Voltage-gated K+ (KV) channels contain a voltage-sensing transmembrane segment that moves in response to changes in membrane potential. A small “gating” current arises from the translocation of positively charged residues within the channel's voltage sensor. These gating currents occur at the microsecond timescale (Gilly and Armstrong, 1980), indicating a highly ordered translocation (or reorganization) of the voltage sensors within the membrane. These initial events trigger additional, complex conformational changes that open, close, or inactivate the channel, thereby providing distinct kinetics to different types of KV channels (Benzanilla, 2008). Initial x-ray crystallography data of a mammalian KV chimeric channel provided the first glimpse of the protein organization of a KV channel when associated with phospholipid in a mixed phospholipid/detergent environment (Long et al., 2007). However, in vivo, KV channels are surrounded by a complex sea of lipids. How do these lipids interact with channels, where are critical sites of interaction, and how do they influence voltage sensing and gating? Answers to these important questions are beginning to be provided by x-ray crystal structures of pore regions, voltage sensors, and cytoplasmic domains of several K+ channels, in combination with biophysical, pharmacological and mutagenesis studies (Jiang et al., 2002; Kuo et al., 2003; Long et al., 2005; 2007; Logothetis et al., 2007; Lundbæk, 2008; Tucker and Baukrowitz, 2008).
A consensus is emerging that the anionic phospholipid, phosphatidylinositol-4,5-bisphosphate (PIP2) has physiologically important interactions with many KV channels and other transmembrane proteins, including ion transporters and ligand-gated K+ channels. In an explosion of interest for this idea, a rapid succession of studies identified a number of K+ channels that are positively regulated by PIP2, including all members of the Kir family of inward rectifiers (Sui et al., 1998; Liou et al., 1999; Zhang et al., 1999; Enkvetchakul et al., 2007; Logothetis et al., 2007), as well as voltage-gated KCNQ channels (Suh and Hille, 2002; Loussouarn et al., 2003; Zhang et al., 2003) and delayed rectifier K+ channels (Oliver et al., 2004). Among the reasons for the interest in PIP2 regulation of channels is that a number of Gq-coupled receptors that stimulate PIP2's breakdown also modulate K+ channel activity, most likely by decreasing the PIP2 available to associate with channels (Xie et al., 1999; Cho et al., 2001; Suh and Hille, 2002).
Hilgemann and Ball (1996) made the initial observation that PIP2 positively regulates the cardiac Na+/Ca2+ exchanger. Addition of ATP to the cytoplasmic side of inside-out (I/O) giant patches of cardiac myocyte membrane increased the Na+/Ca2+ exchange current; in patches pretreated with a bacterial PLC that selectively metabolizes phosphatidylinositol (PI), ATP no longer was able to enhance the current. ATP's stimulatory action recovered if PI was added to the bath before ATP; PI by itself had no effect, whereas PIP2 alone stimulated current in patches pretreated with PLC. These seminal findings suggested that phosphorylation of PI produced PIP2, which then directly interacted with the exchanger to stimulate its activity. To test their model, Hilgemann and Ball (1996) exposed ATP-stimulated patches to a PIP2-selective PLC, and found that current returned to pre-ATP levels, a result consistent with PIP2's actions underlying increased exchanger activity. Using similar approaches, they found an analogous stimulation of KATP channel activity with conditions that increased endogenous PIP2 and decreased activity under conditions that decreased PIP2 levels. Lastly they detected a possible increase in K+ inward rectifier (Kir) current. Soon after, Fan and Makielski (1997) reported for the first time the potentiation of Kir current by PIP2, a finding confirmed by Hilgemann's group (Huang et al., 1998). These early studies suggested that the requirement for PIP2 to elicit channel opening might be a broad principle for a number of K+ channels.
