Archives
br Discussion Agonist and antagonist action at the
Discussion
Agonist and antagonist action at the GluN1/GluN3 172 3 is known to be complex [4], [10]. Recombinant GluN1/GluN3 receptors produce only small excitatory currents in response to glycine, likely because of a combination of activation via GluN3 and desensitization via GluN1 [2], [3], [11]. Indeed, glycine binding to GluN3 alone seems to be sufficient to activate the receptor [3], suggesting that the GluN1 and GluN3 subunits can undergo independent conformational changes. The antagonist DCKA, which binds preferentially to GluN1 [12], likely potentiates receptor function through an inhibition of desensitization. d-serine typically shows little if any measurable agonist activity [2], [3], probably because the desensitization it produces overwhelms its agonist effect. When incubated with isolated ABDs, glycine and d-serine have significantly higher affinities for GluN3A than for GluN1 (650-fold and 10-fold, respectively; [12]). The reported EC50 for glycine acting on the GluN1/GluN3A receptor expressed in Xenopus oocytes was ∼1μM, with rapid desensitization occurring at 3μM and above. At the 10μM concentrations of glycine and d-serine used in the present study, therefore, we would expect the receptor to exist predominantly in the desensitized state, in which the extracellular domain of the receptor is apparently ∼1nm shorter than in the unstimulated state. Given the measured dissociation constants for DCKA binding to GluN1 and GluN3A (0.54μM and 647μM, respectively), we would expect that DCKA would occupy both GluN1 and GluN3A at 1mM but only GluN1 at 10μM. In both cases, the desensitization caused by glycine and d-serine should be prevented, but at the lower concentration activation of GluN3A should still occur. The ability of DCKA to block the effect of glycine on the height of the extracellular domain of the receptor at both concentrations is consistent with this scenario, since in both cases GluN1 should remain ‘tall’.
The height of the extracellular domain of the GluN1/GluN3A receptor (∼7nm) is about 1nm shorter than that reported recently for the GluN1/GluN2A receptor [5], perhaps because of differences in the ‘flattening’ effect of the AFM probe in the two studies, which were carried out using different microscopes. Further, both heights are considerably shorter than the value recently reported for the GluN1/GluN2A NMDA receptor (∼11nm above the membrane; [13], [14]). We suggest that the receptor might be more flexible under physiological conditions than in the crystalline state, leading to an increase in its malleability by the scanning probe. Nevertheless, the ∼1-nm reduction in height in response to agonist activation was seen with both the GluN1/GluN2A NMDA receptor [5] and the GluN1/GluN3A receptor studied here.
The tendency of GluN1/GluN3A receptor to desensitize rapidly in response to glycine has made it difficult to identify in vivo, and there is still some skepticism about whether or not it has a physiological role. Nevertheless, responses to d-serine have been reported in CNS myelin, and these responses were inhibited by the glycine site antagonist 6-cyano-7-nitroquinoxaline-2,3-dione, but not the glutamate antagonist d-2-amino-5-phosphonovalerate [9]. Further, the responses to d-serine were not seen in NR3A-deficient mice. Although its expression may be restricted, therefore, the GluN1/GluN3A receptor does seem to be functionally active in the CNS. Our results represent the first demonstration of the response of this receptor to activation at the single-molecule level.
Conflict of interest
Acknowledgments
Introduction
The glycine receptor (GlyR), expressed throughout the nervous system, is a member of the Cys-loop ligand-gated ion channel superfamily. Most GlyR in adults are composed of α1β heteromers, but α1 subunits can also form functional homomeric receptors in many expression systems (Kuhse et al., 1995). Glycine is thought to be the endogenous ligand for synaptic GlyR, but evidence exists that taurine tonically activates extrasynaptic GlyR (Mori et al., 2002). Taurine concentrations in mammalian cerebrospinal fluid are typically 10–100μM, and in rats are particularly high in the cerebral cortex, olfactory bulb and cerebellum (Huxtable, 1992); GlyR expression has been noted in all these brain regions (Lynch, 2004). For example, taurine may be acting as an endogenous ligand at GlyR to increase dopamine release in the nucleus accumbens, a brain region implicated in reward (Ericson et al., 2006). Taurine efficacy varies substantially depending on the expression system and GlyR subunit composition. Taurine most often acts as a partial agonist, yielding currents that are 5–60% in magnitude relative to the effects produced by maximally-effective concentrations of glycine. These effects are seen in homomeric α1 and α2 GlyR, and heteromeric α1β and α2β GlyR expressed in basolateral amygdala neurons (McCool and Botting, 2000), HEK 293 cells (Lape et al., 2008), and Xenopus oocytes (Schmieden et al., 1999).