Ray-finned fishes (Actinopterygii) are the dominant vertebrate group today (+30 000 species, predominantly teleosts), with great morphological diversity, including their dentitions. How dental morphological variation evolved is best addressed by considering a range of taxa across actinopterygian phylogeny; here we examine the dentition of Polyodon spathula (American paddlefish), assigned to the basal group Acipenseriformes. Although teeth are present and functional in young individuals of Polyodon, they are completely absent in adults. Our current understanding of developmental genes operating in the dentition is primarily restricted to teleosts; we show that shh and bmp4, as highly conserved epithelial and mesenchymal genes for gnathostome tooth development, are similarly expressed at Polyodon tooth loci, thus extending this conserved developmental pattern within the Actinopterygii. These genes map spatio-temporal tooth initiation in Polyodon larvae and provide new data in both oral and pharyngeal tooth sites. Variation in cellular intensity of shh maps timing of tooth morphogenesis, revealing a second odontogenic wave as alternate sites within tooth rows, a dental pattern also present in more derived actinopterygians. Developmental timing for each tooth field in Polyodon follows a gradient, from rostral to caudal and ventral to dorsal, repeated during subsequent loss of teeth. The transitory Polyodon dentition is modified by cessation of tooth addition and loss. As such, Polyodon represents a basal actinopterygian model for the evolution of developmental novelty: initial conservation, followed by tooth loss, accommodating the adult trophic modification to filter-feeding.
Most tooth development models reflect a bias towards morphologically derived vertebrates (e.g. zebrafish, mouse). However, more representative models for the evolution of developmental mechanisms of the dentition are provided by taxa at the base of extant vertebrate phylogenies. The basal actinopterygian order Acipenseriformes includes fossil taxa as well as the American paddlefish Polyodon (family Polyodontidae) and sturgeons (family Acipenseridae, e.g. Acipenser [1,2]) and represents an increasingly used system for addressing developmental questions in an evolutionary context [3–6]. Owing to their basal phylogenetic position, Acipenseriformes are a particularly relevant model to test hypotheses of tooth patterning and evolution. The dentition is lost in adult paddlefish and sturgeon, but present in younger individuals, although details of early stages of tooth development are poorly known [1–3,7]. As pattern order for the forming dentition has previously been described for more derived actinopterygians, comparable data for Polyodon will provide significant information on mechanisms in more phylogenetically basal actinopterygians.
The secreted protein sonic hedgehog (shh) and the TGF-β superfamily member bone morphogenetic protein4 (bmp4) are key dental patterning genes in vertebrates. In situ hybridization assays demonstrate that the transcripts coding for shh/bmp4 are present at the earliest sites of tooth initiation with focused, time specific loci of expression restricted to dental epithelium (shh) [8,9] and co-expression in the underlying condensed mesenchyme (bmp4). Co-expression occurs on each oropharyngeal dentate field, from a diffuse band of dental competence (odontogenic band), to discrete placodes of single tooth initiation. Non-mammalian vertebrates for which this conserved pattern of shh/bmp4 expression has been used to characterize dental patterning include a variety of teleosts (Osteichthyes, Actinopterygii): rainbow trout (Oncorhynchus mykiss [8,9]), Mexican tetra (Astyanax mexicanus ), zebrafish (Danio rerio [11,12]), several Lake Malawi cichlids , the freshwater pufferfish (Monotrete abei ), as well the Queensland lungfish (Neoceratodus forsteri ) and various snakes and lizards (Osteichthyes, Sarcopterygii, [16–18]) and the catshark Scyliorhinus canicula [19,20]. As the only non-teleost actinopterygian yet surveyed, our new data from Polyodon will provide key phylogenetic support for the hypothesis that shh and bmp4 are part of a conserved and ancient gene regulatory network for patterning vertebrate dentitions.
