ABSTRACTUsher syndrome (USH), the leading cause of hereditary combinedhearing and vision loss, is characterized by sensorineural deafnessand progressive retinal degeneration. Mutations in several differentgenes produce USH, but the proximal cause of sensory cell deathremains mysterious. We adapted a proximity ligation assay toanalyze associations among three of the USH proteins, Cdh23,Harmonin and Myo7aa, and the microtubule-based transporter Ift88in zebrafish inner ear mechanosensory hair cells. We found that theproteins are in close enough proximity to form complexes and thatthese complexes preassemble at the endoplasmic reticulum (ER).Defects in any one of the three USH proteins disrupt formation andtrafficking of the complex and result in diminished levels of the otherproteins, generalized trafficking defects and ER stress that triggersapoptosis. ER stress, thus, contributes to sensory hair cell loss andprovides a new target to explore for protective therapies for USH.
INTRODUCTIONUsher syndrome (USH) is a multigenic disease that affects 1 in 6000people in the United States (Kimberling et al., 2010). Individualswith USH present with hearing loss, progressive retinal degenerationand sometimes balance dysfunction. There are three clinical types,USH1, USH2 and USH3, based on the age of onset and severity ofauditory and visual symptoms, with USH1 being the most severe.Individuals with USH1 have profound congenital deafness, balanceproblems and relatively early-onset retinitis pigmentosa.
To date, 11 different genes (13 genetic loci) have been linked toUSH (Bonnet and El-Amraoui, 2012; Jaworek et al., 2012;Puffenberger et al., 2012; Riazuddin et al., 2012). The known USHgenes encode proteins with diverse functions, includingtransmembrane adhesion and signaling, scaffolding, and myosin-motor transport. In vitro binding assays have led to an emergingview that harmonin (USH1C) and whirlin (DFNB31, USH2D) actas scaffold proteins that assemble the Usher proteins into a multi-molecular complex (Adato et al., 2005b; Boëda et al., 2002;Maerker et al., 2008; Reiners et al., 2006; Söllner et al., 2004; vanWijk et al., 2006). We recently showed that PDZD7 serves as agenetic modifier and an additional scaffold for Usher proteins(Ebermann et al., 2010). This multi-molecular-complex model could
RESEARCH ARTICLE
Institute of Neuroscience, University of Oregon, Eugene, OR 97403, USA.
This is an Open Access article distributed under the terms of the Creative CommonsAttribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricteduse, distribution and reproduction in any medium provided that the original work is properlyattributed.
Received 26 August 2013; Accepted 4 March 2014
provide an explanation for how mutations in such a wide variety ofdifferent types of proteins can produce such similar symptoms; if theproteins work together in complexes, then defects in any onecomponent protein might be expected to compromise function of thecomplex as a whole. In mechanosensory hair cells, the USH proteinshave been shown to play structural and developmental roles instereocilia and ribbon synapses (Lefèvre et al., 2008; McGee et al.,2006; Seiler et al., 2005; Söllner et al., 2004; van Wijk et al., 2006;Zallocchi et al., 2012). In the retina, USH proteins are associatedwith the connecting cilium, outer limiting membrane (Gosens et al.,2007) and ribbon synapses of photoreceptors (Kersten et al., 2010;Maerker et al., 2008; Overlack et al., 2008; Phillips et al., 2011).Defects in these proteins are known to disrupt function ofmechanosensory hair cells and photoreceptors; however, theeventual cause of cell death in USH is unknown.
To examine potential interactions among the USH proteins in situ,we adapted a proximity ligation assay (Söderberg et al., 2006) to useon whole-mount zebrafish inner ear mechanosensory hair cells. Weexamined localization and proximity of one of each of the threedifferent types of USH1 proteins: transmembrane (Cdh23), scaffold(Harmonin) and actin-based motor (Myo7a). A defect in any one ofthese proteins produces the severe form of USH (Bonnet and El-Amraoui, 2012), and recent studies have shown that they can forma ternary complex that interacts with membrane phospholipids invitro (Bahloul et al., 2010). We examined the early stages of hair celldevelopment and found that these three USH1 proteins are in closeenough proximity to form a complex. We see that these interactionsoccur in close proximity to COPII proteins, indicating that thecomplex preassembles in the endoplasmic reticulum (ER) before itis trafficked to the Golgi. To characterize the trafficking, weexamined Ift88, an intraflagellar transport protein (Pazour et al.,2002), and found that it, too, is associated with the complex of theseUSH1 and COPII proteins. These close interactions are disrupted bymutations in any one of the genes, the remaining proteins arediminished and mislocalized, and the cells develop generalizedtrafficking defects. Thus, our data demonstrate that the Cdh23,Harmonin, Myo7aa and Ift88 complex is functionally active duringvesicular transport from the ER and acts as a newly identifiedtrafficking regulator in the hair cell. Our results also show thatdefects in the complex produce ER stress that leads to apoptosis.
RESULTSMutations in different USH1 genes and in Ift88 lead tosevere structural mechanoreceptor defectsWe examined the phenotypes resulting from mutations in three ofthe zebrafish USH1 genes, cdh23, ush1c and myo7aa, and in ift88,a gene that encodes an intraflagellar transport protein known tofunction in the development of inner ear mechanosensory hair cells(Jones et al., 2008; Kindt et al., 2012; Tsujikawa and Malicki, 2004).We previously showed that ush1c mutants have hearing and balancedefects (Phillips et al., 2011), and other studies have shown similar
Complexes of Usher proteins preassemble at the endoplasmicreticulum and are required for trafficking and ER homeostasisBernardo Blanco-Sánchez, Aurélie Clément, Javier Fierro, Jr, Philip Washbourne and Monte Westerfield*
Dis
ease
Mod
els
& M
echa
nism
s
548
phenotypes in myo7aa (Ernest et al., 2000) and cdh23 (Söllner et al.,2004) zebrafish mutants. To characterize the cellular basis of thesedefects, we examined the number and the structure of hair cells andstereocilia in the anterior macula of the inner ear. All four mutantsshowed similar significant structural defects in mechanosensory haircells, including bent and/or splayed stereocilia (Fig. 1B-E),consistent with the known requirements of USH1 proteins (Bonnetand El-Amraoui, 2012) and Ift88 (Jones et al., 2008) during hairbundle morphogenesis. Additionally, fewer hair cells (~65%)formed stereocilia in all four mutants (Fig. 1I) compared withcontrols, as previously reported for ush1c mutants (Phillips et al.,2011). We also observed an increase in apoptosis of hair cells incdh23 and ift88 mutants, as well as in ush1c mutants (Fig. 1K), aswe showed previously (Phillips et al., 2011). Despite these defects,the epithelial organization (supplementary material Fig. S1) andnumber of hair cells formed (Fig. 1J) were relatively normal in themutants, indicating that initial specification of hair cell fates doesnot depend upon USH1 or Ift88 function.
Our results demonstrate that these four genes function in thedevelopment of stereocilia. Moreover, the striking similarity of thehair cell phenotypes indicates that these genes might function in the same or overlapping developmental processes, consistent withthe previously proposed model that USH proteins function togetherin a multi-molecular complex (Kremer et al., 2006). This
interpretation is further supported by our observation that all fourproteins had partially overlapping subcellular distributions in haircells (Fig. 1F-H; supplementary material Fig. S2, Fig. S3H-I′).
Because localization of these USH1 proteins in the cell bodies ofhair cells has not been studied previously, we conducted severalcontrol experiments to ensure that the cell body labeling was not dueto non-specific binding (supplementary material Fig. S3H) or cross-reactivity of the avidin/biotinylated enzyme detection complex(supplementary material Fig. S3I′). To substantiate further thecolocalization of Ift88 and the USH proteins, we examined co-labeling of known markers of cellular organelles, Sec23, a markerof the ER-derived vesicles, and GM130, a Golgi marker (Cai et al.,2007; Yu et al., 2006; Zanetti et al., 2011). These markers did notoverlap, as expected (supplementary material Fig. S4). These resultsthus strengthen our interpretation that the three USH1 proteins andIft88 partially overlap in hair cell bodies.
Cdh23 and Harmonin require Ift88 and Myo7aa for normalsubcellular localizationTo learn whether interactions among these proteins are required fortheir proper distributions in hair cells, we used protein-specificantibodies to characterize the subcellular localizations of the threeUSH1 proteins in wild types and in USH1 and ift88 mutants. Inwild-type animals, all four proteins were localized in the cell bodies,particularly in the subapical region (Fig. 2A-D; supplementarymaterial Fig. S3). Ift88 and the USH1 proteins were also observedtogether in the kinocilium, whereas only USH1 proteins were foundin the hair bundle (Fig. 2A; supplementary material Fig. S3), aspreviously observed in several species (Boëda et al., 2002; Grati andKachar, 2011; Jones et al., 2008; Lagziel et al., 2005; Phillips et al.,2011; Siemens et al., 2002; Söllner et al., 2004; Tsujikawa andMalicki, 2004; Verpy et al., 2000).
Mutations in any one of the four genes resulted in a drasticreduction or an almost complete lack of Harmonin or Cdh23immunoreactivity not only in hair bundles, but also in the hair cellbodies (Fig. 2). Ift88 and Myo7aa were less affected by mutationsin cdh23 (Fig. 2G,H) or ush1c (Fig. 2K,L). These resultsdemonstrate that Cdh23, an USH1 transmembrane protein, andHarmonin, the USH1 scaffold protein, require not only the Myo7aaand Ift88 transport proteins, but also each other for proper proteinlevels and localization in hair cells. Lack of one of the proteins couldnegatively impact expression levels or stability of the other proteins.On the other hand, the persistence of Ift88 and Myo7aa proteins incdh23 and ush1c mutants argues that these proteins can be stabilizedand localized subcellularly in hair cells independently of Cdh23 andHarmonin.
