MICROSCOPY RESEARCH AND TECHNIQUE 59:373–395 (2002)

Membranes, Minerals, and Proteins of Developing Vertebrate Enamel THOMAS G.H. DIEKWISCH,1,2,3* BRETT J. BERMAN,4 XOCHITL ANDERTON,2 BRIAN GURINSKY,2 ADAM J. ORTEGA,5 PAUL G. SATCHELL,2 MIA WILLIAMS,2 CHITHRA ARUMUGHAM,2 XIANGHONG LUAN,1,2 JAMES E. MCINTOSH,2 AKIRA YAMANE,6 DAVID S. CARLSON,2 JEAN-YVES SIRE,7 3 AND CHARLES F. SHULER 1

Allan G. Brodie Laboratory for Craniofacial Genetics, University of Illinois at Chicago, Illinois, USA Department of Biomedical Sciences, Baylor College of Dentistry/Texas A&M University System, Dallas, Texas, USA 3 Center for Craniofacial Molecular Biology, USC School of Dentistry, Los Angeles, California, USA 4 Department of Medicine, The University of California, San Diego, California, USA 5 Harvard School of Dental Medicine, Harvard University, Boston, Massachusetts, USA 6 Department of Pharmacology, School of Dental Medicine, Tsurumi University, Japan 7 Equipe “Evolution & De´veloppement du Squelette dermique”, UMR 8570, Universite´ Paris 7, France 2

KEY WORDS

amelogenin; in situ; electron microscopy; evolution; enameloid; adameloid

ABSTRACT Developing tooth enamel is formed as organized mineral in a specialized protein matrix. In order to analyze patterns of enamel mineralization and enamel protein expression in species representative of the main extant vertebrate lineages, we investigated developing teeth in a chondrichthyan, the horn shark, a teleost, the guppy, a urodele amphibian, the Mexican axolotl, an anuran amphibian, the leopard frog, two lepidosauria, a gecko and an iguana, and two mammals, a marsupial, the South American short-tailed gray opossum, and the house mouse. Electron microscopic analysis documented the presence of a distinct basal lamina in all species investigated. Subsequent stages of enamel biomineralization featured highly organized long and parallel enamel crystals in mammals, lepidosaurians, the frog, and the shark, while amorphous mineral deposits and/or randomly oriented crystals were observed in the guppy and the axolotl. In situ hybridization using a full-length mouse probe for amelogenin mRNA resulted in amelogenin specific signals in mouse, opossum, gecko, frog, axolotl, and shark. Using immunohistochemistry, amelogenin and tuftelin enamel proteins were detected in the enamel organ of many species investigated, but tuftelin epitopes were also found in other tissues. The anti-M179 antibody, however, did not react with the guppy and axolotl enameloid matrix. We conclude that basic features of vertebrate enamel/enameloid formation such as the presence of enamel proteins or the mineral deposition along the dentin-enamel junction were highly conserved in vertebrates. There were also differences in terms of enamel protein distribution and mineral organization between the vertebrates lineages. Our findings indicated a correlation between the presence of amelogenins and the presence of long and parallel hydroxyapatite crystals in tetrapods and shark. Microsc. Res. Tech. 59:373–395, 2002. © 2002 Wiley-Liss, Inc. INTRODUCTION The outer covering of vertebrate teeth is a unique structure, both because of its extreme hardness and because of the almost complete lack of cellular components within. This outer, shiny mineral layer is termed enamel in tetrapods (mammals, lepidosaurians, amphibians), and enameloid nontetrapod vertebrates (basal sarcopterygians, actinopterygians, and chondrichthyans) and larval urodele amphibians (Poole, 1967, 1971; Ørvig, 1967). In this volume, we have introduced a third type of vertebrate outer tooth coverings, adameloid, for the palisade-type crystal structures of the outer tooth coverings mostly in chondrichthyans (see Introduction to this issue). In many vertebrates, the enamel/enameloid layer is crystalline and often contains calcium compounds, calcium hydroxyapatite in tetrapods, and calcium carbonate or fluorapatite in chondrichthyans and actinopterygians (Prostak et al., 1993; Slavkin and Diekwisch, 1996, 1997). The organic phase of the developing mammalian ©

2002 WILEY-LISS, INC.

enamel is composed of a variety of proteins, among which amelogenin is the most abundant (90% of the matrix) (Simmer and Fincham, 1995). At the onset of enamel formation, enamel proteins interact in a tissuespecific fashion to organize and control mineral growth in the enamel layer, a process that is called enamel biomineralization (Diekwisch et al., 1993, 1995). During enamel biomineralization, enamel proteins aggregate to form subunit compartments, in which enamel crystals grow (Diekwisch et al., 1993, 1995; Diekwisch, 1998). At the early stage of crystal growth, enamel crystals reach extraordinary long c-axis dimensions

*Correspondence to: Thomas G.H. Diekwisch, D.D.S., Ph.D., Director, Allan G. Brodie Laboratory for Craniofacial Genetics, UIC College of Dentistry (M/C841), 801 South Paulina, Chicago, IL 60612. Received 26 February 2002; accepted in revised form 18 July 2002 Contract grant sponsor: NIH; Contract grant number: DE13378 (to TGHD). DOI 10.1002/jemt.10218 Published online in Wiley InterScience (www.interscience.wiley.com).

374

T.G.H. DIEKWISCH ET AL.

Fig. 1.

375

ENAMEL DEVELOPMENT AND EVOLUTION TABLE 1. Species investigated in this study Species Hornshark (Heterodontus francisci) Guppy (Poecilia reticulata) Australian lungfish (Neoceratodus forsteri) Mexican axolotl (Ambystoma mexicanum) Leopard frog (Rana pipiens) Mediterranean gecko (Hemidactylus turcicus) Green iguana (Iguana iguana) Short-tailed gray opossum (Monodelphis domestica) House mouse (Mus musculus)

Family

Abbreviated Lineage

Heterodontidae

Heterodontiformes—Elasmobranchii—Chondrichthyes

Poeciliidae

Cyprinodontiformes—Neopterygii—Actinopterygii

Ceratodontidae

Ceratodontiformes—Dipnoi—Sarcopterygii

Ambystomatidae

Caudata—Amphibia—Tetrapoda—Sarcopterygii

Ranidae

Anura—Amphibia—Tetrapoda—Sarcopterygii

Gekkonidae

Squamata—Lepidosauria—Tetrapoda—Sarcopterygii

Iguanidae

Squamata—Lepidosauria—Tetrapoda—Sarcopterygii

Didelphidae

Metatheria—Mammalia—Tetrapoda—Sarcopterygii

Muridae

Rodentia—Eutheria—Mammalia—Tetrapoda—Sarcopterygii

(Diekwisch et al., 1995). Eventually, enamel proteins are further degraded and enamel crystal thickness increases to ultimately form a highly mineralized tissue that is almost devoid of proteins: tooth enamel (Simmer and Fincham, 1995). Enamel and Enameloid Traditionally, two types of vertebrate outer tooth coverings are distinguished: “true” enamel, which constitutes the outer layer of tetrapods, and enameloid, which is found in chondrichthyans, actinopterygians, and larval urodele amphibians (Poole, 1967, 1971; Ørvig, 1967; Slavkin and Diekwisch, 1996, 1997). The term enameloid has been created (Poole, 1967; Ørvig, 1967) because developing enameloid demonstrates similarities to dentin matrix and contains proteins (mostly collagen) secreted by odontoblasts (Shellis and Miles, 1974, 1976). In some species, enameloid consists of a densely packed collagen matrix in which clusters of calcium apatite crystals are closely associated with collagen fibers (Kawasaki and Fearnhead, 1983). In contrast to enameloid, enamel is a noncollagenous secretory product of epithelial cells (Smith and Miles,

Fig. 1. Close-up photographs of the animals investigated in the present study. A: Head of a hornshark (Heterodontus francisci). Note the rows of tri-cusped front teeth in the lower jaw. The lowermost row of teeth is continuously exfoliated. B: The guppy (Poecilia reticulata), a familiar cyprinodontid aquarium fish. In contrast to the zebrafish (Danio rerio), a well-known cyprinid species that does not possess teeth on the buccal cavity but only in the pharyngeal region, the guppy features rows of teeth on oral and pharyngeal jaws. C: Upper body of a Mexican axolotl (Ambystoma mexicanum). In contrast to many amphibians, axolotls remain in their larval form (neoteny) and do not undergo metamorphosis, which explains the clearly visible exterior gills in the animal depicted here. D: Leopard frog (Rana pipiens). Most frogs have hundreds of teeth on their upper jaws but they lack teeth on their lower jaws. E: Upper body of a Mediterranean gecko (Hemidactylus turcicus). Geckos and other lizards have teeth in both jaws. F: A green iguana (Iguana iguana), another lepidosaurian investigated in this study. Note the scales covering the outer surface of the head. G: Litter of six 4 days postnatal South American shorttailed gray opossums (Monodelphis domesticus) clinging to their mother’s teats. Even though these opossums are marsupials, they do not possess a pouch and thus the offspring are easily accessible during early stages of development. H: Newborn house mouse (Mus musculus). At this stage, the eyes of the newborn mouse are still closed.