We predict that Polyodon will exhibit the conserved pattern of epithelial shh-positive loci, with comparable mesenchymal expression of bmp4 , observed in other vertebrate taxa. Here we will use expression patterns for these genes, along with other histological and morphological datasets to demonstrate temporal differences in focal localization for each tooth site in Polyodon, mapping position and timing of tooth initiation to demonstrate how pattern order is established through coordinated gene activity. Our hypothesis is that this represents a basal condition of shared genetic regulation of tooth initiation times and topographic order for the Actinopterygii.
2. Material and methods
(a) Animal care and sacrifice
Fertilized Polyodon spathula eggs were obtained from Osage Catfisheries, Inc. (Osage Beach, MO, USA) and raised to desired stages in recirculating, closed freshwater systems mimicking natural conditions (22°C, pH 7.2 ± 0.7, salinity of 1.0 ± _0.2 p.p.t. ). Polyodon were euthanized in a lethal dose of MS-222 (tricaine) and fixed for at least 24 h (dependent of tissue volume) in 4% paraformaldehyde .
(b) Staging of larval Polyodon
Polyodon staging follows [3,21]: lengths for individual specimens for stages 37–46, and other details of the staging, can be obtained from these. Feeding larvae (beyond stage 46) are described as ‘days post-staging’ (dps) and juveniles by standard length (SL). At incubation temperature (22°C), the larval period between hatching (stage 36) and onset of exogenous feeding and yolk exhaustion (stage 46) proceeds at approximately one stage per 24 h period .
(c) In situ hybridization
In situ hybridization used standard protocols  with riboprobes for shh  or bmp4. Bmp4 was cloned from cDNA using the forward primer CGA GGC TAC TTT GTT GCA CA and reverse primer TCC ACG TAC AGT TCG TGT CG. Selected whole larvae (stages 41–45) with shh or bmp4 expression were embedded in 20% gelatin and vibratome-sectioned at 50 μm or, embedded in 30% sucrose, frozen in liquid nitrogen and cryostat sectioned at 20 μm. Numbers of specimens (antisense, comparable number of sense), bmp stages 34–39 n = 7; 40–43 n = 6; 44–46 n = 6. shh stages 36 n = 2; 38 n = 3; 39 n = 3; 40 n = 2; 41 n = 4; 42 n = 2; 43 n = 2; 45 = 6. Photomicrographs were taken with Zeiss Nomarsky optics, or an Olympus SZX16 dissecting microscope equipped with a QImaging RetigaEXi digital camera.
(d) Clearing and staining, CT imaging
Cleared and stained specimens (CS; Alizarin red and Alcian blue ) were dissected and mounted as half-jaws. Older specimens were studied as CS skeletal preps under a stereomicroscope and CT scanned (X-Tek HMX ST CT scanner, Image and Analysis Centre, Natural History Museum, London; MicroCT at Dental Institute, King's College London, GE Locus SP, creating volumes with voxel sizes 6.5 μm) and rendered using the software program Drishti (http://sf.anu.edu.au/Vizlab/drishti).
The terms distal and proximal are used in the upper and lower jaws, with reference to the jaw joint (proximal) and symphysis (distal). The terms rostral and caudal, dorsal and ventral are used with respect to the body axes.
In P. spathula larvae, shh and bmp4 expression reveal both the early events of oral and pharyngeal dental patterning and sequential addition of tooth loci as development proceeds. There are notable differences in the addition of new tooth germs in individual dentate fields, normally caudal, but exceptionally rostrally on the palatopterygoid tooth plate. Concerning timing along the body axis, tooth initiation begins in association with Meckel's cartilage, establishing a spatio-temporal gradient that extends from the oral, through to tooth sites in the pharyngeal cavities (figures 1 and 2; electronic supplementary material, figure S4). Skeletal preparations provide additional data on pattern order; after tooth rows form on the dentary and dermopalatine, they develop on the more caudal palatopterygoids and first hypobranchials (figures 1a,c and 2a,b, respectively). Teeth are later organized into toothed plates, connected by basal bone of attachment, representing functional surfaces of the oropharyngeal dentition (table 1, electronic supplementary material, figure S2c) [1–3].