Previous studies showed that Cdh23 protein localizes relativelynormally in stereocilia of mice lacking Myo7a (Boëda et al., 2002;Lefèvre et al., 2008; Senften et al., 2006). Our observation ofdrastically reduced Cdh23 levels in the hair cell bodies of myo7aamutants seems to contradict these findings from mice. Alternatively,the difference between these studies could indicate that thehomeostatic rate of protein trafficking, rather than the final distributionof proteins, is disrupted. To test this possibility, we examinedlocalization of Cdh23 protein using ‘direct’ detection with a secondaryantibody directly coupled to a fluorochrome (Alexa Fluor 488), asused in the previous mouse studies. Specificity of the antibody wasassayed in cdh23 mutants with the more sensitive AB Complexsystem (supplementary material Fig. S5A,C). We were unable todetect Cdh23 in the cell bodies of either wild-type or myo7aa mutantanimals with the direct labeling method, although the antibody labeledthe hair bundles (supplementary material Fig. S5B,D). This result is
RESEARCH ARTICLE Disease Models & Mechanisms (2014) doi:10.1242/dmm.014068
TRANSLATIONAL IMPACTClinical issueUsher syndrome is a recessively inherited disease of combined deafnessand blindness that is characterized by the loss of sensory cells. Hearingimpairment in people with Usher syndrome is usually apparent from birth,whereas vision loss is slow and progressive, and begins in the first orsecond decade of life. There is a great deal of clinical variability amongpatients with Usher syndrome and, to date, 11 genes that can cause thisdisease when mutated have been identified. These genes encodeproteins that have a wide variety of functions and that can interactphysically with each other to form macromolecular complexes. A betterunderstanding of where in the cell these complexes assemble and howmutations in their component proteins cause cell death will enhance ourunderstanding of the etiology of Usher syndrome, and could contributeto the development of effective treatments.
ResultsIn this paper, the authors describe the assembly of a complex of threeknown Usher proteins – the scaffold protein Harmonin, thetransmembrane adhesion protein Cadherin23 and the actin-based motorMyosin7aa – and the intraflagellar transport protein Ift88, a newlyidentified interactor, in zebrafish inner ear mechanosensory hair cells.The authors report that mutations in any one of these proteins lead tosimilar structural defects in the mechanosensory cells. Then, using an invivo proximity ligation assay, they demonstrate that the protein complexcontaining these four proteins assembles at the level of the endoplasmicreticulum (ER) in mechanosensory cells. Compromising the function ofany member of the complex not only blocks the formation of thecomplex, but also results in severe trafficking defects, ER stress and celldeath. Finally, they report that cell death is reduced by blocking thefunction of Cdk5, a factor that mediates ER-stress-induced apoptosis.
Implications and future directionsThese findings show that ER stress is the underlying cause of sensorycell loss in Usher syndrome and provide a major advance for the fieldby identifying, for the first time, the proximal cause of sensory cell lossin this disease. By demonstrating that the correct assembly andtrafficking of the Usher protein complex are required for ER homeostasis,the authors show a direct link from Usher mutations to ER stress and celldeath. Thus, this study identifies ER stress as a new target to explorefor the development of protective therapies for Usher syndrome.
Dis
ease
Mod
els
& M
echa
nism
s
consistent with the previous studies in mice and further supports ourconclusion that levels of Cdh23 protein are significantly reduced inhair cell bodies of myo7aa mutants, but Cdh23 can nevertheless reachthe stereocilia, albeit in lesser amounts than in wild-type animals.
Cdh23, Harmonin, Ift88 and Myo7aa proteins are in closeenough proximity to form a complexTo examine whether the three USH1 proteins are close enough toeach other to form a complex in situ, we adapted a proximitylabeling technique (Söderberg et al., 2006) to analyze zebrafishmechanosensory hair cells. Ift88 and the USH1 proteins were
recognized by specific primary antibodies, which in turn wererecognized by oligonucleotide coupled secondary antibodies. Whenthe oligonucleotides are in close proximity [estimated to be 40 nmor less (Söderberg et al., 2006)] they can guide the formation ofcircular DNA strands that, in turn, serve as templates for localizedrolling-circle amplification that incorporates fluorochrome-labelednucleotides. This type of assay has the specificity and sensitivity toreveal the precise subcellular localization of stable and transientprotein-protein interactions in situ and has been used to characterizeprotein interactions in brain slices (Trifilieff et al., 2011), retina(Blasic et al., 2012) and photoreceptors (Wang et al., 2012).
549
RESEARCH ARTICLE Disease Models & Mechanisms (2014) doi:10.1242/dmm.014068
Fig. 1. Zebrafish cdh23, ush1c, myo7aa and ift88 mutants have mechanoreceptor structural defects. (A-E) Confocal projections of anterior maculaelabeled with phalloidin in (A) wild-type (WT) sibling, (B) cdh23tj264a, (C) ush1cfh293, (D) ift88tz288b and (E) myo7aaty220d mutants. (F-H) Confocal optical sections ofdouble immunolabeling of (F) Harmonin (green) and Ift88 (red), (G) Harmonin (green) and Cdh23 (red), and (H) Harmonin (green) and Myo7aa (red). Individualhair cells are shown. (I) Quantification of total number of hair bundles. n≥6 analyzed anterior maculae. (J) Quantification of total number of hair cells. n≥5analyzed anterior maculae. (K) Quantification of hair cell death in mutants and wild-type siblings. Number: average number of TUNEL-positive hair cells permacula. n≥16 analyzed anterior maculae. Student’s t-test: **P<0.01, *P<0.05. NS, non significant. Black lines represent ± s.e.m. Scale bars: 5 μm.
Dis
ease
Mod
els
& M
echa
nism
s
550
Analyzing proteins pairwise, we found that all four proteins arein close proximity in the cell bodies of hair cells (Fig. 3A-F),consistent with them forming a multi-molecular complex aspreviously proposed for the USH1 proteins based on in vitro bindingassays (Adato et al., 2005b; Reiners et al., 2006; Siemens et al.,2002; van Wijk et al., 2006). Proximity labeling was greatly reducedor completely lacking in all four mutants (Fig. 3G,Dd), even forprotein pairs where the affected gene encoded one of the otherproteins. This indicated that formation of the complex was blockedby a defect in any one of its members. The one exception was incdh23 mutants, in which some proximity labeling signal remainedfor Cdh23-Harmonin (Fig. 3G), Cdh23-Ift88 (Fig. 3H) andHarmonin-Ift88 (Fig. 3I) protein pairs, but not for Myo7a(Fig. 3J,L). This result shows that an incomplete complex, lackingMyo7a, formed in this mutant. The residual signal could indicate
that the available mutant allele (cdh23tj264a), which has a D166Vmissense mutation (supplementary material Fig. S6), does not resultin complete loss of Cdh23 protein. Although Cdh23 protein wasnearly undetectable in the mutant by single antibody labeling(Fig. 2E), the proximity labeling technique allowed us to detectformation of some Cdh23-Harmonin (Fig. 3G) and Cdh23-Ift88(Fig. 3H) pairs. This interpretation was further supported byknocking down Cdh23 protein with a translation-blockingmorpholino antisense oligonucleotide, which also drasticallyreduced the proximity labeling signals (supplementary material Fig.S7).
Because this proximity assay has not been used previously tostudy zebrafish hair cells, we ran a number of control experiments.To control for non-specific binding between the various secondaryantibodies, we added only one or no primary antibody to the
RESEARCH ARTICLE Disease Models & Mechanisms (2014) doi:10.1242/dmm.014068
Fig. 2. Cdh23 and Harmonin require Ift88 and Myo7aa for normal localization in hair cells. Immunolabeling in (A-D) wild-type siblings or (E-H) cdh23, (I-L) ush1c, (M-P) ift88 and (Q-T) myo7aa mutants of (A,E,I,M,Q) Cdh23, (B,F,J,N,R) Harmonin, (C,G,K,O,S) Ift88 and (D,H,L,P,T) Myo7aa. Confocal views ofanterior macula hair cells are shown. Confocal sections are 0.8 μm thick, whereas the neck and basal region of the hair cell have an estimated diameter of 2-2.5 μm and 5-6 μm, respectively. Thus, not all confocal sections contain a hair bundle, and the observed shape of the hair bundle varies according to themounting angle (see also supplementary material Fig. S3). Some individual hair cells are outlined by dotted lines. WT, wild type. Scale bar: 5 μm.
Dis
ease
Mod
els
& M
echa
nism
s
reaction and observed no signal (supplementary material Fig. S8 anddata not shown). To control for non-specific binding of the primaryantibodies within the cell bodies, we used the proximity assay toexamine the proximity of Harmonin and acetylated tubulin orgamma tubulin (γ-tubulin). Although acetylated tubulin isdistributed throughout the cytoplasm and kinocilium (supplementarymaterial Fig. S9A-C), the proximity signal with Harmonin wasconfined only to the basal region of the cell body at the level of thesynapses (supplementary material Fig. S9D-F), in agreement withprevious reports of Harmonin localization in this region (Gregory etal., 2011; Gregory et al., 2013). Moreover, the proximity signal wasabsent in ush1c mutants (supplementary material Fig. S9G-I),further confirming its specificity. We obtained similar results withγ-tubulin (supplementary material Fig. S9J-O). Together, theseresults demonstrate that the proximity assay reveals specificinteractions among the USH1 proteins and Ift88 in zebrafish haircells.
