Molecular basis of functional voltage-gated K⁺ channel diversity in the mammalian myocardium

Abstract

In the mammalian heart, Ca²⁺-independent, depolarization-activated potassium (K⁺) currents contribute importantly to shaping the waveforms of action potentials, and several distinct types of voltage-gated K⁺ currents that subserve this role have been characterized. In most cardiac cells, transient outward currents, I_to,f and/or I_to,s, and several components of delayed reactivation, including I_Kr, I_Ks, I_Kur and I_K,slow, are expressed. Nevertheless, there are species, as well as cell-type and regional, differences in the expression patterns of these currents, and these differences are manifested as variations in action potential waveforms. A large number of voltage-gated K⁺ channel pore-forming (α) and accessory (β, minK, MiRP) subunits have been cloned from or shown to be expressed in heart, and a variety of experimental approaches are being exploited in vitro and in vivo to define the relationship(s) between these subunits and functional voltage-gated cardiac K⁺ channels. Considerable progress has been made in defining these relationships recently, and it is now clear that distinct molecular entities underlie the various electrophysiologically distinct repolarizing K⁺ currents (i.e. I_to,f, I_to,s, I_Kr, I_Ks, I_Kur, I_K,slow, etc.) in myocyardial cells.

The amplitudes and the durations of action potentials in the mammalian myocardium (Fig. 1) are largely determined by voltage-gated K⁺ channels and, in most cardiac cells, two broad classes of voltage-gated K⁺ channel currents have been distinguished based primarily on differences in time- and voltage-dependent properties: (1) rapidly activating and inactivating, transient outward K⁺ currents, I_to; and (2) slowly inactivating, outwardly rectifying K⁺ currents, I_K (Campbell et al. 1995; Barry & Nerbonne, 1996; Giles et al. 1996; Nerbonne, 1998; Snyders, 1999) (Table 1). These currents serve distinct roles in action potential repolarization: the transient currents (I_to), which activate and inactivate relatively rapidly, for example, underlie the early phase of action potential repolarization (Fig. 1), whereas the delayed rectifiers (I_K), which activate with variable rates and typically inactivate slowly, contribute to the later phase of membrane repolarization, back to the resting potential (Fig. 1). Nevertheless, these (I_to and I_K) are broad classifications, and there are multiple components of I_to and I_K (Table 1) in cardiac cells, with distinct properties and functional roles. In addition, there are differences in the expression patterns of the various voltage-gated K⁺ currents (Table 1) in cardiac cells isolated from different species, as well as in cells from different regions of the heart in the same species (Antzelevitch et al. 1994; Barry & Nerbonne, 1996), and these distinct K⁺ current expression patterns contribute to regional differences in action potential waveforms (Fig. 1).

Figure 1. Action potential waveforms are variable in different regions of the heart.

Schematic representation of the heart; action potential waveforms recorded in different regions of the heart are illustrated. Action potentials are displaced in time to reflect the temporal sequence of propagation through the heart. SA, sino-atrial; AV, atrio-ventricular; RV, right ventricle; LV, left ventricle.

Table 1. Voltagegated K⁺ currents/channels in the mammalian heart.

Current	Activation	Inactivation	Blocker	Tissue*	Species*
Ito
Ito,f	Fast	Fast	Millimolar 4AP	Atrium	Dog, human, mouse, rat
			Flecainide	Ventricle	Cat, dog, ferret, human, mouse, rat
			HaTX
			HpTX
Ito,s	Fast	Slow	Millimolar 4AP	Atrium	Rabbit
				Ventricle	Ferret, human, mouse, rabbit, rat
				Node	Rabbit
IK
IKr	Moderate	Fast	E4031	Atrium	Dog, guinea-pig, human, rat
			Dofetilide	Ventricle	Cat, dog, guineapig, human, mouse, rabbit, rat
			Lanthanum	Node	Rabbit
IKs	Very slow	No	NE10064	Atrium	Dog, guinea-pig, human
			NE10133	Ventricle	Dog, guinea-pig, human
				Node	Guinea-pig, rabbit
IKur	Fast	No	Micromolar 4-AP	Atrium	Dog, human, rat
IKp	Fast	No	Ba2+	Ventricle	Guinea-pig
IK	Slow	Slow	Millimolar TEA	Ventricle	Rat
IK,slow	Fast	Slow	Micromolar 4-AP	Atrium	Mouse
			Millimolar TEA	Ventricle	Mouse
IK,slow	Fast	Very slow	Millimolar 4-AP	Atrium	Human, rat
(IK,DTX)			DTX
Iss	Slow	No	Millimolar TEA	Atrium	Mouse, rat
			Millimolar 4-AP	Ventricle	Dog, human, mouse, rat

^*Currents with these properties have been identified in these tissues/species to date. 4-AP, 4-aminopyridine; HaTX, hanatoxin; HpTX, heteropodatoxin; TEA, tetraethylammonium; DTX, dendrotoxin.

In spite of the marked electrophysiological diversity of repolarizing K⁺ currents, the time- and voltage-dependent properties of the various I_to and I_K currents (Table 1) in different cardiac cell types (and species) are quite similar, observations interpreted as suggesting that the molecular correlates of the underlying K⁺ channels are also the same (Barry & Nerbonne, 1996). A number of voltage-gated K⁺ channel (Kv) pore-forming (α) and accessory (β, minK and MiRP) subunits have now been cloned from or shown to be expressed in heart, and the molecular diversity of Kv channel subunits is even greater than expected (Table 2) based on the recognized electrophysiological heterogeneity of these channels (Table 1). Nevertheless, considerable progress has been made recently in defining the relationships between these subunits and the functional voltage-gated K⁺ currents/channels in cardiac cells. Importantly, the results obtained to date suggest that distinct molecular entities underlie the various electrophysiologically distinct types of voltage-gated K⁺ currents/channels in myocardial cells (Table 2).

