Analysis of novel caudal hindbrain genes reveals different regulatory logic for gene expression in rhombomere 4 versus 5/6 in embryonic zebrafish

Priyanjali Ghosh, Jennifer M. Maurer, Charles G. Sagerström

Previous work aimed at understanding the gene regulatory networks (GRNs) governing caudal hindbrain formation identified morphogens such as Retinoic Acid (RA) and Fibroblast growth factors (FGFs), as well as transcription factors like hoxb1b, hoxb1a, hnf1ba, and valentino as being required for rhombomere (r) r4-r6 formation in zebrafish. Considering that the caudal hindbrain is relatively complex - for instance, unique sets of neurons are formed in each rhombomere segment - it is likely that additional essential genes remain to be identified and integrated into the caudal hindbrain GRN. By taking advantage of gene expression data available in the Zebrafish Information Network (ZFIN), we identified 84 uncharacterized genes that are expressed in r4-r6. We selected a representative set of 22 genes and assayed their expression patterns in hoxb1b, hoxb1a, hnf1b, and valentino mutants with the goal of positioning them in the caudal hindbrain GRN. We also investigated the effects of RA and FGF on the expression of this gene set. To examine whether these genes are necessary for r4-r6 development, we analyzed germline mutants for six of the genes (gas6, gbx1, sall4, eglf6, celf2, and greb1l) for defects in hindbrain development. Our results reveal that r4 gene expression is unaffected by the individual loss of hoxb1b, hoxb1a or RA, but is under the combinatorial regulation of RA together with hoxb1b. In contrast, r5/r6 gene expression is dependent on RA, FGF, hnf1ba and valentino - as individual loss of these factors abolishes r5/r6 gene expression. Our analysis of six mutant lines did not reveal rhombomere or neuronal defects, but transcriptome analysis of one line (gas6 mutant) identified expression changes for genes involved in several developmental processes - suggesting that these genes may have subtle roles in hindbrain development. We conclude that r4-r6 formation is relatively robust, such that very few genes are absolutely required for this process. However, there are mechanistic differences in r4 versus r5/r6, such that no single factor is required for r4 development while several genes are individually required for r5/r6 formation.