Usher syndrome type I (USH1) associates severe congenital deafness, vestibular dysfunction and progressive retinitis pigmentosa leading to blindness. The gene encoding myosin VIIA is responsible for USH1B. Mutations in the murine orthologous gene lead to the shaker-1 phenotype, which manifests cochlear and vestibular dysfunction, without any retinal defect. To address this phenotypic discrepancy, the expression of myosin VIIA in retinal cells was analyzed in human and mouse during embryonic development and adult life. In the human embryo, myosin VIIA was present first in the pigment epithelium cells, and later in these cells as well as in the photoreceptor cells. In the adult human retina, myosin VIIA was present in both cell types. In contrast, in mouse, only pigment epithelium cells expressed the protein throughout development and adult life. Myosin VIIA was also found to be absent in the photoreceptor cells of other rodents (rat and guinea-pig), whereas these cells expressed the protein in amphibians, avians and primates. These observations suggest that retinitis pigmentosa of USH1B results from a primary rod and cone defect. The USH1B/ shaker-1 paradigm illustrates a species-specific cell pattern of gene expression as a possible cause for the discrepancy between phenotypes involving defective orthologous genes in man and mouse. Interestingly, in the photoreceptor cells, myosin VIIA is mainly localized in the inner and base of outer segments as well as in the synaptic ending region where it is co-localized with the synaptic vesicles. Therefore, we suggest that myosin VIIA might play a role in the trafficking of ribbon-synaptic vesicle complexes and the renewal processes of the outer photoreceptor disks.
Usher syndrome (USH) is an autosomal recessive defect which affects both the inner ear and the retina ( 1 ). It is regarded as the most frequent cause of deaf-blindness in humans and accounts for 3 to 6% of deaf children ( 2 , 3 ). Three clinical subtypes have been described, based on the severity of the hearing loss, the presence or absence of balance problems and the age of onset of the retinitis pigmentosa. The most severe form, USH type I (USH1), is characterized by profound congenital sensorineural hearing loss, constant vestibular dysfunction and prepubertal onset of retinitis pigmentosa leading to blindness. At least four loci have been assigned to USH1, three of which, USH1A, 1B and 1C, have already been mapped to chromosomes 14q32, 11q13.5 and 11p15 respectively ( 4–7 ). We have identified the myosin VIIA gene as responsible for USH1B (which accounts for about one half of all USH cases) ( 8 ). Mutations in the murine orthologous gene lead to the shaker-1 phenotype ( 9 ), characterized by hearing impairment and balance problems due to damaged cochlear and vestibular neuroepithelia ( 10 , 11 ). However, none of the six different shaker-1 mutants seems to display any retinal defect ( 10 , 11 ). To explain the discrepancy between the human and mouse retinal phenotypes, several hypotheses can be considered: (i) a difference in the nature of the mutations; (ii) a dissimilarity between mouse and human retinal development ( 12 , 13 ); (iii) a different influence of modifier genes; (iv) the expression of a gene with redundant function in the mouse retina; (v) an accumulation process at the origin of the retinal phenotype, too slow to be deleterious during the mouse lifetime.
The shaker-1 mutations are unlikely to account for the retinal phenotype discrepancy, since at least two of the three mutations identified in shaker-1 mutants are expected to be highly detrimental to the function of myosin VIIA, due to their particular location within the motor head of this protein ( 9 ). Moreover, the possibility that these mice develop an a minima retinal defect during late adult life has been excluded through electroretinographic and electron microscopy studies (K. Steel, pers. comm.). We thus addressed the basis of this phenotypic difference by comparing the retinal cell distribution of myosin VIIA in man and mouse throughout development and in the adult.
Unconventional myosins are motor molecules with a very conserved head domain which contains the ATP-and actin-binding sites. Their tails are highly divergent from one unconventional myosin to another. They are expected to interact with specific membranous compartments which then move along actin filaments ( 14 ). Polyclonal antibodies specific to a synthetic peptide and a histidyl-tagged tail fragment of human myosin VIIA were generated (see Materials and Methods). Since the deduced amino acid sequences of murine and human myosin VIIA are 95% identical ( 8 ), these antibodies were considered likely to detect the protein in the mouse. Indeed, on western blots, they both detected a unique band of the expected molecular weight (220–250 kDa) ( 15 ) in the adult mouse testis, retina, inner ear and kidney ( Fig. 1 A).
