Ubiquitin regulation of neuronal excitability

Abstract

Demonstrating the common mechanism of proteasome-dependent degradation of ion channels, two studies in this issue of Nature Neuroscience show that ubiquitin-dependent protein degradation can modulate neuronal excitability.

Protein modification by ubiquitin and subsequent proteasomal degradation has emerged as a key regulator of neuronal activity¹. However, many of the specific proteins targeted and the underlying molecular mechanisms involved in this process have not yet been determined. Two papers in this issue of Nature Neuroscience report direct ubiquitination-mediated degradation of two types of ion channels critical for neuronal function, namely the L-type voltage-gated calcium channel subunit Cav1.2 and the AMPA receptor (AMPAR) subunit GluR1 (Fig. 1). These papers provide important new insights into how neuronal excitability is modulated in different cellular compartments by different ubiquitin ligases and highlight the crucial roles of ubiquitination and proteasomal degradation at different stages of the biosynthetic pathway.

Figure 1.

Ubiquitin controls neuronal excitability and synaptic transmission by regulating ion channel stability and trafficking. For AMPARs (top), chronically elevated synaptic activity or direct stimulation with ephrin-A1 activates EphA4 receptors. This leads to recruitment of the Cdh1 component of the multiprotein ubiquitin ligase anaphase-promoting complex (APC) that, in turn, binds to and ubiquitinates the GluR1 subunit of AMPARs. These ubiquitinated AMPARs are targeted for degradation in the proteasome, providing a scaling mechanism to downregulate synaptic responsiveness during prolonged periods of synaptic activity. For L-type calcium channels (bottom), the Cav1.2 pore-forming subunit must assemble with the cytosolic β-subunit for correct surface expression and protein stability. In the absence of Cavβ, Cav1.2 binds to and is ubiquitinated by the ER-associated ubiquitin ligase RFP2 and is targeted for proteasomal degradation by ERAD.

For L-type calcium channels, Altier et al.² demonstrate that the accessory subunit Cavβ is necessary for ‘quality control’ during Cav1.2 channel biosynthesis. In the absence of Cavβ, Cav1.2 channels are ubiquitinated, targeted to the endoplasmic reticulum (ER)-associated protein degradation (ERAD) complex and degraded at the proteasome. For AMPARs, Fu et al.³ describe how chronic elevation of synaptic activity increases ephrin signaling, which reduces synaptic strength by invoking direct ubiquitination and proteasomal degradation of the GluR1 subunit of functional AMPARs. These findings increase understanding of the specific mechanisms that determine the trafficking and stability of two proteins essential for regulating neuronal excitability. In the wider context, they provide important new examples of how ubiquitination and proteasomal degradation can regulate proteins in different cell compartments at different stages of their assembly and maturation to control neuronal excitability and synaptic transmission.

L-type calcium channels comprise a poreforming α-subunit (for example, Cav1.2), a membrane-anchored α2δ-subunit and a cytosolic β-subunit. Although it is well established that the α2δ and β subunits act synergistically to promote surface expression, the underlying mechanisms have not been defined. Now Altier et al.² show that expression of Cav1.2 with Cavβ leads to an increase in both total and surface expression of Cav1.2. In the absence of Cavβ there is less Cav1.2, and what remains is almost exclusively confined to the ER. Thus, Cavβ seems to facilitate ER export and either enhances translation or protects Cav1.2 from degradation.

ERAD is a quality control mechanism in which misfolded membrane proteins are translocated from the ER membrane into the cytosol, ubiquitinated by an ER-resident ubiquitin ligase and degraded by the proteasome⁴. Here Altier et al.² show that Cav1.2 is ubiquitinated by the ER-associated ubiquitin ligase RFP2 but in the presence of Cavβ ubiquitination is reduced. In addition, the ERAD proteins Derlin-1 and p97 coimmunoprecipitate with Cav1.2, and this association is also decreased by Cavβ. Thus, it seems that in the absence of Cavβ the Cav1.2 channel is recognized as a misfolded protein, undergoes RFP2-mediated ubiquitination and is targeted for proteasomal degradation by ERAD. Therefore, the Cavβ subunit is a critical determinant of channel stability and function because it allows Cav1.2 to bypass ERAD and promotes exit from the ER. These findings demonstrate that the interplay between Cavβ and the ubiquitin ligase RFP2 defines a trafficking checkpoint to ensure that only correctly assembled Cav1.2 channels reach the surface. Furthermore, several calcium channel subtypes contain Cavβ subunits, potentially implicating this control mechanism as a general determinant of calcium channel trafficking and stability.