Table 2. Kvα subunits: relation to functional voltage-gated cardiac K⁺ currents.

Family	Subunit	Activation*	Inactivation*	Blocker*	Endogenous current
Kv1 (Shaker)	Kv1.1
	Kv1.2†	Fast	Slow	4AP	IK,slow (rat)
				DTX	(IK,DTX)
	Kv1.3
	Kv1.4†	Fast	Slow	4-AP	Ito,s
	Kv1.5†	Fast	Slow	4AP	IKur
					IK,slow (mouse)
	Kv1.6
	Kv1.7†	Fast	Fast	NTX	?
	Kv1.8
Kv2 (Shab)	Kv2.1†	Slow	Very slow	TEA, HaTx	IK,slow (mouse)
	Kv2.2†	Slow	Very slow	TEA	?
Kv3 (Shaw)	Kv3.1†	Fast	Slow	TEA	?
	Kv3.2
	Kv3.3
	Kv3.4
Kv4 (Shal)	Kv4.1
	Kv4.2†	Fast	Fast	4-AP, HaTx, HpTx	Ito,f
	Kv4.3†	Fast	Fast	4-AP,HpTx	Ito,f
Kv5–9
eag family
eag	eag
elk	elk
erg	erg1†	Moderate	Fast	E4031	IKr (with MiRP)
	erg2
	erg3
KvLQT1 family
KCNQ1	KvLQT1†	Very slow	Very, very slow	NE10064	IKs (with minK)
KCNQ2
KCNQ3

^*Properties determined for heterologously expressed subunits; detailed properties, however, do vary with expression system (see text).

^†Subunit expressed in mammalian myocardium. DTX, dendrotoxin; NTX, noxiustoxin; HaTX, hanatoxin; HpTX, heteropodatoxin.

Transient outward K⁺ current (I_to) diversity in cardiac cells

Although two transient outward current components, referred to as I_to1 and I_to2, were originally identified in cardiac cells (Kenyon & Gibbons, 1979a,b; Coraboeuf & Carmeliet, 1982), only I_to1 is a K⁺ current; I_to2 is a Cl⁻-selective, not a K⁺-selective, conductance pathway (Kenyon & Gibbons, 1979a,b; Zygmunt & Gibbons, 1991) and is not discussed further here. The Ca²⁺-independent, depolarization-activated outward K⁺ currents in cardiac cells have been referred to previously as I_to, I_to1 or I_t (Campbell et al. 1995; Giles et al. 1996; Barry & Nerbonne, 1996). Although initially assumed to reflect the same underlying K⁺ channel(s), it has now been clearly demonstrated that there are (at least) two distinct transient outward K⁺ currents, I_to,fast (I_to,f) and I_to,slow (I_to,s), and that these currents are differentially distributed (Brahmajothi et al. 1999; Wang et al. 1999; Wickenden et al. 1999a,b,c; Xu et al. 1999b).

The very rapidly inactivating, transient outward current in cardiac cells is now referred to as I_to,fast or I_to,f (Table 1) after Xu et al. (1999b). Currents that can be classified as I_to,f have been characterized in considerable detail in cat (Furukawa et al. 1990), dog (Litovsky & Antzelevitch, 1988; Tseng & Hoffman, 1989), ferret (Campbell et al. 1993), human (Wettwer et al. 1993, 1994; Konarzewska et al. 1995; Amos et al. 1996; Näbauer et al. 1996, 1998), mouse (Xu et al. 1999b) and rat (Apkon & Nerbonne, 1991; Himmel et al. 1999; Wickenden et al. 1999a,c) ventricular myocytes (Table 1), as well as in atrial cells from dog (Yue et al. 1996a), rat (Boyle & Nerbonne, 1991, 1992), human (Escande et al. 1987; Shibata et al. 1989; Fermini et al. 1992; Amos et al. 1996) and mouse (Xu et al. 1999c; E. Bou-Abboud, H. Li & J. M. Nerbonne, unpublished observations) (Table 1). Although guinea-pig ventricular myocytes reportedly lack I_to,f (Sanguinetti & Jurkiewicz, 1990, 1991), a rapidly activating and inactivating (I_to,f-like) outward K⁺ current is detectable in these cells when extracellular Ca²⁺ is removed (Inoue & Imanaga, 1993), suggesting that functional I_to,f expression may be highly regulated.

The time- and voltage-dependent properties of the currents classified as I_to,f in different cardiac cell types (Table 1) are similar in that activation, inactivation and recovery from steady-state inactivation are all rapid; the pharmacological properties of the currents are also similar (Table 1). The rapid rate of recovery of I_to,f from steady-state inactivation (τ≈ 30–50 ms) has been exploited to distinguish I_to,f and I_to,s (Brahmajothi et al. 1999; Wang et al. 1999; Xu et al. 1999b). It has also been demonstrated that I_to,f in rat (Sanguinetti et al. 1997), ferret (Brahmajothi et al. 1999) and mouse (Xu et al. 1999b) ventricular myocytes is selectively attenuated by heteropodatoxin-2 and −3. The fact that the properties of I_to,f in different cardiac cells are similar (Table 1) led to the suggestion that the molecular correlates of functional I_to,f channels in different cell types and/or species are also the same (Barry & Nerbonne, 1996), and considerable experimental support for this hypothesis has now been provided (see below). Nevertheless, there are some subtle differences in the properties of the currents referred to as I_to,f (Table 1). Comparison of rat atrial and ventricular I_to,f, for example, reveals significant differences in rates of inactivation and recovery (Apkon & Nerbonne, 1991; Boyle & Nerbonne, 1992).

