Evolution and Development of the Mammalian Dentition: Insights from the Marsupial Monodelphis Domestica

DEVELOPMENTAL DYNAMICS 240:232–239, 2011

a PATTERNS & PHENOTYPES

Evolution and Development of the Mammalian Dentition: Insights From the Marsupial Monodelphis domestica

Jacqueline E. Moustakas,1* Kathleen K. Smith,2 and Leslea J. Hlusko3

To understand developmental mechanisms of evolutionary change, we must first know how different morphologies form. The vast majority of our knowledge on the developmental genetics of tooth formation derives from studies in mice, which have relatively derived mammalian dentitions. The marsupial Monodelphis domestica has a more plesiomorphic heterodont dentition with incisors, canines, premolars, and molars on both the upper and the lower jaws, and a deciduous premolar. The complexity of the M. domestica dentition ranges from simple, unicusped incisors to conical, sharp canines to multicusped molars.WeexaminethedevelopmentoftheteethinM. domestica, with a specific focus on the enamel knot, a signaling center in the embryonic tooth that controls shape. We show that the tooth germs of M. domestica express fibroblast growth factor (FGF) genes and Sprouty genes in a manner similar to wild-type mouse molar germs, but with a few key differences. Developmental Dynamics 240:232–239, 2011. VC 2010 Wiley-Liss, Inc.

Key words: tooth development; enamel knot; FGF; Sprouty; heterodont; marsupial; Shh; tribosphenic

Accepted 23 October 2010

INTRODUCTION in tooth type along the dental arcade and patterning in the gray short- (only incisors and molars). Most mam- tailed opossum Monodelphis domes- A central question in evolutionary mals possess a more heterodont denti- tica (Fig. 1a). M. domestica is a small,

Developmental Dynamics morphology and developmental biol- tion, in which there are several dis- easily bred animal that has been used ogy is how diversity in the shape and tinct types of teeth. Tooth shapes in extensively as a research organism size of structures is achieved. The mammals can range from a simple (Keyte and Smith, 2008) and whose mammalian dentition has proved to be conical shape (such as canines) to the genome has been sequenced recently a productive system to study the devel- complex arrangements of cusps seen (reviewed in Mikkelsen et al., 2007). opmental generation of complex struc- in molars. To understand the genera- Importantly, M. domestica has a tures and to model mechanisms of tion of diversity of tooth types across complete heterodont dentition, includ- morphological change. Thus far, most mammals broadly it is critical to study ing incisors, canines, premolars and molecular and genetic studies of den- an animal that possesses a more typi- molars, as well as a deciduous premo- tal development have focused on cal mammalian heterodont dentition. lar (although all premolars are gener- rodents, in particular the mouse Mus In this study, we present data on ally considered to be deciduous (Luck- musculus. Mus, however, has a fairly the expression of major genes known ett, 1993), only the third is replaced uniform dentition, with little variation to be important in tooth development in marsupials). M. domestica also

1Department of Integrative Biology and Museum of Paleontology, University of California, Berkeley California 2Department of Biology, Duke University, Durham, North Carolina 3Department of Integrative Biology, University of California, Berkeley California Grant sponsor: NSF; Grant numbers: BCS-0616308, IBN 0316353. *Correspondence to: Jacqueline E. Moustakas, Institute of Biotechnology, P.O. Box 56 (Viikinkaari 9), FIN-00014 University of Helsinki, Finland. E-mail: jacqueline.moustakas@helsinki.ﬁ DOI 10.1002/dvdy.22502 Published online 24 November 2010 in Wiley Online Library (wileyonlinelibrary.com).

VC 2010 Wiley-Liss, Inc. OPOSSUM TOOTH DEVELOPMENT 233

retains a tribosphenic dentition, con- hog (Shh) expression in the EK of genes (de Maximy et al., 1999), with sidered a key innovation of some of mammalian tooth germs has been Spry genes 1, 2,and4 having expres- the earliest mammals (Luo, 2007), in documented in placental mammals, sion in several organs in the develop- which the cusps of the upper teeth including mouse (Vaahtokari et al., ing mouse embryo (Minowada et al., occlude with a talonid basin formed 1996), vole (Kera¨nen et al., 1998), 1999; Zhang et al., 2001). The expres- by smaller cusps on the lower teeth shrew (Yamanaka et al., 2007), and sion patterns of these Sprouty genes (Fig. 1a). Data from M. domestica ferret (Ja¨rvinen et al., 2009). Evi- suggest roles in epithelial–mesenchy- therefore can provide critical informa- dence from wild-type and mutant mal signaling interactions (Zhang tion on patterning heterodont denti- mice suggests that a functional EK is et al., 2001). tion, deciduous dentition, and the induced/maintained by signaling primitive tribosphenic condition. between epithelium and mesenchyme by means of fibroblast growth factors Comparing Mouse Dental (FGFs; Kettunen et al., 2000) and GENETICS OF MOUSE Developmental Genetics to that this signaling is modulated by Monodelphis DENTAL DEVELOPMENT negative feedback regulators of FGF Early tooth development establishes and other receptor tyrosine kinase Most investigations of the develop- the type of tooth at a specific location (RTK) signaling, encoded by Sprouty mental genetics of tooth morphogene- and the morphology or shape of the genes (Klein et al., 2006). sis have used mice as a model system. tooth (Jernvall, 1995). Tooth develop- Members of the FGF family of Mice are placental mammals with a ment proceeds through a series of secreted intercellular signaling mole- reduced dentition that is composed of morphogenetic movements and sig- cules affect organ development through incisors and molars that are sepa- naling interactions between ectoder- regulating cell proliferation and differ- rated by a toothless diastema. Data mal epithelium and neural crest- entiation (reviewed in Martin, 1998). from studies on rodents with different derived ectomesenchyme. The first FGFs have important roles in the de- molar tooth morphologies (mice; indication of tooth development is the velopment of several vertebrate organs, [Jernvall et al., 1994; Vaahtokari formation of the dental lamina, a including the lungs (Peters et al., et al., 1996; Kettunen and Thesleff, thickening of the epithelium that 1994), the kidneys (Karavanova et al., 1998; Kettunen et al., 2000] and voles marks the future dental arch. A den- 1996), the limbs (reviewed in Wilkie [Kera¨nen et al., 1998]) have shown tal placode forms as an epithelial et al., 2002), the brain (reviewed in that the EKs of the molar tooth pri- thickening along the dental lamina at Iwata and Hevner, 2009), the hair mordia express some of the same de- the future position of the tooth. This (Hebert et al., 1994), the turtle shell velopmental regulatory genes. How- thickening grows and begins to inva- (Cebra-Thomas et al., 2005; Mousta- ever, we know very little regarding ginate into the underlying ectomesen- kas, 2008), the feathers (Widelitz et al., the developmental expression of other chyme, which then condenses around 1996), and the teeth (reviewed in Jern- tooth classes. the epithelium to form a tooth bud. vall and Thesleff, 2000). FGFs exert In this study, we examine expression The epithelium then folds and their biological effects through four patterns of genes shown to be impor-

