PERSPECTIVE

Feeling the hidden mechanical forces in lipidbilayer is an original senseAndriy Anishkina, Stephen H. Loukinb, Jinfeng Tengb, and Ching Kungb,c,1

aDepartment of Biochemistry and Center for Computational Proteomics at the Huck Institute of Life Sciences, Pennsylvania State University,University Park, PA 16802; and bLaboratory of Molecular Biology and cDepartment of Genetics, University of Wisconsin–Madison, Madison,WI 53706

Edited by Roderick MacKinnon, The Rockefeller University and Howard Hughes Medical Institute, New York, NY, and approved April 15, 2014 (received for review January21, 2014)

Life’s origin entails enclosing a compartment to hoard material, energy, and information.The envelope necessarily comprises amphipaths, suchas prebiotic fatty acids, to partition the two aqueous domains. The self-assembled lipid bilayer comes with a set of properties including itsstrong anisotropic internal forces that are chemically or physically malleable. Added bilayer stretch can alter force vectors on embeddedproteins to effect conformational change. The force-from-lipid principle was demonstrated 25 y ago when stretches opened purified Escherichiacoli MscL channels reconstituted into artificial bilayers. This reductionistic exercise has rigorously been recapitulated recently with twovertebrate mechanosensitive K+ channels (TREK1 and TRAAK). Membrane stretches have also been known to activate various voltage-, ligand-,or Ca2+-gated channels. Careful analyses showed that Kv, the canonical voltage-gated channel, is in fact exquisitely sensitive even to very smalltension. In an unexpected context, the canonical transient-receptor-potential channels in the Drosophila eye, long presumed to open by ligandbinding, is apparently opened by membrane force due to PIP2 hydrolysis-induced changes in bilayer strain. Being the intimate medium, lipidsgovern membrane proteins by physics as well as chemistry.This principle should not be a surprise because it parallels water’s paramount role in thestructure and function of soluble proteins.Today, overt or covert mechanical forces govern cell biological processes and produce sensations. At thegenesis, a bilayer’s response to osmotic force is likely among the first senses to deal with the capricious primordial sea.

mechanosensitivity | force sensing | channel gating | bilayer mechanics | touch

To compete successfully, organisms mustevolve effective reactions to earth’s physicalstimuli: heat, force, voltage, and chemicalligands. Emil Fischer’s lock-and-key bindingbetween molecules is well known andexplains such overt senses as vision, smell,and most tastes. We also understand voltagesensing. For example, a charge-bearing helixtraversing the electric field across the lipidbilayer can mechanically drive conforma-tional changes of embedded ion channelsor enzymes (1). Heat and force are universal,governing all matters and reactions. A chal-lenge to current biology is to understandhow Nature chooses and uses her molecularthermometers and force gauges. We reviewthe basis of force sensing, emphasizing theroles of the lipid bilayer mechanics.

Biological Mechanisms Are UltimatelyMechanicsTextbooks describe the energy minimumwhen atoms are brought to an equilibriumdistance (bond length), at which, there is zeroforce between them because their mutualattraction balances repulsion. Thermal oscil-lations or external forces change the bondlength and generate compressive or expan-sive forces in the bond, much like witha spring. External force acting on an atomcan simply move the whole bonded assem-bly. For example, pulling one end of an

α-helix can tilt it. A force can also strain thebond and displace the partners, like stretch-ing or bending the helix. The bond breakswhen work (force times displacement)exceeds the bond energy. For example, ittakes ∼4 pN (4 × 10−12 N) to break a hy-drogen bond, ∼1,600 pN to break a C–Ccovalent bond (Fig. S1). Ligand–protein orprotein–protein binding also mechanicallystrains the bonds between protein do-mains to effect long-range conformationalchanges. The last 100 y of mechanochem-istry have shown how mechanical forcesgovern even inorganic chemical reactions(2). Many steps in biochemical reactions orother molecular processes can therefore besimulated by explicitly stating the New-tonian forces involved and following themotions of the components—an approachbringing invaluable insights into molecularunderpinnings of life. The 2013 ChemistryNobel was awarded for the developmentof molecular dynamics and quantum me-chanical approaches to simulate complexchemical systems. How the paired atomsmove upon a perturbation can be describedby a set of laws and numerical parame-ters called the “force field” (3). In macro-molecules, long-distance Lennard-Jonesand electrostatic forces come into playduring conformational change. Reposition-ing atoms by iteratively calculating the

multiple short- and long-distance forcesamong atoms and feeding them into New-tonian laws of motion is the basis of mo-lecular dynamics simulation (MDS), whichin recent decades has successfully analyzedand explained the many workings of mac-romolecules and their assembly (4, 5). Someforces relevant to biology and used in MDSare listed in Fig. S1. Some findings mightsurprise us. For example, most of the timein catalysis is spent on the diffusion to andfrom the catalytic site, mechanically posi-tioning the substrates and products; whereasthe very stage of chemical transformation, ofwhich we tend to fixate, happens almostinstantaneously.

The Mechanics of the Lipid BilayerA surface tension appears at any liquid–liquidor liquid–gas interface. Recall that at ahexane–water interface, the surface tensionappears because water molecules there aredeprived of half of their hydrogen-bondingpartners they enjoy in the bulk. Thermal

Author contributions: A.A., S.H.L., J.T., and C.K. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

1To whom correspondence should be addressed. E-mail: ckung@wisc.edu.

This article contains supporting information online at www.pnas.org/

lookup/suppl/doi:10.1073/pnas.1313364111/-/DCSupplemental.

7898–7905 | PNAS | June 3, 2014 | vol. 111 | no. 22 www.pnas.org/cgi/doi/10.1073/pnas.1313364111

Dow

nloa

ded

by g

uest

on

July

27,

202

0

dispersion of the surface also makes mole-cules less densely packed in the few inter-facial layers, thus decreasing the repulsionbetween the neighbors (6). Qualitativelysimilar interfacial forces develop in thelipid bilayer. However, there are criticaldistinctions. Molecules at a liquid interfaceare not restrained from freely joining thebulk. Water in a sea of hexane will shrinkinto a sphere. Adding more water increasesthe sphere’s volume but reduces the relativenumber of water at the surface, with nochange in surface tension. In contrast, thepolar lipid head is stuck with its nonpolartail making the bilayer a “surface withoutbulk.” Typical membrane lipids tend tocompact into a bilayer sheet instead ofa droplet. Adding lipids enlarges the sheetbut does not change the exposed-to-hiddensurface ratio per molecule. Water moleculesstrongly bond with the polar atoms of thehead groups or the glycerol, although bar-red from the nonpolar hydrocarbon tails.Thus, a sharp lateral tension develops at thelevel of the lipid neck. As a self-assembledentity at equilibrium, this tension is bal-anced by the large repulsion from interiortail movement and the smaller head–headrepulsion in a relaxed membrane. The rel-ative contribution of the two repulsionsdepends on the head-to-tail width ratio,which corresponds to the spontaneous cur-vature of the lipid leaflet, with a larger headfavoring positive curvature. This layeredinternal force distribution has been calcu-lated from thermodynamics (7, 8) and ex-amined by MDS (9, 10) (Fig. 1B, Upper Left).Thus, unlike soluble proteins, which arebombarded by particles from all directions(Fig. 1A, Left), embedded proteins as a partof the membrane ensemble are subjected toanisotropic push and pull (Fig. 1A, Right),especially sharply focused suction. A stretchwill deform the water sphere in hexaneimagined above, changing the surface-to-volume ratio but not the surface tension.However, a stretch force can only thin thebilayer (until it breaks) and thereby increasethe lateral tension. This general descriptionshould not obscure local variations. Thecontour and hydrophobicity of the proteinsurface need to match that of the lipids.The bilayer structure can therefore bedistorted, e.g., being curved, compressed, orotherwise deformed at the protein-lipid in-terface. The distortion spreads over a fewlayers of annular lipids and even enables amechanical cross-talk between the proteinsor protein domains through the lipid me-dium (11). Added external stretch can thinthe bilayer generating a hydrophobic mis-match at the interface, driving the protein

