Early GABAA Receptor Clustering During the Development of the Rostral Nucleus of the Solitary Tract W

J. Anat. (2003) 202, pp387–396

Blackwell Publishing Ltd. Early GABAA receptor clustering during the development of the rostral nucleus of the solitary tract W. L. Heck,1 A. M. Basaraba,2 A. Slusarczyk2 and L. Schweitzer2 1Departments of Biology and Cell & Molecular Physiology, University of North Carolina-Chapel Hill, Chapel Hill, NC 27599, USA 2Department of Anatomical Sciences and Neurobiology, School of Medicine, University of Louisville, Louisville, KY 40292, USA

Abstract

While there is an abundance of gamma-aminobutyric acid (GABA) in the gustatory zone of the nucleus of the solitary tract of the perinatal rat, we know that GABAergic synapse formation is not complete until well after birth.

Our recent results have shown that GABAB receptors are present at birth in the cells of the nucleus; however, they do not redistribute and cluster at synaptic sites until after PND10. The present study examined the time course of appearance and redistribution of GABAA receptors in the nucleus. GABAA receptors were also present at birth.

However, in comparison to GABAB receptors, GABAA receptors underwent an earlier translocation to synaptic sites.

Extrasynaptic label, for example, of GABAA receptors was non-existent compared to GABAB receptors at PND10 and well-defined clusters of GABAA receptors could be seen as early as PND1. We propose that while GABAA receptors may play an early neurotransmitter role at the synapse, GABAB receptors may play a non-transmitter neurotrophic role. Key words GABA receptors; gustation; inhibition; nucleus solitarius; taste.

Synaptogenesis takes place between postnatal days 1 Introduction and 10 (PND1–10) of the developing rNST (Lasiter & Evidence suggests that the inhibitory neurotransmitter Kachele, 1989). We have demonstrated that during this γ-aminobutyric acid (GABA) plays an important role in phase, GABA is present in many somata and neurites, development and the processing of gustatory informa- but of the few synapses present, less than 20% are tion in the rostral nucleus of the solitary tract (rNST). It GABA-immunoreactive (Brown et al. 2000). The abund- has been demonstrated that 20% of neurones in the ance of GABA and the scarcity of GABAergic terminals adult rNST contain GABA (Lasiter & Kachele, 1988) and suggest that GABA has a non-synaptic role early in GABA-immunoreactive synaptic terminals represent development of the rNST. By weaning (PND 20) the about half of all synaptic terminals within the adult amount of overall GABA decreases, but the number of rNST (Leonard et al. 1999). Furthermore, the application synapses, including GABAergic synapses, in the rNST of GABA inhibits most adult rNST neurones (Liu et al. rises to adult levels (Brown et al. 2000). These anatomical 1993; Wang & Bradley, 1993) and there is evidence that data suggesting synaptic immaturity are supported by the strong inhibitory influences of corticofugal inhibi- electrophysiological brain slice data (Kim et al. 1997). tion are played out through GABAergic mechanisms The rat brainstem cells recorded from during the stage (Smith & Li, 2000). Smith & Li (1998), for example, dem- of synaptogenesis (PND1-PND7) were more sensitive to onstrated that GABA tonically inhibits approximately GABA than were older cells, and responded often with 60% of the taste-responsive neurones in the adult depolarization – a phenomenon not observed in the rNST. adult. Thus, developmental immaturity is demonstrated for the GABAergic system in the rNST of the postnatal rat brainstem. Correspondence Maturation of GABA receptors must be key in the Dr Laura Schweitzer, Department of Anatomical Sciences and Neurobiology, University of Louisville School of Medicine, Louisville, KY transition from developmental, perhaps non-transmitter, 40292, USA. Fax: 502 852 4895; e-mail: [email protected] to transmitter functions of GABA. In the adult, there is Accepted for publication 3 February 2003 ample evidence that many rNST neurones contain

388 GABAA receptor development in rNST, W. L. Heck et al.

both GABAA and GABAB receptors. Many neurones demonstrated at both the light and the electron

respond to both the GABAA receptor agonist muscimol microscopic levels that there is a change of distribution

and the GABAB receptor agonist baclofen by decreasing of GABAB receptors during synapse formation in the input resistance and hyperpolarizing (Wang & Bradley, rNST (Heck et al. 2001) from diffuse (PND1 to PND10, 1993; Bradley et al. 1996). Some rNST neurones respond stage of synaptogenesis) to clustered (PND15 to adult).