Developmental Dynamics extends farther into the mesenchyme, high-affinity tyrosine kinase receptors tant in generating mouse molar teeth forming a cap and then a bell stage (FGFR1–4; reviewed in Wilkie et al., in mice in all tooth types in tooth germ. The tooth itself is com- 2002). Studies on mice mutant for FGF M. domestica. We first examine the posed of both layers with the amelo- receptors have shown that FGFs are timing of dental lamina formation in blasts formed from the epithelium necessary for tooth morphogenesis to M. domestica. Shh and Fgf8 are and the odontoblasts differentiating proceed from the bud stage to the cap expressed in the dental lamina in from the ectomesenchyme. The ame- stage (Celli et al., 1998; De Moerlooze mouse (Bitgood and McMahon, 1995; loblasts deposit enamel and the odon- et al., 2000). Kettunen and Thesleff, 1998). Shh is toblasts secrete dentine (reviewed in Feedback control of biological signal- also expressed in the dental lamina of Jernvall and Thesleff, 2000). An em- ing pathways is an important mecha- snakes (Buchtova et al., 2008), shrews bryonic signaling center that controls nism for ensuring the spatial and tem- (Yamanaka et al., 2007), and catsharks tooth shape, the enamel knot (EK), is poral regulation of cell proliferation (Smith et al., 2009), suggesting a induced at the cap stage (Jernvall and differentiation. Sprouty proteins strong conservation at this early stage et al., 1994; Fig. 1b). The EK is are negative-feedback regulators that of tooth development and a role in the thought to control tooth shape by con- inhibit signaling by RTKs (reviewed in generation of replacement teeth. trolling the differential growth and Kim and Bar-Sagi, 2004). The spry We then focus on genes that regu- folding of the epithelium (Jernvall gene was first identified as a negative late morphogenesis at the cap stage of and Thesleff, 2000). feedback regulator of FGF-mediated tooth development, when shape is The primary EK is a transient, non- tracheal branching in Drosophila being controlled. Specifically, we de- proliferative cluster of epithelial cells (Hacohen et al., 1998) and subsequent scribe the expression patterns of Fgf- that expresses several signaling mole- studies have shown that spry regu- 3, À10, and À4, Shh, and Spry-2 and - cules and has been shown to be essen- lates other RTK signaling pathways as 4 in cap-stage Monodelphis tooth tial for mammalian tooth develop- well (Casci et al., 1999; Kramer et al., germs. FGF signaling is thought to ment (Jernvall et al., 1994; 1999; Reich et al., 1999). The mamma- induce/maintain a functional EK Vaahtokari et al., 1996). Sonic hedge- lian genome contains four Sprouty (Klein et al., 2006) and control cell 234 MOUSTAKAS ET AL.

Fig. 1. a: Skull of Monodelphis domestica (UCMP 197900) showing the adult morphology and relative position of the tooth germs investi- gated, with the exception of the third premolars, which are the replacement rather than the deciduous teeth. M1, first upper molar; P3, third upper premolar; P2, second upper premolar; C1, upper canine; I5, fifth upper incisor; m1, first lower molar; p3, third lower premolar; p2, second lower premolar; c1, lower canine; i4, fourth lower incisor. b: Histological section of a M. domestica cap stage lower molar stained with hematoxylin and eosin, illustrating the morphology of the epithelium and mesenchyme. EK, enamel knot.

Fig. 3. a–e: Gene expression in cap-stage tooth germs of the upper jaw (maxillary arch) of Monodelphis domestica postnatal specimens. Tooth class/germ is indicated at top and gene Fig. 2. Formation of the dental lamina in examined is on left. Fgf3 is expressed in the enamel knot (EK) and mesenchyme of all tooth Monodelphis domestica. a,b: Whole-mount germs (a1–e1). Fgf10 is expressed in the EK and mesenchyme of M1 (a2), dP3 (b2), and P2 embryos assayed for Fgf8 (a) and Shh (b) (c2), in the epithelium and mesenchyme of C1 (d2), and the mesenchyme of I5 (e2). Fgf4 is expression by in situ hybridization, showing expressed in the EK (a3–e3). Shh is expressed in the EK (a4–e4). Spry2 is expressed in the the expression of these genes in the dental epithelium and mesenchyme (a5–e5). Spry4 is expressed in the epithelium and mesenchyme lamina at stage 31. (a6–e6). ep, epithelium; mes, mesenchyme.

proliferation around the EK (Jernvall As FGFs are necessary for the com- neonates (van Nievelt and Smith, and Thesleff, 2000). Sprouty genes are pletion of tooth morphogenesis (Celli 2005), all other teeth develop later.

