Repair of hair bundles in sea anemones by secreted proteins

Repair of hair bundles in sea anemones by secreted proteins

Glen M. Watson *, Patricia Mire, Renee R. HudsonDepartment of Biology, University of Southwestern Louisiana, Lafayette, LA 70504-2451, USA

Received 4 July 1997; accepted 9 October 1997

Abstract

Sea anemones are sessile invertebrates that detect movements of prey using numerous hair bundles located on tentaclessurrounding their mouth. Previously we found that hair bundles of anemones are structurally and functionally similar to those ofvertebrates. After 10^15 min exposure to calcium depleted buffers, hair bundles in chickens suffer moderate damage from which theyrecover in 12 h without requiring new protein synthesis [Zhao, Yamoah and Gillespie, Proc. Natl. Acad. Sci. USA 94 (1996)15469^15474]. We find that after 1 h exposure to calcium free seawater, hair bundles of anemones suffer extensive damage fromwhich they recover in 4 h, apparently because of newly synthesized, secretory proteins called `repair proteins'. Recovery is delayed ina dose dependent fashion by cycloheximide. In the presence of exogenously added repair proteins, recovery occurs within 8 min andis cycloheximide insensitive. Recovery is ascertained by a bioassay performed on intact specimens, by electrophysiology, and bytimelapse video microscopy. Fraction LL, a chromatographic fraction with bioactivity comparable to the complete mixture of repairproteins, consists of complexes having an estimated mass of 2000 kDa. Avidin based cytochemistry suggests that biotinylatedfraction LL binds to damaged hair bundles. SDS-PAGE gel electrophoresis demonstrates that fraction LL contains 8^10 polypeptides of90 kDa or smaller. At least four of these polypeptides apparently are consumed during the repair process. Negatively stained samplesof fraction LL are shown by transmission electron microscopy to include filamentous structures similar in length (150 nm) and width(6 nm) to linkages between stereocilia. The filamentous structures can be associated with globular structures (20 nm in diameter). Amodel is presented wherein repair proteins comprise replacement linkages and enzymes that attach linkages to appropriatemembrane proteins. z 1998 Elsevier Science B.V.

Key words: Tip link; Hair cells ; Signal transduction

1. Introduction

Substantial evidence suggests that linkages betweenstereocilia play a critical role in mechanotransductionof hair cells with recent focus on tip links interconnect-ing the tip of one stereocilium to the adjacent, tallerstereocilium (Pickles et al., 1984; Osborne et al.,1988). According to the model, a tip link attaches tothe gate of a transduction channel embedded in theplasma membrane at one or both ends of the tip link(reviews Howard et al., 1988; Ashmore, 1991; Pickles,1993; Hudspeth, 1985, 1992). De£ecting the bundle to-

wards taller stereocilia produces tension on the tip linkcausing it to pull open the gate of the channel. Cationspass through the channel to generate a positive inwardcurrent that depolarizes the membrane (Hudspeth andCorey, 1977). De£ecting the bundle towards shorterstereocilia allows slack on the tip link, favoring theconformation in which the channel is closed. The mem-brane potential hyperpolarizes (Hudspeth and Corey,1977). Lateral de£ections are without e¡ect (Shotwellet al., 1981). Agents that disrupt tip links also disruptsignal transduction. Within a second of exposing haircells from bullfrogs to calcium depleted bu¡ers, signaltransduction is abolished and tip links disappear (Assadet al., 1991). Hair cells from chickens must be exposedto calcium depleted bu¡ers for 10^15 min to cause tiplinks to disappear (Zhao et al., 1996). Tip links disap-pear from hair cells of guinea pigs after exposing them

0378-5955 / 98 / $19.00 ß 1998 Elsevier Science B.V. All rights reservedPII S 0 3 7 8 - 5 9 5 5 ( 9 7 ) 0 0 1 8 5 - 8

HEARES 2940 9-1-98

* Corresponding author. Tel. : +1 (318) 482-5667;Fax: +1 (318) 482-5660; E-mail: [email protected]

Hearing Research 115 (1998) 119^128

to elastase for 1 min (Preyer et al., 1995). Attempts toisolate tip link proteins and/or to characterize genescoding for tip links are hampered by a low abundanceof the proteins estimated to be about 50 tip links perhair cell (Hudspeth, 1992).

The simplest animals having hair bundle mechanore-ceptors are members of the phylum Cnidaria, includingjelly¢sh, corals, hydra and sea anemones (Barnes,1980). Sea anemones prey upon zooplankton (Shick,1991). Prey movements are detected by hair bundleslocated on tentacles (reviewed in Watson and Mire-Thi-bodeaux, 1994). Detection of vibrations at particularkey frequencies predisposes e¡ector cells called cnido-cytes to discharge cnidae (stinging capsules, Mariscal,1974) in response to contact with test probes. Dischargeof nematocysts (a class of cnidae) into vibrating testprobes is exploited as a quantitative bioassay for vibra-tion sensitivity. The bioassay consistently shows an ap-proximate doubling of nematocysts discharged into testprobes vibrating at key frequencies as compared to oth-er frequencies or to nonvibrating test probes (Watsonand Mire-Thibodeaux, 1994). Under experimental con-ditions in which hair bundles are completely obliterated(e.g. after cytochalasin treatment, Watson and Hessing-er, 1991), discharge into vibrating test probes decreasesto that elicited by nonvibrating test probes. Dischargeinto nonvibrating test probes is una¡ected by cytocha-lasin treatment. Thus, in the anemone system, where-as sensitivity to vibrations requires hair bundles, sensi-tivity to nonvibrating targets does not require hairbundles.

