Currents through voltage-gated sodium channels drive action potential depolarization in neurons and other excitable cells. Smaller currents through these channels are key components of currents that control neuronal firing and signal integration. Changes in sodium current have profound effects on neuronal firing. Sodium channels are controlled by neuromodulators acting through phosphorylation of the channel by serine/threonine and tyrosine protein kinases. That phosphorylation requires specific molecular interaction of kinases and phosphatases with the channel molecule to form localized signalling complexes. Such localization is required for effective neurotransmitter-mediated regulation of sodium channels by protein kinase A. Analogous molecular complexes between sodium channels, kinases and other signalling molecules are expected to be necessary for specific and localized transmitter-mediated modulation of sodium channels by other protein kinases.
Neuronal signals are comprised of action potentials passing along the length of axons and passing from one neuron to the next at synapses. Sodium channels underlie the depolarization of the action potential. In response to depolarization, they activate, allowing ions to flow through their pores, and then inactivate within milliseconds (Figure 1). However, a small fraction of sodium channels remains active between action potentials. This activity contributes to net depolarization, the size and rate of which contributes to control of repetitive neuronal firing. Modulation of these small sodium currents provides a sensitive substrate for neuromodulation.
Sodium channel activation
Molecular structure of voltage-gated sodium channels
Sodium channels are heterotrimeric molecules consisting of an α subunit of approx. 260 kDa that contains the ion-conducting pore and the molecular machinery that senses voltage to gate the channel  (Figure 2). This large α subunit is associated with a β1-like (β1 or β3) subunit of 36 kDa and is disulfide-linked to a β2-like (β2 or β4) subunit of 33 kDa. Nine functional α subunit isoforms have been described, of which four are highly expressed in the brain . The α subunits of sodium channels form part of the ion channel superfamily . Along with calcium channels, they form a subgroup in which the channels consist of four homologous domains. Each homologous domain resembles a voltage-gated potassium channel and contains six α-helical transmembrane segments and an ion-conducting pore. The highly conserved intracellular loop connecting homologous domains III and IV (loop III–IV) forms the inactivation gate [1,4]. The large intracellular loop connecting homologous domains I and II (loop I–II) of brain sodium channels contains a series of consensus phosphorylation sites for PKA (protein kinase A) and PKC.
Molecular structure of voltage-gated sodium channels
Physiological effects of activation of PKA and PKC
Activation of both PKA  and PKC [6,7] reduces sodium currents. Their activation is initiated by G-protein-coupled receptors. D1-like dopamine receptors acting through PKA reduce sodium current in striatal [8–10], prefrontal cortex pyramidal  and hippocampal neurons . Likewise, several receptor types act through PKC to reduce sodium current. Activation of muscarinic receptors in rat hippocampal neurons reduces sodium current by activation of PKC . In cortical pyramidal neurons, activation of 5-HT2A/C (5-hydroxytryptamine) receptors also decreases sodium current via PKC activation . Likewise, in striatal neurons and cholinergic interneurons, activation of D2 dopamine receptors decreases sodium current subsequent to PKC activation . Activation of these pathways, including their effects on sodium channels, results in profound effects on neuronal activity.
The degree of sodium current reduction by PKA depends on membrane potential, being more potent at depolarized potentials . In addition, modulation by PKA is more effective following prior activation of PKC . Recently, it was shown that both of these kinases reduce sodium current by enhancing slow inactivation, an intrinsic voltage-dependent process that reduces sodium current upon prolonged or repetitive depolarization . This was strengthened by the identification of mutations that greatly reduce slow inactivation and that also strongly inhibit functional modulation of the channel by PKA and PKC; mutations that enhance the slow inactivation process potentiate modulation by these kinases .
The reduction of sodium current subsequent to receptor activation has potent effects on neuronal activity. Activation of 5-HT2 receptors acting via PKC in prefrontal cortical neurons results in increased threshold for activation, resulting in reduced sodium channel activity . Likewise, sodium current reduction due to activation of D2 dopamine receptors acting through PKC reduces automaticity in cholinergic interneurons of the striatum .
PKA and PKC as components of signalling complexes
It is becoming increasingly evident that protein kinases frequently form signalling complexes with their substrates . Consistent with this idea, PKA co-purifies with the sodium channel and a 15 kDa AKAP (A-kinase anchoring protein), AKAP15 . AKAP15 specifically associates with loop I–II of NaV1.2 voltage-gated sodium channel α subunits that contain the sites in the channel that are phosphorylated by PKA. This interaction is critical to function since its disruption abolishes sodium current reduction by D1-like dopamine receptors .
