The role of hair cells, cilia and ciliary motility in otolith formation in the zebrafish otic vesicle

Otoliths are biomineralised structures required for the sensation of gravity, linear acceleration and sound in the zebrafish ear. Otolith precursor particles, initially distributed throughout the otic vesicle lumen, become tethered to the tips of hair cell kinocilia (tether cilia) at the otic vesicle poles, forming two otoliths. We have used high-speed video microscopy to investigate the role of cilia and ciliary motility in otolith formation. In wild-type ears, groups of motile cilia are present at the otic vesicle poles, surrounding the immotile tether cilia. A few motile cilia are also found on the medial wall, but most cilia (92-98%) in the otic vesicle are immotile. In mutants with defective cilia (iguana) or ciliary motility (lrrc50), otoliths are frequently ectopic, untethered or fused. Nevertheless, neither cilia nor ciliary motility are absolutely required for otolith tethering: a mutant that lacks cilia completely (MZovl) is still capable of tethering otoliths at the otic vesicle poles. In embryos with attenuated Notch signalling [mindbomb mutant or Su(H) morphant], supernumerary hair cells develop and otolith precursor particles bind to the tips of all kinocilia, or bind directly to the hair cells' apical surface if cilia are absent [MZovl injected with a Su(H)1+2 morpholino]. However, if the first hair cells are missing (atoh1b morphant), otolith formation is severely disrupted and delayed. Our data support a model in which hair cells produce an otolith precursor-binding factor, normally localised to tether cell kinocilia. We also show that embryonic movement plays a minor role in the formation of normal otoliths.     

Otolith matrix proteins OMP-1 and Otolin-1 are necessary for normal otolith growth and their correct anchoring onto the sensory maculae

Fish otoliths are highly calcified concretions deposited in the inner ear and serve as a part of the hearing and balance systems. They consist mainly of calcium carbonate and a small amount of organic matrix. The latter component is considered to play important roles in otolith formation. Previously, we identified two major otolith matrix proteins, OMP-1 (otolith matrix protein-1) and Otolin-1, from salmonid species. To assess the function of these proteins in otolith formation, we performed antisense morpholino oligonucleotide (MO)-mediated knockdown of omp-1 and otolin-1 in zebrafish embryos. We first identified zebrafish cDNA homologs of omp-1 (zomp-1) and otolin-1 (zotolin-1). Whole-mount in situ hybridization then revealed that the expression of both zomp-1 and zotolin-1 mRNAs is restricted to the otic vesicles. zomp-1 mRNA was expressed from the 14-somite stage in the otic placode, but the zOMP-1 protein was detected only from 26-somite stage onwards. In contrast, zotolin-1 mRNA expression became clear around 72 hpf. MOs designed to inhibit zomp-1 and zotolin-1 mRNA translation, respectively, were injected into 1-2 cell stage embryos. zomp-1 MO caused a reduction in otolith size and an absence of zOtolin-1 deposition, while zotolin-1 MO caused a fusion of the two otoliths, and an increased instability of otoliths after fixation. We conclude that zOMP-1 is required for normal otolith growth and deposition of zOtolin-1 in the otolith, while zOtolin-1, a collagenous protein, is involved in the correct arrangement of the otoliths onto the sensory epithelium of the inner ear and probably in stabilization of the otolith matrix.     

Fgf8 and Fgf3 are required for zebrafish ear placode induction,maintenance and inner ear patterning

