Morishita and Sastry, 1995; Bonanno et al., 1997; for reviews, see Bowery, 1997; Deisz, 1997; Bettler et al., 1998). GABABR-mediated postsynaptic responses in GABA neurons have been less well characterized. Nevertheless, this disinhibitory mechanism by other GABA neurons or by recurrent collaterals of the GABA neurons themselves would also reduce GABAergic inhibitory output. Thus, activation of GABABR-mediated events in GABA neurons, either pre- or postsynaptic, would cause increased excitation of the integrated network, contrary to the effects of GABABR activation in excitatory neurons. The present findings suggest that the GABABR-mediated late IPSP and presynaptic autoinhibition of GABA release are differentially disposable to different GABA neurons. Those GABA neurons in the dorsal lateral genicu- Differential expression of GABABR1 mRNA variants in GABA neurons As illustrated schematically in Figure 14, four expression profiles of GABABR1a and/or GABABR1b mRNA variants were distinguished in different GABA neurons. Many GABA neurons in the neocortex, hippocampus and EXPRESSION OF GABAB RECEPTOR-1 mRNA VARIANTS IN RAT CNS late nucleus seemed lacking both mechanisms, in sharp contrast to most GABA neurons in the neocortex and hippocampus. In the context of these results, it is interesting to note that little GABABR-mediated responses have been detected in the intrinsic GABAergic interneurons of dorsal lateral geniculate nucleus of both cats and rats (Pape and McCormick, 1995; Williams et al., 1996). These interneurons, capable of firing trains of action potentials at a frequency exceeding 500 Hz in the cat, seemingly use alternative mechanisms, such as muscarinic acetylcholine receptors, for modulating their capability to gate the transmission of visual information from retina to visual cortex (McCormick and Pape, 1988; Pape and McCormick, 1995). The GABABR1a-positive GABA neurons such as those in the striatum, thalamic reticular nucleus, and cerebellar stellate/basket cells may also have physiological and pharmacologic properties very different from the GABA neurons in the cerebral cortex that expressed both GABABR1a and GABABR1b, or from cerebellar Purkinje cells that mainly expressed GABABR1b. It has been shown previously that GABAB autoreceptors in the cerebral cortex and spinal cord represent pharmacologically distinct receptor subtypes (Bonanno and Raiteri, 1993a). In view of the proposition that the two variants might be differentially located to axon terminals and somatodendritic domains of neurons (see below), it remains to be determined how these findings relate to the functions of individual GABA neuron cell types and to their positioning in the integrated neural network. Margeta-Mitrovic and coworkers (1999) have found very few neurons that coexpressed GABABR1 immunoreactivity and glutamic acid decarboxylase-67, in contrast to the present results showing expression of either or both of the GABABR1a and 1b mRNA variants in GABA neurons of most CNS structures. It remains unknown whether this discrepancy with the present results is caused by extremely low level of GABABR1 mRNA translation, limited sensitivity of the antibody, or unrecognized GABABR variants/subunits in GABA neurons. Of particular relevance to the last point, for example, the antibody used by MargetaMitrovic et al. (1999) would not be able to detect the recently cloned GABABR1d (Isomoto et al., 1998). The present GABABR1b antisense probe, in contrast, would also hybridize with GABABR1c and 1d mRNA variants as the two share the same 58 end with GABABR1b (Isomoto et al., 1998). mRNA variants for receptor subtypes? The differential but partly overlapping expression