Ameloblastic Secretion and Calcification of the Enamel Layer in Shark Teeth NORMAN E

JOURNAL OF MORPHOLOGY 1841215-230 (1985)

Ameloblastic Secretion and Calcification of the Enamel Layer in Shark Teeth NORMAN E. KEMP Division of Biological Sciences, The University of Michigan, Ann Arbor, Michigan 48109

ABSTRACT Tooth primordia at early stages of mineralization in the sharks Negaprion brevirostris and Triaenodon obesus were examined electron micro- scopically for evidence of ameloblastic secretion and its relation to calcification of the enamel (enameloid) layer. Ameloblasts are polarized with most of the mitochondria and all of the Golgi dictyosomes localized in the infranuclear end of the cell toward the squamous outer cells of the enamel organ. Endoplasmic reticular membranes and ribosomes are also abundant in this region. Amelo- blastic vesicles bud from the Golgi membranes and evidently move through perinuclear and supranuclear zones to accumulate at the apical end of the cell. The vesicles secrete their contents through the apical cell membrane in merocrine fashion and appear to contribute precursor material both for the basal lamina and the enameline matrix. The enamel layer consists of four zones: a juxta-laminar zone containing newly polymerized mineralizing fibrils (tubules); a pre-enamel zone of assembly of matrix constituents; palisadal zones of mineralizing fibrils (tubules); and interpalisadal zones containing granular amorphous matrix, fine unit fibrils, and giant cross-banded fibers with a periodicity of 17.9 nm. It seems probable that amorphous, non-mineralizing fibrillar and mineralizing fibrillar constituents of the matrix are all products of ameloblastic secretion. Odontoblastic processes are tightly embedded in the matrix of the palisadal zones and do not appear to be secretory at the stages investigated. The shark tooth enamel layer is considered homologous with that of other vertebrates with respect to origin of its mineralizing fibrils from the inner dental epithelium. The term enameloid is appropriate to connote the histological distinction that the enamel layer contains odontoblastic processes but should not signify that shark tooth enamel is a modified type of dentine. How amelogenins and/or enamelins secreted by ameloblasts in the shark and other vertebrates are related to nucleation and growth of enamel crystallites is still not known.

Like the teeth of all gnathostomes, those of '55). Shortly after the enameline matrix ap- sharks develop from epithelial evaginations pears, it begins to calcify by deposition of of the dental lamina surmounting ectomes- hydroxyapatite crystallites. Although the cell enchymal dental papillae (Gaunt and Miles, bodies of odontoblasts become increasingly '67). As a tooth bud elongates, its epithelium separated from the ameloblasts by accumu- heightens by differentiation of columnar lation of this mineralizing matrix, they ameloblasts, and beneath these a layer of nevertheless extend odontoblastic processes odontoblasts differentiates at the periphery which remain embedded in the enameline of the dental papilla (Kerr, '55; Moss, '70, matrix and become longer as the parent '77; Kemp and Park, '74; Kerebel et al., '77; odontoblasts recede farther from the epithe- Samuel et al., '83). Matrix of the enamel lium. After calcification of the enamel layer layer first accumulates at the tip of the tooth is well under way, deposition and calcifica- between ameloblasts and odontoblasts and tion of the collagenous matrix of dentine then extends over the whole crown (Kerr, ensues.

0 1985 ALAN R. LISS. INC. 216 N.E. KEMP

It is the presence of odontoblastic processes To prove homology of the enamel layers of within the enamel layer of sharks' teeth elasmobranchs and mammals, it would be which accounts for the long-standing contro- necessary to demonstrate that elasmobranch versy about the homology of the tooth enam- ameloblasts actually secrete their amelo- el layer in elasmobranchs and higher verte- genin-enamelin-like proteins into the enam- brates. Is the enameline matrix a product of eline matrix and that these proteins order the overlying ameloblasts as in mammals, or development of enamel crystallites as in rather of the odontoblasts and their penetrat- mammalian teeth. That shark tooth amelob- ing odontoblastic processes; or is it a product lasts are indeed secretory has been demon- of both cell layers? As stated earlier (Kemp strated (Mornstad, '74; Kemp, '74a,b; Kemp and Park, '74), "the case for the homology of and Park, '74; Kerebel et al., '77; Nanci et enamel in elasmobranchs rests squarely on al., '83; Samuel et al., '83). Furthermore, it the question of the origin of the organic ma- has been shown that crystallites of the shark trix." Each of these possibilities has had its tooth enamel layer grow to be large and hex- cadre of supporters (Kemp and Park, '74; agonal like those of mammalian enamel Moss, '77; Schaeffer, '77). If the enameline (Poole and Gillett, '69; Kemp and Park, '74). matrix of shark teeth is ameloblastic in ori- How the secretory product of amelo- gin, we may conclude that it is homologous blasts is related to the pattern of mineraliza- to that of mammalian teeth. If it is secreted tion of the enamel of shark's teeth in compar- by odontoblasts, it is actually a specialized ison to those of other vertebrates is the focus type of dentine. If, however, it is a product of of this paper. both cell types, it is different from either enamel or dentine in mammals. The term MATERIALS AND METHODS enameloid, introduced by Poole ('67) and @r- Developing teeth from two shark species, vig ('67), is commonly used in reference to the lemon shark Negaprion breuirostris and the enamel layer of lower vertebrates but the whitetip shark Triaenodon obesus, were leaves open a decision on which cell layers examined in this investigation. Rectangular produce its mineralizing matrix. blocks the width of a single file of teeth were Mammalian ameloblasts secrete matrix excised from the jaws of a 40-cm specimen of proteins called amelogenins and enamelins Negaprion collected by Dr. I. Kaufman Ar- (Graver et al., '78; Termine et al., '80; Fin- enberg at the Lerner Laboratory, Bimini, Ba- cham et al., '82a,b, '83; Slavkin et al., '82; hamas. Similar blocks were excised from a Belcourt et al., '83; Robinson et al., '83; Zeich- 120-cm specimen of Triaenodon collected at ner-David et al., '83) which control nuclea- the Enewatak Marine Biological Laboratory, tion and growth of the relatively large hex- Enewatak Atoll, Marshall Islands. Negu- agonal crystallites of enamel (Kerebel et al., prion tissue was fixed in 5% glutaraldehyde '79; Arends et al., '83). Odontoblasts, on the buffered at pH 7.4 with shark phosphate other hand, secrete a collagenous matrix buffer (Long et al., '68), that of Triaenodon which fosters development of the smaller in 6.25% glutaraldehyde buffered at pH 7.4 mineral crystallites characteristic of dentine with 0.1 M phosphate buffer. Fixed tissue (Glimcher, '81). By using fluorescent antibod- was transported to Ann Arbor and stored in ies to bovine amelogenin, Herold et al. ('80) the fixative in a refrigerator at 10°C for a and Slavkin et al. ('83b) have demonstrated month to two years before further processing. that shark tooth ameloblasts synthesize an Preparation for transmission electron mi- enamel-like protein. Similarly with antibod- croscopy was continued by