How Does Voltage Open an Ion Channel?

ANRV288-CB22-02 ARI 30 August 2006 17:1

Francesco Tombola,1,4 Medha M. Pathak,2 and Ehud Y. Isacoff1,2,3,#

1Department of Molecular and Cell Biology, University of California, Berkeley, California 94720; email: [email protected] 2Biophysics Graduate Group, University of California, Berkeley, California 94720; email: [email protected] 3Physical Bioscience Division, Lawrence Berkeley National Laboratory, Berkeley, California 94720; email: [email protected] 4Department of Biomedical Sciences, University of Padova, Padova, Italy 35121

Annu. Rev. Cell Dev. Biol. 2006. 22:23–52 Key Words First published online as a Review in potassium channels, voltage gating, coupling, cooperativity, lipid Advance on May 5, 2006

by CAPES on 08/30/07. For personal use only. The Annual Review of Abstract Cell and Developmental Biology is online at http://cellbio.annualreviews.org Neurons transmit information through electrical signals generated by voltage-gated ion channels. These channels consist of a large This article’s doi: superfamily of proteins that form channels selective for potassium, 10.1146/annurev.cellbio.21.020404.145837 sodium, or calcium ions. In this review we focus on the molecular Copyright c 2006 by Annual Reviews. mechanisms by which these channels convert changes in membrane All rights reserved Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org voltage into the opening and closing of “gates” that turn ion con- 1081-0706/06/1110-0023$20.00 ductance on and off. An explosion of new studies in the last year, #To whom correspondence should be including the ﬁrst X-ray crystal structure of a mammalian voltage- addressed. gated potassium channel, has led to radically different interpretations of the structure and molecular motion of the voltage sensor. The interpretations are as distinct as the techniques employed for the studies: crystallography, ﬂuorescence, accessibility analysis, and electrophysiology. We discuss the likely causes of the discrepant results in an attempt to identify the missing information that will help resolve the controversy and reveal the mechanism by which a voltage sensor controls the channel’s gates.

23 ANRV288-CB22-02 ARI 30 August 2006 17:1

They “gate” rapidly, within one or a few mil- Contents liseconds, and when open they selectively conduct speciﬁc ions. The coordinate function of INTRODUCTION...... 24 these channels produces signals (∼one-tenth THE ACTIVATION GATE...... 26 of a volt) in remarkably brief spurts (as short as Where Is the Gate Located? ...... 26 one millisecond) that can repeat at very high Which Parts of the Protein Form rates (up to 1000 per sec) and travel rapidly the Gate? ...... 26 (∼100 meters per sec), even in extremely thin VSD ORGANIZATION AROUND (∼1-micron-diameter) processes, for long dis- THE PORE DOMAIN ...... 28 tances (meters) without decrement. Remark- GATING CHARGE able. How do they do it? We review here the MOVEMENT ...... 29 molecular properties of these channels, focus- VOLTAGE-SENSING ARGININES 30 ing mainly on voltage-gated potassium (Kv) MODELS OF VOLTAGE channels, which have been the objects of the SENSING ...... 31 most extensive investigation in this area and Transporter Model ...... 33 have had a veritable explosion of progress in Helical Screw Model ...... 33 the past year. Paddle Model ...... 33 Kv channels are made of four subunits, Are We Ready for a Uniﬁed each containing six transmembrane segments Model? ...... 34 that are named S1 through S6 (Figure 1a,b). VOLTAGE-GATED CHANNELS Helices S5 and S6 of the four subunits, as HAVE FOUR ADDITIONAL well as the P-helix and the loop that connects PORES ...... 37 the helices, assemble together to form a cen- CREVICES WITHIN THE VSD tral pore domain, which contains the channel’s AND FOCUSING OF THE K+-selective pathway and gates. Four voltage- MEMBRANE ELECTRIC sensing domains (VSDs), each made of he- FIELD ...... 39 lices S1–S4, surround the pore domain and COUPLING OF VOLTAGE control its gates. Subtype-selective assembly SENSING TO GATING ...... 40 of the channel-forming subunits is controlled WORKING TOGETHER TO by an intracellular N-terminal tetrameriza- OPEN THE GATE ...... 42 by CAPES on 08/30/07. For personal use only. tion domain (T1) (Figure 1a,b, green), which LIPID: THE MISSING PLAYER? . . 43 also serves as a scaffold to bind accessory β subunits (Kvβs). Instead of four separate subunits, voltage-gated sodium (Nav) and calcium (Cav) channels have four covalently con- INTRODUCTION nected domains; each domain has a secondary

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org The action potential, the secretion of hor- structure similar to that of a single subunit of mones and neurotransmitters, the heart beat, a potassium channel. Nav and Cav channels the reaction of an egg that prevents fertil- are evolutionarily related to each other and ization by multiple sperm, the contraction of to Kv channels (Yu & Catterall 2004). Bacte- skeletal muscle, and the control of transpira- rial voltage-gated sodium channels are closely tion from the leaves of a plant are diverse bio- related to both eukaryotic Cav and Nav chan- logical phenomena that have at least one thing nels, but they are made of four independent in common: They are all mediated by voltage- subunits, like potassium channels (Koishi et al. gated ion channels. These channels respond 2004, Ren et al. 2001). to changes in the gradient of voltage across The ion-conducting pathway of voltage- the membrane by opening and closing an ion gated channels allows permeation at a high conductance pathway across the membrane. rate, on the order of 106–108 ions per second,

24 Tombola · Pathak · Isacoff ANRV288-CB22-02 ARI 30 August 2006 17:1

Figure 1 (a) The architecture of a Kv channel subunit. Cylinders are helical segments. The pore domain is shown in blue, the voltage-sensing domain (VSD) in red, the S4-S5 linker in purple, and the tetramerization domain in green. (b) A single Kv1.2 subunit color coded as in a. Potassium ions are colored yellow. The Kv1.2 tetramer (c) top view (extracellular side) and (d ) side view. Each subunit is shown in a different color. Potassium ions are colored purple. Coordinates from Long et al. (2005a), PDB ID 2A79. All the molecular drawings have been created using

by CAPES on 08/30/07. For personal use only. Swiss-Pdb viewer (http://www.expasy. org/spdbv/).

and can discriminate between ions with re- cium and sodium channels, and Gouaux & markable efﬁciency. Potassium channels, for MacKinnon (2005) reviewed general princi- + Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org example, have a permeability ratio for K ples of ion transport by channels and pumps. over Na+ of >100:1, and calcium channels The process by which the ion-conducting select for Ca2+ over Na+ with a ratio of pathway of voltage-gated channels opens in >1000:1. Much progress has been recently response to changes in membrane poten- made in understanding the mechanism un- tial is called activation; it is the subject derlying ion permeation in potassium chan- of the present review. The activation gate, i.e., nels. MacKinnon (2003), Armstrong (2003), the element that physically opens and closes and Roux (2005) provide recent reviews on the transmembrane ion conduction pathway, permeation and selectivity of potassium chan- is discussed ﬁrst. Then, the mechanism by nels. French & Zamponi (2005), Sather & which voltage-gated channels detect changes McCleskey (2003), and Yu & Catterall (2003) in membrane potential and control their acti- recently reviewed ion conduction in cal- vation gate is discussed in the context of the

www.annualreviews.org • Voltage-Induced Channel Activation 25 ANRV288-CB22-02 ARI 30 August 2006 17:1

recent crystal structure of the Kv1.2 chan- ies on N-type inactivation in A-type (fast in- nel. The last section of this review deals with activating) potassium channels. Fast inactiva- the interaction between Kv channels and the tion involves the binding of the N-terminal Shaker: fast-inactivating membrane lipid. domain, also referred to as the ball, to its voltage-gated receptor on the intracellular side of the chan- potassium channel nel (Hoshi et al. 1990, Iverson & Rudy 1990, from Drosophila THE ACTIVATION GATE Zagotta et al. 1990). As with internal QA ions, melanogaster that is The pore domain of a voltage-gated ion chan- the N-terminal ball binds to the pore only the best known member of the nel contains the permeation pathway, which when the activation gate is open and acts via Shaker/Kv1 family is opened and closed by two distinct molec- a “foot-in-the-door” mechanism, making it Quaternary ular gates: the activation and slow inactiva- harder, or impossible, for the gate to close ammonium (QA): tion gates. We focus here on the activation while it is bound (Demo & Yellen1991). Thus, class of potassium gate. In most voltage-gated channels at rest- the activation gate appears to be located on channel blockers ing potential (∼−70 mV in neurons), the ac- the intracellular mouth of the ion-conducting NGK2: tivation gate is closed, and membrane depo- pore. voltage-gated larization causes a conformational change in potassium channel the VSDs that is transmitted to the pore do- from a main and results in opening of the gate. It has Which Parts of the Protein Form the neuroblastoma- Gate? glioma hybrid cell been estimated that the conductance of a sin- line. It belongs to the gle Shaker channel drops at least 105 times, Choi et al. (1993) reported that mutations in Shaw/Kv3 subfamily going from the open to the closed state (Soler- the S6 helix alter the internal binding of QA Tetraethyl Llavina et al. 2003) and that the probabil- ions to the pore domain. Lopez et al. (1994) ammonium (TEA): ity for a Shaker channel to be in the open produced a chimeric channel by transplanting a potassium channel state in resting conditions is lower than 10−9 the S6 helix of NGK2 into Shaker and found blocker (Islas & Sigworth 1999). This indicates that that the chimera had single-channel conduc- Methanethio- the block of ion ﬂow by the gate can be ex- tance and sensitivity to internal tetraethyl sulfonate (MTS): a thiol-modifying tremely effective. ammonium (TEA), similar to NGK2. These reagent ﬁndings indicated that the S6 region lines the ion-conducting pathway of voltage-gated Where Is the Gate Located? channels, making it a good candidate to be

by CAPES on 08/30/07. For personal use only. The ﬁrst indication of the existence of a part of the activation gate. gate on the intracellular side of the mem- Liu et al. (1997) found a series of po- brane came from the seminal experiments sitions in the Shaker S6 region that, when of Armstrong (1966, 1969, 1971), in which mutated to cysteine, reacted rapidly with pos- the voltage-dependent potassium current of itively charged methanethiosulfonate (MTS) the giant squid axon was blocked by qua- reagents in the open state, but not in the

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org ternary ammonium (QA) ions applied to the closed state. Some of these positions could axoplasm. Armstrong found that the block- be protected from chemical modiﬁcation by ing ions reached their binding site within the the presence of an intracellular QA blocker. membrane only after the gate was opened by These researchers also found a cysteine mu- depolarization. The gate was shown to open tation in S6 (V476C) that prevented the chan- less readily when the QA ions were bound. nel from closing at negative voltages when Armstrong also found that, after binding, the Cd2+ was applied intracellularly. The remark- gate could be closed by strong hyperpolariza- able effect of Cd2+ on Shaker V476C was later tion fast enough to trap the blocking ions in shown to be due to the formation of bridges the conductive pathway inside the membrane. between the introduced cysteine in one sub- Further evidence for the intracellular lo- unit and the native histidine H486 in another cation of the activation gate came from stud- subunit; these bridges trap the gate in the

26 Tombola · Pathak · Isacoff ANRV288-CB22-02 ARI 30 August 2006 17:1

Figure 2 Comparison of pore domains from three different potassium channels in the open conformation. Only pore-forming helices from two subunits are shown. S5 (TM1) and the P-helix are shown in gray; the GYG motif in the selectivity ﬁlter, in green; and the S6 helix and part of the pore loop, in blue. In the bacterial channels MthK and KvAP, the S6 (TM2) bundle opens at a glycine hinge (red ). In the eukaryotic Kv1.2, the hinge in S6 corresponds to the PVP motif (red ). The gray background represents the membrane. Coordinates from Jiang et al. (2002, 2003a) and Long et al. (2005a).

open conformation (del Camino et al. 2000, in the extracellular half of the membrane. Holmgren et al. 1998). These ﬁndings QA ions can be trapped in the cavity (Zhou strongly suggested that the S6 region forms et al. 2001). The TM2s occlude the conduc- KcsA: the activation gate. tion pathway of KirBac1.1, shutting the ac- proton-activated The crystal structure of the bacterial potas- cess pathway between the cytoplasm and the potassium channel sium channel KcsA (Doyle et al. 1998) pro- cavity even tighter than in KcsA (Kuo et al. from Streptomyces vided detailed information about the position 2003). In MthK and KvAP the bundle of he- lividans and nature of the activation gate, comple- lices appears to have swung open at a glycine MthK: menting what was known from the functional in TM2, which is proposed to serve as a gat- calcium-gated studies mentioned above for the identiﬁca- ing hinge (Figure 2; Jiang et al. 2002, 2003a). potassium channel from the tion of the internal receptor for QA ions The glycine is conserved in bacterial potas- thermophilic and for the N-terminal inactivation domain sium channels as well as in eukaryotic Kvs archaeon

by CAPES on 08/30/07. For personal use only. (Zhou et al. 2001). The comparison of struc- (G83 in MthK, G99 in KcsA, G220 in KvAP, Methanobacterium tural data from bacterial potassium channels G466 in Shaker). The opening of KcsA at the thermoautotrophicum trapped in the open conformation, MthK and glycine hinge is supported by the accessibility KvAP: KvAP ( Jiang et al. 2002, 2003a; Zhou et al. study of Kelly & Gross (2003). However, the voltage-gated 2001), and channels trapped in the closed con- spectroscopic study from Liu et al. (2001) sup- potassium channel from the formation, KcsA and KirBac1.1 (Doyle et al. ports a position for the hinge approximately hyperthermophilic Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org 1998, Kuo et al. 2003), revealed a possible mo- two helical turns downstream of Gly99. The archaeon Aeropyrum tion that the S6 helices may undergo to open reason for this difference is not clear, but it pernix

the gate. may be an indication that in KcsA the bend- KirBac1.1: In KcsA, the second transmembrane he- ing of the TM2s is more gentle than in MthK inward-rectiﬁer lices (TM2s), corresponding to S6 helices in and instead resembles the bending found in potassium channel Kv channels, form a bundle that occludes the open KvAP (Figure 2). from Burkholderia pseudomallei the ion conduction pathway on the intra- Bending of the TM2s and S6 helix at a cellular side of the membrane. The TM2 glycine hinge may well explain gating in bac- helices have an inverted teepee-like arrange- terial potassium channels, but it does not ex- ment, and a large aqueous cavity resides be- plain gating of eukaryotic Kv channels like tween the intracellular bundle crossing of the Shaker. The S6 helix of many eukaryotic Kv TM2s and the selectivity ﬁlter that is located channels has a PXP motif (with X = Vin

www.annualreviews.org • Voltage-Induced Channel Activation 27 ANRV288-CB22-02 ARI 30 August 2006 17:1

