Extreme Positive Selection on a New Highly-Expressed Larval Glycoprotein (LGP) Gene in Galaxias Fishes (Osmeriformes: Galaxiidae)

We describe the intron-exon structure and DNA/protein sequences of a new larval glycoprotein (LGP) gene from nine species of galaxiid fish. The gene has a distant similarity to Danio THP (Tamm-Horsfall urinary glycoprotein; uromodulin) and cichlid SPP120 (seminal plasma glycoprotein) due to conserved features of its zona pellucida (ZP) domain, including eight highly conserved cysteines and a consensus furin cleavage site. Using a combination of 454 sequencing of cDNA and exon-primed intron-spanning sequencing of genomic DNA, we obtained full sequences of the coding region (996 bp) and its intervening sequences (1,459 bp). LGP shows an exceptionally strong signal of positive selection over the entire coding region, as evidenced by d(N)/d(S) values >1. Across nine species of Galaxias, 87/332 (26%) amino acid residues are variable, compared with 9/386 (2%) for mitochondrial cytochrome b (cytb) in the same group of species. Across 36 interspecific pairwise comparisons, genetic distances are in all cases larger for coding region than for introns, by a factor of 2.4-fold on average. Reading frame, gene structure, splice sites, and many ZP motifs are conserved across all species. Together with the fact that the gene is expressed in all species, these results argue clearly against the possibility of a pseudogene. We show by 454 sequencing and quantitative polymerase chain reaction that the transcript is abundant (ca. 0.5%) in newly hatched larvae and appears to be almost absent from a range of adult tissues. We postulate that the strong Darwinian evolution exhibited by this protein may reflect some type of immunoprotection at this vulnerable larval stage.