owen p. hamill and boris martinac 81:685-740, 2001....hamill, owen p., and boris martinac. molecular...

57
81:685-740, 2001. Physiol Rev Owen P. Hamill and Boris Martinac You might find this additional information useful... 476 articles, 203 of which you can access free at: This article cites http://physrev.physiology.org/cgi/content/full/81/2/685#BIBL 98 other HighWire hosted articles, the first 5 are: This article has been cited by [PDF] [Full Text] [Abstract] , August 1, 2007; 582 (3): 977-990. J. Physiol. and S. J. Kim J. H. Nam, H.-S. Lee, Y. H. Nguyen, T. M. Kang, S. W. Lee, H.-Y. Kim, S. J. Kim, Y. E. Earm phosphatidylinositol 4,5-bisphosphate in B lymphocytes Mechanosensitive activation of K+ channel via phospholipase C-induced depletion of [PDF] [Full Text] [Abstract] , August 1, 2007; 21 (10): 2389-2399. FASEB J M. Althaus, R. Bogdan, W. G. Clauss and M. Fronius ion channel open probability Mechano-sensitivity of epithelial sodium channels (ENaCs): laminar shear stress increases [PDF] [Full Text] [Abstract] , September 1, 2007; 293 (3): H1646-H1653. Am J Physiol Heart Circ Physiol K. Yamamoto, N. Shimizu, S. Obi, S. Kumagaya, Y. Taketani, A. Kamiya and J. Ando endothelial cells Involvement of cell surface ATP synthase in flow-induced ATP release by vascular [PDF] [Full Text] [Abstract] , September 1, 2007; 93 (5): 1493-1509. Biophys. J. I. A. Solov'yov and W. Greiner Theoretical Analysis of an Iron Mineral-Based Magnetoreceptor Model in Birds [PDF] [Full Text] [Abstract] , December 1, 2007; 22 (6): 366-372. Physiology K. K. Huynh, J. G. Kay, J. L. Stow and S. Grinstein Fusion, Fission, and Secretion During Phagocytosis on the following topics: http://highwire.stanford.edu/lists/artbytopic.dtl can be found at Medline items on this article's topics Oncology .. Csk Physiology .. Hearing Physiology .. Mechanoreception Biochemistry .. Channel Mouth Chemistry .. Lipid Bilayers Biochemistry .. Membrane Lipids including high-resolution figures, can be found at: Updated information and services http://physrev.physiology.org/cgi/content/full/81/2/685 can be found at: Physiological Reviews about Additional material and information http://www.the-aps.org/publications/prv This information is current as of December 13, 2007 . http://www.the-aps.org/. 20814-3991. Copyright © 2005 by the American Physiological Society. ISSN: 0031-9333, ESSN: 1522-1210. Visit our website at MD published quarterly in January, April, July, and October by the American Physiological Society, 9650 Rockville Pike, Bethesda provides state of the art coverage of timely issues in the physiological and biomedical sciences. It is Physiological Reviews on December 13, 2007 physrev.physiology.org Downloaded from

Upload: others

Post on 31-Dec-2019

5 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

81:685-740, 2001. Physiol RevOwen P. Hamill and Boris Martinac

You might find this additional information useful...

476 articles, 203 of which you can access free at: This article cites http://physrev.physiology.org/cgi/content/full/81/2/685#BIBL

98 other HighWire hosted articles, the first 5 are: This article has been cited by

  [PDF]  [Full Text]  [Abstract]

, August 1, 2007; 582 (3): 977-990. J. Physiol.and S. J. Kim J. H. Nam, H.-S. Lee, Y. H. Nguyen, T. M. Kang, S. W. Lee, H.-Y. Kim, S. J. Kim, Y. E. Earm

phosphatidylinositol 4,5-bisphosphate in B lymphocytesMechanosensitive activation of K+ channel via phospholipase C-induced depletion of 

[PDF]  [Full Text]  [Abstract], August 1, 2007; 21 (10): 2389-2399. FASEB J

M. Althaus, R. Bogdan, W. G. Clauss and M. Fronius ion channel open probability

Mechano-sensitivity of epithelial sodium channels (ENaCs): laminar shear stress increases 

[PDF]  [Full Text]  [Abstract], September 1, 2007; 293 (3): H1646-H1653. Am J Physiol Heart Circ Physiol

K. Yamamoto, N. Shimizu, S. Obi, S. Kumagaya, Y. Taketani, A. Kamiya and J. Ando endothelial cells

Involvement of cell surface ATP synthase in flow-induced ATP release by vascular 

[PDF]  [Full Text]  [Abstract], September 1, 2007; 93 (5): 1493-1509. Biophys. J.

I. A. Solov'yov and W. Greiner Theoretical Analysis of an Iron Mineral-Based Magnetoreceptor Model in Birds

  [PDF]  [Full Text]  [Abstract]

, December 1, 2007; 22 (6): 366-372. PhysiologyK. K. Huynh, J. G. Kay, J. L. Stow and S. Grinstein

Fusion, Fission, and Secretion During Phagocytosis

on the following topics: http://highwire.stanford.edu/lists/artbytopic.dtlcan be found at Medline items on this article's topics

Oncology .. Csk Physiology .. Hearing Physiology .. Mechanoreception Biochemistry .. Channel Mouth Chemistry .. Lipid Bilayers Biochemistry .. Membrane Lipids

including high-resolution figures, can be found at: Updated information and services http://physrev.physiology.org/cgi/content/full/81/2/685

can be found at: Physiological Reviewsabout Additional material and information http://www.the-aps.org/publications/prv

This information is current as of December 13, 2007 .  

http://www.the-aps.org/.20814-3991. Copyright © 2005 by the American Physiological Society. ISSN: 0031-9333, ESSN: 1522-1210. Visit our website at

MDpublished quarterly in January, April, July, and October by the American Physiological Society, 9650 Rockville Pike, Bethesda provides state of the art coverage of timely issues in the physiological and biomedical sciences. It isPhysiological Reviews

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 2: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

Molecular Basis of Mechanotransduction in Living Cells

OWEN P. HAMILL AND BORIS MARTINAC

Physiology and Biophysics, University Of Texas Medical Branch, Galveston, Texas; and

Department of Pharmacology, University Of Western Australia, Nedlands, West Australia

I. Introduction 686A. Basic requirement for mechanosensitivity 687B. Strategy and scope of this review 687

II. Lipid Bilayer Structure and Its Response to Mechanical Deformation 687A. Membrane compression 688B. Membrane area expansion/thinning 688C. Membrane bending/curvature 689D. Membrane extension/shear 689E. Viscous properties and dynamic response of bilayer vesicles 689

III. Mechanical Deformation of the Bilayer by Membrane Protein Insertion 690IV. Simple Peptides That Form Mechanically Gated Channels 691

A. Alamethicin 691B. Gramicidin 692

V. Structure of Prokaryotic Cells 692VI. Mechanically Gated Channels in Bacteria and Archaea 694

A. Identification of the MscL gene/protein 694B. Structure of MscL 694C. Conductive properties of MscL 695D. Is MscL a hexamer or a pentamer? 696E. Origin of MscL mechanosensitivity 696F. Extrinsic and intrinsic factors that affect MscL and other MG channels 698G. Where is the MscL gate? 699H. Mutagenesis studies 699I. Models of MscL mechanosensitivity 699J. Membrane localization and physiological function of MscL 702K. MscS and MscM 703L. MG channels in Archaea 703

M. MG channels in evolution 704VII. The Structure of Animal Cells: Specific Roles of the Cortical Cytoskeleton and Extracellular Matrix

in Mechanosensitivity 705VIII. Mechanically Gated Channels in Animal Cells 706

A. Membrane patch mechanics and morphology 707B. Discrepancy between membrane patch and whole cell mechanosensitivity 709C. MG channel gating: “tethered” versus “bilayer” models 710D. MG channel classification: is there a unifying mechanism for activation and inactivation

of MG channels? 715E. Rapid adaptation of MG channel activity 716F. Structure of eukaryotic MG channels 719

IX. Mechanosensitive Elevation of Intracellular Calcium 723A. MS Ca21 influx mechanisms 723B. MS release of Ca21 from internal Ca21 stores 723

X. Mechanosensitive Release of Transmitter 724A. Historical perspective 724B. Tension-sensitive vesicle recruitment/exocytosis 725C. Stretch-facilitated transmitter release at the vertebrate motor synapse 726D. Mechanosensitive ATP release 726E. Membrane resealing: Ca21-induced vesicle-vesicle fusion and exocytosis 728

XI. Conclusions and Outstanding Issues 728

PHYSIOLOGICAL REVIEWS

Vol. 81, No. 2, April 2001Printed in U.S.A.

http://physrev.physiology.org 6850031-9333/01 $15.00 Copyright © 2001 the American Physiological Society

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 3: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81:685–740, 2001.—The simplest cell-like structure, the lipid bilayer vesicle, can respond to mechanical deformation byelastic membrane dilation/thinning and curvature changes. When a protein is inserted in the lipid bilayer, anenergetic cost may arise because of hydrophobic mismatch between the protein and bilayer. Localized changes inbilayer thickness and curvature may compensate for this mismatch. The peptides alamethicin and gramicidin and thebacterial membrane protein MscL form mechanically gated (MG) channels when inserted in lipid bilayers. Theirmechanosensitivity may arise because channel opening is associated with a change in the protein’s membrane-occupied area, its hydrophobic mismatch with the bilayer, excluded water volume, or a combination of these effects.As a consequence, bilayer dilation/thinning or changes in local membrane curvature may shift the equilibriumbetween channel conformations. Recent evidence indicates that MG channels in specific animal cell types (e.g.,Xenopus oocytes) are also gated directly by bilayer tension. However, animal cells lack the rigid cell wall thatprotects bacteria and plants cells from excessive expansion of their bilayer. Instead, a cortical cytoskeleton (CSK)provides a structural framework that allows the animal cell to maintain a stable excess membrane area (i.e., for itsvolume occupied by a sphere) in the form of membrane folds, ruffles, and microvilli. This excess membrane providesan immediate membrane reserve that may protect the bilayer from sudden changes in bilayer tension. Contractileelements within the CSK may locally slacken or tighten bilayer tension to regulate mechanosensitivity, whereasmembrane blebbing and tight seal patch formation, by using up membrane reserves, may increase membranemechanosensitivity. In specific cases, extracellular and/or CSK proteins (i.e., tethers) may transmit mechanicalforces to the process (e.g., hair cell MG channels, MS intracellular Ca21 release, and transmitter release) withoutincreasing tension in the lipid bilayer.

I. INTRODUCTION

Cells experience a wide variety of mechanical stimuliranging from thermal molecular agitation to potentiallydestructive osmotic pressure gradients. Therefore, fromthe onset, living organisms faced a basic dilemma in theirevolution. As a first priority, they required mechanismsthat would protect their delicate cell membrane frompotentially damaging mechanical stimuli. However, to in-teract with their changing mechanical environment (e.g.,during feeding, escaping, or mating), they needed mech-anosensitivity. Different organisms have solved the di-lemma by different strategies. Bacteria and plants evolveda rigid cell wall that protects their plasma membrane fromexcessive dilation. However, with this strategy they sac-rificed not only cell deformability but also mechanosen-sitivity. In contrast, animals have adopted strategies thatprotect their cell membrane while preserving a high de-gree of cell deformability and mechanosensitivity.

The first mechanosensitive (MS) processes may haveevolved as backup mechanisms for cell protection. Forexample, a large nonselective membrane pore activatedby osmotic swelling will release the cell’s contents andthereby reduce intracellular pressure and membrane ten-sion. Similarly, tension-sensitive fusion of intracellularmembrane vesicles with the cell membrane will act toreduce bilayer tension. These basic mechanisms of me-chanically gated (MG) channels and MS exocytosis mayhave been subsequently refined to participate in cell sig-naling. For example, MG channels and/or MS transmitterrelease are implicated in a myriad of physiological pro-cesses, including touch and pain sensation (46, 292, 403),hearing and vestibular function (148, 190), blood pressurecontrol (45, 61), salt and fluid balance (32), micturition

(36), tissue growth (98), cell volume regulation (301, 308,430), and turgor control (147, 265). Furthermore, abnor-malities in these mechanisms may contribute to neuronal(93) and muscular degeneration (116), cardiac arrhythmia(86, 117, 162), hypertension (224), arteriosclerosis (90),and glaucoma (282).

The external mechanical forces that dominate a cellvary depending on its size and relationship with othercells. For example, unicellular organisms like Esche-

richia coli are constantly jostled by the forces of Brown-ian motion that tend to keep them in suspension. Incontrast, multicellular organisms require specific MSmechanisms that constantly adjust their position in re-sponse to gravity. Furthermore, specific cells, dependingon their location within an organism and association withancillary structures, may be selectively exposed to spe-cific forms of mechanical stimuli, including steady in-dentations, high-frequency vibrations, osmotic pressuregradients, and hemodynamic pressure and fluid shearstresses. All external stimuli act on top of a dynamicbackground of various internally generated forces (e.g.,arising from hydrostatic pressure, cytoskeletal polymer-ization, and molecular motors) that are important in de-termining cell shape, growth, mobility, and adhesion (15,201). To monitor and respond selectively to these differ-ent forces most likely requires multiple, parallel signalingpathways, with each pathway designed to extract specificinformation regarding the “relevant stimulus” while filter-ing out irrelevant stimuli.

Over the last 20 years, the molecular nature of spe-cific MS membrane processes has been identified. Theseinclude MG membrane ion channels (156, 265, 286, 350)and MS receptors (53, 320), enzymes (241, 270), intracel-lular Ca21 release (204) and transmitter release (63). Be-

686 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 4: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

cause each of these elements or processes may interactwith one another, as well as with other non-MS elements,difficulties can arise in distinguishing mechanisms thatare directly or indirectly affected by mechanical forces.Furthermore, given that a single cell may express multiplemechanotransducers, a challenge can arise in determiningwhich transducer mediates a specific MS function (i.e,cause and effect relations). A notable example is verte-brate tactile sensation where the basic distinction be-tween physical and chemical mechanisms of mechano-transduction has yet to be made (cf. Refs. 135, 292). Inprinciple, one should be able to identify the mechano-transducer by comparing its specific properties (i.e., sen-sitivity, kinetics, and pharmacology) with those of the MSfunction.

A. Basic Requirement for Mechanosensitivity

For a membrane protein to be directly MS, it must besensitive to a membrane property that changes with me-chanical deformation. For the specific case of a simpletwo-state channel, a shift in the equilibrium betweenclosed and open channel conformations may be causedby changes in bilayer tension, thickness, or local curva-ture or by direct “tugging” on the protein by cytoskeletalor extracellular tethers. Therefore, a fundamental issue inmechanotransduction is the identification of the mem-brane parameter that actually confers mechanosensitivityon the membrane protein or process.

B. Strategy and Scope of This Review

The membrane of most animal cells is a compositestructure of extracellular (EC), bilayer, and cytoskeletal(CSK) layers. Because of its integrated nature, any exter-nally applied force produces varying tensions and strainsin multiple elements within the three layers (200). For thisreason, it becomes problematic in identifying a singlemembrane property that may be directly involved in themechanotransduction process. To overcome this prob-lem, we adopt a hierarchical approach and consider avariety of membrane preparations, progressing from thesimple artificial bilayer vesicle to increasingly more com-plex cells (i.e., from bacteria to animal cells). The ratio-nale for this approach is that if characteristics of a par-ticular mechanism can be identified (i.e., “finger-printed”)in a simple system, one should be better positioned torecognize its operation in more complex systems. Ourapproach would seem justified by the reoccurring themein evolution in which basic mechanisms that first evolvedin prokaryotes are conserved and refined to carry outmore diverse and specialized functions in eukaryotes.However, some processes such as exocytosis/endocytosisand release of Ca21 from intracellular membrane stores

are unique to eukaryotes (54), and therefore, their mech-anosensitivity must reflect more recently acquired mech-anisms of mechanotransduction.

II. LIPID BILAYER STRUCTURE AND ITS

RESPONSE TO MECHANICAL

DEFORMATION

Because the bilayer is the core structure aroundwhich all other membrane components are arranged, it iscritical to understand its molecular packing and how thispacking may change under steady state and dynamic me-chanical deformation. For example, intrinsic delays orrelaxations in the response of the bilayer to deformationmay be reflected in the functional dynamics (i.e., fre-quency response and adaptive behavior) of MG channelactivities. The bilayer is composed of lipid molecules thatform two monolayers stabilized by van der Waals forcesand the “hydrophobic” effect between the “hidden” acyllipid chains. In addition, water molecules surroundingeach lipid headgroup form hydrogen bonds that furtherstabilize the bilayer (180). Water molecules also penetratedeeper into the bilayer, hopping between acyl chain pack-ing defects, such as trans-gauche kinks. For example, it isestimated that ;4,000 water molecules pass a single phos-pholipid per second compared with 1 Na1 every 70 h (85).At reduced temperatures, lipid bilayers undergo a liquid-gel phase transition in which the acyl chain packing be-comes more ordered (187a). Furthermore, bilayers madeup of more than one phospholipid can undergo lateralphase separations (138, 221). In the case of cell mem-branes, it is generally assumed that their complex lipidand cholesterol makeup ensure a highly fluid state atphysiological temperatures. However, recent studies indi-cate preferential packing of sphingolipids and cholesterolinto floating platforms or rafts of lipid that can be isolatedas detergent-insoluble membrane complexes (369a). Al-though the occurrence of such phase separations willcomplicate bilayer and cell membrane mechanics, theirspecific effects have been little studied.

In principle, it is possible to study the mechanics ofplanar lipid bilayers, typically formed by painting a film ofphospholipid over a hole in a plastic barrier, (e.g., see Ref.112). However, such studies are complicated by the pres-ence of a torus of disordered lipid that can act as a lipidreservoir for formation of new bilayer (336, 433a). Incontrast, the lipid vesicle is a more “cell-like” structure inthat it is a closed system that has its volume set by theosmotic activity of the aqueous environment. How thelipid vesicle responds to mechanical deformation de-pends on both extrinsic (e.g., size and geometry) as wellas intrinsic properties (i.e., material elastic properties)(see Refs. 107, 109). For example, a deflated vesicle filledwith volumes insufficient to form a sphere is highly de-

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 687

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 5: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

formable but somewhat unstable with a tendency to spon-taneously “bud” or vesiculate. In contrast, a vesicle in-flated by osmotic or hydrostatic forces into a sphere isstable but shows limited deformability in that with furtherinflation (i.e., 2–4%) it either ruptures or under specificcircumstances forms pores (452). Transient pore forma-tion by releasing intravesicular pressure will preservevesicle structure and thus may have served as an inbuiltprotection mechanism for primordial cells before theyevolved protein mechanisms.

With the assumption that the lipid bilayer behaves asan elastic solid, its intrinsic mechanical properties can becharacterized by four elasticity constants (or moduli) thatdescribe the response of a unit area of bilayer to com-pression, expansion, bending, and extension (108, 109).The larger the moduli, the greater the resistance to thatform of deformation. Elastic deformations are directlyproportional to and follow instantaneously the applica-tion and removal of external forces. In comparison, vis-coelastic or plastic deformations show time dependence,and one has to take into account the different viscositycoefficients for each type of deformation (108). The elas-tic moduli of bilayer vesicles and human red blood cells(RBCs) have typically been measured with the micropi-pette aspiration technique (106). A critical feature of thistechnique is that there is minimal membrane adherence tothe pipette to ensure reversible and unimpeded move-ment of the aspirated portion of the membrane within thepipette. Under these circumstances, the membrane ten-sions developed during aspiration can be assumed isotro-pic throughout the vesicle, with the membrane protrusiondrawn into the pipette serving as an amplified measure-ment of membrane area changes. In contrast, in patch-clamp recording, the membrane adheres tightly to thewalls of the pipette (153). As long as the membrane-glassadhesion is not disrupted (i.e., the patch boundary re-mains constant), tension changes will be restricted to the“free” membrane area that spans the inside of the pipette(see sect. VIIIA).

A. Membrane Compression

An early study based on the effects of pressure on thelipid bilayer phase transition demonstrated that hydro-static pressures up to ;1 3 107 N/m2 (i.e., 100 atmo-spheres) did not significantly alter the bilayer densitychange associated with the phase transition (378). Basedon this result, Evans and Hochmuth (108) estimated thebilayer compressibility modulus was between 109 and1010 N/m2, similar to that of most “incompressible” fluids.A subsquent study based on X-ray diffraction analysis ofosmotically stressed liposomes gave similar estimates(275, see also Ref. 322). Although higher compressiveforces (i.e., 500–1,500 atm) have since been shown to

increase acyl chain packing density and squeeze water outof pure phospholipid bilayer, these effects are minimizedin bilayers that include cholesterol (16, 334a). Thus onemay assume that the bilayer of the cell membrane isvolumetrically incompressible and will maintain a con-stant density during the mechanical deformations en-countered under physiological conditions. As indicatedbelow, the resistance to volume compression is at least anorder of magnitude larger than the resistance to bilayerthickness and area changes.

B. Membrane Area Expansion/Thinning

The tight lateral packing of lipid molecules in thebilayer underlies its extremely low ion permeability andrelatively low water permeability (85). However, this fea-ture also contributes to the bilayer’s resistance to areaexpansion. This is because even slight additional separa-tion of lipid head groups (i.e., ;2%) will allow more waterto enter between the acyl chains and destabilize the hy-drophobic cohesive structure. Aspiration of spherical ves-icles indicates a simple linear relation between membranetension (t) and the relative area expansion of the bilayer

t 5 KA z DA/A0 (1)

where DA is the increase in surface area, A0 is the originalarea, and KA is the area expansion modulus (108). Typicalvalues of KA range between 102 and 103 mN/m dependingon the cholesterol content of the bilayer, while lytic ten-sions range between 3 and 30 mN/m, consistent withbilayers only being able to be expanded 2–4% beforerupture (108, 296, 298). With the assumption of a bilayercompressibility modulus of 109 N/m2, the bilayer is at least10-fold more compressible in area than in volume (givena KA of 200 mN/m divided by 3 nm for the bilayer thick-ness). Thus, at near lytic tensions, the area may increaseby ;4%, but the volume by ,0.4% so that the thicknesswill decrease by 3.6%. Thus any fractional change in areashould be accompanied by a proportional change in mem-brane thickness (h) so that

DA/A0 5 2Dh/h0 (2)

where h0 is the unstressed membrane thickness and theexpansion and thickness moduli are related according toKA 5 Kh z h0 (108). Estimates of Kh based on capacitancemeasurements of bilayer thickness changes during elec-trocompression indicate values of ;2 3 107 N/m2 (5),which would predict a KA of ;70 mN/m assuming abilayer membrane thickness of 3 nm (108).

Bilayer vesicles generated from membranes of RBC(183, 295, 296) and skeletal muscle (298) give values of KA

of around 500 mN/m, similar to the KA measured for

688 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 6: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

osmotically swollen (i.e, spherical) RBCs (106, 183). Thisagreement has been taken to indicate that the CSK of theRBC does not limit the elastic expansion of the bilayer(108). In studies of other cell types in which significantlylower values of KA have been reported (e.g., 2–20 mN/m,e.g., see Ref. 187), it is not clear that elastic membraneexpansion at constant area was being measured, since thecells were not preswollen to reduce excess membranesurface area. Interestingly, in the case of RBCs, it hasbeen reported that KA values vary by 640% with voltagechanges of 6200 mV (210, 211). Although the mechanismof this polarity-sensitive KA effect remains unknown, itmay reflect electric forces acting on packing of the highlyasymmetrical lipid bilayer of the RBC (458). In this case,it will be interesting to examine the voltage effects onRBCs that have lost their phospholipid asymmetry due tolipid scrambling (152, 458).

C. Membrane Bending/Curvature

The resistance of the bilayer to bending arises be-cause of differential expansion or compression of the twomonolayers within the bilayer and will depend on howtightly the two halves of the bilayer are coupled (i.e.,degree of interdigitation between the acyl chains). If thereis no coupling so that the two monolayers can rapidlyslide past one another, there will be little resistance tobending. Bending rigidity is also dependent on the spon-taneous curvature of the bilayer, which depends on thelipid composition and area of each monolayer (see sect.III) as well as the coupling of the CSK to the bilayer (91,431). In the specific case of a bilayer sealed tightly to thewalls of the patch pipette, the bending rigidity of themembrane patch should be increased because the attach-ment of the outer monolayer to the pipette walls willrestrict its movement relative to the inner monolayer.Although the estimated bending modulus of the bilayer(KB ;10219 N z m) indicates that bending resistance issignificantly less than resistance to expansion (108), it isthe bending rigidity that determines the shape of the lipidvesicle, its elastic response to membrane dimpling, andthe amplitude of thermally induced fluctuations in thevesicle (452a; 276a and references therein).

D. Membrane Extension/Shear

Above the phase transition temperature, the bilayerhas unrestricted internal fluidity and displays negligiblesurface shear rigidity so that it flows like a fluid in re-sponse to shear or extension. Below the phase transition,the shear rigidity increases along with hydrocarbon chainorder (108). Furthermore, bilayers that undergo lipidphase separations may be expected to show heterogene-ities in shear rigidity (138). Similarly, the existence of lipid

rafts in cell membranes (369a) may result in differentialrates of lipid flow in response to shear. However, moreimportant is the cortical CSK that by providing the fluidbilayer with a solid support significantly increases theshear rigidity of the cell membrane and thereby allowselastic responses to membrane extension (see sect. VII).For example, the human RBC has an elastic shear mod-ulus of ;1022 mN/m and can recover rapidly from largeextensions. However, after treatments {e.g., .48°C orincrease in intracellular Ca21 concentration ([Ca21]i)}that disrupt the cortical CSK, the shear modulus is sodiminished the RBC undergoes spontaneous fragmenta-tion (e.g., see Fig. 8 in Ref. 284; Ref. 152).

E. Viscous Properties and Dynamic Response

of Bilayer Vesicles

Elastic solids respond instantaneously to deforma-tion. However, a bilayer vesicle cannot be deformed in-stantaneously because of the inertia of surrounding watermovement and in specific cases possibly due to the vis-cous properties (e.g., bending viscosity) of the bilayeritself (24). For bilayer expansion, hydrodynamic dragrather than expansion viscosity is most likely rate limit-ing. For example, based on fluorescence polarizationmeasurements of the bilayer hydrocarbon interior, theexpansion viscosity (yA) of the bilayer has been estimatedto be 10210 N z s/m (i.e., comparable to a 10-Å-thick layerof olive oil, Ref. 108). In this case, the time constant forarea relaxation (tA) will be 1029 s according to the rela-tion tA 5 yA/KA and assuming a KA of 102 mN/m (108).How does this value compare with experimentally mea-sured kinetics of bilayer expansion? Clearly such mea-surements are limited by current methods. For example,even relatively sophisticated pressure-clamp techniquescan only give pressure steps with a rise time of ;1 ms(272, 273), and these cause membrane patch movements(i.e., expansion) and MG channel activation with millisec-ond latencies (448). In comparison, voltage steps (i.e.,with a rise time ,10 ms) applied to outer hair cells causemembrane patch expansions of ;1% with a t of 100 ms,while the underlying membrane charge movement has a tof 10 ms (119a). The discrepancy may reflect membranedamping by the fluid movement in the pipette and/or theviscous drag of the cortical CSK. The relative contribu-tions may be separated by measurements on blebbed (i.e.,CSK-disrupted) hair cells. In terms of shear relaxation, thenegligibly small shear modulus of the fluid bilayer shouldallow even faster intrinsic kinetics (i.e., ,1029 s). Theshear relaxation of cell membranes may be rate limited bythe larger shear viscosity of the underlying CSK. In thespecific case of bilayer bending, the interfacial drag be-tween the two monolayers may be sufficiently large (i.e.,107 N z s/m3) to slow membrane bending and shape recov-

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 689

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 7: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

ery (110). Evidence of slow bending kinetics may bereflected in the relaxation of fine membrane tethersdrawn from lipid vesicles (181) and possibly the adapta-tion of MG channel activity in liposome membranepatches (see sect. VIIIE).

In summary, the mechanical equilibrium of a lipidvesicle is established by the balance of external forcesapplied to the membrane (i.e., expansion and bending)opposed by the action of membrane tension and the bend-ing rigidity. In considering deformation-sensitive mem-brane parameters that might influence membrane proteinconformational changes, it is often overlooked that dila-tion of the elastic bilayer (i.e., increasing the area occu-pied by lipid molecules) should be accompanied by aproportional decrease in bilayer thickness (i.e., assumingvolume incompressibility). In addition, and as describedbelow, factors that affect spontaneous membrane curva-ture and bending rigidity may influence protein conforma-tional changes. In terms of bilayer dynamics, that rate ofmembrane deformations involving bilayer expansion andextension may be damped by the hydrodynamics of theadjacent water compartments. However, membranebending and relaxation may be rate limited by viscousdrag between the monolayers.

III. MECHANICAL DEFORMATION OF THE

BILAYER BY MEMBRANE PROTEIN

INSERTION

The next level of complexity that can be consideredis how insertion of membrane proteins may mechanicallydistort the bilayer and, in turn, how mechanically inducedbilayer distortions may influence protein conformationalchanges. The central idea to bilayer-protein interaction isthat the hydrophobic thickness of the bilayer immediately

adjacent to the membrane protein will tend to match thelength of the protein’s hydrophobic exterior (Fig. 1; Refs.290, 291). This may be expected to occur because anyuncompensated mismatch will add a high energetic costby exposing hydrophobic groups to water. Because pro-teins are relatively rigid, whereas lipid hydrocarbonchains are flexible, the condition of hydrophobic match-ing can be achieved by stretching, squashing, and/or tilt-ing of the lipid chains (172, 193). Recent direct evidencesupporting protein-induced changes in lipid organizationcomes from the demonstration that hydrophobic a-helicalpeptides, including gramicidin A, can change a bilayerinto a nonlamellar structure, with this transition depen-dent on the degree of hydrophobic mismatch (217). Fur-thermore, insertion of peptides, including alamethicin,into bilayers causes a concentration-dependent thinningof the bilayer as measured by lamellar X-ray diffraction(173).

One of the consequences of the hydrophobic mis-match idea is that proteins will tend to surround them-selves with lipids of matching size and shape so that themechanical strain on the bilayer will be minimized (82,111, 125, 140, 352). Furthermore, if the lipid compositionof the membrane can be altered, one might expect to seeshifts in protein conformational changes. Evidence sup-porting these ideas has come from studies of the effects offoreign phospholipids and lipophilic agents on MG chan-nel activities (53, 251, 253). For example, the oppositeeffects observed with some lipophilic agents on gramici-din and N-methyl-D-aspartate (NMDA) channel kineticsmay be explained by differences in lipid shape, as definedby the ratio of head group size (H) to the acyl tail area (T),on the localized curvature of the bilayer (53, 253). Lipidswith H 5 T will tend to favor neutral curvature, lipids withH . T will favor positive curvature, and lipids with T . H

FIG. 1. A membrane protein in changing conformationalso undergoes changes in hydrophobic mismatch with thelipid bilayer. A: positive hydrophobic mismatch in whichthe protein promotes local positive curvature in the lipidbilayer. B: a protein conformation that promotes local neg-ative curvature. C: neutral mismatch in which the proteindoes not distort the bilayer. [Modified from Fattal andBen-Shaul (111).]

690 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 8: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

will favor negative curvature. As described below, theeffects of lipids of different geometry on MG channelgating provided the initial clue that changes in membranethickness and/or local curvature may underlie one mech-anism of MG channel gating (136, 251). The importance ofthe surrounding lipids on protein function may also bereflected in the enzyme-regulated (i.e., floppase and trans-locase) asymmetrical distribution of phospholipid in eachmonolayer, which is lost with scramblase activation(458).

Figure 1 considers the specific case of a membraneprotein that has three stable conformational states, eachwith different types of hydrophobic mismatch with thebilayer (i.e., A positive, B negative, and C neutral). Inser-tion of foreign lipids or membrane thinning by altering theenergetic cost of membrane deformation should causeshifts in the distribution among these conformations (e.g.,membrane thinning would favor B). However, effectssuch as ligand binding, phosphorylation, or membranepolarization may cause shifts independent of hydrophobicmismatch. In this case, complex interactions may arisebetween mechanical and other forms of stimuli.

IV. SIMPLE PEPTIDES THAT FORM

MECHANICALLY GATED CHANNELS

Alamethicin and gramicidin are the best-character-ized membrane channels in terms of their biophysics (i.e,conductance and gating), structure-activity relations, andmodeling of open-closed channel conformations (6, 47,355, 426). Therefore, the demonstration that both chan-nels display mechanosensitivity in pure lipid bilayers,similar to prokaryotic MG channels, has provided theopportunity to analyze possible underlying molecularmechanisms in extremely well-defined simple systems.

A. Alamethicin

Alamethicin is a 20-amino acid peptide that formsvoltage-gated multi-conductance state channels in lipidbilayers (47, 355). The most commonly evoked model toexplain channel formation is a “barrel-stave” model inwhich each stave of the barrel is formed by a singlea-helical monomer. To explain the voltage sensitivity ofalamethicin, two main classes of mechanisms have beenproposed (47). In one, the channel exists as a preaggre-gate of subunits, and a voltage-dependent conformationalchange results in channel formation. In the other, ala-methicin exists predominantly as monomers and voltage-dependent insertion (or partitioning) with subsequent ag-gregation that leads to channel formation. Because evi-dence exists supporting both mechanisms (47), analysisof the mechanosensitivity of alamethicin channel gatingwas used in an attempt to discriminate between the two

mechanisms (313). Initially, Opsahl and Webb (313) dem-onstrated in patch-clamped bilayers (i.e., using the “tip-dip” method) that increased membrane tension increasedthe probability of the channel occupying higher conduc-tance states. Their quantitative analysis of the relationbetween the state of occupation and applied membranetension (t) was interpreted in terms of the work done inchannel opening W 5 t z DA, where the switch-ing between the adjacent states involves an increase inmembrane occupied area of the channel complex (DA) of;1.2 6 0.10 nm2. Based on these area changes, theyproposed that the mechanosensitivity of switching be-tween different conductance states could be explained bya model involving two tension-sensitive steps. Step 1 in-volved insertion or partitioning of an additional monomer(cross-sectional area ;0.8 nm2) into the channel bilayer.Step 2 involved the subsequent association of the insertedmonomer with the existing channel aggregate resulting inan increase in pore area (;0.4 nm2). In favoring thesubunit-recruitment model, Opsahl and Webb (313)pointed out that the observed area changes were mostlikely too large to be compatible with a model involvingrearrangement (expansion) of the monomers within afixed aggregate (115a). Furthermore, their observationthat the free energy difference between closed and openchannel conformations varied linearly with tension con-firmed that first-order area changes were responsible forthe mechanosensitivity, while second-order (quadratic)effects due to compliance changes in channel states werenot significant (69, 349). An increase in pore area of 0.4nm2 would give an increase in pore volume of ;1.3 nm3

assuming a pore length of 3 nm. This volume change is atleast of the same order predicted based on osmotic ex-periments that indicate channel switching involves up-take of up to 3 nm3 of water (421). Opsahl and Webb (313)demonstrated equivalent tension sensitivity for the threelowest adjacent conductance states in bilayers of fixedcomposition. It will be interesting to see if these samestates as well as the higher states generated in bilayerswith high curvature (214) display the same tension sensi-tivity as might be expected from their model. Finally,Opsahl and Webb (313) did not consider the free energycontribution on changing the environment of one face ofthe recruited alamethicin subunit from that low dielectricof the lipid bilayer to the high dielectric of the aqueouspore. As described later for a bacterial MG channel, thelarge free energy change associated with channel openingmay arise from the energetic cost of exposing pore-lining,hydrophobic residues to water in the open pore (443, seesect. VII).

For alamethicin, the subunit recruitment model re-mains intuitively attractive and appears consistent withmost of the experimental data. However, data on theeffects of lipid composition on alamethicin conductanceswitching (214) may indicate an alternative mechanism

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 691

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 9: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

related to hydrophobic coupling between the peptide andbilayer.

B. Gramicidin

Gramicidin is a 15-amino acid peptide that also formschannels in bilayers. Although a simple molecule, it ex-hibits two different folding motifs: a double helix and ahelical dimer. This polymorphism in structure is manifestin solution, in bilayers, and in the solid state (426). Al-though both folding motifs may form transmembranepores or channels, there is substantial evidence that themost frequently observed channel arises through thedimerization reaction between two nonconducting mono-mers that insert into each monolayer as a-helices (306).The length of the gramicidin dimer exterior is 2.2 nm,which can be compared with ;3 nm for a phopholipidbilayer and 3.2 nm for the long axis of alamethicin (115a).As illustrated in Figure 1B, because of gramicidin’s nega-tive mismatch, the bilayer hydrophobic core will tend tobe compressed (i.e., seen as a negative curvature) tomatch the channel’s hydrophobic exterior surface.

Several groups have previously reported apparent ten-sion sensitivity in the gating of gramicidin channels (101,297, 341). However, interpretation of these early studies wascomplicated because channels were studied in black lipidbilayers where changes in tension remain undefined and insome cases membrane thickness as well as tension wasaltered. In a more recent study, the pipette aspiration tech-nique was used to increase tension in bilayer vesicles (136,see sect. II). With this technique it was demonstrated thattension increased the rate of gramicidin channel formation(2- to 4-fold) and, to a lesser extent, the average channellifetime (136). Note this increase in lifetime was opposite toprevious reports that increased tension decreased gramici-din channel lifetime (101, 297, 341).

For gramicidin, there is little difference in the mem-brane area occupied by monomers and dimers that mightexplain the tension sensitivity of channel gating. Instead,Goulian et al. (136) proposed that increased tension, bycausing bilayer thinning (see sect. IIB), improved the hy-drophobic coupling between the bilayer and the dimer,thereby reducing the membrane deformation energy as-sociated with channel formation and increasing the acti-vation energy associated with channel dissociation. Thesmaller effect of tension on rates of channel dissociationcompared with rates of formation were shown to beconsistent with the smaller displacement (;0.1 nm) ofmonomers necessary for dissociation compared with thelarger mismatch (i.e., ;0.4 nm) between the dimer andbilayer. In contrast to the linear dependence of free en-ergy change of alamethicin channel switching with ten-sion (313), gramicidin displayed a quadratic dependence(136), possibly reflecting the elastic properties of mem-

brane deformation (252). According to Nielsen et al.(300), if a protein conformational change involves anincrease in the hydrophobic mismatch from 0.1 to 0.13nm, there will be a 10-fold shift in the equilibrium distri-bution of protein conformations due to these elastic prop-erties of the membrane.

