INTRODUCTION

TOP ABSTRACT INTRODUCTION RECEPTOR AND NONRECEPTOR... TYPES OF SUPPORTING EVIDENCE REGULATION OF LIGAND-GATED... REGULATION OF K+ CHANNELS... REGULATION OF CA2+ CHANNELS... REGULATION OF NA+ CHANNELS... REGULATION OF RACC AND... REGULATION OF CL CHANNELS... THE CHANNEL-PROTEIN KINASE... CONCLUSIONS AND FUTURE... REFERENCES

ION CHANNELS ARE TARGETS of many intracellular signaling pathways, including protein phosphorylation and dephosphorylation. These processes can modify channel activity and dramatically alter the electrophysiological properties of both excitable and nonexcitable cells. Two well-known examples of this type of regulation are the L-type Ca²⁺ channel and the large conductance Ca²⁺-dependent K⁺ (maxi-K_Ca) channel, proteins that often have opposing physiological functions (157). The L-type Ca²⁺ channel is the primary voltage-dependent Ca²⁺-influx pathway in many excitable cells. Its pore-forming subunit, _1C, contains NH₂- and COOH-terminal cytosolic domains that are potential targets for protein phosphorylation. For example, in cardiac myocytes, activation of ₁-adrenergic receptors leads to an increase in cytosolic cAMP, which activates the serine-threonine kinase protein kinase A (PKA). PKA phosphorylates Ser¹⁹²⁸ of the _1C COOH-terminus, shifting the voltage activation of the Ca²⁺ channel toward more negative potentials and producing an enhancement in whole cell L-type Ca²⁺ current (62). The resulting increase in Ca²⁺ influx contributes to the chronotropy and inotropy induced by sympathetic activation of the heart (140). L-type Ca²⁺ current is also potentiated by agonists of endothelin, ₁-adrenergic, and angiotensin II receptors through an effect on another serine-threonine kinase, protein kinase C (PKC). However, the exact phosphorylation mechanisms and sites of PKC action are unclear (106). The maxi-K_Ca channel is also regulated by serine phosphorylation. This channel is composed of pore-forming -subunits and a regulatory -subunit, with the COOH-termini of the -subunits being cytosolic. The maxi-K_Ca channel is regulated by voltage in a Ca²⁺-dependent manner. Protein kinase G (PKG) phosphorylates the -subunit at Ser¹⁰⁷² near the COOH-terminus (60), shifting the Ca²⁺ sensitivity of the channel and producing hyperpolarization (10, 23). Other putative PKG and PKC phosphorylation sites occur on the -subunit COOH-terminus in splice variants of the maxi-K_Ca channel, which may account for the various phenotypes of the channel observed in different tissues (237).

In addition to the extensive information available about the regulation of ion channels by serine-threonine kinases (for a review, see Ref. 95), an emerging body of evidence suggests that channels are also regulated by phosphorylation on tyrosine residues (for reviews, see Refs. 104, 123, 124, 215, 240). Growth factors, which act through receptor tyrosine kinases, are known to regulate the function and expression of many proteins, including ion channels (14, 178). The literature also indicates that nonreceptor tyrosine kinases can regulate ion channels. Collectively, this evidence suggests a mechanism whereby by growth factors, cell-cell, and cell-substrate interactions acutely regulate cell function through ion channels.

The purpose of this review is to summarize the current state of this field. As will become evident, the majority of voltage-gated, ligand-gated, and second messenger-gated channels are regulated to some degree by tyrosine phosphorylation. We review the evidence to support this concept for most major classes of ion channels, making extensive use of tables to summarize the literature. We examine the probable roles of signaling pathways involving receptor and nonreceptor tyrosine kinases that would be expected to lead to channel regulation. We discuss the critical roles that channel-associated scaffolding proteins may play in localizing protein tyrosine kinases and phosphatases to the vicinity of ion channels. Finally, when possible, we speculate on the physiological significance of ion channel regulation by these processes.

Glossary

AChR	Nicotinic acetylcholine receptor
AKAP	A kinase-associated protein
AMPA	-Amino-3-hydroxy-5-methyl-4-isoxazole proprionate
BDNF	Brain-derived neurotrophic factor
BKCa channel	Large (big) conductance KCa channel
bRET	Bovine rod channel
[Ca2+]i	Intracellular Ca2+ concentration
CaM	Calmodulin
Cav channel	Voltage-activated Ca2+ channel
CFTR	Cystic fibrosis transmembrane conductance regulator
ClC channel	Voltage-gated Cl channel family
CNG channel	Cyclic nucleotide-gated channel
Csk	c-Src kinase
DEP	dsh/egl-10/pleckstrin domain
EGF	Epidermal growth factor
EGFR	EGF receptor
ENaC	Amiloride-sensitive Na+ channel
FAK	Focal adhesion kinase
FGF	Fibroblast growth factor
GABA	-Amino butyric acid
GIRK channel	G protein-coupled inwardly rectifying K+ channel
GK	Guanylate cyclase
GKAP	G kinase-associated protein
GST	Glutathione-S-transferase
HEK	Human embryonic kidney (cell line)
ICRAC	Calcium release-activated current
IGF	Insulin-like growth factor
INAD	Inactivation-no-afterpotential D (protein)
Ins(1,4,5)P3	D-myo-Inositol (1,4,5)-trisphosphate
JAK2	Janus kinase 2
KATP channel	ATP-sensitive K+ channel
KCa channel	Ca2+-dependent K+ channel
Kir channel	Inwardly rectifying K+ channel
Kv channel	Voltage-gated K+ channel
MAGUK	Membrane-associated GK
MAPK	Mitogen-activated protein kinase
MDCK	Madin-Darby canine kidney (cell line)
MuSK	Muscle-specific (tyrosine) kinase
NGF	Nerve growth factor
NGFR	NGF receptor
NMDA	N-methyl-D-aspartate
NT3	Neurotrophin-3
PC12	Rat pheochromocytoma
PDGF	Platelet-derived growth factor
PDZ	PSD-95/SAPSO, Dlg, and 20-1 domain
PH	Pleckstrin homology
PKA	Protein kinase A
PKC	Protein kinase C
PKG	Protein kinase G
PLC	Phospholipase C
Po	Open probability
PTB	Phosphotyrosine binding
PTK	Protein tyrosine kinase
PTP	Protein tyrosine phosphatase
PX	Phox homology
PYK2	Pyruvate kinase 2
RACC	Receptor-activated Ca2+ channel
RPTP	Receptor PTP
RVD	Regulatory volume decrease
rOLF channel	Rat olfactory CNG channel
RPTP	Receptor PTP
SERCA	Sarco(endo)plasmic reticulum Ca2+-ATPase
SH2	Src homology 2
SH3	Src homology 3
Src	Sarcoma virus tyrosine kinase
STOC	Spontaneous transient outward current
Trp	Transient receptor potential
TTX	Tetrodotoxin
VEGF	Vascular endothelial growth factor

	RECEPTOR AND NONRECEPTOR TYROSINE KINASES

TOP ABSTRACT INTRODUCTION RECEPTOR AND NONRECEPTOR... TYPES OF SUPPORTING EVIDENCE REGULATION OF LIGAND-GATED... REGULATION OF K+ CHANNELS... REGULATION OF CA2+ CHANNELS... REGULATION OF NA+ CHANNELS... REGULATION OF RACC AND... REGULATION OF CL CHANNELS... THE CHANNEL-PROTEIN KINASE... CONCLUSIONS AND FUTURE... REFERENCES

To provide a context for understanding how ion channels are regulated by tyrosine phosphorylation, some brief discussion of PTKs and phosphatases seems appropriate.

PTKs were initially discovered as the products of oncogenes from transforming retroviruses (91). PTKs are responsible for transducing key extracellular signals that mediate events such as proliferation, cytoskeletal rearrangement, and coordination of physiological responses. Cellular PTKs can be divided into two primary groups: receptor and nonreceptor PTKs (117).

Members of the receptor PTK family include the insulin receptor and growth factor receptors such as EGF, PDGF, FGF, IGF, and VEGF receptors. Structurally, the receptor PTKs are characterized by an extracellular ligand-binding domain, a transmembrane domain, a kinase-catalytic domain, and cytoplasmic regions responsible for organizing signaling molecules. Signal transduction is initiated by a sequence of growth factor (ligand) binding to the extracellular domain, dimerization of the receptor proteins, and autophosphorylation of the receptor. Receptor autophosphorylation creates phosphorylated tyrosine residues on the cytoplasmic tail of the receptor. These phosphorylated tyrosine residues form docking sites for signaling molecules, and it is the combination of these signaling molecules that determines the specificity of individual receptor PTKs (51).