Developmental Dynamics expressed in developing mouse molar et al., 1998; De Moerlooze et al., The odontogenic epithelium, however, and incisor teeth, as well as in the 2000), we expect that basic conditions is induced in the embryo (Fig. 2) and toothless diastema (Klein et al., 2006, of FGF signaling and its modulation both Fgf8 and Shh are expressed in 2008). In mouse molar tooth germs, are conserved among the different the dental lamina of embryonic stage Spry2 is predominant in the epithe- classes of teeth. However, we hy- 31 (approximately 2 days before birth; lium and Spry4 is predominant in the pothesize that we will find differences Fig. 2a,b). mesenchyme at the cap-stage of tooth in the expression patterns of these development (Klein et al., 2006). Null genes in the different tooth classes mice mutant for Spry genes 2 or 4 de- that are attributable to different Expression of Fgf, Shh, and velop a tooth in the diastema region tooth morphologies. Our study is thus Sprouty Genes in the (Klein et al., 2006). The loss of Sprouty the first to examine these genes, Different M. domestica Tooth gene expression in the diastema buds known to be central in patterning Classes results in an increase in FGF and mammalian tooth form, in the full RTK signaling in the diastema bud ep- range of mammalian tooth classes, as We examined the cap stage of tooth ithelium and the formation of a func- well as in both deciduous and perma- development in postnatal specimens. tional EK and tooth (Klein et al., nent adult teeth. The cap stage occurs at 3 days post- 2006). The development of this mu- natal (3P) for tooth germs dP3 and tant tooth is intriguing because the dp3, 5P for M1 and m1, 7P–8P for C1 morphology of this diastema tooth is RESULTS and c1, 8P for P2 and p2, 9P–11P for similar to that of a premolar and could Formation of the Dental i4, and 11P–13P for I5 (van Nievelt therefore reflect the normal develop- and Smith, 2005; here). We describe Lamina in M. domestica ment of premolars. We describe gene expression patterns of Fgf3, expression in incisor, canine, decidu- The first placodes, or thickenings of Fgf10, Fgf4, Shh, Spry2, and Spry4 in ous and permanent premolar, and the dental lamina, have been observed cap stage germs of different tooth molar teeth. for the deciduous third premolars in classes. OPOSSUM TOOTH DEVELOPMENT 235

Upper and Lower First Molars Fgf3 and Fgf10 are expressed in the mesenchyme and EK of the ﬁrst upper molar tooth germ (M1), whereas Fgf4 is exclusively in the EK (Fig. 3a1-3). Shh is expressed in the EK in a broader pattern than Fgf4 (Fig. 3a4). Spry2 is expressed in the epithelium and the mesenchyme (Fig. 3a5). Spry4 is mostly expressed in the mesenchyme in M1, but also lingually in the epithelium (Fig. 3a6). As in the upper germ, Fgf3 and Fgf10 are expressed in the mesenchyme and EK of the ﬁrst lower molar tooth germ (m1; Fig. 4a1-2). Fgf4 and Shh are expressed in the EK with the expression of Shh being broader than that of Fgf4 (Fig. 4a3-4). Spry2 and Spry4 are expressed in both the epithelium and mesenchyme of m1 (Fig. 4a5-6).

Upper and Lower Premolars Fgf3 and Fgf10 are expressed in the mesenchyme and EK of the upper deciduous third premolar (dP3; Fig. 3b1-2). Fgf4 and Shh are expressed in Fig. 4. a–e: Gene expression in cap-stage tooth germs of the lower jaw (mandibular arch) of the EK, with Shh again having a Monodelphis domestica postnatal specimens. Tooth class/germ is indicated at top and gene broader domain of expression than examined is on left. Fgf3 is expressed in the enamel knot (EK) and mesenchyme of all tooth germs (a1–e1). Fgf10 is expressed in the EK and mesenchyme of m1 (a2), dp3 (b2), and p2 (c2), Fgf4 (Fig. 3b3-4). Spry2 and Spry4 are in the epithelium (and EK) and mesenchyme of c1 (d2), and the mesenchyme of i4 (e2). Fgf4 is both expressed in the epithelium and expressed in the EK (a3–e3). Shh is expressed in the EK (a4–e4). Spry2 is expressed in the mesenchyme of dP3; Spry2 is ex- Developmental Dynamics epithelium and mesenchyme (a5–e5). Spry4 is expressed in the epithelium and mesenchyme pressed throughout the epithelium, (a6–e6). whereas Spry4 expression is stronger in the tips of the epithelium (Fig. 3b5-6).

Fig. 5. The development of the primary enamel knot (PEK) and secondary enamel knots (SEKs) in the tribosphenic molars of Monodelphis domestica. a–d: Tooth and days postnatal are indicated at top and gene examined is on left. e–g: Shh is expressed in the PEK of the lower first molar (m1) from 5P–8P (a–d). Shh is expressed in the SEKs of m1 (e) and M1 (f,g). h–k: Fgf4 is expressed in the PEK of the lower first molar (m1) from 5P–8P. l–n: Fgf4 is expressed in the SEKs of m1 (l) and M1 (m,n). o–q: Fgf10 is expressed in the PEK and the mesenchyme of the lower first molar (m1) from 5P–7P. r–u: Fgf10 is expressed only in the mesenchyme of m1 at 8P (r) and M1 at 11P (s) and 13P (t,u). 236 MOUSTAKAS ET AL.

