-Actin is required for cytoskeletal maintenance but not development

␥-Actin is required for cytoskeletal maintenance but not development Inna A. Belyantsevaa,1, Benjamin J. Perrinb,1, Kevin J. Sonnemannb,1, Mei Zhuc, Ruben Stepanyand, JoAnn McGeee, Gregory I. Frolenkovd,f, Edward J. Walshe, Karen H. Fridericic, Thomas B. Friedmana, and James M. Ervastib,2 aLaboratory

of Molecular Genetics, National Institute on Deafness and Other Communication Disorders/National Institutes of Health, Rockville, MD 20850; of Biochemistry, Molecular Biology and Biophysics, University of Minnesota, Minneapolis, MN 55455; cMicrobiology and Molecular Genetics, Michigan State University, East Lansing, MI 48824; dDepartment of Physiology, University of Kentucky, Lexington, KY 40536; fMolecular Biology and Genetics Section, National Institute on Deafness and Other Communication Disorders/National Institutes of Health, Rockville, MD 20850; and eDevelopmental Auditory Physiology Laboratory, Boys Town National Research Hospital, Omaha, NE 68131 bDepartment

␤cyto-Actin and ␥cyto-actin are ubiquitous proteins thought to be essential building blocks of the cytoskeleton in all non-muscle cells. Despite this widely held supposition, we show that ␥cyto-actin null mice (Actg1ⴚ/ⴚ) are viable. However, they suffer increased mortality and show progressive hearing loss during adulthood despite compensatory up-regulation of ␤cyto-actin. The surprising viability and normal hearing of young Actg1ⴚ/ⴚ mice means that ␤cyto-actin can likely build all essential non-muscle actin-based cytoskeletal structures including mechanosensory stereocilia of hair cells that are necessary for hearing. Although ␥cyto-actin– deficient stereocilia form normally, we found that they cannot maintain the integrity of the stereocilia actin core. In the wild-type, ␥cyto-actin localizes along the length of stereocilia but re-distributes to sites of F-actin core disruptions resulting from animal exposure to damaging noise. In Actg1ⴚ/ⴚ stereocilia similar disruptions are observed even without noise exposure. We conclude that ␥cytoactin is required for reinforcement and long-term stability of F-actin– based structures but is not an essential building block of the developing cytoskeleton. actin 兩 cytoskeleton 兩 hearing

T

here are six genes encoding six vertebrate actins that are classified according to where they are predominately expressed. ␣skeletal-Actin, ␣smooth-actin, ␣cardiac-actin, and ␥smoothactin are primarily found in muscle cells, whereas cytoplasmic ␤cyto-actin and ␥cyto-actin are ubiquitously and highly expressed in non-muscle cells, as reviewed elsewhere (1). Athough ␤cytoactin and ␥cyto-actin differ at only four biochemically similar amino acid residues in their N-termini, several lines of evidence suggest that each protein is functionally distinct. The amino acid sequences of ␤cyto- and ␥cyto-actin are each exactly conserved among avian and mammalian species. In addition, ␤cyto- and ␥cyto-actin proteins are differentially localized (2–5) and posttranslationally modified (6). Finally, although dominant missense mutations in ACTB encoding ␤cyto-actin are associated with syndromic phenotypes including severe developmental malformations and bilateral deafness (7), humans carrying a variety of dominant missense mutations in ACTG1 develop postlingual nonsyndromic progressive hearing loss (DFNA20, OMIM 604717) (8–11). ␥cyto-Actin is widely expressed in the inner ear sensory epithelial cells on which mammalian hearing depends. These cells are organized in rows along the cochlea length: one row of inner hair cells (IHCs) and three rows of outer hair cells (OHCs) (Fig. 2A). IHCs function as auditory receptors, converting sound energy into neuronal signals, whereas OHCs enhance sensitivity to sound stimuli, as reviewed elsewhere (12). The apical surface of a hair cell is topped with actin-rich microvilli-derived protrusions termed stereocilia, which deflect in response to sound stimuli, initiating mechanoelectrical transduction (Fig. 2B). ␤cyto- and ␥cyto-Actin are both thought to be essential components of the stereocilia core (2–4), which consists of a paracryswww.pnas.org兾cgi兾doi兾10.1073兾pnas.0900221106

