Genome size evolution in pufferfish: a comparative analysis of diodontid and tetraodontid pufferfish genomes

Abstract

Smooth pufferfish of the family Tetraodontidae have the smallest vertebrate genomes yet measured. They have a haploid genome size of approximately 400 million bp (Mb), which is almost eight times smaller than the human genome. Given that spiny pufferfish from the sister family Diodontidae and a fish from the outgroup Molidae have genomes twice as large as smooth puffers, it appears that the genome size of smooth puffers has contracted in the last 50-70 million years since their divergence from the spiny puffers. Here we use renaturation kinetics to compare the repetitive nature of the smooth and spiny puffer genomes. We also estimate the rates of small (<400 bp) insertions and deletions in smooth and spiny puffers using defunct non-LTR retrotransposons. We find a significantly greater abundance of a transposon-like repetitive DNA class in spiny puffers relative to smooth puffers, in addition to nearly identical indel rates. We comment on the role that large insertions may play in the evolution of genome size in these two groups.

Figure 1.

Cot curves for a tetraodontid (smooth) puffer, Fugu rubripes, and a diodontid (spiny) puffer, Diodon hystrix. Filled circles indicate F. rubripes data points, and open circles represent D. hystrix data points. Open arrows mark the Cot_1/2 values of the D. hystrix repetitive (left) and single-copy (right) classes, at which half of those components are estimated to have renatured. The filled arrow indicates the Cot_1/2 value of the single-copy component of the F. rubripes genome. The Cot_1/2 value of the F. rubripes repetitive component is estimated to be −3.3, and does not appear on this graph.

Figure 2.

Frequency histogram of deletion sizes for tetraodontid (smooth) and diodontid (spiny) puffers. Represented are 31 deletions from the tetraodontid alignment and 48 deletions from the diodontid alignment.

Figure 3.

Number of terminal indels in individual sequences plotted against the number of terminal nucleotide substitutions in each sequence for the tetraodontid (A) and diodontid (B) data sets. Lines are linear regressions with the intercept fixed at 0. A significant correlation is detected in both plots using Spearman's rank correlation statistic.

Figure 3.

Number of terminal indels in individual sequences plotted against the number of terminal nucleotide substitutions in each sequence for the tetraodontid (A) and diodontid (B) data sets. Lines are linear regressions with the intercept fixed at 0. A significant correlation is detected in both plots using Spearman's rank correlation statistic.

Figure 4.

Plot of insertion and deletion size in the tetraodontid (A) and diodontid (B) data sets. Points above the abscissa are insertions. Points below the abscissa are deletions. The abscissa indicates the midpoint of insertions and deletions on alignment. The tetraodontid alignment is 480 bp in length, and the diodontid alignment is 720 bp in length.

Figure 4.

Plot of insertion and deletion size in the tetraodontid (A) and diodontid (B) data sets. Points above the abscissa are insertions. Points below the abscissa are deletions. The abscissa indicates the midpoint of insertions and deletions on alignment. The tetraodontid alignment is 480 bp in length, and the diodontid alignment is 720 bp in length.