TABLE 1. Summary of Gene Expression Patterns in Mouse and Opossum Cap-Stage Tooth Germs

M. domestica M. domestica M. domestica M. domestica M. musculus molars incisors canines premolars Molars Gene Epithelium Mesenchyme Epi Mes Epi Mes Epi Mes Epi Mes Shh x xxxx Fgf4 x xxxx Fgf3 x x xx xx xx xx Fgf10 x x xx xx xx Spry2 x xxxxxxxx Spry4 x xxxxxxxX

In the lower deciduous third premo- rounding the mesenchyme of C1 and The PEK at the tip of the protoconid lar (dp3), Fgf3 and Fgf10 are c1 (Fig. 3d5,4d5). Spry4 is expressed in M. domestica persists until 8P expressed in both the mesenchyme in the mesenchyme of C1 and c1 and (Jernvall, 1995) and expresses Shh and EK (Fig. 4b1-2), whereas Fgf4 and in the tips of the epithelium (Fig. 3d6, and Fgf4 (Fig. 5a–d,h–k). SEKs of the Shh expression is conﬁned to the EK 4d6). lower molar also express Shh and (Fig. 4b3-4). As in dP3, Spry2 and Fgf4 (Fig. 5e,l). The epithelial expres- Spry4 expression is seen in both the sion of Fgf10 in the PEK persists Upper Fifth and Lower epithelium and mesenchyme of dp3, through 7P (Fig. 5o–q). At later with Spry2 expression seen through- Fourth Incisors stages, Fgf10 is no longer expressed in out the epithelium and Spry4 epithe- Fgf3 is expressed in the EK and mes- the dental epithelium; however, the lial expression being stronger in the enchyme of the upper ﬁfth incisor (I5; mesenchymal expression of Fgf10 is tips of the epithelium (Fig. 4b5-6). Fig. 3e1). maintained (Fig. 5o–s). We also examined the second premo- Fgf10 is expressed in the mesen- Shh and Fgf4 are also expressed in lars that form part of the permanent chyme of I5 (Fig. 3e2). Fgf4 and Shh the SEKs of the upper molar (Fig. dentition. In both the upper (P2) and are expressed in the EK of I5 (Fig. 5f,g,m,n). Fgf10 is expressed in the lower (p2) second premolars, Fgf3 and 3e3-4). Spry2 is expressed in the epi- mesenchyme, but not in the epithe- Fgf10 are expressed in the mesen- thelium and mesenchyme of I5 (Fig. lium, of the developing upper molars chyme and the EK (Fig. 3c1-2,4c1-2). 3e5) and Spry4 is expressed in the during SEK formation (Fig. 5t,u). Fgf4 and Shh are expressed in the EK mesenchyme of I5 (Fig. 3e6). of P2 (Fig. 3c3-4) and p2 (Fig. 4c3-4). In the lower fourth incisor (i4), Fgf3 Spry2 expression is in the epithelium is expressed in the mesenchyme and DISCUSSION and mesenchyme of P2 (Fig. 3c ) and 5 the EK (Fig. 4e1), Fgf10 is expressed Expression of Shh, Fgf, and p2 (Fig. 4c5). Spry4 is expressed in the in the mesenchyme (Fig. 4e2), and

Developmental Dynamics Sprouty Genes in Tooth mesenchyme and weakly in the epithe- Fgf4 and Shh are expressed in the EK Development Is Conserved lium of P2, with a lingual bias as in M1 (Fig. 4e3-4). Spry2 is expressed in the (Fig. 3c6). Spry4 is expressed in the epithelium and mesenchyme of i4 Between Marsupial and mesenchyme and the epithelium of p2, (Fig. 4e5) and Spry4 is expressed Placental Mammals, as Well with the mesenchymal expression strongly in the mesenchyme and as Across Tooth Classes being relatively stronger (Fig. 4c6). weakly in the epithelium of i4 (Fig. We have shown that key genes that 4e6). regulate tooth morphogenesis in Upper and Lower Canines mouse are also expressed in the devel- Fgf3 is expressed in the EK and mes- Primary and Secondary oping tooth germs of the marsupial enchyme of both the upper canine Enamel Knots in the Monodelphis domestica. Many of the (C1; Fig. 3d ) and lower canine (c1; domains of expression of the Fgf, Shh, 1 Tribosphenic Molars Fig. 4d1). Fgf10 is expressed in both and Sprouty genes that we examined the mesenchyme and epithelium of The primary enamel knot at the tip of in M. domestica tooth germs are con- C1 and c1, with the mesenchymal the protoconid in the M. domestica served with the expression patterns expression being stronger than the lower molar persists longer than the seen in mouse (Table 1). Fgf3, Fgf4, epithelial expression (Figs. 3d2,4d2). primary enamel knot of the lower- and Shh are expressed in the enamel The expression of Fgf10 is seen dif- cusped talonid (Jernvall, 1995). We, knots of M. domestica and mouse, and fusely in C1 (Fig. 3d2) and has strong therefore, examined the development Fgf3 and Fgf10 are expressed in the expression in the EK of c1 (Fig. 4d2). of the lower molar primary enamel mesenchymal papillae (Jernvall et al., Fgf4 and Shh are expressed in the EK knot (PEK) through the induction of 1994; Kettunen and Thesleff, 1998; of C1 and c1, with Shh expression the secondary enamel knots (SEK). Kettunen et al., 2000). In mouse inci- being broader than Fgf4 (Fig. 3d3-4, Secondary enamel knots form later in sor and molar tooth germs, Spry2 is 4d3-4). Spry2 is expressed in the mes- tooth development at the cusp tips expressed predominantly in the epi- enchyme and the epithelium sur- (Jernvall et al., 1994). thelium and Spry4 is expressed OPOSSUM TOOTH DEVELOPMENT 237