In contrast to I_to,f, the rate of inactivation and recovery of the transient outward currents in rabbit atrial and ventricular myocytes is slow (Giles & Imaizumi, 1988; Clark et al. 1988; Hiraoka & Kawano, 1989; Fedida & Giles, 1991; Giles et al. 1996; Wang et al. 1999). In mouse ventricular myocytes, a slowly inactivating transient K⁺ current has also been identified and is referred to as I_to,slow or I_to,s (Xu et al. 1999b). Mouse ventricular I_to,s recovers from steady-state inactivation (τ_recovery≈ 1 s) much more slowly than I_to,f (τ_recovery≈ 50 ms) and is insensitive to the heteropodatoxins (Xu et al. 1999b). The properties of I_to (I_t) in rabbit myocytes are similar to those of mouse ventricular I_to,s, suggesting that the transient currents in rabbit should also be termed I_to,s (Table 1). The rates of inactivation and recovery of the transient outward K⁺ currents are significantly slower in endocardial than in epicardial cells isolated from ferret left ventricles, suggesting that distinct K⁺ channel α subunits also underlie the transient currents in these cells (Brahmajothi et al. 1999). The time- and voltage-dependent properties and the pharmacological sensitivities of the currents in ferret epicardial and endocardial left ventricular myocytes are similar to those of the mouse ventricular currents, and these currents are also now referred to as I_to,f and I_to,s, respectively (Table 1). There are also distinct transient outward K⁺ currents in human and in rat ventricular myocytes (Näbauer et al. 1996, 1998; Wickenden et al. 1999a,c) with properties similar to those of mouse and ferret I_to,f anf I_to,s (Table 1).

Delayed rectifier K⁺ current (I_K) diversity in cardiac cells

In most cardiac cells, multiple components of I_K are evident (Table 1). The detailed properties of the delayed rectifier K⁺ currents have been characterized in myocytes isolated from canine (Tseng & Hoffman, 1989; Liu & Antzelevitch, 1995; Yue et al. 1996a,a), feline (Furakawa et al. 1992), guinea-pig (Balser et al. 1990; Horie et al. 1990; Sanguinetti & Jurkiewicz, 1990, 1991; Walsh et al. 1991; Anumonwo et al. 1992; Freeman & Kass, 1993), human (Wang et al. 1994; Li et al. 1996), mouse (Fiset et al. 1997a; London et al. 1998a; Zhou et al. 1998; Xu et al. 1999b,c; Guo et al. 1999), rabbit (Veldkamp et al. 1993) and rat (Apkon & Nerbonne, 1991; Boyle & Nerbonne, 1992) heart. Two prominent components of I_K, referred to as I_Kr (I_K,rapid) and I_Ks (I_K,slow) for example, have been distinguished in guinea-pig atrial and ventricular myocytes. I_Kr activates and inactivates rapidly, displays marked inward rectification, and is selectively blocked by lanthanum and by several class III anti-arrhythmics, including dofetilide, E-4031 and sotalol (Sanguinetti & Jurkiewicz, 1990, 1991). I_Kr and I_Ks can also be distinguished at the microscopic level: in symmetrical K⁺, the single channel conductances of I_Kr and I_Ks channels are10–13 and 3–5 pS, respectively (Balser et al. 1990; Veldkamp et al. 1993). Similar currents have been described in atrial and ventricular myocytes isolated from human (Wang et al. 1993a, 1994; Li et al. 1996) and canine (Liu & Antzelevitch, 1995; Yue et al. 1996a) hearts, and in rabbit ventricular cells (Veldkamp et al. 1993; Salata et al. 1996).

In rat (Boyle & Nerbonne, 1991, 1992), human (Wang et al. 1993a,b) and dog (Yue et al. 1996a,b) atrial myocytes, rapidly activating, non-inactivating K⁺ currents (distinct from I_Kr and I_Ks) have been identified and referred to as I_ss (steady-state), I_sus (sustained) or I_Kur (ultra-rapid). Because the properties of the currents are similar and recent studies suggest that the molecular correlates of the rat and human currents are the same (Feng et al. 1997; Bou-Abboud & Nerbonne, 1999), all are now referred to as I_Kur (Table 1). In guinea-pig ventricular myocytes, a 12–14 pS K⁺-selective channel, distinct from I_Kr, has also been reported and referred to as I_Kp (Yue & Marban, 1988). I_Kp activates rapidly on depolarization, does not inactivate and is sensitive to millimolar concentrations of Ba²⁺ (Backx & Marban, 1993). The properties of this current are very similar to that of dog, human and rat I_Kur, suggesting that I_Kp probably also reflects the expression of I_Kur.

