Biophysical Journal Volume 107 September 2014 1035–1037 1035

New and Notable

Let’s Twist (the S4) Again

Teresa Giraldez*Department of Physiology, School ofMedicine & Centre for BiomedicalResearch of the Canary Islands(CIBICAN), University de La Laguna,La Laguna, Santa Cruz de Tenerife, Spain

When Noda et al. (1) cloned thevoltage-gated sodium channel, theypredicted that a stretch containingcharged residues might adopt a 310helix conformation, which was con-sidered rare in proteins. This structureis more tightly wound, longer, andthinner than an a-helix with the samenumber of residues (2). The chargedregion Noda et al. (1) referred to is infact the S4 transmembrane segment,whose movement across the electricalfield underlies function of the voltage-sensing domains (VSD) to control thegating of the central pore in voltage-gated ion channels (3). Understandingthe structural basis for S4 movementhas been the focus of great debate,and several models have been put for-ward with different degrees of trans-lation and/or rotation. It has beenproposed that the positive S4 gatingcharges are paired with negativelycharged amino acids at neighboringtransmembrane segments and that, asS4 moves, the ion pairs are sequen-tially exchanged (3).

Some excitement came into the fieldwhen, as envisioned by Noda et al. (1),310 conformations were observed inthe available ion channel crystal struc-tures. The characteristics of 310-helicesin this structural context may bringsome harmony into the S4 controversy,because they would make the S4tighter-wound, aligning gating chargesand favoring their transfer (2) (Fig. 1).Thus, a new model was proposed

http://dx.doi.org/10.1016/j.bpj.2014.07.043

Submitted June 9, 2014, and accepted for

publication July 22, 2014.

*Correspondence: giraldez@ull.edu.es

Editor: Michael Pusch.

� 2014 by the Biophysical Society

0006-3495/14/09/1035/3 $2.00

where the S4 would adopt partial andtransient 310-helix conformations (2).In this issue of the Biophysical Jour-nal, Kubota et al. (4) elegantly testthe proposed dynamic switch betweenS4 a and 310 conformations of theVSD by using fluorescent spectro-scopic approaches combined withelectrophysiology. The study is donein the context of the Ciona intestinalisvoltage-sensing phosphatase (Ci-VSP),where the VSD regulates enzymaticfunction, but is thought to retainvoltage-sensing mechanisms similarto those in voltage-gated ion channels(5). In this protein, the VSD ismonomeric (6), which simplifies theinterpretation of the spectroscopic mea-surements.

The relevance of the experimentsperformed by Kubota et al. (4) residesin the fact that, because no crystalstructures have been obtained of theresting and intermediate states thatwould reveal the VSD conformationsduring gating, the transitional 310 hy-pothesis is based mainly on theoreticalmodels. Until now, no experimental ev-idence has been obtained to physicallyvalidate the a-310 dynamic transition.By using lanthanide-based resonanceenergy transfer (LRET) and fluores-cence resonance energy transfer(FRET), the authors address differentquestions about the role of a-to-310transitions in the S4 movement. Doesthe interconversion of the helix occurin the whole S4, leading to simulta-neous alignment of all sensing chargesor, alternatively, does it occur withstepwise a to 310 transitions restrictedto small regions of S4 (2,4,7,8)? Whatis the duration of such transitions?LRET allows Kubota et al. (4) toestimate absolute changes in distancebetween a donor (Tb3þ) and anacceptor (mCherry) that tightly flankCi-VSP S4 in order to calculate theextension of the a-(or 310-)helix andtest for long-lived transitions; FRETis used to further explore the possibilitythat 310 conformations are solely asso-ciated to S4 short-lived transitionalstates.