Cell distribution of myosin VIIA in the mouse cochlea
On mouse tissue sections, myosin VIIA was first detected by immunohistofluorescence in the otic vesicle at embryonic day 11 (E11). In the otocyst, the sensory hair cells were labeled throughout development, in the vestibular utricula and macula, as well as in the cochlea; no myosin VIIA expression was detected in the supporting cells ( Fig. 1 B). In the cochlea, the two types of sensory cells (outer and inner hair cells) were labeled [ Fig. 1 B(3)] and a few positive cells were also detected in the stria vascularis. No immunolabeling was observed upon substitution of pre-immune serum for the purified anti-myosin VIIA antibody or after preadsorption of this antibody with the His 6 -myosin VIIA tail fragment (data not shown).
Cell distribution of myosin VIIA in the retina
The retinal cell distribution of myosin VIIA during the course of development was investigated by immunohistofluorescence in man, mouse as well as several other animal species.
Human . In the human embryo at 6, 9, and 10 weeks, myosin VIIA was detected in the retinal pigment epithelium. At these stages, no labeling was observed in the undifferentiated neural retina ( Fig. 2 A). During the fourth month, the differentiation starts at the inner layers of retina and progresses toward the outer layers, so that the photoreceptor cells are the last to differentiate ( 13 ). At 18 and 19 weeks, myosin VIIA was present in both the pigment epithelium and the photoreceptor cells. Immunolabeling increased in both cell types at 24 ( Fig. 2 B) and 28 weeks; myosin VIIA was mainly localized at the tip of the photoreceptor cells, which at that time are not yet mature (the inner and outer segments are not completely developed). In the adult retina, myosin VIIA was detected in the pigment epithelium cells as well as in the rod and cone photoreceptor cells. In these cells, myosin VIIA was mainly present in the inner segments, the base of the outer segments and the synaptic endings ( Fig. 2 C). No immunolabeling was detected in the bipolar or ganglionic retinal cells ( Fig. 2 A–C).
Mouse . The distribution of myosin VIIA in the mouse retina was investigated from E10 until birth at daily intervals and during post-natal life. Labeling was first detected at E12, in the pigment epithelium, where it increased from E13 onwards. No immunolabeling was observed in the neural retina at any stage of embryonic development ( Fig. 3 A–D). Since the maturation of mouse photoreceptor cells is known to be complete only at 3 weeks postnatal, the distribution of myosin VIIA in the retina was further investigated during adult life. Analysis at day 10, day 21, month 2 and month 6 showed the presence of myosin VIIA only in the pigment epithelium cells ( Fig. 3 D). No immunolabeling was detected in the photoreceptor cells, bipolar or ganglionic retinal cells ( Fig. 3 D). In situ hybridization performed in parallel showed that the photoreceptor cells do not express the myosin VIIA mRNA, whereas the pigment epithelium cells do throughout murine embryonic development and post-natal life (data not shown).
Other animal species . To investigate the interspecies conservation of myosin VIIA expression in the photoreceptor cells, cell distribution of the protein in the retina was analyzed in other vertebrates. In the adult rat [as reported in ( 16 )] and guinea-pig, no immunolabeling was detected in the photoreceptor cells, whereas myosin VIIA was present in the pigment epithelium ( Fig. 4 A, B). In contrast, both photoreceptor and pigment epithelium cells were strongly labeled in the Xenopus , chicken and macaque ( Fig. 4 C–F). In the macaque and chicken embryos, labeling in the photoreceptor cells was concentrated at the cell apical tip (that later differentiates into the inner and outer photoreceptor segments) and in the synaptic endings ( Fig. 4 F, G). In the adult chicken, myosin VIIA was concentrated in the inner segments, basal region of the outer segments ( Fig. 4 D, E) and synaptic endings of the rods and cones ( Fig. 4 D, E). In the E110 macaque ( Fig. 5 A–C) and adult chicken (data not shown), double immunolabeling of myosin VIIA and the synaptic vesicle transmembrane protein (SV2) showed that myosin VIIA was co-localized with the synaptic vesicles in the photoreceptor synaptic ending region.
The human Usher syndrome type 1B (USH1B) and the mouse shaker-1 phenotype are due to a defective myosin VIIA gene ( 8 , 9 ). Both phenotypes involve sensorineural hearing loss and vestibular trouble. However, while USH1B patients manifest progressive retinitis pigmentosa leading to blindness, no retinal defect could be detected in the shaker-1 mouse mutants ( 8–11 ). The present study addresses the basis of the discrepancy between the two phenotypes.
