INTRODUCTION
Cultivation of plants beyond their native ranges has a long history (Hobhouse, 1992; Huxley, 1978; Reichard & White, 2001). Although we often associate this anthropogenic practice with our staple crops, hundreds of plants are currently cultivated. While cultivation provides valuable resources to humans, it can have detrimental effects on the native flora, especially when alien taxa escape their cultivated lifestyle and re-establish themselves in the wild within or outside their native ranges (Todesco et al., 2016; Quilodrán et al., 2020b). Naturalized taxa can for example be highly invasive and outcompete native species (Anderson, Galatowitsch, & Gomez, 2006; Downey & Richardson, 2016). In addition, naturalizations can also increase the opportunities for hybridization and inter-specific gene flow when previously isolated taxa are brought together (Owen et al., 2020).
Hybridization and gene flow are major drivers of speciation and trait evolution in plants (Anderson & Stebbins, 1954; Ellstrand & Schierenbeck, 2000; Morjan & Rieseberg, 2004; Otto & Whitton, 2000; Rieseberg, 1995; Rieseberg & Burke, 2001; Schierenbeck & Ellstrand, 2009; Soltis et al., 2009; Soltis, Marchant, Van de Peer, & Solitis, 2015; Stebbins, 1959; Yakimowski & Rieseberg, 2014). Gene flow between divergent lineages can increase genetic diversity and establish novel stable hybrids that are reproductively isolated from their parental species, by for example allopolyploidization or clonal propagation, and thus can evolve into new species (Ainouche et al., 2009; Ellstrand & Schierenbeck, 2000; Grant, 1981; Mallet, 2007; Morjan & Rieseberg, 2004; Rieseberg & Burke, 2001; Schierenbeck & Ellstrand, 2009; Todesco et al., 2016; Quilodrán et al., 2020a). However, if the isolating barriers between hybrids and parental taxa are weak, continued gene flow between them can result in homogenization and ultimately genetic swamping with the formation of coalescent complexes (hybrid swarms; Beninde, Feldmeier, Veith, & Hochkirch, 2018; Ellstrand & Schierenbeck, 2000; Pinto et al., 2005; Todesco et al., 2016; Quilodrán et al., 2020a). Hybridization can therefore promote the loss of species by merging previously separated taxa (Owens & Samuk, 2020; Todesco et al., 2016; Quilodrán et al., 2020a). Here we explore the genomic and phenotypic consequences of frequent inter-specific gene flow following cultivation and naturalization using mints (Mentha ) as our study system.
Mints are aromatic herbs with cosmopolitan distributions. Several mints are widely cultivated in many parts of the world for culinary use and for extraction of essential oils used as flavoring agents (Gobert, Moja, Colson, & Taberlet, 2002; Singh & Pandey, 2018; Tucker, 2012; Vining et al., 2020). Although Mentha is a relatively small genus, it is highly diverse due to inter-fertility between many taxa and numerous hybrid species are recognized (Gobert et al., 2002; Tucker et al., 1980; Tucker & Naczi, 2007). Most mint hybrids are sterile but can successfully propagate clonally using rhizomes (Harley & Brighton, 1977; Gobert et al., 2002). Exceptions to this are hybrids between the two diploid species M. longifolia (2n=24) and M. suaveolens (2n=24; Chambers & Hummer, 1994; Harley & Brighton, 1977; Sobti, 1965). In Europe, where these two taxa are sympatric, they can spontaneously hybridize forming the partly fertile hybrid complexM. × rotundifolia (Harley & Brighton, 1977). In addition, the widely cultivated M. spicata (spearmint; Singh & Pandey, 2018; Vining et al., 2020) is often described as an allopolyploid (2n=48) formed from hybridization of M. longifolia and M. suaveolens (Chambers & Hummer, 1994; Gobert et al., 2002; Harley & Brighton, 1977; Sobti, 1965; Tucker et al., 1980; Tucker & Naczi, 2007). However, the allopolyploid status of M. spicata has been questioned (Heylen, Debortoli, Marescaux, & Olofsson, 2021) and chromosome counts range between 36 and 72 (Ahmad, Tyagi, Raghuvanshi, & Bahl, 1992; Chambers & Hummer, 1994; Sobti, 1965). Similarly, there are reports of polyploid chromosome counts (2n=36/48) for M. longifolia (Chambers & Hummer, 1994; Harley & Brighton, 1977; Sobti, 1965).
Mentha spicata and its hybrid with M. aquatica , i.e.M. × piperita (peppermint), are the most widely cultivated mints (Salehi et al., 2018). The exact origin of M. spicata is unknown, but some studies suggest that it was formed in cultivation (Harley & Brighton, 1977; Tucker, 2012). Although M. spicata is considered native to most of continental Europe, the Middle East, and southern Asia, it has also spread widely from cultivated plants obscuring its natural range and population size (Gobert et al., 2002; Harley & Brighton, 1977; Vining et al., 2020). Outside of the presumably native range, M. spicata is currently naturalized in most geographic regions with a warm-temperate climate overlapping, and exceeding, that of the two hypothesized parental taxa (Harley & Brighton, 1977; Vining et al., 2020). Despite the reported differences in ploidy, M. spicata frequently forms hybrids with both M. longifolia and M. suaveolens but the offspring (M. ×villosa-nervata and M. × villosa , respectively) are believed to be sterile due to their theoretical triploid status (Chambers & Hummer, 1994; Harley & Brighton, 1977; Gobert et al., 2002). The re-establishment of cultivated M. spicata in its native range creates an excellent opportunity to study the phenotypic and genotypic consequences of increased opportunities for hybridization.
Here we take advantage of the biodiversity stored in herbarium collections that allows for evaluations of specimens collected over large geographic regions (Bieker & Martin, 2018). We first evaluate the morphological spaces of the cultivated M. spicata , the native species (M. longifolia and M. suaveolens ), and their hybrids (M. × rotundifolia ). We then use whole genome sequencing to (1) establish patterns of genetic admixture and gene flow between taxa, (2) infer the genomic histories of M. xrotundifolia and M. spicata , and (3) test the hypothesis that increased opportunities for inter-specific gene flow following the naturalization of the cultivated M. spicata has laid the foundation for the formation of a coalescent complex.