1971; Deutsch et al., 1991). A third group of vertebrate outer tooth coverings, adameloid, is characterized by long and parallel-oriented crystals and occurs in chondrichthyans (see Introduction to this issue for further details). We speculate that adameloid crystal growth and habit is determined by a noncollagenous protein matrix even though the adameloid protein matrix may or may not contain collagens. Among the enamel-possessing lineages, mammalian enamel is best described, notably in primates and in laboratory animals such as rat, mouse, and marsupials (Boyde, 1965, 1976, 1997; Boyde and Lester, 1967; Eiwasa et al., 1995). Reptilian enamel is distinguished from mammalian enamel by the lack of prisms (Sander, 2001). Anuran and adult urodele amphibian teeth are covered by an enamel layer (Smith and Miles, 1971), while teeth in larval urodeles have an enameloid cap (Kerr, 1960; Smith and Miles, 1971; see also Sire et al., pages 408 – 434 in this issue). “True” enamel has also been documented in basal sarcopterygians, e.g., in lungfish (Satchell et al., 2000), in the coelacanth (Smith, 1978), and in basal actinopterygians, polypterids, and lepisosteids, in which the thick tooth collar mineral has been considered enamel (Peyer, 1968; Reif, 1982). While enamel is composed of a single, homogeneous layer, enameloid/adameloid appear in different variations and often include an outer surface layer, a thick layer of parallel structures, and a woven layer in proximity to the dentin-enamel junction (Ro¨ se, 1897; Reif, 1979, 1982; Smith, 1992). Several authors provide evidence that the order of these layers may be reversed in a number of species and that there may be differences in cap and collar enameloid (Ro¨ se, 1897; Sasagawa and Ishiyama, 1988). During vertebrate evolution, a consecutive series of events may have led to the current hard covering tissues in living vertebrates. Smith (1992, 1995) suggested that the evolution of tetrapod enamel begins with enamel being the primitive tissue for osteichthyans, enameloid evolving secondarily in the oral teeth of actinopterygians and enamel emerging as the exclusive outer tooth covering of sarcopterygians (excepted in urodele larvae and, possibly, in larval basal sarcopterygians such as lungfishes: Smith, 1992). Smith (1992) supported the concept of enamel being

376

T.G.H. DIEKWISCH ET AL.

Fig. 2.

ENAMEL DEVELOPMENT AND EVOLUTION

more primitive than enameloid (Kemp, 1985) on the basis of actinopterygian tooth coverings, while Ro¨ se argued on the basis of fossil evidence that dentin was the most primitive vertebrate tissue (Ro¨ se, 1897). From a developmental point of view, enameloid forms prior to dentin in sharks (Risnes, 1990). The debate over the classification of the outer covering in amphibians and sharks goes back more than a century. Owen (1845) believed that the translucent outer layer of amphibian teeth was enamel. In contrast, shark teeth were covered by a tubule-rich outer layer of “vitrodentin” and not by enamel (Owen, 1845). In his studies of newts in southern Germany, Leydig (1867) came to the conclusion that the outer layer of urodele amphibians consisted of a highly mineralized form of dentin. Based on the proximity of the outer tooth layer to the epithelial enamel organ, Hertwig (1874) argued that all vertebrate teeth, including shark and amphibian teeth, were covered by enamel, a view that was influential for decades to come. The accuracy of his observation in the teeth of adult amphibians was later confirmed (Smith and Miles, 1971). Ro¨ se’s investigations of shark teeth under polarized light (Ro¨ se, 1897), however, questioned Hertwig’s view since shark enameloid did not exhibit birefringence as seen in mammalian enamel (not shown in this study, but documented in a polarized photomicrograph in Slavkin and Diekwisch, 1997). This debate over the true nature of the outer tooth coverings in chondrichthyans, actinopterygians, and larval urodele amphibians has continued until today and is anxiously awaiting molecular data. Enamel Proteins The major protein component (90%) of the mammalian enamel protein matrix is amelogenin (Termine et al., 1980; Fincham et al., 1992). While Fincham et al. (1992) in a review still summarized that the function of

Fig. 2. Macrophotographs of developing or newly formed teeth in the animals investigated in this study. A: The overview macrograph of the mandibular dentition of a hornshark demonstrated pointy, tri-cusped, incisor-shaped teeth in the front of the jaw and rounded, elongated, molar-shaped teeth in the back of the jaw. Replacement teeth were continuously formed at the lingual margin of the tooth ark (asterisk). B: The guppy featured a double row of tall, incisor-shaped teeth in upper and lower jaws. Replacement teeth (asterisk) developed continuously. C: The axolotl dentition was characterized by dense rows of conical, incisor-shaped teeth. The tooth rows contained all stages of tooth development, including erupting replacement teeth (asterisk). D: The leopard frog featured a dentition of relatively homogeneously shaped teeth on the upper jaw only (isodont dentition). Developing replacement teeth were frequently found apical of erupted teeth (asterisk). Similar to many amphibians and reptiles, the frog dentition was a typical pleurodont dentition with teeth laterally fused to the jaw. E: A row of isodont gecko teeth. The developing tooth (asterisk) was located between two fully erupted teeth. F: The macrograph illustrates the pleurodont position of iguana teeth in relationship to the jawbone. A developing replacement tooth (asterisk) was positioned directly apical of the root of an erupted tooth. G: Teeth of the South American marsupial Monodelphis domestica at various stages of development. The asterisk points at a tooth organ at the onset of tooth crown development. Opossum teeth exhibit different shapes (heterodont dentition). Their roots are embedded in a cup-like socket of the jawbone by means of a periodontal ligament (thecodont dentition). H: This figure illustrates the position of two mouse (Mus musculus) molar teeth (asterisk) in relationship to the jawbone prior to eruption.

377

amelogenin was not known, a series of genetic, antisense, knockout, and crystal growth studies of the recent decade have established amelogenin’s pivotal role the control of enamel crystal growth and enamel formation (Lagerstrom et al., 1991; Diekwisch et al., 1993; Gibson et al., 2001; Iijima et al., 2002). In contrast to the overwhelming presence and unique functional properties of amelogenin, the nonamelogenin enamel proteins only amount to 10% of the enamel protein matrix. In the past, these nonamelogenin enamel proteins have been termed enamelins (Termine et al., 1980), but since individual protein components have been identified, this terminology is no longer valid. Moreover, one of the high molecular nonamelogenins has been renamed enamelin (Hu et al., 1997). Other nonamelogenin enamel proteins have been cloned and characterized, including ameloblastin, tuftelin, and enamel proteases (Deutsch, 1989; Krebsbach et al., 1996; Bartlett et al., 1996; Hu et al., 1997; Simmer et al., 1998). Although some of these recently identified enamel proteins have been localized in tissues outside of teeth, they might be of relevance to the mechanisms of enamel formation (Deutsch et al., 1991; ZeichnerDavid et al., 1995, 1997; MacDougall et al., 1998). Tuftelin, the nonamelogenin enamel protein investigated in this study, has been sequenced and characterized (Deutsch, 1989; Deutsch et al., 1991; ZeichnerDavid et al., 1995, 1997; Diekwisch et al., 1997; Mao et al., 2001; Satchell et al., 2002). In recent years, our knowledge of the enamel protein composition and function of nonmammalian vertebrates has seen significant progress. Amelogenin sequences from two sauropsids (a crocodile and a snake) and one amphibian (Xenopus) have been published (Ishiyama et al., 1998; Toyosawa et al., 1998) and biochemical and immunohistochemical studies have enhanced our knowledge of enamel protein homologies between different vertebrate species (Graham, 1985; Herold et al., 1980, 1989; Slavkin and Diekwisch, 1996, 1997; Kogaya, 1999; Ishiyama et al., 1999; Satchell et al., 2002). To bridge the gap between morphological and immunological studies, the relationship between specific protein compounds and specific patterns of enamel mineral crystal growth has been a topic of great interest (Slavkin and Diekwisch, 1996, 1997; Delgado et al., 2001) and remains to be understood. A number of studies have provided clues toward a mechanistic explanation of the role of proteins in vertebrate enamel formation. Shellis (1975) demonstrated that enameloid consisted of an ectomesenchymally derived matrix into which epithelial proteins were deposited. The odontoblastic origin of enameloid, in particular in teleost fish, has been documented in autoradiographic studies by Shellis and Miles (1974). In addition, recent immunohistochemical findings have confirmed earlier reports of a predominance of enamelins in shark enameloid/ adameloid (Graham, 1985), compared to a predominance of amelogenins in tetrapod enamel (Satchell et al., 2002). And lastly, reports on pulp-derived amelogenins have changed the view of amelogenin as an exclusively epithelially derived and enamel-specific protein (Nebgen et al., 1999; Veis, 2000; Oida et al., 2002; Satchell et al., 2002). Together, these data appear to provide new pieces of evidence unraveling the puzzle of vertebrate enamel formation.

Fig. 3. Dental epithelium basal lamina in shark, teleost fish, and amphibians. Electron micrographs demonstrated a distinct basal lamina (bl, arrowheads) adjacent to pre-ameloblasts (am) in shark (A), guppy (B), axolotl (C), and frog (D). The orthodentin/dentin featured collagen fibers (coll) in the shark, matrix vesicles (mv), and bundles of fibrils (fib) in the frog, and odontoblast processes (od) in close proximity to the basal lamina in the guppy and in the axolotl. The secre-

tory pole of the pre-ameloblasts contained vesicles (v) in proximity to the basal lamina. The shark basal lamina measured 100 –200 nm in diameter (A) while the basal lamina of the other seven species measured approximately 20 nm in diameter (Figs. 3B–D, 4A–D). Orthodentin/dentin crystals are marked by triple arrows. The scale bar is 200 nm for all electron micrographs in this study.

Fig. 4. Dental epithelium basal lamina in lepidosaurians and mammals. A pronounced basal lamina (bl, arrowheads) was depicted between pre-ameloblasts (am) and adjacent pre-dentin (pd) or mineralized dentin (dt) in gecko (A), iguana (B), opossum (C), and mouse (D). Dentin mineralization nucleation sites in gecko and iguana are

marked by triple arrows. A matrix vesicle (mv) was identified in the iguana. In the opossum and in the mouse, fibrils (fi), mitochondria (mit), odontoblast process (od), and pre-ameloblast secretory vesicles (v) are indicated. Bar ⫽ 200 nm.

380

T.G.H. DIEKWISCH ET AL.

Fig. 5.