Previously, we found several points of similarity be-tween hair bundles in anemones and vertebrates. Struc-turally, anemone hair bundles, like vertebrate hair bun-dles, consist of actin-based stereocilia interconnected bytip links among other linkages (Watson et al., 1997). Inaddition, vibration dependent discharge of nematocystsis inhibited by aminoglycosides at doses known to in-hibit signal transduction in vertebrate hair cells (Kroeseet al., 1989; Watson et al., 1997). Electrophysiologicalrecordings show that de£ecting anemone hair bundleswith transient pu¡s of bu¡er from nearby pipettescauses transient shifts in DC current. These electro-physiological responses of anemone hair bundles arerapidly and reversibly blocked by aminoglycoside anti-biotics, exhibit adaptation to prolonged de£ection, andapparently show an asymmetric sigmoidal stimulus/re-sponse curve (Mire and Watson, 1997). Perhaps mostpertinent to the present study is the observation thatvibration dependent discharge of nematocysts is selec-tively abolished by brief exposure to calcium free sea-water (Watson et al., 1997). In the present study, wepresent evidence that hair bundles in sea anemones selfrepair following exposure to calcium free seawater and,furthermore, that repair is accomplished by proteinssecreted by the anemone.

2. Materials and methods

2.1. Animal culture and maintenance

Clonal specimens of the sea anemone Haliplanellaluciae were reared in natural seawater at a salinity of32 parts per thousand and at 16^18³C according topublished methods (Minasian and Mariscal, 1979).The animals were fed to repletion twice weekly usingfreshly hatched brine shrimp nauplii. Using these meth-ods, we have maintained continuous cultures of ane-mones for 13 years.

2.2. Bioassay for vibration sensitivity

Animals were transferred from mass culture to plas-tic Petri dishes where they were allowed a minimum of1 h to recover from handling. Tentacles were touchedwith test probes consisting of nylon ¢shing line coatedwith 25% gelatin to a thickness of approximately 200Wm. Test probes were inserted into 2 Wl glass pipettesattached to a piezo disk induced to vibrate at 55 Hz bya function generator as described (Watson et al., 1997).The gelatin coated tip of the test probe was brie£yhydrated then touched to tentacles. Test probes were¢xed brie£y in 2.5% glutaraldehyde, then examined at400U total magni¢cation using phase contrast optics.Microbasic p-mastigophore type nematocysts werecounted for a single ¢eld of view for each test probe.For each treatment, a total of eight test probes wasscored, each touched to a di¡erent sea anemone.

2.3. Treatments to abolish vibration sensitivity

Anemones were placed in calcium free seawater forintervals ranging from 1 min to 1 h as dictated by theparticular experiment. Calcium free seawater consistedof (in mM): 447 NaCl, 9 KCl, 24 MgCl2, 25 MgSO4,2.1 NaHCO3, 2 EGTA. When replacing seawater withcalcium free seawater, the calcium free seawater wasrapidly changed three times to ensure removal of cal-cium containing seawater. After exposure to calciumfree seawater, specimens were rinsed three times withcalcium containing arti¢cial seawater to ensure the re-moval of EGTA. Calcium containing arti¢cial seawaterconsisted of (in mM): 423 NaCl, 9 KCl, 24 MgCl2,25 MgSO4, 12 CaCl2, and 2.1 NaHCO3. Specimenswere allowed 10 min to recover from the anesthetice¡ects of calcium free seawater before vibration sensi-tivity was tested using the bioassay described above.

In some experiments anemones were exposed to elas-tase (20 IU/ml, type IV from porcine pancreas (Sigma))for 1 min following the methods of Preyer et al. (1995).The specimens were rinsed three times with seawater toremove elastase and then vibration sensitivity wastested using the bioassay described above.

HEARES 2940 9-1-98

G.M. Watson et al. / Hearing Research 115 (1998) 119^128120

2.4. Collection and processing of repair seawater

Approximately 500 anemones were placed intobeakers containing 50 ml seawater. Seawater was re-placed with calcium free seawater for 1 h, after whichspecimens were exposed to calcium containing seawaterfor 4 h. At this point, the seawater was removed and¢ltered (0.45 Wm pore size). Proteins in the seawaterwere concentrated in a vacuum concentrator (Micro-prodicon, Spectrum, Carson, CA) to form concentratedrepair seawater (CRSW). Concentrated control sea-water (CCSW) was collected in essentially the samefashion except that anemones were not exposed to cal-cium free seawater.

2.5. Liquid chromatography

Three batches of CRSW were pooled and thenpassed through Sepharose 2B, 4B, or 6B columns (Sig-ma). This step was intended to compensate for the pos-sibility of incomplete recovery of protein from the col-umn. Since repair proteins function in seawater, themobile phase consisted of arti¢cial seawater. Biologi-cally active fractions were concentrated as describedabove.