to take a better-matched conformation.Reducing the acyl chain length from PC20 toPC18 converts the reconstituted gramicidinA from a stretch-activated to a stretch-inactivated channel (12). Besides thinning, astretch can also change the magnitude anddirection of the force vectors on the protein.Chemically changing bilayer composition byinserting impurities or through lipid me-tabolism can have the same effect (Fig. 1B).Although these principles are general, wechose to discuss ion channels among mem-brane proteins because their activities canbe examined easily and quantitatively, evenat the level of single proteins and downto a microsecond time resolution (13) byelectric means.

The Force-from-Lipid PrincipleThe force-from-lipid principle was estab-lished some 25 y ago from the study of themechanosensitive channels from Escherichiacoli (14, 15). Their discovery, atomic struc-tures, and biological roles have been peri-odically reviewed (see refs. 16 and 17 andcitations therein). Briefly, the key observa-tion is that purified MscL or MscS can bereconstituted into artificial bilayers andretain their mechanosensitivity, excludingthe need of cytoskeleton, external tether, oraccessory channel subunits as force trans-mitters (18, 19). Because bilayers of phos-pholipids of different head groups (20) orfatty-acid lengths (21) support the mecha-nosensitivity of MscL, a lock-and-key li-gand-gated mechanism need not apply.Judging by the slope of the Boltzmanndistribution of the open probability againstbilayer tension, MscL protein expands by∼20 nm2. This expansion in area (ΔA)requires 220 pN to stretch its 22-nm pe-rimeter, doing 2 × 10−23 J (50 kT) work.Half of the channels are open at 10–12pN/nm, not far from the lytic tension oflipid bilayer (22, 23). The ΔA estimated fromactivities approximates the structural esti-mate from crystallography (24) and elec-tron paramagnetic resonance spectroscopy(25). MscL is a homopentamer of subunitswith two transmembrane helices, M1 andM2. To open, these helices apparently leandown and move outward making a muchthinner structure, as if to meet a thinnedbilayer (25). Bilayers with shorter chainlength favor opening, but tension is stillrequired. However, an asymmetric addi-tion of lysolipids to either one of themonolayers opens the channel in a relaxedmembrane (21), suggesting involvement ofadditional physical parameters, such asa mismatch of tension in the two leafletsand/or bending of annular lipids favored by

positive curvature. Asymmetric insertionsof exogenous amphipaths activate thesechannels (14), consistent with the bilayer-coupling hypothesis (26). The lipid forceprofile and how stretch relates to MscLgating has been examined by MDS (10). Inline with the role of the bilayer’s focusedlateral tension (Fig. 1B), random and scan-ning mutagenesis showed that replacing hy-drophobic with hydrophilic residues in M1or M2 at the outer rim of MscL removesmechanosensitivity in vitro and in vivo (27).The miniprotein antibiotic gramicidin A

from Bacillus brevis readily forms cationchannels in bilayers by pair-wise stacking. Ithas been used to examine the role of vari-ous chemical and physical properties ofthe bilayer in quantitative details (28). Overthe years, the force-from-lipid principle hasgradually been used to understand eukary-otic processes as well. More recently, thisprinciple has been rigorously extended toanimal ion channels as summarized below.

Mechanosensitive Two-Pore-Domain K+

ChannelsMicrobes have more varieties of K+ channelsthan animals (29). The simplest are homo-tetramers of S1-P-S2 subunits (“S” for “seg-ment” in “α-helical segment”; “P” for “pore”in “pore domain”, the filter structure), such asKcsA from the Gram-positive Streptomyces,the first ion channel of known crystal struc-ture (30). A variant of this motif are two-pore-domain K+ channel (K2p) dimersof S1-P1-S2-S3-P2-S4 subunits. Human K2p

family comprises 15 fairly disparate mem-bers, several of which can be stretched openin clamped patches. The structures of twoK2p (31–33) are now known. The mechano-sensitive TRAAK (TWIK-related arachidonicacid-stimulated K+) (34) has been solved inopen states at 3.8 Å (32) (Fig. 2 A and B) andthen at 2.75 Å resolution revealing a domainswap (33). TRAAK has features not foundin KcsA, two of which may be related tomechanosensitivity: Of the four inner helices,the two S2 are long, each having a kink thatmakes the lower portion lying almost flat.This portion is amphipathic with hydro-philic basic residues facing the membrane–cytoplasm interface (Fig. 2A, purple) andhydrophobic residues facing the membraneinterior (green), placing the two peptides atthe membrane–cytoplasm interface of theinner leaflet, rich in acidic lipids in vivo.Interestingly, the inner transmembrane halfof K2p channels (31) is even fenestrated. InTRAAK (32, 33), a sizable portion of chan-nel’s lower half is not enclosed by protein,leaving a 5-Å-wide gap presumably filledwith lipids in vivo, extending from the

Anishkin et al. PNAS | June 3, 2014 | vol. 111 | no. 22 | 7899

PERS

PECT

IVE

Dow

nloa

ded

by g

uest

on

July

27,

202

0

bottom of the filter to the lower edge of thechannel (Fig. 2B, yellow).K2p control the resting membrane poten-

tial by passing leak current (35). They areregulated by disparate stimuli: pH, tempera-ture, lipidic ligands, and membrane stretch.Among the diverse K2p, the mechanosensitiveTREK-1 (TWIK-related K+-1) and TRAAKhave been extensively studied by Patel,Honore, and coworkers (36, 37). Whole-cell TREK-1 current is strongly suppressedby hyperosmolarity. In patches from ex-pressing cells, pressure in either directionactivates TREK-1, although suction is moreeffective. It is also activated by negativelycharged phospholipids such as PIP2, by

lysophospholipids, polyunsaturated fattyacids (PUFAs), volatile general anesthetics(36, 37) and morphine (38).

Reconstitution of VertebrateMechanosensitive K+ Channels, K2p

Although there are antecedents (39, 40), re-cent works generate new excitement becauseof the deep knowledge on the structure andfunction of K2p summarized above. Berrieret al. (41) have recently reconstitutedan enriched fraction of mouse TREK-1into liposomes and examined patches ex-cised from them. They observed the ∼80-pSK+-specific outward-rectifying chlorproma-zine-sensitive conductance. Surprisingly,

these channels displayed spontaneous ac-tivities that could not be further increased bypipet suctions. Positive pipet pressure pulses,however, could proportionally reduce theiractivity to total closure. These authors rea-soned that open probability has been max-imized by inherent tension in the patches(41). Such a tension is likely generatedduring the gigaOhm-seal formation, bond-ing the bilayer lipids to the pipet glass sur-face (42–45).In a recent rigorous study, Brohawn et al.