Shaker and Kv1.2). This motif is not present earlier functional analysis, enables us to get at in bacterial channels. In a study on a Shaker this question. mutant in which the valine of the PVP motif was substituted with cysteine (V474C) and another cysteine was introduced at position VSD ORGANIZATION AROUND 476, Webster et al. (2004) elegantly showed THE PORE DOMAIN that when the activation gate is in the open The crystal structure of Kv1.2 shows an unex- conformation, internal Cd2+ ions can be con- pected position of the four VSDs around the currently coordinated between V474Cs of dif- pore. The VSDs are located at the corners of ferent subunits and between V476C in one the square-shaped pore domain (Figure 1c), subunit and H486 in a neighboring subunit. and their interaction surface with the pore This finding is incompatible with opening at is rather small (Figure 1b,c). As a result, a the glycine hinge, which would bring the cys- large portion of their perimeter is expected teines in position 474 of the four subunits too to face lipid. Based on these findings, Long far away from each other to be able to coor- et al. (2005a) suggested that the VSDs keep dinate cadmium. Opening at a glycine hinge their position at the periphery of the chan- would also result in a distance between H486 nel floating in the membrane and that they and C476C on adjacent subunits incompat- only weakly interact with the pore domain. ible with Cd2+ cross-bridging. The finding This property of VSDs may enable them to of Webster et al. (2004) combined with ear- function as independent transmembrane do- lier evidence from other accessibility stud- mains, which may be transplanted onto other ies (del Camino et al. 2000, del Camino & proteins to confer voltage sensitivity (Long Yellen 2001, Holmgren et al. 1998, Liu et al. et al. 2005b, Lu et al. 2002). Indeed, a re- 1997) call for opening of the S6 bundle at the cently discovered phosphatase, which has an PVP motif of Shaker and similar eukaryotic intracellular enzymatic domain connected to Kv channels. a VSD homolog, has a gating current and The crystal structure of the rat Kv1.2 is voltage dependent (Murata et al. 2005). (Long et al. 2005a,b) shows that the S6 ac- The existence of voltage-gated proton chan- tivation gate is open because of a bend at the nels containing a VSD without a pore domain PVP motif (between positions 473 and 475 in has been also reported (Ramsey et al. 2006,

by CAPES on 08/30/07. For personal use only. Shaker). This wonderful agreement between Sasaki et al. 2006). However, that the inter- the probing analysis on functional channels actions between VSD and pore domain are and the crystal structure ﬁnally makes it pos- not spread over a large surface does not rule sible to say that, after almost 40 years since out functional importance. Li-Smerin et al. Armstrong’s ﬁrst study of potassium current (2000), using perturbation-scanning mutage- block by internal QA ions in the squid giant nesis (systematically substituted tryptophan)

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org axon, the nature of the activation gate of Kv in Shaker, found that some positions have a channels is no longer a mystery. The general low impact on function (i.e., they likely face solution to the problem of opening the gate is water or lipid) and that others have a high im- to bend the inner TM2/S6 helix, and this can pact (i.e., they likely lie at a protein-protein in- occur at either a glycine or a proline motif. terface). In the Kv1.2 crystal structure, part of But how is the activation gate controlled by the high-impact positions in S5 and the pore the four VSDs surrounding the pore domain? helix turn out to face the VSD, whereas high- The recent crystal structure of the rat Kv1.2 impact positions in S4 (Ledwell & Aldrich (Long et al. 2005a) provides new information 1999, Pathak et al. 2005) turn out to face S5. about the organization of the VSD around the In addition, Lai et al. (2005) recently obtained pore and about the connection between S4 evidence, from mutant suppression analysis, and the pore helices that, in combination with for a speciﬁc interaction between S4 and S5 in

28 Tombola · Pathak · Isacoff ANRV288-CB22-02 ARI 30 August 2006 17:1

the hyperpolarization-activated KAT1 potas- the VSDs are wrapped around the pore do- sium channel, whose VSD is homologous to main. The crystal structure of Kv1.2 shows that of depolarization-activated Kv channels. that the S4-S5 linker of one subunit runs un- Thus, a VSD can operate as an independent derneath the neighboring subunit to connect entity floating in the membrane and, in Kv S1-S4 of the VSD to S5-S6. This arrange- channels, has a small interaction surface with ment, which is in agreement with an ear- the pore domain, but the interaction is impor- lier tryptophan-scanning mutagenesis study tant and specific, and its strength remains to on the Shaker pore domain (Li-Smerin et al. be determined. 2000), allows the entire VSD of one subunit In the original crystal structure of KvAP to stand next to the pore helices of the neigh- ( Jiang et al. 2003a), the VSD in the full-length boring subunit (Figure 1c,d ) and provides channel was almost entirely out of contact a substrate for cooperative interactions be- with the pore domain and was in a confor- tween subunits, as we discuss below. Among mation very different from that which was the VSD helices of Kv1.2, S4 is the closest deduced from probing analysis on functional to the pore domain, as previous cross-linking channels and from the structure of Kv1.2. studies on functioning Shaker channels have KvAP was crystallized in complex with an an- found (Broomand et al. 2003, Gandhi et al. tibody Fab fragment. It was speculated that 2003, Laine et al. 2003, Neale et al. 2003). the crystal packing forces of the Fab fragment Exposure to lipid of S2 (see Figure 8, be- bound to the VSD may have been responsi- low) is in agreement with the study by Monks ble for trapping the channel in this unusual et al. (1999), which examined the perturba- conformation. However, a new structure of tions caused by tryptophan-scanning muta- KvAP in absence of antibody showed a simi- genesis of the Shaker S2. Similar mutagen- lar conformation, suggesting that this confor- esis studies were also performed on S1 and S3 mation is actually favored in the absence of a (Hong & Miller 2000). In the Kv1.2 structure membrane (Lee et al. 2005). The loss of na- the side chains of these two VSD segments tive structure in KvAP lacking the surround- are not resolved, making the register of the ing membrane was interpreted by MacKinnon helices uncertain. However, the results from and colleagues to support the notion that the Hong and Miller may well be used to define VSD is loosely attached to the pore domain the right register of S1 and S3.

by CAPES on 08/30/07. For personal use only. (Lee et al. 2005). This conclusion rests on the assumption that the stabilizing effect of the membrane is weak (if it were strong, strong GATING CHARGE MOVEMENT interactions between VSD and pore domain To make the opening of the activation gate also could be seriously perturbed by the re- voltage dependent, the gate must be con- moval of the membrane). At present it is not trolled by a molecular sensor that detects the

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org clear if this assumption has general applicabil- transmembrane potential in real time. This ity. The painfully low rate of success of struc- molecular sensor must contain charges, lo- tural efforts in membrane proteins may sug- cated in the membrane electric ﬁeld, that gest that in general the membrane provides a change their position when the ﬁeld changes, strong stabilization of native structure. as originally proposed by Hodgkin & Huxley In the Kv1.2 structure, the VSD of one (1952). It is now well appreciated that these subunit contacts the pore-forming helices of charges, called gating charges, reside in the the neighboring subunit. This does not come VSDs of voltage-gated channels. The S4 he- as a surprise, as Laine et al. (2003) and Neale lix in each VSD contains four to eight posi- et al. (2003) have shown that the extracellu- tively charged residues, mostly arginines, lo- lar ends of S4 and S5 of adjacent subunits are cated at every third position. The S4 charges close to each other. The real surprise is how are responsible for most of the gating charge

www.annualreviews.org • Voltage-Induced Channel Activation 29 ANRV288-CB22-02 ARI 30 August 2006 17:1

movement during activation. Several authors charge have discrete stopping places along the have reviewed gating charge movement in in-out pathway (Baker et al. 1998, Bezanilla voltage-gated ion channels and the contribu- et al. 1994, Schoppa & Sigworth 1998c, tion of the S4 positively charged residues to Stefani et al. 1994), and moreover, it was the total gating charge (e.g., Bezanilla 2000, debated whether charged arginines could be 2005; Gandhi & Isacoff 2002; Sigworth 1994; located in the hydrophobic core of the bi- Swartz 2004; Yellen 1998). layer (Freites et al. 2005, Grabe et al. 2004, When the gating charges move across the Hessa et al. 2005, Monticelli et al. 2004, membrane electric field, as a result of a change Parsegian 1969). Functional studies on Shaker in the applied membrane voltage, they gen- in cell membranes (discussed in Ahern & erate an electric current called the gating Horn 2004b, Bezanilla 2005, Durell et al. current. The gating current is roughly two 2004) and spectroscopic studies on KvAP in orders of magnitude smaller than the ionic liposomes (Cuello et al. 2004) argued that S4 current flowing through the open channel, arginines face a polar protein crevice. and it is transient. It can be measured only As discussed in more detail below, replace- when the number of channels in the mem- ment of certain S4 arginines with smaller brane is high and when both the ionic current amino acids was found to make the VSDs through the pore and the capacitive current permeable to protons (Starace & Bezanilla required to charge and discharge the mem- 2001, 2004) or metal ions (Sokolov et al. 2005, brane are reduced or eliminated (Bezanilla & Tombolaet al. 2005b) in a manner incompat- Stefani 1998). ible with a location in lipid but compatible Gating currents are a very useful readout with a protein polar crevice. The exact posi- of voltage-sensor movement in the membrane tion of the positively charged residues in the electric field and have played a key role in the S4 helix was found to be essential for volt- development of kinetic models of activation age sensing (Ahern & Horn 2004a), suggest- (Schoppa & Sigworth 1998c, Zagotta et al. ing that the S4 arginines are specifically ori- 1994a). The total amount of charges moved ented within the protein. In addition, some of by the voltage sensors of several voltage-gated the S4 arginines were shown to interact with channels has been determined from measures negatively charged residues in other VSD he- of the gating currents. For example, a charge lices (Papazian et al. 1995, Tiwari-Woodruff

by CAPES on 08/30/07. For personal use only. of ∼13 e0 per channel moved during acti- et al. 2000). vation of the Shaker channel (Aggarwal & In the Kv1.2 structure, some of the side MacKinnon 1996, Schoppa et al. 1992, Seoh chains in the VSD, including the ﬁrst four et al. 1996); each of the four VSDs contributed voltage-sensing arginines (R1–R4), presented

∼3.2 e0. weak electron density, but their approximate positions were deduced (Figure 3). For R1

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org what is thought to be the guanidinium group VOLTAGE-SENSING ARGININES is visible. It would be located at the lipid- Structural studies on KvAP suggested that aqueous interface if the channel were sur- the ﬁrst four arginines of S4, the most im- rounded by membrane. R2 is in an interme- portant ones for voltage sensing (Aggarwal & diate position, between where lipid and the MacKinnon 1996, Seoh et al. 1996), face the pore-forming S5 helix would be. R3 is the hydrophobic core of the lipid bilayer ( Jiang best-resolved S4 arginine; faces S1, S2, and et al. 2003b). This idea was appealing be- S3; and electrostatically interacts with neg- cause it could provide for unhindered mo- atively charged residues, including the con- tion through a greased surrounding to medi- served glutamate 226 (position 283 in Shaker) ate fast voltage sensing. The puzzle was that in S2. R4 is not visible, but its position on earlier evidence had shown that S4 and gating the backbone was deduced from the location

