JOURNAL OF MORPHOLOGY 231:161–174 (1997)

Evidence for Participation of the Epidermis in the Deposition of Superficial Layer of Scales in Zebrafish (Danio rerio): A SEM and TEM Study JEAN-YVES SIRE,* ALEXANDRA QUILHAC, JACQUELINE BOURGUIGNON, AND FRANCOISE ALLIZARD Universite´ Paris 7, 75251 Paris cedex 05, France

ABSTRACT Comparative studies on scale structure and development in bony fish have led to the hypothesis that elasmoid scales in teleosts could be dental in origin. The present work was undertaken to determine whether the scales in zebrafish (Danio rerio), a species widely used in genetics and developmental biology, would be an appropriate focus for further studies devoted to the immunodetection of dental components or to the detection of the expression of genes coding for various dental proteins in fish scales. The superficial region of mature and experimentally regenerated scales and its relationships to the epidermal cover were studied in adult zebrafish using scanning (SEM) and transmission (TEM) electron microscopy. The elasmoid scales are relatively large, thin, and are located in the upper region of the dermis, close to the epidermis. In adults, the surface of the posterior region appears smooth at the SEM level and is entirely covered by the epidermis. During regeneration, the relationship of the epidermal cover to the scale surface is established within 4 days. This interface is easier to study in regenerating than in mature scales because the former are poorly mineralized. TEM revealed that: (1) the epidermis is in direct contact with the scale surface, from which it is separated only by a basement membrane-like structure, (2) there are no dermal elements at the scale surface except at the level of grooves issuing from the focus and crossing the scale surface radially, (3) the mineral crystals located in this superficial region are perpendicular to the scale surface, whereas those located deeper within the collagenous scale matrix are randomly disposed, and (4) when decalcified, the matrix of the superficial region of the scale appears devoid of collagen fibrils but contains thin electron-dense granules, some of which are arranged into layers. The continuous epidermal covering, the absence of dermal elements, as well as the fine structure of the matrix and its type of mineralization, strongly suggest that epidermal products, possibly enamel-like proteins, are deposited at the scale surface and contribute to the thickening of the upper layer in zebrafish scales. J. Morphol. 231:161–174, 1997. r 1997 Wiley-Liss, Inc. In most bony fish the body is protected by a dermal skeleton of scales that may correspond to the ‘‘exoskeleton’’ that covered the early vertebrates (e.g., Heterostracans, Osteostracans) (Schultze, ’66, ’77; Reif, ’82; Smith and Hall, ’90). A clear tendency toward a reduction of the postcranial dermal skeleton is observed in the course of evolution, and only a few tetrapod species (mainly reptiles, some amphibians and mammals) possess dermal sclerifications. In contrast, r 1997 WILEY-LISS, INC.

bony fish (i.e., osteichthyans) show a considerable diversity of their dermal skeleton (various types of scales and bony plates) and of the constituent tissues (various types of bone, of dentine, enamel, enameloid, and highly derived tissues) (Francillon-Vieillot et al., ’90; Zylberberg et al., ’92). For over a *Correspondence to: Dr. Jean-Yves Sire, Equipe de recherche ‘‘Formations Squelettiques,’’ URA CNRS 1137, Universite´ Paris 7, Laboratoire d’Anatomie compare´e, Case 7077, 2, Place Jussieu, 75251 Paris cedex 05, France.

162

J.-Y. SIRE ET AL.

decade, one of our goals has been to elucidate the evolutionary routes that have led from the plesiomorphic structures of the ancestral dermal skeleton to the current diversity of these tissues. A series of comparative studies was devoted to the structure and development of the scaled dermal skeleton in living osteichthyan fish: ganoid scales of polypterids (Sire et al., ’87; Sire, ’89a) and of lepisosteids (Sire, ’94), scutes of armored catfish (Sire and Meunier, ’93; Sire, ’93), elasmoid scales of cichlids (Sire and Ge´raudie, ’83), and, recently, bony plates of syngnathids (Sire, personal observations). From a structural point of view, these studies showed that the elasmoid scales of teleosts could be derived from the ‘‘odontodal’’ tissues (i.e., odontocomplexes; see Ørvig, ’67, ’77) that covered the dermal skeleton in ancestral osteichthyan fish (Sire, ’89a, ’90). This hypothesis was supported by the existence of epidermal products in the upper layer of the scales in cichlids (Sire, ’88) and is consistent with the immunolocalization of mammalian enamel-like antigenic determinants in the skin covering the scales of carp (Krejsa et al., ’84) and of mammalian amelogenins in the ganoine, the ‘‘enamel’’ layer covering the scales of a polypterid (Zylberberg et al., submitted). From a developmental point of view, the comparison of the events (cytodifferentiation and morphogenesis) leading to the initiation and formation of an elasmoid scale (Sire and Ge´raudie, ’83; Sire et al., ’90) with those occurring during tooth and dermal bone formation (Huysseune and Sire, ’92; Sire and Huysseune, ’93) support the hypothesis that scales in teleosts are closer to teeth than to dermal bone. In spite of these results, which strongly suggest the existence of ectodermally derived products in the upper layer of the elasmoid teleost scale, there is still no information on the type of proteins (enamelins, amelogenins), and their presence has not been clearly demonstrated. Before undertaking long-term studies combining immunohistochemistry and molecular biology to detect epidermal products (i.e., enamel-like proteins) and genes coding for them, we first needed to find a fish species with scales that provide favorable material for such investigations. Until now our knowledge on elasmoid scale biology largely has been based on the cichlid fish, Hemichromis bimaculatus (Sire, ’87). However, because of increasing

interest in zebrafish (Danio rerio), a cyprinid, and because a large quantity of molecular tools is now available for this species, we decided to determine whether the zebrafish scale would be suitable to test our hypothesis on the ‘‘odontodal’’ origin of the elasmoid scales. Using scanning and transmission electron microscopy, the present study describes in detail the fine structure of the adult zebrafish scale, with particular attention to its upper region and its relationships to the epidermal cover. The main features of the scale structure in zebrafish were previously reported by Waterman (’70), but he did not describe the epidermis/scale interface. We also used experimentally regenerated scales because scale regeneration, which largely repeats ontogeny, has proven to be a useful tool for investigating similar questions in primitive osteichthyans (Sire et al., ’87; Sire, ’94). Our data strongly support a participation of the epidermis in the deposition of the upper layer of the scale in zebrafish. MATERIALS AND METHODS

