A Marriage of Convenience: β-Subunits and Voltage-dependent K+ Channels*

The movement of ions across cell membranes is essential for a wide variety of fundamental physiological processes, including secretion, muscle contraction, and neuronal excitation. This movement is possible because of the presence in the cell membrane of a class of integral membrane proteins dubbed ion channels. Ion channels, thanks to the presence of aqueous pores in their structure, catalyze the passage of ions across the otherwise ion-impermeable lipid bilayer. Ion conduction across ion channels is highly regulated, and in the case of voltage-dependent K+ channels, the molecular foundations of the voltage-dependent conformational changes leading to the their open (conducting) configuration have provided most of the driving force for research in ion channel biophysics since the pioneering work of Hodgkin and Huxley (Hodgkin, A. L., and Huxley, A. F. (1952) J. Physiol. 117, 500–544). The voltage-dependent K+ channels are the prototypical voltage-gated channels and govern the resting membrane potential. They are responsible for returning the membrane potential to its resting state at the termination of each action potential in excitable membranes. The pore-forming subunits (α) of many voltage-dependent K+ channels and modulatory β-subunits exist in the membrane as one component of macromolecular complexes, able to integrate a myriad of cellular signals that regulate ion channel behavior. In this review, we have focused on the modulatory effects of β-subunits on the voltage-dependent K+ (Kv) channel and on the large conductance Ca2+- and voltage-dependent (BKCa) channel.

nature sequence" (3). The Kv channel ␣-subunit contains six transmembrane regions (TM 3 ; S1-S6), with both N and C termini on the intracellular side of the membrane (a tetrameric 6TM architecture). Although conserving the general structure of Kv channels (i.e. they have a voltage sensor (S1-S4) and pore modules (S5-P-S6)), BK Ca channels are an exception inside the S4 superfamily of ion channels. BK Ca channels contain seven transmembrane segments (S0 -S6) with the N terminus facing the extracellular side (reviewed in Ref. 4). The S4 segment of Kv and BK Ca channels contains positively charged amino acids (Arg or Lys) at every third position and is part of the voltage sensor responsible for voltage-dependent gating (1) (reviewed in Ref. 5).
Potassium channels may be considered the guardians of the cellular electrical homeostasis, and thus K ϩ channel diversity is of great importance in determining the variety of electrical responses of cells when subjected to stimuli. The possible mechanisms that originate the immense voltage-dependent K ϩ channel diversity are: (a) multiple genes, (b) alternative splicing, (c) formation of heteromultimeric channels, and (d) coexpression with accessory subunits.

␤-Subunits of Kv Channels
Kv channel properties can be modified by accessory proteins that regulate their channel gating and/or subcellular distribution (reviewed in Ref. 6). The ␤-subunits of Kv channels (Kv␤) are cytoplasmic proteins that have a mass of ϳ40 kDa. The proteins ␤1, ␤2, and ␤3 are coded by different genes, and additional variability is produced by alternative splicing on the N-terminal region (7,8). The Kv ␤-subunits form a tetrameric structure and are associated in 1:1 ratio with the ␣-subunit (9, 10) (Fig. 1A).

Kv ␤-Subunits Modify the Biophysical Properties of Kv Channels
Two main types of inactivating Kv channels have been described: 1) delayed rectifiers showing slow (second time scale) inactivation and 2) rapidly inactivating (A-type) Kv channels (reviewed in Ref. 7). The co-expression of some Kv ␤-subunits with Kv␣ changes the inactivation kinetics in slow inactivating channels, inducing a fast A-type inactivation (11). In addition, these subunits regulate the surface expression and voltage sensitivity of Kv1 channels (reviewed in Ref. 6; see below).
Kv␤1.1 binds to the N terminus of the Kv1 subfamily but not to Kv2, Kv3, or Kv4, indicating that Kv␤1 interaction with Kv␣ channels is restricted to Kv1 family members. The Kv ␤1.1subunit modifies the rate of inactivation in delayed rectifier channels like Kv1.4, but the voltage dependence of this process remains unchanged (8,11 nels (11). Both Kv ␤1.1and ⌲v␤1.2-subunits produce a leftward shift of the conductance-voltage curves of Kv1.5 channels and increase the rate of inactivation (13). In addition, Kv␤1.3 slows deactivation and modifies the Kv1.5 response to PKA activation. Kv ␤1.3-subunit contains consensus sites for phosphorylation by PKA that induces a response to kinase activation, slowing fast inactivation of the channel. Kv␤2 is unable to induce N-type inactivation by itself, but it increases the rate of inactivation. This subunit also induces an increase in the rate of activation of Kv1.4 channels without any appreciable change in the voltage dependence of activation gating. When co-expressed with Kv1.5, Kv␤2 accelerates inactivation and induces a shift in the activation threshold toward hyperpolarizing potentials (14).