2.6. SDS-PAGE electrophoresis

Gradient gels (4^20%) were poured and electropho-resis performed as described in Bollag and Edelstein(1991). A Western blot of the gel was obtained, stainedwith colloidal gold (Total Protein Stain), and enhancedby silver staining according to manufacturers instruc-tions (Bio-Rad, Hercules, CA). The blot was imagedby video camera, digitized and densitometry performedusing Image1-AT hardware and software (UniversalImaging, West Chester, PA).

2.7. Microscopy

Di¡erential interference contrast (DIC) video micros-copy was performed on excised tentacles as described inWatson and Roberts (1995). Brie£y, an inverted micro-scope was used with imaging of the specimen performedthrough a coverslip. The experimental design was suchthat the specimen (an excised tentacle) was held by asuction pipette, precluding the use of a second coverslipbetween the specimen and condenser. Using a 100Uobjective lens (n.a. = 1.25), a slight deterioration of im-age quality was noted as compared to samples sand-wiched between two coverslips. Nevertheless, the qual-ity of the image was adequate to allow measurements ofthe bundles by computer software (Image1-AT). In oneexperiment, bright ¢eld microscopy of the repair proc-ess was performed using a 40U objective lens(n.a. = 0.55). Before excision of tentacles, anemones

were anesthetized for 2 h in high potassium arti¢cialseawater (KSW) consisting of (mM): 50 KCl, 382NaCl, 24 MgCl2, 26 MgSO4, 12 CaCl2, and 2 NaHCO3.Previously we found that KSW anesthetized anemoneswithout causing bundles to splay (Mire-Thibodeauxand Watson, 1994). The more traditional anestheticconsisting of seawater forti¢ed with MgCl2 (e.g. Wat-son and Mariscal, 1983) caused hair bundles to splay.Images were either saved at intervals during repairdirectly to computer hard drive, or recorded onto videotape using a timelapse video recorder. In this case,the monitor was photographed at intervals during play-back of the video tape. Final prints of light micros-copy images were obtained by photographing the mon-itor.

Anemones were processed for scanning electron mi-croscopy as described in Watson and Hessinger (1991).Anemones were ¢rst anesthetized in KSW for 1 h,transferred to calcium free KSW for 1 h consisting of(mM): 50 KCl, 406 NaCl, 24 MgCl2, 26 MgSO4,2 EGTA and 2 NaHCO3 and then ¢xed immediatelyor at intervals after being returned to KSW.

In order to localize possible binding sites for fractionL on hair bundles, fraction L was biotinylated usingNHS-biotin (Sigma) as described in Bayer and Wilchek(1990). Biotinylated fraction L was puri¢ed on a Se-pharose 4B column and bioactivity con¢rmed usingthe bioassay described above. Specimens were exposedto calcium free KSW for 30 min to damage their hairbundles, after which they were ¢xed in 2.5% glutaralde-hyde for 30 min then incubated with biotinylated frac-tion L for 10 min, followed by avidin rhodamine for10 min. Specimens were rinsed and then imaged usinga cooled CCD camera (Model ST-6, SBIG, Santa Bar-bara, CA).

Transmission electron microscopy was used to exam-ine negatively stained samples of fraction L. Negativestaining of fraction L in 1% uranyl acetate was per-formed as described in Pollard (1982). Grids were ex-amined at 75 kV using an Hitachi H-7000 transmissionelectron microscope.

2.8. Electrophysiology

Electrophysiological recordings were obtained duringbundle de£ections using the cell-attached, loose-patchcon¢guration (Mire and Watson, 1997). Brie£y, the re-cording pipette was attached to the apical surface of acell at the base of a hair bundle (seal resistance = 10^20M6). A pu¡er pipette, positioned near the hair bundle,ejected bu¡er to de£ect the bundle. The pressure andduration of the pu¡ was controlled by a pneumaticpicopump. The hair bundle was de£ected with a trainof 5 pu¡s of 15 psi, 45 ms, separated by approximately1 s between pu¡s. Trains of de£ection stimuli and elec-trical responses were recorded before the addition of

HEARES 2940 9-1-98

G.M. Watson et al. / Hearing Research 115 (1998) 119^128 121

CRSW and then at 5 min intervals after the addition ofCRSW through 30 min.

Experiments on animals reported in this study wereperformed in accordance with the guidelines of the Dec-laration of Helsinki.