The first observation by Kubotaet al. (4) is that Ci-VSP S4 adopts ana-helical conformation that does notchange with voltage (Fig. 1), showingthat a complete a-310 switch to simul-taneously align all gating changes isnot essential for gating to occur. Doesthis mean that no a-310 transitions areassociated to Ci-VSP S4 movement?Before reaching this conclusion, thereare two considerations to be takeninto account:

First, it must be noted that S4 move-ment has been modeled including threestates: resting, activated, and relaxed.In this model, transition from a restingor an activated state to the relaxed stateinvolves an a-310 conversion (9). In thecase of Ci-VSP, where resting and acti-vated states seem to adopt mainly a-helix conformations, an a-310 switchmay occur at the relaxed state. Howev-er, this question remains unansweredbecause, as well noted by the authors,the voltage sensor in their fluorescentconstructs seems not to reach therelaxed state (4).

Second, LRET would only detectlong-lived interconversions betweenhelices. Molecular dynamics studies inpeptide helices have shown that a-to-310 transitions take place very rapidly,on the nanosecond timescale (2). Usingtime-resolved FRET (Fig. 1), Kubotaet al. (4) show that reversible submilli-second a-to-310 transitions seem notto occur. Somewhat surprisingly, theirresults indicate the existence of S4 tiltor rotation after the gating chargeshave been transferred. Although thisconclusion needs further experimentsto be confirmed, it is supported by therelatively high anisotropy values ofboth donor and acceptor. Altogether,Kubota et al. (4) conclude that Ci-VSDS4 does not undergo complete a-to-310interconversions during gating. How-ever, their results are compatible witha model in which the a-310 switch isrestricted to a fraction of the transmem-brane helix (2,4,7,8).

FIGURE 1 Constructs used by Kubota et al. (4) for LRET and FRET experiments. Schematic repre-

sentation of Ci-VSP S4 helix (green) flanked by the lanthanide-binding tag that binds Tb3þ as LRET

donor (yellow circle), and the red fluorescent protein mCherry as LRET acceptor (red). S4 segment

length can be deduced from the total distance between fluorophores, after subtracting the size of lantha-

nide-binding tag (12.1 A), and the total distance from the end of Ci-VSP S4 to the mCherry chromo-

phore (27.7 A) (4). (Blue) Charged residues in Ci-VSD S4. The hypothetical a to 310 transition of

the whole S4 that is tested in this study by Kubota et al. (4) is represented. In response to voltage,

the a-helix (left) would stretch 9.5 A, from 28.5 to 38 A, adopting a 310 conformation (right). The

310 conformation shows a better alignment of charged residues, because the 310 helix has three amino

acids per turn (as opposed to 3.6 in the a-helix). In FRET experiments, the donor (ATTO) is positioned

at residue 222 into the S4 helix (red circle). The acceptor (mCherry) is the same as in LRET ex-

periments. In this scenario, a transitional a-310 switch should lead to a decrease in real-time FRETmea-

surements; however, contrary to this hypothesis, an increase in FRET is observed. See Kubota et al. (4)

for details. To see this figure in color, go online.

1036 Giraldez

The observation that Ci-VSD doesnot undergo large a-310 conversionsduring gating may be a unique featureof Ci-VSP. In fact, multiple scenariosmight be expected, because 310conformations observed in availableion-channel crystal structures covervariable extents. Thus, the prokaryoticNavAb sodium channel shows almostthe entire S4 in 310 conformation,whereas a hybrid a/310 conformationis observed in the Kv1.2/2.1 chimerastructure (2). Similarly to the VSD ofCi-VSP, the Kv7.1 channel S4 showsa pure a-helix conformation (10). Inan attempt to characterize the possibleconformational switch of a voltage-gated ion channel VSD, the authorspush the technique further to test suchconformational change for the NavAb

Biophysical Journal 107(5) 1035–1037

channel S4 (11), which is inserted inthe structural context of Ci-VSP intwo different versions, according totwo possible alignments of chargedresidues. Whereas both constructs pro-duce gating currents, LRET measure-ments with NavAb S4 are consistentwith a pure a-helix without long-livedtransitions to a 310-helix.

