Multiple sequence alignments and phylogenetic trees from: Co-option of the limb patterning program in cephalopod eye development

Background

Across the Metazoa, similar genetic programs are found in the development of analogous, independently evolved, morphological features. The functional significance of this reuse and the underlying mechanisms of co-option remain unclear. Cephalopods have evolved a highly acute visual system with a cup shaped retina and a novel refractive lens in the anterior, important for a number of sophisticated behaviors including predation, mating and camouflage. Almost nothing is known about the molecular-genetics of lens development in the cephalopod.

Results

Here we identify the co-option of the canonical bilaterian limb pattering program during cephalopod lens development, a functionally unrelated structure. We show radial expression of transcription factors SP6-9/sp1, Dlx/dll, Pbx/exd, Meis/hth, and a Prdl homolog in the squid Doryteuthis pealeii, similar to expression required in Drosophila limb development. We assess the role of Wnt signaling in the cephalopod lens, a positive regulator in the developing Drosophila limb, and find the regulatory relationship reversed, with ectopic Wnt signaling leading to lens loss.

Conclusion

This regulatory divergence suggests that duplication of SP6-9 in cephalopods may mediate the co-option of the limb patterning program. Thus our study suggests that the limb network could perform a more universal developmental function in radial pattering and highlights how canonical genetic programs are repurposed in novel structures.

Methods
Genes were preliminarily identified using reciprocal BLAST with Mus musculus and Drosophila melanogaster sequences as bait with the exception of S-Crystallin where previous Doryteuthis opalescens sequences were also used (Altschul et al., 1990). Top hits in the D. pealeii transcriptome were trimmed for coding sequence and translated to amino acid sequences. To find related sequences, BLASTp was used, searching only the RefSeq protein database in NCBI filtered for vertebrate and arthropod models, as well as spiralian models when published annotated sequences could be found. The top hits of each gene name were downloaded and aligned with D. pealeii sequences for each tree using MAFFT in Geneious (Katoh, 2002). To check sequence redundancy and proper outgroups quick trees were made using FastTree. We constructed maximum-likelihood trees on the FASRC Cannon cluster supported by the FAS Division of Science Research Computing Group at Harvard University (Price et al. 2010). Using PTHREADS RAxML v.8.2.10, we ran the option for rapid bootstrapping with searcxh for best maximum likelihood tree, resampling with 1000 bootstrap replicates, the PROTGAMMAAUTO model of amino acid substitution, and otherwise default parameters (Stamatakis, 2014).