only to muscimol and therefore contain only GABAA At the EM level, the amount of extrasynaptic label receptors (Nakagawa et al. 1991). The inhibitory effects drops after rapid synaptogenesis (PND10) in the rNST. of GABA are partly or entirely blocked in most adult Like neurotransmitter receptors in other systems such rNST neurones by equivalent concentrations of the as the neuromuscular junction (Bevan & Steinbach,

GABAA receptor antagonist bicuculline, while super- 1977; Hall & Sanes, 1993), it is evident that synapse

fusion of the GABAB receptor antagonist phaclofen formation is associated with clustering of GABAB receptors depresses membrane responses to GABA only in and an increased specificity at synapses.

some neurones (Nakagawa et al. 1991; Liu et al. 1993; As reviewed above, GABAA receptors may be even

Wang & Bradley, 1993; Bradley et al. 1996), further more important than GABAB receptors to modulating confirming that many rNST neurones contain only neurotransmission in the brainstem taste nucleus rNST.

GABAA receptors. The following paper describes GABAA receptor immu- GABA receptors are functional in the very young rat noreactivity, clustering and synapse formation of this brainstem, including the rNST (Nakagawa et al. 1991; important receptor type, and compares it with what is Grabauskas & Bradley, 1999, 2001). Although functional, known about the physiological responses and our

they are clearly immature. Grabauskas & Bradley (1999) observations relating to GABAB receptors. studied the characteristics of stimulus-evoked IPSPs in rat brainstem slices at 0–7 days and > 55 days (adult) in Materials and methods order to assess developmental changes in inhibitory synaptic activity in the rNST. Dose–response curve of Sprague-Dawley rats were obtained from Harlan

the IPSPs to the GABAA receptor antagonist bicuculline (Indianapolis, IN, USA). All animal manipulations and shifts to the left during development, such that there anaesthetic methods used in our protocol were in is a more than 10-fold change in sensitivity during accordance with the National Institute of Health Guide maturation. for the Care and Use of Laboratory Animals and

There is evidence that GABAA receptors function in approved by the Institutional Animal Care and Use

transmission mode prior to GABAB receptors during Committee. development. Du & Bradley (1998) used whole-cell recordings in current clamp mode to investigate the GABA receptor antibody responses of acutely isolated neurones from the A

rNST of rats aged PND8 to PND24 to GABA and GABA Polyclonal goat anti-GABAA receptor antibody (sc-7363) receptor agonists and antagonists. They only found was obtained from Santa Cruz Biotechnology, Inc. support for GABA responses in the immature rNST (Santa Cruz, CA, USA). This antibody is raised against an

being mediated by GABAA and not GABAB receptors. amino acid sequence mapping at the amino terminus β The GABAA receptor antagonists picrotoxin and bicuc- (N-19) of the GABAA R 2 subunit and reacts with β β β ulline blocked the effect of GABA in a concentration- GABAA R 1, GABAA R 2, and GABAA R 3 of mouse, rat

dependent manner, while GABAB receptor antagonists and human origin. did not (although their number of samples was admit- tedly small). In addition, there appears to be a transient GABA receptor immunohistochemistry for light expression of GABA receptors prior to PND 14. To A C microscopy (LM) recap these results, physiological results support the

contention that GABAA receptors mature prior to Four rats from each of the following ages: PND1, 5, 10,

GABAB receptors. 15, 20 and adult were injected intraperitoneally with an In contrast to this body of knowledge regarding the overdose of 50 mg kg−1 of sodium pentobarbital followed electrophysiological maturation of GABA receptors in by transcardial perfusion with 4% paraformaldehyde the rNST, anatomical studies are lacking. Recently, we in 0.1 M phosphate buffer (PB). Brains were removed