The vertebrate inner ear develops from initially 'simple' ectodermal placode and vesicle stages into the complex three-dimensional structure which is necessary for the senses of hearing and equilibrium. Although the main morphological events in vertebrate inner ear development are known, the genetic mechanisms controlling them are scarcely understood. Previous studies have suggested that the otic placode is induced by signals from the chordamesoderm and the hindbrain, notably by fibroblast growth factors (Fgfs) and Wnt proteins. Here we study the role of Fgf8 as a bona-fide hindbrain-derived signal that acts in conjunction with Fgf3 during placode induction, maintenance and otic vesicle patterning. Acerebellar (ace) is a mutant in the fgf8 gene that results in a non-functional Fgf8 product. Homozygous mutants for acerebellar (ace) have smaller ears that typically have only one otolith, abnormal semi-circular canals, and behavioral defects. Using gene expression markers for the otic placode, we find that ace/fgf8 and Fgf-signaling are required for normal otic placode formation and maintenance. Conversely, misexpression of fgf8 or Fgf8-coated beads implanted into the vicinity of the otic placode can increase ear size and marker gene expression, although competence to respond to the induction appears restricted. Cell transplantation experiments and expression analysis suggest that Fgf8 is required in the hindbrain in the rhombomere 4-6 area to restore normal placode development in ace mutants, in close neighbourhood to the forming placode, but not in mesodermal tissues. Fgf3 and Fgf8 are expressed in hindbrain rhombomere 4 during the stages that are critical for placode induction. Joint inactivation of Fgf3 and Fgf8 by mutation or antisense-morpholino injection causes failure of placode formation and results in ear-less embryos, mimicking the phenotype we observe after pharmacological inhibition of Fgf-signaling. Fgf8 and Fgf3 together therefore act during induction and differentiation of the ear placode. In addition to the early requirement for Fgf signaling, the abnormal differentiation of inner ear structures and mechanosensory hair cells in ace mutants, pharmacological inhibition of Fgf signaling, and the expression of fgf8 and fgf3 in the otic vesicle demonstrate independent Fgf function(s) during later development of the otic vesicle and lateral line organ. We furthermore addressed a potential role of endomesomerm by studying mzoep mutant embryos that are depleted of head endomesodermal tissue, including chordamesoderm, due to a lack of Nodal-pathway signaling. In these embryos, early placode induction proceeds largely normally, but the ear placode extends abnormally to midline levels at later stages, suggesting a role for the midline in restricting placode development to dorsolateral levels. We suggest a model of zebrafish inner ear development with several discrete steps that utilize sequential Fgf signals during otic placode induction and vesicle patterning.     

A screen for genetic defects of the zebrafish ear

To advance the understanding of genetic mechanisms involved in the patterning and the differentiation of the vertebrate auditory system, we screened for mutations affecting ear development in the zebrafish larva. Fifteen recessive mutant alleles have been isolated and analyzed. The phenotypes of these mutants involve abnormalities in ear morphology, otolith formation, or both processes in parallel. Among morphological defects, we found mutations affecting early patterning of the otic vesicle, the morphogenesis of semicircular canals, and the expansion of the ear lumen. The two most severe mutant phenotypes involve the absence of anterior and posterior cristae, as well as a severely misshapen morphology of the ear. In the category of otolith mutants, we found defects in otolith formation, growth, and shape. As it proved to be the case in past screening efforts of this type, these mutant lines represent an asset in the studies of molecular mechanisms that regulate vertebrate ear development.     

Sparc (Osteonectin) functions in morphogenesis of the pharyngeal skeleton and inner ear

Sparc (Osteonectin), a matricellular glycoprotein expressed by many differentiated cells, is a major non-collagenous constituent of vertebrate bones. Recent studies indicate that Sparc expression appears early in development, although its function and regulation during embryogenesis are largely unknown. We cloned zebrafish sparc and investigated its role during development, using a mo rpholino antisense oligonucleotide-based knockdown approach. Consistent with its strong expression in the otic vesicle and developing pharyngeal cartilages, knockdown of Sparc function resulted in specific inner ear and cartilage defects that are highlighted by changes in gene expression, morphology and behavior. We rescued the knockdown phenotypes by co-injecting sparc mRNA, providing evidence that the knockdown phenotype is due specifically to impairment of Sparc function. A comparison of the phenotypes of Sparc knockdown and known zebrafish mutants with similar defects places Sparc downstream of sox9 in the genetic network that regulates development of the pharyngeal skeleton and inner ear of vertebrates.     

Zona pellucida domain-containing protein β-tectorin is crucial for zebrafish proper inner ear development

Background

The zona pellucida (ZP) domain is part of many extracellular proteins with diverse functions from structural components to receptors. The mammalian β-tectorin is a protein of 336 amino acid residues containing a single ZP domain and a putative signal peptide at the N-terminus of the protein. It is 1 component of a gel-like structure called the tectorial membrane which is involved in transforming sound waves into neuronal signals and is important for normal auditory function. β-Tectorin is specifically expressed in the mammalian and avian inner ear.

