DEVELOPMENTAL DYNAMICS 235:1371–1389, 2006

RESEARCH ARTICLE

Expression of the dlx Gene Family During Formation of the Cranial Bones in the Zebrafish (Danio rerio): Differential Involvement in the Visceral Skeleton and Braincase L. Verreijdt,1 M. Debiais-Thibaud,2 V. Borday-Birraux,2 C. Van der heyden,1 J.-Y. Sire,3 and A. Huysseune1*

We have used dlx genes to test the hypothesis of a separate developmental program for dermal and cartilage bones within the neuro- and splanchnocranium by comparing expression patterns of all eight dlx genes during cranial bone formation in zebrafish from 1 day postfertilization (dPF) to 15 dPF. dlx genes are expressed in the visceral skeleton but not during the formation of dermal or cartilage bones of the braincase. The spatiotemporal expression pattern of all the members of the dlx gene family, support the view that dlx genes impart cellular identity to the different arches, required to make arch-specific dermal bones. Expression patterns seemingly associated with cartilage (perichondral) bones of the arches, in contrast, are probably related to ongoing differentiation of the underlying cartilage rather than with differentiation of perichondral bones themselves. Whether dlx genes originally functioned in the visceral skeleton only, and whether their involvement in the formation of neurocranial bones (as in mammals) is secondary, awaits clarification. Developmental Dynamics 235:1371–1389, 2006. © 2006 Wiley-Liss, Inc. Key words: zebrafish; Danio rerio; dermal bones; cartilage bones; dlx genes; braincase; visceral skeleton Accepted 1 February 2005

INTRODUCTION Two fundamentally different units build up the vertebrate skeleton: the endoskeleton (cartilage and cartilage bones, i.e., perichondral and endochondral bones, Daget, 1964; Patterson, 1977) and the dermal skeleton (dermal bones, also called membrane bones). These bones have a separate evolutionary origin (Smith and Hall, 1990; Donoghue

and Sansom, 2002) and cannot replace each other, not phylogenetically, not ontogenetically (Patterson, 1977). During vertebrate evolution, the relative participation of cartilage bones and dermal bones in building the skeleton has changed significantly. For instance, once forming a completely ossified braincase in ostracoderms, endochondral bones form only the skull floor in mam-

1

mals. This regression is compensated for by an expansion of the dermal bones of the cranial vault, which become the essential constituents of the braincase in mammals (Beaumont and Cassier, 1994). A dermal covering consisting of scales or other dermal components has been lost in most tetrapods (De Beer, 1937; Hanken and Hall, 1993), but elements of a postcranial dermal skeleton are

Ghent University, Biology Department, Ghent, Belgium ´ volution, CNRS, UPR 9034, Gif-sur-Yvette Cedex, France Populations, Ge´ne´tique et E 3 Syste´matique, Adaption et Evolution, UMR 7138, CNRS - Universite´ Paris VI - MNHM - IRD, Paris, France Grant sponsor: Ghent University (BOF, VEE); Grant number: 011V1298; Grant sponsor: CNRS; Grant number: UMR7138; Grant number: UPR9034; Grant sponsor: IWT Flanders; Grant number: 178W0300. *Correspondence to: Ann Huysseune, Ghent University, Biology Department, K.L. Ledeganckstraat 35, B-9000, Ghent, Belgium. E-mail: [email protected] 2

DOI 10.1002/dvdy.20734 Published online 13 March 2006 in Wiley InterScience (www.interscience.wiley.com).

© 2006 Wiley-Liss, Inc.

1372 VERREIJDT ET AL.

still present in some amphibians, reptiles, and even mammals (Huysseune and Sire, 1998). The vertebrate craniofacial skeleton is a complex assemblage of cartilages, cartilage bones, and dermal bones, grouped in two units with a distinctive evolutionary origin and function. The visceral skeleton (syn. viscerocranium, splanchnocranium) evolved first and strengthened the lining of the buccopharyngeal cavity. In contrast, the neurocranium arose as a box protecting the enlarged brain and three pairs of sensory organs (olfactory, optic, and otic organs). Both types of bone (cartilage bones and dermal bones) participate in the formation of the neurocranium as well as the viscerocranium. Many of the skeletal and connective tissue components of the head are neural crest-derived. The established view is that the major part of the skeletal tissues of the skull derives from cranial neural crest cells (Le Douarin, 1982; Hall, 1999; Chambers and McGonnell, 2002; Graham et al., 2004). The neural crest typically supplies the branchial arches of the splanchnocranium with mesenchyme from which cartilages and bones differentiate. In the neurocranium, the embryonic origin of the cartilages and bones (whether neural crest or mesoderm-derived) is less clear; some components, e.g., occipital bones, being derived from transformed vertebrae, are obviously mesodermal, whereas others, like the nasal capsules, are without doubt neural crest-derived, at least in jawed vertebrates (Donoghue and Sansom, 2002). Given the separate evolutionary origin of neurocranium versus splanchnocranium, and of cartilage bones versus dermal bones, one can hypothesize that these different categories of bones show differences in the molecular control underlying their differentiation process. The Dlx gene family is one of the multigene families encoding for homeodomain-containing transcription factors that are expressed early in cells of the cranial neural crest and later in craniofacial mesenchyme and skeletal tissues (Dolle´ et al., 1992; Bulfone et al., 1993; Akimenko et al., 1994; Robinson and Mahon, 1994; Simeone et al., 1994; Ellies et al., 1997; Qiu et al., 1997; Yang et al., 1998; Acampora et

al., 1999; Depew et al., 1999). Although Dlx genes are generally associated with ectodermal and neuralcrest derived tissues, they are also expressed in mesoderm (Myojin et al., 2001). Vertebrate Dlx genes share a highly conserved homeobox with the Drosophila distal-less gene (Merlo et al., 2000). To account for the present complement of six mammalian Dlx genes, i.e., Dlx1, Dlx2, Dlx3, Dlx4 (formerly Dlx7), Dlx5, Dlx6, a tandem duplication of an ancestral Dlx gene has been proposed to have been followed by two rounds of genome duplication and a subsequent loss of one Dlx pair (Ellies et al., 1997; Neidert et al., 2001). The zebrafish is known to possess eight dlx genes resulting from an additional genome duplication in the lineage leading to teleosts (Ekker et al., 1992; Akimenko et al., 1994; Stock et al., 1996; Ellies et al., 1997). Six of these dlx genes are found in three convergently transcribed pairs (dlx1adlx2a, dlx3b-dlx4b, dlx5a-dlx6a), as in their mammalian orthologs (the nomenclature used throughout this study follows that proposed by Panganiban and Rubenstein, 2002). These linked genes have extensively overlapping expression patterns in embryos (Akimenko et al., 1994; Ellies et al., 1997; Quint et al., 2000; Solomon and Fritz, 2002), suggesting that they share cis-acting sequences that control their spatiotemporal expression (Ellies et al., 1997). The remaining two (dlx2b, dlx4a) seem not to be linked to other dlx genes (Ellies et al., 1997). Vertebrate Dlx genes are generally expressed in regions of epithelial– mesenchymal interactions, which are typically involved in the formation of the craniofacial skeleton (Noden, 1988; Hanken and Thorogood, 1993; Le douarin et al., 1997). Several functional studies have revealed that Dlx genes are involved in patterning of the branchial arch skeleton (Qiu et al., 1995, 1997; Depew et al., 1999; Ferguson et al., 2000). Involvement of Dlx genes in bone and cartilage formation has been shown through functional studies (Qiu et al., 1997; Price et al., 1998b; Acampora et al., 1999; Kimmel et al., 2001; Ferrari and Kosher, 2002) and has also been inferred from expression data (Simeone et al., 1994;

Ferrari et al., 1995; Ellies et al., 1997; Price et al., 1998a). Expression studies have suggested, furthermore, a possible role of Dlx genes in the formation of integumental appendages requiring epithelial–mesenchymal interactions, such as teeth (Weiss et al., 1994; Bei and Maas, 1998; Price et al., 1998a; Zhao et al., 2000; Zhao and Dhoot, 2000; Jackman et al., 2004; Borday-Birraux et al., in press), hairs (van Steensel et al., 2000), and limbs (Ferrari et al., 1995; Crackower et al., 1996; Ferrari and Kosher, 2002). Expression of vertebrate dlx genes, mostly in relation to the developing skeleton and the nervous system, has been reviewed by Merlo et al. (2000) for chicken, mouse, and human and by Panganiban and Rubenstein (2002) for zebrafish, mouse, and human. The importance of Dlx genes in craniofacial development is shown by the many malformations observed in the case of defective Dlx genes. Mouse Dlx mutants exhibit severe craniofacial deformities, including cleft palate, and dysmorphic middle ear and jaw bones (Depew et al., 2002, 2005). For instance, Dlx5⫺/⫺ mice show a delayed ossification of dermal bones (Merlo et al., 2000), resembling the defects exhibited by mice in which one copy of Cbfa1, a key regulator of osteoblast differentiation, is inactivated (Otto et al., 1997). In Dlx5/6⫺/⫺ mutant mice, there is a homeotic transformation of the mandibular skeleton into a maxillary skeleton (Beverdam et al., 2002; Depew et al., 2005). In zebrafish embryos dlx genes are expressed during gastrulation, and later in the nervous system, the sensory organs and in the fins (Akimenko et al., 1994; Ellies et al., 1997). In the cranial skeleton, expression has only been reported in cells of the visceral arches and their primordia (Akimenko et al., 1994; Ellies et al., 1997). Recently, expression of dlx genes in teeth of larval zebrafish has been reported (Jackman et al., 2004; BordayBirraux et al., 2006). Because research of dlx expression in zebrafish has focused mainly on embryos up to 3 days postfertilization (dPF), and given the rather late development of cartilage bones and dermal bones in the head (from 3 to around 60 dPF; Cubbage and Mabee, 1996), little is known about the dlx expression patterns as-

dlx GENES IN CRANIAL BONES OF THE ZEBRAFISH 1373

sociated with cranial bone formation. Yet, because of their assumed importance in craniofacial development, at least in mammals, dlx genes are obvious candidates to explore potentially different expression patterns in cranial bones of different evolutionary origin. Here, we use dlx genes to test the hypothesis of a separate developmental program for cartilage bones and dermal bones within the neuro- and splanchnocranium by comparing the patterns of expression of all eight dlx genes during the formation of cranial bones in the zebrafish from 1 dPF to 15 dPF. Many of the cartilage bones in the teleost skull do not develop beyond a stage of forming a bone layer at the surface of a cartilaginous template (so-called perichondral bones; Blanc, 1953; Daget, 1964). Only rarely does osteogenesis progress to a stage in which bone invests the medullary cavity arisen after resorption of the cartilage (so-called endochondral bones, Blanc, 1953; Daget, 1964). In addition, endochondral ossification, when present, occurs late in development. In this study, perichondral bones, therefore, substituted for endochondral bones as members of the endoskeleton.

RESULTS Stages Distinguished During Osteogenesis A preliminary account of observations by light microscopy (LM) and transmission electron microscopy (TEM) on a developing dermal and endochondral bone has been given by Verreijdt et al. (2002). Briefly, at the LM level, the different stages can be identified as follows. In dermal bones, mesenchymal cells first accumulate to form a dense condensation. The cells become polarized around an extracellular region first devoid of any structured matrix (stage of condensation of mesenchymal cells). The first matrix is deposited within this extracellular space and recognizable on toluidine blue-stained semithin sections as a faint blue line or patch, depending on the shape the bone will take (stage of synthesis and secretion of extracellular matrix). The matrix is surrounded at least partially by polarized and

highly basophilic osteoblasts. Mineralization of the matrix is visible as a change of staining properties of the matrix, the center becoming deep blue, the osteoid (unmineralized bone) at the edge remaining faint blue. Cartilage bones first differentiate on the surface of a cartilage template; therefore, a stage of condensation of mesenchymal cells is absent. The first stage (differentiation of osteoblasts) can be identified when perichondral cells, similar to chondroblasts, flatten and become basophilic. The edge of the cartilage matrix changes its staining properties (from metachromatic to faint blue staining), reflecting the deposition of osteoid (stage of synthesis and secretion of extracellular matrix), and next takes on the deep blue staining of mineralized bone matrix (stage of mineralization). These observations have been repeatedly confirmed by TEM (data not shown).

not showing dlx expression, i.e., the symplectic, the epibranchial bones 1– 4, the subopercular and preopercular bone, are nevertheless commented upon in the description. Below, dlx expression patterns associated with each of the bones will be described in detail (see Tables 1, 2A,B). Some examples of dlx gene expression are presented in Figure 2. Every description is preceded by histological details on the formation of the bones, compared with the data provided by Cubbage and Mabee (1996), keeping in mind that the start of ossification has been determined by these authors through alizarin redstained and cleared specimens, revealing bones at the start of mineralization only. In contrast, our observations on 1- to 2-␮m-thick toluidine blue-stained sections allowed the earliest stages of bone formation to be identified, which is usually well before the first sign of mineralization.