GABAA receptor development in rNST, W. L. Heck et al. 389 from the skull and post-fixed for approximately 2 h at 4 °C. Analysis of light microscopy data After post-fixing, brains were transferred to a solution of 30% sucrose in 4% paraformaldehyde overnight. The density of GABAA receptor immunoreactivity within Coronal sections, 40 µm thick, were obtained using individual neurones was measured as follows. Area and a frozen sliding microtome and collected in a divided staining density (darkness) measurements of digitized dish filled with phosphate-buffered saline (PBS). images were made using the NIH Image Program available Sections were incubated in an endogenous peroxidase on the Internet at http://rsb.info.nih.gov/nih-image/. blocking solution (3% hydrogen peroxide and 10% The outlines of individual cell somas were traced in methanol in PBS) for 20 min, rinsed and then incubated order to obtain an estimate of cell cross-sectional for 30 min at room temperature in a blocking solution area and then the cell’s staining density (darkness) was of 5% normal rabbit serum (Vector Laboratories, determined. To correct for non-specific background Burlingame, CA, USA) and 0.3% Triton-X (Spectrum staining, the traced outline of the cell was then Chemicals, Gardena, CA, USA). They were then incubated superimposed over an adjacent predominantly stain- overnight at 4 °C in the primary antibody diluted free area, the inferior cerebellar peduncle. The density 1 : 200 in blocking solution. In all experiments, every of staining of this area was subtracted to ‘correct’ for fifth section was processed without the primary variations in background staining between sections. The antibody (omission control) added to the blocking ‘corrected’ density (darkness of labelling per mm2) solution. Yet, in all other regards, they were processed was quantified in 400 neurones (20 of the darkest along with the other sections. No labelling was seen in neurones per each of four rats from each of five age any control section. groups). Differences in the immunoreactivity in rNST On the next day, sections were rinsed with PBS neurones of each age group were compared using an and then incubated for 1 h at room temperature in a individual one-way ANOVA. When the overall ANOVA was biotinylated secondary antibody (rabbit antigoat significant, Fisher post hoc PLSD tests were used to test IgG, Vector Laboratories) diluted 1 : 200 in blocking differences between individual age groups. Alpha solution. Standard methods described in detail in our levels of P < 0.05 were used throughout. paper on GABAB receptors (Heck et al. 2001) were used to react the tissue and visualize the reaction Post-embedding GABA receptor product. A immunohistochemistry for electron microscopy

The methods used are notable for their preservation of Identifying the rNST ultrastructure, maintenance of immunogenicity and We defined the rNST as the rostral one-third of the excellent signal to background label levels, and are NST. This was identified in adult brainstem sections as more extensively described in our methods paper (Heck the rostral 1.33 mm of the nucleus (corresponding to et al. 2002). Ultrastructural localization of GABAA Paxinos & Watson, 1986; atlas plates 60–65 labelled receptors in the rNST was determined in adult SOL). The emerging facial and glossopharyngeal nerve (> PND60) and PND10 rats with electron microscopy. roots served, respectively, as rostral and caudal land- Four rats from each age group were processed. Since marks. The corresponding region in the immature the results were consistent, no additional animals were brainstem sections was identified in our standard needed. reference series of myelin and Nissl-stained brainstems Rats were deeply anaesthetized with an overdose of from rats PND1, 5, 10, 15 and 20 cut in sagittal, horizontal pentobarbital and perfused transcardially first with and coronal planes. The region identified as ‘rostral’ for Tyrode solution (Hayat, 1981), followed by a fixative the purpose of this study corresponds to 0.87 mm of consisting of 4% paraformaldehyde and 0.1% glutaral- the NST as shown in Paxinos et al. (1994); PND0 (day of dehyde in 0.1 M phosphate buffer (PB). The brainstems birth) atlas plate(s) 59–64 (NST is also labelled SOL in were removed from the skull, post-fixed in the perfusate this atlas). The cranial nerve root landmarks identified for 2 h at 4 °C, transferred to 0.1 M PB and incubated in the adult remained in their relative location in the overnight at 4 °C. younger brainstems and helped guide our choice of The next day, coronal 50-µm sections of brainstems sections. were cut using a Vibratome (Leica VT 1000S; Leica

390 GABAA receptor development in rNST, W. L. Heck et al.

Microsystems Inc., Deerfield, IL, USA) and placed in the brainstem at the ultrastructural level. Electron PBS. The free-floating sections were incubated in 0.4% micrographs of synapses and other structures of interest Triton X-100 in PBS for 1 h at room temperature and were taken at a magnification of 28 500× from at rinsed in PB. Each section was then viewed via light least four sections spanning the rostro-caudal extent microscopy to identify those sections containing the (approximately every 300 µm) of the rNST. Approxi- rNST. mately 16 synapses in each section were photographed The following day, sections were placed in 1% randomly within each section without regard to immu-

osmium tetroxide (OsO4) in PB for 30 min at room nogold labelling. For analysis, photomicrographs were temperature, and following standard processes they printed at 1.8× for a final magnification of 51 000×. were flat embedded in LX-112 (Ladd Research Indus- The threshold for particle density (gold particles per tries, Burlington, VT, USA). The plasticized brain stems length) for positive identification of immunoreactive