Animals Ten adult zebrafish, Danio rerio (30 to 40 mm SL), were bred in 40-liter tanks, in controlled conditions of light (12h/12h) and temperature 25°C, and fed daily on Tetramin powder and chironomid larvae. Two specimens were killed by an overdose of MS 222 and used for alizarin red staining. Three fish were anaesthesized in 0.05% MS 222 solution and several scales were removed from the pectoral region of the left flank (Fig. 1) to be prepared for scanning electron microscopical (SEM) observations. These fish were allowed to regenerate their scales for 4 or 7 days. They were then over-anaesthesized and blocks of skin containing regenerated scales were dissected and fixed for transmission electron microscopical (TEM) study. Five other fish were killed and blocks of skin containing scales were dissected and fixed for light and TEM studies. Alizarin red staining The fish were immersed in 10% formaldehyde for a night, then stored in 70% ethanol for 2 days. They were cleared in 1% KOH for 1 day and depigmented in 0.02% H2O2 for 2 h, then placed in a solution of 1% KOH containing 0.5% Alizarin red S (Fluka) for 2 h. After staining, the fish were cleared in a mixture of 1% KOH and glycerol (v/v) for 3

FINE STRUCTURE OF ZEBRAFISH SCALE

163

Fig. 1. Squamation in zebrafish, Danio rerio (drawing from a 32 mm SL alizarin red stained and cleared specimen).

days, then stored in pure glycerol and observed with a binocular microscope. SEM To remove the soft tissues at their surfaces, isolated scales were immersed in 8% sodium hypochlorite for several minutes while monitored with a binocular microscope. The scales were then rinsed, dehydrated, dried, glued on an aluminum support, and covered with a thin layer of gold/ palladium prior to observation with a JEOL 35 SEM. TEM Blocks of skin or isolated scales were immersed in a fixative solution containing 1.5% glutaraldehyde and 1.5% paraformaldehyde in 0.1 M cacodylate buffer (pH 7.4) for 2 h at room temperature. After a short rinse in the buffer containing 10% sucrose, the samples were postfixed for 2 h in 1% OsO4 in 0.1 M cacodylate buffer to which 8% sucrose was added. The samples were subsequently rinsed in the buffer, dehydrated in a graded series of ethanol, and embedded in Epon 812. Some samples were decalcified for at least 7 days in the fixative solution to which 0.1 M EDTA was added (changed every 2 days), then processed as described above. Histology was studied from 1-µm-thick, toluidine blue-stained sections. For ultrastructural descriptions, thin sections were contrasted with uranyl acetate and lead citrate and observed in a 201 Philips EM operating at 80 kV. RESULTS

In zebrafish, scales are thin collagenous plates (,100 µm thick) that are large compared to the size of the fish. They are imbricated, i.e., their anterior region is covered by the preceding scale from the same longitudinal row and the lateral regions are covered

by two scales of the neighboring rows (Fig. 1). The squamation forms a regular pattern on the body. On both sides of the caudal peduncle, four scale rows are present to which two rows are added, dorsally and ventrally, forming a roof and a keel (Fig. 1). Seven scale rows cover both sides of the anterior region to the level of the pelvic fins. Two other rows cover the belly, from the anterior region of the anal fin to the region below the pectoral fin where several other rows of smaller scales occur. The fins are not covered by scales except for the basal part of the caudal fin. Details on the development of the squamation pattern in zebrafish are given elsewhere (Sire et al., in press). Scale surface ornamentation (SEM) In the pectoral region of a 37 mm SL adult zebrafish, the scales are roughly rectangular (1.7 mm wide, 2.1 mm high in Fig. 2) with a convex posterior region. The scale surface is separated into two regions: a small anterior region, which is overlapped by the preceding and neighboring scales, and a large posterior region, which overlaps the scale behind and is covered by the epidermis. These regions are characterized by different ornamentations: thin ridges in the anterior region, and grooves and undulations (here called ripples) in the posterior region (Fig. 2). The ridges, also called circuli, are not numerous (,20 circuli on the scale in Fig. 2). They are thin elevations (1.5 µm) that are disposed around the center of the scale, the focus, and are parallel to one another and to the scale margin (Fig. 3). They are tightly but regularly spaced on the most anterior region of the scale and more widely spaced in the lateral parts. The crests of the circuli are ornamented by numerous, small serrations, ,0.5 µm wide (Fig. 4). In the scales of adult specimens described here, the focus is located closer to the anterior rim than to the

164

J.-Y. SIRE ET AL.

