Journ

alof

Cell

Scie

nce

The composition and role of cross links inmechanoelectrical transduction in vertebrate sensoryhair cells

Carole M. Hackney1 and David N. Furness2,*1Department of Biomedical Science, University of Sheffield, Western Bank, Sheffield, South Yorkshire, S10 2TN, UK2School of Life Sciences and Institute for Science and Technology in Medicine, Keele University, Staffordshire, ST5 5BG, UK

*Author for correspondence (d.n.furness@keele.ac.uk)

Journal of Cell Science 126, 1–11� 2013. Published by The Company of Biologists Ltddoi: 10.1242/jcs.106120

SummaryThe key components of acousticolateralis systems (lateral line, hearing and balance) are sensory hair cells. At their apex, these cells have

a bundle of specialized cellular protrusions, which are modified actin-containing microvilli, connected together by extracellularfilaments called cross links. Stereociliary deflections open nonselective cation channels allowing ions from the extracellularenvironment into the cell, a process called mechanoelectrical transduction. This produces a receptor potential that causes the release ofthe excitatory neurotransmitter glutamate onto the terminals of the sensory nerve fibres, which connect to the cell base, causing nerve

signals to be sent to the brain. Identification of the cellular mechanisms underlying mechanoelectrical transduction and of some of theproteins involved has been assisted by research into the genetics of deafness, molecular biology and mechanical measurements offunction. It is thought that one type of cross link, the tip link, is composed of cadherin 23 and protocadherin 15, and gates the

transduction channel when the bundle is deflected. Another type of link, called lateral (or horizontal) links, maintains optimal bundlecohesion and stiffness for transduction. This Commentary summarizes the information currently available about the structure, functionand composition of the links and how they might be relevant to human hearing impairment.

Key words: Acousticolateralis systems, Cadherin 23, Lateral links, Mechanoelectrical transduction, Protocadherin 15, Stereocilia, Tip links

IntroductionHair cells are the sensory receptors of the vertebrate

acousticolateralis system. These mechanosensors have an apical

bundle of cellular protrusions (the ‘hairs’), which are actually

modified microvilli and are called stereocilia (Fig. 1A). These

are stiff structures filled with a longitudinal array of actin

filaments. In most sensory hair bundles, the exception being

those of the adult mammalian cochlea, there is also a single

cilium, called the kinocilium, located to one side of the bundle. In

all vertebrate hair-cell epithelia, the stereocilia on individual hair

cells have a precise organization within the bundle, forming rows

that increase in height from short to tall across it. Displacement

of the bundle modulates the opening of mechanoelectrical

transduction (MET) channels; deflection in the direction of the

tallest row depolarizes the hair cell by means of K+ and Ca2+ ions

entering through the MET channels. (Other cations such as Na+

and Mg2+ can pass through the MET channels but are not the

main cations found around the bundle in vivo.) Deflections in the

direction towards the shortest row close the channels and reduce

the flow of ions, resulting in hyperpolarization of the cell. Thus,

the MET channel is a mechanically gated nonselective cationic

channel (see review by Furness and Hackney, 2006).

The operating mechanism is ascribed to a mechanical element

in the bundle that is attached to the MET channel, the gating

spring (Corey and Hudspeth, 1983), which, in effect, ‘pulls open’

the channel during positive deflections. The current produced

then declines in two phases when the deflection is maintained, a

process that is Ca2+ sensitive; the first phase is a rapid adaptation

that might reflect an intrinsic property of the channel itself (Ricci

et al., 1998; Beurg et al., 2006) or, potentially, in some hair cells,

requires myosin XVa, an unconventional myosin (Stepanyan and

Frolenkov, 2009). (Myosins are molecules that interact with actin,

for instance in muscles, and are generally involved in motile

processes in cells; unconventional myosins are those found in

nonmuscle cells.) The second phase is a slower adaptation

(Stauffer and Holt, 2007) for which both myosin XVa, in

cochlear inner hair cells (Stepanyan and Frolenkov, 2009), and

myosin 1c (see review by Gillespie and Cyr, 2004) are candidates.

Although defects in myosin VIIa also affect adaptation they do not

abolish it (Kros et al., 2002), so this myosin is not considered to be

a candidate for the adaptation motor.

A fundamental feature of all hair bundles is the presence of

extracellular filaments that hold the stereocilia together

(Fig. 1B–G) and also link the tallest stereocilia to the

kinocilium. The precise distribution of these filaments varies

between hair cells of different types, but there are two main

categories, tip links and lateral links. The tip link comprises a

filament that connects each of the shorter stereocilium tips

with the neighbouring longer stereocilium located behind it

along the excitatory axis of the bundle (Pickles et al., 1984;

Furness and Hackney, 1985). The lateral links are horizontally

directed links connecting the rows of stereocilia, both along a

row and between rows (Pickles et al., 1984; Furness and

Hackney, 1985) and also connecting the bundle to the

Commentary 1

JCS Advance Online Article. Posted on 2 May 2013

Journ

alof

Cell

Scie

nce

kinocilium (Takumida et al., 1988). It has been suggested that

the gating spring is represented by the tip links because they

are situated in such a way that depolarizing deflections would

stretch them, thereby putting tension on the MET channel gate,

and opposing deflections would relax them, releasing tension

(Box 1). The function of the lateral links is more likely to be in

holding the stereocilia together. However, there are still

uncertainties with this basic view of the role of tip links and

lateral links, such as the nature of the connection between the

channel and the tip link, and there is also the idea that its

structure and composition might indicate it is too stiff to be the

spring (Kachar et al., 2000; Sotomayor et al., 2005). This

article will discuss the structure and function of stereociliary

linkages and what is currently known about their molecular

composition, concentrating primarily on tip links. It will also

indicate some human and animal mutations that have provided

further information about the molecular basis of the linkages

and their role in hearing.

Structure and function of tip links and laterallinksIn this section, the main anatomical features of the tip links and

lateral links, as described primarily from ultrastructural studies,

will be discussed along with suggestions for their respective

functions. The basic features of the hair bundle and links are

illustrated in Fig. 1A–F and summarized in Fig. 1G.

Tip links

The idea that an extracellular link between stereocilia could

represent the MET channel gating spring stems from suggestions

by Thurm (Thurm, 1981) and Hudspeth (Hudspeth, 1982). This

led to anatomical studies attempting to identify specific structures

in the hair bundle that could act as a gating spring, and a study

using high-resolution scanning electron microscopy identified the

tip link in adult hair bundles, a single filament (the upwards-

pointing link) connecting shorter to taller stereocilia in the row

behind (Pickles et al., 1984). During early stages of hair bundle

development, however, the tips of stereocilia have many

filaments connecting them to the adjacent taller stereocilia, and

it is thus not always possible to identify an individual tip link. As

development progresses, the number of links at the tips

diminishes to one main tip link (Furness et al., 1989). In

vestibular organs, multiple links can still be observed at the tips

in some adult hair cells (Furness and Hackney, 2006), making

individual tip links difficult to identify. Thus, it is possible that

the tip link is not absolutely distinct from other links.

Nevertheless in many hair bundles, a single tip link is present

in an ideal location to represent a gating spring, provided the

MET channels are located near one or other end of the link.

Electrophysiological studies (Beurg et al., 2006) and high-speed

Ca2+ imaging (Beurg et al., 2009) indicate that there are 1 or 2

MET channels located at the tips of the shorter stereocilia, thus at

the lower end of the tip link (Box 1) but this technique is of

Box 1. Simplified tip-link model of mechanotransduction

The most widely accepted model for the mechanism underlying the opening of the mechanotransduction channels is that the MET channels are

located at the ends of the tip links with recent evidence favouring the lower end, as shown in the figure (Beurg et al., 2009). The schematic

illustration below shows the links that connect the tall (T) to the intermediate (I), and the intermediate to the short (S) row of stereocilia. Deflection

of the bundle in the excitatory direction (indicated by the arrow) puts tension on the tip link (represented as a gating spring), which mechanically

pulls open the channel (depicted in red in the inset) to allow K+ and Ca2+ ions to flow in (arrow, inset) and depolarise the cell. Opposing

deflections relax the spring and allow the channels to close. Figure adapted from Furness et al., 2010.

Deflection Tiplink

K+

Ca2+

Journal of Cell Science 126 (0)2

Journ

alof

Cell

Scie

nce

insufficient resolution to indicate the precise location. To address

this question using higher resolution electron microscopy, an

approach based on detecting the presence of amiloride-binding

sites on the channels was attempted. Although of relatively low

sensitivity compared with other amiloride-sensitive channels, the

MET channel is blocked by amiloride, an effect which was

exploited by using antibodies to amiloride-binding proteins and

immunoelectron microscopy to localize the MET channel. The

data suggest that the MET channel is located below the tip, ,50–

100 nm away from the tip link (Hackney et al., 1992; Furness

et al., 1996). Kinematic modeling suggests that channels in this

location would still be efficiently activated by bundle deflections

(Furness et al., 1997) but not necessarily by the tip link, although

if the latter stretched the membrane at the stereociliary tip, it

might pull on MET channels that are in this location. As the

identity of the MET channels remains uncertain, there are no

specific reagents such as antibodies that are available to

determine their exact localization and distribution. There have

been a number of candidate proteins suggested as possible MET

channels, most recently two that are known to form channels,

transmembrane channel-like protein 1 (TMC1) and TMC2

(Kawashima et al., 2011).