were examined under a light microscope in order to GABAA receptors was set by comparing the density locate the rNST. Four representative coronal sections of colloidal gold particles at apparently labelled pre- spanning the rostro-caudal axis of the nucleus were synaptic terminals or post-synaptic targets (synaptic label selected for each animal. After verifying inclusion of = signal) with equivalent measurements in adjacent the rNST on 1-µm-thick sections, thin sections were non-synaptic regions of the same size (extrasynaptic then cut at 60–80 nm with a diamond knife and placed label = background). The adjacent region selected was on Formvar-coated nickel grids. a set distance and direction away from the synapse as The method used to determine the ultrastructural long as this placement did not encroach on another localization of GABA receptors in the rNST was modified synapse. The density of label was measured by counting from that described by Phend et al. (1992) and the number of colloidal gold particles and dividing the described in detail in our methods paper (Heck et al. count by the length of the synaptic density determined 2002). Incubation in the primary antibody diluted with the NIH Image software. Differences between the 1 : 200 in TBST pH 7.6 occurred for 18 h at 4 °C. A number of colloidal gold particles per unit length in pre-absorption control was performed on some sections. synaptic processes and neighbouring non-synaptic These particular sections were incubated in goat regions at each age were compared with one-way

anti-GABAA receptor polyclonal antibody that was ANOVAs. neutralized by a blocking peptide (sc-7363p, Santa Cruz Biotechnology, Inc.). For neutralization, the antibody was combined with five-fold excess (by weight) of Results blocking peptide for 2 h at room temperature. Next, Light microscopy the antibody/peptide mixture was diluted 1 : 200 and

this mixture was applied to sections on grids in the An abundance of GABAA receptor-immunoreactive identical manner to the primary antibody. Only a few neurones were found in the rostral nucleus of the gold particles were seen in any of these sections, and solitary tract (rNST), confirming previous observations

those present were not specifically associated with that GABAA receptors are numerous in this brainstem

synaptic profiles. nucleus (Terai et al. 1998). GABAA receptor immuno- The next day, all grids were rinsed in buffer and then reactivity in the adult rNST was seen in both cell somata incubated for 1 h in a 15-nm gold-particle secondary and processes (Fig. 1D). There was an interesting

antibody solution (EMRAG15 rabbit antigoat IgG, uneven distribution of GABAA receptor immunoreactivity Research Diagnostics Inc., Flanders, NJ, USA) diluted that was present in most labelled cells at this age that 1 : 25 in TBST pH 8.2. The grids were then rinsed, is best described as ‘clustered’. In some cases distinct stained for 30 min with uranyl acetate in methanol, puncta of darkened label can be seen. This adult-like followed by staining with 0.1% lead citrate. pattern of label was well established by PND10 (Fig. 1C). Somatic staining for GABA receptors was present Analysis of electron microscopy data A throughout development of the rNST. Most cells at

A Phillips CM10 transmission electron microscope was PND1 had diffuse labelling for GABAA receptors

used to examine GABAA receptor immunoreactivity in (Fig. 1A), although a few (approximately 1% of

GABAA receptor development in rNST, W. L. Heck et al. 391

Fig. 1 Light microscopic images of immunoreactivity for GABAA receptors in PND1 (A), PND 5 (B), PND 10 (C) and adult (D) rat rNST. Immunoreactivity is present in somata (asterisks) and extends into the dendritic processes (arrows) at all ages. Immunoreactivity is present right after birth in the rat in many cells, but the clusters (arrowheads) do not become common until PND5. On PND1, a few large neurones show the clustered pattern of immunoreactivity (one example is seen in the centre of this figure). The distinct clustering of immunoreactivity is present in each labelled cell by PND10 and remains in the adult.

labelled neurones, usually one to two large cells per neurones. By PND5 (Fig. 1B), the clustering of GABAA section) neurones did have a distinct pattern of cluster- receptor immunoreactivity had a much more common ing even at this earliest postnatal age (Fig. 1A, largest pattern and, as mentioned above, the adult pattern cell in the centre). The clustering extended from the cell and predominance of the clustered appearance soma well out into the neuritic processes in these select emerged by PND10 (Fig. 1C).