The positive proximity labeling signal for Ift88 with Cdh23,Harmonin and Myo7aa (Fig. 3B,C,F) was unexpected. With theexception of Spag5 (Kersten et al., 2012), previous studies have notreported direct binding between USH proteins and components ofthe microtubule transport system. Because the proximity assayindicates that proteins are near each other, but not necessarilybinding directly, we used co-immunoprecipitation to see whetherIft88 can interact with the USH1 proteins. We expressed HA-taggedHarmonin, HA-tagged Ift88 and GFP-tagged Cdh23 in Madin-Darby canine kidney (MDCK) cells. Protein complexes wereimmunoprecipitated using anti-Harmonin or anti-GFP antibodies,and protein extracts were analyzed on western blots. Harmonin was
able to bind to a membrane-targeted intracellular domain of Cdh23(mbn-Cdh23; Fig. 4A, lanes 9 and 12), but neither of these USHproteins immunoprecipitated with Ift88 (Fig. 4A, lanes 3, 6, 12 and15). Consistent with these results, immunolabeling of MDCK cellsshowed extensive overlap of mbn-Cdh23 with Harmonin (Fig. 4B),but little overlap with Ift88 (Fig. 4C). We obtained essentially thesame results when the vectors were transfected into COS7 cells(supplementary material Fig. S10 and data not shown). Interestingly,the intracellular Cdh23 domain, without membrane tethering, didnot co-precipitate with Harmonin (Fig. 4A, lane 15), consistent withthe interpretation that association of Cdh23 with the membrane isrequired for USH complex formation. These results show that Ift88does not co-immunoprecipitate with Cdh23 or Harmonin andsuggest that Ift88 might participate in the complex with the threeUSH1 proteins indirectly.
The Cdh23, Harmonin, Ift88 and Myo7aa protein complexassembles at the ERThe proximity labeling signal was very strong in the perinuclearregion of hair cells (Fig. 3A-F). The ER is a major constituent of theperinuclear region, and the proximity labeling in this regionsuggested that the complex of Ift88 and the three USH1 proteinsforms at the time of their synthesis in the ER. Membrane proteins,including Cdh23, are synthesized in the rough ER and packaged intoCOPII vesicles at the ER exit site (ERES). Thus, we examined theproximity of the USH proteins to Sec23 or Sec13, majorcomponents and markers of COPII vesicles (Zanetti et al., 2011).
Using pairwise labeling, we found that Sec23 is in closeproximity to Cdh23 (Fig. 5A). Harmonin (Fig. 5B) and Myo7aa
551
RESEARCH ARTICLE Disease Models & Mechanisms (2014) doi:10.1242/dmm.014068
Fig. 3. Cdh23, Harmonin, Ift88 and Myo7aa proteins are in close proximity. Proximity labeling (PL) of (A,G,M,S,Y) Cdh23 and Harmonin, (B,H,N,T,Z)Cdh23 and Ift88, (C,I,O,U,Aa) Harmonin and Ift88, (D,J,P,V,Bb) Cdh23 and Myo7a, (E,K,Q,W,Cc) Harmonin and Myo7a, and (F,L,R,X,Dd) Ift88 and Myo7. (A-F) Wild-type (WT) siblings; (G-L) cdh23, (M-R) ush1c, (S-X) ift88 and (Y-Dd) myo7aa mutants. Confocal views of anterior macula hair cells. Some individualhair cells are outlined by dotted lines. Scale bar: 5 μm.
Dis
ease
Mod
els
& M
echa
nism
s
552
(Fig. 5D) were also in close proximity to Sec23, and Ift88 was closeto Sec13 (Fig. 5C). The proximity of these proteins was alsoindicated by standard double labeling with antibodies specific foreach protein (supplementary material Fig. S11A-C). This result wassurprising, because the two USH1 cytoplasmic proteins would notnormally be expected to associate with transport vesicles. Together,these results indicate that the three USH1 proteins form a complexthat is packaged in COPII vesicles. The inclusion of Ift88 in thiscomplex (Fig. 3B,C,F and Fig. 5C) indicates that COPII vesiclescarrying the USH complex might be trafficked via microtubulesfrom the ERES (Gupta et al., 2008).
In cdh23 mutants, association of Sec23 or Sec13 with the othertwo USH1 proteins and Ift88 was lost (Fig. 5F-I). Presumably,misfolding of the mutant Cdh23 protein precludes association of theUSH1 and Ift88 cytoplasmic proteins with COPII vesicles. Also, theCdh23-Sec23 proximity labeling signal was completely absent inmyo7aa mutants (Fig. 5U), indicating that interaction with Myo7aais required for translocation of Cdh23 to the ERES. Thisinterpretation was further supported by the apparent formation of anincomplete complex, lacking Myo7aa, in cdh23 mutants (Fig. 3J-L).Interactions with the Sec proteins were also reduced in ush1cmutants (Fig. 5K-N). In ift88 mutants, by contrast, assembly of thecomplex was completely disrupted (Fig. 3S-X), although Cdh23(Fig. 5P) and Harmonin (Fig. 5Q) could still associate with Sec23,albeit at diminished levels. This shows that, in ift88 mutants,individual Sec23 COPII vesicles contain Cdh23 or Harmonin, butnot both. Together, these observations indicate that Myo7aa isrequired for Cdh23 translocation to the ERES, where Ift88-dependent stabilization and secretion of the USH1 protein complexnormally occurs. The binding of Harmonin to the membrane-bound
form of Cdh23 (Fig. 4A, lanes 9 and 12) but not the cytoplasmicform (Fig. 4A, lane 15) provides additional support that Cdh23needs to be transported through the ERES for assembly of thecomplex.
The loss of Sec23 or Sec13 association with the three USH1proteins or Ift88 in USH1 or ift88 mutants indicated that assemblyof the complex is necessary for targeting the complex to the ERESand subsequent secretion. To test this interpretation further, weexamined formation of the ER-Golgi intermediate complex(ERGIC), as indicated by Trappc3 labeling (Cai et al., 2007; Yu etal., 2006). In wild-type hair cells, some Trappc3 labeling overlappedwith that of each of the three USH1 proteins (supplementarymaterial Fig. S11D-F), indicating that the USH1 proteins trafficthrough the ERGIC. The total amount of Trappc3 labeling wasreduced in all three USH1 mutants and in ift88 mutants(Fig. 5E,J,O,T,Y; supplementary material Fig. S12). This result isconsistent with the trafficking defects seen by proximity labelingwith Sec23 or Sec13 (Fig. 5), and further demonstrates that defectsin assembly or trafficking of the complex produce a global reductionof the ERGIC. Thus, defects in Ift88 or any one of the three USH1proteins disrupt formation of the complex, trafficking, andbiogenesis of the ERGIC (Fig. 6).
The Cdh23, Harmonin, Myo7aa and Ift88 protein complex isrequired for trafficking of USH2 proteins and can trigger ERstress when defectiveGiven the importance of close association of Cdh23, Harmonin andMyo7aa for assembly and intracellular trafficking of this complex of USH1 proteins and the requirement of Myo7a for the subcellular localization of Gpr98, whirlin and usherin in mouse
RESEARCH ARTICLE Disease Models & Mechanisms (2014) doi:10.1242/dmm.014068
Fig. 4. Harmonin and Cdh23 bind directly to each other,but not to Ift88. (A) Co-immunoprecipitation. MDCK cellswere transfected as indicated by (+) with combinations of DNAvectors encoding Harmonin-HA, mbn-Cdh23-GFP, Cdh23-GFPor IFT88-HA. Input lanes: 1,4,7,10,13. Unbound fraction lanes:2,5,8,11,14. Bound fraction lanes: 3,6,9,12,15.Immunoprecipitation was performed with an antibody againstHarmonin (α-Harm) or GFP (α-GFP). (B,C) Confocal sectionsof transfected MDCK cells. (B) Double immunolabeling of mbn-Cdh23-GFP (green) and Harmonin (red). (C) Doubleimmunolabeling of mbn-Cdh23-GFP (green) and Ift88 (red).Harmonin-HA: full-length zebrafish Harmonin isoform A fusedto HA-tag at the C-terminal. mbn-Cdh23-GFP: zebrafishCdh23 membrane bound cytoplasmic domain fused to GFP atthe C-terminal. Cdh23-GFP: zebrafish Cdh23 cytoplasmicdomain fused to GFP at the C-terminal. Ift88-HA: full-lengthzebrafish Ift88 fused to HA-tag at the C-terminal. Scale bar:7.5 μm.
Dis
ease
Mod
els
& M
echa
nism
s
mechanoreceptors (Michalski et al., 2007), we also examined the cellbody localization of the USH2 proteins Ush2a, Gpr98 and Whirlinb.In addition to their localization in hair bundles (supplementarymaterial Fig. S13), consistent with mouse and human (Adato et al.,2005a; Bonnet and El-Amraoui, 2012; Mburu et al., 2003; Michalskiet al., 2007), all three proteins were localized subapically in wild-typemechanosensory hair cells (Fig. 7A-C). In cdh23 (Fig. 7D-F), ift88(Fig. 7J-L) and myo7aa (Fig. 7M-O) mutants, all three USH2proteins were greatly reduced in the cell body. In ush1c mutants(Fig. 7G-I), the USH2 proteins were concentrated in the perinuclearregion instead of subapically (supplementary material Fig. S13G-I,Fig. S14F-H). These results indicate that trafficking of the USH2proteins from the ER was disrupted.
