Identification of downstream targets of zebrafish Hoxb1a

Despite 30 years of Hox gene study we have a remarkably limited knowledge of the downstream target genes that Hox transcription factors regulate to confer regional identity. Here, we have used a microarray approach to identify genes that function downstream of a single vertebrate Hox gene, zebrafish hoxb1a. This gene plays a critical and conserved role in vertebrate hindbrain development, conferring identity to hindbrain rhombomere 4. For example, zebrafish Hoxb1a, similar to mouse Hoxb1, is required for the migration of r4-derived facial branchiomotor neurons into the posterior hindbrain. We have screened microarrays carrying more than 16,000 expressed sequence tags (ESTs) for genes that are differentially regulated in normal versus Hoxb1a-deficient rhombomere 4 tissue. Using this approach, we have identified both positively and negatively regulated candidate Hoxb1a target genes. We have used in situ hybridization to validate twelve positively regulated Hoxb1a targets. These downstream targets are expressed in a variety of subdomains within r4, with one gene, a novel prickle homolog (pk1b), expressed specifically within the facial branchiomotor neurons. Using morpholino knock-down we show that the Hoxb1a target Pk1b is required for facial neuron migration, a single aspect of rhombomere 4 identity. Keywords: Comparison of normal and Hox-deficient tissue We used microarray analysis to compare transcripts in Hoxb1a-positive versus Hoxb1a-deficient zebrafish hindbrain tissue from 19-20 hour post fertilization embryos. To focus our analysis specifically on the region of hoxb1a expression, we microdissected Hoxb1a-positive r4 tissue from unmanipulated embryos, and Hoxb1a-deficient r4 tissue from Hoxb1a-Morpholino injected embryos. As a secondary source of Hoxb1a-deficient tissue, we also dissected r2 from unmanipulated embryos. For each tissue sample we performed a total of three replicates. Our first experiment (Replicate 1) compared normal r4 and Hoxb1a-deficient r4. Our second experiment (Replicate 2) compared r4, Hoxb1a-deficient r4, and r2. Our third experiment (Replicate 3) compared r4, Hoxb1a-deficient r4, and two separate r2 samples.