Methodology/Principal Findings

We identified and cloned the gene encoding zebrafish β-tectorin. Through whole-mount in situ hybridization, we demonstrated that β-tectorin messenger RNA was expressed in the otic placode and specialized sensory patch of the inner ear during zebrafish embryonic stages. Morpholino knockdown of zebrafish β-tectorin affected the position and number of otoliths in the ears of morphants. Finally, swimming behaviors of β-tectorin morphants were abnormal since the development of the inner ear was compromised.

Conclusions/Significance

Our results reveal that zebrafish β-tectorin is specifically expressed in the zebrafish inner ear, and is important for regulating the development of the zebrafish inner ear. Lack of zebrafish β-tectorin caused severe defects in inner ear formation of otoliths and function.     

Cadherin expression in the inner ear of developing zebrafish

Cadherins are cell adhesion molecules that have been implicated in development of a variety of organs including the ear. In this study we analyzed expression patterns of three zebrafish cadherins (Cadherin-2, -4, and -11) in the embryonic and larval zebrafish inner ear using both in situ hybridization and immunocytochemical methods. All three Cadherins exhibit distinct spatiotemporal patterns of expression during otic vesicle morphogenesis. Cadherin-2 and Cadherin-4 proteins and their respective mRNAs were detected mainly in the sensory patches and the statoacoustic ganglion (SAg), respectively. In contrast, cadherin-11mRNA was widely expressed earlier in the otic placode, and later became restricted to a subset of cells in the inner ear, including hair cells.     

Role of Bone morphogenetic protein 4 in zebrafish semicircular canal development

Bone morphogenetic proteins (BMPs) are known to play roles in inner ear development of higher vertebrates. In zebrafish, there are several reports showing that members of the BMP family are expressed in the otic vesicle. We have isolated a novel zebrafish mutant gallery, which affects the development of the semicircular canal. Gallery merely forms the lateral and the immature anterior protrusion, and does not form posterior and ventral protrusions. We found that the expression of bmp2b and bmp4, both expressed in the normal optic vesicle at the protrusion stage, are extremely upregulated in the otic vesicle of gallery. To elucidate the role of BMPs in the development of the inner ear of zebrafish, we have applied excess BMP to the wild-type otic vesicle. The formation of protrusions was severely affected, and in some cases, they were completely lost in BMP4-treated embryos. Furthermore, the protrusions in gallery treated with Noggin were partially rescued. These data indicate that BMP4 plays an important role in the development of protrusions to form semicircular canals.     

Requirements for FGF3 and FGF10 during inner ear formation

Members of the fibroblast growth factor (FGF) gene family control formation of the body plan and organogenesis in vertebrates. FGF3 is expressed in the developing hindbrain and has been shown to be involved in inner ear development of different vertebrate species, including zebrafish, Xenopus, chick and mouse. In the mouse, insertion of a neomycin resistance gene into the Fgf3 gene via homologous recombination results in severe developmental defects during differentiation of the otic vesicle. We have addressed the precise roles of FGF3 and other FGF family members during formation of the murine inner ear using both loss- and gain-of-function experiments. We generated a new mutant allele lacking the entire FGF3-coding region but surprisingly found no evidence for severe defects either during inner ear development or in the mature sensory organ, suggesting the functional involvement of other FGF family members during its formation. Ectopic expression of FGF10 in the developing hindbrain of transgenic mice leads to the formation of ectopic vesicles, expressing some otic marker genes and thus indicating a role for FGF10 during otic vesicle formation. Expression analysis of FGF10 during mouse embryogenesis reveals a highly dynamic pattern of expression in the developing hindbrain, partially overlapping with FGF3 expression and coinciding with formation of the inner ear. However, FGF10 mutant mice have been reported to display only mild defects during inner ear differentiation. We thus created double mutant mice for FGF3 and FGF10, which form severely reduced otic vesicles, suggesting redundant roles of these FGFs, acting in combination as neural signals for otic vesicle formation.     