Problems with protein targeting to and secretion from the EREScan lead to ER stress (McGuckin et al., 2010). Because our analysisof USH1 mutants indicated defects in trafficking from the ER, weexamined markers of ER stress, including hspa5 expression(Kozutsumi et al., 1988; Kroeger et al., 2012), and the size of theER as indicated by the area of the cell occupied by proteins thatcontain the signal peptide KDEL (Munro and Pelham, 1987).Expression of hspa5 was significantly increased in ush1c and cdh23mutants (Fig. 7P), consistent with increased ER stress. In myo7aamutants, however, hspa5 expression was relatively unaffected,which might indicate partial compensation by the myo7ab duplicate
gene. Similarly, the amount of ER membrane was expanded inush1c mutants as indicated by increased KDEL labeling (Fig. 7Q;supplementary material Fig. S14E) and increased perinuclearlocalization of USH2 proteins (Fig. 7G-I; supplementary materialFig. S14F-H), which is further indication that the hair cells wereundergoing ER stress in this mutant. Interestingly, although theamount of ER membrane apparently increased in ush1c mutants(Fig. 7Q), the amount of ERGIC decreased (Fig. 5E,O). Thus, theflow of membrane from the ER through the ERGIC might becompromised in USH mutants, and this could contribute to thereduced numbers of stereocilia we observed in mutant hair cells(Fig. 1). ER stress also attenuates translation (McGuckin et al.,2010), perhaps accounting for the decreased levels of USH proteinswe observed (Fig. 2).
Chronic ER stress can lead to apoptosis mediated by the Cdk5-Mekk1-JNK pathway (Kang et al., 2012). Because apoptosis is amajor symptom of Usher syndrome, we examined whether there isa link between ER stress and cell death in USH mutants. We usedTUNEL labeling to assay apoptosis of hair cells. Consistent with ourprevious study (Phillips et al., 2011), we observed increased levelsof cell death in ush1c mutants relative to wild-type siblings(Fig. 7R). Knock-down of Cdk5 activity by morpholino injectionblocked the increased cell death in the mutants, demonstrating thatER stress in these cells leads to apoptosis.
553
RESEARCH ARTICLE Disease Models & Mechanisms (2014) doi:10.1242/dmm.014068
Fig. 5. The Cdh23, Harmonin, Ift88 and Myo7aa protein complex is present at the ER and associated vesicles. Proximity labeling (PL) of (A-D) wild-type(WT) siblings, and (F-I) cdh23, (K-N) ush1c, (P-S) ift88 and (U-X) myo7aa mutants. Proximity labeling of (A,F,K,P,U) Sec23 and Cdh23, (B,G,L,Q,V) Sec23 andHarmonin, (C,H,M,R,W) Sec13 and Ift88, and (D,I,N,S,X) Sec23 and Myo7a (recognizes Myo7aa and possibly Myo7ab; see Materials and Methods).(E,J,O,T,Y) Immunolabeling of Trappc3 in (E) WT, and (J) cdh23, (O) ush1c, (T) ift88 and (Y) myo7aa mutants. Confocal views of anterior macula hair cells.Scale bar: A-D,F,I,K-N,P-S,U-X: 7.5 μm; E,J,O,T,Y: 5 μm.
Dis
ease
Mod
els
& M
echa
nism
s
554
DISCUSSIONOur studies show that Cdh23, Harmonin, Myo7aa and Ift88 are inclose enough proximity to form a multi-molecular protein complex,and this interaction occurs at the level of the ER in zebrafish innerear mechanosensory hair cells. Assembly of this complex is requiredfor trafficking of USH2 proteins and for biogenesis of the ERGIC.Defects in complex formation lead to ER stress and apoptosis.Together, our results demonstrate that normal formation andtrafficking of this complex are required for proper function of thecommon secretory pathway and subsequent development of hairbundles.
The use of an enhancement step during antibody detectionallowed us to study variations in protein levels in the cytoplasm, aswell as hair bundles, in various genetic backgrounds. Thelocalization of the USH proteins at the level of the hair bundle inzebrafish at 5 days post-fertilization (dpf) corresponds to that inmouse during embryonic stages, just before the spatial restriction ofthe USH proteins along the stereocilia occurs (Adato et al., 2005a;Boëda et al., 2002; Lagziel et al., 2005). In early postnatal stages,whirlin has been localized near both the tip and ankle links (Grati etal., 2012), and also along the stereocilia (Mburu et al., 2003). Gpr98has been studied primarily after birth, so its distribution during earlydevelopment is less well known in mice.
Assembly of the Cdh23, Harmonin, Myo7a and Ift88 complexcould couple membrane targeting to the actin polymerization andstabilization machinery required for stereocilia growth. The initiallocalization of Cdh23, Harmonin and Myo7a at the kinocilium(Boëda et al., 2002; Lagziel et al., 2005; Lefèvre et al., 2008)indicates that USH1 proteins engage in ciliary trafficking. Ift88might facilitate the movement of proteins like Cdh23, Harmonin andMyo7a across the ciliary barrier (Nachury et al., 2010) towards theapical region during formation of stereocilia (Tanimoto et al., 2011;Tilney et al., 1992). Consistent with this idea, Cdh23 has beenreported at the level of the centrosome (Lagziel et al., 2005), astructure that could be used as a cargo loading point for vesicularciliary trafficking.
Our results show that, although assembly of the USH1 proteincomplex and subsequent trafficking are seriously disrupted in USH1mutants, some USH1 protein eventually localizes properly in thedeveloping stereocilia (supplementary material Fig. S5D). Thus,failure of USH1 protein complex assembly does not cause ‘all ornone’ effects, but rather differentially alters protein distributions.Similarly, mutant Myo7a, Ush1c and Cdh23 mice form fewermechanoreceptor lateral links (Lefèvre et al., 2008), consistent withdecreased trafficking of these components. It is still unclear,however, precisely how failure of assembly of this USH1 complexaffects protein trafficking routes in the hair cell.
At least three hypotheses could explain the underlying traffickingmechanism affected by disruption of the USH1 complex. First, thecomplex could function as a cargo adaptor. Although SEC23A, thecore element of the secretory pathway, is ubiquitously expressed,human patients and zebrafish with SEC23A mutations have tissue-and cell-specific phenotypes with no reported mechanosensorydysfunction (Fromme et al., 2007; Lang et al., 2006) (and data notshown). This suggests that, in SEC23A-deficient hair cells, USHproteins are trafficked correctly and that the trafficking phenotypewe have observed is specific to USH mutants. Previous authors haveproposed that the trafficking efficiency of the secretory pathway ismodulated in a cell-type- and tissue-specific manner by specificinteractions between cargo and core elements of the secretorypathway (Fromme et al., 2007). Our discovery of the requiredfunction of USH proteins in trafficking thus suggests that theassembled protein complex of Cdh23, Harmonin, Myo7aa and Ift88,rather than the individual components, functions as a new cargoadaptor that sorts transmembrane proteins among membranetransport carriers.
Second, the Cdh23, Harmonin, Myo7aa and Ift88 complex mightfunction as a cargo effector required for vesicle transport. It has beenpreviously shown that the vesicular recruitment and motor activityof Myo5a, an atypical myosin (like Myo7aa), depends on thefunction of distinct Rab GTPases (Wu et al., 2001; Wu et al., 2002).We found that the USH complex can interact with Sec23, and that
RESEARCH ARTICLE Disease Models & Mechanisms (2014) doi:10.1242/dmm.014068
Fig. 6. The Cdh23, Harmonin, Ift88 andMyo7aa protein complex preassembles at theER. Model for assembly and trafficking of USHproteins. Under normal conditions in wild-type(WT) animals, Cadherin23, Harmonin, Ift88 andMyo7aa preassemble into a complex at the levelof the ER. Interaction between Cdh23 andMyo7aa is required for Cdh23 translocation tothe ERES. All four proteins are required forassembly of a functional complex. Onceassembled, the complex translocates to ER exitsites (ERES), engages components of the COPIIsystem (Sec23 and Sec13), and enters thesecretory pathway where it is trafficked to theER-Golgi intermediate complex (ERGIC) enroute to final destinations. Loss-of-functionmutations in cdh23, ush1c, ift88 and myo7aadisrupt assembly of the complex in various ways.RER, rough endoplasmic reticulum.
Dis
ease
Mod
els
& M
echa
nism
s
disruption of complex assembly leads to defects in the biogenesis ofthe ERGIC, as assayed by Trappc3 labeling. Previous studiesshowed that Trappc3 interacts physically with Sec23, and that thisinteraction is required for homotypic fusion of COPII vesicles (Caiet al., 2007; Yu et al., 2006). Trappc3 is an essential subunit of theTRAPP I complex, which acts as a guanine exchange factor forRab1 (Cai et al., 2007; Jones et al., 2000; Kim et al., 2006; Wang etal., 2000). Thus, in hair cells, the Cdh23, Harmonin, Myo7aa andIft88 complex might play a structural role during biogenesis of the
ERGIC by providing a physical platform for the required interactionbetween Trappc3 and Sec23 during homotypic fusion of COPIIvesicles.
