Repeated evolution of photoperiodic plasticity by different genetic architectures during recurrent colonizations in a butterfly

Abstract

In cases of recurrent colonizations of similar habitats from the same base population, it is commonly expected that repeated phenotypic adaptation is caused by parallel changes in genetic variation. However, it is becoming increasingly clear that similar phenotypic variation may also evolve by alternative genetic pathways. Here, we explore the repeated evolution of photoperiodic plasticity for diapause induction across Swedish populations of the speckled wood butterfly, Pararge aegeria. This species has colonized Scandinavia at least twice, and population genomic results show that one of the candidate regions associated with spatial variation in photoperiodism is situated on the Z-chromosome. Here, we assay hybrid crosses between several populations that differ in photoperiodic plasticity for sex-linked inheritance of the photoperiodic reaction norm. We find that while a cross between more distantly related populations from the two different colonization events shows strong sex-dependent inheritance of photoperiodic plasticity, a cross between two more closely related populations within the oldest colonization range shows no such effect. We conclude that the genotype-phenotype map for photoperiodic plasticity varies across these populations and that similar local phenotypic adaptation has evolved during recurrent colonization events by partly non-parallel genetic changes.

Keywords: Z-linkage; colonization; diapause; genotype-phenotype map; parallel evolution; plasticity.

Figure 1.

(a) Map of Sweden highlighting the sample sites. The green points are where adult P. aegeria have been observed from January 2017 to May 2023 according to citizen science data (Artportalen.se). The orange points are the sample sites that have been used for population crosses in the present study (Stockholm and Gotland) while the purple points are the sample sites from Pruisscher et al.'s 2018 study [42] (Sundsvall and Skåne). The grey bubble covering southern Sweden and parts of Denmark indicates populations from the most recent colonization event. (b) A Treemix plot showing the evolutionary relationship of seven populations of P. aegeria in Sweden based on population genomic data where branch length indicates amount of genetic drift (based on figure 1b in [62]).

Figure 2.

Illustration of expected results when analysing diapause proportions of reciprocal crosses with (a) no sex-linkage or (b) sex-linkage explaining part of the difference in the diapause response between the two original populations. The F1 male reciprocal crosses are expected to have a reaction norm that is intermediate to the two original populations. The same is expected of the reciprocal female crosses if there is no sex-linkage. If the reciprocal female crosses differ, it would indicate that genetic variation at the Z-chromosome influences the difference in the critical photoperiod of the two original populations.

Figure 3.

Proportion of diapausing individuals at the photoperiods 17, 17.4, 18.2, 18.2 and 18.6 h (left to right). All cabinets had a constant temperature of 18°C. Bars show 95% CI of actual proportions and functions show model-predicted responses. (a) Larval diapause response of the two original populations (G × G, Gotland, males and females, dark blue; S × S, Stockholm, males and females, dark green) and the reciprocal hybrid crosses pooled (G × S + S × G, males and females, black solid line). Means are depicted as rectangles based on 9–16 individuals for S × S and G × G and 75–80 individuals for G × S + S × G. (b) Same results but for the pupal decision. Means are based on 9–16 individuals for S × S and G × G and 75–80 individuals for G × S + S × G. (c) Larval diapause response for the two reciprocal crosses (G × S, father from Stockholm, light green; S × G, father from Gotland, light blue). Males have solid lines with means depicted as triangles, while females have stippled lines with means depicted as dots. Means are based on 11–19 individuals for G × S females, 18–21 individuals for G × S males, 15–21 individuals for S × G females and 21–26 individuals for S × G males. (d) Same results but for pupal decision. Means are based on 11–19 individuals for G × S females, 18–21 individuals for G × S males, 15–21 individuals for S × G females and 20–26 individuals for S × G males.

Figure 4.

Proportion of diapausing individuals in reciprocal hybrid crosses from the present (orange) and the study by Pruisscher et al. [42] (purple). Within the studies, individuals are classified by father’s origin (Stockholm = north, Gotland = south; Sundsvall = north, Skåne = south) and by sex (females, stippled lines and dots; males, solid lines and triangles). Bars show 95% CI of actual proportions. (a) Larval decision. Means are based on 81 individuals for G × S females, 98 for G × S males, 96 for S × G females, 115 for S × G males, 38 for Sk × Su females, 48 for Sk × Su males, 48 for Su × Sk females and 37 for Su × Sk males. (b) Pupal decision. Means are based on 81 individuals for G × S females, 98 for G × S males, 96 for S × G females, 112 for S × G males, 38 for Sk × Su females, 48 for Sk × Su males, 47 for Su × Sk females and 36 for Su x Sk males.