talline array of unidirectionally oriented actin filaments (Fig. 2C) (13–15). In the mammalian organ of Corti, the precise architecture of stereocilia is preserved for the life of the organism. Meanwhile, the stereocilia actin core is reported to undergo renewal by continuous actin polymerization at filament barbed ends and depolymerization at pointed ends, which is precisely coupled to maintain stereocilia length (15, 16). The speed of stereocilia treadmilling is reported to be the same for all stereocilia of the same row and is proportional to stereocilia length (17). Immuno-electron microscopy shows that in wild-type hair cells ␤cyto-actin is largely restricted to stereocilia, their rootlets, and the cuticular plate (2, 3, 18), whereas ␥cyto-actin is reported to have more broad localization, including hair cell stereocilia and their rootlets, the cuticular plate in which stereocilia are anchored, adherens junctions, and outer hair cell lateral walls (2, 3, 18). Hair cells and their stereocilia are thus an attractive model to study the structural consequences of perturbing actin isoform composition. ␤cyto- and ␥cyto-Actin are among the most abundant proteins in every mammalian cell, leading to the common assumption that both cytoplasmic actins are essential for function and viability. To test this supposition and to uncover the unique function of ␥cyto-actin, we generated a whole-body ␥cyto-actin knockout mouse (Actg1⫺/⫺). We show here that mice completely lacking ␥cyto-actin can survive to adulthood. Interestingly, Actg1⫺/⫺ mice initially have normal hearing but develop progressive hearing loss during adulthood that is characterized by stereocilia actin core disruptions and stereocilia degradation. These findings led us to conclude that ␥cyto-actin is not necessary for the formation of actin-based structures required for organogenesis and development, but is essential for maintenance of the hair cell actin cytoskeleton. Results ␥cyto-Actin Null Mice Are Viable. To determine whether there is a unique function of ␥cyto-actin that cannot be compensated by the other actin family members, we generated a ␥cyto-actin null (Actg1⫺/⫺) mouse. Mice entirely devoid of ␥cyto-actin were viable, but born at one-third of the expected Mendelian ratio, indicating that the absence of ␥cyto-actin caused some embryonic or perinatal lethality. Although the overall development of

Author contributions: I.A.B., B.J.P., K.J.S., R.S., J.M., G.I.F., E.J.W., K.H.F., T.B.F., and J.M.E. designed research; I.A.B., B.J.P., K.J.S., M.Z., R.S., J.M., and G.I.F. performed research; I.A.B., B.J.P., K.J.S., M.Z., R.S., J.M., G.I.F., E.J.W., K.H.F., T.B.F., and J.M.E. analyzed data; and I.A.B., B.J.P., K.J.S., G.I.F., K.H.F., T.B.F., and J.M.E. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. 1I.A.B., 2To

B.J.P., and K.J.S. contributed equally to this work.

whom correspondence should be addressed. E-mail: [email protected].

This article contains supporting information online at www.pnas.org/cgi/content/full/ 0900221106/DCSupplemental.

PNAS 兩 June 16, 2009 兩 vol. 106 兩 no. 24 兩 9703–9708

CELL BIOLOGY

Edited by Carl Frieden, Washington University School of Medicine, St. Louis, MO, and approved April 24, 2009 (received for review January 8, 2009)

␥cyto-Actin Null Mice Show Progressive Loss of Hearing. We assessed hearing in wild-type and Actg1⫺/⫺ mice by measuring auditory brainstem response (ABR) thresholds. ABR objectively measures synchronous electrical activity generated by the neurons in the ascending auditory system and can be recorded from scalp electrodes by averaging responses to short tone bursts (19, 20). We found that Actg1⫺/⫺ mice up to 6 weeks of age had near-normal ABR thresholds (Fig. 1D). However, 16-week-old Actg1⫺/⫺ mice demonstrated a marked hearing impairment at each frequency tested, and by 24 weeks of age were profoundly deaf (Fig. 1D). This progressive hearing loss was not found in Actg1⫹/⫺ littermates, which exhibited wild-type ABR thresholds up to 24 weeks of age (Fig. S2) despite expressing only 50% of wild-type levels of ␥cyto-actin (Fig. 1C).

40

A Mass (grams)

35 30 25 20

Actg1 +/+ Actg1 +/Actg1 -/-

15 10 0 0

50

100

150

200

250

300

Age (days)

B

100

Actg1 +/-

Survival (%)

80 60

Differential Localization of ␤cyto- and ␥cyto-Actin in Developing and Adult Mouse Hair Cells Revealed Delayed Appearance of ␥cyto-Actin in Stereocilia. Consistent with previous reports in postnatal chicken

40

Actg1 -/-

20 0

0

100

200

300

50 0

in ct

cy to a

ct

ct β

cy to a

γ

D

in

in

-

Ac tg 1+ /+ Ac tg 1+ /Ac tg 1 -/

α-tubulin

100

la

total actin

Actg1 +/+ Actg1 +/Actg1 -/-

150

ta

βcyto-actin

200

to

γcyto-actin

Protein level (% of wild-type)

Age (days)

C

Level (dB SPL)

120 100

A ctg1 +/+ 6 weeks A ctg1 +/+ 16 weeks A ctg1 +/+ 24 weeks A ctg1 - /- 6 weeks A ctg1 - /- 16 weeks A ctg1 - /- 24 weeks

80 60 40 20 0 4

6

10

16

25

Frequency (kHz)