posterior one. This is not the case in young specimens (personal observations). The focus has an irregular surface from which issue radial grooves (Figs. 2, 3), also called radii, which either radiate directly from the focus or, in the lateral regions, at some distance from it. They cross the posterior region of the scale to reach its margin perpendicularly. Short ‘‘extra’’ radii are also present close to the posterior margin, dividing the space between two radii (Fig. 2). The radii are thin (10 µm wide), roughly straight grooves (Figs. 2, 3, 5) in which the superficial layer of the scale is lacking. Each groove is bordered on both sides by a narrow (5 µm wide), convex crest with an irregular, rough surface (Fig. 6). The undulations, or ripples, are only visible in the posterior region of the scale (Fig. 2). They are small, regularly disposed, parallel one another and to the scale margin. Their surface seems smooth at low magnification (Fig. 7), but at a higher magnification appears covered by numerous small granules of variable diameter (100–500 nm) (Fig. 8). Scale location in the skin The zebrafish scales are located in the outer part of the dermis, and roughly parallel the skin surface (Figs. 9–11). They are inserted into spaces in the dermis (scale pockets) completely bounded by organized mesenchymal tissues except posteriorly, on the outermost surface, where the basal layer of the epidermis contacts the scale directly. Hence, a great percentage of the upper scale surface is covered by the epidermis, which moreover forms a fold around the free posterior margin of the scale (Fig. 9). On the upper surface, the epidermis is several cell layers thick with numerous specialized cells, mainly mucous and club cells. In contrast, the epidermal fold located at the deep scale surface is only two cell layers thick and devoid of specialized cells (Figs. 9, 10). The dense dermis (5stratum compactum) forms a relatively thin layer (50–100 µm thick) in the deep region of the skin facing the lateral musculature and below the scales. Sheets of loose dermis (5stratum laxum) emanate from the dense dermis and separate the scale pockets from one another. Posteriorly, these dermal sheets are connected to the epidermal folds. A thin layer of loose connective tissue is also present between the deep surface of the scale and the epidermal fold in the posterior region (Fig. 10). The stratum laxum is composed of a loose network of

collagen fibrils in which elongated fibroblasts, pigment cells, and capillary blood vessels are located. Only the very anterior region of the scale is surrounded by dense dermis. The bottom of each scale pocket is covered by a thin cellular sheet, called the scale pocket lining (Whitear et al., ’80), which separates the pocket from the dermis below (Fig. 10). This ‘‘epithelium’’ is composed of a single layer of flat fibroblast-like cells linked by desmosomes, and it is covered on both sides by a thin basement membrane, as previously described in other teleost fish (Whitear et al., ’80; Sire, ’89b). The scale pocket lining is lacking in the posterior region facing the epidermal fold (Fig. 10). The deep surface of the scale is covered by flat scaleforming cells, which form a thin ‘‘epithelium’’ (5hyposquama of Waterman, ’70). Pigment cells are frequently seen in the space between the hyposquama and the scale pocket lining cells (Fig. 10). In the anterior region, isolated scale-forming cells (5episquama of Waterman, ’70) cover the scale surface where it is at a distance from the epidermis or covered by the neighboring scales. The episquamal cells are not seen at

Figs. 2–8. Scanning electron micrographs of Danio rerio scale surface after sodium hypochlorite treatment. Scales from the pectoral region of a 37 mm SL specimen, in the region indicated by an asterisk in Figure 1. Anterior margin is to the left. Fig. 2. The surface of the zebrafish scale is separated into two regions by the anterior limit of the epidermal cover (dotted line). The anterior region is characterized by circular ridges, the circuli (c). The posterior region is ornamentated with radial grooves, the radii (r) and semi-circular undulations, or ripples (s). f, focus. Fig. 3. Anterior region showing the circuli (c) and the focus (f), from which some radii originate (r). Fig. 4. The surface of a circulus (c) is irregular and shows small serrations (arrows). Fig. 5. In the posterior region, several radii (r) interrupt the semi-circular ripple(s). Fig. 6. A radius is composed of a central groove surrounded by two convex crests showing an irregular surface. Fig. 7. The ripples are slight undulations of the scale surface without any prominent ornamentation. Fig. 8. High magnification of the scale surface in the posterior region showing numerous small, rounded granules. Bars 5 200 µm (345; Fig. 1), 100 µm (3120, Fig. 5), 50 µm (3250, Fig. 3), 10 µm (31,100, Fig. 7), 2 µm (35,000, Fig. 4, 34,000, Fig. 6), 1 µm (38,500, Fig. 8).

FINE STRUCTURE OF ZEBRAFISH SCALE

Figures 2–8

165

166

J.-Y. SIRE ET AL.

Fig. 9. Schematical drawing from 1-µm-thick, serial, longitudinal sections of the skin in adult Danio rerio (female, 37 mm SL) showing the relationships of the scales with the surrounding tissues. d, dermis; e, epidermis; l, lipid; m, muscle; sp, scale pocket.

the scale surface where it is covered by the epidermis (Fig. 10). Scale structure Descriptions of the scale structure proper can be found in Waterman (’70), but the general organization is briefly recalled here for better understanding. Both in transverse and longitudinal sections, the zebrafish scale appears to be clearly constituted of two regions: a superficial, thin, well-mineralized, woven-fibered layer, called the external layer, and a deep, thick, partially mineralized, lamellar layer, called isopedine (Figs. 10, 11). In the latter the collagen fibrils are regularly arranged into layers forming a plywood-like structure in which the direction of the fibrils changes from one layer to another by ,90°. The mineralization progresses only a short distance from the external layer downward and a large part of the isopedine is unmineralized, even in adults. Cells are never observed in the scale tissues. The scale organization is interrupted at regular intervals by radial grooves. Here, the well-mineralized external layer is replaced by an unmineralized, loose, collagenous tissue, whereas the isopedine below is entirely unmineralized and its collagenous matrix devoid of background substance (Figs. 10, 11). The surface of the radii is covered by a roughly convex dermal space containing fibroblast-like cells and blood vessels embedded within a loose matrix. These grooves provide the only places where dermal components are located at the scale surface in areas where it is covered by the

