Role of Heteromer Formation in GABAB Receptor Function

See allHide authors and affiliations

Science  01 Jan 1999:
Vol. 283, Issue 5398, pp. 74-77
DOI: 10.1126/science.283.5398.74


Recently, GBR1, a seven-transmembrane domain protein with high affinity for γ-aminobutyric acid (GABA)B receptor antagonists, was identified. Here, a GBR1-related protein, GBR2, was shown to be coexpressed with GBR1 in many brain regions and to interact with it through a short domain in the carboxyl-terminal cytoplasmic tail. Heterologously expressed GBR2 mediated inhibition of adenylyl cyclase; however, inwardly rectifying potassium channels were activated by GABAB receptor agonists only upon coexpression with GBR1 and GBR2. Thus, the interaction of these receptors appears to be crucial for important physiological effects of GABA and provides a mechanism in receptor signaling pathways that involve a heterotrimeric GTP-binding protein.

GABAB receptors play a critical role in the fine-tuning of central nervous system synaptic transmission (1) and are attractive targets for the treatment of epilepsy, anxiety, depression, cognitive deficits, and nociceptive disorders (2). Their effects are brought about by multiple signaling cascades involving adenylyl cyclase, inwardly rectifying potassium channels (GIRKs), and voltage-dependent Ca2+ channels (1). Recently, the cDNA for a seven-transmembrane domain (7TM) protein, termed GABABreceptor 1 (GBR1), which exists in two NH2-terminal splice forms (A and B) and has high affinity for GABAB receptor antagonists, was identified. GBR1 can account for some, but not all, of the functional properties of native GABAB receptors (3).

We used the yeast two-hybrid system (Y2H) (4) to look for intracellular proteins that mediate signaling events downstream of GBR1 activation. The COOH-terminal intracellular region of GBR1 (amino acids 857 to 960 in GBR1A) (3) was used as a bait (5) to screen a rat brain cDNA library (6). Our search through 2 × 106 recombinant clones yielded five positive clones, all of which encoded overlapping fragments of the COOH-terminus of a previously unidentified protein. Full-length cloning (7) of the identified cDNA revealed an open reading frame for a protein of 940 amino acids with an NH2-terminal 40-residue signal sequence and seven internal hydrophobic segments characteristic for 7TM proteins (Fig. 1A) (8). A public database search for related sequences (8) revealed GBR1B with the best score (36% amino acid sequence identity across 804 amino acids). On the basis of this similarity, we termed the protein GBR2.

Figure 1

Sequence comparison and interaction analysis of GBR proteins. (A) Amino acid sequence (21) alignment of GBR1A, GBR1B, and GBR2. The overall sequence similarities are 25.5% between GBR1A and GBR2, and 29.9% between GBR1B and GBR2. Identical residues are shaded in black, and conserved changes are shaded in gray. The transmembrane domains (TM1 to TM7) are indicated by thin lines. The amino acid sequences responsible for the interaction between GBR1 and GBR2 are indicated by gray bars above (GBR1) and below (GBR2) the sequences. The cDNA sequence of GBR2 was deposited in the GenBank database (accession numberAF109405). A second cDNA clone for GBR2 encoded an additional proline residue at position 19. (B) GBR1 and GBR2 interact through amino acids in their COOH-terminal tails. The bait, two of the preys and further deletion constructs of GBR1 and GBR2 are depicted in alignment with their COOH-terminal regions. The ability of the individual GBR1 constructs to bind to the GBR2 prey 2 (upper panel) and the ability of the individual GBR2 constructs to interact with the GBR1 bait (lower panel) is indicated on the right. The regions common to all constructs scoring positive in the Y2H interaction assay are shaded in gray. They (GBR1Δ7 and GBR2Δ4) were also able to interact with each other. Their location is depicted in a membrane topology scheme of the two GBRs. The constructs were as follows. Bait, amino acids (aa) 857–960; R1Δ1, aa857–911; R1Δ2, aa908–960; R1Δ3, aa857–937; R1Δ4, aa887–960; R1Δ5, aa857–921; R1Δ6, aa898–960; and R1Δ7, aa887–921 of the immature polypeptide GBR1A. Prey1, aa754–940; prey2, aa785–940; R2Δ1, aa785–881; R2Δ2, aa785–837; R2Δ3, aa785–821; R2Δ4, aa785–816; R2Δ5, aa785–797; and R2Δ6, aa798–940 of the immature polypeptide GBR2. (C) The COOH-terminus of GBR2, but not of GBR1, can be pulled down by GST–GBR1-COOH-terminus fusion proteins (10). The left panel is a Coomassie blue–stained gel of the bacterially expressed GST fusion proteins GST (lane1), GST–GBR1-CT (lane 2, GBR1-CT is equivalent to GBR1 bait) and GST-GBR1Δ7 (lane 3) used for pull-down experiments. The right panel shows immunoblots of Flag-tagged COOH-termini of GBR2 (Flag–GBR2-CT, lanes 4 to 7) and GBR1 (Flag–GBR1-CT, lanes 8 to 11) recovered from cytoplasmic extracts of HEK293 cells (lanes 4 and 8) by interaction with GST (lanes 5 and 9), GST-GBR1-CT (lanes 6 and 10), and GST-GBR1Δ7 (lanes 7 and 11). The amount of extracts used for lanes 4 and 8 was 10% of the amount of extracts used for lanes 5 to 7 and lanes 9 to 11.