Fig. 1. Characterization of live-born homozygous mutant Actg1⫺/⫺ mice. (A) Body mass growth curve of Actg1⫹/⫹ (wild-type, closed squares), Actg1⫹/⫺ (heterozygous, open circles) and Actg1⫺/⫺ (homozygous mutant, open triangles) mice from P28 until P300 (n ⫽ 12 Actg1⫹/⫹, 18 Actg1⫹/⫺, 11 Actg1⫺/⫺, mean ⫾ SEM). (B) Kaplan-Meier survival curve of Actg1⫹/⫺ and Actg1⫺/⫺ mice from P0 to P350, (n ⫽ 31 for each genotype). (C) Representative immunoblots of SDS extracts from Actg1⫹/⫹, Actg1⫹/⫺ and Actg1⫺/⫺ cochlear extracts probed with antibodies specific for ␥cyto-actin, ␤cyto-actin, pan-actin, or tubulin antibody. Protein levels were quantified and are expressed relative to the wild-type level (mean ⫾ SEM). (D) Actg1⫺/⫺ mice develop progressive hearing loss. Auditory brainstem response (ABR) thresholds were determined for Actg1⫹/⫹ and Actg1⫺/⫺ mice at 6, 16, and 24 weeks of age using stimulus frequencies from 4 to 22 kHz, presented at half-octave steps (n ⬎ 5, mean ⫾ SEM).

surviving Actg1⫺/⫺ mice appeared largely normal, their body weight was ⬇20% lower than wild-type (Actg1⫹/⫹) and heterozygous (Actg1⫹/⫺) littermates (Fig. 1A). In addition, some Actg1⫺/⫺ mice died prematurely of unknown cause(s) (Fig. 1B). To investigate whether the observed effects were caused by a general depletion of cellular actin, we analyzed actin isoform expression in wild-type, Actg1⫹/⫺ and Actg1⫺/⫺ tissues by Western blot. We observed gene dose-dependent expression of ␥cyto-actin and compensatory up-regulation of other actin family members to maintain the total actin level in all tissues examined (Fig. 1C and [supporting information (SI) Fig. S1]). Therefore, the actin composition, but not the concentration, was altered in Actg1⫺/⫺ mice. 9704 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0900221106

and mature guinea pig or rat, both ␤cyto- and ␥cyto-actin were detected in stereocilia (Fig. 2 D and E) and the cuticular plate of adult wild-type mouse hair cells. The three independently generated ␥cyto-actin-specific antibodies used did not stain any structures in Actg1⫺/⫺ hair cells (Fig. 2F), demonstrating the specificity of these antisera for ␥cyto-actin. We found that during embryonic development of wild-type mice, ␤cyto-actin appeared in the body of hair cells and subsequently in stereocilia earlier than ␥cyto-actin, which accumulated first in supporting cells and only later appeared in hair cells (Fig. 2 G–P). We observed ␤cyto-actin in auditory hair cell stereocilia as soon as they appear around E16.5 (Fig. 2 G–I) in the basal turn of the cochlea. The first appearance of ␥cyto-actin within stereocilia was detected after stereocilia emerged at approximately E18.5 (Fig. 2 O–P). These data are consistent with ␤cyto-actin primarily contributing to the formation of the actin cytoskeleton of developing stereocilia, whereas ␥cyto-actin may be important for cytoskeleton maintenance and/or reinforcement. Although both actins are found in mouse stereocilia, we observed differential localization within the stereocilia, again consistent with ␤cyto- and ␥cyto-actin having disparate functions. In the adult wild-type mouse stereocilia, ␤cyto-actin staining overlapped completely with rhodamine-phalloidin staining, whereas ␥cyto-actin was concentrated more toward the periphery of the stereocilia actin core, often only partially overlapping with rhodamine-phalloidin staining (Fig. 2 Q–T). Phalloidin-Negative Gaps in F-Actin Stereocilia Cores Contain Core Components. In the course of characterizing the localization of

the cytoplasmic actins, we observed occasional gaps in phalloidin staining of F-actin cores of vestibular hair cell stereocilia (Fig. 3 A–C). The gaps were most frequently observed at the base and along the length of stereocilia in the tallest row (Fig. 3D). Using our ␥cyto-actin specific antibodies, which recognize both globular (G) and filamentous (F) forms of actin (see SI Text), we found that gaps were enriched in ␥cyto-actin. This actin population is likely to be predominantly monomeric, because phalloidin recognizes only filamentous actin (Fig. 3 A–D). Usually gap staining was much more intense relative to that along the length of stereocilium (Figs. 3 A–D and 3 F–M), which may be caused by enhanced antibody accessibility within the gaps. Alternatively, intense gap staining could be partially explained by the redistribution of ␥cyto-actin to F-actin gaps from a pool of available non-filamentous actin within a stereocilium. A similar redistribution to F-actin gaps was also seen for ␤cyto-actin (Fig. S3). It is likely that ␤cyto-actin is also recruited to the gaps from a pool of non-filamentous actin, as staining intensity along a stereocilium with a gap was not different from the intensity of staining along a stereocilium without gaps (Fig. S3B). The same pattern of staining was also observed for DNase I (Fig. 3E), a marker for Belyantseva et al.

