Homoplastic evolution of proximal podomere patterning across Arthropoda

The jointed arthropod leg is often credited as a key innovation spurring the evolutionary success of the phylum. Across lineages, each territory of the appendage has served as the substrate for myriad adaptations and structural innovations. While much of the available data has supported a conserved suite of transcription factors responsible for the initiation and patterning of the arthropod appendage, recent works have begun to highlight instances of lineage-specific shifts in expression dynamics, particularly amongst the chelicerates, correlated both with the emergence of novel structures and developmental system drift in the specification of plesiomorphic structures. These works confound efforts to align or homologize the leg segments across arthropod lineages, an effort already challenged by ground plan differences in segment number between the subphyla. An even more unusual architecture of leg segments occurs in sea spiders (Pycnogonida), which alone among Arthropoda are considered by systematists to bear three coxae (proximal segments), in contrast to the single coxa of the remaining arthropods. Here, we examined proximal podomere patterning in the developing appendages of a sea spider and an arachnid. We find that a particular spatial arrangement of Lim1 and SoxNeuro characterizes the developing coxa and trochanter of the harvestman Phalangium opilio (Opiliones) throughout appendage formation, with RNAi depletions of Lim1 yielding deletions of coxa and trochanter. Comparison of expression patterns across taxa shows that Lim1 and SoxNeuro have a similar spatial arrangement in the harvestman and sea spider, but these suggest conflicting interpretations of leg segment homology. We show that this is attributable to homoplastic patterns of evolution in SoxNeuro across Arthropoda. Expanding our surveys of prominent PD axis-patterning transcription factors supports additional instances of homoplasy in the domains of Sp6-9 and Irx2. The pervasiveness of homoplasy in gene expression patterns demands measured skepticism of homology statements and careful comparisons of multiple patterning genes, toward achieving falsifiability in the homology of structures.