(a) Timing of shh expression in whole mounts maps sequential tooth initiation (stages 37–43)
Spatial expression of shh occurs as focal loci, with changes in intensity coincident with each stage of tooth germ morphogenesis, mapping location and developmental timing for each tooth position (figures 1, 3 and 4). This pattern of spatio-temporal expression identifies new tooth germs added relative to preexisting ones, in precise locations at sequential times, from one dentate region to another (table 1).
Shh expression is first observed in the odontogenic fields beginning at stage 37 (figure 1d; electronic supplementary material, figure S4b). Strong expression loci on the odontogenic band occur first as focused placodes (stages 39–41; figure 1a–e), then expression as a cap around the cone of the tooth tip (figure 3p2 and 4b; electronic supplementary material, figure S4c–h). These loci mark tooth positions within one row (figure 1; electronic supplementary material, figure S4i–p). Shh expression is next upregulated at alternate (second) tooth positions, within this same row (figure 1a–c,e, arrows). By stage 43, shh is downregulated in epithelial cells of older tooth germs around tooth cones. Accurate counts of tooth number from shh expression at these later stages relies on seeing tooth cones (using Nomarsky optics). Nevertheless, differences in total number between upper and lower jaws are observed (table 1); for example, at stages 40 and 42 there are more tooth loci on the dentary than dermopalatine (compare electronic supplementary material, figure S4g,h (new parasymphysial tooth on dentary) with figure 4e,f and i–m (new loci added distally) with n).
Given this recognizable developmental sequence of epithelial shh expression, sites of tooth initiation can be identified along the rostro-caudal body axis. In both jaws at stages 39–40, there are four to five tooth buds in each dentary and dermopalatine field, contrasting with lack of tooth buds in more caudal toothed sites (electronic supplementary material, figure S4c–h). Later, at stage 41 the dentary and dermopalatine have seven tooth positions with alternating higher intensity of shh expression, and a new distal and proximal tooth germ, all in the same tooth row (figure 1e,i, arrows). As well, two shh-positive tooth loci are present on the first hypobranchial and the palatopterygoids (figure 1a–c,e, arrowheads). In stages 42–43, these shh expression sites are intense caps around the tooth cone (electronic supplementary material, figures S4la and lb), forming rings in later stages where shh is downregulated in cap cells (electronic supplementary material, figure S4o–p, further details see §3c,d and figure 4).
(b) bmp4 expression maps timing of cooperative activity during tooth morphogenesis (stages 40–45, 1dps)
All stages show bmp4 expression associated with each tooth locus (figure 1f–i; electronic supplementary material, figure S5). When compared to stage-matched specimens stained for shh, intense expression of bmp4 appears associated with mesenchyme of the newest forming tooth loci (figure 1g,i, arrows). Notably, stage 41 and 45 bmp4 expression shows upregulation in alternate positions of (second) tooth germs within the tooth row, on the dentary and dermopalatine, while the most rostral (first) tooth germs are dentine cones with bmp4 downregulated in the papilla. Note these show strong papillary expression in more caudal sites, indicating that these are younger (figure 1f,g,i; asterisk versus arrows, respectively). However, on the palatopterygoid, the intense papillary bmp4 expression of the younger loci is rostral to the dentine cones, as observed in the expression pattern for shh (i.e. an opposite second tooth addition pattern to the dentary and dermopalatine, electronic supplementary material, figure S5h’, st 42, 5l’, st 45, arrows.