Despite the difficulties of localizing the MET channel, there is

developmental and physiological evidence that it is operated by

the tip link. Identifiable tip links appear during development

concurrently with the ability of hair cells to transduce (Geleoc

and Holt, 2003). Treatment with the Ca2+-chelating agent 2-bis-

(o-aminophenoxy)ethane-N,N,N9,N9-tetra-acetic acid (BAPTA)

Fig. 1. Organisation of stereocilia and cross links in hair bundles. (A) Scanning electron micrograph (SEM) of the stereociliary bundle of an inner hair cell

from a rat cochlea. The stereocilia form well-defined rows that increase in height along the excitatory axis of the bundle. (B) SEM showing that shorter stereocilia

are connected to taller stereocilia in the adjacent row by fine filaments called tip links (arrows). (C) Close up of a tip link. These links frequently appear to

bifurcate towards the upper attachment site. (D) Transmission electron microscopy (TEM) of a tip link from a guinea-pig cochlear hair cell, showing the fine

filament (arrow) connecting the short and tall stereocilia, within each of which is a dense patch at either end of the link. Note also the region of close approach

(contact region) between stereocilia (arrowhead). (E) TEM of a horizontal section of a group of rat hair-cell stereocilia showing filamentous lateral links (arrows)

connecting those in adjacent rows and within a row. (F) TEM of the three rows of guinea-pig stereocilia, in which the different cross links are visible (left) and a

schematic illustration of interrelationship between the different cross links (right). The contact region is shown, an area that may coincide with a region of ‘top

connectors’ but in mammals also appears to label with antibodies to amiloride sensitivity sodium channels (Hackney et al., 1992). Another region of top

connectors is the lateral links at the tip of the tallest stereocilia, some of which extend into the overlaying tectorial membrane (Verpy et al., 2011). Scale bars:

A53 mm; B5600 nm; C550 nm; D5100 nm; E and F5250 nm.

Stereociliary links 3

Journ

alof

Cell

Scie

nce

destroys tip links and abolishes transduction (Assad et al., 1991)and causes hair-bundle stiffness to decline substantially, which is

thought to be due to loss of the gating spring (Marquis andHudspeth, 1997). Loss of tip links and their regeneration in vitro

is accompanied by loss and recovery of transduction (Zhao et al.,1996). However, one study suggests that destroying the tip link

leaves the MET channel open rather than closed, prompting theidea that the gating spring is actually a distinct structure thatconnects the MET channel to the internal cytoskeleton (Meyer

et al., 1998). In fact, this idea may well be consistent with theproposed molecular model of the tip link whereby the link mightnot itself be the gating spring (see below), an idea supported by

mathematical models (Powers et al., 2012). Determining thestructure and molecular composition of the tip link and how it iscoupled to the MET is therefore crucial to determining themechanism of transduction.

Structural studies in guinea pig cochlea suggest that the tip linkhas a mean length of about 185 nm under conditions of relativelylow Ca2+ concentrations (50 mM) (Furness et al., 2008), similar

to those found in endolymph, the fluid that bathes the apex of thehair cell. In conditions of higher Ca2+, the tip-link length isshorter, ,150 nm. The same study showed that tip-link lengths in

turtle cochlear hair cells are similar to those of mammals whenthey are fixed under similar conditions. However, based on dataobtained from bullfrog saccular hair cells, it has been suggested

that tip links fall into two populations that have different lengths(Auer et al., 2008). These differences might reflect speciesdifferences, differences in location of the hair cells (in thecochlea, the bundles vary systematically in shape and height

along the length of the hair-cell epithelium, the organ of Corti) ordifficulties in finding a sufficient amount of intact tip links forquantitative analysis. However, when we measured a greater

number of tip links than did Auer et al. (Auer et al., 2008), weonly found one population of tip links with regard to their length(Furness et al., 2008).

A number of morphological studies suggest that the tip link isnot a single strand but that it is composed of helically woundstrands (Kachar et al., 2000; Tsuprun and Santi, 2000). The linkalso frequently bifurcates partway along its length so that two

filaments connect to the membrane of the taller stereocilia(Fig. 1C; Hackney and Furness, 1995; Kachar et al., 2000). Thebifurcation has been observed in tip links from all vertebrate

classes except teleost fish (roach) hair cells where onlyunbifurcated tip links have so far been observed (Furness andHackney, 2006). This may be resolved by examining more

samples or different fish species. In addition, not all links in anysample are bifurcated. The molecular architecture of thebifurcation also remains uncertain but its presence or absence

may depend on physiological conditions. For example, itspresence may be affected by Ca2+ sensitivity or changes intension on the tip link. The number of strands that form the link isalso uncertain as some studies suggest that there are two (Kachar

et al., 2000), whereas one study suggests possibly three (Tsuprunand Santi, 2000). The more likely model is that the tip linkconsists of two spirally wound strands, which perhaps unwind

into two separate strands near the top of the link; see below forevaluation of the evidence supporting this model. Alternatively, ithas been suggested that one of the branches is a thinner auxiliary

link (Auer et al., 2008; Rzadzinska and Steel, 2009).

At both the lower (Pickles et al., 1984) and upper terminationsof the tip link, electron-dense patches can be observed beneath

the membrane in transmission electron microscopy (Fig. 1D,F;

Furness and Hackney, 1985; Little and Neugebauer, 1985).Immunocytochemical data suggest these patches contain avariety of proteins. The patch at the lower end of the tip link

lies over the end of the core of actin filaments and containswhirlin (Delprat et al., 2005), a protein encoded by the deafnessgene DFNB31 (Mburu et al., 2003; Mburu et al., 2006), themouse ortholog of which produces a whirling behaviour (Holme

et al., 2002). Myosin XVa (Belyantseva et al., 2003), myosinVIIa, twinfilin, an actin binding protein that inhibits actinpolymerization (Rzadzinska et al., 2009), and EPS8, an actin

regulating protein (Manor et al., 2011; Zampini et al., 2011), arealso present. The patch at the tip-link’s upper end is thought tocontain myosin 1c (Steyger et al., 1998), myosin VIIa, scaffold

protein containing ankyrin repeats and sterile alpha motif domain(SANS) (Grati and Kachar, 2011) and harmonin isoform b(harmonin-b) (Grillet et al., 2009), (Fig. 2), although there is also

evidence for SANS being present at the lower end of the tip link(Caberlotto et al., 2011), where harmonin-b is also found duringdevelopment (Lefevre et al., 2008).

The upper electron-dense plaque appears to have a regular

arrangement of subunits, which might reflect some of theseproteinaceous components (Furness et al., 2008). Both patchregions are enriched in calmodulin (Furness et al., 2002), which

could be associated with the myosins therein and modulate theiractivity by Ca2+ binding and unbinding. For instance, myosinlocated in the upper attachment could change tension on the tiplinks by means of moving along the actin core, hence producing

adaptation (Howard and Hudspeth, 1987).

The tip links appear to insert through the membrane and mightbe anchored into the electron-dense plaques as the link remains in

contact with both attachment sites after detergent treatment,which removes the membrane (Furness et al., 2008). Although, atthe upper end, the attachment might be direct, at the lower end anumber of fine strands have been detected (Osborne et al., 1988;

Kachar et al., 2000; Furness et al., 2008), which potentiallyconnect the tip link from the membrane to the dense material thatcaps the actin core of the stereocilium.

Lateral links

There are various types of horizontal cross links or lateral linksbetween the shafts of the stereocilia (Fig. 1E–G), which project

in all directions, attaching the nearest neighbours to each other(Pickles et al., 1984; Furness and Hackney, 1985). The preciselocation and appearance of these links differs in different species

and types of hair cells (see review by Furness and Hackney,2006). Nevertheless, there are particular regions of horizontalconnections (top connectors) between the tops of adjacent tallstereocilia (another category of top connectors occurs just below

the tips of shorter stereocilia) and close to the lower tip linkinsertion (Verpy et al., 2011). Occasionally a region of closeapproach (the contact region) between the short and taller

stereocilia is seen in this location (Fig. 1; Hackney et al., 1992).Lower down, shaft links exist with a medial stripe where theyoverlap (Fig. 1E), and these link stereocilia both between rows

and within rows (Furness and Hackney, 1985). The medial stripemight represent the joining of links that emanate separately fromeach stereociliary shaft. The extent of these shaft links varies

even within one organ. A particular set of links occurs at theankles of the stereocilia that are present only during developmentin mammalian cochlear hair bundles (Goodyear et al., 2005) but

Journal of Cell Science 126 (0)4

Journ

alof

Cell

Scie

nce

are retained in adult non-mammalian vertebrates, such as chicken(Goodyear and Richardson, 1999), as well as in adult mammalianvestibular hair cells (Jeffries et al., 1986).