B

D

G

C

F

E

H

I

J

K

L

M

N

O

P

Q

R

S

T

Fig. 2. Differential localization of ␤cyto- and ␥cyto-actin in the mouse organ of Corti (OC). (A) The OC has three rows of outer hair cells (OHCs) and one row of inner hair cells (IHCs). Each hair cell is surrounded by non-sensory supporting cells. (B) Scanning electron microscopy images of OHC and IHC stereocilia bundles. (C) Stereocilium core consists of tightly packed unidirectional actin filaments (F-actin). In (D–T), rhodamine-phalloidin highlights F-actin (red), and actin stained by antibodies (green). Isoform-specific antibodies detect ␤cyto-actin (D) and ␥cyto-actin (E) along the length of adult wild-type (wt) OHC and IHC stereocilia. (F) Absence of ␥cyto-actin (green) in 6-week-old Actg1⫺/⫺ OC. (G–L) At E16.5, ␤cyto-actin immunoreactivity follows rhodamine-phalloidin labeling in wt hair cells (G–I), whereas ␥cyto-actin is detected in supporting cells but not in hair cells (J–L). (M–P) At E18.5, ␤cyto-actin is present in all stereocilia of hair cells throughout the cochlea (M, N), whereas ␥cyto-actin begins to appear only in stereocilia of more developed basal turn of the cochlea (O, P). (Q–T) ␤cyto-Actin immunoreactivity (Q, R) overlaps with rhodamine-phalloidin staining, whereas ␥cyto-actin (S, T) is concentrated toward the periphery of the IHC stereocilia F-actin core in adult wt mice. Scale bars (B, Q–T), 2 ␮m; scale bars (D–P), 5 ␮m.

monomeric actin (G-actin) (21), and espin (Fig. 3F), an actin bundling protein essential for stereocilia formation, which is reported to have both F- and G-actin binding sites (22). Interestingly, only proteins that are either actin core components or directly involved in actin turnover were found to accumulate in the phalloidin-negative gaps. For example, actin-associated proteins cadherin-23, protocadherin-15-CD1, myosin-VIIa, and myosin-XVa are not present in gaps (Fig. S4 and data not shown). In contrast, cofilin, which was implicated in both severing and nucleation of F-actin (23), selectively accumulates in stereocilia gaps (Fig. S4). Together, these data suggest that (i) gaps have a different structural arrangement than the stereocilia actin core, (ii) gaps are enriched for ␤cyto- and ␥cyto-actin along with other core components, and (iii) cofilin may mediate ongoing actin remodeling in the gap to facilitate repair of local damage of the F-actin core. Belyantseva et al.

Fig. 3. ␥cyto-Actin concentrates at the sites of stereocilia core disruptions. (A–C) ␥cyto-Actin antibody highlights gaps (segments of F-actin depolymerization; arrows) in wild-type (wt) mouse vestibular hair cell (VHC) stereocilia. In all panels, rhodamine-phalloidin highlights F-actin in red, and labeling with antibodies is in green. (D) ␥cyto-Actin at the base and within the F-actin gaps of longest stereocilia in wt mouse VHC (arrows). (E) DNase I stains globular actin within F-actin gaps of VHC stereocilia (arrows). (F) Espin concentrates in gaps of wt VHC stereocilia. (G) Uniform distribution of ␥cyto-actin along adult guinea pig IHC stereocilia not exposed to damaging noise. (H) Redistribution of ␥cyto-actin in noise-damaged guinea pig IHC stereocilia. ␥cyto-Actin absent from the tips and evenly distributed along stereocilia which appear unaffected (inset: second and fifth stereocilium from the left). (I–K) ␥cyto-Actin concentrates at sites of F-actin damage (arrows) and at tips of shortened stereocilia (asterisks, inset in H) in a noise-damaged bundle from (H). (L) The F-actin gaps in IHC stereocilia from Actg1⫺/⫺ mouse (arrows). (M) ␤cyto-Actin concentrates in the F-actin gap of Actg1⫺/⫺ IHC stereocilium. ␤cyto-Actin staining along stereocilia is barely visible because of intense gap staining. (N–P) Espin concentrates in gaps of Actg1⫺/⫺ VHC stereocilia. Scale bars, 2 ␮m.

␥cyto-Actin Localizes to Phalloidin-Negative Gaps That Form in Response to Damage. In contrast to vestibular hair cell stereocilia,

gaps were not observed in undamaged auditory hair cell stereocilia of wild-type mice. Previously, F-actin gaps were reported in guinea pig auditory hair cell stereocilia after noise damage (24), suggesting that gaps develop in response to stereocilia damage. To assess whether damage-induced gaps are also enPNAS 兩 June 16, 2009 兩 vol. 106 兩 no. 24 兩 9705

CELL BIOLOGY

A

6 weeks old

BB

C

D

Actg1+/+

Actg1-/-

16 weeks old

Immunofluorescence Confocal Microscopy and Scanning Electron Microscopy Analyses Reveal an Unexpected Pattern of Degeneration of Actg1ⴚ/ⴚ Stereocilia. Interestingly, F-actin gaps were occasionally