Kv ␤-Subunit Pharmacology
The ␣-dendrotoxin (␣-DTX) block a class of fast inactivating aminopyridine-sensitive K ϩ channels. By sedimentation analysis of ␣-DTX acceptors isolated from bovine cortex, two species are identified: a large subunit (␣) and a "novel" subunit (␤) (9). The Kv ␤-subunits have been related with changes in blockade induced by anesthetics. Bupivacaine induces internal and external blockade in the Kv␣1.5 channels. The internal blockade induced by bupivacaine decreases (ϳ4-fold) when Kv␣1.5 is assembled with Kv␤1.3. Quinidine is also less potent (ϳ8-fold) in blocking channels formed by Kv␣1.5/Kv␤1.3 than channels consisting only of ␣-subunits (15). In dorsal root ganglion neurons, the Kv ␤1-subunit decreased the Kv1.1 sensitivity to the local anesthetic n-butylp-aminobenzoate used in treatment of chronic pain (16).

Structure and Redox Properties of Kv ␤-Subunit
The structure of the isolated Kv␤1 N terminus (amino acids 1-62) was solved using NMR spectroscopy (17). The N terminus of Kv␤1.1 does not exhibit a well defined, unique, threedimensional structure, indicating a fast conformational equilibrium between weakly structured substrates. The lack of a well defined structure can be an advantage in view of the long trajectory that is followed by the N terminus before reaching its blocking site (10,18). The crystallization of the Kv ␤2-subunit showed that it forms a 4-fold symmetric tetramer composed of four triose-phosphate isomerase barrels, each having eight parallel ␤-strands that form a central core and intervening ␣-helices encircling the perimeter of the barrel (19). At the front face of each Kv ␤2-subunit, there is a tightly bound NADP ϩ molecule. The crystal structure of the Kv1.2 channel in complex with the Kv ␤2-subunit shows that the N terminus of the Kv1.2 ␣-subunit forming the T1 domain is like a docking platform for the Kv ␤2-subunit (10). As observed in the isolated structure of the Kv ␤2-subunit (19), the active site contains an NADP ϩ cofactor and catalytic residues for the hydride transfer. Therefore, Kv ␤-subunits can be important for catalytic function behaving as a redox sensor and allowing direct coupling of membrane electrical activity to the redox state of the cell (20 -22). Some of the main properties of Kv ␤-subunits are summarized in Table 1.

␤-Subunits of BK Ca Channels
Regulatory ␤-subunits of BK Ca channels (BK Ca ␤) contain 191-235 amino acids sharing a predicted membrane topology, with the N and C termini oriented toward the cytoplasm (Fig.  1B). They have two putative TM segments connected by a 112-123-residue extracellular "loop" that contains three or four putative glycosylation sites and disulfide linkages arising from conserved cysteine residues (23). At present, four BK Ca ␤-subunits have been cloned in mammals (␤1-␤4, coded by genes KCNMB1-4) (Reviewed in Ref. 23). ⌻he BK Ca ␤3 family comprises four distinct subunits (␤3ad) that arise as a consequence of alternative splicing of a single gene.