(c) Cellular expression of shh during tooth germ morphogenesis, stage 45
The exact location of expression within the epithelial tooth germ is shown in more detail in serial, parasagittal sections than in whole mount in situ (figure 3; electronic supplementary material, figure S4), while the mesenchyme of the dental papilla shows complimentary bmp4 expression (electronic supplementary material, figure S6). Gene expression changes are associated with different tooth germ morphologies through development (figures 3p and 4), where different intensities are associated with specific timing of morphogenesis at each tooth site in the oropharyngeal cavity, including first locations of the sites on the branchial arches. These demonstrate a rostro-caudal activation gradient of tooth initiation for each dentate field. Initially, the placode shows intense shh and bmp4 expression and is superficial (no dental lamina), with shh located to the middle epithelial cells (figure 3d,i,p1). In the cap stage, shh is more intense in all epithelia, surrounding the papilla (figure 3g,p2; bmp4, electronic supplementary material, figure S6b). After dentine histogenesis, shh is downregulated in the cap cells but is strongly expressed in the epithelium as a collar around the tooth cone (cone + collar stage, figures 3c,p3, 4c). Subsequently, shh is downregulated around the whole tooth cone (figure 3j,n,p4), but within the adjacent dental epithelium (not the inner dental epithelium), shh is upregulated as an intense focal expression, attributed to an incipient, successive tooth germ (figures 3j,n, 4d; electronic supplementary material, figure S6a,c,d). In the second, alternate tooth position the same steps of shh expression are observed, including cap and cone + collar stages (figure 3f,h,k).
Serial sections show these expression stages simultaneously throughout the oropharyngeal cavity. Loci of shh expression occur dorsally on the dermopalatine and palatopterygoid (figure 3a–d,g), and ventrally on the dentary and 1st hypobranchial (figure 3e,h,i; electronic supplementary material, figure S6a), along with a focal spot on the infrapharyngobranchials dorsally and 1st and 2nd hypobranchials ventrally (figure 3e,h,i; electronic supplementary material, figure S6a,d), but a field of expression on the more caudal branchial arches (figure 3e). When dentine is present in the first dentary teeth, as a collar plus translucent cone, the more caudal, second tooth germ is only at the placode stage (figure 3f). In other sections, the first tooth appears as a translucent dentine cone with a second tooth at cap, or collar stage (figure 3h). All these observations show a staggered time difference in each second tooth germ, as well as the first (bmp4 data, electronic supplementary material, figure S6c,d). Similar staggered stages are seen in the dermopalatine tooth germs, and those of the palatopterygoid relative to the dermopalatine (figure 3m,n,g).
The restriction of shh expression to an intense focal locus (placode) forms first in the evaginated epithelium above the cartilage on the 2nd, as in the 1st, hypobranchial (figure 3l). The placode is superficial (i.e. forms without a dental lamina; figure 3n,o,p1), but also evaginated at the cone-cap stages (figure 3h,p2), then just within the expanded dental epithelium at cone + collar stage (figure 3k,n,p3). When shh is downregulated in all dental epithelium around the tooth there is an upregulated intense locus of shh expression next to this first tooth, in the dental epithelium, ‘cone + bud’, not evaginated but located in the epithelium adjacent to the dentine cone. Papillae with taste buds on the inner oral epithelium always exhibit faint shh expression, similar in intensity to the downregulated collar epithelium (figures 3o, arrow and 4c, sensory papilla with differentiated cells), while bmp4 expression is absent (electronic supplementary material, figure S6).
(d) Skeletal preparations show tooth addition positions in 7dps larvae
(i) Tooth development on upper jaw, dorsal branchial skeleton
Tooth rows are present ventral to the upper jaw cartilage, both rostrally on the dermopalatine bone and caudally on the palatopterygoid. The dermopalatine has 17–20 ankylosed teeth, while the latter lacks an independent ossification at this stage, with teeth conjoined by the individual bone of attachment of each tooth (translucent rings, figure 2a,c,g,h). Caudal to the palatopterygoid are two paired patches of teeth, the first associated with the hyoid arch with six teeth, joined only by their bases (figure 2c, black arrow, white box, k). The second is associated with the second infrapharyngobranchial, possessing four teeth (figure 2c, white box, j). The dermopalatine bone represents the most developmentally advanced in the upper jaw with new unattached teeth being added caudal to tooth positions 2 and 4, as well as parasymphysially (arrows, figure 2g,i). On the ventral surface of the palatopterygoid cartilage, the oldest teeth are joined together via attachment bone (dentine cones expanded into cylinders), with 11 teeth on the right side, nine on the left. As opposed to the caudal tooth addition associated with the dermopalatine, two new teeth (lacking bony rings; figure 1g,h, arrows) are rostral to the attached (older) teeth.