severing the oral ies to mouse amelogenins Slavkin et al. ('83a) mucosa with attached teeth from the block of have shown that hagfish tooth epithelium jaw skeleton, washing with 0.1 M phosphate evidently secretes the same type of protein. buffer, and post-fixing in ice-cold 1% Os04 in Enamel proteins evolved early in vertebrate the same buffer at pH 7.4. Fixed files of teeth phylogeny and apparently have not changed were dehydrated in a graded series of etha- markedly up to the present (Slavkin et al., nols, cleared in propylene oxide, and embed- '83a, '84; Kemp, '84). Shark tooth dentine ded in Epon 812 in flat containers. After closely resembles mammalian dentine in the polymerization at 60°C the tooth files were relationship between collagen fibrils and cut from the flat block and mounted in a slot mineral crystallites (Garant, '70; Kemp and at the end of a round stub of Epon previously Park, '74). Dentine as a type of mineralized polymerized in a size-00 gelatin capsule. Tis- tissue appears to be unequivocally homolo- sue was oriented and trimmed so that young, gous in all vertebrate groups (Kemp, '84). newly calcifying teeth at the posterior end of SHARK TOOTH ENAMEL LAYER 217 a tooth file were exposed for sectioning. Thin The elongated nuclei of secretory amelo- sections were cut with a diamond knife on an blasts lie approximately midway along the LKB Ultratome and mounted on 200-mesh length of the cell. Cytoplasmic morphology is copper grids. They were stained with uranyl polarized so that the infranuclear (proximal acetate and lead citrate, and micrographs or basal) end is preferentially endowed with were taken with an RCA EMUSE electron organelles indicative of protein synthesis and microscope operating at 50 kV. production of secretory vesicles-mitochondria, Golgi dictyosomes, endoplasmic reticu- RESULTS lar membranes, and ribosomes. From the Morphology of arneloblasts standpoint of ameloblastic function this po- larization of synthetic machinery is impor- Tooth primordia of Negaprion described in tant, for it indicates that it is in this region the present investigation were from the most where the ameloblastic vesicles destined for posterior row showing visible mineraliza- secretion are produced. Such vesicles, appar- tion. Before osmication these young teeth ently budded off from the Golgi membranes, were white at the tip, but they became black- are abundant in the infranuclear zone (Figs. ened after osmication. At this stage calcifi- 2-4). cation was in progress in the tooth tip and Nuclei of ameloblasts have prominent nu- also around the sides (collar) of the tooth in cleoli (Fig. 51, indicating that they have the the enamel layer; however, differentiation of potential for synthesis of ribosomal RNA. ameloblasts and of underlying organic ma- Perinuclear cytoplasm contains cisternae and trix was still under way. Tooth sections vesicles of endoplasmic reticulum as well as showed an inner dental epithelium of colum- ribosomes and occasional mitochondria, but nar ameloblasts surrounding the enamel no Golgi elements. Intercellular boundaries (enameloid) matrix covering the dental pa- at nuclear levels are characterized predomi- pilla. At this stage the enamel organ also nately by smooth gap junctions with occa- includes two layers of squamous epithelial sional desmosomes. There is also some cells enclosing the ameloblasts, namely, the protrusion and interdigitation of cell pro- stratum intermedium and outer dental epi- cesses at this level (Fig. 5). thelium (Fig. 1). Apical to the nucleus the supranuclear cy- Nuclei of cells in both outer dental epithe- toplasm (Fig. 6) resembles that of the peri- lium and stratum intermedium are flattened nuclear region. Endoplasmic reticulum and so that in cross section they appear elongated ribosomes are abundant, but mitochondria circumferentially around the tooth. Nuclei of are scarce and Golgi elements are missing. the ameloblasts, on the other hand, are ori- Moderate interdigitation of cell membranes ented radially with respect to the long axis continues for some distance apical to the nu- of the tooth. Heterochromatin is concen- cleus, but the degree of interdigitation pro- trated along the nuclear membranes and in gresses still more apically to the extent that irregular patches within the nucleoplasm of long, microvillous processes become cells in all three layers of the enamel organ. interpenetrated like interlocking fingers Mitochondria and endoplasmic reticular (Figs. 6, 7, 9). membranes are widely dispersed in cells of Bordering the basal lamina, the cell mem- the intermediate layer, but these organelles brane at the apical end of ameloblasts is ir- are largely eliminated and only a thin rim of regularly folded (Figs. 7-10). Depressions cytoplasm remains in the outer dental between surface folds are sac-like invagina- epithelium. tions which appear round or oval in cross The border between ameloblasts and inter- section (Fig. 9) and enclose inward exten- mediate cells is distinguished by peg-like sions of the basal lamina. Gaps between the protrusions of the ameloblasts (Figs. 1, 2). apical ends of ameloblasts are also filled with Elsewhere along this border gap junctions basal lamina material (Fig. 10). Since com- and occasional desmosomes bind ameloblasts ponents of the basal lamina are considered to overlying cells. Lateral borders of amelo- to be derived chiefly from epithelial cells, blasts at levels basal to the nucleus are bound distribution of the lamina probably reveals by gap junctions and occasional desmosomes, the extent of ameloblastic surface participat- and also tight junctions near their outer ends ing in secretion of ameloblastic products. En- (Figs. 1, 3). Cell membranes may be sepa- largement of the basal lamina shows that its rated enough to allow outgrowth of short, lamina densa is granular and variable in microvillous processes (Fig. 1). thickness (Fig. 8). Granular and filamentous 2 18 N.E. KEMP SHARK TOOTH ENAMEL LAYER 2 19 components of enameline matrix make direct secretory ameloblastic vesicles than does that contact with the under side of the basal of highly interdigitated cells (cf. Figs. 9, 10). lamina. These differences-less interdigitation and Secretory ameloblastic vesicles accumulate more accumulated vesicles-suggest that the toward the apical end of an ameloblast and lateral ameloblasts are at an earlier stage of may become clustered near the cell mem- ameloblastic secretion than are more medial brane (Fig. 10). Secretion is accomplished in ones. Highly interdigitated cells like those the usual merocrine fashion, i.e., by fusion of illustrated in Figures 7 and 9 may be beyond a vesicle to the cell membrane, rupture, and the peak of ameloblastic secretion and may discharge of vesicular contents. Materials for have lost cell volume as a result of consider- the basal lamina and for transport through able secretion. Interdigitation may be a re- it into the enameline matrix are thus made sponse to decreased volume of the supra- available. nuclear zone, analogous to the change in As tooth primordia begin to mineralize, shape of an accordion from its relatively they flatten antero-posterially and the base smooth inflated shape to its highly pleated expands laterally; thus in cross section they deflated shape. are lens-shaped. Around the angle at the lateral border of a tooth the ameloblasts appear Morphology of enameline matrix to be stretched in