Extensive synteny conservation of holocentric chromosomes in Lepidoptera despite high rates of local genome rearrangements

Abstract

The recent assembly of the silkworm Bombyx mori genome with 432 Mb on 28 holocentric chromosomes has become a reference in the genomic analysis of the very diverse Order of Lepidoptera. We sequenced BACs from two major pests, the noctuid moths Helicoverpa armigera and Spodoptera frugiperda, corresponding to 15 regions distributed on 11 B. mori chromosomes, each BAC/region being anchored by known orthologous gene(s) to analyze syntenic relationships and genome rearrangements among the three species. Nearly 300 genes and numerous transposable elements were identified, with long interspersed nuclear elements and terminal inverted repeats the most abundant transposable element classes. There was a high degree of synteny conservation between B. mori and the two noctuid species. Conserved syntenic blocks of identified genes were very small, however, approximately 1.3 genes per block between B. mori and the two noctuid species and 2.0 genes per block between S. frugiperda and H. armigera. This corresponds to approximately two chromosome breaks per Mb DNA per My. This is a much higher evolution rate than among species of the Drosophila genus and may be related to the holocentric nature of the lepidopteran genomes. We report a large cluster of eight members of the aminopeptidase N gene family that we estimate to have been present since the Jurassic. In contrast, several clusters of cytochrome P450 genes showed multiple lineage-specific duplication events, in particular in the lepidopteran CYP9A subfamily. Our study highlights the value of the silkworm genome as a reference in lepidopteran comparative genomics.

Fig. 1.

Syntenic relationship at the CYP332A locus. Schemes at scale representing, from top to bottom, H. armigera, S. frugiperda, and B. mori. Arrows represent genes predicted by KAIKOGAAS. Only valid genes are shown (Fig. S1). Synteny links are shown with black dotted lines. In text boxes, gene ID, or for other genes, HP, unknown protein (UP; presence of a match to an EST with a threshold of 10⁻⁴⁰ by BlastN). Synteny blocks (as defined in text) are shown as colored boxes spanning genes arrows in the case of H. armigera and S. frugiperda, below and above gene arrows in the case of S. frugiperda and B. mori, respectively. See Fig. S2 for gene name abbreviations.

Fig. 2.

Number of genes per synteny block. Black and white striped bars: between H. armigera and S. frugiperda (average size 2.04). Black bars: between B. mori and S. frugiperda (average size 1.29). White bars: between B. mori and H. armigera (average size 1.32).

Fig. 3.

Correlation between synteny breaks and TE copy density. TE copy density is shown as a function of BAC length. 3A, S. frugiperda CYP4M region; 3B, H. armigera EcR region; dotted lines, TE copy density. Thick lines, regions corresponding to synteny breaks; thin lines. gene density.

Fig. 4.

Evolution of the CYP9A gene cluster. The CYP9A genes in the three species are shown in their correct orientation and order, but not relative distance, on the BACs or scaffold. From top to bottom: Sf, S. frugiperda; Ha, H. armigera; Bm, B. mori. The phylogenetic tree based on alignments of the CYP proteins is superimposed with its correct topology, but with branch lengths modified for clarity of the figure. Gene duplication events (●), B. mori/noctuid split (i.e., ancient speciation rather than duplication event) (■), and the S. frugiperda/H. armigera split (▲) are indicated. The sequence of events for the CYP9A15, -26, and -27 genes is unresolved. The relative orientation of the genes indicates at least two inversions in addition to the duplication events. Recognized flanking genes [alcohol dehydrogenases (ADH) and BmETS transcription factor] are shown (see Fig. S2 for details).