epidermis. Indeed, elsewhere on its whole posterior region, the scale surface is only separated from the epidermal basal layer cells by a 100-nm-thick layer that looks like a basement membrane (Figs. 12, 13), which is fairly ‘‘typical’’ on Figure 17. When decalcified, the woven-fibered external layer of a mature scale is seen to be covered by a thin layer (1 µm thick in average) that lacks collagen fibrils but is rich in an electron-dense, fine, granular material (Fig. 13). This layer is similar to the limiting layer previously described in cichlid scales (Scho¨nbo¨rner et al., ’79; Sire, ’85), and we have chosen to conserve this name. Some granules are aligned within this limiting layer, and they form dense lines parallel to the scale surface (Fig. 13). The surface of this layer is in direct contact with the basement membrane-like structure immediately below the epidermal basal layer cells. The mineral crystals located in the limiting layer of the scale are oriented perpendicularly to its surface, and this contrasts with the random disposition of the mineral crystals within the woven-fibered external layer below (Fig. 12; see also Fig. 17). The crystals appear to be attached directly to the deep surface of the basement membrane. In 4-day regenerated scales the epidermis is already in contact with the upper surface in several regions (Fig. 14). Mineralization has started in this region of the scale but only a few crystals are deposited (Fig. 15). This loose organization of the crystals facilitates sectioning and allows better observation of the features at the scale surface than

FINE STRUCTURE OF ZEBRAFISH SCALE

167

Fig. 10. Detail of the framed region in Figure 9, schematically drawn from thin sections. The rectangle is detailed on Figures 12 and 13. e, epidermis; ef, epidermal fold; ld, loose dermis; pc, pigment cell; r, radius; sc, scale; sfc, scale-forming cell; sp, scale pocket; spl, scale pocket lining.

in mature, well-mineralized scales. Patches of thin granular matrix, organized into small spherules or ovoid structures, are observed at the scale surface, in the region immediately below the epidermal basal layer (Fig. 15). These patches contain radially oriented mineral crystals forming ‘‘urchin-like’’ structures that differ from the randomly disposed crystals in the collagen matrix of the external layer of the scale (Fig. 16). This repre-

sents the anlage of the limiting layer. Facing the scale surface, the epidermal basal layer cells are cuboidal and juxtaposed. Their cytoplasm is rich in organelles such as RER cisternae, mitochondria, Golgi saccules, and small vesicles, some of them fusing with the cell membrane, which has irregular contours. There is no lamina densa of the basement membrane interposed between the epidermal basal cells and the scale surface. In

168

J.-Y. SIRE ET AL.

Fig. 11. One-µm-thick section of the skin from a 35 mm SL Danio rerio (undecalcified sample). Scales are parallel to the skin surface and their well-mineralized

superficial layer is closely apposed to the epidermal (e) basal layer cells (arrows). Radial grooves are visible (arrowheads). Bar 5 50 µm (3300).

Figs. 12, 13. Danio rerio. Transmission electron micrographs of the region indicated by a rectangle in Figure 10. The asterisks in these figures indicate the same locations. Fig. 12. Undecalcified sample. The mineral crystals in the limiting layer at the surface of the superficial region (*) are densely packed. This organization contrasts with the randomly disposed mineral crystals in the external layer below. e, epidermis. Fig. 13. EDTA decalcified sample. The limiting layer of the scale (*) is rich in fine granules that in some areas are linearly arranged, forming dense layers (arrowheads). This matrix contrasts with that of the external layer(s) below. Moreover the scale surface is in direct contact with the epidermal (e) basal layer cells (arrows). Bars 5 250 nm (340,000, Fig. 12; 360,000, Fig. 13). Fig. 14. Danio rerio. One-µm-thick section of the skin containing a 4-day-regenerated scale. It is already well formed and in some areas, the epidermis (e) is in contact with their surface (arrows). d, dermis. Bar 5 50 µm (3300). Figs. 15–17. Danio rerio. Transmission electron micrographs of 4 day- (Figs. 15, 16) and 7-day- (Fig. 17)

regenerated scales in the region covered by the epidermis (e). Undecalcified samples. Fig. 15. Cuboidal, juxtaposed epidermal basal layer cells are directly facing (arrows) the superficial layer(s) of the scale that shows rounded patches of granular matrix. i, isopedine. Fig. 16. Detail of the epidermis/scale limit as in Figure 15. Patches of thin granular matrix (arrows) are located immediately adjacent to invaginations (arrowheads) of the plasmalemma of the epidermal cells. These patches are invaded by mineral crystals that are radially arranged. There is no basement membrane between the epidermal cells and the scale matrix. Fig. 17. Detail of the upper region of a well-formed scale during the initial phase of matrix mineralization. A lamellar basement membrane-like structure (arrowheads) constitutes the only interface between scale and epidermis. The arrow points to an urchin-like structure located close to the epidermal basal layer cells. Elsewhere at the scale surface, the crystals are perpendicular to the deep surface of the basement membrane. Bars 5 1 µm (39,000, Fig. 15); 250 nm (360,000, Figs. 16, 17).

170

J.-Y. SIRE ET AL.