How the different types of lateral link are segregated in thebundle is also another interesting question. According to thedisposition of membranes with different structural and functional

properties, there are three different bundle domains (Zhao et al.,2012), which could be associated with where and how differentkinds of lateral links are distributed.

Lateral links probably tie the stereocilia together to enable thebundle to act as a unit, increasing its stiffness and preventing it

from splaying. Measurements of bundle motion imply that thestereocilia are tightly constrained so that there is very little lagbetween the motion of stereocilia on either side of the bundle andthey all move coherently resulting in maximally efficient gating

of the MET channels (Kozlov et al., 2007).

Approximately 48% of the bundle stiffness in chick vestibular

hair cells has been attributed to shaft and upper shaft links, and,43% to ankle and tip links (Bashtanov et al., 2004). Adultmammalian cochlear hair cells appear to lack the ankle links and,

therefore, an even greater proportion of their stiffness mightdepend on the shaft and upper shaft links. However, in mammalsa substantial amount of stiffness has been attributed to the

rootlets that anchor the stereocilia in to the apex of the hair cellbased on studies of a mouse engineered to lack the actin-bundlingprotein TRIOBP (trio and actin binding protein), in which therootlets fail to form. Defects in TRIOBP underlie human deafness

type DFNB28 (Kitajiri et al., 2010).

Nevertheless, mathematical modeling suggests that variation in

the quantity of lateral links within the bundle is likely to impacton stiffness (Silber et al., 2004) and that upper shaft connectorsmight contribute most to bundle stiffness (Cotton and Grant,

2004). This might explain observations that loss of topconnectors substantially affects cochlear non-linearity (Verpyet al., 2011). It has also been reported that the top connectors may

function to reduce viscous drag on the bundle (Kozlov et al.,2011). The extent to which links contribute to stiffness andbundle motion therefore deserves further investigation in avariety of bundle types; for example, a candidate protein

composing the shaft connectors PTPRQ, is absent in maturebasal (high frequency) hair cells of the cochlea, but present in theapical (low frequency) ones (Goodyear et al., 2003). Whether the

lateral links contribute to transduction more directly remains tobe discovered, but there is some evidence that the MET channelsmight be associated with horizontal connections or top

connectors below the tips of shorter stereocilia (Hackney et al.,1992). In addition, bundles lacking the staircase arrangement,such as in myosin-XVa-deficient mice, which do not possessidentifiable tip links, still transduce mechanical stimuli

(Stepanyan and Frolenkov, 2009).

Molecular composition of the linksTip links

Evidence from immunocytochemistry, mutations and knockoutsconfirms that members of the cadherin superfamily are

components of tip links. Cadherin 23 (CDH23) (Sollner et al.,2004) and protocadherin 15 (PCDH15) (Ahmed et al., 2006) havebeen proposed to interact to form tip links (Kazmierczak et al.,

2007) and both proteins are mutated in Usher type I syndrome, anautosomal recessive condition that affects human hearing andvision (Di Palma et al., 2001; Alagramam et al., 2001b).

The twisted and forked appearance of the tip link gives some clues

to its possible structure and composition. Immunocytochemicallabeling suggests that tip links are composed of a homodimer ofCDH23, which forms the upper part of the tip link, and a homodimer

of PCDH15, which forms the lower part (Box 2; Kazmierczak et al.,2007). CDH23 has a short intracellular domain, a transmembranedomain and an ectodomain composed of 27 cadherin repeats(Kazmierczak et al., 2007). PCDH15 has cytoplasmic and

transmembrane domains with an ectodomain that consists of 11cadherin repeats (Kazmierczak et al., 2007). This hypothesis issupported by the fact that regeneration of the links following

chemical disruption in immature hair cells, but not in adults, can beinhibited by PCDH15 and CDH23 fragments (Lelli et al., 2010).There is a Ca2+-binding motif in each cadherin repeat (Sotomayor

et al., 2005; Sotomayor et al., 2010) and the terminal regions ofCDH23 and PCDH15 probably also interact with each other bymeans of a Ca2+-sensitive binding site (Kazmierczak et al., 2007;

Sotomayor et al., 2010; Elledge et al., 2010).

The model of CDH23 being coupled to PCDH15 at the N-termini predicts a length for the tip link of about 180 nm, basedon measurements of the molecules in trans, which is in close

agreement with measurements made by electron microscopy ofguinea pig and turtle hair-cell tip links (Furness et al., 2008) andsimilar to those measured for chick tip links (150–200 nm)

(Tsuprun et al., 2004). The model is also supported byobservations that CDH23 forms dimers in trans, which canpartially unwind into two separate branches at one end if mediumcontaining lower Ca2+ levels are used (Kazmierczak et al., 2007).

On the basis of the biochemical data (Kazmierczak et al.,2007) and our own anatomical data, which show that increasingCa2+ levels decrease the length of the tip link, we have proposed

the hypothesis that the bifurcation represents a varying degree ofunwinding of the two interwoven strands of the CDH23 dimer(Furness et al., 2008). This hypothesis could be tested byanalyzing the dimensions of the forked part of the tip link under

different Ca2+ conditions, which, however, is technically difficultto achieve.

As noted earlier, myosin VIIa and harmonin-b are concentrated

in the stereociliary tips of developing hair bundles of mice. Resultsfrom immunofluorescence and confocal microscopy suggest that,soon after birth, harmonin-b relocates to the position of the upperend of the tip link and that this relocalisation does not occur in

CDH23- or PCDH15-deficient mice (Lefevre et al., 2008). Myosin1c receptors in the stereocilia are also absent in the absence ofCDH23 and after Ca2+ chelation (Phillips et al., 2006), which is

known to destroy the tip links and eliminate transduction (Assadet al., 1991). Another protein that binds to CDH23 is themembrane-associated guanylate kinase MAGI1, which shows a

punctate staining pattern corresponding to that shown by CDH23on the stereocilia both during development and in adults. Thisprotein might be involved in anchoring the tip link to the

cytoskeleton (Xu et al., 2008). During development, the relatedmembrane-associated guanylate kinase (MAGUK) p55 is localizedto the stereociliary tips region, suggesting that these proteins arepart of the regulatory process that forms the actin core and the

transduction apparatus (Mburu et al., 2006).

At the C-terminal ends of both cadherins, there aretransmembrane regions (Kazmierczak et al., 2007), which allow

the two sets of homodimers to pass through the membrane andanchor the tip link to the cytoskeleton (Furness et al., 2008). Atthe upper end of the tip link, CDH23 is likely to interact with

Stereociliary links 5

Journ

alof

Cell

Scie

nce

myosins, harmonin-b and SANS, with harmonin-b linking

CDH23 to the actin filaments in the stereocilium core (Grati

and Kachar, 2011). At the lower end, it is assumed that PCDH15

interacts somehow with the MET channel, enabling it to be gated

by tension in the link. The linkers between the lower end of the

tip link and the dense material at the top of the actin cytoskeleton

that have been reported in a number of studies (Osborne et al.,

1988; Kachar et al., 2000; Furness et al., 2008) might allow the

tip link to interact with the MET channels or might simply anchor

the channels and/or the tip link to the lower electron-dense patch.

Recent studies of a stereociliary tetraspan protein (tetraspan

membrane protein of hair cell stereocilia, or THMS), mutations

in which cause deafness in humans (Shabbir et al., 2006) and

defects in tip-link development and transduction in mice, suggest

that this protein couples PCDH15 to the channel (Box 2; Xiong

et al., 2012). THMS appears to bind preferentially with the

transmembrane domain and a conserved membrane proximal

domain on the cytoplasmic side of PCDH15 so that it could

occupy the region between the latter and the channel.

The evidence suggesting that the tip link is composed of two

homodimers of PCDH15 and CDH23 is fairly strong, at least in

mammals; in avians (chick), a similar distribution of the two

homodimers has been indicated by immunogold labeling

(Goodyear et al., 2010). Whether such a composition is also

present in other vertebrates remains to be determined, but there

are indications that the tip link structure is highly conserved. For

example, turtle tip links appear morphologically identical to

those in mammals (Furness et al., 2008), and CDH23 has been

detected in fish tip links (Sollner et al., 2004). Fish also have

PCDH15a, which is similar to mammalian PCDH15 and is

necessary for hearing (Seiler et al., 2005).

