Population genetic insights into establishment, adaptation, and dispersal of the invasive quagga mussel across perialpine lakes

Abstract

Human activities have facilitated the invasion of freshwater ecosystems by various organisms. Especially, invasive bivalves such as the quagga mussels, Dreissena bugensis, have the potential to alter ecosystem function as they heavily affect the food web. Quagga mussels occur in high abundance, have a high filtration rate, quickly spread within and between waterbodies via pelagic larvae, and colonize various substrates. They have invaded various waterbodies across the Northern Hemisphere. In Central Europe, they have invaded multiple large and deep perialpine lakes with first recordings in Lake Geneva in 2015 and 2016 in Lake Constance. In the deep perialpine lakes, quagga mussels quickly colonized the littoral zone but are also abundant deeper (>80 m), where they are often thinner and brighter shelled. We analysed 675 quagga mussels using ddRAD sequencing to gain in-depth insights into the genetic population structure of quagga mussels across Central European lakes and across various sites and depth habitats in Lake Constance. We revealed substantial genetic differentiation amongst quagga mussel populations from three unconnected lakes, and all populations showed high genetic diversity and effective population size. In Lake Constance, we detected no genetic differentiation amongst quagga mussels sampled across different sites and depth habitats. We also did not identify any convincing candidate loci evidential for adaptation along a depth gradient and a transplant experiment showed no indications of local adaptation to living in the deep based on investigating growth and survival. Hence, the shallow-water and the deep-water morphotypes seem to be a result of phenotypic plasticity rather than local adaptation to depth. In conclusion, our ddRAD approach revealed insight into the establishment of genetically distinct quagga mussel populations in three perialpine lakes and suggests that phenotypic plasticity and life history traits (broadcast spawner with high fecundity and dispersing pelagic larvae) facilitate the fast spread and colonization of various depth habitats by the quagga mussel.

Keywords: Dreissena; ddRADseq; dispersal; phenotypic plasticity; population genetics.

FIGURE 1

Map of quagga mussel sampling sites across Switzerland including the far German sampling site and outlining the fine‐scale sampling approach in Lake Constance (dark blue). At each site across Switzerland (1–9), mussels were collected by snorkelling. At sites of the fine‐scale sampling in Lake Constance (a–j), two transects were sampled (~500 m apart). Each transect sampled the following depths: 1, 30, 60, and 80 m. In the shallower part in Lower Lake Constance (sites i, j, and k), only 1, 10 m, and max. depth were sampled. Sampling sites: (1) Lower Lake Constance: Radolfzell; (2) Lower Lake Constance: Reichenau; (3) Upper Lake Constance: Uttwil; (4) Upper Lake Constance: Altenrhein; (5) River Rhine: Wallbach; (6) Lake Neuchâtel: St. Aubin; (7) Lake Neuchâtel: Grandson; (8) Lake Geneva: Rivaz; (9) Lake Geneva: St. Prex; (10) Germany: Löebejün. Sites within Lake Constance (a) Altenrhein, (b) Bregenz, (c) Langenargen, (d) Fischbach, (e) Uttwil, (f) Konstanz, (g) Sipplingen, (h) Wallhausen, (i) Reichenau, (j) Radolfzell, (k) Berlingen. Further details about Lake Constance sites in Table S1.

FIGURE 2

Genetic differentiation between quagga mussels from Lake Geneva, Lake Neuchâtel, and the Constance regions (Upper and Lower Lake Constance, Rhine, and Germany). (a) PCA plot showing the genetic differences between 127 quagga mussels based on 81,197 loci. Populations are coloured differently (see legend). Quagga mussels from Upper and Lower Lake Constance are not distinct from River Rhine mussels, while Lake Geneva and Neuchâtel populations form separate groups. (b) Quagga mussels' genetic composition (admixture proportions estimated by LEA). Lake Constance and River Rhine populations show similar ancestry proportions. Each bar represents an individual (ordered by lakes), and the colours represent the proportion of each of the four genetic contributions.

FIGURE 3

No genetic differentiation between quagga mussels sampled across depth gradients in Lake Constance. (a) PCA plot showing the genetic similarities between 549 quagga mussels based on 4939 loci. Colours indicate different collection depths. (b) Quagga mussels have similar admixture over all depths (admixture proportions estimated by LEA). Each bar (ordered by sampling depth) represents an individual, and the three colours represent the proportion of each of the three genetic contributions.

FIGURE 4

Genome scan detecting three candidate loci detected potentially shaped by selection along a depth gradient. F _st values of loci along the chromosome as estimated by outflank. Detected candidate loci are marked in pink and represent the highest F _st values.

FIGURE 5

Reciprocal transplant experiment revealed no evidence for local adaptation at different depth. (a) Predicted survival probability of quagga mussels in the pelagic in 60 m depth in Lake Constance (green) and Lake Geneva (purple). Dashed lines show the survival of transplanted quagga mussels from a shallow environment (10 m), and the solid line shows the survival of mussels collected in 60 m in the depth of destination of 60 m. All lines show fitted values of AFT survival models (Weibull regressions). (b) Mussel size changes as shell size differences (in mm, mean ± 95% confidence interval) between start and end of the reciprocal transplant experiment in Lake Constance (for 98 days in green) and in Lake Geneva (for 104 days in purple).