their supranuclear zones Beneath the basal lamina is the enameline so that they are not highly interdigitated matrix, which can be subdivided into four (Fig. lo), in contrast to ameloblasts surround- zones (Figs. 7,9); Immediately under the lam- ing more medial portions of the tooth rim ina is a juxta-laminar zone of granules and where interdigitation is extensive (Figs. 7,9). fibrils which have begun to mineralize by Moreover, the apical cytoplasm of these more deposition of accompanying hydroxyapatite lateral ameloblasts appears to contain more crystallites. Below this narrow calcifying zone is a non-calcified zone of preenamel, which seems to be an assembly area for the precursor materials of the two deeper-lying zones, the mineralized palisades (Garant, Figs. 1-14. Electron micrographs of dental epithe- '701, and the non-mineralized interpalisadal lium and underlying enameline matrix in cross sections zones. of teeth of Negaprion breuirostris (Figs. 1-13) and Triaen- odon obesus (Fig. 14). Granular material of the enameline matrix is in direct continuity with the granular sub- Fig. 1. Ameloblasts (A), oriented perpendicular to stance of the basal lamina (Fig. 8) and thus tooth surface, are covered by squamous cells of stratum gives the latter an irregular contour. This intermedium (SI)and outer dental epithelium (OE). Peg- like protrusions of ameloblasts penetrate cortex of inter- morphology supports the view that the amor- mediate cells (arrowhead), Baso-lateral ends of amelo- phous granular component of the juxta-lami- blasts adhere closely (arrow), but lateral cell borders nar zone is derived from materials which may be separated by intercellular spaces into which migrate through the basal lamina from its short microvilli (mv) extend. The infranuclear zone of cytoplasm basal to the nucleus (N) contains abundant outer to its inner side. Fibrillogenesis of min- mitochondria (m), Golgi dictyosomes (g), and profiles of eralizing fibrils evidently begins almost as both rough and smooth endoplasmic reticulum (er). soon as precursor materials breach the basal x 7,330. lamina, for such fibrils are often attached to Fig. 2. Infranuclear cytoplasm of ameloblast (A) and the lamina. In fact, the fibrils sometimes ap- adjacent cell of stratum intermedium @I). Peg-like pro- pear anchored in the basal lamina (Fig. 8) as trusions of ameloblasts (arrow) stud surface and enhance though polymerization of fibrils had begun adhesion to SI cells. Junctions between cells are gap even before precursor substance had com- junctions and occasional desmosomes (d). m, mitochon- drion; g, Golgi dictyosome; rer, cisterna of rough endo- pleted its passage through the membrane. plasmic reticulum; ser, vesicle of smooth endoplasmic Primordia of the giant, cross-banded fibers reticulum. ~20,035. which develop within the pre-enamel and interpalisadal zones may also attach directly Fig. 3. Infranuclear cytoplasm of two adjacent ameloblasts (A) showing generally smooth gap junctions to the under side of the basal lamina. Crys- along their borders and a tight junction toward basal tallogenesis of enamel crystallites begins surface (arrow). x 13,360. early in fibrillogenesis, for mineralizing fibrils contacting the basal lamina or embed- Fig. 4. Enlarged view of Golgi regions in Figure 3, showing accumulation of ameloblastic vesicles (av), ap- ded in it may already be accompanied by parently budded from Golgi dictyosomes. X 13,030. crystallites (Figs. 7, 8). 220 N.E. KEMP SHARK TOOTH ENAMEL LAYER 221

Within the pre-enamel zone there are collagen fibers (average 64.0 nm); hence it is amorphous granular material, fine non-min- not likely that they are collagenous. eralizing filaments called “unit fibrils,” and Except for the thin layer of mineralized primordia of the large banded fibers called fibrils in the juxta-laminar zone, sharks’ “giant fibres” (Kemp and Park, ’74). Also in teeth mineralize in blocks of matrix called this zone there are clusters of mineralizing palisades because of their cross-sectional re- fibrils which appear to be emigrating toward semblance to a line of promontories. Odon- the mineralizing palisades (Figs. 7, 9). No toblastic processes extend into the palisades collagen fibers are found in this region. from odontoblasts along the outer border of Morphology of the interpalisadal zones is the dental papilla and are embedded tightly similar to that of the pre-enamel zone. Amor- within the matrix of mineralizing fibrils (Fig. phous material, unit fibrils, and giant fibers 11). There is no accumulation of unmineral- are present, but the latter reach their full ized matrix around these processes; hence potential for growth here. These large, un- they do not appear to be secretory at the mineralized fibers have room to grow radi- stage when mineralization is in progress. At ally between the palisades and may enlarge this time they appear merely to be en- to the full width of the space between adja- trapped. Conceivably odontoblastic processes cent palisades (Figs. 11,lZ). As they increase might help to organize the pattern of pali- either in length or girth, they recruit amor- sades and interpalisadal zones by providing phous and fine fibrillar components of the protoplasmic cores around which enameline matrix. What is most striking about these matrix is deposited. Since there may be sev- fibers is the periodicity of their banding, 17.9 eral odontoblastic processes within a single nm. It is markedly different from that of palisade, however, it seems improbable that individual odontoblasts are responsible for the observed pattern. Alternatively, individual ameloblasts may impose the palisadal- interpalisadal pattern. Although there is no Fig. 5. Elongated nuclei (N) of ameloblasts. Cell obvious one-to-one relation between the api- boundaries at this level are moderately interdigitated cal ends of ameloblasts and the palisadal (arrow). h, heterochromatin; ncl, nucleolus; d, desmo- blocks, the diameters of ameloblasts and pal- some; er, cisterna of endoplasmic reticulum; m, mito- isades are similar in magnitude. chondrion. x 13,055. Crystallites of the enamel layer develop Fig. 6. Cytoplasm at supranuclear level of amelo- within the interior of fibrils, which thus are blasts. Here endoplasmic reticular profiles (er), includ- actually tubules (Kemp and Park, ’74).Dense ing small vesicles thought to be secretory ameloblastie crystallites develop within electron-lucid vesicles (av), are abundant, but mitochondria are sparse and Golgi elements absent at this level. Interdigitation sheaths of organic matrix (Figs. 13,141. Some of cell surfaces is moderate at levels toward nucleus tubules appear to lack interior crystallites. (arrow) but is more pronounced (double arrow) at more Whether tubule formation is a necessary pre- apical levels. ~8,710. requisite for crystallite nucleation, or vice Fig. 7. Border between ameloblasts (A) and enamel versa is a response to the mechanics of crys- layer matrix, consisting of juxta-laminar zone (JL)con- tallite initiation, is an important, unan- taining mineralizing fibrils, pre-enamel zone (PE), mtn- swered question. eralizing palisades (P) and non-mineralizing in- Enamel crystallites in a developing tooth terpalisadal zones (IP). Re-enamel zone contains amorphous granular material, fine fibrils and profiles of po- of the whitetip shark Triaenodon (Fig. 14) lymerizing giant fibrils (gD.Apical surface of ameloblast illustrate