Nonreceptor PTKs are found in both the cytoplasm and nuclei of cells. The largest family of cytoplasmic PTKs is the Src family (for a review, see Ref. 1). The Src family consists of eight members: Src, Fyn, and Yes (which are ubiquitously expressed) and Lck, Hck, Fgr, Lyn, and Blk (which exhibit a more restricted expression pattern). Src family PTKs are characterized by NH₂-terminal sequences that direct myristolation or pamitoylation, resulting in membrane localization. Other sequences include a nonconserved unique domain, an SH3 domain that directs binding to polyproline-rich sequences, an SH2 domain that binds phosphotyrosine, a catalytic domain, and a short COOH-terminal tail. Regulation of Src family members is highly conserved. In particular, there are two tyrosine residues: one in the kinase domain and one in the COOH-terminal tail, whose phosphorylation state is important for Src activation. Autophosphorylation of the kinase domain tyrosine leads to increased kinase activity, whereas phosphorylation of the COOH-terminal tail region by Csk represses activity (164). Hence, activation can occur by dephosphorylation of the COOH-terminal tyrosine by a phosphatase such as RPTP or phosphorylation of the kinase domain tyrosine (69). Many stimuli, including receptor PTKs, G protein-coupled receptors, and integrins, have been implicated in Src activation, making this family of kinases a key point of integration for many signal transduction pathways. It is now clear that autophosphorylation of receptor PTKs can form binding sites on the receptor for Src family members and that these nonreceptor PTKs are important for some of the downstream signaling events initiated by receptor binding (233).

The other class of nonreceptor PTKs that is important for this discussion includes FAK and Pyk2/RAFTK (for a review, see Ref. 12). pp125^FAK is a 125-kDa PTK that is discretely localized to cellular focal adhesions and has been shown to colocalize with integrins. FAK is a major substrate for integrin-dependent tyrosine phosphorylation, and, upon phosphorylation, it becomes enzymatically active and serves as a scaffold for the binding and localization of other proteins to the focal adhesion (201). Pyk2/RAFTK is characterized by a more diffuse cytoplasmic localization and can be activated by stress signals (122) and by a variety of extracellular signals that elevate [Ca²⁺]_i. FAK and Pyk2/RAFTK exhibit ~48% amino acid identity and have a similar organization: a unique NH₂-terminal region, a centrally located PTK domain, and two proline-rich regions at the COOH-terminus. Neither kinase contains SH2 or SH3 domains.

Activation of the Src family members FAK and Pyk2/RAFTK serves as a central integration mechanism for a number of extracellular signaling pathways. For example, when integrins on the cell surface are engaged, signal transduction pathways are triggered. Because integrins lack intrinsic enzymatic activity, they rely on the activation of a number of cytoplasmic signaling molecules, including FAK and Src. The two proteins are colocalized, and their interaction leads to autophosphorylation of FAK, creation of a binding site for the Src SH2 domain, and ultimately to Src activation (198). Src then serves to phosphorylate additional sites on FAK, allowing binding of other signaling molecules and scaffolding proteins (179, 200). This process leads to the assembly of complex signaling molecules at the focal adhesion site and organizes signaling events further downstream. There is increasing evidence for synergy between receptor PTKs and integrin signaling pathways. The Src-FAK pathways are thought to be key sites for the integration of signals from integrin pathways (203), and evidence for this integration with respect to ion channel regulation will be apparent in subsequent sections of this review.

PTPs are composed of two groups: one that recognizes phosphoserine, phosphothreonine, and phosphotyrosine residues and another that specifically recognizes phosphotyrosines (174, 249). The latter group can be subdivided into intracellular (nonreceptor) or receptor PTPs. PTPs have a conserved catalytic domain that recognizes the VIHCSAGXGRXG motif. Receptor PTPs contain extracellular domains with structural motifs similar to those found in cell adhesion molecules and two catalytic domains on their intracellular COOH-terminus. Nonreceptor PTPs contain only a single conserved phosphatase domain flanked by noncatalytic segments that mediate interactions with adapter molecules through specific protein-protein interaction domains (see THE CHANNEL-PROTEIN KINASE REGULATORY COMPLEX). Although much is known about the structure and function of PTPs (174), the evidence for involvement of specific PTPs in the regulation of ion channels is sparse.

	TYPES OF SUPPORTING EVIDENCE

TOP ABSTRACT INTRODUCTION RECEPTOR AND NONRECEPTOR... TYPES OF SUPPORTING EVIDENCE REGULATION OF LIGAND-GATED... REGULATION OF K+ CHANNELS... REGULATION OF CA2+ CHANNELS... REGULATION OF NA+ CHANNELS... REGULATION OF RACC AND... REGULATION OF CL CHANNELS... THE CHANNEL-PROTEIN KINASE... CONCLUSIONS AND FUTURE... REFERENCES

Initial evidence that ion channels were regulated by tyrosine phosphorylation came largely from pharmacological studies using soluble inhibitors of PTKs. Electrophysiological and molecular biology methods have subsequently provided insight into the mechanisms of channel regulation. This progression of methodology can be traced using the L-type Ca²⁺ channel as an example and is detailed below.

In vascular smooth muscle, broad-spectrum PTK inhibitors such as genistein and herbimycin attenuate depolarization-induced tone (64, 236) as well as myogenic tone (135, 165). Genistein inhibits the rise in [Ca²⁺]_i induced by KCl depolarization (64), suggesting that it acts on a Ca²⁺ entry mechanism. Electrophysiological studies confirm that soluble PTK inhibitors block whole cell Ca²⁺ current (265), including studies that utilize inhibitors such as PP1 and PP2, which more specifically target certain PTKs, e.g., Src (48, 270). Other studies have employed strategies involving intracellular dialysis with antibodies or peptides directed at specific kinases. For example, L-type Ca²⁺ current in smooth muscle is increased by intracellular dialysis with a pipette solution containing a c-Src-activating peptide (263) or constitutively active Src (270) but is inhibited by monoclonal antibodies against c-Src (84, 270). Studies combining electrophysiological and molecular biology strategies have provided the best evidence for specific channel-kinase interactions. Expression of recombinant L-type Ca²⁺ channels in a mammalian cell line, along with kinase-dead Src to compete with endogenous Src, results in blockade of the otherwise potentiating action of IGF-1 on Ca²⁺ current (16). Subsequently, these electrophysiological data have been supported by biochemical evidence for Src-Ca²⁺ channel association using immunoprecipitation and GST fusion protein assays (16, 84). In addition, kinase assays show that purified Src kinase can phosphorylate L-type Ca²⁺ channels in vitro (16). Collectively, this combination of methods provides convincing support for the idea that the L-type Ca²⁺ channel is regulated by tyrosine phosphorylation.

As noted in our data tables, the majority of evidence for regulation of ion channels by tyrosine phosphorylation is based primarily on the effects of soluble PTK inhibitors. A major flaw in this evidence is that these antagonists are reported to have nonspecific effects (3), including direct interactions with channels (85, 166, 216), inhibition of serine-threonine kinases (2), interactions with Ca²⁺ indicators (17), and other nonkinase actions (114, 278). Thus approaches that depend solely on the use of these inhibitors must be interpreted with caution. A stronger case for regulation by tyrosine phosphorylation can be made if multiple PTK inhibitors acting through different mechanisms produce the same effect on a channel or if structurally related inactive forms are used. However, because some soluble PTK inhibitors are relatively specific for certain kinase families (e.g., herbimycin A, PP1, and PP2), the lack of an effect on a given channel may simply indicate that a certain class of PTK is not involved in regulation of that channel. Ideally, supportive evidence would be provided by an opposite effect resulting from inhibition of PTPs. Unfortunately, few selective tools are available to selectively inhibit PTPs. The most commonly used phosphatase blockers, pervanadate and sodium orthovanadate, are also inhibitory for serine-threonine phosphatases (72). The purification and use of specific PTPs and their inhibitors will obviously allow more specific testing of the role of these enzymes in ion channel regulation.