Expression of dlx Genes in the Cranial Bones

Parasphenoid Bone

We found comparable expression patterns for the dlx genes on wholemount in situ hybridizations and on hybridized sections. Below, we will describe the results without further reference to the technique used. In the zebrafish, expression of dlx genes is found in association to certain bones: the parasphenoid in the roof of the buccal cavity, several bones associated with the visceral arches, and the opercular bones. Other bones, which develop during the period of examination, surprisingly show no expression of any of the dlx genes in their vicinity. In the neurocranium, bones lacking expression are the lateral ethmoid, the pterosphenoid, the sphenotic, the pterotic, the pro-otic, the basioccipital, the exoccipital and the supraoccipital bone, all of which are cartilage bones, and the post-temporal and the infraorbital bone 1, the latter two being canal bones. In the splanchnocranium, bones not showing dlx expression are the metapterygoid, the basihyal, the ventral hypohyal, the anterior ceratohyal (ceratohyal in Cubbage and Mabee, 1996), and the posterior ceratohyal bone (epihyal in Cubbage and Mabee, 1996), all of which are cartilage bones (Table 2C; Fig. 1). Other splanchnocranial bones

The parasphenoid is a dermal bone, which is first observed at 3 dPF (3.4 mm notochord length [NL]) as an elongate, dorsoventrally flattened sliver of bone matrix in the connective tissue of the interorbital septum (Cubbage and Mabee, 1996, and Verreijdt, personal observations). It is pointed anteriorly and has bifurcate spinous processes posteriorly at 4.5 mm NL. Only a few flattened, active osteoblasts surround the parasphenoid bone. At the site where the trabecula communis splits into the paired trabeculae cranii, a faint expression signal is visible ventromedial to the trabeculae, at 3, 4, and 5 dPF for dlx2b and dlx3b and at 3 dPF and 4 dPF for dlx5a (Tables 1, 2A). Expression of these dlx genes is located in the matrix-depositing osteoblasts surrounding the parasphenoid bone (Figs. 2, 3, 5, 6). The other dlx genes are not expressed in this region. At 3 and 4 dPF, dlx3b is also expressed in the epithelium underlying the parasphenoid bone (i.e., the epithelium of the roof of the buccal cavity). Before 3 dPF, we find no expression for any of the dlx genes, suggesting that dlx genes are not involved during early stages of parasphenoid bone development, i.e., condensation

Fig. 1. Semischematic representation of the cranial bones of a 30-day-old zebrafish, as observed on alizarin red-stained and cleared specimens. Note that, due to the advanced stage, more bones are shown than are present and examined within this study. Cartilages are not shown. A: Neurocranial bones, red; splanchnocranial bones, green, with opercular series dark green; bones of the pectoral girdle, gray. The urohyal, being an intratendinous ossification, is not color coded. Dermal bones are assigned to either neurocranium or splanchnocranium based on their evolutionary relationship to elements of these two units (Daget, 1964). Note that canal bones of the supra- and infraorbital series are also assigned a neurocranial identity. Several bones of the branchial series are not visible (covered by the opercular bones: basibranchial bones 1 to 3, ceratobranchial bones 1 to 4, and epibranchial bones 1 to 3). Nomenclature follows Cubbage and Mabee (1996), except for the anterior ceratohyal (ceratohyal in Cubbage and Mabee, 1996), posterior ceratohyal (epihyal in Cubbage and Mabee, 1996), and epiotic (epioccipital in Cubbage and Mabee, 1996). 1, Kinethmoid bone; 2, Supra ethmoid bone; 3, Lateral ethmoid bone; 4, Infraorbital bone 1; 5, Supraorbital bone; 6, Orbitosphenoid bone; 7, Frontal bone; 8, Pterosphenoid bone; 9, Sphenotic bone; 10, Parietal bone; 11, Pro-otic bone; 12, Pterotic bone; 13, Post-temporal bone; 14, Epiotic bone; 15, Supraoccipital bone; 16, Exoccipital bone; 17, Basioccipital bone; 18, Premaxillary bone; 19, Maxillary bone; 20, Palatine bone; 21, Vomerine bone; 22, Parasphenoid bone; 23, Entopterygoid bone; 24, Ectopterygoid bone; 25, Quadrate bone; 26, Metapterygoid bone; 27, Dentary bone; 28, Anguloarticular bone; 29, Coronomeckelian bone; 30, Retroarticular bone; 31, Basihyal bone; 32, Ventral Hypohyal bone; 33, Anterior Ceratohyal bone; 34, Posterior Ceratohyal bone; 35, Symplectic bone; 36, Hyomandibular bone; 37, Urohyal bone; 38, Ceratobranchial bone 5; 39, Branchiostegal ray 3; 40, Branchiostegal ray 2; 41, Branchiostegal ray 1; 42, Preopercular bone; 43, Interopercular bone; 44, Opercular bone; 45, Subopercular bone. B: Bones for which at least one dlx gene was found to be expressed in association with their early development. One or two dlx genes expressed, light blue; all dlx genes expressed, dark blue; intermediate number of dlx genes expressed, intermediate blue.

dlx GENES IN CRANIAL BONES OF THE ZEBRAFISH 1375

TABLE 1. Summary of Spatiotemporal Patterns of dlx Gene Expressiona dlx expression Bones Dermal bones Parasphenoid bone Dentary bone Anguloarticular bone Premaxillary bone Maxillary bone Entopterygoid bone Branchiostegal ray 1 Branchiostegal ray 2 Branchiostegal ray 3 Interopercular bone Opercular bone Cartilage bones Retroarticular bone Quadrate bone ° Hyomandibular bone Symplectic bone ° Ceratobranchial bones 1-4°, 5 Epibranchial bone 5 °

dlx1a

dlx2a

dlx2b

dlx3b

dlx4a

dlx4b

dlx5a

dlx6a

— 3-5dPF 3-5dPF — 6dPF — — 5-6dPF 4dPF — 3-5dPF

— 3-5dPF 3-4dPF — 4-6dPF — 15dPF 5-6dPF 3-4dPF 15dPF 3-6/15dPF

3-5dPF 4-5dPF — — 4-5dPF 4-5dPF — — 3dPF — 3-5dPF

3-5dPF 3-4dPF — — 4-5dPF — 9-15dPF 5-7dPF 3-4dPF 9-15dPF 3-15dPF

— 3-6dPF 3-7dPF 14dPF 4-7dPF — — 5-7dPF 3-4dPF — 3-7dPF

— 3-4dPF 3-5dPF — — — — — 3-4dPF — 3-5dPF

3-4dPF 3-4dPF 3-7dPF — 4dPF 3-4dPF — 5-7dPF 3-4dPF — 3-6/9-10dPF

— 4-5dPF 3-5dPF — 4-5dPF — — 5dPF 3-4dPF — 3-5dPF

3-4dPF 3-5dPF 4-5dPF 3-5dPF —

3-4dPF 3-5dPF 4-6dPF 3/5dPF —

3dPF 4dPF 4-5dPF 5dPF —

3-4dPF 4-5dPF 4-5dPF 3-4/6dPF —

3-7dPF 4-7dPF 4-5dPF 3-7dPF —

3-5dPF 3-5dPF 4-5dPF 4-5dPF 3-4dPF

3-7dPF 3-4dPF 4dPF 3-4/6-7/9-10dPF 3-4dPF

3-5dPF 4-5dPF 4dPF 4-5dPF —

—

—

—

3-6dPF

—

3-4dPF

—

—

a

Summary of the spatiotemporal patterns of expression of dlx genes in relation to cranial bone development in zebrafish between 1 and 15 days postfertilization (dPF), ordered by gene. 3-5 dPF, expression continuing from 3 dPF over 4 dPF until 5 dPF; 3/5 dPF, expression present from 3 to 5 dPF but interrupted at 4 dPF; °, expression present long before ossification of the bone.

of mesenchymal cells and differentiation of osteoprogenitor cells. The expression site at 3 dPF extends over an anteroposterior distance of approximately 110 ␮m. According to the reference series, the parasphenoid bone at 3 dPF is approximately 260 ␮m long, meaning that the bone extends more rostral and more caudal than the expression site does. This is also the case at 4 and 5 dPF. Transcripts of these dlx genes are no longer detected after 5 dPF.

Visceral Arches Ventral part of the mandibular arch. The ventral part of the mandibular arch consists of Meckel’s cartilage and all the cartilage and dermal bones associated with it. Dentary bone. The dentary bone is a dermal bone that first appears around the anterior end of Meckel’s cartilage at 10 dPF (5.1 mm NL; Cubbage and Mabee, 1996). Toluidine blue-stained sections reveal the presence of bone already at 4 dPF, ventro-

laterally to Meckel’s cartilage, as a partly parachondral (dermal) and partly perichondral ossification. The latter represents the continuation of the mentomeckelian bone (as described in a cichlid by Huysseune and Sire, 1992). Condensation of mesenchymal cells is first seen at 3 dPF, and by 4 dPF extracellular matrix has been deposited and the first signs of mineralization appear. The parachondral dentary bone is separated from Meckel’s cartilage by a single cell row at early stages of development. Later, two or three cell rows separate the cartilage from the bone. Toward the posterior part of Meckel’s cartilage, the anguloarticular bone is interposed between the dentary bone and Meckel’s cartilage. There is usually only one row of osteoblasts between the dentary and the ventral epidermis. Expression of dlx genes is seen in the condensed mesenchymal cells ventral and lateral to the rostral end of Meckel’s cartilage, starting at 3 dPF and continuing up to 6 dPF (around the mineralized bone; Table 1). Moreover, once osteoid is being deposited,

expression is situated medial to the dentary bone in the few cells between the cartilage and the bone, ventral to the bone near the ventral epidermis, and lateral to the bone (Figs. 4, 7, 8, 11). All eight dlx genes are expressed in these regions during these stages. However, only dlx4a is still expressed at 6 dPF. After 6 dPF, there is no longer dlx expression in connection to the dentary bone. All dlx genes of the three pairs show an almost identical expression pattern in the osteoblasts associated with the dentary bone. The genes dlx2b and dlx6a are not expressed during condensation of mesenchymal cells. dlx5a, the linkage member of dlx6a, is expressed, however, during early developmental stages as well as at the start of mineralization. In contrast to the other dlx genes, dlx3b is expressed in the ventral epidermis as well as in mesenchymal cells. Anguloarticular bone. The anguloarticular bone is a dermal bone that ossifies initially on the posterolateral surface of Meckel’s cartilage, near the

1376 VERREIJDT ET AL.

TABLE 2A. Expression Patterns of dlx Genes in Dermal Bonesa

Dermal bones Parasphenoid bone Dentary bone

Anguloarticular bone Premaxillary bone Maxillary bone

Entopterygoid bone Branchiostegal ray 1 Branchiostegal ray 2 Branchiostegal ray 3

Condensation of mesenchymal cells

Synthesis and secretion of extracellular matrix

Mineralization of matrix

1dPF — 3dPF dlx1a, 2a, 3b, 4a, 4b, 5a 3dPF dlx1a, 2a, 4a, 4b, 5a, 6a 9dPF — 3dPF —

2dPF — 4dPF dlx1a, 2a, 2b, 3b, 4a, 4b, 5a, 6a 4dPF dlx1a, 2a, 4a, 4b, 5a, 6a 9dPF — 3dPF —