of the two mRNA variants in various neuronal cell types and brain structures as shown in the present study, may contribute to the production of diverse subtypes of GABAB receptors. In line with this molecular diversity, variable physiological and pharmacologic profiles of GABABR agonist and antagonist actions have been demonstrated previously in different CNS structures and cell types (for reviews see, Bowery, 1997; Deisz, 1997; Bettler et al., 1998). However, the differential and variable expression patterns of GABABR1a and GABABR1b mRNA variants in both GABA and non-GABA neurons deny the possibility of separation of each variant into either GABA or non-GABA category. Differences in intracellular signal transduction cascades and effector systems may greatly enhance the functional heterogeneity and specificity derived from varia- 493 tions in GABABR variant expression in different neuronal cell types. It has been recently proposed for the cerebellar cortex and retina that GABABR1a and GABABR1b variants might target pre- and postsynaptic GABABR subtypes, respectively (Bettler et al., 1998; Kaupmann et al., 1998b; Zhang et al., 1998). Although the present study showed diverse patterns of expression of the two variants in the central nervous system, a detailed correlation of the present data with previous electrophysiological and pharmacologic studies seems to extend the postulation to several other brain regions. The present study demonstrated, for example, the predominant GABABR1a expression in the striatum, ventral lateral geniculate nucleus, and the thalamic reticular nucleus. In the same brain regions, electrophysiological studies have failed to detect significant late IPSPs mediated by postsynaptic GABABRs (Soltesz et al., 1989b; Calabresi et al., 1991; Crunelli and Leresche, 1991; Nisenbaum et al., 1993; Ulrich and Huguenard, 1996; SanchezVives and McCormick, 1997). Thus, it seemed conceivable that the mature translation product of GABABR1a mRNA in neurons of these structures was mainly for locations at presynaptic axon terminals. Both pre- and postsynaptic GABABR-mediated events have been demonstrated for the neocortex, hippocampus, and most dorsal thalamic nuclei (Bowery, 1997; Deisz, 1997; Bettler et al., 1998). Again, this agrees with the present findings of coexpression of the two variants by these structures. In accordance with the present results showing abundant expression of GABABR1b and very weak expression of GABABR1a mRNA by cerebellar Purkinje cells, previous studies have reported high levels of postsynaptic GABABRs on Purkinje cell dendrites in molecular layer of the cerebellar cortex and low levels of presynaptic GABABRs on Purkinje axon terminals in deep cerebellar nuclei (Bowery et al., 1987; Chu et al., 1990; Turgeon and Albin, 1993; Morishita and Sastry, 1995; Mouginot et al., 1998). Likewise, it may be deduced that most brain stem and spinal neuronal cell types, which mainly expressed the GABABR1a variant as shown in the present study, would manifest primarily a presynaptic GABABR-mediated effect. Except in the retina, no definite evidence has been presented for presynaptic GABABR1 immunoreactivity, either on GABA or non-GABAergic axon terminals, despite the vast number of electrophysiological reports on the matter. A direct immunohistochemical demonstration of GABABR1a peptide distribution in the CNS, for example, has not been possible because of the lack of suitable specific antibodies (Koulen et al., 1998; Margeta-Mitrovic et al., 1999; Fritschy et al., 1999). In view of the recent cloning of GABABR1c/1d and the possibility of other unidentified GABABR1 variants, the subtraction strategy (GABABR1p ⫺ GABABR1b ⫽ GABABR1a distribution) as used by Fritschy and coworkers (1999) is no longer tenable. Because peptides are mostly synthesized in the cell bodies and subsequently transported to their final destinations, immunoreactivity detected in the cell bodies and proximal dendrites might represent immature, preassembling forms of the receptor subunit and, therefore, could not rule out that the functioning form is located at axon terminals. It awaits further studies, especially those at electron microscopic level by using variant-specific antibodies, to resolve the veracity of the postulation. 494 F. LIANG ET AL. ACKNOWLEDGMENTS We thank Dr. Katsuyoshi Ishii for many helpful discussions and Kiseko Shionoya, Takumi Agaki, Yasuhiro Yamazaki, Chishiro Wakabayashi, and Yuki Hasegawa for technical assistance. LITERATURE CITED Andrade R, Malenka RC, Nicoll RA. 1986. A G protein couples serotonin and GABABR receptors to the same channels in hippocampus. Science 234:1261–1265. Bettler B, Kaupmann K, Bowery N. 1998. GABAB receptors: drugs meet clones. Curr Opin Neurobiol 8:345–350. Bonanno G, Raiteri M. 1993a. ␥-Aminobutyric acid (GABA) autoreceptors in rat cerebral cortex and spinal cord represent pharmacologically distinct subtypes of the GABAB receptor. J Pharmacol Exp Ther 265:765–770. Bonanno G, Raiteri M. 1993b. Multiple GABAB receptors. Trends Pharmacol Sci 14:259–261. Bonanno G, Fassio A, Schmid G, Severi P, Sala R, Raiteri M. 1997. Pharmacologically distinct GABAB receptors that mediate inhibition of GABA and glutamate release in human neocortex. Br J Pharmacol 120:60–64. Bowery NG. 1997. Pharmacology of mammalian GABAB receptors. 2nd Ed. In: Enna SJ, Bowery NG, editors. Totowa: Humana Press. p 209–236. Bowery NG, Hill DR, Hudson AL, Doble A, Middlemiss DN, Shaw J, Turnbull M. 1980. (-) Baclofen decreases neurotransmitter release in the mammalian CNS by an action at a novel GABA receptor. Nature 283:92–94. Bowery NG, Hudson AL, Price GW. 1987. GABAA and GABAB receptor site distribution in the rat central nervous system. Neuroscience 20:365– 383. Bowery NG, Knott C, Moratalla R, Pratt GD. 1990. GABAB receptors and their heterogeneity. Adv Biochem Psychopharmacol 46:127–139. Brucato FH, Levin ED, Mott DD, Lewis DV, Wilson WA, Swartzwelder HS. 1996. Hippocampal long-term potentiation and spatial learning in the rat: effects of GABAB receptor blockade. Neuroscience 74:331–339. Caddick SJ, Hosford DA. 1996. The role of GABAB mechanisms in animal models of absence seizures. Mol Neurobiol 13:23–32. Calabresi P, Mercuri NB, De Murtas M, Bernardi G. 1991. Involvement of GABA systems in feedback regulation of glutamate-and GABAmediated synaptic potentials in rat neostriatum. J Physiol (Lond) 440:581–599. Castelli MP, Ingianni A, Stefanini E, Gessa GL. 1999. Distribution of GABA(B) receptor mRNAs in the rat brain and peripheral organs. Life Sci 64:1321–1328. Chu DC, Albin RL, Young AB, Penney JB. 1990. Distribution and kinetics of GABAB binding sites in rat central nervous system: a quantitative autoradiographic study. Neuroscience 34:341–357. Connors BW, Malenka RC, Silva LR. 1988. Two inhibitory postsynaptic potentials, and GABAA and GABAB receptor-mediated responses in neocortex of rat and cat. J Physiol (Lond) 406:443–468. Crunelli V, Leresche N. 1991. A role for GABAB receptors in excitation and inhibition of thalamocortical cells. Trends Neurosci 14:16–21. Davies CH, Collingridge GL. 1993. The physiological regulation of synaptic inhibition by GABAB autoreceptors in rat hippocampus. J Physiol (Lond) 472:245–265. Davies CH, Collingridge GL. 1996. Regulation of EPSPs by the synaptic activation of GABAB autoreceptors in rat hippocampus. J Physiol (Lond) 496:451–470. Davies CH, Starkey SJ, Pozza MF, Collingridge GL. 1991. GABAB autoreceptors regulate the induction of LTP. Nature 349:609–611. Deisz RA. 1997. Electrophysiology of GABAB receptors. 