the fact that as crystallites grow adjoining basal lamina (BL) is extensively infolded by they assume the hexagonal outline charac- pocket-like invaginations (arrow) lined by basal laminar teristic of apatite minerals (Kemp, ’84).These material. Apico-lateral surfaces of adjacent ameloblasts are highly interdigitated (double arrow). X 21,510. of Triaenodon are from a larger and presum- ably older tooth primordium than those of Fig. 8. Enlarged view of a portion of the border be- Negaprion illustrated in Figure 13. Sheaths tween the ameloblast (A) and enamelin matrix illus- surround most of the Triaenodon crystallites, trated in Figure 7. Granular components of juxta-laminar zone (JL)make direct contact with lamina densa of basal but these investments are either very thin or lamina (arrow). Mineralized fibrils sometimes appear absent from the large crystallites. Appar- anchored in basal lamina (double arrow). Primordium of ently reduction or removal of organic matrix giant fiber (g0 incorporates amorphous granular mate- accompanies crystallite growth. Growing rial and fine, unit fibrils (D during growth. Mineral crystallites (c) are aligned along mineralizing fibrils. crystallites become aligned in cables, within ~75,845. which they are oriented in the same direc- 222 N.E. KEMP SHARK TOOTH ENAMEL LAYER 223 tion. The crystallites extend for long dis- Differentiation of ameloblasts tances without breaks, but their actual During amelogenesis of mammalian teeth lengths have not been determined. Tubules the ameloblasts undergo a series of morpho- (sheaths) around the crystallites limit their logical and functional changes from their cu- growth in width while allowing extensive boidal, preameloblastic, presecretory phase growth in length. How mineral and organic to their tall, columnar, secretory phase and components interact to regulate crystallite then to reduced height again in their matu- growth is another intriguing, but unan- ration phase. Reith ('67, '70) has distin- swered question. guished five stages in amelogenesis of rat DISCUSSION molar teeth: 1) secretory phase; 2) transitional stage when cells begin to diminish in In the long march of time since vertebrates height; 3) preabsorptive phase; 4) early ma- diverged from their protochordate ancestors turation stage when organic materials are (Gans and Northcutt, '83; Northcutt and reabsorbed from the matrix; and 5) late ma- Gans, '83), two separate modes of phosphatic turation stage when water is probably re- calcification of their sclerified tissues have sorbed from the matrix. Maturation stages of evolved (Kemp, '84). The first is collagen- ameloblasts in rat incisor enamel develop- associated mineralization exemplified by the ment show morphological specializations in- development of relatively small, needle- or dicative of resorption of organic matrix and lath-shaped crystallites of hydroxyapatite water from maturing enamel, as well as aligned along collagen fibrils in calcified car- transfer of iron pigment back into it (Kallen- tilage, bone, dentine, and cementum. The bach, '68). Secretory ameloblasts of mice ac- second mode is amelogenin-enamelin-associ- cumulate secretion granules in their Tomes' ated mineralization, characterized by the de- processes (Garant and Nalbandian, '68). velopment of relatively large, hexagonal Differentiating shark tooth ameloblasts crystallites of apatite within a sheath com- pass through stages similar to those of mam- posed of the glycoproteins amelogenin or malian teeth. Kerebel et al. ('77) have de- enamelin in the enamel (enameloid) layer of scribed three stages in development of tooth integumentary scales or teeth. Both modes enamel in the dogfish shark Scyllwrhinus were evidently utilized in development of the canicula, applicable to three other species dermal odontodes and endoskeletal tesserae they studied-Prionace glauca, Squalus of the earliest fossil ostracoderms (Prvig, '77; acanthias and Scymnorhinus lichas. At stage Schaeffer, '77; Kemp and Westrin, '79) and 1 the ameloblasts have centrally located nu- have been retained, probably with little bio- clei, mitochondria are randomly distributed, chemical change, throughout the course of and Golgi dictyosomes are localized at the vertebrate phylogeny (Kemp, '84). basal pole of the cell. At this time the apical border of the cell adjacent to the basal lamina is smooth. At stage 2 the nucleus has shifted toward the basal end of the elongat- Fig. 9. Border between ameloblasts (A) and enamel ing cell and has an irregular surface contour, layer matrix, showing extensive interdigitation of apico- lateral surfaces of adjacent cells (double arrow). Invagi- mitochondria have elongated, and the apical nations of apical surface appear in section as rounded border of the cell has invaginations which profiles (arrow) enclosing basal laminar material. Be- increase its surface area. Secretion from the neath the basal lamina (BL) the enameline matrix con- ameloblasts results in thickening of the basal sists of juxta-laminar zone (&), pre-enamel zone (PEL palisades (P)and interpalisadal zones (IP). er, cisterna of lamina. At stage 3, considered to be a period endoplasmic reticulum; av, ameloblastic vesicle; gf, pri- of maturation, mineralization of the enamel mordium of giant fiber polymerizing from amorphous layer is under way. Mineral crystallites first and finely fibrillar matrix. X22,150. form in the matrix immediately beneath the Fig. 10. Border between ameloblasts (A) and enamel- basal lamina, but only rarely do they contact ine matrix (E) at lateral margin of lens-shaped cross the collagenfiberswhich project into the enam- section of tooth. Arneloblastic vesicles (av) accumulate in eline matrix from its inner side adjacent to apical cytoplasm and fuse with cell surface (arrow) be- odontoblasts. Collagen fibers extending into fore discharging their contents extracellularly. Material of basal lamina (BL) extends into spaces between sepa- the developing enameline layer diminish as rated apical ends of ameloblasts. Basal lamina here is maturation proceeds. relatively thin, and mineralized palisadal matrix abuts According to the description of Kerebel et directly against it; thus, juxta-laminar zone and pre- al. ('771, ameloblasts differentiate in basal- enamel zone (Figs. 7-91 are absent. ~27,870. 