predominantly in the mesenchyme at hypothesize that the expression of the teeth considered deciduous or per- the cap-stage of tooth development Fgf10 in the epithelium of canine tooth manent. However, we describe one (Klein et al., 2006, 2008). In contrast, germs and in the primary enamel critical difference from previous stud- Spry genes 2 and 4 are more broadly knots of premolar and molar tooth ies, observed here for the first time. expressed in many of the tooth germs germs controls the faster rate of epi- Teeth that have exceptionally sharp of M. domestica than mouse, having thelial growth in these teeth in M. and tall cusps, specifically the canine, both epithelial and mesenchymal domestica and that this has led to dif- premolars, and molars, also have epi- domains of expression. Because the ferences in tooth morphology, such as thelial expression in the primary epithelial–mesenchymal cross-talk cusp height and sharpness. enamel knots of Fgf10. In mice, and in between these Fgfs and Sprouty genes tooth germs that form relatively low induces/maintains Shh expression in Sproutys 2 and 4 Are cusps in Monodelphis (the incisors), the EK at the cap stage of tooth devel- Expressed in Both the Fgf10 expression is limited to the mes- opment (Klein et al., 2006), we enchyme. We propose that the epithe- hypothesize that the function of FGFs Epithelium and Mesenchyme lial expression of Fgf10 leads to in cap stage tooth germs is conserved of M. domestica increased epithelial proliferation and between marsupials and placentals. In mice incisor and molar cap-stage therefore to increased height. tooth germs, Spry2 is expressed in the In addition, our data may contrib- Fgf10 Has Epithelial Domains epithelium and Spry4 is expressed in ute to discussions on the generation of the mammalian dental diastema. of Expression in M. domestica the mesenchyme (Klein et al., 2006, 2008). Because Spry2-null and Spry4- Whereas previous studies suggest In contrast to the many genes with null mice develop a tooth in the dia- that the expression of Spry2 and conserved expression, we found dif- stema, immediately anterior to the Spry4 is important in suppressing ferences in Fgf10 expression patterns lower first molar (Klein et al., 2006), dental development in the diastema in M. domestica tooth germs. In mice, it would be reasonable to hypothesize region, we observed these genes in all Fgf10 is only expressed in the mesen- that these genes contribute to the tooth germs. Our data thus suggest chyme (Kettunen et al., 2000); for M. suppression of premolar teeth and the that Sprouty expression does not nec- domestica, we observed Fgf10 expres- formation of the diastema in mice, as essarily lead to diastema formation. sion in the enamel knot and mesen- Spry2-null and Spry4-null mice de- Finally, our results emphasize the chymal papilla of the canine, premo- velop a tooth in the diastema, imme- need for comparative data, and in lar, and molar teeth. This is the first diately anterior to the lower first particular data on the plesiomorphic observation of epithelial Fgf10 ex- molar (Klein et al., 2006). Our results condition, before we can fully under- pression in tooth development, and are not consistent with this hypothe- stand the generation and evolution of given its variation in degree of sis, though, as we see Spry2 and tooth shape in mammals. expression, we hypothesize that it Spry4 expression in both the epithe- underlies variation in tooth shape. lial and mesenchymal compartments The enamel knot is an embryonic of all tooth classes. Therefore, Spry2 EXPERIMENTAL Developmental Dynamics signaling center that controls tooth or Spry4 expression does not neces- PROCEDURES shape (Jernvall et al., 1994) and is sarily lead to a diastema or suppres- Specimens essential for tooth morphogenesis. sion of premolar teeth. However, FGF10 has been shown to stimulate expression patterns for Spry2 and M. domestica is a member of the family cell proliferation in cultured mouse Spry4 in mice and M. domestica do Didelphidae and is found throughout molar dental epithelium (Kettunen differ. Unlike mice, in M. domestica the northern two-thirds of South et al., 2000) and to regulate cell sur- both genes are expressed in both the America (Streilein, 1982; Nowak, vival of incisor epithelium (Ohuchi mesenchyme and epithelium, and it is 1999; Macrini, 2004). The specimens et al., 2000; Harada et al., 2002). Pre- possible that these expression pat- used in this study were collected from vious studies (Jernvall, 1995) have terns may be important in maintain- a captive breeding colony maintained shown that the epithelium of M. ing a full heterodont dentition. at Duke University by K.K.S. (Keyte domestica lower molar tooth germs and Smith, 2008). M. domestica young grows faster relative to the mesen- are born after 14.5 days of gestation, CONCLUSION chyme. Experimental evidence, mod- begin to detach from the teat 10–12 els of pattern formation, and simula- This study is the first study of gene days after birth, and are weaned tion analyses have shown that expression patterns in the complete between 50 and 60 days after birth differences in tooth morphology result dentition of a mammal with a fully (Smith, 2006). The age of specimens is from differences in epithelial growth heterodont dentition, and provides given in embryonic stages (Mate et al., (Kera¨nen et al., 1998; Jernvall et al., critical information about the genera- 1994) or days postnatal with day of 2000; Salazar-Ciudad and Jernvall, tion of dental diversity. First, we show birth considered day 0P. The basic ele- 2002, 2004). In particular, the sharp- that the role of many of the genes stud- ments and timing of tooth develop- ness of tooth cusps has been proposed ied appears to be conserved, not only ment in M. domestica have been to be affected by changes in the rate evolutionarily but across tooth classes described in detail (van Nievelt and of epithelial growth (Salazar-Ciudad of different dental types. Furthermore, Smith, 2005). Postnatal specimens and Jernvall, 2010). Therefore, we we see little difference in patterning in used in this study ranged from three 238 MOUSTAKAS ET AL.