In addition to the effects of membrane tension, a varietyof other experimental maneuvers have been shown to altergramicidin channel gating. These effects were also inter-preted as arising through changes in membrane deformationenergy. The treatments include incorporation of membranelipids, detergents, and cholesterol that tend to alter themembrane local curvature, thereby either stabilizing thechannel, in the case of agents that promote positive curva-ture, or destabilizing the channel, in the case of agents thatpromote negative curvature (251, 253). Similarly, conditionssuch as elevated Ca21 that decrease electrostatic energy ofthe bilayer by screening surface charge are proposed topromote channel dissociation by a mechanism involvingchanges in the membrane local curvature or thickness(254). Note that although the gramicidin channel isaxially symmetrical, its sensitivity to the sign of localcurvature of the adjacent lipid is expected in terms of themismatch model (see Fig. 1). One would also predict thatchanges in membrane voltage, by altering membranethickness through electroconstriction (5), would also al-ter membrane deformation energy and thereby gramicidinchannel gating. However, voltage-sensitive gating ofgramicidin has not been reported (6).

In summary, increased bilayer tension promotes dimer-ization of gramicidin and higher conductance states of ala-methicin. It may be that different mechanisms underlie themechanosensitivity of the two channels. For alamethicin, asubunit recruitment model has been favored over the fixedaggregate model. However, there is evidence that indicatesthe switching between alamethicin conductance states isdependent on lipid composition, similar to gramicidin ex-cept that the lipids that promote gramicidin dimerizationfavor lower conductance states of alamethicin (214, 251).Because other evidence indicates that membrane deforma-tion energy may be the major driving force for the alamethi-cin insertion transition (see Ref. 171), the hydrophobic mis-match model may also contribute to the mechanosensitivityof alamethicin. Finally, the existence of polymorphic struc-tures, ambiguities, and unresolved issues with such simplechannel-forming peptides is a useful reminder of the com-plications that lie ahead for modeling larger and more com-plicated channel proteins as described in the next section.

V. STRUCTURE OF PROKARYOTIC CELLS

The discovery of Archaea (formerly achaebacteria)has clearly shown that the old prokaryote/eukaryote di-chotomy is obsolete by largely oversimplifying diversity

692 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 10: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

of prokaryotic microbes in all its aspects including pro-karyotic cell envelopes (436). Nevertheless, from the per-spective of the MG channel gating mechanism, bacterialas well as archaeal cells may be considered the next levelof complexity after a bilayer vesicle. The cytoplasmicmembrane of bacteria is a fragile structure composed ofphospholipids and proteins enclosed by a cell wall (i.e.,outer membrane) that provides a strong, rigid structuralcomponent able to withstand the osmotic pressurescaused by the intracellular concentration of various os-moticants in the cell (423). Without the mechanical sup-port of the cell wall, a bacterial cell would behave as atiny dialysis bag that would take up water from the envi-ronment, swell, and burst. Bacterial cell walls (with theexception of the mycoplasma) have a structural compo-nent called peptidoglycan that provides the rigidity nec-essary to maintain cell integrity. The major buildingblocks of peptidoglycan are N-acetylglucosamine and N-acetylmuramic acid that are unique to bacterial cells. Onthe basis of the Gram stain, a differential staining tech-nique invented in 1884 by Christian Gram, bacteria can bedivided into two large groups: Gram positive and Gramnegative, which differ in the peptidoglycan content oftheir cell wall (i.e., ;5 times larger in Gram-positive cells)(Fig. 2A) than in Gram-negative cells (see Fig. 2B). Inaddition, Gram-negative bacteria have a second chemi-cally distinct outer membrane attached to the peptidogly-can layer on its external side. The lipid bilayer of the outermembrane is made of phospholipids in the inner leaflet ofthe bilayer and lipopolysaccharides (LPS) in its outermonolayer. Gram-positive cells lack the second outermembrane (300b).

The situation with Archaea is more complicated,likely reflecting the large diversity of extreme habitats towhich Archaea have adapted (318). Archaea lack pepti-doglycan, which was one of the features used originally todefine prokaryotes. Two types of archaeal cells, which todate were found to harbor MG channels in their cellmembranes, may illustrate this. The cell wall in halophilicarchaea, such as Haloferax volcanii, whose cell mem-brane was the first to be examined for the presence of MGchannels (239), consists of the S layer formed by a hex-agonal arrangement of a glycoprotein (238). The proteinappears to be anchored in the cytoplasmic membrane bya hydrophobic stretch found near the COOH terminus ofthe H. volcanii glycoprotein sequence (392). Thermo-

plasma volcanii is the second archaeon found to haveMG channels in its cell membrane (220). This thermo-philic archaeon has no cell wall, but instead contains anouter meshlike lattice of elements similar to nuclearlamins that is reminiscent of the CSK in animal cells(179). In addition, its cell membrane contains ether lipidsbased on 40-carbon, isopranoid-branched diglycerol tethra-ethers (361). In general, lipid bilayers of cell membranesof all archaea consist of diphytanylglycerol-diether or-tetraether or both (92). It is worth mentioning that nei-ther bacteria nor archaea have a CSK in the eukaryoticsense. To describe the structural diversity of prokaryoticcell envelopes in its entirety would go over the scope ofthis review. What matters from the perspective of pro-karyotic MG channels is that despite this diversity, it isalways the lipid bilayer that is the tension-bearing ele-ment. The cell wall functions as a parallel viscoelasticstructure that constrains the bilayer from excessive dila-

FIG. 2. Cell envelope of Gram-positive (A) and Gram-negative bacteria (B). [From Volk (423).]

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 693

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 11: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

tion and in this way may reduce MG channel activation(41, 265, 267). The practical implication of the MG chan-nel’s sensitivity to bilayer tension is that, as discussedlater, they retain their mechanosensitivity when reconsti-tuted in liposomes (239, 388). The physiological implica-tion is that they regulate cellular turgor by responding tobilayer stretch caused by osmotic swelling (2, 244; seesect. VIJ).

VI. MECHANICALLY GATED CHANNELS

IN BACTERIA AND ARCHAEA

Since their initial discovery in giant spheroplasts ofE. coli (267), the existence of MG channels in Gram-negative and Gram-positive bacteria has been amply doc-umented (17, 18, 26, 27, 268, 387, 455, 456). The best-characterized are the MG channels of the Gram-negativebacterium E. coli, which have been studied by the patch-clamp technique in various giant spheroplasts (41, 76, 229,266, 267) and in reconstituted membrane fractions fusedwith liposomes (18, 87). Based on their conductance andsensitivity to applied pressure, three types of mechano-sensitive channels (Msc) can be distinguished: MscM (Mfor mini), MscS (S for small), and MscL (L for large) (17).The higher the conductance, the higher their activationpressure as evident in both in situ or in vitro recordings.Also, in Gram-positive S. faecalis and B. subtilis (397,398, 455), membrane stretch results in the activation of awhole array of conductances, ranging from 100 pS to upto several nS.

A. Identification of the MscL Gene/Protein

The property of E. coli MG channels of being acti-vated by mechanical force transmitted via the lipid bilayer(266) allowed for detergent solubilization and functionalreconstitution of these channels into artificial liposomesamenable to patch clamp (388). Furthermore, it allowedapplication of a unique strategy of fractionation of E. coli

membrane constituents by column chromatography andfunctional examination of the individual fractions for MGchannel activity by patch clamp. This unusual approachled to the identification of a membrane protein underlyingthe activity of MscL (386, 388). The MscL protein waspartially sequenced, which enabled the cloning of thecorresponding mscL gene (384). The expression of themscL gene alone in a heterologous and in an in vitrotranscription/translation system demonstrated that themscL gene alone was necessary and sufficient for MscLactivity.

B. Structure of MscL

The mscL gene encodes a 15-kDa protein comprising136 amino acid residues corresponding roughly to the

17-kDa protein band on a SDS-PAGE originally identifiedas the MscL protein (386, 388). Manipulation of the mscL

gene by recombinant DNA techniques enabled productionof the MscL protein on a preparative scale, which inaddition to functional studies allowed also for higherorder structural studies of MscL. Two methods have beenused for simple purification of the MscL recombinantproteins. The first method employs the glutathione S-transferase (GST) protein fusion technique to expressMscL attached to a cleavable GST domain (168), whereasthe second method uses a 6xHis polyhistidine tag forpurification of the recombinant MscL protein on a Ni1-NTA column (27). Both methods yielded functional MGchannels in patch-clamp experiments. The 6xHis-taggedMscL protein was used for secondary structure analysisby employing both transmission Fourier transform infra-red spectroscopy (FTIR) and circular dichroism (CD) (7).The MscL secondary structure includes two a-helicaltransmembrane domains (M1 and M2) connected by aperiplasmic loop (Fig. 3, A and B). Thus MscL belongs tothe family of structurally related ion channels with twomembrane segments that include the epithelial sodiumchannel (ENaC), the inward-rectifier potassium channel(Kir), and the ATP-gated (P2X) cation channel (37, 302).Using the PhoA fusion technique, Blount et al. (27) coulddemonstrate that the NH2 as well as the COOH terminusof MscL were located within the cytoplasm.

The higher order structure of MscL was first assessedin cross-linking studies. In one study, the multimericstructure of MscL was examined either by cross-linkingthe protein in situ and visualizing the cross-linked prod-ucts by Western blotting using MscL antibodies against aCOOH-terminal peptide or by cross-linking purified 6xHis-tagged MscL in which purified proteins were cross-linked.Both approaches indicated that MscL might form homo-hexameric channels (26, 27). In another study, variouscross-linkers were applied directly to bacterial cells inwhich the MscL protein was radiolabeled by [35S]methi-onine (170) and left open the possibility that MscL formsmultimers of higher order other than a hexamer. How-ever, recent reevaluation of cross-linking experiments us-ing various cross-linking reagents indicates that MscL is ahomopentamer (389, 270a). Preparative scale productionof milligram amounts of the MscL protein allowed em-ployment of electron crystallography to study the struc-tural assembly of MscL. The crystallographic analysis oftwo-dimensional MscL crystals at 1.5-nm resolution indi-cated that MscL forms homohexameric channels in lipidbilayers (353). However, the three-dimensional X-raycrystallographic study by Chang et al. (60) solved theoligomeric structure of the MscL homolog from Mycobac-

terium tuberculosis (Tb-MscL) to 0.35-nm resolution andindicated this homolog is organized as a homopentamer(Fig. 3C) (for discussion of homohexamer versus homo-pentamer, see sect. VID).

694 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 12: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

C. Conductive Properties of MscL

MscL forms nonselective ion channels of a very largeconductance. The absence of any cation/anion selectivity

or saturation in channels conductance up to 2 M KClindicates MscL forms a large water-filled pore (75, 390).The reported values for MscL conductance range from 0.9to 4.4 nS. Some of the differences may reflect the different

FIG. 3. Large mechanosensitive chan-nel (MscL) membrane-spanning models. A:membrane topology of E. coli MscL mono-mer proposed by Sukharev et al. (387). B:drawing of the MscL pentamer (left) andmonomer (right) from M. tuberculosis ac-cording to the 3-dimensional structuralmodel proposed by Chang et al. (60). [Mod-ified from Oakley et al. (305).]

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 695

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 13: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

ionic composition of the recording solutions used forpatch-clamp experiments. However, the variations inMscL conductance were also observed in recording solu-tions of the same or similar ionic composition (27, 75, 226,283, 384). The two extreme conductance values reportedfor MscL (in a recording solution containing 200 mM KClplus 40 mM MgCl2 and 5 or 10 mM HEPES, pH 6.0 or 7.0)are 2.5 and 3.8 nS, although the exact reason for thesediffferences remains unclear. As discussed in the nextsection, one explanation is that MscL can form more thatone type of homomultimeric channel, perhaps analogousto the channel and pore forms of gramicidin (see Ref.426).

D. Is MscL a Hexamer or a Pentamer?

It is important to establish the multimeric organiza-tion of MscL because, as previously discussed for ala-methicin and gramicidin, this feature can have importantimplications for gating mechanisms. As indicated above,homohexameric (Eco-MscL) and homopentameric (Tb-MscL) structures have been reported from crystallo-graphic analysis (60, 353). In addition, several reportsbased on cross-linking experiments contributed to theMscL structural controversy by showing that MscL mayform hexamers (27) as well as pentamers (389). More-over, tandems of two MscL monomers expressed as asingle dimer protein formed functional channels in giantE. coli spheroplasts, indicating that the functional chan-nels can be made from an even number of monomers(27). On this last point, it would be interesting to obtaintwo- or three-dimensional crystals of channels made ofthe MscL protein dimers.

Is MscL a hexamer or a pentamer? Probably it is safeto say that MscL forms pentameric channels taking intoaccount the resolution of 0.35 nm at which the Tb-MscLthree-dimensional structure was obtained (60). Moreover,at the resolution of 1.5 nm for the two-dimensional Eco-MscL crystals, a pentameric MscL structure is just aslikely to be judged as a hexamer (353). A resolution of atleast 1.0 nm would be required to conclusively distinguishbetween hexameric and pentameric structures of thetwo-dimensional MscL crystals (J.-L. Rigaud and J-.J.Lacaperre, personal communication). Nonetheless, it isworthwhile pointing out several peculiar details of theTb-MscL structure and the experimental conditions atwhich the three-dimensional structure was obtained. Ac-cording to the Protein Data Base (PDB) summary report,the Matthews coefficient (Vm) of 5.98 for MscL is highcompared with known structures of other proteins in thedata base, indicating that the MscL structure is an outlier.The coefficient is usually in the range between 1.5 and 4.0for tightly and loosely packed proteins, respectively. Also,the Ramachandran Z-score of 25.671 for MscL appearsvery low and suggests a very unusual backbone confor-

mation of Tb-MscL that is probably due to some localuncertainties in the backbone side chain conformation(A. Oakley, personal communication). In addition, theTb-MscL crystallization experiments were performed at avery low pH between 3.6 and 3.8. This pH is below the pKa

value of 4.25 of the glutamic acid residue E104 in theCOOH-terminal domain of Tb-MscL, which probably hadto be protonated to stabilize the MscL pentameric struc-ture. At higher pH one might expect the five E104 residuesof the pentamer to become negatively charged causingdestabilization of the multimeric structure, unless there ischarge compensation by neighboring basic residues orcations present in the surrounding medium (305). Changein pH is known to modulate significantly the pressuresensitivity of MG channels (as discussed later) and alsoinduce structural changes in proteins (42). Therefore, atpH ;7 the MscL tertiary structure might differ from thereported pentameric one.

Obviously, it is not possible to compare directly thetwo- or three-dimensional structural data of MscL withthe electrophysiological conductance measurements,since the crystallographic structures are depicting closedchannels, whereas electrophysiology can detect onlyopen conducting channels. For example, the three-dimen-sional crystal structures of the Tb-MscL pentamer re-vealed a transmembrane pore that narrows from 3.6 to 0.4nm in diameter in going from the extracellular to thecytoplasmic surface (60) (Fig. 3B). In contrast, patch-clamp permeation studies indicated the open MscL has achannel pore of ;4.0 nm diameter (75). Consequently, amajor structural rearrangement occurs when MscL“switches” between the two conformations (14). To fullyunderstand the process will most likely require the crystalstructure of the open MscL.

E. Origin of MscL Mechanosensitivity

Bacterial and archaeal MG channels provide a cleardemonstration that microbial MG channels can sensemembrane tension directly. The tension develops in thelipid bilayer alone and directly gates these channels asdescribed by the bilayer model. Since it was first pro-posed in relation to bacterial MG channels (261, 266), thebilayer model has become, along with the tethered model,one of the two mechanisms used to describe MG channelgating (see sect. VIIIC). In the case of MscL, the validity ofthe bilayer model has been amply documented (28, 168,384, 387, 388, 390). Interestingly, the well-studied stretch-activated cation (SA-CAT) channel endogenous to Xeno-

pus oocytes also appears to be gated by bilayer tension(448).

As the amount of negative pressure (i.e., suction)applied to a patch pipette increases, the MscL channelopen probability also increases (Fig. 4A). Activation ofMscL by pressure can be described by a Boltzmann dis-

696 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 14: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

tribution function for the channel open probability (Po)(Fig. 4B)

Po /~1 2 Po! 5 exp@a~p 2 p1/2!# (3)

where p is the applied negative pressure, p1/2 is the suc-tion at which the channel is open half the time, and a isthe slope of the plot ln [Po/(1 2 Po)]. Because the Boltz-

mann function is very often used to characterize themechanosensitivity of MG channels in the literature, wewill discuss the meaning of the terms p1/2 and a, first inrelation to MscL and then other MG channels. The follow-ing exercise should be useful to laboratories studying MGchannels that do not routinely image membrane patchmovements and calculate the membrane tension changes.

As shown previously (147, 374, 375, 390, 448), mostMG channels respond to mechanical forces along theplane of the cell membrane (membrane tension), and notpressure perpendicular to it. According to the model ofHoward et al. (190), the free energy (DG) is a linearfunction of membrane tension t

DG 5 t z DA 2 DGo (4)

where DGo is the difference in free energy between theclosed and open conformations of the channel in theabsence of the externally applied membrane tension andDA is the assumed difference in membrane area occupiedby an open and closed channel at a given membranetension, whereas tDA is the work required to keep an MGchannel open by external mechanical force at the openprobability of 0 , Po , 1. Consequently, in this model, thesize of DA is considered the sole parameter determiningthe mechanosenitivity of channel gating (i.e., the largerDA the more sensitive the channel). Evidence indicatesthat for non-MG channels, like the ACh receptor channel(417) and specific voltage-gated K1 channels (56, 130),the movement of transmembrane helices is quite small. Incontrast, large movements of MscL are needed to accountfor the large pore formation and the steep tension-re-sponse relation of MscL (390).

Using the expression in Equation 4, the Boltzmannfunction can be rewritten as

Po /~1 2 Po! 5 exp@~t z DA 2 DGo!/kT# (5)

In the case of MscL, which is activated by membranetension near the lytic strength of the bilayer (i.e., maximalcurvature), it has been demonstrated that in that activat-ing pressure range and assuming an elastic membranewith a KA 5 102 mN/m (see Refs. 373, 390), the tensionwill be nearly proportional to the pressure (i.e., there willbe little change in curvature). Using a version of Laplace’slaw

t 2 t1/2 5 ~p 2 p1/2!~r/2! (6)

where r is the radius of curvature of the liposome mem-brane patch under external negative pressure p applied tothe patch pipette. Thus, under these conditions, p1/2 andt1/2 , the characteristic negative pressure and membranetension, respectively, at which the channel is open 50% ofthe time can be assumed to remain constant for a mem-

FIG. 4. Activation of MscL by negative pressure applied to thepatch-clamp pipette. A: current traces of the reconstituted recombinant6xHis-MscL recorded from an isolated liposome patch at various nega-tive pressures applied to a patch pipette. Trace is a ;100-s-long record-ing at a pipette potential of 130 mV. C denotes the closed state, and On

denotes the open state of n number of channels. Nine channels wereactive in the patch shown. B: curve is a Boltzmann distribution functioncalculated for a single MscL shown in A. The function relates negativepressure (suction) and open channel probability (Po). Boltzmann distri-bution for a single MS channel has a form Po/(1 2 Po) 5 exp [a(p 2 p1/2],where p is the applied negative pressure, p1/2 is the suction at which thechannel is open 50% of the time (Po 5 0.5), and a is the slope of the plotln [Po/(1 2 Po)]. The Boltzmann function is very often described in theMS channel literature in terms of p1/2 and a. For the 6xHis-MscL in theparticular recording shown in A, p1/2 was estimated to be ;77 mmHg,whereas the sensitivity to pressure 1/a was ;5 mmHg for e-fold changein Po, which corresponds to DGo ' 15kT (Eq. 5). Overall, DGo for therecombinant 6xHis-MscL was estimated to be 17.0 6 1.5kT (SE; n 5 4)(A. Kloda and B. Martinac, unpublished data).

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 697

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 15: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

brane patch during an experiment. Because the free en-ergy difference DG (Eq. 4) is equal to zero when the openprobability Po 5 0.5 (P 5 p1/2 and t 5 t1/2), consequentlyt1/2 5 DG0 /DA and p1/2 5 2DG0 /rDA, whereas a 5 rDA/2kT. Thus it follows that at least in liposome or blebexperiments in which parallel CSK elements are not vari-ables in supporting bilayer tension, the slope term willprovide a direct estimate of molecular rearrangements ofa MG channel, as long as the diameter and shape of patchpipettes remain nearly constant throughout the experi-ments. A convenient expression can be obtained by mul-tiplying p1/2 and a

GMGC 5 p1/2 z a 5 DGo /kT (7)

showing that the product GMGC is independent of thepatch geometry. GMGC provides a direct estimate of theenergy difference DGo between the closed and open stateof an MS channel, and thus it presents the very charac-teristic of any type of MG channel reconstituted into adefined lipid membrane. For example, according to sev-eral reports, a negative pressure (p1/2) ranging between;7.7 and 10.3 kN/m2 (i.e., ;58 and 77 mmHg) is requiredto activate MscL 50% of the time in liposome patches (3,141, 246a, 220b), whereas the sensitivity to pressure (1/a)of the MscL channels was found to vary between 0.6 and0.8 kN/m2 (i.e., 4.5 and 6.0 mmHg) with a ;1.67 and 1.25(kN/m2)21 (i.e., 0.22 and 0.17 mmHg21), respectively (3,75, 168, 220b). It follows that on average GMscL ;14, andconsequently, the average energy required for openingMscL is DGo ;14kT. This value is in reasonable agreementwith an independent estimate of ;18.6kT obtained forDGo from Po-membrane tension curves for MscL (390).Clearly, the interpretation of Equation 7 becomes morecomplicated for cell membranes in which the associatedcortical CSK (i.e., an extrinsic factor) can alter the tensionseen by the bilayer and thereby change the shape of theBoltzmann independent of the intrinsic properties of thechannel protein (see Ref. 447 and sect. VII).

F. Extrinsic and Intrinsic Factors That Affect

MscL and Other MG Channels

The apparent sensitivity of MG channels to mem-brane tension may be experimentally altered by the fol-lowing treatments.

1) Lysozyme, which disrupts the peptidoglycan of thebacterial cell wall, irreversibly increases the pressuresensitivity of MG channels in E. coli giant spheroplasts(41, 268).

2) Cytochalasin, which disrupts actin microfilaments,increases the pressure sensitivity of MG channels in chickmuscle and snail neurons (142, 370).

3) Mechanical decoupling of the CSK from theplasma membrane (i.e., blebbing) in Xenopus oocytesdecreases the MG channel mechanosensitivity (154,160, 447).

4) Amphipatic (amphiphilic) compounds can in-crease mechanosensitivity of bacterial (266) and eukary-otic MG channels (323, 372).

5) Deletion, substitution, or single-site mutations caneither increase or decrease the native MscL mechanosen-sitivity (25, 28, 170, 316, 317).

6) Changes in pH can affect MG channel sensitivity inboth ways, with alkaline pH increasing (143; Martinac,unpublished data) and acidic pH decreasing the channel’spressure sensitivity (28; Martinac, unpublished data).

These examples illustrate how the apparent mech-anosensitivity of a channel may be altered by extrinsic,intrinsic, or possibly by a combination of mechanisms.Specifically, examples 1–4 may involve alteration in theway mechanical force is delivered to the channel (i.e., byalteration of CSK or bilayer properties) without alteringthe intrinsic properties of the channel protein (see alsoadaptation, sect. VIIIE), example 5 may involve changes inthe protein’s intrinsic mechanosensitivity, and example 6

may result from changes in both extrinsic (bilayer) andintrinsic (protein) properties.

In the specific case of MscL reconstituted into lipo-somes, the following example illustrates how, for a MGchannel activated by lipid bilayer tension, the measuredchanges in p1/2 and a can provide an estimate of thechannel’s intrinsic physical properties and identify thecontribution of specific structural domains to the gatingmechanism. Specifically, Ajouz et al. (3) studied the ef-fects of various proteases on pressure-dependent gatingof MscL in an attempt to identify molecular domains ofMscL responsible for mechanosensitivity. They demon-strated that both parameters, a and p1/2, were dramati-cally affected by protease treatment, with the slope terma significantly increased and p1/2 decreased (i.e., mech-anosensitivity increased). The quantitative changes in p1/2

and a caused by protease treatment were such that GMscL

(Eq. 7) ranged between 12 and 18 (i.e., DG0 ;12–18kT)compared with ;14 before protease treatment. BecauseGMscL and DG0 were hardly affected by cutting the ex-tramembraneous domains, it was concluded that neitherthe cytoplasmic termini (i.e., the NH2 and COOH termini)nor the S2-S3 periplasmic loop (see Fig. 3A) contributecritically to the mechanical gating of MscL. Instead, it wasproposed that these extramembranous domains resist themovement of the transmembrane helices that underlie thecritical event in mechanical gating of MscL. Note thisresult disagrees with the electromechanical model de-scribed below that proposes the NH2 termini gate MscL(141).

698 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 16: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

G. Where Is the MscL Gate?

Based on a mutagenesis study in which Gly-22 (E. coli

MscL) was changed to all other 19 amino acid residues,Yoshimura et al. (443) concluded that by analogy with Ala-20of Tb-MscL, Gly-22 should surround the Eco-MscL gate.

Indeed, the Tb-MscL has been found to be extremelystiff when expressed in E. coli and examined by the patchclamp. Specifically, Moe et al. (283a) found that the chan-nel required twice the membrane tension needed to gatethe Eco-MscL. In addition, their study showed that aminoacid substitutions at the neighboring residue V21 hadsevere effects on the channel mechanosensitivity, indi-cating that besides G22, V21 also participates in the en-ergy barrier of MscL gating (283a). By using the programVOIDOO/FLOOD (222), Oakley et al. (305) estimated thatin the closed conformation MscL could be filled withwater molecules to a point 2.4 nm below the periplasmicsurface of the channel. Also, the second cavity at thecytoplasmic face of the channel was found to be wateraccessible. Consequently, most of the inside of the closedMscL contains water except for a hydophobic stretch of0.8 nm that includes the Gly-22 residue and forms a water-tight occlusion. Thus the hydrophobic channel gate isquite thin compared with the 3- to 3.5-nm-thick membranebilayer (60). If we accept that in the open state of MscL,the hydrophobic gate becomes exposed to water, as orig-inally proposed by Cruickshank et al. (75) and later reit-erated by Yoshimura et al. (443), then can the value of18.6kT forDG0 be explained solely by this process. Takinginto account that 17 mJ/m2 is necessary to transfer ahydrophobic protein from an organic solvent into anaqueous environment (66), it follows that 18.6kT sufficesto expose a hydrophobic area of 4.42 nm2 to water(18.6kT 5 7.521 3 10220 J). This area is relatively smallcompared with the total membrane-associated area ofMscL (;140 nm2) and roughly corresponds to a half sur-face of 5 a-helices, each having a diameter 2r 5 0.68 nm(396), with a height l 5 0.8 nm corresponding to the heightof the hydrophobic gate (i.e., 2rpl ;5/2 5 4.27 nm2).Consequently, this result indicates that most, if not all, ofthe channel opening energy of 18.6kT is used to expose arelatively small buried hydrophobic surface of the TM1helices to the ionic environment of the bulk solutionsurrounding the channel. This may explain why Ajouz etal. (3) never observed a permanently open MscL and whythe proteases did not affect the gating mechanism of thechannel.

H. Mutagenesis Studies

Several studies have used recombinant DNA tech-niques (25, 27, 28, 169, 283) or a genetic approach (316) todissect the MscL molecule in an attempt to identify the

essential functional domains responsible for MscL mech-anosensitivity. Blount et al. (27) showed that a shortdeletion of 3 amino acids from the NH2 terminus (14residues long) and the larger deletion of 27 amino acidsfrom the COOH-terminal region (36 residues long) hadlittle or no effect on the channel properties. However,when 33 COOH-terminal residues were deleted that in-cluded a charged cluster RKKEEP of 6 more residues, thechannel activity was abolished (387). Furthermore, thesestudies demonstrated that single residue substitutions inthe first transmembrane domain TM1 of MscL, in partic-ular, lysine K31 (28) and glycine G22 (316, 443), led tomajor changes in the MscL activities that could be corre-lated with phenotypes with major impairments in re-sponse to osmotic stress. In addition to its structuralconservation throughout bacterial species, these studiesemphasized the importance of TM1 in the MS propertiesof MscL.

Hase et al. (169) dissected MscL systematically byintroducing deletions, additions, or large substitutionsinto the extramembranous domains. They reconstitutedthe mutated proteins into liposomes and examined theirfunction. Consistent with previous findings (27, 387),large NH2-terminal deletions or changes to the NH2-termi-nal amino acid sequence affecting the overall charge ofthe NH2 terminus were poorly tolerated and resulted inchannels that exhibited altered pressure sensitivity andgating. Hase et al. (169) also confirmed the result ofBlount et al. (28) showing that deletion of the chargedcluster RKKEE in the COOH-terminal domain abolishedchannel activity, possibly indicating MscL was no longeractivatable by sublytic membrane tensions.

I. Models of MscL Mechanosensitivity

There is nothing in the three-dimensional structure ofthe closed conformation of Tb-MscL that unambiguouslypoints toward a mechanism of how mechanical forcegates the channel. However, several models based onstructure-function relations are discussed below.

1. Multimerization model

One of the first models used to explain MscL gatingwas a tension-sensitive recruitment of MscL monomersinto a pore-forming channel multimer (i.e., analogous tothe alamethicin recruitment model in sect. IVA). This ideawas based on pressure-response relations seen in manyliposomes, as well as native membrane patches, in whichchannel activity (single-channel open probability andnumber of channels) continued to increase without satu-ration until the patch ruptured (168, but see Ref. 390).Furthermore, this “multimerization model” seemed to besupported by in situ cross-linking studies that indicatedMscL existed predominantly as a monomer (of ;15 kDa)

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 699

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 17: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

in the E. coli cell envelope (170) and the observation thatfunctional channels could be formed from reconstitutedMscL monomers electroeluted from SDS gels (Saint andMartinac, unpublished data). However, in contradiction ofthe model, it was found that a much larger functionalchannel complex (;70–100 kDa) could be identified ingel filtration experiments (27) and that covalently linkedtandem subunits produced the same size complex andMscL activity (27). We now know that the tandem resultmay occur because either the extra subunit does notparticipate in the functional pentameric channel or it isincorporated into a neighboring pentameric channel (e.g.,see Fig. 7 in Ref. 389). Another result considered incon-sistent with multimerization was the rapid opening tran-sitions (i.e., ;0.2 ms) of single MscL channels that wereindependent of MscL concentration and applied steady-state pressure (383). It was argued that fast transitionsexcluded diffusion-limited assembly of MscL monomersinto functional multimers. However, it remains possiblethat there is an initial tension-triggered assembly of dis-persed monomers into a multimer that is then gated co-operatively under steady pressure. To exclude this possi-bility, the latency for MscL opening in response to fastpressure steps (i.e., ;1 ms) should be measured as afunction of MscL concentration and step size.

2. Open-channel model

Based on a functional study in which large organicmolecules were used to estimate the size of the MscLpore, an open-channel model of MscL was proposed thatenvisages 12 membrane-spanning a-helices of a hexa-meric channel lining the pore of ;4.0 nm diameter (75).Since the three-dimensional Tb-MscL structure indicatesthat MscL may be a pentamer rather than a hexamer (60),10 transmembrane helices may line the pore of a smallerdiameter channel that would have to be shorter to main-tain the observed large conductance. The crystal struc-ture of the closed MscL (Fig. 3B, Ref. 60) and the perme-ability properties of the open MscL indicate that MscLmust undergo a large conformational change. For exam-ple, in order for MscL to open and let through moleculesthe size of thioredoxin (radius ;3.5 nm) (2), both TM1and TM2 helices would need to shift radially from thefivefold axis of symmetry so that the pore size at thecytoplasmic end would match the MscL diameter of ;3.6nm at the periplasmic end of the channel (60) (Fig. 3B).Such a radial shift of helices corresponds to a substantialrearrangement of the channel structure that could ac-count for the large DA of ; 6 nm2 calculated from the Po

versus tension relations (390) [the difference in area cal-culated from the radii of the pore at the periplasmic rp 51.8 nm and cytoplasmic side rc 5 0.2 nm of the pentam-eric channel DA 5 p (rp

2 2 rc2) ;10 nm2]. The same study

calculated the free energy of transition from the closed to

the open states to be 46.3 kJ/mol. This is comparable tothe folding free energy of ;54 kJ/mol required to com-pletely unfold the 136 residues of the MscL monomer(;0.4 kJ z mol21 z residue21) (305). Therefore, MscLmust experience a radical structural change in channelopening that requires large molecular energies of theorder expected during protein denaturation unless someunknown low-energy transition pathway exists betweenthe two channel states.

3. Electromechanical coupling model

By recognizing the overall importance of chargedresidues, in particular in the NH2- and COOH-terminaldomain, Gu et al. (141) proposed an electromechanicalcoupling (EMC) model for gating MscL. In essence, themodel proposes that the pore region of MscL is present inthe closed channel and that the NH2 termini of the fivesubunits interact electrostatically with one another andpore regions to gate the channel (i.e., 5 swinging gates).Membrane tension by altering the tilt of the transmem-brane helices is proposed to interfere with the balance ofcoulombic forces between the various domains of thechannel molecule. The actual channel gating is accom-plished by the flexible NH2 termini of each subunit pivot-ing around glycine G14 (i.e., located at the bilayer-waterinterface between the NH2 terminus and the TM1 trans-membrane helix) to occlude the pore. Although the EMCmodel could account for some differences in mechano-sensitivity of several MscL mutant channels (141), in itsoriginal form and with its underlying assumptions it suf-fers basic limitations. For example, one assumption wasthe fractional increase in pore area (DA) and patch areaduring channel gating were the same (i.e., equal proteinand patch elasticity) so that DA was calculated (using Eq.

4) as only 0.78 nm2 with a free energy difference (DGo) of2kT. However, direct estimates DA and DGo from Po-tension curves give values of 6.5 nm2 and 18.6kT, respec-tively (390). The model also did not take into accountelectrolyte screening of charged residues that would tendto negate any electrostatic interactions between the NH2

termini and the pore over the necessary long distances(i.e., ;1 nm). Furthermore, the model predicts that MscLgating should be strongly dependent on ionic strength andmembrane potential. However, no significant differencein MscL gating is observed in salt solutions ranging from0.05 to 1 M or as a function of voltage (390). Furthermore,despite EMC model expectations, MscL function is notcritically dependent on the length or charge of the NH2

terminal, since the removal of 3, substitution of 8, oraddition of 20 new residues to this domain does notsignificantly alter MscL gating (26, 28, 169). Similarly,proteolytic cleavage of either the NH2 and/or COOH ter-mini, rather than producing a permanently open channelas predicted, increases channel mechanosensitivity with-

700 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 18: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

out altering channel conductance (3). Finally, the three-dimensional crystal structure of the closed state of Tb-MscL indicates the channel gate involves a narrow poreregion (;0.2 nm in diameter) near the cytoplasmic endsof the TM1 domains (60). Despite these limitations, thereare specific MscL structure-function relations thatseemed to support the EMC model and have yet to beexplained by other models. For example, the importanceof charged residues in MscL gating has been well demon-strated by the deletion of the charged cluster RKKEE inthe COOH-terminal domain that renders the D104 MscLmutant nonfunctional within the range of sublytic mem-brane tensions corresponding to pressures of 200 mmHg.The EMC model predicts the D104 mutant should only befunctional at negative pressures of ;300 mmHg or more(141). Also, the finding that very low pH (i.e., between 3.6and 3.8) was a prerequisite for the successful three-di-mensional crystallization of Tb-MscL (60) supports theimportance of charged residues. In particular, the glu-tamic acid residue E104 with a pKa 4.25 may need to beprotonated to stabilize the closed configuration of theMscL pentameric structure (305). One might expect thatat a higher pH the electrostatic repulsion between theE104 negatively charged residues would destabilize theclosed MscL structure, thus contributing in some way tothe channel opening. Nevertheless, evidence now indi-cates that rather than being involved in directly gatingMscL, the NH2 termini may modulate the MscL mechano-sensitivity, possibly by resisiting the movement of thetransmembrane helices during channel opening as well asinterfering with ion permeation to produce multiconduc-tance state behavior (3). Therefore, the EMC model aswell as another recently proposed molecular model forgating transitions of MscL (387a) are most likely incorrectby assuming that the NH2 termini function as the channelgate. However, the mobility of the NH2 termini remains avalid assumption in both models. It may also turn out thatthe EMC-type model and its assumptions (i.e., small porearea change and a mechanoelectrical “triggered” swinginggate) will be applicable to eukaryotic MG channels.

4. Five-state kinetic model

Based on an analysis of multiple conducting states ofthe single MscL currents, a five-state linear kinetic modelwas proposed (390). The linear scheme of four conduct-ing and one closed state C1 7 S2 7 S3 7 S4 7 O5 withsequentially increasing substate conductances was foundto be the simplest model to estimate the transition rateconstants kij 5 k0exp(a/kT) and their dependence onmechanical force. In the expression for kij, k0 is the com-ponent independent of membrane tension, whereas a de-termines the tension sensitivity. According to the model,k12 between the closed C1 and the first subconductingstate S2 is the only tension-dependent rate constant lim-

iting MscL opening. The large energy of 18.6kT (46.3kJ/mol) calculated for DG0 accounts for MscL being al-most always shut in the resting membrane. (Energy dif-ference of 18.6kT can be calculated using Eq. 7 by divid-ing the tension of 11.8 mN/m required for the MscL openprobability of 50% with the maximal slope sensitivity of0.63 mN z m21 z e-fold21 obtained from the Boltzmanndistribution, Ref. 390). However, the model is probablyincorrect by describing the MscL gating in terms of onlyfive conducting states. According to the molecular dynam-ics simulations of protein conformations, one would ex-pect the channel to exist in many relatively stable confor-mations on a very short time scale on the order of micro-to nanoseconds that would be well below the resolutionof patch-clamp experiments. Thus only relatively long-lasting conducting states ($1024 s) characterized by suf-ficiently large currents would be detected. This is in ac-cordance with another study proposing that the minimalnumber of the MscL conducting states was seven with oneclosed, five subconducting, and a fully open channel state(247). The study determined that the number of subcon-ducting states varied with the applied pipette voltage (Liuand Martinac, unpublished data), which is expected if theoccupancy of different subconducting levels changes withvoltage. This may explain why in the study by Sukharev etal. (390) fewer number of conducting states were de-tected, since the analysis was based on MscL currentsrecorded at a single membrane potential of 120 mV thatwas probably too low to resolve more than four conduct-ing states. Nevertheless, the five-state model provides acorrect estimate of energies involved in MscL gating byshowing that the energy DG0 necessary to open MscL islargely used to open the pore of 6 nm2.