A third hypothesis is that the observed trafficking defects aresecondary consequences of ER stress and induction of the unfolded-protein response (UPR) pathway. The combination of both responsescould induce trafficking failure. Structural integrity of the ERGIChas been used as a read-out of functional trafficking through thesecretory pathway. Disrupted cycling of the cargo receptors Surf4,
555
RESEARCH ARTICLE Disease Models & Mechanisms (2014) doi:10.1242/dmm.014068
Fig. 7. The Cdh23, Harmonin andMyo7aa protein complex is requiredfor proper trafficking of USH2proteins and can trigger ER stresswhen defective. (A-O) Immunolabelingof USH2 proteins in (A-C) wild-type(WT) siblings, and (D-F) cdh23, (G-I)ush1c, (J-L) ift88 and (M-O) myo7aamutants with antibodies against Ush2a(A,D,G,J,M), Gpr98 (B,E,H,K,N) andWhirlinb (C,F,I,L,O). (P) Levels of hspa5expression based on in situhybridization technique in whole-mountlarvae. For ush1c siblings n=57;ush1c−/− n=51; cdh23 siblings n=35; forcdh23−/− n=33; myo7aa siblings n=24;for myo7aa−/− n=24 analyzed anteriormaculae. (Q) Quantification ofpercentage area per hair cell containingKDEL expression. For ush1c siblingsn=85; ush1c−/− n=125; cdh23 siblingsn=66; for cdh23−/− n=60; myo7aasiblings n=60; myo7aa−/− n=54 analyzedhair cells for each genotype. (R) Knock-down of cdk5 rescues ER-stress-induced cell death in the ush1c mutant.Bars show average number of TUNEL-positive cells. For ush1c siblings n=52;ush1c siblings + 4 ng cdk5-MO n=18;ush1c−/− n=70; ush1c−/− + 4 ng cdk5-MOn=70 analyzed anterior maculae.Average ± s.d. Statistics wereconducted with Student’s t-test.**P<0.01. NS, non significant. Someindividual hair cells are outlined bydotted lines. MO, morpholino. Scale bar:5 μm.
Dis
ease
Mod
els
& M
echa
nism
s
556
ERGIC53, p25 and Syntaxin17 affects biogenesis of the ERGIC(Mitrovic et al., 2008; Muppirala et al., 2011) owing todisequilibrium of membrane flow among secretory-pathwaycompartments. Contrary to this hypothesis, our observation thatmyo7aa and ift88 mutants do not develop elevated hspa5 levels orER distension, at least by 5 dpf (Fig. 7), indicates that traffickingdefects do not necessarily cause ER stress. Thus, further study isrequired to provide an understanding of the underlying mechanismsthat will ultimately explain how disruption of the USH1 proteincomplex affects membrane flow from the ER.
The rough ER is a major site of secreted and membrane-integrated protein synthesis, including, presumably, the USHtransmembrane proteins. Post-translational modification and foldingof proteins occurs in the ER. Protein folding is regulated by weak,primarily hydrophobic, interactions and, thus, a proportion of allproteins normally misfold. Because of the likelihood of misfolding,the ER has mechanisms that are part of routine housekeeping torecognize misfolded proteins and remove them for degradation(Vembar and Brodsky, 2008). A range of molecules within the ER,including chaperones, are essential for appropriate biosynthesis andcorrect folding. Chaperones disengage from proteins once thecorrect conformation is achieved, and, conversely, chaperonesaccumulate when proteins misfold. If protein misfolding isexcessive, ER homeostasis is not maintained, leading to ER stress.Chaperone production is also upregulated in response to increasedmisfolding, and measurement of increased chaperone production,particularly HSPA5 (also called BIP or GRP78), is commonly usedas a marker of ER stress (Kozutsumi et al., 1988; Kroeger et al.,2012). ER stress is often transient and well controlled; however,prolonged ER stress can seriously affect protein production andother cellular functions, ultimately triggering apoptosis (Doyle et al.,2011).
Our observations of defects in protein trafficking, proteinretention in the ER, upregulation of hspa5 chaperone transcriptionand increased size of the ER in USH1 mutants indicate that thesemutations lead to ER stress. Our studies also show that some USHproteins preassemble into a complex at the ER and that defects inone or more components of the complex can activate the UPR.Perhaps when one member of the complex is defective or missing,surfaces on other members of the complex are abnormally exposedand might be recognized as ‘misfolded’, thus triggering the UPR.With prolonged absence or misfolding of one component of theUSH complex, the UPR system could be overwhelmed, leading toER stress. ER stress also activates the Cdk5-Mekk1-JNK cell-deathpathway (Kang et al., 2012). Thus, in USH, a long-termconsequence of USH gene mutations is ER stress that ultimatelyleads to apoptosis.
Recent studies of neurodegenerative diseases (Doyle et al., 2011)such as Alzheimer’s, Parkinson’s and some retinopathies (Kang etal., 2012; Kroeger et al., 2012; Kunte et al., 2012; Shinde et al.,2012) have implicated ER stress as a probable cause of apoptosis.Our current and previous studies (Ebermann et al., 2010; Phillips etal., 2011) add to this view by demonstrating both ER stress andapoptosis in zebrafish USH gene mutants, thus providing the firstevidence that ER stress is a proximal cause of sensory cell loss inUSH. The link between ER stress and apoptosis raises the possibilitythat therapeutics that are being developed for the treatment of otherneurodegenerative diseases (Doyle et al., 2011) will be helpful inmanaging the progression of symptoms in USH patients. Althoughhearing defects are typically congenital in USH1 owing to defectsin the mechanoreceptors, hair cells ultimately die in USH1 mutants.In all forms of USH, vision loss is progressive and photoreceptors
degenerate over decades. Treatments that delay or reduce cell losswill provide time to patients, while therapies that address the defectscan be developed and applied to the remaining, non-degeneratingcell populations.
MATERIALS AND METHODSAnimalsZebrafish strains were AB wild-type, cdh23tj264a (Söllner et al., 2004),ush1cfh293 (Phillips et al., 2011), ift88tz288b (Tsujikawa and Malicki, 2004)and myo7aaty220d (Ernest et al., 2000). All mutations are recessive alleles;we refer to animals with homozygous mutant alleles as mutants.Phenotypically wild-type siblings were used as controls. Animals wereraised in a 10-hour dark and 14-hour light cycle and maintained aspreviously described (Westerfield, 2007). Larvae were staged according tothe standard series (Kimmel et al., 1995). Results presented were obtainedat 5 dpf. All animal-use protocols were IACUC approved.
ConstructsTo construct Ush1c-HA and Ift88-HA, the coding sequence of the HA-tagwas fused to the 3′ ends of ush1c (Phillips et al., 2011) and ift88 (clonedusing the following PCR primers, forward: 5′-ATGGAGAATGTGCATC -TTGTC-3′; and reverse: 5′-CTCTGATCAAATATAGTTCAGCATATT-3′),respectively. For mbn-Cdh23-GFP, the sequence encoding the signal peptideand membrane-targeting domain of syncam2 (Doyle et al., 2011) was fusedto the 5′ end of the sequence that encodes the intracellular domain of cdh23,which includes exon 68. The GFP coding sequence was then fused to the 3′end of the construct. For Cdh23-GFP, only the GFP sequence was fused tothe cdh23 3′ end. The hspa5 clone was obtained from the ZIRC EST bank(cb865) and then used as a template to make a full-length antisense RNAprobe for in situ hybridization. All gene and protein names are approvedZFIN (http://zfin.org) nomenclature.
Whole-mount immunohistochemistry and proximity ligationassayLarvae were anesthetized and fixed at 5 dpf in BT fix (4% PFA, 4% sucrose,0.15 mM CaCl2 in PBS) for 48 hours for immunolabeling or 24 hours forthe proximity ligation assay, washed in PBST (PBS + 0.01% Tween20), andprocessed for immunolabeling. Samples were permeabilized in 5% Tween20in PBS for 9 hours, fixed for 20 minutes in BT-fix, then blocked for at least2 hours in PBS blocking solution (PBS + 0.01% Tween20, 10% NGS, 10%BSA), followed by an overnight incubation at 4°C in primary antibodysolution (PBS blocking + primary antibody). For double immunolabeling,both primary antibodies were added simultaneously. Unbound antibodieswere removed by five serial incubations of 30 minutes in PBST. Sampleswere fixed for 20 minutes in BT-fix, re-incubated in PBST for short periods,blocked, and incubated overnight at 4°C in secondary antibody solution(PBS blocking + secondary antibody). Unbound secondary antibodies wereremoved as described above.
To detect signals, reagents A and B (AB complex) from the VectastainABC Elite kit (Vector Laboratories) were diluted at 1:100 each in blockingsolution (AB solution) and pre-incubated for 20 minutes. Samples were thenincubated for 25 minutes in AB solution, followed by at least fourincubations of 5 minutes in PBST. Tyramide from the TSA kit (Perkin-Elmer) was diluted 1:50 in pre-warmed buffer reagent and added to samplesfor 20 minutes following the manufacturer’s recommendations. Thedetection reaction was stopped by adding cold TNT (0.1 M Tris-HCl, pH7.5, 0.15 M NaCl, 0.1% Tween20) to the samples, followed by several TNTwashes. For single labeling, the last TNT wash was replaced by Vectashieldsolution (Vector Labs). For double labeling, samples were fixed in TNT-fix(1% PFA, 0.15 mM CaCl2, 4% sucrose in TNT), washed in TNT, and thenincubated in pre-warmed buffer reagent to quench HRP. Samples were thenincubated sequentially for 1 hour with reagent A followed by reagent B ofthe avidin/biotin blocking kit (Invitrogen). After blocking with TNT Block(TNT + 10% NGS, 10% BSA) for 2 hours, samples were incubatedovernight at 4°C in TNT secondary antibody solution (secondary antibodydiluted 1:250 in TNT Block). Signal was detected as described above. Afterdetection, samples were immersed in Vectashield reagent (Vector Labs).