In some cardiac cell types, there are additional components of I_K with properties different from I_Ks, I_Kr and I_Kur (Table 1). In mouse ventricular and atrial myocytes, for example, two voltage-gated K⁺ currents have been identified and are referred to as I_K,slow and I_ss (Fiset et al. 1997a; London et al. 1998a; Zhou et al. 1998; Xu et al. 1999b,c; Guo et al. 1999; E. Bou-Abboud, H. Li & J. M. Nerbonne, unpublished observations). In spite of the terminology, I_K,slow, which is rapidly activating and slowly inactivating and is blocked selectively by micromolar concentrations of 4-aminopyridine (4-AP), is distinct from I_Ks (Fiset et al. 1997a; London et al. 1998a; Zhou et al. 1998; Xu et al. 1999b). The current remaining at the end of long depolarizing voltage steps, I_ss, is slowly activating and, although 4-AP insensitive, is blocked by tetraethylammonium (TEA) (Xu et al. 1999a,b). In rat ventricular and atrial myocytes, there are also novel delayed rectifier K⁺ currents, referred to as I_K and I_K,slow, respectively (Table 1) (Apkon & Nerbonne, 1991; Boyle & Nerbonne, 1991, 1992). In addition to differences in inactivation, I_K and I_K,slow can be distinguished pharmacologically: I_K is blocked selectively by TEA (Apkon & Nerbonne, 1991), whereas I_K,slow is blocked by millimolar concentrations of 4-AP (Boyle & Nerbonne, 1991, 1992) and by nanomolar concentrations of the dendrotoxins (Van Wagoner et al. 1996). For this reason, rat atrial I_K,slow is now referred to as I_K,DTX (Table 1).

Regional differences in voltage-gated K⁺ channel expression

In some species, there are marked differences in the densities of the K⁺ currents expressed in different myocardial cell types. In rat, for example, I_to,f density is higher in atrial than in ventricular myocytes (Boyle & Nerbonne, 1991, 1992), whereas in mouse, I_to,f density is higher in ventricular than in atrial cells (Xu et al. 1999c). The density of I_to,f also varies in different regions of the ventricles in canine (Litovsky & Antzelevitch, 1988; Liu et al. 1993), cat (Furukawa et al. 1992), ferret (Brahmajothi et al. 1999), human (Wettwer et al. 1994; Näbauer et al. 1996, 1998), mouse (Xu et al. 1999b; Guo et al. 1999) and rat (Clark et al. 1993; Wickenden et al. 1999a) heart. In canine left ventricles, for example, I_to,f density is 5- to 6-fold higher in epicardial and midmyocardial than in endocardial cells (Liu et al. 1993). I_to,f density is also reportedly higher in guinea-pig atrial epicardium than endocardium (Wang et al. 1991). I_to,s density is significantly higher in rabbit atrial than ventricular myocytes (Giles & Imaizumi, 1988), and the density of I_to,s varies markedly in ferret left ventricular myocytes (Brahmajothi et al. 1999). In mouse ventricles, I_to,f and I_to,s are also differentially distributed (Xu et al. 1999b; Guo et al. 1999): in cells isolated from the right ventricle or the apex of the left ventricle, only I_to,f is present, whereas cells isolated from the left ventricular septum express either I_to,s alone (∼20 % of the cells) or both I_to,f and I_to,s (∼80 % of the cells) (Xu et al. 1999b). In addition, when expressed, I_to,f density is significantly higher in apex than in septum cells (Xu et al. 1999b).

The densities of I_Ks and I_Kr in different myocardial cell types are also variable. In guinea-pig, for example, I_Kr and I_Ks densities are 2-fold higher in atrial than in ventricular myocytes (Sanguinetti & Jurkiewicz, 1991). In dog heart, I_Ks density is higher in epicardial and endocardial cells than in midmyocardial cells (Liu & Antzelevitch, 1995). There are also regional differences in I_Kr and I_Ks expression in guinea-pig left ventricle (Bryant et al. 1998; Main et al. 1998). In cells isolated from the guinea-pig left ventricular free wall, for example, the density of I_Kr is reportedly higher in subepicardial than in midmyocardial or subendocardial myocytes (Main et al. 1998), whereas at the base, the densities of I_Kr and I_Ks are lower in endocardial than in midmyocardial or epicardial cells. It seems certain that differences in voltage-gated K⁺ current densities contribute to the variations in action potential waveforms recorded in atrial and ventricular myocytes, as well as in cells isolated from different regions (apex versus base) and layers (epicardial, midmyocardial and endocardial) of the atria and ventricles (Fig. 1).

Molecular diversity of voltage-gated K⁺ channel pore-forming α subunits

The first voltage-gated K⁺ channel (Kv) pore-forming (α) subunit was cloned from the Shaker locus in Drosophila and this was followed quickly by the cloning of three homologous Kv α subunit subfamilies, referred to as Shab (Kv2), Shaw (Kv3) and Shal (Kv4) (Butler et al. 1989; Wei et al. 1990). Each Kv α subunit protein has six membrane-spanning regions, a K⁺-selective pore and a highly charged S4 domain (Fig. 2A), placing the Kv α subunits in the ‘S4 superfamily’ of voltage-gated ion channels (Jan & Jan, 1992; Pongs, 1992). Further diversity results from the presence of multiple members of each subfamily (Table 2) and through alternative splicing of transcripts (Pongs, 1992; Barry & Nerbonne, 1996; Deal et al. 1996; Coetzee et al. 1999). Heterologous expression of Kv α subunit cDNAs reveals voltage-gated K⁺ channel currents with distinct time- and voltage-dependent properties (Fig. 2B). Because functional voltage-gated K⁺ channels comprise four α subunits (Fig. 2C) (MacKinnon, 1991), further diversity could arise through the formation of heteromultimeric channels between two or more Kv α subunit proteins (Covarrubias et al. 1991).