Changes in Biophysical Properties of BK Ca Channels Induced by BK Ca ␤-Subunits
The BK Ca ␤1-subunits induce an increase of the apparent Ca 2ϩ sensitivity, a decrease of the voltage dependence of the channel, and slowing of the macroscopic kinetics (reviewed in Refs. 4 and 23). The BK Ca ␤1 effects seem to result from a stabilization of voltage sensor activation both when the channel is closed and when open. Ca 2ϩ sensitivity in these channels is increased because, at all voltages, less Ca 2ϩ -binding energy is necessary to open the channel (24).
BK Ca ␤2 increases the Ca 2ϩ and voltage sensitivity of BK Ca channels and slows the kinetics of the channel (25,26). Moreover, this subunit induces fast and complete inactivation (27). The N terminus of the BK Ca ␤2-subunit (residues 1-45, BK Ca ␤2N) blocks the BK Ca channel via interaction with a receptor site in the ␣-subunit, which becomes accessible once the channel is in the open state. BK Ca ␤2N structure was studied by NMR spectroscopy and consists of two domains connected by a flexible linker (Glu 17 -Arg 19 ) (28) (Fig. 1B). Orio et al. (29) suggested that N-and C-terminal domains from BK Ca ␤1 and BK Ca ␤2-subunits modulate the steady-state and kinetic parameters of BK Ca channels.
BK Ca ␤3ac induce channel inactivation to BK Ca currents and also produce an outward rectification of the open channel currents. The inactivation process is faster than BK Ca ␤2-induced inactivation albeit incomplete (25). BK Ca ␤3b-subunit induces a small and consistent decrease in activation time constants at all Ca 2ϩ concentrations, and it does not affect channel deactivation (25). ␤3b-Subunit confers a non-linearity on instantaneous current-voltage properties that is regulated by extracellular segment of this ␤-subunit (30).
The human BK Ca ␤4-subunit has a complex Ca 2ϩ concentration-dependent effect on BK Ca channel current. This subunit decreases apparent Ca 2ϩ sensitivity at low Ca 2ϩ concentrations but induces an increase in the apparent sensitivity at high Ca 2ϩ concentrations (25,31,32). Human BK Ca ␤4 also slows down activation and deactivation kinetics (25,31).

BK Ca ␤-Subunit Pharmacology
Charybdotoxin (ChTX), a toxin isolated from the scorpion Leiurus quinquestratus (33), made possible the isolation and purification of the first BK Ca ␤-subunit reported (reviewed in Ref. 34). By the inhibition of 125 I-ChTX binding to BK Ca channels, a natural product identified as dehydrosoyasaponin was discovered. Dehydrosoyasaponin is a triterpene glycoside that increases the mean open time of BK Ca channels but only when it is added into the intracellular face of the channel and when the ␤-subunit is present (35). Iberiotoxin (IbTx), a scorpion toxin isolated from the scorpion Buthus tamulus, is another potent BK Ca channel blocker with the advantage of being highly selective for BK Ca (34). BK Ca ␤1-, ␤2-, and ␤4-subunits altered ChTx and IbTx binding in electrophysiological and biochemical studies (36 -38). BK Ca channels are modulated by external binding of 17 ␤-estradiol. The presence of 17 ␤-estradiol elicits an increase in the currents recorded in patches expressing ␣and ␤-subunits but not in those expressing only the ␣-subunit (39). The chemotherapeutic xenoestrogen tamoxifen also increased the BK Ca probability of opening only in the BK Ca ␤1-subunit presence (40). Cells expressing BK Ca ␣␤4 channels confer particular sensitivity to the adrenal glucocorticoids cortisol and corticosterone and are potentiated to a lesser degree by other sex and stress steroids (41). Fatty acids such as arachidonic acid (AA) alter BK Ca ␤-subunit modulation of BK Ca channel inactivation. Currents induced by channels formed by ␣ϩ␤2 and ␣ϩ␤3 were affected by AA (42). BK Ca channel inactivation may be a specific mechanism by which AA and other unsaturated fatty acids influence neuronal death/survival in neuropathological conditions (42). Recently, the fluorescent dye voltage-sensitive DiBAC 4 (3) was reported as a BK Ca channel selective activator only when the regulatory rBK Ca ␤1 or rBK Ca ␤4, but not rBK Ca ␤2, were co-expressed with rBK Ca ␣ in HEK293 cells (43). Some of the main properties of BK Ca ␤-subunits are summarized in Table 1.