3dPF dlx5a 9dPF dlx3b 5dPF dlx1a, 2a, 3b, 4a, 5a, 6a 2dPF —

3dPF dlx5a 9dPF dlx3b 5dPF dlx1a, 2a, 3b, 4a, 5a, 6a 2dPF —

10dPF dlx3b 2dPF —

11dPF dlx3b 2dPF —

3dPF dlx2b, 3b, 5a 4dPF dlx1a, 2a, 2b, 3b, 4a, 4b, 5a, 6a 5dPF dlx1a, 4a, 4b, 5a, 6a 10dPF dlx4a (14dPF) 4dPF dlx1a(at 6dPF), 2a, 2b, 3b, 4a, 5a, 6a 4dPF dlx2b, 5a 10dPF dlx2a(at 15dPF), 3b 6dPF dlx1a, 2a, 3b, 4a, 5a 3dPF dlx2a, 2b, 3b, 4a, 4b, 5a, 6a (and dlx1a at 4dPF) 12dPF dlx2a(at 15dPF), 3b 3dPF dlx1a, 2a, 2b, 3b, 4a, 4b, 5a, 6a

Interopercular bone Opercular bone

a

Time table of the stages distinguished during the development of the cranial bones of the visceral skeleton in the zebrafish, between 1 and 15 days postfertilization (dPF), with a summary of the expression patterns of dlx genes during each of these stages. A, dermal bones; B, perichondral bones. The different stages of bone formation are not considered as discrete events in time. When condensation of mesenchymal cells is reported to take place at 3 dPF, and synthesis and secretion of extracellular matrix at 4 dPF, then in fact, the two stages overlap each other somewhere between 3 and 4 dPF. C, timing in dPF of the different stages of bone formation of neurocranial and splanchnocranial bones, not expressing dlx genes, that develop during the period examined.

articulation with the quadrate, at 12 dPF (5.5 mm NL; Cubbage and Mabee, 1996). A posterior process forms thereby creating a synovial joint with the quadrate. Ossification spreads anteriorly over the lateral surface of Meckel’s cartilage. Reference series show that the anguloarticular bone starts to develop much earlier, at 3 dPF (as a small condensation of mesenchymal cells lateral and near to Meckel’s cartilage). At 5 dPF, the first signs of mineralization are present (a tiny piece of mineralized bone surrounded by a small number of osteoblasts). Expression patterns of dlx genes overlap each other in time (Tables 1, 2A). The expression signals are situated posterolaterally to Meckel’s cartilage near the articulation with the palatoquadrate cartilage (Fig. 9, 12). All pairs of dlx genes are expressed,

except for dlx3b, the pair member of dlx4b. Of the nonlinked genes, dlx4a is expressed, but dlx2b is not. The gene dlx2a is expressed during early development in a patch of cells lateral to the posterior end of Meckel’s cartilage (i.e., during condensation of mesenchymal cells, synthesis and secretion of extracellular matrix; thus, in preosteoblasts and osteoblasts) and dlx1a, its linked gene, shows the same expression pattern, but is additionally expressed at the start of mineralization. The pair members dlx5a and dlx6a differ in their pattern of expression: dlx5a is expressed, from 3 to 7 dPF, during early steps of bone formation (condensation of mesenchymal cells and synthesis and secretion of extracellular matrix) as well as at the start of and during mineralization, much like dlx4a; dlx6a expression on the other hand is restricted to early

steps of bone formation and to the start of mineralization, as for dlx1a and dlx4b. After 7 dPF, dlx genes are no longer expressed in association with the anguloarticular bone. Coronomeckelian bone. The coronomeckelium (sesamoid articular) is a dermal bone that ossifies at 14 dPF, where the tendon from the adductor mandibulae A3 inserts on Meckel’s cartilage (Cubbage and Mabee, 1996). This insertion occurs on the dorsal side of Meckel’s cartilage, just anterior to the articulation with the palatoquadrate cartilage. On toluidine blue-stained sections the coronomeckelian bone cannot be seen as separate from the far-larger anguloarticular bone during the stages examined here. In the region where osteoblasts are expected to differentiate for the

dlx GENES IN CRANIAL BONES OF THE ZEBRAFISH 1377

TABLE 2B. Expression Patterns of dlx Genes in Cartilage Bonesa

Cartilage bones

Non-ossified cartilage

Differentiation of osteoblasts

Retroarticular bone — Quadrate bone dlx1a, 2a, 2b, 3b, 4a, 4b, 5a, 6a Hyomandibular bone — Symplectic bone dlx1a, 2a, 2b, 3b, 4a, 4b, 5a, 6a Ceratobranchial bones 1-4 dlx4b, 5a Ceratobranchial bone 5 — Epibranchial bone 5 dlx3b, 4b a

Synthesis and secretion of extracellular matrix

Mineralization of matrix

3dPF dlx1a, 2a, 2b, 3b, 4a, 4b, 5a, 6a 10dPF —

4dPF dlx1a, 2a, 3b, 4a, 4b, 5a, 6a 11dPF —

5dPF dlx4a, 4b, 5a, 6a

4dPF dlx1a, 2a, 2b, 3b, 4a, 4b, 5a, 6a 11dPF —

5dPF dlx1a, 2a, 2b, 3b, 4a, 4b 11dPF —

6dPF dlx2a 12dPF —

11dPF — 2dPF — 12dPF —

12dPF — 2dPF — 13dPF —

13dPF — 3dPF dlx4b, 5a 14dPF —

12dPF —

For details, see the footnote to Table 2A.

TABLE 2C. Timing of Stages of Bone Formation for Bones not Expressed in dlx Genesa

Condensation of mesenchymal cells/differentiation of osteoblasts

Synthesis and secretion of extracellular matrix

Mineralization of matrix

12 12 11 11 9 9 6 14 14 14

13 13 12 12 10 10 6 15 15 15

14 14 13 13 11 11 7 16 16 16

14 10 9 6 9 11 15 13 4

15 11 10 6 10 11 15 14 4

16 12 11 7 11 12 16 14 5

Neurocranium Lateral ethmoid bones Pterosphenoid bone Sphenotic bone Pterotic bone Pro-otic bone Basioccipital bone Exoccipital bone Supraoccipital bone Post-temporal bone Infraorbital bone 1 Splanchnocranium Metapterygoid bone Basihyal bone Ventral hypohyal bone Ant. ceratohyal bone Post. ceratohyal bone Symplectic bone Epibranchial bones 1-4 Preopercular bone Subopercular bone a

For details, see the footnote to Table 2A.

coronomeckelian bone, expression of dlx genes was observed, which could not be distinguished however as separate from the signals connected to the formation of the anguloarticular bone.

Retroarticular bone. The retroarticular bone is a perichondral bone that forms at the posteroventral tip of Meckel’s cartilage at 10 dPF (5.1 mm NL), where the interoperculoman-

dibular ligament attaches (Cubbage and Mabee, 1996). In toluidine bluestained semithin sections, the retroarticular bone is already present at 5 dPF as a small piece of mineralized

1378 VERREIJDT ET AL.

Figs. 2–23.

dlx GENES IN CRANIAL BONES OF THE ZEBRAFISH 1379

bone ventral to Meckel’s cartilage, and very near to the ventral epidermis, with only a few osteoblasts surrounding the bone. At this stage, the bone extends over a length of approximately 25 ␮m. All dlx genes are expressed during the formation of the retroarticular bone in a pattern (Table 2B) that is quite similar to that observed for the dentary bone (compare with Table 2A). The linked genes dlx1a and dlx2a are coexpressed in the osteoblasts at 3 and 4 dPF, i.e., at the early stages of bone formation (differentiation of osteoblasts and synthesis and secretion of extracellular matrix) but are then down-regulated (Figs. 9, 12). The expression pattern of dlx3b and dlx4b is similar to the above-mentioned one, except that dlx3b is also expressed in

the ventral epidermis. The pattern of expression of the third pair (dlx5adlx6a) differs. Although dlx6a is expressed in an area and for a duration similar to the other pairs, dlx5a is expressed during an extended period until 7 dPF (much as is the case for dlx4a). dlx2b is expressed at 3 dPF only. No dlx expression is associated with the retroarticular bone beyond 7 dPF.

Dorsal part of the mandibular arch. The dorsal part of the mandibular arch consists of the palatoquadrate cartilage and associated cartilage and dermal bones. Premaxillary bone. The dermal premaxillary bone is first visible as a

thin sliver in the anterodorsal region of the upper lip, anterior to the dorsal tip of the maxillary bone at 12 dPF (5.5 mm NL; Cubbage and Mabee, 1996). The ascending process forms as a thin, posterodorsally directed sheet of bone. Reference series show a mesenchymal condensation at 9 dPF and the anlage of the mineralized premaxillary bone at 10 dPF. Only one dlx gene, dlx4a, is expressed around the already mineralized premaxillary bone, at 14 dPF (Tables 1, 2A). Maxillary bone. The dermal maxillary bone is first visible as a small dorsoventrally oriented rod of bone in the lateral region of the upper jaw. At 9 dPF (4.8 mm NL), the maxillary bone becomes S-shaped through ossi-

dlx gene expression during cranial bone development in zebrafish larvae in Figures 2–23. Transverse, 6-␮m-thick sections of whole-mount hybridized specimens (Figs. 2, 3, 4, 8, 9, 10, 14, 15, 16, 20, 21) with their respective 2-␮m-thick toluidine blue-stained reference section (5, 6, 7, 11, 12, 13, 17, 18, 19, 22, 23). Legends to all reference sections are identical to those of the hybridized sections. BC, buccal cavity; CB4, ceratobranchial cartilage 4; CB5, ceratobranchial cartilage 5; CH, ceratohyal cartilage; E, eye; HM, hyomandibular cartilage; MC, Meckel’s cartilage; O, opercular bone; PQ, palatoquadrate cartilage; TC, trabecula communis; TCR, trabecula cranii. Scale bars ⫽ 20 ␮m. Fig. 2. At 3 days postfertilization (dPF). Expression of dlx3b in the osteoblasts surrounding the parasphenoid bone (arrowhead). Fig. 3. At 3 dPF. Expression of dlx5a in the osteoblasts surrounding the parasphenoid bone (white arrowhead). Black arrowhead, expression in the condensation of mesenchmyal cells dorsal to the palatoquadrate cartilage, where the entopterygoid bone will form. Fig. 4. At 3 dPF. Expression of dlx5a in the mesenchymal condensations (white asterisk) surrounding Meckel’s cartilage. The dentary bone will form in the condensation lateral and ventral to the cartilage. Fig. 5. Toluidine blue-stained reference section for Figure 2, showing the mineralized parasphenoid bone (arrowhead). Fig. 6. Toluidine blue-stained reference section for Figure 3. The white arrowhead indicates the mineralized parasphenoid bone, and the black arrowhead shows the mesenchymal cell condensation which will give rise to the entopterygoid bone. Fig. 7. Toluidine blue-stained reference section for Figure 4, showing a mesenchymal cell condensation (asterisk) lateral to Meckel’s cartilage. Fig. 8. At 3 dPF. Expression of dlx3b in the mesenchyme lateral to Meckel’s cartilage (white arrowhead) and in the ventral epidermis (black arrowhead). Fig. 9. At 4 dPF. Expression of dlx1a in association with the anguloarticular (white arrowhead) and retroarticular bone (black arrowhead) during synthesis and secretion of extracellular matrix. Fig. 10. At 4 dPF. Expression of dlx2b in the osteoblasts surrounding the mineralized maxillary bone (arrowhead). Fig. 11. Toluidine blue-stained reference section for Figure 8, indicating the sites corresponding to dlx3b expression (arrowheads). Fig. 12. Toluidine blue-stained reference section for Figure 9, showing the anguloarticular (white arrowhead) and retroarticular bone (black arrowhead), both not yet mineralized. Fig. 13. Toluidine blue-stained reference section for Figure 10. The arrowhead indicates the mineralized maxillary bone. Fig. 14. At 3 dPF. Expression of dlx5a in the mesenchmyal cell condensation where the entopterygoid bone will form (white arrowhead). Fig. 15. At 4 dPF. Expression of dlx1a in the mesenchymal cells (arrowhead) before the formation of the quadrate bone. Fig. 16. At 7 dPF. Expression of dlx4a in osteoblasts located at the ventral tip of the opercular bone (white arrowhead). dlx4a is also expressed in the mesenchyme of the gills of the fourth and fifth branchial arch (black arrowheads). Fig. 17. Toluidine blue-stained reference section for Figure 14. The arrowhead shows a mesenchymal cell condensation dorsal to the palatoquadrate cartilage. Fig. 18. Toluidine blue-stained reference section for Figure 15. The arrowhead indicates the site where the quadrate bone will form. Fig. 19. Toluidine blue-stained reference section for Figure 16. The white arrowhead indicates mesenchymal cells ventral to the mineralized opercular bone, and black arrowheads indicate mesenchymal cells belonging to the fourth branchial arch, which is situated more medial compared with the section in Figure 16. Fig. 20. At 3 dPF. Expression of dlx5a in the mesenchymal cells (black arrowhead) ventral to the hyomandibular cartilage where the perichondral hyomandibular bone will form. Lateral to the hyomandibula, dlx5a expression is strong in the osteoblasts surrounding the opercular bone (white arrowhead). Strong expression is also observed in mesenchymal cells of the fourth branchial arch. Probably this specimen is slightly advanced, given the observation that this expression connected to the hyomandibular usually only starts at 4 dPF. Fig. 21. At 3 dPF. Expression of dlx3b in the branchiostegal membrane. Transcripts are located in the osteoblasts surrounding branchiostegal ray 3 (black arrowhead). This gene is also expressed in mesenchymal cells of the fourth and fifth branchial arch. Fig. 22. Toluidine blue-stained reference section for Figure 20. The white arrowhead indicates the mineralized opercular bone, at this stage still very small. The black arrowhead indicates a group of mesenchymal cells ventral to the hyomandibular cartilage. Fig. 23. Toluidine blue-stained reference section for Figure 21. The black arrowhead indicates the mineralized branchiostegal ray 3, and the white arrowhead indicates the mesenchymal cell condensation later giving rise to the second branchiostegal ray.

