Genetic analysis of tissue interactions required for otic placode induction in the zebrafish.

Development of the vertebrate inner ear begins during gastrulation with induction of the otic placode. Several embryonic tissues, including cephalic mesendoderm, notochord, and hindbrain, have been implicated as potential sources of otic-inducing signals. However, the relative contributions of these tissues have not been determined, nor have any genes affecting placode induction been identified. To address these issues, we analyzed otic placode induction in zebrafish mutants that are deficient in prospective otic-inducing tissues. Otic development was monitored by examining mutant embryos for morphological changes and, in some cases, by visualizing expression patterns of dlx-3 or pax-2.1 in preotic cells several hours before otic placode formation. In cyclops (cyc-) mutants, which develop with a partial deficiency of prechordal mesendoderm, otic induction is delayed by up to 1 h. In one-eyed pinhead (oep-) mutants, which are more completely deficient in prechordal mesendoderm, otic induction is delayed by 1.5 h, and morphology of the otic vesicles is abnormal. Expression of marker genes in other regions of the neural plate is normal, suggesting that ablation of prechordal mesendoderm selectively inhibits otic induction. In contrast, the timing and morphology of otic development is not affected by mutations in no tail (ntl) or floating head (flh), which prevent notochord differentiation. Similarly, a mutation in valentino (val), which blocks early differentiation of rhombomeres 5 and 6 in the hindbrain, does not delay otic induction, although subsequent patterning of the otic vesicle is impaired. To test whether inductive signals from one tissue can compensate for loss of another, we generated double or triple mutants with various combinations of the above mutations. In none of the multiple mutants do the flh or val mutations exacerbate delays in placode induction, although val does contribute additively to defects in subsequent patterning of the otic vesicle. In contrast, mutants homozygous for both oep and ntl, which interact synergistically to disrupt differentiation of cephalic and axial mesendoderm, show a delay in otic development of about 3 h. These data suggest that cephalic mesendoderm, including prechordal mesendoderm and anterior paraxial mesendoderm, provides the first otic-inducing signals during gastrulation, whereas chordamesoderm plays no discernible role in this process. Because val- mutants are deficient for only a portion of the hindbrain, we cannot rule out a role for that tissue in otic placode induction. However, if the hindbrain does provide otic-inducing signals, they apparently differ quantitatively or qualitatively from the signals required for vesicle patterning, as val disrupts only the latter.