7-day regenerated scales the limiting layer is thicker and more mineralized than previously. The urchin-like structures are rare and the mineral crystals are now perpendicularly oriented towards the scale surface. The plasmalemma of the epidermal basal layer cells lining the scale matrix is more regular than in the previous stage and the anlage of the basement membrane-like structure is appearing (Fig. 17). DISCUSSION

Our SEM study of the scale surface and the detailed description of the upper layer of the scale using TEM completes and adds information to the previous description of the scale structure in zebrafish (Waterman, ’70). Moreover, the present work brings new data on the relationships between the scale surface and the epidermis in teleost fish. Indeed the observations presented above strongly suggest that the epidermal basal layer cells produce substances that are directly incorporated into the superficial region of the scale. Scale structure and location In zebrafish, a cyprinid, the scales are relatively large compared to the size of the fish, they are superficially located in the skin, they are easy to manipulate, and they regenerate as rapidly as do the scales of cichlid fish (Sire and Ge´raudie, ’84). Moreover they obviously belong to the elasmoid type as defined by Bertin (’44), and their structure is similar to that found in all elasmoid scales described until now in teleosts (review in Meunier, ’83; Whitear, ’86; Sire, ’87). These findings, and the recent description of developmental stages (Sire et al., in press), demonstrate that zebrafish scales are favorable material for further investigations dealing with scale biology in general, and scale development in particular. In zebrafish scales, the surface of the posterior region is relatively smooth except along the radial grooves. These grooves are thought to have two important roles, i.e., mechanical and trophic. First, the absence of mineralization at their level allows the large posterior region of the scale to be more flexible; second, nutrients can be brought to the epidermis through the dermal components which are located at their surface. Elsewhere, the epidermis is in direct contact with the scale surface, from which it is only

separated by a basement membrane. This surface lacks the typical ornamentations that have been described in this region for other elasmoid scales (e.g., Hughes, ’81; Sire, ’86; Lippitsch, ’92). The ornamentation of the scale surface seems to be related to the distance from the epidermis. The surface is smooth in zebrafish where the epidermis is in close contact with the scale, and it is ornamented where the epidermis is at a short distance as, e.g., in cichlids (Sire, ’86). Such a relationship is also illustrated by the scales in some Clupeiformes. They have a smooth surface in the region covered by the epidermis, and this corresponds to a tight covering of the scale by epidermis (Sire, personal observations). In the ornamented scales of cichlids, tubercles develop around anchoring fibers that attach the epidermis to the scale surface (Sire, ’85, ’86). Such anchoring fibers are not present in zebrafish, and one can ask how the epidermis maintains its relation to the scale surface. A similar question has been raised in primitive osteichthyans, the polypterids and the lepisosteids, where the epidermis also is in close contact with the surface of ganoid scales (Sire et al., ’87; Sire, ’94). Here, a characteristic basement membrane-like structure separates the well-mineralized scale surface, ganoine (an enamel), from the epidermal basal layer (Zylberberg et al., ’85). This unmineralized layer has been called the ganoine membrane (Sire, ’94), and it probably contains adhesive substances that allow the epidermis to remain in contact with the scale surface during swimming activity (Zylberberg et al., ’85; Sire, ’94). The thin basement membrane-like structure described in the present study, at the epidermis-scale interface in zebrafish, could have a similar function. From a comparison of the data available on the fine structure of the elasmoid scales in teleosts (e.g., Brown and Wellings, ’69; Yamada, ’71; Kobayashi et al., ’72; Lanzing and Wright, ’76; Scho¨nbo¨rner et al., ’79; Zylberberg and Meunier, ’81; Sire and Ge´raudie, ’83; Zylberberg et al., ’84), it appears that the extent of the so-called limiting layer at the external layer surface is directly related to the epidermal covering, i.e., the limiting layer exists only on the scale surface, that is directly covered by the epidermis or lies at a small distance from it (see also Sire, ’85). Elsewhere, i.e., in the anterior part of the scale not covered by the epidermis, the

FINE STRUCTURE OF ZEBRAFISH SCALE

external layer proper is only present. The degree of development of the epidermal covering also depends on the position of the scale in the dermis and on the degree of imbrication of the scales, both being related to the ecological adaptation of the species, mainly with regard to a mechanical protection against the substratum (Burdak, ’79). In this way, thick scales, a large extent of mineralization and a thick limiting layer, are interpreted as a reinforcement of the protection of the skin (Sire, ’85, ’86). It would appear that zebrafish scales, which are superficially located in the dermis, are thin, incompletely mineralized, and have a poorly developed superficial layer, probably constitute a poor mechanical protection. A general tendency toward reduction of the dermal skeleton is observed during the evolution of several fish groups, and it is largely admitted that such a reduction permitted the exploitation of a wide range of ecological niches (e.g., Schultze, ’77; Meunier, ’83). Zebrafish scales could represent a type of reduction of the scale cover that results in superficialization in a small-size fish, in contrast to another type of reduction that results in deep embedding in the dermis (as in Anguilla anguilla, e.g.) (Zylberberg et al., ’84). Scale-epidermis relationships The present ultrastructural study of the relationships of the epidermis to the scale surface in zebrafish scales shows that (1) in adult specimens, the epidermis is in direct contact with the whole outer surface of the posterior region of the scale, from which it is separated only by a thin basement-membrane like structure, (2) there are no dermal cells in between except opposite the radial grooves, (3) the limiting layer is devoid of collagen fibrils, but it is rich in fine granules arranged into electron-dense lines suggesting a periodical deposit, (4) in this layer, the mineral crystals are oriented perpendicularly to the scale surface, and (5) in regenerating scales the epidermal basal layer cells are directly in contact with the scale surface within 4 days after scale removal, and they show evidence of protein synthesis in the vicinity of the scale surface. The latter has thickened after 7 days of regeneration. These morphological observations strongly suggest the existence of epidermal products that are deposited at the surface of the external layer of the scales in zebrafish, thus constituting a thin limiting layer.

