Low genetic and phenotypic divergence in a contact zone between freshwater and marine sticklebacks: gene flow constrains adaptation

Abstract

Background: Distinct hybrid zones and phenotypic and genomic divergence is often observed between marine and freshwater threespine sticklebacks (Gasterosteus aculeatus). Nevertheless, cases also exist where marine-freshwater divergence is diffuse despite seemingly similar environmental settings. In order to assess what characterizes these highly different outcomes, we focused on the latter kind of system in the Odder River, Denmark. Here, a previous study based on RAD (Restriction site Associated DNA) sequencing found non-significant genome-wide differentiation between marine and freshwater sticklebacks. In the present study, we analyzed samples on a finer geographical scale. We assessed if the system should be regarded as panmictic, or if fine-scale genetic structure and local selection was present but dominated by strong migration. We also asked if specific population components, that is the two sexes and different lateral plate morphs, contributed disproportionally more to dispersal.

Results: We assessed variation at 96 SNPs and the Eda gene that affects lateral plate number, conducted molecular sex identification, and analyzed morphological traits. Genetic differentiation estimated by F_ST was non-significant throughout the system. Nevertheless, spatial autocorrelation analysis suggested fine scale genetic structure with a genetic patch size of 770 m. There was no evidence for sex-biased dispersal, but full-plated individuals showed higher dispersal than low- and partial-plated individuals. The system was dominated by full-plated morphs characteristic of marine sticklebacks, but in the upstream part of the river body shape and frequency of low-plated morphs changed in the direction expected for freshwater sticklebacks. Five markers including Eda were under possible diversifying selection. However, only subtle clinal patterns were observed for traits and markers.

Conclusions: We suggest that gene flow from marine sticklebacks overwhelms adaptation to freshwater conditions, but the short genetic patch size means that the effect of gene flow on the most upstream region must be indirect and occurs over generations. The occurrence of both weak unimodal and strong bimodal hybrid zones within the same species is striking. We suggest environmental and demographic factors that could determine these outcomes, but also highlight the possibility that long-term population history and the presence or absence of genomic incompatibilities could be a contributing factor.

Keywords: Adaptive divergence; Body shape; Eda; Gene flow; Hybrid zone; Sex-biased dispersal; Single nucleotide polymorphisms; Spatial autocorrelation.

Fig. 1

Sampling sites in the Odder River (freshwater) and Norsminde Fjord (marine). The three major groups of samples are as follows: (1) OddUa and OddUb are the two upstream localities in Odder River, pooled into the OddU group for some analyses. OddM denotes the group of samples that were collected continuously in the lower part of the river (as shown in the inserted box). Nor represents marine sticklebacks, sampled at three neighbouring localities and pooled into a single sample

Fig. 2

a The position of the 16 landmarks used. LM1: Anterior tip of the upper lip, LM2: Posterior extent of supraoccipital, LM3: Anterior insertion of first dorsal spine, LM4: Anterior insertion of second dorsal spine, LM5: Anterior insertion of dorsal fin on the dorsal midline (DML), LM6: Posterior insertion of dorsal fin on DML, LM7: Origin of the caudal fin membrane on DML, LM8: Posterior extent of caudal peduncle on lateral midline, LM9: Origin of the caudal fin membrane on ventral midline (VML), LM10: Posterior insertion of anal fin, LM11: Anterior insertion of anal fin, LM12: Caudal tip of the posterior process of the pelvic girdle on the VML, LM13: Posterior tip of the ectocoracoid, LM14: Anterior tip of the ectocoracoid, LM15: Posterior edge of the angular, LM16: Anterior tip of the upper lip (placed same place as LM1). b PCA scatter plot of the first and second, first and third, and second and third principal components that account for most of the variation in body shape. Marine individuals (Nor, red, plus sign), the downstream group (OddM, blue, empty squares) and the upstream group (OddU, pink, full squares) are shown in convex hulls. c Thin-plate spline transformation grid showing the body shape changes between the mean shapes of Nor to OddM (top), mean shapes of Nor to OddU (middle), and means shapes of OddM to OddU (bottom). The colored grids show the color-coded Jacobian expansion factors which measure the degree of local expansion or contraction of the grid: yellow to red for factors >1, indicating expansion; light to dark blue for factors between 0 and 1 indicating contraction

Fig. 3

Corellograms showing the results of spatial autocorrelation analyses for the individuals sampled continuously from the river mouth and upstream. In all plots the shading along lines denotes 95% confidence intervals of r values. The analyses were conducted using GenAlEx 6.5 [61, 102]. a Analysis involving all individuals. The dashed lines denote the upper and lower confidence intervals for the case of no structure (i.e. correlation coeffificent r = 0; not shown in the other plots). b Separate corellograms for full-plated and low + partial-plated individuals. c Separate correlograms for male and females

Fig. 4

Cline analysis of the small scale-data set (from OddM only), showing results for the loci 2620, 11,996, 35,236, 16,548, 32,060, 32,994 and Eda and for the following quantitative traits: numbers of plates, centroid size, PC1, PC2, PC3 (of body shape variation), parasites in trunk and black spot disease. Plots show the maximum-likelihood cline and observed frequency data over the associated fuzzy cline region (95% credible cline region, indicated by grey). The distance indicates the distance to the sea (Nor = 0, see Table 1 for details). For each locality the mean was calculated and plotted