Lateral links

The main categories of lateral links – top connectors, horizontal

links along the shaft and ankle links (and kinociliary links when a

kinocilium is present) – appear to contain different molecules (Box

3). There are a number of candidate molecules, although more might

yet be discovered. Among the lateral links, a subset of top

connectors has been found in mammalian hair cells of the organ of

Corti, the outer hair cells (OHC), that contain the protein stereocilin,

which was first described in a search for candidate deafness genes in

mice that are linked to the human deafness locus DFNB16 (Verpy

et al., 2001; Verpy et al., 2011). Stereocilin is a member of the

mesothelin protein superfamily, which contains proteins that

have been suggested to function as superhelical lectins that

bind to carbohydrate moieties of extracellular glycoproteins

(Sathyanarayana et al., 2009). The region of the top connectors

stains with antibodies against stereocilin and, in a stereocilin-

knockout mouse, the OHC hair bundle shape is normal until

postnatal day 9 (P9), but because the top connectors fail to form

properly, bundle cohesion progressively deteriorates after P9 and tip

links disappear after P15, so that by P60, the stereocilia are no longer

Box 2. A putative model of the molecular composition of the tip link

Studies by immunoelectron microscopy and in trans observations of PCDH15 and CDH23 suggest that the upper part of the tip link is composed

of a dimer of CDH23 (coloured green), wrapped into a helix, and the lower part of a dimer of PCDH15 (coloured blue), also wound helically. The

Ca2+-dependent splitting of the CDH23 dimer (Kazmierczak et al., 2007) is a possible explanation of the bifurcation seen in EM studies of the tip

link. At the upper end, the electron-dense plaque (blue) consists of a number of subunits (dark blue area) that might represent several of the

known protein constituents, which are listed at the upper attachment point of the link. At the lower end, the link is shown attached to two channels

linked to PCDH15 via THMS (inset) as proposed by Xiong et al. (Xiong et al., 2012). The attachment of the TMHS is here depicted as

intracellular, but it is not known where the channel gate is located. Proteins known to be present in the lower tip link dense plaques are also listed

(some of these such as harmonin-b and MAGUK are in involved in development of the stereocilia).

Lower tip link density(EPS8, harmonim b, myosin VIIa,whirlin, twinfilin,calmodulin)

Journal of Cell Science 126 (0)6

Journ

alof

Cell

Scie

nce

linked together. There is also evidence that these stereocilin links

connect the tall OHC stereocilia to the overlying gelatinous flap, the

tectorial membrane. This flap in the organ of Corti contacts the

stereocilia to stimulate these hair cells. Stereocilin thus appears to be

required to form the top connectors and, in their absence, the bundle

cannot retain a normal shape. This in turn affects the ability of the

hair cell to maintain intact tip links.

The shaft links appear to contain protein tyrosine phosphatase

receptor type Q, in the absence of which the lateral links become

elongated (Nayak et al., 2011). As noted above, the extent of

PTPRQ, and therefore shaft links, differs both between different

organs and in different frequency regions of the mammalian

cochlea.

A component of the ankle links is the very large G-protein-

coupled receptor (VLGR1; McGee et al., 2006), which has been

found in the ankle regions together with the transmembrane

proteins usherin and vezatin, and the submembrane protein

whirlin (Kussel-Andermann et al., 2000; Michalski et al., 2007).

VLGR1 is defective in type IIc Usher syndrome, whereas usherin

is defective in type IIa Usher syndrome. Together with vezatin

and whirlin, these proteins might be targeted to ankle links by

myosin VIIa during the development of hair cells (Michalski

et al., 2007). Usherin is also present in adult vestibular hair cells,

which retain ankle links (Adato et al., 2005).

Another protein that might be involved in the development of

the lateral links is transient receptor potential mucolipin 3

(TRPML3), a member of the TRP ion channel protein family. Its

loss in the varitint-waddler mouse causes stereociliary splaying

(Atiba-Davies and Noben-Trauth, 2007).

Box 3. Composition of different categories of laterallink in mature bundles

The diagram illustrates the three categories of inter-stereociliary

lateral link presently recognized as structurally and compositionally

distinct from studies on a variety of different bundle types. Not all

categories are present in all bundles; for example, ankle links are

present in all bundles examined except mature mammalian

cochlear hair bundles. Note that the top connectors on tall

stereocilia in mammalian cochlea connect to the overlying

tectorial membrane. The tip links are shown to illustrate their

relationship with the other links. Other than the aforementioned

mammalian cochlear ankle links, transiently present components

of lateral links (protocadherin 15 and cadherin 23) are not

illustrated. The molecular details of these links are significantly

less well known than that of the tip link.

Box 4. Hearing loss caused by gene defects

Genetic hearing loss contributes to ,50% of the cases of hearing

loss in humans, with the remainder being attributed to

environmental factors (Bitner-Glindzicz, 2002). Studies of genetic

hearing loss have led to the identification of a number of candidate

proteins that have been found to be crucial for mechanotransduction

and forming or maintaining the links between stereocilia (El Amraoui

and Petit, 2005). Usher syndrome, a disease with three clinical

subtypes, Usher types 1–3, each based on separate defects in

several genes, combines deafness and blindness (see review by

Williams, 2008) and thus is a major disability. The most common

form, Usher syndrome type IIa, results in people developing

moderate to severe hearing impairment and arises from mutations

in the USH2A gene that encodes usherin, which has previously

been considered to be a basement membrane protein

(Bhattacharya et al., 2002). However, in rodents, several larger

USH2A transcripts encode transmembrane isoforms of different

lengths, one of which has a cytoplasmic form that is predominantly

expressed in the inner ear but not in the retina, where it is also

typically found. Usherin has been found to localize to the base of

stereocilia, where it binds to harmonin and whirlin, two additional

proteins whose defects lead to the genetic deafness types,

nonsyndromic deafness and Usher type I, respectively (Adato

et al., 2005).

Another protein with relevance for deafness is stereocillin, which

is found at the other end of the stereocilia compared with where

usherin is located and has been localised immunocytochemically

to the top connector region where it links to the tectorial

membrane. It is encoded by the STRC gene, which is defective

in the human deafness type DFNB16 (Verpy et al., 2001). DFNB16

is an autosomal recessive progressive condition with hearing

impairment that appears in childhood and which, in mouse models,

results in a lack of cohesiveness in the hair bundle after P9 (Verpy

et al., 2011).

Usher syndrome type 1D, in which deafness is also

accompanied by blindness, is caused by null alleles of CDH23

that presumably affect both the tip links and also the lateral links

(which contain CDH23 during development). In addition, various

missense mutations in CDH23 result in nonsyndromic deafness

(so-called DFNB12) and, according to studies of salsa mice, which

model this mutation, might cause a weakening or greater

susceptibility of the tip link to damage and affect tip-link

maintenance, thus explaining the progressive hearing loss seen

with this syndrome (Schwander et al., 2009). Other mutations in

CDH23 and PCDH15 produce non-syndromic deafness that have

been identified in Turkish (Duman et al., 2011) and Japanese

families (Nakanishi et al., 2010), confirming that these genes play

a vital role in maintaining normal auditory function. It should be

noted that nonsyndromic deafness, in which hearing loss is not

accompanied by other symptoms, is thought to result from

missense mutations that leave some residual function in the

affected proteins, whereas mutations underlying other types of

deafness cause the loss of the respective protein, resulting in both

hearing loss and other symptoms (Schultz et al., 2011).

Stereociliary links 7

Journ

alof

Cell

Scie

nce

As well as being part of the tip link, both PCDH15 and CDH23are also present in lateral links at some stages of development. As

with other bundle proteins, interaction between PCDH15 andmyosin VIIa is necessary for hair bundles to develop properly(Senften et al., 2006). It has also been reported that a set of lateral

links that contain CDH23 is transiently present duringdevelopment (Michel et al., 2005). These lateral links mightbe recruited into tip links, for example, to repair themechanotransducing apparatus if it is damaged or has been

regenerated (Zhao et al., 1996).

The kinociliary links also contain CDH23 and PCDH15 in an

asymmetrical arrangement, as in the tip link but in the reversedirection (Goodyear et al., 2010). Such links are not present inadult mammalian cochlear hair cells, which lack the kinocilium,but they do occur in mammalian vestibular hair cells and those of

non-mammalian vertebrates. It is possible that these kinociliarylinks might represent a parallel transducer mechanism, separatefrom the usual tip link, or a mechanism for transmitting forces to

the hair bundle from the kinocilium, given that the latter isusually the point of coupling to an overlying accessory structurethat delivers a stimulus to the bundle (see discussion in Goodyear

et al., 2010).