Signalling by PKC also involves specific signalling molecules. Reduction of sodium current by the PKC activator OAG (oleoylacetylglycerol) was blocked by inhibitors that specifically block the activity of PKCϵ. In addition, a peptide inhibitor of anchoring of PKCϵ blocked sodium channel modulation by OAG. Consistent with these findings, modulation by OAG was greatly reduced in hippocampal neurons in mice in which PKCϵ had been removed by targeted deletion . Thus, as with PKA, modulation by PKC involves specific molecular components subject to specific targeting for successful modulation.
Sites of phosphorylation and regulation by PKA and PKC
The sites phosphorylated by PKA and PKC are found in loop I–II, which is a large intracellular loop connecting domains I and II that is present in brain sodium channels, but is absent or greatly altered in muscle sodium channels. NaV1.2 α subunits are phosphorylated by PKA on Ser573, Ser610, Ser623 and Ser687 [24,25]. These sites are also selectively dephosphorylated by phosphoprotein phosphatases calcineurin and phosphatase 2A .
PKC phosphorylates three sites that overlap those phosphorylated by PKA as well as additional sites . One is Ser1506 in loop III–IV that forms the inactivation gate [27,28]. The other two sites are Ser576 and Ser610 in loop I–II, the same loop that contains the four sites phosphorylated by PKA . Ser610 is phosphorylated by both PKA and PKC in vitro.
The sites that are required for regulation of the channel are a smaller subset of the sites that are phosphorylated biochemically. Ser573 is required for full reduction of current by both PKA [28,29] and PKC . In addition, Ser554 and Ser1506 also participate in the physiological effects of PKC [7,28,30]. Two other sites of PKA phosphorylation are needed for the increased effects of PKA in the presence of PKC activation, Ser576 and Ser687 .
Sodium channel regulation by tyrosine phosphorylation
Tyrosine kinases have also been implicated in regulation of sodium channels. Evidence was initially provided by Hilborn et al. . They showed that sodium channel inactivation was changed by activation of tyrosine kinases in PC12 cells, and implicated an Src-family tyrosine kinase in that modulation. Recent studies have examined the molecular basis of that regulation, which may differ, depending on the sodium channel α subunit isoform. For the cardiac Nav1.5 sodium channel α subunit, a tyrosine in the inactivation gate has been shown to be phosphorylated by Fyn kinase . Phosphorylation of this tyrosine shifts the voltage dependence of inactivation towards more depolarized potentials, increasing the number of channels available for activation at a particular potential. Conversely, tyrosine phosphorylation of brain NaV1.2 sodium channels shifts the voltage dependence of inactivation towards more negative potentials, suggesting actions that may be molecularly different .
The β1 subunit has also been implicated as a site of tyrosine phosphorylation. A critical tyrosine on the intracellular cytoplasmic tail is a site of tyrosine phosphorylation . This tyrosine controls the interaction of this β subunit with the cytoskeletal protein, ankyrin. In cardiac myocytes, control of this interaction determines whether the β1 subunit is found in the intracellular transverse tubules or, alternatively, at the intercalated discs connecting two myocytes .
Voltage-gated sodium channels also interact with a tyrosine phosphatase, receptor tyrosine phosphatase β . Both extra- and intra-cellular portions of the phosphatase interact with the channel. Expression of the catalytic portion of the phosphatase resulted in a positive shift in the voltage dependence of inactivation of the channel. Conversely, inhibitors of tyrosine phosphatases resulted in a negative shift in the voltage dependence of inactivation. Thus the effects of this phosphatase counteract those of tyrosine kinases on the rat brain sodium channel.
Regulation of sodium channels by serine/threonine and tyrosine kinases provides a potent means of short- and long-term regulation of sodium channels and neuronal activity. The kinases and phosphatases involved in signalling often exist in specific signalling complexes with the sodium channel, and these molecular interactions are required for functional modulation. The requirement for these complexes is likely to underlie specificity in signalling. Further elucidation of these complexes, their physiological effects and the molecular mechanisms underlying those effects provides fertile ground for future progress in understanding a key control point in neuronal signalling.
International Symposium on Neurodegeneration and Neuroprotection: Independent Meeting held at University of Münster, Germany, 23–27 July 2006. Organized and Edited by S. Klumpp and J. Krieglstein (Münster, Germany).