3. Results

3.1. Recovery of anemone hair bundles

Relatively brief exposure to calcium free seawater

(1^5 min) abolishes vibration sensitivity in anemoneswithout causing obvious damage to the hair bundlesat the light microscopic level (Watson et al., 1997).After the animals are returned to calcium containingseawater, vibration sensitivity recovers. The recoveryis fastest in animals exposed to calcium free seawaterfor 1 min, followed by 5 min (Fig. 1A). After 1 h ex-posure to calcium free seawater, anemone hair bundlesare profoundly disorganized (Fig. 1B). Within 4 h afterreturning test animals to calcium containing seawater,both vibration sensitivity and normal morphology ofhair bundles are restored (Fig. 1A,C). Recovery of vi-

HEARES 2940 9-1-98

Fig. 1. Recovery of morphology of hair bundles and vibration dependent discharge of nematocysts after treatment in calcium free seawater. A:Time course of recovery for vibration dependent discharge of nematocysts (vibration sensitivity) following exposure to calcium free seawaterfor (solid squares) 1 min, (empty triangles) 5 min, or (solid circles) 1 h. The animals were returned to calcium containing seawater where theirtentacles were touched with gelatin coated probes vibrating at 55 Hz, a preferred frequency. See Section 2 for details. Brie£y, microbasic p-mas-tigophore nematocysts discharged into the probes were counted for a single ¢eld of view for eight probes, each touched to a di¡erent anemone.Data points represent the mean number of nematocysts counted at 400U magni¢cation þ S.E.M. (n = 8). B: Scanning electron microscopy ofthe tentacle surface in anemones ¢xed immediately after 1 h exposure to calcium free KSW. Hair bundles are unrecognizable. C: Scanning elec-tron microscopy of the tentacle surface in anemones ¢xed after 1 h exposure to calcium free KSW followed by 4 h in calcium containingKSW. Hair bundles (hb) are abundant. Scale bar = 1 Wm. D: E¡ects of cycloheximide on the recovery of vibration sensitivity. Following 1 h incalcium free seawater (with cycloheximide at the dose speci¢ed below), animals were returned to calcium containing seawater with (solid circles)no cycloheximide, (empty triangles) 1036 M cycloheximide, (solid squares) 1035 M cycloheximide, or (empty circles) 1034 M cycloheximide. Vi-bration sensitivity was tested as described in panel A. E: E¡ects of `repair seawater' on the recovery of vibration sensitivity. Following 1 h incalcium free seawater, animals were returned either to (solid circles) `repair seawater' (calcium containing seawater in which recovery of vibra-tion sensitivity had previously occurred for a separate set of anemones), or (empty circles) boiled repair seawater (boiled for 5 min). Vibrationsensitivity was assayed as described above.


bration sensitivity is delayed in a dose dependent fash-ion by cycloheximide, an inhibitor of protein synthesis.For example, in the presence of 1034 M cycloheximide,vibration sensitivity is restored in 11 h (2.75 times aslong as in the absence of cycloheximide) (Fig. 1D).Assuming that tip links and other linkages are requiredto maintain proper structure and function of hair bun-dles, molecules responsible for repairing bundles mightbe expected to occur in the extracellular £uid. To ex-amine this possibility, test animals were exposed to cal-cium free seawater for 1 h and then placed in `repairseawater' (RSW, calcium containing seawater in whichrecovery had previously occurred for a separate set ofanimals). Under these experimental conditions, vibra-tion sensitivity is restored within 15 min (Fig. 1E), the¢rst time point tested. In boiled RSW, vibration sensi-tivity is restored in 4 h (Fig. 1E), comparable to controlanimals not placed in RSW. RSW was concentratedand dialyzed to form CRSW. Control seawater was

similarly processed to form CCSW. CRSW restores vi-bration sensitivity in approximately 8 min (Fig. 2A).This time course is not signi¢cantly a¡ected by 1034

M cycloheximide (Fig. 2A). After trypsin treatment ofCRSW, bioactivity is lost. CCSW lacks bioactivity(data not shown).

Exposure to calcium free KSW for 30 min causesstereocilia of the bundles to splay as evidenced by awidening of the tip and middle regions of the hair bun-dles (Fig. 2B), but does not cause such extensive dam-age (as does 1 h exposure) to render the bundles unrec-ognizable. In order to investigate e¡ects of CRSW onmorphology of individual hair bundles using timelapsemicroscopy, it was necessary that the damaged bundlesbe recognizable. Hence, a 30 min exposure to calciumfree KSW was used for these experiments. Once thebundles are returned to calcium containing KSW, theaddition of CRSW causes stereocilia of splayed hairbundles to become more tightly packed together within

HEARES 2940 9-1-98

Fig. 2. E¡ects of concentrated repair seawater (CRSW) on recovery of morphology of hair bundles and vibration sensitivity after treatment incalcium free seawater. A: E¡ects of CRSW on time course of recovery for vibration sensitivity. Animals were exposed to calcium free seawaterfor 1 h (without or with cycloheximide) then transferred to calcium containing seawater (solid squares) alone or with 1034 M cycloheximide(empty circles). After a 10 min recovery, CRSW was added (20 Wl in 6 ml seawater) and vibration sensitivity assayed as described above. B: Ef-fects of CRSW on hair bundle morphology. The width of hair bundles was measured in calcium containing KSW (CON) using video micro-scopy (DIC optics and a 100U objective lens) at the (solid squares) base, (solid triangles) middle and (solid circles) tip regions. After 30 minexposure to calcium free KSW (CaF), the same hair bundles were measured. The calcium free KSW was replaced with calcium containingKSW and CRSW was added to the same dose as described above. Hair bundles were measured at time 0 and at 5 min intervals afterwards.To facilitate comparisons among hair bundles, data were normalized to control values. Data points indicate means þ S.E.M. (n = 6). C: DICvideo microscopy of a hair bundle (left) after 30 min exposure to calcium free KSW. Notice that the stereocilia are splayed. After 4 min inCRSW (center), the hair bundle is less splayed. After 10 min in CRSW (right), the hair bundle exhibits morphology comparable to a healthyhair bundle. Scale bar = 1 Wm. D: Bright ¢eld video microscopy of a hair bundle imaged during the repair process. Imaging (a 40U objective)begins immediately after CRSW is added to the preparation (left). Notice that in the ¢rst image (left) the tip of the hair bundle is splayed asevidenced by a pronounced cleft at its tip. As imaging continues (1 min at center panel and 3 min at right panel), the cleft (indicated by thedistance between arrows) progressively shortens. Scale bar = 3 Wm.