The same approach that implicated PIP2 as a regulator of Na+/Ca2+ exchangers and KATP channels led to negative outcomes in studies on the large conductance, Ca2+-activated K+ channel (also known as slo1, maxi-K+, or BK channel), in skeletal muscle and cardiac myocytes (Hilgemann et al., 2001; Hilgemann, D.W., personal communication). Throughout the body, maxi-K+ channels integrate intracellular Ca2+ signals with changes in membrane potential, often providing an important negative feedback mechanism that limits depolarization-driven Ca2+ influx. Previous findings demonstrated that metabolic products of PIP2 and downstream targets such as PLC, ryanodine receptors, kinases, and free fatty acids modulate BK channel activity (Dopico et al., 1994; Kirber et al., 1992; Salkoff et al., 2006). From beautifully executed experiments, Alex Dopico and his colleagues (see article by Vaithianathan et al., p. 13) now provide evidence that PIP2 can indeed interact with BK channels, adding another type of K+ channel to the ever-growing list of PIP2-sensitive channels. However, in order to observe enhanced activity, BK channels must coexpress with a specific accessory β subunit giving rise to tissue-specific PIP2 effects.
BK channels arise from a single mammalian gene (KCNMA1) with multiple splice sites providing for diverse expression patterns and differing biophysical properties (Salkoff et al., 2006). Not surprisingly, BK channels differ in topological organization from the conventional KV channels, which are comprised of six transmembrane segments (S1–S6) with both N and C termini located intracellularly. Distinct regions within S1–S4 (S3b and S4 in particular) appear responsible for voltage sensing while a P-loop spans segments S5 and S6 to contribute to the channel's pore (Long et al., 2005; 2007). Four of these subunits form a functional channel that has a central pore and four voltage sensors. BK channel structure deviates in several ways from this organization. First, the BK channel has an additional S0 segment preceding S1–S6 that results in the N terminus ending extracellularly. S0 may interface with voltage-sensing areas, introducing additional regulation of voltage sensor movements (Liu et al., 2008). Second, each S0 interacts with an accessory β subunit that contains two transmembrane segments. Encoded by distinct genes, expression of different β subunits (β1–β4) in different tissues adds complexity to BK gating, kinetics, and pharmacology. Third, as the name indicates, this Ca2+-sensitive maxi-K+ channel opens in response to rises in intracellular Ca2+ levels, membrane depolarization, or a combination of the two. BK's long C terminus tail contains a variety of functional domains involved in divalent sensing, such as the Ca2+ bowl, and two RCK (regulators of conductance of K+) domains, conferring Ca2+ sensitivity to the channel, yet the molecular underpinnings are not fully understood (Salkoff et al., 2006).
To test whether PIP2 may interact directly with BK channels, Dopico and colleagues performed several experiments in the presence of basal levels of Ca2+ (300 nM), where exogenous PIP2 robustly and reproducibly increased steady-state activity (NPo) of native BK channels in I/O patches from freshly isolated cerebral vascular myocytes, with no changes in the unitary current amplitude. In contrast, little change in NPo occurred following extracellular PIP2 application to cell-attached or outside-out patches, indicating that the putative site of interaction is most accessible from the inner leaflet, as would be expected because that is where endogenous PIP2 is found (Laux et al., 2000). The authors then used a dizzying array of pharmacological conditions to examine the effect of endogenous PIP2 on BK current in I/O patch and in perforated-patch recordings where the myocytes remained intact. Every experiment that favored increased endogenous PIP2 levels yielded increased channel activity. Dopico and colleagues then took the investigation a further step. They pressurized endothelium-free cerebral arteries that develop myogenic tone in order to determine whether changes in PIP2 levels could alter vessel diameter. Under conditions that minimized PIP2 metabolism, arterial diameter significantly increased, indicating that elevated levels of PIP2 increase BK channel activity. The resulting increase in BK-driven, outward current then hyperpolarizes myocyte membrane, promoting muscle relaxation and vessel dilation. They confirmed this hypothesis by showing that when BK activity was blocked with paxilline, arteries no longer relaxed under conditions that increased PIP2 levels. This in vivo demonstration is the first report documenting the consequences of a PIP2 interaction with an ion channel on organ function.
