INTRODUCTION
Local adaptation of populations to low-quality habitats, such as species range edges, is crucial for within-species genetic diversity and range expansion/contraction dynamics, with consequences for long-term species persistence (Wright 1943; Pulliam 1988; Venail et al. 2008; Edelaar & Bolnick 2012). The typically asymmetric migration between core and edge habitats (source-sink dynamics) can play a crucial role in local adaptation to the latter. Immigrants into the edge populations may fuel local adaptation by supplementing genetic variation (Holt et al. 2004; Perron et al. 2007), or retard it by competing with the locally adapted genotypes or disrupting locally adaptive alleles via recombination (Figure 1a, 1b) (García-Ramos & Kirkpatrick 1997; Fedorkaet al. 2012; Eriksson & Rafajlović 2021). Theory suggests that the net effect of immigration on local adaptation will depend on a number of factors including migration rate; and both positive and negative effects have been reported in previous empirical studies (Perron et al. 2007, 2010; Tigano & Friesen 2016; Mirrahimi & Gandon 2020).
Coevolving natural enemy species have recently been recognized as important players for population diversification and species biogeography (Engelkes et al. 2008; Ricklefs 2010; Ricklefs & Jenkins 2011; Betts et al. 2018). Here we investigate how the presence of enemy species may limit abiotic adaptation in edge habitats with and without source-sink dynamics. In an isolated edge habitat, enemy species can reduce victim population sizes and thus the supply of genetic variation and the efficiency of natural selection (Hudsonet al. 1998; Bohannan & Lenski 2000); and may also drive the evolution of defenses that trade-off with growth traits underlying abiotic adaptation (Figure 1c) (Kraaijeveld & Godfray 1997; Webster & Woolhouse 1999; Brockhurst et al. 2004; Agrawal et al.2010).
When the enemy species are far more mobile than the victim (as in many predator-prey and plant-herbivore systems), immigration of the enemy individuals from the core habitats would further reduce victim population size, particularly when the core habitats also function as evolutionary hotspots that promote enemy-victim arms race coevolution (Figure 1d) (Hochberg & Van Baalen 1998; Lopez Pascua et al.2012; Gorter et al. 2016). This negative effect of the source populations on the sink populations is an example of enemy-mediated intraspecific apparent competition (Holt 1977; Morris et al.2004; Ricklefs 2010; Allen et al. 2018). Enemy migration may also coincide with victim migration, particularly in host-parasite systems. Here, the net effect for abiotic local adaptation is less predictable: Immigration of victim species itself may promote local adaptation by increasing genetic variation, but it is equally possible that the co-occurring enemy-victim migration leads to repeated sweeps of sink populations by immigrants and thus prevents local adaptation (Figure 1e) (Zhang & Buckling 2016; Poulin & de Angeli Dutra 2021). Some more complex but less realistic scenarios are not considered here, e.g. enemy species being present in the core, but not edge habitats; or enemy species having much lower migration rate than victim species.
Here we experimentally test the above hypotheses using a model bacterium-phage system, Pseudomonas fluorescens SBW25 and its lytic phage SBW25Φ2 (Buckling & Rainey 2002). This is a host-parasitoid system as the phages both develop within the bacterial cells and kill the bacterial cells after replication (Lenski 1984; Buckling & Rainey 2002; Forde et al. 2004). Therefore, the bacterium and the phage are victim and enemy species, respectively. Due to the experimental amenability of this system, we are able to study all the five scenarios illustrated above (Figure 1). We considered a low temperature environment as an edge habitat for the bacterium and an optimal temperature its core habitat.