C E

DF

G

H

Actg1+/+

Actg1-/-

J

I 120 100

Number of stereocilia per hair cell

80

*

60 40

K

20

Actg1-/-

Actg1+/+

0

Actg1+/+

observed in auditory hair cell stereocilia of Actg1⫺/⫺ mice without exposure to damaging noise. These gaps lacked ␥cytoactin (Fig. 3L) but contained ␤cyto-actin (Fig. 3M) and espin (Fig. 3 N–P). The staining of ␤cyto-actin within gaps of Actg⫺/⫺ stereocilia was so intense that we had to turn down the gain on the confocal microscope so that the signal within gaps was not saturated (Fig. 3M and Fig. S3A). As a consequence, the ␤cyto-actin signal along the lengths of stereocilia was reduced to a barely detectable level (compare Fig. 3M and Fig. S3A with Fig. S3B). The presence of F-actin gaps in the auditory hair cell stereocilia of hearing-impaired, ␥cyto-actin–deficient mice led us to investigate whether a stereocilia maintenance/repair mechanism is defective in Actg1⫺/⫺ mice. To define the structural changes associated with a ␥cyto-actin-deficiency, we characterized the morphology of auditory hair cells. We examined outer hair cells of 6-week-old Actg1⫺/⫺ mice by scanning electron microscopy and found that ␥cyto-actin deficient stereocilia were indistinguishable from OHC stereocilia of wild-type littermates (Fig. 4 A–D). However, Actg1⫺/⫺ stereocilia deteriorated progressively as the animals aged (Fig. 4 E–H). By 16 weeks of age, ⬇50% of stereocilia within a hair bundle were degraded or absent in Actg1⫺/⫺ mice (Fig. 4I). Across all three rows within the hair bundle, we observed missing or shortened (Fig. 4 J and K) stereocilia, although the remaining stereocilia looked normal.

A

Actg1-/-

riched in ␥cyto-actin, we compared ␥cyto-actin localization in stereocilia from control and noise-damaged guinea pigs. Consistent with previous studies (2, 18), ␥cyto-actin was distributed along the length of control stereocilia (Fig. 3G). After a damaging noise exposure, ␥cyto-actin was enriched in both the tips and in phalloidin-negative gaps observed along the length of noise-damaged stereocilia, which were often abnormally shorter than neighboring normal appearing stereocilia of the same row (Fig. 3 H–K). In the same bundle, some stereocilia that appeared unaffected by noise still had ␥cyto-actin evenly distributed along their length and did not have an accumulation of ␥cyto-actin at the tips (Fig. 3H, inset).

6 weeks 16 weeks

Discussion Actg1⫺/⫺ mice survive to birth and beyond, demonstrating that ␥cyto-actin is not strictly required for mammalian development or viability. Our observations of stereocilia from Actg1⫺/⫺ mice instead indicate that ␥cyto-actin is necessary to maintain cytoskeletal integrity and function. The phenotype of Actg1⫺/⫺ stereocilia is unique; we observed stereocilia defects ranging from simple shortening to complete loss of individual stereocilia across all three rows within the hair bundle indicating selected stereocilia disassembly (Fig. 4 E–K), whereas the remaining stereocilia within the same bundle appeared intact. The apparent independent nature of this phenomenon (affected stereocilia surrounded by normal stereocilia) indicates that the disassembly process is initiated and/or regulated at the level of an individual stereocilium. Therefore, this disassembly process appears different from actin treadmilling that normally occurs simultaneously in all stereocilia of the same row (17). Rather, our results suggest that ␥cyto-actin strengthens stereocilia F-actin cores, preventing stereocilia core damage, and/or is required to repair the damaged core. Consistent with this view, in developing wild-type mice ␥cyto-actin appeared and accumulated in stereocilia a few days before the onset of hearing function, perhaps preparing stereocilia to withstand the rigors of acoustical stimulation. Furthermore, damaging noise induces the appearance of ␥cyto-actin–enriched, phalloidin-negative gaps in the stereocilia core of wild-type rodent hair cells. These gaps 9706 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0900221106

Fig. 4. Morphology of stereocilia bundles in adult wild-type (Actg⫹/⫹) and ␥cyto-actin deficient (Actg1⫺/⫺) mice. (A–D) Scanning electron micrographs of stereocilia from (A, B) 6-week-old Actg1⫹/⫹ and (C, D) 6-week-old Actg1⫺/⫺ mice. (E–H) Scanning electron microscopy images of OHC stereocilia from 16-week-old Actg1⫹/⫹ (E, F) and 16-week-old Actg1⫺/⫺ mice (G, H). There is a loss of individual stereocilia from all three rows of OHC hair bundle from Actg1⫺/⫺ mice. Images are from the middle turn of the cochlea. (I) Box and whisker plot (whiskers, maximum and minimum; box, 5th–95th percentile; line, mean) of the number of individual stereocilia in individual OHC bundles from Actg1⫹/⫹ or Actg1⫺/⫺ mice at 6 and 16 weeks of age, *P ⬍ 0.005. (J–K) enlargements of image in (H) with arrows indicating missing and shortened stereocilia. Scale bars (A, C, E, and G), 5 ␮m; scale bars (B, D, F, H, J, and K), 1 ␮m.

were not present in control stereocilia but were observed in Actg1⫺/⫺ mouse auditory stereocilia indicating that stereocilia core damage was more frequent or more slowly repaired in the absence of ␥cyto-actin. Finally, Actg1⫺/⫺ stereocilia progressively deteriorated, demonstrating that ␥cyto-actin is required to maintain these structures. ␥cyto- and ␤cyto-Actin are nearly identical, featuring only four biochemically similar residue substitutions in the N terminus, suggesting likely compensation between these proteins. Indeed, ␤cyto-actin protein levels were elevated in Actg1⫺/⫺ mice and the total actin level was equivalent in Actg1⫺/⫺ and wild-type mice Belyantseva et al.