Mutations of Cdh23 and Pcdh15 in mouse and zebra fishand their effects on hair-cell function

A variety of mutations and transgenic animals have allowed thecomposition of the links and their function to be further

investigated. Gene mutations have been found in humans thathave been investigated by using murine and zebrafish geneticexperiments. Usher syndrome, as noted above, provided some

initial clues. CDH23 was found to be one of the proteins affectedin Usher syndrome (Bork et al., 2001) and there is a mouseorthologue of this protein (Di Palma et al., 2001). Moreover, a

mutation in CDH23 was found in zebrafish, and zebrafish withthis mutation have defects in balance and hearing that involveimpaired mechanotransduction, and no tip links in their hairbundles (Sollner et al., 2004).

Studies of Usher syndrome have also implicated mutations ofPCDH15 in the disease (Alagramam et al., 2001b; Ahmed et al.,

2008), for which there is the mouse model Ames waltzer(Alagramam et al., 2001a). Genomic studies of Pcdh15 led to thediscovery of three isoforms in mice with variations in thecytoplasmic domains, CD1, CD2 and CD3, and CD1 and CD3

antibody staining of the stereocilia suggests that these might beassociated with the tip link (Ahmed et al., 2006). Knockout micethat lack CD1 and CD3 still retain tip links and hearing,

suggesting that when one isoform is absent, the others cancompensate. However, in the absence of the CD2 isoform, tiplinks remain but hearing is lost (Webb et al., 2011). Lack of CD2

also results in defects in polarity of the hair bundles, which mightexplain deafness. In these mice, vestibular function is stillretained (Webb et al., 2011).

Cdh23 and Pcdh15 mutations have been investigated in otherrodent models for their effects on hair-cell structure and function.The waltzer mouse has several mutations in CDH23 (Di Palma

et al., 2001; Holme and Steel, 2002, Holme and Steel, 2004), Anexample is the v2J waltzer variant, in which the hair bundlesshow developmental defects and become disturbed [although in

early development, stereociliary ranking (i.e. the increase in rowheight from the modular side to the strial side of the bundle incochlear hair cells) can still be present in portions of the bundle].

The transducer currents measured in organ cultures from theseP2–P3 mice are also abnormal in that the cells respond with

depolarisation to displacements in the normally negative(inhibitory) direction (Alagramam et al., 2011).

The nature of the putative tip links in the waltzer mice is

controversial. One study in early postnatal mice with the v alleleof CDH23 reported that the tip link is still present in some formand, on the basis of this observation, the authors implied that

CDH23 is not required for tip link formation (Rzadzinska andSteel, 2009). The tip links observed were not bifurcated, as theywere in normal wild-type adult mouse controls (Rzadzinska and

Steel, 2009). In the study of the v2J allele, very short non-bifurcated links were also found, and these could not be classedunequivocally as tip links, which run from the tips of the shorterstereocilia to sides of the taller row stereocilia in the next row,

whereas in age-matched controls, bifurcated tip links wereevident (Alagramam et al., 2011). It is possible to reconcile thesedata if the tip link in waltzer mice is incomplete, i.e. lacks the

upper forked part that consists of CDH23.

Another mouse mutant, salsa, suffers from hearing loss due toa missense mutation of CDH23. In contrast to waltzer mice, in

which hair cell development is abnormal, salsa mice haverelatively normal bundles, but progressively lose tip links(Schwander et al., 2009). At P7–P8, the hair cells only have a

relatively mild impairment of transducer function. Hearing inthese mice is also progressively lost; it is impaired at 3 weeks andworsens by the time the mice are 2 months old. Therefore, the

hearing loss has been primarily attributed to a loss of tip links.However, the tip links do not appear to be affected in vestibularhair cells in salsa mice. It is possible that the mutation weakensthe tip links, resulting in their damage and loss in the cochlea

(where the bundles are stimulated at greater frequencies and havedifferent morphologies than those in the vestibular system).

There are several variants of another mouse, the Ames waltzer(av), which have different PCDH15 mutations. In av6J, there isan in-frame deletion that affects the ninth cadherin repeat ofPCDH15 and in av3J, a point mutation that is expected to abolish

expression of PCDH15. Studies of these two mice mutationsshow that av6J mice have an almost normal phenotype, with tiplinks present, albeit in a reduced number compared with the av3J

mice, which have severely affected hair bundles with breaks inthe rows of stereocilia and substantial loss of tip links(Alagramam et al., 2011). The reduced number of tip links in

av6J suggests that the mutation does not prevent tip linkformation but might, as suggested for salsa mice, cause the linksto be more susceptible to damage. The hair cells of the av6J

mouse still show evidence of transduction, although this isreduced in amplitude, consistent with the presence of fewer tiplinks than in controls. The av3J mouse shows a more significantloss of mechanotransduction, consistent with the even greater

reduction in tip links than in av6J mice. Thus, tip links are lesslikely to form in the absence of PCDH15 expression, and the lossof transduction indicates that PCDH15 is likely to be essential for

channel gating.

Future perspectivesThe recent findings from mouse models and human deafness withregard to the composition and structure of tip links and lateral

links discussed above have led to a greater understanding ofmechanoelectrical transduction and to the molecular factorsinvolved. The current working model of the tip link is that it

Journal of Cell Science 126 (0)8

Journ

alof

Cell

Scie

nce

consists of two conjoined homodimers of CDH23 and PCDH15,

with a possible bifurcation as the result of Ca2+-dependent

untwisting of the CDH23 dimer, which is likely to be the basis of

further investigation. Although the function of the tip link as the

gating spring of the MET channel is not indisputable, as there are

still unexplained aspects of gating [such as the structural data of

the tip link, which suggests that it may not have sufficient

elasticity to reflect that measured for the gating spring (Kachar

et al., 2000)], the biochemical and genetic evidence underlying

this model go a long way towards clarifying the role of the tip

link. Taken together they show that: (1) tip link regeneration can

be inhibited by CDH23 and PCDH15 fragments, which results in

concomitant failure of transduction; and (2) various mutations in

these proteins produce hair bundle and cross-link anomalies with

associated deafness (El-Amraoui and Petit, 2005). The location of

the MET channel, which is likely to be at the lower end of the tip

link, needs to be confirmed by directly identifying the channel at

high resolution, for instance using electron microscopy combined

with immunocytochemical methods. Further evidence about how

the channel is coupled to the tip link, perhaps via THMS, remains

to be obtained and the precise role of myosins found at the lower

and upper tip-link insertion points require further investigations.

Answering these questions is necessary to resolve important

aspects of mechanotransduction, such as the molecular

mechanisms of gating and the way in which adaptation is

achieved.

The lateral links and their contribution to transduction remain

even less well understood. For example, if their main role is in

providing stiffness to the bundle, their distribution and

composition might show systematic changes that are associated

with the different frequencies detected by hair cells in different

inner ear organs and in different cochlear locations because

cochlear hair cells detect a wide range of frequencies from a few

hundred to 100,000 Hz or more, across different mammalian

species. This possible correlation is a further avenue of research

that is worthwhile exploring.

From the human perspective, the importance of tip and lateral

links lies in their potential contribution to hearing and balance

disorders if they are lost or do develop not properly. Some of

the genes associated with deafness clearly have a role in the

establishment and maintenance of the cross links, but

the prevalence of such disorders is still not fully appreciated;

the combination of genetic studies of human deafness and

analysis of mouse models is likely to continue to bring new

insights into the structure and function of hair bundle linkages.

FundingThe authors’ research has been supported by Deafness Research UK,the Henry Smith Charity, the Wellcome Trust, the MRC, RNID (nowAction on Hearing Loss) and the Grand Charity.

ReferencesAdato, A., Lefevre, G., Delprat, B., Michel, V., Michalski, N., Chardenoux, S., Weil,

D., El-Amraoui, A. and Petit, C. (2005). Usherin, the defective protein in Usher

syndrome type IIA, is likely to be a component of interstereocilia ankle links in the

inner ear sensory cells. Hum. Mol. Genet. 14, 3921-3932.

Ahmed, Z. M., Goodyear, R., Riazuddin, S., Lagziel, A., Legan, P. K., Behra, M.,

Burgess, S. M., Lilley, K. S., Wilcox, E. R., Riazuddin, S. et al. (2006). The tip-link

antigen, a protein associated with the transduction complex of sensory hair cells, is

protocadherin-15. J. Neurosci. 26, 7022-7034.

Ahmed, Z. M., Riazuddin, S., Aye, S., Ali, R. A., Venselaar, H., Anwar, S.,

Belyantseva, P. P., Qasim, M., Riazuddin, S. and Friedman, T. B. (2008). Gene

structure and mutant alleles of PCDH15: nonsyndromic deafness DFNB23 and type 1

Usher syndrome. Hum. Genet. 124, 215-223.

Alagramam, K. N., Murcia, C. L., Kwon, H. Y., Pawlowski, K. S., Wright, C. G. and

Woychik, R. P. (2001a). The mouse Ames waltzer hearing-loss mutant is caused bymutation of Pcdh15, a novel protocadherin gene. Nat. Genet. 27, 99-102.