171

Epidermis/scale surface relationships In adult zebrafish, the epidermal basal layer cells are always seen in contact with the posterior region of the scale surface, but they do not seem to be synthetically active, judged by a cytoplasm that is poor in organelles and rich in microfilaments (probably cytokeratins). An intimate epidermal covering has been reported for this species by Waterman (’70), but without further description or comment. Experimentally regenerated scales have been found to be a useful tool, because at a given time they repeat the events that occur during ontogeny, albeit on a larger scale (Sire and Ge´raudie, ’84; Sire et al., ’87; Sire, ’94). This technique allowed us to show that the epidermal basal layer cells rapidly contact the scale surface and that they are involved in the deposition of some substance on it. This conclusion is based on the fact that (1) they have a cytoplasmic content rich in organelles known to be involved in protein synthesis, (2) they immediately contact the scale surface with an irregular contour of the cell membrane, after 4 days, and (3) the limiting layer is seen to be thicker after 7 days without any major change in the epidermal cells. In adult specimens, the ‘‘inactive’’ aspect of the epidermal cells is undoubtedly related to a slowing down, or an arrest, in the deposition of epidermal products at the scale surface, because such substances probably are only deposited periodically (see below). Recent observations of the scale surface in some Clupeiformes have also revealed a direct covering by the epidermis (Sire, personal observations), but in the other elasmoid scales described until now (see reviews in Whitear, ’86; Sire, ’87), the outer layer of the posterior region was always found to be separated from the epidermis by a narrow dermal space. However, in the cichlid Hemichromis bimaculatus there are indications of possible epidermal basal layer cell participation in scale material deposit (Sire, ’88). Moreover, the fine structure of the scale is known only for a few teleosts compared to the 23,000 or more species that exist (Nelson, ’94). Among these works little attention is paid to the relationship between the epidermis and the scale surface (Sire, ’85, ’88), because these studies are generally concerned with systematics (e.g., Hughes, ’81; Lippitsch, ’92, ’93). Until now, direct epidermal contact has only been described in the ganoid scales of

172

J.-Y. SIRE ET AL.

polypterids and lepisosteids. Consequently, this would be a plesiomorphic character for teleosts. The superficial layer of the ganoid scales, ganoine, has recently been demonstrated to be entirely formed by the inner epidermal cells, and consequently to be ‘‘true’’ enamel (Sire et al., ’87; Sire, ’94, ’95) as in tetrapod teeth. In zebrafish, the events occurring at the scale surface during regeneration are similar to those described during the deposit of the ganoine: active, roughly cuboidal cells are in direct contact with the scale surface, deposition of substances, then formation of a basement membrane-like layer at the interface, and a rest phase. The only difference we could find is related to the amount of material produced in the surfaces of the two scales: a thin (,1 µm in adult) limiting layer in zebrafish compared to a thick (,100 µm) ganoine layer in polypterids. Formation of the limiting layer In elasmoid scales, as first reported by Schmidt (’51) and Lerner (’53) using polarized light, the limiting layer covering the posterior region of the scale is characterized by the organization of the mineral crystals perpendicular to the scale surface. Using TEM, Scho¨nbo¨rner et al. (’79) described this superficial layer as poor in collagen fibrils and called it the external limiting layer. Detailed studies of this layer in cichlids have shown that it is thick, devoid of collagen fibrils, and it develops preferentially around the numerous anchoring fibers that form at the scale surface and reach the epidermaldermal boundary (Sire, ’85, ’86). Moreover, calcified spherules, containing substances that are probably epidermal in origin, have been seen to contribute to the periodical thickening of the limiting layer (Sire, ’88). In zebrafish, the limiting layer is thinner than in cichlid scales (Sire, ’85), but the other characteristics are similar: absence of collagen fibrils; matrix composed of thin granules, some being arranged in layers; and mineral crystals perpendicular to the scale surface. Recently, a similar matrix also has been described to constitute a thick superficial layer in the scutes of armoured catfish. This layer was called hyaloine (Sire, ’93). Moreover, the first elements of the matrix deposited in the limiting layer of the zebrafish scale, and the mineralization process, look similar to those described for gan-

oine in ganoid scales (Sire, ’95), namely, patches of matrix and mineral crystals with radial or urchin-like organization are deposited first. Then, these patches fuse with one another as matrix is added to form a layer in which the mineral crystals are oriented perpendicularly to the scale surface. Also, the layers of granules within the limiting layer strongly support the existence of a periodical deposition of the matrix, as is the case in the limiting layer of cichlid scales (Sire, ’85, ’88), in the hyaloine of armored catfish (Sire, ’93), and for the deposition of ganoine on the surface of the ganoid scales (Sire, ’94). Evolutionary implications All the comparative data presented above show that: (1) the limiting layer at the scale surface in zebrafish is similar in structure to the limiting layers in other elasmoid scales and to the hyaloine in callichthyid scutes, and (2) the matrix deposit and the mineral organization in the limiting layer of zebrafish scales look similar to ganoine deposit in ganoid scales. Thus, on the one hand, the limiting layer of the zebrafish scale can be considered as representative of this layer in elasmoid scales in teleosts, and on the other hand, it appears closer to the ganoine of the scales in primitive actinopterygians than previously suspected. Given: (1) that it is widely believed that elasmoid scales are derived from ganoid-like (rhombic) scales (Schultze, ’77; Reif, ’82; Smith and Hall, ’90), (2) that the elasmoid scales, at least in part, are probably derived from ‘‘odontodal’’ tissues that covered these rhombic scales (Sire, ’89a, ’90), (3) that in zebrafish scales, epidermal substances are deposited in the limiting layer, and (4) that this layer forms like ganoine, then the limiting layer of the elasmoid scales can be derived from ganoine covering the dermal skeleton of the more plesiomorphic actinopterygians, and it should contain enamel-like proteins. ACKNOWLEDGMENTS

We are indebted to Mary Whitear (Tavistok, UK) and Ann Huysseune (Gand, Belgium) for their constructive remarks and English corrections. We thank Olivier Babiar for his technical assistance in rearing zebrafish. TEM and the photographic work have been done at the CIME (Centre Interuniversitaire de Microscopie Electronique, Universite´s P6/P7, Paris).

