Homecoming of the estranged GluD channels

Abstract

A recent paper by Carrillo and colleagues demonstrates that GluD proteins form bona fide ligand-gated ion channels when their intrinsic flexibility is constrained by interactions with protein partners. Therefore, Delta receptors resemble all other members of the ionotropic glutamate receptor family not only by sequence and structural homology, but also by functional dynamics.

The genomic structure, expression, and three-dimensional architecture of Delta receptors have been described in detail [1,2]. In contrast, their operation and roles in synaptic physiology have remained a matter of debate for several decades [2,3]. Particularly puzzling was the fact that although Delta receptors share salient structural and functional features with the other sixteen members of the ionotropic glutamate receptor (iGluR) family, they appeared to lack the class-defining ligand-gated ionotropy [3]. In a recent paper, Carrillo and colleagues have shown that Delta receptors function as amino acid-gated excitatory channels when their conformational flexibility is restricted by endogenous interacting proteins [4]. This observation reconciles a large number of molecular, structural, and functional findings and provides a coherent foundation necessary to further dissect the many roles of GluD receptors in health and disease.

Two GluD proteins, GluD1 and GluD2, were identified in the mid-1990s by homology screening of mouse brain libraries with probes derived from iGluR proteins [1]. Their substantial sequence homology with iGluR subunits indicates common evolutionary history, and suggests similar molecular structures and signaling mechanisms [5]. The other members of the family, GluA(1–4), GluK(1–6), and GluN(1–3), form the better-known AMPA, Kainate, and NMDA receptors, respectively (Figure 1 top). Like these relatives, Delta receptors are tetrameric transmembrane proteins and adopt similar architectures. The Delta receptor ectodomain consists of the homologous globular domains of the four component protomers, the membrane-distal N-terminal domains (NTD), and the membrane-proximal ligand-binding domains (LBD). The LBDs of Delta receptors contain high-affinity (low micromolar) ligand-binding sites specific for glycine (or D-serine) and connect directly to the trans-membrane domain (TMD), which encloses a cation-permeable pore [1]. Lastly, the TMD continues in the cytoplasm with an unstructured C-terminal domain (CTD) that anchors the protein in the postsynaptic density and hosts recognition sites for cellular signaling pathways [3].

Figure 1. Structure and ligand-gated ionotropic function of Delta receptors.

Top: Members of the mammalian iGluR family (left), and overall architecture of GluD receptors (right). GluD subunits share sequence homology, membrane topology, and three-dimensional architecture with members of the iGluR family [1]. Bottom: In a recent study, Carrillo et al. [4] demonstrated that GluD2 receptors are ligand-gated excitatory channels. From left to right: glycine (or D-serine) does not gate currents from wild-type GluD2 receptors; GluD2 receptors carrying the Lurcher mutation (A654T) have constitutively open pores that leak excitatory currents and can be blocked with pentamidine; glycine gates excitatory currents from GluD2 receptors whose NTDs are stabilized by chemical cross-linking across the NTDs; glycine gates excitatory currents from wild-type GluD2 receptors embedded in a trans-cellular complex with NRX1 and Clb1. This figure was created using BioRender.com.

The LBDs of all iGluR subunits resemble the Venus flytrap in shape and operation. A series of small diffusible ligands can bind in the cleft formed by two hinged lobes and change the thermodynamics and overall shape of this globular domain. Ligands that induce cleft closure stabilize less dynamic, more compact conformations, and can initiate intramolecular rearrangements that culminate with opening the channel gate to allow ionic flux through the transmembrane pore. In Delta receptors, glycine or D-serine bind in the LBD cleft and stabilize a more compact conformation. However, this movement alone is not sufficient to produce ionic flux [6,7]. For this reason they are often referred to as ‘orphan’ receptors, implicitly assuming that their ionotropic function may be gated by a yet unknown diffusible agonist [8].

It is also possible that during evolution, GluD receptors have lost the ligand-dependent gating function altogether, and like metabotropic receptors, they transduce the energy gained by binding glycine solely as conformational change, with no role for the ion permeable pore. Alternatively, GluD receptors may simply serve a structural role that is altogether independent of their ability to bind glycine. This latter hypothesis builds on a wealth of evidence that in the cerebellum, postsynaptic Delta receptors form a ternary trans-synaptic complex with the secreted protein cerebellin 1 (Cbln1), and with the presynaptic transmembrane protein neurexin (NRX). Notably, the GluD-Cbln1-NRX complex is required for the normal development and function of cerebellar synapses [2]. Carrillo and colleagues integrate all three hypotheses by demonstrating that when assembled in a trans-cellular GluD2-Cbln1-NRX ternary complex, Delta receptors function as glycine-gated excitatory channels.

In support of this unifying concept, Carrillo and colleagues show that glycine gates excitatory currents from interconnected cells that co-express recombinant GluD2, Cbln1, and NRX. In contrast, glycine fails to gate currents in cells that lack either of these proteins or in cells that are lifted from the cellular lawn, and thus mechanically separated from neighboring cells (Figure 1, bottom). To our knowledge, this is the first report of ionic currents elicited by glycine from wild-type GluD receptors. Importantly, the results show pharmacological and biophysical properties consistent with the existing literature on GluD proteins. The half-maximal effective concentrations for glycine and D-serine align well with results from binding studies, and from functional assays with competitive and non-competitive inhibitors [6,7]. Similarly, the unitary channel conductance measured here in wild-type receptors matches well the values reported for constitutively active GluD variants [9].

The recent work by Carrillo and colleagues is notable in the mechanistic insight it offers. The authors propose that in wild-type GluD2 proteins, glycine-dependent gating is prevented by the intrinsic flexibility of NTDs, and transcellular tethering permits ligand-induced ionotropy by restricting the conformational freedom of the NTDs. The authors support this scenario with three lines of evidence. First, fluorescence resonance energy transfer measurements show a more compact conformation of the NTD in cells that co-express Cbln1 and NRX, relative to cells in which one of the partners is omitted. Second, chemical crosslinking of the NTDs confers glycinergic ionotropy onto recombinant GluD receptors and as expected, the channel open probability increases with the decreasing length of the covalent bridge. Lastly, GluD proteins that lack the entire NTD layer have the highest open channel probability, thus demonstrating that the NTD layer exerts an inhibitory effect onto the ionotropic function of GluD2 receptors.

Given that mechanical stability conferred by intercellular connections is critical for revealing the ionotropic function of GluD2 receptors, it is highly likely that interactions with cytosolic proteins will also prove influential in modulating their glycinergic currents. The area of mechanobiology, including mechanotransduction in the CNS, is experiencing a growth spurt, and GluD receptors appear well-positioned and well-endowed to serve as mechanotransducers.

In light of the profound effects of auxiliary proteins on the ionotropic function of both AMPA and kainate receptors, the fact that GluD receptors require endogenous partners (Cbln1, NRX) should not be surprising [10]. Yet, this new clarity and the demonstrated broad expression of GluD proteins in neurons and in non-neuronal cells throughout the body is sure to ignite a race to identify tissue-specific regulatory partners, endogenous modulators, and biological roles; to reveal and examine the biophysical properties and structural determinants of the GluD-mediated glycinergic currents; and to explore GluD receptors as potential novel therapeutic targets of neurological disease.