Figure 2. Structure, expression and assembly of voltage-gated K⁺ channel pore-forming (α) and accessory (minK and β) subunits.

A, Kv α subunits are integral membrane proteins with six transmembrane domains, intracellular N- and C-termini and a positively charged S4 region, placing them in the ‘S4 superfamily’ of voltage-gated ion channels. B, schematic representation of outward K⁺ current waveforms produced on heterologous expression of a Kv α subunit of the Shal subfamily, Kv4.2, which gives rise to K⁺ currents that activate and inactivate rapidly on membrane depolarization. C, schematic diagram of a voltage-gated K⁺ channel illustrating the assembly of four Kv α subunits to produce a K⁺-selective pore. D, schematic diagram of functional voltage-gated cardiac K⁺ channels with Kv α subunits depicted in blue, cytosolic β (or KChAP) subunits in yellow and minK (or MiRP1) subunits in red.

Additional Kv α subunit subfamilies, Kv5.x to Kv9.x (Table 2), have also been identified (Drewe et al. 1992; Hugnot et al. 1996; Salinas et al. 1997). Although heterologous expression of these subunits (alone) does not reveal functional voltage-gated K⁺ channel currents, coexpression with Kv2.x subfamily members attenuates the amplitude of the Kv2.x-induced currents (Salinas et al. 1997). These observations, together with sequence similarities to the Kv2.x subfamily suggest that Kv5.x to Kv9.x may be regulatory (Kv α) subunits of the Kv2.x subfamily (Castellano et al. 1997), rather than distinct subfamilies (Salinas et al. 1997). The role of these subfamilies in the generation of functional cardiac K⁺ channels is presently unclear.

Homologous subfamilies of voltage-gated K⁺ channel α subunit genes were revealed with the cloning of the Drosophila ether-à-go-go (eag) locus (Warmke et al. 1991) and with the identification of the eag-like gene (elk) and the eag-related gene (erg) in brain (Warmke & Ganetzky, 1994). Although the overall amino acid identity is low (10-15 %), the S4 and pore regions of the proteins encoded by these genes are homologous (Warmke et al. 1991) and the predicted membrane topologies of the proteins are similar to the Kv α subunits (Fig. 2A). erg1 is the locus of mutations leading to one form of familial long QT-syndrome, LQT2, and heterologous expression of erg1 reveals inwardly rectifying K⁺ currents (Sanguinetti et al. 1995; Trudeau et al. 1995) similar to cardiac I_Kr. Additional members of the erg subfamily, erg2 and erg3, have also been identified (Table 2), although these are not expressed in heart (Shi et al. 1997). Another subfamily of voltage-gated K⁺ channel α subunits was identified with the cloning of KvLQT1, the locus of mutations in LQT1. Heterologous expression of KvLQT1 produces rapidly activating K⁺ currents, whereas coexpression of KvLQT1 with minK (see below) produces slowly activating K⁺ currents that resemble cardiac I_Ks (Barhanin et al. 1996; Sanguinetti et al. 1996). Additional members of the KvLQT (KCNQ) subfamily, KCNQ2 and KCNQ3, have been cloned from brain and identified as loci of mutations that lead to benign familial neonatal convulsions (Biervert et al. 1998; Wang et al. 1998; Schroeder et al. 1998). Expression of KCNQ2 or KCNQ3 in Xenopus oocytes produces K⁺-selective currents that deactivate very slowly on membrane repolarization, and it has been suggested that KCNQ2-KCNQ3 heteromultimers underlie neuronal M currents (Wang et al. 1998). Neither KCNQ2 nor KCNQ3, however, appears to be expressed in the heart (Schroeder et al. 1998; Wang et al. 1998).

Accessory subunits of voltage-gated K⁺ channels

Further K⁺ channel diversity arises through the association of accessory subunits, such as minK, with the pore-forming K⁺ channel α subunits. Although heterologous expression of minK, which is a single transmembrane protein, in Xenopus oocytes was shown to result in voltage-gated K⁺ currents (Murai et al. 1989; Folander et al. 1990; Goldstein & Miller, 1991), subsequent studies revealed that these K⁺ currents arise from coassembly of minK with the Xenopus homologue of KvLQT1 (Sanguinetti et al. 1996); minK does not, by itself, form functional voltage-gated K⁺ channels. Rather, minK coassembles with KvLQT1 (Fig. 2D) to form K⁺ channels similar to cardiac I_Ks (Barhanin et al. 1996; Sanguinetti et al. 1996).

Another type of voltage-gated K⁺ channel accessory subunit was revealed with the molecular cloning of low molecular mass (∼45 kDa) accessory β subunits (β1 and β2) from brain (Rettig et al. 1994). Homologous Kv β subunits, Kv β₁, Kv β₂ and Kv β₃, and alternatively spliced transcripts of Kv β_1.x, have also been identified (Castellino et al. 1995; England et al. 1995; Majumder et al. 1995; Morales et al. 1995; Deal et al. 1996). These β subunits are cytosolic proteins that interact with the intracellular domains of Kv α subunits in assembled voltage-gated K⁺ channels (Fig. 2D). In heterologous expression systems, coassembly with β subunits can affect the time- and voltage-dependent properties of Kv α subunit-induced currents (Castellino et al. 1995; England et al. 1995; Majumder et al. 1995; Morales et al. 1995; Accili et al. 1997), as well as the functional expression of voltage-gated K⁺ channels on the cell surface (Shi et al. 1996; Accili et al. 1997). Another Kv β subunit, Kv β₄, increases the functional expression of Kv2.x channels (Fink et al. 1996). Although Kv β₄ is not expressed in heart (Fink et al. 1996), there certainly may be Kv β subunits in heart that interact selectively with Kv2.x subfamily members. Because heterologous coexpression studies suggest that the effects of the Kv β subunits are subfamily specific (Fink et al. 1996; Nakahira et al. 1996; Sewing et al. 1996), there are likely also to be β subunits that associate with the Kv4, as well as the erg1 and/or KCNQ, subfamilies of voltage-gated K⁺ channel α subunits, and further studies aimed at exploring this possibility directly are clearly warranted.