5. Hydrophobic surface match model

The hydrophobic surface match model derives fromthe original studies of the MS gating of gramicidin (seesect. IVB) and represents an additional mechanism thatmay explain MscL. This model assumes that the hydro-phobic match between the hydrocarbon bilayer thicknessand the length of the hydrophobic surface of MscL isimportant in determining the stability of different channelconformations. The feasibility of the model will requireinformation on the open MscL conformation so that esti-mates can be made of the free energy contributions ofdifferences in hydrophobic mismatch between MscLstates and the bilayer before and during stretch (i.e.,bilayer thinning). Because it is already known that MscLgates close to lytic tensions, the bilayer must be near itsmaximal expansion and minimal thickness (i.e., 2–4%according to DA/A 5 2Dd/d; see sect. IIB). A 2–4% changein bilayer with a thickness of 3.5 nm would thin themembrane ;0.1 nm. If the thinned bilayer better matchesthe open channel than the closed channel by only 0.1 nm

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 701

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 19: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

(i.e., because it gets shorter as well as wider), this wouldproduce an improved surface mismatch between MscLand bilayer of ;1.6 nm2, assuming the diameter of MscLis 5 nm (60). Because the free energy increase for trans-ferring a hydrophobic protein surface from an organicsolvent to an aqueous environment is ;17 mJ/m2 (66), theenergy corresponding to the improved MscL/bilayermatch should be DG ;2.7–4.8 3 10220 J or ;7–12kT. Thisenergy is of the same order as the free energy differenceof 18.6kT estimated between closed and open states ofMscL (390). Thus hydrophobic mismatch could presum-ably contribute to the mechanosensitiviy of MscL. Indeed,it was recently found that when placed in a thinner bilayerMscL required less energy for activation by membranetension; the opposite was true in a thicker bilayer (B.Martinac, unpublished observations).

In summary, the present experimental, structural,and theoretical evidence support the following simplifiedview of the MscL gating. The NH2 terminus, althoughmobile, does not function as the channel gate. The gate ismost likely formed by the hydrophobic constriction of theTM1 helices at the cytoplasmic end of the closed channel.The mobile NH2 termini probably serve to stabilize theopen channel conformation(s) and may interfere with thepassage of ions, thus leading to channel subconductancestates. The COOH termini possibly play a role in stabiliz-ing the closed configuration of the channel, whereas theperiplasmic loop may function as an elastic spring resist-ing the opening of the channel by membrane tension.

J. Membrane Localization and Physiological

Function of MscL

MscL is a MG ion channel for which a physiologicalfunction can be correlated with its membrane localizationand molecular properties. To begin with, two independentstudies (27, 170) have associated MscL activity with theinner cytoplasmic membrane of E. coli (i.e., the tension-bearing membrane). First, Blount et al. (27), using su-crose gradient centrifugation to separate total membraneof E. coli into inner and outer membrane fractions, dem-onstrated the MscL protein was associated exclusivelywith the inner membrane fraction. Second, Hase et al.(170) combined [35S]methionine radiolabeling of MscLwith the detection of the NADH oxidase activity in thecytoplasmic membrane fraction and showed 90% of35S-MscL was associated with the NADH oxidase-richmembrane fraction. These results confirm the early studyby Berrier et al. (18) that demonstrated MG channel ac-tivities were localized predominantly in the inner mem-brane fraction.

Bacteria exposed to an hyposmotic shock rapidlyrelease cytoplasmic contents into the surrounding me-dium (38, 225, 344, 359). Moreover, osmotically induced

efflux of lactose and ATP from E. coli as well as ATP fromS. faecalis can be blocked by Gd31 (19). Gd31 blocks MGchannels in E. coli and other cells (158) and is proposedto act by modifying the mechanical properties of thebilayer rather than binding to the different channel pro-teins (105). Furthermore, two other types of MG channelsfound in E. coli, MscS and MscM, can also be activated bydifferences in osmotic pressure in patch-clamp experi-ments (77, 268). Although these results implicate MGchannels in bacterial osmoregulation, what was lackingwas a bacterial phenotype that would directly supportsuch a role. Blount et al. (25) identified such a phenotypewith several site-directed mutations (e.g., K31E or K31Din TM1) that led to dramatic changes in MG channelactivities and also a “slow or no growth” phenotype thatcould be partially reversed by increasing osmolarity of thegrowth medium (25). Furthermore, these mutants showeda correlation between the severity of phenotype and theseverity of “abnormalities” in ion channel activities.

A similar study by Ou and coworkers (316) found thatwhen a randomly mutagenized mscL gene was expressedin an mscL2 strain, which had no detectable phenotypeon growth plates, most of the gain-of-function mutants,characterized by slow or no growth, could be associatedwith significant changes in the MscL channel kinetics andpressure sensitivities. Moreover, the most severe muta-tions occurred when amino acid residues between R13and K31 in the first transmembrane TM1 segment of MscLwere mutated. Significantly, the TM1 domain is highlyconserved among bacterial species (283, 387).

Finally, Ajouz et al. (2) demonstrated that MscLopens in vivo during an osmotic downshock. Specifically,they investigated osmotically induced effluxes in wild-type and mscL2 mutant cells of small osmolytes such aspotassium glutamate, trehalose, and glycine betaine thatserve as osmoprotectants in bacteria. Although no differ-ence was found between the wild-type and the mscL2

cells in the release of these osmoprotectants during os-motic challenge, thioredoxin, a small cellular protein,which is also excreted from E. coli upon osmotic down-shock, was completely released from the wild-type cellsbut retained by the mscL2 cells. This result indicates thatMscL is not required for the excretion of osmoprotectantsbut is apparently activated in vivo to facilitate the efflux ofthioredoxin. The effect of efflux of thioredoxin and pos-sibly other small cytosolic proteins during osmotic down-shock remains to be determined. Note the apparent ex-cess MscL (i.e., 50–100 copies) expressed in singlebacterial cells (26, 169), together with the fact that open-ing of single MscL channel would suffice to dissipateosmotic gradients within 1 ms, does not preclude theirprimary role in microbial osmoregulation or osmoprotec-tion. This is simply because biological designs, particularthose underlying safeguard mechanisms, are not onlybased on economy but also redundancy.

702 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 20: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

K. MscS and MscM

MscS was the first MG channel activity described inbacteria (267) and was also the first MG channel shown torespond to bilayer tension (266). It has a conductance of;1 nS in 200 mM KCl/40 MgCl2 (388) and requires ;0.7the tension needed to open MscL (28, 387), but like MscL,it can also be activated by osmotic pressures (77, 268).MscS differs from MscL and MscM in that it exhibits avoltage dependence characterized by e-fold change in theopen probability per ;15-mV depolarization (267). MscSis relatively nonselective, displaying a slight preferencefor anions over cations (267, 388) and is blocked bysubmillimolar Gd31 (17, 19). Recently, Booth and co-workers (31, 244) found that MscS activity in bacterialprotoplasts is abolished by null mutations in two loci onE. coli chromosome, yggB and kefA. The MG channelsaffected by the kefA and yggB null mutations have similarthresholds of activation by pressure, and both channelsexhibit a conductance of ;1 nS. The activity of the YggBchannel is encountered in almost 100% of protoplastpatches and is characterized by a large number of chan-nels gating simultaneously that inactivate rapidly withsustained pressure. The KefA activity is less frequentlyencountered (70% of the patches) and is characterized byfewer channels that do not inactivate. Whereas YggB is asmall membrane protein of 286 amino acids, KefA is alarge, multidomain 120-kDa membrane protein (1,120amino acid residues). Interestingly, the amino acid se-quence of YggB resembles highly the sequence of the lasttwo domains of the KefA protein.

MscM is less frequently encountered in membranepatches of giant E. coli spheroplasts compared with MscS(76, 77) and MscL that are typically found in every patch(387). The conductance of MscM is about half that ofMscS and exhibits a slight preference for cations (17, 18),indicating it is molecularly distinct from MscS and MscL.However, MscM is also blocked by Gd31 and activated byhyposmotic stress (77). The different pressure sensitivi-ties of the three channels (i.e., their activation pressuresincrease with their conductance) may indicate the chan-nels are activated sequentially to provide a graduation ofefflux pathways (124). However, apparently the activationof either MscS or MscL can maintain the integrity of E.

coli during hyposmotic challenge, since single deletionmutants lacking either channel remain fully functional(31). In contrast, double mutants die, indicating thatMscM alone is not able to protect bacteria from osmoticdownshock.

L. MG Channels in Archaea

Archaea (formerly archaebacteria) are unicellularprokaryotes like eubacteria. However, they constitute a

separate domain on the phylogenetic tree different fromthose of Bacteria and Eukarya (Fig. 5A) (318, 436). Ar-

chaea encompass several distinct groups of microorgan-isms adapted to extreme environments such as super-hotocean hydrothermal vents characterized by extreme tem-peratures or Dead Sea containing extreme salt concentra-tions (13, 436). The existence of ion channels in archaealcell membranes has not been documented until the recentreports of porins (20) and MG channels (239) in thehalophilic archaeon H. volcanii.

Two types of MG channels have been identified inH. volcanii (239) and have been named MscA1 (i.e., Ar-

chaeon1) and MscA2. Both have large conductances anddisplay a similar mechanosensitivity as the E. coli MGchannels, but they differ in their distinct rectificationproperties. MscA1 has a conductance of 380 pS at 140 mVand of 680 pS at 240 mV, whereas MscA2 has a conduc-tance of 850 pS at 140 mV and of 490 pS at 240 mV. Likethe bacterial MG channels, both channels are activated bymechanical force transmitted via the lipid bilayer, andboth are blocked by submillimolar Gd31 (19).

With the use of the same functional approach as usedto identify the MscL protein (384, 386), a 15-kDa mem-brane protein of the thermophilic archaeon Thermo-

plasma volcanium was found to correspond to activitiesof a novel MG channel (220). In symmetric 200 mM KClplus 40 mM MgCl2, the channel has a conductance of ;1.5nS. Twenty NH2-terminal amino acid residues of the 15kDa were determined by microsequencing and werefound to match with 75% identity the start of the openreading frame of a gene of unknown function in thegenome of the related Thermoplasma acidophilum

(Kloda and Martinac, unpublished data). Computer-as-sisted secondary structure analysis of the protein en-coded by the T. acidophilum gene revealed two putativea-helical membrane-spanning regions, suggesting a struc-tural similarity with MscL. Helical wheel alignment of thetwo helices revealed that the first helix is amphipathic incharacter, has a cluster of five charged residues on oneside of the wheel, and has a mixture of hydrophobic andhydrophilic residues on the other side and thus mostlikely lines the pore of the putative channel. This wouldbe consistent with the three-dimensional crystal structureof Tb-MscL that indicates TM1 creates the bulk of thepore of the pentameric channel (60). The second helix hasthe usual a-helix with hydrophobic and hydrophilic resi-dues on both sides of the wheel.

Recently, a new MG channel was identified in thearchaeon Methanococcus jannashii by using the TM1transmembrane domain of MscL as a genetic probe tosearch the microbial genomic database for MscL ho-mologs (220a, 269). A hypothetical protein MJ0170 in theMethanococcus genome was found to contain a sequencethat shares 38.5% identity with the TM1 of Eco-MscL. Thesame protein was also found to share a high homology

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 703

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 21: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

with the YggB protein underlying the activity of MscS inE. coli. This may indicate that MJ0170, which has beenrenamed to MscMJ for the MS channels of M. jannashii,is a hybrid of MscL and MscS, which may have evolved asa result of gene duplication of an ancestral mscL-likegene. Importantly, the alignment of sequences of MscL,MscS, and MscMJ homologs revealed that bacterial andarchaeal channels form a phylogenetic tree composed ofthree main branches of prokaryotic MG channels (see Fig.5B), indicating that the common ancestor of the prokary-otic MG channels most likely resembled MscL. Whenexpressed in E. coli and examined in giant spheroplasts orafter reconstitution into liposomes, the MJ0170 (MscMJ)protein expressed a channel with a conductance of ;270

pS in 200 mM KCl and a cation selectivity (PK/PCl ;6)somewhat similar to eukaryotic SA-CAT channels (158,265, 286).

M. MG Channels in Evolution

The finding of MG channels in organisms belongingto all three domains of the phylogenetic tree (Fig. 5A)points toward their early evolutionary origins. A basicquestion is whether the mechanisms of mechanical gatingthat originated in bacterial channels have been conservedin eukaryotic MG channels? Certain properties of thebacterial MG channels, including their extremely high

FIG. 5. Evolutionary aspects of mechanically gated(MG) channels. A: universal phylogenetic tree based onsmall subunit rRNA sequences. Representative organismsin each of the three domains with documented MG chan-nel activities are shown. [Modified from Pace (318).] B:family of prokaryotic MG channels represented by thephylogenetic tree of aligned sequences of MJ0170 andMscL homologs. The homologs are clustered into threemain branches based on their sequence similarity, suggest-ing that the common ancestor of all prokaryotic MG chan-nels resembled MscL. The mscL-like gene duplication wasmost likely followed by subsequent gene divergence sim-ilar to the evolution of voltage-gated ion channels. [Mod-ified from Kloda and Martinac (220a).]

704 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 22: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

conductance, general lack of ion selectivity, and require-ment of near-lytic activation tensions, have been taken asevidence that a different type of gating mechanism under-lies the more MS and ion-selective eukaryotic MG chan-nels. However, we now know from mutational and pro-teolytic cleavage studies that MscL can be made as MS(i.e., gated at or near zero applied pressures) as anyeukaryotic MG channel studied in patch-clamp experi-ments. Furthermore, although a tethered mechanism (i.e.,involving direct CSK- and/or EC-channel interactions) isoften evoked for eukaryotic MG channels (see sect. VIIIC),there is growing evidence that at least some eukaryoticchannels (e.g., the SAT-CAT channel in Xenopus oocytes)are gated by tension developed in the bilayer. As yet nosequence homologs for MscL have been identified in eu-karyotes. However, there are striking similarities in thebasic structure and membrane topology of MscL and theputative MG channels (MEC-4 and MEC-10) identified inC. elegans (403). This observation alone may indicatecommon biophysical principles confer mechanosensitiv-ity on the two classes of channels. In the following sectionwe discuss some of the additional cellular adaptationsand specializations of animal cells (i.e., extrinsic factors)that play an important role in influencing how animal cellssense and respond to mechanical inputs.

VII. THE STRUCTURE OF ANIMAL CELLS:

SPECIFIC ROLES OF THE CORTICAL

CYTOSKELETON AND EXTRACELLULAR

MATRIX IN MECHANOSENSITIVITY

One of the most important structures in terms ofmodifying the animal cell’s response to mechanical defor-mation is the cortical CSK (102, 108). This structure, byproviding bilayer protection while preserving cell deform-ability, allows dramatic changes in cell shape and sizeduring growth and differentiation and permits rapid ani-mal movements. Furthermore, by removing the require-ment of a thick cell wall, the cortical CSK allows theanimal cell to maintain a more intimate contact with itsmechanical surroundings. The basic function of the cor-tical CSK is to structurally support the fluid bilayer,thereby providing the cell membrane with a shear rigiditythat is lacking in simple bilayer vesicles or spheroplasts/protoplasts. Furthermore, by acting as a membrane scaf-folding, the CSK allows the cell to assume nonsphericalgeometries. As a consequence, animal cells are able tomaintain a stable, excess membrane surface area beyondthat required to enclose their volume as a smooth sphere.The biconcave RBC has 40% excess membrane area (107),whereas other cells (e.g., lymphocytes, skeletal muscle,mast cells, hepatocytes, astrocytes, and oocytes) mayhave between 100 and 1,000% in the form of membranefolds, microvilli, and/or cavaeolae (Fig. 8; Refs. 95a, 107,

338a, 375a, 444, 448). This allows the cell to increase involume without new membrane insertion (312a, 338a,375a, 448). Thus the additional surface area serves as animmediate membrane reserve providing compatibility be-tween the highly expandable CSK network (KA of ;1022

mN/m) and the nonexpandable lipid bilayer (KA ;102

mN/m) (107, 108, 284). During mechanical deformations(e.g., inflation, indentation, stretching), the CSK networkmay be expanded and the excess membrane smoothedout before significant tension develops in the bilayer (223,330, 375a, 448). Several lines of evidence directly supportthis notion. First, scanning electron micrograph images ofcultured mammalian cells indicate that osmotic swellingresults in the surface microvilli “unfolding” as cells almostdouble in diameter (see Fig. 1 in Ref. 223). Second, mastcells can be inflated to nearly four times their volume withlittle increase in membrane capacitance (Cm) (i.e., ,10%)(375a, see also Refs. 312a, 338a). Third, the force requiredto pull tethers (i.e., CSK-free membrane strands) fromfibroblasts is independent of tether length up to ;5 mm(330). Fourth, direct inflation or osmotic swelling of oo-cytes to twice their normal diameter does not activate MGchannels currents (448, 449). In contrast, only slight infla-tion (,10%) of CSK-deficient plasma membrane vesicles(PMVs) (i.e., formed from blebbed oocytes) activates MGchannels (447).

Another effect the cortical CSK may have on bilayermechanics is through organizing the bilayer into localdomains with smaller radii of curvature (rc) than that ofthe cell (an extreme case is the microvilli, rc ;0.1 mm, seebelow). According to Laplace’s law, this will effectivelyreduce the membrane tension (t) that develops in thelocal bilayer domain for a given osmotic or hydrostaticpressure (p) (i.e., t 5 p z rc /2, see Ref. 145). Finally, theexistence of contractile as well as load-bearing elementsin the CSK allows the maintenance of resting cortical CSKtension that can actively resist membrane dilation causedby internally generated forces (15, 200).

The structure of the cortical CSK has been studied ingreatest detail in the human RBC where it shows up inspread membranes as a hexagonal network of spectrintetramers interconnected at their ends to actin, tropomy-osin, and other proteins to form a multi-protein network(284). In resting RBCs, this hexagonal arrangement isobscured and appears in electron micrographs as a densefilamentous network, indicating the links in the networkare normally folded (or bent) but can be extended withapplied tension (144, 284, 377). For example, the distancebetween actin filaments in the resting RBC is ;70 nm butcan reach 200 nm at full extension. Several molecularmodels have been proposed to underlie this large elasticextension, including random coil (377) and helical springmodels (144, 274). The cortical CSK is anchored to thelipid bilayer by interactions between ankyrin and otherCSK proteins with various integral membrane proteins

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 705

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 23: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

(e.g., the band 3 anion exchanger). Although in noneryth-roid cells there are many variations in both the specificcortical CSK elements (e.g., dystrophin in muscle) andthe CSK-bilayer connections, the general organizationalthemes seen in the RBC appear to be conserved (250).The major difference between the anucleated human RBCand nucleated cells is that tension generated within thecortical CSK is not focused exclusively on the membranebut is also shared (or resisted) by the three-dimensionalCSK network that suspends and transmits forces to theelastic nucleus. This CSK meshwork is highly integratedand has been proposed to behave as a tensegrity structurein which any mechanical deformation is transmitted glo-bally throughout the network (200, 201). However, a re-cent study has demonstrated that fibroblasts show highlylocalized responses to mechanical deformation (171a).Also, mechanosensory neurons and receptors “sense” lo-cally and transmit global responses electrically.

Although animal cells lack the rigid cell wall of bac-teria, many animal cells (i.e., in tissues) possess a com-plex extracellular matrix (ECM) that serves as externalscaffolding through which external forces can be filteredor focused on the cell or distributed throughout the tis-sue. For example, the physical coupling between ECMmolecules (e.g., fibronectin) and the CSK through trans-membrane proteins (e.g., integrins) allows forces to befocused on CSK elements that may directly or indirectlyinteract with membrane proteins (63, 327, 429). The vis-coelastic elements in the ECM may also filter out steady-state forces while allowing the transmission of fast me-chanical oscillations possibly through direct physicalconnections with membrane channels. Finally, special-ized mechanosensory cells often possess unique struc-tures (e.g., the tip links of stereocilia, Refs. 190, 326) andancillary cellular layers (e.g., the Pacinian cell capsule)that confer high directional and vibrational sensitivity onthe mechanotransduction process.

VIII. MECHANICALLY GATED CHANNELS

IN ANIMAL CELLS

The original concept of MG channels arose fromwhole cell recordings of specialized mechanosensoryneurons (89, 212, 249). In these early studies a physicalmechanism of channel gating was often favored over anindirect chemical mechanism because of the typical shortlatency (i.e., a few ms) and high-frequency response (kHz)of the mechanotransduction process (70, 135, 242, 409).Tight-seal patch-clamp recording (153) allowed the directmeasurement of single MG channels currents (37a, 142,151), while pressure-clamp techniques (271–273, 347)and, in particular, the ability to apply fast pressure steps,demonstrated that MG channels in specific cell types (e.g.,Xenopus oocytes) could be activated with millisecond

latencies (271, see Fig. 6). However, not all cells show fastMG channel activation. For example, pressure steps ap-plied to cell-attached patches on chick skeletal muscle(375) and snail neurons (370) activate MG channels withdelays of 1–30 s. Furthermore, in snail neurons, physicalor chemical disruption of the CSK abolishes the delayedactivation. This indicates that the intact cortical CSKprevents or slows tension development in these cells (seesect. VII). This behavior illustrates how factors extrinsic tothe channel protein can dramatically modify MG channelactivity.

Patch-clamp studies over nearly 20 years indicatethat MG channels are widely expressed in both sensoryand nonsensory cells and in cells from species spanningthe full evolutionary spectrum (see sect. VIM). This ubiq-

FIG. 6. Rapid activation of MG channel currents in cell-attachedmembrane patches on Xenopus oocytes. Top trace: the waveform of asuction step (;15 mmHg) applied to the patch pipette. The crosses onthe rising phase indicate the time delay in reaching 5 mmHg (1.5 ms) and10 mmHg (4 ms), which represents the range of minimal suctionsrequired to activate the Xenopus MG channels with standard size patchpipettes (i.e., ;2 mm internal tip diameter). Middle trace: a membranepatch in which 5 MG channels were activated by the suction step.Bottom trace: another pipette slightly larger in tip diameter (i.e., ;3 mm)in which over 20 MG channels were activated. The estimated latency forfirst channel activation varied between 0.5 and 2.5 ms. [Modified fromMcBride and Hamill (271).]

706 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 24: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

uity has led to the idea that MG channels play a role ingeneral cellular functions such as cell volume regulation(151, 156, 231, 232, 265, 286, 348, 350, 358, 400). However,the possibilility has also been raised that membranechanges induced by tight seal formation induce mechano-sensitivity in specific channels. The “artifact” idea origi-nally arose from a discrepancy observed between mem-brane patch and whole cell mechanosensitivity in snailneurons (287). Specifically, it was reported that despitethe consistent ability to activate single MG K1 channels inmembrane patches, macroscopic K1 currents could notbe mechanically activated in the whole cell (287). Al-though this discrepancy has been shown not to be gener-alized to all cell types (55, 77, 84, 147, 192, 356, 362, 419,433, 453, 445a, 454), it has remained a high profile con-troversy in the field. The issue has most recently been

addressed in studies of Xenopus oocytes, where specificattention was focused on differences in the geometry andmechanics of the cell membrane in the patch and wholecell recording configurations (446–449).

A. Membrane Patch Mechanics and Morphology

The tightly sealed membrane patch spans the insideof the pipette with its circumference or boundary fixed tothe walls of the pipette (153, 354). High-resolution videoimages indicate that in the absence of stresses normal tothe plane of the membrane (i.e., due to hydrostatic orosmotic pressures) the patch appears to be pulled flat andperpendicular to the walls of the pipette (Fig. 7, Refs. 314,373, 374, 448). This is consistent with Laplace’s law if

FIG. 7. High-resolution video images of a cell-attachedoocyte membrane patch and membrane currents activatedduring brief steps of pressure and suction. A and B: videoimages of a membrane patch before (0 ms), during (50ms), and after (250 ms) steps (i.e., of 100 ms duration) ofsuction (A) and pressure (B). C and D: the membranepatch currents recorded in the same patch imaged in A

and B in response to the suction and pressure steps. Thepipette tip (not visible) had a tip diameter of 4 mm, and thepatch was located ;20 mm from the tip. [From Zhang andHamill (448).]

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 707

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 25: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

there is a resting tension in the patch. Presumably, theadhesive forces at the membrane-glass interface and, inparticular, where the membrane is bent at right anglesgenerates this resting tension (314, 373, 448). With appliedsuction (negative pressure) the patch flexes out (Fig. 7A)and with positive pressure flexes in (Fig. 7B). The patchappears optically smooth, and electrical membrane ca-pacitance measurements indicate a patch area consistentwith a flat membrane disk (448, see below). Furthermore,the expansion of the patch (i.e., ,10%, Ref. 448) associ-ated with rapid activation of MG channels (Fig. 7, C andD) indicates that little smoothing of the membrane isrequired before membrane tension develops. These ob-servations are also consistent with electron micrographimages of oocyte patches in pipette tips that indicate atrilaminar structure with no evidence of membrane foldsor microvilli (345).

In contrast to the simple membrane geometry of thepatch, the plasma membrane of the oocyte and otheranimal cells displays a far more complex geometry. Forexample, transmission and scanning electron micro-

graphs of Xenopus oocytes (Fig. 8) indicate extensivemembrane folding and a high density of microvilli (29,324, 444, 447). The quantitative analysis of freeze-fractureimages of the oocyte indicate that membrane folding dou-bles the surface area while microvilli may further increaseit by fivefold (444). This analysis is consistent with mem-brane capacitance measurements that indicate a mem-brane area that is 5–10 times larger than predicted for asmooth sphere (448). In the electron micrographic analy-sis, the estimated microvilli density was 6–7 microvilli/mm2 (with a microvillus length of 1.4 mm and a diameterof 0.12 mm). If the membrane folds and microvilli werepreserved during tight seal formation, a patch with anapparent geometric area of 50 mm2 (i.e., see Fig. 7) wouldhave ;300 microvilli and an actual area between 250 and500 mm2. However, Cm measurements indicate an area of;50 mm2 (448), again consistent with electron micro-graph patch images that show no evidence of microvilli(345).

There are several scenarios by which tight-seal for-mation may alter membrane geometry (Fig. 8C). In one,

FIG. 8. Microvilli on the surface of Xe-

nopus oocytes and their possible fates dur-ing tight seal formation. A: transmissionelectron microscopy of an oocyte showingprominent microvilli containing dark cyto-plasmic material. B: scanning electron mi-croscopy indicating the high density of mi-crovilli on the oocyte surface. [Modifiedfrom Zhang et al. (447).] C: cartoon showingpossible fates of microvilli during tight sealformation. Top panel shows a patch pipettepressed against the oocyte surface beforeapplication of suction. Middle left panel

shows suction has drawn the membrane andtended to smooth out the microvilli both inand immediately outside the pipette. Middle

right panel indicates suction shearing off (orexcising) the microvilli in the patch. Bottom

panel shows a tightly sealed flat patch afterremoval of suction.

708 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 26: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

suction applied during sealing results in the lipid bilayerbeing decoupled from the underlying cortical CSK anddragged into the patch where it seals tightly to the pipettewalls (281). This may be analogous to the drawing out ofthin bilayer tethers from RBCs and other cells (181, 182).However, such tethers are much thinner in diameter thanthe membrane patch (i.e., ;0.2 vs. ;2 mm) and are com-pletely free of CSK (182). In contrast, there is good evi-dence that a plug of CSK is drawn into the pipette alongwith the membrane (154, 345, 374, 447). Although, oncethe seal has formed the membrane cap can be mechani-cally decoupled from the underlying CSK to form a mem-brane bleb (154, 160, 374, 447). Thus a more likely sce-nario may involve the smoothing out of surface folds andmicrovilli (Fig. 8C) while retaining interactions with areorganized CSK within the patch. In a third scenario, themicrovilli are sheared off by the initial pressure/suctionapplied during sealing. In this case, one might expect themicrovilli to be pinched off at their base with the mem-brane rapidly resealing or fusing. This may be analogousto the situation with an inside-out or outside-out patchthat seals as the patch pipette is withdrawn from the cell(153). At least consistent with this idea is the observationthat vesicles can be seen to be swept from the cell surfaceduring tight-seal formation on chick skeletal muscle(374).

On the specific issue of membrane-CSK coupling,evidence indicates that both cell aspiration and tight-sealformation may alter the relationship between the bilayerand the cortical CSK. For example, in experiments inwhich RBCs were aspirated into pipettes without allow-ing tight-seal formation, fluorescent labeling of the CSKindicates a steep decrease in density of the actin andspectrin network along the aspirated membrane projec-tion (91). Interestingly, this density gradient was main-tained for the duration of the deformation (i.e., .30 min)but recovered rapidly (seconds) with release of suction.Similar changes may occur with “gentle” tight-seal forma-tion that results in expansion, but not irreversible disrup-tion, of the CSK network. As a consequence, the densityof the network may decrease and membrane wrinkles andfolds become smoothed out. However, stronger suctionor pressure applied after the tight seal has formed may“overdilate” the network and thereby irreversibly disruptlinks within the network as well as those between thenetwork and the membrane (154, 374, 447). In this case,the membrane patch would lose its shear rigidity andelastic response to mechanical deformation. As a conse-quence it would behave more like a fluid flowing into thepipette to form a CSK-free bilayer bleb (e.g., see Fig. 10 inRef. 447). Interestingly, Sokabe et al. (375) reported thatsuction steps caused slow progressive membrane patchmovements with latencies of ;200 ms and a rise time of1 s. These slow movements clearly contrast with the fastpatch movements seen in Figure 7 and may be associated

with decoupling of the membrane from the underlyingCSK (see also Ref. 447). As a consequence of this decou-pling, one might expect the ability to develop membranetension in the patch would be reduced as more membraneflows into the patch (i.e., from along the sides of thepipette and the cell). Consistent with this idea is thereported decrease in mechanosensitivity (i.e., shift inthe stimulus-response relations to higher pressures) inoverstimulated patches and in blebbed membranes fromXenopus oocytes (375, 447, see Fig. 14A). However, asmentioned earlier, the mechanosensitivity of MG chan-nels in snail neurons and chick skeletal muscle is in-creased after CSK disruption (370, 375). This may indicatethat the CSK in these cells has a much higher resistance toexpansion (i.e., .KA) than the oocyte CSK and is moreeffective in preventing tension development in the bilayer.Such differences indicate that the viscoelastic propertiesof the CSK in each cell may be “tuned” to the mechanicalrequirements (functions) of the cell.

B. Discrepancy Between Membrane Patch

and Whole Cell Mechanosensitivity

The above analysis provides some clues to the dis-crepancy between membrane patch and whole cell mech-anosensitivity (370, 448). Specifically, the simple flat ge-ometry of the oocyte membrane patch (i.e., before anymechanical decoupling) allows for rapid tension develop-ment and MG channel activation (Fig. 7). In comparison,the excess membrane area of the oocyte serves to buffertension development and MG channel activation (Fig. 8).As mentioned previously, this explanation is also sup-ported by the observation that while macroscopic MGcurrents cannot easily be activated in whole oocytes, theycan be activated in oocyte PMVs (447). The PMVs lack acortical CSK and take on a spherical geometry with noexcess membrane area (i.e., as judged by electron micros-copy) so that even slight inflation (1–2%) should increasebilayer tension. A similar explanation may account forwhy macroscopic currents can be activated in sphero-plasts formed from E. coli, yeast, and other microbialcells (77, 146, 147, 454). To form sphereoplasts, the parentcells are enzymatically treated to remove the cell walland, as a consequence, the membrane blebs assume asimple spherical geometry so that even slight inflationcauses MG channel activation (77, 146, 147, 454). Simi-larly, to evoke whole cell MS responses in smooth musclecells, the cells apparently must be inflated to the pointthat they form whole cell membrane blebs or ghosts (seeFig. 1 in Ref. 362). One would also predict that in snailneurons, if the CSK can be disrupted (i.e., the cellblebbed) without causing cell rupture, then whole cell MGK1 currents would be activated similar to the single MGK1 channels in mechanically traumatized patches (see

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 709

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 27: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

Ref. 428). It remains to be determined whether the appliedmechanical stimuli cause changes in membrane-CSK in-teractions in other cell types, such as vascular smoothmuscle (84), urinary bladder myocytes (433), and cardiacmyocytes (192), where whole cell MS currents have beenreported.

Another cell type where membrane geometry may beimportant in determining mechanosensitivity is the kid-ney proximal tubule cell. In this highly polarized cell, twoclasses of MG channels have been recognized in mem-brane patches, a weakly cation-selective MG channel thatis localized to the apical microvilliated surface and aK1-selective MG channel that is localized on the relativelysmooth proximal cell surface (see Ref. 351). Significantly,osmotic swelling or direct inflation of the cell can activatea whole K1 conductance without activating the cation-selective conductance (55, 419). This differential sensitiv-ity may reflect the location of the channels in regions ofmembrane with different geometry (i.e., different radii ofcurvature) such that they experience different tensionsfor the same applied pressure according to Laplace’s law(448). One can also imagine that other cells (e.g., mech-anosensory neurons) possess localized regions of bilayerthat are prestressed by CSK and/or ECM interactions toenable rapid tension increase in response to mechanicalstress (i.e., analogous to the patch).

C. MG Channel Gating: “Tethered” Versus

“Bilayer” Models

Bilayer reconstitution experiments have provided un-equivocal evidence for a bilayer model of mechanicalgating of alamethicin, gramicidin, and bacterial MG chan-

nels (see sects. IV and VI). However, it was apparently notintuitively obvious to early investigators (170a) that thefluid lipid bilayer with its low tolerance to dilation coulddevelop and maintain tensions sufficient to influencemembrane proteins. Indeed, in bilayer reconstitution ex-periments, phospholipid and cholesterol content are oftenadjusted to increase the expansion modulus of the bilayerand thereby shift the lytic tensions to higher values thanthose displayed by cell membranes (296). Furthermore, asdiscussed above, the excess membrane area of animalcells tends to buffer rapid increases in bilayer tension thatotherwise might rupture the cell (79, 80, 330, 448). Theseapparent biases against the bilayer model would seemample reason to consider an alternative model in whichmechanical force is transmitted directly to the channelprotein through CSK and/or ECM tethers. As discussedbelow, there are several lines of evidence that indicate a“tethered” mechanism may gate specific MG channels.However, as yet, there is no single experimental result(i.e., analogous to liposome reconstitution) that providesunequivocal support for this class of model.

1. Physical and chemical disruption

of putative tethers

One strategy to identify a tethered mechanism is toshow that disruption of the putative tethers abolishes MGchannel activity. This strategy has been most successfullyused in studies of audio-vestibular hair cells where extra-cellular “tip links” connecting stereocilia tips are hypoth-esized to act as external gating springs for MG channels(71, 326; Fig. 9). This model proposes that hair bundledisplacement stretches the tip links and thereby influ-ences the transition between the closed and open states

FIG. 9. Diagram of a pair of stereocilia illustrating twodifferent hypothesized sites of mechanotransduction. A:the tip-link between the stereocilia (a) is proposed to exerta gating tension on MG channels located at each attach-ment point (Pickles-Hudspeth model). Horizontal connec-tions (exaggerated at b) between the stereocilia mem-branes where they come into closest contact are proposedto exert a shear displacement and gate MG channels lo-cated in the region of membrane abutment (Furness-Hack-ney model). B: force applied to the stereocilia (arrow) maycause distortion (stretch) of the membrane at a and b.

710 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 28: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

of the MG channel (69, 71, 73, 188–190, 263, 326, see Ref.260 for recent review). Consistent with the model is thedemonstration that tip-link disruption with external solu-tions in which Ca21 is reduced by the presence of tet-racarboxylate chelators (e.g., BAPTA) also abolishesmechanotransduction (10, 97, 263). Furthermore, afterregeneration of the tip links (i.e., that takes several hoursafter disruption and chelator removal), mechanotransduc-tion can be restored, although without the strong adapta-tion seen before disruption (451).

In the original version of the gating spring model, atwo-state channel model was proposed (71, 188–190).However, this simple model (150) could not account forseveral features of hair cell mechanotransduction (seeRef. 260 for review). Therefore, a three-state model in-volving two closed states and a single open state (C17C2 7 O) was suggested (73, 260). In the first closed stateC1, it was assumed that a portion of its gating spring wasimmobilized so that part of the spring’s tension was nottransmitted to the gate. As a consequence, this “latched”closed state would have a lower compliance than eitherthe unlatched closed state (C2) or the open state (O). Itwas assumed that the energy of channel opening de-pended linearly on hair bundle displacement such that thegating force was a constant independent of bundle posi-tion (69, 260). However, because the gating springs resistextension but not compression, they should slacken atnegative displacement such that the channel’s total en-ergy should become independent of negative displace-ment. This model could reproduce the observed asymmet-rical displacement-response relation and account for themeasured differences in bundle stiffness at negative andpositive displacements as well as provide a good fit toexperimental data from vestibular and cochlear hair cells(260). In this model, it was originally assumed that thegating springs were linearly elastic. However, recently ithas been shown that their stiffness increases with ten-sion, similar to other elastic materials (263).