(46) purified zebrafish TREK-1 and humanTRAAK to homogeneity, as evidenced bymonodispersion of column-elution profiles(Fig. 2C), and also reconstituted them intoliposomes made of mixed-length phophati-dylcholine extracted from soybean azolectin.In excised patches with multiple channels,they found the channels to respond in-stantaneously to applied force (Fig. 2D).These channels are also activated by arach-idonic acid (the PUFA embedded in thename TRAAK). These K2p respond propor-tionally to either negative or positive pres-sures much like the bacterial MscL in asimilar setting (21). This is also true whenTRAAKs are expressed in cultured cells,where they should be in the same orientation(Fig. 2 E and F). This indifference to pressuredirection leaves no doubt that it is the addedbilayer tension that gates the channels. Be-yond demonstrating the force-from-lipidbiophysical principle for animal channels,these authors also expressed TRAAK in cul-tured mouse Neuro2A cells. Prodding themwith a probe evokes a K2p-dependent out-ward current that counteracts the native in-ward current through Piezo1. Details aside,both sets of reconstitution experimentsclearly showed that vertebrate K2p, much likeMscL and MscS, are themselves sensitive tostretch force from the lipid bilayer, requiringno additional subunits, cytoskeleton, or ex-tracellular matrix tethers.Precisely how lipid forces operate TREK-1

or TRAAK is not yet clear. Like many K+

channels, including the budding-yeast K2p

(47), TREK-1 apparently uses a dynamic fil-ter, where several gain-of-function mutationsare found (48). Its quaternary ammonium(QA)-binding site seems accessible from theinside before channel activation, suggestinga static open inner gate (49), although QAscould reach the site through the fenestrationnow shown in the atomic structure (32, 33)(Fig. 2B). TREK-1 is activated by negativelycharged phospholipids such as PIP2, andinhibited by polycations or agents that causePIP2 hydrolysis (36, 37). These observationscan be explained by the interaction of thenegatively charged inner leaflet with a

A

B

Fig. 1. Anisotropic forces of the lipid bilayer and their changes that can reshape embedded proteins. (A) Adiagram comparing a protein in the cytoplasm, near-isotropically compressed by the bombardment of neigh-boring particles (Left) and one embedded in a bilayer, subjected to large lateral tension at the two polar–nonpolar interfaces opposed by compression forces at the other levels of the bilayer (Right). (B) The lateraltension at the interface is balanced by the pressure in the head groups and hydrophobic tails (Upper Left) andthis equilibrium can match a certain protein conformation, say, the closed state of a channel protein (UpperRight). External stretch force (Lower, broad red arrows) and/or amphipaths with positive (red triangle) ornegative (green triangles) spontaneous curvature asymmetrically added to one leaflet can thin and bend thebilayer at the protein– lipid interface, changing the force vectors (small arrows) and therefore prefer a bettermatched protein conformation, say, the open state.

7900 | www.pnas.org/cgi/doi/10.1073/pnas.1313364111 Anishkin et al.

Dow

nloa

ded

by g

uest

on

July

27,

202

0

proximal peptide segment, rich in basicamino acids, which immediately follows theend of the inner helix S4. The bilayercrenators (trinitrophenol, lysolecithin) acti-vate, whereas cup formers (chlorpromazine,tetracaine) inhibit TREK-1 (36, 37), consis-tent with the bilayer-coupling hypothesis (14,26). TREK-1 is estimated to expand by ∼2 to4 nm2 in the membrane plane when opens(45). It remains reasonable that the innerhelices (S2 and/or S4) converge to close anddiverge to open like other K+ channels as wassupposed by earlier homology models (50),even if the inner gate cooperates with thefilter gate. The two S2 inner helices, havingamphipathicity (Fig. 2 A and B) that matchesthe inner leaflet naturally rich in acidicphospholipids and being at the level wherepeak lateral tension is expected, would pre-dict that they are subjected to that tensionand its changes during membrane stretch.

The Voltage-Dependent K+ Channel Is asSensitive to Bilayer Force as It Is toVoltageThe celebrated voltage-dependent K+ chan-nel, Kv, is a tetramer of S1-S2-S3-S4-S5-P-S6subunits in essence (1). Like the S1-P-S2 ofKcsA, the four S5-P-S6 helical pairs cometogether to form the pore domain housingthe filter and the gate. The four S1-S4 formthe four peripheral domains, each bearinga voltage sensor. Driven by the resting

membrane potential, the voltage sensors,acting through the S4-S5 linkers, mechanicallyconstrict the lower gate comprising the fourS6. Depolarization moves the sensors andlinkers, releasing the constriction (51). Theseactions likely also make use of the intimatelyattached lipids (52). Here the channel opensnot by removing a plug, but by a lateralexpansion of the cytoplasmic side of thechannel with a ΔA. If Kv is built only forinteractions between its protein domains,one might expect the domains to be denselypacked to minimize the contact with themembrane—a potential source of noise.However, the association between the poreand the peripheral domains is limited to theinner S4-S5 link and a contact between S1 andS6 toward the outer side (52, 53). As a result,all domains, including the gate-bearing coreare in extensive contact with lipids. Thisunusual design implies roles of lipids in Kvfunction. Indeed, the action of the voltagesensor with its positive arginines (the gatingcharges) engages the negative phosphateoxygens at the lipid neck (54, 55). However,why should the pore domain also be sur-rounded by lipids?The free-energy change associated with the

bilayer deformation is estimated to be com-parable to that of the voltage-dependent partof the total gating energy (56). Although notcommonly known, Kv (e.g., the Kv1.2 paddlechimera of known crystal structure) (52)

shows drastically different behavior in wholecells, in patches, or upon reconstitution intoplanar bilayers. MacKinnon and coworkersfound the differences best explained bychanges in tension on the membrane, heldlow when it is locally constrained by sub-cortical cytoskeleton, but increases when themembrane is detached and adheres to thesurface of the glass electrode (43, 44). Forexample, the tendency to open, as indicatedby the Boltzmann slope and midpoint volt-age, is much higher for channels in an on-cellpatch than those in the rest of the samecell after the patch is broken (Fig. 3 A–D).Quantitative analyses of kinetic schemes in-dicate that the tension-sensitive step is thefinal opening of the gate after the four voltagesensors have moved, i.e., at the point whenΔA takes place. Thus, tension mainly acts oncoupling the sensors to the gate, presumablythrough lipids, and not on the sensors them-selves. The tension in the patch generatedby lipid adhesion to glass is only about 0.5–4 mN/m (42). Free-energy changes, as es-timated from the changes of the rate ofopening step with or without added ten-sion, indicate a ΔA of some 3–4 nm2,consistent with structural information (44).Because this tension-dependent behavior isobserved in excised patches and varies withlipid compositions upon Kv reconstitution(43), the tension must come from the lip-ids. The effect of this small tension is nottrivial. “Kv channel is as much a mechano-sensitive channel as it is a voltage-dependentchannel” (44, p. 10356); open probabilityrises from 0 to 0.65 between −60 and +50mV, but rises from 0 to 1.0 upon suction atmidrange voltages. Kv, being a model forNav, Cav, and TRPs, indicates that the lipid-tension effect is much more general than wecommonly thought. The generality of channelmechanosensitivity has been advocated byMorris and coworkers (ref. 57 and refer-ences therein).

