A rebellious last-born

NMDA (N-methyl-d-aspartate) receptors are heteromeric, Ca²⁺-permeable channels that require two agonists for their activation — glutamate and glycine. Glycine binds to the so-called NR1 subunit, whereas glutamate binds to any of four different NR2 subunits — NR2A–D. In 1995, a new subunit, NR3A, was independently cloned by Ciabarra et al. and Lipton and colleagues. The function of NR3A is not fully clear, although we know that its co-expression with other NMDA receptor subunits modifies the channel properties. Now Lipton and colleagues have identified NR3B, a new NR3 subunit that, judging by the human genome, is probably the final member of the NMDA receptor family. NR3B sheds some light on the function of the NR3 subunits, but opens just as many intriguing questions.

The main peculiarity of NR3B is that it can form heteromers with NR1, giving rise to cation-selective channels that need only glycine for their activation. Although glycine has its own set of receptors, they are selective for anions and have pharmacological properties that are quite different from the NR3B-containing heteromers. In fact, the pharmacology of NR3B is also atypical in the context of NMDA receptors. So, NR3B is unresponsive to glutamate agonists and antagonists and, although it recognizes NR1-selective glycine agonists, these drugs behave as NR3B antagonists. Moreover, NR3B-containing receptors are not permeable to Ca²⁺ and show little of the Mg²⁺-dependent blockade that is characteristic of NMDA receptors. Last, the authors revisited the properties of the previously identified NR3 subunit and found that it could also form glutamate-insensitive, glycine-gated channels when co-expressed with NR1.

What is the functional significance of NR3B in vivo? As a first step, the authors found that glycine can evoke action potentials in cultured neurons, an effect that shared, to some degree, the pharmacological profile of the NR1/NR3A heteromers. But as NR3B is expressed in ventral horn motor neurons, it will be interesting to test directly whether it participates in transmission in the spinal cord, a region in which glycine has long been known to have synaptic actions.