The contrast between modalities 1 and 2 is particularly striking because both have been seen with recombinant expression of GPR55 in HEK293 cells. There are now multiple examples in the superfamily of GPCRs where ligand specificity is modified by interaction with other proteins. For example, Family B GPCRs can associate with single transmembrane proteins known as receptor activity modifying proteins (RAMPs). In the case of the calcitonin receptor, association with RAMP1 confers the ability to be activated by amylin (Hay et al., 2006). Within Family A, several pairs of GPCRs have been shown to heterodimerize with resultant effects on ligand specificity (Milligan, 2007). Thus the A-opioid receptor and the sensory neuron-specific receptor-4 (SNSR-4), both of which are expressed in dorsal root ganglion cells, appear to heterodimerize when recombinants are coex-pressed. In coexpression, SNSR-4 signaling is retained, but that to A-opioid receptors are masked, and selective agonists for the latter receptor fail to activate the dimer (Breit et al., 2006). Perhaps GPR55 in a similar fashion interacts with a secondary factor that modifies its ligand-binding profile.
The correlation between ligand specificity in yeast, where GPR55 is isolated from other mammalian proteins, with that described by Henstridge et al. (2009) in GPR55-HEK293 cells suggests interaction with AM251, LPI, and CP55,940 (Modality 1) is an intrinsic property of GPR55. By inference, interaction with the wider array of cannabinoids (Modality 2) could result from the interaction of GPR55 with a secondary modifying factor. The defining property of this factor would be that it confers on GPR55 affinity for diverse cannabinoid ligands. Given the propensity of GPCRs to heterodimerize we should consider the possibility that cannabi-noid receptors combine with GPR55 to yield the specificity observed by Ryberg et al. (2007). This is attractive because several of the ligands differing between modalities 1 and 2 are ofcourse validated ligands at CB1 receptors, for example, anandamide. Moreover, this hypothesis is testable by combinatorial expression of CB1 with GPR55. Signaling of CB1 in this notional GPR55/CB1 heterodimer is presumably masked, similar to the A-opioid receptor/SNSR-4 heterodimer, since Ryberg et al. (2007) observed no activation of Gai, and WIN55,212-2 was inactive. HEK293 cells, used to express GPR55 by both Ryberg et al. (2007) and Johns et al. (2007), are not generally regarded as having endogenous cannabinoid pharmacology, so contingent on this hypothesis would be the ability of GPR55 to upregulate expression ofCB1 receptors. A clue about GPR55-interacting factors may be contained in the fascinating observation that antibodies directed against fi1 or b3 integrins, or the integrin ligand fibronectin, can block the anandamide-evoked Ca2+ response in endothelial cells (Waldeck-Weiermair et al., 2008). If this Ca2+ response is indeed mediated by GPR55, this raises immediate questions of whether the GPR55-integrin interaction involves direct protein— protein contact and, if so, is it sufficient to modify the ligand specificity of GPR55.
Was this article helpful?