ENAMEL DEVELOPMENT AND EVOLUTION

In order to investigate whether evolutionary trends in protein composition and mineral patterning would provide significant insights into possible mechanisms of the evolution of vertebrate enamel formation, we performed a careful ultrastructural analysis of stages of enameloid/enamel formation in tandem with a molecular analysis of enamel protein gene expression and protein localization. Specifically, we asked the question whether individual enamel proteins were correlated with the emergence of long and parallel-oriented enamel crystals as found in mammalian and lepidosaurian teeth. Our analysis revealed a stage-wise progression of enamel formation, including the following four key stages: 1) the presence of a basal lamina and its subsequent degradation in nonchondrichthyan vertebrate lineages; 2) the stage of amorphous mineral deposition and crystal nucleation; 3) formation of small and randomly oriented crystals; and 4) the stage of crystal elongation and fusion—presence of long and parallel-oriented crystals. We found enamel-related protein epitopes, amelogenin and tuftelin, to be highly conserved among species, even though tuftelin also had a significant presence outside of the developing tooth organ. There was, however, a lack of amelogenin reactivity in the enamel organ using the recombinant M179 antibody in the guppy and the Mexican axolotl. In situ hybridization with the full-length mouse amelogenin probe did not reveal a specific signal for guppy teeth, while this probe reacted in all other species investigated. We speculate that, while enamel proteins are widely conserved, differences in enamel/adameloid/ enameloid mineral patterning might be related to evolutionary trends in enamel protein expression and distribution. MATERIALS AND METHODS Selection of Species and Tissue Preparation The following experimental animals were used in this study (Table 1): a chondrichthyan, the hornshark

Fig. 5. Initial enameloid/enamel crystal formation in shark, teleost fish, and amphibians. Short and randomly oriented initial mineral crystals were visualized in all four species investigated. In the hornshark (A), short initial crystals were densely packed to form a 50-nmthick layer (triple arrows) parallel to the basal lamina. Besides this layer of short and dense crystals, the initial adameloid layer (demarked by a dotted line) also contained 10-nm diameter crystal platelets, either isolated or arranged in bundles (arrowheads). The ameloblast secretory pole (am) and the orthodentin collagen fibers (coll) are labeled for orientation. In the guppy (B), the interface between ameloblasts (am) and orthodentin (dtoid) can be divided into two regions. Immediately adjacent to the ameloblast surface was an enameloid region (enoid, between dotted lines) with dense 10-nm crystal platelets and slightly elongated crystallites (triple arrows). Between the enameloid and the orthodentin was a homogeneous region of lesser mineral content than the enameloid zone, but free of collagen-type substructures. Lastly, the orthodentin (dtoid) contained arrangements of fine mineral nuclei of approximately 5 nm size, presumably in the periphery of collagen fibers (arrowheads). In the axolotl, two regions of different mineralization were distinguished in the recently deposited enameloid matrix (C). The region (here labeled enoid) close to the ameloblast secretory pole (am) contained bundles of, or isolated thin, 50-nm long crystals between collagen fibers. Between this region (enoid) and the orthodentin (dtoid) surface was an interface (if) of densely arranged mineral platelets (triple arrowheads). In the frog (D) bundles of randomly oriented 50 –100 nm initial enamel crystals were mixed with crystal platelets (triple arrows and arrowheads) located between the ameloblast surface (am) and dentin (dt). Bar ⫽ 200 nm.

381

(Heterodontus francisci), a teleost, the guppy (Poecilia reticulata), a urodele amphibian, the Mexican axolotl (Ambystoma mexicanum), an anuran amphibian, the leopard frog (Rana pipiens), two lepidosaurians, a gecko (Hemidactylus turcicus) and an iguana (Iguana iguana), two mammals, the South American shorttailed gray opossum, a marsupial (Monodelphis domesticus), and a newborn house mouse (Mus musculus). Species were selected as representative of the main living vertebrate lineages possessing teeth. The Pacific hornshark feeds on hard prey such as sea urchins using pointy front teeth to grasp their prey and molar-like teeth in the back of jaws for crushing. The guppy was chosen as an easily accessible omnivorous teleost. In contrast to the well-known teleost model, zebrafish, (Danio rerio), in which teeth are hardly accessible because they are only located in the pharyngeal region, guppies possess oral and pharyngeal teeth. The axolotl was chosen as a neotenic urodele, i.e., that does not undergo a metamorphosis and that remains, for a number of characters, in its larval stage throughout its adult life. However, previous studies have detected differences in enamel protein localization between larval and adult amphibian (Herold et al., 1989). The leopard frog was selected as the anuran counterpart of the urodele. Frogs carry teeth only in their upper jaws. Within the lepidosaurians, we chose the Mediterranean gecko as an insectivore and the common iguana as an herbivore. Birds and turtles were omitted because they are toothless. Among mammals, the house mouse was an easy first choice because it is a reference animal for tooth developmental research. Our second mammalian selection was another popular laboratory animal, the South American short-tailed opossum. Marsupials are considered basal among the mammalian lineage. Apart from the hornshark, all species were available from breeding colonies. The hornshark (34 cm head-to-tail length) was caught in southern California and provided by Pacific Biomarine (Los Angeles, CA). Animals were sacrificed according to Baylor College of Dentistry Animal Care Regulations, or material was obtained from specimen preserved in paraffin. In addition, larval lungfish (Neoceratodus forsteri) sections from a previous study (Satchell et al., 2000) were used to include a basal sarcopterygian in our study. Lungfish mandible sections were used for the amelogenin/tuftelin double stain. The lineages for the nine species used in this study are summarized in Table 1. Tissue Processing For amelogenin immunohistochemistry, tissues were fixed with 10% buffered formalin, decalcified in 4% EDTA for 2 days, and dehydrated in a graded series of ethanols. Specimens were embedded in paraffin and cut at 5 ␮m thickness. For tuftelin immunohistochemistry, tissues were fixed in cold 100% acetone and transferred into paraffin via a 1:1 mixture of ethanol and ether. Paraffin blocks were cut into sections at 5 ␮m. Sections were mounted on coated glass slides. For electron microscopy, tissues were fixed in Karnovsky’s fixative for 4 hours, transferred into sodium cacodylate buffer for 30 minutes, and then osmicated in 2% OsO4 for 2 hours (Diekwisch et al., 1993, 1995).

Fig. 6. Initial enamel crystal formation in lepidosaurians and mammals. The forming enamel layer (en) contained a mixture of 100 –200 nm long, thin, and more or less randomly oriented mineral crystals (triple arrows) and 20-nm diameter crystal platelets. The enamel layer was positioned between ameloblasts (am) and dentin (dt). Bar ⫽ 200 nm.

ENAMEL DEVELOPMENT AND EVOLUTION

Whole-Mount Preparations Tissues were immersed in a solution of 95% ethanol and 0.1% Alcian blue and allowed to stain overnight, rehydrated in a reverse series of ethanols, and then transferred into a mixture of 0.1% potassium hydroxide with 0.2% Alizarin red, in which they were kept for 5 hours. Electron Microscopy The details of our electron microscopy technique have been published (Diekwisch et al., 1993, 1995; Diekwisch, 1998). Specimen blocks were embedded in Epon 812. Polymerized blocks were cut at 80 nm thickness. Sections were contrasted in 1% uranyl acetate followed by Reynold’s lead citrate for 15 minutes each and observed using a JEOL 1200EX transmission electron microscope at 80 kV. Crystal measurements were established as approximate crystal dimensions per cross-section and did not represent entire crystal length (Diekwisch et al., 1993, 1995). For comparative purposes, the stage of initial enameloid/enamel layer formation was called “stage of crystal initiation.” The “stage of crystal elongation” was the stage in which the entire enamel layer was deposited, yet crystals were still developing. In Situ Hybridization Sections were deparaffinized, treated with proteinase K, and incubated with 35S-labeled amelogenin antisense or sense riboprobe. Following posthybridization washes, slides were dipped in Kodak NTB-3 solution and developed in Kodak Dektol developer. The 35Slabeled antisense and sense riboprobes were synthesized with T3 and T7 polymerase, respectively, using an RNA transcription kit (Stratagene, La Jolla, CA) and purified on G-50 Sephadex minicolumns (5 PRIME – 3 PRIME, Boulder, CO). Shorter probe fragments (200 – 400 bp) were derived by alkaline hydrolysis (1/10 volume 1 M sodium carbonate, pH 10.2 at 65°C, 10 minutes). Amelogenin positive reaction products were recorded on brightfield micrographs as black on white. Further details of this technique have been published (Diekwisch et al., 1999). Immunohistochemistry Immunoreactions were performed using polyclonal primary antibodies against a recombinant M179 amelogenin produced in E. coli (courtesy of Dr. Simmer, Department of Pediatric Dentistry, University of Texas Health Science Center at San Antonio, TX) and a polypeptide derived from the tuftelin sequence (courtesy of Dr. Zeichner-David, University of Southern California School of Dentistry, Los Angeles, CA). In order to avoid misinterpretations caused by tissue autofluorescence, an indirect immunoperoxidase technique was applied to detect signals (Diekwisch et al., 1997). Briefly, sections were treated against endogenous peroxidase using methanol and 3% hydrogen peroxide and then blocked using 10% goat serum for 10 minutes. Sections were incubated with primary antibody (amelogenin or tuftelin) for 2 hours. Primary antibodies were diluted in phosphate-buffered saline (PBS). The dilution of the primary antibody was determined in preliminary experiments. The following controls

