Sparc (Osteonectin), a matricellular glycoprotein expressed by many differentiated cells, is a major non-collagenous constituent of vertebratebones. Recent studies indicate that Sparc expression appears early in development, although its function and regulation during embryogenesis arelargely unknown. We cloned zebrafish sparc and investigated its role during development, using a mo rpholino antisense oligonucleotide-basedknockdown approach. Consistent with its strong expression in the otic vesicle and developing pharyngeal cartilages, knockdown of Sparc functionresulted in specific inner ear and cartilage defects that are highlighted by changes in gene expression, morphology and behavior. We rescued theknockdown phenotypes by co-injecting sparc mRNA, providing evidence that the knockdown phenotype is due specifically to impairment ofSparc function. A comparison of the phenotypes of Sparc knockdown and known zebrafish mutants with similar defects places Sparc downstreamof sox9 in the genetic network that regulates development of the pharyngeal skeleton and inner ear of vertebrates.

Background: The inner ear arises from a specialized set of cells, the otic placode, that formsat the lateral edge of the neural plate adjacent to the hindbrain. Previous studies indicated thatfibroblast growth factors (Fgfs) are required for otic induction; in zebrafish, loss of both Fgf3 andFgf8 results in total ablation of otic tissue. Furthermore, gain-of-function studies suggested thatFgf signaling is not only necessary but also sufficient for otic induction, although the amount ofinduced ectopic otic tissue reported after misexpression of fgf3 or fgf8 varies among differentstudies. We previously suggested that Foxi1 and Dlx3b may provide competence to form theear because loss of both foxi1 and dlx3b results in ablation of all otic tissue even in the presenceof a fully functional Fgf signaling pathway.Results: Using a transgenic line that allows us to misexpress fgf8 under the control of thezebrafish temperature-inducible hsp70 promoter, we readdressed the role of Fgf signaling andotic competence during placode induction. We find that misexpression of fgf8 fails to induceformation of ectopic otic vesicles outside of the endogenous ear field and has differentconsequences depending upon the developmental stage. Overexpression of fgf8 from 1-cell tomidgastrula stages leads to formation of no or small otic vesicles, respectively. Overexpressionof fgf8 at these stages never leads to ectopic expression of foxi1 or dlx3b, contrary to previousstudies that indicated that foxi1 is activated by Fgf signaling. Consistent with our results we findthat pharmacological inhibition of Fgf signaling has no effect on foxi1 or dlx3b expression, butinstead, Bmp signaling activates foxi1, directly and dlx3b, indirectly. In contrast to early activationof fgf8, fgf8 overexpression at the end of gastrulation, when otic induction begins, leads to muchlarger otic vesicles. We further show that application of a low dose of retinoic acid that doesnot perturb patterning of the anterior neural plate leads to expansion of foxi1 and to a massiveFgf-dependent otic induction.

The vertebrate inner ear arises from an ectodermalthickening, the otic placode, that forms adjacent to thepresumptive hindbrain. Previous studies have suggestedthat competent ectodermal cells respond to Fgf signalsfrom adjacent tissues and express two highly related pairedbox transcription factors Pax2a and Pax8 in the developingplacode. We show that compromising the functions of bothPax2a and Pax8 together blocks zebrafish ear development,leaving only a few residual otic cells. This suggests thatPax2a and Pax8 are the main effectors downstream of Fgfsignals. Our results further provide evidence that pax8expression and pax2a expression are regulated by twoindependent factors, Foxi1 and Dlx3b, respectively.Combined loss of both factors eliminates all indications ofotic specification. We suggest that the Foxi1-Pax8 pathwayprovides an early ‘jumpstart’ of otic specification that ismaintained by the Dlx3b-Pax2a pathway.

The vertebrate inner ear develops from the otic placode,an ectodermal thickening that forms adjacent to thepresumptive hindbrain. Previous studies have suggestedthat competent ectodermal cells respond to signals fromadjacent tissues to form the placode. Members of the Fgffamily of growth factors and the Dlx family of transcriptionfactors have been implicated in this signal-responsepathway. We show that compromising Fgf3 and Fgf8signaling blocks ear development; only a few scattered oticcells form. Removal of dlx3b, dlx4b and sox9a genestogether also blocks ear development, although a fewresidual cells form an otic epithelium. These cells fail toform if sox9b function is also blocked. Combined loss of Fgfsignaling and the three transcription factor genes, dlx3b,dlx4b and sox9a, also completely eliminates all indicationsof otic cells. Expression of sox9a but not dlx3b, dlx4b orsox9b requires Fgf3 and Fgf8. Our results provide evidencefor Fgf3- and Fgf8-dependent and -independent geneticpathways for otic specification and support the notion thatFgf3 and Fgf8 function to induce both the otic placode andthe epithelial organization of the otic vesicle.

