Background: Although the vertebrate skeleton arose in the sea 500 million years ago, our understanding of the molecular fingerprints of chondrocytes and osteoblasts may be biased because it is informed mainly by research on land animals. In fact, the molecular fingerprint of teleost osteoblasts differs in key ways from that of tetrapods, but we do not know the origin of these novel gene functions. They either arose as neofunctionalization events after the teleost genome duplication (TGD), or they represent preserved ancestral functions that pre-date the TGD. Here, we provide evolutionary perspective to the molecular fingerprints of skeletal cells and assess the role of genome duplication in generating novel gene functions. We compared the molecular fingerprints of skeletogenic cells in two ray-finned fish: zebrafish (Danio rerio)--a teleost--and the spotted gar (Lepisosteus oculatus)--a "living fossil" representative of a lineage that diverged from the teleost lineage prior to the TGD (i.e., the teleost sister group). We analyzed developing embryos for expression of the structural collagen genes col1a2, col2a1, col10a1, and col11a2 in well-formed cartilage and bone, and studied expression of skeletal regulators, including the transcription factor genes sox9 and runx2, during mesenchymal condensation.
Results: Results provided no evidence for the evolution of novel functions among gene duplicates in zebrafish compared to the gar outgroup, but our findings shed light on the evolution of the osteoblast. Zebrafish and gar chondrocytes both expressed col10a1 as they matured, but both species' osteoblasts also expressed col10a1, which tetrapod osteoblasts do not express. This novel finding, along with sox9 and col2a1 expression in developing osteoblasts of both zebrafish and gar, demonstrates that osteoblasts of both a teleost and a basally diverging ray-fin fish express components of the supposed chondrocyte molecular fingerprint.
Conclusions: Our surprising finding that the "chondrogenic" transcription factor sox9 is expressed in developing osteoblasts of both zebrafish and gar can help explain the expression of chondrocyte genes in osteoblasts of ray-finned fish. More broadly, our data suggest that the molecular fingerprint of the osteoblast, which largely is constrained among land animals, was not fixed during early vertebrate evolution.