We first confirmed that myosin VIIA was present in the cochlear sensory hair cells during mouse embryonic development, as previously described in the adult guinea-pig cochlea ( 16 ). We also report an expression of myosin VIIA in the vestibular hair cells throughout mouse development. Myosin VIIA was found to be also restricted to sensory hair cells in the developing human otic vesicle at 6 weeks of gestation (unpublished results). This may be correlated to the vestibular and cochlear dysfunctions resulting in balance problems and hearing impairment observed in both Usher patients and shaker-1 mouse mutants. We then studied the retinal cell distribution of myosin VIIA in man and mouse and found a striking difference between the two species. While the pigment epithelium cells expressed myosin VIIA in both species, the protein was present only in the human photoreceptor cells, from the early stages of differentiation into late adult life. Moreover, myosin VIIA was found to be absent from the photoreceptor cells of other rodents (rat and guinea-pig), whereas these cells expressed the protein in amphibians, avians and primates. Therefore, with respect to the observed difference in retinal phenotypes between USH1B and shaker-1 , the simplest hypothesis is that the absence of retinal defects in shaker-1 mutants is due to the lack of myosin VIIA expression in the murine photoreceptor cells; we thus propose that defective myosin VIIA in human rods and cones accounts for the retinitis pigmentosa in USH1B patients.
All the genes responsible for human retinitis pigmentosa identified so far encode proteins involved in the phototransduction cascade or outer photoreceptor disk structural integrity, thus resulting, when mutated, in a primary defect of the photoreceptor cells ( 17–23 ). However, the study of a rat model of retinal degeneration (RCS, Royal College of Surgeons) has revealed that a primary pigment epithelium cell defect can also be responsible for retinitis pigmentosa ( 24 ). Although, the present study provides no evidence for a role of the pigment epithelium cells in the retinal defect of USH1B patients, the possibility that these cells may somehow contribute to the pathogenesis of retinal degeneration cannot be ruled out.
Unconventional myosins have been found to be involved in several membrane-associated processes, including secretory vesicles transport, vacuole formation, plasma membrane extension, transport of mitochondria and in membrane movements accompanying the pseudofurrow formation ( 25 , 26 ). However, very little is known about their role in the retina. The conservation of myosin VIIA expression in the photoreceptor cells from Xenopus to man suggests that myosin VIIA has an important function in these cells. Interestingly, this protein is highly expressed in the synaptic regions of the photoreceptor cells as well as in the cochlear [also reported in ( 16 )] and vestibular hair cells. It is worth noting that the synapses in these cells have specific structural and functional features. They are characterized by a dense body, the synaptic ribbon, anchored to the presynaptic membrane and covered with a layer of synaptic vesicles ( 27–30 ). In the photoreceptor synaptic endings, myosin VIIA was found to be co-localized with a synaptic vesicle transmembrane protein either in the E110 macaque or adult chicken. This led us to suggest that myosin VIIA could be involved in the formation and trafficking of the ribbon-synaptic vesicle complexes which characterize these sensory cells. In addition, the presence of myosin VIIA in the photoreceptor inner segment and base of the outer segment suggests a function for this protein in the formation and folding of the outer photoreceptor disks. The impairment of myosin VIIA function in the photoreceptor cells might lead to the apoptosis phenomena which characterize the retinal degeneration process. These phenomena have been studied in several mouse models involving defects in the rhodopsin, phosphodiesterase β-subunit or peripherin genes ( 31–33 ).
With the identification of an increasing number of mouse models for human diseases, resulting from defects in orthologous genes, phenotypic differences between the two species has become a frequent observation ( 34 , 35 ). To our knowledge, this is the first report of a human/mouse difference in the cell distribution of a protein that is likely to account for such a phenotypic discrepancy. The USH1B/shaker–1 paradigm points out the necessity of exploring protein expression in primate tissues in order to develop appropriate hypotheses for the pathogenesis of a human disease ( 36 ). Indeed, based on the restricted expression of myosin VIIA in the adult rat retinal pigment epithelium cells, observations limited to this species had previously led to the suggestion that retinitis pigmentosa in USH1B patients results from a defect of the pigment epithelium cells ( 16 ).
Since the conservation of myosin VIIA expression in the photoreceptor cells of distant vertebrates suggests an important role for this protein in these cells, it is expected that another unconventional myosin with similar function is present in the rodent photoreceptor cells. The identification of this myosin would open up the possibility of generating a genuine mouse model for the retinitis pigmentosa of USH1B.
