A comparative examination of odontogenic gene expression in both toothed and toothless amniotes

Abstract

A well-known tenet of murine tooth development is that BMP4 and FGF8 antagonistically initiate odontogenesis, but whether this tenet is conserved across amniotes is largely unexplored. Moreover, changes in BMP4-signaling have previously been implicated in evolutionary tooth loss in Aves. Here we demonstrate that Bmp4, Msx1, and Msx2 expression is limited proximally in the red-eared slider turtle (Trachemys scripta) mandible at stages equivalent to those at which odontogenesis is initiated in mice, a similar finding to previously reported results in chicks. To address whether the limited domains in the turtle and the chicken indicate an evolutionary molecular parallelism, or whether the domains simply constitute an ancestral phenotype, we assessed gene expression in a toothed reptile (the American alligator, Alligator mississippiensis) and a toothed non-placental mammal (the gray short-tailed opossum, Monodelphis domestica). We demonstrate that the Bmp4 domain is limited proximally in M. domestica and that the Fgf8 domain is limited distally in A. mississippiensis just preceding odontogenesis. Additionally, we show that Msx1 and Msx2 expression patterns in these species differ from those found in mice. Our data suggest that a limited Bmp4 domain does not necessarily correlate with edentulism, and reveal that the initiation of odontogenesis in non-murine amniotes is more complex than previously imagined. Our data also suggest a partially conserved odontogenic program in T. scripta, as indicated by conserved Pitx2, Pax9, and Barx1 expression patterns and by the presence of a Shh-expressing palatal epithelium, which we hypothesize may represent potential dental rudiments based on the Testudinata fossil record.

Fig. 1. Conserved expression domains of early tooth development genes in the red-eared slider turtle T. scripta

(a-d) For reference, normal facial development in T. scripta from Y13-16. (e-h) Expression of Pitx2. (e-g) From Y13-Y15, Pitx2 is expressed broadly throughout the epithelium, (h) but by Y16 its expression is limited to a continuous band in the jaws. (i-l) Expression of Barx1. (i-j) From Y13-Y14, Barx1 is expressed proximally in oral region of both the upper and lower jaws as well as in the proximal, aboral region of the upper jaws. (k) At Y15, Barx1 expression is lost from the proximal oral region of the jaws, but persists in the proximal aboral region of the upper jaw as well as on the edges of the closing choanae. (l) By Y16, Barx1 expression continues to be prominent in the proximal outer upper jaw as well as on the edges of the closing choanae. (m-p) Expression of Pax9. (m-n) From Y13-Y14, Pax9 is expressed in the proximal region of the upper and lower jaws, as well as in the distal region of the frontonasal prominence. (o-p) From Y15-Y16, Pax9 is expressed broadly throughout the oral cavity. Scale bar = 1mm.

Fig. 2. Bmp4, Msx1, and Msx2 expression is missing from the proximal region, and Fgf8 expression is missing from the distal region, of the T. scripta mandible during the putative initiation period of odontogenesis

(a-d) Expression of Bmp4. (a-c) mRNA transcripts of Bmp4 are found in the distal-most region of the developing mandible only. (d) By Y16, there is diffuse Bmp4 expression throughout the mandible. (e-h) Expression of Msx1. (e-f) At Y13 and Y14, Msx1 expression is limited to the distal-most region of the developing mandible. (g) By Y15, the Msx1 domain has become more diffuse and spread throughout the entire lower jaw; (h) however, by Y16, mandibular Msx1 expression has largely disappeared. (i-l) Expression of Msx2. (i-k) From Y13-Y15, Msx2 is expressed only in the distal-most region of the mandible. (l) By Y16, Msx2 is expressed more broadly and diffusely in the distal mandible. (m-p) Expression of Fgf8. (m-n) At Y13 and Y14, Fgf8 expression is found only in the most proximal regions of the developing mandible. (o-p) By Y15 and Y16, Fgf8 expression has disappeared from the lower jaw. Scale bar = 1mm.

Fig. 3. A limited Bmp4 and Fgf8 domain is present in embryonic opossum and alligator mandibles, respectively, despite that both taxa possess teeth as adults

(a-e) Comparative expression of Bmp4 across stage- matched amniotes. Expression of Bmp4 is limited proximally in Y14 T. scripta (a), HH22 G. gallus (b) and e30 M. domestica (d), in comparison to the broader Bmp4 domain found in both E10.5 M. musculus (e) and F13 A. mississippiensis (c). (f-j) Comparative expression of Fgf8 across stage-matched amniotes. Expression of Fgf8 is limited distally in Y14 T. scripta (f) and F13 A. mississipiensis (h). Fgf8 is expressed broadly in the proximal mandible of HH22 G. gallus (g), e30 M. domestica (i) and E10.5 M. musculus (j). Phylogenetic relationships after Murphy et al. (2001) and Hedges and Poling (2002). Scale bar = 1mm.

Fig. 4. Msx domains in embryonic opossum and alligator mandibles differ markedly from those found in mice

(a-e) Comparative expression of Msx1 across stage-matched amniotes. Expression of Msx1 is limited proximally in (a) Y14 T. scripta, (b) HH22 G. gallus and (c) F13 A. mississipiensis. (d) Msx1 is expressed broadly along the proximal-to-distal axis of the e30 M. domestica mandible. (e) Msx1 is expressed in the distal mandible of e30 M. musculus. (f-j) Comparative expression of Msx2 across stage-matched amniotes. (f) Msx2 expression is limited proximally in Y14 T. scripta. (g) Msx2 expression is limited proximally in HH22 G. gallus. (h) Msx2 is expressed broadly in the distal F13 A. mississipiensis mandible. (i) Msx2 is missing from the entire odontogenic region of the e30 M. domestica mandible, although its expression appears prominently in the proximal mandible underlying the odontogenic region. (j) Msx2 is expressed broadly in the distal E10.5 M. musculus mandible. Scale bar = 1mm.

Fig. 5. Epithelium of Y17 T. scripta palate is marked by Shh expression

(b, d, f, h, j) Shh expression in a whole mount Y17 T. scripta embryo that was subsequently dehydrated, paraffin-sectioned, and stained with Eosin Y. (c, e, g, i) Morphologically comparable sections clipped from a μCT scan of a Y17 T. scripta embryo accompany the gene expression sections (μCT video available in the supplementary material). (a) Gross morphology of a Y17 T. scripta embryonic head photographed from an anterior angle. (b) Shh expression marks the developing palate in a whole mount Y17 T. scripta embryo photographed from an anterior angle. (c-d) Shh expression marks the edges of the open choanae epithelium as well as two localizations of palatal epithelium. (e-f) Shh expression marks the epithelium where the choanae have closed as well as two patches of palatal epithelium labial to the choanae. (g-h) Shh expression marks the epithelium of the palate in a continuous line. (i-j) Although the accompanying μCT scan image reveals invaginations of palatal epithelium, Shh expression is missing from this region. Scale bar = 500um.

Fig. 6. The labial-to-lingual sequential loss of tooth rows in the turtle lineage

Evidence in the paleontological record suggests that the turtle lineage became edentulous in a stepwise fashion: first losing the outer-most row of maxillary, premaxillary and dentary teeth, last recorded in (a) Odontochelys semitestacea, 220 Mya (Li et al., 2008; figure adapted from same reference); then losing rows from the vomer and palatine bones, as shown in (b) Proganochelys quenstedti, 210 Mya (Gaffney and Meeker, ‘83; Gaffney and Jenkins, ‘90; figure adapted from Gaffney and Meeker, ‘83), and finally losing the innermost pterygoid teeth, present in (c) Kayentachelys aprix, ~174-201Mya (Gaffney et al., ‘87; figure adapted from Gaffney and Jenkins, 2009) and Paleochersis talampayensis, ~201-235Mya (not shown, Rougier et al., ‘95). From at least the late Jurassic, all turtle fossils described to date have been edentulous (Meredith et al., 2013), such as the (d) Chelydra serpentina specimen pictured here (Creative Commons).