Search for a command to run...
Abstract The origin of paired fins was a major event in early vertebrate history that fuelled the adaptive radiation of the gnathostome clade. Evidence for the conservation of ancient regulatory systems between paired and median fins supports a model in which appendage-patterning mechanisms first evolved in the midline and were later redeployed to a new embryonic context, the flank. However, while paired fins and limbs are asymmetric along the dorso-ventral (DV) axis, the equivalent axis does not exist in median fins, which are symmetric across the midline, raising questions of how DV patterning evolved. Here, we combine genetic tools in zebrafish with comparative expression analyses in representative ray-finned and cartilaginous fishes to show that a subset of limb dorsoventral (DV) patterning genes are expressed during median fin development. Using genetic lineage tracing of the canonical limb dorsalizing factor lmx1bb , we further demonstrate that, although lmx1bb -derived cells occupy distinct spatial domains across fin types, a subset of this lineage is fated to differentiate into fin-ray osteoblasts in both paired and median fins. To test the potential conservation of regulatory modules across fin types, we generated a series of deletion mutants for two known lmx1b mouse limb enhancers, LARM1 and LARM2, that we show are partially conserved upstream of zebrafish lmx1bb . These experiments revealed that LARM function is essential for dorsoventral patterning in the paired pectoral and pelvic fins of zebrafish but is dispensable for formation of the unpaired dorsal and anal fins. To determine if the paired fin specificity of LARM -mediated lmx1b regulation is a feature of more basal gnathostomes, we used a single cell multiomics (RNAseq + ATACseq) approach in the epaulette shark Hemiscyllium ocellatum , a representative chondrichthyan and outgroup to osteichthyans. Strikingly, these analyses demonstrated linkages between the LARM1 enhancer and lmx1b in the unpaired dorsal fins of epaulette sharks, as well as the deployment of a Lmx1b-mediated regulatory network whose core downstream components are conserved between dorsal fins, paired fins, and limbs. Collectively, these data support the ancient origin of a “DV” regulatory module in median fins that was redeployed during the early evolution paired fins, facilitating the assembly of the DV patterning axis. Intriguingly, multiome-based inference of global fin network architecture in epaulette shark also revealed reduced regulatory complexity in pectoral relative to median fins, suggesting that co-option as a mechanism for anatomical innovation can proceed through selective streamlining of ancestral network components.