Multiple reversals of strand asymmetry in molluscs mitochondrial genomes, and consequences for phylogenetic inferences

Abstract

Strand asymmetry in nucleotide composition is a remarkable feature of animal mitochondrial genomes. The strand-specific bias in the nucleotide composition of the mtDNA has been known to be highly problematic for phylogenetic analyses. Here, the strand asymmetry was compared across 140 mollusc species and analyzed for a mtDNA fragment including twelve protein-coding genes. The analyses show that almost all species in Gastropoda (except Heterobranchia) and all species in Bivalvia present reversals of strand bias. The skew values on individual genes for all codon positions (P123), third codon positions (P3), and fourfold redundant third codon positions (P4FD) indicated that CG skews are the best indicators of strand asymmetry. The differences in the patterns of strand asymmetry significantly influenced the amino acid composition of the encoded proteins. These biases are most striking for the amino acids Valine, Cysteine, Asparagine and Threonines, which appear to have evolved asymmetrical exchanges in response to shifts in nucleotide composition. Molluscs with strong variability of genome architectures (ARs) are usually characterized by a reversal of the usual strand bias. Phylogenetic analyses show that reversals of asymmetric mutational constraints have consequences on the phylogenetic inferences, as taxa characterized by reverse strand bias (Heterobranchia and Bivalvia) tend to group together due to long-branch attraction (LBA) artifacts. Neutral Transitions Excluded (NTE) model did not overcome the problem of heterogeneous biases present in molluscs mt genomes, suggested it may not be appropriate for molluscs mt genome data. Further refinement phylogenetic models may help us better understand internal relationships among these diverse organisms.