224 N.E. KEMP SHARK TOOTH ENAMEL LAYER 225 coronal sequence as tooth primordia elon- now recognized that adult amphibian teeth gate. Those toward the broad base of a miner- have a covering of true enamel (Gillette, ’55; alizing tooth bud are in stage 1, those Meredith Smith and Miles, ’69, ’71). Since somewhat higher in stage 2, and those over the cap layer of larval urodele teeth appears the rest of the crown of the tooth in stage 3. to be different from that of adults, the latter Although ameloblasts at the latter stage are authors theorized that the phylogenetic tran- perhaps beyond their period of most active sition from enameloid to enamel may have secretion, they still contain secretory ame- come about with the acquisition of metamor- loblastic vesicles and are probably still se- phosis in the ontogeny of ancestral amphibi- creting them. Ameloblasts described in this ans. In contemporary usage the term investigation were at Kerebel stage 3, al- enameloid is commonly used to denote the though those around the lateral edge of a cap layer of fish teeth (Moss, ’70). In a gen- developing tooth were evidently at a younger eral sense, the term merely implies that there stage than those over the flattened surfaces. is a difference between the cap layer of fish In Negaprion the ameloblasts show charac- teeth and the ectodermal enamel of higher teristic features of secretory cells-polariza- vertebrates. About the nature of the differ- tion of cytoplasmic organelles, accumulation ence it is non-committal. Conceivably the cap and secretion of ameloblastic vesicles, and layer could be derived from ectoderm, from cell surface changes which include lateral ectomesenchyme, or from both sources interdigitations and apical invaginations. (Schaeffer, ’77). Used more specifically, the term enameloid is often used as a substitute Enamel uersus enameloid for the older terms vitrodentine, mesodermal The term enameloid was originally intro- enamel, and durodentine, denoting the cap duced to distinguish the tooth cap of lower layer as modified dentine and therefore of vertebrates from the ectodermally derived ectomesenchymal origin (Poole, ’71; Mere- “true enamel” of reptiles and mammals dith Smith and Miles, ’71; Meinke and (Poole, ’67; grvig, ’67), because it was contro- Thomson, ’83).Some authors consider the cap versial whether the outer tooth layer of fishes layer in either elasmobranch or teleost teeth and amphibians was derived from amelo- to be of mixed origin, i.e., both epithelial and blasts or was a modified type of dentine de- ectomesenchymal (Shellis and Miles, ’74, ’76; rived from mesenchymal odontoblasts. It is Shellis, ’78; Reif, ’79). For some fish groups though, it has been demonstrated that the cap layer in reality is a product of the inner dental epithelium and therefore is true enamel. Whenever the question of origin remains unsettled, the term enameloid is a useful Fig. 11. Enameline matrix below pre-enamel zone, designation. showing odontoblastic processes (op) embedded in pali- Herold (’74) has concluded that the thin sade (P). Non-mineralizingfibrils (0 have polymerized in interpalisadal zone (IP). x 17,195. surface “enameloid” layer covering teeth of the great northern pike Esox lucius is true Fig. 12. Interpalisadal zone (IP) between adjacent pal- ectodermal enamel homologous to that of isades (PI, showing giant fiber (gfl polymerized from mammalian enamel. Its matrix is secreted amorphous material (a) and unit fibrils (0. Banding periodicity is 17.9 nm. ~55,850. by the inner dental epithelial cells (ameloblasts); its organic component is a granular, Fig. 13. Small, dense crystallites of enameline matrix homogeneous, non-fibrillar substance; and its in a mineralizing palisade of Neguprion breuirostris, sur- mineralized component is calcified through rounded by less-dense sheath (arrow) of mineralizing fibrils (tubules). Some fibrils appear to be unmineral- ionic transport of calcium from dental epithe- ized. ~91,030. lial cells. Enameloid in the angler fish, Le phius piscatorius, appears to be like that of Fig. 14. Crystallites of enameline matrix from tooth Esox (Kerebel and Le Cabellec, ’80). Garant’s of Triaenodon obesus, larger than that of the Neguprion specimen utilized for the illustration in Figure 13. Hex- (’70) opinion that the surface layer of teeth agonal shape of larger crystallites is clearly apparent. in the blue gourami Helostoma temmincki is Younger crystallites are surrounded by sheath (arrow); modified dentine has led Herold (‘74) to spec- older ones have thinner sheath or have lost it. Crystal- ulate that the ability to produce ectodermal lites run in bundles within which they are oriented in the same direction, as in the groups along left and right enamel may have evolved during “the evo- sides and the curving central group of this micrograph. lutionary period of the fish.” Since Esox and ~88,060. 226 N.E. KEMP Lophius are more primitive than Helostoma, the view that the mineralized components of one explanation for the proposed differences the shark tooth cap layer are of ectodermal in their cap layers might be that primitive origin and therefore homologous with mam- actinopterygians possessed ectodermal enam- malian enamel. In a study on comparative el, whereas some more advanced members of staining rt?actions of shark tooth enameloid the group may have lost it. On the basis of and bovinu enamel, Everett and Miller (’81) staining reactions in tissue sections, Shellis found differences sufficient to make their ho- (’75) has concluded that the mineralizing mology questionable. They speculated, how- enameloid matrix of four teleost species ever, that there might be an ectodermal (eel,wrasse, bass, and sea bream) receives a contribution to shark tooth enameloid, local- contribution of ectodermal protein from the izing particularly in the transitional zone be- inner dental epithelium. tween that layer and dentine. Fish at the Qrvig (‘78a,b,c) has examined the dermal chondricht hyan level of evolution evidently odontodes and teeth of several fossil palaeo- had already evolved true enamel. Elasmo- nisciform actinopterygian fishes, which have branch teeth do differ from those of tetrapods cap layers of hypermineralized matrix called by the presence of odontoblastic processes ex- ganoin or acrodin. He concludes that al- tending into the enamel layer. To connote though the origin of these layers is still un- this histological difference the term enamel- certain they resemble the “enameloid” of oid is useful, but it should not imply that the elasmobranch teeth. They probably received mode of mineralization of the enamel layer an ectodermal contribution, although they necessarily differs in these groups (Kemp, ’84). might have received a mesenchymal one as Was ectodermal enamel an invention of the well (Qrvig, ’78c). Chondrichthyes, the first gnathostome ver- Considering the Dipnoi and Crossopterygii tebrates? Not likely, for the enameloid of der- as sister groups with the Actinopterygii mal odontodes in the earliest known among the Osteichthyes (Gardiner, ’73), these agnathan vertebrates, e.g., the heterostracan groups must have evolved from a common ostracoderm Astraspis, appears to have been ancestor. Among the crossopterygians the ex- like that of elasmobranch dermal denticles tant actinistian coelacanth Latimeria cha- and teeth (Denison ’67; Qrvig, ’67; Halstead, lumnae has teeth with ectodermal enamel ’73). The presence of odontoblastic processes (Miller, ’69; Grady, ’70; Shellis and Poole, penetrating the enamel layer of the odon- ‘78; Smith, ’78). The fossil rhipidistian fish todes in osfxacodermsjustifies use of the term Onychodus sp. also had teeth with “pseudo- enameloid for the reason discussed above, prismatic enamel” considered to be homolo- but the source of the mineralizing matrix for gous with mammalian enamel (Smith, ’79). this layer was most probably ectoderm rather Dermal denticles in the fossil coelacanth than mesenchyme. It follows from this inter- Spermatodus pustulosus (Meinke, ’82) were pretation that ectodermal enamel as a type capped with ectodermal enamel. From this of mineralized tissue ranks with calcified emerging pool of information about the enam- cartilage, bone, and dentine among the most el layer in crossopterygians, Meinke and primitive vertebrate hard tissues (Moss, ’69, Thomson (’83)have concluded that enamel in ’70, ’77; Kemp and Park, ’74; Kemp and Wes- tetrapods is a primitive character derived trin, ’79). from their lobe-finned ancestors. The enamel According to Reif (’791, the enamel layer layer of dipnoan teeth also appears to be (enameloid) has evolved structurally from a derived from inner dental epithelium (Be- condition of random orientation of crystals mis,’84). Available facts point to the supposi- (“single-crystallite enameloid”), as found in tion that all osteichthyan fish have inherited ostracoderrns and fossil elasmobranchs, to a from a common ancient progenitor the pro- more highly ordered arrangement (“parallel- clivity for coating their teeth with ectoder- structured enameloid”) in Euselachii and te- ma1 enamel. The same holds for Teleostomi trapods. Nevertheless it is doubtful that the in general, since the Acanthodii probably basic homology of the enameloid layer has shared common ancestry with the Ostei- been altered by such pattern changes. As chthyes (Gardiner, ’73). expressed by Moss (’691, tooth tissues result- Evidence presented here and elsewhere ing from epithelio-mesenchymal interactions (Moss, ’70; Kemp, ’74; Kemp and Park, ’74; may “differ in detail but utilize similar de- Kerebel et al., ’77; Nanci et al., ’83)supports velopmentd processes.” SHARK TOOTH ENAMEL LAYER 227 Fibrillogenesis and crystallogenesis in the Previously it was conjectured that the min- enameline matrix eralizing fibrils of shark tooth enamel are a product of ameloblastic secretion, but that In an earlier publication (Kemp and Park. the non-mineralizing giant fibers of the in-, '74) the significance of three kinds of fibrils terpalisadal zones might be an odontoblastic in the enamel layer matrix was discussed. product (Kemp and Park, '74). Although the Conceivably their protein precursors could banding periodicity of these large fibers (17.9 be all of ameloblastic origin, all of odonto- nm) differs markedly from that of conven- blastic origin, or mixed in their cells of ori- tional collagen (64.0 nm), it is possible that gin. The fine "unit fibrils" and large, cross- their precursor is secreted by odontoblasts banded "giant" fibers of the interpalisadal and as a protein different from that of the zones appear to be successive stages of poly- mineralizing fibrils segregates to form the merization of the same precursor. Both may interpalisadal zones. A priori, however, one be designated as non-mineralizing fibrils. would expect the fibrillar product of odontob- The fibrils which become mineralized, on the lasts to be collagenous. Kerebel et al. ('77) other hand, develop as hollow tubules consti- have shown that collagen fibrils may be pres- tuting the sheaths for the mineral crystal- ent between inner dental epithelium and lites. Whether the non-mineralizing fibrils dental papilla cells at the beginning of ame- and the mineralizing tubules are different logenesis, but that they recede toward the states of polymerization of a common precur- interior of the tooth as the mineral layer is sor or develop from two distinct kinds of pro- deposited. Conventional collagen is no longer tein is not known. present in the mineralizing enamel matrix. Mammalian enamel crystallites develop Observations in this investigation support within a sheath, first described at the elec- the supposition that all three kinds of fibrils tron microscopic level by Scott and Nylen in the enameline matrix-unit fibrils, giant ('62) from decalcified sections. In cross sec- fibers, and mineralizing fibrils (tubules)-are tion this sheath is oval, correlated with the products of ameloblastic secretion, polymer- observation that mammalian enamel crys- izing from a single protein pool or from two tallites are greater in width than in thick- segregating pools. ness (Travis and Glimcher, '64; Jessen, '68). Mammalian amelogenins have a high-mo- These early workers envisioned the organic lecular-weight precursor of about 40,000 dal- matrix as organized into hollow tubules tons during early amelogenesis, and this may which control the orientation and alignment be depolymerized to proteins of lower molec- of crystallites in their prisms and inter- ular weight as development proceeds (Fin- prisms. Daculsi and Kerebel ('781, however, cham et al., '82a,b). Enamelins are evidently have questioned the reality of crystallite more stable and retained longer during tooth sheaths, speculating that they could be arti- maturation. Slavkin et al. ('84) suggest that facts of fixation, staining, or electron-beam enamelins are the predominant type of enam- damage. el layer proteins in aquatic vertebrates, but Shark tooth enamel crystallites do develop that evolution has selected amelogenins as within sheaths. Protein for the tubular the predominant enameline protein in ter- sheath is probably amelogenin andor enam- restrial vertebrates. So far enamel layer pro- elin, since it is known that these precursors teins extracted from elasmobranchs have not synthesized by mammalian ameloblasts are been characterized directly. also produced in shark teeth (Herold et al., Apatite minerals typically exhibit a hex- '80; Nanci et al., '83). These glycoproteins are agonal mode of mineralization (McConnell, probably secreted together with constituents '73). Thus it is not surprising that the hy- of the basal lamina, including type 4 colla- droxyapatite crystallites of enamel are hex- gen, laminin and proteoglycans, but fibronec- agonal in all vertebrates from sharks to man. tin for the basal lamina is considered to be In humans the crystallites grow to be about an odontoblastic product (Slavkin et al., '84). twice as wide as they are thick (Daculsi and In sharks the basal lamina appears to persist Kerebel, '78). They grow in two stages. Ini- during secretory activity of ameloblasts, tially they grow more rapidly in width than whereas in mammalian teeth it characterist- in thickness and during this time are sur- ically disperses during that period (Kallen- rounded by a sheath. After a time they ap- bach and Piesco, '78). pear to lose their sheath, remain about the 228 N.E. KEMP same in width, but continue to grow in thick- remineralisation. In S. A. Leach and W. M. Edgar (eds): ness. Within prisms crystallites become Demineralisation and Remineralisation of the Teeth. Oxford: IRL Press Ltd., pp. 155-163. aligned end-to-end and oriented along the Belcourt, A.B., F.G. Fincham, and J.D. Termine (1983) length of the prism (Kerebel et al., '79). Shark Bovine high molecular weight amelogenin proteins. tooth enamel crystallites are hexagonal but Calcif. Tissue Int. 35:111-114. tend to become equilateral with growth Bemis, W.E. (1984) Morphology and growth of lepidosi- renid lungfish tooth plates (Pisces: Dipnoi). J. Morphol. (Kemp and Park, '74; Kerebel et al., '77; 179:73-93. Kemp, '84), possibly because of their high Daculsi, G., and B. Kerebel (1978) High-resolution elec- fluoride content (Ripa et a1 ., '72; Daculsi and tron microscope study of human enamel crystallites: Kerebel, '80). Within palisades crystallites size, shape, and growth. J. Ultrastruct. Res. 65:163- 172. are organized into bundles with individual Daculsi, G., :tnd L.M. Kerebel (1980) Ultrastructural crystallites running in parallel. study and comparative analysis of fluoride content of How fibrils of the enameline matrix are enameloid in sea-water and fresh-water sharks. Arch. related to precursor proteins, in particular Oral Biol. 25:145-151. Denison, R.H. (1967) Ordovician vertebrates from west- the amelogenins and enamelins, remains to ern United States. Fieldiana (Geol.) 16131-192. be elucidated. In addition, how fibrillogene- Doi, Y., E.D. Eanes, H. Shimokawa, and J.D. Termine sis is related to crystallogenesis is still an (1984)Inhibition of seed growth of enamel apatite crys- unsolved mystery (Swancar et al., '71). Fi- tals by amelogenin and enamelin protein in uitro. J. Dent. Res., 6.3:98-105. brillar sheaths appear to direct orientation Everett, M.M., and W.A. Miller (1981)Histochemistry of and alignment of crystallites, but in later lower vertebrate calcified structures. I. Enamel of the stages of crystal growth the organic matrix dogfish Squcdus acanthias compared with mammalian is degraded (Stack, '67) and may serve only enamel and homologous dentine. J. Morphol. 