Separate Na,K-ATPase genes are required for otolith formation and semicircular canal development in zebrafish

We have investigated the role of Na,K-ATPase genes in zebrafish ear development. Six Na,K-ATPase genes are differentially expressed in the developing zebrafish inner ear. Antisense morpholino knockdown of Na,K-ATPase alpha1a.1 expression blocked formation of otoliths. This effect was phenocopied by treatment of embryos with ouabain, an inhibitor of Na,K-ATPase activity. The otolith defect produced by morpholinos was rescued by microinjection of zebrafish alpha1a.1 or rat alpha1 mRNA, while the ouabain-induced defect was rescued by expression of ouabain-resistant zebrafish alpha1a.1 or rat alpha1 mRNA. Knockdown of a second zebrafish alpha subunit, alpha1a.2, disrupted development of the semicircular canals. Knockdown of Na,K-ATPase beta2b expression also caused an otolith defect, suggesting that the beta2b subunit partners with the alpha1a.1 subunit to form a Na,K-ATPase required for otolith formation. These results reveal novel roles for Na,K-ATPase genes in vestibular system development and indicate that different isoforms play distinct functional roles in formation of inner ear structures. Our results highlight zebrafish gene knockdown-mRNA rescue as an approach that can be used to dissect the functional properties of zebrafish and mammalian Na,K-ATPase genes.     

Mutated <Emphasis Type="Italic">otopetrin 1</Emphasis> affects the genesis of otoliths and the localization of Starmaker in zebrafish

Otoliths in bony fishes and otoconia in mammals are composite crystals consisting of calcium carbonate and proteins. These biominerals are part of the gravity and linear acceleration detection system of the inner ear. Mutations in otopetrin 1 have been shown to result in lack of otoconia in tilted and mergulhador mutant mice. The molecular function of Otopetrin 1, a novel protein that contains ten predicted transmembrane domains, however, has remained elusive. Here we show that a mutation in the orthologous gene in zebrafish is responsible for the complete absence of otoliths in backstroke mutants. We examined the localization of Starmaker, a secreted protein that is highly abundant in otoliths in backstroke mutants. Starmaker protein accumulated within cells of the otic epithelium, indicating a possible defect in secretion. Our data suggest that Otopetrin 1 in zebrafish may be involved in the protein trafficking of components required for formation of biominerals in the ear.     

Fgf and Hh signalling act on a symmetrical pre-pattern to specify anterior and posterior identity in the zebrafish otic placode and vesicle

Specification of the otic anteroposterior axis is one of the earliest patterning events during inner ear development. In zebrafish, Hedgehog signalling is necessary and sufficient to specify posterior otic identity between the 10 somite (otic placode) and 20 somite (early otic vesicle) stages. We now show that Fgf signalling is both necessary and sufficient for anterior otic specification during a similar period, a function that is completely separable from its earlier role in otic placode induction. In lia(-/-) (fgf3(-/-)) mutants, anterior otic character is reduced, but not lost altogether. Blocking all Fgf signalling at 10-20 somites, however, using the pan-Fgf inhibitor SU5402, results in the loss of anterior otic structures and a mirror image duplication of posterior regions. Conversely, overexpression of fgf3 during a similar period, using a heat-shock inducible transgenic line, results in the loss of posterior otic structures and a duplication of anterior domains. These phenotypes are opposite to those observed when Hedgehog signalling is altered. Loss of both Fgf and Hedgehog function between 10 and 20 somites results in symmetrical otic vesicles with neither anterior nor posterior identity, which, nevertheless, retain defined poles at the anterior and posterior ends of the ear. These data suggest that Fgf and Hedgehog act on a symmetrical otic pre-pattern to specify anterior and posterior otic identity, respectively. Each signalling pathway has instructive activity: neither acts simply to repress activity of the other, and, together, they appear to be key players in the specification of anteroposterior asymmetries in the zebrafish ear.     

First steps of otolith formation of the zebrafish: role of glycogen?

The first steps of otolith formation were studied by electron microscopy in zebrafish embryos at different postfertilization (PF) time intervals. Between 19 and 22 h PF, the otic cavity contains glycogen particles derived by an apocrine process from the apical portions of the epithelial cells of the inner ear. The particles are arranged in parallel arrays, then in pseudocrystalloid structures, and finally in concentric arrays to form dense clusters referred to as "spherules". At 23 h PF, a group of "globules", consisting of modified aggregated "spherules" surrounded by several free "spherules", forms the nascent otolith. At 30 h PF, fused globules form a roughly spherical otolith. Spherules undergoing their process of modification and aggregation, are located in its central part, and constitute the so-called "nucleus". At 50 h PF, the otolith is a flattened hemisphere. It is made up of fused globules surrounded by two concentric layers whose organization is similar to that observed in the otolith of the adult fish. At this stage, calcium may be detected in the otolith except in its nucleus. We suggest that glycogen molecules found in the nascent otolith might allow the insertion of molecules such as glycoproteins (collagens) which are known to fix calcium. As a result, glycogen might play a key role in initiating the formation of otoliths and possibly that of other calcified tissues.     