10 min (Fig. 2B,C) with little further e¡ects discernablethrough 25 min. Timelapse microscopy of a bundlesplayed mostly at the tip shows that CRSW restoresits normal morphology in a base to tip direction suchthat the cleft separating stereocilia progressively short-ens in a fashion reminiscent of a garment being `zippedup' (Fig. 2D). Since the length of the hair bundle doesnot change (Fig. 2D), the disappearance of the cleftcannot be explained by a reorientation of the bundletip out of the plane of focus.

Electrophysiological recordings from a hair bundlewith obviously splayed stereocilia show that de£ectionsof the bundle (15 psi and 45 ms duration) fail to elicitresponses readily distinguishable from backgroundnoise (Fig. 3A; only the ¢rst two of ¢ve total responses

are shown). After 5 min in CRSW, an obvious negativeshift in DC current is observed with each de£ection ofthe bundle. The amplitude of the negative response in-creases with duration of exposure to CRSW and thensaturates at about 25 min (Fig. 3F).

3.2. Characterization of repair proteins

Fractionation of CRSW by liquid chromatographyon Sepharose 4B columns yields bioactivity in thevoid volume (active fraction K), and in a single otherfraction (active fraction L). Fractionation of K on Se-pharose 2B columns yields bioactivity in three frac-tions: K1,K2, and K3. Active fraction L, the smallestof the active fractions, has a molecular mass of approx-imately 2000 kDa based on the coelution of L and bluedextran (Sigma). SDS-PAGE gels of fraction L show8^10 bands of 90 kDa or less (Fig. 4A). To determinewhether any of the polypeptides in fraction L are con-sumed during the repair process, successive sets of ani-mals having damaged hair bundles were exposed to thesame batch of fraction L for 10 min each. After ¢ve setsof animals (a total of 60 animals), the solution of Lfailed to accomplish repair. The `exhausted' L solutionwas concentrated and examined by SDS-PAGE electro-phoresis. Four bands present in the control L are miss-ing from the exhausted L, apparently having been con-sumed during repair (Fig. 4A). On SDS-PAGE gels, Kfractions exhibit major bands between 45 and 66 kDaand minor bands at less than 45 kDa (Fig. 4B).

Following 1 h exposure to calcium free seawater,fractions K1 and K2 are about 1/7 as e¤cient asCRSW and fraction K3 is 1/15 as e¤cient as CRSWin restoring vibration sensitivity (Fig. 4C). Surprisingly,fraction L is as e¤cient as CRSW in restoring vibrationsensitivity (Fig. 4C). Only fraction L restores vibrationsensitivity in the presence of cycloheximide (Fig. 4D).Moreover, cycloheximide does not a¡ect the e¤ciencyof fraction L (Fig. 4D).

Following 1 min exposure to elastase, both CRSWand fraction L restore vibration sensitivity in approxi-mately 30 min (Fig. 5). However, the recovery con-ferred by CRSW or L is cycloheximide sensitive in elas-tase treated animals. Following elastase treatment,vibration sensitivity is restored within 2.5 h in the ab-sence of exogenously added CRSW (data not shown).

3.3. Cytochemistry and microscopy of fraction L

Biotinylated fraction L binds to damaged hair bun-dles (Fig. 6). Because fraction L was incubated withdamaged hair bundles after ¢xation, only initial bindingof fraction L is revealed by this experiment. Whereasbundles that are severely damaged by exposure to cal-cium free seawater tend to show £uorescence near thebase of the bundle (Fig. 6A), less severely damaged

HEARES 2940 9-1-98

Fig. 3. E¡ects of CRSW on electrical responses of hair bundles tode£ection. A splayed hair bundle was de£ected by pressure ejectedfrom a pu¡er pipette positioned adjacent to the tip of the hair bun-dle. A recording pipette was attached to the apical surface of a cellat the base of the hair bundle using the cell-attached, loose-patchrecording method. The bar depicts the opening and closing of thesolenoid valve of the pneumatic picopump used to deliver pressureto the pu¡er pipette. Current recordings depict e¡ects of de£ectingthe hair bundle (con¢rmed by video microscopy during the experi-ment) (A) before the addition of CRSW; (B) 5 min after the addi-tion of CRSW; (C) 10 min; (D) 15 min; (E) 20 min; (F) 25 min;or (G) 30 min after the addition of CRSW to the preparation (20 Wlper 6 ml KSW).


bundles show £uorescence more distally along theirlength (Fig. 6B,C).