2nd Ed. In: Enna SJ, Bowery NG, editors. Totowa: Humana Press. p 157–207. Deisz RA, Prince DA. 1989. Frequency-dependent depression of inhibition in guinea-pig neocortex in vitro by GABAB receptor feed-back on GABA release. J Physiol (Lond) 412:513–541. Deisz RA, Billard JM, Zieglgansberger W. 1997. Presynaptic and postsynaptic GABAB receptors of neocortical neurons of the rat in vitro: differences in pharmacology and ionic mechanisms. Synapse 25:62–72. Doze VA, Cohen GA, Madison DV. 1995. Calcium channel involvement in GABAB receptor-mediated inhibition of GABA release in area CA1 of the rat hippocampus. J Neurophysiol 74:43–53. Dutar P, Nicoll RA. 1988. Pre- and postsynaptic GABAB receptors in the hippocampus have different pharmacological properties. Neuron 1:585– 591. Floran B, Silva I, Nava C, Aceves J. 1988. Presynaptic modulation of the release of GABA by GABAA receptors in pars compacta and by GABAB receptors in pars reticulata of the rat substantia nigra. Eur J Pharmacol 150:277–286. Fraser DD, Mudrick-Donnon LA, MacVicar BA. 1994. Astrocytic GABA receptors. Glia 11:83–93. Fritschy JM, Meskenaite V, Weinmann O, Honer M, Benke D, Mohler H. 1999. GABAB-receptor splice variants GB1a and GB1b in rat brain: developmental regulation, cellular distribution and extrasynaptic localization. Eur J Neurosci 11:761–768. Fukuda A, Mody I, Prince DA. 1993. Differential ontogenesis of presynaptic and postsynaptic GABAB inhibition in rat somatosensory cortex. J Neurophysiol 70:448–452. Giralt MT, Bonanno G, Raiteri M. 1990. GABA terminal autoreceptors in the pars compacta and in the pars reticulata of the rat substantia nigra are GABAB. Eur J Pharmacol 175:137–144. Harrison NL, Lange GD, Barker JL. 1988. (-)-Baclofen activates presynaptic GABAB receptors on GABAergic inhibitory neurons from embryonic rat hippocampus. Neurosci Lett 85:105–109. Hatanaka Y, Jones EG. 1998. Early region-specific gene expression during tract formation in the embryonic rat forebrain. J Comp Neurol 395:296– 309. Hausser MA, Yung WH. 1994. Inhibitory synaptic potentials in guinea-pig substantia nigra dopamine neurones in vitro. J Physiol (Lond) 479:401– 422. Hosford DA, Lin FH, Kraemer DL, Cao Z, Wang Y, Wilson JT Jr. 1995. Neural network of structures in which GABAB receptors regulate absence seizures in the lethargic (lh/lh) mouse model. J Neurosci 15:7367–7376. Hosli E, Hosli L. 1990. Evidence for GABAB-receptors on cultured astrocytes of rat CNS: autoradiographic binding studies. Exp Brain Res 80:621–625. Hosli L, Hosli E, Redle S, Rojas J, Schramek H. 1990. Action of baclofen, GABA and antagonists on the membrane potential of cultured astrocytes of rat spinal cord. Neurosci Lett 117:307–312. Howe JR, Sutor B, Zieglgansberger W. 1987. Baclofen reduces post-synaptic potentials of rat cortical neurones by an action other than its hyperpolarizing action. J Physiol (Lond) 384:539–569. Isaacson JS, Hille B. 1997. GABA(B)-mediated presynaptic inhibition of excitatory transmission and synaptic vesicle dynamics in cultured hippocampal neurons. Neuron 18:143–152. Isomoto S, Kaibara M, Sakurai-Yamashita Y, Nagayama Y, Uezono Y, Yano K, Taniyama K. 1998. Cloning and tissue distribution of novel splice variants of the rat GABAB receptor. Biochem Biophys Res Commun 253:10–15. Jarolimek W, Misgeld U. 1997. GABAB receptor-mediated inhibition of tetrodotoxin-resistant GABA release in rodent hippocampal CA1 pyramidal cells. J Neurosci 17:1025–1032. Jones KA, Borowsky B, Tamm JA, Craig