Alagramam, K. N., Yuan, H., Kuehn, M. H., Murcia, C. L., Wayne, S., Srisailpathy,C. R., Lowry, R. B., Knaus, R., Van Laer, L., Bernier, F. P. et al. (2001b).Mutations in the novel protocadherin PCDH15 cause Usher syndrome type 1F. Hum.

Mol. Genet. 10, 1709-1718.

Alagramam, K. N., Goodyear, R. J., Geng, R., Furness, D. N., van Aken, A. F.,Marcotti, W., Kros, C. J., Richardson, G. P. (2011). Mutations in protocadherin 15and cadherin 23 affect tip links and mechanotransduction in mammalian sensory haircells. PLoS One 6, e19183.

Assad, J. A., Shepherd, G. M. and Corey, D. P. (1991). Tip-link integrity andmechanical transduction in vertebrate hair cells. Neuron 7, 985-994.

Atiba-Davies, M. and Noben-Trauth, K. (2007). TRPML3 and hearing loss in thevaritint-waddler mouse. Biochim. Biophys. Acta 1772, 1028-1031.

Auer, M., Koster, A. J., Ziese, U., Bajaj, C., Volkmann, N., Wang, N. and Hudspeth,A. J. (2008). Three-dimensional architecture of hair-bundle linkages revealed byelectron-microscopic tomography. J. Assoc. Res. Otolaryngol. 9, 215-224.

Bashtanov, M. E., Goodyear, R. J., Richardson, G. P. and Russell, I. J. (2004). Themechanical properties of chick (Gallus domesticus) sensory hair bundles: relativecontributions of structures sensitive to calcium chelation and subtilisin treatment.J. Physiol. 559, 287-299.

Belyantseva, I. A., Boger, E. T. and Friedman, T. B. (2003). Myosin XVa localizes tothe tips of inner ear sensory cell stereocilia and is essential for staircase formation ofthe hair bundle. Proc. Natl. Acad. Sci. USA 100, 13958-13963.

Beurg, M., Evans, M. G., Hackney, C. M. and Fettiplace, R. (2006). A large-conductance calcium-selective mechanotransducer channel in mammalian cochlearhair cells. J. Neurosci. 26, 10992-11000.

Beurg, M., Fettiplace, R., Nam, J. H. and Ricci, A. J. (2009). Localization of innerhair cell mechanotransducer channels using high-speed calcium imaging. Nat.

Neurosci. 12, 553-558.

Bhattacharya, G., Miller, C., Kimberling, W. J., Jablonski, M. M. and Cosgrove,D. (2002). Localization and expression of usherin: a novel basement membraneprotein defective in people with Usher’s syndrome type IIa. Hear. Res. 163, 1-11.

Bitner-Glindzicz, M. (2002). Hereditary deafness and phenotyping in humans. Br. Med.

Bull. 63, 73-94.

Bork, J. M., Peters, L. M., Riazuddin, S., Bernstein, S. L., Ahmed, Z. M., Ness, S. L.,Polomeno, R., Ramesh, A., Schloss, M., Srisailpathy, C. R. et al. (2001). Ushersyndrome 1D and nonsyndromic autosomal recessive deafness DFNB12 are causedby allelic mutations of the novel cadherin-like gene CDH23. Am. J. Hum. Genet. 68,26-37.

Caberlotto, E., Michel, V., Foucher, I., Bahloul, A., Goodyear, R. J., Pepermans, E.,

Michalski, N., Perfettini, I., Alegria-Prevot, O., Chardenoux, S. et al. (2011).Usher type 1G protein sans is a critical component of the tip-link complex, a structurecontrolling actin polymerization in stereocilia. Proc. Natl. Acad. Sci. USA 108, 5825-5830.

Corey, D. P. and Hudspeth, A. J. (1983). Kinetics of the receptor current in bullfrogsaccular hair cells. J. Neurosci. 3, 962-976.

Cotton, J. and Grant, W. (2004). Computational models of hair cell bundle mechanics:II. Simplified bundle models. Hear. Res. 197, 105-111.

Delprat, B., Michel, V., Goodyear, R., Yamasaki, Y., Michalski, N., El-Amraoui, A.,Perfettini, I., Legrain, P., Richardson, G., Hardelin, J. P. et al. (2005). MyosinXVa and whirlin, two deafness gene products required for hair bundle growth, arelocated at the stereocilia tips and interact directly. Hum. Mol. Genet. 14, 401-410.

Di Palma, F., Holme, R. H., Bryda, E. C., Belyantseva, I. A., Pellegrino, R., Kachar,B., Steel, K. P. and Noben-Trauth, K. (2001). Mutations in Cdh23, encoding a newtype of cadherin, cause stereocilia disorganization in waltzer, the mouse model forUsher syndrome type 1D. Nat. Genet. 27, 103-107.

Duman, D., Sirmaci, A., Cengiz, F. B., Ozdag, H. and Tekin, M. (2011). Screening of38 genes identifies mutations in 62% of families with nonsyndromic deafness inTurkey. Genet. Test. Mol. Biomarkers 15, 29-33.

El-Amraoui, A. and Petit, C. (2005). Usher I syndrome: unravelling the mechanismsthat underlie the cohesion of the growing hair bundle in inner ear sensory cells. J. Cell

Sci. 118, 4593-4603.

Elledge, H. M., Kazmierczak, P., Clark, P., Joseph, J. S., Kolatkar, A., Kuhn, P. andMuller, U. (2010). Structure of the N terminus of cadherin 23 reveals a new adhesionmechanism for a subset of cadherin superfamily members. Proc. Natl. Acad. Sci. USA

107, 10708-10712.

Furness, D. N. and Hackney, C. M. (1985). Cross-links between stereocilia in theguinea pig cochlea. Hear. Res. 18, 177-188.

Furness, D. N. and Hackney, C. M. (2006). The structure and composition of thestereociliary bundle of vertebrate hair cells. In Vertebrate Hair Cells (ed. R. A.Eatock, R. R. Fay and A. N. Popper), pp. 95-153. New York, NY: Springer Scienceand Business Media Inc.

Furness, D. N., Richardson, G. P. and Russell, I. J. (1989). Stereociliary bundlemorphology in organotypic cultures of the mouse cochlea. Hear. Res. 38, 95-109.

Furness, D. N., Hackney, C. M. and Benos, D. J. (1996). The binding site on cochlearstereocilia for antisera raised against renal Na+ channels is blocked by amiloride anddihydrostreptomycin. Hear. Res. 93, 136-146.

Furness, D. N., Zetes, D. E., Hackney, C. M. and Steele, C. R. (1997). Kinematicanalysis of shear displacement as a means for operating mechanotransductionchannels in the contact region between adjacent stereocilia of mammalian cochlearhair cells. Proc. Biol. Sci. 264, 45-51.

Stereociliary links 9

Journ

alof

Cell

Scie

nce

Furness, D. N., Karkanevatos, A., West, B. and Hackney, C. M. (2002). An

immunogold investigation of the distribution of calmodulin in the apex of cochlearhair cells. Hear. Res. 173, 10-20.

Furness, D. N., Katori, Y., Nirmal Kumar, B. and Hackney, C. M. (2008). The

dimensions and structural attachments of tip links in mammalian cochlear hair cellsand the effects of exposure to different levels of extracellular calcium. Neuroscience

154, 10-21.

Furness, D. N., Hackney, C. M. and Evans, M. G. (2010). Localisation of themechanotransducer channels in mammalian cochlear hair cells provides clues to their

gating. J. Physiol. 588, 765-722.

Geleoc, G. S. and Holt, J. R. (2003). Developmental acquisition of sensory transductionin hair cells of the mouse inner ear. Nat. Neurosci. 6, 1019-1020.

Gillespie, P. G. and Cyr, J. L. (2004). Myosin-1c, the hair cell’s adaptation motor.Annu. Rev. Physiol. 66, 521-545.

Goodyear, R. J. and Richardson, G. P. (1999). The ankle-link antigen: an epitopesensitive to calcium chelation associated with the hair-cell surface and the calycal

processes of photoreceptors. J. Neurosci. 19, 3761-3772.

Goodyear, R. J., Legan, P. K., Wright, M. B., Marcotti, W., Oganesian, A., Coats,

S. A., Booth, C. J., Kros, C. J., Seifert, R. A., Bowen-Pope, D. F. et al. (2003). Areceptor-like inositol lipid phosphatase is required for the maturation of developingcochlear hair bundles. J. Neurosci. 23, 9208-9219.

Goodyear, R. J., Marcotti, W., Kros, C. J. and Richardson, G. P. (2005).Development and properties of stereociliary link types in hair cells of the mousecochlea. J. Comp. Neurol. 485, 75-85.