Although molecular approaches can potentially be used to pinpoint the site and mechanism of action of a specific PTK, these approaches are not without their own limitations. Expression of specific channel subunits in Xenopus oocytes or in mammalian cell systems (e.g., HEK 293 and COS-7) can result in uncertainties about proper membrane targeting of the channel. In many cases, the complete array of functions for accessory subunits are unknown, so that it cannot be determined a priori which channel subunits need to be expressed. Some expression systems may not possess the adequate intracellular signaling machinery to reconstitute a specific PTK-ion channel regulatory pathway. Also, some channels may undergo posttranslational processing in native cells to alter associations between channel subunits (63), but this processing may not be possible in recombinant expression systems due to lack of the appropriate enzymatic machinery. These inherent difficulties can lead to substantial differences in the behavior and regulation of expressed versus endogenous channels. Therefore, it is necessary to keep these limitations in mind when interpreting such studies, and the conclusions must ultimately be confirmed using the endogenous channel and PTK(s) in the native cell system.

	REGULATION OF LIGAND-GATED CHANNELS BY PTKS

TOP ABSTRACT INTRODUCTION RECEPTOR AND NONRECEPTOR... TYPES OF SUPPORTING EVIDENCE REGULATION OF LIGAND-GATED... REGULATION OF K+ CHANNELS... REGULATION OF CA2+ CHANNELS... REGULATION OF NA+ CHANNELS... REGULATION OF RACC AND... REGULATION OF CL CHANNELS... THE CHANNEL-PROTEIN KINASE... CONCLUSIONS AND FUTURE... REFERENCES

Ligand-gated ion channels are among the most precisely localized of membrane proteins through their extensive associations with specific scaffolding and cytoskeletal proteins at the postsynaptic density (207, 208). PTKs are also highly localized in this region (227). It is therefore not surprising that several types of ligand-gated channels are associated with, and/or regulated by, PTKs. We will review the evidence for regulation of three distinct types of ligand-gated channels by tyrosine phosphorylation. Other ligand-gated channels, or their scaffolding protein partners, are also known to be associated with PTKs (25), e.g., the AMPA receptor is associated with the nonreceptor PTK Lyn (71). However, this section will focus on those channels for which there is the best functional evidence for this type of regulation (see Table 1).

AChR

Regulation of AChR localization is coordinated by a number of accessory proteins through tyrosine phosphorylation events. During synaptogenesis, the neurally released factor agrin causes the clustering of AChRs. Agrin is an extracellular heparin sulfate proteoglycan capable of binding growth factors (36) and, along with at least one accessory protein, stimulates the muscle-specific receptor PTK MuSK. The -subunit of AChR is tyrosine phosphorylated by MuSK or by an intermediate PTK, perhaps Fyn or Src (59), through a process involving one or more scaffolding proteins, including rapsyn (36). The - and -subunits of AChR are also tyrosine phosphorylated (89). Fyn and Fyk PTKs are known to associate through their SH2 domains with the YFNI sequence of the -subunit (228). Of note, tyrosine phosphorylation of the -subunit results in its association with the adaptor protein Grb2 in the AChR complex (35). Grb2 is a key upstream player in the Ras-MAPK signaling pathway, leaving open the possibility that this pathway regulates the channel. However, the extent to which these accessory proteins and PTKs play a role in functional regulation of AChR, e.g., channel sensitization or desensitization, is not known.

NMDA Receptor

NMDA receptor currents are potentiated by PTK activation and inhibited by PTP activation (112, 253). For example, the Src-activating peptide EPQ(pY)EEIPIA applied to the intracellular surface of the NMDA receptor increases P_o of the native channel in cultured neurons from the rat spinal dorsal horn (128). Another peptide containing the amino acid sequence 40-58 of Src blocks the action of EPQ(pY)EEIPIA. P_o is reduced by either of two Src antibodies but increased by recombinant pp60^c-Src (128, 253). Importantly, the effects of these peptides are also evident on the NMDA component of miniature excitatory postsynaptic currents (128). Further evidence for a physiological effect is the correlation between enhanced tyrosine phosphorylation of NMDA receptors and long-term potentiation (191, 192). Src and Fyn kinases lose their potentiating effect on recombinant NR₁-NR_2A channels if the COOH-terminus of NR_2A is deleted (112). Coimmunoprecipitation of NR₂ subunits and Src family kinases suggests a physical association between the receptor and the kinase (115), although intermediate kinase(s) may also be involved. Indeed, the inhibitory action of Src(40-58) peptide, which does not itself interact with the Src catalytic domain (128), points to additional protein-protein interactions required for tyrosine phosphorylation of the receptor.

GABA Receptor

Initial studies indicated that the GABA_A receptor, the most widely distributed subunit, is modulated by tyrosine phosphorylation of the ₂- and ₁-subunits. GABA_A receptors expressed in oocytes are inhibited by genistein and tyrphostin B-44, and this effect is associated with tyrosine phosphorylation of ₂- and ₁-subunits by the nonreceptor PTK pp60^c-Src (243). If ₁-, ₁-, and ₂-subunits are expressed in HEK 293 cells along with a constitutively active form of Src, v-Src, two tyrosine residues on ₂ are phosphorylated and whole cell GABA current is enhanced (154). Mutation of residues Tyr^365/367 to Phe^365/367 on ₂ blocks Src-induced phosphorylation and enhancement of current. Mutation of residues Tyr^384/386 to Phe^384/386 on ₁ blocks Src-induced phosphorylation without any obvious effect on current, suggesting that tyrosine phosphorylation of ₂ is critical to regulation (154). The function of ₁ tyrosine phosphorylation remains to be determined. In superior cervical ganglion neurons, native GABA_A channels are modulated by tyrosine kinase/phosphatase inhibitors and by dialysis with c-Src in a way consistent with physiological potentiation of GABA_A current by tyrosine phosphorylation (154).

A similar regulation of GABA_A is observed in spinal dorsal horn neurons. GABA_A currents are reversibly inhibited by genistein but potentiated by recombinant pp60^c-Src (250). Both ₂/₃-subunits are tyrosine phosphorylated by an endogenous PTK(s); however, genistein appears to produce comparable inhibition of GABA current in heterologously expressed ₁₂-subunits compared with ₁₂₂-subunits. This suggests that ₂, not ₂, is required for this regulation (250). The functional significance of ₂ tyrosine phosphorylation is not yet known.

In contrast, tyrosine phosphorylation of frog melanotroph ₂/₃-subunits inhibits GABA-evoked current (29). Soluble PTK inhibitors potentiate GABA currents in inside-out patches from these cells and reduce endogenous phosphorylation of ₂/₃-subunits. Furthermore, single channel currents are inhibited by recombinant pp60^c-Src or the PTP inhibitor sodium orthovanadate. The reasons for these disparate results (29, 250) are unknown, but they possibly reflect the actions of multiple interacting tyrosine phosphorylation events, different endogenous PTKs, phosphorylation of different GABA subunits, or involvement of different channel-associated scaffolding proteins.

	REGULATION OF K+ CHANNELS BY PTKS

TOP ABSTRACT INTRODUCTION RECEPTOR AND NONRECEPTOR... TYPES OF SUPPORTING EVIDENCE REGULATION OF LIGAND-GATED... REGULATION OF K+ CHANNELS... REGULATION OF CA2+ CHANNELS... REGULATION OF NA+ CHANNELS... REGULATION OF RACC AND... REGULATION OF CL CHANNELS... THE CHANNEL-PROTEIN KINASE... CONCLUSIONS AND FUTURE... REFERENCES

Kv Channels

Initial evidence for Kv channel regulation by tyrosine phosphorylation was based on the effects of pharmacological inhibitors of tyrosine kinases and phosphatases. The evidence was obtained using a variety of native cells as well as recombinant channels in heterologous expression systems. The majority of these studies are summarized in Table 2. Our discussion focuses primarily on combined electrophysiological and molecular biology studies that provide the most compelling evidence.

Regulation of Shaker family Kv channels by tyrosine phosphorylation is firmly established. Coexpression of the M₁ (muscarinic) AChR and the delayed rectifier K⁺ channel, Kv1.2, in oocytes results in suppression of K⁺ current after activation of the receptor (87). This inhibition is mediated by tyrosine phosphorylation of the channel subsequent to activation of PLC-1 and PKC. Suppression of Kv current is relieved by mutation of a single tyrosine residue, Tyr¹³² to Phe¹³², on the NH₂-terminus of the channel. This mutation reduces both basal and carbachol-induced tyrosine phosphorylation of the channel, as assessed from immunoprecipitation protocols. In contrast, mutation of another NH₂-terminus site, Tyr⁷⁶ to Phe⁷⁶, or two potential serine phosphorylation sites for MAPK on the COOH-terminus have no effect on the inhibition caused by muscarinic agonists (87).