392 GABAA receptor development in rNST, W. L. Heck et al.

Fig. 2 Histogram of the mean density (darkness of label per unit area) of GABAA receptor immunoreactivity in rNST neurones

during postnatal development. Somatic staining for GABAA receptors is highest at PND1, decreases to PND5 (P < 0.0001) and increases again to be maintained at adult levels by PND10 (P < 0.0001). This graph could be contrasted with Fig. 5 in Heck et al.

(2001), the analogous graph for GABAB receptors in the rat rNST. The same pattern is seen; however, GABAB receptor

development is delayed compared to GABAA. For GABAB receptors the decrease in overall immunoreactivity occurs between

PND10 and PND15 and the increase to adult levels occurs after weaning (PND20). GABAA receptors mature earlier than GABAB receptors. Vertical bars indicate the standard errors of the mean. Horizontal arrows with stars indicate significant differences between the groups as determined with Fisher post hoc PLSD tests.

These qualitative observations were reinforced by was excellent at both ages and labelling was easily quantitative density measurements. There was an detected. overall significant pattern of change in the density Electron micrographs from adult rNST showed excel-

(darkness) of GABAA receptor immunoreactivity in lent preservation and easily detectable labelling of the

individual rNST neurones depending on age (Fig. 2; GABAA receptor antibody and the methods used F = 5.743, d.f. = 5,234, P < 0.0001). The density of (Fig. 3A–C). Very little background label is seen, and

GABAA receptor immunoreactivity was actually the only select, presumably GABAergic, synapses are labelled. highest at PND1, followed by a significant decrease in The morphology of these labelled synapses is consistent

the density of GABAA receptor immunoreactivity from with previous descriptions of GABAergic synapses in PND1 to PND5. This drop occurred at the time of the the rNST including their vesicles of varying shapes clustering of label on the EM level. From PND5 to PND10, (Leonard et al. 1999).

there was a significant increase in the density of GABAA There was clearly more colloidal gold label for

receptor immunoreactivity to its adult level. Between GABAA receptors present in synaptic interfaces than in PND10 and adulthood, there were no significant the general neuropil in both age groups (Table 1).

changes in the density of GABAA receptor immunoreac- One-way ANOVAs confirmed that in the adult there was

tivity (also compare Fig. 1C and D). significantly more GABAA receptor immunoreactivity in

Table 1 Synaptic and extrasynaptic label Electron microscopy No. particles The ultrastructural localization of GABA receptors A (per µm synaptic in the rNST of adult and PND10 Sprague-Dawley rats thickening) Adult PND10 was studied. PND10 was chosen to be consistent ± ± with our earlier paper on the distribution of GABA Synaptic 19.92 0.5 0.73 12.24 0.5 0.65 B Extrasynaptic 0.74 ± 0.5 0.05 0.37 ± 0.5 0.03 receptors (Heck et al. 2001). Ultrastructural preservation

Fig. 3 Electron micrographs of GABAA receptor labelling in the adult (A,B,C) and PND10 (D,E) rat rNST. The ultrastructure is well preserved in this tissue despite the fact that it was fixed to preserve immunogenicity. Background label is very scarce and labelled synapses (arrows) are easy to detect. Label is found either in the cleft or in association with the post-synaptic density. The targets of synapses shown are a large dendrite (A) and two small dendrites (B,C). Even at PND10, the ultrastructure is well preserved (D). Labelled synapses appear very mature, although the density of vesicles is somewhat low and the post-synaptic targets are quite large (and are probably quite proximal to the cell body, see text). Panel E shows a clathrin-coated vesicle pinching off near the synapse. Such structures have been noted before in association with post-synaptic membranes of developing GABAergic terminals (see Brown et al. 2000). synaptic profiles (usually 19.92 or more colloidal gold P < 0.0001). The distributions of label densities for particles per micrometre, Table 1) than in nearby synaptic and extrasynaptic profiles did not overlap. extrasynaptic areas (usually 0.74 or less colloidal The lowest density of label at synapses was 7.51 gold gold particles per micrometre, F = 689.44; d.f. = 1,30; particles per micrometre and the highest density of

label in adjacent extrasynaptic neuropil was 1.12 gold In contrast to GABAB receptors, overall GABAA receptor particles per micrometre. Therefore, structures in the immunogenicity drops earlier (between PND1 and

adult rNST were considered ‘labelled for GABAA receptors’ PND5). Like GABAB receptors, this drop is also associ- if they had 3 or more colloidal gold particles per ated with the clustering of immunoreactivity in the micrometre. neurones (although a few select neurones seem to In PND10 rats, there was also significantly more have this clustered appearance even at PND1 for