Transmission electron microscopy of negativelystained fraction L shows globular structures measuring22 þ 4 nm in diameter (mean þ standard deviation;n = 50) occurring in clusters (Fig. 7A). In samples offraction L prepared in the presence of elevated levelsof ATP or AMP-PNP, an ATP analog that is nonhy-drolyzable by most ATPases, ¢lamentous structures arerevealed that measure approximately 146 þ 15 nm inlength and 6 þ 1 nm in width (n = 10) (Fig. 7B). The¢lamentous structures commonly appear to bebranched at one or both ends, but are relatively devoidof globular structures. Globular structures measuring20 þ 3 nm in diameter (n = 20) can occur separatefrom the ¢lamentous structures (Fig. 7B). Occasionally,¢lamentous structures are observed in association with

HEARES 2940 9-1-98

Fig. 4. Liquid chromatographic fractions of CRSW and their e¡ects on vibration dependent discharge of nematocysts. A: SDS-PAGE gel offraction L. Whereas the solid line indicates fully functional L, the dashed line depicts `exhausted' L from the same batch. The exhausted L wasobtained by exposing L brie£y to a su¤cient number of animals until bioactivity was lost. Peaks appearing in the control L and not in the ex-hausted L are indicated by asterisks. The gel was transferred to nitrocellulose paper, stained with colloidal gold followed by silver enhancement(see Section 2). The blot was imaged by video camera. B: SDS-PAGE gel of fractions K1, K2, and K3. The gel was directly silver stained andimaged. C: E¡ects of liquid chromatographic fractions of CRSW on the time course of recovery for vibration sensitivity. Animals were exposedto calcium free seawater for 1 h then transferred to calcium containing seawater. After a 10 min recovery from anesthetic e¡ects of calciumfree seawater, a fraction of CRSW was added as follows: (empty triangles) fraction L, (solid triangles) fraction K1, (empty circles) fraction K2,(solid squares) fraction K3. Vibration sensitivity was assayed as described above. D: E¡ects of cycloheximide and fractions of CRSW on thetime course of recovery for vibration sensitivity. Anemones were exposed to calcium free seawater for 1 h (with cycloheximide), transferred tocalcium containing seawater containing 1034 M cycloheximide. After a 10 min recovery from calcium free seawater, the appropriate fraction ofCRSW was added and vibration sensitivity assayed as described above.

Fig. 5. E¡ects of CRSW or fraction L on vibration sensitivity fol-lowing exposure to elastase. Anemones were exposed to 20 units/mlelastase for 1 min (see Section 2), then rinsed three times in freshseawater. At this point, (solid circles) CRSW or (empty circles) frac-tion L was added. Vibration sensitivity was assayed as describedabove.


globular structures similar in size to those describedabove (Fig. 7C).

4. Discussion

4.1. Evidence for repair of anemone hair bundles

It appears that anemone hair bundles recover fromextensive trauma caused by 1 h exposure to calcium freeseawater. This conclusion is based on results from stud-ies of morphology of the hair bundles (Fig. 1C,Fig. 2C,D) from a bioassay performed on intact speci-mens (Fig. 1A) and from electrophysiology (Fig. 3).Recovery is substantially delayed in the presence ofcycloheximide (Fig. 1D), suggesting that newly synthe-sized proteins are involved in the recovery process. Inthe presence of seawater in which recovery has previ-ously occurred for a separate set of animals (repair sea-water), the rate of recovery is shortened by at least16-fold (from 4 h to 15 min; Fig. 1E). Thus, proteinsinvolved in the repair process may be secretory pro-teins. Exogenously added proteins concentrated fromrepair seawater (repair proteins) permit recovery to oc-cur within 8 min (Fig. 2A). If the repair proteins areexposed to trypsin, or if the repair seawater is boiled(Fig. 1E), bioactivity is lost. The recovery conferred byrepair proteins appears to be una¡ected by cyclohexi-

mide (Fig. 2A) suggesting that the repair proteins aloneare su¤cient to accomplish repair.

4.2. Time course for repair

In the presence of CRSW (concentrated repair sea-water), estimates for the time course for recovery rangefrom 8 min based on the bioassay performed on intactspecimens (Fig. 2A) to 25 min based on electrophysiol-ogy (Fig. 3). Microscopic measurements of hair bundlesshow that splayed bundles narrow considerably within10 min of exposure to CRSW, but then may remainslightly larger in diameter than control dimensionsthrough 25 min (Fig. 2B).

Inasmuch as the degree to which bundles are splayedafter exposure to calcium free seawater varies consider-ably (e.g. compare left panels of Fig. 2C,D), we suspectthat the actual time course of recovery varies consider-ably from one bundle to another. Nevertheless, resultsfrom the bioassay are consistent and highly reproduci-ble, indicating that a su¤cient proportion of hair bun-dles are fully repaired or at least adequately repairedwith 8 min exposure to CRSW to permit a normalorganismal response to vibrating stimuli (Fig. 2A). In-deed the electrophysiological data indicate a substantialrecovery occurs within 5 min exposure to CRSW withfull recovery requiring 25 min exposure to CRSW(Fig. 3).