Physics or Physiology?Any entity that deforms in the directionof force is, by physical definition, mechano-sensistive. So does the mechanosensitivity ofKv simply reflect its necessary expansion inthe membrane plane under stress, or doesit have any biological significance? As de-scribed above, heterologous TRAAK can passoutward current that counteracts in piezo1inward current in cells responding to externalforce (46). Given evolution’s opportunism, itseems unlikely that this exquisite sensitivityis not exploited. Hypoosmotic swelling opensKv in cultured cells (44), presumably bystretching small membrane domains confinedby cytoskeletons (Fig. 3D). Thus, osmotic

C

D

out

in

090

A B

E F

negativepressure

positivepressure

Cur

rent

(pA

)

2000

1000

00 20 40 60

Absolute value of applied pressure (mmHg)

500

1500

Fig. 2. The atomic structures and the reconstituted activities of TRAAK. (A and B) TRAAK structures from twoperspectives showing the two S2 amphipathic peptides [hydrophobic (green) and charged (purple) residues] and oneof the two prominent fenestrations (yellow) (from ref. 32). (C ) Monodispersion of column eluate indicates the purity ofthe TRAAK protein produced. (D) Such protein is reconstituted into liposomes. TRAAK channels in a patch excised fromsuch a proteoliposome are activated (upper traces) in proportion to the suction applied to patch (bottom traces). (E )TRAAKs in a patch excised from an expressing CHO cell, in which they should have the same orientation, respond(upper traces) to either suction (black) or pressure (red). (F ) Activation by pressures in either direction are quantitativelysimilar, indicating that it is the membrane stretch that activates (modified from ref. 46).

Anishkin et al. PNAS | June 3, 2014 | vol. 111 | no. 22 | 7901

PERS

PECT

IVE

Dow

nloa

ded

by g

uest

on

July

27,

202

0

changes, muscle contractions, or lipid-com-position changes (below) may activate ormodulate Kv, Cav, etc., at least in principle.Mechanosensitivity is most “palpable” inthe sense of touch. Interestingly, a recentreport shows that, when certain dorsal-root-ganglion neurons are prodded, channels withKv1.1 subunits pass K+ outward, acting asa brake (58) to oppose the depolarization bythe inward current through other touch-sensitive channels yet to be defined. It isreasoned that these Kv1.1 subunit-containingchannels therefore set the threshold of thetouch-induced action potential or tune thefiring adaptation. Mice expressing a domi-nant negative allele of Kv1.1 suffer severemechanical allodynia, interpreting touchas hurt (58) (Fig. 3E).Evolution also exploits Kv’s mechano-

sensitivity negatively. The Chilean rose ta-rantula paralyzes its preys with venom thatincludes the toxin VSTx1, which enters themembrane to bind the Kv2.1 sensor at 10−8

M specificity (59). Strangely, VSTx1 has noeffect on Kv in slack membranes, and causesthe channel in the tense bilayer to behave asin a slack membrane (43), as if the toxinperturbs the local lipid arrangement to blockvoltage sensor-gate coupling and/or the finalgate opening.

The Photomechanics of Fly VisionIn the 1970s, the Pak laboratory and others(60) began a systematic genetic dissection ofthe Drosophila visual transduction. Blind flieswere sorted by complementation groupingand by the wave of discharges fromthe compound eye during a light flash(electroretinography). Mutants display dif-ferent corruptions of the wild-type wave-forms. For examples: “norpA” for “noreceptor Potential, complementation groupA” later proves to encode phospholipase C;“innaE” for “inactivation, no afterpotential,group E” encodes a diacylglycerol lipase,etc., implicating the involvement of lipidmetabolism.“trp” stands for “transient receptor poten-

tial” (61), which was the first blind fly genecloned (62). trp and its homolog trpl (trp-like) were later shown to encode pore-form-ing channel subunits (63). Much researchimplicates the cascade of rhodopsin-Gq-phospholipase C (PLC, from norpA)hydrolyses PIP2 (phosphatidylinositol4,5-bisphosphate) into DAG (diacylglycerol),although how these lipids activate the chan-nels was unclear. Thinking being chemical inour era, one tends to assume that these lipidsbind a specific site to control the channels. Itis therefore a great surprise when Hardie andFranze (64) showed that there is a global

physical change in the membrane. Aston-ishingly, light induces a near-micrometershrinkage of the ommatidia (units in thecompound eye) that is directly visible undera light microscope and measured with anatomic force microscope (Fig. 4 C and D).This contraction begins before the channelcurrent. Deleting the PLC enzyme (norpA−/−)takes away the contraction; deleting thechannels does not. The authors reason thatbeheading PIP2 to make the smaller DAGincreases tension in the inner membraneleaflet to pull open channels (Fig. 4A). Totest, the well-established mechanosensitivechannel, gramicidin A (28) was added to thephotoreceptor cell and found to pass currentproportional to light intensity. Note that go-ing from PIP2 to DAG is not an isometricshrinkage. That PIP2 rod becoming the DAGcone (Fig. 4B) should increase bilayer asym-metry and curvature, and would changethe bilayer force profile (Fig. 1). Althoughwe do not know the exact changes in mem-brane thickness, curvature, or forces in thechannel’s vicinity, diverse cationic amphi-paths that are expected to enter and crowdthe inner leaflet (thus reducing lateral tensionthere) did inhibit light-induced current. Thecase of membrane forces controlling TRP/TRPL therefore seems convincing. Still, H+

(65) and PUFAs (products of innaE) (66)are also involved, so photomechanics maynot be the whole story.The Drosophila work coined the term

“TRP,” which is honored as the founder ofthe TRPC subfamily (below), with “C” being“canonical”. Canonical or not, is this a specialcase? Rods and cones do not work this way:The rhodopsin-Gt-phosphodiesterase cascadeleads to a cGMP drop that closes the cGMP-gated channel, shutting off dark current.However, the vertebrate retina also containsnon–image-forming but intrinsically photo-sensitive ganglion cells for circadian entrain-ment. Here the molecular pathway directlyparallels that of the insect, beginning withmelanopsin and ending in the activation ofTRPC6/7 (67). The conservation is striking,with melanopsin being far more similar tothe insect rhodopsin than vertebrate rho-dopsin (68). Similar pathways may be usedfor other TRP channels in melanocyte (69)and keratinocyte (70) of the skin.That the ∼5-Å PIP2-to-DAG head shrink

can become a near-micrometer contractionhas to do with the unusual anatomy of theinsect photoreceptor cell, which sums theminute shrinkage of the thousands of stackedmicrovilli it bears (Fig. 4E). Potentially, thereis a direct mechanical link between the ac-cumulation of DAG and the size of a stack ofmicrovilli. For example, increased content of

this lipid with negative spontaneous curva-ture in the inner monolayer would favorhigher curvature, thinner microvilli, andlateral-stack shrinkage. Elsewhere, isolatedmicrovilli, cilia, filopodia, or dendritic spinesare common, where such physical move-ment cannot easily be detected. Nonethelessthey house many signal-transduction path-ways that deploy receptors, G proteins, PLC,and TRPs. Such pathways are commonlyexplored according to the current dominantparadigm, with the ultimate explanationbeing protein phosphorylations due to PLC’sactivation of protein kinase C (PKC) (71).One wonders how much the PLC-inducedchanges in bilayer tension discussed here infact participate in these pathways (72).