383

were used to determine the specificity of the antibodies: 1) tissue controls—the specificity of the antibody will be evaluated in tissues with known immunoreactivity; 2) antibody controls by using a dilution series; 3) controls with preadsorbed antibody to exclude unspecific binding; 4) controls with preimmune serum to control for binding to serum components; 5) omission of primary antibody as a systemic control. Sections were washed three times in PBS and subsequently incubated for 10 minutes with biotinylated rabbit IgGs as secondary antibodies. After washing in PBS (three times), sections were exposed to the streptavidin-peroxidase conjugate for 10 minutes and then washed again in PBS (three times). Signals were detected using an AEC Substrate-Chromogen mixture. Sections were counterstained using hematoxylin and mounted with GVA-mount. RESULTS Tooth Development and Replacement in Eight Vertebrate Species For this study, we investigated the developing tooth organs of eight vertebrate species containing a developing outer tooth mineral layer (Table 1, Figs. 1, 2). New tooth organs were formed at the lingual edge of the hornshark mandible (Heterodontus francisci) (Fig. 2A), in between the tooth rows of both jaws of the guppy (Poecilia reticulata) (Fig. 2B), within the dense rows of the conical axolotl teeth (Ambystoma mexicanum) (Fig. 2C), within the upper jaw dentition of the leopard frog (Rana pipiens) (Fig. 2D), between already erupted gecko teeth (Hemidactylus turcicus) (Fig. 2 E), apical of fully erupted iguana teeth (Iguana iguana) (Fig. 2F), during neonatal development of the opossum dentition (Fig. 2G), and during the embryonal and neonatal development of the mouse dentition (Fig. 2H). Enamel Epithelium Basal Lamina Conserved in All Species Investigated In all species investigated a distinct basal lamina separated the inner enamel epithelium or its equivalent from the ectomesenchymal portion of the tooth, namely predentin and orthodentin in the present study (Figs. 3A–D, 4A–D). The shark basal lamina measured 100 –200 nm in diameter (Fig. 3A) and was approximately 10-fold as thick as the basal lamina in all other species investigated (Figs. 3B–D, 4A–D). Initial Short and Randomly Oriented Crystals in All Species Investigated There was a distinct enamel/enameloid mineral layer between the mesenchymally derived predentin/ dentin/orthodentin complex and the secretory ameloblasts (Figs. 5, 6) in all eight species investigated. In all of them, the initial stage of enamel/enameloid crystal formation contained a characteristic arrangement and patterning of the mineral phase (Figs. 5, 6). In all eight species, with the possible exception of the guppy (Fig. 5B), the initial enamel/enameloid layer contained a dense arrangement of both polygonal 10 –20 nm diameter platelets and 50 –200 nm long crystallites (Figs. 5A,C,D, 6A–D) (measurements as length per section). In the guppy, the initial mineral phase was restricted to 5–10 nm platelets only (Fig. 5B). In all eight species, with the exception of the guppy (Fig. 5B) and the axo-

384

T.G.H. DIEKWISCH ET AL.

Fig. 7.

ENAMEL DEVELOPMENT AND EVOLUTION

lotl (Fig. 5C), initial enamel/enameloid platelets and crystallites were densely packed and demonstrated a unidirectional arrangement (Figs. 5A,D, 6A–D). In the guppy and the axolotl, initial crystallites were polygonal and randomly arranged (Fig. 5B,C). In the shark, the early adameloid crystallites formed a narrow 50 nm band parallel to the basal lamina with perpendicular crystal orientation (Fig. 5A). In the frog as well as in the reptiles and mammals, initial enamel crystals measured 100 –200 nm in cross-sectional length and 5 nm in diameter. In contrast, 5–20 nm diameter polygonal mineral platelets were of lesser staining intensity and randomly arranged within the enamel/enameloid space of all species investigated (Figs. 5, 6). Diverse Patterns of Crystal Elongation in Many Species At later stages of development, the enamel/enameloid/adameloid layer was more prominent compared to the initiation stage (Figs. 7, 8). Thickness and mineral content of the enamel/enameloid/adameloid layer were increased (not shown). Most notably, the enamel layer of the frog (Fig. 7D), the gecko (Fig. 8A), the iguana (Fig. 8B), the opossum (Fig. 8C), and the mouse (Fig. 8D) contained up to 1 ␮m long (length within micrograph of section), thin (approximately 5 nm), and parallel-oriented enamel crystals. Besides those extremely long, thin, and parallel crystals, short and polygonal mineral platelets were documented in all the above species (Figs. 7D, 8A–D). In contrast, hornshark adameloid crystals at this stage also measured approximately 1 ␮m in length (within micrograph of thin section), but were 10-fold thicker (approximately 50 nm diameter) and were initially randomly oriented (Fig. 7A). At later stages, hornshark adameloid crystals were long and parallel-oriented and grouped in prisms.

Fig. 7. Enameloid/enamel crystal elongation in shark, teleost fish, and amphibians. At the crystal elongation stage, the shark adameloid (A) featured 50-nm diameter thick crystals (triple arrowheads) with crystals in cross-section measuring approximately 1 ␮m in length. Note the apparently random orientation of the crystals in this electron micrograph compared to fig. 1B of the Introduction to this issue. Considering comparable average lengths per ultrathin cross-section, shark adameloid crystals were approximately 5–10 times as thick as amphibian, lepidosaurian, and mammalian enamel crystals (Figs. 7D, 8A–D). The micrograph illustrates the 200-nm diameter basal lamina (bl) between ameloblasts (am) and adameloid (adoid) as well as the outlines of an adameloid collagen fiber (coll) of similar dimensions in cross section. B: Electron micrograph of advanced guppy mineralization, remarkably similar to the micrograph of early axolotl mineralization (Fig. 6C). In both cases, a collagen-rich enameloid was visualized adjacent to the ameloblast cell bodies. Nucleation sites of mineralization containing densely packed aggregates of 50-nm long crystals (triple arrowheads) were grouped between collagen bundles and detected sporadically. Only the interface between orthodentin (dtoid) and enameloid was more densely mineralized (triple arrows, dotted line). C: Electron micrograph of the densely packed axolotl enameloid mineral layer (enoid, triple arrows) between ameloblasts (am) and orthodentin (dtoid). The interface between enameloid and orthodentin (triple arrowheads, dotted line) was significantly more electron-dense, possibly resembling a dentin-enamel junction or mantle dentin. D: Electron micrograph of frog enamel, remarkably similar to the micrographs shown in Figure 8 of lepidosaurian and mammalian enamel. In immediate proximity to secretory ameloblasts (am), the frog enamel contained long and parallel oriented enamel crystals (en, triple arrows) next to individual mineral platelets (triple arrowheads). Bar ⫽ 200 nm.

385

Shark adameloid prisms were less interlaced than mammalian enamel prisms (Introduction to this issue, fig. 1B). The electron micrograph of advanced guppy mineralization (Fig. 7B) was remarkably similar to the micrograph of early axolotl mineralization (Fig. 5C). The guppy enameloid layer contained mineralization foci of densely packed and randomly arranged aggregates of 50 nm long crystals (Fig. 7B). The axolotl enameloid featured a densely packed mineral layer at this stage (Fig. 7C). In both guppy and axolotl there was a highly electron-dense mineral layer recorded at the interface between enameloid and orthodentin (Fig. 7B,C). Amelogenin mRNA Signals Recorded in Many Species, Including Shark The full-length mouse amelogenin probe recorded positive in situ hybridization signals in all species investigated, except the guppy. In the mouse and the opossum, the probe recognized an extremely strong signal in the ameloblast layer (Fig. 9A,B). There was also a weak but positive signal in the odontoblast layer and in the dental papilla (Fig. 9A,B). In the gecko, frog, and axolotl the probe recognized strong signals in the ameloblasts, general lamina, oral epithelium, and dental papilla (Fig. 9C,D,E). In the guppy, there was no specific message above background (Fig. 9F). In contrast, the shark exhibited strong signals in the ameloblasts lining the crown adameloid as well as in the oral epithelium and no signal was recorded in the basal lamina and the dental papilla (Fig. 9G). Amelogenin and Tuftelin Protein Epitopes Were Detected in Several, But Not All Species In many investigated animals, immunoreactions indicated the presence of amelogenin and tuftelin antigenic sites (Figs. 10 –12). The antibody against the recombinant mouse M179 amelogenin generated in E. coli recognized positive signals for amelogenin in the developing enamel/adameloid layer of the shark (Fig. 10A), frog (Fig. 10D), gecko (Fig. 10E), iguana (Fig. 10F), opossum (Fig. 10G), and mouse (Fig. 10H). In the mouse and in the iguana, the anti-amelogenin antibody also reacted with the ameloblast cell bodies (Fig. 10F,H). In addition, the mouse stellate reticulum (sr) stained positively for amelogenin (Fig. 10H). The antibody against the recombinant M179 amelogenin generated in E. coli did not react in the axolotl (Fig. 10C). In contrast, this antibody labeled the entire oral epithelium and ameloblasts in the guppy, but did not stain the guppy enameloid (Fig. 10B). The shark (Fig. 11A) featured a distinct reaction for tuftelin in ameloblasts, a less intense reaction in the orthodentin, and weak reaction products in the oral epithelium. In the guppy (Fig. 11B), tuftelin reaction products were detected in the upper half of the mineralized tooth. The axolotl (Fig. 11C) featured a strong reaction for tuftelin at the ameloblast secretory pole, in the enamel layer, and in the extracellular space between ameloblasts and cells of the outer enamel epithelium. The frog contained strong tuftelin signals in the ameloblast cell walls and in Hertwig’s epithelial root sheath. In the lepidosaurians, gecko (Fig. 11E), and iguana (Fig. 11F), the tuftelin reaction was almost exclusively restricted to the enamel layer, apart from a

Fig. 8. Enamel crystal elongation in lepidosaurians and mammals. All four electron micrographs document long and parallel oriented enamel crystals (en, triple arrows) next to individual mineral platelets (triple arrowheads) adjacent to the secretory ameloblast pole (am). The micrographs were from gecko (A), iguana (B), monodelphis (C), and mouse (D). Bar ⫽ 200 nm.