Despite the necessary added assumptions, the tip-link/gating-spring model remains the main theoreticalframework for hair cell studies. However, more recentobservations have challenged the model (278). Specifi-cally, it was reported that tip-link disruption by eitherBAPTA or elastase resulted in a sustained inward currentthat could be blocked by dihydrostreptomycin (100 mM),amiloride (300 mM), or Gd31 (1 mM) (278). Because thesethree agents are known to block MG channels (althoughnonspecifically, see Ref. 158), it was proposed that thesustained inward current was carried through perma-nently open MG channels. This was considered inconsis-tent with the tip-link model because without the tip links,MG channels should be either closed or only show a lowPo consistent with their displacement independent energyevident at negative displacements (260). An alternativemodel referred to as the “abutment” model (Fig. 9) was

proposed to explain the permanently open MG channels(119, 148, 278). This model originally arose from immu-nocytochemical localization of the putative MG channelsnear, but not at, the points of tip-link attachment (148).The putative MG channels were identified using a poly-clonal antibody for the rat kidney amiloride-sensitive Na1

channel (rENaC) based on the observation that amilorideblocks both channels (158). However, amiloride blocksMG channels with a much lower affinity, and amiloridederivatives block with a different order of potency com-pared with the rENaC (343), raising questions regardingthe specificity of the ENaC antibody for the MG channel.Furthermore, recent studies using a-ENaC knockout miceindicate that the a-subunit is not required for hair cellmechanotransduction (342). However, this may only indi-cate that other non-aENaC subunits make up the hair cellMG channel (see sect. VIF2).

In the abutment model, the MG channels are locatedat the junctionlike structures where the stereocilia ad-here. The resultant shearing forces at the membrane junc-tions during bundle displacement are proposed to causechannel opening, analogous to stretch-activated channels(see Refs. 142, 161). Recent kinematic analysis indicatesthat the putative channels at the abutment region could beoperated as well as if they were located at the tip links(119). In particular, a linear change in shear displacement(i.e., at the contact region between the stereocilia) occursas one deflects a sterocilia pair (within the physiologicalrange of deflections), and these changes are comparableto the predicted tip link elongations (see Fig. 4 in Ref.119). Furthermore, if the disruption of the tip links resultsin membrane junctions being exposed to increased ten-sion, due for example to splaying of the stereocilia, thenthe abutment model might also explain why MG channelsare permanently open (278).

It should be pointed out that others have reportedeven larger sustained inward currents following BAPTAtreatment (10) [i.e., .400 pA (10) cf. with 40 pA (278)].However, the currents were interpreted as arising fromCa21-dependent shifts in voltage-gated Ca21 channel ac-tivation (i.e., to more negative potentials) and increasedmonovalent cation flux through the Ca21 channels in theabsence of external Ca21. Significantly, the three agentsused to implicate MG channels also block Ca21 channels(158). On the other hand, there is no evidence that elas-tase could act this way on Ca21 channels to produce thesustained current. Clearly, more selective agents will berequired to unequivocally identify the nature of the sus-tained current and its relationship with MG channels.However, discrimination between the two models willprobably only come with determination of the exact lo-calization and nature of interactions between MG chan-nels and specific stereocilia structures.

Another MG channel initially proposed to operate bya tethered mechanism is the endogenous SA-CAT channel

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 711

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 29: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

in Xenopus oocytes (154). This idea arose because repet-itive mechanical stimulation of oocyte patches irrevers-ibly abolished rapid adaptation and reduced patch mech-anosensitivity (154, see Fig. 14). On this basis it wasproposed that critical CSK elements involved in gatingand adaptation of MG channels were selectively de-coupled from the MG channel. Consistent with this ideawas the observation that during mechanical stimulation aclear space developed between the membrane and theunderlying cytoplasmic structures (154, 447). However,subsequent studies found that MG channel activity wasretained in membrane blebs and PMVs that lack any or-ganized CSK based on immunocytochemical and electronmicroscopic evidence. This activity did not display adap-tation, and the stimulus-response relation was shifted tohigher pressures (447). On the basis of these new find-ings, it has been suggested that tension in the bilayeractually gates the oocyte MG channel but that the corticalCSK modulates the development and relaxation of bilayertension (447). Interestingly, recent studies indicate thatthe CSK modulates the functional properties of non-MGchannels (365, 399). Specifically, two independent groupshave shown that mechanical-induced decoupling of theoocyte membrane patch from the underlying CSK (asdescribed above) can alter the inactivation of voltage-gated Na1 channels by irreversibly switching the gatingmode from a slow to a fast inactivation. The mechanismof this effect remains unknown. However, the Na1 chan-nel is not stretch sensitive, since suction alone cannotactivate the channel nor modify its voltage-dependentactivation (cf., Refs. 365, 399).

Concerning attempts to biochemically disrupt CSKtethers, it has been demonstrated that cytochalasin notonly does not block MG channel activation, it increasesthe mechanosensitivity of specific MG channels (142, 158,370, 375). This indicates that actin microfilaments maynormally constrain the development of tension in eitherthe bilayer or other CSK proteins. Often spectrin (i.e.,fodrin) has been proposed to act as a tether for MGchannels (156, 348, 350). However, this idea has comefrom exclusion of other CSK proteins rather than anydirect evidence, since there are no reagents that selec-tively disrupt spectrin. On the other hand, the recentdemonstration that oocyte MG channel activity is retainedin PMVs devoid of spectrin (i.e., based on electron micro-graph images, Ref. 447) indicates that neither spectrin norany other CSK protein is required to gate this channel.

The lack of effect of colchicine on MG channel ac-tivity in skeletal muscle (142) and Xenopus oocytes (158)and on tactile sensation in the cockroach (233) would alsoseem to rule out microtubules as gating tethers. However,colchicine treatment does abolish touch sensitivity in thenematode C. elegans and the cricket Acheta domesticus

(59, 104). Furthermore, genetic and electron microscopicstudies indicate that microtubules may be part of a com-

plex that mediates touch sensitivity in nematodes andinsects (161, 403, 409, 410).

2. Genetic disruption of putative tether proteins

Different genetic mutants provide evidence both forand against the involvement of specific EC and CSK proteinsin MG channel gating. For example, the dystrophic (mdx)mouse, an animal model of Duchenne muscular dystrophy(DMD), is characterized by the absence of dystrophin, alarge CSK protein normally expressed in vertebrate skeletalmuscle. If dystrophin transmits force directly to the MGchannels in skeletal muscle, then MG channels in mdx andDMD should be mechanically insensitive. However, stretch-inactivated Ca21-permeable channels are upregulated inmdx muscle (116), and stretch-activated Ca21-permeablechannels can still be activated, although they may displayabnormally slow closing (271). Therefore, although the ab-sence of dystrophin may contribute to the elevation of[Ca21]i (i.e., via MG channels) that contribute to muscledegeneration (129, 416), the fact that MG channel activity isretained indicates dystrophin is not required for mechanicalgating.

In studies of touch-insensitive mutants of C. elegans,Chalfie and colleagues have shown that mutations in thegenes encoding a specific collagen in the mantle (i.e.,mec-5; Ref. 95) and the a- and b-microtubulin subunits inthe CSK (mec-7 and mec-12; Refs. 195, 357) block touchsensation. This has led to the idea that proteins in theextracellular and cytoplasmic domains are tethered to theputative MG channels (i.e., MECs/DEGs, see sect. VIIIF2).Interestingly, these studies indicate little redundancy interms of transmitting mechanical force to the channel,since a single mutation in any one element is sufficient toabolish touch sensitivity. Although various models havebeen put forward for how these proteins may be orga-nized (see Fig. 10A, Refs. 120, 161, 403), direct evidence ofMG channel activation (i.e., from patch-clamp studies) isstill lacking. It remains possible that forces transmittedthrough ECM and CSK proteins do not directly gate thechannel but instead stretch the bilayer. For example, earlyelectron microscopic studies of the campaniform recep-tor of the fly by Thurm et al. (411) indicated a complexinvolving microtubules, membrane cones, and extracellu-lar elements proposed to interact with MG channels anal-ogous to the Chalfie model (Fig. 10B). However, it may bethat the channels rather than being directly connected tothe microtubules through the membrane cones are actu-ally localized in the intervening bilayer regions that un-dergo increase in tension when the dendritic sheaf andmembrane cones are compressed.

3. Identification of CSK protein binding sequences

The basic idea of the tethered model is that mole-cules in the CSK and/or ECM domains directly interact

712 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 30: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

with the MG channel protein. In this case, one wouldexpect to find specific consensus regions or domains inthe membrane channel protein that would allow suchinteractions. For example, COOH-terminal cysteine-richregions (208) and NH2-terminal repetitive ankyrin repeats(425), identified in two recently cloned MG channels can-didates, Mid1 and NompC (see sect. VIIIF), may mediatethe protein-protein interactions of a tethered mechanism.On the other hand, they may serve to localize or clusterthe MG channels in specialized regions of the cell ratherthan mechanically gate the channel. Certainly, there areother membrane transport proteins, including the RBCCl2 exchanger (band 3) and various ligand- and voltage-gated channels that interact with that CSK but are gener-ally not thought to be mechanosensitive. Furthermore, a

mutant form of the NMDA receptor channel in which theCOOH terminus that normally links the channel to theCSK was deleted still displayed the same stretch sensitiv-ity as the wild-type NMDA channels (53). This result,together with the observation that amphiphilic com-pounds modulate the NMDA channel’s stretch sensitivity,indicates a bilayer rather than a tethered model of gating.

4. Voltage sensitivity of patch mechanics

and MG channels

Initial studies of membrane patch mechanics wereinterpreted as indicating that gating tension was devel-oped in the CSK rather than the lipid bilayer (375). How-ever, more recent studies indicate that the bilayer can

FIG. 10. Models of the supramolecu-lar complexes proposed to underlie MGchannel activation in the touch-sensitiveneurons of C. elegans and A. domesticus

(house fly). A: individual molecular com-ponents identifed in touch-insensitive C.

elegans mutants. The uppermost struc-ture in each panel represents the man-tle (sheath) that requires the presenceof MEC-1 (unidentified) and MEC-5 (aunique collagen). MEC-9 is also an extra-cellular protein but does not form themantle. Instead, it is proposed to form theextracellular link (i.e., gating spring) be-tween the MG channel and the mantle.Here the MG channel is formed from fivesubunits involving the membrane pro-teins MEC-4, MEC-10, and MEC-6. MEC-2is also an integral membrane protein thatshows strong homology to stomatin, a redblood cell protein that binds to the cy-toskeleton and regulates cation conduc-tance. In this model MEC-2 is proposed toconnect the MG channel to the microtu-bules composed of a- and b-tubulin andencoded by mec-12 and mec-7, respec-tively. The proposed shearing force (ar-row) on the mantle results in opening ofthe MG channel. [From Hamill andMcBride (161).] B: schematized version ofmicrotubule-membrane complex visual-ized by electron microscopy of the flycampaniform sensilla. Cone-shaped mem-brane bridges ;18 nm long are connectedat their tips with microtubules that areembedded in an amorphous material. Inthis model, each cone-membrane com-plex is proposed to represent a single MGchannel unit, and the indentation of theneuronal dendrite compresses this struc-ture, thereby activating MG channels.[Modified from Thurm et al. (411).]

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 713

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 31: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

develop tension (4). In the initial study, prolonged pres-sure steps (i.e., ;5 s) were shown to increase membranepatch area beyond the expected elastic limit of the bilayer(i.e., .10%). This increase occurred with an initial delayof several 100 ms followed by a slow exponential (i.e.,t ;1 s) rise during the pressure step and a slow exponen-tial fall after the pressure step. These changes in areaindicated KA values of ;50 mN/m, compared with valuesof ;500 mN/m for lipid bilayers (see sect. IIB). The lowerKA values were interpreted as reflecting the more elastic(expandable) properties of the CSK network (107, 375).However, the authors also concluded that influx of newlipids from stores along the walls of the pipette couldcontribute to the large increase in patch area. The lattereffect would also tend to lower estimates of KA, but it wasassumed that the bilayer alone could not produce anelastic recovery after the stimulus was removed (but seebelow). In the more recent study (4), the effects of mem-brane potential on patch breakdown were examined. Inthe case of lipid bilayers, it was already known thatbreakdown occurs when the sum of the energy due totension-induced thinning and electrically induced com-pression exceeds a critical value (295). Therefore, it wasargued that if the cell membrane breakdown was alsovoltage sensitive it would indicate the bilayer supports asignificant fraction of the membrane tension (4). Indeed,patch-clamp results indicated that ;40% of the membranetension was supported by the bilayer. To explain thereversible movements of the patch after the pressuresteps (375), it was proposed that restoring forces weregenerated by the tendency of the membrane to be drawnback into the cell (4). This elastic recoil may be analogousto that displayed by thin lipid tethers that are pulled backby the lowered energy state associated with reestablish-ing interactions with the CSK (182).

Reports indicating membrane polarization causesmembrane patch movement and activates MG channelsare also consistent with bilayer tension gating the MGchannel (127, 159, 177, 368, 449). However, although dif-ferent groups agree that depolarization causes membranepatch movements, they disagree on the direction andreversibility of the movements and their contribution toMG channel activation. Specifically, Gil et al. (126) re-ported that depolarization (150 mV) causes a slow(10–60 s) displacement of the membrane plug up thepipette (i.e., 1–2 mm away from the cell) as would beexpected with applied suction. Furthermore, they pro-posed that this patch movement caused the slow depolar-ization-induced activation of MG channels. However, itwas not reported whether the plug could return to itsoriginal position to turn off the channels (see Figs. 1 and2 in Ref. 126). In contrast, Zhang and Hamill (449) foundthat depolarization caused the patch to flex inward to-ward the cell (i.e., as occurs with applied pressure) with-out disrupting the patch boundary. This movement was

reversible but did not consistently activate MG channels.Further studies on the role of patch geometry in thedifferent patch pipettes may resolve the basis for thediscrepancies. However, it does not appear to be due todifference in glass composition (126, 449). Both groupsagree that the depolarization of the whole oocyte does notcause MG channel activation, presumably because theexcess membrane area of the oocyte reduces the ability ofelectromechanical forces to create bilayer stress in thewhole cell membrane (448).

5. Effects of lipophilic compounds

Recent studies indicate that the same lipophilic com-pounds that gate gramicidin (i.e, lysophospholipids andarachidonic acid; Ref. 251) and MscS (i.e., trinitrophenoland chlorpromazine; Ref. 266) in lipid bilayers also gateSA-CAT channels and MG K1 channels (i.e., TREK/TRAAK) as well as modulate NMDA-R channels (53, 257,323, 372; see sect. VIIIF3). This indicates that these eukary-otic MG channels may also be gated by bilayer tensionsimilar to prokaryotic channels. However, because thecompounds may act on the channel protein itself or onCSK proteins, liposome reconstitution of MG channel ac-tivity will be required to confirm a bilayer mechanism.

6. Reconstitution of purified proteins in lipid bilayers

The single criterion that can disprove the tetheredmodel is demonstration that the purified MG channelprotein retains mechanosensitivity when reconstitutedinto lipid bilayers. At this time, the only eukaryotic chan-nel that this criterion has been applied is the ENaC (seesect. VIIIF2). ENaC was reconstituted into planar lipidbilayers (PLBs) and a hydrostatic pressure gradient ap-plied to stretch the bilayer (11, 202). This protocol wasreported to reveal an unusual mechanosensitivity involv-ing MS relief from Ca21 block (203). In the past, the abilityto develop tension in the PLB has been questioned be-cause of the presence of a lipid torus (158, 350). Althoughit was argued the torus may be exhausted by large bilayerexpansions (203), liposome patch-clamp studies of theENaC (i.e., similar to MscL) should be carried out toconfirm any tension sensitivity. More significantly, thereported mechanosensitivity seen in PLBs is not recapit-ulated in cell membranes (e.g., see Ref. 12). This leavesopen the possibility that contaminant proteins form chan-nels or that purification and reconstitution of the ENaCintroduces novel properties due to incorrect foldingand/or oligomerization. On the other hand, it may indicatethat interactions with the CSK suppress the propertiesseen in the bilayer (51, 340). Studies of the ENaC inCSK-deficient PMVs from Xenopus oocytes (447) maydiscriminate between these possibilities. The PMV prep-aration should also be helpful in testing the feasibility of

714 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 32: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

reconstituting other MG channels heterologously ex-pressed in the oocyte.

D. MG Channel Classification: Is There a Unifying

Mechanism for Activation and Inactivation

of MG Channels?

Like voltage- and receptor-gated channels, MG chan-nels can be classified according to differences in their ionselectivity, conductance, and/or pharmacology. However,a more meaningful classification relates to their responseto mechanical stimulation. For example, different MGchannels vary in their sensitivity and response to pressureand suction applied to the membrane patch. A commonclassification is whether they are opened (i.e., stretch-activated channels; SAC) or closed (i.e., stretch-inacti-vated channels; SIC) by membrane stretch (143, 288). Thisclassification refers to steady-state responses and ignoresdynamic behaviors such as adaptation in which sustainedmechanical stimulation may close SACs and open SICs(see below). Both SACs and SICs respond symmetricallyto pressure and suction (i.e., PASA and PISI, respectively,see Fig. 11), indicating that it is membrane tension thatactually activates or deactivates the channel. This sym-

metrical behavior may be mediated by tension developedin the bilayer or in CSK/EC elements that lie parallel to theplane of the bilayer. However, there are also MG channelsthat respond asymmetrically to pressure and suction andtherefore may constitute additional classes. The firstchannel reported to display asymmetrical responses wasa MG cation channel in rat astrocytes that was activatedby pressure and inactivated by suction (PASI) (33–35).Subsequently, other groups reported similar behavior forMG cation channels in toad smooth muscle cells (178) andrat endothelial cells (191, 259). Recently, a possible fourthclass of channel has been identified (PISA) in which suc-tion activates a neuronal K1 channel but pressure haslittle effect (257, 323).

Figure 11 summarizes the four basic types of channelbehavior in terms of idealized Boltzmann distributions.The asymmetrical responses to suction and pressure (Fig.11, C and D) may arise with either a bilayer or tetheredmodel of gating. For example, in the case of audio-vestib-ular hair cells, displacement of the stereocilia in oppositedirections with respect to the graduated stereocilia axiseither opens or closes the MG channels. This polarity-dependent behavior has been explained in terms ofstretching or relaxing the extracellular gating spring (i.e.,

FIG. 11. Idealized Boltzmann distributions for differentclasses of channel behavior seen in response to suction andpressure applied to the membrane patch. A: the stretch-activated channel (SAC) is symmetrically activated by pres-sure and suction (i.e., PASA), indicating it is responding tomembrane tension. B: the stretch-inactivated channel (SIC)is symmetrically inactivated by suction and pressure (PISI).C: a channel that is activated by pressure but inactivated bysuction (PASI). D: a channel inactivated by pressure andactivated by suction (PISA).

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 715

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 33: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

tip links, see sect. VIIIB2). Perhaps analogous CSK/ECstructures that lie perpendicular to the plane of the bi-layer, and are stretched in one direction and compressedin the other direction, are responsible for the asymmetri-cal responses to pressure and suction. However, asym-metrical effects may also arise in a bilayer model if pres-sure and suction cause differential expansion andcompression of the inner and outer monolayer duringpatch movements. The following simple calculation indi-cates that this source of asymmetry may be significant. Ifsuction deforms the patch from a flat circular disk to ahemisphere with a radius of curvature of 1 mm, then for amembrane thickness of 5 nm, the radii of curvature of theouter and inner monolayers will be 1 and 0.995 mm,respectively. This will result in monolayer areas of 6.28and 6.22 mm2, respectively (the reverse changes would beexpected with a pressure-induced deformation). If themonolayers are coupled so that they cannot slide past oneanother and the number of lipid molecules in each mono-layer remains fixed, then this would translate into propor-tional differences in both the area and the thickness ofeach monolayer. The difference, although only 1%, may besignificant considering that the bilayer can rupture withdilations of ;2%. If it is further assumed that the mech-anosensitivity of the channel arises because of differ-ences in protein bilayer mismatch associated with differ-ent conformations (see sect. III), then a unifyingmechanism may explain all forms of MG channel behaviorseen in Figure 11. For example, the symmetrical SAC (Fig.11A) and SIC (Fig. 11B) channel activity may arise be-cause positive and negative mismatch that is symmetricalwith respect to each monolayer favors opposite but sym-metrical shifts in the closed-open distribution withchanges in membrane curvature. The symmetry in themismatch (i.e., at the external and internal interface)means that the channel would respond the same to eithersign of curvature, since in one direction one monolayer iscompressed and the other monolayer expanded, and viceversa in the other direction. On the other hand, a channelthat displays asymmetrical mismatch with respect to eachmonolayer might be expected to respond differentiallydepending on the direction of membrane curvature (Fig.11, C and D). In one case it could be activated by positivecurvature and inactivated by negative curvature, and viceversa in another case. It may be possible to test thisunified hypothesis by examining specific lipids that pro-duced either positive or negative curvature when they areonly allowed to partition into one monolayer.

E. Rapid Adaptation of MG Channel Activity

Adaptation to sustained stimulation is an importantfeature of many sensory receptors and is critical in allow-ing a receptor to ignore continuous or static stimuli and

respond to transient or dynamic stimuli. Specific mech-anoreceptor functions such as the pulsatile pressure sen-sitivity of arterial baroreceptors, the movement (acceler-ation) detection by vestibular hair cells, and the highvibration sensitivity of certain tactile receptors all dependon adaptation to maintain their high dynamic sensitivityover a broad stimulus domain. Although adaptation canarise at any stage of the transduction process, it is clearthat single MG channel activity can display adaptation tosustained mechanical stimulation (35, 147, 154, 155, 168,226, 285, 382). Figure 12 illustrates rapidly adapting MGchannel currents recorded from an inside-out patch of aXenopus oocyte (Fig. 12A), an isolated liposome patch inwhich MscL was reconstituted (Fig. 12B), and from amouse utricle hair cell (Fig. 12C). The decay of the oocyteMG current is well approximated by an exponential witha time constant of ;100 ms similar to the initial fast decayof the hair cell current but 10 times faster that the decayof MscL activity.

Because mechanical gating arises from the channelprotein being sensitive to some mechanical-induced de-formation (i.e., either in the bilayer or in CSK/ECM ele-ments), then adaptation could arise because of a relax-ation in the force causing the deformation or a relaxationin the sensitivity to that deformation. Consider the sim-plest case of a two-state channel in which the rate con-stants for channel opening (b) and closing (a) are dis-placement sensitive (i.e., for a tethered MG channel) ortension sensitive (i.e., for bilayer-gated MG channel). Theprobability of the channel being open (Po) will be given by

Po 5 1/~1 1 K ! (8)

where

K 5 b/a (9)

or in terms of displacement

K 5 K0es~xo2x! (10)

where K0 is the equilibrium constant when the displace-ment x is equal to the set point x0 and determines thenumber of channels open at zero relative displacement,and s is the sensitivity to the relative displacement change(x0 2 x). For a bilayer-gated channel, we could substitutedisplacement with area change. An exponential time re-laxation in either s or x0 can produce the same adaptingMG currents (see Ref. 155). However, in one case therewill be a reduction in sensitivity (i.e., a change in the slopeof the response-displacement, Po-x, Boltzmann) while inthe other case there will be a shift along the x-axis with nochange in slope. [Note that in the simple and probablyunrealistic two-state kinetic scheme (150, 260), the shift

716 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 34: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

and shape changes are clearly independent. However, in athree-state scheme, a change in set point may change theshape as well as produce a shift in the relation.] The lattermechanism is true adaptation because sensitivity is main-tained, whereas the other mechanism is more akin toreceptor desensitization or voltage-gated channel inacti-vation, where the stimulus must be removed for sensitiv-ity to recover. For MG channel activity in the hair cell, thepredominant effect of adaptation is a shift in the Po-xcurve along the x-axis (8). Similarly, double steps of suc-tion or pressure (Fig. 13) indicate that after oocyte MGchannels have fully adapted they retain the same sensi-tivity to reactivation (154).

A number of factors have been shown to alter therate of adaptation of MG channel activity in hair cells andoocytes. In both cell types, membrane depolarizationcauses a slowing of adaptation, and in the specific case ofthe oocyte MG channel, the slowing is a monotonic func-tion of voltage with ;150 mV depolarization causing ane-fold decrease in the rate of decay. However, unlike

adaptation in audio-vestibular hair cells (9, 74, 98a), theoocyte adaptation and its voltage dependence is Ca21

independent (154). The oocyte adaptation mechanismmay be located in the bilayer where it senses the electricfield or alternatively coupled to a membrane protein thattransmits electric field effects through a conformationalchange. Another notable difference between the hair celland oocyte MG channel adaptation is the former’s direc-tional sensitivity. For example, negative displacement ofthe hair bundle (i.e., toward the short-ended stereocilia)tends to turn off the channels, but then when the stimulusis removed, a rebound activation of channels occurs (Fig.12C). This phenomenon presumably arises because adap-tation results in the Po-x relation being transiently shiftedto the left so that more channels will be open at zerodisplacement. In contrast, the oocyte MG channel showsno directional sensitivity to suction/pressure stimuli andno rebound channel reopening after the stimulus is re-moved. This difference most likely reflects the absence inthe oocyte of the hair cell Ca21-sensitive myosin motor

FIG. 12. Adaptation of MG channel ac-tivity in response to sustained stimulationmeasured in different membrane systems.A: inside-out patch recording from a Xeno-

pus oocyte in response to a 10-mmHg suc-tion step measured at a patch potential of2130 mV. The MG channel activity turnedoff with a time constant of 163 ms. B: MscLactivity in a liposome patch in response to asuction step of 40 mmHg measured at apatch potential of 210 mV. The MscL activ-ity turned off with a time constant of ;2 sto leave some residual activity. [Modifiedfrom Hase et al. (168).] C: whole cell re-cording from a mouse utricular hair cellthat was deflected (;1 nm) with a fluid jetunder pressure-clamp control. There wasan initial fast decay of current (i.e., over in,50 ms) followed by a sustained current.When the pressure pulse was turned off,there was a rebound outward current, indi-cating the MG channels at rest had beenturned off. [Modified from Holt et al.(183a).]

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 717

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 35: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

that is proposed to actively reset the gating spring tension(see below).

For the hair cell, the voltage sensitivity of adaptationderives from voltage-dependent Ca21 influx through theMG channel (9, 74, 189). However, within this framework,two distinctly different molecular mechanisms have beenevoked to explain how Ca21 might mediate adaptation. Inone mechanism, internal Ca21 interacts directly with thechannel protein causing it to favor the initial closed stateconformation C1 (i.e., in the kinetic scheme C17 C27 O,see Ref. 260), thus reducing the tension sensitivity be-cause it would require more tension to put the channel inC2. As the channels close, Ca21 sequestering processes

lower internal Ca, and the C2 conformation is favored(73). This model has been recently revised to specificallyaccount for the initial fast phase of adaptation (t ; 1 ms)that displays symmetrical kinetics (i.e., mirrored images)for small positive and negative displacements and is in-sensitive to myosin ATPase inhibitors such as vanadate(440). Furthermore, the fast kinetics of this phase con-trast with the slower cycle time of a myosin ATPasemotor that may contribute to the slow phase (t ; 50 ms)of adaptation.

In the second mechanism, the internal Ca21 reducesthe gating tension by causing a movement (i.e., by slip-page of a myosin motor) of the tip link anchoring point tothe MG channel, thus relaxing the tip link spring tensionand thereby favoring the closed states. As MG channelsclose and internal Ca21 falls, active motoring of the an-choring point retensions the gating spring. Consistentwith this model, it has been shown that myosin ATPaseinhibitors block the slower phase of adaptation (196, 440).It therefore seems possible that a combination of distinctmechanisms may contribute to adaptation. Overall evi-dence would seem to favor the second model in which thepredominant effect of voltage, as with adaptation, is toshift the Po-x curve along the x-axis rather than reduce itsslope (i.e., sensitivity) (8). On the other hand, evidenceindicates that Ca21 may directly interact with the channelin a number of ways to either block the channel or inducechannel conformational changes (see Ref. 155). Most re-cently, it has been demonstrated that external Ca21 af-fects tip link elastic properties as well as their integrity(263). In contrast to MG channel adaptation in the oocyte,there is no intrinsic voltage sensitivity in either model ofhair cell adaptation (i.e., it comes from voltage-dependentCa21 influx). However, both oocyte and hair cell MGchannels show evidence of an intrinsic voltage-dependentconformational change that is proposed to underlie thevoltage-dependent amiloride channel block (234, 343).The relationship between this voltage-dependent confor-mation change and other aspects of mechanotransductionhas yet to be determined.

In oocyte patches, it is possible to irreversibly abol-ish adaptation of MG channel activity by either repetitiveapplication of moderate stimuli (Fig. 14A) or by applica-tion of a single strong stimulus. This apparent fragility ofadaptation (to sealing and stimulation protocols) explainswhy some groups have reported stationary kinetic behav-ior for the Xenopus oocyte MG channel (154, 442). Inter-estingly, it illustrates an example where a single MGchannel may be changed from a highly phasic receptorinto a tonic receptor most likely by a change in theextrinsic properties of the membrane (see below). On thebasis of the observation that during the loss of adaptationthe membrane patch could be seen to be decoupled fromthe underlying cytoskeletal structures (154), it was sug-gested that viscous elements (dashpots) in the cytoskel-

FIG. 13. After adaptation, MG channel activity in oocyte membranepatches can be reactivated by increase in suction or pressure. A: MGchannel activity in response to an initial 8-mmHg suction step com-pletely adapted but could be reactivated by a 16-mmHg suction step.Pipette solution contained (in mM) 100 KCl, 2 CaCl2, and 10 HEPES,with a patch potential of 2100 mV. [Modified from Hamill and McBride(154).] B: MG channel activity in another patch completely adapted to a10-mmHG pressure but could be reactivated by an increase to 20 mmHg.Pipette solution contained (in mM) 100 KCl, 10 EGTA, and 10 HEPES,with a patch potential of 2130 mV. Note a larger tip pipette was used inA (;4 mm) than in B (;2 mm). This accounts for the difference in peakcurrents. [Modified from McBride and Hamill (271).]

718 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 36: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

eton become frozen or decoupled without disconnectingthe gating springs (154, see also Ref. 35). Consistent withthis idea was that MG channel activity recorded in CSK-deficient plasma membrane blebs or vesicles displayedlittle or no adaptation to sustained mechanical stimula-tion (447).

The above studies point to EC/CSK proteins tetheredto the channel as being critical for adaptation of MGchannel activity. However, recent studies of MscS in E.

coli protoplasts (226) and MscL in liposomes (168) indi-cate that channel activities may display adaptation in theabsence of CSK proteins (Fig. 12C). The adaptation of theE.coli channel activity is slower than that in Xenopus

oocytes or hair cells (i.e., t ; 1 s vs. 50–200 ms at similarvoltages), is not voltage sensitive, and is not abolished bystrong repetitive mechanical stimulation of the patch.Limited proteolysis applied to the cytoplasmic side of thepatch reduces the number of MscS without removingadaptation, whereas stronger proteolysis abolishes mech-anosensitivity (226). Note this proteolytic inhibition of

MscS activity is opposite to the potentiation of MscLactivity (3). Given that a bilayer rather than a tetheredmechanism gate MscS, it was proposed that adaptationmight be associated with insertion of the cytoplasmicdomain of MscS in the bilayer (226). An alternative expla-nation is that the expansive force in the bilayer (i.e., dueto bilayer bending, see sect. VIIID) relaxes as the mono-layers slip past one another or lipids move from onemonolayer to the other may (see Ref. 350). In this case,adaptation should be sensitive to the lipid make up of themonolayers and the degree of coupling between the twomonolayers.

F. Structure of Eukaryotic MG Channels

Recent progress has been made in identifying severaleukaryotic genes that encode MG or putative MG chan-nels. However, so far none of the structurally identifiedchannels has been characterized in the same detail as

FIG. 14. The voltage-dependent adaptation of MG channel activity in the Xenopus oocyte is mechanically fragile. A:repetitive pulses of suction of 30 mmHg applied to the patch 30 s apart results in a progressive loss in adaptation andirreversible reduction in the MG channel current. [Modified from McBride and Hamill (271a).] B: loss of adaptation dueto repetitive suction pulses also dramatically alters the voltage-sensitive response of the MG channel currents. The initialresponse to the suction step (i1) shows highly asymmetrical currents at 2100 mV/100 mV due to adaptation only at thenegative potential (reversal potential was ;0 mV). The response to the third suction step (i3) shows a more symmetricalcurrent due to the irreversible loss of adaptation. The larger current noise at 2100 mV compared with 100 mV reflectsmore rapid closing of MG channels at the negative potential (i.e., shorter channel lifetime). This voltage dependence isnot fragile and presumably reflects an intrinsic property of the channel protein. [Modified from Hamill and McBride(154).]

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 719

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 37: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

MscL, and in some cases (e.g., MECs, NOMPC), patch-clamp measurements have yet to confirm single MG chan-nel activity. Nevertheless, the preliminary findings indi-cate that a number of different molecular designs andmechanisms have evolved to confer mechanosensitivityon membrane ion channels. Below we discuss their keystructural (i.e., topology, subunit stoichiometry, and pro-tein-protein consensus domains) and mechanistic (i.e.,bilayer vs. tethered) features.

1. A Ca21-permeable cation MG channel in yeast

Mid1 was originally identified from a mutant screenof yeast (Saccharomyces cerevisiae) induced to undergoCa21-dependent mating-induced differentiation (199).Cloning and sequencing of mid1 indicated an integralmembrane protein of 548 amino acids with at least 4 (andpossibly 6) transmembrane domains (H1-H4) and 2COOH-terminal cysteine-rich regions that may be in-volved in protein-protein interactions (i.e., that may local-ize and/or activate the channel) (208). Although Mid1does not appear closely related to other membrane pro-teins, its H4 sequence shows similarities to the S3/H3hydrophobic segment of voltage-gated Ca21 and Na1

channels (208 and references therein). Heterologous ex-pression of mid1 in Chinese hamster ovary cells wasreported to increase resting Ca21 membrane permeabil-ity, whereas stretch of the cell substrate (i.e., a siliconelastic membrane) resulted in elevation in [Ca21]i (208,but see Refs. 207, 208a for corrections). Patch-clamp stud-ies indicate that Midl is a SA-CAT channel with a highCa21 permeability (PCa/PK ;7). Although the Mid1 chan-nel conductance (32 pS in 150 mM CsCl) is similar to thatof a SAC previously characterized in S. cerevisiae proto-plasts (36 pS in 170 mM CsCl), the protoplast channeldisplays a lower Ca21 permeability (PCa/PK ;0.54) andsignificant anion permeability (147). Furthermore, recentresults (X. L. Zhou, C. Palmer, and C. Kung, personalcommunication) indicate that protoplasts prepared fromS. cerevisiae with the mid1 gene deleted still express theMG channel activity of the wild-type yeast protoplast(147). Although no homologs of Mid1 have been identifiedin higher eukaryotes, a potential Mid1 homolog, yam8,has been cloned from the fission yeast S. pombe anddemonstrated to partially rescue S. cerevisiae from themid1 mutant phenotype (402).

2. MECs: putative MG channels in C. elegans

Members of the Mec (mechanosensory abnormal)family in C. elegans were originally identified in mutantscreens for touch-insensitive animals (57, 58, 59, 93, 122,161, 403). The initial clue that specific mec genes mayencode MG channels was that while recessive mutationsin mec-4 resulted in touch insensitivity, dominant muta-tions in the same gene resulted in swelling-induced de-

generation and lysis of the mechanosensory neurons (i.e.,consistent with continuously open channels) (93). Thecloning of mec-4 demonstrated that it was homologous todeg-1 that mediates swelling induced degeneration inother neurons (403). Together, Mec-4, Mec-10, and Deg-1were proposed to belong to a protein superfamily calleddegenerins (DEGs). Other members of the MEC/DEGfamily in C. elegans include unc-8 (uncoordinated), ex-pressed in motor neurons and required for normal loco-motion; unc-105, expressed in muscle and required forstretch sensitivity; and flr-1 (fluoride resistance) requiredfor normal defecation rhythm (248, 401, 404).

As yet, no MEC family member has been directlydemonstrated to form MG channels. However, when unc-105 genes with gain-of-function mutations (i.e., predictedto cause constitutive channel activation) are expressed inXenopus oocytes, they form spontaneous opening cationchannels that display a variety of conductances (i.e., 2–30pS) with no Ca21 permeability (121). Amiloride blocks theUnc-105 channels, apparently by binding to a single site inthe pore. This is similar to the amiloride block of ENaC(123) but different from the block of MG channels inXenopus oocytes and hair cells where a voltage-depen-dent conformational change is proposed to expose mul-tiple amiloride binding sites outside the membrane field(234, 343).

On the basis of sequence similarities, several otheramiloride-sensitive Na1 channels have been classified asMEC/DEG family members. These include the ENaCs (50,123), acid-sensing ion channels (ASIC) (424a), molluscanFMRFamide-gated channels, and Drosophila Na1 chan-nels expressed in gonads [dGNaC1 or RPK (for rippedpocket) and in multiple dendritic neurons dmdNaC1 orPPK (for pickpocket) (1, 424a)]. The RPK and PPK wereidentified in Drosophila database searches that initiallyrecognized a sequence homologous with a conserved re-gion (i.e., in M2) in other MEC/DEG genes (1). All mem-bers of the MEC/DEG family are characterized by 1) twotransmembrane domains (M1 and M2) with a single P-loop structure believed to line the channel pore, 2) intra-cellular NH2 and CCOH termini, and 3) a large extracel-lular loop (e.g., see Refs. 49, 403). In terms of secondarystructure and membrane topology (but not primary aminoacid sequence), the DEG/MEC family members are similarto MscL from prokaryotes (see sect. IVB) and also to theATP-gated (P2X receptor) channels (302).

Of all the DEG/MEC family members, the ENaC hasbeen studied in greatest detail (50, 123). ENaC in mostepithelia is composed of three homologous subunits: a, b,and g, with each subunit proposed to contribute to thepore walls (357a). A fourth subunit, d, mainly expressed inthe testis and ovaries, shows properties similar to a andmay form heteromultimers with b and g in these tissues(82a). Controversy exists regarding the subunit stoichi-ometry of the ENaCs, with some studies indicating a2bg

720 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 38: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

(113) and others a3b3g3 (371a). Because the discrepancymay relate to different detergents used to solubilize andpurify the ENaC protein, the results of a recent study thatdid not attempt to purify the channel complex are signif-icant (105a). Instead, freeze-fracture electron microscopywas used to visualize the ENaC complexes expressed onthe surface of Xenopus oocytes. The individual ENaCcomplex was seen as a square particle of ;24 nm2, indi-cating 17 6 2 transmembrane a-helices (i.e., assuming 1.4helix/nm2) or 8 or 9 subunits (105a).

