RE so UR c E Molecular architecture of the chick vestibular hair bundle Jung-Bum Shin1,2,8, Jocelyn F Krey2,8, Ahmed Hassan3, Zoltan Metlagel3, Andrew N Tauscher3, James M Pagana1,2, Nicholas E Sherman4, Erin D Jeffery4, Kateri J Spinelli2, Hongyu Zhao2, Phillip A Wilmarth5, Dongseok Choi6, Larry L David5, Manfred Auer3 & Peter G Barr-Gillespie2,7 Hair bundles of the inner ear have a specialized structure and protein composition that underlies their sensitivity to mechanical stimulation. Using mass spectrometry, we identified and quantified >1,100 proteins, present from a few to 400,000 copies per stereocilium, from purified chick bundles; 336 of these were significantly enriched in bundles. Bundle proteins that we detected have been shown to regulate cytoskeleton structure and dynamics, energy metabolism, phospholipid synthesis and cell signaling. Three-dimensional imaging using electron tomography allowed us to count the number of actin-actin cross-linkers and actin- membrane connectors; these values compared well to those obtained from mass spectrometry. Network analysis revealed several hub proteins, including RDX (radixin) and SLC9A3R2 (NHERF2), which interact with many bundle proteins and may perform functions essential for bundle structure and function. The quantitative mass spectrometry of bundle proteins reported here establishes a framework for future characterization of dynamic processes that shape bundle structure and function. An outstanding example of a specialized organelle devoted to a single ­identified those proteins selectively targeted to bundles. Many bundle- purpose, the vertebrate hair bundle transduces mechanical signals enriched proteins are expressed from deafness-associated genes, for the inner ear, converting sound and head movement to electrical confirming their essential function in the inner ear. We also imaged signals that propagate to the central nervous system. Protruding stereocilia using electron tomography and counted actin-actin cross- from the apical surface of a sensory hair cell, a bundle typically con- linkers and actin-membrane connectors; those counts compared fav­ sists of 50–100 actin-filled stereocilia and, at least during develop­ orably to mass-spectrometric estimates for cross-linker and connector ment, an axonemal kinocilium1. A bundle enlists ~100 transduction proteins. To place the bundle’s proteome into a network of functional Nature America, Inc. All rights reserved. Inc. Nature America, 3 channels, which are mechanically gated by tip links as external and structural interactions, we assembled an interaction map that forces oscillate the bundle; the opening and closing of the channels highlights the central roles in hair-bundle function played by actin, 2+ © 201 in turn modulates the hair cell’s membrane potential, controlling phosphatidylinositol-4,5-bisphosphate (PtdIns(4,5)P2), Ca and neurotransmitter release. CALM (calmodulin). Moreover, we identified two other key hub pro- Because hair bundles have a reduced protein complement and carry teins: the ezrin-radixin-moesin (ERM) family member RDX (radixin), out a specialized task, once we know which proteins are present—as important in hair-bundle function5, and SLC9A3R2 (NHERF2; solute well as their concentrations and interactions—understanding bundles’ carrier family 9 member 3 regulator 2), a PDZ-domain adaptor pro- assembly and operation seems possible. Although genetics studies tein that couples RDX to many transmembrane proteins6. The com- have identified many proteins essential for bundle function2, others prehensive view offered by quantitative mass spectrometry reveals may have escaped detection because they are essential during develop­ functional pathways in hair bundles and, on the basis of the absence ment or, in some cases, can be compensated for by paralogs. To dis- of key protein families, also rules out alternative mechanisms. cover these additional proteins, biochemical strategies are essential. Although bundles are scarce, quantitative mass spectrometry3 has RESULTS the sensitivity and accuracy to detect and quantify the bundle’s Mass spectrometry of purified hair bundles protein complement. Using liquid-chromatographic tandem mass spectrometry (LC- Our previous analysis of hair-bundle proteins using mass spec- MS/MS), we identified proteins from hair bundles and epithelia of trometry detected 59 proteins, including several that are critical for utricles (Supplementary Fig. 1), vestibular organs that detect linear bundle function4. Here, using a more sensitive mass spectrometer, acceleration, from embryonic day 20–21 chicks; at this age, utricles we detected over 1,100 proteins from chick vestibular bundles and are functional7. Bundles (BUN) were enriched 40-fold, to ~80% purity 1Department of Neuroscience, University of Virginia, Charlottesville, Virginia, USA. 2Oregon Hearing Research Center, Oregon Health & Science University, Portland, Oregon, USA. 3Life Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, California, USA. 4W.M. Keck Biomedical Mass Spectrometry Laboratory, University of Virginia, Charlottesville, Virginia, USA. 5Department of Biochemistry & Molecular Biology, Oregon Health & Science University, Portland, Oregon, USA. 6Department of Public Health & Preventive Medicine, Oregon Health & Science University, Portland, Oregon, USA. 7Vollum Institute, Oregon Health & Science University, Portland, Oregon, USA. 8These authors contributed equally to this work. Correspondence should be addressed to P.G.B.