ENAMEL DEVELOPMENT AND EVOLUTION

small area of tuftelin-positive epitopes in the iguana ameloblasts. The opossum (Fig. 11G) featured distinct tuftlin reaction products in ameloblasts, enamel, and odontoblasts. Lastly, the mouse demonstrated a strong, tuftelin-specific reaction in the enamel. Tuftelin reaction products were also detected in many tissues outside of the developing tooth organ (not shown). Our double-staining procedure for amelogenin and tuftelin resulted in distinct signals within the developing enamel organ of mouse, frog, and lungfish (Fig. 12). In the mouse (Fig. 12A), the anti-amelogenin antibody reacted with the enamel layer and ameloblast secretory vesicles, while the anti-tuftelin antibody recognized ameloblast secretory vesicles and vesicle membranes. In the frog (Fig. 12B), the anti-amelogenin antibody reacted with the secretory pole of central ameloblasts, the enamel layer, and part of the dentin surface. The anti-tuftelin antibody strongly labeled the enamel layer at the dentin-enamel junction. In the lungfish (Fig. 12C), the anti-amelogenin antibody stained the enamel and parts of the dentin layer, while the antituftelin antibody reacted distinctly with the enamel layer. DISCUSSION In the present study, we investigated significant events and features during tooth enamel development in selected species representative of the main living vertebrate lineages. We confirm previous findings that teeth in teleost fish and juvenile urodele were covered with enameloid, chondrichthyan teeth with a new category of outer tooth coverings, adameloid, while lungfish and tetrapod teeth featured a “true” enamel surface layer as the outer tooth crown covering. Using this wide variety of vertebrate outer tooth coverings as a model system, we analyzed and compared cellular features, patterns of crystal formation, as well as the expression of amelogenin gene and the localization of the amelogenin and tuftelin products. Inner Enamel Epithelium Basal Lamina Electron micrographs indicated that prior to enamel/ enameloid/adameloid mineralization the basal lamina of the inner enamel epithelium was present in all species investigated. This finding essentially confirms existing knowledge on the presence of the epithelial basal lamina during odontogenesis (Kallenbach and Piesco, 1978) and provides verification for the species investigated in the present study. Similarly, the thickness of the shark basal lamina and its persistence during adameloid mineralization have been demonstrated earlier (Garant, 1970; Kemp and Park, 1974; Kallenbach and Piesco, 1978), as have been the disruption and eventual disappearance of the basal lamina in other vertebrates (Reith, 1967; Kallenbach, 1971, 1976; Kallenbach and Piesco, 1978; Hurmerinta and Thesleff, 1981). A new finding related to the role of the basal lamina of the inner dental epithelium during amelogenesis was the disappearance of the guppy basal lamina prior to enameloid mineralization. This finding was not only interesting because there have been few studies on guppy teeth but also because it provides evidence that the basal lamina persistence during adameloid/ enameloid/enamel mineralization varies among vertebrates; the basal lamina appears to persist in chon-

387

drichthyans, while it disappears during the course of enameloid and enamel formation in teleosts and dipnoans (Satchell et al., 2000). However, in teleosts both basal lamina disappearance and persistence have been reported (Prostak and Skobe, 1984; Sasagawa, 1997). Two Types of Mineral in Vertebrate Outer Tooth Coverings Two types of mineral shapes were documented during initial enamel/adameloid/enameloid mineralization: Type I were 5–20 nm diameter platelets with polygonal shape and of lower electron density and Type II were highly oriented long (200 nm) and thin (5–10 nm) electron-dense crystals. In most species, except the guppy and the axolotl, which only contained Type I platelets, both types of mineral shapes coexisted next to each other. Enameloid in the guppy only contained Type I platelets, while in the axolotl teeth the platelets were longer and more electron-dense, but not as polarized and oriented. Frog, lepidosaurian, and mammalian enamel featured long, thin, and parallel-oriented crystals, while there was a layer of short, thin, and oriented crystals contained within the shark basal lamina. The presence of Type I platelets in developing enamel has been reported previously and attributed to a conversion from amorphous calcium phosphate to hydroxyapatite (Diekwisch et al., 1995; Diekwisch, 1998), as documented for other biological systems (Eanes et al., 1973). In previous publications we have provided evidence to support our hypothesis that long and parallel-oriented enamel crystals grow by fusion of Type I crystal platelets (Diekwisch et al., 1993, 1995; Diekwisch, 1998). This contention is further supported by our electron micrographs of developing frog, lizard, Monodelphis, and mouse enamel, demonstrating perfectly aligned Type I crystal platelets in close proximity to and continuity of fused elongated Type II crystals. It appears, however, that during early enamel mineralization a gemisch (mix) of both types, Type I platelets and Type II crystals, coexists. The function of this gemisch is not known at this point: whether Type I calcium phosphate platelets serve as a necessary precursor for Type II elongated hydroxyapatite crystals or whether the presence of Type I is a remnant from basal stages of enamel development. Both might be the case considering that the conversion of amorphous calcium phosphates to hydroxyapatite is normal within biological systems (Eanes et al., 1973). Basal Lamina as a Template of Enamel Mineral Nucleation Another new finding related to the question of initial enameloid/adameloid/enamel mineral formation was the role of the basal lamina to form a template for mineral deposition and crystal formation. This was particularly obvious in the shark and in the guppy, where parts of the basal lamina became mineralized, but also in the frog and in the gecko, in which initial enamel crystals appeared to arise from remnants of the basal lamina. This finding supports the role of interfaces, either between different proteins, or, in the case of basal laminae, of lipid–protein interfaces, as potential nucleators for mineralization, a theme that has been highlighted in three recent Gordon Conferences for Biomineralization (1998, 2000, 2002). It further

388

T.G.H. DIEKWISCH ET AL.

Fig. 9.

ENAMEL DEVELOPMENT AND EVOLUTION

389

Basal Vertebrate Outer Tooth Coverings in Nontetrapod Vertebrates: Enameloid, Adameloid, and Enamel At the stage of crystal elongation, long and paralleloriented crystals were recorded in mammals, reptiles, and in the frog. In the shark adameloid, crystals were long, thick, and parallel oriented in certain areas (see Introduction to this issue, fig. 1B) while long, thick, and randomly oriented in other areas (Fig. 7A). In the guppy enameloid crystals were short and often associated with mineralized aggregates. The axolotl enameloid featured a densely packed mineral layer. Together these data suggest distinct differences between the developing outer tooth coverings in the different vertebrate lineages. There are obviously a number of different ways of looking at this diversity, but most authors (Poole 1967, 1971; Ørvig, 1967; Kawasaki and Fearnhead, 1983; Slavkin and Diekwisch, 1996, 1997) have distinguished between classic enamel profiles in tetrapods, with the exception of the larval urodeles, and the multitude of forms and compositions found in nontetrapod lineages and in urodele larvae, usually termed enameloid, but possibly containing several different forms of mineralization and organic matrix organization (Reif, 1979; Smith, 1992). On an ultrastructural and developmental level, our findings support earlier observations by Reif (1979) on the triple-layer structure of vertebrate outer tooth cov erings. In all three species—shark, guppy, and larval axolotl—we detected a dense mineralization zone in proximity to the ameloblast cell membrane, a second intermediate zone that was less mineralized, and a third highly mineralized zone in close proximity to the orthodentin/enameloid junction (zones for Heterodontus not shown but recorded). The triple-layer structure

of enameloid/adameloid might have functional implications to provide high bending strength, as suggested by Preuschoft et al. (1974) and Reif (1979). It might also have evolved as a simple consequence of increased rates of mineralization at interfaces, such as the protein–protein interface at the orthodentin– enameloid junction or the lipid–protein interface in proximity to the ameloblast cell membrane, the inner enamel epithelial basement membranes, or remnants thereof. Another interesting finding in this set of data were the similarities between guppy and axolotl enameloid. In the guppy, thin enameloid layers were characterized by minute mineral platelets in close proximity to the ameloblast cell membrane. Thick guppy enameloid and thin layers of developing axolotl enameloid had a remarkably similar appearance, in which both featured isolated crystal aggregates amidst collagen fibers and a distinct mineral aggregation at the dentin– enamel junction. Only in a few regions of the axolotl enameloid was the condensed mineral matrix featured in Figure 7C detected. The axolotl is a classic example of pedomorphism since it remains in its larval state throughout the entire adult life. Consequently, axolotl teeth are typical urodelian larvae teeth covered by enameloid (Smith and Miles, 1971). Even though axolotl and guppy enameloid were very similar, our investigations of multiple sites and sections revealed that the axolotl enameloid was not only more mineralized but also a stage further advanced than the guppy enameloid. In the axolotl, sites with minute crystal platelets close to the cell membrane, as observed in the guppy, were found in thin enameloid layers, while in areas of maximum thickness the entire enameloid matrix was mineralized. This difference in the appearance of patterns of mineralization may be interpreted as evidence for heterochrony between key events in the guppy and axolotl. In the present study, however, without a direct correlation to events of gene expression and protein composition at each stage, such correlations are merely hypothetical. Therefore, the implications of our ultrastructural studies on crystal formation will be discussed at the end of this article in conjunction with genetic and biochemical findings.