1380 VERREIJDT ET AL.

fication anteriorly, with its anteriormost tip directed medially (Cubbage and Mabee, 1996). Toluidine bluestained sections reveal the presence of a large patch of condensed mesenchymal cells at 3 dPF and a small mineralized anlage of the maxillary bone at 4 dPF. Two to three cell layers separate the bone from the basal lamina of the lateral epidermis and of the buccal epithelium, respectively. With the exception of dlx4b, all dlx genes are expressed in association with the formation of the maxillary bone. The two genes within a pair are not expressed simultaneously (Table 1). dlx1a is transiently expressed at 6 dPF, when the maxillary bone is well mineralized, whereas its linked gene, dlx2a, is expressed from the first signs of mineralization (4 dPF) until 6 dPF. dlx3b is expressed during the first 2 days of mineralization, whereas dlx4b is not expressed. dlx5a and dlx6a are both expressed during the first steps of mineralization, as is dlx2b (Figs. 10, 13). Expression of dlx4a lasts the longest, from 4 dPF until 7 dPF; after 7 dPF, none of the dlx genes are expressed. The expression signals are always located close to the bone surface but not in the more peripheral mesenchymal cells. For none of the dlx genes, expression is present in the condensation phase or during synthesis and secretion of extracellular matrix of the maxillary bone. Entopterygoid bone. The dermal entopterygoid bone develops as a long sliver of bone in the mesenchyme dorsomedial to the palatoquadrate cartilage at 4 dPF (Cubbage and Mabee, 1996). Our reference series showed condensation and secretion stages at 3 dPF and mineralized bone at 4 dPF. The bone follows the border of the buccal epithelium on its medial side and the curvature of the eye on its lateral side. Only a small number of cells surround the bone on both sides. Only two dlx genes are expressed in connection with the formation of this bone (Tables 1, 2A). dlx2b is expressed just before and at the start of mineralization, whereas dlx5a is also expressed during earlier steps of development (condensation of mesenchymal cells and synthesis and secretion of

extracellular matrix; Figs. 14, 17). Both genes are expressed lateral, medial, and dorsal to the bone. The signals are not restricted to the welldifferentiated central regions (in anteroposterior direction) of the bone, but also present in peripheral regions. Expression is down-regulated after 5 dPF. Quadrate bone. Ossification of the quadrate bone, a perichondral bone, begins as a posteriorly projecting membranous spine between the symplectic cartilage and the preopercular bone at 10 dPF (5.1 mm NL; Cubbage and Mabee, 1996). The body of the quadrate bone ossifies first at the ball and socket articulation with the anguloarticular bone at 16 dPF. Reference sections show that condensation of mesenchymal cells occurs around 10 dPF and that the quadrate bone is well mineralized by 12 dPF. The quadrate bone is situated very close to the ventral epidermis; it is separated from it by only a few cell layers. All dlx genes are expressed in patterns that overlap in time (Tables 1, 2B). The three pairs are all expressed ventral to the palatoquadrate cartilage before the development of the quadrate bone (before the differentiation of osteoblasts; Figs. 15, 18). dlx3b, apart from being expressed in the mesenchyme, is also expressed in the ventral epidermis at 4 dPF. The pattern of expression of the two other genes, dlx2b and dlx4a, differs from that shown by the linked genes. dlx2b is expressed at 4 dPF, dlx4a is expressed up to 7 dPF, both at the stage of nonossified cartilage. After 7 dPF, dlx genes are no longer expressed in this region; thus, expression of dlx genes is down-regulated well before the first signs of perichondral bone formation.

Hyoid Arch Several perichondral bones, associated with the hyoid arch (the basihyal, ventral hypohyal, anterior ceratohyal, and posterior ceratohyal bones) show no dlx expression during the period examined.

Hyomandibular bone. The hyomandibular bone begins ossifying along the hyosymplectic carti-

lage near the hyomandibular foramen at 8 dPF (4.6 mm NL; Cubbage and Mabee, 1996). Ossification spreads along the cartilage, and sheets of membrane bone develop from the cartilage anteriorly and posteriorly at 16 dPF (6.6 mm SL). According to our reference series, mesenchymal cells start to differentiate into osteoblasts around the hyosymplectic cartilage at 4 dPF. At 6 dPF, perichondral bone flanked by osteoblasts is clearly present lateral and ventral to the cartilage, near the hyomandibular foramen. The sheets of membrane bone start forming at 13 dPF. In each of the three dlx pairs, the genes share a similar expression pattern (Tables 1, 2B). Expression is restricted to cells surrounding the hyomandibular cartilage, during the stages of osteoblast differentiation and synthesis and secretion of extracellular matrix (Figs. 20, 22), except for dlx2a, which is still expressed during onset of mineralization. The genes dlx5a and dlx6a are no longer expressed after 4 dPF. dlx2b and dlx4a are also expressed in the early stages of bone development. Expression of dlx genes is down-regulated after 6 dPF. There is no dlx expression associated with the formation of the sheets of membrane bone that start forming from the perichondral bone.

Symplectic bone. The symplectic bone is a perichondral bone, which ossifies in the middle of the anterior portion of the hyosymplectic cartilage, medial to the posterior tip of the quadrate spine, at 13 dPF (5.8 mm SL). The ossification progresses along the cartilage anteriorly and posteriorly (Cubbage and Mabee, 1996). According to our reference series, the symplectic bone starts its mineralization at 12 dPF as a perichondral ring of bone surrounded by osteoblasts. All dlx genes are expressed in patterns overlapping in time in the region surrounding the symplectic cartilage at the place where the symplectic bone will form but well before ossification of this bone (Tables 1, 2B). The pair dlx1a-dlx2a is expressed from 3 to 5 dPF. The pair dlx3b-dlx4b is expressed in a comparable manner, although dlx3b is expressed until 6 dPF. The expression of the two genes from

dlx GENES IN CRANIAL BONES OF THE ZEBRAFISH 1381

the third pair differs: dlx6a is expressed in the same stages as dlx4b, but dlx5a is almost continuously expressed until 10 dPF, just before the start of osteoblast differentiation. Finally, dlx2b is expressed at 5 dPF only, and dlx4a from 3 to 7 dPF. Expression of all dlx genes ends after 10 dPF.

Branchial Arches Ceratobranchial bones. The ceratobranchial bones ossify as perichondral bones around the ceratobranchial cartilages. Ceratobranchial bone 5 ossifies first at 3 dPF. Ossification spreads along the cartilage in both the anterior (proximal) and posterior (distal) direction. Perichondral bone appears around the center of each of the ceratobranchials 1– 4 at 13 dPF (5.8 mm SL) and progresses proximally and distally (Cubbage and Mabee, 1996). Reference series confirm the presence of the mineralized ceratobranchial bone 5 at 3 dPF; the other ceratobranchial bones start to appear approximately simultaneously at 11 dPF. The anlagen of the ceratobranchial cartilages express all eight dlx genes except dlx2b from 1 to 2 dPF. Given the location of the start of perichondral ossification for each of the ceratobranchials, early dlx expression associated with the formation of these perichondral bones is expected to be situated halfway the longitudinal axis of each cartilage. Both dlx4b and dlx5a are expressed at 3 and 4 dPF at this level along the ceratobranchial cartilages 1–5 but only along ceratobranchial cartilage 5 have cells differentiated into osteoblasts in association with ceratobranchial bone 5. The other dlx genes are not expressed in the vicinity of the cartilage. Several dlx genes are expressed in the pharyngeal teeth associated with the fifth ceratobranchials. This finding has been documented by Jackman et al. (2004) and Borday-Birraux et al. (2006).

Epibranchial bones. Epibranchial bones 1– 4 ossify around the epibranchial cartilages in a posterior to anterior progression. Perichondral ossifications appear around the

centers of the cartilages at 15 dPF and progress medially and laterally (6.4 mm SL; Cubbage and Mabee, 1996). In our reference sections, epibranchial bone 5 appears at 14 dPF posterior to the lateral tip of epibranchial cartilage 4 and immediately anterior to the upper end of ceratobranchial 5. Genes of the pair dlx3b-dlx4b are the only dlx genes expressed between 3 and 6 dPF and 3 to 4 dPF, respectively, in mesenchymal cells near the roof of the pharyngeal cavity at the site of epibranchial bone 5 formation. The expression of these two genes is thus present long before the bone is formed.

Opercular Series Branchiostegal rays. Zebrafish have three branchiostegal rays. These dermal bones are formed in the branchiostegal membrane ventral to the anterior ceratohyal cartilage. The third ray is the first to ossify at 3 dPF (3.4 mm NL), followed by the second, and finally the first, which ossifies at 12 dPF (5.5 mm NL; Cubbage and Mabee, 1996). Our reference series show that branchiostegal ray 3 is present at 3 dPF, as a tiny piece of dermal bone surrounded by densely packed, active, secretory cells. Condensed mesenchymal cells for the second ray are visible at 5 dPF, and the anlage of the mineralized bone is present at 6 dPF. For branchiostegal ray 1, condensation occurs at 9 dPF and mineralization starts at 10 dPF. The three rays lie in close vicinity to the epidermis and are completely surrounded by osteoblasts. The eight dlx genes are expressed in the osteoblasts surrounding ray 3 during the mineralization stage (i.e., at 3 dPF; Figs. 21, 23) with two exceptions: dlx1a is not expressed at 3 dPF (only at 4 dPF), although its linked gene, dlx2a, is. dlx2b is only expressed at 3 dPF (Tables 1, 2A). Branchiostegal ray 2 shows expression for the pair dlx1a-dlx2a at 5 and 6 dPF. Remarkably, the dlx genes of the two other pairs differ in their expression pattern: dlx3b is expressed both during early and late developmental stages of the second ray, but its linked gene, dlx4b, is not expressed. Similarly, dlx5a is expressed during condensation and at the start of mineralization

of ray 2, while dlx6a is only expressed during condensation of mesenchymal cells and early secretion stages. The expression of dlx genes during branchiostegal ray 1 development is weaker than for the two other rays. Only dlx3b is expressed, but its expression is maintained up to the age of 15 dPF, at which stage dlx2a is expressed as well.