GABAA receptor immunoreactivity in the synaptic GABAA receptors). For both types of receptors, clustering profiles (usually 12.24 or more colloidal gold articles and a reduction in overall immunolabelling occur per micrometre, Table 1) than in nearby extrasynaptic simultaneously; however, both of these phenomena

areas (usually 0.37 or fewer colloidal gold particles per occur earlier for GABAA receptors. micrometre, F = 331.658; d.f. = 1,30; P < 0.0001). Again, The electron microscopic data confirm and extend

the distributions were non-overlapping (synaptic these results. Unlike GABAB receptors where much minimum = 6.36 gold particles per micrometre and extrasynaptic label was seen at PND10 with regression extrasynaptic maximum = 1.01 gold particles per to synaptic sites thereafter, extrasynaptic label was

micrometre) and to be consistent, structures in the virtually never seen with the GABAA receptor antibody,

PND10 rNST were considered to be ‘labelled for GABAA suggesting that by PND10, clustered receptors had receptors’ if they contained 3 or more colloidal gold already selectively associated themselves with synapses. particles per micrometre. There was no extrasynaptic In fact, the only indication of immaturity noted label even at PND10, the youngest age analysed. between PND10 and adults was that labelled synapses Electron micrographs reveal that synapses labelled at PND10 were most often associated with large even at PND10 were remarkably mature (Fig. 3D,E). dendritic profiles rather than the smaller post-synaptic However, a few immaturities were noted, including targets of these terminals at a later age. This could be their association only with large dendritic profiles and coincidental to their location (on proximal dendrites the appearance of clathrin-coated pits associated with close to the cell soma which are large in diameter) or a some synapses (Fig. 3E). Clathrin-coated pits were developmental phenomenon since most distal dendrites noted in association with the post-synaptic membranes that are smaller in diameter have yet to develop at this of very immature synapses in our study of the morpho- age (Renehan et al. 1997; Lasiter et al. 1989). logical development of GABAergic synapses (Brown Clustering of receptors into functional synapses has et al. 2000) and are thought to be involved in recycling been the subject of much recent research (reviewed by of membranes and endocytosis during synaptogenesis. Brandon et al. 1999). There is evidence that the clustering of GABA receptors is not due to the action of GABA itself (Kannenberg et al. 1999). Rather, specific mole- Discussion cules are selectively associated with clustered GABAA These results suggest an expedited developmental receptors such as gephyrin and GABARAP. Anti-sense

course for GABAA receptors in the rNST compared to oligonucleotides that disrupt gephyrin expression

that for GABAB receptors (Heck et al. 2001). Clustering disrupt clustering of GABAA receptors in cultured cortical

of GABAB receptors occurred between PND10 and 15 neurones (Essrich et al. 1998). In addition, chemical

with a concomitant drop of overall immunoreactivity disruption of gephyrin leads to a reduction in GABAA density (see Fig. 5 in Heck et al. 2001). Prior to PND10, clusters (Kneussel et al. 1999), the proposed mechanism

the amount of extrasynaptic GABAB receptor labelling of gephyrin being an anchoring mechanism with the was significant, suggesting a non-synaptic function of post-synaptic cytoskeleton. In contrast, there is evidence

GABA working at these receptors. GABA has been that GABAC receptors are clustered by MAP-1B implicated in developmental phenomenon other than (Tachibana & Kaneko, 1987; Koulen et al. 1998; Hanley neurotransmission (McLaughlin et al. 1975; Taylor et al. et al. 1999), suggesting that specific GABA receptor 1990; Belhage et al. 1998; Lauder et al. 1998) and we subtypes are clustered by different molecules. The time would speculate here that these developmental roles course of expression and presence of these different

may be played out by GABAB receptors in extrasynaptic clustering molecules could serve as the substrate for

regions of the neurone and that GABAA receptors may the separate time courses revealed in our studies for

take on an earlier synaptic role. GABAA and GABAB receptors.