RESEARCH ARTICLE Disease Models & Mechanisms (2014) doi:10.1242/dmm.014068
Dis
ease
Mod
els
& M
echa
nism
s
For the proximity ligation assay (Duo-Link), larvae were prepared asabove and treated according to the manufacturer’s recommendations (OlinkBioscience) with some minor modifications. Primary and PLA probe-coupled antibodies were diluted in PBS blocking solution, and samples wereincubated overnight at 4°C. Unbound antibodies were removed through fiveserial PBST rinses of 30 minutes each. Samples were incubated in PBST at37°C for 1 hour. They were then incubated at 37°C for 4 hours in Duo-Linkligation solution, followed by overnight incubation at 37°C with Duo-Linkamplification solution. Samples were washed in PBST and then immersedin Vectashield for imaging.
Primary antibodies used were rabbit anti-Trappc3 (Sigma); goat, rabbitand chicken anti-Cdh23 (Santa Cruz, Sigma, and custom made against C-terminal region, respectively); rabbit anti-Cip98a (Whirlinb; SDIX); rabbitanti-GFP (Torrey Pines Biolabs); mouse anti-Gm130 (BD TransductionLaboratories); rabbit anti-Gpr98 (SDIX); goat, rabbit and guinea pig anti-Harmonin (SDIX and custom made against the C-terminal region,respectively); rabbit anti-Ift88 (Krock and Perkins, 2008); mouse anti-KDEL(Calbiochem); guinea pig anti-Myo7aa (custom made against C-terminalregion); rabbit anti-Sec23 (Santa Cruz); goat anti-Sec13 (Santa Cruz); mouseanti-acetylated tubulin, used at 1:100 (Santa Cruz); mouse anti-γ-tubulin,used at 1:100 (Sigma); and rabbit anti-Ush2a (SDIX). We also used mouseanti-Myo7a (Sigma), which we showed cross-reacts with zebrafish Myo7aa(supplementary material Fig. S15). Given the sequence similarity betweenzebrafish Myo7aa and Myo7ab, it is possible that the antibody alsorecognizes Myo7ab. Secondary antibodies used were biotinylated anti-mouse, biotinylated anti-goat, biotinylated anti-rabbit, biotinylated anti-chicken and biotinylated anti-guinea pig (Vector Labs).
For whole-mount antibody labeling with the AB complex, antibodies werediluted at 1:500; secondary antibodies were used at a dilution of 1:250. Fordirect detection, primary anti-Cdh23 antibody was diluted at 1:250;secondary antibody coupled to a fluorochrome (Molecular Probes) wasdiluted at 1:250. For antibody dilutions used in cell culture immunolabeling,see ‘Cell culture and immunolabeling’ section below. For the proximityligation assay, primary antibodies were co-incubated at 1:500, except for theanti-acetylated tubulin and γ-tubulin that were used at 1:100. The proximityligation assay probe-coupled antibodies used were anti-goat plus, anti-goatminus, anti-rabbit plus, anti-rabbit minus, anti-mouse plus and anti-mouseminus. All proximity ligation assay probe-coupled antibodies were dilutedfollowing the manufacturer’s instructions (1:5). For immunolabeling of F-actin, FITC-phalloidin was used as described previously (Haddon andLewis, 1996).
TUNEL labelingApoptag plus Peroxidase In Situ Apoptosis Detection Kit was used toexamine cell death following the manufacturer’s instructions with minormodifications. Larvae were fixed and permeabilized as described (Thisseand Thisse, 2008). After proteinase K treatment, larvae were incubated inacetone and ethanol solution (1:2 mixture) for 8 minutes at −20°C,followed by PBS washes. Larvae were equilibrated for 1 hour at roomtemperature in equilibration buffer, then incubated in reaction solution(16 μl of TdT enzyme + 38 μl of reaction buffer) for 90 minutes at 37°C.The reaction was stopped by adding 200 μl of stop buffer. The reactionproduct was detected with an anti-DIG coupled to HRP with DAB assubstrate.
Cell culture and immunolabelingMDCK cells or COS7 cells were cultured in Dulbecco’s modified Eagle’smedium supplemented with 10% fetal bovine serum and penicillin (40U/ml)/streptomycin (40 μg/ml) at 37°C under 5% CO2. Cells were transientlytransfected with 1 μg of plasmids of interest using Lipofectamine 2000.
For cell-culture labeling, cells were cultured on poly-L-lysine-coated glasscoverslips and transfected. 24 hours after transfection, cells were fixed with4% PFA in 4% sucrose-PBS for 10 minutes and then permeabilized with0.25% Triton X-100 in PBS. Detection was performed using mouse anti-GFP (Abcam, 1:500), rabbit anti-Harmonin (SDIX, 1:100) and rabbit anti-Ift88 (Krock and Perkins, 2008) antibodies. Secondary antibodies were goatanti-mouse Alexa-Fluor-488-conjugated and goat anti-rabbit Alexa-Fluor-568-conjugated (Molecular Probes, 1:250).
Co-immunoprecipitation and western blottingAt 16 hours after transfection (see above), cells were lysed in 125 μl lysisbuffer (150 mM NaCl, 50 mM Tris-HCl, 0.5 mM EDTA, 0.1% Triton, andprotease inhibitor tablets) for 20 minutes at 4°C with agitation. The mixturewas centrifuged at 5000 rpm for 5 minutes, and the supernatant wascollected. 40 μl of the supernatant was kept for input, and the rest wasincubated overnight at 4°C with either rabbit anti-GFP (Torrey PinesBiolabs) or rabbit anti-Harmonin (SDIX) antibodies at a concentration of 1ng/μl. Proteins were precipitated with protein A-sepharose beads, washedthree times in lysis buffer, and eluted with sample buffer. The eluted proteinswere boiled, separated by SDS-polyacrylamide gel electrophoresis, andtransferred to PVDF membranes. Detection was performed using rabbit anti-GFP (1:500, Torrey Pines Biolabs) or mouse anti-HA (1:500, Covance)antibodies. Secondary antibodies were donkey anti-rabbit and donkey anti-mouse coupled to HRP (1:5000, Jackson ImmunoResearch Laboratories).
Morpholino injectionWe used a Narishige microinjector to inject one-cell-stage zebrafish embryoswith 2 ng of translation blocker cdh23_morpholino (Söllner et al., 2004) or4 ng of the translation blocker cdk5_morpholino (5′-CCAGCTTCTCAT -ACTTTTGCATGGT-3′) (Easley-Neal et al., 2013). Efficiency of thecdk5_morpholino was evaluated as previously described (Tanaka et al.,2012). Injection volumes were estimated using a micrometer.
Fluorescence in situ hybridizationWhole-mount in situ hybridization was carried out as previously described(Thisse and Thisse, 2008) with the following modifications: Digoxigenin-labeled probes were prepared according to the manufacturer’s instructions(Roche). Probe signal was detected using an anti-DIG POD conjugatedantibody and TSA (Perkin-Elmer).
ImagingVectashield immersed larvae were mounted laterally in a slide chamber andimaged with Zeiss LSM5 or Bio-Rad Radiance 2100 confocal microscopesusing a 63× objective. For whole-mount and cell-culture labeling, opticalsections were 0.8 μm thick. Optical sections for proximity ligationexperiments were 2.5 μm thick. Images were assembled using AdobePhotoshop and Illustrator. ImageJ (v 1.43u) was used to measure hspa5 pixelintensity and to calculate the size of KDEL-positive cell areas.
AcknowledgementsWe wish to thank Chris Doe, Jennifer Phillips, Tom Stevens and Sabrina Toro fortheir comments on the manuscript, Judy Peirce and Jeremy Wegner for technicalassistance, ZIRC and Teresa Nicholson for providing animals, and Brian Perkinsfor sharing reagents.
Competing interestsThe authors declare no competing financial interests.
Author contributionsB.B.-S., A.C. and J.F. conducted experiments; all contributed to design and writing.
FundingSupported by the National Institutes of Health (NIH) DC004186, DC010447,OD011195 and HD22486.
Supplementary materialSupplementary material available online athttp://dmm.biologists.org/lookup/suppl/doi:10.1242/dmm.014068/-/DC1
D., El-Amraoui, A. and Petit, C. (2005a). Usherin, the defective protein in Ushersyndrome type IIA, is likely to be a component of interstereocilia ankle links in theinner ear sensory cells. Hum. Mol. Genet. 14, 3921-3932.
Adato, A., Michel, V., Kikkawa, Y., Reiners, J., Alagramam, K. N., Weil, D.,Yonekawa, H., Wolfrum, U., El-Amraoui, A. and Petit, C. (2005b). Interactions inthe network of Usher syndrome type 1 proteins. Hum. Mol. Genet. 14, 347-356.
Bahloul, A., Michel, V., Hardelin, J. P., Nouaille, S., Hoos, S., Houdusse, A.,England, P. and Petit, C. (2010). Cadherin-23, myosin VIIa and harmonin, encodedby Usher syndrome type I genes, form a ternary complex and interact withmembrane phospholipids. Hum. Mol. Genet. 19, 3557-3565.
557
RESEARCH ARTICLE Disease Models & Mechanisms (2014) doi:10.1242/dmm.014068
Dis
ease
Mod
els
& M
echa
nism
s
558
Blasic, J. R., Jr, Brown, R. L. and Robinson, P. R. (2012). Phosphorylation of mousemelanopsin by protein kinase A. PLoS ONE 7, e45387.