Interopercular bone. The interopercular bone is a dermal bone that ossifies in the interoperculomandibular ligament, which attaches to the retroarticular. Ossification begins in the mesenchyme lining the future dorsal edge of the bone at 10 dPF (5.1 mm NL; Cubbage and Mabee, 1996). In our reference series, the interopercular bone is visible as a distinct condensation of mesenchymal cells at 10 dPF. dlx2a is expressed in the osteoblasts surrounding the bone at 15 dPF (Table 1). dlx3b is expressed from 9 to 15 dPF, before and during condensation of mesenchymal cells, during synthesis and secretion of extracellular matrix, and during mineralization stages.

Opercular bone. Ossification of the opercular bone, a dermal bone, begins in the mesenchyme near its future ball and socket articulation with the hyomandibular cartilage at 3 dPF (3.0 mm NL; Cubbage and Mabee, 1996). At 13 dPF, bone formation continues in the mesenchyme posteriorly along a horizontal strut and ventral to it in a thin sheet. Our reference series show that the opercular bone is mineralized before 3 dPF. Osteoblasts are present along the lateral, ventral, and medial side. Dorsal to the opercular bone lie the dilatator and adductor operculi muscles (Schilling and Kimmel, 1997). The opercular bone, furthermore, is separated from the epidermis by only two or three rows of osteoblasts, preosteoblasts, and mesenchymal cells. No dlx expression is detected in the region of the opercular bone before 3 dPF, i.e., during condensation stage and synthesis and secretion of extracellular matrix. At 3, 4, and 5 dPF, all dlx genes are expressed in the osteoblasts surrounding the opercular

1382 VERREIJDT ET AL.

bone, in a rather late developmental stage (Tables 1, 2A). Expression signals are strongest along the ventral and medial side of the opercular bone, the lateral expression is very faint or nearly absent. Transcripts are never located along the dorsal side of the opercular bone. The pair dlx1a-dlx2a is no longer expressed after 5 and 6 dPF, respectively, except for the unique expression of dlx2a at 15 dPF. As for branchiostegal ray 1 and the interopercular bone, dlx3b is expressed until 15 dPF (Table 1). Its linked gene, dlx4b, however, is only expressed until 5 dPF. All other dlx genes are expressed between 3 dPF and 10 dPF at the latest, i.e., at a time the opercular bone is well mineralized (Figs. 16, 19).

Subopercular and preopercular bone. The subopercular bone is a dermal bone that ossifies along its future dorsal edge, in a region ventral to the opercular bone at 10 dPF (5.1 mm NL; Cubbage and Mabee, 1996). The subopercular bone starts to form at 4 dPF according to our reference series. Given its close proximity to the opercular bone, it is difficult to distinguish dlx expressions possibly associated with this bone, from those associated with the opercular bone. The preopercular bone is a crescentshaped dermal bone with dorsal (posterior) and ventral (anterior) arms ossifying in the mesenchyme posterior to the hyosymplectic cartilage. Condensation of mesenchymal cells starts at 13 dPF, and the preopercular bone is mineralized at 14 dPF. However, none of the dlx genes is expressed in the area of preopercular bone formation in the period examined.

DISCUSSION In the Zebrafish, dlx Genes Are Expressed in the Visceral Skeleton Only Summarizing our results, dlx expression is observed during the development of several dermal as well as cartilage bones, in cells that surround the bone anlagen: preosteoblasts and osteoblasts (in this study, we have not distinguished between the two cell types). Based on previous studies in

chicken and mice (reviewed in Merlo et al., 2000), expression of different dlx genes could be expected in the cranial skeleton of zebrafish. Surprisingly, all expression domains are connected with the development of cartilage bones and dermal bones associated with the visceral skeleton: the bones associated with the upper and lower jaws, the bones associated with the hyoid and branchial arches, and the opercular series. It is well known that the visceral arch skeleton is evolutionarily older than the neurocranium. A visceral arch skeleton, albeit of noncollagenous nature and non–neural crest-derived, is already present in nonvertebrate chordates such as Amphioxus, which lack a braincase (De Beer, 1937; Donoghue and Sansom, 2002). Only later, in the earliest vertebrates, was the splanchnocranium enlarged and its development taken over by neural crest. A neurocranium arose in basal vertebrates (Donoghue and Sansom, 2002), largely from mesodermal mesenchyme. Some elements, although functionally related to the neurocranium, have been argued to be evolutionary related to the splanchnocranium, in particular the trabeculae, and the prevomer and parasphenoid bone (De Beer, 1937; Daget, 1964; Kuratani et al., 1997). The parasphenoid is supposedly derived from arcual plates associated with the mandibular and hyoid arch (Daget, 1964; Schultze, 1993). The expression domains that we find associated with the parasphenoid indeed could support a splanchnocranial origin of this element. Available fossil data suggest that the dermal skeleton arose apparently as a single event, within the ostracoderms, early jawless vertebrates (Donoghue and Sansom, 2002), and the dermal bones became associated with splanchnocranial or neurocranial elements. In our data, we see a clear distinction between dlx-positive dermal bones associated with the splanchnocranium, and dlx-negative bones associated with the neurocranium. These data can be interpreted in two possible ways. If dlx expression is a hallmark of neural crest cells, then they could suggest that dermal bones associated with the neurocranium are non–neural crest-derived. This choice is an unlikely explanation,

given the widespread acceptance that all dermal skeleton is neural crest-derived (Smith and Hall, 1990). Alternatively, assuming that these bones are neural crest-derived, a possible explanation for the lack of dlx expression could lie in the so-called “neural crest plasticity and independent gene regulation model” (Trainor et al., 2003): neural crest cells can respond and adapt to the environment in which they migrate. Thus, when dermal bones first appeared, cells fated to form bones within the splanchnocranium (already populated by neural crest-derived and dlx-expressing ectomesenchyme) naturally could have adopted a similar expression profile; within the neurocranium, neural crest cells migrated into mesodermal mesenchyme thereby possibly being forced to change their expression program. The involvement of Dlx5 in the formation of calvarial (i.e., dermal) bones in birds and mammals could be secondarily acquired and unrelated to the function of these genes in the visceral skeleton, as discussed in the next sections. Based on expression studies in the primitive chordate Amphioxus, it has been proposed that an ancient function of Dll/Dlx was in specifying or patterning the neural crest (Holland et al., 1996). In vertebrates, all Dlx genes except zebrafish dlx2b, are expressed in ectomesenchymal cells derived from the cranial neural crest (Dolle´ et al., 1992; Bulfone et al., 1993; Akimenko et al., 1994; Robinson and Mahon, 1994; Simeone et al., 1994; Qiu et al., 1997; Yang et al., 1998; Davideau et al., 1999; Myojin et al., 2001; Neidert et al., 2001). The migratory neural crest cells populate the branchial arches, which in turn give rise to much of the craniofacial skeleton and connective tissue (Depew et al., 2002). In the mouse, members of the Dlx gene family are expressed early in cranial neural crest cells and later in craniofacial mesenchyme (Dolle´ et al., 1992; Bulfone et al., 1993; Robinson and Mahon, 1994; Simeone et al., 1994; Qiu et al., 1997; Yang et al., 1998; Acampora et al., 1999; Depew et al., 1999). Dlx gene expression is subsequently found in differentiating skeletal tissues (such as the skeletal derivatives of all pharyngeal arches but also bones composing the

dlx GENES IN CRANIAL BONES OF THE ZEBRAFISH 1383

base and roof of the skull; Acampora et al., 1999). Dlx5⫺/⫺ mutant mice exhibit malformations in both neural crest-derived bones of the visceral skeleton and in non– crest-derived neurocranial bones located along the midline of the base of the cranium, like the basioccipital bone. Defects are also present in dermatocranial bones forming the roof of the skull, such as the parietal, interparietal, and superoccipital bones (Acampora et al., 1999). Assuming that the skeletogenic mesenchyme of the visceral arches is exclusively neural crest-derived in the zebrafish, our results seem to be in contrast to studies reporting that, particularly in mice, Dlx5 and Dlx6 are broadly expressed in neural crest-derived as well as in mesodermal skeletal tissues (Simeone et al., 1994; Zhao et al., 1994; Ferrari et al., 1995; Chen et al., 1996; Ryoo et al., 1997; Yang et al., 1998; Acampora et al., 1999; Depew et al., 1999; Xu et al., 2001). However, Dlx5 (and perhaps Dlx6) might well have taken a novel function in skeletogenic mesenchyme within the tetrapod lineage, as discussed further. The basioccipital bone is the only endochondral bone present in zebrafish before 15 dPF (i.e., during the period of observation), although still in its perichondral phase. The basioccipital bone shows no dlx expression. We have no a priori reason why expression of dlx genes would not be expected in endochondral bones. In the mouse, Dlx5 and Dlx6 expression in developing endochondral bone was first described by Simeone et al. (1994). Dlx5 expression was found in the perichondrium, periosteum, and in osteoblasts of developing endochondral bones (both in the neurocranium and in the limbs; Acampora et al., 1999; Zhao et al., 1994). In the zebrafish, the absence of any signal in the basioccipital bone could well be explained by its location in the neurocranium, as a bone supporting the rear part of the braincase.

dlx Genes Show Different Expression Patterns in Dermal Vs. Cartilage Bones In contrast to the finding that dlx expression is consistently associated with the visceral skeleton, the set of genes expressed, and the spatiotempo-

ral patterns of expression, differ both within and between dermal and cartilage bones. Before further commenting on these differences, it is necessary to emphasize that, because the expression patterns are not always restricted to a well-delimited number of cells but are usually quite diffuse, it was sometimes impossible to assign the expression signals to specific bones. This is, e.g., the case for the coronomeckelian and anguloarticular bones: the expression domain observed between 3 and 5 dPF dorsal to Meckel’s cartilage, near the articulation with the palatoquadrate, could be associated with the presumptive condensation for the coronomeckelian bone, but could just as well be interpreted as a broadened expression domain associated with the anguloarticular bone. Similar interpretation problems were encountered for the opercular and subopercular bones. On the other hand, we should emphasize that serial sections reveal bone formation processes at a much earlier stage than alizarin red-stained specimens do (cf. our results compared with those of Cubbage and Mabee, 1996).

Dermal bones. During the period of observation, from 1 to 15 dPF, dlx genes are expressed during the development of the following dermal bones of the visceral skeleton: parasphenoid, dentary, anguloarticular, premaxillary, maxillary, entopterygoid, interopercular and opercular bones, and all three branchiostegal rays. Thus, all the dermal bones present in the skull at the stages examined, except for the urohyal, the infraorbital 1 and post-temporal bones, express dlx genes at some stage in their development. These three exceptions are bones that appear late, at respectively 12, 15, and 15 dPF. However, the lack of signal is probably not due to this late appearance, but to the fact that both the infraorbital 1 and post-temporal bones are canal bones associated with the neurocranium, and the urohyal bone is an intratendinous ossification and, therefore, can be assumed to develop differently in comparison to the true dermal bones. Although all the dermal bones that develop in the visceral skeleton during the period of observation show expres-

sion of dlx genes, not all of them express the same dlx genes in the same time frame. Sometimes expression of one or several genes extends throughout the different stages of differentiation of the bone (condensation of mesenchymal cells, synthesis and secretion of matrix, and mineralization), as is the case for the dentary, anguloarticular, entopterygoid, and interopercular bones and branchiostegal rays 1 and 2; alternatively expression can be restricted to one particular phase, the mineralization stage (e.g., the parasphenoid, premaxillary, maxillary, branchiostegal ray 3, and opercular bones). Overlapping expression patterns might suggest that there exist both redundant and distinct functions of the dlx genes in the morphogenesis of the visceral skeleton. Expression analysis alone does not allow us to confirm or refute this hypothesis, but previous studies have confirmed the existence of redundant and distinct function in the morphogenesis of the skeleton by analysis of the Dlx1, Dlx2, Dlx1/Dlx2, Dlx5, Dlx2/Dlx5 and Dlx5/ Dlx6 mouse mutants (Qiu et al., 1995, 1997; Acampora et al., 1999; Depew et al., 1999, 2005; Robledo et al., 2002). Thus, in zebrafish, the development of the dermal bones in the visceral skeleton appears to be associated with combinatorial patterns of dlx expression, which differ according to the bones considered. Bones of the first arch, at least those associated with the lower jaw (dentary and anguloarticular bone), share an expression pattern that is distinct from that of bones associated with the second arch (e.g., opercular bone, branchiostegal ray 3; Table 2A). Focusing on the mandibular arch, in the lower jaw, at least five dlx genes are expressed through all three stages of osteogenesis, the expression profile corresponding very well in the two bones (dentary and anguloarticular bone, Table 2A). Bones of the upper jaw (premaxillary, maxillary, and entopterygoid bone), on the other hand, show expression only late, at mineralization stage or beyond (dlx5a being the only one expressed also earlier, in the entopterygoid bone, but this gene is discussed further). Yet, at mineralization stage, the expression profile in the maxillary bone resembles that in lower jaw