30 Tombola · Pathak · Isacoff ANRV288-CB22-02 ARI 30 August 2006 17:1

of its neighbors. Like R3, R4 seems to face other VSD helices. Thus, three out of four arginines are likely to be in a hydrophilic environment, and R2 is at the lipid-protein interface, with its charged guanidinium situated in an environment that will strongly depend on the exact orientation of the side chain. In the Shaker channel in the activated conformation (i.e., in a conformation likely similar to that of Kv1.2 in the crystal), cysteines replacing the first two S4 arginines can readily react with charged MTS reagents added to the extracellular side of the membrane, indicating that R1 and R2 are located in a hydrophilic environment in contact with the extracellular solution (Larsson et al. 1996). Recent molecular dynamics simulations (Freites et al. 2005) on a helix similar to the Figure 3 N-terminal part of the S4 segment immersed The structural organization of the Kv1.2 VSD, shown with a top view in a lipid membrane suggest that the four (extracellular perspective). Helices from two adjacent subunits are shown in arginines facing lipid may alter the bilayer’slo- two different colors: Pore-forming helices of one subunit are blue, and the cal structure to ensure that each guanidinium VSD of the other subunit is red. The side chains of the first four S4 group is salt bridged by a negatively charged arginines, R1–R4, and the S2 glutamate in position 226 (283 in Shaker) are displayed in the space-filling CPK scheme. The side chains of R4 and E226 phosphate group from a lipid molecule. This, were created by Swiss-Pdb viewer on the basis of partial crystallographic combined with intrusion of water in the prox- coordinates. Complete coordinates are not available for these two residues imity of the salt-bridged arginines, may min- (see discussion regarding R4 in text). imize the contact of the charged stripe of arginines with the membrane’s hydrophobic mine their movement in three-dimensional core. Although it is interesting to see that the space. The only electrically measurable com- bare S4 segment can stay in a stable posi- ponent of voltage-sensor motion within the

by CAPES on 08/30/07. For personal use only. tion within the membrane for the simulated membrane is that associated with charge 8 ns, it is not clear what would happen over a movement along the electric ﬁeld. Although time comparable with the timescale of gating electric potential drops pretty much lin- charge movement, which is in the millisecond early across a lipid bilayer, inhomogene- range. In the crystal structure of Kv1.2, only a ity in dielectric in the protein and at the portion of S4 is exposed to lipid, and the helix lipid-protein interface can result in distor-

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org is significantly far from the position perpen- tions near the voltage sensor, where the field dicular to the membrane plane considered in may be locally focused and its direction may the simulations; nevertheless, the findings of not be perpendicular to the membrane. To Freites and colleagues support the notion that characterize the motion of the voltage sen- R2 may be stabilized at the protein-lipid inter- sor associated with gating charge displace- face by salt bridging a lipid phosphate group. ment, a host of different techniques have been used. These techniques include elec- trophysiological recordings, scanning mu- MODELS OF VOLTAGE SENSING tagenesis, site-specific accessibility studies, The arginines that constitute the gating cross-linking, fluorometric techniques em- FRET: Forster¨ charge have been known since the mid- to ploying environment-sensitive dyes, and dis- resonance energy transfer late 1990s, yet it has been difficult to deter- tance measurements using FRET. Figure 4

www.annualreviews.org • Voltage-Induced Channel Activation 31 ANRV288-CB22-02 ARI 30 August 2006 17:1 by CAPES on 08/30/07. For personal use only. Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org Figure 4 Models of voltage sensing. Red cylinders represent S4 unless otherwise indicated. Protein surrounding S4 is colored gray. Lipid is colored green. The ﬁrst four S4 arginines are represented as blue spheres when they are in the foreground; when they are behind the cylinder, they are shown as empty circles. To keep drawings simple, in some of the models the arginines are arranged on the same face of the S4 helix. In each model, the S4 resting position is shown on the left, and the activated position on the right. The extent of S4 transmembrane motion involved is reported on the side of each model. In the helical screw and moving oriﬁce models, the thickness of the septum separating extracellular and intracellular crevices/vestibules is the one reported in Baker et al. (1998). An alternative thickness of the septum (Yang et al. 1996) is shown as a dotted line. The extent of S4 transmembrane movement indicated for the helical screw model ranges from ∼5 Ato˚ ∼13 A˚ because, depending on the tilt of the helix, a 13 A˚ movement of S4 along the axis of the helix can have a smaller transmembrane component (see Figure 6e).

32 Tombola · Pathak · Isacoff ANRV288-CB22-02 ARI 30 August 2006 17:1

shows the basic conceptual models that have residues in other VSD helices. The model has been proposed to describe the movement of changed over time to account for evidence ob- the voltage sensor during activation. tained by a large number of studies (recently discussed in Ahern & Horn 2004b, Durell et al. 2004, Gandhi & Isacoff 2002, Keynes Transporter Model & Elinder 1999, and Lecar et al. 2003). In the According to the transporter model (Bezanilla version of the helical screw model represented 2002, 2005; Chanda et al. 2005; Starace & in Figure 4, the S4 charges move ∼13 A˚ along Bezanilla 2004), in the resting state the S4 the axis of the helix (to account for the ob- charges are located in a crevice in contact served solution exposure change) and rotate with the intracellular solution. During acti- 180◦ (deduced both from the FRET analysis vation the charges move into another crevice, and from the pitch of the helix’s “threads,” this time in contact with the extracellular so- which include the S4 arginines at every third lution. Tilt change and rotation of S4 as well position that slowly wind like the stripes of a as crevice reshaping allow the charges to move barbershop poll along the helix). S4 was origi- from one crevice to the other (Figure 4). nally proposed to move through a gating pore In the resting state the membrane potential [also referred to as S4 channel, gating canal, or drops to the largest extent across the sep- canaliculus (Goldstein 1996, Yang et al. 1996, tum around R1, separating the intracellular 1997)] made of both VSD helices and pore crevice from the extracellular solution. In the helices and completely surrounding S4 to inactivated state the membrane potential drops sulate it from the lipid. In the version of the across the septum around R4. The transporter model shown in Figure 4, the gating pore has model involves a very small transmembrane been substituted by an omega pore, as dis- movement of S4 (2–3 A),˚ and thus it is an cussed in detail below, and S4 is tilted to as- evolved version of two previous models of sume a position across the membrane more charge translocation, “rotation in place” (Cha similar to the one the helix has in the Kv1.2 et al. 1999) and “moving oriﬁce” (Yang et al. crystal structure. 1996), in which S4 does not move across the membrane (Figure 4). Paddle Model

by CAPES on 08/30/07. For personal use only. According to the paddle model, inspired by Helical Screw Model the crystal structure of the bacterial chan- In the helical screw model, the S4 helix ro- nel KvAP ( Jiang et al. 2003a,b), S4 and the tates and, at the same time, translates along most extracellular part of S3 (S3b) form a he- its axis upon depolarization to move the gat- lical hairpin, or paddle, that moves across the ing charges across the transmembrane electric membrane as a unit. In the resting state, S4 is

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org ﬁeld. As a result, the S4 charges change their located deep in the membrane, not far from exposure from the intracellular to the extra- the interface with the intracellular solution, cellular solution. The extents of rotation and and S3 is positioned on top of S4. Upon depo- translation both vary in different versions of larization, S4 and S3 translocate to the outer the helical screw model. In the version that membrane surface. The net transmembrane Guy & Seetharamulu (1986) and Catterall movement of the paddle voltage sensor is 15– (1986) originally forwarded, S4 was proposed 20 A˚ (Ruta et al. 2005). The fundamental char- to rotate 60◦ and move 4.5–5 A˚ toward the acteristic that differentiates the paddle model extracellular side of the membrane during ac- from both the transporter and helical screw tivation. If rotation accompanies translation, models is that the S4 arginines are directly a changing subset of the arginines can always exposed to lipid in both resting and activated form salt bridges with ﬁxed negatively charged states, whereas in the other models they move

www.annualreviews.org • Voltage-Induced Channel Activation 33 ANRV288-CB22-02 ARI 30 August 2006 17:1

enclosed in a protein environment. Although structure of the VSD of KvAP docked in dif- an earlier model of full translocation shown ferent ways onto the pore domain (Broomand in Figure 4 (Aggarwal & MacKinnon 1996) et al. 2003, Chanda et al. 2005, Elliott et al. cannot be considered a precursor of the paddle 2004, Gandhi et al. 2003, Jiang et al. 2003b, model because it involves S4 partial uncoiling, Laine et al. 2004, Shrivastava et al. 2004). But in this early model, as in the paddle model, the can a single model account for all of the ex- S4 arginines are directly exposed to lipid. In perimental evidence? Several recent reviews the eukaryotic Kv1.2, the entire S3 helix has have re-examined the new evidence and ear- been proposed to be part of the paddle (Long lier findings (Ahern & Horn 2004b, Bezanilla et al. 2005b). 2005, Cohen et al. 2003, Horn 2005, Swartz In the transporter, helical screw, and pad- 2004) without arriving at a unified model. dle models, none of the first four charged Since the appearance of these reviews, there residues in S4 fully translocates across the en- have emerged significant new data, includ- tire membrane hydrophobic core, which is ing a crystal structure of the full-length Kv1.2 thought to be ∼30–32 A˚ thick. Because two channel that is largely compatible with func- of the residues (R2 and R3) must move across tional probing analyses (Long et al. 2005a), a the entire electric field to account for two of series of functional studies that significantly the ∼3.2 gating charges per subunit, all three expanded earlier measurements of S4 trans- models require some focusing of the electric membrane motion in eukaryotic Kv chan- field acting on S4. What varies between the nels (Chanda et al. 2005, Phillips et al. 2005, models is the extent of focusing. Focusing Posson et al. 2005) and in the bacterial KvAP is very strong in the transporter model, in- channel (Ruta et al. 2005), and a novel mea- termediate in the helical screw model, and sure of the distance traveled by the Shaker mild in the paddle model. Evidence that S4 first S4 arginine across the membrane electric is, at least partially, in contact with lipid has field (Ahern & Horn 2005). The new data still been accumulating over time (Cuello et al. challenge efforts at reconciliation, and the au- 2004, Elinder et al. 2001, Jiang et al. 2003b, thors of the studies favor radically divergent Lee et al. 2005, Lee & MacKinnon 2004, views of S4 motion, ranging from ∼2 Atoas˚ Li-Smerin et al. 2000, Long et al. 2005a, much as 20 A.˚ However, the actual measure- Schonherr et al. 2002) and has led this fea- ments may not be as incompatible as they first

by CAPES on 08/30/07. For personal use only. ture to be incorporated into all three models appear. (lipid is shown in green in Figure 4). There is Selvin and colleagues (Posson et al. 2005) a difference, though, in the extent of lipid ex- and Bezanilla and colleagues (Chanda et al. posure in the three models. The transporter 2005) used FRET to assay the transmembrane and helical screw models accept exposure to distance traveled by the Shaker S4 during acti- lipid of roughly one-third of the S4 circum- vation. Selvin and colleagues used FRET be-

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org ference, pointing the buried arginines into tween a donor ﬂuorophore attached to one of a polar protein environment, whereas in the a number of single cysteines in S4 and an ac- paddle model most of the S3/S4 hairpin is in ceptor on the “back” of a pore-blocking pep- contact with lipid, including the arginines. tide toxin. They estimated a small change in distance between S4 and the toxin and deduced from this that if S4 moved in a pure Are We Ready for a Uniﬁed Model? perpendicular motion, it could move by no Since the crystal structure of KvAP was solved more than ∼2 A.˚ Bezanilla and colleagues also ( Jiang et al. 2003a), several different models placed a donor on S4 and used as an acceptor of voltage-sensor movement, compatible with dipicrylamine (DPA), a lipid-soluble organic either the transporter, helical screw or paddle anion that localizes in the lipid and shuttles model, have been proposed on the basis of the between the polar headgroups of the inner

34 Tombola · Pathak · Isacoff ANRV288-CB22-02 ARI 30 August 2006 17:1

and outer leaflets. Although their study was voltage-sensor paddle and the membrane sur- qualitative, their measurements indicated that face cannot be larger than 8 A.˚ a pure perpendicular motion would have to be Ahern & Horn (2005) used an original ap- small. proach to determine the distance traveled by MacKinnon and colleagues (Ruta et al. S4 across the membrane electric field during 2005) deduced the extent of KvAP paddle activation. They reduced the gating charge (S3b/S4) movement across the membrane by moved by the Shaker voltage sensor by sub- assaying the accessibility to avidin of biotin stitution of the first S4 arginine with cysteine attached to different positions in the paddle. (R1C). They then measured the amount of When attachment was via a linker length of gating charge restored by modification of the 17 A,˚ biotin anywhere on S4 could be grabbed R1C channels with cysteine-reactive probes from either side of the membrane by avidin. bearing a positively charged group at the end The voltage dependence of the accessibility of hydrocarbon linkers of varying lengths. was not systematically examined, so the ex- The shorter the linker, the deeper the charge posure could not be directly associated with moved with S4 into the electric field. These state-dependent conformation. But in transit- investigators found that an increase in linker ing from resting to activated conformations, length of 4 A˚ eliminated the contribution to S4 exposed the biotin, at some point, to avidin gating charge. They thus suggested that this in the solution on both sides of the membrane. is enough to stick the charge permanently out Unlike what was seen with S4, sites in the pore of the electric field and argued that the depth domain could only be reached by avidin from of the electric field between the outside and one side of the membrane. The motion of S3b the location of R1 at rest is ∼4 A.˚ was intermediate, between that of S4 and the The ∼8 A˚ maximal extent of voltage- pore domain. This suggests that S4 shuttles sensor movement across the membrane de- through the span of the membrane by 15– termined by Swartz and colleagues (Phillips 20 A,˚ changing tilt angle in such a way that S3b et al. 2005) and Ahern & Horn’s (2005) moves less, whereas the pore-domain protein ∼4 A˚ estimate for the motion of R1 across stands still. the electric field can be compatible with both Swartz and colleagues (Phillips et al. 2005) the small transmembrane movement mea- used the peptide hanatoxin to track the move- sured by Selvin, Bezanilla, and colleagues and

by CAPES on 08/30/07. For personal use only. ment of the Kv2.1 voltage sensor during the larger movement measured by MacKin- activation. This tarantula toxin inhibits Kv non and colleagues, as discussed elsewhere channels by binding to their voltage sen- (Tombola et al 2005a). But how can the sors, mainly to S3, and stabilizing the rest- ∼2 A˚ transmembrane motion obtained by en- ing conformation (Lee et al. 2003). Swartz ergy transfer measurements in Shaker be rec- and colleagues showed that after partitioning onciled with the 15–20 A˚ motion determined