The fact that ENaC is a member of the MEC/DEGprotein family has naturally led to the hypothesis that it isalso a MG channel. Although some electrophysiological(11, 202, 203, 219) and immunocytochemical localizationstudies (94, 118, 148a) support this hypothesis, variousmethodological concerns and problems with interpreta-tion have been raised (see Refs. 12, 158, 339, 350, 427).For example, the reported mechanosensitivity of ENaCsreconstituted into bilayers (11, 202, 203 but see sect.VIIIC6) is not evident in the ENaCs expressed in cellmembranes (12, 319). Furthermore, the report that heter-ologous expression of a-ENaC in LM(TK) fibroblasts re-sults in MG channels (219) appears problematic becausethat cell line expresses endogenous MG channels withsimilar properties to the presumed a-ENaC (428). Finally,the results indicating that the abg-ENaC subunits trans-fected in oocytes express a volume-sensitive inward cur-rent (cf., Refs. 11, 205) indicate, if anything, that the ENaCis a SIC, since cell shrinkage activates, while oocyte swell-ing inactivates the current (32, 288). The contribution of arecently reported endogenous shrinkage-activated cur-rent that is not mediated by the SA-CAT channel remainsunclear (448, 449).

Despite the lack of compelling biophysical evidencefor the ENaC being a MG channel, there are severalimmunocytochemical studies that have localized ENaCsubunits in mechanosensory cells (94, 118, 148, 148a, butsee Refs. 128, 342). One study reported expression of theb- and g-ENaC subunits, but not the a-subunit, in barore-ceptor nerve terminals (94). Because b- and g-subunitscannot form channels in the absence of the a-subunit(50), another unidentified subunit must combine with theENaC subunits to form the baroreceptor MG channel (94).The unidentified subunit may also confer Ca21 permeabil-ity on the channel (e.g., see Refs. 363, 391), since theabg-ENaC is Na1 selective (123). Interestingly, a morerecent study has reported that a-, b-, and g-ENaC subunitsare all localized in the perikarya of the trigeminal mech-anosensory nerve terminals innervating the rat vibrissalfollicle sinus complex (118). Furthermore, the subunitsare colocalized with stomatin, a RBC protein homologousto C. elegans Mec-2 proposed to form the molecular linkbetween the Mec4/Mec10 channel and the CSK (Fig. 10A,Refs. 195, 258). Significantly, the ENaC subunits and sto-matin are not expressed in other specialized mechanore-

ceptors, although they are expressed in autonomic nervecells (118). This may indicate that other MG channels(e.g., NOMPC, see below) or mechanisms are involved inmediating mechanotransduction in different mechanore-ceptors or that the ENaC has another role in specificneurons not directly related to mechanotransduction.

3. NOMPC: putative MG channels in Drosophilaand C. elegans

The gene nompC was identified in mechanorecep-tive-defective (i.e., uncoordinated) Drosophila mutantsthat also showed an absence or reduction in the mech-anoreceptor potentials (i.e., no mechanosensory poten-tial) recorded from external sensory bristles (215, 425).NompC predicts a protein of 1,619 amino acids of which1,150 NH2-terminal residues consist of 29 ankyrin (ANK)repeats. ANK repeats have previously associated withforming complexes between membrane and CSK proteins(425) and may couple the channel to a CSK protein thatgates and/or anchors (i.e., clusters) the channel in a sub-cellular region of the dendrite. The remaining 469 resi-dues of NOMPC share some sequence identity (;20%) tothe transient receptor potential (TRP) and TRP-like(TRPL) class of ion channels (67, 166a). The TRP chan-nels also share structural similarities with vertebrate volt-age-gated Ca21 and Na1 channels, including six trans-membrane domains (S1-S6) and a predicted pore regionbetween S5 and S6. However, they lack the positivecharged amino acids in the S4 region that confers voltagesensitivity on the channels.

Three nompC mutants with severely reduced MGcurrents had single nucleotide changes that introducedpremature termination codons into their sequence. How-ever, a fourth mutant (nompC4) with normal-amplitudeMG currents but accelerated adaptation involved a cys-teine to tyrosine substitution at amino acid residue 1400(i.e., between the S3 and S4 domains). This may be acritical residue underlying protein-protein or protein-lipidinteractions that influence adaptation kinetics. Interest-ingly, screening of a C. elegans library revealed a homolog(Ce-NOMPC) that showed 40% identity with NOMPC andincluded the S1-S6 domains and the 29 ANK repeats.Furthermore, NOMPC and Ce-NOMPC were shown to beselectively expressed in ciliated mechanoreceptors (425,see also Ref. 209), whereas MEC/DEGs are expressedonly in nonciliated touch cells (1, 403).

Heterologous expression and patch-clamp studieshave yet to demonstrate that NOMPC actually forms a MGchannel. Therefore, it remains unclear whether NOMPCcan exist as a functional homomultimeric complex orwhether it must interact with other protein subunits toform a MG channel (i.e., analogous to the MEC/ENaCs).The similar properties of mechanotransduction in Dro-

sophila bristles and vertebrate hair cells (i.e., submilli-

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 721

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 39: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

second latencies, high directional sensitivity, and fast adap-tation) may indicate a common tethered mechanism ofgating (Figs. 9 and 10B, Ref. 411). However, liposome recon-stitution or expression in CSK-deficient PMVs will ultimatelybe required to exclude a bilayer type mechanism.

4. SIC: a stretch-inactivated channel cloned

from rat kidney

The capsaicin or vanilloid receptor (VR1) originallycloned from dorsal root ganglia (DRG) encodes a Ca21-permeable, nonselective cation channel and is a member ofthe TRP family (53a). In addition to being activated bycapsaicin, it is also activated by heat, consistent with a rolein nociception. Recently, a homolog of VR1 was cloned fromrat kidney and shown to encode a large-conductance chan-nel (250 pS) that was weakly cation selective and permeableto Ca21 (393). Furthermore, the channel appeared to be aSIC because it was activated by cell shrinkage and inhibitedby cell swelling (i.e., of sic transfected Chinese hamsterovary cells). The sic gene encodes a 563-amino acid proteinwith 6 transmembrane domains and a single pore regioncommon to VR1. However, the NH2 terminal of the SIC isshorter than the VR1. Although it was reported that SIC wasturned off by increased suction applied to the patch, itshowed particularly slow recovery of activity (.1 min) afterremoval of the suction. The effect of increased pressure onchannel activity was not reported (393). Other SICs in snailneurons (288), skeletal muscle (116), smooth muscle (178),and supraoptic neurons (311, 350a) display a significantlylower conductance (,50 pS), so it remains to be demon-strated whether they are structurally related to the rat kid-ney SIC.

5. TREK-1 and TRAAK: MG K1 channels

in mammalian neurons

Recently, a new protein superfamily of weakly inwardrectifying K1 channels has been identified that is character-ized by four transmembrane domains and two pore-formingregions called P domains. The first family member identifiedwas TWIK-1 (i.e., for tandem of P domains in a weaklyinward rectifying K1 channel) (243). Subsequently, twoother members, TREK-1 (related to TWIK-1) and TRAAK(opened by arachidonic acid), have been demonstrated toform MG K1 channels (257, 323). For example, whenTREK-1 is expressed in oocytes, it shows a comparablesensitivity to suction as the endogenous channel but is lesssensitive to pressure activation (323). Unlike the endoge-nous oocyte MG channel (448, 449), osmotic swelling andshrinkage increase and decrease, respectively, whole cellTREK currents. This may indicate the two channels arelocalized in different membrane regions (e.g., nonvilliatedmembrane vs. microvilli, respectively) so that they experi-ence different tensions (i.e., due to the different radii ofcurvature, see Ref. 448). Alternatively, the two channels may

involve different mechanisms of activation. For example,while the endogenous channel is most likely bilayer gated(448, 447), TREK-1 may be tethered. However, neither cy-tochalasin nor colchicine treatment was found to alterTREK activity (323, see below for TRAAK). Arachidonic acidand trinitrophenol (TNP) also open TREK-1. TNP is an an-ionic crenator of RBCs that activates bacterial MG channels.It may be that these amphipaths act on MG channels byaltering the mechanics of the bilayer. However, direct ef-fects on the channel proteins have yet to be excluded. It isalso possible that stretch activation is mediated by stretch-induced release of arachidonic acid (e.g., via phospholipaseactivation). However, the lack of effect of delipidated bovineserum albumin, which should intercept fatty acid transmis-sion (i.e., unless the release and receptor sites are extremelyclose), indicates stretch more likely acts by mechanicaldeformation of the bilayer. Kinetic measurements and thedemonstration that mechanosensitivity is preserved in ex-cised patches would reinforce this idea. Structure-activitystudies of TREK-1 indicate that the COOH-terminal region isnecessary but not sufficient for mechanosensitivity. Evenmore interesting, a charge cluster region resembling theCOOH terminus of the MscL protein (RKKEE) appears cru-cial for the mechanosensitivity of both MG channels, indi-cating that a protein domain important in bacterial MGchannel gating has been conserved in the eukaryotic chan-nel (323).

TRAAK is similar to TREK-1 in that it can be activatedby arachidonic acid, but it is only activated by suction in thecell-attached patch configuration and positive pressure inthe outside-out configuration (257). This asymmetrical sen-sitivity indicates that a specific membrane curvature (i.e.,convex) may be required for the channel to be activated (seesect. VIIID). Consistent with this idea is the observation thatexternal but not internal application of TNP activatesTRAAK. TNP is negatively charged and presumably acts bypartitioning into and expanding the less negative externalmonolayer to induce a convex curvature. Unlike TREK-1(323), it was reported that both colchicine and cytochalasinenhance TRAAK activity, indicating that these CSK elementsmay constrain tension development in the bilayer. As withTREK-1 (and MscL), there is a charged cluster region in theCOOH terminus that is critical for both arachidonic acidactivation and mechanosensitivity. Finally, although TRAAKis expressed in DRG neurons, indicating a possible role inmechanotransduction (257), it is also widely expressed inthe brain, spinal cord, and retina, indicating a more generalfunction in neuronal excitability (257a).

6. Orbiter, mercury, and gemini:

mechanotransduction mutants in zebrafish

Specific genes encoding MG channels in specializedmechanoreceptors of vertebrates have yet to be identi-fied. However, recent studies of zebrafish have identified

722 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 40: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

a group of mutants that show defects in balance andswimming patterns (299). Of particular interest are threemutants, orbiter, mercury, and gemini, that have normalhair cell morphology (i.e., no signs of hair cell degenera-tion or hair bundle disorganization) and normal synaptictransmission but do not express a “microphonic” currentwhen the hair bundle is mechanically defected. Conse-quently, it was suggested that these genes may encodeMG channels or ancillary proteins associated with trans-mitting force to the channels (299).

IX. MECHANOSENSITIVE ELEVATION

OF INTRACELLULAR CALCIUM

Changes in [Ca21]i regulate a wide variety of cellularprocesses, including cell growth and differentiation, cellmotility and contraction, intercellular coupling, synaptictransmission, fertilization, apoptosis, and necrosis. It istherefore significant that mechanical stimulation triggerselevation in [Ca21]i in many cells, including cardiac (366)and smooth muscle cells (218), fibroblasts (22), glia (62),osteoblasts (327), vascular endothelial cells (VECs) (88,293, 294, 307, 364, 367), epithelial cells (30, 240, 435), haircells (196), and neurons (65, 132, 332, 363, 391). In prin-ciple, an increase in [Ca21]i may be caused by one or acombination of mechanisms involving 1) increased Ca21

influx from the external solution, 2) increased Ca21 mo-bilization from internal stores, and/or 3) decreased Ca21

efflux from the cell. Thus mechanical stimulation mayincrease [Ca21]i by 1) gating MG channels located in theplasma or internal membranes, 2) affecting MS enzymesthat regulate second messengers that in turn modulateCa21 channels, 3) causing MS release of transmitter (e.g.,ATP) that then activates Ca21-permeable channels, and/or4) reducing Ca21 efflux through specific pathways (e.g.,Na1/Ca21 exchanger). Direct evidence exists for the firstthree mechanisms.

A. MS Ca21 Influx Mechanisms

An early patch-clamp study of VECs identified aCa21-permeable MG channel (235). Subsequently, MGCa21 channel activities have been proposed to mediateMS elevation of [Ca21]i in heart cells (366), sensory neu-rons (134, 322, 363), hair cells (163 ), GH3 cells (65),fibroblasts (22), and keratocytes (240). Apart from directpatch recording of channel activity, the other criteria usedto evoke the MG Ca21 channel mechanism has been theblock of [Ca21]i elevation by removal of external Ca21 orby addition of Gd31 (293, 294, 312, 366, 367, but seebelow). In addition to mediating direct Ca21 influx, theMG channel may also indirectly activate voltage-gatedCa21 channels by causing depolarization (65, 218). TheCa21 entry through MG and voltage-gated Ca21 channels

may then trigger either Ca21-induced Ca21 release (294)or activate phospholipase C activity and inositol 1,4,5-trisphosphate (IP3)-sensitive Ca21 release (30, 39, 270),thereby further amplifying the increase in [Ca21]i. In turn,Ca21 store emptying may produce a sustained Ca21 influxthrough store depletion-activated (i.e., capacitative) Ca21

channels in the plasma membrane that refills the emptystores (321).

B. MS Release of Ca21 From Internal Ca21 Stores

Mechanical stimulation, even in the complete ab-sence of external Ca21, can elevate [Ca21]i by promotingCa21 release from internal stores (30, 62, 88, 197, 204,300a, 307, 435). Two classes of mechanisms, indirect anddirect, have been proposed to underlie this MS Ca21

release. Examples of indirect mechanisms include themechanical activation of either a MS phospholipase C (30,39, 270) or a phospholipase A2 (206, 241, 307) that resultsin activation of the IP3-sensitive Ca21 release channel.Interestingly, addition of Ca21 channel blockers (i.e.,Gd31, Ni21, nifedipine, and nimodipine) in the absence ofexternal Ca21 causes a larger MS increase in [Ca21]i,presumably because they block Ca21 efflux through MGchannels, voltage-gated Ca21 channels (30), and/or storedepletion-activated Ca21 channels in the plasma mem-brane (307). Although it is possible that permeation ofGd31 (and Ni21) through open MG channels could acti-vate the fura 2 used to monitor [Ca21]i and thereby pro-duce an apparent rise in [Ca21]i, this mechanism couldhardly explain the similar potentiation caused by dihydro-pyridines. Another type of indirect mechanism involvesMS release of transmitter (e.g., ATP) and activation ofreceptor-gated Ca21-permeable channels (see sect. XD).This mechanism has recently been shown to mediate themechanically triggered spread of Ca21 waves in prostatecancer cells (356a).

Evidence for a direct mechanical activation of Ca21

release from internal stores has come from the responseof VECs to osmotic swelling (204) and mechanical acti-vation by twisting magnetic beads attached to the cellsurface (300a). In the first case, Jena et al. (204) identifieda novel internal Ca21 store that can still release Ca21 inresponse to hypotonic stress even after the plasma mem-brane’s permeability barrier had been selectively dis-rupted with the detergent saponin. Under these circum-stances, mechanical forces are presumably transmitted tothe Ca21 channel in the endoplasmic reticulum (ER) bydirect osmotic swelling of the ER. It was also demon-strated that under these conditions removal of externalCa21 can rapidly deplete internal Ca21 stores, and highconcentrations of Gd31 can block the internal Ca21 re-lease channel. These last results indicate the need forcaution in using these criteria alone to implicate the Ca21

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 723

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 41: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

influx mechanism. In the second case, Niggel et al. (300a)demonstrated that VECs, astrocytes, and C6 glioma cellsdisplayed a fast transient increase in [Ca21]i in responseto twisting magnetic beads attached to the plasma mem-brane (but see below). Because the Ca21 increase alsooccurred in “sugar water” (i.e., ion-free solution), it didnot involve Ca21 influx. Instead, it was proposed thatCa21 was released from IP3-sensitive stores by differentcell-specific mechanisms. For example, in VECs, the Ca21

release appeared to be mediated by an MS increase in IP3

(based on inhibitor efffects), whereas in C6 glioma cells itappeared to be mediated by direct mechanical activationof MG channels in the ER. The exact pathway by whichforces are transmitted to the ER remains unknown. Inter-nal Ca21 release has been shown to be dependent onintegrins (327), microtubules (406), and/or actin micro-filaments (227). It has also been shown that the ER can beintimately associated with the plasma membrane (e.g.,Ref. 122a). At least in C6 cells, neither cytochalasin D norcolchicine was found to block MS Ca21 release (300a). Afinal question that arises concerns whether cells evernormally experience the mechanical stresses generatedby twisting beads that attached to their cell surface. Forexample, if the beads undergo phagocytosis and/or causereorganization of the surface membrane or CSK, theninternal membrane Ca21 stores may be made hypermech-anosensitive, analogous to MG channels in tightly sealedmembrane patches. For this reason, it is important toconfirm that other forms of mechanical stimuli (e.g., fluidshear stress, cell stretch, or osmotic swelling) producesimilar responses in the same cells.

X. MECHANOSENSITIVE RELEASE

OF TRANSMITTER

Although the major focus of research on mechano-sensitivity over the last 15 years has been on the ubiqui-tous MG channel, it has recently become clear that MSrelease of transmitter, most notably ATP, is as equallyubiquitous in eukaryotic cells (45, 63, 139, 280, 292). Fur-thermore, in specific cell types such as the Xenopus oo-cyte, where both MG channels and MS ATP release areexpressed, it is the MS ATP release that is more sensitiveto mechanical stimuli (262, 292, 448). Perhaps most sur-prisingly is that MS ATP release has been implicated invertebrate touch and stretch sensation (68, 292), two MSprocesses commonly assumed to be mediated by MGchannels. Although MS ATP release has yet to be con-firmed in mechanosensory neurons in situ, the questionnaturally arises how more than 50 years of research hasfailed to make the fundamental distinction between phys-ical and chemical mechanisms of transduction.

A. Historical Perspective

From a historical perspective, one can see how earlystudies of touch sensors may have been biased towardone mechanism or the other depending on the disciplineof study. For example, the anatomists Spencer andSchaumburg studying the ultrastructure of Pacinian cor-puscles (376) came to the conclusion that “a simple phys-ical hypothesis of transduction does not take account ofthe elaborate array of organelles present at the base ofeach axon process. These organelles include numerousclear-core vesicles having the appearance and propor-tions of synaptic vesicles.” They further noted “that theseclear-core vesicles in sensory axon terminals contain asubstance which is released during the postulated me-chanical distortion of an axon process and is then able toaffect the ionic conductance of the axolemma, is an in-teresting speculation.” On the other hand, electrophysi-ologists in favoring a physical mechanism appeared tobe more impressed by the speed and high-frequencyresponse of specific tactile receptors. For example,Gottschaldt and Vahle-Hinz (135) studying Merkel cellsconcluded “the characteristics of the vibratory responsescan hardly be explained in terms of chemosynaptic trans-mission. It is unlikely that a transmitter release mecha-nism could operate in phase with a vibratory stimulus of1,200 Hz or more. Also, at such high frequencies an accu-mulation of released transmitter would be likely.” Thatthe physical view has often prevailed would appear welljustified in specific sensory systems such as hearingwhere the frequency response can be as high as 100 kHz(70, 196). Indeed, mechanically and ATP-activated cur-rents in mouse outer hair cells have been shown to beindependent and differentially blocked by D-tubocurarine(132). Nevertheless, the presumption that moderately fastkinetics (;10 kHz) must necessarily mean a physicalrather than a chemical mechanism can no longer be jus-tified by experimental evidence. For example, it has beendemonstrated that transmitter release can be activatedwith latencies as short as ;150 ms (i.e., after the start ofthe action potential) and rise times in transmitter concen-tration can be as fast as 60 ms (40, 346). Furthermore, theidea that accumulation of transmitter may severely limitthe frequency response needs to be revised in the light ofrapid mechanisms for terminating the action of neuro-transmitters, most notably ATP. Specifically, it has beendemonstrated that nerve stimulation not only releasesATP but also soluble nucleotidases that would facilitatethe ATP breakdown (413). Although the kinetics of trans-mitter release and breakdown are still not rapid enough toaccount for the .20-kHz frequency response of auditorytransduction, they could be fast enough to mediate atactile vibration sensitivity in the 1–2 kHz range (135).Below, we examine the kinetics and other features of

724 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 42: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

mechanisms that may mediate chemical forms of mech-anotransduction.

B. Tension-Sensitive Vesicle

Recruitment/Exocytosis

Tension-sensitive vesiculation and recruitment canoccur in artificial bilayer vesicles and specific cell types.For example, a flaccid bilayer vesicle will tend to losesmaller membrane vesicles by a process of budding andvesiculation until it becomes a sphere (i.e., the lowestenergy state for a symmetrical structure under centripe-dal tension, Ref. 107). Similar membrane vesiculation/fragmentation also occurs in human and frog RBCs afterdisruption of their CSK (152, 284). Depending on a varietyof factors that determine the spontaneous curvature ofthe bilayer (see sect. IIC), the parent vesicle (or RBC) mayeither eject vesicles into the external solution or injectthem internally to produce an internal membrane reser-voir. Increased membrane tension will tend to inhibit thebudding (vesiculation) process by inhibiting the mem-brane invagination and neck formation between the par-ent membrane and the vesicle (78). On the other hand,increased tension should promote membrane fusion be-tween the shed internal vesicles and the parent membranebecause fusion (i.e., membrane area recruitment) willlower the energy (tension) of the system (see exceptionbelow). In other words, the surface tension (area) of theparent vesicle may be adjusted according to the freeenergy difference per lipid molecule in the reservoir andparent membrane (437).

Membrane tension homeostasis has been describedin isolated plant protoplasts (184, 185, 213, 230, 438, 457)and more recently evoked to explain membrane reorga-nization in animal cells (78, 79, 149, 174, 279, 331, 337, 427;for recent review, see Ref. 286a). In plant cells, turgorpressure ensures the resting membrane is fully distended(i.e., no excess membrane) so that a rapid regulatoryfeedback loop exists between tension and surface arearegulation. However, most animal cells maintain a stableexcess membrane area that buffers rapid fluctuations intension (see sect. VII and below). Membrane area adjust-ments in response to osmotic swelling and shrinkage havebeen directly monitored by changes in membrane capac-itance (80, 185, 427, 457). For example, swelling andshrinking of guard cell protoplasts cause discrete upwardand downward deflecting capacitance (2 fF) steps thathave been interpreted as reflecting osmotically drivenunitary membrane fusion (exocytotic) and fission (endo-cytotic) of vesicles of ;300 nm (185). However, in somestudies, membrane conductance changes have been re-ported to exactly mirror Cm changes (457). Although suchparallel changes may reflect the addition and removal ofchannels to the plasma membrane corresponding to mem-

brane insertion and retrieval events, they may also arisebecause of incorrect phase settings in which the imagi-nary capacitance changes (i.e., due to conductance in-crease) contribute to the apparent Cm (129a). Pharmaco-logical agents that block the conductance but not the Cm

increase may be used to exclude this form of artifact.Certainly, there are examples where osmotic swelling canactivate a conductance increase without accompanyingCm increase (137, 338a), and vice versa (80, 129a, 427).

Tension-sensitive vesicle recruitment also results inexocytosis, since the vesicle will release its contents uponfusion with the plasma membrane. Stretching or inflatingsome cells has been reported to promote or facilitateexocytosis/release (63, 149, 262, 309, 352a, 435). However,the response is not universal (375a). For example, infla-tion of mast cells (i.e., to ;4 times their volume) actuallycauses a reversible block of exocytosis that recoversrapidly with deflation (375a). In this cell, exocytosis mayrequire initial membrane dimpling to form the fusion porewith the mast cell granule, and tension presumably inhib-its this dimpling (375a). Xenopus oocytes can also beinflated (e.g., to more that twice their volume) withoutincreasing Cm and/or activating MG channels (448). Bothoocytes and mast cells display a large excess membranearea (i.e., $500%) that most likely protects their bilayerfrom tension changes except under the most extreme(pathological) conditions. In contrast, cells that respondto frequent changes in passive stretch and/or osmoticswelling (e.g., fibroblasts, urinary bladder epithelial cells,and mechanosensory neurons) may lack such a largeexcess membrane and thus rely on vesicle recruitment toaccommodate stretch and cell volume changes (245b,286a, 330). For example, during expansion of the urinarybladder, the epithelial cells undergo an initial smoothingout of surface folds on the urine-facing (i.e., apical) mem-brane followed by fusion of cytoplasmic vesicles with theapical membrane (245a). The membrane fusion/retrieval(i.e., as reflected in Cm changes) is reversible and occurswithin 5 min (245b). In comparison, snail neurons showan increase (;10%) in Cm that takes between 0.5 and 3.0min after exposure to a 50% hypotonic solution (80), andplant protoplasts undergo a pressure-induced increase inCm that occurs with a latency of ;100 ms and a rise timeof several minutes (457). In this case, membrane recruit-ment may be rate limited by random collisions betweenthe vesicle and the membrane. In both cells, the decreasein Cm after return to isotonic (or hypertonic) solution (80)or release of pressure took 5–10 min. Although thesekinetics may be adequate for preserving cell membraneintegrity during relatively slow and sustained changes inmembrane tension, they are too slow to enable responsesto rapid cell deformations or movements and certainlycould not transduce rapid oscillations in mechanical stim-ulation (i.e., necessary for tactile sensation). In contrast,stretch-facilitated transmitter release at synapses can dis-

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 725

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 43: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

play very fast kinetics (,1 ms, Ref. 63, see below), pre-sumably because the vesicles are docked next to the cellmembrane ready for rapid fusion. Another difference be-tween synaptic exocytosis and tension-dependent vesiclerecruitment is that the latter shows little or no Ca21

dependence (see Refs. 174, 184).

C. Stretch-Facilitated Transmitter Release

at the Vertebrate Motor Synapse

It has long been recognized that stretching skeletalmuscle promotes transmitter release from the motor syn-apse (198, 415). This stretch facilitation is functionallysignificant because it can amplify the spinal stretch reflex.Most recently, Chen and Grinell (63) have shown thatstretching frog muscle in the physiological range resultsin a 10% increase in release per 1% muscle stretch that isreflected in both increased frequency of miniature end-plate potentials (mepps) and amplitude of evoked end-plate potentials. The kinetics of the effect are extremelyrapid, with the development and decay of the enhance-ment occurring in ,1–2 ms. The stretch-induced enhance-ment of mepps is not dependent on Ca21 influx and is notblocked by Gd31 (63). At this stage, the enhancementmechanism remains unclear. However, the fast and sym-metrical on-off kinetics as well as the low temperaturesensitivity (i.e., Q10 ; 1) seem most consistent with adirect physical mechanism, although biochemicalchanges occurring within a highly confined space cannotbe excluded. The fact that the facilitation still occurswhen incremental changes in intraterminal Ca21 havebeen blocked (i.e., “clamped” at 100 nM) rules out stretch-induced Ca21 release (63). Because integrin antibodiesand binding peptides (arginine-glycine-aspartic acid,RGD) block the facilitation, mechanical forces may betransmitted intracellularly (via integrins) to alter the po-sition or conformation of molecules controlling release(see Fig. 12 in Ref. 63). Although the RGD peptides blockfacilitation, they do not prevent stretch-induced elonga-tion of the nerve terminal, indicating that other interac-tions between the muscle and the nerve remain intact butare incapable of transmitting stretch to the release mech-anism. Interestingly, RGD peptides also block MS ATPrelease from Xenopus oocytes (262, see below). However,although ATP been shown to be coreleased with acetyl-choline at the motor synapse (369) and a recent reportindicates evoked ATP release from sensory neurons(381), evidence described below indicates that severalmechanisms may contribute to MS ATP release.

D. Mechanosensitive ATP Release

Burnstock and colleagues (43–46) have advocatedATP’s role as a neurotransmitter for over 30 years. More

recently, ATP has been implicated in mechanosensationand cell volume regulation, by the demonstration thatmechanical stimuli (including cell inflation, direct inden-tation, and osmotic swelling) can cause nonlytic ATP (andUTP) release from excitable and nonexcitable cells (139,236, 237, 280, 292, 420, 430). The MS release of ATP (orUTP) may in turn induce an electrical or biochemicalsignal by activating purinergic receptors on the same cell(autocrine) and/or neighboring cells (paracrine). Twobroad purinergic receptor families have been identified(303). One is the ionotropic receptor (P2X) family, inwhich the ion channel is an integral part of the receptorprotein (37, 64, 245). The other is the metabotropic recep-tor (P2Y) family, in which the receptor is indirectly cou-pled to membrane ion channels either via soluble secondmessenger pathways or via membrane-delimited path-ways (255, 289). Clearly, the class of receptor will influ-ence the kinetics of channel activation and inactivationand thereby set limits on the frequency response of thetransducer. This may vary from submilliseconds for thefastest ionotropic receptor channel to several minutes formetabotropic receptors (167, 176, 256, 264).

A key question for ATP release mechanisms is thenature of the driving force for ATP efflux. ATP in theextracellular milieu is normally kept extremely low (i.e.,,1 mM) by extracellular ectonucleotidases (133). For ex-ample, Forrester (115) has measured basal levels of ATPof ,2 3 1028 M in human venous plasma that may beincreased ;50-fold (i.e., to ;1 mM) by partial arterialocclusion and exercise. Although ATP may be still higherin the interstitial fluid and even higher at the membranesurface near vesicular release sites, there should alwaysbe a steep gradient for ATP efflux, given that intracellular[ATP] is in the 1–5 mM range (360). The fact that ATP isalso stored in cytoplasmic vesicles (424) indicates theexistence of distinct pools of ATP that may be released bydifferent mechanisms. For example, any stimulus that isstrong enough to cause tissue damage will result in mas-sive ATP release. It is presumably this mechanism thatreleases the ATP that mediates pain sensation. However,it is unlikely that ATP release caused by gentle mechani-cal stimulation arises from cell damage. For example, MSATP release can occur without associated membrane con-ductance changes (i.e., in high impedance cells) and canbe blocked by specific agents (262, 292, 448). Electroneu-tral ATP release can be monitored by the luciferin-lucif-erase luminescence assay or by the use of endogenous orheterologously expressed purinergic receptors as biosen-sors (139, 262, 292, 405, 430). A recent study using theluminescence assay has shown that collagenase treat-ment, RGD peptides, and cytochalasin can block MS ATPrelease from Xenopus oocytes. Furthermore, the ATPrelease saturates with repetitive mild mechanical stimuli(i.e., puffs of solution) at levels four orders of magnitudebelow those reached immediately after oocyte damage

726 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 44: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

(262; R. Maroto and O. P. Hamill, unpublished observa-tions).

Electroneutral release of ATP may be mediated by anATP electroneutral transporter and/or by vesicular re-lease. Specific ATP transporters have been identified inthe ER and the Golgi apparatus that translocate cytoplas-mic ATP (176a, 328a). It is unknown whether these arefunctionally expressed in the plasma membrane. If ATP isreleased by exocytosis, one may be able to correlate itwith changes in Cm if there is a net increase in membranearea (see Ref. 129a). However, when exocytosis and en-docytosis are exactly balanced, vesicular release can pro-ceed with no detectable change in Cm. This is likely thecase for the Xenopus oocyte that undergoes high rates ofvesicle fusion (exocytosis) and fission (endocytosis) (i.e.,2,000–16,000 vesicles/s corresponding to 60–500 um2

membrane/s) capable of replacing the cell membrane ev-ery 24 h (444). This membrane turnover, which proceedsunder basal (unstimulated) conditions, is associated withthe delivery via vesicle trafficking from the Golgi appara-tus and removal of membrane proteins. The vesicle traf-ficking between the Golgi apparatus and the plasma mem-brane can be blocked by brefeldin A (BFA) (118a).Interestingly, BFA also blocks basal and MS ATP releasefrom Xenopus oocytes (262). Although membrane traf-ficking/ATP release does not require Ca21 influx, recentevidence indicates it may be increased by mechanicalstimulation (262). The interesting implication here is thatthe protein and lipid composition of the surface mem-brane may be rapidly altered under specific mechanicalenvironments. Because most eukaryotic cells share thisvesicle trafficking pathway, it may account for the ubiq-uitous nature of ATP release from animal cells.

It has also been proposed that ATP permeatesthrough specific membrane ion channels, most notablythe cystic fibrosis transmembrane conductance regulator(CFTR) (52, 335) and the multidrug resistance (MDR) Cl2

channels (4a). This idea remains somewhat controversial(360). However, the original argument that ATP was toolarge to permeate these Cl2 channels appears invalid(246). On the other hand, there is no evidence to indicatethat either CFTR or MDR are MG channels. Another chan-nel that might mediate ATP release is the hemi-gap-junc-tional channel (72). However, mechanical stimulation ap-pears to close rather than open this class of channel (448,450). Furthermore, it has been shown that submillimolarconcentrations of Gd31 block the hemi-gap channel, yeteven higher concentrations of Gd31 (1–10 mM) fail toblock either basal or MS ATP release (262, Maroto andHamill, unpublished data). Most recently, it has beenreported that ATP is carried by a membrane conductanceactivated by strong hyperpolarizations of around 2200mV (29a). However, others have concluded that this con-ductance reflects reversible dielectric breakdown of themembrane, based on its lack of ion selectivity, failure to

saturate, and slow (up to 10 min) recovery upon repolar-ization (449).

ATP release in the oocyte may be important in thecross talk with the surrounding follicular cells that ex-press purinergic receptor-gated Cl2 channels (324a). Toaccount for the ubiquitous expression of ATP releasefrom mammalian cells, it has been proposed that externalATP plays some role in establishing a set point for signaltransduction pathways, in particular, those involved inchanging [Ca21]i, cAMP, and activating protein kinases(315). This set point may be important in influencingspecific steps in cell development and differentiation. Forexample, ATP released from developing sensory neuronsdelays the terminal differentiation of surroundingSchwann cells and nerve myelination until they are ex-posed to axon-derived signals (381).

Nakamura and Strittmatter (292) have hypothesizedthat vertebrate touch sensation is mediated by an ATP-dependent mechanism of mechanotransduction. Theypropose that mechanical stimulation of the afferent nerveterminal causes ATP release that then results in autocrineactivation of P2Y1 receptors. In support of this hypothesis,they demonstrated that touch-induced action potentials infrog sensory nerve were increased in frequency by sub-cutaneous injection of ATP, and this increase could beblocked by injection within the receptive field of a P2

purinoceptor antagonist (suramin) or a ATP-degradingenzyme (apyrase). Subsequently, Svichar et al. (394, 395)demonstrated that ATP induces Ca21 release from IP3-sensitive Ca21 stores in large DRG but not in small DRGneurons. Interestingly, they also found that ATP activateda large transient inward current that was over before the[Ca21]i began to rise, ruling out the transient currentbeing activated by IP3-sensitive Ca21 release. AnotherDRG study indicated a cation conductance activated byrelease of intracellular Ca21 stores (333). However, thisconductance was also activated by heat and most likelymediates pain sensation. It remains unknown whether arelated Ca21-sensitive conductance coupled to P2Y1 recep-tor activation mediates the generator (receptor) currentin specific touch receptors.

Although the ATP hypothesis is attractive, severalkey elements of the hypothesis lack experimental sup-port. To begin with, MS ATP release and its underlyingmechanism need to be determined in the same large fiberDRG neurons that show the P2Y responses. Second, therelationship of this release process with Ca21-dependentexocytosis in DRGs (e.g., see Ref. 194) needs to be deter-mined. For example, does the mechanosensitivity derivefrom Ca21 influx through MG cation channels? Third, thespecific channel or conductance mechanism coupled tothe P2Y1 receptor that presumably mediates the generatorpotential needs to be identified. Fourth, the underlyingkinetics and sensitivity of both ATP release, channel ac-tivation, and ATP removal must be demonstrated to be

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 727

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 45: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

compatible with the fast kinetics, high mechanosensitiv-ity, and in some cases rapid adaptation of mechanotrans-duction in specific touch receptors. Fifth, it must be dem-onstrated that all the phenomena operate in the intactDRG nerve endings/specializations (see Ref. 379). Finally,the alternative hypothesis that tactile sensation is medi-ated directly by MG channels (e.g., MECs/DEGs and/orNOMPC homologs) needs to be rigorously examined atthe functional level in different classes of tactile recep-tors. On the last issue, although it has recently beendemonstrated that mechanical stimulation of the soma(273a, 400a) and growth cones (199a) of DRG neurons caninduce conductance changes, they have yet to be relatedto specific single MG channel activities.

E. Membrane Resealing: Ca21-Induced

Vesicle-Vesicle Fusion and Exocytosis

Apart from releasing cytoplasmic contents (e.g.,ATP), a more fundamental response to membrane dam-age in terms of cell survival is the activation of mecha-nisms that repair or reseal the damaged membrane. Mem-brane repair/resealing may represent the most primitiveform of cellular mechanotransduction. Studies in severaldifferent cell types have led to different models of mem-brane repair/resealing. One model, the “vesicle plug”model, has been used to explain the resealing oftransected squid and crab axons (23, 100, 276), where it isproposed that Ca21-dependent formation and accumula-tion of endocytotic vesicles fuse to form a vesicular plugthat reseals the transected axon. The resealing processrequires high [Ca21]i (i.e., .100 mM) and calpain activa-tion and occurs over a relatively slow time scale (i.e.,minutes to hours) (23). A second model referred to as the“exocytotic” model has been used to explain resealing ofmicropunctures in sea urchin eggs and 3T3 fibroblasts(21). The resealing requires high [Ca21]i, is completedwithin seconds, and has been correlated with bursts ofexocytosis near the wound site (21). It is proposed thatthe vesicle fusion events supply the extra membraneneeded to reseal the membrane wound. Unlike the firstmodel, the vesicles are most likely present in the cyto-plasm, possibly even docked close to the plasma mem-brane. Furthermore, because the resealing can be inhib-ited by neurotoxins that selectively interact with theSNARE complex proteins (e.g., synaptobrevin and syn-taxin), the delivery, docking, and fusion of these vesicleswith the membrane surrounding the wound site may in-volve a process similar to transmitter release (21, 382c).More recently, it has been reported that faciliation ofresealing of 3T3 cells in response to a second disruptionis blocked by BFA and cytochalasin, indicating that vesi-cle trafficking from the Goli apparatus is involved (413a).A third model, representing a composite of the plug and

exocytosis models, has been used to explain the resealingof large disruptions of marine eggs and oocytes (407).Here, Ca21 elevation induces preexisting vesicles to fusewith one another to form a large single “wound vesicle”that then somehow fuses with the discontinuous bilayer(407). In both the exocytosis and composite models, themembrane repair may result in the release of signalingmolecules (i.e., stored in the preformed vesicles) thatmediate further responses, including cellular hyperplasiaand/or hypertrophy that are commonly associated withmechanical stress (276). In the specific case of maturedeggs, it is interesting that membrane damage (e.g., needlepricking) can promote the fertilization response, whichinvolves Ca21-activated fusion of cortical granules withthe plasma membrane (416a). Clearly, the secondary re-sponses related to membrane repair will depend on thenature of the vesicles and their intracellular contents. Forexample, preliminary studies indicate that the yolk plate-lets, rather than cortical granules or endosomal vesicles,may mediate resealing of sea urchin eggs (407). Thefeature common to all models is that elevation of intra-cellular Ca21 triggers the resealing mechanism. It may bethat sublytic membrane tension, by promoting Ca21 influxthrough SA-CAT channels, elevates local [Ca21]i andthereby primes the Ca21-sensitive repair mechanisms. Inthis case, the MG channel would act as a safety devicesomewhat analogous to MscL in E. coli. Patch-clamprecordings indicate the tensions that activate MG chan-nels and cause patch rupture are within the same order ofmagnitude.