DA, Durkin MM, Dai M, Yao WJ, Johnson M, Gunwaldsen C, Huang LY, Tang C, Shen Q, Salon JA, Morse K, Laz T, Smith KE, Nagarathnam D, Noble SA, Branchek TA, Gerald C. 1998. GABA(B) receptors function as a heteromeric assembly of the subunits GABA(B)R1 and GABA(B)R2. Nature 396:674–679. Kaupmann K, Huggel K, Heid J, Flor PJ, Bischoff S, Mickel SJ, McMaster G, Angst C, Bittiger H, Froestl W, Bettler B. 1997. Expression cloning of GABA(B) receptors uncovers similarity to metabotropic glutamate receptors. Nature 386:239–246. Kaupmann K, Malitschek B, Schuler V, Heid J, Froestl W, Beck P, Mosbacher J, Bischoff S, Kulik A, Shigemoto R, Karschin A, Bettler B. 1998a. GABA(B)-receptor subtypes assemble into functional heteromeric complexes. Nature 396:683–687. Kaupmann K, Schuler V, Mosbacher J, Bischoff S, Bittiger H, Heid J, Froestl W, Leonhard S, Pfaff T, Karschin A, Bettler B. 1998b. Human ␥-aminobutyric acid type B receptors are differentially expressed and regulate inwardly rectifying K⫹ channels. Proc Natl Acad Sci USA 95:14991–14996. Kim U, Sanchez-Vives MV, McCormick DA. 1997. Functional dynamics of GABAergic inhibition in the thalamus. Science 278:130–134. Knott C, Maguire JJ, Moratalla R, Bowery NG. 1993. Regional effects of pertussis toxin in vivo and in vitro on GABAB receptor binding in rat brain. Neuroscience 52:73–81. EXPRESSION OF GABAB RECEPTOR-1 mRNA VARIANTS IN RAT CNS Koulen P, Malitschek B, Kuhn R, Bettler B, Wassle H, Brandstatter JH. 1998. Presynaptic and postsynaptic localization of GABA(B) receptors in neurons of the rat retina. Eur J Neurosci 10:1446–1456. Kuner R, Kohr G, Grunewald S, Eisenhardt G, Bach A, Kornau HC. 1999. Role of heteromer formation in GABAB receptor function. Science 283:74–77. Lambert NA, Wilson WA. 1993. Heterogeneity in presynaptic regulation of GABA release from hippocampal inhibitory neurons. Neuron 11:1057– 1067. Le Feuvre Y, Fricker D, Leresche N. 1997. GABAA receptor-mediated IPSCs in rat thalamic sensory nuclei: patterns of discharge and tonic modulation by GABAB autoreceptors. J Physiol (Lond) 502:91–104. Liang F, Jones EG. 1997. Differential and time-dependent changes in gene expression for type II calcium/calmodulin-dependent protein kinase, 67 kDa glutamic acid decarboxylase, and glutamate receptor subunits in tetanus toxin-induced focal epilepsy. J Neurosci 17:2168–2180. Liang F, Hatanaka Y, Hashikawa T. 1998. Patterns of GABAB receptor gene expression in GABAergic neurons of rat brain. Neurosci Res Suppl 22:S83–S83. Lu XY, Ghasemzadeh MB, Kalivas PW. 1999. Regional distribution and cellular localization of ␥-aminobutyric acid subtype 1 receptor mRNA in the rat brain. J Comp Neurol 407:166–182. Malitschek B, Ruegg D, Heid J, Kaupmann K, Bittiger H, Frostl W, Bettler B, Kuhn R. 1998. Developmental changes of agonist affinity at GABAB1 receptor variants in rat brain. Mol Cell Neurosci 12:56–64. Margeta-Mitrovic M, Mitrovic I, Riley RC, Jan LY, Basbaum AI. 1999. Immunohistochemical localization of GABA(B) receptors in the rat central nervous system. J Comp Neurol 405:299–321. McCormick DA, Pape HC. 1988. Acetylcholine inhibits identified interneurons in the cat lateral geniculate nucleus. Nature 334:246–248. Metherate R, Ashe JH. 1994. Facilitation of an NMDA receptor-mediated EPSP by paired-pulse stimulation in rat neocortex via depression of GABAergic IPSPs. J Physiol (Lond) 481:331–348. Misgeld U, Bijak M, Jarolimek W. 1995. A physiological role for GABAB receptors and the effects of baclofen in the mammalian central nervous system. Prog Neurobiol 46:423–462. Morishita W, Sastry BR. 1995. Pharmacological characterization of preand postsynaptic GABAB receptors in the deep nuclei of rat