A region of the sex chromosome associated with population differences in diapause induction contains highly divergent alleles at clock genes

Abstract

Developmental plasticity describes the capacity of individuals with the same genotype to induce permanent change in a phenotype depending on a specific external input. One well-studied example of adaptive developmental plasticity is the induction of facultative diapause in insects. Studies investigating the inheritance of diapause induction have suggested diverse genetic origins. However, only few studies have performed genome-wide scans to identify genes affecting the induction decision. Here we compare two populations of the butterfly Pieris napi that differ in the propensity to enter diapause, and despite showing a low genome-wide divergence, we identify a few genomic regions that show high divergence between populations. We then identified a single genomic region associated with diapause induction by genotyping diapausing and directly developing siblings from backcrosses of these populations. This region is located on the Z chromosome and contained three circadian clock genes, cycle, clock, and period. Additionally, period harbored the largest number of SNPs showing complete fixation between populations. We conclude that the heritable basis of between-population variation in the plasticity that determines diapause induction resides on the Z chromosome, with the period gene being the prime candidate for the genetic basis of adaptive plasticity.

Keywords: Crosses; Pieris napi; diapause induction; genes; local adaptation.

Figure 1

Genome‐wide divergence measured as FST in 50 kb nonoverlapping windows between Barcelona and Abisko. The solid horizontal line indicates the 95th percentile at FST = 0.087. The 25 chromosomes are indicated by number, and modScaffolds are indicated by a dash on the X‐axis. Chromosome 1 is the Z chromosome.

Figure 2

(A) Genome‐wide divergence measured as FST in 50 kb nonoverlapping windows between direct development and diapause pools of five backcrosses. (B) Venn diagram of the overlap of the most divergent regions between crosses. Of the total of 253 unique regions, 180 (71%) were shared between all five families, and 213 (84%) were common to at least four families.

Figure 3

QTL results and genome‐wide divergence in 50 kb windows for (A) Chromosome 1, (B) modScaffold_17_1, and (C) modScaffold_95_1. The black line indicates population differentiation (FST) as measured in 50 kb windows. The colored dots represent the FST values of the comparisons between direct and diapausing individuals in the five crosses: family 106 is blue, 110 is red, 115 is green, 128 is yellow, and 135 is orange, and the corresponding lines are smoothed means and their standard deviations. The gray dots on the X‐axis represent the 180 outlier windows shared between the five crosses, and red dots represent the 46 outlier windows shared between comparisons of phenotypes within the crosses and the population comparisons. Black vertical bars below the X‐axis represent gene models present on the scaffolds. (D) Synteny plot showing orthology between P. napi Chromosome 1, modScaffolds 17_1 and 95_1, and Zerene cesonia Z chromosome.