Boëda, B., El-Amraoui, A., Bahloul, A., Goodyear, R., Daviet, L., Blanchard, S.,Perfettini, I., Fath, K. R., Shorte, S., Reiners, J. et al. (2002). Myosin VIIa,harmonin and cadherin 23, three Usher I gene products that cooperate to shape thesensory hair cell bundle. EMBO J. 21, 6689-6699.
Bonnet, C. and El-Amraoui, A. (2012). Usher syndrome (sensorineural deafness andretinitis pigmentosa): pathogenesis, molecular diagnosis and therapeuticapproaches. Curr. Opin. Neurol. 25, 42-49.
Cai, H., Yu, S., Menon, S., Cai, Y., Lazarova, D., Fu, C., Reinisch, K., Hay, J. C. andFerro-Novick, S. (2007). TRAPPI tethers COPII vesicles by binding the coat subunitSec23. Nature 445, 941-944.
Doyle, K. M., Kennedy, D., Gorman, A. M., Gupta, S., Healy, S. J. and Samali, A.(2011). Unfolded proteins and endoplasmic reticulum stress in neurodegenerativedisorders. J. Cell. Mol. Med. 15, 2025-2039.
Easley-Neal, C., Fierro, J., Jr., Buchanan, J. and Washbourne, P. (2013). Laterecruitment of synapsin to nascent synapses is regulated by cdk5. Cell Rep. 3, 1199-1212.
Ebermann, I., Phillips, J. B., Liebau, M. C., Koenekoop, R. K., Schermer, B.,Lopez, I., Schäfer, E., Roux, A. F., Dafinger, C., Bernd, A. et al. (2010). PDZD7 isa modifier of retinal disease and a contributor to digenic Usher syndrome. J. Clin.Invest. 120, 1812-1823.
Ernest, S., Rauch, G. J., Haffter, P., Geisler, R., Petit, C. and Nicolson, T. (2000).Mariner is defective in myosin VIIA: a zebrafish model for human hereditarydeafness. Hum. Mol. Genet. 9, 2189-2196.
Fromme, J. C., Ravazzola, M., Hamamoto, S., Al-Balwi, M., Eyaid, W., Boyadjiev,S. A., Cosson, P., Schekman, R. and Orci, L. (2007). The genetic basis of acraniofacial disease provides insight into COPII coat assembly. Dev. Cell 13, 623-634.
Gosens, I., van Wijk, E., Kersten, F. F., Krieger, E., van der Zwaag, B., Märker, T.,Letteboer, S. J., Dusseljee, S., Peters, T., Spierenburg, H. A. et al. (2007). MPP1links the Usher protein network and the Crumbs protein complex in the retina. Hum.Mol. Genet. 16, 1993-2003.
Grati, M. and Kachar, B. (2011). Myosin VIIa and sans localization at stereocilia uppertip-link density implicates these Usher syndrome proteins in mechanotransduction.Proc. Natl. Acad. Sci. USA 108, 11476-11481.
Grati, M., Shin, J. B., Weston, M. D., Green, J., Bhat, M. A., Gillespie, P. G. andKachar, B. (2012). Localization of PDZD7 to the stereocilia ankle-link associates thisscaffolding protein with the Usher syndrome protein network. J. Neurosci. 32, 14288-14293.
Gregory, F. D., Bryan, K. E., Pangršič, T., Calin-Jageman, I. E., Moser, T. and Lee,A. (2011). Harmonin inhibits presynaptic Cav1.3 Ca2+ channels in mouse inner haircells. Nat. Neurosci. 14, 1109-1111.
Gregory, F. D., Pangrsic, T., Calin-Jageman, I. E., Moser, T. and Lee, A. (2013).Harmonin enhances voltage-dependent facilitation of Cav1.3 channels andsynchronous exocytosis in mouse inner hair cells. J. Physiol. 591, 3253-3269.
Gupta, V., Palmer, K. J., Spence, P., Hudson, A. and Stephens, D. J. (2008).Kinesin-1 (uKHC/KIF5B) is required for bidirectional motility of ER exit sites andefficient ER-to-Golgi transport. Traffic 9, 1850-1866.
Haddon, C. and Lewis, J. (1996). Early ear development in the embryo of thezebrafish, Danio rerio. J. Comp. Neurol. 365, 113-128.
Jaworek, T. J., Bhatti, R., Latief, N., Khan, S. N., Riazuddin, S. and Ahmed, Z. M.(2012). USH1K, a novel locus for type I Usher syndrome, maps to chromosome10p11.21-q21.1. J. Hum. Genet. 57, 633-637.
Jones, S., Newman, C., Liu, F. and Segev, N. (2000). The TRAPP complex is anucleotide exchanger for Ypt1 and Ypt31/32. Mol. Biol. Cell 11, 4403-4411.
Jones, C., Roper, V. C., Foucher, I., Qian, D., Banizs, B., Petit, C., Yoder, B. K. andChen, P. (2008). Ciliary proteins link basal body polarization to planar cell polarityregulation. Nat. Genet. 40, 69-77.
Kang, M. J., Chung, J. and Ryoo, H. D. (2012). CDK5 and MEKK1 mediate pro-apoptotic signalling following endoplasmic reticulum stress in an autosomaldominant retinitis pigmentosa model. Nat. Cell Biol. 14, 409-415.
Kersten, F. F., van Wijk, E., van Reeuwijk, J., van der Zwaag, B., Märker, T., Peters,T. A., Katsanis, N., Wolfrum, U., Keunen, J. E., Roepman, R. et al. (2010).Association of whirlin with Cav1.3 (alpha1D) channels in photoreceptors, defining anovel member of the usher protein network. Invest. Ophthalmol. Vis. Sci. 51, 2338-2346.
Kersten, F. F., van Wijk, E., Hetterschijt, L., Baubeta, K., Peters, T. A., Aslanyan,M. G., van der Zwaag, B., Wolfrum, U., Keunen, J. E., Roepman, R. et al. (2012).The mitotic spindle protein SPAG5/Astrin connects to the Usher protein networkpostmitotically. Cilia 1, 2.
Kim, Y. G., Raunser, S., Munger, C., Wagner, J., Song, Y. L., Cygler, M., Walz, T.,Oh, B. H. and Sacher, M. (2006). The architecture of the multisubunit TRAPP Icomplex suggests a model for vesicle tethering. Cell 127, 817-830.
Kimberling, W. J., Hildebrand, M. S., Shearer, A. E., Jensen, M. L., Halder, J. A.,Trzupek, K., Cohn, E. S., Weleber, R. G., Stone, E. M. and Smith, R. J. (2010).Frequency of Usher syndrome in two pediatric populations: Implications for geneticscreening of deaf and hard of hearing children. Genetics in Medicine 12, 512-516.
Kimmel, C. B., Ballard, W. W., Kimmel, S. R., Ullmann, B. and Schilling, T. F.(1995). Stages of embryonic development of the zebrafish. Dev. Dyn.203, 253-310.
Kindt, K. S., Finch, G. and Nicolson, T. (2012). Kinocilia mediate mechanosensitivityin developing zebrafish hair cells. Dev. Cell 23, 329-341.
Kozutsumi, Y., Segal, M., Normington, K., Gething, M. J. and Sambrook, J. (1988).The presence of malfolded proteins in the endoplasmic reticulum signals theinduction of glucose-regulated proteins. Nature 332, 462-464.
Kremer, H., van Wijk, E., Märker, T., Wolfrum, U. and Roepman, R. (2006). Ushersyndrome: molecular links of pathogenesis, proteins and pathways. Hum. Mol.Genet. 15, R262-R270.
Krock, B. L. and Perkins, B. D. (2008). The intraflagellar transport protein IFT57 isrequired for cilia maintenance and regulates IFT-particle-kinesin-II dissociation invertebrate photoreceptors. J. Cell Sci. 121, 1907-1915.
Kroeger, H., Messah, C., Ahern, K., Gee, J., Joseph, V., Matthes, M. T., Yasumura,D., Gorbatyuk, M. S., Chiang, W. C., LaVail, M. M. et al. (2012). Induction ofendoplasmic reticulum stress genes, BiP and chop, in genetic and environmentalmodels of retinal degeneration. Invest. Ophthalmol. Vis. Sci. 53, 7590-7599.
Kunte, M. M., Choudhury, S., Manheim, J. F., Shinde, V. M., Miura, M., Chiodo, V.A., Hauswirth, W. W., Gorbatyuk, O. S. and Gorbatyuk, M. S. (2012). ER stress isinvolved in T17M rhodopsin-induced retinal degeneration. Invest. Ophthalmol. Vis.Sci. 53, 3792-3800.
Lagziel, A., Ahmed, Z. M., Schultz, J. M., Morell, R. J., Belyantseva, I. A. andFriedman, T. B. (2005). Spatiotemporal pattern and isoforms of cadherin 23 in wildtype and waltzer mice during inner ear hair cell development. Dev. Biol. 280, 295-306.
Lang, M. R., Lapierre, L. A., Frotscher, M., Goldenring, J. R. and Knapik, E. W.(2006). Secretory COPII coat component Sec23a is essential for craniofacialchondrocyte maturation. Nat. Genet. 38, 1198-1203.
Lefèvre, 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.
Maerker, T., van Wijk, E., Overlack, N., Kersten, F. F., McGee, J., Goldmann, T.,Sehn, E., Roepman, R., Walsh, E. J., Kremer, H. et al. (2008). A novel Usherprotein network at the periciliary reloading point between molecular transportmachineries in vertebrate photoreceptor cells. Hum. Mol. Genet. 17, 71-86.
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 in whirlin, aPDZ domain molecule involved in stereocilia elongation, cause deafness in thewhirler mouse and families with DFNB31. Nat. Genet. 34, 421-428.
