Developmental Roles of Pufferfish Hox Clusters and Genome Evolution in Ray-Fin Fish

Abstract

The pufferfish skeleton lacks ribs and pelvic fins, and has fused bones in the cranium and jaw. It has been hypothesized that this secondarily simplified pufferfish morphology is due to reduced complexity of the pufferfish Hox complexes. To test this hypothesis, we determined the genomic structure of Hox clusters in the Southern pufferfish Spheroides nephelus and interrogated genomic databases for the Japanese pufferfish Takifugu rubripes (fugu). Both species have at least seven Hox clusters, including two copies of Hoxb and Hoxd clusters, a single Hoxc cluster, and at least two Hoxa clusters, with a portion of a third Hoxa cluster in fugu. Results support genome duplication before divergence of zebrafish and pufferfish lineages, followed by loss of a Hoxc cluster in the pufferfish lineage and loss of a Hoxd cluster in the zebrafish lineage. Comparative analysis shows that duplicate genes continued to be lost for hundreds of millions of years, contrary to predictions for the permanent preservation of gene duplicates. Gene expression analysis in fugu embryos by in situ hybridization revealed evolutionary change in gene expression as predicted by the duplication-degeneration-complementation model. These experiments rule out the hypothesis that the simplified pufferfish body plan is due to reduction in Hox cluster complexity, and support the notion that genome duplication contributed to the radiation of teleosts into half of all vertebrate species by increasing developmental diversification of duplicate genes in daughter lineages.

The duplication of the Hox gene clusters in teleost fishes*1

“…Apart from these two differences, which we believe are due to incomplete data, all three pufferfish species, Tetraodon nigroviridis, Takifugu rubripes, and Spheroides nephalus [4], appear to have the same Hox gene complement with the exception of an intact HoxB7a in Spheroides nephalus that has become a pseudogene in the other two pufferfish species. Evidence for a third HoxA cluster in Takifugu rubripes was reported in [4] based on a a survey of the version 2.0 assemby of the fugu genome. We were not able to find hints for such a cluster in either the version 3.0 assembly of the fugu or in the tetraodon sequence data.…”

“…Recently, a thorough study on the Hox clusters of two pufferfishes, Spheroides nephalus and Takifugu rubripes was published [4]. Both species have a very similar structure of their Hox clusters as one would expect from their close phylogenetic relationship.…”

“…If a gene is named in the literature but the sequence is not available the gene is marked with (most of the spheroides sequences reported in [4]). and denote known pseudogenes, -indicates the known absence of a gene.…”

“…In the case of the HoxA and HoxB clusters there is ample evidence that they were indeed duplicated prior to the split of the euteleost and the ostariophysian lineages: Both the amino acid sequences of exon 1 of the HoxA2, HoxA10, HoxA11, and HoxA13 proteins (for which both first order paralogs were retained) and the conserved non-coding DNA within the cluster show that the cluster duplication preceeded the split of the fugu and the zebrafish lineage [9]. Phylogenetic analyis of HoxA9, HoxA13, HoxB1, and HoxB6 in [4] also support the duplication-first scenario. The similarity of the gene-complements of the duplicated cluster pairs, Tab.…”

“…The four cluster situation is retained in the sarcopterygian lineage (data are available for a number of mammalia, Xenopus tropicalis, and both known coelacanth species [28,4], in basal actinopterygians (bichir Polypterus senegalus [9]), and (presumably) also in condrichtya (horn shark, Heterodontus francisci [27]). Higher ray-finned fishes, however, have 6 or 7 Hox clusters that arose by means of duplication from the ancestral gnathostome clusters [3].…”