Belyantseva et al.

body knockout was widely presumed to be unviable. This murine model exhibited normal muscle development followed by progressive myopathy and muscle cell necrosis (29). Both muscle cells and outer hair cells must resist force generated by contractility or electromotility, respectively. Although otherwise clearly disparate in structure and function, these mechanically challenged cells seem to have a particularly evident requirement for the specialized properties of ␥cyto-actin necessary for cellular maintenance. Additional ␥cyto-actin deficiency-based cytoskeletal pathologic conditions may exist in other organs and tissues of Actg1⫺/⫺ mice that could affect long-term function. Indeed, the lower body mass of Actg1⫺/⫺ mice and their occasional premature death suggests a hidden, slowly developing or partially compensated pathologic condition. We conclude that ␥cyto-actin is not necessary to build actin cytoskeletal structures required for organogenesis and development but, instead, functions primarily to reinforce and/or repair the actin cytoskeleton. Methods Generation of Actg1-Null Mice. A targeting vector in which loxP sites flank exons 2 and 3 of the murine Actg1 gene was described previously (29). Embryonic stem cell targeting, screening, blastocyst injections, and subsequent EIIa-cre breeding were performed to generate Actg1⫹/⫺ mice (29). Actg1⫹/⫺ mice were backcrossed to C57BI/6 for 10 generations before N10 Actg1⫹/⫺ X Actg1⫹/⫺ breedings were arranged to obtain Actg1⫺/⫺ mice. All genotypes were determined as previously described (29). Animals were housed and treated in accordance with the standards set by the University of Minnesota Institutional Animal Care and Use Committee. Immunoblot Analysis. Brain, lung, kidney, and cochlea were dissected from mice of the indicated genotypes, frozen in liquid nitrogen, ground into powder, boiled in buffer (1% SDS, EGTA, PMSF, benzamidine, leupeptin), and centrifuged to remove insoluble material. Protein concentration in the resulting lysate was determined with either a colorimetric assay (DC assay, BioRad) or by A280 measurement. Equal amounts of protein were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE), transferred to nitrocellulose membranes, and probed with the indicated antibodies (␥cytoactin; mAb 2– 4 or pAb7577 (29); ␤cyto-actin, AC15 (Sigma); pan-actin, C4, gift of J. Lessard, University of Cincinnati; ␥smooth-actin, B4 (J. Lessard, University of Cincinnati); ␣smooth-actin, 1A4 (Sigma); ␣-tubulin B512 (Sigma). Fluorescently labeled secondary antibodies were detected and quantified from three separate experiments blotted in triplicate using an Odyssey infrared scanner and software (Li-Cor Biosciences). Auditory Brainstem Responses. ABRs were collected as previously described (30) or using a Tucker-Davis Technologies System3 to generate sound stimuli and to amplify and record brainstem potentials as described in the SI Text. Antibody Validation and Immunostaining. Polyclonal antibody pAb7577 against cytoplasmic ␥cyto-actin (ACTG1) was generated in the laboratory of J. Ervasti as described, and the specificity was verified (29). The second anti-␥cytoactin polyclonal antibody was a gift from C. Bulinski and was characterized previously (31). The third anti-␥cyto-actin polyclonal antibody was developed in the laboratory of K. Friderici by immunizing rabbits with a peptide of the N-terminal 15 residues (Princeton Biomolecules) and affinity purified as described in the SI Text. The immunostaining is described in the SI Text. All animal care and experimental procedures were approved by the NINDS/NIDCD ACUC. Animals and Noise Exposure. Noise exposure methods are described in the SI Text. Scanning Electron Microscopy. Cochlea were rapidly dissected and fixed by perfusing 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer pH 7.3 with 1 mM CaCl2, through the round window, followed by immersion in the same solution for 2 hours. After microdissection to reveal hair cell stereocilia, cochlea were incubated in 2% arginine-HCl, glycine, glutamic acid, and sucrose followed by treatment with 2% tannic acid and 2% guanidine-HCl and were postfixed in 1% aqueous osmium tetroxide. Specimens were dehydrated in ethanol, critical point dried, sputter coated, and imaged using a cold field emission gun scanning electron microscope (Hitachi S-4700). PNAS 兩 June 16, 2009 兩 vol. 106 兩 no. 24 兩 9707