Coexpression of M₁ AChR, Kv1.2, and EGFR in HEK 293 cells results in suppression of K⁺ current after receptor activation (87, 239). Application of EGF alone results in suppression of Kv1.2 current and phosphorylation of Kv1.2 by EGFR, but the "specific" EGFR kinase inhibitor tyrphostin AG1478 blocks carbachol-mediated inhibition of Kv1.2 current (239). Therefore, M₁ AChR appears to transactivate EGFR to modulate the K⁺ channel. Because AG1478 blocks only ~50% of this response, it leaves open the possibility that other PTKs may also be involved. Indeed, in PC12 cells, another PTK responsible for phosphorylation of Kv1.2 is the nonreceptor tyrosine kinase PYK2 (122). PYK2 activity is stimulated by depolarization, by PKC, and by agonists of G protein-coupled receptors (e.g., bradykinin), all of which stimulate increases in [Ca²⁺]_i and subsequently inhibit Kv1.2 current (122).

In T lymphocytes, Kv1.3 is tyrosine phosphorylated after stimulation of Fas (an apoptosis-related protein), which leads to apoptotic cell death. Phosphorylation appears to be mediated by the tyrosine kinase p56^Lck because deficiency in that kinase abolishes inhibition of current by Fas, whereas reconstitution of p56^Lck restores the effect (229). Whether Lck interacts directly with the channel has not been determined. When Kv1.3 is coexpressed in HEK 293 cells with EGFR or v-Src, the channel becomes tyrosine phosphorylated and current is suppressed by >95% (24, 78). PTK inhibitors relieve the current suppression by v-Src. The action of v-Src is not mediated by an effect on Kv1.3 expression but by modulation of channel activation and inactivation kinetics (78). On the basis of mutational analysis, regulation of Kv1.3 kinetics by Src is determined by multiple interacting phosphorylation sites, including Tyr¹³⁷ on the Kv1.3 NH₂-terminus and Tyr⁴⁴⁹ on the COOH-terminus (24). Interestingly, inhibition of Kv1.3 by EGF, which is predominantly through speeding of C-type inactivation, is mediated at a different site, Tyr⁴⁷⁹. Collectively, these observations imply that Kv1.3 can be differentially regulated by various PTKs.

The NH₂-terminus of the human K⁺ channel, hKv1.5, contains a proline-rich SH3 domain known to interact with Src (79). The channel and enzyme colocalize in cell adhesion zones in the myocardium (138). Coexpression of Src and hKv1.5 in HEK 293 cells results in tyrosine phosphorylation of the channel and suppression of channel current (79). This is consistent with inhibition of Kv current by EGF in cardiac myocytes (24).

In contrast to the recombinant K⁺ channel studies cited above, tyrosine phosphorylation of native Kv current in mouse Schwann cells leads to current activation (219). This action is apparently mediated by the Src-family kinase p55^Fyn because cell dialysis with recombinant Fyn kinase results in potentiation of a slowly inactivating K⁺ current component that is blocked by herbimycin A. Both Kv1.5 and Kv2.1 -subunits are constitutively tyrosine phosphorylated (219). Furthermore, coimmunoprecipitation and double-labeling experiments suggest a direct association between the p55^Fyn and K⁺ channels in these cells (219). Kv2.1 associates with Fyn through an SH2 domain (219), in contrast to Kv1.5, which associates with Src family kinases through its SH3 domain (79). The effect of tyrosine phosphorylation on K⁺ current in this study is noteworthy because nearly all other Kv channel studies suggest that the end result of tyrosine phosphorylation is inactivation of the channel. Sobko et al. (219) proposed that this discrepancy may be related to the absence of regulatory channel subunits in some heterologous expression systems or to the possibility that heteromultimers formed with Kv2.1 subunits in native cells confer this type of regulation by PTKs (219). In support of this idea, at least two other studies presented indirect evidence for activation of native K⁺ current after stimulation of receptor PTKs (104, 173, 181). It is also possible that multiple receptor and nonreceptor PTKs are involved in the normal regulation of Kv channels, e.g., Src (49, 77, 125), EGFR (239), PYK2 (122), and JAK2 (181). Resolution of these important issues may completely alter conclusions about the physiological role of Kv channel tyrosine phosphorylation.

Kir Channels

Kir channels are also regulated by tyrosine phosphorylation, as summarized in Table 2. Native and recombinant Kir2.1 currents are suppressed by perorthovanadate inhibition of PTPs, and inhibition is reversed by genistein (268). Both effects are prevented by mutation of Tyr²⁴² to Phe²⁴² in the Kir2.1 COOH-terminus (268). When Kir2.1 is coexpressed with NGFR or EGFR, application of the appropriate ligand leads to suppression of current. Other growth factors (NT3 or BDNF) similarly suppress current, either by crossreaction with NGFR or by interaction with endogenously expressed receptor PTKs (268). Additional potential tyrosine phosphorylation sites have been identified on Kir2.1 (171), but their role in regulating the channel is not known. Tyr²⁴² is present in other members of Kir2 at the equivalent position, so receptor PTKs have the potential to regulate other Kir2 channels by this mechanism. The equivalent tyrosine site is not found in GIRK channels (Kir3.x subfamily), but mutation of either of two other tyrosine residues in Kir3.1 (Tyr¹² or Tyr⁶⁷) or in Kir3.4 (Tyr³² or Tyr⁵³) blocks BDNF-induced inhibition of current (189). The sites more proximal to the NH₂-terminus appear to play the more critical role in regulating the BDNF effect. Kir3.2 channels, which lack a tyrosine residue at this site, acquire BDNF sensitivity if an Asp⁴¹ to Tyr⁴¹ mutation is made in the proximal position (189). To achieve this effect, it is also necessary to add an additional basic charge near this site, apparently to complete the general motif characteristic of PTK phosphorylation sites, i.e., charged amino acids on the NH₂ side of the tyrosine and a hydrophobic residue on the COOH side (220). The specific PTK involved in regulation of Kv3.x is not known, although Kv3.1 and Src do not coimmunoprecipitate (275).

Maxi-K_Ca Channels

Maxi-K_Ca is also regulated by tyrosine phosphorylation (Table 2). In smooth muscle, STOCs, which reflect the activity of maxi-K_Ca channels in that cell type, are activated by EGF and inhibited by genistein (70). In neurons, NT3 and NGF activate paxilline-sensitive BK_Ca channels (76). When the pore-forming -subunit of the BK_Ca channel, mSlo (the mammalian homologue of slo), is coexpressed with c-Src in HEK 293 cells, the channel undergoes tyrosine phosphorylation (125). Src has no effect on channel gating in the absence of Ca²⁺, but at free cytosolic Ca²⁺ levels >4 µM, channel P_o is higher at any given voltage when active Src is coexpressed. In contrast, expression of kinase-dead Src prevents tyrosine phosphorylation of the channel and the enhancement in Ca²⁺ sensitivity (125). The site of regulation appears to be Tyr⁷⁶⁶ on the COOH-terminus because a Tyr⁷⁶⁶ to Phe⁷⁶⁶ mutation prevents phosphorylation of the channel by Src and the subsequent shift in Ca²⁺ sensitivity (125). These experimental approaches show convincingly that Src is required for tyrosine phosphorylation of the maxi-K_Ca channel but, as with other studies, it is not certain if the channel is directly phosphorylated by Src or by an intermediate PTK.

	REGULATION OF CA2+ CHANNELS BY PTK

TOP ABSTRACT INTRODUCTION RECEPTOR AND NONRECEPTOR... TYPES OF SUPPORTING EVIDENCE REGULATION OF LIGAND-GATED... REGULATION OF K+ CHANNELS... REGULATION OF CA2+ CHANNELS... REGULATION OF NA+ CHANNELS... REGULATION OF RACC AND... REGULATION OF CL CHANNELS... THE CHANNEL-PROTEIN KINASE... CONCLUSIONS AND FUTURE... REFERENCES

Voltage-gated Ca²⁺ channels are composed of ₁ (190-250 kDa)-, ₂ (125 kDa)-, and  (52-62 kDa)-subunits (47). The ₁-subunit forms the voltage-gated Ca²⁺-selective pore, with some 2,000 amino acids organized in four homologous domains. Each domain contains six transmembrane segments with a membrane reentrant pore loop in each segment (141). Members of the dihydropyridine class of Ca²⁺ antagonists interact with the ₁-subunit (224).  and ₂ channel subunits increase functional expression and therefore are likely to play roles in membrane targeting of the channel complex. These subunits also acutely modulate channel function. The -subunit is cytoplasmic and contains a highly conserved domain that interacts with the I-II loop of ₁. The -subunit is an integral membrane protein, whereas ₂ is an extracellular glycolsylated protein (106). A -subunit (25 kDa) is found in some tissues, notably skeletal muscle, but its function is uncertain.