We mapped the domains mediating the interaction between GBR1 and GBR2. As identified by a series of deletion constructs and subsequent analysis in the Y2H system (4, 5), the interaction was mediated by two short domains centrally located in the intracellular COOH-termini of GBR1 and GBR2 (Fig. 1B). These domains span 35 amino acids in GBR1 and 32 amino acids in GBR2 and are predicted to be α-helical (8). Interaction was found between GBR1 and GBR2 but not between GBR1 molecules or between GBR2 molecules (9), suggesting a requirement for a heteromeric assembly of the two 7TM proteins.

Glutathione S-transferase (GST) pull-down assays from HEK293 cell extracts (10) confirmed the interaction between the COOH-termini of GBR1 and GBR2. Only the GST fusion proteins containing the GBR1–COOH-terminus or the GBR1 heteromerization domain (see GBR1Δ7 in Fig. 1B), but not GST alone, were able to bind to the GBR2–COOH-terminus. No dimerization between GBR1–COOH-termini was detected (Fig. 1C). Thus, the COOH-termini of GBR1 and GBR2 can interact in both yeast cell nuclei and the cytoplasm of mammalian cells.

The GBR2 mRNA was expressed only in the brain (Fig. 2A) (11), as has been described for GBR1 (3). To assess if GBR1 and GBR2 have the potential to interact in the brain, we compared the expression patterns of the mRNAs for the two proteins. As revealed by in situ hybridization (12) in serial rat brain sections, GBR1 and GBR2 are widely expressed and show considerable overlap (for example, in the cerebellum, cortex, and medial habenula, Fig. 2, B to E). Thus, GBR1 and GBR2 have the potential to interact in many neuronal populations. Indeed, the pattern as well as the strength of expression of the mRNAs is largely consistent with the distribution and density of GABAB receptor binding sites in the brain (13). Despite the large overlap in the expression of GBR1 and GBR2, we found spatial and temporal differences. GBR1 mRNA expression was more widespread than that of GBR2 (for example, in the striatum, olfactory bulb, and lateral habenula, Fig. 2, B, D, and E) and had an earlier onset (Fig. 2F). Thus, GABAB receptor isoforms lacking GBR2 may be present in the brain, and the delayed developmental expression of GBR2 may contribute to a maturation of GABAB receptors.

Figure 2

Expression profile of rat GBR1 and GBR2. (A) A multiple tissue Northern (RNA) blot hybridized with a radiolabeled GBR2 cDNA probe revealed a strong signal for a 5.5-kb transcript in brain and a very weak signal (difficult to see on reproduction) for a shorter transcript in testis. (B to F) In situ hybridization of GBR1 and GBR2 mRNA in adult rat brain sections revealed high signal densities of both transcripts in various thalamic nuclei (B), hippocampus (B and E), cerebellar purkinje cells (C), and medial habenula (D). Both mRNA transcripts are moderately expressed in the cerebral cortex (B) and certain anterioventral thalamic nuclei (9) and are undetectable in white matter (9). Expression of GBR1 is enriched in comparison to GBR2 in the striatum (B), the olfactory bulb (B), the septal nuclei (B), the lateral habenula (D), the CA2-CA1 hippocampal subfields [(B and E); arrowheads in (E) indicate the approximate boundary between CA3 and CA2 subfields], and in the neuroepithelial cells of the ventricular zone (indicated by asterisks) of embryonic (day 17) brain (F). On the day of birth, GBR1 and GBR2 are coexpressed in the hippocampus and the thalamic nuclei and diffusely expressed in the cortex, but only GBR1 is detected in the developing cerebellum (9). Sense probes yielded no signal (9). Abbreviations: Br-brain, Bs-brainstem, Cb-cerebellum, Ctx-cortex, Fb-forebrain, G-cerebellar granule cell layer, He-heart, Hi-hippocampus, In-intestine, Li-liver, Lu-lung, M-molecular layer, Mh-medial habenula, Ki-kidney, Lh-lateral habenula, Olf-olfactory bulb, Pu-cerebellar purkinje cells, Se-septal nuclei, Str-striatum, Te-testis, and Th-thalamus.