Materials and Methods
Producing a tail fragment of human myosin VIIA in E. coli
A 459 bp fragment [position 2805 to 3263 ( 15 )], encoding the N-terminal portion of the human myosin VIIA tail domain was subcloned into the expression vector pQE31 (Qiagen) and overexpressed in E. coli according to the supplier's instructions. The histidyl-tagged (His 6− ) protein was purified by immobilized metal affinity chromatography; the protein was eluted with 250–300 mM imidazole in phosphate buffered saline (PBS) containing 0.5% CHAPS, 1 mM PMSF, 20 µg/ml leupeptin and 20 µg/ml pepstatin. Fractions containing the protein were concentrated on a Centricon-3 column (Amicon) and quantitated by the Bradford's assay (Bio-Rad).
Rabbit immune sera were generated against a synthetic peptide YTRRPLKQPLL YHDDEGDQ-C [amino acids 1017–1036 of the myosin VIIA deduced amino acid sequence ( 15 )] and against the purified His 6 -myosin VIIA tail fragment (see above). Male New Zealand white rabbits were primed with 300 µg of the protein and received four additional boosts (100 µg each) at 4-week intervals. Antibodies to myosin VIIA were purified by incubating the sera with a nitrocellulose strip blotted either with the His 6 -tail fragment or the myosin VIIA immunoreactive band from a mouse testis extract. The polyclonal antibody to the myosin VIIA tail fragment was used for the immunohisto-fluorescence.
Monoclonal antibodies to human rhodopsin and to the transmembrane synaptic vesicle protein SV2 were obtained from D. Hicks (Clinique Ophtalmologique, Strasbourg, France) and T. Galli (Institut Curie, Paris, France), respectively.
Protein samples were homogenized in the extraction buffer (150 mM NaCl, 1% NP-40, 10 mM HEPES, pH 7.4) supplemented with 1 mM PMSF, 20 µg/ml leupeptin and 20 µg/ml pepstatin. Five µg of protein samples (3 µg for testis extract) were subjected to SDS-PAGE and transferred to nitrocellulose ( 37 ). Blots were incubated in TNT buffer (150 mM NaCl, 0.25% Triton X-100, 50 mM Tris-HCl, pH 8), plus 5% non-fat dry milk, for 1 h at room temperature, and then with the purified anti-myosin VIIA antibody (10 µg/ml) overnight at 4°C. After four washes in TNT buffer, blots were incubated with horseradish peroxidase-conjugated goat anti-rabbit IgG (Boehringer Mannheim) in TNT-dry milk solution for 1 h at room temperature. The immunoreactive bands were visualized using the enhanced chemiluminescence system (Amersham).
Whole mouse embryos (sampled daily from day 10 until birth) and adult mouse enucleated eyes were fixed by immersion in 2% paraformaldehyde (pH 7.4) overnight at 4°C. After three PBS rinses, they were immersed in 20% sucrose-PBS for 12 h at 4°C and then frozen in O.C.T. embedding medium (Miles, Elkhart, USA). Eyes obtained from human fetuses and adult individuals were treated as described above. For adult frog, rat, guinea-pig, E110 macaque, E14 and adult chicken, the eye cups (after removal of cornea and lens) were fixed in 4% paraformaldehyde for 4–6 h.
Cryostat sections (10–14 µm) were post-fixed in 4% paraformaldehyde for 5 min, dehydrated and stored at-80°C until use.
After two PBS rinses, 15 min incubation in 50 mM NH 4 Cl and 1 h incubation in 3% bovine serum albumin (BSA)-PBS, sections were incubated overnight at 4°C with the first antibody diluted in 1% BSA-PBS, either polyclonal antibody to myosin VIIA (20 µg/ml) or monoclonal antibody to rhodopsin or SV2. After three PBS rinses, sections were incubated for 1 h at room temperature with the second antibody (Boehringer Mannheim), either FITC-conjugated anti-rabbit IgG or rhodamine-conjugated anti-mouse IgG. Sections were analyzed by either conventional epifluorescent or laser scanning confocal microscopy.
We wish to thank J.-P. Hardelin for helpful discussions and critical review of the manuscript, D. Weil and J. Levilliers for helpful advice, K. Steel for sharing unpublished information, V Kalatzis, G. Levy, P. Kussel and R. Legouis for critical comments, R. Hellio for assistance with confocal microscopy analysis.
This work was supported by grants from A. and M. Suchert, Association Française Retinitis Pigmentosa and Association Entendre.