HEARES 2940 9-1-98

Fig. 6. Avidin based cytochemistry of biotinylated fraction L. Paired transmitted and £uorescent microscopic images of damaged hair bundles.In each transmitted image (upper panel), the base (b) and tip (arrowhead) of the bundle is indicated. Scale bar = 3 Wm. The £uorescent avidinpresumably indicating the location of bound, biotinylated L is shown in the lower panel for each of three bundles. A: A severely damaged bun-dle with £uorescence located at the base of the bundle. B: A moderately damaged bundle with £uorescence located midway along the length ofthe bundle at the junction of splayed stereocilia. C: A bundle that appears to have normal morphology with £uorescence located at the tip ofthe bundle. A grid is provided to help indicate the position of the £uorescence in relation to the position of the bundle.


Fraction L, a chromatographic fraction having anestimated mass of 2000 kDa, has comparable bioactiv-ity to the complete repair protein mixture (Fig. 4C,D).Negative staining TEM observations of fraction L re-veal ¢lamentous structures that appear to be associatedwith globular structures (Fig. 7). Although the signi¢-cance of these structures to the repair process is not yetclear, it is interesting to note that the ¢lamentous struc-tures in fraction L have dimensions that are comparableto linkages, including tip links, normally interconnect-ing stereocilia of the hair bundle (Watson et al., 1997).

Furthermore, it appears that the globular structuresmay dissociate from the ¢lamentous structures depend-ing on the local availability of ATP or structural ana-logs of ATP. The e¡ects of ATP on the repair processwill be considered in a separate manuscript.

4.3. Model for mechanism of repair

According to our model, CRSW contains replace-ment linkages of all types (the ¢lamentous structures)as well as accessory and enzymatic components (theassociated globular structures) that participate in at-taching linkages to the appropriate integral proteins inthe stereocilium membrane. Attachment of the linkagesto the target integral proteins would require that thestereocilia are spaced closely together. Thus, in bundleswhere stereocilia of the damaged hair bundle are visiblysplayed at their tips, reattachment or replacement oflinkages must occur in a base to tip direction. Time-lapse microscopy shows narrowing of some splayed hairbundles in a base to tip direction during the repairprocess (Fig. 2D), an observation consistent with thisproposed mechanism of repair. Once the linkage is at-tached to the target integral proteins, the accessory andenzymatic components of the linkage complex woulddissociate from the linkage. Thus, the model predictsthat some, but not all, components of the repair proteinmixture are consumed during repair. We tested themodel by exposing successive sets of animals havingdamaged hair bundles to the same batch of fraction Lfor 10 min each. After 5 sets of animals (a total of60 animals), the solution of L failed to accomplish re-pair within 10 min. Four bands present in the control Lare missing from the exhausted L, apparently havingbeen consumed during repair (Fig. 4A). Presumably,the consumed bands correspond to replacement link-ages and the bands not consumed correspond to acces-sory proteins and enzymes. An alternative explanationwould be that the polypeptides of L that are not con-sumed are unrelated to the repair process. Ongoing ef-forts are aimed at resolving this issue.

4.4. E¡ects of elastase on hair bundles

Elastase is suspected of digesting tip links in hairbundles of guinea pigs (Preyer et al., 1995). In thepresent study, we found that elastase causes a tempo-rary loss of vibration sensitivity in sea anemones. Inter-estingly, CRSW or fraction L (obtained from animalsexposed to calcium free seawater) speeds the rate ofrecovery by 5-fold (Fig. 5). On the other hand, therecovery conferred by the repair proteins remains cyclo-heximide sensitive in elastase treated animals indicatingthat other proteins are required for recovery. It shouldbe interesting to identify the additional proteins re-quired following exposure to elastase.

HEARES 2940 9-1-98

Fig. 7. Negative stained fraction L. A: In 10312 M ATP, globularstructures are observed in clusters. Scale bar = 50 nm. B: In 1033 MAMP-PNP, ¢lamentous structures are observed. Scale bar = 50 nm.C: Filamentous structures are observed in an apparent associationwith globular structures in preparations containing either mM ATPor AMP-PNP. Scale bar = 50 nm.


The present report provides evidence that the mostprimitive animals that possess hair bundle mechanore-ceptors utilize a remarkable mechanism to repair thesemechanoreceptors. Such an e¤cient and well developedrepair mechanism may have evolved and/or been con-served in anemones because their hair bundles are ex-posed to the external environment (seawater) and thusare more vulnerable to damage than vertebrate hairbundles. The ¢nding of an inducible repair mechanismin anemones may make it feasible to genetically char-acterize the mechanotransduction apparatus of hairbundle mechanoreceptors.

Acknowledgments

We gratefully acknowledge support from NIHR01GM52234 and NSF MCB9505844.

References

Ashmore, J.F., 1991. The electrophysiology of hair cells. Annu. Rev.Physiol. 53, 465^476.