FINE STRUCTURE OF ZEBRAFISH SCALE

LITERATURE CITED Bertin, L. (1944) Modifications propose´es dans la nomenclature des e´cailles et des nageoires. Bull. Soc. Zool. Fr. 69:198–202. Brown, G.A., and S.R. Wellings (1969) Collagen formation and calcification in teleost scales. Z. Zellforsch. 93:571–581. Burdak, V.D. (1979) Morphologie fonctionnelle du te´gument e´cailleux des poissons. La Pense´e Scientifique, Kiev (en russe), traduction franc¸aise, vol. spe´cial Cybium 10, 1986, 147 pp. Francillon-Vieillot, H., V. de Buffre´nil, J. Castanet, F.J. Meunier, J.Y. Sire, L. Zylberberg, and A. de Ricqle`s (1990) Microstructure and mineralization of vertebrate skeletal tissues. In J.G. Carter (ed.): Skeletal Biomineralization: Patterns, Processes and Evolutionary Trends. Short Course in Geology, vol. 5, Pt II. New York: Van Nostrand & Reinhold, pp. 175–234. Hughes, D.R. (1981) Development and organization of the posterior field of ctenoid scales in the Platycephalidae. Copeia 1981:596–606. Huysseune, A., and J.Y. Sire (1992) Bone and cartilage resorption in relation to tooth development in the anterior part of the mandible in cichlid fish: A light and TEM study. Anat. Rec. 234:1–14. Kobayashi, S., J. Yamada, K. Maekawa, and K. Ouchi (1972) Calcification and nucleation in fish-scales. Biomineralization 6:84–90. Krejsa, R.J., N. Samuel, C. Bessem, and H. Slavkin (1984) Immunogenetic and phylogenetic comparisons between teleost scale and dental enameloid with mammalian enamel antigens. In Tooth Morphogenesis and Differentiation, Colloques de l’INSERM, Vol. 125, pp. 369–376. Lanzing, W.J.R., and R.G. Wright (1976) The ultrastructure and calcification of the scales of Tilapia mossambica (Peters). Cell Tissue Res. 167:37–47. Lerner, H. (1953) Polarisationsoptische Beitrage zur Kenntnis der Verkalkung der Knochenfischschuppen. Z. Zell. Anat. Mikr. 39:36–73. Lippitsch, E. (1992) Squamation and scale character stability in cichlids, examined in Sarotherodon galilaeus (Linnaeus, 1758) (Perciformes, Cichlidae). J. Fish Biol. 41:355–362. Lippitsch, E. (1993) A phyletic study on lacustrine haplochromine fishes (Perciformes, Cichlidae) of East Africa, based on scale and squamation characters. J. Fish Biol. 42:903–946. Meunier, F.J. (1983) Les tissus osseux des Oste´ichthyens. Structure, gene`se, croissance et e´volution The`se de Doctorat e`s Sciences, Univ. Paris VII. Arch. Doc., Micro-e´d., Inst. Ethnol., Mus. Natl. Hist. Nat., Paris, SN 82 600 328, 200 pp. Nelson, J.S. (1994) Fishes of the World, 3rd edit. John Wiley & Sons, Inc. 600 p. Ørvig, T. (1967) Phylogeny of tooth tissues: Evolution of some calcified tissues in early vertebrates. In A.E.W. Miles (ed.): Structural and Chemical Organization of Teeth. New York: Academic Press, pp. 45–110. Ørvig, T. (1977) A survey of odontodes (‘‘dermal teeth’’) from developmental, structural, functional, and phyletic points of view. In S.M. Andrews, R.S. Miles, and A.D. Walker (eds): Problems in Vertebrate Evolution. Linnean Soc., Symp. 4 New York: Academic Press, pp. 52–75. Reif, W.E. (1982) Evolution of dermal skeleton and dentition in vertebrates. The odontode regulation theory. In M.K. Hecht, B. Wallace, and G.T. Prauce (eds.). Evolutionary Biology. New York: Plenum Press, pp. 287–368.

