**6. Discussion**

198 Biochemistry

Fig. 7. Bar graphs comparing the relative distributions of A1R(A1)/P2Y1R(Y1)

the mean of three independent experiments.

immunoreactive elements in Purkinje cells (A) Hippocampal pyramidal cells (B), and A1R(A1)/P2Y2R(Y2) immunoreactive elements in Purkinje cells (C), Hippocampal pyramidal cells (D) and Cortical neurons (E). P2YR-P2YR, A1R-A1R and A1R-P2YR oligomers are indicated by Y1-Y1 (Y2-Y2), A1-A1 and A1-Y1 (A1-Y2), respectively. The total number of immunoreactive gold particles on the cell surface was defined as 100%. Each column represents the average frequency (± SD) from three cells. Data describing the average numbers of gold particles are shown in tables under the graphs. Data represent

Previous reports describe electron microscopic studies of plasma membranes for homooligomeric B1 Bradykinin receptor complexes (Kang *et al*., 2005), heteromericoligomerization of GABAB R1 and R2 receptors (Charara *et al*., 2004), and the localization of A1R with caveolin-3 in rat ventricular cardiomyocytes (Lasley *et al*., 2000). An immunological study suggested that A1R forms oligomers the cortex of the pig brain (Ciruela *et al*., 1995), and a FRET study demonstrated the oligomerization of P2Y2R in transfected HEK293 cells (Kotevic *et al*., 2005). The hetero-oligomerization of A1R-P2Y1R on postsynaptic neurons was also analyzed by IEM (Tonazzini *et al*., 2007). The present study provides the first detailed evidence of an interaction between endogenous A1R and P2Y2R in brains using IEM.

The homo- oligomerization of A1R and its structural profile were previously analyzed in our laboratory by computational prediction, co-immunoprecipitation, and BRET analysis with differently tagged A1Rs (Suzuki *et al*., 2009); homo-oligomers and monomers were easily distinguished by IEM. This particular study confirmed the existence of homo-oligomers (A1R-A1R and P2Y2R-P2Y2R) using IEM. Interestingly, the percentage of A1R homooligomers was higher than that of P2Y2R in both rat brain and transfected HEK293T cells (Namba *et al*., 2010). By contrast, the ratio of heteromeric gold-particle clusters were different in the cortex, hippocampus, and cerebellum. Importantly, both homo-oligomeric and hetero-oligomeric gold-particles were reduced in number at inner cytoplasmic membranes than at the cell surface (data not shown). In general, most GPCRs oligomers have been observed at the cell surface (Minneman, 2007; Bulenger *et al*., 2005).

While the frequencies of A1R and P2Y1R homo-oligomers and monomers were similar in the cerebellum (Fig. 5) and in transfected HEK293T cells (Fig. 4B), the ratio of the different receptor oligomers occurred in different patterns in each of the three brain areas (Fig. 7). Total numbers of hetero-oligomers observed on the cell surface and in the cytoplasm were clearly different (Fig. 6A, B) and may reflect the process of receptor maturation and association of the A1R-P2YR complex. However, hetero-oligomers were unmistakably detected at the cell surface by IEM (Fig. 3C, Fig. 4B).

As a signaling pathway, P2Y1R and P2Y2R display different ligand specificities. As ligands, ATP and UTP fully activate P2Y2R. However, UTP is not an agonist for P2Y1R. In addition, ADP is a strong agonist for P2Y1R but not P2Y2R (Abbracchio *et al*., 2006). Many previous studies suggest that A1R-P2Y1R and A1R-P2Y2R hetero-oligomers exhibit general pharmacological profiles, possibly because of differences in the conformational changes induced by oligomerization (Nakata *et al*., 2010). The hetero-oligomerization of A1R-P2Y1R inhibits adenylyl cyclase activity via the Gi/o protein linked effector. The heterooligomerization of A1R-P2Y2R resulted in an increase in intracellular Ca2+ levels induced by P2Y2R activation of Gq/11 which was synergistically enhanced by the simultaneous addition of an A1R agonist in the co-expressing cells (Suzuki *et al*., 2006). Differences in the amounts of hetero-oligomerization between A1R-P2Y1R and A1R-P2Y2R were observed (Fig. 7). Assuming that the number of hetero-oligomers formed is functionally dominant, the dominancy of the signaling via A1R-P2Y1R may be generated by competitive antagonism in pharmacology between P2Y1R and P2Y2R in order to oligomerize with A1R. This hypothesis, however, requires further investigation.

In our previous study, the hippocampal hetero-oligomerization of A1R and P2Y2R was far more pronounced than in other regions of the brain (Namba *et al*., 2010). Another research group suggested that the hetero-oligomerization, or cross-talk between A1R and P2Y1R is involved in regulation of glutamate release in the hippocampus (Tonazzini *et al*., 2007). The relative distributions of immunoreactivity for GABAB R2 and GABAB R1 were also different in the basal ganglia and globus pallidus/substantia nigra, which suggests the possible coexistence and hetero-oligomerization of the two types of receptors at various pre- /postsynaptic sites (Charara *et al*., 2004). From the present study, it can be speculated that the A1R/P2Y2R hetero-oligomer might be responsible for down regulation, via hippocampal Ca2+ secretion, of synaptic functions (Safiulina *et al*., 2006). Furthermore, the abundant formation of A1R/P2Y1R or A1R/P2Y2R hetero-oligomers in the cerebellum revealed in this present study supports the idea that the unique signal transduction generated by heterooligomerization, including the enhancement of Ca2+ signaling via Gq/11, observed in transfected cells, also occurs in the cerebellum.

There are many families of GPCRs expressed in whole brain, most of which remain a mystery. However, it is clear that GPCR hetero-oligomerization is common in the brain and exhibits unique pharmacology in this region, thus implying that associated signal transduction pathways can be anticipated in this region. The methodology described here using immunogolod particles is one of the most influential techniques available to elucidate the ingenious mechanism underlying GPCR hetero-oligomerization.
