The canonical binding sites to which or contribute are
The canonical RG 7204 synthesis to which α2, α3, or α5 contribute are highly similar. Therefore, differences in ligand affinity will not be large even if a ligand makes optimal use of the small differences in the pockets. As a possible alternative approach to achieve separation of compound effects, ligands with similar binding affinity but different allosteric effects at different receptor subtypes (functionally selective ligands) would be desirable. Ideally, a functionally selective ligand should bind silently (antagonist-like) at all subtypes except for one, where it should exert positive or negative modulatory effects. While this concept has led to some promising results 65, 66, it remains to be explored whether ‘silent’ binding is indeed physiologically silent and will not lead to unwanted long-term adaptive changes in the nervous system. Moreover, structure-guided development of such ligands is made difficult for two reasons: (i) functional assays are much more time-consuming than binding assays, and (ii) structures of the benzodiazepine binding site in the apo- and the positively allosteric stimulated states are not known.
Non-Canonical Binding Sites May Be Interesting Targets As outlined above, in addition to the high-affinity site for benzodiazepines, several non-canonical sites are present at subunit interfaces and in the transmembrane part of the receptor. Owing to the high similarity, for example between γ2 and γ3, or between α2 and α3, their high-affinity benzodiazepine binding sites are not suitable for reliable selective targeting, nor for binding or functional selectivity. The use of other allosteric binding sites might be a valuable alternative. For the extracellular α+/β− binding sites, β1-selective ligands with potency in the nM range have already been reported . The sites in the transmembrane domains generally are less well suited for subtype-selective ligands owing to the much higher sequence and structural conservation in this part of the subunits . However, some specific subunits do feature unique pocket segments in this group of allosteric binding sites. For example, the α2 and α3 subunits, that have nearly identical extracellular plus sides (thus leading to very similar high-affinity benzodiazepine pharmacology), contain distinct M3 and M4 transmembrane segments  that can potentially be targeted individually by appropriate ligands. So far, limited data (which are restricted to diazepam) make it difficult to predict in detail drug action via the non-canonical sites. Mice carrying a point mutation that renders the canonical binding site diazepam-insensitive in all four DS α subunits were studied . These mice were protected from diazepam-induced muscle relaxation and motor impairment. These mice showed, under treatment, reduced locomotor activity that was relatively prominent at higher doses. This indicates that at least part of this residual response to diazepam could be mediated by non-canonical sites. It should be noted that minor effects may easily be missed in behavioral experiments.
Concluding Remarks Third, rational design of drugs with better-defined subtype profiles is urgently needed. While an enormous number of small molecules from ∼100 chemotypes have been identified as ligands of the canonical benzodiazepine binding sites, for the majority of them the subtype profiles are only partially known. In days of high-throughput and big data science, a structured repository and standardized protocols for the determination of full electrophysiological characterization of compound effects on a large panel of receptor isoforms and their respective benzodiazepine binding sites should be feasible. Such data would then assist in identifying the most promising compounds for subsequent lead optimization to obtain ligands with novel or improved subtype preferences. methods of drug discovery such as structure-based pharmacophore screening of libraries, augmented with experimental data , would further accelerate the development from lead compounds to useful research tools or even novel therapeutic drugs.