Using a yeast two-hybrid screen, another K⁺ channel accessory protein, KChAP, which is a 574 amino acid protein with no transmembrane domains and no homology to Kv α or β subunits or to minK, was identified (Wible et al. 1998). Coexpression of KChAP with Kv2.1 (or Kv2.2) markedly increases functional Kv2.x-induced current densities without measurably affecting the time- and/or the voltage-dependent properties of the currents (Wible et al. 1998). Yeast two-hybrid assays also revealed that KChAP interacts with the N-termini of Kv1.xα subunits, as well as with the C-termini of Kv β_1.x subunits. Coexpression in Xenopus oocytes of KChAP with Kv1.5, however, has no measurable effect on the properties of Kv1.5-induced currents (Wible et al. 1998). These results suggest either that KChAP and Kv1.5 do not associate in oocytes or, alternatively, that the interaction between KChAP and Kv1.5 does not influence functional (Kv1.5) K⁺ channel expression. Further experiments aimed at delineating the roles of β subunits and KChAP in controlling the properties and influencing the functional expression of cardiac K⁺ channels will clearly be of interest.

Molecular correlates of cardiac transient outward K⁺ currents

Heterologous expression of the voltage-gated K⁺ channel α subunits Kv1.4, Kv4.2 and Kv4.3 reveals rapidly activating, 4-AP-sensitive K⁺ currents (Fig. 2B) that qualitatively resemble cardiac transient outward K⁺ currents (Tseng-Crank et al. 1990; Blair et al. 1991; Roberds & Tamkun, 1991; Tamkun et al. 1991; Dixon et al. 1996). Initially Kv1.4, either as a homomultimer (Tseng-Crank et al. 1990; Tamkun et al. 1991), or as a heteromultimer with other Kv1.x subunits (Po et al. 1993), was considered the likely correlate of the fast transient outward K⁺ current I_to,f (Table 1) in cardiac cells. The finding that the Kv4.2 mRNA level varies through the thickness of the ventricular wall (in rat), however, led Dixon & McKinnon (1994) to postulate that Kv4.2, rather than Kv1.4, underlies rat ventricular I_to,f. Consistent with this hypothesis, Kv4.2 protein expression in rat heart is high, whereas Kv1.4 is barely detectable (Barry et al. 1995; Xu et al. 1996). In addition, the time- and voltage-dependent properties and the pharmacological sensitivities of the currents produced on heterologous expression of Kv4 α subunits more closely resemble I_to,f than do Kv1.4-induced currents (Yeola & Snyders, 1997; Franqueza et al. 1999; Peterson & Nerbonne, 1999; Wickenden et al. 1999c).

Considerable experimental evidence has now been provided documenting a role for Kv α subunits of the Kv4 subfamily in the generation of cardiac I_to,f. Fiset et al. (1997b), for example, reported that I_to,f is attenuated in rat ventricular myocytes exposed to antisense oligodeoxynucleotides (AsODNs) targeted against Kv4.2 or Kv4.3. In experiments on rat atrial myocytes, in contrast, an AsODN targeted against Kv4.2 attenuated I_to,f, whereas a Kv4.3 AsODN was without effect (Bou-Abboud & Nerbonne, 1999). These results suggest a molecular explanation for the differences in the kinetic properties of rat atrial and ventricular I_to,f (Apkon & Nerbonne, 1991; Boyle & Nerbonne, 1992), i.e. that both Kv4.2 and Kv4.3 contribute to rat ventricular I_to,f, whereas only Kv4.2 contributes to rat atrial I_to,f.

Rat ventricular I_to,f density is also reduced in cells exposed to adenoviral constructs encoding a truncated Kv4.2 subunit, Kv4.2ST (Johns et al. 1997). I_to,f is also attenuated in transgenic mice expressing this mutant (Wickenden et al. 1999b). In contrast, in ventricular myocytes isolated from transgenic mice expressing a pore mutant of Kv4.2 (Kv4.2W362F) that functions as a dominant negative, I_to,f is eliminated (Barry et al. 1998). I_to,f is also undetectable in atrial myocytes from Kv4.2W362F-expressing animals (Xu et al. 1999c). Given the similarities in the properties of I_to,f in different cell types and species (Table 1), it seems reasonable to speculate that Kv4 α subunits underlie I_to,f in all cardiac cells (Table 2). In dog and human, the candidate subunit is Kv4.3 (Dixon et al. 1996), and it has been demonstrated that I_to,f in human atrial myocytes is attenuated in cells exposed to a Kv4.3 AsODN (Wang et al. 1999). Interestingly, two splice variants of Kv4.3 have been identified in human (Kong et al. 1998) as well as in rat (Takimoto et al. 1997; Ohya et al. 1997) heart. Neither the relative expression levels of the two Kv4.3 proteins nor the contribution of these subunits to the generation of functional cardiac I_to,f channels, however, has been determined.