Assad, J.A., Shepherd, G.M.G., Corey, D.P., 1991. Tip link integrityand mechanotransduction in vertebrate hair cells. Neuron 7, 985^994.

Barnes, R.D., 1980. Invertebrate Zoology, 4th edn. Saunders College/Holt, Rinehart and Winston, Philadelphia, PA, 1089 pp.

Bayer, E.A., Wilchek, M., 1990. Protein biotinylation. Methods En-zymol. 184, 138^160.

Bollag, D.M., Edelstein, S.J., 1991. Protein Methods. Wiley-Liss, NewYork, 230 pp.

Howard, J., Roberts, W.M., Hudspeth, A.J., 1988. Mechanoelectricaltransduction by hair cells. Annu. Rev. Biophys. Biophys. Chem.17, 99^124.

Hudspeth, A.J., 1985. The cellular basis of hearing: the biophysics ofhair cells. Science 230, 745^752.

Hudspeth, A.J., 1992. Hair-bundle mechanics and a model for mecha-noelectrical transduction by hair cells. In: Corey, D.P., Roper,S.D. (Eds.), Sensory Transduction, Rockefeller University Press,New York, pp. 357^370.

Hudspeth, A.J., Corey, D.P., 1977. Sensitivity, polarity, and conduc-tance change in the response of vertebrate hair cells to controlledmechanical stimuli. Proc. Natl. Acad. Sci. USA 74, 2407^2411.

Kroese, A.B.A., Das, A., Hudspeth, A.J., 1989. Blockage of the trans-

duction channels of hair cells in the bullfrog's sacculus by amino-glycoside antibiotics. Hear. Res. 37, 203^218.

Mariscal, R.N., 1974. Nematocysts. In: Muscatine, L., Lenho¡, H.M.(Eds.), Coelenterate Biology: Reviews and New Perspectives. Aca-demic Press, New York, pp. 129^178.

Minasian, L.L., Jr., Mariscal, R.N., 1979. Characteristics and regu-lation of ¢ssion activity in clonal cultures of the cosmopolitan seaanemone, Haliplanella luciae (Verrill). Biol. Bull. 157, 478^493.

Mire, P., Watson, G.M., 1997. Mechanotransduction of hair bundlesarising from multicellular complexes in anemones. Hear. Res. (inpress).

Mire-Thibodeaux, P., Watson, G.M., 1994. Cyclical morphodynamicsof hair bundles in sea anemones: Second messenger pathways.J. Exp. Zool. 270, 517^526.

Osborne, M.P., Comis, S.D., Pickles, J.O., 1988. Further observationson the ¢ne structure of tip links between stereocilia of the guineapig cochlea. Hear. Res. 35, 99^108.

Pickles, J.O., 1993. A model for the mechanics of the stereociliarbundle on acousticolateral hair cells. Hear. Res. 68, 159^172.

Pickles, J.O., Comis, S.D., Osborne, M.P., 1984. Cross-links betweenstereocilia in the guinea pig organ of Corti, and their possiblerelation to sensory transduction. Hear. Res. 15, 103^112.

Pollard, T.D., 1982. Assays for myosin. Methods Enzymol. 85,123^130.

Preyer, S., Hemmert, W., Zenner, H.P., Gummer, A.W., 1995. Abo-lition of the receptor potential response of isolated mammalianouter hair cells by hair-bundle treatment with elastase: a test ofthe tip-link hypothesis. Hear. Res. 89, 187^193.

Shick, J.M., 1991. A Functional Biology of Sea Anemones. Chapmanand Hall, New York, 395 pp.

Shotwell, S.L., Jacobs, R., Hudspeth, A.J., 1981. Directional sensitiv-ity of individual vertebrate hair cells to controlled de£ection oftheir hair bundles. Ann. NY Acad. Sci. 374, 1^10.

Watson, G.M., Hessinger, D.A., 1991. Chemoreceptor-mediated elon-gation of stereocilium bundles tunes vibration-sensitive mechano-receptors on cnidocyte-supporting cell complexes to lower frequen-cies. J. Cell Sci. 99, 307^316.

Watson, G.M., Mariscal, R.N., 1983. Comparative ultrastructure ofcatch tentacles and feeding tentacles of the sea anemone, Halipla-nella. Tissue Cell 15, 939^953.

Watson, G.M., Mire-Thibodeaux, P., 1994. The cell biology of nem-atocysts. Int. Rev. Cytol. 156, 275^300.

Watson, G.M., Mire, P., Hudson, R.R., 1997. Hair bundles of seaanemones as a model system for vertebrate hair bundles. Hear.Res. 107, 53^66.

Watson, G.M., Roberts, J., 1995. Chemoreceptor-mediated polymer-ization and depolymerization of actin in hair bundles of sea ane-mones. Cell Motil. Cytoskel. 30, 208^220.

Zhao, Y., Yamoah, E.N., Gillespie, P.G., 1996. Regeneration of brok-en tip links and restoration of mechanical transduction in haircells. Proc. Natl. Acad. Sci. USA 94, 15469^15474.

HEARES 2940 9-1-98


Repair of hair bundles in sea anemones by secreted proteins

Documents

Transcript of Repair of hair bundles in sea anemones by secreted proteins