Fig. 7

Gene duplication of patristacins preceded the evolution of diverse expression patterns related to male pregnancy. Patristacins are unique, tandemly arrayed C6 astacin-like metalloprotease genes presumably co-opted during the evolution of male pregnancy [18]. a A maximum likelihood gene tree inferred from astacin-like metalloprotease amino acid sequences, representing five fish genomes, is rooted assuming Meprin1b proteins as an outgroup. Different protein subfamily clades (colored by clade and including terminology from Kawaguchi et al. [89]) mostly correspond to conserved syntenic regions. Clade support values are SH-aLRT, but see Figure S8 (in Additional file 1) for bootstrap values and tip accession numbers. Zebrafish sequences with annotated Ensembl gene names are labeled for reference. Patristacins comprise a monophyletic group nested within the Zc6ast1-4 clade, suggesting pipefish or syngnathid lineage-specific duplication events. Note the absence of pipefish orthologs from the Zc6ast5-6 clade (colored gray). In medaka, orthologs from this group are expressed exclusively in the developing jaw [89]. Also note the red asterisk in the hatching enzyme clade, which corresponds to intron loss in the pipefish lineage. b The physical arrangement of patristacins in the Gulf pipefish genome, with two other genes in the region (small text). Arrows indicate the direction of the sense strand and vertical bars reflect coding exons. Note that the status of "pastn-like orf" as a gene is uncertain, so it is depicted by open bars and a question mark where three missing exons would normally be. c Patristacin expression levels from RNA-seq data for six non-pregnant male brood pouch samples (blue), six pregnant pouch samples not including embryos (orange), and a pooled embryo library (black). Y-axis values are copies per million (cpm) on a log scale. Individual data points and boxplots are shown. Note the extreme expression differences between pastn1 and pastn2