Fig. 3. The mechanosensitivity of Kv. (A–C ) The Kv1.2paddle chimera expressed in an Sf-9 cell responds tovoltage steps ranging from −100 to 40 mV. Theresponses in on-cell mode (A) are drastically differentfrom those in the whole-cell mode (B). The normalizedconductances are graphically compared in C. As inter-preted in D, the large differences originate from thehigh pipet tension in the on-cell patch due to lipid–glassadhesion, but the lack of it in the whole-cell membrane,where the cortical cytoskeleton prevents the localmembranes from being stretched (from ref. 44). (E ) Therole of Kv1.1’s mechanosensitivity in countering im-pact-induced excitation in vivo is made evident whena copy of a dominant-negative allele of Kv1.1 isexpressed in the mouse (mceph+/−) causing it to beoverly sensitive to gentle poking, as indicated by the lowlimb-withdrawal threshold of prod strength (from ref.58). ns, not significant; **P < 0.01.

7902 | www.pnas.org/cgi/doi/10.1073/pnas.1313364111 Anishkin et al.

Dow

nloa

ded

by g

uest

on

July

27,

202

0

TRP SuperfamilyTRPs form a loosely connected family: Themammalian subtype TRP-A, -C, -M, -ML,-N, -P, and -V are only similar in the S5-P-S6core sequence. Although known since the1970s, it is the 1997 discovery of the “pepperchannel” (TRPV1) being a molecular ther-mometer (73) that caught imagination andpopularized TRP research, resulting in co-pious and complex findings that defy anycomprehensive review today. Besides thesubfamily diversities, TRP channels arepolymodal, meaning that each can receiveand integrate multiple inputs, includingcytoplasmic ligands, pH, Ca2+, membranelipids and other amphipaths, temperature,osmolarity, and mechanical force. TRPC1,C6, M3, M4, M7, P2, V1, V2, and V4 arejudged as mechanosensitive, by widely dif-ferent criteria. The risk seems not that theirmechanosensitivity is in doubt but thatother TRPs are not so assigned, given thatthe mechanosensitivity can so easily bemasked, as in Kv and the Drosophila TRP.For example, TRPV4 has long been thoughtto be activated by a lipid ligand (74) but canin fact be stretched open (75, 76).Besides mammals and insects, TRPs are

found in protists, algae, diatoms, fungi, etc.,although not in prokaryotes (16). The ex-tensively studied TRPY1 (77) in the vacuolar

membrane of the budding yeast presents a300-pS inward rectifier, favoring the flow ofcations including vacuolar Ca2+ into the cy-toplasm, observed upon hyperosmotic shock.It opens to membrane stretch in excisedpatches. Forward and reverse genetics pointto the importance of aromatic residues, whichtend to locate at the polar–nonpolar interfacesof the bilayer. The S5-P-S6 core appears to bethe force sensor because insertion of amor-phous peptides covalently disrupting theconnection between the core from the S1-S4peripheral domains has little effect on chan-nel mechansoensitivity (78).The atomic structure of TRPV1 has re-

cently been solved by cryoelectron micros-copy showing a Kv-like basic structure withseveral differences (79, 80). It has also beenpurified and reconstituted into lipid bilayers,showing that this minimal assembly is suffi-cient for temperature sensing (81). Therehas been no report on TRP channel re-constitution experiments to directly test theprinciple of lipid-force sensing. However,indirect evidence is accumulating, arguingagainst force transmission from partnerproteins. For example, rat TRPV4 has beenexpressed in budding yeast and remainsresponsive to osmotic stimulation (82) andin Xenopus oocytes displaying mechano-sensitive unitary conductances (75, 76).

Given the divergence among yeast, toad,and rat, it is difficult to imagine how het-erologous cytoskeleton or other tether canreconnect to transmit force.

AccessoriesBesides the definitive reconstitution experi-ments, there are other evidences that generalphysical properties instead of specific chem-ical binding govern mechanosensitivity. Byvarious criteria, reported mechanosensitivechannels include MscL, MscS, Mec-4/10,TRP, Piezo, K2p, Kv, KCa, Nav, NMDAreceptor, CFTR, AChR, etc. (see ref. 83 forreferences). These proteins have no simi-larities and no recognizable “force-sensingdomain” comparable to voltage sensor orligand-binding sites. We emphasize thenonspecific physical interaction between em-bedded protein and the bilayer here partlybecause it is often ignored. There are, ofcourse, examples of specific binding of cer-tain lipids (e.g., PIP2), anesthetics, or drugs,that act on membrane proteins.The cortical cytoskeleton protects the

membrane from being stretch (Fig. 3 A–D) inanimal cells in general. Treatment with cyto-chalasins, actin-filament dissociating agents,ease the stretch activation of myocyte’s nativeunitary conductances, the first such activationever described (84). Similar observations havealso been made on mechanosensitive chan-nels of known molecular identity. For exam-ple, treatment with latrunculin, another suchagent, greatly enhances the suction-inducedcurrent through TREK-1–expressed in COScells (85). Mutating β-spectrin, which linksthe membrane to the cytoskeleton, reducesthe force needed to pull membrane tetherfrom cultured touch-receptor neurons of theworm with an atomic force microscope (86).The elaborate anatomy of the vertebrate

cochlear or the insect Johnston’s organillustrates how additional structures are builtto enhance mechanosensations. At the levelof molecules, channels are accessorized toharness or amplify the signals. Some chan-nels clearly choose their own lipids, e.g., thenegatively charged PIP2 as deduced by itseffects. The binding can even be tight enoughfor resolution in crystal structure, such as Kir

(87). Tightly bound lipids can be consideredsubunits of such proteins. Many channelsseem to preferentially reside in raft-likemicrodomains. Mechanosensitive channelsare found to be in ordered and thickerdomains with cholesterol and sphingolipids(83). Accessory proteins such as cadherins orintegrins also form tethers to transmit forceinward. There is little doubt that the trans-mitted force gate mechanosensitive channels,as in the case of the hair-cell transduction

in

out

A

C

B

D

E

Fig. 4. The photomechanical transduction pathway in the Drosophila compound eye. (A) A diagram showing thepathway, where the G protein-activated PLC converts PIP2 to DAG, concentrated in the inner leaflet (upper half of thebilayer in this diagram). (B) Lipid structures emphasize that the beheading of the rod-shaped PIP2 converts it intothe cone-shaped DAG, yielding change in both the volume of lipid and spontaneous curvature in the inner leaflet(modified from ref. 64). (C ) The arrangement of the cantilever of an atomic force microscope on a retina. (D) Briefflashes (bump on the center line) leads to shrinkages as measured by the atomic force microscope, proportional tolight intensity (lower curves). The shrinkage precedes the phototransduction current (upper curves) (from ref. 64).E shows the structure of an ommatidium, the unit of the compound eye, which comprises thousands of microvilliwhere the phototransduction takes place. It is the sum of the small shrinkage in each microvillus that is measured bythe atomic force microscope (modified from ref. 100).