The properties of I_to,s in rabbit atrial and ventricular myocytes are distinct from those of I_to,f and, in fact, appear similar to those of heterologously expressed Kv1.4 (Tseng-Crank et al. 1990; Petersen & Nerbonne, 1999). Support for a role for Kv1.4 in the generation of I_to,s was provided recently with the demonstration that rabbit atrial I_to,s is attenuated following exposure to an AsODN targeted against Kv1.4 (Wang et al. 1999). In addition, in myocytes isolated from the left ventricular septum of mice with a targeted deletion in the Kv1.4 gene (London et al. 1998b), I_to,s is undetectable, revealing that Kv1.4 also underlies I_to,s in mouse ventricle (Guo et al. 1999). As noted above, the time- and voltage-dependent properties of the transient outward K⁺ currents in ferret left ventricular epicardial and endocardial myocytes are distinct, and there are regional differences in the expression of Kv1.4 and Kv4.2/Kv4.3 in ferret left ventricles (Brahmajothi et al. 1999). These observations suggest that Kv1.4 and Kv4.2/Kv4.3 underlie I_to,s and I_to,f in ferret left ventricular endocardial and epicardial myocytes, respectively (Brahmajothi et al. 1999). Regional differences in the transient outward K⁺ currents (I_to,f and I_to,s) in rat ventricles are also attributed to the differential expression of Kv4.2/Kv4.3 and Kv1.4 (Wickenden et al. 1999a).

Molecular correlates of cardiac delayed rectifiers

Considerable progress has also been made in identifying the molecular correlates of several components of I_K, including I_Kr, I_Ks, I_Kur and I_K,slow (Table 2), in cardiac cells. Expression of erg1, identified as the locus of LQT2 mutations, for example, reveals inwardly rectifying K⁺-selective currents similar to I_Kr, suggesting that erg1 underlies cardiac I_Kr (Sanguinetti et al. 1995; Trudeau et al. 1995). It has also been reported, however, that AsODNs targeted against minK attenuate I_Kr in AT-1 (an atrial tumour line) cells (Yang et al. 1995), and that heterologously expressed erg1 and minK coimmunoprecipitate (McDonald et al. 1997). Although these observations were interpreted as suggesting that functional I_Kr channels reflect the coassembly of erg1 and minK, more recent work has revealed that a minK homologue, MiRP (rather than minK) contributes to I_Kr (Abbott et al. 1999). Alternatively processed forms of erg1 with unique N- and C-termini have also been identified (Lees-Miller et al. 1997; London et al. 1997; Kuperschmidt et al. 1998) and suggested to be important in the generation of functional I_Kr channels. Western blot analysis with anti-erg1 antibodies, however, revealed that only full-length erg1 proteins are detected in rat, mouse and human heart, suggesting that alternatively spliced erg1 variants do not contribute to functional cardiac I_Kr channels (Pond et al. 2000).

Although heterologous expression of KvLQT1, the locus of mutations in LQT1, reveals rapidly activating K⁺ currents, coexpression of KvLQT1 with minK produces slowly activating K⁺ currents similar to cardiac I_Ks (Barhanin et al. 1996; Sanguinetti et al. 1996). The fact that mutations in the transmembrane domain of minK alter the properties of I_Ks channels further suggests that the transmembrane segment of minK contributes to the K⁺-selective pore of I_Ks channels (Goldstein & Miller, 1991; Takumi et al. 1988; Wang et al. 1996; Tai & Goldstein, 1998). Nevertheless, direct biochemical evidence for coassembly of KvLQT1 and minK in the mammalian heart has not been provided to date, and the stoichiometry of functional I_Ks channels is not known. In addition, the role of the N-terminal splice variant of KvLQT1, which functions as a dominant negative (Jiang et al. 1997), in the functional expression of I_Ks channels remains to be determined.

The properties of the currents produced on heterologous expression of Kv1.5 are similar to those of I_Kur in rat (Boyle & Nerbonne, 1992), human (Wang et al. 1993a,b) and canine (Yue et al. 1996a,b) atrial myocytes. In addition, Kv1.5 mRNA (Tamkun et al. 1991; Fedida et al. 1993) and protein (Barry et al. 1995; Van Wagoner et al. 1997) are abundant in rat and human heart. Direct evidence for a role for Kv1.5 in the generation of I_Kur in human atrial myocytes was provided with the demonstration that I_Kur in these cells is selectively attenuated following exposure to AsODNs targeted against Kv1.5, whereas AsODNs against Kv1.4 were without effect (Feng et al. 1997). Rat atrial I_Kur is also selectively reduced following treatment with Kv1.5 AsODNs (Bou-Abboud & Nerbonne, 1999). Interestingly, rat atrial I_K,slow (I_K,DTX) is selectively attenuated by AsODNs targeted against Kv1.2 (Bou-Abboud & Nerbonne, 1999). These results clearly demonstrate that rat atrial I_K,slow (I_K,DTX) is a unique molecular entity, distinct from both I_to,f and I_Kur, and that Kv1.2 and Kv1.5 do not coassemble to form heteromultimeric K⁺ channels (in rat atria) in vivo (Bou-Abboud & Nerbonne, 1999). The molecular correlate of I_Kur in canine atrial myocytes (Yue et al. 1996a,b), however, remains to be identified.