cerebellar slices. Neuroscience 68:1127–1137. Mott DD, Lewis DV. 1991. Facilitation of the induction of long-term potentiation by GABAB receptors. Science 252:1718–1720. Mott DD, Lewis DV. 1994. The pharmacology and function of central GABAB receptors. Int Rev Neurobiol 36:97–223. Mouginot D, Gahwiler BH. 1996. Presynaptic GABAB receptors modulate IPSPs evoked in neurons of deep cerebellar nuclei in vitro. J Neurophysiol 75:894–901. Mouginot D, Kombian SB, Pittman QJ. 1998. Activation of presynaptic GABAB receptors inhibits evoked IPSCs in rat magnocellular neurons in vitro. J Neurophysiol 79:1508–1517. Munoz A, Huntsman MM, Jones EG. 1998. GABA(B) receptor gene expression in monkey thalamus. J Comp Neurol 394:118–126. 495 Newberry NR, Nicoll RA. 1984. A bicuculline-resistant inhibitory postsynaptic potential in rat hippocampal pyramidal cells in vitro. J Physiol (Lond) 348:239–254. Newberry NR, Nicoll RA. 1985. Comparison of the action of baclofen with ␥-aminobutyric acid on rat hippocampal pyramidal cells in vitro. J Physiol (Lond) 360:161–185. Nisenbaum ES, Berger TW, Grace AA. 1993. Depression of glutamatergic and GABAergic synaptic responses in striatal spiny neurons by stimulation of presynaptic GABAB receptors. Synapse 14:221–242. Pape HC, McCormick DA. 1995. Electrophysiological and pharmacological properties of interneurons in the cat dorsal lateral geniculate nucleus. Neuroscience 68:1105–1125. Pitler TA, Alger BE. 1994. Differences between presynaptic and postsynaptic GABAB mechanisms in rat hippocampal pyramidal cells. J Neurophysiol 72:2317–2327. Sanchez-Vives MV, McCormick DA. 1997. Functional properties of perigeniculate inhibition of dorsal lateral geniculate nucleus thalamocortical neurons in vitro. J Neurosci 17:8880–8893. Seabrook GR, Evans ML, Howson W, Benham CD, Lacey MG. 1991. Presynaptic inhibition mediated by GABAB receptors in rat striatal brain slices. Ann N Y Acad Sci 635:495–496. Soltesz I, Crunelli V. 1992. GABAA and pre- and post-synaptic GABAB receptor-mediated responses in the lateral geniculate nucleus. Prog Brain Res 90:151–169. Soltesz I, Lightowler S, Leresche N, Crunelli V. 1989a. On the properties and origin of the GABAB inhibitory postsynaptic potential recorded in morphologically identified projection cells of the cat dorsal lateral geniculate nucleus. Neuroscience 33:23–33. Soltesz I, Lightowler S, Leresche N, Crunelli V. 1989b. Optic tract stimulation evokes GABAA but not GABAB IPSPs in the rat ventral lateral geniculate nucleus. Brain Res 479:49–55. Tecott LH, Eberwine JH, Barchas JD, Valentino KL. 1994. Methodological considerations in the utilization of in situ hybridization. New York: Oxford University Press. p 3–23. Turgeon SM, Albin RL. 1993. Pharmacology, distribution, cellular localization, and development of GABAB binding in rodent cerebellum. Neuroscience 55:311–323. Ulrich D, Huguenard JR. 1996. GABAB receptor-mediated responses in GABAergic projection neurones of rat nucleus reticularis thalami in vitro. J Physiol (Lond) 493:845–854 Vigot R, Batini C. 1997. GABA(B) receptor activation of Purkinje cells in cerebellar slices. Neurosci Res 29:151–160. White JH, Wise A, Main MJ, Green A, Fraser NJ, Disney GH, Barnes AA, Emson P, Foord SM, Marshall FH. 1998. Heterodimerization is required for the formation of a functional GABA(B) receptor. Nature 396:679–682. Williams SR, Turner JP, Anderson CM, Crunelli V. 1996. Electrophysiological and morphological properties of interneurones in the rat dorsal lateral geniculate nucleus in vitro. J Physiol (Lond) 490:129–147. Zhang C, Bettler B, Duvoisin RM. 1998. Differential localization of GABA(B) receptors in the mouse retina. Neuroreport 9:3493–3497.