Anishkin et al. PNAS | June 3, 2014 | vol. 111 | no. 22 | 7903

PERS

PECT

IVE

Dow

nloa

ded

by g

uest

on

July

27,

202

0

channel. However, it is unclear whether thechannels bind the tether directly or indirectlythrough other proteins to receive the force.The channel can even receive the force froma lipid microdomain, which is pulled by thetether (83). The hair-cell tip link producestenting that appears to tense up the mem-brane (88). For cells less specialized than thehair cells, a common conception is that thechannel is pulled by the cytoskeleton. Asdescribed above, the cortical cytoskeletonprevents rather than activates mechano-sensitive channels (44) (Fig. 3 A–D). Al-though tugging on the stress fibers canindeed open certain channels (89), it isagain unclear whether we are pulling thechannel or the surrounding lipids.If a channel with a small free-energy dif-

ference between its closed and open state hasto expand a sizable gate to allow ion passage,it must be very sensitive to membrane ten-sion, as in the case of Kv. Additional struc-tures may well be needed to prevent ac-cidental opening. The peripheral domainsin Kv and its related tetrameric channelsmay well be such structures. A major role ofthe cholesterol-rich stiffen platform is likelyto enforce order and suppress mechanicalnoise (83).

The OriginHomo sapiens is but ∼4 × 105-y-old, whereasAustralian slime-mat fossils date back to∼4 × 109 y, near the earth’s formation at∼4.5 × 109 y. Life’s origin requires an earlystep to segregate and concentrate the self-replicators, presumably RNAs, and theirmetabolites. Segregation of an aqueouscompartment requires enclosure with anamphipathic material. Thus, life’s abiotic or-igin is not just a chicken-and-egg problem,but also a yolk-and-shell problem. Besides theRNA-world, there is in fact a lipid-worldtheory that argues for the early appearanceof vesicles and protocells (90, 91). Venterand coworkers’ Mycoplasma laboratoriumexperiment (92) or Michael Crichton’sJurassic Park fiction both require placingremodeled or renewed DNAs into a preex-isting cell. To completely create life in vitro,synthetic RNAs need to be wrapped insidea man-made bilayer vesicle, as being at-tempted in the Szostak laboratory (93, 94)and elsewhere (95). Mother Nature took hertime with available material. Abiotic Miller–Urey experiments produce amino acids butalso long-chain hydrocarbons, their deriva-tives, and even lipid-like amphipaths. Somemeteorites carry much more fatty acidthan amino acids. Such lipid-like mole-cules undergo spontaneous noncovalentlylinked aggregations to form micelles, bilayers,

or vesicles through entropy-driven hydro-phobic interactions. This spontaneity is thekey, requiring no additional material or in-formation. The self-assembled bilayer comeswith its physical properties and abilities.Even in a naked lipid bilayer, lateral forcesprofile respond to deformation (10), heat(96), pH (97), ions (98), chemical mod-ifications by light (99), etc. It could workas an elementary sensor in response tochanges, including osmotic changes, inthe violent primordial soup. The responsesmay appear crude now, but allow for evo-lutionary improvements by the embedmentof impurities, such as peptides. We arenow at awe by the sophistication of modernmembranes, which house the key energyconverters: photosynthesis and oxidativephosphorylation. However, at the most basiclevel, we are still struggling to understandhow physical forces govern various cell bi-ological, genetic, and developmental pro-cesses. The 21st century remains surprisinglyignorant of mechanosensitive processes.We know not the molecular bases of touch,

hearing, the Bayliss effect, electromechani-cal feedback of the heart, the monitoring ofblood pressure. Nor do we know how organ-isms sense gravitation, wind, waves, drought,and flood. The growing literature on theimportance of lipids and amphipaths leadsus to advocate a paradigm that combinesthe chemistry and the mechanics of thebilayer beyond lock-and-key thinking. Wereview here the recent extension of theforce-from-lipid principle from bacterial toanimal ion channels. This extension showedthat membrane proteins are universally gov-erned by the mechanical properties of sur-rounding lipids throughout evolution. Thisreview on the extension of force-from-lipidprinciple is meant to remind us that certainphysical principles are so basic that theyunderlie all molecules, whether they residein bacteria, however lowly, or in humans,however lofty.

ACKNOWLEDGMENTS. Work in our laboratories is sup-ported by the Huck Institute of Life Sciences (A.A.) andNational Institutes of Health Grant GM096088 and the VilasTrust of the University of Wisconsin–Madison (to C.K.).

1 Long SB, Campbell EB, Mackinnon R (2005) Crystal structure of

a mammalian voltage-dependent Shaker family K+ channel. Science

309(5736):897–903.2 Beyer MK, Clausen-Schaumann H (2005) Mechanochemistry: The

mechanical activation of covalent bonds. Chem Rev 105(8):

2921–2948.3 Wang W, Donini O, Reyes CM, Kollman PA (2001) Biomolecular

simulations: Recent developments in force fields, simulations of

enzyme catalysis, protein-ligand, protein-protein, and protein-nucleic

acid noncovalent interactions. Annu Rev Biophys Biomol Struct

30:211–243.4 Khalid S, Bond PJ (2013) Multiscale molecular dynamics

simulations of membrane proteins. Methods Mol Biol 924:635–657.5 Stansfeld PJ, Sansom MS (2011) Molecular simulation approaches

to membrane proteins. Structure 19(11):1562–1572.6 Alejandre J, Tildesley DJ, Chapela GA (1995) Molecular dynamics

simulation of the orthobaric densities and surface tension of water.

J Chem Phys 102(11):4574–4583.7 Cantor RS (1997) Lateral pressures in cell membranes: A

mechanism for modulation of protein function. J Phys Chem B

101:1723–1725.8 Cantor RS (1999) The influence of membrane lateral pressures on

simple geometric models of protein conformational equilibria. Chem

Phys Lipids 101(1):45–56.9 Lindahl E, Edholm O (2000) Spatial and energetic-entropic

decomposition of surface tension in lipid bilayers from molecular

dynamics simulations. J Chem Phys 113(9):3882–3893.10 Gullingsrud J, Schulten K (2004) Lipid bilayer pressure profiles

and mechanosensitive channel gating. Biophys J 86(6):3496–3509.11 Phillips R, Ursell T, Wiggins P, Sens P (2009) Emerging roles for

lipids in shaping membrane-protein function. Nature 459(7245):

379–385.12 Martinac B, Hamill OP (2002) Gramicidin A channels switch

between stretch activation and stretch inactivation depending on

bilayer thickness. Proc Natl Acad Sci USA 99(7):4308–4312.13 Shapovalov G, Lester HA (2004) Gating transitions in bacterial ion

channels measured at 3 microns resolution. J Gen Physiol 124(2):

151–161.14 Martinac B, Adler J, Kung C (1990) Mechanosensitive ion

channels of E. coli activated by amphipaths. Nature 348(6298):

261–263.15 Markin VS, Martinac B (1991) Mechanosensitive ion channels as

reporters of bilayer expansion. A theoretical model. Biophys J 60(5):

1120–1127.16 Kung C, Martinac B, Sukharev S (2010) Mechanosensitive

channels in microbes. Annu Rev Microbiol 64:313–329.

