A macroevolutionary role for chromosomal fusion and fission in Erebia butterflies

Abstract

The impact of large-scale chromosomal rearrangements, such as fusions and fissions, on speciation is a long-standing conundrum. We assessed whether bursts of change in chromosome numbers resulting from chromosomal fusion or fission are related to increased speciation rates in Erebia, one of the most species-rich and karyotypically variable butterfly groups. We established a genome-based phylogeny and used state-dependent birth-death models to infer trajectories of karyotype evolution. We demonstrated that rates of anagenetic chromosomal changes (i.e., along phylogenetic branches) exceed cladogenetic changes (i.e., at speciation events), but, when cladogenetic changes occur, they are mostly associated with chromosomal fissions rather than fusions. We found that the relative importance of fusion and fission differs among Erebia clades of different ages and that especially in younger, more karyotypically diverse clades, speciation is more frequently associated with cladogenetic chromosomal changes. Overall, our results imply that chromosomal fusions and fissions have contrasting macroevolutionary roles and that large-scale chromosomal rearrangements are associated with bursts of species diversification.

Fig. 1.. Sample distribution and relationships within the Palearctic genus Erebia.

(A) Map of the Northern Hemisphere indicating sampling locations of Erebia specimens used, colored by clade. (B) Known chromosome numbers of Erebia species, grouped by clade [from (21)]. (C) Time-calibrated phylogeny of Erebia calculated in MrBayes. The fossil V. amerindica was used to calibrate the root of the tree (i.e., stem lineage of Satyrini), while the fossil L. corbieri was placed at the crown node of Satyrini. Clade names are based on (21) with the exception of medusa and pluto. For each clade a representative phenotype is shown. From top to bottom, these are Erebia medusa, Erebia pluto, Erebia pronoe, Erebia epiphron, Erebia tyndarus, Erebia ligea, Erebia magdalena, Erebia embla, and Erebia parmenio.

Fig. 2.. Summary of the chromosome evolution model for Erebia, implemented in ChromoSSE.

Estimated ancestral chromosome numbers for Erebia, inferred using the phylogenomic topology of fig. S1. Chromosome numbers are indicated proportionally by the color of the pie charts at the branch nodes. The pie charts at the “shoulders” of each node represent the inferred chromosomal state immediately after a speciation event. The two inferred chromosome numbers with highest posterior distribution of the ancestor of each clade are depicted atop or below the node that starts that clade. For deeper nodes within the tree, the reconstructed ancestral character states are likewise presented. For each extant species in the phylogeny, the karyotype is shown before the name. When the karyotype of a species is not known, the species is denoted with a “?.” Clades are named as in Fig. 1. The posterior probabilities of reconstructed ancestral chromosome states are visualized in fig. S7.

Fig. 3.. A summary of the posterior distributions of inferred ChromoSSE parameters for the overall Erebia analysis.

Shown are the posterior densities per parameter, the raw data points, and the 95% HPD interval as presented by the bold black line underneath each density plot (table S2). All parameter rates are expressed in events per species per million years (Ma).

Fig. 4.. Summary of ana- and cladogenetic ancestral chromosome estimations for six Erebia clades.

(A) Overview of the phylogeny of Erebia with relevant clades highlighted. (B to H) Distributions of the posterior probability estimates for (B) anagenetic fusions, (C) anagenetic fissions, (D) relative extinction rates, (E) cladogenetic changes unrelated to chromosomal change, (F) cladogenetic chromosomal fusions, (G) cladogenetic chromosomal fissions, and (H) total speciation rates. For (B) and (C) and (E) to (G), comparisons between standardized parameter estimations of clade specific state-dependent birth-death models (ChromoSSE) and birth-death independent models (ChromEVOL) are shown. Each dot indicates the mean of the posterior probability space. All parameter rates are expressed in events per species per million years. An alternative visualization, per clade, not per parameter, is given in fig. S8.