This early maturation of the synaptic function of Bradley RM, King MS, Wang L, Shu X (1996) Neurotransmitter and neuromodulator activity in the gustatory zone of the GABAA receptors is consistent with physiological nucleus tractus solitarius. Chem. Senses 21, 377–385. findings. Early sensitivity to the GABA receptor antag- A Brandon NJ, Bedford FK, Connolly CN, Couve A, Kittler JT, onists bicuculline (Grabauskas & Bradley, 1999) and Hanley JG, et al. (1999) Synaptic targeting and regulation picrotoxin (Du & Bradley, 1998) and a lack of sensitivity of GABA (A) receptors. Biochem. Soc. Transactions 27, 527– 530. to GABAB receptor antagonists confirms a precocious development of synaptic GABA receptors. Brown M, Renehan WE, Schweitzer L (2000) Changes in A GABA-immunoreactivity during development of the rostral There is evidence that GABA itself stimulates formation subdivision of the nucleus of the solitary tract (rNST). Neu- (as apposed to clustering) of GABA receptors (Schousboe roscience 100, 849–859. & Redburn, 1995) and the data presented here are Du J, Bradley RM (1998) Effects of GABA on acutely isolated consistent with that. Early in development there is an neurons from the gustatory zone of the rat nucleus of the solitary tract. Chem. Senses 23, 683–688. abundance of GABA in the rat brainstem taste nucleus Essrich C, Lorez M, Benson JA, Fritschy JM, Luscher B (1998) rNST, but only few synapses and even fewer GABAergic Postsynaptic clustering of major GABAA receptor subtypes synapses (Brown et al. 2000). GABAergic synapses rise requires the gamma 2 subunit and gephyrin. Nature Neuro- science 1, 541–543. to adult levels by PND10. Both GABAA and GABAB Grabauskas G, Bradley RM (1999) Potentiation of GABAergic receptors are present at birth, but assemble themselves synaptic transmission in the rostral nucleus of the solitary into synapses after birth. Shortly following the birth tract. Neuroscience 94, 1173–1182. of the rat, GABAA receptors begin clustering at the Grabauskas G, Bradley RM (2001) Postnatal development of inhibitory synaptic transmission in the rostral nucleus of the developing GABAergic synapses, while GABAB receptors remain dispersed through the cell. Later, after PND10, solitary tract. J. Neurophysiol. 85, 2203–2212. Hall ZW, Sanes JR (1993) Synaptic structure and development: GABA receptors also assume this more restricted B the neuromuscular junction. Neuron 10, 99–121. synaptic distribution. This pattern of results reinforces Hanley JG, Koulen P, Bedford F, Gordon-Weeks PR, Moss SJ the concept that GABA has a developmental function (1999) The protein MAP-1B links GABA (C) receptors to the cytoskeleton at retinal synapses. Nature 397, 66–69. outside of neurotransmission and that GABAA receptors Hayat MA (1981) Principles and Techniques of Electron Micro- are the predominant receptors for neurotransmission scopy: Biological Applications, Vol. 1, 2nd edn. Baltimore: even in the youngest infant rats as they are thought to University Park Press. be in the adults. Heck WL, Renehan WE, Schweitzer L (2001) Redistribution and

increased specificity of GABAB receptors during development of the rostral nucleus of the solitary tract. Int. J. Dev. Acknowledgments Neuroscience 19, 503–515. Heck WL, Slusarczyk A, Basaraba AM, Schweitzer L (2002) We would like to thank Ms. Betsey Wiegman for help Subcellular localization of GABA receptors in the central with the data and statistical analyses. This research was nervous system using post-embedding immunohistochemis- supported by NIH Grant #DC01074 to L. S. W. H. com- try. Brain Res. Protocols 9, 173–180. Kannenberg K, Sieghart W, Reuter H (1999) Clusters of pleted the majority of the study while a graduate GABAA receptors on cultured hippocampal cells correlate student and then postdoctoral fellow at the University only partially with functional synapses. Eur. J. Neurosci. 11, of Louisville. The writing was completed when she 1256–1264. moved to become a fellow in the Seeding Postdoctoral Kim M, Bradley RM, Mistretta CM (1997) Effects of GABA on developing neurons in rat gustatory nucleus of solitary Innovators in Research and Education (SPIRE) program tract. Int. Symp Olfaction Taste XII Achems XIX, 15. supported by the Minority Opportunities in Research Kneussel M, Brandstatter JH, Laube B, Stahl S, Muller U,