17 Kung C (2005) A possible unifying principle for

mechanosensation. Nature 436(7051):647–654.18 Sukharev SI, Blount P, Martinac B, Kung C (1997)

Mechanosensitive channels of Escherichia coli: The MscL gene,

protein, and activities. Annu Rev Physiol 59:633–657.19 Nomura T, et al. (2012) Differential effects of lipids and lyso-lipids

on the mechanosensitivity of the mechanosensitive channels MscL

and MscS. Proc Natl Acad Sci USA 109(22):8770–8775.20 Moe P, Blount P (2005) Assessment of potential stimuli for

mechano-dependent gating of MscL: Effects of pressure, tension,

and lipid headgroups. Biochemistry 44(36):12239–12244.21 Perozo E, Kloda A, Cortes DM, Martinac B (2002) Physical

principles underlying the transduction of bilayer deformation forces

during mechanosensitive channel gating. Nat Struct Biol 9(9):

696–703.22 Sukharev SI, Sigurdson WJ, Kung C, Sachs F (1999) Energetic and

spatial parameters for gating of the bacterial large conductance

mechanosensitive channel, MscL. J Gen Physiol 113(4):525–540.23 Chiang CS, Anishkin A, Sukharev S (2004) Gating of the large

mechanosensitive channel in situ: Estimation of the spatial scale of

the transition from channel population responses. Biophys J 86(5):

2846–2861.24 Chang G, Spencer RH, Lee AT, Barclay MT, Rees DC (1998)

Structure of the MscL homolog from Mycobacterium tuberculosis: A

gated mechanosensitive ion channel. Science 282(5397):2220–2226.25 Perozo E, Cortes DM, Sompornpisut P, Kloda A, Martinac B

(2002) Open channel structure of MscL and the gating mechanism of

mechanosensitive channels. Nature 418(6901):942–948.26 Sheetz MP, Singer SJ (1974) Biological membranes as bilayer

couples. A molecular mechanism of drug-erythrocyte interactions.

Proc Natl Acad Sci USA 71(11):4457–4461.27 Yoshimura K, Nomura T, Sokabe M (2004) Loss-of-function

mutations at the rim of the funnel of mechanosensitive channel

MscL. Biophys J 86(4):2113–2120.28 Lundbaek JA, Collingwood SA, Ingólfsson HI, Kapoor R,

Andersen OS (2010) Lipid bilayer regulation of membrane protein

function: Gramicidin channels as molecular force probes. J R Soc

Interface 7(44):373–395.29 Martinac B, Saimi Y, Kung C (2008) Ion channels in microbes.

Physiol Rev 88(4):1449–1490.30 Doyle DA, et al. (1998) The structure of the potassium channel:

Molecular basis of K+ conduction and selectivity. Science 280(5360):

69–77.31 Miller AN, Long SB (2012) Crystal structure of the human two-

pore domain potassium channel K2P1. Science 335(6067):432–436.

7904 | www.pnas.org/cgi/doi/10.1073/pnas.1313364111 Anishkin et al.

Dow

nloa

ded

by g

uest

on

July

27,

202

0

32 Brohawn SG, del Mármol J, MacKinnon R (2012) Crystal structureof the human K2P TRAAK, a lipid- and mechano-sensitive K+ ionchannel. Science 335(6067):436–441.33 Brohawn SG, Campbell EB, MacKinnon R (2013) Domain-swapped chain connectivity and gated membrane access in a Fab-mediated crystal of the human TRAAK K+ channel. Proc Natl AcadSci USA 110(6):2129–2134.34 Maingret F, Fosset M, Lesage F, Lazdunski M, Honoré E (1999)TRAAK is a mammalian neuronal mechano-gated K+ channel. J BiolChem 274(3):1381–1387.35 Enyedi P, Czirják G (2010) Molecular background of leak K+currents: Two-pore domain potassium channels. Physiol Rev 90(2):559–605.36 Patel AJ, Lazdunski M, Honoré E (2001) Lipid and mechano-gated 2P domain K(+) channels. Curr Opin Cell Biol 13(4):422–428.37 Honoré E (2007) The neuronal background K2P channels: Focuson TREK1. Nat Rev Neurosci 8(4):251–261.38 Devilliers M, et al. (2013) Activation of TREK-1 by morphineresults in analgesia without adverse side effects. Nat Commun4:2941.39 Kloda A, Lua L, Hall R, Adams DJ, Martinac B (2007) Liposomereconstitution and modulation of recombinant N-methyl-D-aspartatereceptor channels by membrane stretch. Proc Natl Acad Sci USA104(5):1540–1545.40 Maroto R, et al. (2005) TRPC1 forms the stretch-activated cationchannel in vertebrate cells. Nat Cell Biol 7(2):179–185.41 Berrier C, et al. (2013) The purified mechanosensitive channelTREK-1 is directly sensitive to membrane tension. J Biol Chem288(38):27307–27314.42 Opsahl LR, Webb WW (1994) Lipid-glass adhesion in giga-sealedpatch-clamped membranes. Biophys J 66(1):75–79.43 Schmidt D, MacKinnon R (2008) Voltage-dependent K+ channelgating and voltage sensor toxin sensitivity depend on the mechanicalstate of the lipid membrane. Proc Natl Acad Sci USA 105(49):19276–19281.44 Schmidt D, del Mármol J, MacKinnon R (2012) Mechanistic basisfor low threshold mechanosensitivity in voltage-dependent K+

channels. Proc Natl Acad Sci USA 109(26):10352–10357.45 Honoré E, Patel AJ, Chemin J, Suchyna T, Sachs F (2006)Desensitization of mechano-gated K2P channels. Proc Natl Acad SciUSA 103(18):6859–6864.46 Brohawn SG, Su Z, MacKinnon R (2014) Mechanosensitivity ismediated directly by the lipid membrane in TRAAK and TREK 1 K+

channels. Proc Natl Acad Sci USA 111(9):3614–3619.47 Loukin SH, Saimi Y (2002) Carboxyl tail prevents yeast K(+)channel closure: Proposal of an integrated model of TOK1 gating.Biophys J 82(2):781–792.48 Bagriantsev SN, Peyronnet R, Clark KA, Honoré E, Minor DL, Jr.(2011) Multiple modalities converge on a common gate to controlK2P channel function. EMBO J 30(17):3594–3606.49 Piechotta PL, et al. (2011) The pore structure and gatingmechanism of K2P channels. EMBO J 30(17):3607–3619.50 Milac A, et al. (2011) Structural models of TREK channels andtheir gating mechanism. Channels (Austin) 5(1):23–33.51 Long SB, Campbell EB, Mackinnon R (2005) Voltage sensor ofKv1.2: Structural basis of electromechanical coupling. Science309(5736):903–908.52 Long SB, Tao X, Campbell EB, MacKinnon R (2007) Atomicstructure of a voltage-dependent K+ channel in a lipid membrane-likeenvironment. Nature 450(7168):376–382.53 Lee SY, Banerjee A, MacKinnon R (2009) Two separate interfacesbetween the voltage sensor and pore are required for the function ofvoltage-dependent K(+) channels. PLoS Biol 7(3):e47.54 Schmidt D, Jiang QX, MacKinnon R (2006) Phospholipids and theorigin of cationic gating charges in voltage sensors. Nature444(7120):775–779.