270:95- 111. as a vehicle for diffusion of mineral ions in Fincham, A.Cr., A.B. Belcourt, D.M. Lyaruu, and J.D. solution. In vitro experiments on the proper- Termine (1982a) Comparative protein biochemistry of ties of amelogenins and enamelins relative developing dental enamel matrix from five mamma- to fibrillogenesis and crystallogenesis should lian species. Calcif. Tissue Int., 34:182-189. be instructive. Doi et al. ('84) have shown Fincham, A.G., A.B. Belcourt, and J.D. Termine (1982b3) Changing patterns of enamel matrix proteins in the that in culture both of these glycoproteins developing bovine tooth. Caries Res. Ifi:64-71. inhibited growth of seeded apatite crystals. Fincham, A.G., A.B. Belcourt, J.D. Termine, W.T. Butler, Enamelins were more inhibitory than ame- and W.C. Cochran (1983) Amelogenins: sequence ho- logenins. Further experimental analysis will mologies in enamel-matrix proteins from 3 mammalian species. Biochem. J. 211:149-154. help to clarify how these proteins control nu- Cans, C., and R.G. Northcutt (1983)Neural crest and the cleation, early growth and orientation of origin of vertebrates: a new head. Science 220:268- crystallites. 274. Garant, P.R. (1970) An electron microscopic study of the ACKNOWLEDGMENTS crystal-matrix relationship in the teeth of the dogfish Squalus acanthias. J. Ultrastruct. Res. 30:441-449. This research was supported by NSF Grant Garant, P.R., and J. Nalbandian (1968) Observations on 4317, USPHS Grant AM13745, and grants the ultrastructure of ameloblasts with special refer- from the University of Michigan Cancer Re- ence to the Golgi complex and related complexes. J. search Institute. I thank Dr. I. Kaufman Ar- Ultrastruct. Res. 2.3427-443. Gardiner, B.G. (1973) Interrelationships of teleostomes. enberg for donation of Negaprion tissue he In P. H. Greenwood, R.S. Miles, and C. Patterson (eds): obtained during a period of research at the Interrelationships of Fishes. Zool. J. Linn. Soc., Suppl. Lerner Laboratory, Bimini, Bahamas in 1, Vol. 53. London, New York: Academic Press, pp. 105- 1972. Triaemdon tissue was obtained by the 135. Gaunt, W.A., and A.E.W. Miles (1967) Fundamental as- author at Enewatak Atoll, Marshall Islands, pects of tooth morphogenesis. In A.E.W. Miles (ed): during a research project at the Mid-Pacific Structural and Chemical Organization of Teet.h. Vol. I. Marine Laboratory there in 1972, sponsored New York: Academic Press, pp. 151-197. by the U.S. Department of Energy under an Gillette, R. (1955)The dynamics of continuous succession of teeth in the frog (Rampipiens). Am. J. Anat. 96:l- institutional grant to the Hawaii Institute of 36. Marine Biology, University of Hawaii. I am Glimcher, M.J. (1981) On the form and function of bone: grateful to Professor John E. Bardach, then from molecules to organs. Wolff's law revisited, 1981. Director of the HIMB, for that opportunity. I In A. Veis (ed): The Chemistry and Biology of Miner- am especially indebted to Sandra K. Westrin alized Connective Tissues. New York: Elsevier-North Holland, pp. 617-675. for technical assistance in electron micros- Grady, J.E. (1970) Tooth development in Latimeria cha- COPY. lumnae (Smith). J. Morphol. 132:377-388. LITERATURE CITED Graver, H.T., R.C. Herold, T.-Y. Chung, P.J. Christner, C. Pappas, and J. Rosenbloom (1978) Immunofluores- Arends, J., W.L. Jongebloed, and J. Schuthof (1983) The cent localization of amelogenins in developing bovine ultrastructure of surface enamel in relation to de- and teeth. Dev. Biol. 63:390-401. SHARK TOOTH ENAMEL LAYER 229

Halstead, L.B. (1973) The heterostracan fishes. Biol. Rev. Meredith Smith, M., and A.E.W. Miles (1971) The ultra- 48:279-332. structure of odontogenesis in larval and adult uro- Herold, R.C.B. (1974) Ultrastructure of odontogenesis in deles; differentiation of the dental epithelial cells. 2. the pike (Esox Zucius). Role of dental epithelium and Zellforsch. 121:470-498. formation of enameloid layer. J. Ultrastruct. Res. Miller, W.A. (1969) Tooth enamel of Latirnerzu chalum- 48435-454. ROA (Smith). Nature 221:1244. Herold, R.C., H.T. Graver, and P. Christner (1980) Immu- Mornstad, H. (1974) On the histogenesis of shark enam- nohistochemical localization of amelogenins in enamel. Odontolog. Revy 25: 317-325. eloid of lower vertebrate teeth. Science 2071357-1358. Moss, M.L. (1969) Phylogeny and comparative anatomy Jessen. H. (1968) Elliutical tubules as unit structure of of oral ectodermal-ectomesenchymal inductive inter- forming enamel matrix in the rat. Arch. Oral Biol. actions. J. Dent. Res. 48:732-737. 13:351-352...~~.~ Moss, M.L. (1970) Enamel and bone in shark teeth with Kallenbach, E. (1968) Fine structure of rat incisor ame- a note on fibrous enamel in fishes. Acta Anat. 77161- loblasts during enamel maturation. J. Ultrastruct. Res. 187. 2290-119. Moss, M.L. (1977) Skeletal tissues in sharks. Am. 2001. Kallenbach, E., and N.D. Piesco (1978) The changing 17:335-342. morphology of the epithelium-mesenchyme interface Nanci, A,, P. Bringas, Jr., N. Samuel, and H.C. Slavkin in the differentiation zone of growing teeth of selected (1983)Selachian tooth development. 3. Ultrastructural vertebrates and its relationship to possible mecha- features of secretory amelogenesis in Squalus acan- nisms of differentiation. J. Biol. Buccale 6:229-240. thias. J. Craniofac. Genet. Dev. Biol. 353-74. Kemp, N.E. (1974a) Ultrastructure of ameloblasts and Northcutt, R.G., and C. Gans (1983) The genesis of neural enamel in shark teeth. In C.J. Arceneaux (ed): Proc. crest and epidermal placodes: a reinterpretation of ver- Electron Micr. Soc. America, 32nd Annual Meeting, tebrate origins. Q. Rev. Biol. 58:l-28. 1974. Baton Rouge, Louisiana: Claitor's Publ. Div., pp. Prvig, T. (1967) Phylogeny of tooth tissues. Evolution ~~ of 76-77 (abstract). some calcified tissues in early vertebrates. In A.E.W. Kemp, N.E. (1974b) Ameloblastic secretion in calcifying Miles (ed): Structural and Chemical Organization of teeth of sharks. In J.V. Sanders and D.J. Goodchild Teeth. Vol. I. New York Academic Press, pp. 45-110. (eds): Electron Microscopy 1974. Vol. I1 Biological. 8th Qrvig, T. (1977) A survey of odontodes ('dermal teeth') Intern. Cong. Electron Micros. Canberra: Australian from developmental, structural, functional, and phy- Acad. Sci., pp. 458-459 (abstract). letic points of view. In S.M. Andrews, R.S. Miles, and Kemp, N.E. (1984) Organic matrices and mineral crys- A.D. Walker (eds): Problems in Vertebrate Evolution. tallites in vertebrate scales, teeth and skeletons. Am. New York: Academic Press, pp. 53-75. Zool. 24:965-976. Prvig, T. (1978a) Microstructure and growth of the der- Kemp, N.E., and J.H. Park (1974) Ultrastructure of the mal skeleton in fossil actinopterygian fishes: Eirgeria enamel layer in developing teeth of the shark Carcha- and Scanilepis. Zool. Scripta 733-56. rhinus menisorrah. Arch. Oral Biol. 19:633-644. Qrvig, T. (1978b) Microstructure and growth of the der- Kemp, N.E., and S.K. Westrin (1979) Ultrastructure of mal skeleton in fossil actinopterygian fishes: Boreoso calcified cartilage in the endoskeletal tesserae of mus, Plegmolepis and Gyrolepis. 