Cilia-driven fluid flow as an epigenetic cue for otolith biomineralization on sensory hair cells of the inner ear

Ciliary motility is necessary for many developmental and physiological processes in animals. In zebrafish, motile cilia are thought to be required for the deposition of otoliths, which comprise crystals of protein and calcium carbonate, on hair cells of the inner ear. The identity of the motile cilia and their role in otolith biogenesis, however, remain controversial. Here, we show that the ear vesicle differentiates numerous motile cilia, the spatial distribution of which changes as a function of the expression pattern of the ciliogenic gene foxj1b. By contrast, the hair cells develop immotile kinocilia that serve as static tethers for otolith crystallization. In ears devoid of all cilia, otoliths can form but they are of irregular shapes and sizes and appear to attach instead to the hair cell apical membranes. Moreover, overproduction of motile cilia also disrupts otolith deposition through sustained agitation of the precursor particles. Therefore, the correct spatial and temporal distribution of the motile cilia is crucial for proper otolith formation. Our findings support the view that the hair cells express a binding factor for the otolith precursors, while the motile cilia ensure that the precursors do not sediment prematurely and are efficiently directed towards the hair cells. We also provide evidence that the kinocilia are modified motile cilia that depend on Foxj1b for their differentiation. We propose that in hair cells, a Foxj1b-dependent motile ciliogenic program is altered by the proneural Atoh proteins to promote the differentiation of immotile kinocilia.     

Immunohistochemical localization of two otolith matrix proteins in the otolith and inner ear of the rainbow trout,<Emphasis Type="Italic"> Oncorhynchus mykiss</Emphasis>: comparative aspects between the adult inner ear and embryonic otocysts

The fish otolith consists mainly of calcium carbonate and organic matrices, the latter of which may play important roles in the process of otolith formation. We previously identified two otolith matrix proteins, named otolith matrix protein-1 (OMP-1) and otolin-1, from the rainbow trout, Oncorhynchus mykiss, and the chum salmon, O. keta. In this study, recombinant proteins corresponding to OMP-1 and otolin-1 were synthesized using yeast and bacterial expression systems, respectively, to produce specific antibodies against each protein. Immunohistochemical analysis using these antisera revealed that in the otoliths of adult fish, OMP-1 and otolin-1 were colocalized along the daily rings possibly formed by alternate deposition of calcium carbonate and organic matrices. In the adult inner ear, OMP-1 was produced at most of the saccular epithelium, while otolin-1 was produced at a limited part of cylindrical cells located at the marginal zone of the sensory epithelium. In the embryonic inner ear, these proteins had already existed in the otolith primordia when calcification had commenced. In addition, otolin-1 was localized in the fibrous materials connecting otolith primordia and sensory epithelium at this stage. These results indicate that these proteins are required as essential components for otolith formation and calcification.     

Development of utricular otoliths, but not saccular otoliths, is necessary for vestibular function and survival in zebrafish

We have been studying the consequences of embryonic vestibular dysfunction caused by the monolith (mnl) mutation in zebrafish. mnl is a dominant mutation that specifically inhibits formation of utricular otoliths. However, briefly immobilizing mnl/mnl embryos in agarose with the otic vesicle orientated at certain angles selectively induces or prevents formation of utricular and/or saccular otoliths. With this noninvasive technique, we generated six phenotypic classes of mnl/mnl mutants, designated S-S, U-U, U-S, S-US, U-US, and US-US, depending on which otoliths are present on each side (U, utricular otolith; S, saccular otolith). All mnl/mnl larvae survived through day 10 of development. Thereafter, S-S larvae showed a rapid decline, probably because of starvation, and none survived to adulthood. Survival rates in all other classes of mnl/mnl larvae (those having at least one utricular otolith) were close to normal. The presence or absence of utricular otoliths also correlated with vestibular function during early larval development, as measured by three criteria: First, unlike wild-type larvae, S-S mutant larvae showed almost no detectable counter-rotation of the eyes when tilted tail up or tail down. Second, 95% of S-S mutant larvae never acquired the ability to maintain a balanced dorsal-up posture. Third, although most wild-type larvae responded to gentle prodding by swimming in a straight line, S-S larvae responded by swimming in rapid circles, showing sudden and frequent changes in direction ("zigzagging"), and/or rolling and spiraling. All other phenotypic classes of mnl/mnl larvae behaved normally in these assays. These data demonstrate that bilateral loss of utricular otoliths disrupts the ability to sense gravity, severely impairs balance and motor coordination, and is invariably lethal. The presence of a utricular otolith in at least one inner ear is necessary and sufficient for vestibular function and survival. In contrast, saccular otoliths are dispensable for these functions.     