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org into the membrane, hanatoxin binds tightly by biotin accessibility in KvAP? In one pos- to the voltage-sensor paddle and follows it sibility, the bacterial channel may undergo a in its transmembrane movement. They added movement more extreme than does its eukary- the toxin to membranes containing bromi- otic cousin, as suggested by a recent model- nated lipids with bromine attached to differ- ing study based on the structures of Kv1.2 ent positions of the lipid hydrocarbon tails. and KvAP (Yarov-Yarovoy et al. 2006). KvAP By quenching the ﬂuorescence from the toxin does differ in several respects from the eukary- tryptophan with bromine situated at different otic channels ( Jiang et al. 2003a; Long et al. depths in the lipid, they determined that the 2005a): (a) The KvAP S3 is broken into two toxin penetrates no deeper than ∼8 A˚ from the helices, whereas in Kv1.2 it is one continuous outer surface of the membrane. This means helix; (b) the secondary structure of S4 differs that the distance between the top of the Kv2.1 between the channels; (c) the regular interval

www.annualreviews.org • Voltage-Induced Channel Activation 35 ANRV288-CB22-02 ARI 30 August 2006 17:1

a 0 0 Z FRET 10 10 ZZ 20 20 -150 mV +50 mV

30 30

b 0 0

10 10 BA Z Z 20 20 +A i +Ao 30 30

Figure 5 FRET and biotin accessibility look at the transmembrane motion of S4 through opposite ends of the telescope (see text for details). (a) Shift in the average z-position of S4 during activation as estimated by FRET ( zFRET). (b) Capture of tethered biotin by extracellular (Ao) or intracellular (Ai) avidin (gray Pac-Man). The transmembrane motion inferred by the biotin accessibility assay (zBA) is related to the width of the z-position distribution of S4 conformations during gating. The individual distributions for

by CAPES on 08/30/07. For personal use only. resting and activated state are colored in two shades of gray. The reference point for the z-position distribution is indicated as a purple dot on the S4 helix. The ﬁrst four S4 arginines are shown as blue spheres. Biotin is represented as a gray sphere attached through a linker to a location on S4 coinciding with the reference point for the distribution. The shortest linkers compatible with avidin capture of the tethered biotin (i.e., in its maximally z-axis-extended position) on each side of the membrane are shown.

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org of S4 arginines found in the eukaryotic chan- different z-axis positions in the membrane. It nels is disrupted after R4 in KvAP; and (d ) the could spend most of its time in a preferred activation gate opens at the glycine hinge in (most stable) conformation but from time to KvAP, whereas it opens at the PXP motif in time visit one of a broad set of less stable posi- Shaker-like eukaryotic channels. tions. Figure 5a shows a simple view of such Alternatively, if the voltage sensor were a a broad distribution in the z-position at neg- flexible structural element, as its strong mo- ative voltage (bell-shaped surface). When the bility (Ruta et al. 2005, Lee et al. 2005) may membrane is depolarized to open the chan- suggest, the conflicting results on the extent nel (Figure 5a, right panel), the average posi- of its transmembrane motion could be action of the voltage sensor and the peak of the counted for in another way (Figure 5). A flex- distribution shift toward the extracellular side ible voltage sensor might fluctuate between of the membrane (z = 0). How would such

36 Tombola · Pathak · Isacoff ANRV288-CB22-02 ARI 30 August 2006 17:1

distributions in S4 position look to the FRET gauge how exposed the biotin is at each site. and tethered biotin accessibility studies? Instead, it is possible to say only whether bi- In the FRET experiments, all the confor- otin is ever exposed during the course of the mations of the voltage sensor within the dis- experiment. This implies that not only sites tribution of positions contribute to the energy close to the membrane edge can have bind- transfer between donor and acceptor, and the ing; so can deep sites from where the biotin FRET reﬂects a weighted average of all of only occasionally pokes out of the membrane the distances: (a) The conformational states but which, once bound, is not released by the at the peak of the distribution contribute to tightly binding avidin during the experiment. the average more than do the conformations If one gauges position in the z-axis, based on at the edges, which are visited less often, and the deepest sites from where the biotin can (b) the states with z close to zero contribute reach the avidin, then one is biased toward more energy transfer, and thus weigh more the far edge of the distribution (red and blue heavily in the average, than do states with circles in Figure 5b) and thus infers an over- a large z (green gradients on gray distribu- estimated transmembrane motion. tions in Figure 5a). This uneven weighing Thus, if S4 really does ﬂuctuate in a broad causes some underestimation of the distance. distribution of z-positions, then its voltage- In Figure 5a the distribution resulting from sensing motion between its most stable resting overweighting conformations with small z is and activated positions may be larger than the represented as a dotted gray line. The ex- ∼2 A˚ calculated by FRET and shorter than tent of underestimation depends on the ex- the ∼20 A˚ deduced from the binary tethered act shape of the distribution, on the donor- biotin accessibility assay.

acceptor distance, and on the R0 (Forster¨ radius) for the donor-acceptor pair. Posson et al. (2005) and Tombolaet al. (2005a) discuss VOLTAGE-GATED CHANNELS other potential sources of underestimation. HAVE FOUR ADDITIONAL On the other hand, the tethered biotin PORES accessibility measurements have the opposite A major challenge in the design of a voltage tendency, i.e., that of overestimating the dis- sensor is how to allow S4 arginines to traverse tance traveled by the voltage sensor. The ex- the membrane rapidly but to do so in a se-

by CAPES on 08/30/07. For personal use only. pectations are that in the activated state, bi- lective manner that prevents leakage through otin attached to S4 is more accessible to avidin the arginine permeation pathway of solution on the outside of the membrane and that the ions. Bezanilla and colleagues (Starace et al. cysteine attachment positions closer to the N- 1997) found that the VSDs of Shaker chan- terminal end of S4 spend a larger fraction of nels in which the second or third S4 argi- the time with the biotin exposed to the avidin nine is substituted with histidine behave as

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org and thus have a faster binding reaction. Thus, proton transporters, shuttling protons across to gauge the degree of voltage-dependent bi- the membrane every time the VSDs transi- otin exposure, it is necessary to compare the tion between the resting and activated con- binding rates of the tethered biotin in the ac- formations. They also found that changing tivated and resting states. This is also how R4 to histidine creates a proton pore in the one measures voltage-dependent cysteine ac- VSD in the activated state, whereas changing cessibility to water-soluble thiol reagents. Al- R1 to histidine creates a proton pore in the ternatively, the binding assay can be used to resting state (Starace & Bezanilla 2001, 2004). provide a binary outcome of accessibility or These ﬁndings suggest that the S4 arginines inaccessibility of the tethered biotin to avidin move through a hydrophilic pathway when (Ruta et al. 2005). In this case, binding rates the VSD changes conformation with volt- are not measured, and it is not possible to age. That arginines translocate across a

www.annualreviews.org • Voltage-Induced Channel Activation 37 ANRV288-CB22-02 ARI 30 August 2006 17:1

Figure 6 (a) Omega- and proton-conducting Shaker channels. (b) Omega-conducting Nav1.2a. Pore region shown in blue. VSDs permeable to solution cations colored red. Nonpermeable VSDs in the sodium channel colored green. (c) S4 helix surrounded by the gating pore. The ﬁrst four S4 arginines are represented as blue spheres. (d ) S4 arginines occlude the omega pore. Substitution of the ﬁrst arginine with smaller uncharged residues in the Shaker S4 allows solution cations ( purple spheres) to permeate the omega pore in the resting state. (e) Tilt of the S4 helix reduces the transmembrane component ( green arrows)ofthe helical screw movement that brings R4 into the place of R1 during activation.

hydrophilic pore ﬁts well with the concept et al. 2005b). If R1 pointed into the lipid, the of a gating pore, a proteinaceous pathway in hydrophobic substitution would be expected the span of the membrane through which S4 to disrupt a water-boundary layer. If, on the

by CAPES on 08/30/07. For personal use only. had been proposed to move during activation other hand, R1 pointed into a protein path- and deactivation to account for its voltage- way, then reducing side-chain bulk would be dependent changes in internal and external expected to increase conductance and perhaps exposure (Goldstein 1996, Larsson et al. 1996, let larger ions through. Indeed, substitution of Yanget al. 1996, 1997). The advent of the pad- R1 with Ala, Ser, Cys, or Val led to a cation- dle model ( Jiang et al. 2003b), which predicts selective current, referred to as omega cur-

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org that S4 is not enclosed in protein but instead rent to distinguish it from the potassium al- faces lipid, however, posed a problem for the pha current ﬂowing through the pore domain interpretation of the proton pores, leading to (Figure 6a, Tombola et al. 2005b). The cur- two possible explanations: Either (a) the pro- rent was larger for the polar (Ser) and smaller tons moved through a small boundary layer of side chains (Ala), but substitution with His did water (a water wire) lining the arginine face not permit metal ions to pass. Omega current of S4 at the protein-lipid interface or (b) VSD ﬂowed only in the resting state (Figure 6d ), architecture is different, and the S4 arginines when the R1 position is deep in the span of actually face a polar protein pathway. the membrane (Baker et al. 1998, Larsson An ensuing study in Shaker tested these et al. 1996, Wang et al. 1999). The omega two possibilities by the substitution of R1 pore was able to conduct ions up to the size with smaller, uncharged residues (Tombola of guanidinium, the positively charged group

38 Tombola · Pathak · Isacoff ANRV288-CB22-02 ARI 30 August 2006 17:1

of the arginine side chain. The omega pore protein in the heart of the VSD (Figure 3) was therefore proposed to serve normally (Long et al. 2005a), consistent with what is as the arginine-conducting pathway. As the expected for the activated state of Shaker- guanidinium group of the arginine side chains like channels. The lipid exposure on one face permeates this pathway, one of them always of S4 and short length of interaction at the occupies its narrowest part. This stabilizes S4 buried arginines may be essential for rapid conformation (and is compatible with a ratch- (minimally hindered) transmembrane motion eting motion) and prevents flux by solution of the voltage sensor. Study of the relationship ions. between (a) the omega pathway of Shaker and Catterall and colleagues found an omega Nav1.2a and (b) the proton pathway in the re- current also in Nav1.2a with arginine muta- cently discovered voltage-gated proton chan- tions in the S4 of the second domain (Sokolov nels containing a VSD but lacking the pore et al. 2005). These researchers saw omega cur- domain (Ramsey et al. 2006, Sasaki et al. 2006) rent in the resting conformation when both is likely to provide important clues about the the first and second S4 arginines were substi- conformational changes occurring in the VSD tuted with glutamine (Figure 6b). Mutating during voltage sensing. only R1 was not sufficient. The requirement for the paired substitutions suggests that, unlike in Shaker, at any given time at least two CREVICES WITHIN THE VSD S4 arginines occlude the omega pore of the AND FOCUSING OF THE sodium channel. This is consistent with the MEMBRANE ELECTRIC FIELD fact that an omega current could be recorded Several pieces of evidence suggest that the from the sodium channel in the activated con- S4 helix in the VSD is surrounded by water- formation when the mutated arginines were filled crevices/vestibules on both sides of the R2 and R3 instead of R1 and R2. The need membrane. Accessibility studies with small, of double-arginine substitution in the sodium charged, cysteine-reactive probes that can fit channel as compared with single substitution in polar crevices showed that a large por- in Shaker may indicate that the omega pore in tion of S4 is in contact with either the ex- Shaker is shorter than in Nav1.2a. Also possi- tracellular or the intracellular solutions; only ble is that the voltage-sensor movement dif- a ten-residue stretch of the helix is inacces-

by CAPES on 08/30/07. For personal use only. fers to some extent in the two channels, so that sible (Baker et al. 1998, Larsson et al. 1996, in the resting state Shaker S4 places only its Wang et al. 1999, Yang et al. 1996, Yusaf et al. first arginine in the omega pore, whereas the 1996). A stretch of ten consecutive residues in Nav1.2a S4 places both R1 and R2. an α-helix corresponds to approximately 13 A˚ Taken together, these findings show that in the transmembrane direction (z) if the he- voltage-gated channels have five pores: one lix is placed perpendicularly to the membrane