(ii) Tooth development on lower jaw, ventral branchial skeleton
Tooth rows are present dorsally on Meckel's cartilage, with 22 left and 21 right teeth fused to the dentary bone via bone of attachment with new, unattached teeth caudal to the attached (older) teeth and at proximal and distal ends of the row (Mc, figure 2d,e, arrows). Other toothed plates are caudal to Meckel's cartilage in the pharyngeal cavity, on the hypobranchials (first, 11 teeth; second, three teeth). Hypobranchial teeth are not ankylosed to bone but older teeth are joined at their bases via their individual bone of attachment (figure 2d,f,l). Three new teeth (not joined by bone of attachment) on left hypobranchial 1 are added caudally (figure 1f, arrows). By later functional stages, with increasing tooth numbers at all sites, pharyngeal teeth are arranged in radial rows (four to five teeth in each), differing from the oral dentition (electronic supplementary material, figure S2a,b).
Combined data from ontogenetic stages of P. spathula establishes sequences of gene expression and tooth morphogenesis in the oropharyngeal cavity, allowing spatio-temporal patterns of tooth initiation and development to be documented; tooth rows form on the dentary and dermoplatine before the more caudal palatopterygoids and first hypobranchials (figures 1a,c and 2a,b, respectively). Teeth are later organized into toothed plates, connected together by basal bone of attachment, independently of the membrane bone, representing early functional surfaces of the oropharyngeal dentition (electronic supplementary material, figure S2c). Skeletal whole mounts show where new teeth are added to individual dentate fields, while post-larval stages indicate that tooth addition slows and teeth are lost (electronic supplementary material).
These observations indicate progressive rostral–caudal and ventro-dorsal tooth initiation/addition gradients within the oropharyngeal cavity: tooth addition occurs first on Meckel's cartilage, showing alternate patterns of gene expression along the tooth row, prior to the dermopalatine (stage 40, electronic supplementary material, figure S4g,h; stage 42, electronic supplementary material, figure S4i–m versus n). At 7dps, a larger number of teeth are present on the dentary (figure 2) and at later juvenile stages the dentary shows substantial toothless areas of membrane bone relative to other dentate regions in the oropharyngeal cavity, due to tooth-related loss of attachment bone (electronic supplementary material, figures S1f–i and S3c–f, asterisk). With respect to a rostral–caudal gradient of tooth addition, the dentary and dermopalatine develop tooth germs with a cone of dentine before the palatopterygoid (electronic supplementary material, figure S5), while teeth in the oral cavity develop before those in the pharyngeal cavity. There is also a rostral–caudal progression in the pharyngeal cavity with the placode stage attained in hypobranchial 1, versus field expression on hypobranchials 2. The former has the most teeth; caudally hypobranchials 3 and 4 never show upregulated tooth loci. With respect to rostro-caudal tooth addition on each oral site, new tooth buds are initiated caudally on the dermopalatine and the dentary (figure 1i, electronic supplementary material, figure S5m,n), but new teeth form rostrally on the palatopterygoid (figure 2g,h).
Our results show that shh and bmp4 expression data during Polyodon tooth initiation follows the same spatio-temporal order observed in all other non-mammalian vertebrate species assayed to date [8,9,14–17,24]; however, our observations on the ordered sequence of timing of tooth germ initiation in oral and pharyngeal tooth sets also reveal directed rostro-caudal and ventro-dorsal patterns. This graded progression has not previously been reported for actinopterygians, or for gnathostome oropharyngeal dentitions. Nevertheless, tooth patterning, at least with respect to tooth initiation and differentiation appears evolutionarily stable and highly conserved among gnathostomes. For example, no differences in collocation of shh and bmp4 expression were detected between developing oral and pharyngeal teeth in Polyodon, comparable to a variety of other taxa. Along with the ordered tooth initiation sequence, this implies that tooth germs in all regions are equivalent and conserved modular vertebrate units.
