Invasive species represent a significant threat to global biodiversity (Early et al., 2016; Dueñas et al., 2021). They are an important driver of species extinction (Bellard et al., 2017) and can have a strong negative impact on native species abundance (Emery-Butcher et al., 2020). The strength of the impact of invasive species on native populations and communities depends on their abundance and trophic level, with invasive predators typically having the most substantial impact (Bradley et al., 2019). The extent of a biological invasion in a geographic location will be determined by the limits of habitat suitability in the area (Liu et al., 2020). Environmental gradients can be important barriers restricting invasive species’ unchecked advance (Mothes et al., 2019). For example, the range of the invasive Asian clam in North America and Europe is currently limited by its inability to tolerate minimum temperatures lower than -10°C and by altitudes higher than 2000m (Crespo et al., 2015). Thus, impacted native species that can tolerate a broader range of environmental parameters than the invaders may have access to refuge habitats free from invaders (Chapman et al., 2002; Reid et al., 2013). However, across large spatial scales or strong environmental gradients, it is unclear whether this type of ’physiological refugia’ results from universally broad physiological tolerance in native species or local adaptation of populations experiencing distinct environmental conditions. Moreover, invasive predators might impose additional divergent selection on native species because local populations that overlap with the invader could experience selection for anti-predator traits (Strauss et al., 2006; Brookes & Rochette, 2007). Indeed, among natural and anthropogenic stressors, invasive species are one of the strongest drivers of phenotypic change in native populations (Sanderson et al., 2022).
Local adaptation could interact with demographic processes to facilitate or hinder the co-existence of native species with an invasive predator by providing demographic subsidies, genetic rescue, or by introducing maladaptive alleles to recipient habitats. While some sites can have environmental conditions more conducive to invasion and therefore suffer stronger biotic selection from invasion, other sites can have environmental conditions that exclude or reduce the density of invasive predators, thereby acting as uninvaded refuges with lowered ecological impacts from the invaders (Derry et al., 2013; Reid et al., 2013; Astorg et al., 2021; Morissette et al., 2023). In this scenario, refuge populations could also potentially serve as a demographic subsidy of individuals for invaded populations experiencing population decline (Foppen et al., 2000; With et al., 2006), and prevent their extinction through demographic rescue (Hufbauer et al., 2015). Genetic rescue can additionally occur when migrants prevent the extinction of declining populations through increased genetic diversity that reduces inbreeding depression (Carlson et al., 2014; Whiteley et al., 2015; Fitzpatrick et al., 2016); if this genetic variation includes adaptive alleles, then genetic rescue can also lead to an evolutionary rescue, i.e. the avoidance of extinction via adaptation (G. Bell & Gonzalez, 2011; Sexton et al., 2011; Hufbauer et al., 2015).
Conversely, gene flow could instead bring maladapted alleles into invaded populations if populations have experienced strong divergent selection across the environmental gradient (Bolnick & Nosil, 2007). Migration from refuges could consequently pose some risks if recipient and source populations are divergent due to local adaptation, which can cause genetic incompatibilities in hybrids and lead to temporary reductions in fitness (i.e., outbreeding depression, Fenster & Galloway, 2000; Edmands, 2007; Frankham et al., 2011; D. A. Bell et al., 2019). Thus, there could be a tension between demographic rescue (immigrants from refuge populations providing individuals to bolster the shrinking populations in invaded habitat) versus these immigrants potentially being maladapted, thereby reducing the mean fitness of the invaded populations. It is thus important to understand how these two sources of adaptation (to the refuge vs. to the predator) can interact with demographic processes to either facilitate or hinder the ability of a native species to avoid population decline from an invasive predator.
Genomic methods are increasingly used to understand species and populations’ responses to sudden environmental changes induced by anthropogenic activities such as invasive species (Stern & Lee, 2020) and are an important tool for informing conservation (Willi et al., 2022; Bernatchez et al., 2023). They can enable the assessment of population connectivity, investigate demographic and genetic changes, and detect the potential for genetic adaptation (e.g., Marques et al., 2019). Reconstructing demographic changes can help assess potential population declines induced by invasive species. Additionally, inference of gene flow can identify the source and recipient populations in a metapopulation impacted by the invaders. Finally, assessing genetic adaptation can determine if source and recipient populations are divergent because of local adaptation (Cure et al., 2017), thus altering the likelihood of genetic and/or evolutionary rescue from genetically differentiated populations. Hence, knowledge of evolutionary forces, which can be elucidated through genomic tools, is critical for understanding the overall response of native species to the impact of biological invasions.