days postnatal to thirteen days post- Transcriptase Kit (QIAGEN, Valencia, Klein, Amy Maxmen, Alexis Lainoff, natal to capture the cap stage of all CA). Polymerase chain reaction (PCR) and Chris Conroy. J.E.M. thanks dental classes (van Nievelt and Smith, was used to isolate genes using plati- Kevin Padian. L.J.H. and K.K.S. were 2005). All animal care was approved num Pfx DNA polymerase (Invitrogen, funded by NSF grants. This is UCMP by the Duke University Institutional Carlsbad, CA). Fragments were cloned contribution no. 2017. Animal Care and Use Committee in into the pCR4Blunt-TOPO vector accordance with the established guide- (Invitrogen) using the Zero Blunt lines (National Research Council, TOPO PCR Cloning Kit for Sequenc- REFERENCES 1996). ing (Invitrogen). Orthologous genes Bitgood MJ, McMahon AP. 1995. Hedge- The upper dentition is denoted with were identified by BLAST and hog and Bmp genes are coexpressed at a capital letter, the lower dentition sequence alignment comparison in Se- many diverse sites of cell-cell interac- with a lower-case letter. For example, Al v2.0a11 (Rambaut, 1996). The tion in the mouse embryo. Dev Biol 172:126–138. the first upper molar is denoted ‘‘M1’’ aligned data matrices included orthol- Buchtova M, Handrigan GR, Tucker AS, and the second lower premolar is ogous and paralogous sequences of Lozanoff S, Town L, Fu K, Diewert VM, denoted ‘‘p2’’. Although all premolars multigene families to identify homo- Wicking C, Richman JM. 2008. Initia- in marsupials are generally consid- logues. The sequences for Fgf8 (387 tion and patterning of the snake dentition are dependent on Sonic hedgehog ered to be of the deciduous generation base pairs), Fgf4 (381 bp), Fgf3 (145 signaling. Dev Biol 319:132–145. (Luckett, 1993), the first and second bp), Spry2 (369 bp), and Spry4 (535 bp) Casci T, Vino´s J, Freeman M. 1999. are not replaced and contribute to the are deposited in GenBank under the Sprouty, an intracellular inhibitor of adult dentition. In marsupials, only following accession numbers: Fgf8 Ras signaling. Cell 96:655–665. the third premolar has a deciduous GU984788, Fgf4 GU984791, Fgf3 Cebra-Thomas J, Tan F, Sistla S, Estes E, Bender G, Kim C, Riccio P, Gilbert SF. tooth that is later replaced by a sec- GU984787, Spry2 GU984789, and 2005. How the turtle forms its shell: a ond generation. Deciduous teeth are Spry4 GU984790. paracrine hypothesis of carapace forma- denoted by the prefix, ‘‘d’’. tion. J Exp Zool B Mol Dev Evol 304: M. domestica retains the primitive 558–569. In Situ Hybridization marsupial dental formula consisting Celli G, LaRochelle WJ, Mackem S, Sharp R, Merlino G. 1998. Soluble dominant- of five upper and four lower incisors, RNA probes were labeled with digoxi- negative receptor uncovers essential one upper and one lower canine, three genin-11-UTP (Roche, Indianapolis, roles for fibroblast growth factors in upper and three lower premolars, and IN). M. domestica Fgf10 and Shh plas- multi-organ induction and patterning. four upper and four lower molars in mids were kindly provided by A. Keyte EMBO J 17:1642–1655. de Maximy AA, Nakatake Y, Moncada S, each jaw quadrant (I 5/4, C 1/1, P 3/3, (GenBank accession numbers Fgf10 Itoh N, Thiery JP, Bellusci S. 1999. Clon- M 4/4; Fig. 1a). The incisors in M. GU593350 and Shh GU593352). Em- ing and expression pattern of a mouse domestica are simple, unicusped teeth bryonic specimens were fixed in 4% homologue of Drosophila sprouty in the that are flattened medio-laterally. paraformaldehyde in phosphate-buf- mouse embryo. Mech Dev 81:213–216. The canines are the tallest teeth of fered saline, dehydrated in methanol, De Moerlooze L, Spencer-Dene B, Revest À JM, Hajihosseini M, Rosewell I, Dick- the dental arcade and are formed by a and stored at 20 C. Whole-mount in son C. 2000. An important role for the