Having demonstrated physiological significance for PIP2's actions on BK channel activity, the authors examined the structural requirements for increased channel activity. They found that more water soluble forms of PIP2 (diC4 and diC8) also increased NPo but to a lesser degree than the longer-chain form, and reversed with washout more easily than PIP2. This result suggested that the longer fatty acid tails on the one hand facilitate increased partitioning of PIP2 into the lipid environment and hence access to channels and, on the other, slow the off rate from the bilayer to the aqueous phase. Additionally, the authors found that channel activation positively correlated with the number of anionic head group charges (−1 to −4). Moreover, when they included PIP2 antibodies or the anion scavenger poly-l-lysine along with PIP2, the normally robust increase in NPo was significantly reduced. Lastly, the authors tested whether 1,2-dipalmtoyl-sn-glycero-3-phospho-L-serine (PS) versus 1,2-dipalmtoyl-sn-glycero-3-phosphoinositol (PI) are equally effective in increasing BK activity. Both phospholipids were matched for charge (−1) and “flavor” of fatty acid tails. PI increased NPo 2.5 times greater than PS. This finding is especially important since it indicates that not only is charge critical, but so too is the structure of the moiety in the sn-3 position of the phospholipid. Taken together, the structural specifications for BK activation, which include the importance of long-chain fatty acid tails, the phospho-moiety in the sn-3 position, and the number of negative charges are consistent with the negatively charged phosphoinositol head group directly interacting with a specific binding site on the channel.
From these findings with native BK channels in cerebral vascular myocytes, Dopico and colleagues turned to the oocyte heterologous expression system to search for a potential PIP2 binding site in recombinant BK channels, with surprising results. Following expression of cvb1 (encoded by KCNMA1), a BK splice variant prevalent in cerebral artery myocytes (Liu, J., P. Liu, M. Asuncion-Chin, and A. Dopico. 2005. Soc. Neurosci. Abstr. Online. 960.913), PIP2 robustly increased the current recorded from I/O patches, establishing that cbv1, together with its immediate lipid environment, was sufficient for PIP2 to enhance current. They then identified a three–amino acid sequence (RKK) in the S6–S7 cytosolic linker of cbv1 that, when mutated to AAA, blunted increases in NPo following PIP2 application, indicating that this sequence may serve as a site of interaction with PIP2. In contrast, mutation of a positively charged amino acid (K239A), located in the S4–S5 cytosolic loop, had no effect on cvb1 sensitivity to PIP2. Taken together, these studies demonstrate that loss of PIP2's actions does not depend on all positive residues contained within cytoplasmic loops, but rather, on specific positive residues in the S6–S7 segment. Moreover, current enhancement by PIP3 occurred in wt and K239A mutant but not in the RKK to AAA substitution, indicating that PIP2 and PIP3 act at a common site of action.
Cvb1's RKK sequence is most likely a PIP2 interaction site because it shares certain characteristics with PIP2 binding sites in other K+ channels, which normally contain at least two (though often more than five) positively charged residues (arginine and lysine). At least one of the residues must be arginine. As with cvb1, critical sequences for PIP2 binding have been located to the proximal portion of the C-terminal tail of bacterial and mammalian Kir channels (Huang et al., 1998; Shyng and Nichols, 1998; Logothetis et al., 2007), KCNQ1, the subunit that confers PIP2 sensitivity to KCNQ1/KCNE1 channels (Loussouarn et al., 2003), and KCNQ2, the subunit that when coexpressed with KCNQ3 gives rise to M-current (Zhang et al., 2003). For certain channels, additional positively charged residues in the proximal N terminus also confer channel sensitivity to PIP2. Interestingly, the RKK mutation in cbv1 decreased but did not ablate BK channel sensitivity to PIP2, raising the possibility that additional nearby residues participate in PIP2 binding and/or additional sites of interaction mediate further enhancement; these sites await future discovery.