XI. CONCLUSIONS AND OUTSTANDING ISSUES

The response of both simple bilayer vesicles and cellsto mechanical stimulation is determined by both extrinsic(e.g., size and surface area to volume) and intrinsic (i.e.,material properties) properties. The deformation-sensi-tive membrane parameters that may likely influence mem-brane protein conformational changes include membranedilation (i.e., increased area occupied by lipid molecules),the accompanying membrane thinning (i.e., assumingmembrane incompressibility), and local changes in mem-brane curvature or bending. When a membrane protein isinserted in a bilayer it produces a mechanical deformationthat depends on the coupling between the hydrophobicregions of the protein and the bilayer. A protein confor-mational change that involves a change in hydrophobicmismatch will be sensitive to both membrane thicknessand local membrane curvature, each of which may bealtered by changing either the lipid bilayer composition orby mechanically stretching or bending the membrane. Thetwo simple channel-forming peptides, alamethicin andgramicidin, display MS channel gating that have beenexplained by two different classes of mechanisms. The

728 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 46: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

important distinction between the two is that whereasalamethicin channel formation may involve relativelylarge changes in membrane occupied area (i.e., associatedwith subunit recruitment into a barrel-staved complex),gramicidin channel formation involves insignificant areachanges (i.e., a dimerization between monomers in eachmonolayer). As a consequence, the energy of gating ofalamethicin but not gramicidin may reflect the product ofthe membrane tension (membrane dilation) and the oc-cupied area change associated with channel opening. Onthe other hand, increased tension may act on gramicidinindirectly by causing membrane thinning, thereby affect-ing the free energy components that depend on bilayer-gramicidin hydrophobic coupling.

The purification, cloning, and recent determinationof the crystal structure of the MscL protein from E. coli

and M. tuberculosis have provided a rich environment formodel building and testing. The present evidence favors apentameric channel complex with the channel gate mostlikely formed by the hydrophobic constriction formed bythe TM1 helices at the cytoplasmic end of the channel inthe closed configuration. The large energy required toopen the channel (;18 kT) may be needed to compensatefor the exposure of the hydrophobic region to water thatenters the open channel. The NH2 terminus, althoughmobile, does not function as the channel gate but insteadserves to stabilize the open conformation(s) of the chan-nel and may interfere with the passage of ions, thusleading to the appearance of the channel subconductinglevels. The COOH termini play a role in stabilizing theclosed configuration of the channel, whereas the periplas-mic loop may function as an elastic spring resisting the

opening of the channel by membrane tension. Clearly,future studies that provide structural information on theopen conformation of MscL (e.g., by recently employedcysteine scanning mutagenesis combined with electro-paramagnetic resonance spectroscopy to probe the struc-ture of MscL, Ref. 269a) will further refine this model.

Growing evidence in the form of amphipath activa-tion and MG channel activity in CSK-deficient PMVs indi-cates that MG channels in specific animals cells (e.g.,Xenopus oocytes) are gated by increase in bilayer tensionsimilar to prokaryotic MG channels. However, structuralstudies of several eukaryotic MG and putative MG chan-nels (Table 1) indicate that a variety of molecular mech-anisms may confer mechanosensitivity on membrane pro-teins. In some cases, a tethered mechanism of gating isfavored (e.g., hair cell MG channels, MECs/DEGs, andNOMPC), but unequivocal evidence is lacking. Such evi-dence may come with the demonstration that 1) deletionof CSK binding sequences abolish mechanosensitivity, 2)rearrangement of CSK/membrane proteins (e.g., mea-sured with fluorescence energy transfer) correlates withchannel activation, and 3) mechanosensitivity can be re-stored in channel proteins reconstituted into liposomesby pulling on CSK/ECM labeled microspheres (i.e., thatact as molecular handles).

In addition to the intrinsic properties of MG channelproteins, there are also various extrinsic mechanisms thatmodify the animal cell’s response to mechanical stimula-tion. For example, the excess membrane area of animalcells in the form of microvilli, membrane folds, and in-vaginations (i.e., caveolae) acts as a membrane reservoirto buffer sudden and/or large changes in bilayer tension

TABLE 1. A list of MG and putative MG channels with details regarding their biophysics,

structure, and favored gating mechanisms

Channel Protein (Organism)Conductance/Ion

SelectivityAminoAcids

TransmembraneDomains Oligomerization

MS GatingMechanism Reference No.

Gramicidin (B. brevis) 15 pS/cation 15 1 2 Bilayer 6, 426Alamethicin (T. viride) 20–200 pS/

nonselective20 1 $3 Bilayer 47, 355

MsCL (E. coli) ;3 nS/nonselective 136 2 5 Bilayer 60, 387MscS (YggB) (E. coli) ;1 nS/nonselective 286 Bilayer 31, 244MscMJ (MJ0170)

(M. jannashii)270 pS/cation 350 3 Bilayer 220a, 269

MID1 (Yeast) 35 pS/cation 548 4 199, 208TRAAK (Human) 30 pS (0 mV)/K1 300 4/2P Bilayer 257TREK-1/-2 (Mammalian

neurons)48 pS/68 pS/K1 538 4/2P Bilayer† 12a, 323

SIC (Rat kidney) 250 pS/cation 563 6 393ENaC* (Rat) 10 pS/Na1 638 2 5 or 9 Bilayer

(MS Ca21 block)50, 105a, 113,

203, 371aMEC-4 (C. elegans) 768 2 Tethered† 93, 403NOMPC (Drosophila) 1,619 6 Tethered† 215, 425MGXO (Xenopus occyte) ;80 pS/cation Bilayer 277, 447MGHC (Vertebrate hair

cell);100 pS/cation Tethered† 74, 260

* Mechanosensitivity has been reported for abg-epithelial Na1 channel (ENaC) reconstituted in bilayers but not in cells. † Evidence for themechanism is less compelling than liposome reconstitution. MG, mechanically gated.

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 729

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 47: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

that might otherwise rupture the cell (330, 448). As aconsequence, bilayer-gated MG channels detected intight-seal patch-clamp recording that tends to smooth outthe membrane patch may not be exposed to activatingtensions in the cell membrane except under special cir-cumstances (447). For example, specialized mechanosen-sors may possess highly localized regions of bilayer thatare prestressed (i.e., by CSK/ECM elements) to enablerapid response to mechanical deformation. Furthermore,cells that undergo changes in membrane geometry duringgrowth and differentiation (96) or experience membranereorganization or blebbing during specific cellular andphysiological processes, including mitosis (328), apopto-sis (216), cell locomotion (103), cell spreading (240), andorgan distension (245b), may use up their membranereserves and thus expose their bilayer to increased ten-sion. However, some channels that display mechanosen-sitivity in patch-clamp recordings [e.g., NMDA and S-typeK1 (TREK/TRAAK) channels] may be of biophysicalrather than physiological significance (287). Clearly, theidentification of toxins that selectively block MG channels(e.g., the peptide toxin from Grammastola spatulata spi-der venom, Ref. 382b) and the development of genetic MGchannel knock-outs should prove useful in demonstratingthe function of specific MG channels (380, 434). It will beparticularly interesting to use these tools to determine thepossible relationship between MG channels and the MSprocesses controlling membrane turnover, vesicle traf-ficking, and membrane repair.

We thank Peter Gage for suggesting we write this review.We also thank him for his insight, input, and support. We thankOlaf Andersen, Evan Evans, Anna Kloda, Jean-JacquesLacaperre, Rosario Maroto, David Needham, Aaron Oakley, andJean-Louis Rigaud for advice and input. Finally, we thank PaulBlount, Paul Moe, Cathy Morris, and Ching Kung for sharingunpublished information.

Our research is supported by the National Institute ofArthritis and Musculoskeletal and Skin Diseases Grant RO1-AR-42782, the National Science Foundation, the Muscular Dystro-phy Association, the National Health and Medical ResearchCouncil of Australia Grant 960591, the Australian ResearchCouncil Grants A09701150 and A23387847, and the Raine Med-ical Foundation.

Address for correspondence: O. P. Hamill, Physiology andBiophysics, UTMB, Galveston, TX 77555 (E-mail: [email protected]).

REFERENCES

1. ADAMS CM, ANDERSON MG, MOTTO DG, PRICE MP, JOHNSON WA, AND

WELSH MJ. Ripped pocket and pickpocket, novel Drosophila DEG/ENaC subunits in early development and mechanosensory neu-rons. J Cell Biol 140: 143–152, 1998.

2. AJOUZ B, BERRIER C, GARRIGUES A, BESNARD M, AND GHAZI A. Releaseof thioredoxin via mechanosensitive channel MscL during osmoticdownshock of Escherichia coli cells. J Biol Chem 273: 26670–26674, 1998.

3. AJOUZ B, BERRIER C, BESNARD M, MARTINAC B, AND GHAZI A. Contri-butions of the different extramembraneous domains of the mech-anosensitive ion channel MscL to its response to membrane ten-sion. J Biol Chem 275: 1015–1022, 2000.

4. AKINLAJA J AND SACHS F. The breakdown of cell membranes byelectrical and mechanical stress. Biophys J 75: 247–254, 1998.

4a.AL-AWQATI Q. Regulation of ion channels by ABC transporters thatsecrete ATP. Science 269: 805–806, 1995.

5. ALVAREZ O AND LATORRE R. Voltage-dependent capacitance in lipidbilayers made from monolayers. Biophys J 21: 1–17, 1978.

6. ANDERSEN OS AND KOEPPE RE. Molecular determinants of channelfunction. Physiol Rev 72 Suppl: S89–S158, 1992.

7. ARKIN IT, SUKHAREV SI, BLOUNT P, KUNG C, AND BRUNGER AT. Helicity,membrane incorporation, orientation and thermal stability of thelarge conductance mechanosensitive ion channel from E. coli.

Biochim Biophys Acta 1369: 131–140, 1998.8. ASSAD JA AND COREY DP. An active motor model for adaptation by

vertebrate hair cells. J Neurosci 12: 3291–3309, 1992.9. ASSAD JA, HACOHEN N, AND COREY DP. Voltage dependence of adap-

tation and active bundle movement in bullfrog saccular hair cells.Proc Natl Acad Sci USA 86: 2918–2922, 1989.

10. ASSAD JA, SHEPHARD GMG, AND COREY DP. Tip-link integrity andmechanical transduction in vertebrate hair cells. Neuron 7: 985–994, 1991.

11. AWAYDA MS, ISMAILOV II, BERDIEV BK, AND BENOS DJ. A cloned renalepithelial Na1 channel protein displays stretch activation in planarlipid bilayers. Am J Physiol Cell Physiol 268: C1450–C1459, 1995.

12. AWAYDA MS AND SUBRAMANYAM M. Regulation of the epithelial Na1

channel by membrane tension. J Gen Physiol 112: 97–111, 1998.12a.BANG H, KIM Y, AND KIM D. TREK-2, a new member of the mech-

anosensitive tandem-pore K1 channel family. J Biol Chem 275:17412–17419, 2000.

13. BARINAGA M. Molecular evolution. Archaea and eukaryotes growcloser. Science 264: 1251, 1994.

14. BATIZA AF, RAYMENT I, AND KUNG C. Channel gate! Tension, leak anddisclosure. Struct Fold Design 7: R99–R103, 1999.

15. BEREITER-HAHN J. Mechanical principles of architecture of eucary-otic cells. In: Cytomechanics: The Mechancial Basis of Cell Form

and Structure, edited by Bereiter-Hahn J, Anderson O, and ReifW-E. Berlin: Springer-Verlag, 1987, p. 3–30.

16. BERNSDORFF C, WOLF A, WINTER R, AND GRATTON E. Effects of hy-drostatic pressure on water penetration and rotational dynamics inphospholipid-cholesterol bilayers. Biophys J 72: 1264–1277, 1997.

17. BERRIER C, BESNARD M, AJOUZ B, COULOMBE A, AND GHAZI A. Multiplemechanosensitive ion channels from Escherichia coli, activated atdifferent thresholds of applied pressure. J Membr Biol 151: 175–187, 1996.

18. BERRIER C, COULOMBE A, HOUSSIN C, AND GHAZI A. A patch-clampstudy of inner and outer membranes and of contact zones of E. coli,

fused into giant liposomes Pressure-activated channels are local-ized in the inner membrane. FEBS Lett 259: 27–32, 1989.

19. BERRIER C, COULOMBE A, SZABO I, ZORATTI M, AND GHAZI A. Gadolin-ium ion inhibits loss of metabolites induced by osmotic shock andlarge stretch-activated channels in bacteria. Eur J Biochem 206:559–565, 1992.

20. BESNARD M, MARTINAC B, AND GHAZI A. Voltage-dependent porin-likeion channels in the archaeon Haloferax volcanii. J Biol Chem 272:992–995, 1997.

21. BI GQ, ALDERTON JM, AND STEINHARDT RA. Calcium-regulated exo-cytosis is required for cell membrane resealing. J Cell Biol 131:1747–1758, 1995.

22. BIBBY KJ AND MCCULLOCH CAG. Regulation of cell volume and[Ca21] in attached human fibroblasts responding to anisosmoticbuffers. Am J Physiol Cell Physiol 266: C1639–C1649, 1994.

23. BITTNER GD AND FISHMAN HM. Axonal sealing following injury. In:Nerve Regeneration, edited by Ingola N and Murray M. New York:Dekker, 2000, chapt. 12, p. 1–30.

24. BLOOM M, EVANS E, AND MOURISTEN OG. Physical properties of thefluid lipid-bilayer component of cell membrane: a perspective. Q

Rev Biophys 24: 293–397, 1991.25. BLOUNT P, SCHROEDER M, AND KUNG C. Mutations in a bacterial

mechanosensitive channel change the cellular response to osmoticstress. J Biol Chem 272: 32150–32157, 1997.

730 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 48: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

26. BLOUNT P, SUKHAREV SI, MOE P, MARTINAC B, AND KUNG C. Mechano-sensitive channels in bacteria. Methods Enzymol 294: 458–482,1999.

27. BLOUNT P, SUKHAREV SI, MOE P, SCHROEDER GUY MJ, AND KUNG C.Membrane topology and multimeric structure of a mechanosensi-tive channel protein of Escherichia coli. EMBO J 15: 101–108. 1996.

28. BLOUNT P, SUKHAREV SI, SCHROEDER MJ, NAGLE SK, AND KUNG C.Single residue substitutions that change gating properties of amechanosensitive channel in Escherichia coli. Proc Natl Acad Sci

USA 93: 11652–11657, 1996.29. BLUEMINK JG, HAGE WJ, VAN DEN HOEF MH, AND DICTUS WJ. Freeze-

fracture electron microscopy of membrane changes in progester-one-induced maturing oocytes and eggs of Xenopus laevis. Eur

J Cell Biol 31: 85–93, 1983.29a.BODAS E, ALEU J, PUJOL G, MARTIN-SATUE M, MARSAL J, AND SOLSONA

C. ATP crossing the cell plasma membrane generates an ioniccurrent in Xenopus oocytes. J Biol Chem 275: 20268–20273, 2000.

30. BOTIANO S, SANDERSON MJ, AND DIRKSEN ER. A role for Ca21 con-ducting ion channels in mechanically induced signal transductionin airway epithelial cells. J Cell Sci 107: 3037–3044, 1994.

31. BOOTH IR AND LOUIS P. Managing hypoosmotic stress: aquaporinsand mechanosensitive channels in Escherichia coli. Curr Opin

Micro 2:166–169, 1999.32. BOURQUE CW AND OLIET SHR. Osmoreceptors in the central nervous

system. Annu Rev Physiol 59: 601–619, 1997.33. BOWMAN CB, DING JP, SACHS F, AND SOKABE M. Mechanotransducing

ion channels in astrocytes. Brain Res 584: 272–286, 1992.34. BOWMAN CB AND LOHR JW. Curvature sensitive mechanosensitive

ion channels and osmotically evoked movements of the patchmembrane (Abstract). Biophys J 70: A365, 1996.

35. BOWMAN CB AND LOHR JW. Mechanotransducing ion channels in C6glioma cells. Glia 18: 161–176, 1996.

36. BRADING AF. The physiology of the mammalian urinary outflowtract. Exp Physiol 84: 215–221, 1999.

37. BRAKE AJ, WAGENBACH MJ, AND JULIUS D. A new structural motif forligand-gated ion channels defined by an ionotropic ATP receptor.Nature 371: 519–523, 1994.

37a.BREHM P, KULLBERG R, AND MOODY-CORBETT F. Properties of non-junctional acetylcholine receptor channels on innervated muscle ofXenopus laevis. J Physiol (Lond) 350: 631–648, 1984.

38. BRITTEN RJ AND MCCLURE FT. The amino acid pool of Escherichia

coli. Bacteriol Rev 26: 292–300, 1962.39. BROPHY CM, MILLS I, ROSALLES O, ISALES C, AND SUMPIO BE. Phos-

pholipase C: a putative mechanotransducer for endothelial cellresponse to acute hemodynamic changes. Biochem Biophys Res

Commun 109: 576–581, 1993.40. BRUNS D AND JAHN R. Real-time measurement of transmitter release

from single synaptic vesicles. Nature 377: 62–65, 1995.41. BUECHNER M, DELCOUR AH, MARTINAC B, ADLER J, AND KUNG C. Ion

channel activities in the Escherichia coli outer membrane. Bio-

chim Biophys Acta 1024: 111–119, 1990.42. BULLOUGH PA, HUGHSON F, SKEHEL JJ, AND WILEY DC. Structure of

infuenza haemagglutinin at the pH of membrane fusion. Nature

371: 37–43, 1994.43. BURNSTOCK G. Purinergic nerves. Pharm Rev 24: 509–581, 1972.44. BURNSTOCK G. Purinergic mechanisms. Ann NY Acad Sci 603: 1–19,

1990.45. BURNSTOCK G. Release of vasoactive substances from endothelial

cells by shear stress and purinergic mechanosensory transduction.J Anat 194: 335–342, 1999.

46. BURNSTOCK G AND WOOD JN. Purinergic receptors: their role innociception and primary afferent neurotransmission. Curr Opin

Neurobiol 6: 526–532, 1996.47. CAFISO DS. Alamethicin: a peptide model for voltage gating and

protein-membrane interactions. Annu Rev Biophys Biomol Struct

23: 141–165, 1994.48. CANESSA CM, HORISBERGER JD, AND ROSSIER BC. Epithelial sodium

channel related to proteins involved in neurodegeneration. Nature

361: 467–470, 1993.49. CANESSA CM, MERILLAT AM, AND ROSSIER BC. Membrane topology of

the epithelial sodium channel. Am J Physiol Cell Physiol 267:C1682–C1690, 1994.

50. CANESSA CM, SCHILD L, BUELL G, THORENS B, GAUTSCHI I, HORISBERGER

JD, AND ROSSIER BC. Amiloride sensitive epithelial Na1 channel ismade of three homologous subunits. Nature 367: 463–467, 1994.

51. CANTIELLO HF. Role of the actin cytoskeleton on epithelial Na1

channel regulation. Kindey Int 48: 970–984, 1995.52. CANTIELLO HF, JACKSON GR JR, GROSMANN CF, PRAT AG, BORKAN CS,

WANG Y, REISIN IL, O’RIORDAN CR, AND AUSIELLO DA. Electrodiffu-sional ATP movement through the cystic fibrosis transmembraneconductance regulator. Am J Physiol Cell Physiol 274: C799–C809,1998.

53. CASADO M AND ASCHER P. Opposite modulation of NMDA receptorsby lysophospholipids and arachidonic acid: common features withmechanosensitivity. J Physiol (Lond) 513: 317–330, 1998.

53a.CATERINA MJ, SCHUMACHER MA, TOMINAGA M, ROSEN TA, LEVINE JD,AND JULIUS D. The capsaicin receptor: a heat-activated ion channelin the pain pathway. Nature 389: 816–824, 1997.

54. CAVALIER-SMITH T. The origin of eukaryotic and archaebacterialcells. Ann NY Acad Sci 503: 17–54, 1987.

55. CEMERIKIC D AND SACKIN H. Substrate activation of mechanosensi-tive, whole cell currents in renal proximal tubule. Am J Physiol

Renal Fluid Electrolyte Physiol 264: F697–F714, 1993.56. CHA A, SNYDER GE, SELVIN PR, AND BEZANILLA F. Atomic scale

movement of the voltage-sensing region in a potassium channelmeasured via spectroscopy. Nature 402: 809–813, 1999.

57. CHALFIE M. Touch receptor development and function in Caeno-

rhabditis elegans. J Neurobiol 24: 1433–1441, 1993.58. CHALFIE M AND AU M. Genetic control of differentiation of the

Caenorabditis elegans touch receptor neurons. Science 243: 1027–1033, 1989.

59. CHALFIE M AND THOMSON JN. Structural and functional diversity inthe neuronal microtubules of Caenorhabditis elegans. J Cell Biol

93: 15–23, 1982.60. CHANG G, SPENCER R, LEE A, BARCLAY M, AND REES C. Structure of the

MscL homologue from Mycobacterium tuberculosis: a gated mech-anosensitive ion channel. Science 282: 2220–2226, 1998.

61. CHAPLEAU MW Cardiovascular mechanoreceptors. Adv Comp En-

viron Physiol 10: 138–164, 1992.62. CHARLES AC, MERRILL JE, DIRKSEN ER, AND SANDERSON MJ. Intercel-

lular signalling in glial cells: calcium waves and oscillations inresponse to mechanical stimulation and glutamate. Neuron 6: 983–992, 1991.

63. CHEN BM AND GRINNELL AD. Kinetics, Ca21 dependence and bio-physical properties of integrin mediated mechanical modulation oftransmitter release from frog motor nerve terminals. J Neurosci 17:904–916, 1996.

64. CHEN CC, AKOPLAN AN, SIVILOTTI L, COLQHOUN D, BURNSTOCK G, AND

WOOD JN. A P2x purinoceptor expressed by a subset of sensoryneurons. Nature 377: 428–431, 1995.

65. CHEN Y, SIMASKO SM, NIGGEL J, SIGURDSON WJ, AND SACHS F. Ca21

uptake in GH3 cells during hypotonic swelling: the sensory role ofstretch-activated ion channels. Am J Physiol Cell Physiol 270:C1790–C1798, 1996.

66. CHOTHIA C. Hydrophobic bonding and accessible surface area inproteins. Nature 248: 338–339, 1974.

67. COLBERT HA, SMITH TL, AND BARGMANN CI. Osm-9, a novel proteinwith structural similarity to channels, is required for olfaction,mechanosensation, and olfactory adapation in Caenorhabditis el-

egans. J Neurosci 17: 8259–8269, 1997.68. COOK SP, VULCHANOVA L, HARGRAEVES KM, ELDE R, AND MCCLESKEY

EW. Distinct ATP receptors on pain-sensing and stretch-sensingneurons. Nature 387: 505–508, 1997.

69. COREY DP AND HOWARD J. Models for ion channel gating withcompliant states. Biophys J 66: 1254–1257, 1994.

70. COREY DP AND HUDSPETH AJ. Response latency of vertebrate haircells. Biophys J 26: 499–506, 1979.

71. COREY DP AND HUDSPETH AJ. Kinetics of the receptor current inbullfrog saccular hair cells. J Neurosci 3: 962–976, 1983.

72. COTRINA ML, LIN JHC, ALVES-RODRIGUES A, LIU S, LIU S, LI J, AZMI-GHADIMI H, KANG J, NAUS CCG, AND NEDERGAARD M. Connexinsregulate calcium signaling by controlled ATP release. Proc Natl

Acad Sci USA 95: 15735–15740, 1998.73. CRAWFORD AC, EVANS MG, AND FETTIPLACE R. Activation and adap-

tation of turtle hair cells. J Physiol (Lond) 419: 405–434, 1989.74. CRAWFORD AC, EVANS MG, AND FETTIPLACE R. The actions of calcium

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 731

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 49: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

on the mechano-electrical transducer current of turtle hair cells.J Physiol (Lond) 434: 369–398, 1991.

74a.CRUICKSHANK CC. Estimation of the Pore Size of the Large-Conduc-

tance Mechanosensitive Ion Channel (MscL) of E. coli (Honorsthesis). Nedlands: Univ. of Western Australia, 1996.

75. CRUICKSHANK CC, MINCHIN R, LE DAIN A, AND MARTINAC B. Estimationof the pore size of the large-conductance mechanosensitive ionchannel of Escherichia coli. Biophys J 73: 1925–1931, 1997.

76. CUI C AND ADLER J. Effect of mutation of potassium-efflux system,KefA, on mechanosensitive channels in the cytoplasmic membraneof Escherichia coli. J Membr Biol 150: 143–152, 1996.

77. CUI C, SMITH DO, AND ADLER J. Characterization of mechanosensi-tive channels in Escherichia coli cytoplasmic membrane by whole-cell patch-clamp recording. J Membr Biol 144: 31–42, 1995.

78. DAI J AND SHEETZ MP. Regulation of endocytosis, exocytosis, andshape by membrane tension. Cold Spring Harbor Symp Quant Biol

60: 567–571, 1995.79. DAI J AND SHEETZ MP. Membrane tether formation from blebbing

cells. Biophys J 77: 3363–3370, 1999.80. DAI J, SHEETZ MP, WAN X, AND MORRIS CE. Membrane tension in

swelling and shrinking molluscan neurons. J Neurosci 18: 6681–6692, 1998.

81. DAI J, TING-BEALL HP, SHEETZ MP. The secretion-coupled endocyto-sis correlates with membrane tesnion changes in RBL 2H3 cells.J Gen Physiol 110: 1–10, 1997.

82. DAN N AND SAFRAN SA. Effect of lipid characteristics on the struc-ture of transmembrane proteins. Biophys J 75: 1410–1414, 1998.

82a.DARBOUX I, LINGUEGLIA E, CHAMPIGNY G, COSCOY S, BARBRY P, AND

LAZDUNSKI M. dGNaCl, a gonad-specific amiloride sensitive Na1

channel. J Biol Chem 273: 9424–9429, 1998.83. DAVIES PF. Flow-mediated endothelial mechanotransduction.

Physiol Rev 75: 519–560, 1995.84. DAVIS MJ, DONOVITZ JA, AND HOOD JD. Stretch-activated single-

channel and whole cell currents in vascular smooth muscle cells.Am J Physiol Cell Physiol 262: C1083–C1088, 1992.

85. DEAMER DW AND BRAMHALL J. Permeability of lipid bilayers to waterand ionic solutes. Chem Phys Lipids 40: 167–188, 1986.

86. DEAN JW AND LAB MJ. Arrhythmia in heart failure: role of mechan-ically induced changes in electrophysiology. Lancet 1: 1309–1312,1989.

87. DELCOUR AH, MARTINAC B, ADLER J, AND KUNG C. Modified reconsti-tution method used in patch-clamp studies of Escherichia coli ionchannels. Biophys J 56: 631–635, 1989.

88. DEMER LL, WORTHAM CM, DIRKSEN ER, AND SANDERSON MJ. Mechan-ical stimulation induces intracellular calcium signaling in bovineaortic endothelium cells. Am J Physiol Heart Circ Physiol 264:H2094–H2102, 1993.

89. DETWEILER PB. Sensory transduction. In: Textbook of Physiology:

Excitable Cells and Neurophysiology, edited by Patton HD, FuchsAF, Hille B, Sher AM, and Streiner R. Philadelphia, PA: Saunders,1989, p. 98–129.

90. DIMMELER S, HERMANN C, AND ZEIHER AM. Apoptosis of endothelialcells. Contribution to the pathophysiology of atherosclerosis? Eur

Cytokine Network 9: 698–698, 1998.91. DISCHER DE, MOHANDAS N, AND EVANS EA. Molecular maps of red cell

deformation: hidden elasticity and “in situ” connectivity. Science

266: 1032–1035, 1994.92. DOOLITTLE WF. Phylogenetic classification and the universal tree.

Science 284: 2124–2128. 1999.93. DRISCOLL M AND CHALFIE M. The mec-4 gene is a member of a family

of Caenorhabditis elegans genes that can mutate to induce neuro-nal degeneration. Nature 349: 588–593, 1991.

94. DRUMMOND HA, PRICE MP, WELSH MJ, AND ABBOUD FM. A molecularcomponent of the arterial baroreceptor mechanotransducer. Neu-

ron 21: 1435–1441, 1998.95. DU H, GU G, WILLIAM CM, AND CHALFIE M. Extracellular proteins

needed for C. elegans mechanosensation. Neuron 16: 183–194,1996.

95a.DULHUNTY AF AND FRANZINI-ARMSTRONG C. The relative contributionsof the folds and caveolae to the surface membrane of frog skeletalmuscle fibers at different sarcomere lengths. J Physiol (Lond) 250:513–539, 1975.

96. DUMONT JN. Oogenesis in Xenopus laevis (Daudin). I. Stages of

oocyte development in laboratory maintained animals. J Morphol

136: 153–180, 1972.97. DUNCAN RK, DYCE OH, AND SAUNDERS JC. Low calcium abolishes tip

links and alters stereocilia motion in chick cochlear hair cells. Hear

Res 124: 69–77, 1998.98. DUNCAN RL AND TURNER CH. Mechanotransduction and the func-

tional response of bone to mechanical strain. Calcif Tissue Int 57:344–358, 1995.

98a.EATOCK RA. Adaptation in hair cells. Annu Rev Neurosci 23: 285–314, 2000.

99. EATOCK RA, COREY DP, AND HUDSPETH AJ. Adaptation of mechano-sensitive transduction in hair cells of the bull frog’s sacculus.

J Neurosci 7: 2821–2836, 1987.100. EDDLEMAN CS, BALLINGER ML, SMYERS ME, GODELL CM, FISHMAN HM,

AND BITTNER GD. Repair of plasmalemmal lesions by vesicles. Proc

Natl Acad Sci USA 94: 4745–4750, 1997.101. ELLIOT JR, NEEDHAM D, DILGER JP, AND HAYDON DA. The effect of

bilayer thickness and tension on gramicidin single channel lifetime.Biochim Biophys Acta 735: 95–103, 1983.

102. ELSON EL. Cellular mechanics as an indicator of cytoskeletal struc-ture and function. Annu Rev Biophys Biophys Chem 17: 397–430,1988.

103. ERIKSON CA AND TRINKAUS JP. Microvilli and blebs as sources ofreserve surface membrane during cell spreading. Exp Cell Res 99:375–384, 1976.

104. ERLER G. Reduction of mechanical sensitivity in an insect mech-anoreceptor correlated with destruction of the tubular body. Cell

Tissue Res 234: 451–461, 1983.105. ERMAKOV YA, AVERBAKH AZ, ARBUZOVA AB, AND SUKHAREV SI. Lipid

and cell membranes in the presence of gadolinium and other ionswith high affinity to lipids. 2. A dipole component of the boundarypotential on membranes with different surface charge. Membr Cell

Biol 12: 411–426, 1998.105a.ESKANDARI S, SNYDER PM, KREMAN M, ZAMPIGHI GA, WELSH MJ, AND

WRIGHT EM. Number of subunits comprising the epithelial sodiumchannel. J Biol Chem 274: 27281–27286, 1999.

106. EVANS E. Structure and deformation properties of red blood cells:concepts and quantitative methods. Methods Enzymol 173: 3–35,1989.

107. EVANS E. Composite membranes and structured interfaces: fromsimple to complex designs in biology. In: Biomembranes Structure

and Function—The State of the Art, edited by Gaber BP andEaswaran KRK. New York: Adenine, 1992, p. 81–101.

108. EVANS E AND HOCHMUTH RM. Mechanical properties of membranes.In: Topics in Membrane and Transport, edited by Kleinzeller A andBronner F. New York: Academic, 1978, vol. 10, p. 1–64.

109. EVANS E AND SKALAK R. Mechanics and thermodynamics of mem-branes. CRC Crit Rev Bioeng 3: 181–418, 1980.

110. EVANS E, YEUNG A, WAUGH R, AND SONG J. Dynamic coupling andnonlocal curvature elasticity in bilayer membranes. In: The Struc-

ture and Conformation of Amphiphilic Membranes. Spriner Pro-

ceedings in Physics, edited by Lipowsky R, Richter D, and KramerK. Berlin: Springer-Verlag, 1992, vol. 66, p. 148–153.

111. FATTAL DR AND BEN-SHAUL A. A molecular model for lipid-proteininteraction in membranes: the role of hydrophobic mismatch. Bio-

phys J 65: 1795–1809, 1993.112. FETTIPLACE R, ANDREWS DM, AND HAYDON DA. The thickness, com-

position and structure of some lipid bilayers and natural mem-branes. J Membr Biol 5: 277–296, 1971.

113. FIRSOV D, GAUTSCHI I, MERILLAT AM, ROSSIER BC, AND SCHILD L. Theheterotetrameric architecture of the epithelial sodium channel(ENaC) EMBO J. 17: 344–352, 1998.

114. FISHMAN HM, TEWARI KP, AND STEIN PG. Injury-induced vesiculationand membrane redistribution in squid giant axon. Biochim Bio-

phys Acta 1023: 421–435, 1990.115. FORRESTER T. An estimate of adenosine triphosphate release into

the venous effluent from exercising human forearm muscle.J Physiol (Lond) 244: 611–628, 1972.

115a.FOX RO AND RICHARDS FM. A voltage-gated ion channel modelinferred from the crystal structure of alamethicin at 1.5 A resolu-tion. Nature 300: 325–330, 1982.

116. FRANCO A AND LANSMAN JB. Calcium entry through stretch-inacti-vated channels in mdx myotubes. Nature 344: 670–673, 1990.

732 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 50: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

117. FRANZ MR, CIMA R, WANG D, PROFITT D, AND KURZ R. Electrophysi-ological effects of myocardial stretch and mechanical determinantsof stretch-activated arrhythmias. Circulation 86: 968–978, 1992.

118. FRICKE B, LINTS R, STEWART G, DRUMMOND H, DODT G, DRISCOLL M,AND VON DURING M. Epithelial Na1 channels and stomatin areexpressed in rat trigeminal mechanosensory neurons. Cell Tissue

Res 299: 327–334, 2000.118a.FUJIWARA T, ODA K, YOKOTA S, TAKATSUKI A, AND IKEHARA Y. Brefeldin

A causes disassembly of the Golgi complex and accumulation ofsecretory protein in the endoplasmic reticulum. J Biol Chem 263:18545–18552, 1998.

119. FURNESS DN, ZETES DE, HACKNEY CM, AND STEELE CR. Kinematicanalysis of shear displacement as a means for operating mechano-transduction channels in the contact region between adjacent ste-reocilia of mammalian cochlear hair cells. Proc R Soc Lond B Biol

Sci 264: 45–51, 1997.119a.GALE JE AND ASHMORE JF. An intrinsic frequency limit to the co-

chlear amplifier. Nature 389: 63–66, 1997.120. GARCIA-ANOVERNOS J AND COREY DP. The molecules of mechanosen-

sation. Annu Rev Neurosci 20: 567–594, 1997.121. GARCIA-ANOVERNOS J, GARCIA JA, LIU JD, AND COREY DP. The nema-

tode degenerin UNC-105 forms ion channels that are activated bydegeneration- or hypercontraction-causing mutations. Neuron 20:1231–1241, 1988.

122. GARCIA-ANOVERNOS J, MA C, AND CHALFIE M. Regulation of Caeno-

rhabditis elegans degenerin proteins by a putative extracellulardomain. Curr Biol 5: 441–448, 1995.

122a.GARDINER DM AND GREY RD. Membrane junctions in Xenopus eggs:their distribution suggests a role in calcium regulation. J Cell Biol

96: 1159–1163, 1983.123. GARTY H AND PALMER LG. Epithelial sodium channels; function,

structure and regulation. Physiol Rev 77: 359–396, 1997.124. GHAZI A, BERRIER C, AJOUZ B, AND BESNARD M. Mechanosensitive ion

channels and their mode of activation. Biochimie 80: 357–362,1998.

125. GIL T, IPSEN JH, MOURITSEN OG, SABBRA MC, SPEROTTO MM, AND

ZUCKERMANN MJ. Theoretical analysis of protein organization inlipid membranes. Biochim Biophys Acta 1376: 245–266, 1998.

126. GIL Z, MAGLEBY KL, AND SILBERBERG SD. Membrane-pipette interac-tions underlie delayed voltage activation of mechanosensitivechannels in Xenopus oocytes. Biophys J 76: 3118–3127, 1999.

127. GIL Z, SILBERBERG SD, AND MAGLEBY KL. Voltage-induced membranedisplacement in patch pipettes activates mechanosensitive chan-nels. Proc Natl Acad Sci USA 96: 14594–14599, 1999.

128. GILLESPIE PG. Molecular machinery of auditory and vestibulartransduction. Curr Opin Neurobiol 5: 449–455, 1995.

129. GILLIS JM. Membrane abnormalities and Ca homeostasis in musclesof the mdx mouse, an animal model of the Duchenne musculardystrophy: a review. Acta Physiol Scand 156: 397–406, 1996.