At each end of the ₁ channel protein are large cytosolic NH₂-terminal and COOH-terminal domains that are potential targets for protein phosphorylation. In cardiac and brain isoforms, _1C undergoes posttranslational proteolytic processing, after which a 30- to 50-kDa fragment of the COOH-terminus becomes colocalized with the body (190 kDa) of the _1C- and -subunits (63). The COOH-terminus is involved in CaM binding and Ca²⁺-dependent inactivation. Ser¹⁹²⁸ near the COOH-terminus is a major site of PKA-mediated phosphorylation (38). The mechanism of regulation of _1C by PKC is still unclear (106) but requires the NH₂-terminus (212). However, when recombinant channels are expressed in HEK 293 cells, phosphorylation of Thr²⁷ and Thr³¹ by PKC results in inhibition (not activation, as might be expected) of current (141). Because the NH₂- and COOH-terminal domains may constitutively serve as partially independent but interacting inhibitory gates (212), it is possible that PKC and other kinases, including PTKs, act by relieving this inhibition.

Several lines of evidence link PTKs to the regulation of the pore-forming subunits of voltage-gated calcium channels, including _1C (132), _1D (222), and _1E (11). Much of this work is summarized in Table 3.

At least two isoforms of the L-type Ca²⁺ channel are regulated by tyrosine phosphorylation. In vascular myocytes, Ca²⁺ current (_1C-b, Cav1.2b) is reduced by PTK inhibition (127, 260), whereas PTP inhibition increases current (264). PDGF, which stimulates tyrosine phosphorylation of multiple proteins, enhances L-type Ca²⁺ current (84, 261). A key role for Src kinase in this process is supported by observations that Ca²⁺ current is 1) increased by intracellular application of constitutively active Src kinase (262), 2) increased by c-Src-activating peptide (263), and 3) inhibited by a monoclonal antibody for c-Src (84). In visceral smooth muscle, PDGF potentiates L-type Ca²⁺ current and increases tyrosine phosphorylation of _1C (84). Additionally, _1C coimmunoprecipitates with c-Src, and PDGF regulation of current is inhibited by dialysis with antibodies to FAK or Src (84).

Cav1.2b is also regulated by at least three integrins. Whole cell recordings from single arteriolar myocytes show that soluble (nonclustering) ligands of v₃-integrin, such as RGD peptides, inhibit L-type Ca²⁺ current (271). Interestingly, bound ligands of v₃-integrin, which lead to integrin clustering (146), produce a similar inhibitory effect (271). In contrast, soluble ligands of the laminin receptor, ₄₁, enhance current (248), whereas only clustered ₅₁-integrin ligands enhance current (271). Therefore, regulation of this channel in smooth muscle occurs by multiple intracellular pathways downstream from v₃-, ₅₁-, and ₄₁-integrins, and at least two of these pathways do not appear to require integrin clustering. As for PDGF-potentiated currents, targeting of FAK or Src in these vascular smooth muscle cells with specific antibodies blocks potentiation of current by ₅₁-integrin ligands (270). Antibodies to two integrin-associated cytoskeletal protein targets, paxillin and vinculin, also block regulation of L-type Ca²⁺ current by ₅₁ ligands. The ability of vinculin and paxillin antibodies to do this is likely due to their interference with the assembly of Src or another PTK on an intracellular scaffold of focal adhesion proteins rather than a direct interaction with the channel. SH2 and SH3 domains in these proteins enable them to associate with PTKs (247). As mentioned above, the COOH-terminus of Cav1.2b contains a proline-rich domain that has been identified as a mediator of membrane association. This region interacts with SH3 domains in Src, Lyn, and Hck tyrosine kinases (63), and its deletion results in increased channel current, suggesting that it is constitutively involved in channel inhibition (258). It is not yet known whether phosphorylation of the _1C COOH-terminus by any of these PTKs alters this inhibitory property.

Depolarization recruits the neuronal L-type Ca²⁺ channel (Cav1.2c) and results in increased tyrosine phosphorylation of multiple cytoplasmic proteins, including FAK and vinculin. PP1 blocks calcium influx through the channel and reduces phosphorylation of these proteins. Conversely, the phosphatase inhibitor sodium orthovanadate mimics the depolarization effect and recruits the channel (48). In molluscan neurons, NGF acutely enhances high voltage-activated Ca²⁺ currents (266). Interestingly, the extracellular matrix protein fibronectin produces a similar effect and, in addition, enhances the frequency of action potential firing (W. Wildering, personal communication). Collectively, these observations suggest that Cav1.2c is functionally regulated by tyrosine phosphorylation.

Definitive evidence for regulation of Cav1.2c by the growth factor IGF-1, through Src, was recently presented. In cerebellar granule neurons, _1C is tyrosine phosphorylated in response to IGF-1, resulting in potentiation of L-type Ca²⁺ current (21, 204). This effect can be duplicated in recombinant _1C-c channels expressed in SH-SY5Y cells. Potentiation occurs primarily as a result of increasing the rate of activation at hyperpolarized potentials and likely reflects altered gating of the ₁-subunit. c-Src mediates this response because expression of kinase-dead Src, or application of the soluble Src inhibitor PP2, blocks potentiation of current (16). Kinase assays using lysates from neuroblastoma cells expressing _1C-c show that purified Src kinase phosphorylates Tyr²¹²² of the _1C-c COOH-terminus. Furthermore, point mutation of Tyr²¹²² to Phe²¹²² (but not of a nearby Tyr residue) prevents tyrosine phosphorylation and prevents IGF-1 potentiation of current (16). Because Cav1.2b and Cav1.2c share identical sequences in this region, it is likely that Tyr²¹²² [rather than other postulated sites (84, 111)] also mediates PDGF potentiation of current in smooth muscle. In addition to this tyrosine phosphorylation site, the intracellular loop of _1C contains two SH3 domains that may be involved in docking of Src. Upstream of Tyr²¹²² are several amino acid residues that correspond to known sequences of Src substrates, but the various docking, adaptor, or scaffolding proteins that may be required for tyrosine phosphorylation of this channel are not yet known.

T-type Ca²⁺ current in spermatogenic cells is potentiated by 50% after strong depolarization (e.g., a prepulse to +60 mV) or high-frequency stimulation (11). Unlike voltage-dependent potentiation of current in other Ca²⁺ channel types, the phenomenon in this tissue is not consistent with a Ca²⁺- or G protein-mediated mechanism. The same degree of potentiation is produced by PTK inhibitors, whereas PTP inhibitors block prepulse potentiation of current (11). Therefore, tyrosine phosphorylation/dephosphorylation may play a role in acute modulation of this channel.

	REGULATION OF NA+ CHANNELS BY PTKS

TOP ABSTRACT INTRODUCTION RECEPTOR AND NONRECEPTOR... TYPES OF SUPPORTING EVIDENCE REGULATION OF LIGAND-GATED... REGULATION OF K+ CHANNELS... REGULATION OF CA2+ CHANNELS... REGULATION OF NA+ CHANNELS... REGULATION OF RACC AND... REGULATION OF CL CHANNELS... THE CHANNEL-PROTEIN KINASE... CONCLUSIONS AND FUTURE... REFERENCES

Evidence for the regulation of two major classes of Na⁺ channels, voltage-gated and epithelial Na⁺ channels, is summarized in Table 4.