It could be concluded that, in anative conformation, the VSD ofNavAb does not adopt a 310 conforma-tion. Nevertheless, although it is incor-porated into a functional protein in thecontext of the plasma membrane,NavAb does not necessarily adopt itsnative conformation in the experimentspresented by Kubota et al. (4). The au-thors are conscious of this limitation,because the isolated NavAb-S4 is

transplanted into Ci-VSP and thusmight lack the specific stereochemicalenvironment required for the NavAb-S4 to adopt its 310 structure. Anotherrelevant difference might be the factthat Ci-VSP is monomeric, whereasNavAb channels are tetramers. How-ever, regardless of the helical confor-mation that NavAb S4 adopts in theCi-VSP protein context, gating cur-rents are observed, suggesting againthat an a-310 transition of the wholeS4 is not a requirement for gating tooccur. Interestingly, 310-helices arealso found in a ligand-gated channelin which the S4 lacks the crucialcharges and is thought not to functionas a voltage sensor (2).

The contribution of Kubota et al. (4)constitutes the first attempt to physi-cally measure dynamic S4 a-to-310transitions from functional VSDs inthe membrane. Their main finding isthat an a-310 switch of the whole S4is not mandatory for gating to occur.Nonetheless, many exciting questionsremain about the role of 310 helices inthe S4 structural movement. In fact, asingle model may not apply to allvoltage-gated ion channels, as hintedby the varying a/310 conformationsobserved in the crystal structures.Kubota et al. (4) give the existingmodels a new twist, setting the stagefor further experiments to explore therole of dynamic S4 a-310 transitionsin voltage-gated ion channels function.

REFERENCES

1. Noda, M., S. Shimizu, ., S. Numa. 1984.Primary structure of Electrophorus electri-cus sodium channel deduced from cDNAsequence. Nature. 312:121–127.

2. Vieira-Pires, R. S., and J. H. Morais-Cabral.2010. 310 helices in channels and othermembrane proteins. J. Gen. Physiol. 136:585–592.

3. Catterall, W. A. 2010. Ion channel voltagesensors: structure, function, and pathophysi-ology. Neuron. 67:915–928.

4. Kubota, T., J. J. Lacroix, ., A. M. Correa.2014. Probing a-310 transitions in avoltage sensing S4 helix. Biophys. J. 107:1117–1128.

5. Murata, Y., H. Iwasaki, ., Y. Okamura.2005. Phosphoinositide phosphatase activity

New & Notable 1037

coupled to an intrinsic voltage sensor.Nature. 435:1239–1243.

6. Kohout, S. C., M. H. Ulbrich, ., E. Y.Isacoff. 2008. Subunit organization andfunctional transitions in Ci-VSP. Nat. Struct.Mol. Biol. 15:106–108.

7. Schwaiger, C. S., P. Bjelkmar, ., E.Lindahl. 2011. 310-Helix conformation facil-itates the transition of a voltage sensor S4

segment toward the down state. Biophys. J.100:1446–1454.

8. Long, S. B., X. Tao, ., R. MacKinnon.2007. Atomic structure of a voltage-depen-dent Kþ channel in a lipid membrane-likeenvironment. Nature. 450:376–382.

9. Villalba-Galea, C. A., W. Sandtner, ., F.Bezanilla. 2008. S4-based voltage sensorshave three major conformations. Proc.Natl. Acad. Sci. USA. 105:17600–17607.

10. Peng, D., J. H. Kim,., C. R. Sanders. 2014.Purification and structural study of thevoltage-sensor domain of the humanKCNQ1 potassium ion channel. Biochem-istry. 53:2032–2042.

11. Payandeh, J., T. M. Gamal El-Din,., W. A.Catterall. 2012. Crystal structure of avoltage-gated sodium channel in twopotentially inactivated states. Nature. 486:135–139.

Biophysical Journal 107(5) 1035–1037