1384 VERREIJDT ET AL.

bones (Table 2A). Also, the opercular bone, which starts its development rather dorsally, shares an expression pattern with that of branchiostegal ray 3 (the ray that forms simultaneously with the opercular bone in a rather ventral position). Furthermore, there is no connection between the stage in which the dlx genes are expressed, and which dlx genes in particular are expressed. This finding is a somewhat surprising result, considering that the bones go through similar stages of differentiation, involving the same type of cells (preosteoblasts and osteoblasts). Taken together, there is evidence that, rather than being involved in the different steps of differentiation of the bones, dlx genes impart cellular identity to the mesenchymal cells of the visceral skeleton. There is, however, little support so far for the idea that patterns of dlx expression in the zebrafish define within-arch identity, in a way that nested patterns of Dlx expression define proximodistal identity within the mouse branchial arches (Depew et al., 2002, 2005). These authors showed that simultaneous loss of Dlx5 and Dlx6 results in a homeotic transformation of the lower jaw into an upper jaw (a mirror-image duplication), where the lower jaw contained dermatocranial derivatives nearly identical in shape and size to the upper jaw, the most striking being the transformation of the dentary into a maxilla (Depew et al., 2002, 2005). At the very least, the distinctive expression patterns between upper jaw bones in the zebrafish calls for a precise fate mapping of the embryonic tissues contributing to each of these elements (Cerny et al., 2004). Somewhat in contradiction to the idea that expression has a role in assigning cellular identity, but is not specifically linked to the ossification process, Merlo et al. (2000) show that Dlx5 expression in mice is found in all limb bones (thus endochondral bones) during osteoblast differentiation and disappears in fully differentiated osteocytes. However, as emphasized before, Dlx5 might be a special case and is discussed further in more detail. An at first sight remarkable tendency that emerges from our observations is that, the earlier a bone starts to develop, the more dlx genes are ex-

pressed. The later in ontogeny it starts to develop, the less dlx genes are expressed (see Tables 1, 2). All eight dlx genes are expressed in the dentary, the branchiostegal ray 3, and the opercular bone, which start their development before or at 3 dPF. Bones that start to develop later, such as the premaxillary bone (9 dPF) and the interopercular bone (10 dPF), show expression of only one or two dlx genes. Intermediate numbers of genes being expressed are found associated with bones that develop between these two extremes. The two exceptions to this tendency are the parasphenoid and the entopterygoid bone, which only express a small subset of dlx genes, although they develop early. It is unlikely that these results are artefactual, for different reasons. First, the decrease in number of genes being expressed starts already at a stage in which probe penetration problems in whole-mounts are not yet expected to occur. Besides, whole-mount results were backed up by results of hybridization on sections. Second, expression of some genes is found up to the age of 15 dPF (branchiostegal ray 1, interopercular bone, opercular bone). Furthermore, we were able to confirm previously reported results on early expression of dlx genes in the zebrafish: the expression of dlx3b and dlx4b in the otic vesicle (Ekker et al., 1992; Solomon and Fritz, 2002); the expression of both dlx5a and dlx6a, but not dlx2b and dlx4a, in the otic vesicle at 48hPF (Akimenko et al., 1994; Ellies et al., 1997); and the expression of all eight zebrafish dlx genes except dlx2b in the mesenchyme forming the cartilaginous branchial arches (Akimenko et al., 1994; Ellies et al., 1997). So, rather than being artefactual, we believe that this decreasing number of genes being expressed concomitant with later developing bones, is related to our interpretation that dlx expression imparts cellular identity to the mesenchyme of the arch and that this process is accomplished early rather than late in development. In addition, rapid development in early stages could lead to redundancy, whereas finetuned transcriptional regulation in later stages could be responsible for the observed decrease in number of genes expressed. Of the two excep-

tions to this general tendency, at least the parasphenoid bone is consistent with this interpretation: this bone is not linked to a specific arch, but possibly derives evolutionary from the mandibular and hyoid arch. The case of the entopterygoid bone is more difficult to explain. With the exception of the premaxillary bone, the branchiostegal ray 1, and interopercular bone, dlx5a is transcribed in all mineralization stage dermal bones. The mammalian ortholog of dlx5a, Dlx5, is also expressed in differentiating dermal (intramembranous) bones (Depew et al., 1999) and Dlx5⫺/⫺ mice show a delayed ossification of dermal bones (Merlo et al., 2000). In humans, DLX5 gene expression is primarily detected in the mandible at 6 weeks and afterward in the maxilla (Davideau et al., 1999). Later on the expression of DLX5 becomes restricted to progenitor cells of both the developing tooth germs and of the bones and cartilages of mandible and maxilla.

Perichondral bones. Expression of dlx genes during the formation of cartilage bones in the visceral skeleton shows remarkable differences with that observed during dermal bone formation. First, unlike the dermal bones in the visceral skeleton, all of which show dlx expression at some stage, not all cartilage bones in the visceral skeleton show dlx expression. Transcripts of dlx genes are observed in association with the retroarticular, quadrate, hyomandibular, and symplectic bones, ceratobranchial bones 1–5, and epibranchial bone 5. No signals are detected in other cranial cartilage bones, such as the lateral ethmoids, the pterosphenoid, the sphenotic, the pterotic, the pro-otic, the basioccipital, the exoccipital, and the supraoccipital bones. In line with our previous interpretation, we believe that the absence of expression associated with these bones is connected to their position within the neurocranium. dlx expression is also lacking for the metapterygoid and the perichondral ossifications of the hyoid arch, i.e., the basihyal, anterior ceratohyal, posterior ceratohyal, and hypohyal bones, but this finding is more difficult to understand. Possibly this finding is because all these bones, ex-

dlx GENES IN CRANIAL BONES OF THE ZEBRAFISH 1385

cept the anterior ceratohyal bones (6 dPF), develop rather late (between 8 and 16 dPF), and the signals for their development might have been given much earlier through the specific combination of expression patterns in the mesenchyme of the arches, preceding the formation of these bones. A second difference concerns the time frame in which dlx genes are expressed. Unlike in dermal bones, and irrespective of the timing of development of the bone (whether early or late), dlx genes are predominantly expressed around the nonossified cartilage, well before the onset of perichondral ossification. Only two bones express dlx genes during all three stages of perichondral ossification: the retroarticular and hyomandibular bone. These are the only two bones that show such an early development. Finally, the dlx genes in each pair seem to behave quite similarly, unlike what is seen during the development of dermal bones, where differences in expression within a given dlx pair appear to be more frequent, with often only one gene of a pair being expressed. Given the above results, the early expression of dlx genes in presumptive areas of perichondral bone formation is more likely to be associated with ongoing differentiation of the cartilage than with early signals for bone formation. Alternatively, the early expression of these genes could well be part of the same combination of spatiotemporal patterns of expression that imparts cellular identity upon the cells that later differentiate into osteoprogenitor cells. However, that the patterns differ from those observed in the dermal bones of the corresponding arch (compare, e.g., the entopterygoid to quadrate bone, both of the same arch) argues in favor of the former possibility.

Dlx3 and Dlx5 in Bone Formation The genes dlx5a and dlx3b are the only ones expressed over a considerable length of time, well beyond the time other dlx genes are down-regulated (Table 1). Could this finding be an indication that, unlike the other dlx genes, dlx3b and dlx5a are in-

volved in bone formation itself and not in early patterning? In mice, Dlx3 expression is associated with new bone formation and regulation of osteoblast differentiation. Dlx3 is expressed in the osteoblasts, and overexpression of this gene in osteoprogenitor cells promotes, whereas specific knockdown of Dlx3 by RNA interference inhibits, induction of osteogenic markers (Hassan et al., 2004). In accordance with this result, we find strong expression of dlx3b during the different stages of ossification in almost all bones investigated: condensation of mesenchymal cells (the osteoprogenitors)/differentiation of osteoprogenitors into osteoblasts, synthesis and secretion of extracellular matrix by the osteoblasts, and matrix mineralization (Table 2A,B). Similarly, in the avian skull Dlx5 is expressed at several stages of osteoblast development (Holleville et al., 2003). In the mouse, Dlx5 stimulates osteoblast differentiation in bones formed either by endochondral (cartilage) or by membranous (dermal) ossification (Simeone et al., 1994; Ryoo et al., 1997; Acampora et al., 1999; Krauss and Lufkin, 1999; Miyama et al., 1999; Benson et al., 2000; Merlo et al., 2000; Erceg et al., 2003). In the case of endochondral bones, Dlx5 has also been shown to be expressed during chondrocyte differentiation (Ferrari et al., 1995; Depew et al., 1999; Ferrari and Kosher, 2002). Later, it is expressed in the zone of hypertrophy and in proliferating chondrocytes that are poised to differentiate (Bendall et al., 2003). Krauss and Lufkin (1999) have shown Dlx5 to be expressed in the perichondral region of all the developing fetal skeletal elements starting as early as cartilage initiation and continuing through the period of mineralization. In mice, Dlx5 occupancy increased over Dlx3 in mature osteoblasts at the mineralization stage of differentiation (Hassan et al., 2004). Dlx5 mutants appear to have a delay in the maturation of specific dermal bones (Depew et al., 1999, 2005). The mechanisms underlying these histological defects are unknown, although there is tissue culture evidence that the Dlx genes can regulate skeletal development through controlling the expression of two collagen

genes and osteocalcin (Dodig et al., 1996; Ryoo et al., 1997; Xu et al., 2001). It is clear that functional studies will be required to assign a role of dlx3b and dlx5a in the ossification process in the zebrafish. Through the relatively late development of the bones, knockdown experiments using morpholinos are uninformative, but advantage could be taken from in vitro cultures using head explants, as described by Van der heyden et al. (2005).

Epithelial Expression of dlx3b Unlike all other dlx genes, dlx3b is expressed also in epithelia, more precisely in the roof of the buccal cavity, and the ventral epidermis. These epithelial expression domains adjoin the areas in the mesenchyme where the parasphenoid, and the dentary, retroarticular, and quadrate bones form, respectively. These bones (dermal and perichondral bones) lie in the close vicinity of the epithelium, very near to the basal lamina. Epithelial expression of dlx genes is not unusual in se, and several studies have reported the epithelial expression (Ekker et al., 1992; Akimenko et al., 1994; Ellies et al., 1997; Solomon and Fritz, 2002; Borday-Birraux et al., 2006). Sites of epithelial–mesenchymal signaling are known to be the sites of dlx expression for vertebrates (Stock et al., 1996). However, not all epithelia adjacent to sites of bone formation (e.g., opercular bone, branchiostegal rays) exhibit dlx3b expression. Thus it is unlikely that dlx3b is involved in epithelial– mesenchymal signaling required in many instances of bone formation (Hall, 2005). Of interest, Zerucha et al. (1997) report that overexpression of truncated forms of dlx3b were found to result in perturbations in dlx5a expression in zebrafish and that cotransfection experiments indicate the ability of dlx3b to activate transcription of dlx5a. This might suggest that dlx5a is one of the target genes of dlx3b in embryos. However, because we found no epithelial dlx5a expression, the possible role of epithelial expression of dlx3b remains to be clarified.