CELL BIOLOGY

(Fig. 1C). However, ␥cyto-actin–deficient stereocilia progressively deteriorated despite the localization of ␤cyto-actin to gaps in the F-actin core of Actg1⫺/⫺ mouse auditory stereocilia (Fig. 3M and Fig. S3). This surprising result indicates that ␥cyto-actin has at least some functions that are unique and cannot be compensated for by ␤cyto-actin. One possibility is that ␥cyto-actin brings to the site of damage a unique ␥cyto-actin protein partner that is necessary for ␤cyto-actin, ␥cyto-actin, or for ␤cyto- and ␥cyto-actin together, to repair damage to the core. Consistent with different functions of ␥cyto- and ␤cyto-actin, we observed distinct localization patterns for each protein within wild-type auditory stereocilia. ␤cyto-Actin localized to stereocilia cores, exactly overlaying with phalloidin staining, whereas ␥cyto-actin was concentrated toward the periphery of stereocilia cores. Because the ␥cyto-actin antibody detects both monomeric and filamentous actin whereas phalloidin detects only filamentous actin, there appears to be a pool of monomeric ␥cyto-actin at the periphery of stereocilia cores. Alternatively, phalloidin-negative actin may still be filamentous but unable to bind phalloidin because of a particular F-actin conformation, which was observed in nuclear actin as previously reviewed (25), different paracrystal filament packing that excludes phalloidin, or masking by actin binding proteins. In any case, the peripheral population of ␥cyto-actin is distinct and may be used for stereocilia core remodeling and repair, perhaps redistributing to F-actin gaps that form as a result of stereocilia core damage. Based on ␥cyto-actin localization and Actg1⫺/⫺ stereocilia degradation, we envision two models of ␥cyto-actin function. First, ␥cyto-actin may have a specific role involving annealing of broken filaments or de novo polymerization, perhaps depending on an unknown actin-binding protein with specificity for ␥cytoactin. Alternatively, ␥cyto-actin-containing filaments may have distinct biophysical and biochemical properties, such as different polymerization rates or polymer stability, which protect stereocilia from mechanical stress. Deficient repair and/or diminished structural integrity then result in the eventual loss of Actg1⫺/⫺ stereocilia. Interestingly, the gaps observed in auditory stereocilia of noise-damaged animals and untreated Actg1⫺/⫺ mice (24) (Fig. 3) resemble discontinuities in actin-rich developing Drosophila bristles (26, 27). In these structures, gaps are observed both during formation, as short modules of F-actin are cross-linked to form fibers, and during disassembly, as the fibers are broken down into the original modules (26). Although mammalian stereocilia are not thought to be composed of cross-linked F-actin modules, elements of Drosophila actin regulation may nonetheless be conserved in mammalian stereocilia. Stereocilia gaps may arise through physical damage that cause F-actin bundle breakage or through the action of a protein that senses damage and severs and depolymerizes actin filaments, generating gaps similar to those that occur during modular disassembly of Drosophila bristles (28). The accumulation of ␥cyto-actin at sites of damage in wild-type hair cell stereocilia after noise exposure (Figs. 3 H–K), together with the disassembly and subsequent loss of individual stereocilia in Actg1⫺/⫺ hair cells (Fig. 4 G–I), are consistent with ␥cyto-actin being required for maintenance and/or repair of stereocilia in adult hair cells. However, ␥cyto-actin seems to be entirely dispensable for the proper development and functional maturation of hair cells. Indeed, the viability of Actg1⫺/⫺ embryos and the normal lifespan of at least one-third of all live-born Actg1⫺/⫺ mice demonstrate that ␥cyto-actin is not crucial for general organogenesis and thus is not necessary for the formation of actin-based structures in general. Consistent with this idea, wild-type and Actg1⫺/⫺ mice intestinal brush border microvilli are morphologically indistinguishable (Fig. S5). We previously characterized a muscle-specific ␥cyto-actin knockout mouse that was generated precisely because a whole-

ACKNOWLEDGMENTS. We thank J. Bartles and C. Bulinski for providing anti-espin antibody and one of the anti-␥cyto-actin antibodies, respectively; D. Catts and P. Diers for technical assistance; D. Drayna, R. Chadwick, N. Gavara, A. Griffith, J. Bird, and R. Morell for critically reading the manuscript; P. Belyantsev for Fig. 1 drawing; and K. Prins, S. Ikeda, and A. Ikeda for prelim-

inary analysis of Actg1⫺/⫺ mice. The work was supported by National Institutes of Health (NIH) intramural funds 1 Z01 DC000048 –11 LMG (to T.B.F.), NIH intramural funds Z01-DC-000060 (to Andrew J. Griffith), funds from the DRF and NOHR Foundation (to G.I.F.), and NIH grants DC004568 (to K.H.F.), F32 DC009539 (to B.J.P.), and a R01 AR049899 (to J.M.E.).