Fig. 9. In situ hybridization detection of mouse amelogenin mRNA. These are brightfield micrographs in which the amelogenin-posivite reaction products appear black while the background appears white. In situ hybridization reactions demonstrated positive signals in all but one species using a full-length mouse amelogenin probe. In the mouse (A), the full-length amelogenin probe recognized an extremely strong signal in the ameloblast layer (amel) causing the entire layer to turn black. Note the weak but possibly positive signal in the odontoblast layer (od). In the opossum (B), the amelogenin-specific signal in the ameloblasts was similarly strong as in the mouse. Again, there was some weaker signal in the odontoblasts (od) and in the dental papilla (pl). A somewhat different but congruent amelogenin mRNA recognition pattern was documented in gecko (C), frog (D), and axolotl (E). In all three species, the probe recognized strong signals in the ameloblasts (am), general lamina (gl), oral epithelium (oe), and dental papilla (pl). In the guppy (F), no specific message was detectable, even though there was a diffuse reaction in the oral epithelium and in the ameloblasts (am). In the shark, however (G), signals in the oral epithelium (oe) were distinct and there was a strong signal in the ameloblasts lining (am) of the shark crown adameloid. The line in the lower left of the radiograph marked by an asterisk was an artifact created by tissue folding.

Enameloid/Adameloid Proteins Our antibody against the recombinant M179 amelogenin clearly did not label guppy and axolotl enameloid, while it reacted extensively with many epithelia in the guppy. The significance of these findings, while consistent in our experiments, was somewhat dampened by an in situ hybridization labeling of axolotl ameloblasts using the full-length amelogenin mouse probe and other, unpublished immunoreactions with a polyclonal anti-amelogenin slab gel extract antibody that yielded distinct signals in axolotl, guppy enameloid, as well as in the enamel of other vertebrates (data not shown). One explanation for the discrepancy between antibody and in situ hybridization data in the axolotl might be a failure of an immunologically recognizable amelogenin to be translated. Another explanation for false-positive findings in nontetrapod vertebrates is the possibility that our mouse probes might have bound to a highly conserved osteonectin-related amelogenin fragment, similar to the one that has recently been identified by Delgado et al.

supports our hypothesis that numerous interfaces in the enamel layer, including the ameloblast cell membrane, the dentin– enamel junction proximity (and not the dentin crystals per se), as well as protein aggregates within the enamel layer may serve as nucleators for enamel crystal formation (Diekwisch et al., 1995; Diekwisch, 1998).

390

T.G.H. DIEKWISCH ET AL.

Fig. 10. Immunohistochemical M179 recombinant amelogenin protein localization. A distinct red color signal for amelogenin reaction products was recorded in the enamel/ enameloid layer (en) of many species investigated, including shark (A), frog (D), gecko (E), iguana (F), opossum (G), and mouse (H). In the mouse and in the iguana, the anti-M179 recombinant amelogenin antibody also reacted with the ameloblast cell bodies (am) (F,H). In addition, the mouse stellate reticulum (sr) stained positively for amelogenin (H). In the shark, the anti-M179 recombinant antibody reacted with the tip of the crown (en) as well as with the surface layer of the shaft/collar region (sc). In the guppy, the anti-M179 recombinant amelogenin antibody labeled the oral epithelium (oe) and the ameloblast (am) layer, while the enamel layer (en) did not react (Fig. 10B). Note the black pigment cells (pigm). The antiM179 amelogenin antibody did not label any epitopes in the axolotl (Fig. 10C). For the guppy and the axolotl, several stages of development were included in the photomicrograph to demonstrate the absence of a positive staining for amelogenin at several stages of development. Ameloblasts (am), dentin (dt), oral epithelium (oe), and outer enamel epithelium (oee) were labeled for orientation purposes.

ENAMEL DEVELOPMENT AND EVOLUTION

Fig. 11. Immunohistochemical tuftelin protein localization in all species investigated. The shark (A) featured a distinct reaction for tuftelin in ameloblasts (am), a less intense reaction in the orthodentin (dt), and weak reaction products in the oral epithelium (oe). The tip-ofthe crown adameloid (asterisk) was devoid of tuftelin reaction products. In the guppy (B), tuftelin reaction products were detected in the upper half of the mineralized tooth (asterisk), but not in the orthodentin (dt). The axolotl (C) featured a strong reaction for tuftelin at the ameloblast (am) secretory pole and in the enamel layer (en, arrowheads). In the axolotl, tuftelin was also localized in the extracellular space between ameloblasts and cells of the outer enamel epithelium (triple arrows) as well as in areas outside of the tooth organ (asterisk). The frog contained strong tuftelin signals in the ameloblast (am) cell walls (triple arrowheads) and in Hertwig’s epithelial root sheath (hers). Enamel and dentin were free of reaction products. The distinct tuftelin marked area indicated by the asterisk might represent a tangentially cut tooth organ. In the gecko (E) and iguana (F), the tuftelin reaction was almost exclusively restricted to the enamel layer (en), apart from a restricted area of tuftelin-positive epitopes in the iguana ameloblasts (am, triple arrows). Dentin (dt), outer enamel epithelium (dt), general lamina (gl), and alveolar bone (alv) were labeled for orientation purposes. The opossum (G) featured strong tuftelin reaction products in ameloblasts (am), enamel (en), and odontoblasts (od), while the dentin (dt) was devoid of tuftelin epitopes. Lastly, the mouse demonstrated a strong, tuftelin-specific reaction in the enamel (en), while ameloblasts (am) and dentin (dt) did not react.

391

392

T.G.H. DIEKWISCH ET AL.

Fig. 12. Colocalization of amelogenin and tuftelin epitopes in developing teeth. Double-staining for M179 recombinant amelogenin (bluish-grayish color) and tuftelin (red color) in mouse (A), leopard frog (B), and lungfish (C). In the mouse (A), the anti-amelogenin antibody (bluish-grayish color) reacted with the enamel layer (en) and ameloblast (am) secretory vesicles (ves). The anti-tuftelin antibody (red color) recognized ameloblast secretory vesicles (ves) and vesicle membranes. In the frog (B), the anti-amelogenin antibody (bluish-

grayish color) reacted with the secretory pole of central ameloblasts (am), the enamel layer (en), and part of the dentin surface (dent, tangential section). The anti-tuftelin antibody (red color) strongly labeled the enamel layer at the dentin-enamel junction (dej). In the lungfish (C), the anti-amelogenin antibody stained the enamel (en) and parts of the dentin (de) layer. The anti-tuftelin antibody strongly reacted with the enamel layer (en). The position of the ameloblasts (am) and papilla/pulp (plp) are indicated.

(2001). Nevertheless, our data using the antibody against the recombinant M179 amelogenin suggest distinct differences between teleost and axolotl enameloid, chondrichthyan adameloid, and tetrapod enamel, including a preference for amelogenins in tetrapodian enamel matrices and shark adameloid, and a lack thereof in enameloid-carrying vertebrates such as the guppy and the axolotl. According to our in situ hybridization and recombinant M179 amelogenin antibody data, there appears to be an exception from this rule in shark adameloid. The full-length mouse amelogenin probe recognized a strong signal in the ameloblasts covering the hornshark adameloid, and the antibody against the recombinant mouse M179 amelogenin strongly reacted with the adameloid matrix. This indicated a probable presence of amelogenins in the shark enamel organ. This result confirms previous studies using mammalian-

raised antibodies (Herold et al., 1980; Slavkin et al., 1982, 1983; Slavkin and Diekwisch, 1996, 1997). However, amino acid analyses from shark tooth extracts (Clement, 1984) and immunohistochemical stainings using monoclonal antibodies (Herold et al., 1989) did not reveal any evidence for amelogenins in nontetrapod vertebrates, including shark. Enamelin might be another protein that might play an enhanced role in shark adameloid (Graham, 1985; Herold and Rosenblum, 1989; Satchell et al., 2002). There may not be a single answer to the role of amelogenins in chondrichthyans because differences have been found among shark species (Gurinsky and Diekwisch, 2000). In the latter study we reported the presence of amelogenins together with long and parallel crystals in the Pacific hornshark, Heterodontus francisci, while amelogenin was absent and crystals were short and randomly oriented in the banded catshark Chiloscyllium puncta-

ENAMEL DEVELOPMENT AND EVOLUTION

tum. Moreover, in related studies we found that the enameloid layer of an actinopterygian, the reedfish Polypterus, reacted with the recombinant amelogenin antibody while also featuring long and parallel-oriented crystals (Gurinsky et al., 1999). This finding contrasts well with our current data in the guppy, a more derived actinopterygian species in which an absence of amelogenin goes along with short and random crystal structures. Together, our recent and current studies on amelogenin and crystal structure in the outer covering tissues of nontetrapod vertebrates allow the conclusion that wherever amelogenins were detected, we also recorded Type II long and parallel crystal structures via electron microscopy. Based on this distinction, we propose to introduce a change in enameloid nomenclature: adameloid is an enameloid-like tissue containing long and parallel-oriented crystals and a predominantly noncollagenous protein matrix (as in shark teeth), while enameloid sensu stricto contains short and randomly oriented crystals and a rich collagenous protein matrix (as in guppy teeth) (see Introduction to this issue). Another part of our study addressed the role of another enamel-related non-amelogenin protein, tuftelin. Our data demonstrated that tuftelin was loosely associated with the enamel organ in all species investigated, but clearly also with many other tissues and organs, thus suggesting a role in enamel formation or differentiation as well as in other developing tissues. Apart from the enamel proteases, enamelin was the other significant enamel-related protein according to our recent, comparative immunohistochemical study (Satchell et al., 2002). Enamelin demonstrated a significant distribution pattern in the developing teeth of the hornshark, Heterodontus francisci. Interestingly, another frequently discussed enamel protein, ameloblastin (amelin and sheathelin are synonyms) did not exhibit a specific localization pattern in nonmammalian vertebrates in our studies (data not shown). The question of vertebrate enamel, adameloid, and enameloid protein sequences is not of academic interest alone. The pursuit of this topic will also most likely significantly enhance our understanding of the mechanisms of enamel formation and evolution. A relatively recent series of studies has shed some light onto the area of amelogenin evolutionary genomics. Bonass et al. (1994) have portrayed the LRAP as a highly conserved functional amelogenin fragment. Delgado et al. (2001) have established a similarity between amelogenin and osteonectin exon 2 and presented evidence to document the first occurrence of amelogenin-related genomic DNA in the Proterozoic (⬃630 MYA). A systematic study by Toyosawa et al. (1998) revealed a relatively high conservation of exons 2, 3, 5, and 7 in mammals, caiman, and the African clawed toad (Xenopus). Exon 6 was partially truncated in Xenopus, while exon 4 was absent in Xenopus and in the caiman, as in several mammals. It would be of highest interest to follow the path of amelogenin genomics throughout vertebrate evolution and to correlate genomic studies with functional studies on crystal morphology. In this fashion, evolutionary biology would contribute to the deciphering of the functional domains of the amelogenin gene as related to enamel crystal formation.