An in vivo dominant-negative strategy, involving cardiac-specific expression of a truncated Kv1.1 subunit (Kv1.1N206Tag), was exploited by London and colleagues (1998a) in experiments focused on determining whether Kv1.5 (or other Kv1 α subunits) contribute to the outward K⁺ currents in mouse ventricle. Electrophysiological recordings from ventricular myocytes isolated from Kv1.1N206Tag-expressing transgenics revealed that I_K,slow is selectively attenuated (London et al. 1998a). More recent studies, however, suggest there are two distinct components of I_K,slow in mouse ventricular (and atrial) myocytes and that both Kv1.5 and Kv2.1 contribute to these two conductance pathways (Xu et al. 1999a,b E. Bou-Abboud, H. Li & J. M. Nerbonne, unpublished observations). In addition, although the molecular correlate of mouse ventricular I_ss has not been determined, it has recently been suggested that Kv2.1 contributes to mouse atrial I_ss (E. Bou-Abboud, H. Li & J. M. Nerbonne, unpublished observations).

The results of the various studies described above have allowed the identification of the α subunits contributing to most of the voltage-gated K⁺ currents expressed in cardiac cells (Table 2). It seems reasonable to suggest that in vivo and in vitro experimental strategies similar to those used in these studies can be exploited to reveal the identities of the α subunits contributing to the other voltage-gated K⁺ channels in cardiac cells (Table 2), as well as to define the roles of silent Kv α subunits (e.g. Kv5.x to Kv9.x) and accessory (β, KChAP) subunits in the formation and functional expression of these channels.

Summary and future directions

Electrophysiological studies have clearly documented the presence of multiple types of voltage-gated K⁺ channels/currents in cardiac cells isolated from different species, as well as in cells isolated from different regions of the heart in the same species (Table 1). In most myocardial cells, several voltage-gated K⁺ channels are expressed, and each channel type contributes importantly to shaping the waveforms of action potentials, as well as influencing automaticity and refractoriness. Molecular cloning has revealed an even greater potential for generating voltage-gated K⁺ channel diversity than was expected based on the electrophysiology, with the identification of many K⁺ channel pore-forming (α) (Table 2) and accessory (minK, MiRP1, β and KChAP) subunits in the heart. A variety of experimental approaches have been exploited recently to probe the relationship(s) between these subunits and the functional voltage-gated K⁺ channels in myocardial cells. Important insights into these relationships have been provided through molecular genetics and the application of techniques that allow functional voltage-gated K⁺ channel expression to be manipulated in vitro and in vivo. The results of these efforts have led to the identification of the α subunits contributing to the formation of most of the voltage-gated K⁺ channels (I_to,f, I_to,s, I_Kr, I_Ks, I_Kur, I_K,slow) expressed in cardiac cells (Table 2). It seems reasonable to suggest that similar experimental approaches will reveal the identities of the Kv α subunits underlying the other voltage-gated K⁺ channels in cardiac cells (Table 1).

In contrast to the progress made in identifying the Kv α subunits that underlie the various voltage-gated cardiac K⁺ channels, much less is presently known about the role of accessory transmembrane (minK or MiRP1) or cytosolic (β and KChAP) subunit proteins in the formation or in the functioning of voltage-gated K⁺ channels in cardiac (or other) cells. Defining the roles of these subunits in the generation, regulation and modulation of voltage-gated cardiac K⁺ channels will be an important area for future research efforts. Similarly, very little information is available about the post-translational processing of K⁺ channel subunit proteins and/or about the role of these modifications in determining the properties of functional voltage-gated K⁺ channels in cardiac cells. There are a number of potential post-translational mechanisms, including phosphorylation, glycosylation, myristoylation, palmitoylation, etc., that could affect the properties and/or the functional expression levels of voltage-gated myocardial K⁺ channels (Nerbonne, 2000). Studies focused on examining post-translational processing of voltage-gated myocardial K⁺ channel α and β subunits, as well as those focused on determining the functional consequences of this processing and the underlying molecular mechanisms controlling post-translational processing, are likely to be major areas of future research interests and efforts.

It is now well-documented that there are marked changes in the functional expression of voltage-gated K⁺ channels and K⁺ channel-forming α subunits during normal (fetal and postnatal) cardiac development (Wetzel & Klitzner, 1996; Nerbonne, 1998), as well as in the damaged or diseased myocardium (Van Wagoner et al. 1997; Näbauer & Kääb, 1998; Carmeliet, 1999; Tomaselli & Marbán, 1999; Nerbonne, 2000). In addition, it is clear that electrical remodelling occurs in the damaged/diseased mammalian heart, presumably in direct response to changes in cardiac electrical activity or cardiac output, and that at least some of these changes can be attributed to alterations in the expression and/or the properties of voltage-gated K⁺ channels (Van Wagoner et al. 1997; Näbauer & Kääb, 1998; Tomaselli & Marbán, 1999; Nerbonne, 2000). There are numerous possible transcriptional (Levitan & Takimoto, 1998), translational and post-translational (Nerbonne, 2000) mechanisms that could be involved in regulating the expression and the properties of functional voltage-gated cardiac K⁺ channels. At present, however, very little is known about the molecular mechanisms mediating the changes in K⁺ channel expression/properties that are evident during normal development, as well as in conjunction with myocardial damage, disease and/or electrical remodelling. It seems certain that a major focus of future research will be on exploring these mechanisms in detail.