Genetic variation in populations of the earthworm, Lumbricus rubellus, across contaminated mine sites

Abstract

Background: Populations of the earthworm, Lumbricus rubellus, are commonly found across highly contaminated former mine sites and are considered to have under-gone selection for mitigating metal toxicity. Comparison of adapted populations with those found on less contaminated soils can provide insights into ecological processes that demonstrate the long-term effects of soil contamination. Contemporary sequencing methods allow for portrayal of demographic inferences and highlight genetic variation indicative of selection at specific genes. Furthermore, the occurrence of L. rubellus lineages across the UK allows for inferences of mechanisms associated with drivers of speciation and local adaptation.

Results: Using RADseq, we were able to define population structure between the two lineages through the use of draft genomes for each, demonstrating an absence of admixture between lineages and that populations over extensive geographic distances form discrete populations. Between the two British lineages, we were able to provide evidence for selection near to genes associated with epigenetic and morphological functions, as well as near a gene encoding a pheromone. Earthworms inhabiting highly contaminated soils bare close genomic resemblance to those from proximal control soils. We were able to define a number of SNPs that largely segregate populations and are indicative of genes that are likely under selection for managing metal toxicity. This includes calcium and phosphate-handling mechanisms linked to lead and arsenic contaminants, respectively, while we also observed evidence for glutathione-related mechanisms, including metallothionein, across multiple populations. Population genomic end points demonstrate no consistent reduction in nucleotide diversity, or increase in inbreeding coefficient, relative to history of exposure.

Conclusions: Though we can clearly define lineage membership using genomic markers, as well as population structure between geographic localities, it is difficult to resolve markers that segregate entirely between populations in response to soil metal concentrations. This may represent a highly variable series of traits in response to the heterogenous nature of the soil environment, but ultimately demonstrates the maintenance of lineage-specific genetic variation among local populations. L. rubellus appears to provide an exemplary system for exploring drivers for speciation, with a continuum of lineages coexisting across continental Europe, while distinct lineages exist in isolation throughout the UK.

Keywords: Adaptation; Arsenic; Earthworms; Ecotoxicology; Lead; Population genomics; RADseq.

Fig. 1

Locations across the UK where L. rubellus were collected (a) and a PCA of log metal concentrations demonstrating comparable site soil characteristics for each of the sites (b). Chemical symbols are used to refer to specific metals and the amount of variation explained by PC1 and PC2 is 44.4% and 24.6%, respectively. Site name abbreviations are: Carrock Fell (CF), Cwmystwyth (CWM) and Devon Great Consols (DGC)

Fig. 2

Principle component analyses of RADseq data for populations of L. rubellus collected across the UK, representing all individuals (a, n = 128), those from lineage A (b, n = 66) and Lineage B (c, n = 62). Triangles represent individuals originating from former mine sites, while circles represent those from nearby control sites. The amount of variance explained by each component is noted on their respective axes

Fig. 3

Structure results from RADseq data for L. rubellus collected from former mine sites and nearby control sites across the UK. Each bar represents an individual, with the proportion of colours reflecting affiliation with differentiated populations. a represents all individuals sampled (n = 128), where K = 2, and highlights the distinctions between lineages A and B, coloured red and blue, respectively. b and c demonstrate the best supported clustering of individuals belonging to lineage A (n = 66, K = 3) and B (n = 62, K = 2), respectively

Fig. 4

Lineage-specific summary genetic statistics in populations of L. rubellus from control and mine sites, showing inbreeding coefficient (F_IS, Top) and nucleotide diversity (Pi, Bottom). Error bars represent standard error. No marker for F_IS was calculated at “B CFM”, as this represents a single individual