Voltage-Gated Na+ Channels

In the rat PC12 neuronal cell line, the growth factors NGF, PDGF, EGF, and basic FGF acutely inhibit VGNa (75). Inhibition is achieved through a leftward shift of the Na⁺ channel inactivation curve, rendering the channel more inactivated at physiological potentials. Inhibition depends on tyrosine kinase activity of the associated growth factor receptors because the receptor PTK inhibitors tyrphostin AG9 and tyrphostin AG879 almost completely reverse this effect. Furthermore, the inhibition of Na⁺ current by NGF and PDGF is not additive, suggesting a common signaling pathway (but this conclusion may also depend on whether maximal doses of the growth factors were used). Although the specific PTKs involved have not been determined, the Src inhibitor PP1 decreases the response to PDGF in wild-type cells, and PDGF receptors with mutations in their Src-binding domains are less effective in mediating the PDGF response (75).

A recent study of neuronal Na⁺ channels is one of the few to show a direct association of a PTP with an ion channel. Receptor PTP-, a receptor PTP whose mRNA is highly expressed in the brain and kidney (195), interacts with neuronal Na⁺ channel - and ₁-subunits through intracellular domains of the proteins (187). In HEK 293 cells, the recombinant -subunit is tyrosine phosphorylated under basal conditions, and the level of phosphorylation increases after treatment with the phosphatase inhibitor sodium pervanadate. Coexpression of -subunits with receptor PTP- intracellular phosphatase domains results in an increase in Na⁺ current due to a rightward shift in voltage-dependent inactivation. Thus interaction of the phosphatase and the channel leads to a reversal of the inhibition by tyrosine phosphorylation (75, 187). The extent to which the actions of endogenous tyrosine kinases and phosphatases differentially modulate VGNa under basal conditions, or in response to hormonal modulation, is not yet known.

Epithelial Na+ Channels

Many studies have documented the effects of growth factors on enhancing expression of ion channels, including ENaC, in transporting epithelia. A few of these studies have also noted that PTK inhibitors have apparently acute effects on channels involved in osmoregulation (234). For example, in renal A6 cells, genistein and tyrphostin A23 abolish macroscopic ENaC currents that are elicited by hyposmotic swelling (159). However, this effect is mimicked by an inhibitor of protein translocation, brefeldin A, suggesting that a PTK might not be regulating the channel directly but regulating its insertion into the plasma membrane. It should be noted that interpretation of some of these studies is complicated by the fact that several types of channels interact to determine epithelial water and solute transport, and, in many cases, direct measurements from specific channels have not been made. For example, the RVD of epithelial cells in response to osmotic swelling is sensitive to PTK inhibition (234), but RVD depends on the summed interaction of multiple channels. In addition, these channels functionally interact with transporters, several of which are also known to be sensitive to PTK inhibitors (134, 158).

	REGULATION OF RACC AND CATION CHANNELS BY PTKS

TOP ABSTRACT INTRODUCTION RECEPTOR AND NONRECEPTOR... TYPES OF SUPPORTING EVIDENCE REGULATION OF LIGAND-GATED... REGULATION OF K+ CHANNELS... REGULATION OF CA2+ CHANNELS... REGULATION OF NA+ CHANNELS... REGULATION OF RACC AND... REGULATION OF CL CHANNELS... THE CHANNEL-PROTEIN KINASE... CONCLUSIONS AND FUTURE... REFERENCES

In many types of cells, Ca²⁺-permeable channels can be activated by agonist-receptor interaction, where the receptor protein is distinct from the channel protein. This broad group of channels varies widely in conductance, voltage sensitivity, ion selectivity (e.g., Ca²⁺ vs. monovalent cation permeability), mechanism of activation, and Ca²⁺-induced inactivation. Most of these channels have not yet been cloned, so it is difficult to classify them and to define the precise mechanisms by which they are regulated. Barritt (15) has proposed that the commonly used term "receptor-activated Ca²⁺ channel" (RACC) be used to include non-voltage-gated, Ca²⁺-permeable channels that are activated by Ca²⁺ store depletion or second messengers, including Ca²⁺, Ins(1,4,5)P₃, cyclic nucleotides, and G proteins. We will adhere to this convention.

A complete description of the evidence for regulation of RACCs by tyrosine phosphorylation is difficult to compile and categorize, and the reader is referred to recent reviews for more information (15, 53, 167, 240). Table 5 summarizes much of the available data, although it should be noted that, in many cases, electrophysiological measurements were not made and the identity of the specific Ca²⁺ or cation channel is uncertain.

Store-Operated Channels

Stimulation of receptor or nonreceptor PTKs is a well-known stimulator of Ca²⁺ influx in nonexcitable cells (19, 202), and there are numerous reports that PTK inhibitors interfere with capacitative Ca²⁺ influx (152). The majority of these studies are documented in Table 5.

The first electrophysiological description of a store-operated Ca²⁺ current was of a small inward current in mast cells, termed I_crac (82). I_crac is activated after store depletion by Ins(1,4,5)P₃, EGTA, or SERCA inhibitors, is highly Ca²⁺ selective, voltage independent, and has an extremely low single channel conductance that can only be estimated from noise analysis (83, 283). I_crac is distinct from other cation-permeable channels, and no specific pharmacological antagonist has been identified (121, 167). Some investigators have distinguished at least two different non-voltage-gated Ca²⁺ currents in the same cell type (52, 55), including I_crac. In some nonexcitable cells, store-operated Ca²⁺ currents appear to be less selective for divalent over monovalent cations than I_crac (54, 205, 242) and occasionally display measurable single channel currents (242). Some investigators maintain that these differences represent two distinct types of currents rather than a single family of store-operated currents (167).

At least one report suggests that I_crac may be regulated, perhaps weakly, by PTKs (168). Resolution of this issue is limited by the fact that the molecular nature of the native store-operated channel has not been definitively resolved. The best candidates for a store-operated channel are the Trp family of proteins isolated from Drosophila photoreceptors (150) and from their human homologs (259, 281). Recombinant Trp channels share many, but not all, characteristics of I_crac (149, 167). However, heterologous combinations of Trp isoforms in native cells may account for the unique characteristics of endogenous store-operated currents (93, 176). The mechanism of activation of store-operated channels is highly controversial, but recent evidence links their activation with possible conformational coupling of the Ins(1,4,5)P₃ receptor to Trp channels (130, 169). This process requires the cytoskeleton because it is disrupted by cytochalasin or by cortical condensation of actin (169, 190). One Trp isoform has even been shown to contain intrinsic kinase activity (193).

Soluble PTK inhibitors such as genistein, lavendustin A, tyrphostin 47, and MDHC impair capacitative Ca²⁺ entry (8, 99, 163, 168, 175, 196, 197, 206, 280), whereas vanadate enhances capacitative Ca²⁺ entry (280). On the basis of these effects, it is proposed that the levels of cytosolic and stored Ca²⁺ antagonistically control tyrosine phosphorylation of a store-operated channel and/or protein involved in regulation of that channel (246). Indeed, in platelets and fibroblasts, store-operated Ca²⁺ influx is associated with tyrosine phosphorylation of a 130-kDa protein (116, 196, 246). Unfortunately, few of these studies have used electrophysiological methods to measure store-operated current (except Refs. 168, 175, and 206) and instead infer properties of the underlying channel(s) from measurements of cytosolic [Ca²⁺]. Thus a common concern with these studies is that Ca²⁺ influx may have changed with changes in membrane potential due to effects of the inhibitors on other channels. In fact, it is likely that these inhibitors have effects of K⁺ channels because of the rigorous evidence supporting regulation of K⁺ channels by tyrosine phosphorylation (Table 2). The observation that in several of these studies some, but not most, soluble PTK inhibitors inhibit capacitative Ca²⁺ influx (8, 167, 175) raises concern about inhibitor specificity. Alternatively, inhibitors that do not affect store-operated Ca²⁺ influx or current may simply be acting on PTKs that are not involved the regulation of store-operated channels. Thus a definitive role for tyrosine phosphorylation in the regulation of store-operated Ca²⁺ channels, including I_crac, remains to be proven.

However, there is substantial evidence that a nonselective cation current in vascular smooth muscle is regulated by tyrosine phosphorylation. In cells isolated from the rabbit portal vein, dialysis with a Src-activating peptide or bath application of insulin activates a 20-pS inward cation current (4). This current has the same characteristics as the current activated by noradrenaline, and the noradrenaline current is inhibited by soluble tyrosine kinase inhibitors. Inhibition of PTPs by sodium orthovanadate also activates the cation current (4). Thus receptor and nonreceptor tyrosine kinases appear to regulate a cation current in this tissue. A recent study provides evidence that this channel may represent a combination of one or more Trp isoforms (93).