Division of the National Institute of General Medical Betz H (1999) Loss of postsynaptic GABAA receptor clus- Sciences (Grant #GM000678). tering in gephyrin-deficient mice. J. Neurosci. 19, 9289– 9297. Koulen P, Brandstatter JH, Enz R, Bormann J, Wassle H (1998) References Synaptic clustering of GABA (C) receptor rho-subunits in the rat retina. Eur. J. Neurosci. 10, 115–127. Belhage B, Hansen GH, Elster L, Schousboe A (1998) Effects Lasiter PS, Kachele DL (1988) Organization of GABA and of GABA on synaptogenesis and synaptic function. Perspectives GABA-transaminase containing neurons in the gustatory Dev. Neurobiol. 5, 235–246. zone of the nucleus of the solitary tract. Brain Res. Bull. 21, Bevan S, Steinbach JH (1977) The distribution of α-bungarotoxin 623–636. binding sites on mammalian skeletal muscle developing Lasiter PS, Kachele DL (1989) Postnatal development of in vivo. J. Physiol. 267, 195–213. protein P38 (‘synaptophysin’) immunoreactivity in pontine

and medullary gustatory zones of rat. Dev. Brain Res. 4, amino acid neurotransmitters. J. Histochem. Cytochem. 40, 27–33. 1011–1020. Lasiter PS, Wong DM, Kachele DL (1989) Postnatal develop- Renehan WE, Massey J, Zhigao J, Zhang X, Liu Y-Z, Schweitzer L ment of the rostral solitary nucleus in rat: dendritic mor- (1997) Developmental changes in the dendritic architecture phology and mitochondrial enzyme activity. Brain Res. Bull. of salt-sensitive neurons in the nucleus of the solitary tract. 22, 313–321. Dev. Brain Res. 102, 231–246. Lauder JM, Liu J, Devaud L, Morrow LA (1998) GABA as a Schousboe A, Redburn DA (1995) Modulatory actions of trophic factor for developing monoamine neurons. Perspec- gamma aminobutyric acid (GABA) on GABA type A receptor tives Dev. Neurobiol. 5, 247–259. subunit expression and function. J. Neurosci. Res. 41, 1–7. Leonard NL, Renehan WE, Schweitzer L (1999) Structure Smith DV, Li C-S (1998) Tonic GABAergic inhibition of taste- and function of gustatory neurons in the nucleus of the responsive neurons in the nucleus of the solitary tract. solitary tract. IV. The morphology and synaptology of GABA- Chem. Senses 23, 159–169. immunoreactive terminals. Neuroscience 92, 151–162. Smith DV, Li C-S (2000) GABA-mediated corticofugal inhibition Liu H, Behbehani MM, Smith DV (1993) The influence of GABA of taste-responsive neurons in the nucleus of the solitary of on cells in the gustatory region of the hamster solitary tract. Brain Res. 858, 408–415. nucleus. Chem. Senses 18, 285–305. Tachibana M, Kaneko A (1987) Gamma-aminobutyric acid McLaughlin BJ, Wood JG, Saito K, Roberts E, Wu J (1975) The exerts a local inhibitory action on the axon terminal of bipolar fine structural localization of glutamate decarboxylase in cells: evidence for negative feedback from amacrine cells. developing axonal processes and presynaptic terminals of Proc. Natl Acad. Sci. USA 84, 3501–3505. rodent cerebellum. Brain Res. 85, 355–371. Taylor J, Docherty M, Gordon-Weeks PR (1990) GABAergic Nakagawa T, Wakamori M, Shirasaki T, Nakaye T, Akaike N growth cones: Release of endogenous gamma-aminobutyric (1991) Gamma-aminobutyric acid-induced response in acid precedes the expression of synaptic vesicle antigens. J. acutely isolated nucleus solitarii neurons of the rat. Am. J. Neurochem. 54, 1689–1699. Physiol. 260, C745–C749. Terai K, Tooyama I, Kimura H (1998) Immunohistochemical Paxinos G, Ashwell KWS, Tork I (1994) Atlas of the Developing localization of GABA receptors in comparison with GABA- Rat Nervous System, 2nd ed. San Diego: Academic Press. immunoreactive structures in the nucleus tractus solitarii of Paxinos G, Watson C (1986) The Rat Brain in Stereotaxic Coor- the rat. Neuroscience 82, 843–852. dinates, 2nd edn. San Diego: Academic Press. Wang L, Bradley RM (1993) Influence of GABA on neurons of Phend KD, Weinberg RJ, Rustioni A (1992) Techniques to the gustatory zone of the rat nucleus of the solitary tract. optimize post-embedding single and double staining for Brain Res. 616, 144–153.