Gastropods have been widely used to study adaptation in response to predation (Brookes & Rochette, 2007; Hooks & Padilla, 2021), with abiotic factors such as calcium concentration modulating this response through changes in shell morphology and behavior (Rundle et al., 2004; Bukowski & Auld, 2014). As such, they are a useful biological study model for addressing evolutionary responses to biological invasions.Amnicola limosus is a small dominant freshwater gastropod species with a wide geographical distribution in the USA and Canada (www.gbif.org/species/5192461). This gastropod does not have a pelagic larval phase: egg masses are deposited on the substrate, and juveniles move from the substrate to the macro-algal substrate (Pinel-Alloul & Magnin, 1973). Part of the range of A. limosus has been invaded by the round goby (Neogobius melanostomus ), a molluscivorous fish, from the lower Great Lakes and running downstream throughout the Upper St. Lawrence River (Hickey & Fowlie, 2005). Amnicola limosus is commonly found in the stomach contents of round gobies, and following the goby invasion of Lake Saint-Louis, A. limosuspopulations experienced a 0.5-1 mm reduction in shell size (Kipp et al., 2012). Because the mean gape size of the round goby is larger than the maximum size of A. limosus , round gobies do not have to crush the snail, which suggests that shell size reduction is likely to be due to reduced predation pressure on smaller and less visible individuals (round gobies are visual predators; Kipp et al., 2012). A considerable reduction in small gastropod abundance (down to 2-5% of the original population size, with A. limosus being the most abundant species) and species richness in the Upper St. Lawrence River were also reported since the invasion of round gobies in this ecosystem in 2005 (Kipp et al., 2012). However, round gobies cannot tolerate low calcium concentrations (Baldwin et al., 2012; Iacarella & Ricciardi, 2015), and have not invaded the Ottawa River (Ca2+ concentrations below 22 mg/L; Sanderson et al., 2021; Morissette et al., 2023) at its junction with the Upper St. Lawrence River. On the contrary, this low calcium concentration is not a physiological limit for A. limosusembryonic development (> 1.1 mg/L; Shaw & Mackie, 1990), and Pinel-Alloul & Magnin (1973) showed that A. limosus was present in the Ottawa river before the invasion of gobies, indicating that this species can tolerate the calcium concentration found in the Ottawa river. These calcium-poor waters are thus acting as a refuge from goby predation in this system (Astorg et al., 2021; Morissette et al., 2023). Calcium-poor waters could potentially provide demographic subsidies for the native populations at invaded sites (e.g., amphipods; Derry et al., 2013).
This study aims to investigate the potential adaptation of A. limosus to the water calcium gradient and the presence of round goby invasion in the Upper St. Lawrence River, as well as the demographic and genetic consequences of the goby invasion on the native gastropod. We tested for evidence of local adaptation via 1) genome scans for SNPs associated with calcium concentration and round goby presence and 2) a laboratory reciprocal transplant of wild A. limosus individuals to measure survival and fecundity in factorially-crossed treatments of water calcium concentration and round goby chemical cue. For the local adaptation to the low calcium and high goby predation environmental conditions, we expected to find outlier SNPs associated with one or both of these covariables. In the transplant experiment, we also expected populations to show a home versus transplant advantage in life-history traits relative to the calcium concentration and presence of goby cues. For demography, because of the large decrease in the population size of gastropods observed in the Upper St. Lawrence River following round goby invasion (Kipp et al., 2012), we hypothesized that A. limosuspopulations in invaded habitats could have undergone a similar decrease in abundance, perhaps accompanied by a genetic bottleneck. If so, uninvaded populations could potentially provide demographic, genetic, and evolutionary rescue for invaded populations (Hufbauer et al., 2015; Whiteley et al., 2015). However, this possibility would depend on the level of gene flow between habitats and the extent of adaptive differentiation. Because the life history of A. limosus does not involve a pelagic larval phase (Pinel-Alloul & Magnin, 1973), we expected to observe low gene flow in the absence of strong water currents. Low gene flow impedes demographic and genetic rescue, and strong adaptive differentiation will likely hinder evolutionary rescue. As such, we hypothesized that local adaptation to the distinct habitat types might have led to reduced gene flow (isolation by environment; Wang & Bradburd, 2014). Our paper provides a rare empirical study to address how spatial heterogeneity in both abiotic conditions and invasive predator presence can interact with demographic processes to shape the response of a native species to biological invasion. Freshwater environments are deeply impacted by invasive species (Gallardo et al., 2016). It is important to not only consider ecological impacts of invasion but also the evolutionary and demographic responses in native species that can help foster invasive-native species coexistence in invaded freshwater ecosystems.