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.
McGuckin, M. A., Eri, R. D., Das, I., Lourie, R. and Florin, T. H. (2010). ER stressand the unfolded protein response in intestinal inflammation. Am. J. Physiol. 298,G820-G832.
Michalski, N., Michel, V., Bahloul, A., Lefevre, G., Barral, J., Yagi, H., Chardenoux,S., Weil, D., Martin, P., Hardelin, J. P., et al. (2007). Molecular characterization ofthe ankle-link complex in cochlear hair cells and its role in the hair bundlefunctioning. J. Neurosci. 27, 6478-6488.
Mitrovic, S., Ben-Tekaya, H., Koegler, E., Gruenberg, J. and Hauri, H. P. (2008).The cargo receptors Surf4, endoplasmic reticulum-Golgi intermediate compartment(ERGIC)-53, and p25 are required to maintain the architecture of ERGIC and Golgi.Mol. Biol. Cell 19, 1976-1990.
Munro, S. and Pelham, H. R. (1987). A C-terminal signal prevents secretion of luminalER proteins. Cell 48, 899-907.
Muppirala, M., Gupta, V. and Swarup, G. (2011). Syntaxin 17 cycles between the ERand ERGIC and is required to maintain the architecture of ERGIC and Golgi. Biol.Cell 103, 333-350.
Nachury, M. V., Seeley, E. S. and Jin, H. (2010). Trafficking to the ciliary membrane:how to get across the periciliary diffusion barrier? Annu. Rev. Cell Dev. Biol. 26, 59-87.
Overlack, N., Maerker, T., Latz, M., Nagel-Wolfrum, K. and Wolfrum, U. (2008).SANS (USH1G) expression in developing and mature mammalian retina. VisionRes. 48, 400-412.
Pazour, G. J., Baker, S. A., Deane, J. A., Cole, D. G., Dickert, B. L., Rosenbaum, J.L., Witman, G. B. and Besharse, J. C. (2002). The intraflagellar transport protein,IFT88, is essential for vertebrate photoreceptor assembly and maintenance. J. CellBiol. 157, 103-114.
Phillips, J. B., Blanco-Sanchez, B., Lentz, J. J., Tallafuss, A., Khanobdee, K.,Sampath, S., Jacobs, Z. G., Han, P. F., Mishra, M., Titus, T. A. et al. (2011).Harmonin (Ush1c) is required in zebrafish Müller glial cells for photoreceptorsynaptic development and function. Dis. Model. Mech. 4, 786-800.
Puffenberger, E. G., Jinks, R. N., Sougnez, C., Cibulskis, K., Willert, R. A., Achilly,N. P., Cassidy, R. P., Fiorentini, C. J., Heiken, K. F., Lawrence, J. J. et al. (2012).Genetic mapping and exome sequencing identify variants associated with five noveldiseases. PLoS ONE 7, e28936.
Reiners, J., Nagel-Wolfrum, K., Jürgens, K., Märker, T. and Wolfrum, U. (2006).Molecular basis of human Usher syndrome: deciphering the meshes of the Usherprotein network provides insights into the pathomechanisms of the Usher disease.Exp. Eye Res. 83, 97-119.
Riazuddin, S., Belyantseva, I. A., Giese, A. P., Lee, K., Indzhykulian, A. A.,Nandamuri, S. P., Yousaf, R., Sinha, G. P., Lee, S., Terrell, D. et al. (2012).Alterations of the CIB2 calcium- and integrin-binding protein cause Usher syndrometype 1J and nonsyndromic deafness DFNB48. Nat. Genet. 44, 1265-1271.
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.
RESEARCH ARTICLE Disease Models & Mechanisms (2014) doi:10.1242/dmm.014068
Dis
ease
Mod
els
& M
echa
nism
s
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 andfunctional interaction between protocadherin 15 and myosin VIIa in mechanosensoryhair cells. J. Neurosci. 26, 2060-2071.
Shinde, V. M., Sizova, O. S., Lin, J. H., LaVail, M. M. and Gorbatyuk, M. S. (2012).ER stress in retinal degeneration in S334ter Rho rats. PLoS ONE 7, e33266.
Siemens, J., Kazmierczak, P., Reynolds, A., Sticker, M., Littlewood-Evans, A. andMüller, U. (2002). The Usher syndrome proteins cadherin 23 and harmonin form acomplex by means of PDZ-domain interactions. Proc. Natl. Acad. Sci. USA 99,14946-14951.
Söderberg, O., Gullberg, M., Jarvius, M., Ridderstråle, K., Leuchowius, K. J.,Jarvius, J., Wester, K., Hydbring, P., Bahram, F., Larsson, L. G. et al. (2006).Direct observation of individual endogenous protein complexes in situ by proximityligation. Nat. Methods 3, 995-1000.
Söllner, C., Rauch, G. J., Siemens, J., Geisler, R., Schuster, S. C., Müller, U.,Nicolson, T.; Tübingen 2000 Screen Consortium (2004). Mutations in cadherin 23affect tip links in zebrafish sensory hair cells. Nature 428, 955-959.
Tanaka, H., Morimura, R. and Ohshima, T. (2012). Dpysl2 (CRMP2) and Dpysl3(CRMP4) phosphorylation by Cdk5 and DYRK2 is required for proper positioning ofRohon-Beard neurons and neural crest cells during neurulation in zebrafish. Dev.Biol. 370, 223-236.
Tanimoto, M., Ota, Y., Inoue, M. and Oda, Y. (2011). Origin of inner ear hair cells:morphological and functional differentiation from ciliary cells into hair cells inzebrafish inner ear. J. Neurosci. 31, 3784-3794.
Thisse, C. and Thisse, B. (2008). High-resolution in situ hybridization to whole-mountzebrafish embryos. Nat. Protoc. 3, 59-69.
Tilney, L. G., Tilney, M. S. and DeRosier, D. J. (1992). Actin filaments, stereocilia,and hair cells: how cells count and measure. Annu. Rev. Cell Biol. 8, 257-274.
Trifilieff, P., Rives, M. L., Urizar, E., Piskorowski, R. A., Vishwasrao, H. D.,Castrillon, J., Schmauss, C., Slättman, M., Gullberg, M. and Javitch, J. A.(2011). Detection of antigen interactions ex vivo by proximity ligation assay:endogenous dopamine D2-adenosine A2A receptor complexes in the striatum.Biotechniques 51, 111-118.
Tsujikawa, M. and Malicki, J. (2004). Intraflagellar transport genes are essential fordifferentiation and survival of vertebrate sensory neurons. Neuron 42, 703-716.
van Wijk, E., van der Zwaag, B., Peters, T., Zimmermann, U., Te Brinke, H.,Kersten, F. F., Märker, T., Aller, E., Hoefsloot, L. H., Cremers, C. W. et al. (2006).The DFNB31 gene product whirlin connects to the Usher protein network in thecochlea and retina by direct association with USH2A and VLGR1. Hum. Mol. Genet.15, 751-765.
Vembar, S. S. and Brodsky, J. L. (2008). One step at a time: endoplasmic reticulum-associated degradation. Nat. Rev. Mol. Cell Biol. 9, 944-957.
Verpy, E., Leibovici, M., Zwaenepoel, I., Liu, X. Z., Gal, A., Salem, N., Mansour, A.,Blanchard, S., Kobayashi, I., Keats, B. J. et al. (2000). A defect in harmonin, aPDZ domain-containing protein expressed in the inner ear sensory hair cells,underlies Usher syndrome type 1C. Nat. Genet. 26, 51-55.
Wang, W., Sacher, M. and Ferro-Novick, S. (2000). TRAPP stimulates guaninenucleotide exchange on Ypt1p. J. Cell Biol. 151, 289-296.
Wang, J., Morita, Y., Mazelova, J. and Deretic, D. (2012). The Arf GAP ASAP1provides a platform to regulate Arf4- and Rab11-Rab8-mediated ciliary receptortargeting. EMBO J. 31, 4057-4071.
Westerfield, M. (2007). The Zebrafish Book: A Guide for the Laboratory Use ofZebrafish (Danio rerio), 5 edn. Eugene, OR: University of Oregon Press.
Wu, X., Rao, K., Bowers, M. B., Copeland, N. G., Jenkins, N. A. and Hammer, J. A.,III (2001). Rab27a enables myosin Va-dependent melanosome capture by recruitingthe myosin to the organelle. J. Cell Sci. 114, 1091-1100.
Wu, X. S., Rao, K., Zhang, H., Wang, F., Sellers, J. R., Matesic, L. E., Copeland, N.G., Jenkins, N. A. and Hammer, J. A., III (2002). Identification of an organellereceptor for myosin-Va. Nat. Cell Biol. 4, 271-278.
Yu, S., Satoh, A., Pypaert, M., Mullen, K., Hay, J. C. and Ferro-Novick, S. (2006).mBet3p is required for homotypic COPII vesicle tethering in mammalian cells. J. CellBiol. 174, 359-368.
Zallocchi, M., Meehan, D. T., Delimont, D., Rutledge, J., Gratton, M. A., Flannery,J. and Cosgrove, D. (2012). Role for a novel Usher protein complex in hair cellsynaptic maturation. PLoS ONE 7, e30573.
Zanetti, G., Pahuja, K. B., Studer, S., Shim, S. and Schekman, R. (2011). COPII andthe regulation of protein sorting in mammals. Nat. Cell Biol. 14, 20-28.
559
RESEARCH ARTICLE Disease Models & Mechanisms (2014) doi:10.1242/dmm.014068