393

In a simplified fashion, our data confirmed that a high degree of homology to mammalian amelogenin was associated with species that featured long and parallel-oriented enamel crystals such as lepidosaurians, some amphibians, and even chondrichthyans. It is probable that the near future of sequence analysis and genomic research will reveal the presence of enamel proteins in sharks and teleosts as well as further information on the sequence and evolution of these proteins. Follow-up studies will evaluate to what extent these proteins are represented in the developing enamel/adameloid/enameloid protein matrix of each species. Eventually, a precise relationship between the enamel protein composition and the resulting crystal structure in individual vertebrate teeth will be established. This information will significantly enhance our knowledge of the role of proteins in vertebrate enamel biomineralization. REFERENCES Bartlett JD, Simmer JP, Xue J, Margolis HC, Moreno EC. 1996. Molecular cloning and mRNA tissue distribution of a novel matrix metalloproteinase isolated from porcine enamel organ. Gene 183: 123–128. Bonass WA, Kirkham J, Brookes SJ, Shore RC, Robinson C. 1994. Isolation and characterization of an alternatively-spliced rat amelogenin cDNA: LRAP — a highly conserved, functional alternatively-spliced amelogenin? Biochim Biochphys Acta 1219:690 – 692. Boyde A. 1965. The structure of developing mammalian enamel. In: Stark MG, Fearnhead RW, editors. Tooth enamel. Bristol. p 163– 194. Boyde A. 1976. Amelogenesis and the structure of enamel. In: Cohen B, Kramer IR, editors. Scientific foundations of dentistry. London: William Heinsmann Medical Books. p 335–352. Boyde A. 1997. Microstructure of enamel. In: Chadwick I, Cardew G, editors. Dental enamel. Ciba Foundation Symposium 205. p 18 –31. Boyde A, Lester KS. 1967. The structure and development of marsupial enamel tubules. Zeitschr Zellfor 82:558 –576. Clement JG. 1984. Changes to structure and chemistry of chondrichtyan enameloid during development. In: Fearnhead RW, Suga S, editors. Tooth enamel IV. Amsterdam: Elsevier. p 422– 426. Delgado S, Casane D, Bonnaud L, Laurin M, Sire JY, Girondot M. 2001. Molecular evidence for Precambrian origin of amelogenin, the major protein of vertebrate enamel. Mol Biol Evol 18:2146 –2153. Deutsch D. 1989. Structure and function of enamel gene products. Anat Rec 224:189 –210. Deutsch D, Palmon A, Fisher LW, Kolodny N, Termine JD, Young MF. 1991. Sequencing of bovine enamelin (“tuftelin”) a novel acidic enamel protein. J Biol Chem 266:16021–16028. Diekwisch TG. 1998. Subunit compartments of secretory stage enamel matrix. Connect Tissue Res 38:101–111; discussion 139 – 145. Diekwisch T, David S, Bringas P Jr, Santos V, Slavkin HC. 1993. Antisense inhibition of AMEL translation demonstrates supramolecular controls for enamel HAP crystal growth during embryonic mouse molar development. Development 117:471– 482. Diekwisch TGH, Berman BJ, Gentner S, Slavkin HC. 1995. Initial enamel crystals are not spatially associated with mineralized dentine. Cell Tissue Res 279:149 –167. Diekwisch TG, Ware J, Fincham AG, Zeichner-David M. 1997. Immunohistochemical similarities and differences between amelogenin and tuftelin gene products during tooth development. J Histochem Cytochem 45:859 – 866. Diekwisch TGH, Marches F, Williams A, Luan X. 1999. Cloning, gene expression, and characterization of CP27, a novel gene in mouse embryogenesis. Gene 235:19 –30. Eanes ED, Termine JD, Nylen MU. 1973. An electron microscopic study of the formation of amorphous calcium phosphate and its transformation to crystalline apatite. Calc Tissue Res 12:143–158. Eiwasa Y, Suzuki K, Kozawa Y. 1995. The structure and development of the enamel tubules in a marsupial, the opossum (Monodelphis domestica). Proc 10th Int Symp Dent Morphol. p 106 –109. Fincham AG, Lau EC, Simmer J, Zeichner-David M. 1992. Amelogenin biochemistry — form and function. In: Slavkin H, Price P, editors. Chemistry and biology of mineralized tissues. Amsterdam: Excerpta Media. p 187–201.

394

T.G.H. DIEKWISCH ET AL.

Garant PR. 1970. An electron microscopic study of the crystal-matrix relationship in the teeth of the dogfish, Squalus acanthias. J Ultrastruct Res 30:441:449. Gibson CW, Yuan Z-A, Hall B, Longenecker G, Chen E, Thyagarajan T, Sreenath T, Wright JT, Decker S, Piddington R, Harrison G, Kulkarni AB. 2001. Amelogenin-deficient mice display an amelogenesis imperfecta phenotype. J Biol Chem 276:31871–31875. Graham EE. 1985. Isolation of enamelinlike proteins from blue shark (Prionace glauca) enameloid. J Exp Zool 234:185–191. Gurinsky B, Diekwisch TGH. 2000. Distinct differences in enameloid formation between two shark species Heterodontus francisci and Chiloscyllium punctatum. J Dent Res 79:303. Gurinsky B, Satchell P, Diekwisch TGH. 1999. Teeth of the reedfish Polypterus are covered by an outer layer different from enamel but containing enamel-like features such as enamel proteins and long and parallel crystals. J Dent Res 78:174. Herold RC, Graver HT, Christner P. 1980. Immunohistochemical localization of amelogenins in enameloid of lower vertebrate teeth. Science 207:1357–1358. Herold R, Rosenbloom J, Granovsky M. 1989. Phylogenetic distribution of enamel proteins: immunohistochemical localization with monoclonal antibodies indicates the evolutionary appearance of enamelins prior to amelogenins. Calcif Tissue Int 45:88 –94. ¨ ber den Bau und die Entwickelung der PlacoidHertwig O. 1874. U schuppen und der Za¨ hne der Selachier. Jen Zeitschr Naturw 8:331– 404. Hu CC, Fukae M, Uchida T, Qian Q, Zhang CH, Ryu OH, Tanabe T, Yamakoshi Y, Murakami C, Dohi N, Shimizu M, Simmer JP. 1997. Cloning and characterization of porcine enamelin mRNAs. J Dent Res 76:1720 –1729. Hurmerinta K, Thesleff I. 1981. Ultrastructure of the epithelial-mesenchymal interface in the mouse tooth germ. J Craniofac Genet Dev Biol 1:191–202. Iijima M, Moriwaki Y, Wen HB, Fincham AB, Moradian-Oldak J. 2002. Elongated growth of octacalcium phosphate crystals in recombinant amelogenin gels under controlled ionic flow. J Dent Res 81:69 –73. Ishiyama M, Mikami M, Shimokawa H, Oida S. 1998. Amelogenin protein in tooth germs of the snake Elaphe quadrivirgata, immunohistochemistry, cloning and cDNA sequence. Arch Histol Cytol 61:467– 474. Ishiyama M, Inage T, Shimokawa H. 1999. An immunocytochemical study of amelogenin proteins in the developing tooth enamel of the gar-pike, Lepisosteus oculatus (Holostei, Actinopterygii). Arch Histol Cytol 62:191–197. Kallenbach E. 1971. Electron microscopy of the differentiating rat inscisor ameloblast. J Ultrastruct Res 35:508 –531. Kallenbach E. 1976. Fine structure of differentiating ameloblasts in the kitten. Am J Anat 145: 285–318. Kallenbach E, Piesco NP. 1978. The changing morphology of the epithelium-mesenchyme interface in the differentiation zone of growing teeth of selected vertebrates and its relationship to possible mechanisms of differentiation. J Biol Buccale 6:229 –240. Kawasaki K, Fearnhead RK. 1983. Comparative immunohistochemistry of tooth enamel and enameloid. In: Suga S, editor. Mechanism of tooth enamel formation. Tokyo: Quintessence. p 229 –238. Kemp NE. 1985. Ameloblastic secretion and calcification of the enamel layer in shark teeth. J Morphol 184:215–230. Kemp NE, Park JH. 1974. Ultrastructure of the enamel layer in developing teeth of the shark Carcharhinus menisorrah. Arch Oral Biol 19:633– 644. Kerr T. 1960. Development and structure of some actinopterygian and urodele teeth. Proc Zool Soc Lond 133:401– 422. Kogaya Y. 1999. Immunohistochemical localisation of amelogeninlike proteins and type I collagen and histochemical demonstration of sulphated glycoconjugates in developing enameloid and enamel matrices of the larval urodele (Triturus pyrrhogaster) teeth. J Anat 195:455– 464. Krebsbach PH, Lee SK, Matsuki Y, Kozak CA, Yamada KM, Yamada Y. 1996. Full-length sequence, localization, and chromosomal mapping of ameloblastin. A novel tooth-specific gene. J Biol Chem 271:4431– 4435. Kvam T. 1946. Comparative study of the ontogenetic and phylogenetic development of dental enamel. Norske Tannlaegeforen Tid 56(Suppl):1–198. Lagerstrom M, Dahl N, Nakahori Y, Nakagome Y, Backman B, Landegren U, Pettersson U. 1991. A deletion in the amelogenin gene (AMG) causes X-linked amelogenesis imperfecta (AIH1). Genomics 10:971–975. ¨ ber die Molche der wu¨ rttembergischen Fauna. Arch Leydig F. 1867. U Naturgesch 33:163–282.

