A mechanistic framework for studying Kv channel clustering

Abstract

 

Action potential (AP) generation, propagation and transmission rely on precisely timed events associated with activation and inactivation gating transitions of voltage-dependent Kv and Nav channels¹ aggregated in numerous copies at unique membrane sites, such as the initial segment of an axon, nodes of Ranvier or at the post-synaptic density (PSD).² Despite the emerging importance of channel clustering for efficient electrical signaling,³ not much is known of the clustering process, in particular in the case of Kv channels. Its been demonstrated that Kv channel clustering at the PSD is mediated by binding to members of the family of synapse-associated scaffold proteins, such as the PSD-95 protein.⁴ Yet, the mechanism by which this elementary binding event leads to clustering of Kv channel molecules in a restricted area of membrane remains unclear. Indeed, a slew of questions remain unanswered. How many ion channel molecules reside in such clusters? Is there interaction between ion channel molecules within such clusters? Are Kv channel clusters stable over time or do they undergo activity-related dynamic changes? How is Kv channel clustering regulated in both the spatial and the temporal dimensions? How are Kv channel clusters formed, considering the multiple interaction partners of the PSD-95 hub protein? In absence of a molecular mechanism describing the Kv channel-PSD-95 interaction, answers for many of these questions have been hard to attain. Recent studies, however, have begun to provide some insight into Kv channel clustering.

We have lately demonstrated that the Kv channel-PSD-95 interaction occurs according to a ‘ball and chain’ mechanism,⁵ analogous to the ‘ball and chain’ mechanism that describes fast N-type channel inactivation (Fig. 1).⁶ Similar to its N-terminal-based counterpart (upper panel), the C-type mechanism relies on a cytoplasmic C-terminal channel tail comprising an extended ‘chain’, bearing a conserved PDZ-recognition motif at its tip (the ‘ball’), able to bind to the PDZ domains of PSD-95 (middle panel). According to this mechanism, the length of the extended chain provides the degrees of freedom needed to seek and then bind the PSD-95 scaffold protein partner. Considering the stoichiometry of this ‘ball and chain’ interaction, dictated by the 3 PDZ interaction modules of PSD-95 and the 4 C-terminal tails of the oligomeric Kv channel and the ability of PSD-95 to multimerize channel clustering results (lower panel). Three lines of evidence supported the proposed ‘ball and chain’ mechanism.⁵ First, hydrodynamic and spectroscopic analyses demonstrated that the isolated C-terminal tail of the Shaker Kv channel is an intrinsically disordered random chain. Second, surface plasmon resonance and isothermal titration calorimetery showed the Kv channel-PSD-95 interaction to be entropy-controlled, as determined by C-terminal tail length. Finally, tail length was found to time Kv channel-PSD-95 complex formation without affecting complex lifetime. These observations mirror evidence obtained for the case of fast channel inactivation⁶ and indicate that the Kv channel N- and C-terminal tails function as entropic clocks^5-7 to respectively time Kv channel fast inactivation and binding to the PSD-95 scaffold protein.

Figure 1.

‘Ball and chain’ mechanisms for Kv channel inactivation and clustering. Schematic representations of the ‘ball and chain’ mechanisms for channel fast inactivation (upper panel) and clustering (middle and lower panels). In the fast inactivation intra-molecular mechanism, the open Kv channel pore inactivates in a precisely timed manner, as determined by N-terminal chain length, upon binding of the ‘chain’-tethered ‘ball’ to a receptor site within the inner cavity of the open pore. In the inter-molecular clustering ‘ball and chain’ mechanism, Kv channel interaction with the membrane-associated PSD-95 scaffold protein is precisely timed, as determined by C-terminal chain length, upon binding of the ‘chain’-tethered peptide ‘ball’ to PSD-95 PDZ domain(s). Considering the stochiometery of the interaction and the ability of PSD-95 to aggregate, channel clustering results. The membrane-embedded portion corresponds to the channel voltage-sensor and pore domains, while the rectangular shape corresponds to the T1 assembly domain. The crescent, box and rectangular shapes represent the PDZ, SH3 and guanylate kinase-like domains of the PSD-95 protein, respectively.

The analogy between the channel fast inactivation and clustering ‘ball and chain’ mechanisms allows generalization to be made. In such a mechanism, entropic and enthalpic contributions to the binding reaction can be exclusively attributed to distinct sequence modules. Whereas the ‘chain’ module primarily controls the change in entropy (ΔS) of the binding reaction, the ‘chain’-tethered ‘ball’ sequence primarily determines the change in enthalpy (ΔH) of the reaction.⁵ These ‘chain’ and ‘ball’ sequence modules were further found to determine the respective association and dissociation kinetics of the recognition motif to and from its receptor site.^5,6 Based on these observations, we suggested a thermodynamic signature of entropy-controlled ‘ball and chain’ mechanisms, whereby the enthalpy and entropy changes of the binding reaction are invariant to each other upon changes in ‘chain’ length.⁵

The physiological relevance of Kv channel ‘ball and chain’ mechanisms is clearly demonstrated by the fact that alternative splicing of the Shaker Kv channel gene only occurs at regions encoding either the N- or C-terminal tails, leading to the appearance of chains of different lengths.⁸ Indeed, N-terminally spliced variants exhibited differences in channel fast inactivation kinetics,⁶ a modulation expected to affect AP shape, propagation and firing frequency. Similarly, C-terminally spliced channel variants exhibited differences in affinity to PSD-95, in cell surface expression levels and in PSD-95-mediated clustering phenotypes that correlated with differences in channel C-terminal tail length. Thus, molecular distinctions reflected in the differential affinity of the tail variants to PSD-95 translated into functional differences in the context of cellular channel clustering. Such splicing-based modulation of Kv channel clustering⁵ underlies changes in AP propagation and transmission.³

The ‘ball and chain’ description for Kv channel binding to a scaffold protein provides a mechanistic framework for studying Kv channel clustering, in a manner analogous to that employed for the study of Kv channel fast inactivation. As such, we can now rephrase the questions listed above in more mechanistic terms so as to directly define experiments that should be performed, at both the molecular and the cellular levels. Specifically, one can ask whether the C-terminal ‘chain’ plays an active or passive role in channel clustering. Does chain length systematically affect Kv channel cluster area size? Is an optimum ‘chain’ length expected? Does heterologous assembly of C-terminal tail-spliced variant subunits serve to control channel clustering? What is the relation between the timing property of the Kv channel PSD-95 interaction and the dynamics of channel cluster formation? Does chain length control the kinetics of cluster formation? What regulation strategies occur at the ‘chain’-level and/or the ‘ball’-level to control channel binding to PSD-95? Does crosstalk exist between the inactivation and clustering ‘ball and chain’ mechanisms? Is chain length a primary factor that determines the binding specificity of PSD-95 toward its multiple membrane protein partners, all bearing intrinsically disordered C-terminal tails harboring PDZ-binding motifs? Research designed to answer several of these questions is currently underway.

On a final note, with the diverse functional roles assumed by intrinsically disordered proteins (or protein segments) in almost all cellular processes described in recent years,⁷ the sequence-structure-function narrative of biochemistry has been extended to also include the unstructured protein dimension. In this respect, the channel fast inactivation and clustering functions of the Kv channel protein, described by a ‘ball and chain’ mechanism, nicely exemplify the entropic clock function of intrinsically disordered protein segments, in this case, in the context of electrical signaling.