CNG Cation Channels

One type of CNG channel is a voltage-gated cation channel found in bag cell neurons of Aplysia. In this cell, elevations in cAMP trigger a depolarization that leads to repetitive action potential firing. The cation channel is regulated by an endogenous PTP (267), and there is an interesting interaction between serine phosphorylation and tyrosine phosphorylation. The tyrosine phosphorylation state of the channel determines the response of the channel to PKA phosphorylation, i.e., whether PKA will produce an increase or decrease in P_o (267).

In expressed CNG channels, PTK inhibitors, but not serine-threonine inhibitors, enhance channel sensitivity to cGMP (148). The -subunit of bRET contains a critical phosphorylation site, Tyr⁴⁹⁸, on the COOH-terminus that is required for altering cGMP sensitivity (147). Dephosphorylation of this tyrosine residue enhances channel P_o. Mutation of this site, Tyr⁴⁹⁸ to Phe⁴⁹⁸, but neither of two other COOH-terminal tyrosine residues nor any of several NH₂-terminal residues required for CaM regulation, abolishes ~75% of the regulation by cGMP. Moreover, introduction of a tyrosine residue at a comparable site, Phe⁴⁷⁷ to Tyr⁴⁷⁷, on the rOLF channel confers partial cGMP sensitivity to the otherwise insensitive native rOLF channel (147). It is interesting that the predicted position of this tyrosine residue is between the ligand-binding site and the rest of the channel protein, i.e., in a strategic position to modulate allosteric gating transitions (147). These results firmly support the idea that the CNG channel is regulated by tyrosine phosphorylation, probably at multiple sites, because mutation of this COOH-terminal tyrosine determines most, but not all, of the cGMP sensitivity. The role of other possible tyrosine phosphorylation sites on this channel and the importance of this mechanism in vivo remain to be determined.

	REGULATION OF CL CHANNELS BY PTK

TOP ABSTRACT INTRODUCTION RECEPTOR AND NONRECEPTOR... TYPES OF SUPPORTING EVIDENCE REGULATION OF LIGAND-GATED... REGULATION OF K+ CHANNELS... REGULATION OF CA2+ CHANNELS... REGULATION OF NA+ CHANNELS... REGULATION OF RACC AND... REGULATION OF CL CHANNELS... THE CHANNEL-PROTEIN KINASE... CONCLUSIONS AND FUTURE... REFERENCES

Non-ligand-gated Cl channels have been grouped into four families: 1) Ca²⁺-activated Cl channels, 2) voltage-dependent Cl channels, 3) ClC voltage-gated Cl channels, and 4) the CFTR channel (58, 244). Many of the first two classes of Cl channels have not yet been purified and cloned, so identification is based completely on their ion selectivity, pharmacology, current-voltage relationship, and Ca²⁺ sensitivity. In contrast, much is known about the structure and function of ClC and CFTR channels (100, 110).

ClC channels have membrane topography quite different from voltage-gated cation channels in that they contain 11 transmembrane domains. Yet, like cation channels, both NH₂- and COOH-termini are intracellular. ClC-0 channels have two identical, yet independent, conductance levels, leading to the conclusion that there are two pores in each channel (244). ClC-2 and ClC-3 are thought to conduct the swelling-activated Cl current observed in many cell types. A serine residue (S⁵¹) in the ClC-3 NH₂-terminus has recently been shown to be a consensus PKC phosphorylation site that determines volume sensitivity (46).

CFTR is a cAMP-regulated Cl channel that controls salt and water transport across epithelia. It is composed of five membrane-spanning domains and two nucleotide-binding domains linked to a unique, regulatory (R) domain (211). The cytosolic R domain of CFTR contains multiple serine phosphorylation sites, at least five of which are phosphorylated by PKA to stimulate channel activity. PKC phosphorylation may play a facilatory role in PKA phosphorylation of CFTR (211). The NH₂-terminus of CFTR binds to the scaffolding protein syntaxin 1A, which appears to exert a tonic inhibitory role on the channel. The COOH-terminus of CFTR binds to several PDZ domain proteins that could potentially link to PTKs and cytoskeletal proteins. It has been proposed that this region of the channel may be involved in formation of a multiprotein signaling complex (110).

At least one study has suggested that CFTR gating is regulated by tyrosine phosphorylation independent of PKA (56). pp60^Src was implicated in channel regulation, but a putative tyrosine phosphorylation site was not identified. Furthermore, it is not known if the CFTR NH₂- or COOH-termini contain a Src recognition site. It is possible that a PTK only indirectly regulates CFTR function, e.g., by phosphorylating PKC, in a similar manner to the mechanism by which tyrosine phosphorylation of PLC- modulates Cl current in mesangial cells (132). This would be consistent with the inability of some investigators to detect phosphotyrosines in biochemical assays of CFTR (177).

Specific mechanisms by which other Cl channels may be regulated by tyrosine phosphorylation are largely unknown (58, 108). As with other channels, a number of studies have shown that Cl currents are modulated by soluble PTK and PTP inhibitors. This literature is summarized in Table 6.

View this table: [in this window] [in a new window]   Table 6.   Cl and CFTR channels

One series of studies stands out. Lepple-Wienhues et al. (118, 119) showed that a native swelling-activated Cl channel in lymphocytes (presumably a ClC family member) is regulated by p56^Lck tyrosine kinase (118). This channel is activated by osmotic swelling and by cAMP, and the process is blocked by genistein or lavendustin A. Furthermore, osmotic swelling produces an increase in p56^Lck activity (118). Addition of purified p56^Lck activates a Cl conductance in the absence of swelling. Osmotic activation of the channel is defective in lymphocytes from p56^Lck knockout animals but rescued after transfection with cDNA for p56^Lck. Although it is not known if the channel is tyrosine phosphorylated by p56^Lck or in response to swelling, the collective evidence strongly suggests that Lck regulates this Cl channel. In a subsequent study, it was shown that osmotic swelling or purified p56^Lck can activate the same channel in lymphocytes from patients with cystic fibrosis, suggesting that the channel may partially compensate for defective CFTR (119). Regulation of other ClC channels by PTKs, or even whether other ClC channels contain potential tyrosine phosphorylation sites on their NH₂- or COOH-termini, are issues that remain to be resolved.

	THE CHANNEL-PROTEIN KINASE REGULATORY COMPLEX

TOP ABSTRACT INTRODUCTION RECEPTOR AND NONRECEPTOR... TYPES OF SUPPORTING EVIDENCE REGULATION OF LIGAND-GATED... REGULATION OF K+ CHANNELS... REGULATION OF CA2+ CHANNELS... REGULATION OF NA+ CHANNELS... REGULATION OF RACC AND... REGULATION OF CL CHANNELS... THE CHANNEL-PROTEIN KINASE... CONCLUSIONS AND FUTURE... REFERENCES

It is becoming clear that regulation of an ion channel by a protein kinase requires the formation of a multiprotein complex. While pore-forming -subunits of many channels bind to auxiliary channel subunits, they also associate with scaffolding proteins that play essential roles in channel localization and activity (124). Scaffolding proteins link signaling enzymes, substrates, and potential effectors (such as channels) into a multiprotein signaling complex that may be anchored to the cytoskeleton. Besides an obvious role in targeting the channel to a particular location on the cell membrane, there are at least three advantages to having an ion channel in a multiprotein complex. First, there is a large increase in the efficiency of the kinetic reaction when an enzyme is localized with its substrate and effector in a microenvironment with restricted diffusion (188). Second, the anchoring of enzyme complexes to some channels may be necessary for the extremely rapid transmission of signals required to regulate some channels (273). Third, compartmentalization may be essential to determining specificity in signal transduction pathways (37).

There appear to be many families of scaffolding and adaptor proteins that could potentially be involved in organizing ion channels into signaling complexes and regulating function by coupling the channels to protein kinases. For serine-threonine kinases, the prominent families are the MAGUK protein, AKAP, and GKAP families (73). Another family of proteins that specifically associates with Kv4.x channels has also recently been described (73). Many scaffolding protein-channel interactions have been identified using yeast two-hybrid screens followed by subsequent immunoprecipitation and GST-fusion protein strategies. In most cases, however, the evidence for an anatomic relationship between the channel and other proteins is stronger than the evidence for a functional relationship. It is beyond the scope of this review to exhaustively discuss the evidence for channel-associated proteins in ion channel regulation (207), but examples from several classes of scaffolding proteins are worth mentioning because PTKs are also associated with some of the same proteins.