129a.GILLIS KD. Techniques for membrane capacitance measurements.In: Single Channel Recording (2nd ed.), edited by Sakmann B andNeher E. New York: Plenum, 1995, p. 155–198.

130. GLAUNER KS, MANNUZZU LM, GANDHI CS, AND ISACOFF EY. Spectro-scopic mapping of voltage sensor movement in the Shaker potas-sium channel. Nature 402: 813–817, 1999.

131. GLOCAUER M, FERRIER J, AND MCCULLOCH CAG. Magnetic fields ap-plied to collagen-coated ferric oxide beads induced stretch-acti-vated Ca21 flux in fibroblasts. Am J Physiol Cell Physiol 269:C1093–C1093, 1995.

132. GLOWATZKI E, RUPPERSBERG JP, ZENNER HP, AND RUESCH A. Mechan-ically and ATP-induced currents of mouse outer hair cells areindependent and differentially blocked by D-tubocurarine. Neuro-

pharmacology 36: 1269–1275, 1997.133. GORDON JL. ATP: effects, sources and fate. Biochem J 233: 309–319,

1986.134. GOTOH H AND TAKAHASHI A. Mechanical stimuli induce intracellular

calcium response in a subpopulation of cultured rat senory neu-rons. Neuroscience 92: 1323–1329, 1999.

135. GOTTSCHALDT KM AND VAHLE-HINZ C. Merkel cell receptors: structureand transducer function. Science 214: 183–186, 1981.

136. GOULIAN M, MESQUITA ON, FYGENSON DK, NIELSEN C, ANDERSEN OS,AND LIBCHABER A. Gramicidin channel kinetics under tension. Bio-

phys J 74: 328–337, 1998.

137. GRAF J, RUPNIK M, ZUPANCIC G, AND ZOREC R. Osmotic swelling ofhepatocytes increases membrane conductance but not membranecapacitance. Biophys J 68: 1359–1363, 1995.

138. GRANT CWM, WU SHW, AND MCCONNELL HM. Lateral phase separa-tions in binary lipid mixtures: correlations between spin labels andfreeze-fracture electron microscopic studies. Biochim Biophys

Acta 363: 151–158, 1974.139. GRYGORCZYK R AND HANRAHAN JW. CFTR-independent ATP release

from epithelial cells triggered by mechanical stimuli. Am J Physiol

Cell Physiol 272: C1058–C1066, 1997.140. GRUNER SM. Intrinsic curvature hypothesis for biomembrane lipid

composition: a role for nonbilayer lipids. Proc Natl Acad Sci USA

82: 3665–3669, 1985.141. GU L, LIU W, AND MARTINAC B. Electromechanical coupling model of

gating the large mechanosensitive ion channel (MscL) of Esche-

richia coli by mechanical force. Biophys J 74: 2889–2902, 1998.142. GUHARAY F AND SACHS F. Stretch-activated single ion channel cur-

rents in tissue cultured embryonic chick skeletal muscle. J Physiol

(Lond) 352: 685–701, 1984.143. GUHARAY F AND SACHS F. Mechanotransducer ion channels in chick

skeletal muscle: the effects of extracellular pH. J Physiol (Lond)

363: 119–134, 1985.144. GRUM VL, LI D, MACDONALD RI, AND MONDRAGON A. Structures of two

repeats of spectrin suggest models of flexibility. Cell 98: 523–535,1999.

145. GUREN DWR AND WOLFE J. Lateral tensions and pressures in mem-branes and lipid monolayers. Biochim Biophys Acta 688: 572–580,1982.

146. GUSTIN MC, SACHS F, SIGURDSON W, RUKNUDIN A, BOWMAN C, MORRIS

CE, AND HORN R. Single-channel mechanosensitive currents. Sci-

ence 253: 800, 1991.147. GUSTIN MC, ZHOU XL, MARTINAC B, AND KUNG C. A mechanosensitive

ion channel in the yeast plasma membrane. Science 242: 762–765,1988.

148. HACKNEY CM AND FURNESS DN. Mechanotransduction in vertebratehair cells: structure and function of the stereociliary bundle. Am J

Physiol Cell Physiol 268: C1–C13, 1995.148a.HACKNEY CM, FURNESS DN, BENOS DJ, WOODLEY JF, AND BARRATT J.

Putative immunolocalization of the mechano-electrical transduc-tion channels in mammalian cochlear hair cells. Proc R Soc Lond B

Biol Sci 248: 215–221, 1992.149. HAGMANN J, DAGAN D, AND BURGER MM. Release of endosomal

content induced by plasma membrane tension: video image inten-sification time-lapse analysis. Exp Cell Res 198: 298–304, 1992.

150. HAMILL OP. Membrane ion channels. In: Topics in Molecular Phar-

macology, edited by Burgen A and Roberts AK. Amsterdam:Elsevier, 1983, vol. 3, p. 183–205.

151. HAMILL OP. Potassium and chloride channels in red blood cells. In:Single-Channel Recording, edited by Sakmann B and Neher E.New York: Plenum, 1983, p. 451–471.

152. HAMILL OP, HOLUBEC KV, AND GAO F. Calcium-activated apoptosis infrog (Xenopus laevis) red blood cells (Abstract). J Physiol (Lond)

527P: 46P, 2000.153. HAMILL OP, MARTY A, NEHER E, SAKMANN B, AND SIGWORTH FJ. Im-

proved patch-clamp techniques for high-resolution current record-ing from cells and cell-free membrane patches. Pflugers Arch 391:85–100, 1981.

154. HAMILL OP AND MCBRIDE DW JR. Rapid adaptation of single mech-anosensitive channels in Xenopus oocytes. Proc Natl Acad Sci USA

89: 7462–7466, 1992.155. HAMILL OP AND MCBRIDE DW JR. Molecular mechanisms of mech-

anoreceptor adaptation. News Physiol Sci 9: 53–59, 1994.156. HAMILL OP AND MCBRIDE DW JR. Mechanoreceptive membrane ion

channels. Am Sci 83: 30–37, 1995.157. HAMILL OP AND MCBRIDE DW JR. The cloning of a mechano-gated

membrane ion channel. Trends Neurosci 17: 439–443, 1994.158. HAMILL OP AND MCBRIDE DW JR. The pharmacology of mechano-

gated membrane ion channels. Pharmacol Rev 48: 231–252, 1996.159. HAMILL OP AND MCBRIDE DW JR. Membrane voltage and tension

interactions in the gating of the mechano-gated cation channel inXenopus oocytes (Abstract). Biophys J 70: A348, 1996.

160. HAMILL OP AND MCBRIDE DW JR. Induced membrane hypo/hyper-

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 733

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 51: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

mechanosensitivity: a limitation of patch-clamp recording. Annu

Rev Physiol 59: 621–631, 1997.161. HAMILL OP AND MCBRIDE DW JR. A supramolecular complex under-

lying touch sensitivity. Trends Neurosci 19: 258–261, 1996.162. HANSEN DE, CRAIG CS, AND HONDEGHEM LM. Stretch induced arrhyth-

mias in the isolated canine ventricle. Circulation 81: 1094–1105,1990.

163. HARADA N, ERNST A, AND ZENNER HP. Hyposmotic activation hyper-polarizes outer hair cells of guinea pig cochlea. Brain Res 614:205–211, 1993.

164. HARRIS N, PRESNELL S, AND COHEN FE. Four helix bundle diversity inglobular proteins. J Mol Biol 236: 1356–1368, 1994.

165. HARROUN TA, HELLER WT, WEISS TM, YANG L, AND HUANG HW.Experimental evidence for hydrophobic matching and membrane-mediated interactions in lipid bilayers containing gramicidin. Bio-

phys J 76: 937–945, 1999.166. HARROUN TA, HELLER WT, WEISS TM, YANG L, AND HUANG HW.

Theoretical analysis of hydrophobic matching and membrane-me-diated interactions in lipid bilayers containing gramicidin. Biophys

J 76: 3176–3185, 1999.166a.HARTENECK C, PLANT TD, AND SHULTZ G. From worm to man: three

subfamilies of TRP channels. Trends Neurosci 23: 159–166, 2000.167. HARTZELL HC. Activation of different Cl2 currents in Xenopus oo-

cytes by Ca21 liberated from stores and by capacitive Ca influx.J Gen Physiol 108: 157–175, 1996.

168. HASE CC, LE DAIN AC, AND MARTINAC B. Purification and functionalreconstitution of the recombinant large mechanosensitive ionchannel (MscL) of Escherichia coli. J Biol Chem 270: 18329–18334,1995.

169. HASE CC, LE DAIN AC, AND MARTINAC B. Molecular dissection of thelarge mechanosensitive ion channel (MscL) of Escherichia coli:mutants with altered channel gating and pressure sensitivity. J

Membr Biol 157: 17–25, 1997.170. HASE CC, MINCHIN RF, KLODA A, AND MARTINAC B. Cross-linking

studies and membrane localization and assembly of radiolabelledlarge mechanosensitive ion channel (MscL) of Escherichia coli.

Biochim Biophys Res Commun 232: 777–782, 1997.170a.HAYDON DA. A critique of the black lipid film as a membrane model.

In: Permeability and Function of Biological Membranes, edited byBolis L, Katchalsky A, Keynes RD, Lowenstein WR, and Pethica BA.Amsterdam: North-Holland, 1970.

171. HE K, LUDTKE SJ, HELLER WT, AND HUANG HW. Mechanism of alam-ethicin insertion into lipid bilayers. Biophys J 71: 2669–2679, 1996.

171a.HEIDEMANN SR, KAECH S, BUXBAUM RE, AND MATUS A. Direct obser-vations of the mechanical behaviors of the cytoskeleton in livingfibroblasts. J Cell Biol 145: 109–122, 1999.

172. HELFRICH P AND JAKOBSSON E. Calculation of deformation energiesand conformations in lipid membranes containing gramicidin chan-nels. Biophys J 57: 1075–1084, 1990.

173. HELLER WT, WARING AJ, LEHRER RI, HARROUN TA, WEISS TM, YANG L,AND HUANG HW. Membrane thinning effect of the b-sheet antimi-crobial protegrin. Biochemistry 39: 139–145, 2000.

174. HERRING TL, SLOTIN IM, BALTZ JM, AND MORRIS CE. Neuronal swellingand surface area regulation: elevated intracellular calcium is not arequirement. Am J Physiol Cell Physiol 274: C272–C278, 1998.

175. HILLE B. Pharmacological modification of the sodium channels offrog nerve. J Gen Physiol 51: 199–219, 1968.

176. HILLE B. G-protein-coupled mechanisms and nervous signaling.Neuron 9: 187–195, 1992.

176a.HIRSCHBERG CB, ROBBINS PW, AND ABEIJON C. Transporters of nucle-otide sugars, ATP and nucleotide sulphate in the endoplasmicreticulum and Golgi apparatus. Annu Rev Biochem 67: 49–69, 1998.

177. HISADA T, ORDWAY RW, WALSH JV, AND SINGER JJ. Hyperpolarization-activated cationic channels in smooth muscle cells. Pflugers Arch

417: 493–499, 1991.178. HISADA T, WALSH JV, AND SINGER JJ. Stretch-inactivated cationic

channels in single smooth muscle cells. Pflugers Arch 422: 393–396,1993.

179. HIXON WG AND SEARCY DG. Cytoskeleton in the archaebacteriumThermoplasma acidophilum? Viscosity increase in soluble ex-tracts. Biosystems 29: 151–160, 1993.

180. HO C, SLATER SJ, AND STUBBS CD. Hydration and order in lipidbilayers. Biochemistry 34: 6188–6195, 1995.

181. HOCHMUTH RM AND EVANS EA. Extensional flow of erythrocytemembrane from cell body to elastic tether. I. Analysis. Biophys J

39: 71–81, 1982.182. HOCHMUTH RM, SHAO J, DAI J, AND SHEETZ MP. Deformation and flow

of membrane into tethers extracted from neuronal growth cones.Biophys J 70: 358–369, 1996.

183. HOCHMUTH RM AND WAUGH RE. Erythrocyte membrane elasticityand viscosity. Annu Rev Physiol 49: 209–219, 1987.

183a.HOLT JR, COREY DP, AND EATOCK RA. Mechanoelectrical transduc-tion and adaptation in hair cells of the mouse utricle, a low-frequencey vestibular organ. J Neurosci 17: 8739–8748, 1997.

184. HOMANN U. Fusion and fission of plasma-membranes material ac-comodates for osmotically induced changes in the surface area ofguard-cell protoplasts. Planta 206: 329–333, 1998.

185. HOMANN U AND THIEL G. Unitary exocytotic and endocytotic eventsin guard-cell protoplasts during osmotically driven volumechanges. FEBS Lett 460: 495–499, 1999.

186. HONG K AND DRISCOLL M. A transmembrane domain of the putativechannel subunit MEC-4 influences mechanotransduction and neu-rodegeneration in C. elegans. Nature 367: 470–473, 1994.

186a.HONG K, MANO I, AND DRISCOLL M. In vivo structure-function analysisof Caenorhabditis elegans MEC-4, a candidate mechanosensoryion channel subunit. J Neurosci 20: 2575–2588, 2000.

187. HORIE H, IKUTA S, AND TAKENAKA T. Membrane elasticity of mousedorsal root ganglion neurons decreases with aging. FEBS Lett 269:23–25, 1990.

187a.HOUSLAY MD AND STANLEY KK. Dynamics of Biological Membranes.New York: Wiley, 1982, p. 1–330.

188. HOWARD J AND HUDSPETH AJ. Mechanical relaxation of the hairbundle mediates adaptation in mechanoelectrical transduction bythe bullfrog’s saccular hair cell. Proc Natl Acad Sci USA 84: 3064–3068, 1987.

189. HOWARD J AND HUDSPETH AJ. Compliance of the hair bundle associ-ated with gating of mechanoelectrical transduction channels in thebullfrog’s saccular hair cell. Neuron 1: 189–199, 1988.

190. HOWARD J, ROBERTS WM, AND HUDSPETH AJ. Mechanoelectrical trans-duction by hair cells. Annu Rev Biophys Chem 17: 99–124, 1988.

191. HOYER J, KOEHLER R, HAASE W, AND DISTLER A. Up-regulation ofpressure-activated Ca21 permeable cation channel in intact vacularenothelium of hypertensive rats. Proc Natl Acad Sci USA 93: 11253–11258, 1996.

192. HU H AND SACHS F. Single channel and whole cell studies of mech-anosensitive channels in the chick heart. J Membr Biol 154: 205–216, 1996.

193. HUANG HW. Deformation free energy of bilayer membrane and itseffect on gramicidin channel lifetime. Biophys J 50: 1061–1070,1986.

194. HUANG LYM AND NEHER E. Ca21-dependent exocytosis in the somataor dorsal root ganglion neurons. Neuron 17: 135–145, 1996.

195. HUANG M, GU G, FERGUSON EL, AND CHALFIE M. A stomatin likeprotein necessary for mechanosensation in C. elegans. Nature 378:292–295, 1995.

196. HUDSPETH AJ AND GILLESPIE PG. Pulling strings to tune transduction:adaptation by hair cells. Neuron 12: 1–9, 1994.

197. HULSMANN S, MUSSHOF U, MADEJA M, FISCHER B, AND SPECKMANN EJ.Characterization of ion channels elicited by a stream of fluid duringspontaneous and ligand-induced chloride current oscillations inXenopus laevis oocytes. Pflugers Arch 436: 49–55, 1988.

198. HUTTER OF AND TRAUTWEIN W. Neuromuscular facilitation by stretchof motor nerve endings. J Physiol (Lond) 133: 610–625, 1956.

199. IIDA H, NAKAMURA H, ONO T, OKUMURA M, AND ANRAKU Y. Mid1, anovel Sacharomyces cerevisiae gene encoding a plasma membraneprotein, is required for Ca21 influx and mating. Mol Cell Biol 14:8259–8271, 1994.

199a.IMAI K, TATSUMI H, AND KATAYAMA Y. Mechanosensitive chloridechannels on the growth cones of cultured rat dorsal root ganglionneurons. Neuroscience 97: 347–355, 2000.

200. INGBER DE. Cell tensegrity: defining new rules of biological designthat govern the cytoskeleton. J Cell Sci 104: 613–627, 1993.

201. INGBER DE. Tensegrity: the architectural basis of cellular mechano-transduction. Annu Rev Physiol 59: 575–599, 1997.

202. ISMAILOV II, AWAYDA MS, BERDIEV BK, BUBIEN KK, LUCAS JE, FULLER

734 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 52: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

CM, AND BENOS DJ. Triple barrel organization of ENaC, a clonedepithelial Na1 channel. J Biol Chem 271: 807–816, 1996.

203. ISMAILOV II, BERDIEV BK, SHLYONSKY VG, AND BENOS DJ. Mechanosen-sitivity of an epithelial Na1 channel in planar lipid bilayers: releasefrom Ca21 block. Biophys J 72: 1182–1192, 1997.

204. JENA M, MINORE JF, AND O’NEILL WC. A volume-sensitive, IP3-insen-sitive Ca21 store in vascular smooth muscle. Am J Physiol Cell

Physiol 273: C316–C322, 1997.205. JI HL, FULLER CM, AND BENOS DJ. Osmotic pressure regulates abg-

rENaC expressed in Xenopus oocytes. Am J Physiol Cell Physiol

275: C1182–C1190, 1998.206. JUKKA Y, LEHTONEN A, AND KINNUNEN PKJ. Phospholipase A2 as a

mechanosensor. Biophys J 68: 1888–1894, 1995.207. KANZAKI M. Molecular identification of a eukaryotic, stretch-acti-

vated nonselective cation channel (correction). Science 285: 1493,1999.

208. KANZAKI M, NAGASAWA M, KOJIMA I, SATO C, NARUSE K, SOKABE M, AND

IIDA H. Molecular identification of a eukaryotic, stretch-activatednonselective cation channel. Science 285: 882–886, 1999.

208a.KANZAKI M, NAGASAWA M, KOJIMA I, SATO C, NARUSE K, SOKABE M, AND

IIDA H. Report clarification. Science 288: 1347, 2000.209. KAPLAN JM AND HORVITZ HR. A dual mechanosensory and chemo-

sensory neuron in Caenorhabditis elegans. Proc Natl Acad Sci USA

90: 2227–2231, 1993.210. KATNIK C AND WAUGH R. Electric fields induce reversible changes in

the surface to volume ratio of micropipette-aspirated erythrocytes.Biophys J 57: 865–875, 1990.

211. KATNIK C AND WAUGH R. Alterations of the apparent expansivitymodulus of red blood cell membrane by electric fields. Biophys J

57: 877–882, 1990.212. KATZ B. Depolarization of sensory terminals and the initiation of

impulses in the muscle spindle. J Physiol (Lond) 111: 261–282,1950.

213. KELL A AND GLASER RW. On the mechanical and dynamic propertiesand plant cell membranes: their role in growth, direct gene transferand protoplast fusion. J Theor Biol 160: 41–62, 1993.

214. KELLER SI, BEZRUKOV SM, GUNER SM, TATE MW, VODYANOY I, AND

PARSEGIAN VA. Probability of alamethicin conductance states varieswith nonlamellar tendencey of bilayer phospholipids. Biophys J 65:23–27, 1993.

215. KERNAN M, COWAN D, AND ZUCKER C. Genetic dissection of mech-anosensory transduction of mechanosensory-defective mutationsof Drosophila. Neuron 12: 1195–1206, 1994.

216. KERR JFR, WILEY AH, AND CURRIE AR. Apoptosis: a basic phenom-enon with wide ranging implications in tissue kinetics. Br J Cancer

26: 239–357, 1972.217. KILLIAN JA. Hydrophobic mismatch between proteins and lipids in

membranes. Biochim Biophys Acta 1376: 401–415, 1998.218. KIRBER MT, GUERRERO-HERNANDEZ A, BOWMAN DS, FOGARTY KE, TUFT

RA, SINGER JJ, AND FAY FS. Multiple pathways responsible for thestretch-induced increase in Ca21 concentration in toad stomachsmooth muscle cells. J Physiol (Lond) 524: 3–17, 2000.

219. KIZER N, GUO XL, AND HRUSKA K. Reconstitution of stretch-activatedcation channels by expression of the a-subunit of the epithelialsodium channel cloned from osteoblasts. Proc Natl Acad Sci USA

94: 1013–1018, 1997.220. KLODA A AND MARTINAC B. A novel mechanosensitive ion channel in

the thermophilic cell wall-less archaeon Thermoplasma volcanium

(Abstract). Biophys J 76: A201, 1999.220a.KLODA A AND MARTINAC B. Molecular identification of a mechano-

sensitive ion channel in Archaea. Biophys J 80: 229–240, 2001.221. KLEEMANN W, GRANT CW, AND MCCONNELL HM. Liquid phase sepa-

rations and protein distribution in membranes. J Supramol Struct

2: 609–616, 1974.222. KLEYWEGT GJ AND JONES TTA. Detection, delineation, measurement

and display of cavities in macromolecular structures. Acta Cryst

50: 178–185, 1994.223. KNUTTON S, KACKSON D, GRAHAM JM, MICKLEM KJ, AND PASTERNAK CA.

Microvilli and cell swelling. Nature 262: 52–53, 1976.224. KOHLER R, DISTLER A, AND HOYER J. Increased mechanosensitive

currents in aortic endothelial cells from genetically hypertensiverats. J Hypertens 17: 365–371, 1999.

225. KOO SP, HIGGINS SF, AND BOOTH IR. Regulation of compatible solute

accumulation in Salmonella typhimurium: evidence for glycinebetaine efflux system. J Gen Microbiol 137: 2617–2625, 1991.

226. KOPROWSKI P AND KUBALSKI A. Voltage-independent adaptation ofmechanosensitive channels in Escherichia coli protoplasts. J

Membr Biol 164: 253–262, 1998.227. KRAUS-FRIEDMANN N. Signal transduction and calcium: a suggested

role for the cytoskeleton in inostol 1,4,5-trisphospahte action. Cell

Motil Cytoskelet 28: 279–284, 1994.228. KUBALSKI A. Generation of giant protoplasts of Escherichia coli and

an inner-membrane anion selective conductance. Biochim Biophys

Acta 1238: 177–182, 1995.229. KUBALSKI A, MARTINAC B, LING KY, ADLER J, AND KUNG C. Activities of

a mechanosensitive ion channel in an E. coli mutant lacking themajor lipoprotein. J Membr Biol 131: 151–160, 1993.

230. KUBITSCHECK U, HOMANN U, AND THIEL G. Osmotically evoked shrink-ing of guard-cell protoplasts causes vesicular retrieval of plasmamembrane into the cytoplasm. Planta 210: 423–431, 2000.

231. KUNG C AND SAIMI Y. Solute sensing vs. solvent sensing, a specula-tion. J Eukaryot Microbiol 42: 199–200, 1995.

232. KUNG C, SAIMI Y, AND MARTINAC B. Mechanosensitive ion channels inmicrobes and the early evolutionary origin of solvent sensing. In:Current Topics in Membranes and Transport, edited by Claudio T.New York: Academic, 1990, chapt. 36, p. 9451–9455.

233. KUSTER JE, FRENCH AS, AND SANDERS EJ. The effects of microtubuledissociating agents on the physiology and cytology of the sensoryneuron in the femoral tactile spine of the cockroach, Periplaneta

americana L. Proc R Soc Lond B Biol Sci 219: 397–412, 1983.234. LANE JW, MCBRIDE DW JR, AND HAMILL OP. Amiloride block of the

mechanosensitive cation channel in Xenopus oocytes. J Physiol

(Lond) 441: 347–366, 1991.235. LANSMAN JB, HALLAM TJ, AND RINK TJ. Single stretch-activated ion

channels in vascular endothelial cells as mechanotransducers? Na-

ture 325: 811–813, 1987.236. LAZAROWSKI ER AND HARDEN TK. Quantitation of extracellular UTP

using a sensitive enzymatic assay. Br J Pharmacol 127: 1272–1278,1999.

237. LAZAROWSKI ER, HOMOLYA L, BOUCHER RC, AND HARDEN TK. Directdemonstration of mechanically-induced release of cellular UTP andits implication for uridine nucleotide receptor activation. J Biol

Chem 272: 24348–24354, 1997.238. LECHNER J AND WEILAND F. Structure and biosynthesis of prokary-

otic glycoproteins. Annu Rev Biochem 58: 173–194, 1989.239. LE DAIN AC, SAINT N, KLODA A, GHAZI A, AND MARTINAC B. Mechano-

sensitive ion channels of the archaeon Haloferax volcanii. J Biol

Chem 273: 12116–12119, 1998.240. LEE J, ISHIHARA A, OXFORD G, JOHNSON B, AND JACOBSON K. Regulation

of cell movement is mediated by stretch-activated calcium chan-nels. Nature 400: 382–386, 1999.

241. LEHTONEN JYA AND KINNUNEN PKJ. Phospholipase A2 as a mech-anosensor. Biophys J 68: 1888–1894, 1995.

242. LELE PP, SINCLAIR DC, AND WEDDELL G. The reaction time to touch.J Physiol (Lond) 123: 187–203, 1954.

243. LESAGE F, GUILLEMARE E, FINK M, DUPRAT F, LAZDUNSKI M, ROMEY G,AND BARHANIN J. Twik-1, a ubiquitous human weakly inward recti-fying K1 channel with a novel structure. EMBO J 15: 1004–1011,1996.

244. LEVINA N, TOTEMEYER S, STOKES NR, LOUIS N, JONES MA, AND BOOTH

IR. Protection of Escherichia coli cells against extreme turgor byactivation of MscS and MscL mechanosensitive channels: identifi-cation of genes required for MscS activity. EMBO J 18: 1730–1737,1999.

245. LEWIS C, NEIDHART S, HOLY C, NORTH RA, BUELL A, AND SURPRENANT

A. Coexpression of P2X2 and P2X3 receptor subunits can accountfor ATP-gated currents in Sensory neurons. Nature 377: 432–435,1995.

245a.LEWIS SA. Everything you wanted to know about the bladder epi-thelium but were afraid to ask. Am J Physiol Renal Physiol 278:F867–F874, 2000.

245b.LEWIS SA AND DE MOURA JL. Apical membrane area of rabbit urinarybladder increases by fusion of intracellular vesicles: an electro-physiologcial study. J Membr Biol 82: 123–136, 1984.

246. LINSDELL P AND HANRAHAN JW. Adenosine trisphosphate-dependentasymmetry of anion permeation in the cystic fibrosis transmem-

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 735

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 53: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

brane conductance regulator chloride channel. J Gen Physiol 111:601–614, 1998.

246a.LIU W. The Role of the N-terminal Domain S1 in Gating of the

Large Mechanosensitive Ion Channel (MscL) of E. coli by Mechan-

ical Force (PhD thesis). Nedlands: Univ. of Western Australia, 1998.247. LIU W, DEITMER JW, AND MARTINAC B. Multiple conductances in the

wild-type and wild-type/DG14 MscL hybrid channels (Abstract).Biophys J 76: A203, 1999.

248. LIU J, SCHRANK B, AND WATERSTON RH. Interaction between a puta-tive mechanosensory membrane channel and a collagen. Science

273: 361–364, 1996.249. LOEWENSTEIN WR. The generation of electric activity in a nerve

ending. Ann NY Acad Sci 81: 367–387, 1959.250. LUNA EJ AND HITT AL. Cytoskeleton-plasma membrane interactions.

Science 258: 955–964, 1992.251. LUNDBAEK JA AND ANDERSEN OS. Lysophospholipids modulate chan-

nel activity by altering the mechanical properties of the bilayer.J Gen Physiol 104: 645–673, 1994.

252. LUNDBAEK JA AND ANDERSEN OS. Spring constants for channel-in-duced lipid bilayer deformations estimates using gramicidin chan-nels. Biophys J 76: 889–895, 1999.

253. LUNDBAEK JA, BIRN P, GIRSHMANN J, HANSEN AJ, AND ANDERSEN OS.Membrane stiffness and channel function. Biochemistry 35: 3825–3830, 1996.

254. LUNDBAEK JA, MAER AM, AND ANDERSEN OS. Lipid bilayer electro-static energy, curvature stress, and assembly of gramicidin chan-nels. Biochemistry 36: 5695–5701, 1997.

255. LUSTIG KD, SHIAU AK, BRAKE AJ, AND JULIUS D. Expression cloning ofan ATP receptor from mouse neuroblastoma cells. Proc Natl Acad

Sci USA 90: 5113–5117, 1993.256. MACONOCHIE DJ AND STEINBACH JH. The channel opening rate of

adult- and fetal-type mouse muscle nicotinic receptors activated byacetylcholine. J Physiol (Lond) 506: 53–72, 1998.

257. MAINGRET F, FOSSET M, LESAGE F, LAZDUNSKI M, AND HONORE E. Traakis a mammalian neuronal mechano-gated K1 channel. J Biol Chem

274: 1381–1387, 1999.257a.MAINGRET F, PATEL AJ, LESAGE F, LAZDUNSKI M, AND HONORE E.

Lysophospholipids open the two-pore domain mechano-gated K1

channels TREK-1 and TRAAK. J Biol Chem 275: 10128–10133, 2000.258. MANNSFELDT AG, CARROLL P, STUCKY CL, AND LEWIN GR. Stomatin, a

MEC-2 like protein, is expressed by mammalian sensory neurons.Mol Cell Neurosci 13: 391–404, 1999.

259. MARCHENKO SM AND SAGE SO. A novel mechanosensitive cationicchannels from the endothelium of rat aorta. J Physiol (Lond) 498:419–425, 1997.

260. MARKIN VS AND HUDSPETH AJ. Gating-spring models of mechanoelec-trical transduction by hair cells of the internal ear. Annu Rev

Biophys Biomol Struct 24: 59–83, 1995.261. MARKIN VS AND MARTINAC B. Mechanosensitive ion channels as

reporters of bilayer expansion. A theoretical model. Biophys J 60:1120–1127, 1991.

262. MAROTO R AND HAMILL OP. Integrin-dependent mechanosensitiverelease of ATP from Xenopus oocytes is blocked by Brefeldin A(Abstract). J Physiol (Lond) 527P: 45P, 2000.

263. MARQUIS RE AND HUDSPETH AJ. Effects of extracellullar Ca21 con-centration on hair-bundle stiffness and gating-spring integrity inhair cells. Proc Natl Acad Sci USA 94: 11923–11928, 1997.

264. MARRION NV AND TAVALIN SJ. Selective activation of Ca21-activatedK1 channels by co-localized Ca21 channels in hippocampal neu-rons. Nature 395: 900–905, 1998.

265. MARTINAC B,. Mechanosensitive ion channels: biophysics and phys-iology. In: Thermodynamics of Membrane Receptors and Chan-

nels, edited by Jackson MB. Boca Raton, FL: CRC, 1993, p. 327–351.266. MARTINAC B, ADLER J, AND KUNG C. Mechanosensitive ion channels

of E. coli activated by amphipaths. Nature 348: 261–263, 1990.267. MARTINAC B, BUECHNER M, DELCOUR AH, ADLER J, AND KUNG C.

Pressure-sensitive ion channel in Escherichia coli. Proc Natl Acad

Sci USA 84: 2297–2301, 1987.268. MARTINAC B, DELCOUR AH, BUECHNER M, ADLER J, AND KUNG C.

Mechanosensitive ion channels in bacteria. In: Comparative As-

pects of Mechanoreceptor Systems, edited by Ito F. Berlin: Spring-er-Verlag, 1992, p. 3–18.

269. MARTINAC B AND KLODA A. Mechanosensitive (MS) ion channel of the

archaeon Methanococcus jannashii with structural and functionalhomology to bacterial (MS) channels (Abstract). Biophys J 78:A124, 2000.

269a.MARTINAC B, KLODA A, AND PEROZO E. Structural dynamics of MscLfirst transmembrane segment. A site directed spin-labeling study(Abstract). Biophys J 78: A805, 2000.

270. MATSUMOTO H, BARON CB, AND COBURN RF. Smooth muscle stretch-activated phospholipase C activity. Am J Physiol Cell Physiol 268:C458–C465, 1995.

270a.MAURER JA, ELMORE DE, LESTER HA, AND DOUGHERTY DA. Comparingand contrasting E. coli and M. tuberculosis mechanosensitivechannels (MscL). New gain of function mutations in the loopregion. J Biol Chem 275: 22238–22244, 2000.

271. MCBRIDE DW JR AND HAMILL OP. Pressure-clamp: a method for rapidstep perturbation of mechanosensitive channels. Pflugers Arch 421:606–612, 1992.

272. MCBRIDE DW JR AND HAMILL OP. A fast pressure clamp technique forstudying mechanogated channels. In: Single Channel Recording

(2nd ed.), edited by Sakmann B and Neher E. New York: Plenum,1995, p. 329–340.

273. MCBRIDE DW JR AND HAMILL OP. A simplified fast pressure clamptechnique for studying mechanically gated channels. Methods En-

zymol 294: 482–489, 1999.273a.MCCARTER GC, REICHLING DB, AND LEVINE JD. Mechanical transduc-

tion by dorsal root ganglion neurons in vitro. Neurosci Lett 273:179–182, 1999.

274. MCGOUGH A. Membrane skeleton: how to build a molecular shockabsorber. Curr Biol 9: R887–R889, 1999.

275. MCINTOSH TJ AND SIMON SA. Hydration force and bilayer deforma-tion: a reevaluation. Biochemistry 25: 4058–4066, 1986.

276. MCNEIL PL AND STEINHARDT RA. Loss, restoration, and maintenanceof plasma membrane integrity. J Cell Biol 137: 1–4, 1997.

276a.MELEARD P, GERBEAUD C, BARDUSCO P, JEANDAINE N, MITOV MD, AND

FERNANDEZ-PUENTE L. Mechanical properties of model membranesstudied from shape tranformations of giant vesicles. Biochimie 80:401–413, 1998.

277. METHFESSEL C, WITZEMANN V, TAKAHASHI T, MISHINA M, NUMA S, AND

SAKMANN B. Patch clamp measurements on Xenopus laevis oocytes:currents through endogenous channels and implanted acetylcho-line receptor and sodium channels. Pflugers Arch 407: 577–588,1986.

278. MEYER J, FURNESS DN, ZENNER HP, HACKNEY CM, AND GUMMER AW.Evidence for opening of hair-cell transducer channels after tip-linkloss. J Neurosci 18: 6748–6756, 1998.

279. MILLS LR AND MORRIS CE. Neuronal plasma membrane dynamicsevoked by osmomechanical perturbations. J Membr Biol 166: 223–235, 1998.

280. MILNER P, BODIN P, LOESCH A, AND BURNSTOCK G. Rapid release ofendothelin and ATP from isolated aortic endothelial cells exposedto increased flow. Biochem Biophys Res Commun 170: 649–658,1990.

281. MILTON RL AND CALDWELL JH. Membrane blebbing and tight sealformation: are there hidden artifacts in single-channel patch clamprecordings? Comments Theoret Biol 3: 265–284, 1994.

282. MITCHELL CH, CARRE DA, MCGLINN AM, STONE RA, AND CIVAN MM. Arelease mechanism for stored ATP in ocular ciliary epithelial cells.Proc Natl Acad Sci USA 95: 7174–7178, 1998.

283. MOE PC, BLOUNT P, AND KUNG C. Functional and structural conser-vation in the mechanosensitive channel MscL implicates elementscrucial for mechanosensation. Mol Microbiol 28: 583–592, 1998.

283a.MOE PC, LEVIN G, AND BLOUNT P. Pursuing the roots of mechanosen-sation: a structure-based genetic analysis of the M. tuberculosis

MscL channel (Abstract). Biophys J 78: 137A, 2000.284. MOHANDAS N AND EVANS E. Mechanical properties of the red cell

membrane in relation to molecular structure and defects. Annu

Rev Biophys Biomol Struct 23: 787–818, 1994.285. MOODY WJ AND BOSMA MM. A nonselective cation channel activated

by membrane deformation in oocytes of the Ascidian Boltenia

villosa. J Membr Biol 107: 179–188, 1989.286. MORRIS CE. Mechanosensitive ion channels. J Membr Biol 113:

93–107, 1990.286a.MORRIS CE AND HOMANN U. Cell surface area regulation and mem-

brane tension. J Membr Biol. In press.

736 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 54: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

287. MORRIS CE AND HORN R. Failure to elicit neuronal macroscopicmechanosensitive currents anticipated by single-channel studies.Science 251: 1246–1249, 1991.

288. MORRIS CE AND SIGURDSON WJ. Stretch inactivated ion channelscoexist with stretch activated channels. Science 243: 807–809,1989.

289. MOSBACHER J, MAIER R, FAKLER B, GLATZ A, CRESPO J, AND BILBE G.P2y receptor subtypes differentially couple to inwardly-rectifyingpotassium channels. FEBS Lett 436: 104–110, 1998.

290. MOURITSEN OG AND BLOOM M. Mattress model of lipid-protein inter-actions in membranes. Biophys J 46: 141–153, 1984.

291. MOURITSEN OG AND BLOOM M. Models of lipid-protein interactions inmembranes. Annu Rev Biophys Biomol Struct 22: 145–171, 1993.

292. NAKAMURA F AND STRITTMATTER SM. P2y1, purinergic receptors insensory neurons: contribution to touch-induced impulse genera-tion. Proc Natl Acad Sci USA 93: 10465–10470, 1996.

293. NAKAO M, ONO K, FUJISAWA S, AND IIJIMA T. Mechanical stress-induced Ca21 entry and Cl2 currents in cultured human aorticendothelial cells. Am J Physiol Cell Physiol 276: C238–C249, 1999.

294. NARUSE K AND SOKABE M. Involvement of stretch-activated ion chan-nels in Ca21 mobilization to mechanical stretch in endothelial cells.Am J Physiol Cell Physiol 264: C1037–C1044, 1993.

295. NEEDHAM D AND HOCHMUTH RM. Electro-mechanical permeabiliza-tion of lipid vesicles. Biophys J 55: 1001–1009, 1989.

296. NEEDHAM D AND NUNN RS. Elastic deformation and failure of lipidbilayer membranes containing cholesterol. Biophys J 58: 997–1009,1990.

297. NEHER E AND EIBL HJ. The influence of phospholipid polar groupson gramicidin A channels. Biochim Biophys Acta 464: 37–44, 1977.

298. NICHOL. JA AND HUTTER OF. Tensile strength and dilatational elas-ticity of giant sarcolemmal vesicles shed from rabbit muscle.