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org selective pore in the center of the tetramer and plane (Figure 6e). This stretch corresponds four omega pores, one in each VSD. Although to a shorter distance if the helix is tilted, as the gating pore was originally proposed to it appears to be in the Kv1.2 crystal structure surround S4 completely (Figure 6c), more (Long et al. 2005a). recent studies have proposed that S4 can be Islas & Sigworth (2001) studied how the exposed partially (Durell et al. 1998, Elinder reduction of ionic strength of the intracellu- et al. 2001, Li-Smerin et al. 2000, Schonherr lar or extracellular solutions affects the gat- et al. 2002) or extensively (Cuello et al. 2004; ing charge moved by S4 in the Shaker chan- Jiang et al. 2003b, 2004; Lee et al. 2005; Lee nel. Their data and continuum electrostatic & MacKinnon 2004; Ruta et al. 2005) to lipid. calculations are consistent with the presence In the crystal structure of Kv1.2, a portion of of solvent-accessible vestibules in the VSD. the S4 helix faces lipid, but R3 and R4 face They proposed that the vestibules on the two

www.annualreviews.org • Voltage-Induced Channel Activation 39 ANRV288-CB22-02 ARI 30 August 2006 17:1

sides of the membrane are separated by a focus the field on S4. Continuum electrostatic thin septum of 3–7 A˚ and that the intracel- calculations suggest that they need to be as lular vestibule is much deeper than the extra- large as 10–20 A˚ (Chanda et al. 2005, Grabe cellular one. The septum separating the two et al. 2004, Islas & Sigworth 2001). vestibules, or crevices, may well be the location where the omega/proton pore resides. Because their calculations of the septum thick- COUPLING OF VOLTAGE ness depended on the dielectric constant of SENSING TO GATING the surrounding protein, and the dielectric We now have some understanding of the constant may be higher than adopted in their molecular identity and physical mechanism of model, given the charged residues in S2 and gating (see The Activation Gate, above) and of S3 that face the buried arginines in Kv1.2, the the mechanism of voltage sensing (see Mod- estimate of the septum thickness may need to els of Voltage Sensing, above), but we still be revised upward. do not understand how motion of the volt- The presence of crevices in the VSD has age sensor opens the gate. The coupling be- important consequences for the mechanism tween the voltage sensor and gate may be al- of voltage sensing. These crevices reduce the losteric (i.e., the activated conformation of the distance across which the membrane poten- voltage sensor stabilizes the open state of the tial drops on S4. The shorter the distance, the gate) or obligatory (i.e., the gate cannot open stronger the local electric field becomes in the unless the voltage sensor is in the activated S4 region connecting the crevices. Like lenses conformation). It is important to appreciate focusing light on an object, the crevices focus the difference between these two models be- the electric field on S4, which needs to travel a cause they would require different mechan- shorter distance across the membrane to move ical connections between the voltage sensor the same amount of gating charge across the and gate. Moreover, an allosteric scheme of field. Asamoah et al. (2003), using a fluoro- coupling would lend the channel to spon- metric technique, attempted to analyze the taneous openings even at very negative po- electric field profile within the Shaker protein tentials, i.e., voltage-independent openings, and obtained data consistent with electric field but an obligatory model would not. Islas & focusing on S4. More recently, Ahern & Horn Sigworth (1999) determined the probability

by CAPES on 08/30/07. For personal use only. (2005) estimated that the distance over which of spontaneous openings, using the limit- the electric ﬁeld is focused in the Shaker VSD ing slope method (Almers 1978) in combi- is 4–8 A,˚ a distance similar to that deduced by nation with recent advances in measurements Islas & Sigworth (2001) for the septum sepa- of very small open probabilities at negative rating internal and external VSD vestibules. voltages (Hirschberg et al. 1995) and in es- Data from molecular dynamics simula- timation of charge movement per channel

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org tions of a model peptide resembling S4 im- from gating current measurements (Sigg & mersed in a lipid membrane suggest that a Bezanilla 1997). They found that the prob- proteinaceous environment is not absolutely ability for voltage-independent openings is required to form polar crevices around S4 negligible (<10−9), indicating that the chan- (Freites et al. 2005). In the proximity of the nel’s gate is very tightly coupled to gating charged helix, the lipid bilayer can be discharge movement and supporting an obliga- torted enough to allow lipid phosphate groups tory coupling mechanism. Studies on sodium to salt bridge the S4 arginine residues and al- channels found a similar low probability of low water to penetrate partially in the mem- <10−7 (Hirschberg et al. 1995). brane span. However, crevices that are so What parts of the channel are involved in narrow (only a few As˚ wide) are not likely to this coupling? Several studies on Shaker and signiﬁcantly bend the electric ﬁeld lines and related channels suggest that the S4-S5 linker

40 Tombola · Pathak · Isacoff ANRV288-CB22-02 ARI 30 August 2006 17:1

is involved (Caprini et al. 2005; Chen et al. internal end of S6 interact in a manner impor- 2001; Decher et al. 2004; Isacoff et al. 1991; tant for coupling. Lu et al. 2001, 2002; McCormack et al. 1991; In agreement with the above analyses, two HCN channels: Sanguinetti & Xu 1999; Slesinger et al. 1993; extensive studies by Lu & colleagues (Lu hyperpolarization- Tristani-Firouzi et al. 2001). Isacoff et al. et al. 2001, 2002) involving the Shaker, KcsA, activated (1991) found that mutations at ﬁve highly con- and Kv2.1 channels demonstrate that a se- cyclic-nucleotide- served residues on the S4-S5 linker affect the ries of residues in the S4-S5 linker and the gated potassium single channel conductance and/or stability of C-terminal end of S6 are jointly required for channels the ball-inactivated state in full-length Shaker voltage gating. Chimeric channels in which NaChBac: channels, suggesting that this part of the chan- these portions of the channel come from dif- one-domain, voltage-gated nel lies near the permeation pathway for ions ferent channels do not show any voltage de- sodium channel from and serves as a receptor for the inactivating pendence to gating, but voltage-dependent Bacillus halodurans ball, which had been proposed to be associ- gating is preserved when these two parts come ated with the activation gate (Armstrong & from the same channel. The crystal structure Bezanilla 1977, Bezanilla & Armstrong 1977). of Kv1.2 demonstrates this interaction beauti- Moreover, single amino acid substitutions fully (Figure 7a). The distal end of the S4-S5 at several uncharged residues in the S4-S5 linker comes close to the internal end of S6 linker result in a big shift in the conductance- below the PVP motif, with extensive contacts voltage (G-V) relationship of the channel even between side chains of the two regions. The for substitutions as conservative as that of a static picture of what is likely to be the acti- leucine to valine (McCormack et al. 1991). vated conformation suggests that to close the Several of the high-impact leucines are spaced channel’s gate, the S4-S5 linker would have at intervals consistent with a leucine heptad to move inward, clamping shut the S6 gate repeat, suggesting that this motif is impor- (Long et al. 2005b). This model is consistent tant for coupling (McCormack et al. 1991). A with an obligatory mechanism of coupling. well-known leucine mutant in this category, Coupling could be disrupted if the angle at the so-called V2 mutation (L382V in Shaker) which the S6 gate bends opened changes, for (McCormack et al. 1991), alters a late coop- example, owing to a mutation in the PVP re- erative transition in the activation pathway in gion. Some mutations in the S6 region—such which voltage sensing and gating both occur as P475D and P475Q (which alter the second

by CAPES on 08/30/07. For personal use only. (Schoppa et al. 1992, Schoppa & Sigworth P of the PVP motif) in Shaker (Sukhareva et al. 1998b). 2003), which render the channel almost com- Mutations at the internal end of the Shaker pletely constitutively conducting, and L226P S6 also reveal a role of this region in coupling. in the NaChBac channel (Zhao et al. 2004), Mutations at F484 in Shaker show a large neg- which converts this depolarization-activated ative shift in the gating charge-voltage (Q-V) channel into a hyperpolarization-activated

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org relationship and faster gating current on re- channel—may exert their effect by altering or polarization, suggesting that they reduce the eliminating this coupling. “load” on S4 and allow it to move more freely As discussed above, the S4-S5 linker and (Ding & Horn 2002, 2003). Studies on a HCN the C-terminal end of S6 are critical for cou- channel showed a similar disruption of gat- pling. In addition, studies on a mutant of the ing with mutations of charged residues in Shaker channel, the ILT channel (Ledwell & the S4-S5 linker or in a region downstream Aldrich 1999, Smith-Maxwell et al. 1998b), of the S6 segment that connects the chan- show that a motion of S4 distinct from its nel to a nucleotide-binding domain (Chen voltage-sensing motion is associated with the et al. 2001, Sanguinetti & Xu 1999). Takento- ﬁnal opening step (Pathak et al. 2005), sug- gether, these studies about Shaker and HCN gesting that S4 itself plays a role in cou- channels suggest that the S4-S5 linker and the pling voltage sensing with gating. The ILT

www.annualreviews.org • Voltage-Induced Channel Activation 41 ANRV288-CB22-02 ARI 30 August 2006 17:1

Figure 7 (a) Intrasubunit interactions between helix S6 (red residues) and the S4-S5 linker ( green residues) in Kv1.2. Helices S1 through S3 are shown in pale red, S4 and pore-forming helices in gray, and the S4-S5 linker in purple. Potassium ions are colored yellow. (b) Interactions between the VSD and pore-forming helices of adjacent Kv1.2 subunits. Only the VSD of one subunit and the pore-forming helices of the adjacent subunit are shown. Residues at the interface between VSD and pore domain that were shown to have high impact on gating in a tryptophan-scanning mutagenesis study on Shaker are shown in orange (Shaker L398, F402, I405, L409, S412, F433) (Li-Smerin et al. 2000). The ILT residues in S4 (Shaker V369, I372, S376) are colored red. The V2 position on the S4-S5 linker (Shaker L382) is colored green. The Sh5 position in S5 (Shaker F401) is colored dark blue. Helices are color coded as in a, except for the pore-forming helices (of the adjacent subunit), which are shown in blue.

channel has three conservative substitutions at WORKING TOGETHER TO uncharged residues in the Shaker S4 (V369I, OPEN THE GATE I372L, S376T), which isolate the final open- How do four separate voltage sensors to- ing step without altering the molecular mech- gether control the conformation of the chan- anisms of channel activation (Ledwell & nel’s gate? The VSDs are not in direct contact Aldrich 1999, Pathak et al. 2005, Webster with one another, so they could, in prin- et al. 2004). Gating current measurements ciple, operate independently. However, evi- show a small charge movement associated by CAPES on 08/30/07. For personal use only. dence for cooperativity comes from studies with the final opening step (Ledwell & Aldrich of potassium channels in which kinetics and 1999, Smith-Maxwell et al. 1998b), and fluo- steady-state behavior of wild-type (Bezanilla rescence measurements from the outer end et al. 1994; Hoshi et al. 1994; Schoppa & of the ILT S4 show that S4 movement is as- Sigworth 1998a,c; Sigworth 1994; Stefani sociated with this step (Pathak et al. 2005). et al. 1994; Zagotta et al. 1994a,b), mutant S4 thus appears to have two distinct phases Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org (Kanevsky & Aldrich 1999; McCormack et al. of motion: the voltage-sensing motion (dis- 1991; Schoppa et al. 1992; Schoppa & Sig- cussed above; see Models of Voltage Sensing) worth 1998b,c), chimeric (Ledwell & Aldrich and an additional motion termed the gating 1999; Shieh et al. 1997; Smith-Maxwell et al. motion. This gating motion of S4 carries little 1998a,b), and heterotetrameric (Hurst et al. charge and unlike the voltage-sensing motion 1992, Tytgat & Hess 1992) channels were of S4 is highly cooperative between the four measured, and kinetic models developed, to S4s and can drive the conformational changes simulate this behavior (for a detailed discus- in the coupling machinery that open the sion, see review by Fedida & Hesketh 2001). channel’s gate. Immediately below we discuss Although the details of the channels and ki- further the role of cooperativity in channel netic models of the studies differ, a common opening. feature of all the models is that some form

42 Tombola · Pathak · Isacoff ANRV288-CB22-02 ARI 30 August 2006 17:1

of cooperativity was needed to account for as important to cooperativity as the interac- the experimental data. The generally accepted tion between the S4-S5 linker and S6 is to cou- idea now is that activation involves multi- pling. Further molecular details of how coop- Sh5: a mutant step movement of the voltage sensors. The erativity is generated remain to be worked out, Shaker allele that early steps, which carry the majority of the but with increasing structural information and induces the fly nerve charge, occur in a manner that is mostly in- a wealth of mutagenesis data, we should see to fire rapid bursts of dependent in the four subunits, and the late answers coming in over the next few years. action potentials steps, which have smaller voltage dependence, are highly cooperative (Schoppa & Sigworth 1998c, Zagotta et al. 1994a). LIPID: THE MISSING PLAYER? What is the molecular basis of the cooper- A quite remarkable feature of the Kv1.2 struc- ative interactions in the channel? Mutations ture is that the perimeter of the channel pro- that disrupt the cooperative steps in chan- tein is very uneven and presents deep grooves nel activation carry some clues. The L382V at the four edges of the channel (Figure 8). mutation in the S4-S5 linker discussed above The four VSDs, positioned at the corners of (see Coupling of Voltage Sensing to Gat- the square-shaped pore domain, protrude lat- ing) and the F401I mutation in S5 that is erally in the membrane, creating hydropho- responsible for the Sh5 phenotype (Gautam bic grooves that are expected to fill with lipid. & Tanouye 1990, Kanevsky & Aldrich 1999, The way the lipid is packed in the grooves Lichtinghagen et al. 1990) both seem to reduce the cooperative stabilization of the open state without altering either the total charge moved by the channel or the voltage dependence of the early transitions. Their effects can be explained by altering the final opening transition(s) (Kanevsky & Aldrich 1999, Schoppa & Sigworth 1998c), suggesting that these residues are involved in important interactions in the open state. The ILT mutations seem to exert their effect differently and

by CAPES on 08/30/07. For personal use only. appear to stabilize a state close to, but distinct from, the open state, so that entry into the resting state requires a stronger negative voltage and the cooperative motion of S4 that opens the gate requires stronger depolarization (Ledwell & Aldrich 1999, Pathak et al.