The molecular genetic mechanisms of cartilageconstruction are incompletely understood. Zebrafishembryos homozygous for jellyfish (jef) mutations showcraniofacial defects and lack cartilage elements of theneurocranium, pharyngeal arches, and pectoral girdlesimilar to humans with campomelic dysplasia. We showthat two alleles of jef contain mutations in sox9a, one of twozebrafish orthologs of the human transcription factorSOX9. A mutation induced by ethyl nitrosourea changed aconserved nucleotide at a splice junction and severelyreduced splicing of sox9a transcript. A retrovirus insertioninto sox9a disrupted its DNA-binding domain. Inhibitingsplicing of the sox9a transcript in wild-type embryos withsplice site-directed morpholino antisense oligonucleotidesproduced a phenotype like jef mutant larvae, and causedsox9a transcript to accumulate in the nucleus; thisaccumulation can serve as an assay for the efficacy of amorpholino independent of phenotype. RNase-protectionassays showed that in morpholino-injected animals, thepercent of splicing inhibition decreased from 80% at 28hours post fertilization to 45% by 4 days. Homozygousmutant embryos had greatly reduced quantities of col2a1message, the major collagen of cartilage. Analysis of dlx2expression showed that neural crest specification andmigration was normal in jef (sox9a) embryos. Confocalimages of living embryos stained with BODIPY-ceramiderevealed at single-cell resolution the formation ofprecartilage condensations in mutant embryos. Besides thelack of overt cartilage differentiation, pharyngeal archcondensations in jef (sox9a) mutants lacked three specificmorphogenetic behaviors: the stacking of chondrocytesinto orderly arrays, the individuation of pharyngealcartilage organs and the proper shaping of individualcartilages. Despite the severe reduction of cartilages,analysis of titin expression showed normal musclepatterning in jef (sox9a) mutants. Likewise, calcein labelingrevealed that early bone formation was largely unaffectedin jef (sox9a) mutants. These studies show that jef (sox9a)is essential for both morphogenesis of condensations andovert cartilage differentiation.

A zebrafish Ftz-F1 homologue, zFF1A (zebrafish Ff1a or Nr5a2, a member of nuclear receptor superfamily) and its C-terminally truncated variant (zFF1B) were previously identified. Due to lack of the identity box (I-box) and activation function 2 (AF-2) domain, zFF1B lacks transactivation function and fails to synergize with estrogen receptor (ER) in regulating promoters. It was speculated that the I-box might be involved in the zFF1A/ER interaction. In the present study, the function of the I-box was examined. In the absence of the I-box or with an altered heptad 9, the AF-2 of zFF1A was not functional, either in the presence or absence of ER. The GST pull-down assay showed that zFF1A and its mutants exerted similar physical contacts with ER-LBD, suggesting that the “dimerization” domain (I-box) is essential for the transcriptional activity of zFF1A. Moreover, nuclear receptor coactivator selectively activated zFF1 with the I-box but exerted no effect on zFF1B, indicating that the I-box is able to interact with the coactivators. By deletion study and analysis of the identified domains in GAL4-DNA binding domain, other regions of zFF1A critical for its AF were also delineated. Consistent with the mutation analysis, AF-2 was active only in the presence of the I-box. We also identified a novel AF domain (AF-3) located in the hinge region (amino acids 155–267), although the activity of AF-3 was inhibited by its flanking region. We suggest that the D and E regions of zFF1A possess both positive and negative transactivation functions, and interdomain “cross-talk” may confer the full transcriptional activity of the protein.

Steroidogenic factor 1, a member of the fushitarazu factor 1 (FTZ-F1) subfamily of nuclear receptors,is a key regulator in mammalian reproduction.From an embryonic complementary DNA library,the zebrafish homolog of FTZ-F1 (zFF1A) andan alternatively spliced variant (zFF1B) were isolated.zFF1B represented a C-terminally truncatedversion of zFF1A. Whole mount in situ hybridizationand reverse transcriptase-PCR analysis revealedthat both zFF1A and B transcripts were present inthe developing pituitaries, adult fish brain, gonads,and liver, albeit zFF1B messenger RNA was absentin testis. Comparison of the primary sequences ofzFF1 with those of other FTZ-F1 subfamily membersshowed a close structural relationship betweenthe mouse liver receptor homolog, whichactivated the a1-fetoprotein gene in rodent liver.However, similar to mouse steroidogenic factor 1,zFF1A regulated chinook salmon gonadotropin IIbsubunit gene expression. On the contrary, zFF1B,which could bind a consensus gonadotrope-specificelement with an affinity similar to that ofzFF1A, lacked both the trans-activation functionand synergistic interaction with the estrogen receptor.Furthermore, cotransfection studies inHeLa cells showed that zFF1B was a strong competitorfor the action of zFF1A on the chinooksalmon gonadotropin IIb subunit gene promoter.Our investigation suggests that 1) zFF1 representsan ancestor protein of the vertebrate FTZ-F1 homologs;2) the antagonistic relationship betweenzFF1A and -B may dictate the expression of theFTZ-F1 target genes in a variety of tissues, includingthe pituitary; and 3) the naturally occurringzFF1B provides evidence that the C-terminal portionof zFF1A (80 amino acid residues) contains amajor trans-activation function and a protein-proteininterface. (Molecular Endocrinology 11: 877–890, 1997)