MAGUK proteins contain three types of domains that function as sites for specific protein-protein interactions: PDZ, SH3, and GK-like domains. Adjacent GK and SH3 domains, in which the GK domains appear to lack enzymatic activity (180), are defining features of MAGUKs. SH3 domains are proline-rich amino acid sequences found in many signaling and cytoskeleton-associated molecules (170). PDZ domains contain 80-100 amino acids and have been found in many proteins, including 30% of G protein-coupled receptors (186). Although MAGUK proteins are involved in directing protein associations apart from ion channels, there is evidence to support their role in the intracellular localization of several types of channels. Members of this family appear to be essential for proper organization of ion channels at neuronal synapses. For example, colocalization of Shaker family K⁺ channels with NMDA receptors in neuronal synapses requires PSD-95, SAP97, Chapsyn-110, SAP102, and Dlg, among other scaffolding proteins (208). There are many other examples encompassing both voltage-gated and ligand-gated families of channels (180, 207).

Importantly, the available evidence suggests that the association of scaffolding proteins with channels is part of a general mechanism that not only directs channel localization within a cell but is also essential for regulation of channel function. For example, localization of Shaker Kv1.4c is mediated by the binding of two PDZ domains of PSD-95 to the COOH-terminal cytoplasmic tail of the channel (109). Coexpression of PSD-95 with Kir2.3 in HEK 293 cells suppresses the activity of that channel by >50% (156). SAP97 is required for normal G protein sensitivity of Kir3.2c, an effect mediated through the GK domain of SAP97 (73). Syntaxin, which is abundant in nerve terminals, interacts with voltage-gated Ca²⁺ channels (18, 209, 210, 269) to shift their voltage dependence of inactivation in a direction that would inhibit presynaptic Ca²⁺ entry (18). One example that is not specific to the neuronal synapse is syntaxin 1A. That protein interacts with CFTR to tonically inhibit the channel (155). Future studies will likely reveal roles for these or related scaffolding/adaptor proteins in the normal functioning of other channels.

Some channel-associated proteins containing PDZ domains do not belong to the MAGUK protein family. Examples of these include GRIP (44) and Homer (26), which are important for localization of glutamate-gated channels. Most notably, INAD is a Drosophila protein containing five PDZ domains that link together most of the proteins involved directly in phototransduction. This multiprotein complex includes rhodopsin, CaM, the putative store-operated Ca²⁺ channels Trp and Trpl, and the protein kinases PLC and PKC (88, 273). INAD may serve as a template for understanding how other channels are regulated in a multiprotein complex.

Although the association of other ion channels with PDZ domain proteins and protein kinases is not as firmly established as for INAD, several ion channel-associated PDZ domain proteins do contain catalytic domains (180). Additionally, some of these same proteins contain other signaling domains, e.g., SH3, PH, PTB, PX, and DEP (180), implying that they also associate with other regulatory proteins. As mentioned above, SH3 domains are found in many cytoskeletal and focal adhesion proteins and in most PTKs (247). The ₁-subunit of L- and N-type Ca²⁺ channels has a conserved sequence of proline-rich residues (Prp²⁰³⁹-Pro²¹⁹⁴) containing a number of potential SH3-binding motifs (PxxP). This region encompasses part of the channel known to be involved in G protein regulation by -subunits (183) and has been shown to bind Src, Lyn, and Hck tyrosine kinases, although the functional significance of these interactions (except for Src, see Ref. 16) is not yet known (63). However, mutant _1C lacking the proline-rich domain exhibits enhanced cation current conduction, consistent with an autoinhibitory role for the COOH-terminus (258). Another sequence at the COOH-terminus of _1B contains a conserved DxxC-COOH motif that allows interaction with PDZ domain proteins such as the scaffolding protein Mint1 through a class III PDZ domain of the latter (137). Furthermore, a proline-rich domain elsewhere in Mint1 interacts with the SH3 domain of CASK, an adaptor protein that contains a calmodulin kinase domain (137). These observations point to the likely possibility that combinations of PDZ, SH3, and other signaling domain sequences may direct protein-protein interactions between ion channels and the kinases that regulate them.

Evidence that scaffolding proteins can mediate kinase-channel interactions is more firmly established for the AKAP family of proteins (66). AKAPs serve to localize both kinases and phosphatases to multiprotein effector complexes that include K⁺ and Ca²⁺ channels (41). Evidence for AKAP participation in ion channel regulation has been obtained using both molecular biology and electrophysiology approaches. A general scheme is that a conserved AKAP-anchoring motif directs dimerized PKA subunits to a particular subcellular target (102). For example, forskolin and cAMP potentiate ROMK1 (Kir1.x) current in native renal secretory cells (251). Activity is preserved after patch excision, in which the excised membrane is known to retain channel-associated proteins. However, this potentiation is lost when ROMK1 is expressed in oocytes but restored if AKAP79 is coexpressed with the channel (251). AKAP15/18 is required for PKA potentiation of L-type Ca²⁺ channels in cardiac (62), skeletal (103), and vascular smooth muscle (60). In skeletal muscle, AKAP15 also aids in localizing L-type Ca²⁺ channels to transverse tubules (65). AKAPs have also been implicated in the regulation of K_Ca channels (256) and CFTR (225). AKAP79 associates with both the ₂-adrenergic receptor and MAGUK proteins, and the latter are coupled to glutamate-gated ion channels. Although it has not yet been proven, AKAP79 has the potential to form part of a scaffold on which PKA and PP2B may dually regulate the coupling of ₂-adrenergic receptors and glutamate channels (41). Thus AKAPs appear to be capable of assembling signaling complexes by virtue of their associations with other scaffolding proteins. It is possible that AKAPs and INAD represent a general scheme for kinase regulation of channels and that similar families of adaptor proteins associated with other kinases, perhaps PTKs, will soon be identified.

	CONCLUSIONS AND FUTURE DIRECTIONS

TOP ABSTRACT INTRODUCTION RECEPTOR AND NONRECEPTOR... TYPES OF SUPPORTING EVIDENCE REGULATION OF LIGAND-GATED... REGULATION OF K+ CHANNELS... REGULATION OF CA2+ CHANNELS... REGULATION OF NA+ CHANNELS... REGULATION OF RACC AND... REGULATION OF CL CHANNELS... THE CHANNEL-PROTEIN KINASE... CONCLUSIONS AND FUTURE... REFERENCES

It should be apparent that there is strong collective evidence to support a role for regulation of ion channels by tyrosine phosphorylation, particularly for ligand-gated channels and voltage-gated cation channels. Our knowledge of the importance of this mechanism for the regulation of other channels will depend largely on the isolation, purification, and cloning of those channels.

On the basis of the evidence described herein, several remaining questions can be identified. What specific PTKs, receptor and nonreceptor, are directly responsible for tyrosine phosphorylation of a particular channel? Can more than one PTK control phosphorylation of a given tyrosine residue on a channel? Are there interactions between tyrosine phosphorylation sites on a channel (5)? Are there interactions between tyrosine phosphorylation sites and serine-threonine phosphorylation sites (267)? What specific endogenous tyrosine phosphatases are involved in ion channel regulation (187)?

Little information is currently available regarding scaffolding and adaptor proteins that might specifically be required for regulation of ion channels by tyrosine phosphorylation. This field lags considerably behind that for serine-threonine phosphorylation systems, but AKAPs, CFTR, and INAD systems may serve as useful models to guide future study.

Finally, a major issue needing to be resolved for virtually every class of ion channel is the physiological role for tyrosine phosphorylation. Many questions remain to be answered. Is there a significant impact of tyrosine phosphorylation on channel function, membrane potential, and/or [Ca²⁺]_i regulation? Does tyrosine phosphorylation produce the same effect on native channels in vivo as it does on recombinant channels in vitro? Is this type of regulation determined by or modified by auxiliary channel subunits (219)? We discussed evidence for physiological relevance, when known, in each of the above sections dealing with specific families of channels. In general, acute regulation of an ion channel by a growth factor (in addition to any effect on channel protein expression) may represent an adaptive mechanism that could be enhanced or unmasked in processes such as tissue repair or remodeling. Moreover, because signaling pathways for nonreceptor tyrosine kinases and growth factors overlap, and in some cases converge on ion channels, it is possible that the adhesion molecules linked to those kinases play a physiological role in the regulation of the channels. This implies that tyrosine phosphorylation may be one mechanism for transduction of mechanical forces through adhesion molecules to ion channels.