173

Scho¨nbo¨rner, A.A., G. Boivin, and C.A. Baud (1979) The mineralization process in teleost fish scales. Cell Tissue Res. 202:203–212. Schmidt, W.J. (1951) Polarisationsoptische Analyse der Schuppen des Knochenfisches Capros aper. Zeits. Wiss. Mikrosk. 60:1–15. Schultze, H.P. (1966) Morphologische und histologische Untersuchungen an Schuppen mesozoischer Actinopterygier. Ubergang von Ganoid-zu Rund-Schuppen. Neue Jahrb. Geol. Palaontol. 126:232–314. Schultze, H.P. (1977) Ausgansform und Entwicklung der rhombischen Schuppen der Osteichthyes (Pisces). Palae¨ontol. Zeitschrift 51:152–168. Sire, J.Y. (1985) Fibres d’ancrage et couche limitante externe a` la surface des e´cailles du Cichlidae Hemichromis bimaculatus (Te´le´oste´en, Perciforme): donne´es ultrastructurales. Ann. Sci. Nat. Zool., Paris 13: 163–180. Sire, J.Y. (1986) Ontogenic development of surface ornamentation in the scales of Hemichromis bimaculatus (Cichlidae). J. Fish Biol. 28:713–724. Sire, J.Y. (1987) Structure, formation et re´ge´ne´ration des e´cailles d’un poisson te´le´oste´en, Hemichromis bimaculatus (Perciforme, Cichlide´). The`se de Doctorat e`s-Sciences, Univ. Paris VII, Arch. Doc., Micro-e´d., Inst. Ethnol., Mus. Nat. Hist. Nat., Paris, SN 87 600 449, 261 pp. Sire, J.Y. (1988) Evidence that mineralized spherules are involved in the formation of the superficial layer of the elasmoid scale in the cichlids Hemichromis bimaculatus and Cichlasoma octofasciatum (Pisces, Teleostei): An epidermal active participation? Cell Tissue Res. 253:165–172. Sire, J.Y. (1989a) The scales in young Polypterus senegalus are elasmoid: new phylogenetic implications. Am. J. Anat. 186:315–323. Sire, J.Y. (1989b) The same cell lineage is involved in scale formation and regeneration in the teleost fish Hemichromis bimaculatus. Tissue Cell 21:447–462. Sire, J.Y. (1990) From ganoid to elasmoid scales in the actinopterygian fishes. Neth. J. Zool., 40:75–92. Sire, J.Y. (1993) Development and fine structure of the bony scutes in Corydoras arcuatus (Siluriformes, Callichthyidae). J. Morphol. 215:225–244. Sire, J.Y. (1994) Light and TEM study of nonregenerated and experimentally regenerated scales of Lepisosteus oculatus (Holostei) with particular attention to ganoine formation. Anat. Rec. 240:189–207. Sire, J.Y. (1995) Ganoine formation in the scales of primitive actinopterygian fishes, Lepisosteids and Polypterids. Conn. Tissue Res. 33:213–222. Sire, J.Y., and J. Ge´raudie (1983) Fine structure of the developing scale in the cichlid Hemichromis bimaculatus (Pisces, Teleostei, Perciformes). Acta Zool. 64:1–8. Sire, J.Y., and J. Ge´raudie (1984) Fine structure of the regenerating scale and their associated cells in the Cichlid Hemichromis bimaculatus (Gill). Cell Tissue Res. 237:537–547. Sire, J.Y., and A. Huysseune (1993) Fine structure of the developing frontal bone and scales of the cranial vault in the cichlid fish Hemichromis bimaculatus (Teleostei, Perciformes). Cell Tissue Res. 273:511–524. Sire, J.Y., and F.J. Meunier (1993) Ornementation superficielle et structure des plaques osseuses dermiques de quelques Siluriformes cuirasse´s (Loricariidae, Callichthyidae, Doradidae). Ann. Sci. Nat., Zool. 14:101– 123. Sire, J.Y., F. Allizard, O. Babiar, J. Bourguignon, and A. Quilhac (in press) Scale development in zebrafish (Danio rerio).

174

J.-Y. SIRE ET AL.

Sire, J.Y., H. Boulekbache, and C. Joly (1990) Epidermaldermal and fibronectin cell-interactions during fish scale regeneration: immunofluorescence and TEM studies. Biol. Cell 68:147–158. Sire, J.Y., J. Ge´raudie, F.J. Meunier, and L. Zylberberg (1987) On the origin of the ganoine: histological and ultrastructural data on the experimental regeneration of the scales of Calamoichthys calabaricus (Osteichthyes, Brachyopterygii, Polypteridae). Am. J. Anat. 180:391–402. Smith, M.M., and B.K. Hall (1990) Development and evolutionary origins of vertebrate skeletogenic and odontogenic tissues. Biol. Rev. 65:277–373. Waterman, R.E. (1970) Fine structure of scale development in the teleost, Brachydanio rerio. Anat. Rec. 168:361–380. Whitear, M. (1986) The skin of fishes including cyclostomes. Epidermis. Dermis. In J. Bereiter-Hahn, A.G. Matoltsy, and K.S. Richards (eds.): Biology of the Integument, 2, Vertebrates. Berlin: Springer Verlag, pp. 8–64. Whitear, M., A.K. Mittal, and E.B. Lane (1980) Endothelial layers in fish skin. J. Fish Biol. 17:43–65. Yamada, J. (1971) A fine structural aspect of the development of scales in the chum salmon fry. Bull. Jap. Soc. Scient. Fish. 37:18–29.

Zylberberg, L., and F.J. Meunier (1981) Evidence of denticles and attachment fibres in the superficial layer of scales in two fishes: Carassius auratus and Cyprinus carpio (Cyprinidae, Teleostei). J. Zool. Lond. 195:459–471. Zylberberg, L., J. Ge´raudie, J.Y. Sire, and F.J. Meunier (1985) Mise en e´vidence ultrastructurale d’une couche organique entre l’e´piderme et la ganoı¨ne du dermosquelette des Polypteridae. CR. Acad. Sci. Paris III, 10:517–522. Zylberberg, L., J. Ge´raudie, F.J. Meunier, and J.Y. Sire (1992) Biomineralization in the integumental skeleton of the living lower vertebrates. In B.K. Hall (ed.): Bone-Bone Metabolism and Mineralization, vol. 4. Boca Raton, Ann Arbor, Boston: CCR Press, pp. 171–224. Zylberberg, L., F.J. Meunier, F. Escaig, and S. Halpern (1984) Donne´es nouvelles sur la structure et la mine´ralisation des e´cailles d’Anguilla anguilla (Osteichthyes, Anguillidae). Can. J. Zool. 62:2482–2494. Zylberberg, L., J.Y. Sire, and A. Nanci (submitted) Immunodetection of amelogenin-like proteins in the ganoine of the scales of Calamoichthys calabaricus, a primitive actinopterygian fish, and evolutionary implications.