Goodyear, R. J., Forge, A., Legan, P. K. and Richardson, G. P. (2010). Asymmetricdistribution of cadherin 23 and protocadherin 15 in the kinocilial links of avian

sensory hair cells. J. Comp. Neurol. 518, 4288-4297.

Grati, M. and Kachar, B. (2011). Myosin VIIa and sans localization at stereocilia upper

tip-link density implicates these Usher syndrome proteins in mechanotransduction.Proc. Natl. Acad. Sci. USA 108, 11476-11481.

Grillet, N., Xiong, W., Reynolds, A., Kazmierczak, P., Sato, T., Lillo, C., Dumont,

R. A., Hintermann, E., Sczaniecka, A., Schwander, M. et al. (2009). Harmoninmutations cause mechanotransduction defects in cochlear hair cells. Neuron 62, 375-387.

Hackney, C. M. and Furness, D. N. (1995). Mechanotransduction in vertebrate haircells: structure and function of the stereociliary bundle. Am. J. Physiol. 268, C1-C13.

Hackney, C. M., Furness, D. N., Benos, D. J., Woodley, J. F. and Barratt, J. (1992).Putative immunolocalization of the mechanoelectrical transduction channels in

mammalian cochlear hair cells. Proc. Biol. Sci. 248, 215-221.

Holme, R. H. and Steel, K. P. (2002). Stereocilia defects in waltzer (Cdh23), shaker1(Myo7a) and double waltzer/shaker1 mutant mice. Hear. Res. 169, 13-23.

Holme, R. H. and Steel, K. P. (2004). Progressive hearing loss and increasedsusceptibility to noise-induced hearing loss in mice carrying a Cdh23 but not a Myo7a

mutation. J. Assoc. Res. Otolaryngol. 5, 66-79.

Holme, R. H., Kiernan, B. W., Brown, S. D. and Steel, K. P. (2002). Elongation of

hair cell stereocilia is defective in the mouse mutant whirler. J. Comp. Neurol. 450,94-102.

Howard, J. and Hudspeth, A. J. (1987). Mechanical relaxation of the hair bundlemediates adaptation in mechanoelectrical transduction by the bullfrog’s saccular haircell. Proc. Natl. Acad. Sci. USA 84, 3064-3068.

Hudspeth, A. J. (1982). Extracellular current flow and the site of transduction byvertebrate hair cells. J. Neurosci. 2, 1-10.

Jeffries, D. J., Pickles, J. O., Osborne, M. P., Rhys-Evans, P. H. and Comis, S. D.

(1986). Crosslinks between stereocilia in hair cells of the human and guinea pigvestibular labyrinth. J. Laryngol. Otol. 100, 1367-1374.

Kachar, B., Parakkal, M., Kurc, M., Zhao, Y. and Gillespie, P. G. (2000). High-resolution structure of hair-cell tip links. Proc. Natl. Acad. Sci. USA 97, 13336-13341.

Kawashima, Y., Geleoc, G. S., Kurima, K., Labay, V., Lelli, A., Asai, Y.,

Makishima, T., Wu, D. K., Della Santina, C. C., Holt, J. R. et al. (2011).Mechanotransduction in mouse inner ear hair cells requires transmembrane channel-

like genes. J. Clin. Invest. 121, 4796-4809.

Kazmierczak, P., Sakaguchi, H., Tokita, J., Wilson-Kubalek, E. M., Milligan, R. A.,

Muller, U. and Kachar, B. (2007). Cadherin 23 and protocadherin 15 interact toform tip-link filaments in sensory hair cells. Nature 449, 87-91.

Kitajiri, S., Sakamoto, T., Belyantseva, I. A., Goodyear, R. J., Stepanyan, R.,

Fujiwara, I., Bird, J. E., Riazuddin, S., Riazuddin, S., Ahmed, Z. M. et al. (2010).Actin-bundling protein TRIOBP forms resilient rootlets of hair cell stereociliaessential for hearing. Cell 141, 786-798.

Kozlov, A. S., Risler, T. and Hudspeth, A. J. (2007). Coherent motion of stereociliaassures the concerted gating of hair-cell transduction channels. Nat. Neurosci. 10, 87-

92.

Kozlov, A. S., Baumgart, J., Risler, T., Versteegh, C. P. and Hudspeth, A. J. (2011).

Forces between clustered stereocilia minimize friction in the ear on a subnanometrescale. Nature 474, 376-379.

Kros, C. J., Marcotti, W., van Netten, S. M., Self, T. J., Libby, R. T., Brown, S. D.,

Richardson, G. P. and Steel, K. P. (2002). Reduced climbing and increased slippingadaptation in cochlear hair cells of mice with Myo7a mutations. Nat. Neurosci. 5, 41-47.

Kussel-Andermann, P., El-Amraoui, A., Safieddine, S., Nouaille, S., Perfettini, I.,

Lecuit, M., Cossart, P., Wolfrum, U. and Petit, C. (2000). Vezatin, a novel

transmembrane protein, bridges myosin VIIA to the cadherin-catenins complex.EMBO J. 19, 6020-6029.

Lefevre, G., Michel, V., Weil, D., Lepelletier, L., Bizard, E., Wolfrum, U., Hardelin,J. P. and Petit, C. (2008). A core cochlear phenotype in USH1 mouse mutantsimplicates fibrous links of the hair bundle in its cohesion, orientation and differentialgrowth. Development 135, 1427-1437.

Lelli, A., Kazmierczak, P., Kawashima, Y., Muller, U. and Holt, J. R. (2010).Development and regeneration of sensory transduction in auditory hair cells requiresfunctional interaction between cadherin-23 and protocadherin-15. J. Neurosci. 30,11259-11269.

Little, K. F. and Neugebauer, D.-C. (1985). Interconnections between the stereovilli ofthe fish inner ear. II. Systematic investigation of saccular hair bundles from Rutilusrutilus (Teleostei). Cell Tissue Res. 242, 427-432.

Manor, U., Disanza, A., Grati, M., Andrade, L., Lin, H., Di Fiore, P. P., Scita,

G. and Kachar, B. (2011). Regulation of stereocilia length by myosin XVa andwhirlin depends on the actin-regulatory protein Eps8. Curr. Biol. 21, 167-172.

Marquis, R. E. and Hudspeth, A. J. (1997). Effects of extracellular Ca2+ concentrationon hair-bundle stiffness and gating-spring integrity in hair cells. Proc. Natl. Acad. Sci.

USA 94, 11923-11928.

Mburu, P., Mustapha, M., Varela, A., Weil, D., El-Amraoui, A., Holme, R. H.,

Rump, A., Hardisty, R. E., Blanchard, S., Coimbra, R. S. et al. (2003). Defects inwhirlin, a PDZ domain molecule involved in stereocilia elongation, cause deafness inthe whirler mouse and families with DFNB31. Nat. Genet. 34, 421-428.

Mburu, P., Kikkawa, Y., Townsend, S., Romero, R., Yonekawa, H. and Brown,S. D. (2006). Whirlin complexes with p55 at the stereocilia tip during hair celldevelopment. Proc. Natl. Acad. Sci. USA 103, 10973-10978.

McGee, J., Goodyear, R. J., McMillan, D. R., Stauffer, E. A., Holt, J. R., Locke,K. G., Birch, D. G., Legan, P. K., White, P. C., Walsh, E. J. et al. (2006). The verylarge G-protein-coupled receptor VLGR1: a component of the ankle link complexrequired for the normal development of auditory hair bundles. J. Neurosci. 26, 6543-6553.

Meyer, J., Furness, D. N., Zenner, H. P., Hackney, C. M. and Gummer, A. W.

(1998). Evidence for opening of hair-cell transducer channels after tip-link loss.J. Neurosci. 18, 6748-6756.

Michalski, N., Michel, V., Bahloul, A., Lefevre, G., Barral, J., Yagi, H.,

Chardenoux, S., Weil, D., Martin, P., Hardelin, J. P. et al. (2007). Molecularcharacterization of the ankle-link complex in cochlear hair cells and its role in the hairbundle functioning. J. Neurosci. 27, 6478-6488.

Michel, V., Goodyear, R. J., Weil, D., Marcotti, W., Perfettini, I., Wolfrum, U.,

Kros, C. J., Richardson, G. P. and Petit, C. (2005). Cadherin 23 is a component ofthe transient lateral links in the developing hair bundles of cochlear sensory cells.Dev. Biol. 280, 281-294.

Nakanishi, H., Ohtsubo, M., Iwasaki, S., Hotta, Y., Takizawa, Y., Hosono, K.,Mizuta, K., Mineta, H. and Minoshima, S. (2010). Mutation analysis of theMYO7A and CDH23 genes in Japanese patients with Usher syndrome type 1. J. Hum.

Genet. 55, 796-800.