55 Li Q, Wanderling S, Sompornpisut P, Perozo E (2014) Structuralbasis of lipid-driven conformational transitions in the KvAP voltage-sensing domain. Nat Struct Mol Biol 21(2):160–166.56 Reeves D, Ursell T, Sens P, Kondev J, Phillips R (2008) Membranemechanics as a probe of ion-channel gating mechanisms. Phys Rev EStat Nonlin Soft Matter Phys 78(4 Pt 1):041901.57 Morris CE (2011) Voltage-gated channel mechanosensitivity: Factor friction? Front Physiol 2:25.58 Hao J, et al. (2013) Kv1.1 channels act as mechanical brake in thesenses of touch and pain. Neuron 77(5):899–914.59 Lee SY, MacKinnon R (2004) A membrane-access mechanism ofion channel inhibition by voltage sensor toxins from spider venom.Nature 430(6996):232–235.60 Pak WL (1979) Study of photoreceptor function using Drosophilamutants. Neurogenetics, Genetic Approaches to the NervousSystem, ed Breakfile X (Elsevier North-Holland Publishing Co.,Amsterdam), pp 67–99.61 Minke B, Wu C-F, Pak WL (1975) Induction of photoreceptorvoltage noise in the dark in Drosophila mutant. Nature 258(5530):84–87.62 Montell C, Rubin GM (1989) Molecular characterization of theDrosophila trp locus: A putative integral membrane protein requiredfor phototransduction. Neuron 2(4):1313–1323.63 Hardie RC, Minke B (1992) The trp gene is essential for a light-activated Ca2+ channel in Drosophila photoreceptors. Neuron 8(4):643–651.64 Hardie RC, Franze K (2012) Photomechanical responses inDrosophila photoreceptors. Science 338(6104):260–263.65 Huang J, et al. (2010) Activation of TRP channels by protons andphosphoinositide depletion in Drosophila photoreceptors. Curr Biol20(3):189–197.66 Lev S, Katz B, Tzarfaty V, Minke B (2012) Signal-dependenthydrolysis of phosphatidylinositol 4,5-bisphosphate withoutactivation of phospholipase C: Implications on gating of DrosophilaTRPL (transient receptor potential-like) channel. J Biol Chem 287(2):1436–1447.67 Hughes S, Hankins MW, Foster RG, Peirson SN (2012)Melanopsin phototransduction: Slowly emerging from the dark. ProgBrain Res 199:19–40.68 Sexton T, Buhr E, Van Gelder RN (2012) Melanopsin andmechanisms of non-visual ocular photoreception. J Biol Chem 287(3):1649–1656.69 Bellono NW, Kammel LG, Zimmerman AL, Oancea E (2013) UVlight phototransduction activates transient receptor potential A1 ionchannels in human melanocytes. Proc Natl Acad Sci USA 110(6):2383–2388.70 Moore C, et al. (2013) UVB radiation generates sunburn pain andaffects skin by activating epidermal TRPV4 ion channels andtriggering endothelin-1 signaling. Proc Natl Acad Sci USA 110(34):E3225–E3234.71 Newton AC (2010) Protein kinase C: Poised to signal. Am JPhysiol Endocrinol Metab 298(3):E395–E402.72 Liman ER (2012) Cell signaling. Putting the squeeze onphototransduction. Science 338(6104):200–201.73 Caterina MJ, et al. (1997) The capsaicin receptor: A heat-activated ion channel in the pain pathway. Nature 389(6653):816–824.74 Watanabe H, et al. (2003) Anandamide and arachidonic acid useepoxyeicosatrienoic acids to activate TRPV4 channels. Nature424(6947):434–438.75 Loukin S, Zhou X, Su Z, Saimi Y, Kung C (2010) Wild-type andbrachyolmia-causing mutant TRPV4 channels respond directly tostretch force. J Biol Chem 285(35):27176–27181.76 Loukin S, Su Z, Kung C (2011) Increased basal activity is a keydeterminant in the severity of human skeletal dysplasia caused byTRPV4 mutations. PLoS ONE 6(5):e19533.

77 Su Z, et al. (2009) The use of yeast to understand TRP-channelmechanosensitivity. Pflugers Arch 458(5):861–867.78 Su Z, Anishkin A, Kung C, Saimi Y (2011) The core domain as theforce sensor of the yeast mechanosensitive TRP channel. J GenPhysiol 138(6):627–640.79 Cao E, Liao M, Cheng Y, Julius D (2013) TRPV1 structures indistinct conformations reveal activation mechanisms. Nature504(7478):113–118.80 Liao M, Cao E, Julius D, Cheng Y (2013) Structure of the TRPV1ion channel determined by electron cryo-microscopy. Nature504(7478):107–112.81 Cao E, Cordero-Morales JF, Liu B, Qin F, Julius D (2013) TRPV1channels are intrinsically heat sensitive and negatively regulated byphosphoinositide lipids. Neuron 77(4):667–679.82 Loukin SH, Su Z, Kung C (2009) Hypotonic shocks activate ratTRPV4 in yeast in the absence of polyunsaturated fatty acids. FEBSLett 583(4):754–758.83 Anishkin A, Kung C (2013) Stiffened lipid platforms at molecularforce foci. Proc Natl Acad Sci USA 110(13):4886–4892.84 Guharay F, Sachs F (1984) Stretch-activated single ion channelcurrents in tissue-cultured embryonic chick skeletal muscle. J Physiol352:685–701.85 Lauritzen I, et al. (2005) Cross-talk between the mechano-gatedK2P channel TREK-1 and the actin cytoskeleton. EMBO Rep 6(7):642–648.86 Krieg M, Dunn AR, Goodman MB (2014) Mechanical control ofthe sense of touch by β-spectrin. Nat Cell Biol 16(3):224–233.87 Hansen SB, Tao X, MacKinnon R (2011) Structural basis of PIP2activation of the classical inward rectifier K+ channel Kir2.2. Nature477(7365):495–498.88 Assad JA, Shepherd GM, Corey DP (1991) Tip-link integrity andmechanical transduction in vertebrate hair cells. Neuron 7(6):985–994.89 Hayakawa K, Tatsumi H, Sokabe M (2008) Actin stress fiberstransmit and focus force to activate mechanosensitive channels. J CellSci 121(Pt 4):496–503.90 Segré D, Ben-Eli D, Deamer DW, Lancet D (2001) The lipid world.Orig Life Evol Biosph 31(1-2):119–145.91 Chen IA, Walde P (2010) From self-assembled vesicles toprotocells. Cold Spring Harb Perspect Biol 2(7):a002170.92 Gibson DG, et al. (2010) Creation of a bacterial cell controlled bya chemically synthesized genome. Science 329(5987):52–56.93 Adamala K, Szostak JW (2013) Nonenzymatic template-directedRNA synthesis inside model protocells. Science 342(6162):1098–1100.94 Service RF (2013) The life force. Science 342(6162):1032–1034.95 Black RA, et al. (2013) Nucleobases bind to and stabilizeaggregates of a prebiotic amphiphile, providing a viable mechanismfor the emergence of protocells. Proc Natl Acad Sci USA 110(33):13272–13276.96 Simon SA, Advani S, McIntosh TJ (1995) Temperaturedependence of the repulsive pressure between phosphatidylcholinebilayers. Biophys J 69(4):1473–1483.97 Gong K, Feng S-S, Go ML, Soew PH (2002) Effects of pH on thestability and compressibility of DPPC/cholesterol monolayers at theair–water interface. Colloids Surf A Physicochem Eng Asp 207(1–3):113–125.98 Ermakov YA, Kamaraju K, Sengupta K, Sukharev S (2010)Gadolinium ions block mechanosensitive channels by altering thepacking and lateral pressure of anionic lipids. Biophys J 98(6):1018–1027.99 Girotti AW (2001) Photosensitized oxidation of membrane lipids:Reaction pathways, cytotoxic effects, and cytoprotectivemechanisms. J Photochem Photobiol B 63(1-3):103–113.100 Montell C (2012) Drosophila visual transduction.Trends Neurosci 35(6):356–363.

Anishkin et al. PNAS | June 3, 2014 | vol. 111 | no. 22 | 7905

PERS

PECT

IVE

Dow

nloa

ded

by g

uest

on

July

27,

202

0