Sonic hedgehog regulates otic capsule chondrogenesis and inner ear development in the mouse embryo

Development of the cartilaginous capsule of the inner ear is dependent on interactions between otic epithelium and its surrounding periotic mesenchyme. During these tissue interactions, factors endogenous to the otic epithelium influence the differentiation of the underlying periotic mesenchyme to form a chondrified otic capsule. We report the localization of Sonic hedgehog (Shh) protein and expression of the Shh gene in the tissues of the developing mouse inner ear. We demonstrate in cultures of periotic mesenchyme that Shh alone cannot initiate otic capsule chondrogenesis. However, when Shh is added to cultured periotic mesenchyme either in combination with otic epithelium or otic epithelial-derived fibroblast growth factor (FGF2), a significant enhancement of chondrogenesis occurs. Addition of Shh antisense oligonucleotide (AS) to cultured periotic mesenchyme with added otic epithelium decreases levels of endogenous Shh and suppresses the chondrogenic response of the mesenchyme cells, while supplementation of Shh AS-treated cultures with Shh rescues cultures from chondrogenic inhibition. We demonstrate that inactivation of Shh by targeted mutation produces anomalies in the developing inner ear and its surrounding capsule. Our results support a role for Shh as a regulator of otic capsule formation and inner ear development during mammalian embryogenesis.     

The developing lamprey ear closely resembles the zebrafish otic vesicle: otx1 expression can account for all major patterning differences

The inner ear of adult agnathan vertebrates is relatively symmetric about the anteroposterior axis, with only two semicircular canals and a single sensory macula. This contrasts with the highly asymmetric gnathostome arrangement of three canals and several separate maculae. Symmetric ears can be obtained experimentally in gnathostomes in several ways, including by manipulation of zebrafish Hedgehog signalling, and it has been suggested that these phenotypes might represent an atavistic condition. We have found, however, that the symmetry of the adult lamprey inner ear is not reflected in its early development; the lamprey otic vesicle is highly asymmetric about the anteroposterior axis, both morphologically and molecularly, and bears a striking resemblance to the zebrafish otic vesicle. The single sensory macula originates as two foci of hair cells, and later shows regions of homology to the zebrafish utricular and saccular maculae. It is likely, therefore, that the last common ancestor of lampreys and gnathostomes already had well-defined otic anteroposterior asymmetries. Both lamprey and zebrafish otic vesicles express a target of Hedgehog signalling, patched, indicating that both are responsive to Hedgehog signalling. One significant distinction between agnathans and gnathostomes, however, is the acquisition of otic Otx1 expression in the gnathostome lineage. We show that Otx1 knockdown in zebrafish, as in Otx1(-/-) mice, gives rise to lamprey-like inner ears. The role of Otx1 in the gnathostome ear is therefore highly conserved; otic Otx1 expression is likely to account not only for the gain of a third semicircular canal and crista in gnathostomes, but also for the separation of the zones of the single macula into distinct regions.     

The zebrafish gene claudinj is essential for normal ear function and important for the formation of the otoliths

We have identified a mutation in the zebrafish gene claudinj generated by retroviral integration. Mutant embryos display otoliths severely reduced in size, no response to tapping stimulus, and an inability to balance properly suggesting vestibular and hearing dysfunction. Antisense in situ hybridization to the cldnj gene showed expression first in the otic placode and later asymmetric expression in the otic vesicle. Morpholino inhibition of claudinj expression showed similar defects in otolith formation. Phylogenetic analysis of claudin sequences from multiple species demonstrates that claudinj was part of a gene expansion that began in the common ancestor of fish and humans, but additional fish specific gene duplications must have also occurred.