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org 2005). The effect of the ILT mutations was proposed to be due to interactions between a hydrophobic face of S4 and the neighboring subunit’s S5 (Pathak et al. 2005). This idea is supported by functional evidence that the Figure 8 outer ends of S4 and a neighboring subunit’s Surface representation of the Kv1.2 channel surrounded by lipid (top view, S5 are close to each other in the open state extracellular perspective). Each of the four subunits is shown in a different of the wild-type channel (Laine et al. 2003); color. White dotted lines encircle Kv1.2 positions S392 and A395 in each this is consistent with the crystal structure of subunit. The effect of bupivacaine on Kv channels is altered by mutations at these positions (Nilsson et al. 2003). The yellow dots indicate position K388 Kv1.2 (Long et al. 2005a) (Figure 7b). These in each subunit. See explanation in text. The S2 region in each subunit “daisy-chain interactions” between the S4 of proposed to be exposed to lipid by Monks et al. (1999) is colored gray. Blue one subunit and the S5 of its neighbor may be triangles point to the approximate location of the binding site for hanatoxin.

www.annualreviews.org • Voltage-Induced Channel Activation 43 ANRV288-CB22-02 ARI 30 August 2006 17:1

may be different from the way it is packed 1994, Oliver et al. 2004, Villarroel & Schwarz in the rest of the membrane, and it may be 1996), but at present the mechanism of ac- in slow exchange with the bulk. This may tion of these molecules on Kv channels is not have important consequences for the reg- understood. ulation of Kv channels by lipid and lipid- The KcsA channel can be cocrystallized soluble factors. Several lipophilic compounds with four negatively charged lipid molecules, are known to modulate the activity of Kv one bound to each subunit (Valiyaveetil et al. channels. For example, phosphatidylinositol- 2002). Negatively charged lipids have been

4,5-bisphosphate (PIP2) has been shown to shown to be important for channel function in act as a docking platform for the N-terminal KcsA (Heginbotham et al. 1998, Valiyaveetil “ball” domain of fast-inactivating Kv chan- et al. 2002). Although the presence in Kv1.2

nels (Oliver et al. 2004). PIP2 on the mem- of a binding site for lipid similar to that found brane’s intracellular leaﬂet sequesters the ball in KcsA has not been yet investigated, of in- domain and inhibits fast inactivation. It had terest is that the location of the homologous been a mystery how this worked. Does the binding site would be deep in the Kv1.2 lateral ball dangle on a chain that is long enough for groove (Figure 8, yellow dots). it to reach the lipid? The N-terminal balls of Gating modiﬁers like the peptide hana- the four subunits are believed to reach their toxin bind to the voltage sensor of Kv channels binding site in the intracellular mouth of the after partitioning into the membrane (Lee & pore through four lateral windows located in MacKinnon 2004, Phillips et al. 2005). The the “hanging gondola” between the channel blue triangles in Figure 8 point to the ap- transmembrane region and the tetrameriza- proximate location of the hanatoxin-binding tion domain; the same windows allow potas- sites (Swartz & MacKinnon 1997). Concen- sium ions to access the channel pore from the trations of toxin in the 100 nM range effec- intracellular side. For ball inactivation (block tively inhibit the Kv2.1 channel, suggesting of internal mouth of the pore) to occur within strong binding (Lee et al. 2003, Phillips et al. one or two milliseconds, the ball cannot wan- 2005, Swartz & MacKinnon 1997) and an ex- der far from the mouth of the channel, and tensive interacting surface between channel

yet for PIP2 modulation it also would have and toxin. However, only a few residues in to be near the lipid. It turns out that these the channel have been identiﬁed as making

by CAPES on 08/30/07. For personal use only. two requirements are not mutually exclusive. contact with hanatoxin (Swartz & MacKinnon The lateral grooves (and lipid) in the trans- 1997). To explain this apparent paradox, the membrane region of Kv1.2 are positioned toxin has been proposed to concentrate in the right above the lateral windows (Figure 8, red lipid membrane (Lee & MacKinnon 2004, arrows). Phillips et al. 2005). The binding to the chan- Nav and Kv channels are inhibited by lo- nel may be weaker than is indicated by the

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org cal anesthetics (Hille 2001). Lipophilic local apparent dissociation constant, and still the anesthetics, such as lidocaine and bupivacaine, fraction of channels bound to the toxin may can reach the pore domain after partition- be high because of the high concentration of ing into the membrane. The location of two toxin in the lipid. Nevertheless, in a recent residues on the pore domain that have been study Swartz and colleagues showed that the

proposed to modulate binding of bupivacaine low value of the hanatoxin Kd is primarily due to Kv channels (Nilsson et al. 2003) is shown to strong protein binding, not to preconcen- in Figure 8 (white dotted line). This location tration in the membrane (Phillips et al. 2005), can be reached from the lipid only through the so the puzzle of strong binding despite the few lateral grooves in the channel protein. Some channel-toxin contacts remains. An intriguing Kv channels are inhibited by polyunsaturated explanation may be that hanatoxin may not fatty acids and anandamide (e.g., Honore et al. recognize the channel by itself but rather a

44 Tombola · Pathak · Isacoff ANRV288-CB22-02 ARI 30 August 2006 17:1

combination of channel and ordered lipid that et al. 2003a, Lee et al. 2005) highlights the is clustered at the channel’s periphery. Thus, importance of the lipid in holding the dif- the lipid clusters or patches may provide an ferent domains of the channel together. Al- extra interacting surface for the binding. The though much still needs to be learned about channel’s lateral grooves may be ideal places the interactions between voltage-gated chan- to accommodate such lipid clusters. nels and lipid, especially where the S4 helix The ﬁnding that KvAP assumes a nonna- is concerned, such interactions are likely to tive conformation when it is removed from play an important role in the modulation of the lipid membrane (Cuello et al. 2004, Jiang channel activity.

SUMMARY POINTS 1. The activation gate of voltage-gated ion channels is created by the four S6 helices (transmembrane segments 2 in 2 TM channels) of the pore domain. A PVP motif in S6 is the hinge of the activation gate in Shaker and related eukaryotic Kv channels. In bacterial Kvs, in which the PVP motif is absent, the gating hinge is at a conserved glycine. 2. In the Kv1.2 structure, the four VSDs are positioned at the corners of the square- shaped pore domain. S4 interacts on one side with S5 in the pore domain, is exposed to lipid on its second side, and on its third side faces into the heart of the VSD. 3. Water-ﬁlled vestibules in the VSD focus the membrane electric ﬁeld on the S4 arginines. The arginines move across pores connecting internal and external VSD vestibules. When an arginine is shortened by mutation, the pathway that normally conducts its guanadinium group can conduct solution ions. 4. There are three major conceptual models of voltage sensing: the transporter model, the helical screw model, and the paddle model. They differ in the extent and nature of S4 transmembrane movement and in terms of whether the S4 arginines are exposed to the lipid hydrophobic core. 5. The S4-S5 linker and C-terminal end of S6 in the same subunit are important for

by CAPES on 08/30/07. For personal use only. coupling voltage sensing to gating. The crystal structure of Kv1.2 shows that side chains of these regions form extensive contacts with each other. 6. The S4 of one subunit contacts the S5 of the adjacent subunit in the crystal structure of Kv1.2. These intersubunit contacts may mediate known cooperative interactions in the channel.

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org 7. The irregularity of the perimeter of Kv1.2 exposure to lipid poses interesting questions about the organization of the lipid around the channel protein.

FUTURE ISSUES 1. The conformation of the VSD in the resting state remains to be determined. 2. There remains a discrepancy between the apparently small voltage-sensor movement of eukaryotic Kv channels and the apparently large movement measured in KvAP. Two possible reconciliations are suggested, but new experiments will be required to resolve the question.

www.annualreviews.org • Voltage-Induced Channel Activation 45 ANRV288-CB22-02 ARI 30 August 2006 17:1

3. There still remains a need for molecular analysis to explain how voltage-sensor motion is coupled to the motion of the channel gates and how cooperative interactions between subunits are mediated. 4. A better understanding of the interactions between channel protein and membrane lipid, both in terms of stabilization of the channel structure and in terms of regulation of channel activity, is required.

ACKNOWLEDGMENTS We are grateful to Bruce Cohen, Arnd Pralle, Susy Kohout, and Gautam Agarwal for insightful comments on the manuscript. The authors’ research on voltage-gated potassium channels is supported by the National Institutes of Health (grant 2R01NS35549-9A2 to E.I.) and by the American Heart Association WSA (postdoctoral fellowship to F.T.).

LITERATURE CITED Aggarwal SK, MacKinnon R. 1996. Contribution of the S4 segment to gating charge in the + Shaker K channel. Neuron 16:1169–77 Ahern CA, Horn R. 2004a. Specificity of charge-carrying residues in the voltage sensor of potassium channels. J. Gen. Physiol. 123:205–16 Ahern CA, Horn R. 2004b. Stirring up controversy with a voltage sensor paddle. Trends Neurosci. 27:303–7 Ahern CA, Horn R. 2005. Focused electric field across the voltage sensor of potassium channels. Neuron 48:25–29 Almers W. 1978. Gating currents and charge movements in excitable membranes. Rev. Physiol. Biochem. Pharmacol. 82:96–190 Armstrong CM. 1966. Time course of TEA+-induced anomalous rectification in squid giant

by CAPES on 08/30/07. For personal use only. axons. J. Gen. Physiol. 50:491–503 Armstrong CM. 1969. Inactivation of the potassium conductance and related phenomena caused by quaternary ammonium ion injection in squid axons. J. Gen. Physiol. 54:553– 75 Armstrong CM. 1971. Interaction of tetraethylammonium ion derivatives with the potassium channels of giant axons. J. Gen. Physiol. 58:413–37

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org Armstrong CM. 2003. Voltage-gated K channels. Sci. STKE 2003:re10 Armstrong CM, Bezanilla F. 1977. Inactivation of the sodium channel. II. Gating current experiments. J. Gen. Physiol. 70:567–90 Asamoah OK, Wuskell JP, Loew LM, Bezanilla F. 2003. A ﬂuorometric approach to local electric ﬁeld measurements in a voltage-gated ion channel. Neuron 37:85–97 Baker OS, Larsson HP, Mannuzzu LM, Isacoff EY. 1998. Three transmembrane conformations and sequence-dependent displacement of the S4 domain in shaker K+ channel gating. Neuron 20:1283–94 Bezanilla F. 2000. The voltage sensor in voltage-dependent ion channels. Physiol. Rev. 80:555– 92 Bezanilla F. 2002. Voltage sensor movements. J. Gen. Physiol. 120:465–73 Bezanilla F. 2005. Voltage-gated ion channels. IEEE Trans. Nanobiosci. 4:34–48

46 Tombola · Pathak · Isacoff ANRV288-CB22-02 ARI 30 August 2006 17:1

Bezanilla F, Armstrong CM. 1977. Inactivation of the sodium channel. I. Sodium current experiments. J. Gen. Physiol. 70:549–66 Bezanilla F, Perozo E, Stefani E. 1994. Gating of Shaker K+ channels. II. The components of gating currents and a model of channel activation. Biophys. J. 66:1011–21 Bezanilla F, Stefani E. 1998. Gating currents. Methods Enzymol. 293:331–52 Broomand A, Mannikko R, Larsson HP, Elinder F. 2003. Molecular movement of the voltage sensor in a K channel. J. Gen. Physiol. 122:741–48 Caprini M, Fava M, Valente P, Fernandez-Ballester G, Rapisarda C, et al. 2005. Molecular compatibility of the channel gate and the N terminus of S5 segment for voltage-gated channel activity. J. Biol. Chem. 280:18253–64 Catterall WA. 1986. Molecular properties of voltage-sensitive sodium channels. Annu. Rev. Biochem. 55:953–85 Cha A, Snyder GE, Selvin PR, Bezanilla F. 1999. Atomic scale movement of the voltage-sensing region in a potassium channel measured via spectroscopy. Nature 402:809–13 Chanda B, Asamoah OK, Blunck R, Roux B, Bezanilla F. 2005. Gating charge displacement in voltage-gated ion channels involves limited transmembrane movement. Nature 436:852– 56 Chen J, Mitcheson JS, Tristani-Firouzi M, Lin M, Sanguinetti MC. 2001. The S4-S5 linker couples voltage sensing and activation of pacemaker channels. Proc. Natl. Acad. Sci. USA 98:11277–82 Choi KL, Mossman C, Aube J, Yellen G. 1993. The internal quaternary ammonium receptor site of Shaker potassium channels. Neuron 10:533–41 Cohen BE, Grabe M, Jan LY. 2003. Answers and questions from the KvAP structures. Neuron 39:395–400 Cuello LG, Cortes DM, Perozo E. 2004. Molecular architecture of the KvAP voltage- + dependent K channel in a lipid bilayer. Science 306:491–95 Decher N, Chen J, Sanguinetti MC. 2004. Voltage-dependent gating of hyperpolarization- activated, cyclic nucleotide-gated pacemaker channels: molecular coupling between the S4-S5 and C-linkers. J. Biol. Chem. 279:13859–65 del Camino D, Holmgren M, Liu Y, Yellen G. 2000. Blocker protection in the pore of a + by CAPES on 08/30/07. For personal use only. voltage-gated K channel and its structural implications. Nature 403:321–25 del Camino D, Yellen G. 2001. Tight steric closure at the intracellular activation gate of a + voltage-gated K channel. Neuron 32:649–56 Demo SD, Yellen G. 1991. The inactivation gate of the Shaker K+ channel behaves like an open-channel blocker. Neuron 7:743–53 Ding S, Horn R. 2002. Tail end of the S6 segment: role in permeation in Shaker potassium

