Moths passing in the night: Phenological and genomic divergences within a forest pest complex

Abstract

Temporal separation of reproductive timing can contribute to species diversification both through allochronic speciation and later reinforcement of species boundaries. Such phenological differences are an enigmatic component of evolutionary divergence between two major forest defoliator species of the spruce budworm complex: Choristoneura fumiferana and C. occidentalis. While these species interbreed freely in laboratory settings, natural hybridization rates have not been reliably quantified due to their indistinguishable morphology. To assess whether temporal isolation is contributing to reproductive isolation, we collected adult individuals throughout their expected zone of sympatry in western Canada at 10-day intervals over two successive years, assigning taxonomic identities using thousands of single nucleotide polymorphisms. We found unexpectedly broad sympatry between C. fumiferana and C. occidentalis biennis and substantial overlap of regional flight periods. However, flight period divergence was much more apparent on a location-by-location basis, highlighting the importance of considering spatial scale in these analyses. Phenological comparisons were further complicated by the biennial life cycle of C. o. biennis, the main subspecies of C. occidentalis in the region, and the occasional occurrence of the annually breeding subspecies C. o. occidentalis. Nonetheless, we demonstrate that biennialism is not a likely contributor to reproductive isolation within the species complex. Overall, interspecific F₁ hybrids comprised 2.9% of sequenced individuals, confirming the genomic distinctiveness of C. fumiferana and C. occidentalis, while also showing incomplete reproductive isolation of lineages. Finally, we used F _ST-based outlier and genotype-environment association analyses to identify several genomic regions under putative divergent selection. These regions were disproportionately located on the Z linkage region of C. fumiferana, and contained genes, particularly antifreeze proteins, that are likely to be associated with overwintering success and diapause. In addition to temporal isolation, we conclude that other mechanisms, including ecologically mediated selection, are contributing to evolutionary divergence within the spruce budworm species complex.

Keywords: allochrony; ecological speciation; periodicity; two‐year cycle spruce budworm.

FIGURE 1

Collection locations (a) and phenology (b) of adult Choristoneura across the study area. Moths were pheromone‐trapped at each location during ~10 days before collection. Histograms show numbers of adult Choristoneura trapped in each collection interval, with traps being emptied on: 1 = 11–15 June, 2 = 21–27 June, 3 = 1–7 July, 4 = 11–17 July, 5 = 21–27 July, 6 = 31 July ‐ 6 August, 7 = 10–17 August, 8 = 20–24 August, 9 = 30 August ‐ 2 September (Tables S2 and S3 show days since 1 January for each location). Collection intervals without number digits from 2017 indicate traps were not deployed, while numbered collection intervals without histogram bars indicate no moths were trapped. Inset map of North America shows rectangular study area in Alberta and British Columbia with C. fumiferana (green) and C. occidentalis (blue) ranges (adapted from Dupuis et al., 2017). Map was prepared in QGIS version 3.6.0 using base maps obtained from https://github.com/stamen/terrain‐classic and https://github.com/CartoDB/basemap‐styles

FIGURE 2

Principal component analysis, structure barplot, and NewHybrids barplot of SNPs for 260 genotyped Choristoneura. Individuals with <90% assignment to either cluster in the structure analysis are labeled “F₁ hybrids” based on NewHybrids assignment

FIGURE 3

Boxplots for median collection dates (a) and accumulated degree‐days (b) for adult Choristoneura fumiferana and C. o. biennis in region of overlap (locations d–j) during 2017 and 2018. Both had significant differences in maximum flight abundance (Wilcoxon–Mann–Whitney tests: collection date: U = 5151, p < 0.001, r = 0.67; accumulated degree‐days: U = 1683, p < 0.001, r = 0.42). Error bars depict standard error

FIGURE 4

Pooled numbers of genotyped adult Choristoneura fumiferana and C. occidentalis from zone of sympatry (locations d–j). Histograms show numbers of adult Choristoneura genotyped from each collection interval, with traps being emptied in the date ranges listed for Figure 1 (Tables S2 and S3). Traps were not deployed for 2017 collection intervals 1 and 9

FIGURE 5

Number of genotyped adult Choristoneura from each collection location. Moths were pheromone‐trapped at each location during ~10 days before collection. Histograms show numbers of adult Choristoneura trapped during each collection interval, with traps being emptied in the date ranges listed for Figure 1 (Tables S2 and S3). Collection intervals without number digits from 2017 indicate traps were not deployed, while numbered collection intervals without histogram bars indicate no moths were trapped