Developmental Dynamics single cusp. The lower premolars are situ hybridization was performed fol- IIIb isoform of fibroblast growth factor composed of the paraconid, protoco- lowing Sive et al. (2000) with modifica- receptor 2 (FGFR2) in mesenchymal- nid, and metaconid, where the proto- tions found in Moustakas (2008). Post- epithelial signalling during mouse orga- conid is the tallest cusp. The upper natal specimens were fixed in a nogenesis. Development 127:483–492. Hacohen N, Kramer S, Sutherland D, premolars have three cusps, the para- formol-alcohol fixative (Lillie, 1965; Hiromi Y, Krasnow MA. 1998. sprouty cone, protocone, and metacone, where ethanol, formaldehyde, glacial acetic encodes a novel antagonist of FGF sig- the paracone is the largest cusp. The acid), dehydrated in ethanol, and naling that patterns apical branching of deciduous third premolar on both the stored at À20C. Marsupial neonates the Drosophila airways. Cell 92: lower and upper jaws is molariform. have an extremely thick epidermis, 253–263. Harada H, Toyono T, Toyoshima K, Ohuchi In the lower molars, the cusps of the which was removed before specimens H. 2002. FGF10 maintains stem cell talonid basin (hypoconid, entoconid, were embedded in paraffin wax. Post- population during mouse incisor devel- and hypoconulid) are relatively lower natal specimen heads were sectioned opment. Connect Tissue Res 43:201–204. than the cusps of the trigonid (proto- at 7–8 mm and the in situ hybridiza- Hebert JM, Rosenquist T, Go¨tz J, Martin GR. 1994. FGF5 acts as a regulator of conid, paraconid, and metaconid). The tion on paraffin-embedded tissue sec- the hair growth cycle: evidence from upper molars are composed of three tions was performed following Lescher targeted and spontaneous mutations. trigon cusps, the paracone (the larg- et al. (1998) with modifications in Cell 78:1017–1025. est cusp), protocone, and metacone, Moustakas (2008). Hematoxylin and Iwata T, Hevner RF. 2009. Fibroblast and four stylar cusps. eosin staining was performed follow- growth factor signaling in development of the cerebral cortex. Dev Growth Dif- ing Presnell and Schreibman (1997). fer 51:299–323. Ja¨rvinen E, Tummers M, Thesleff I. 2009. Isolation of Opossum Genes The role of the dental lamina in mam- Total RNA extraction was performed ACKNOWLEDGMENTS malian tooth replacement. J Exp Zool 312B:281–291. using the Aurum Total RNA kit (Bio- We thank Anna Keyte, Heather Derby, Jernvall J. 1995. Mammalian molar cusp Rad, Hercules, CA). cDNA was gener- Jukka Jernvall, Pat Holroyd, Bill patterns: Developmental mechanisms of ated using the Omniscript Reverse Clemens, Gregory Wilson, Ophir diversity. Acta Zool Fennica 198:1–61. OPOSSUM TOOTH DEVELOPMENT 239

Jernvall J, Thesleff I. 2000. Reiterative Luckett WP. 1993. An ontogenetic assess- receptor blocks branching morphogene- signaling and patterning during mam- ment of dental homologies in therian sis and epithelial differentiation of the malian tooth morphogenesis. Mech Dev mammals. In: Szalay FS, Novacek MJ, mouse lung. EMBO J 13:3296–3301. 92:19–29. McKenna MC, editors. Mammal phylog- Presnell JK, Schreibman MP. 1997. Jernvall J, Kettunen P, Karavanova I, eny. Vol. 1. Mesozoic differentiation, Humason’s animal tissue techniques. Martin LB, Thesleff I. 1994. Evidence multitubercultes, monotreams, early Baltimore: The Johns Hopkins Univer- for the role of the enamel knot as a con- therians, and marsupials. New York: sity Press. trol center in mammalian tooth cusp Springer. p 182–204. Rambaut A. 1996. Se-Al Sequence Align- formation: non-dividing cells express Luo Z-X. 2007. Transformation and diver- ment Editor. Version 1.0 alpha 1. growth stimulating Fgf-4 gene. Int J sification in early mammal evolution. Oxford: University of Oxford. Dev Biol 38:463–469. Nature 450:1011–1019. Reich A, Sapir A, Shilo B. 1999. Sprouty Jernvall J, Kera¨nen SV, Thesleff I. 2000. Macrini TE. 2004. Monodelphis domes- is a general inhibitor of receptor tyro- Evolutionary modification of development tica. Mamm Spec 760:1–8. sine kinase signaling. Development in mammalian teeth: quantifying gene Martin GR. 1998. The roles of FGFs in 126:4139–4147. expression patterns and topography. Proc the early development of vertebrate Salazar-Ciudad I, Jernvall J. 2002. A Natl Acad Sci U S A 97:14444–14448. limbs. Genes Dev 12:1571–1586. gene network model accounting for de- Karavanova I, Dove LF, Resau JH, Peran- Mate KE, Robinson ES, VandeBerg JL, Pe- velopment and evolution of mammalian toni AO. 1996. Conditioned medium from dersen RA. 1994. Timetable of in vivo teeth. Proc Natl Acad Sci U S A 99: a rat ureteric bud cell line in combination embryonic development in the grey 8116–8120. with bFGF induces complete differentia- short-tailed opossum (Monodelphis Salazar-Ciudad I, Jernvall J. 2004. How tion of isolated metanephric mesen- domestica). Mol Reprod Dev 39:365–374. different types of pattern formation chyme. Development 122:4159–4167. Mikkelsen TS, Wakefield MJ, Aken B, mechanisms affect the evolution of form Kera¨nen SV, Aberg T, Kettunen P, The- Amemiya CT, Chang JL, Duke S, and development. Evol Dev 6:6–16. sleff I, Jernvall J. 1998. Association of Garber M, Gentles AJ, Goodstadt L, Salazar-Ciudad I, Jernvall J. 2010. A developmental regulatory genes with Heger A, Jurka J, Kamal M, Mauceli E, computational model of teeth and the the development of different molar Searle SMJ, Sharpe T, Baker ML, Bat- developmental origins of morphological tooth shapes in two species of rodents. zer MA, Benos PV, Belov K, Clamp M, variation. Nature 464:583–586. Dev Genes Evol 208:477–486. Cook A, Cuff J, Das R, Davidow L, Dea- Sive HL, Grainger RM, Harland RM. Kettunen P, Thesleff I. 1998. Expression kin JE, Fazzari MJ, Glass JL, Grabherr 2000. Early development of Xenopus and function of FGFs-4, -8, and -9 sug- M, Greally JM, Gu W, Hore TA, Huttley laevis: a laboratory manual. New York: gest functional redundancy and repetitive GA, Kleber M, Jirtle RL, Koina E, Lee Cold Spring Harbor Laboratory Press. use as epithelial signals during tooth JT, Mahony S, Marra MA, Miller RD, Smith KK. 2006. Craniofacial develop- morphogenesis. Dev Dyn 211:256–268. Nicholls RD, Oda M, Papenfuss AT, ment in marsupial mammals: develop- Kettunen P, Laurikkala J, Ita¨ranta P, Parra ZE, Pollock DD, Ray DA, Schein mental origins of evolutionary change. Vainio S, Itoh N, Thesleff I. 2000. Asso- JE, Speed TP, Thompson K, VandeBerg Dev Dyn 235:1181–1193. ciations of FGF-3 and FGF-10 with sig- JL, Wade CM, Walker JA, Waters PD, Smith MM, Fraser GJ, Chaplin N, Hobbs naling networks regulating tooth Webber C, Weidman JR, Xie X, Zody C, Graham A. 2009. Reiterative pattern morphogenesis. Dev Dyn 219:322–332. MC, Platform BIGS, Team BIWGA, of sonic hedgehog expression in the cat- Keyte AL, Smith KK. 2008. Opossum Graves JAM, Ponting CP, Breen M, shark dentition reveals a phylogenetic (Monodelphis domestica): a marsupial Samollow PB, Lander ES, Lindblad-Toh template for jawed vertebrates. Proc R developmental model. In: Emerging K. 2007. Genome of the marsupial Soc B 276:1225–1233. model organisms: a laboratory manual. Monodelphis domestica reveals innova- Streilein KE. 1982. Behavior, ecology, and Cold Spring Harbor, NY: Cold Spring tion in non-coding sequences. Nature distribution of South American marsu- Harbor Press. p 557–575. 447:167–177. pials. Special Publication of the Pymatun-