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org channels. J. Gen. Physiol. 120:87–97 Ding S, Horn R. 2003. Effect of S6 tail mutations on charge movement in Shaker potassium channels. Biophys. J. 84:295–305 Doyle DA, Morais Cabral J, Pfuetzner RA, Kuo A, Gulbis JM, et al. 1998. The structure of the + potassium channel: molecular basis of K conduction and selectivity. Science 280:69–77 Durell SR, Hao Y, Guy HR. 1998. Structural models of the transmembrane region of voltage- + gated and other K channels in open, closed, and inactivated conformations. J. Struct. Biol. 121:263–84 Durell SR, Shrivastava IH, Guy HR. 2004. Models of the structure and voltage-gating mech- + anism of the shaker K channel. Biophys. J. 87:2116–30 Elinder F, Mannikko R, Larsson HP. 2001. S4 charges move close to residues in the pore domain during activation in a K channel. J. Gen. Physiol. 118:1–10

www.annualreviews.org • Voltage-Induced Channel Activation 47 ANRV288-CB22-02 ARI 30 August 2006 17:1

Elliott DJ, Neale EJ, Aziz Q, Dunham JP, Munsey TS, et al. 2004. Molecular mechanism of voltage sensor movements in a potassium channel. EMBO J. 23:4717–26 Fedida D, Hesketh JC. 2001. Gating of voltage-dependent potassium channels. Prog. Biophys. Mol. Biol. 75:165–99 Freites JA, Tobias DJ, von Heijne G, White SH. 2005. Interface connections of a transmembrane voltage sensor. Proc. Natl. Acad. Sci. USA 102:15059–64 French RJ, Zamponi GW. 2005. Voltage-gated sodium and calcium channels in nerve, muscle, and heart. IEEE Trans. Nanobiosci. 4:58–69 Gandhi CS, Clark E, Loots E, Pralle A, Isacoff EY. 2003. The orientation and molecular + movement of a K channel voltage-sensing domain. Neuron 40:515–25 Gandhi CS, Isacoff EY. 2002. Molecular models of voltage sensing. J. Gen. Physiol. 120:455–63 Gautam M, Tanouye MA. 1990. Alteration of potassium channel gating: molecular analysis of the Drosophila Sh5 mutation. Neuron 5:67–73 Goldstein SA. 1996. A structural vignette common to voltage sensors and conduction pores: canaliculi. Neuron 16:717–22 Gouaux E, Mackinnon R. 2005. Principles of selective ion transport in channels and pumps. Science 310:1461–65 Grabe M, Lecar H, Jan YN, Jan LY. 2004. A quantitative assessment of models for voltage- dependent gating of ion channels. Proc. Natl. Acad. Sci. USA 101:17640–45 Guy HR, Seetharamulu P. 1986. Molecular model of the action potential sodium channel. Proc. Natl. Acad. Sci. USA 83:508–12 Heginbotham L, Kolmakova-Partensky L, Miller C. 1998. Functional reconstitution of a + prokaryotic K channel. J. Gen. Physiol. 111:741–49 Hessa T, White SH, von Heijne G. 2005. Membrane insertion of a potassium-channel voltage sensor. Science 307:1427 Hille B. 2001. Ion Channels of Excitable Membranes. Sunderland, MA: Sinauer Assoc. Inc. Hirschberg B, Rovner A, Lieberman M, Patlak J. 1995. Transfer of twelve charges is needed + to open skeletal muscle Na channels. J. Gen. Physiol. 106:1053–68 Hodgkin AL, Huxley AF. 1952. A quantitative description of membrane current and its appli- cation to conduction and excitation in nerve. J. Physiol. 117:500–44 + by CAPES on 08/30/07. For personal use only. Holmgren M, Shin KS, Yellen G. 1998. The activation gate of a voltage-gated K channel can be trapped in the open state by an intersubunit metal bridge. Neuron 21:617–21 + Hong KH, Miller C. 2000. The lipid-protein interface of a Shaker K( ) channel. J. Gen. Physiol. 115:51–8 Honore E, Barhanin J, Attali B, Lesage F, Lazdunski M. 1994. External blockade of the major + cardiac delayed-rectiﬁer K channel (Kv1.5) by polyunsaturated fatty acids. Proc. Natl. Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org Acad. Sci. USA 91:1937–41 Horn R. 2005. How ion channels sense membrane potential. Proc. Natl. Acad. Sci. USA 102:4929–30 Hoshi T, Zagotta WN, Aldrich RW. 1990. Biophysical and molecular mechanisms of Shaker potassium channel inactivation. Science 250:533–38 Hoshi T, Zagotta WN, Aldrich RW. 1994. Shaker potassium channel gating. I: Transitions near the open state. J. Gen. Physiol. 103:249–78 Hurst RS, Kavanaugh MP, Yakel J, Adelman JP, North RA. 1992. Cooperative interactions among subunits of a voltage-dependent potassium channel. Evidence from expression of concatenated cDNAs. J. Biol. Chem. 267:23742–45 Isacoff EY, Jan YN, Jan LY. 1991. Putative receptor for the cytoplasmic inactivation gate in + the Shaker K channel. Nature 353:86–90

48 Tombola · Pathak · Isacoff ANRV288-CB22-02 ARI 30 August 2006 17:1

Islas LD, Sigworth FJ. 1999. Voltage sensitivity and gating charge in Shaker and Shab family potassium channels. J. Gen. Physiol. 114:723–42 Islas LD, Sigworth FJ. 2001. Electrostatics and the gating pore of Shaker potassium channels. J. Gen. Physiol. 117:69–89 Iverson LE, Rudy B. 1990. The role of the divergent amino and carboxyl domains on the inactivation properties of potassium channels derived from the Shaker gene of Drosophila. J. Neurosci. 10:2903–16 Jiang QX, Wang DN, MacKinnon R. 2004. Electron microscopic analysis of KvAP voltage- + dependent K channels in an open conformation. Nature 430:806–10 Jiang Y, Lee A, Chen J, Cadene M, Chait BT, MacKinnon R. 2002. Crystal structure and mechanism of a calcium-gated potassium channel. Nature 417:515–22 Jiang Y, Lee A, Chen J, Ruta V, Cadene M, et al. 2003a. X-ray structure of a voltage-dependent + K channel. Nature 423:33–41 Jiang Y, Ruta V, Chen J, Lee A, MacKinnon R. 2003b. The principle of gating charge movement + in a voltage-dependent K channel. Nature 423:42–48 Kanevsky M, Aldrich RW. 1999. Determinants of voltage-dependent gating and open-state stability in the S5 segment of Shaker potassium channels. J. Gen. Physiol. 114:215–42 Kelly BL, Gross A. 2003. Potassium channel gating observed with site-directed mass tagging. Nat. Struct. Biol. 10:280–84 Keynes RD, Elinder F. 1999. The screw-helical voltage gating of ion channels. Proc. Biol. Sci. 266:843–52 Koishi R, Xu H, Ren D, Navarro B, Spiller BW, Shi Q, Clapham DE. 2004. A superfamily of voltage-gated sodium channels in bacteria. J. Biol. Chem. 279:9532–38 Kuo A, Gulbis JM, Antcliff JF, Rahman T, Lowe ED, et al. 2003. Crystal structure of the potassium channel KirBac1.1 in the closed state. Science 300:1922–26 Lai HC, Grabe M, Jan YN, Jan LY. 2005. The S4 voltage sensor packs against the pore domain in the KAT1 voltage-gated potassium channel. Neuron 47:395–406 Laine M, Lin MC, Bannister JP, Silverman WR, Mock AF, et al. 2003. Atomic proximity between S4 segment and pore domain in Shaker potassium channels. Neuron 39:467–81 Laine M, Papazian DM, Roux B. 2004. Critical assessment of a proposed model of Shaker.

by CAPES on 08/30/07. For personal use only. FEBS Lett. 564:257–63 Larsson HP, Baker OS, Dhillon DS, Isacoff EY. 1996. Transmembrane movement of the + shaker K channel S4. Neuron 16:387–97 Lecar H, Larsson HP, Grabe M. 2003. Electrostatic model of S4 motion in voltage-gated ion channels. Biophys. J. 85:2854–64 Ledwell JL, Aldrich RW. 1999. Mutations in the S4 region isolate the ﬁnal voltage-dependent

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org cooperative step in potassium channel activation. J. Gen. Physiol. 113:389–414 Lee HC, Wang JM, Swartz KJ. 2003. Interaction between extracellular Hanatoxin and the resting conformation of the voltage-sensor paddle in Kv channels. Neuron 40:527–36 Lee SY, Lee A, Chen J, Mackinnon R. 2005. Structure of the KvAP voltage-dependent K+ channel and its dependence on the lipid membrane. Proc. Natl. Acad. Sci. USA 102:15441– 46 Lee SY, MacKinnon R. 2004. A membrane-access mechanism of ion channel inhibition by voltage sensor toxins from spider venom. Nature 430:232–35 Li-Smerin Y, Hackos DH, Swartz KJ. 2000. α-helical structural elements within the voltage- + sensing domains of a K channel. J. Gen. Physiol. 115:33–50 Lichtinghagen R, Stocker M, Wittka R, Boheim G, Stuhmer W, et al. 1990. Molecular basis of altered excitability in Shaker mutants of Drosophila melanogaster. EMBO J. 9:4399–407

www.annualreviews.org • Voltage-Induced Channel Activation 49 ANRV288-CB22-02 ARI 30 August 2006 17:1

Liu Y, Holmgren M, Jurman ME, Yellen G. 1997. Gated access to the pore of a voltage- + dependent K channel. Neuron 19:175–84 Liu YS, Sompornpisut P, Perozo E. 2001. Structure of the KcsA channel intracellular gate in the open state. Nat. Struct. Biol. 8:883–87 Long SB, Campbell EB, Mackinnon R. 2005a. Crystal structure of a mammalian voltage- + dependent Shaker family K channel. Science 309:897–903 Long SB, Campbell EB, Mackinnon R. 2005b. Voltage sensor of Kv1.2: structural basis of electromechanical coupling. Science 309:903–8 Lopez GA, Jan YN, Jan LY. 1994. Evidence that the S6 segment of the Shaker voltage-gated + K channel comprises part of the pore. Nature 367:179–82 Lu Z, Klem AM, Ramu Y. 2001. Ion conduction pore is conserved among potassium channels. Nature 413:809–13 Lu Z, Klem AM, Ramu Y. 2002. Coupling between voltage sensors and activation gate in + voltage-gated K channels. J. Gen. Physiol. 120:663–76 MacKinnon R. 2003. Potassium channels. FEBS Lett. 555:62–65 McCormack K, Tanouye MA, Iverson LE, Lin JW, Ramaswami M, et al. 1991. A role for + hydrophobic residues in the voltage-dependent gating of Shaker K channels. Proc. Natl. Acad. Sci. USA 88:2931–35 Monks SA, Needleman DJ, Miller C. 1999. Helical structure and packing orientation of the + S2 segment in the Shaker K channel. J. Gen. Physiol. 113:415–23 Monticelli L, Robertson KM, MacCallum JL, Tieleman DP. 2004. Computer simulation of the KvAP voltage-gated potassium channel: steered molecular dynamics of the voltage sensor. FEBS Lett. 564:325–32 Murata Y, Iwasaki H, Sasaki M, Inaba K, Okamura Y. 2005. Phosphoinositide phosphatase activity coupled to an intrinsic voltage sensor. Nature 435:1239–43 Neale EJ, Elliott DJ, Hunter M, Sivaprasadarao A. 2003. Evidence for intersubunit interactions between S4 and S5 transmembrane segments of the Shaker potassium channel. J. Biol. Chem. 278:29079–85 Nilsson J, Madeja M, Arhem P. 2003. Local anesthetic block of Kv channels: role of the S6 helix and the S5-S6 linker for bupivacaine action. Mol. Pharmacol. 63:1417–29

by CAPES on 08/30/07. For personal use only. Oliver D, Lien CC, Soom M, Baukrowitz T, Jonas P, Fakler B. 2004. Functional conversion + between A-type and delayed rectiﬁer K channels by membrane lipids. Science 304:265–70 Papazian DM, Shao XM, Seoh SA, Mock AF, Huang Y, Wainstock DH. 1995. Electrostatic + interactions of S4 voltage sensor in Shaker K channel. Neuron 14:1293–301 Parsegian A. 1969. Energy of an ion crossing a low dielectric membrane: solutions to four relevant electrostatic problems. Nature 221:844–46

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org Pathak M, Kurtz L, Tombola F, Isacoff E. 2005. The cooperative voltage sensor motion that gates a potassium channel. J. Gen. Physiol. 125:57–69 Phillips LR, Milescu M, Li-Smerin Y, Mindell JA, Kim JI, Swartz KJ. 2005. Voltage-sensor activation with a tarantula toxin as cargo. Nature 436:857–60 Posson DJ, Ge P, Miller C, Bezanilla F, Selvin PR. 2005. Small vertical movement of a K+ channel voltage sensor measured with luminescence energy transfer. Nature 436:848–51 Ramsey IS, Moran MM, Chong JA, Clapham DE. 2006. A voltage-gated proton-selective channel lacking the pore domain. Nature 440:1213–16 Ren D, Navarro B, Xu H, Yue L, Shi Q, Clapham DE. 2001. A prokaryotic voltage-gated sodium channel. Science 294:2372–75 + Roux B. 2005. Ion conduction and selectivity in K channels. Annu. Rev. Biophys. Biomol. Struct. 34:153–71