Analysis of macroscopic currents showed that PIP2 causes a parallel leftward shift in voltage sensitivity of activation without altering the effective valence of BK channels. Thus the negative charge of the PIP2 head group does not appear to alter critical aspects of voltage gating. Further analysis of unitary gating kinetics by the authors revealed increases in both open and closed dwell time distributions caused by PIP2 amplification of Ca2+i-driven gating. Indeed, when channels were opened with voltage in the absence of Ca2+, PIP2 had little effect on BK NPo, whereas in the presence of 300 nM or greater, PIP2 robustly enhanced NPo. Another observation highlights the importance of PIP2 control of Ca2+-driven gating. The authors found that increases in cbv1 channel NPo following PIP2 was markedly less than that observed in I/O patches from cerebral artery myocytes. One notable difference between the two sets of experiments was the absence of β1 subunits in the heterologous expression system, whereas cvb1 tightly associates with β1 subunits in cerebral vascular myocytes (Liu et al., 2005; Salkoff et al., 2006). Therefore, the authors tested whether β1 might potentiate PIP2's actions when coexpressed with cvb1. As suspected, PIP2's actions now recapitulated the remarkably large increase in NPo of vascular myocyte BK channels. In contrast, coexpression of β4 had no potentiating effect. Native BK channels in skeletal myocytes, cells known to barely express β1, responded to PIP2 similarly to homomeric slo1 consistent with β1, but not β4, facilitating PIP2's effects. These findings add a new wrinkle to PIP2 actions on channels where the accessory β1 subunit, known to robustly increase the apparent Ca2+ sensitivity of the BK channel complex, confers an additional amplification mechanism that controls the final effect of PIP2.
Though increasing evidence indicates that PIP2 regulates channel availability and Po of other K+ channels, how this occurs remains ill defined. Nonspecific deformation in lipid packing can alter channel gating (Lundbæk, 2008). However, the first x-ray crystal structures of the initial shell of lipid surrounding a chimeric voltage-sensing paddle yielded evidence that lipids may alter K+ channel conformational changes due to specific electrostatic interactions (Schmidt et al., 2006; Long et al., 2007). In particular, the finding of a phospholipid “wedged” between the voltage sensor and the S4–S5 linker, a domain of the channel that couples voltage sensor movement to S5–S6 channel gating (Long et al., 2007), suggests specific molecular interactions between individual amino acid residues and phospholipid molecules. The specific phospholipids interacting with the channels in mammalian cell membranes remain unidentified. However, molecular modeling of PIP2 with Kir 3.4 (Logothetis et al., 2007) or Kir 6.2 (Haider et al., 2007) find PIP2 localized to a location comparable to the phospholipid “wedge” between the S4–S5 linker in the crystal structures (Long et al., 2007). Similarly, PIP2's binding to the proximal region of the BK channel's cytoplasmic tail is consistent with an analogous orientation of the phosphoinositol head, embedded in BK's large tail region like an anchor; PIP2's long hydrophobic fatty acid tails serve as tethers by extending into transmembrane regions of the channel to influence gating. Whether additional molecular interactions of the maxi-K+ channel (perhaps via β1 subunit and Ca2+) reorient PIP2 within the bilayer and increase coupling between voltage sensors and gating regions within S5–S6 remains to be tested.
In summary, the results of Dopico and colleagues (Vaithianathan et al., 2008) provide an important addition to our understanding of the role of PIP2 in regulating transmembrane proteins. The similar changes occurring across multiple preparations make it unlikely that PIP2's interaction with the proteolipid environment around the BK channel complex indirectly mediates changes in channel gating. More likely the cbv1 + β1 complex is sufficient to support channel activation by PIP2. The fact that a specific β subunit appears to confer endogenous PIP2 sensitivity to a specific BK channel complex, which exhibits tissue-specific coexpression, strongly suggests that modulation by PIP2 is physiologically significant. In particular, the novel action of PIP2 in increasing channel sensitivity to Ca2+-dependent regulation of gating is reminiscent of PIP2's ability to increase sensitivity to G-protein subunits and Na+ binding observed with other K+ channels (Logothetis et al., 2007). The experimental findings in this report advance the importance of PIP2's interaction with the cytoplasmic region following S6 in a number of K+ channels.
Abbreviations used in this paper: I/O, inside-out; PIP2, phosphatidylinositol-4,5-bisphosphate.
© 2008 Rittenhouse This article is distributed under the terms of an Attribution–Noncommercial–Share Alike–No Mirror Sites license for the first six months after the publication date (see http://www.jem.org/misc/terms.shtml). After six months it is available under a Creative Commons License (Attribution–Noncommercial–Share Alike 3.0 Unported license, as described at http://creativecommons.org/licenses/by-nc-sa/3.0/).