-G. ([email protected]). Received 20 August 2012; accepted 17 December 2012; published online 20 January 2013; doi:10.1038/nn.3312 NATURE NEUROSCIENCE ADVANCE ONLINE PUBLICATION 1 RE so UR c E Figure 1 Quantitative analysis of chick a b c hair-bundle proteins. ( ) Top, proteins 0 0 a Epithelium –1 –1 GAPDH identified in bundles and epithelium (two or 2,753 proteins 8/8 CKB –2 8/8 –2 CALM FSCN2 more experiments). Bottom left, representation (BUN) ANXA5 riBAQ –3 6/8 –3 FSCN2 of bundle proteins as bundle-enriched, 10 (UTR) Bundles –4 mole fraction –4 10 immunoblot MYO1C unenriched and epithelium-enriched (by Log 1,095 proteins –5 –5 2/8 mass spectrometry protein frequency). Bottom right, same as Log –6 –6 left except with the mole fractions of proteins –6 –5 –4 –3 –2 –1 0 –6 –5 –4 –3 –2 –1 0 30 bundle–only Log mole fraction Log mole fraction in each class summed. (b) Calibration curve 10 10 >3 Fold spiked proteins mass spectrometry relating mole fraction of human 0.3–3 enrichment <0.3 d e protein standards spiked into E. coli extract 500 200 143 Epithelium to riBAQ value. The number of identified 46% 400 150 level proteins is indicated for each data point 300 Bundles 100 (mean s.d.). The points corresponding to 527 455 ± 47% 200 Contamination Number −2 −3 −4 7% Number 50 mole fractions of 10 , 10 and 10 100 were fit with a line constrained through the 0 0 origin (y = 1.02x; R = 0.999). (c) Calibration Proteins detected Protein mole fraction –8 –6 –4 –2 0 –6 –4 –2 0 2 4 Log (riBAQ) Log (enrichment) curve relating mole fraction determined from 10 2 Bundles Epithelium riBAQ values to mole fraction measured f g Epithelium Bundles Bundles only (1,125 proteins) (2,753 proteins) SLC1A3 ACTG1 using quantitative immunoblots with purified 1.00 GLG1 OCM –0 UGGT1 FSCN2 PLOD2 TPI1 proteins as standards; data points are RPL4 1,090 proteins (97% of total) SNRNP200 YWHAE PRPF8 CRABP1 mean ± s.e.m. and are fit by y = 0.98x –1 SF3B1 CALM 2,559 proteins (93%) LRPPRC RDX VILL PLS1 (R = 0.97). Data for CKB and GAPDH were COL14A1 XIRP2 ZNF326 0.75 CLIC5 CA7 –2 COL8A2 ADH1B from ref. 4. Dotted lines in a,b correspond to CTNNA1 ALDOA CPD SLC9A3R2 SERPINF1 28 proteins (2%) DNAJC10 LTF unity. (d) Abundance distribution of bundle NOMO1 GSTO1 PSMD11 156 proteins (6%) –3 ANK3 CEP290 MRC2 ENO1 TNC PSMB7 MIM1 MYH9 ESPN CHAD and epithelium proteins; single Gaussian fits. HSPA8 CLINT1 0.50 FSCN2 DNAJC13 PLS2 GRXCR1 GAPDH mole fraction MYEF2 FYCO1 ESPNL (e) Enrichment distribution of proteins TUBB4B –4 PLXNA4 SRM SERPINB7 10 TUBA5 6 proteins (0.5%) NCBP1 ATP8B1 ARL13B OCM SMARCC1 PPID detected in bundles and epithelium; single TOP2B SET PDZD7 CKB 29 proteins (1%) Log PITRM1 INTS7 CAB39L –5 SMARCA2 EIF3J Gaussian fit. (f) Cumulative protein molar CTNND1 PFN2 ACTG1 CYFIP2 PTPRF HYI 0.25 GAPDH DHX30 SPUNK WDR16 abundance, from highest to lowest. The most Cumulative protein abundance CKB CHD4 FBXO44 TRIM23 AGR3 –6 CNOT1 NUDC 0.8 H32 GAK ATIC abundant proteins in bundles and epithelium 1 protein (0.1%) GPD2 UNC119B TUBB4B OTOGL Power H2AFJ DMD USH1C 9 proteins (0.3%) MYO15A coefficient are indicated. (g) Mole fractions of proteins in ACTG1 GOLGA4 HIST1H4E LIMCH1 MYO1H AKAP9 PCDH15 –0.8 epithelium (left) and bundle (right); the slope 500 1,000 1,500 2,000 2,500 –3 –2 –1 –0 of the line connecting them represents Ranked proteins (by abundance) Log10 bundle purity bundle-to-epithelium enrichment. Proteins most highly enriched in the epithelium are indicated at left, those most highly enriched in bundles at right. Hue represents relative enrichment (power coefficient of fit connecting points) for each protein. Far right, proteins detected only in bundles. (see below), using the twist-off technique4,8. To obtain utricular were detected, demonstrating the limitations in detecting proteins at epithelia (UTR), we used an eyelash to peel the hair-cell and low mole fractions. We carried out a linear regression (log10 riBAQ = Nature America, Inc. All rights reserved. Inc. Nature America, 3 −2 −3 −4 supporting-cell layer from the underlying stroma layer (Supplementary 1.02 ± 0.01 × log10 mole fraction) with the 10 , 10 and 10 mole Fig. 1b). Four experiments each of BUN and UTR were analyzed. fraction data points (Fig. 1b). Although the 10−5 data point did not © 201 We identified proteins using an Orbitrap mass spectrometer, ana- fall on the regression line, only two of eight proteins were detected. We lyzing data with the Andromeda search engine and MaxQuant9,10. conclude that the correspondence between riBAQ and mole fraction Proteins that shared more than 20% of their detected peptides were is nearly exact, at least above to a mole fraction of ~10−5 (Fig.