␤-Subunit Kv and BK Ca Channel Trafficking
␤-Subunits of Kv channels, in addition to modulating the channel activity at the cell surface, control the surface expression of the ␣-subunit (44). The interaction of Kv1 ␣-subunit and Kv ␤-subunit polypeptides is an early event in Kv1 biosynthesis, occurring in the endoplasmic reticulum (ER) (44, 45) ( Fig. 2A). Despite dramatic differences in their effects on channel gating, each of the Kv ␤-subunits displays robust trafficking effects. Kv ␤1.1-, ␤1.2-, ␤2-, and ␤3-subunits increase the membrane expression and the mature form of Kv1.2 when they are co-expressed (44, 46 -48). The interaction with Kv ␤2-subunits results in increased stability of Kv1.2 ␣-subunits. There is a dramatic difference in the degradation rates of the free Kv1.2 pool (non-bonded to Kv ␤2-subunit, t1 ⁄ 2 ϳ3 h) and the Kv1.2 associated with Kv␤2 (t1 ⁄ 2 ϳ15 h) (44). Therefore, although some cytoplasmic Kv1 channel ␤-subunits affect the inactivation kinetics of ␣-subunits, a more general and perhaps more  (46). Regarding BK Ca ␤-subunit trafficking, two reports have appeared indicating that BK Ca ␤1 and BK Ca ␤2 are able to reach the plasma membrane when they are expressed alone in HEK293 cells (50,51). Co-expressing BK Ca ␤1 with BK Ca ␣-subunit reduces steady-state BK Ca channels surface expression levels by means of an endocytic mechanism. This result shows that BK Ca ␤1 can also regulate BK Ca surface expression levels (50) (Fig. 2B). In addition, the co-expression of BK Ca ␤1-subunit with BK Ca ␣-subunit splice variant SV1 (that is retained in ER) exhibits dominant-negative properties on BK Ca ␤1 surface expression. This study provides important insights into BK Ca subunit assembly and suggests the early assembly of BK Ca and ␤1-subunits in the ER (52).

Knock-out Models
Mouse genetic models have played an important role in the elucidation of molecular pathways underlying human disease; gene deletions have also underscored the physiological relevance of Kv and BK Ca channel ␤-subunits. This approach has been used to determine the effects of Kv␤1 (53) and Kv␤2 (54) removal on Kv1 family Kv currents. Kv␤1.1-deficient mice show normal synaptic plasticity, but they show impaired learning, indicating that the Kv ␤1.1-subunit contributes to certain types of learning and memory (53). In aged mice, the deletion of the auxiliary potassium channel subunit Kv␤1.1 resulted in increased neuronal excitability, synaptic plasticity, and learning (55). The phenotype of Kv␤2-null mice includes reduced life spans, occasional seizures, and cold swim-induced tremors similar to that observed in Kv1.1-null mice (54).
Regarding BK Ca channels, deletion of the smooth muscle BK Ca ␤1-subunit causes slight hypertension and increased contractile response to vasoactive agonists (56,57). BK Ca ␤4 knock-out mice, on the other hand, display abnormal neuronal firing properties and temporal lobe seizures, indicating that the gating properties conferred by the ␤4-subunits are essential to normal neuronal function (58).

Coda
The properties of native Kv and BK Ca channels are profoundly influenced by associated ␤-subunits controlling their subcellular distribution and channel gating. Here we have reviewed the most important examples of Kv ␤and BK Ca ␤-subunit modulation of channel gating, pharmacological properties, and channel trafficking. ␤-Subunits are expressed in many tissues, and in some cases, they are tissue-specific, allowing the involvement of K ϩ channels in a variety of different physiological processes. Three different genes that code for Kv␤ (KCNAB1-3) and four genes that code for BK Ca ␤ (KCNMB1-4) have been reported. Additional diversity of ␤-subunits is produced by alternative splicing. ␤-Subunits are, however, only one piece of a protein network that is associated with ion channels. The pore-forming subunits of Kv and BK Ca channels are components of large protein complexes in the plasma membrane. It is very important to know the changes in Kv and BK Ca channel function induced by partners sharing the same protein complex. Identification of these partners and determination of their influence in channel properties will not only provide us with new insights about channel function but can also lead us to unravel new disguises of these molecular machines in cell physiology and pathophysiology.