Chromosome-Level Assemblies of the Pieris mannii Butterfly Genome Suggest Z-Origin and Rapid Evolution of the W Chromosome

Abstract

The insect order Lepidoptera (butterflies and moths) represents the largest group of organisms with ZW/ZZ sex determination. While the origin of the Z chromosome predates the evolution of the Lepidoptera, the W chromosomes are considered younger, but their origin is debated. To shed light on the origin of the lepidopteran W, we here produce chromosome-level genome assemblies for the butterfly Pieris mannii and compare the sex chromosomes within and between P. mannii and its sister species Pieris rapae. Our analyses clearly indicate a common origin of the W chromosomes of the two Pieris species and reveal similarity between the Z and W in chromosome sequence and structure. This supports the view that the W in these species originates from Z-autosome fusion rather than from a redundant B chromosome. We further demonstrate the extremely rapid evolution of the W relative to the other chromosomes and argue that this may preclude reliable conclusions about the origins of W chromosomes based on comparisons among distantly related Lepidoptera. Finally, we find that sequence similarity between the Z and W chromosomes is greatest toward the chromosome ends, perhaps reflecting selection for the maintenance of recognition sites essential to chromosome segregation. Our study highlights the utility of long-read sequencing technology for illuminating chromosome evolution.

Keywords: Lepidoptera; genome assembly; homology; long-read sequencing; sequence alignment; sex chromosome.

Fig. 1.

Alternative evolutionary pathways from an ancestral Z-/ZZ sex chromosome configuration (top row) to a ZW/ZZ system through Z chromosome fusion (A) and through B chromosome recruitment (B). In (A), the Z chromosome fuses with an autosome (middle). After extensive evolution of the neosex chromosomes thus formed, the W may retain some sequence homology and structural similarity (synteny) to formerly autosomal segments of the Z (bottom, indicated by connecting lines). In (B), a B chromosome (nonessential, often small chromosome occurring sporadically within a species and not segregating with the ordinary A chromosomes) may gain sequence copies from the Z chromosome (middle). The sequence homology acquired in this way (bottom) may allow this chromosome to become a W chromosome segregating with the Z, although similarity in chromosome structure is expected to be low.

Fig. 2.

Comparison of individual chromosome lengths between (A) the P. mannii male and female v2 genome assemblies and (B) the female P. mannii and P. rapae assemblies. Autosomes are shown in light gray, the Z chromosome in dark gray, and the W chromosome in blue. In (C), the positions of sequence tags from the P. mannii W chromosome (not repeat-masked) are plotted against their corresponding alignment positions on the P. rapae W chromosome (n = 226 tags with unique alignment). An analogous analysis based on 226 sequence tags drawn at random from a representative autosome (chromosome 1) is shown in (D).

Fig. 3.

Female to male RAD locus read depth ratio along the Z chromosome, and along an exemplary autosome of approximately similar length, for P. mannii and P. rapae. RAD loci with a balanced read depth between the sexes (ratio around 1) reflect sequences present on the two parental chromosomes in each sex. In contrast, a reduced read depth in females (ratio around 0.5) along most of the Z chromosome indicates segments hemizygous in females, hence missing on the W chromosome. The horizontal lines represent the average read depth ratio observed across the autosomes. The number of individuals is 19 and 6 per sex for P. mannii and P. rapae, and the number of RAD loci on the Z and chromosome 2 is 2,501 and 2,138 for P. mannii and 2,057 and 1,817 for P. rapae.

Fig. 4.

Distinctive signatures of the P. mannii W chromosome include an exceptionally low proportion of nonrepeated DNA, a low density of loci to which RAD sequences align uniquely (expressed as loci per Mb), and a high GC content, relative to the autosomes (A) and the Z chromosome. The autosomal values are the medians across all 24 autosomes, with the error bars representing the associated 95% bootstrap compatibility intervals.

Fig. 5.

Exploring W chromosome similarity based on the alignment of sequence tags within and between Pieris species. The upper row shows the chromosome-specific density (alignments per Mb) of P. mannii and P. rapae sequence tags extracted from the repeat-masked W chromosome when aligned to the conspecific genome lacking the W (in P. mannii the male genome and in P. rapae the female genome with the W chromosome excluded). The lower row shows the alignment density of the same W sequence tags aligned to the female genome (including the W) of the sister species. Autosomes are plotted in light gray, Z chromosomes in dark gray, and W chromosomes in blue.

Fig. 6.

Structural similarity between the Z and W chromosomes in P. mannii and P. rapae. The left graphs show the Spearman coefficients (r_S) for the correlations between W sequence tag positions on their source chromosome (i.e., the repeat-masked W) and on all target chromosomes to which at least 120 tags aligned uniquely. The Z chromosome is shaded dark gray and autosomes light gray. In the right graphs, the positions of W sequence tags are connected to their alignment positions on the Z chromosome. The sequence tags were extracted from the repeat-masked W and were required to map uniquely to the Z when aligning them to the genome lacking the W (in P. mannii the male genome and in P. rapae the female genome with the W chromosome excluded). Sample size is n = 260 and 979 sequence tags for P. mannii and P. rapae. The sex chromosomes are drawn to scale, and the tick marks delimit 2 Mb intervals. The P. rapae W chromosome was reversed relative to its orientation in the reference genome.

Fig. 7.

(A) Physical position of all P. mannii W chromosome sequence tags aligning to the Z chromosome. The left graph is based on sequence tags derived from the repeat-masked W (hence uses the same data as in fig. 6, top right). The right graphs represent close-ups of the two regions showing extended Z–W sequence similarity, based on sequence tags extracted from the W chromosome not repeat-masked. In (B), analogous data are shown for chromosome 4.