1386 VERREIJDT ET AL.

CONCLUSIONS We started this study with the aim to search for differences in the developmental program between cartilage bones and dermal bones, within the two major functional units in the skull, the neurocranium and viscerocranium. Irrespective of the type of bone (cartilage bone or dermal bone), we obtained a clear-cut demarcation between neurocranial bones on the one hand, none of which show expression of any of the dlx genes, and bones associated to the visceral skeleton on the other hand, which show patterns of expression apparently related to the identity of the arch to which they are associated. The spatiotemporal pattern of expression of all the members of the dlx gene family within the visceral skeleton seems to support the view that the genes impart cellular identity to the different arches, allowing dermal bones to differentiate according to their position. Patterns of dlx expression seemingly associated with perichondral ossifications in the visceral arches, in contrast, are assumed to be related to the ongoing differentiation of the cartilage mould, on which these bones will eventually be deposited. Possibly, dlx genes acquired a patterning function in the visceral skeleton concomitant with the takeover of the visceral skeleton by the neural crest, which occurred in the earliest vertebrates (Donoghue and Sansom, 2002). This primitive involvement of dlx genes in the visceral skeleton has been conserved in the actinopterygian lineage leading to the zebrafish. The involvement of Dlx genes in the formation of neurocranial bones (endoskeletal and associated dermal bones), reported in mammals, is limited to some members of the Dlx family and could reveal a novel function, specifically related to the ossification process. This novel function would then have been secondarily acquired within the tetrapod (or perhaps sarcopterygian) lineage. Further research is needed to clarify the relationship between bones whose formation is characterized by patterns of dlx expression, and a neural crest origin of the osteoprogenitor cells.

EXPERIMENTAL PROCEDURES Zebrafish were laboratory reared. Fertilized eggs were allowed to develop in 3l tanks under a light period of 14 hr/day at 28°C. Embryos and larvae of 1 to 15 dPF were killed using an overdose of the anesthetic MS-222. Before fixation, they were staged according to a standard method (Kimmel et al., 1995) and measured to the nearest 0.1 mm (notochord length, NL, the standard measure of preflexion larvae [Ahlstrom et al., 1976; Leis and Trnski, 1989], or standard length, SL, from the anterior end of the upper jaw to the posterior end of the hypurals [Cubbage and Mabee, 1996]).

Light Microscopy To obtain reference series, specimens between 1 and 15 dPF were fixed in a mixture of 1.5% paraformaldehyde and 1.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4) for 2 hr at room temperature. Starting from 6 dPF onward, the specimens were decalcified by adding 0.1 M ethylenediaminetetraacetic acid (EDTA) to the fixative solution for 5 to 10 days, according to the level of calcification, at 4°C. Decalcifying solution was changed daily. After fixation, the specimens were rinsed in 0.1 M cacodylate buffer with 10% sucrose, then postfixed for 2 hr in 1% osmium tetroxide solution in 0.1 M cacodylate buffer to which 8% sucrose was added. After rinsing, the specimens were dehydrated in a graded series of ethanol and embedded in Epon (Huysseune and Sire, 1992). Serial semithin sections (1 or 2 ␮m thick) were cut using a diamond knife, stained with toluidine blue (1% toluidine blue in 1% borax) on a hot plate for approximately 30 sec, rinsed, dried, and mounted with DPX.

Preparation of RNA Probes The following fragments were used as templates for the synthesis of the antisense riboprobes: dlx1a, 775-bp polymerase chain reaction fragment; dlx2a, 1,700-bp cDNA fragment; dlx2b, 830-bp cDNA fragment; dlx3b, 1,500-bp cDNA fragment; dlx4a, 550-bp cDNA fragment; dlx4b, 1,000-bp cDNA fragment; dlx5a, 1,123-bp cDNA fragment; dlx6a, 777-bp cDNA fragment (Akimenko et

al., 1994; Ellies et al., 1997). Digoxigenin (DIG) -labeled antisense RNA probes were prepared using in vitro transcription and stored at ⫺80°C. For each of the genes, sense probes were used as controls, but they yielded no specific signals.

In Situ Hybridization on Sections The protocol for in situ hybridization was adapted from Strahle et al. (1994). Briefly, 1 to 15 dPF embryos and larvae were fixed at daily intervals in 4% paraformaldehyde (PFA) in phosphate buffered saline (PBS); when necessary (starting from 6 dPF), decalcified in 0.2 M Na2-EDTA in 4% PFA in PBS during one to several weeks, depending on the size of the larvae; dehydrated in methanol; and stored at ⫺20°C. Subsequently, they were rehydrated in a methanol/ PBS series and transferred into a 30% sucrose in PBS solution until the specimens were saturated with this solution. Specimens were embedded in a cryo-embedding compound (Prosan) using a standard freezing protocol. Cryosections were made at 16 ␮m in a cryotome and placed onto Superfrost/Plus slides, coated with a 1 mg/ml poly-L-lysine hydrobromide solution. Sections were stored dry at ⫺20°C. Hybridization on sections was performed with a probe concentration of approximately 500 ng/␮l at a temperature of 65°C overnight. Excess probe was washed away; heat-inactivated calf serum was used to prevent nonspecific binding of the antibody. Sections were incubated with an alkaline phosphatase linked to an anti-DIG antibody at 4°C overnight. Excess antibody was then washed away and staining performed with nitroblue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate (NBT/BCIP). After staining, the sections were post-fixed using 4% PFA in PBS and the slides were mounted with aquamount (Gurr) for light microscopic observation. Replicates of each experiment were usually carried out at least twice, sometimes even up to six times for the same age and the same gene.

dlx GENES IN CRANIAL BONES OF THE ZEBRAFISH 1387

Whole-Mount In Situ Hybridization Methanol-stored larvae of 1 to 7 dPF were subjected to whole-mount in situ hybridization according to the protocol described in Thisse and Thisse (1998). At least 20 to 30 larvae were used for each stage. Hybridized specimens were next embedded in Epon, serially sectioned at 6 ␮m and the slides were mounted with Epon. Light microscopic observation of hybridized sections and sectioned wholemounts was performed under both phase contrast and Nomarski illumination. Photomicrographs were taken under phase contrast illumination.

ACKNOWLEDGMENTS The authors thank Dr. M.-A. Akimenko and Dr. M. Ekker for providing the plasmids. F. Allizard, N. Van Damme, and M. Soenens provided expert technical assistance. This study benefited from COST Action B23 on “Oral facial development and regeneration.”

REFERENCES Acampora D, Merlo GR, Paleari L, Zerega B, Postiglione MP, Mantero S, Bober E, Barbieri O, Simeone A, Levi G. 1999. Craniofacial, vestibular and bone defects in mice lacking the Distal-less related gene Dlx5. Development 126:3795–3809. Ahlstrom EH, Butler JL, Sumida BY. 1976. Pelagic stromateoid fishes (Pisces, Perciformes) of the eastern pacific: kinds, distribution, and early life histories and observations of five of these from the northwest Atlantic. Bull Mar Sci 26:285–402. Akimenko MA, Ekker M, Wegner J, Lin W, Westerfield M. 1994. Combinatorial expression of three zebrafish genes related to distal-less: part of a homeobox gene code for the head. J Neurosci 14:3475– 3486. Beaumont A, Cassier P. 1994. Biologie animale. Les Corde´s, anatomie compare´e des Verte´bre´s. 6th ed. Paris: Dunod. 648 p. Bei M, Maas R. 1998. FGFs and BMP4 induce both Msx-1-independent and Msx-1-dependent signaling pathways in early tooth development. Development 125:4325– 4333. Bendall AJ, Hu G, Levi G, Abate-Shen C. 2003. Dlx5 regulates chondrocyte differentiation at multiple stages. Int J Dev Biol 47:335–344. Benson MD, Bargeon JL, Xiao G, Thomas PE, Kim A, Cui Y, Franceschi RT. 2000. Identification of a homeodomain binding element in the bone sialoprotein gene promoter that is required for its osteo-

blast-selective expression. J Biol Chem 275:13907–13917. Beverdam A, Merlo GR, Paleari L, Mantero S, Genova F, Barbieri O, Janvier P, Levi G. 2002. Jaw transformation with gain of symmetry after Dlx5/Dlx6 inactivation: mirror of the past? Genesis 34: 221–227. Blanc M. 1953. Contribution sur l’e´tude de l’oste´ogene`se chez les poissons te´le´oste´ens. Me´m Mus natn Hist nat, Se´rie A, 7:1–146. Borday-Birraux V, Van der heyden C, Bediais-Thibaud M, Verreijdt L, Stock DW, Huysseune A, Sire J-Y. 2006. Expression of Dlx genes during the development of the zebrafish pharyngeal dentition: evolutionary implications. Evol Dev 8:130 – 141. Bulfone A, Kim HJ, Puelles L, Porteus MH, Grippo JF, Rubenstein JLR. 1993. The mouse Dlx-2 (Tes-1) gene is expressed in spatially restricted domains of the forebrain, face and limbs in midgestation mouse embryos. Mech Dev 40:129 –140. Cerny R, Lwigale P, Ericsson R, Meulemans D, Epperlein HH, Bronner-Fraser M. 2004. Developmental origins and evolution of jaws: new interpretation of “maxillary” and “mandibular”. Dev Biol 276:225–236. Chambers D, McGonnel IM. 2002. Neural crest: facing the facts of head development. Trends Genet 18:381–384. Chen X, Li X, Wang W, Lufkin T. 1996. Dlx5 and Dlx6: an evolutionary conserved pair of murine homeobox genes expressed in the embryonic skeleton. Ann N Y Acad Sci 785:38 –47. Crackower MA, Scherer SW, Rommens JM, Hui CC, Poorkaj P, Soder S, Cobben JM, Hudgins L, Evans JP, Tsui LC. 1996. Characterization of the split hand/ split foot malformation locus SHFM1 at 7q21.3-q22.1 and analysis of a candidate gene for its expression during limb development. Hum Mol Genet 5:571–579. Cubbage CC, Mabee PM. 1996. Development of the cranium and paired fins in the zebrafish Danio rerio (Ostariophysi, Cyprinidae). J Morphol 229:121–160. Daget J. 1964. Le craˆne des Te´le´oste´ens. Me´m Mus Natn Hist nat, Se´rie A, Zool 31:163–341. Davideau JL, Demri P, Gu TT, Simmons D, Nessman C, Forest N, MacDougall M, Berdal A. 1999. Expression of DLX5 during human embryonic craniofacial development. Mech Dev 81:183–186. De Beer GR. 1937. The development of the vertebrate skull. Oxford: Oxford University Press. 552 p. Depew MJ, Liu JK, Long JE, Presley R, Meneses JJ, Pedersen RA, Rubenstein JL. 1999. Dlx5 regulates regional development of the branchial arches and sensory capsules. Development 126:3831– 3846. Depew MJ, Lufkin T, Rubenstein JLR. 2002. Specification of jaw subdivision by Dlx genes. Science 298:381–384. Depew MJ, Simpson CA, Morasso M, Rubenstein JLR. 2005. Reassessing the Dlx code: the genetic regulation of

branchial arch skeletal pattern and development. J Anat 207:501–561. Dodig M, Kronenberg MS, Bedalov A, Kream BE, Gronowicz G, Clark SH, Mack K, Liu Y-H, Maxon R, Pan ZZ, Upholt WB, Rowe DW, Lichtler AC. 1996. Identification of a TAATcontaining motif required for high level expression of the COL1A1 promoter in differentiated osteoblasts of transgenic mice. J Biol Chem 271:16422–16429. Dolle´ P, Price M, Duboule D. 1992. Expression of the murine Dlx-1 homeobox gene during facial, ocular and limb development. Differentiation 49:93–99. Donoghue PCJ, Sansom IJ. 2002. Origin and early evolution of vertebrate skeletonization. Microsc Res Tech 59:352– 372. Ekker M, Akimenko MA, Bremiller R, Westerfield M. 1992. Regional expression of three homeobox transcripts in the inner ear of zebrafish embryos. Neuron 9:27–35. Ellies DL, Stock D, Hatch G, Giroux G, Weiss KM, Ekker M. 1997. Relationship between the genomic organization and the overlapping embryonic expression patterns of the zebrafish dlx genes. Genomics 45:580 –590. Erceg I, Tadic T, Kronenberg MS, Marijanovic I, Lichter AC. 2003. Dlx5 regulation of mouse osteoblast differentiation mediated by avian retrovirus vector. Croat Med J 44:407–411. Ferguson CA, Tucker AS, Sharpe PT. 2000. Temporospatial cell interactions regulating mandibular and maxillary arch patterning. Development 127:403–412. Ferrari D, Kosher RA. 2002. Dlx5 is a positive regulator of chondrocyte differentiation during endochondral ossification. Dev Biol 252:257–270. Ferrari D, Sumoy L, Gannon J, Sun H, Brown AM, Upholt WB, Kosher RA. 1995. The expression pattern of the Distal-less homeobox-containing gene Dlx-5 in the developing chick limb bud suggests its involvement in apical ectodermal ridge activity, pattern formation, and cartilage differentiation. Mech Dev 52:257–264. Graham A, Begbie J, McGonnel I. 2004. Significance of the cranial neural crest. Dev Dyn 229:5–13. Hall BK. 1999. The neural crest in development and evolution. New York: Springer Verlag. 313 p. Hall BK. 2005. Bones and cartilage. Developmental and evolutionary skeletal biology. San Diego: Elsevier. 760 p. Hanken J, Hall BK. 1993. Evolution of the vertebrate skeleton. Nat Hist 92:28 –39. Hanken J, Thorogood P. 1993. Evolution and development of the vertebrate skull: the role of pattern formation. Trends Ecol Evol 8:9 –15. Hassan MQ, Javed A, Morasso MI, Karlin J, Montecino M, van Wijnen AJ, Stein GS, Stein JL, Lian JB. 2004. Dlx3 transcriptional regulation of osteoblast differentiation: temporal recruitment of Msx2, Dlx3 and Dlx5 homeodomain pro-