50 Tombola · Pathak · Isacoff ANRV288-CB22-02 ARI 30 August 2006 17:1

Ruta V, Chen J, MacKinnon R. 2005. Calibrated measurement of gating charge arginine + displacement in the KvAP voltage-dependent K channel. Cell 123:463–75 Sanguinetti MC, Xu QP. 1999. Mutations of the S4-S5 linker alter activation properties of HERG potassium channels expressed in Xenopus oocytes. J. Physiol. 514(Pt. 3):667–75 Sasaki M, Takagi M, Okamura Y. 2006. A voltage sensor-domain protein is a voltage-gated proton channel. Science 312:589–92 Sather WA, McCleskey EW. 2003. Permeation and selectivity in calcium channels. Annu. Rev. Physiol. 65:133–59 Schonherr R, Mannuzzu LM, Isacoff EY, Heinemann SH. 2002. Conformational switch between slow and fast gating modes: allosteric regulation of voltage sensor mobility in the + EAG K channel. Neuron 35:935–49 Schoppa NE, McCormack K, Tanouye MA, Sigworth FJ. 1992. The size of gating charge in wild-type and mutant Shaker potassium channels. Science 255:1712–15 Schoppa NE, Sigworth FJ. 1998a. Activation of Shaker potassium channels. I. Characterization of voltage-dependent transitions. J. Gen. Physiol. 111:271–94 Schoppa NE, Sigworth FJ. 1998b. Activation of Shaker potassium channels. II. Kinetics of the V2 mutant channel. J. Gen. Physiol. 111:295–311 Schoppa NE, Sigworth FJ. 1998c. Activation of Shaker potassium channels. III. An activation gating model for wild-type and V2 mutant channels. J. Gen. Physiol. 111:313–42 Seoh SA, Sigg D, Papazian DM, Bezanilla F. 1996. Voltage-sensing residues in the S2 and S4 + segments of the Shaker K channel. Neuron 16:1159–67 Shieh CC, Klemic KG, Kirsch GE. 1997. Role of transmembrane segment S5 on gating of + voltage-dependent K channels. J. Gen. Physiol. 109:767–78 Shrivastava IH, Durell SR, Guy HR. 2004. A model of voltage gating developed using the KvAP channel crystal structure. Biophys. J. 87:2255–70 Sigg D, Bezanilla F. 1997. Total charge movement per channel. The relation between gating charge displacement and the voltage sensitivity of activation. J. Gen. Physiol. 109:27–39 Sigworth FJ. 1994. Voltage gating of ion channels. Q. Rev. Biophys. 27:1–40 Slesinger PA, Jan YN, Jan LY. 1993. The S4-S5 loop contributes to the ion-selective pore of potassium channels. Neuron 11:739–49

by CAPES on 08/30/07. For personal use only. Smith-Maxwell CJ, Ledwell JL, Aldrich RW. 1998a. Role of the S4 in cooperativity of voltage- dependent potassium channel activation. J. Gen. Physiol. 111:399–420 Smith-Maxwell CJ, Ledwell JL, Aldrich RW. 1998b. Uncharged S4 residues and cooperativity in voltage-dependent potassium channel activation. J. Gen. Physiol. 111:421–39 Sokolov S, Scheuer T, Catterall WA. 2005. Ion permeation through a voltage-sensitive gating pore in brain sodium channels having voltage sensor mutations. Neuron 47:183–89

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org Soler-Llavina GJ, Holmgren M, Swartz KJ. 2003. Deﬁning the conductance of the closed state + in a voltage-gated K channel. Neuron 38:61–67 Starace DM, Bezanilla F. 2001. Histidine scanning mutagenesis of basic residues of the S4 + segment of the shaker K channel. J. Gen. Physiol. 117:469–90 Starace DM, Bezanilla F. 2004. A proton pore in a potassium channel voltage sensor reveals a focused electric ﬁeld. Nature 427:548–53 Starace DM, Stefani E, Bezanilla F. 1997. Voltage-dependent proton transport by the voltage + sensor of the Shaker K channel. Neuron 19:1319–27 Stefani E, Toro L, Perozo E, Bezanilla F. 1994. Gating of Shaker K+ channels: I. Ionic and gating currents. Biophys. J. 66:996–1010 Sukhareva M, Hackos DH, Swartz KJ. 2003. Constitutive activation of the Shaker Kv channel. J. Gen. Physiol. 122:541–56

www.annualreviews.org • Voltage-Induced Channel Activation 51 ANRV288-CB22-02 ARI 30 August 2006 17:1

Swartz KJ. 2004. Towardsa structural view of gating in potassium channels. Nat. Rev. Neurosci. 5:905–16 Swartz KJ, MacKinnon R. 1997. Mapping the receptor site for hanatoxin, a gating modifier of + voltage-dependent K channel. Neuron 18:675–82 Tiwari-Woodruff SK, Lin MA, Schulteis CT, Papazian DM. 2000. Voltage-dependent struc- + tural interactions in the Shaker K channel. J. Gen. Physiol. 115:123–38 Tombola F, Pathak MM, Isacoff EY. 2005a. How far will you go to sense voltage? Neuron 48:719–25 Tombola F, Pathak MM, Isacoff EY. 2005b. Voltage-sensing arginines in a potassium channel permeate and occlude cation-selective pores. Neuron 45:379–88 Tristani-Firouzi M, Chen J, Mitcheson JS, Sanguinetti MC. 2001. Molecular biology of K+ channels and their role in cardiac arrhythmias. Am. J. Med. 110:50–59 Tytgat J, Hess P. 1992. Evidence for cooperative interactions in potassium channel gating. Nature 359:420–23 Valiyaveetil FI, Zhou Y, MacKinnon R. 2002. Lipids in the structure, folding, and function of + the KcsA K channel. Biochemistry 41:10771–77 Villarroel A, Schwarz TL. 1996. Inhibition of the Kv4 (Shal) family of transient K+ currents by arachidonic acid. J. Neurosci. 16:1016–25 Wang MH, Yusaf SP, Elliott DJ, Wray D, Sivaprasadarao A. 1999. Effect of cysteine substitutions on the topology of the S4 segment of the Shaker potassium channel: implications for molecular models of gating. J. Physiol. 521(Pt. 2):315–26 Webster SM, Del Camino D, Dekker JP, Yellen G. 2004. Intracellular gate opening in Shaker + K channels defined by high-affinity metal bridges. Nature 428:864–68 Yang N, George AL Jr, Horn R. 1996. Molecular basis of charge movement in voltage-gated sodium channels. Neuron 16:113–22 Yang N, George AL Jr, Horn R. 1997. Probing the outer vestibule of a sodium channel voltage sensor. Biophys. J. 73:2260–68 Yarov-Yarovoy V, Baker D, Catterall WA. 2006. Voltage sensor conformations in the open and + closed states in ROSETTA structural models of K channels. Proc. Natl. Acad. Sci. USA 103:7292–97 Yellen G. 1998. The moving parts of voltage-gated ion channels. Q. Rev. Biophys. 31:239–95 by CAPES on 08/30/07. For personal use only. Yu FH, Catterall WA. 2003. Overview of the voltage-gated sodium channel family. Genome Biol. 4:207(1–7) YuFH, Catterall WA. 2004. The VGL-chanome: a protein superfamily specialized for electrical signaling and ionic homeostasis. Sci. STKE 2004: re15 Yusaf SP, Wray D, Sivaprasadarao A. 1996. Measurement of the movement of the S4 segment during the activation of a voltage-gated potassium channel. Pflügers Arch. 433:91–97 Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org Zagotta WN, Hoshi T, Aldrich RW. 1990. Restoration of inactivation in mutants of Shaker potassium channels by a peptide derived from ShB. Science 250:568–71 Zagotta WN, Hoshi T, Aldrich RW. 1994a. Shaker potassium channel gating. III: Evaluation of kinetic models for activation. J. Gen. Physiol. 103:321–62 Zagotta WN, Hoshi T, Dittman J, Aldrich RW. 1994b. Shaker potassium channel gating. II: Transitions in the activation pathway. J. Gen. Physiol. 103:279–319 Zhao Y, Scheuer T, Catterall WA. 2004. Reversed voltage-dependent gating of a bacterial sodium channel with proline substitutions in the S6 transmembrane segment. Proc. Natl. Acad. Sci. USA 101:17873–78 Zhou M, Morais-Cabral JH, Mann S, MacKinnon R. 2001. Potassium channel receptor site for the inactivation gate and quaternary amine inhibitors. Nature 411:657–61

52 Tombola · Pathak · Isacoff Contents ARI 12 September 2006 8:3

Annual Review of Cell and Developmental Contents Biology Volume 22, 2006

From Nuclear Transfer to Nuclear Reprogramming: The Reversal of Cell Differentiation J.B. Gurdon ppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppp1 How Does Voltage Open an Ion Channel? Francesco Tombola, Medha M. Pathak, and Ehud Y. Isacoff ppppppppppppppppppppppppppppppp 23 Cellulose Synthesis in Higher Plants Chris Somerville pppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppp 53 Mitochondrial Fusion and Fission in Mammals David C. Chan ppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppp 79 Agrobacterium tumefaciens and Plant Cell Interactions and Activities Required for Interkingdom Macromolecular Transfer Colleen A. McCullen and Andrew N. Binns ppppppppppppppppppppppppppppppppppppppppppppppp101 Cholesterol Sensing, Trafficking, and Esterification Ta-Yuan Chang, Catherine C.Y. Chang, Nobutaka Ohgami, and Yoshio Yamauchi ppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppp129 by CAPES on 08/30/07. For personal use only. Modification of Proteins by Ubiquitin and Ubiquitin-Like Proteins Oliver Kerscher, Rachael Felberbaum, and Mark Hochstrasser ppppppppppppppppppppppppppp159 Endocytosis, Endosome Trafficking, and the Regulation of Drosophila Development

Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org Janice A. Fischer, Suk Ho Eun, and Benjamin T. Doolan pppppppppppppppppppppppppppppppp181 Tight Junctions and Cell Polarity Kunyoo Shin, Vanessa C. Fogg, and Ben Margolis pppppppppppppppppppppppppppppppppppppppp207 In Vivo Migration: A Germ Cell Perspective Prabhat S. Kunwar, Daria E. Siekhaus, and Ruth Lehmann pppppppppppppppppppppppppppp237 Neural Crest Stem and Progenitor Cells Jennifer F. Crane and Paul A. Trainor pppppppppppppppppppppppppppppppppppppppppppppppppppp267

vii Contents ARI 12 September 2006 8:3

Of Extracellular Matrix, Scaffolds, and Signaling: Tissue Architecture Regulates Development, Homeostasis, and Cancer Celeste M. Nelson and Mina J. Bissell ppppppppppppppppppppppppppppppppppppppppppppppppppppp287 Intrinsic Regulators of Pancreatic β-Cell Proliferation Jeremy J. Heit, Satyajit K. Karnik, and Seung K. Kim pppppppppppppppppppppppppppppppppp311 Epidermal Stem Cells of the Skin Cédric Blanpain and Elaine Fuchs pppppppppppppppppppppppppppppppppppppppppppppppppppppppppp339 The Molecular Diversity of Glycosaminoglycans Shapes Animal Development Hannes E. Bülow and Oliver Hobert ppppppppppppppppppppppppppppppppppppppppppppppppppppppp375 Recognition and Signaling by Toll-Like Receptors A. Phillip West, Anna Alicia Koblansky, and Sankar Ghosh pppppppppppppppppppppppppppppp409 The Formation of TGN-to-Plasma-Membrane Transport Carriers Frédéric Bard and Vivek Malhotra pppppppppppppppppppppppppppppppppppppppppppppppppppppppp439 Iron-Sulfur Protein Biogenesis in Eukaryotes: Components and Mechanisms Roland Lill and Ulrich Mühlenhoff pppppppppppppppppppppppppppppppppppppppppppppppppppppppp457 Intracellular Signaling by the Unfolded Protein Response Sebastián Bernales, Feroz R. Papa, and Peter Walter ppppppppppppppppppppppppppppppppppppp487 The Cellular Basis of Kidney Development Gregory R. Dressler ppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppp509 Telomeres: Cancer to Human Aging Sheila A. Stewart and Robert A. Weinberg pppppppppppppppppppppppppppppppppppppppppppppppp531

by CAPES on 08/30/07. For personal use only. The Interferon-Inducible GTPases Sascha Martens and Jonathan Howard pppppppppppppppppppppppppppppppppppppppppppppppppppp559 What Mouse Mutants Teach Us About Extracellular Matrix Function A. Aszódi, Kyle R. Legate, I. Nakchbandi, and R. Fässler pppppppppppppppppppppppppppppppp591 Caspase-Dependent Cell Death in Drosophila Annu. Rev. Cell Dev. Biol. 2006.22:23-52. Downloaded from arjournals.annualreviews.org Bruce A. Hay and Ming Guo ppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppp623 Regulation of Commissural Axon Pathfinding by Slit and its Robo Receptors Barry J. Dickson and Giorgio F. Gilestro pppppppppppppppppppppppppppppppppppppppppppppppppp651 Blood Cells and Blood Cell Development in the Animal Kingdom Volker Hartenstein ppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppp677 Axonal Wiring in the Mouse Olfactory System Peter Mombaerts pppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppp713

viii Contents