1. Herman IM (1993) Actin isoforms. Curr Opin Cell Biol 5:48 –55. 2. Furness DN, Katori Y, Mahendrasingam S, Hackney CM (2005) Differential distribution of beta- and gamma-actin in guinea-pig cochlear sensory and supporting cells. Hear Res 207:22–34. 3. Hofer D, Ness W, Drenckhahn D (1997) Sorting of actin isoforms in chicken auditory hair cells. J Cell Sci 110:765–770. 4. Slepecky NB, Savage JE (1994) Expression of actin isoforms in the guinea pig organ of Corti: Muscle isoforms are not detected. Hear Res 73:16 –26. 5. Yao X, Chaponnier C, Gabbiani G, Forte JG (1995) Polarized distribution of actin isoforms in gastric parietal cells. Mol Biol Cell 6:541–557. 6. Karakozova M, et al. (2006) Arginylation of beta-actin regulates actin cytoskeleton and cell motility. Science 313:192–196. 7. Procaccio V, et al. (2006) A mutation of beta-actin that alters depolymerization dynamics is associated with autosomal dominant developmental malformations, deafness, and dystonia. Am J Hum Genet 78:947–960. 8. Morell RJ, et al. (2000) A new locus for late-onset, progressive, hereditary hearing loss DFNA20 maps to 17q25. Genomics 63:1– 6. 9. Rendtorff ND, et al. (2006) A novel missense mutation in ACTG1 causes dominant deafness in a Norwegian DFNA20/26 family, but ACTG1 mutations are not frequent among families with hereditary hearing impairment. Eur J Hum Genet 14:1097–1105. 10. van Wijk E, et al. (2003) A mutation in the gamma actin 1 (ACTG1) gene causes autosomal dominant hearing loss (DFNA20/26). J Med Genet 40:879 – 884. 11. Zhu M, et al. (2003) Mutations in the gamma-actin gene (ACTG1) are associated with dominant progressive deafness (DFNA20/26). Am J Hum Genet 73:1082–1091. 12. Dallos P (1992) The active cochlea. J Neurosci 12:4575– 4585. 13. DeRosier DJ, Tilney LG, Egelman E (1980) Actin in the inner ear: The remarkable structure of the stereocilium. Nature 287:291–296. 14. Flock A, Cheung HC (1977) Actin filaments in sensory hairs of inner ear receptor cells. J Cell Biol 75:339 –343. 15. Tilney LG, Derosier DJ, Mulroy MJ (1980) The organization of actin filaments in the stereocilia of cochlear hair cells. J Cell Biol 86:244 –259. 16. Schneider ME, Belyantseva IA, Azevedo RB, Kachar B (2002) Rapid renewal of auditory hair bundles. Nature 418:837– 838. 17. Rzadzinska AK, Schneider ME, Davies C, Riordan GP, Kachar B (2004) An actin molecular treadmill and myosins maintain stereocilia functional architecture and self-renewal. J Cell Biol 164:887– 897.

18. Furness DN, Mahendrasingam S, Ohashi M, Fettiplace R, Hackney CM (2008) The dimensions and composition of stereociliary rootlets in mammalian cochlear hair cells: Comparison between high- and low-frequency cells and evidence for a connection to the lateral membrane. J Neurosci 28:6342– 6353. 19. Biacabe B, Chevallier JM, Avan P, Bonfils P (2001) Functional anatomy of auditory brainstem nuclei: Application to the anatomical basis of brainstem auditory evoked potentials. Auris Nasus Larynx 28:85–94. 20. Liberman MC, et al. (2002) Prestin is required for electromotility of the outer hair cell and for the cochlear amplifier. Nature 419:300 –304. 21. Mannherz HG, Leigh JB, Leberman R, Pfrang H (1975) A specific 1:1 G-actin:DNAase i complex formed by the action of DNAase I on F-actin. FEBS Lett 60:34 –38. 22. Loomis PA, et al. (2003) Espin cross-links cause the elongation of microvillus-type parallel actin bundles in vivo. J Cell Biol 163:1045–1055. 23. Andrianantoandro E, Pollard TD (2006) Mechanism of actin filament turnover by severing and nucleation at different concentrations of ADF/cofilin. Mol Cell 24:13–23. 24. Avinash GB, Nuttall AL, Raphael Y (1993) 3-D analysis of F-actin in stereocilia of cochlear hair cells after loud noise exposure. Hear Res 67:139 –146. 25. Jockusch BM, Schoenenberger CA, Stetefeld J, Aebi U (2006) Tracking down the different forms of nuclear actin. Trends Cell Biol 16:391–396. 26. Guild GM, Connelly PS, Ruggiero L, Vranich KA, Tilney LG (2003) Long continuous actin bundles in Drosophila bristles are constructed by overlapping short filaments. J Cell Biol 162:1069 –1077. 27. Tilney LG, Connelly P, Smith S, Guild GM (1996) F-actin bundles in Drosophila bristles are assembled from modules composed of short filaments. J Cell Biol 135:1291–1308. 28. Tilney LG, Connelly PS, Ruggiero L, Vranich KA, Guild GM (2003) Actin filament turnover regulated by cross-linking accounts for the size, shape, location, and number of actin bundles in Drosophila bristles. Mol Biol Cell 14:3953–3966. 29. Sonnemann KJ, et al. (2006) Cytoplasmic gamma-actin is not required for skeletal muscle development but its absence leads to a progressive myopathy. Dev Cell 11:387– 397. 30. McGee J, et al. (2006) The very large G-protein-coupled receptor VLGR1: A component of the ankle link complex required for the normal development of auditory hair bundles. J Neurosci 26:6543– 6553. 31. Otey CA, Kalnoski MH, Lessard JL, Bulinski JC (1986) Immunolocalization of the gamma isoform of nonmuscle actin in cultured cells. J Cell Biol 102:1726 –1737.

9708 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0900221106

Belyantseva et al.