1388 VERREIJDT ET AL.

teins to chromatin of the osteocalcin gene. Mol Cell Biol 24:9248 –9261. Holland N, Panganiban G, Henyey E, Holland LZ. 1996. Sequence and developmental expression of AmphiDll, an amphioxus Distal-less gene transcribed in the ectoderm, epidermis and nervous system: insights into the evolution and forebrain of craniates. Development 122: 2911–2920. Holleville N, Quilhac A, Bontoux M, Monsoro-Burq A-H. 2003. BMP signals regulate Dlx5 during early avian skull development. Dev Biol 257:177–189. Huysseune A, Sire J-Y. 1992. Development of cartilage and bone tissues of the anterior part of the mandible in cichlid fish: a light and TEM study. Anat Rec 233:357– 375. Huysseune A, Sire J-Y. 1998. Evolution of patterns and processes in teeth and tooth-related tissues in non-mammalian vertebrates. Eur J Oral Sci 106:437–481. Jackman WR, Draper BW, Stock DW. 2004. Fgf signaling is required for zebrafish tooth development. Dev Biol 274: 139 –157. Kimmel CB, Balard WW, Kimmel SR, Ullman B, Schilling T. 1995. Stages of embryonic development of the zebrafish. Dev Dyn 203:253–310. Kimmel CB, Miller CT, Moens CB. 2001. Specification and morphogenesis of the zebrafish larval head skeleton. Dev Biol 233:239 –257. Kraus P, Lufkin T. 1999. Mammalian Dlx homeobox gene control of craniofacial and inner ear morphogenesis. J Cell Biochem Suppl 32/33:133–140. Kuratani S, Matsuo I, Aizawa S. 1997. Developmental patterning and evolution of the mammalian viscerocranium: genetic insights into comparative morphology. Dev Dyn 209:139 –155. Le Douarin N. 1982. The neural crest. New York: Cambridge University Press. 259 p. Le Douarin NM, Catala M, Batini C. 1997. Embryonic neural chimeras in the study of vertebrate brain and head development. Int Rev Cytol 175:241–309. Leis JM, Trnski T. 1989. The larvae of Indo-Pacific shorefishes. Sydney: New South Wales University Press; and Honolulu: University of Hawaii Press. 371 p, 83 plates. Merlo GR, Zerega B, Paleari L, Trombino S, Mantero S, Levi G. 2000. Multiple functions of Dlx genes. Int J Dev Biol 44:619 –626. Miyama K, Yamada G, Yamamoto TS, Takagi C, Miyado K, Sakai M, Ueno N, Shibuya H. 1999. A BMP-inducible gene, dlx5, regulates osteoblast differentiation and mesoderm induction. Dev Biol 208: 123–133. Myojin M, Ueki T, Sugahara F, Murakami Y, Shigetani Y, Aizawa S, Hirano S, Kuratani S. 2001. Isolation of Dlx and Emx gene cognates in an agnathan species, Lampetra japonica, and their expression patterns during embryonic and larval

development: conserved and diversified regulatory patterns of homeobox genes in vertebrate head evolution. J Exp Zool 291:68 –84. Neidert AH, Virupannavar V, Hooker GW, Langeland JA. 2001. Lamprey Dlx genes and early vertebrate evolution. Proc Natl Acad Sci U S A 98:1665–1670. Noden DM. 1988. Interactions and fates of avian craniofacial mesenchyme. Development Suppl 103:121–140. Otto F, Thornell AP, Crompton T, Denzel A, Gilmour KC, Rosewell IR, Stamp GWH, Beddington RSP, Mundlos S, Olsen BR, Selby PB, Owen MJ. 1997. Cbfa1, a candidate gene for cleidocranial dysplasia syndrome, is essential for osteoblast differentiation and bone development. Cell 89:75–771. Panganiban G, Rubenstein JLR. 2002. Developmental functions of the Distal-less/ Dlx homeobox genes. Development 129: 4371–4386. Patterson C. 1977. Cartilage bones, dermal bones and membrane bones, or the exoskeleton versus the endoskeleton. In: Andrews S, Miles R, Walker A, editors. Problems in vertebrate evolution. London: Academic Press. p 77–121. Price JA, Bowden DW, Wright JT, Pettinati MJ, Hart TC. 1998a. Identification of a mutation in DLX3 associated with tricho-dentoosseous (TDO) syndrome. Hum Mol Genet 7:563–569. Price JS, Wright JT, Kula K, Bowden DW, Hart TC. 1998b. A common DLX3 gene mutation is responsible for tricho-dentoosseous syndrome in Virginia and North Carolina families. J Med Genet 35:825– 828. Qiu M, Bulfone A, Martinez S, Meneses JJ, Shimamura K, Pedersen RA, Rubenstein JLR. 1995. Null mutation of Dlx-2 results in abnormal morphogenesis of proximal first and second branchial arch derivatives and abnormal differentiation in the forebrain. Genes Dev 9:2523– 2538. Qiu M, Bulfone A, Ghattas I, Meneses JJ, Christensen L, Sharpe PT, Presley R, Pedersen RA, Rubenstein JLR. 1997. Role of the Dlx homeobox genes in proximodistal patterning of the branchial arches: mutations of Dlx-1, Dlx-2, Dlx-1 and Dlx-2 alter morphogenesis of proximal skeletal and soft tissue structures derived from the first and second arches. Dev Biol 185:165–184. Quint E, Zerucha T, Ekker M. 2000. Differential expression of orthologous Dlx genes in zebrafish and mice: implications for the evolution of the Dlx homeobox gene family. J Exp Zool 288:235–241. Robinson GW, Mahon KA. 1994. Differential and overlapping expression domains of Dlx-2 and Dlx-3 suggests distinct roles for Distal-less homeobox genes in craniofacial development. Mech Dev 48:199 – 215. Robledo RF, Rajan L, Li X, Lufkin T. 2002. The Dlx5 and Dlx6 homeobox genes are essential for craniofacial, axial and ap-

pendicular skeletal development. Genes Dev 16:1089 –1101. Ryoo HM, Hoffmann HM, Beumer T, Frenkel B, Towler DA, Stein GS, Stein JL, van Wijnen AJ, Lian JB. 1997. Stagespecific expression of Dlx-5 during osteoblast differentiation: involvement in regulation of osteocalcin gene expression. Mol Endocrinol 11:1681–1694. Schilling TF, Kimmel CB. 1997. Musculoskeletal patterning in the pharyngeal segments of the zebrafish embryo. Development 124:2945–2960. Schultze H-P. 1993. Patterns of diversity in the skulls of jawed fishes. In: Hanken H, Hall BK, editors. The skull. Vol. 2. Chicago: University of Chicago Press. p 189 –254. Simeone A, Acampora D, Pannese M, D’Esposito M, Stornaiuolo A, Gulisano M, Mallamaci A, Kastury K, Druck T, Huebner K, Bonicelli E. 1994. Cloning and characterization of two members of the vertebrate Dlx gene family. Proc Natl Acad Sci U S A 91:2250 –2254. Smith MM, Hall BK. 1990. Development and evolutionary origins of vertebrate skeletogenic and odontogenetic tissues. Biol Rev 5:277–373. Solomon K, Fritz A. 2002. Concerted action of two dlx paralogs in sensory placode formation. Development 129:3127–3136. Stock DW, Ellies DL, Zhao Z, Ekker M, Ruddle FR, Weiss KM. 1996. The evolution of the vertebrate Dlx gene family. Proc Natl Acad Sci U S A 93:10858 – 10863. Stra¨hle U, Blader P, Adam J, Ingham PW. 1994. A simple and efficient procedure for non-isotopic in situ hybridisation to sectioned material. Trends Genet 10:75– 76. Thisse C, Thisse B. 1998. High resolution whole-mount in situ hybridization. In: Westerfield M, editor. Zebrafish science monitor Vol. 5. Eugene: University of Oregon Press. p 8 –9. [Zfin database: http:// www.igbmc.u-strasbg.fr/zf-info/zbook/ chapt9/9.82.html]. Trainor PA, Melton KR, Manzanares M. 2003. Origins and plasticity of neural crest cells and their roles in jaw and craniofacial evolution. Int J Dev Biol 47: 541–553. Van der heyden C, Allizard F, Sire J-Y, Huysseune A. 2005. Tooth development in vitro in two teleost fish, the cichlid Hemichromis bimaculatus and the cyprinid Danio rerio. Cell Tissue Res 321: 375–389. van Steensel MAM, Happle R, Steijlen PM. 2000. Molecular genetics of the hair follicle: the state of the art. Proc Soc Exp Biol Med 223:1–7. Verreijdt L, Vandervennet E, Sire J-Y, Huysseune A. 2002. Developmental differences between cranial bones in the zebrafish (Danio rerio): some preliminary light and TEM observations. Connect Tissue Res 43:109 –112.

dlx GENES IN CRANIAL BONES OF THE ZEBRAFISH 1389

Weiss KM, Bollekens J, Ruddle FH, Takashita K. 1994. Distal-less and other homeobox genes in the development of the dentition. J Exp Zool 270:273–284. Xu SC, Harris MA, Rubenstein JL, Mundy GR, Harris SE. 2001. Bone morphogenetic protein-2 (bmp-2) signaling to the col2alpha1 gene in chondroblasts requires the homeobox gene dlx-2. DNA Cell Biol 20:359 –365. Yang L, Zhang H, Hu G, Wang H, AbateShen C, Shen MM. 1998. An early phase

of embryonic Dlx5 expression defines the rostral boundary of the neural plate. J Neurosci 18:8322–8330. Zerucha T, Muller JP, Chartrand N, Ekker M. 1997. Cross-interactions between two members of the Dlx family of homeobox-containing genes during zebrafish development. Biochem Cell Biol 75:613–622. Zhao W, Dhoot GK. 2000. Both smooth and skeletal muscle precursors are present in foetal mouse oesuphagus and they follow

different differentiation pathways. Dev Dyn 218:587–602. Zhao GQ, Zhao S, Zhou X, Eberspaecher H, Solursh M, de Crombrugghe B. 1994. RDlx, a novel distal-less-like homeoprotein is expressed in developing cartilages and discrete neuronal tissues. Dev Biol 164:37–51. Zhao Z, Stock D, Buchanan A, Weiss K. 2000. Expression of Dlx genes during the development of the murine dentition. Dev Genes Evol 210:270 –275.