Nayak, G., Goodyear, R. J., Legan, P. K., Noda, M. and Richardson, G. P. (2011).Evidence for multiple, developmentally regulated isoforms of Ptprq on hair cells ofthe inner ear. Dev. Neurobiol. 71, 129-141.

Osborne, M. P., Comis, S. D. and Pickles, J. O. (1988). Further observations on thefine structure of tip links between stereocilia of the guinea pig cochlea. Hear. Res. 35,99-108.

Phillips, K. R., Tong, S., Goodyear, R., Richardson, G. P. and Cyr, J. L. (2006).Stereociliary myosin-1c receptors are sensitive to calcium chelation and absent fromcadherin 23 mutant mice. J. Neurosci. 26, 10777-10788.

Pickles, J. O., Comis, S. D. and Osborne, M. P. (1984). Cross-links between stereociliain the guinea pig organ of Corti, and their possible relation to sensory transduction.Hear. Res. 15, 103-112.

Powers, R. J., Roy, S., Atilgan, E., Brownell, W. E., Sun, S. X., Gillespie, P. G. and

Spector, A. A. (2012). Stereocilia membrane deformation: implications for the gatingspring and mechanotransduction channel. Biophys. J. 102, 201-210.

Ricci, A. J., Wu, Y.-C. and Fettiplace, R. (1998). The endogenous calcium buffer andthe time course of transducer adaptation in auditory hair cells. J. Neurosci. 18, 8261-8277.

Rzadzinska, A. K. and Steel, K. P. (2009). Presence of interstereocilial links in waltzermutants suggests Cdh23 is not essential for tip link formation. Neuroscience 158, 365-368.

Rzadzinska, A. K., Nevalainen, E. M., Prosser, H. M., Lappalainen, P. and Steel,

K. P. (2009). MyosinVIIa interacts with Twinfilin-2 at the tips of mechanosensorystereocilia in the inner ear. PLoS ONE 4, e7097.

Sathyanarayana, B. K., Hahn, Y., Patankar, M. S., Pastan, I. and Lee, B. (2009).Mesothelin, Stereocilin, and Otoancorin are predicted to have superhelical structureswith ARM-type repeats. BMC Struct. Biol. 9, 1.

Schultz, J. M., Bhatti, R., Madeo, A. C., Turriff, A., Muskett, J. A., Zalewski, C. K.,King, K. A., Ahmed, Z. M., Riazuddin, S., Ahmad, N. et al. (2011). Allelichierarchy of CDH23 mutations causing non-syndromic deafness DFNB12 or Ushersyndrome USH1D in compound heterozygotes. J. Med. Genet. 48, 767-775.

Schwander, M., Xiong, W., Tokita, J., Lelli, A., Elledge, H. M., Kazmierczak, P.,Sczaniecka, A., Kolatkar, A., Wiltshire, T., Kuhn, P. et al. (2009). A mouse modelfor nonsyndromic deafness (DFNB12) links hearing loss to defects in tip links ofmechanosensory hair cells. Proc. Natl. Acad. Sci. USA 106, 5252-5257.

Seiler, C., Finger-Baier, K. C., Rinner, O., Makhankov, Y. V., Schwarz, H.,

Neuhauss, S. C. and Nicolson, T. (2005). Duplicated genes with split functions:independent roles of protocadherin15 orthologues in zebrafish hearing and vision.Development 132, 615-623.

Journal of Cell Science 126 (0)10

Journ

alof

Cell

Scie

nce

Senften, M., Schwander, M., Kazmierczak, P., Lillo, C., Shin, J. B., Hasson, T.,

Geleoc, G. S., Gillespie, P. G., Williams, D., Holt, J. R. et al. (2006). Physical and

functional interaction between protocadherin 15 and myosin VIIa in mechanosensory

hair cells. J. Neurosci. 26, 2060-2071.

Shabbir, M. I., Ahmed, Z. M., Khan, S. Y., Riazuddin, S., Waryah, A. M., Khan,

S. N., Camps, R. D., Ghosh, M., Kabra, M., Belyantseva, I. A. et al. (2006).

Mutations of human TMHS cause recessively inherited non-syndromic hearing loss.

J. Med. Genet. 43, 634-640.

Silber, J., Cotton, J., Nam, J. H., Peterson, E. H. and Grant, W. (2004).

Computational models of hair cell bundle mechanics: III. 3-D utricular bundles.

Hear. Res. 197, 112-130.

Sollner, C., Rauch, G. J., Siemens, J., Geisler, R., Schuster, S. C., Muller, U.,

Nicolson, T.; Tubingen 2000 Screen Consortium (2004). Mutations in cadherin 23

affect tip links in zebrafish sensory hair cells. Nature 428, 955-959.

Sotomayor, M., Corey, D. P. and Schulten, K. (2005). In search of the hair-cell gating

spring elastic properties of ankyrin and cadherin repeats. Structure 13, 669-682.

Sotomayor, M., Weihofen, W. A., Gaudet, R. and Corey, D. P. (2010). Structural

determinants of cadherin-23 function in hearing and deafness. Neuron 66, 85-100.

Stauffer, E. A. and Holt, J. R. (2007). Sensory transduction and adaptation in inner and

outer hair cells of the mouse auditory system. J. Neurophysiol. 98, 3360-3369.

Stepanyan, R. and Frolenkov, G. I. (2009). Fast adaptation and Ca2+ sensitivity of the

mechanotransducer require myosin-XVa in inner but not outer cochlear hair cells.

J. Neurosci. 29, 4023-4034.

Steyger, P. S., Gillespie, P. G. and Baird, R. A. (1998). Myosin Ibeta is located at tip

link anchors in vestibular hair bundles. J. Neurosci. 18, 4603-4615.

Takumida, M., Wersall, J. and Bagger-Sjoback, D. (1988). Stereociliary glycocalyx

and interconnections in the guinea pig vestibular organs. Acta Otolaryngol. 106, 130-

139.

Thurm, U. (1981). Mechano-electric transduction. Biophys. Struct. Mech. 7, 245-246.

Tsuprun, V. and Santi, P. (2000). Helical structure of hair cell stereocilia tip links in

the chinchilla cochlea. J. Assoc. Res. Otolaryngol. 1, 224-231.

Tsuprun, V., Goodyear, R. J. and Richardson, G. P. (2004). The structure of tip linksand kinocilial links in avian sensory hair bundles. Biophys. J. 87, 4106-4112.

Verpy, E., Masmoudi, S., Zwaenepoel, I., Leibovici, M., Hutchin, T. P., Del Castillo,I., Nouaille, S., Blanchard, S., Laine, S., Popot, J. L. et al. (2001). Mutations in anew gene encoding a protein of the hair bundle cause non-syndromic deafness at theDFNB16 locus. Nat. Genet. 29, 345-349.

Verpy, E., Leibovici, M., Michalski, N., Goodyear, R. J., Houdon, C., Weil, D.,

Richardson, G. P. and Petit, C. (2011). Stereocilin connects outer hair cellstereocilia to one another and to the tectorial membrane. J. Comp. Neurol. 519, 194-210.

Webb, S. W., Grillet, N., Andrade, L. R., Xiong, W., Swarthout, L., Della Santina,C. C., Kachar, B. and Muller, U. (2011). Regulation of PCDH15 function inmechanosensory hair cells by alternative splicing of the cytoplasmic domain.Development 138, 1607-1617.

Williams, D. S. (2008). Usher syndrome: animal models, retinal function of Usherproteins, and prospects for gene therapy. Vision Res. 48, 433-441.

Xiong, W., Grillet, N., Elledge, H. M., Wagner, T. F., Zhao, B., Johnson, K. R.,

Kazmierczak, P. and Muller, U. (2012). TMHS is an integral component of themechanotransduction machinery of cochlear hair cells. Cell 151, 1283-1295.

Xu, Z., Peng, A. W., Oshima, K. and Heller, S. (2008). MAGI-1, a candidatestereociliary scaffolding protein, associates with the tip-link component cadherin 23.J. Neurosci. 28, 11269-11276.

Zampini, V., Ruttiger, L., Johnson, S. L., Franz, C., Furness, D. N., Waldhaus, J.,Xiong, H., Hackney, C. M., Holley, M. C., Offenhauser, N. et al. (2011). Eps8regulates hair bundle length and functional maturation of mammalian auditory haircells. PLoS Biol. 9, e1001048.

Zhao, Y., Yamoah, E. N. and Gillespie, P. G. (1996). Regeneration of broken tip linksand restoration of mechanical transduction in hair cells. Proc. Natl. Acad. Sci. USA

93, 15469-15474.Zhao, H., Williams, D. E., Shin, J. B., Brugger, B. and Gillespie, P. G. (2012). Large

membrane domains in hair bundles specify spatially constricted radixin activation.J. Neurosci. 32, 4600-4609.

Stereociliary links 11