J Physiol (Lond) 493: 187–198, 1996.299. NICOLSON T, RUSCH A, FRIEDRICH RW, GRANATO M, RUPPERSBERG JP,

AND NUSSLEIN-VOLHARD C. Genetic analysis of veretebrate sensoryhair cell mechanosensation: the zebrafish circler mutants. Neuron

20: 271–283, 1998.300. NIELSEN C, GOULIAN M, AND ANDERSEN OS. Energetics of inclusion-

induced bilayer deformations. Biophys J 74: 1966–1983, 1998.300a.NIGGEL J, SIGURDSON W, AND SACHS F. Mechanically-induced calcium

movements in astrocytes, bovine aortic endothelial cells and C6glioma cells. J Membr Biol 174: 121–134, 2000.

300b.NIKAIDO H. Outer membrane. In: Escherichia coli and Salmonella.Cellular and Molecular Biology (2nd ed.), edited by Neidhard FC.Washington, DC: ASM, 1996, vol. I.

301. NILIUS B, EGGERMONT J, VOETS T, BUYSE G, MANOLOPOULOS V, AND

DROOGMANS G. Properties of volume-regulated anion channels inmammalian cells. Prog Biophys Mol Biol 68: 69–119, 1997.

302. NORTH RA. Families of ion channels with two hydrophobic seg-ments. Curr Opin Cell Biol 8: 474–483, 1996.

303. NORTH RA AND BARNARD EA. Nucleotide receptors. Curr Opin Neu-

robiol 7: 346–357, 1997.304. OAKLEY AJ, LO BELLO M, RICCI G, FEFERICI G, AND PARKER MW.

Evidence for an induced-fit mechanism operating in Pi class gluta-thione transeferases. Biochemistry 37: 9912–9917, 1998.

305. OAKLEY AJ, MARTINAC B, AND WILCE MCJ. Structure and function ofthe bacterial mechanosensitive channel of large conductance. Pro-

tein Sci 8: 1915–1921, 1999.306. O’CONNELL AM, KOEPPE RE, AND ANDERSEN OS. Kinetics of gramici-

din channel formation in lipid bilayers: transmembrane monomerassociation. Science 250: 1256–1260, 1990.

307. OIKE M, DROOGMANS G, AND NILIUS B. Mechanosensitive Ca21 tran-sients in endothelial cells from human umbilical vein. Proc Natl

Acad Sci USA 91: 2940–2944, 1994.308. OKADA Y. Volume expansion-sensing outward-rectifier Cl2 channel:

fresh start to the molecular identity and volume sensor. Am J

Physiol Cell Physiol 273: C755–C789, 1997.309. OKADA Y, HAZAMA A, HASHIMOTO A, MARUYAMA Y, AND KUBO M. Exo-

cytosis upon osmotic swelling in human epithelial cells. Biochim

Biophys Acta 1107: 201–205, 1992.310. OLESEN SP, CLAPHAM DE, AND DAVIES PF. Haemodynamic shear

stress activates a K1 current in vascular endothelial cells. Nature

331: 168–170, 1988.

311. OLIET SHR AND BOURQUE CW. Mechanosensitive channels transduceosmosensitivity in supraoptic neurons. Nature 364: 341–343, 1993.

312. OLIVER JA AND CHASE HS JR. Changes in luminal flow rate modulatebasal and bradykinin-stimulated cell [Ca21] in aortic endothelium.J Cell Physiol 151: 37–40, 1992.

312a.OLSON JE AND LI GZ. Increased potassium, chloride and taurineconductances in astrocytes during osmotic swelling. Glia 20: 254–261, 1997.

313. OPSAHL L AND WEBB WW. Transduction of membrane tension by theion channel alamethicin. Biophys J 66: 71–74, 1994.

314. OPSAHL LR AND WEBB WW. Lipid-glass adhesion in giga-sealed patchclamped membranes. Biophys J 66: 75–79, 1994.

315. OSTROM RS, GRAGORIAN C, AND INSEL PA. Cellular release of andresponse to ATP as key determinants of the set-point of signaltransduction pathways. J Biol Chem 275: 11735–11739, 2000.

316. OU X, BLOUNT P, HOFFMAN R, KUSANO A, AND KUNG C. Randommutagenesis of a mechanosensitive channel identifies regions ofthe protein crucial for normal function (Abstract). Biophys J 72:A139, 1997.

317. OU X, BLOUNT P, HOFFMAN RJ, AND KUNG C. One face of a transmem-brane helix is crucial for mechanosensitive channel gating. Proc

Natl Acad Sci USA 95: 11471–11475, 1998.318. PACE NR. A molecular view of microbial diversity and the bio-

sphere. Science 276: 734–740, 1997.319. PALMER LG AND FRIDT G. Gating of Na channels in the rat cortical

collecting tubule: effects of voltage and membrane stretch. J Gen

Physiol 107: 35–45, 1996.320. PAOLETTI P AND ASCHER P. Mechanosensitivity of NMDA receptors in

cultured mouse central neurons. Neuron 13: 645–655, 1994.321. PAREKH AB AND PENNER R. Store depletion and calcium influx.

Physiol Rev 77: 902–930, 1997.322. PARSEGIAN VA, FULLER N, AND RAND RP. Measured work of defor-

mation and repulsion of lecithin bilayers. Proc Natl Acad Sci USA

76: 2750–2754, 1979.323. PATEL A, HONORE E, MAINGRET F, LESAGE F, FINK M, DUPRAT F, AND

LAZDUNSKI M. A mammalian two pore domain mechano-gated S-likeK1 channel. EMBO J 17: 4283–4290, 1998.

324. PERES A AND BERNARDINI G. The effective membrane capacity ofXenopus eggs: its relations with membrane conductance and cor-tical granule exocytosis. Pflugers Arch 404: 266–272, 1985.

324a.PEREZ-SAMARTIN AL, MILEDI R, AND ARELLANO RO. Activation of vol-ume-regulated Cl2 channels by ACh and ATP in Xenopus follicles.J Physiol (Lond) 525.3: 721–734, 2000.

325. PEROZO E, CORTES DM, AND CUELLO LG,. Structural rearrangementsunderlying K1-channel activation gating. Science 285: 73–78, 1999.

326. PICKLES JO, COMIS SD, AND OSBORNE MP. Cross-links between stereo-cilia in the guinea pig organ of Corti, and their possible relation tosensory transduction. Hear Res 15: 103–112, 1984.

327. POMMERENKE H, SCHREIBER E, DURR F, NEBE B, HAHNEL C, MOLLER W,AND RYCHLY J. Stimulation of integrin receptors using a magneticdrag force device induces an intracellular free calcium response.Eur J Cell Biol 70: 157–164, 1996.

328. PORTER KR, PRESCOTT D, AND FRYE J. Changes in the surface mor-phology of Chinese hamster cells during the cell cycle. J Cell Biol

57: 815–836, 1973.328a.PUGLIELLI L, MANDON EC, AND HIRSCHBERG CB. Identification, purifi-

cation and characterization of the rat liver Golgi membrane ATPtransporter. J Biol Chem 274: 12665–12669, 1999.

329. RAND RP AND PARSEGIAN VA. The forces between interacting bilayermembranes and the hydration of phospholipid assemblies. In: The

Structure of Biological Membranes, edited by Yeagle P. BocaRaton, FL: CRC, 1992, p. 251–306.

330. RAUCHER D AND SHEETZ MP. Characteristics of a membrane reser-voir buffering membrane tension. Biophys J 77: 1992–2002, 1999.

331. RAUCHER D AND SHEETZ MP. Membrane expansion increases endo-cytosis rate during mitosis. J Cell Biol 144: 497–506, 1999.

332. RAYBOULD HE, GSCHOSSMAN JM, ENNES H, LEMBO T, AND MAYER EA.Involvement of stretch-sensitive calcium flux in mechanical trans-duction in visceral afferents. J Auton Nerv Syst 75: 1–6, 1999.

333. REICHLING DB AND LEVINE JD. Heat transduction in rat sensoryneurons by calcium-dependent activation of a cation channel. Proc

Natl Acad Sci USA 94: 7006–7011, 1997.334. REIFARTH FW, CLAUSS W, AND WEBER WM. Stretch-independent ac-

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 737

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 55: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

tivation of the mechanosensitive cation channel in oocytes of Xe-

nopus laevis. Biochim Biophys Acta 1417: 63–76, 1999.334a.REIS O, WINTER R, AND ZERDA TW. The effect of high external

pressure on DPPC-cholesterol multilamellar vesicles: a pressure-tuning Fourier transform infrared spectroscopy study. Biochim

Biophys Acta 1279: 5–16, 1996.335. REISIN IL, PRAT AG, ABRAHAM EH, AMARA RJ, GREGORY DA, AUSIELLO

DA, AND CANTIELLO HF. The cystic fibrosis transmembrane conduc-tance regulator is a dual ATP and chloride channel. J Biol Chem

269: 20584–20591, 1994.336. REQUENA J, HAYDON DA, AND HLADKY SB. Lenses and the compres-

sion of black lipid membranes by an electric field. Biophys J 15:77–81, 1975.

337. REUZEAU C, MILLS LR, HARRIS JA, AND MORRIS CE. Discrete andreversible vacuole-like dilations induced by osmomechanical per-turbation of neurons. J Membr Biol 145: 33–47, 1995.

338. REYES R, LAURITZEN I, LESAGE F, ETTAICHE M, FOSSET M, AND LAZDUN-SKI M. Imunolocalization of the arachidonic acid and mechanosen-sitive baseline TRAAK potassium channel in the nervous system.Neuroscience 95: 893–901, 2000.

338a.ROSS PE, GARBER SS, AND CAHALAN MD. Membrane chloride conduc-tance and capacitance in Jurkat T lymphocytes during osmoticswelling. Biophys J 66: 169–178, 1994.

339. ROSSIER BC. Mechanosensitivity of the epithelial sodium channel(ENaC): controversey or pseudocontroversey? J Gen Physiol 112:95–96, 1998.

340. ROTIN D, BAR-SARGI D, O’BRODOVITCH H, MERILAINEN J, LEHTO VP,CANESSA CM, ROSSIER BC, AND DOWNEY GP. An SH3 binding region inthe epithelial Na1 channel (aENaC) mediates its localization at theapical membrane. EMBO J 13: 4440–4450, 1994.

341. RUDNEV VS, ERMISHKIN LN, FONINA LA, AND ROVIN YU. The depen-dence of the conductance and life time of gramicidin channels onthe thickness and tension of lipid bilayers. Biochim Biophys Acta

642: 196–202, 1981.342. RUESCH A AND HUMMLER E. Mechano-electrical transduction in mice

lacking the a-subunit of the epithelila sodium channel. Hear Res

131: 170–176, 1999.343. RUESCH A, KROS CJ, AND RICHARDSON GP. Block by amiloride and its

derivatives of mechano-electrical transduction in outer hair cells ofmouse cochlear cultures. J Physiol (Lond) 474: 75–86, 1994.

344. RUFFERT S, LAMBERT C, PETER H, WENDISH F, AND KRAMER R. Efflux ofcompatible solutes in Corynebacterium glutamicum mediated byosmoregulated channel activity. Eur J Biochem 247: 572–580, 1997.

345. RUKNUDIN A, SONG MJ, AND SACHS F. The ultrastructure of patch-clamped membranes: a study using high voltage electron micros-copy. J Cell Biol 112: 125–134, 1991.

346. SABATINI BL AND REGEHR WG. Timing of neurotransmission at fastsynapses in the mammalian brain. Nature 384: 170–172, 1996.

347. SACHS F. Baroreceptor mechanism at the cellular level. Federation

Proc 46: 12–16, 1987.348. SACHS F. Mechanical transduction in biological systems. CRC Crit

Revs Biomed Eng 16: 141–169, 1988.349. SACHS F AND LECAR H. Stochastic models for mechanical transduc-

tion. Biophys J 59: 1143–1145, 1991.350. SACHS F AND MORRIS CE. Mechanosensitive ion channels in nonspe-

cialized cells. Revs Physiol Biochem Pharmacol 132: 1–77, 1998.350a.SACHS F, MORRIS CE, HAMILL OP, BOURQUE CW, AND CHAFKE Y. Does

a stretch-inactivated cation channel integrate osmotic and pepti-dergic signals? Nature 3: 847–848, 2000.

351. SACKIN H. Mechanosensitive channels. Annu Rev Physiol 57: 333–353, 1995.

352. SACKMANN E. Molecular and global structure and dynamics of mem-branes and lipid bilayers. Can J Physiol 68: 999–1012, 1990.

352a.SADOSHIMA J, XU Y, SLAYTER HS, AND IZUMO S. Autocrine release ofangiotension II mediates stretch-induced hypertrophy of cardiacmyocytes in vitro. Cell 75: 977–984, 1993.

353. SAINT N, LACAPERE JJ, GU L, GHAZI A, MARTINAC B, AND RIGAUD JL. Ahexameric transmembrane pore revealed by two-dimensional crys-tallization of the large mechanosensitive ion channel (MscL) ofEscherichia coli. J Biol Chem 273: 14667–14669, 1998.

354. SAKMANN B AND NEHER E. Geometric parameters of pipettes andmembrane patches. In: Single Channel Recording, edited by Sak-mann B and Neher E. New York: Plenum, 1983, p. 31–51.

355. SANSOM MSP. The biophysics of peptide models or ion channels.Prog Biophys Mol Biol 55: 139–235, 1991.

356. SASAKI N, MITSUIYE T, AND NOMA A. Effects of mechanical stretch onmembrane currents of single ventricular myocytes of guinea pigheart. Jpn J Physiol 42: 957–970, 1992.

356a.SAUER H, HESCHELER J, AND WARTENBERG M. Mechanical strain-induced Ca21 waves are propagated via ATP release and purinergicreceptor activation. Am J Physiol Cell Physiol 279: C295–C307,2000.

357. SAVAGE C, XUE Y, MITANI S, HALL D, ZAKHARY R, AND CHALFIE M.Mutations in the Caenorhabditis elegans b-tubulin gene mec-7:effects on microtubule assembly and stability and on tubulin auto-regulation. J Cell Sci 107: 2165–2175, 1994.

357a.SCHILD L, SCHNEEBERGER E, GAUTSCHI I, AND FIRSOV D. Identificaion ofamino acid residues in the alpha, beta, and gamma subunits of theepithelial sodium channel (ENaC) involved in amiloride block andion permeation. J Gen Physiol 109: 15–26, 1997.

358. SCHUTT W AND SACKIN H. A new technique for evaluating volumesensitivity of ion channels. Pflugers Arch 433: 368–375, 1997.

359. SCHLEYER M, SCHMID R, AND BAKKER EP. Transient, specific andextremely rapid release of osmolytes from growing cells of Esch-

erichia coli. Arch Microbiol 160: 424–431, 1993.360. SCHWIEBERT EM. ABC transporter-facilitated conductive transport.

Am J Physiol Cell Physiol 276: C1–C8, 1999.361. SEGERER A, LANAWORTHY TA, AND STETTER KO. Thermoplasma aci-

dophilum and Thermoplasma volcanium sp. from Solfatara FieldsSystem. Appl Microbiol 10: 161–171, 1988.

362. SETOGUCHI M, OHYA Y, ABE I, AND FUJISHIMA M. Stretch-activatedwhole-cell currents in smooth muscle cells from mesenteric resis-tance artery of guinea-pig. J Physiol (Lond) 501: 343–353, 1997.

363. SHARMA RV, CHAPLEAU MW, HAAJDUCZOK G, WACHTEL RE, WAITE LJ,BHALLA RC, AND ABBOUD FM. Mechanical stimulation increases in-tracellular calcium concentration in nodose sensory neurons. Neu-

roscience 666: 433–441, 1995.364. SHEN J, LUSCINSKAS FW, CONNOLLY A, DEWEY CF JR, AND GIMBRONE MA

JR. Fluid shear stress modulates cytosolic free calcium in vascularendothelial cells. Am J Physiol Cell Physiol 262: C384–C390, 1992.

365. SHCHERBATKO A, ONO F, MANDEL G, AND BREHM P. Voltage-dependentsodium channel function is regulated through membrane mechan-ics. Biophys J 77: 1945–1959, 1999.

366. SIGURDSON W, RUKNUDIN A, AND SACHS F. Calcium imaging of me-chanically induced fluxes in tissue cultured chick heart: role ofstretch-activated channels. Am J Physiol Heart Circ Physiol 262:H1110–H1115, 1992.

367. SIGURDSON WJ, SACHS F, AND DIAMOND SL. Mechanical perturbationof cultured human endothelial cells causes rapid increases of in-tracellular calcium. Am J Physiol Heart Circ Physiol 264: H1745–H1752, 1993.

368. SILBERBERG SD AND MAGLEBY KL. Voltage-induced slow activationand deactivation of mechanosensitive channels in Xenopus oo-cytes. J Physiol (Lond) 505: 551–569, 1997.

369. SILINSKY EM AND REDMAN RS. Synchronous release of ATP andneurotransmitter within milliseconds of motor nerve impulse in thefrog. J Physiol (Lond) 492: 815–822, 1996.

369a.SIMONS K AND IKONEN E. Functional rafts in cell membranes. Nature

387: 569–572, 1997.370. SMALL DL AND MORRIS CE. Delayed activation of single mechano-

sensitive channels in Lymnaea neurons. Am J Physiol Cell Physiol

267: C598–C606, 1994.371. SMART OS, BREED J, SMITH GR, AND SANSOM MS. A novel method for

structure-based prediction of ion channel conductance properties.Biophys J 72: 1109–1126, 1997.

371a.SNYDER PM, CHENG C, PRINCE LS, ROGERS JC, AND WELSH MJ. Elec-trophysiological and biochemical evidence that DEG/ENaC cationchannels are composed of nine subunits. J Biol Chem 273: 681–684,1998.

372. SOKABE M, HASEGAWA N, AND YAMAMORI K. Blockers and activatorsfor stretch-activated ion channels of chick skeletal muscle. Ann NY

Acad Sci 707: 417–420, 1993.373. SOKABE M, NARUSE K, AND NUNOGAKI K. Mechanosensitive ion chan-

nels: single channel vs. whole cell activities. Prog Cell Res 6:139–149, 1997.

374. SOKABE M AND SACHS F. The structure and dynamics of patch

738 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 56: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

clamped membrane: a study using differential interference contrastmicroscopy. J Cell Biol 111: 599–606, 1990.

375. SOKABE M, SACHS F, AND JING Z. Quantitative video microscopy ofpatch clamped membranes—stress, strain, capacitance and stretchchannel activation. Biophys J 59: 722–728, 1991.

375a.SOLSONA C, INNOCENTI B, AND FERNANDEZ JM. Regulation of exocy-totic fusion by cell inflation. Biophys J 74: 1061–1073, 1998.

376. SPENCER PS AND SCHAUMBURG HH. An ultrastructural study of theinner core of the Pacinian corpuscle. J Neurocytol 2: 217–235, 1973.

377. SPEICHER DW AND MARCHESI VT. Erythrocyte spectrin is comprisedof many homologous triple helical segments. Nature 311: 177–180,1984.

378. SRINIVASAN KR, KAY RL, AND NAGLE JF. The pressure dependence ofthe lipid bilayer phase transition. Biochemistry 13: 3494–3496,1974.

379. STEBBING MJ, MCCLACHLAN EM, AND SAH P. Are there functional P2X

receptors on cell bodies in intact dorsal root ganglia of rats?Neuroscience 86: 1235–1244, 1998.

380. STEFFENSEN I, BATES WR, AND MORRIS CE. Embryogenesis in thepresence of blockers of mechanosensitive ion channels. Dev

Growth Differ 33: 437–442, 1991.381. STEVENS B AND FIELDS RD. Response of Schwann cells to action

potentials in development. Science 287: 2267–2271, 2000.382. STOCKBRIDGE LL AND FRENCH AS. Stretch-activated cation channels

in human fibroblasts. Biophys J 59: 722–728, 1988.382a.STRONG M, CHANDY KG, AND GUTMAN GA. Molecular evolution of

voltage sensitive ion channel genes: on the origins of electricalexcitability. Mol Biol Evol 10: 221–242, 1993.

382b.SUCHYNA TM, JOHNSON JH, HAMER K, LEYKAM JF, GAGE DA, CLEMO HF,BAUMGARTEN CM, AND SACHS F. Identification of a peptide toxin fromGrammostola spatulata spider venom that blocks cation-selectivestretch-activated channels. J Gen Physiol 115: 583–598, 2000.

382c.SUDHOF TC. The synaptic vesicle cycle. Nature 375: 645–653, 1995.383. SUKHAREV SI. Dynamics of bacterial mechanosensitive channel

complexes (MscL) as viewed by time-resolved current records(Abstract). Biophys J 74: A324, 1998.

384. SUKHAREV SI, BLOUNT P, MARTINAC B, BLATTNER FR, AND KUNG C. Alarge mechanosensitive channel in E. coli encoded by mscL alone.Nature 368: 265–268, 1994.

385. SUKHAREV SI, BLOUNT P, MARTINAC B, GUY HB, AND KUNG C. MscL: amechanosensitive ion channel of Escherichia coli. In: Organellar

Ion Channels and Transporters, edited by Clapham DE and Ehr-lich B. New York: Rockefeller Univ. Press, 1996, vol. 51, p. 133–141.(Soc Gen Physiol Ser)

386. SUKHAREV SI, BLOUNT P, MARTINAC B, AND KUNG C. Functional recon-stitution as an assay for biochemical isolation of channel proteins:application to the molecular identification of a bacterial mechano-sensitive channel. Methods 6: 51–59. 222, 1994.

387. SUKHAREV SI, BLOUNT P, MARTINAC B, AND KUNG C. Mechanosensitivechannels of Escherichia coli: the MscL gene, protein, and activities.Annu Rev Physiol 59: 633–657, 1997.

387a.SUKHAREV SI, DURELL SR, AND GUY HR. Structural models of thegating mechanism of MscL channels (Abstract). Biophys J 78:473A, 2000.

388. SUKAHREV SI, MARTINAC B, ARSHAVSKY VY, AND KUNG C. Two types ofmechanosensitive channels in the E. coli cell envelope: solubiliza-tion and functional reconstitution. Biophys J 65: 177–183, 1993.

389. SUKHAREV SI, SCHROEDER MJ, AND MCCASLIN DR. Stoichiometry of thelarge conductance bacterial mechanosensitive channel of E. coli: abiochemical study. J Membr Biol 171: 183–193, 1999.

390. SUKHAREV SI, SIGURDSON WJ, KUNG C, AND SACHS F. Energetic andspatial parameters for gating of the bacterial large conductancemechanosensitive channel, MscL. J Gen Physiol 113: 525–539, 1999.

391. SULLIVAN MJ, SHARMA RV, WACHTEL RE, CHAPLEAU MM, WAAITE LJ,BHALLA RC, AND ABBOUD FM. Non-voltage-gated Ca21 influx throughmechanosensitive ion channels in aortic baroreceptor neurons.Circ Res 80: 861–867, 1997.

392. SUMPER M, BERG E, MENGELE R, AND STROBEL I. Primary structureand glycosylation of the S-layer protein of Haloferax volcanii. J

Bacteriol 172: 7111–7118, 1990.393. SUSUKI M, SATO J, KUTSUWADA K, OOKI G, AND IMAI M. Cloning of a

stretch-inhibitable nonselective cation channel. J Biol Chem 274:6330–6335, 1999.

394. SVICHAR N, SCMIGOL A, VERKHRANTSKY A, AND KOSTYUK P. IP3-inducedCa21 release in dorsal root ganglion neurones. Neurosci Lett 227:107–110, 1997.

395. SVICHAR N, SCMIGOL A, VERKHRANTSKY A, AND KOSTYUK P. ATP inducesCa21 release from IP3-sensitive Ca21 stores exclusively in largeDRG neurons. Neuroreport 8: 1555–1559, 1997.

396. SYBESMA C. An Introduction to Biophysics. New York: Academic,1977, p. 54.

397. SZABO I, PETRONILLI V, AND ZORATTI M. A patch-clamp study ofBacillus subtilis. Biochim Biophys Acta 1112: 29–38, 1992.

398. SZABO I, PETRONILI V, AND ZORATTI M. A patch-clamp investigation ofthe Streptococcus faecalis membrane. J Membr Biol 131: 203–218,1993.

399. TABAREAN LV, JURANKA P, AND MORRIS CE. Membrane stretch affectsgating modes of a skeletal muscle sodium channel. Biophys J 77:758–774, 1999.

400. TAGLIETTI V AND TOSELLI M. A study of stretch-activated channels inthe membrane of frog oocytes: interactions with Ca21 ions.J Physiol (Lond) 407: 311–328, 1988.

400a.TAKAHASHI A AND GOTOH H. Mechanosensitive currents in culturedsomatosensory neurons. Brain Res 869: 225–230, 2000.

401. TAKEUCHI M, KAWAKAMI M, ISHIHARA T, AMANO T, KONDO K, AND

KATSURA I. An ion channel of the degenerin/epithelial sodium chan-nel superfamily controls the defecation rhythm in Caenorhabditis

elegans. Proc Natl Acad Sci USA 95: 11775–11789, 1998.402. TASAKA Y, NAKAGAWA Y, SATO C, MINO M, UOZUMI N, MARATA N, MUTO

S, AND IIDA H. yam81, a Schizosaccharomyces pombe gene, is apotential homologue of the Saccharomyces cerevisiae MID1 geneencoding a stretch-activated Ca21-permeable channel. Biochem

Biophys Res Commun 269: 265–269, 2000.403. TAVERNARAKIS N AND DRISCOLL M. Molecular modelling of mechano-

transduction in the nematode Caenorhabditis elegans. Annu Rev

Physiol 59: 659–689, 1997.404. TAVERNARAKIS N, SHREFFLER W, WANG S, AND DRISCOLL M. Unc-8, a

DEG/ENaC family member, encodes a subunit of a candidate me-chanically gated channel that modulates C. elegans locomotion.

Neuron 18: 107–119, 1997.405. TAYLOR AL, KUDLOW BA, MARRS KL, GRUENERT DC, GUGGINO WB, AND

SCHWIEBERT EM. Bioluminescence detection of ATP release mech-anisms in epithelia. Am J Physiol Cell Physiol 275: C1391–C1406,1998.

406. TERASAKI M, CHEN LB, AND FUJIWARA K. Microtubules and the endo-plasmic reticulum are highly interdependent structures. J Cell Biol

103: 1557–1568, 1986.407. TERASAKI M, MIYAKE K, AND MCNEIL PL. Large plasma membrane

disruptions are rapidly resealed by Ca21-dependent vesicle-vesiclefusion events. J Cell Biol 139: 63–74, 1997.

408. THIEL G, SUTTER JU, AND HOMANN U. Ca21-sensitive and Ca21-insen-sitive exocytosis in maize coleoptile protoplasts. Pflugers Arch 439:R152–R153, 2000.

409. THURM U. Mechanoreceptors in the cuticle of the honeybee: finestructure and stimulus mechanism. Science 145: 1063–1065, 1964.

410. THURM U. Mechano-electrical transduction. In: Biophysics, editedby Hoppe W, Lohman W, Markl H, and Ziegler H. Berlin: Springer,1983, p. 666–671.

411. THURM U, ERLER G, GOEDDE J, KASTRUP H, KEIL TH, VOELKER W, AND

VOHWINKEL B. Cilia specialized for mechanoreception. J Submi-

crosc Cytol 15: 151–155, 1983.412. TODOROV AT, PETROV AG, AND FENDLER JH. First observation of the

converse flexoelectric effect in lipid bilayers. J Phys Chem 98:3076–3079, 1994.

413. TODOROV LD, MIHAYLOVA-TODOROVA S, WESTFALL TD, SNEDDON P,KENNEDY C, BJUR RA, AND WESTFALL DP. Neuronal release of solublenucleotidases and their role in neurotransmitter inactivation. Na-

ture 387: 76–79, 1997.413a.TOGO T, ALDERTON JM, BI G-Q, AND STEINHARDT RA. The mechanism

of facilitated cell membrane resealing. J Cell Sci 112: 719–731,1999.

414. TRINKHAUS JP. Formation of protrusions of the cell surface duringcell movement. Prog Clin Biol Res 41: 887–906, 1980.

415. TURKANIS SA. Effects of muscle stretch on transmitter release atend-plates of rat diaphragm and frog sartorius muscle. J Physiol

(Lond) 230: 391–403, 1973.

April 2001 MECHANISMS OF MECHANOTRANSDUCTION 739

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from

Page 57: Owen P. Hamill and Boris Martinac 81:685-740, 2001....Hamill, Owen P., and Boris Martinac. Molecular Basis of Mechanotransduction in Living Cells. Physiol Rev 81: 685–740, 2001.—The

416. TURNER PR, FONG P, DENETCLAW WF, AND STEINHARDT RA. Increasedcalcium influx in dystrophic muscle. J Cell Biol 115: 1701–1712,1991.

416a.UEHARA T AND YANAGIMACHI R. Activation of hamster eggs by prick-ing. J Exp Biol 199: 269–274, 1977.

417. UNWIN N. Acetylcholine receptor channel imaged in the open state.Nature 373: 37–43, 1995.

418. VALERA S, HUSSY N, EVENS RJ, ADAMI N, NORTH RA, SURPRENANT A,AND BUELL G. A new class of ligand-gated ion channel defined by P2x

receptor for extracellular ATP. Nature 371: 516–519, 1994.419. VANOYE CG AND REUSS L. Stretch-activated single K1 channels ac-

count for whole-cell currents elicited by swelling. Proc Natl Acad

Sci USA 96: 6511–6516, 1999.420. VIZI ES, SPERLAGH B, AND BARANYI M. Evidence that ATP, released

from the postsynaptic site by noradrenaline, is involved in mechan-ical responses of guinea pig vas deference: cascade transmission.Neuroscience 50: 455–465, 1992.

421. VODYANOY I, BEZRUKOV SM, AND PARSEGIAN VA. Probing alamethicinchannels with water soluble polymers. Size modulated osmoticaction. Biophys J 65: 2097–2105, 1993.

422. VOLK T, FROMTER E, AND KORBMACHER C. Hypertonicity activatesnonselective cation channels in mouse cortical collecting ductcells. Proc Natl Acad Sci USA 92: 8478–8482, 1995.

423. VOLK WA. Basic Microbiology (7th ed). New York: Harper Collins,1992.

424. VOLKNANDT W AND ZIMMERMANN H. Acetylcholine, ATP, and proteo-glycan are common to synaptic vesicles isolated from the electricorgans of electric eel and electric catfish as well as from ratdiaphragm. J Neurochem 47: 1449–1462, 1986.

424a.WALDMANN R AND LAZDUNSKI M. H1-gated cation channels; neuronalacid sensors in the NaC/Deg family of ion channels. Curr Opin

Neurobiol 8: 418–424, 1998.425. WALKER RG, WILLINGHAM AT, AND ZUKER CS. A Drosophilia mech-

anosensory transduction channel. Science 287: 2229–2234, 2000.426. WALLACE BA. Recent advances in the high resolution structural

studies of bacterial channels: gramicidin A. J Struct Biol 121:123–141, 1998.

427. WAN X, HARRIS JA, AND MORRIS CE. Responses of neurons to ex-treme osmomechanical stress. J Membr Biol 145: 21–31, 1995.

428. WAN X, JURANKA P, AND MORRIS CE. Activation of mechanosensitivecurrents in traumatized membrane. Am J Physiol Cell Physiol 276:C318–C327, 1999.

429. WANG N, BUTLER JP, AND INGBER DE. Mechanotransduction acrossthe cell surface and through the cytoskeleton. Science 260: 1124–1127, 1993.

430. WANG Y, ROMAN R, LIDOFSKY SD, AND FITZ JG. Autocrine signallingthrough ATP release represents a novel mechanism for cell volumeregulation. Proc Natl Acad Sci USA 93: 12020–12025, 1996.

431. WAUGH RE. Elastic energy of curvature-driven bump formation onred blood cell membrane. Biophys J 70:1027–1035, 1996.

432. WEBB TE, SIMON J, KRISHEK BJ, BATESON AN, SMART TG, KING BF,BURNSTOCK G, AND BARNARD EA. Cloning and functional expressionof a brain G-protein-coupled ATP receptor. FEBS Lett 324: 219–225, 1993.

433. WELLNER MC AND ISENBERG G. Stretch effects on whole-cell currentsof guinea-pig urinary bladder myocytes. J Physiol (Lond) 480:439–448, 1994.

433a.WHITE SH. Analysis of the torus surrounding planar lipid bilayermembranes. Biophys J 12: 432–445, 1972.

434. WILKINSON NC, GAO F, AND HAMILL OP. Effects of mechanogatedchannel blockers on oocyte growth, fertilization and embryogene-sis. J Membr Biol 165: 161–74, 1998.

435. WIRTZ H AND DOBBS LG. Calcium mobilization and exocytosis afterone mechanical stretch of lung epithelial cells. Science 250: 1266–1269, 1990.

436. WOESE CR. There must be a prokaryote somewhere: microbiology’ssearch for itself. Microbiol Rev 58: 1–9, 1994.

437. WOLFE J, DOWGERT MF, AND STEPONKUS PL. Mechanical study of thedeformation and rupture of the plasma membranes of protoplastsduring osmotic expansions. J Membr Biol 93: 63–74, 1986.

438. WOLFE J AND STEPONKUS PL. The stress-strain relation of the plasmamembrane of isolated plant protoplasts. Biochim Biophys Acta

643: 663–668, 1981.439. WU G, MCBRIDE DWJ, AND HAMILL OP. Mg21 block and inward

rectification of mechanosensitive channels in Xenopus oocytes.Pflugers Arch 435: 572–574, 1998.

440. WU Y, RICCI AJ, AND FETTIPLACE R. Two components of transduceradaptation in auditory hair cells. J Neurophysiol 82: 2171–2181, 1999.

441. YANG XC AND SACHS F. Block of stretch-activated ion channels inXenopus oocytes by gadolinium and calcium ions. Science 243:1068–1071, 1989.

442. YANG XC AND SACHS F. Characterization of stretch-activated ionchannels in Xenopus oocytes. J Physiol (Lond) 431: 103–122, 1990.

443. YOSHIMURA K, BATIZA A, SCHROEDER M, BLOUNT P, AND KUNG C.Hydrophilicity of a single residue within MscL correlates withincreased mechanosensitivity. Biophys J 77: 1960–1972, 1999.

444. ZAMPIGHI GA, KREMAN M, BOORER KJ, LOO DDF, BEZANILLA F, CHANDY

G, HALL JE, AND WRIGHT EM. A method for determining the unitaryfunctional capacity of cloned channels and transporters expressedin Xenopus laevis oocytes. J Membr Biol 148: 65–78, 1995.

445. ZAMPIGHI GA, LOO DDF, KREMAN M, ESKANDARI S, AND WRIGHT EM.Functional and morphological correlates of connexin50 expressedin Xenopus laevis oocytes. J Gen Physiol 113: 507–523, 1999.

445a.ZENG T, BETT GCL, AND SACHS F. Stretch-activated whole cell cur-rents in adult cardiac myocytes. Am J Physiol Heart Circ Physiol

278: H548–H557, 2000.446. ZHANG Y. Membrane Patch and Whole Cell Response of Xenopus

Oocytes to Mechanical, Electrical and Osmotic Stimulation (PhDdissertation). Galveston: Univ. of Texas Medical Branch, 1998.

447. ZHANG Y, GAO G, POPOV VL, WEN JW, AND HAMILL OP. Mechanicallygated channel activity in cytoskeleton deficient plasma membraneblebs and vesicles. J Physiol (Lond) 523: 117–129, 2000.

448. ZHANG Y AND HAMILL OP. On the discrepancy between whole celland membrane patch mechanosensitivity of Xenopus oocytes.J Physiol (Lond) 523: 101–115, 2000.

449. ZHANG Y AND HAMILL OP. Calcium, voltage and osmotic stress sensitivecurrents in Xenopus oocytes and their possible relationship to singlemechanically-gated channels. J Physiol (Lond) 523: 83–99, 2000.

450. ZHAO HB AND SANTOS-SACCHI J. Effect of membrane tension on gapjunctional conductance of supporting cells in Corti’s organ. J Gen

Physiol 112: 447–455, 1998.451. ZHAO YD, YAMOAH EN, AND GILLESPIE PG. Regeneration of broken tip

links and restoration of mechanical transduction in hair cells. Proc

Natl Acad Sci USA 94: 15469–15474, 1996.452. ZHELEV DV AND NEEDHAM D. Tension-stabilized pores in giant vesi-

cles: determination of pore size and pore line tension. Biophys

Biochim Acta 1147: 89–104, 1993.452a.ZHELEV DV, NEEDHAM D, AND HOCHMUTH RH. A novel micropipette

method for measuring the bending modulus of vesicle membranes.Biophys J 67: 720–727, 1994.

453. ZHOU XL AND KUNG C. A mechanosensitive channel in Schizosac-

charomyces pombe. EMBO J. 11: 2869–2875, 1992.454. ZHOU XL, STUMPF MA, HOCH HC, AND KUNG C. A mechanosensitive

channel in whole cells and in membrane patches of the fungusUromyces. Science 253: 1415–1417, 1991.

454a.ZIGANSHIN AU, ZIGANSHINA LE, KING BF, PINTOR J, AND BURNSTOCK G.Effects of P2-purinoceptor antagonists on degradation of adeninenucleotides by ecto-nucleotidases in folliculated oocytes of Xeno-

pus laevis. Biochem Pharmacol 51: 897–901, 1996.455. ZORATTI M AND PETRONILLI V. Ion-conducting channels in a Gram-

positive bacterium. FEBS Lett 240: 105–109, 1988.456. ZORATTI M, PETRONILI V, AND SZABO I. Stretch-activated composite

ion channels in Bacillus subtilis. Biochem Biophys Res Commun

168: 443–450, 1990.457. ZOREC R AND TESTER M. Rapid pressure driven exocytosis-endocy-

tosis cycle in a single plant cell: capacitance measurements inaleurone protoplasts. FEBS Lett 333: 283–286, 1993.

458. ZWAAL RFA AND SCHROIT AJ. Pathophysiologic implications of mem-brane phospholipid asymmetry in blood cells. Blood 89: 1121–1132,1997.

740 OWEN P. HAMILL AND BORIS MARTINAC Volume 81

on Decem

ber 13, 2007 physrev.physiology.org

Dow

nloaded from