MacDougall M, Simmons D, Dodds A, Knight C, Luan X, ZeichnerDavid M, Zhang C, Ryu OH, Qian Q, Simmer JP, Hu CC. 1998. Cloning, characterization, and tissue expression pattern of mouse tuftelin cDNA. J Dent Res 77:1970 –1978. Mao Z, Shay B, Hekmati M, Fermon E, Taylor A, Dafni L, Keikinheimo K, Lustmann J, Fisher LW, Young MF, Deutsch D. 2001. The human tuftelin gene: cloning and characterization. Gene 279:181– 196. Nebgen DR, Inoue H, Sabsay B, Wei K, Ho CS, Veis A. 1999. Identification of the chondrogenic-inducing activity from bovine dentin (bCIA) as a low-molecular-mass amelogenin polypeptide. J Dent Res 78:1484 –1494. Oida S, Nagano T, Yamakoshi Y, Ando H, Yamada M, Fukae M. 2002. Amelogenin gene expression in porcine odontoblasts. J Dent Res 81:103–108. Ørvig T. 1967. Phylogeny of tooth tissues: evolution of some calcified tissues in early vertebrates. In: Miles AEW, editor. Structural and chemical organization of teeth; New York: Academic Press. p 45– 110. Owen R. 1845. Odontography; or a treatise on the comparative anatomy of the teeth; their physiological relations, mode of development and microscopic structure in the vertebrate animals. 2 vols. Hyppolyte. London: Baillie`re. Peyer B. 1968. Comparative odontology. Chicago: University of Chicago Press. Poole DFG. 1956. The structure of the teeth of some mammal-like reptiles. Q J Microsc Sci 97:303–312. Poole DFG. 1967. The phylogeny of tooth tissues; enamel and enameloid in recent vertebrates. In: Miles AEW, editor. Structural and chemical organization of teeth I. New York: Academic Press. p 111–149. Poole DFG. 1971. An introduction to the phylogeny of calcified tissues. In: Dahlberg AA, editor. Dental morphology and evolution. Chicago: Chicago University Press. p 65– 80. Preuschoft H, Reif WE, Mu¨ ller WH. 1974. Funktionsanpassungen von Haifischza¨ hnen in Form und Struktur. Z Anat Entwicklungsgesch 143:315–344. Prostak K, Skobe Z. 1986. Ultrastructure of the dental epithelium during enameloid mineralization in a teleost fish, Cichlasoma cyanoguttatum. Arch Oral Biol 31:73– 85. Prostak K, Seifert P, Skobe Z. 1993. Enameloid formation in two tetraodontiform fish species with high and low fluoride contents in enameloid. Arch Oral Biol 38:1031–1044. Rasmussen AS, Arnason U. 1999. Phylogenetic studies of complete mitochondrial DNA molecules place cartilaginous fishes within the tree of bony fishes. J Mol Evol 48:118 –123. Reif WE. 1979. Structural convergences between enameloid of actinopterygian teeth and of shark teeth. Scan Electr Microsc II:547– 554. Reif WE. 1982. Evolution of dermal skeleton and dentition in vertebrates: the odontode regulation theory. Evol Biol 15:287–368. Reith EJ. 1967. The early stages of amelogenesis as observed in molar teeth of young rats. J Ultrastruct Res 17:503–526. Risnes S. 1990. Shark tooth morphogenesis. And SEM and EDX analysis of enameloid and dentin development in various shark species. J Biol Buccale 18:237–248. Ro¨ se C. 1897. Ueber die verschiedenen Aba¨ nderungen der Hartgewebe bei niederen Wirbeltieren. Anat Anzeiger 14:33– 69. Sander PM. 2001. Prismless enamel in amniotes: terminology, function, and evolution. In: Teaford MF, Smith MM, Ferguson MWJ, editors. Development, function and evolution of teeth. Cambridge, UK: Cambridge University Press. p 92–106. Sasagawa I. 1997. Fine structure of the cap enameloid and of the dental epithelial cells during enameloid mineralisation and early maturation stages in the tilapia, a teleost. J Anat 190:589 – 600. Sasagawa I, Ishiyama M. 1988. The structure and development of the collar enameloid in two teleost fishes, Halichoeres poecilopterus and Pagrus major. Anat Embryol 178:499 –511. Satchell PG, Shuler CF, Diekwisch TG. 2000. True enamel covering in teeth of the Australian lungfish Neoceratodus forsteri. Cell Tissue Res 299:27–37. Satchell PG, Anderton X, Ryu OH, Luan X, Ortega AJ, Opamen R, Berman BJ, Witherspoon, DE, Gutmann JL, Yamane A, ZeichnerDavid M, Simmer JP, Shuler CF, Diekwisch TGH. 2002. Conservation and variation in enamel protein distribution during tooth development across vertebrates. Mol Dev Evol J Exp Zool 294:91–106. Schaeffer B. 1977. The dermal skeleton in fishes. In: Andrews SM, Miles RS, Walker AD, editors. Problems in vertebrate evolution. Linnean Society Symposium Series 4. London: Academic Press. p 25–52.

ENAMEL DEVELOPMENT AND EVOLUTION Shellis RP. 1975. A histological and histochemical study of the matrices of enameloid and dentine in teleost fishes. Arch Oral Biol 20: 183–187. Shellis RP, Miles AEW. 1974. Autoradiographic study of the formation of enameloid and dentine matrices in teleost fishes using tritiated amino acids. Proc R Soc Lond B 185:51–72. Shellis RP, Miles AEW. 1976. Observations with the electron microscope on enameloid formation in the common eel (Anguilla anguilla); (Teleostei). Proc R Soc Lond B 194:253–269. Simmer JP, Fincham AG. 1995. Molecular mechanisms of dental enamel formation. Crit Rev Oral Biol Med 6:84 –108. Simmer JP, Fukae M, Tanabe T, Yamakoshi Y, Uchida T, Xue J, Margolis HC, Shimizu M, DeHart BC, Hu CC, Bartlett JD. 1998. Purification, characterization, and cloning of enamel matrix serine proteinase 1. J Dent Res 77:377–386. Slavkin HC, Diekwisch T. 1996. Evolution in tooth developmental biology: of morphology and molecules. Anat Rec 245:131–50. Slavkin HC, Diekwisch TGH. 1997. Molecular strategies of tooth enamel formation are highly conserved during vertebrate evolution. Ciba Found Symp 205:73– 80; discussion 81– 84. Slavkin HC, Zeichner-David M, MacDougall M, Bringas P, Bessem C, Honig LS. 1982. Antibodies to murine amelogenins: localization of enamel proteins during tooth organ development in vitro. Differentiation 23:73– 82. Slavkin HC, Brownell AG, Bringas P Jr, MacDougall M, Bessem C. 1983a. Basal lamina persistence during epithelial-mesenchymal interactions in murine tooth development in vitro. J Craniofac Genet Dev Biol 3:387– 407. Slavkin HC, Samuel N, Bringas P Jr, Nanci A, Santos V. 1983b. Selachian tooth development. II. Immunolocalization of amelogenin polypeptides in epithelium during secretory amelogenesis in Squalus acanthias. J Craniofac Genet Dev Biol 3:43–52. Smith MM. 1978. Enamel in the oral teeth of Latimeria chalumnae (Pisces, Actinistia): A scanning electron microscope study. J Zool Lond 185:355–369.

395

Smith MM. 1992. Microstructure and evolution of enamel amongst osteichthyan fishes and early tetrapods. In: Smith P, Tchernov E, editors. Structure, function, and evolution of teeth. London: Freund Publishing. p 73–101. Smith MM, Miles AEW. 1969. An autoradiographic investigation with the light microscope of [3H] proline incorporation during tooth development in the crested newt (Triturus cristatus). Arch Oral Biol 14:479 – 490. Smith MM, Miles AEW. 1971. The ultrastructure of odontogenesis in larval and adult urodeles; differentiation of the dental epithelial cells. Zeitschr Zellfors 121:470 – 498. Termine JD, Belcourt AB, Christner PJ, Conn KM, Nylen MU. 1980a. Properties of dissociatively extracted fetal tooth matrix proteins. I. Principal molecular species in developing bovine enamel. J Biol Chem 255:9760 –9768. Termine JD, Belcourt AB, Miyamoto MS, Conn KM. 1980b. Properties of dissociatively extracted fetal tooth matrix proteins. II. Separation and purification of fetal bovine dentin phosphoprotein. J Biol Chem 255:9769 –9772. Toyosawa S, O’Huigin C, Figueroa F, Tichy H, Klein J. 1998. Identification and characterization of amelogenin genes in monotremes, reptiles, and amphibians. Proc Natl Acad Sci USA 95:13056 –13061. Veis A, Tompkins K, Alvares K, Wei K, Wang L, Wang XS, Brownell AG, Jengh SM, Healy KE. 2000. Specific amelogenin gene splice products have signaling effects on cells in culture and in implants in vivo. J Biol Chem 275:41263– 41272. Zeichner-David M, Diekwisch T, Fincham A, Lau E, MacDougall M, Moradian-Oldak J, Simmer J, Snead M, Slavkin HC. 1995. Control of ameloblast differentiation. Int J Dev Biol 39:69 –92. Zeichner-David M, Vo H, Tan H, Diekwisch T, Berman B, Thiemann F, Alcocer MD, Hsu P, Wang T, Eyna J, Caton J, Slavkin HC, MacDougall M. 1997. Timing of the expression of enamel gene products during mouse tooth development. Int J Dev Biol 41:27–38.