Developmental Dynamics Kim HJ, Bar-Sagi D. 2004. Modulation of Minowada G, Jarvis LA, Chi CL, Neu- ing Laboratory of Ecology 6:231–250. signalling by Sprouty: a developing bu¨ ser A, Sun X, Hacohen N, Krasnow Vaahtokari A, Aberg T, Jernvall J, Kera- story. Nat Rev Mol Cell Biol 5:441–450. MA, Martin GR. 1999. Vertebrate nen S, Thesleff I. 1996. The enamel Klein OD, Minowada G, Peterkova R, Sprouty genes are induced by FGF sig- knot as a signaling center in the devel- Kangas A, Yu BD, Lesot H, Peterka M, naling and can cause chondrodysplasia oping mouse tooth. Mech Dev 54:39–43. Jernvall J, Martin GR. 2006. Sprouty when overexpressed. Development 126: van Nievelt AFH, Smith KK. 2005. To genes control diastema tooth develop- 4465–4475. replace or not to replace: the signiﬁ- ment via bidirectional antagonism of Moustakas JE. 2008. Development of cance of reduced functional tooth epithelial-mesenchymal FGF signaling. the carapacial ridge: implications for replacement in marsupial and placental Dev Cell 11:181–190. the evolution of genetic networks in mammals. Paleobiology 31:324–346. Klein OD, Lyons DB, Balooch G, Marshall turtle shell development. Evol Dev 10: Widelitz RB, Jiang TX, Noveen A, Chen GW,BassonMA,PeterkaM,BoranT, 29–36. CW, Chuong CM. 1996. FGF induces Peterkova R, Martin GR. 2008. An FGF National Research Council. 1996. Guide new feather buds from developing avian signaling loop sustains the generation of for the care and use of laboratory ani- skin. J Invest Dermatol 107:797–803. differentiated progeny from stem cells in mals. Washington, DC: National Acad- Wilkie AOM, Patey SJ, Kan S-H, van den mouse incisors. Development 135:377–385. emy Press. Ouweland AMW, Hamel BCJ. 2002. Kramer S, Okabe M, Hacohen N, Kras- Nowak RM. 1999. Walker’s mammals of FGFs, their receptors and human limb now MA, Hiromi Y. 1999. Sprouty: a the world. Baltimore, MD: Johns Hop- malformations: clinical and molecular common antagonist of FGF and EGF kins University Press. correlations. Am J Med Genet 112: signaling pathways in Drosophila. De- Ohuchi H, Hori Y, Yamasaki M, Harada 266–278. velopment 126:2515–2525. H, Sekine K, Kato S, Itoh N. 2000. Yamanaka A, Yasui K, Sonomura T, Lescher B, Haenig B, Kispert A. 1998. FGF10 acts as a major ligand for FGF Uemura M. 2007. Development of heter- sFRP-2 is a target of the Wnt-4 signal- receptor 2 IIIb in mouse multi-organ odont dentition in house shrew (Suncus ing pathway in the developing meta- development. Biochem Biophys Res murinus). Eur J Oral Sci 115:433–440. nephric kidney. Dev Dyn 213:440–451. Commun 277:643–649. Zhang S, Lin Y, Ita¨ranta P, Yagi A, Vainio Lillie RD. 1965. Histopathologic technic Peters K, Werner S, Liao X, Wert S, Whit- S. 2001. Expression of Sprouty genes 1, and practical histochemistry. New York: sett J, Williams L. 1994. Targeted 2 and 4 during mouse organogenesis. McGraw-Hill. expression of a dominant negative FGF Mech Dev 109:367–370.