Fu et al. show that AMPAR degradation is controlled by direct ubiquitination of the GluR1 subunit³. Synaptic scaling is the process by which neurons adjust their excitability to compensate for changes in network activity by regulating synaptic AMPARs⁵. One mechanism for reducing synaptic responsiveness upon prolonged elevation of activity involves direct ubiquitination and proteasome-dependent degradation of the AMPAR subunit GluR1 by means of a pathway dependent on the ephrin receptor EphA4 (ref. 3). Eph receptors are receptor tyrosine kinases activated by membrane-bound ephrin ligand proteins expressed on the surface of adjacent cells, allowing the initiation of bidirectional signaling cascades⁶. Although ubiquitination of AMPAR-associated proteins has been reported previously^7,8, the study by Fu and colleagues shows for the first time that direct ubiquitination of GluR1 leads to its proteasomal degradation³.

Fu et al.³ show that ephrin-A1 activation of one of its cognate receptors, EphA4, causes a reduction in AMPAR miniature excitatory postsynaptic current frequency and amplitude, indicating fewer functional synapses and fewer AMPAR at individual synapses, respectively. Consistent with this, a reduction in total and surface GluR1 and a reduction in spine number in response to ephrin-A1 were observed in corresponding biochemical experiments. To define the physiological relevance of this ephrin signaling, Fu et al.³ induced synaptic scaling by prolonged incubation of cortical neurons with the GABA_A antagonist bicu culline, which enhances network activity. As expected, this elevated synaptic activity activated the EphA4 receptor and reduced total expression of GluR1 in wild-type neurons. In Epha4^–/– neurons, bicuculline had no effect on GluR1 levels. These findings suggest that chronic elevation of synaptic activity results in the activation of ephrin-EphA4 signaling, which, in turn, downregulates AMPAR expression, resulting in homeostatic scaling.

The EphA4-dependent downregulation of GluR1 requires both clathrin-mediated endocytosis and proteasomal activity. To investigate the molecular mechanisms underlying this effect, Fu et al.³ performed yeast two-hybrid screens for EphA4 interactors and isolated Cdh1, a component of the multiprotein ubiquitin ligase anaphase-promoting complex (APC). Of key importance, ephrin-A1 increases the interaction between APC and EphA4 in neurons, indicating that APC is recruited to EphA4 in a ligand-dependent manner. Furthermore, in heterologous cells, APC can bind to and ubiquitinate GluR1, decreasing both its surface and total expression. Consistent with this, knockdown of Cdh1 in neurons prevents both the ephrin-A1– and bicuculline-induced decrease in GluR1. Overall these results demonstrate that APC-mediated ubiquitination of GluR1 represents a physiologically important regulator of synaptic strength.

Although the actual lysine(s) that are ubiquitinated GluR1 remain to be defined, a surprising observation was that the site of APC binding to GluR1 is located in the extracellular N-terminal region of GluR1. Although the authors do not provide direct evidence, they speculate that the extracellular N-terminal region of GluR1 interacts with components of the APC ubiquitin ligase complex after retrotranslocation into cytoplasm. Indeed, this retrotranslocation event seems mechanistically analogous to the Cav1.2 translocation during ERAD reported by Altier et al.², but how this occurs for GluR1 is unclear.

Both papers present findings that significantly advance understanding of the roles of ubiquitin in controlling membrane protein stability and trafficking events that, in turn, regulate neuronal excitability and synaptic transmission. As always, however, these new findings raise further questions. Most immediate among these is identification of the specific lysine residues that are ubiquitinated in both Cav1.2 and GluR1 so that nonubiquitinatable mutants can be made and tested. Further work is also required to define exactly how Cavβ prevents the ERAD machinery recognizing Cav1.2. Possible mechanisms include steric hindrance and/or a role in folding of Cav1.2. For GluR1, what are the mechanisms for retrotranslocation, as this apparently occurs before ubiquitination, and how does EphA4-mediated recruitment of APC lead it to ubiquitinate GluR1?

Protein ubiquitination is now recognized as a key regulator of neuronal responsiveness. Clearly, these new studies raise questions concerning roles for ubiquitination in the regulated trafficking and stability of other proteins. Are equivalent processes required for other members of the voltage-gated ion channel and ligand-gated ion channel superfamilies, and more generally for other neuronal membrane proteins? For example, given the role of ERAD in the quality control of Cav1.2, it certainly seems plausible that an analogous mechanism could regulate AMPAR surface expression, possibly through one or more of the AMPAR accessory subunits or binding proteins such as TARPs⁹. Similarly, as shown for GluR1, ubiquitin may regulate the degradation of surface-expressed Cav1.2 channels. Much remains to be discovered, but it is already evident that ubiquitin and ubiquitin-like post-translational modification of neuronal proteins represents a crucial and, as yet, poorly understood regulatory system. The findings presented in this issue by Altier et al.² and Fu et al.³ provide further impetus to this exciting and expanding area of research.