2001. Scripta 7125- sharks. J. Morphol. 160:75-102. 144. Kerebel, L.-M., and M.-T. Le Cabellec (1980) Enameloid Qrvig, T. (1978~)Microstructure and growth of the der- in the teleost fish Lophius: an ultrastructural study. mal skeleton in fossil actinopterygian fishes: Nephro- Cell Tissue Res. 206211-223. tus and Colobodus, with remarks on the dentition in Kerebel, B., G. Daculsi, and S. Renaudin (1977) Ultra- other forms. Zool. Scripta 7297-326. structure des ameloblastes au cows de la formation de Poole, D.F.G. (1967) Phylogeny of tooth tissues: enamel- l'emailoide des Slaciens. Biol. Cell. 28:125-130. oid and enamel in recent vertebrates, with a note on Kerebel, B., G. Daculsi, and L.-M. Kerebel(1979) Ultra- the history of cementum. In A.E.W. Miles (ed): Struc- structural studies of enamel crystallites. J. Dent. Res. tural and Chemical Organization of Teeth. Vol. I. New 58:844-850. York: Academic Press, pp. 111-149. Kerr, T. (1955) Development and structure of the teeth Poole, D.F.G. (1971)An introduction to the phylogeny of in the dogfish Squalus acanthias (L.)and Scyliorhinus calcified tissues. In A.A. Dahlberg (edj: Dental Mor- curzLculus (I,.).Proc. Zool. SOC.Lond. 12595-114. phology and Evolution. Chicago and London: Univer- Long, D.M., T.S. Bodenheimer, J.F. Hartmann, and I. sity of Chicago Press, pp. 65-79. Klatzko (1968) Ultrastructural features of the shark Poole, D.F.G., and M.S. Gillett (1969) Collagen fibers in brain. Am. J. Anat. 122:209-238. immature enameloid of dogfish teeth. J. Dent. Res. McConnell, D. (1973) Apatite: Its Crystal Chemistry, 48:1119 (abstract). Mineralogy, Utilization, and Geologic and Biologic Oc- Reif, W.-E. (1979) Structural convergences between currences, New York: Springer-Verlag. enameloid of actinopterygian teeth and of shark teeth. Meinke, D. (1982) A light microscope and scanning elec- In 0. Johari and R.P. Becker (edsj: Scanning Electron tron microscope study on the dermal skeleton of Sper- Microscopy/l979/II. A.M.F. O'Hare, Illinois: Scanning matodus pustulosus (Pisces: Coelacanthini) and the Electron Microscopy, Inc., pp. 547-554. evolution of the dermal skeleton in coelacanths. J. Reith, E.J. (1967) The early stage of amelogenesis as Paleontol. 56:620-630. observed in molar teeth of young rats. J. Ultrastruct. Meinke, D., and K.S. Thomson (1983) The distribution Res. 17503-526. and significance of enamel and enameloid in the der- Reith, E.J. (1970) The stages of amelogenesis as observed mal skeleton of osteolepiform rhipidistian fishes. Pa- in molar teeth of yocng rats. J.- Ultrastruct. Res. leobiology 9313s ?. 30:lll-151. Meredith Smith, M., and A.E.W. Miles (1969) An auto- Ripa, L.W., A.J. Gwinnett, C. Guzman, and D. Legler radiographic investigation with the light microscope (1972) Microstructural and microradiographic quali- of proline-H3 incorporation during tooth development ties of lemon shark enameloid. Arch. Oral Biol. 17:165- in the crested newt (Triturus cristatus). Arch. Oral 173. Biol. 14:479-490. 230 N.E. KEMP

Robinson, C., H.D. Briggs, J. Kirkham, and P.J. Atkin- Slavkin, H.C., N. Samuel, P. Bringas, Jr., A. Nanci, and son (1983) Changes in the protein components of rat V. Santos (1983b) Selachian tooth development. 2. Im- incisor enamel during tooth development. Arch. Oral munolocalization of amelogenin polypeptides in epi- Biol. %8:993-1000. thelium during secretory amelogenesis in Squalus Samuel, N., P. Bringas, Jr., V. Santos, A. Nanci, and acanthias. J. Craniofac. Gen. Dev. Biol. 3:43-52. H.C. Slavkin (1983) Selachian tooth development. 1. Slavkin, H.C., M.L. Snead, M. Zeichner-David, P. Brin- Histogenesis, morphogenesis and anatomical features gas, Jr., and G.L. Greenberg (1984) Amelogenin gene in Squalus acanthias. J. Craniofac. Genet. Dev. Biol. expression during epithelial-mesenchymal interac- 3:29-42. tions. In R.1,. Trelstad (ed): The Role of Extracellular Schaeffer, B. (1977) The dermal skeleton in fishes. In Matrix in Development. New York: Alan R. Liss, Inc., S.M. Andrews, R.S. Miles, and A.D. Walker (eds): Prob- pp. 221-253 lems in Vertebrate Evolution. New York Academic Smith, M.M. t 1978)Enamel in the oral teeth of Latimeria Press, pp. 25-52. chalumnae [Pisces: Actinistia): a SEM study. J. Zool. Scott, D.B., and M.U. Nylen 11962) Organic-inorganic Gond.) 185:355-370. interrelationships in enamel and dentin-a possible Smith, M.M. (19791 SEM of the enamel layer in oral key to the mechanism of caries. Int. Dent. J. 12:417- teeth of fossil and extant crossopterygian and dipnoan 442. fishes. In 0.Johari and R.P. Becker (eds): Scanning Shellis, R.P. (1975) A histological and histochemical Electron Microscopy/l979/II. A.M.F. O’Hare, Illinois: study of the matrices of enameloid and dentine in Scanning Electron Microscopy, Inc., pp. 483-489. teleost fishes. Arch. Oral Biol. 20:183-188. Stack, M.V. (1967) Chemical organization of the organic Shellis, R.P. (1978) The role of the inner dental epithe- matrix of enamel. In A.E.W. Miles (ed): Structural and lium in the formation of the teeth in fish. In P.M. Chemical Clrganization of Teeth. Vol. 11. New York: Butler and K.A. Joysey (eds): Development, Function Academic Press, pp. 317-346. and Evolution of Teeth. London, New York: Academic Swancar, J.R., D.B. Scott, J.W. Simmelink, and T.J. Press, pp. 30-42. Smith (1971) The morphology of enamel crystals. In Shellis, R.P., and A.E.W. Miles (1974) Autoradiographic R.W. Fearnliead and M.W. Stack (eds): Tooth Enamel study of the formation of enameloid and dentine ma- 11. Its Composition, Properties and Fundamental trices in teleost fishes using tritiated amino acids. Roc. Structure. Ekistol: John Wright & Sons Ltd., pp. 233- R. Soc. Lond. 03) 185:51-72. 239. Shellis, R.P., and A.E.W. Miles (1976) Observations with Termine, J.D.. A.B. Belcourt, P.J. Christner, K.M. Conn, the electron microscope on enameloid formation in the and M.V.Ny len (1980) Properties of dissociatively ex- common eel (Anguilla anguilla: Teleostei). Proc. R. Soc. tracted fetal tooth matrix proteins. I. Principal molec- Lond. CB) 194253-269, ular species in developing bovine enamel. J. Biol. Shellis, R.P., and D.F.G. Poole (1978) The structure of Chem. 255:0760-9768. the dental hard tissues of the coelacanthid fish Lati- Travis, D.F., and M.J. Glimcher (1964)The structure and meria chalumnae Smith. Arch. Oral Biol. 23:1105-1113. organization of, and the relationship between the or- Slavkin, H.C., M. Zeichner-David, D., M. MacDougall, P. ganic matrix and the inorganic crystals of embryonic Bringas, C. Bessem, and L.S. Honig (1982) Antibodies bovine enamel. J. Cell Biol. 23447-497. to murine amelogenins: localization of enamel proteins Zeichner-David, M., H.C. Slavkin, D.M. Lyaruu, and J.D. during tooth organ devlcopmcnt in vitro. Differentia- Termine (1983) Biosynthesis and secretion of enamel tion 23:73-82. proteins during hamster tooth development. Calcif. Slavkin, H.C., E. Graham, M. Zeichner-David, and W. Tissue Int. 65:366-371. Hildemann (1983a) Enamel-like antigens in hagfish; possible evolutionary significance. Evolution 37:404- 412.