In keeping with the properties of native GABABreceptors (14), activation of heterologously expressed GBR2 elicited a decrease in forskolin-stimulated adenosine 3′,5′-monophosphate (cAMP) production (Fig. 3B) (15). This decrease, in our hands, was significant in contrast to the effect of GBR1 activation on forskolin-stimulated cAMP production (Fig. 3A). Coexpression of both proteins decreased forskolin-stimulated cAMP production to the same extent as GBR2 alone (Fig. 3C). The GBR2-mediated decrease in cAMP production was sensitive to pertussis toxin, suggesting the involvement of the Gi/Go class of heterotrimeric GTP-binding proteins (G proteins). Thus, heteromeric assembly of GBR1 and GBR2 does not seem to be required for inhibition of adenylyl cyclase.

Figure 3

Coupling of GBR1 and GBR2 expressed in HEK293 cells to effector systems. Forskolin (2 μM) treatment for 20-min stimulated intracellular cAMP concentrations by about a factor of 10 (normalized to 100%, filled bars). Treatment of parallel samples with (R)-baclofen (500 μM, unfilled bars) or GABA (1 mM) (9) attenuated forskolin-stimulated cAMP production in cells transfected with GBR2 alone (B) (P = 0.001) or with GBR1 and GBR2 (C) (P = 0.019), but not in GBR1-transfected cells (A) or in untransfected cells (9). Pretreatment of transfected cells with pertussis toxin (10 ng/ml) for 10 hours abolished the GBR2-mediated decrease in forskolin-stimulated cAMP production (114 ± 22%; P = 0.001) (9). Data are presented as mean ± SEM from at least three experiments performed in quadruplicates and were analyzed by analysis of covariance with post hoc Dunnett's test. (R)-baclofen (50 μM) or GABA (100 μM) (9) did not increase barium (Ba2+, 1 mM)-sensitive GIRK currents (224 ± 46 pA,n = 23, measured at −140 mV) when either (D) GBR1 (a factor of 1.01 ± 0.01 over control,n = 4) or (E) GBR2 (a factor of 0.95 ± 0.05 over control, n = 3) was coexpressed with GIRK1 and GIRK2 in HEK293 cells. Thus, the current traces overlapped in the absence (control) and presence of (R)-baclofen. (F) Upon coexpression of GBR1 and GBR2, (R)-baclofen (50 μM) or GABA (100 μM) (9) reversibly increased GIRK currents by a factor of 2.2 ± 0.3 (n = 13) or 3.3 ± 0.8 (n = 4) over control values, respectively. Pretreatment of transfected cells with pertussis toxin (500 ng/ml) for 24 hours abolished the effect of GBR1 and GBR2 activation on GIRK currents (n = 3) (9). Current traces represent average currents of five voltage ramps (−150 to +5 mV over 300 ms).

A crucial physiological effect mediated by native GABAB receptors is the activation of outward potassium currents through GIRKs (16). Reconstitution of GBR1 or of GBR2 with GIRK1 and GIRK2 in HEK293 cells failed to mediate GIRK activation (Fig. 3, D and E) (17). When coexpressed, GBR1 and GBR2 mediated a robust increase in potassium conductance through GIRK activation (Fig. 3F) in a pertussis toxin–sensitive manner. Thus, the physical interaction between GBR1 and GBR2 appears to be essential for the coupling of GABAB receptors to GIRKs. Given the importance of GIRK activation in the generation of late inhibitory postsynaptic potentials at inhibitory synapses (16), the interaction between GBR1 and GBR2 is likely to play a pivotal role in modulation of neurotransmission.

Thus, we have identified a COOH-terminal interaction between two GABAB receptor proteins. The resultant heteromeric assembly adds an element of complexity to G protein–mediated signaling mechanisms. Monomers of GBR1, of GBR2, and of putative additional members of this receptor family and heteromers thereof may provide a molecular basis for the different pharmacological and functional subtypes of GABAB receptors (1, 18), thereby opening therapeutic avenues.

  • * To whom correspondence should be addressed. E-mail: kornau{at}


Stay Connected to Science

Navigate This Article