We have demonstrated cellular partitioning for shh and bmp4 expression and sequential stages of tooth germ morphogenesis from ‘placode’, ‘cap’, ‘cone + collar’ to ‘cone + bud’ (figure 4). This is based on expression intensity that changes in a characteristic sequence within the dental epithelium, for each developing tooth germ. Notably, a new locus for strong expression forms alongside the developed, functional tooth (‘cone + bud’). We interpret this as the incipient tooth germ representing what we term a successional tooth. This is distinct from superficial, initial tooth ‘placodes' and is consistent with observations that in actinopterygian fish, successional teeth form from the older tooth and not from a dental lamina . In some actinopterygian taxa (Cyprinidae, derived teleosts), functional and replacement teeth can be retained as a pair, particularly during larval stages, although the functional tooth is eventually lost with the replacement tooth moving into place . In Polyodon, by comparison, the functional tooth is retained and not lost in response to the presence of the successional tooth; the latter should therefore not be considered a replacement tooth per se. Tooth loss occurs much later in Polyodon in what appears to be a general reduction and loss in the oropharyngeal cavity. This suggests that the more typical osteichthyan dentition pattern, with tooth replacement, never happens and is altered at this early ontogenetic stage.
Despite the enormous diversity, the presence of teeth organized into a functional dentition is a shared feature among jawed vertebrates, undoubtedly one reason for their evolutionary success, allowing a variety of feeding niches to be exploited. This diversity is underpinned by a high degree of developmental genetic conservation, particularly in early development, in taxa such as trout [7,8], cichlids [10,24] and the pufferfish ; these early patterns are also seen in sarcopterygian fish Neoceratodus  as well as the shark Scyliorhinus [17,26]. This conservation is also present in the dentition of P. spathula, with modifications early in development, including tooth retention and lack of replacement teeth. Tooth addition slows, while in Acipenser, teeth are lost, entirely linked to suction feeding adaptations [1,2]. However, we currently lack information on candidate genes involved in tooth regeneration that may change, or be missing in Polyodon and Acipenser ; other basal taxa, such as Polypterus, show full dentitions with tooth replacement . New analysis of genes directed towards key transitions from tooth initiation to replacement in P. spathula will offer insight into the evolution of tooth regeneration strategies and dental diversity. Modifications to the dentition that occur later in ontogeny, allow the diversity of vertebrate dentitions to be expressed , and are the precursor steps to the development of drastically different modes of feeding among the gnathostomes.
All animal care, feeding and euthanization protocols were in accordance with an approved IACUC (Institutional Animal Care and Use Committee) Animal Care Protocol [KSU #12-001; NSF IOS 1144965].
Drishti files for scans of Polyodon spathula are available at http://chondrichthyes.myspecies.info/.
We thank the Natural Environmental Research Council (NERC grant nos. NE/K01434X1, NE/K014595/1 and NE/K0122071/1) and National Science Foundation (IOS 1144965 to M.C.D.) for financial support.
We would like to thank the following curators for access to specimens in their collections: Oliver Crimmen and James MacLaine (NHM, London); Radford Arindell and Barbara Brown (AMNH). We would also like to thank Dan Sykes, and Monique Welten (NHM), and Christopher Healy (KCL) for help with CT scanning and James Massey (KSU) for help with bmp4 specimen preparation. Additionally, we thank the Kahrs family and Osage Beach Catfisheries for continued support of paddlefish research.
- Received November 3, 2014.
- Accepted February 19, 2015.
© 2015 The Authors. Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited.