INTRODUCTION
Identifying mechanisms involved in rapid adaptation to novel environmental conditions is a central theme in evolutionary biology and a pressing concern in the context of global changes characterising the Anthropocene (Malhi, 2017). The vast majority of studies investigating mechanisms involved in rapid adaptation to new environments have focused on phenotypic plasticity on the one hand and on genetic responses to selection on the other hand. At their crossroad, recent work underlines the potential role of epigenetics in rapid adaptation to new environments (Liu, 2013). In particular, environmental variations can induce differences in DNA methylation patterns and hence modulate gene expression and upper-level phenotypes (Duncan, Gluckman, & Dearden, 2014; Jaenisch & Bird, 2003). Such methylation-linked phenotypic variation can occur during an individual’s lifetime, especially early on during the organism’s development (Waterland & Jirtle, 2003; Weaver et al., 2004). Although methylation changes acquired across an individual’s lifetime may often be non-heritable (7 ) but see (8,9 ), epigenetically induced phenotypic shifts may nevertheless enhance individual fitness in new environments. Moreover, during the course of evolution, divergent genetic variants regulating epigenetic modifications may also under selection, hence promoting the evolution of divergent epigenotypes and epigenetically-linked phenotypic variation (Richards, Bossdorf, & Pigliucci, 2010). While epigenetic studies focused on human diseases and medical topics are now abundant, studies in an ecological context are still rare (Derks et al., 2016). A few epigenetic studies in natural populations revealed that DNA methylation shifts might play a determinant role in local adaptation to environmental variation (Foust et al., 2016). There is hence an urgent need for further empirical investigations of simultaneously rapid genetic and epigenetic evolution in response to environmental change (Danchin et al., 2011).
Urbanization rapidly and irreversibly changes natural habitats into human-made environments and is considered a major threat to biodiversity (Brondizio, Settele, & Díaz, 2019). For species who appear to cope with urbanisation, urban habitats present a myriad of novel environmental conditions compared to the habitat where they evolved, including high levels of chemical, light and sound pollution, high proportion of impervious surfaces, high habitat fragmentation, low vegetation cover and high human densities (Grimm et al., 2008; Szulkin, Garroway, Corsini, Kotarba, & Dominoni, 2020). Such extreme environmental changes compared to natural areas are expected to result in numerous novel selection pressures for city-dwelling species (Szulkin, Munshi-South, & Charmantier, 2020). Accordingly, rates of recent phenotypic change, concerning multiple types of traits related to behaviour, morphology, phenology and physiology, were found greater in urban areas than in any other habitat types, including non-urban anthropogenic contexts (Alberti et al., 2017; K. A. Thompson, Rieseberg, & Schluter, 2018). The exploration of the molecular mechanisms implicated in urban-driven phenotypic changes has only begun, with both genetic (Mueller, Partecke, Hatchwell, Gaston, & Evans, 2013; Perrier et al., 2018; Salmón et al., 2020), and epigenetic investigations (McNew et al., 2017; Riyahi, Sánchez-Delgado, Calafell, Monk, & Senar, 2015; Watson, Powell, Salmón, Jacobs, & Isaksson, 2021). For instance, DNA methylation variations have been associated in vertebrates with high levels of traffic-related air pollution (Ding et al., 2017). Yet, epigenetic studies have been performed at relatively small genomic resolution. In addition, very little is known about the level of parallelism and hence of the predictability of genetic and epigenetic evolution in response to urbanisation in distinct cities (Rivkin et al., 2018; Santangelo, Rivkin, & Johnson, 2018). So far, a small number of studies have provided evidence for a range of situations: from local adaptation despite strong gene flow (e.g. in the red-tailed bumblebee Bombus lapidaries Theodorou et al., 2018) to restricted gene flow and independent colonization in different cities by a few founders, followed by adaptation (e.g. in the burrowing owl Athene cunicularia,Jakob C. Mueller et al., 2018). Providentially, recent genomic tools of high resolution and the multitude of cities around the globe offer unique opportunities to compare simultaneously individuals’ genomic and epigenomic responses in several cities and thereby study the parallelism and predictability in molecular mechanisms implicated in rapid adaptation to urbanization (Perrier, Caizergues, & Charmantier, 2020; Santangelo et al., 2020).
In this study, we used both genome-wide and epigenome-wide sequencing approaches to compare genetic and epigenetic responses among three pairs of great tit Parus major populations in urban and forest habitats. At the European level, population monitoring of Great tits revealed parallel phenotypic shifts in city birds compared to their forest conspecifics, with in particular smaller and lighter urban birds laying earlier and smaller clutches (Biard et al., 2017; Caizergues et al., 2021; Chamberlain, Hatchwell, & Gaston, 2009; Corsini et al., 2020). In addition, genomic analyses showed patterns of genome-wide differentiation between urban and forest birds (Perrier et al., 2018) while a large scale analysis revealed some parallel footprints of adaptation to urbanization across nine European cities (Salmón et al., 2020). At the epigenetic level, a preliminary Great tit study recently described methylation shifts associated with urbanization (Watson et al., 2021). However, this analysis focused on a single location (n = 6 urban and 6 forest males) and hence could not test for a potential parallelism in the urban-related epigenomic response. Moreover, the search for DMSs (Differentially Methylated Sites) rather than DMRs (differentially methylated regions) in this recent study most probably weakened the power to detect significantly differentially methylated regions of the genome. In order to advance our understanding of the genome-wide and epigenome-wide responses to urbanization, and its putative spatial parallelism, we here searched for genomic footprints of divergent selection and for DMRs between three pairs of urban-forest populations across Europe. Our results show that despite limited genetic differentiation and few genomic footprints of divergent selection between forest and urban populations, urban life was associated with numerous differentially methylated regions notably associated with neural development, behaviour and immunity. Hence, this study shows the potential role of an epigenetic response in rapid adaptation to urbanization. Importantly, we found little parallelism between cities in both the genomic and the epigenomic responses to urbanization, possibly confirming the hypothesis that multiple evolutionary ways exist to independently cope with similar novel environmental conditions.