Introduction

Fagus L. (Fagaceae) is a small but ecologically and economically important genus of about ten monoecious tree species occurring in three isolated regions of the Northern Hemisphere: East Asia, western Eurasia, and (eastern) North America (Shen, 1992; Peters, 1997; Denk, 2003; Fang & Lechowicz, 2006). The genus is monophyletic and might have originated at high latitudes during the Paleocene (western Greenland, northeast Asia; Fradkina et al., 2005; Denk & Grimm, 2009; Grímsson et al., 2016). It is currently subdivided into two informal subgenera corresponding to reciprocally monophyletic lineages (Shen, 1992; Denk et al., 2005). “Subgenus Engleriana” currently includes three (North-)East Asian species:Fagus engleriana Seemen ex Diels, F. multinervis Nakai, and F. japonica Maxim. “Subgenus Fagus” includes five or more Eurasian species: Fagus sylvatica L. s.l. (including F. orientalis Lipsky) in western Eurasia; F. crenata Blume in Japan; F. lucida Rehder & E.H.Wilson, F. longipetiolata Seemen and F. hayatae Palib. in central and southern China and Taiwan, and a single North American species, F. grandifolia Ehrh. (including F.mexicana [Martínez] A.E.Murray). These two lineages diverged by the early Eocene, as inferred from a fossilized birth–death (FBD) model with an extensive set of 53 fossils (Renner et al., 2016), including the oldest unambiguous macrofossil record of the genus (Manchester & Dillhoff, 2004). While the lineage leading to the modern genus is at least 82–81 myrs old (Grímsson et al., 2016), the extant species are the product of ~50 myrs of trans-continental range expansion and phases of fragmentation leading to diversification (Denk, 2004; Denk & Grimm, 2009). These dynamic migration and speciation histories left multifarious morphological and molecular imprints on modern members of the genus.
Intra- and inter-specific phylogenetic relationships within Fagushave been difficult to resolve (Denk et al., 2002, 2005; Renner et al., 2016). In western Eurasia, where beech is the dominant climax species in mid-altitude forests, the number and rank of several taxa is still controversial (e.g. Gömöry et al., 2018). Difficulties arise from a low inter-specific but high intra-specific morphological diversity (e.g. Denk, 1999a, b) and equally complex inter- and intra-specific genetic differentiation in both the nucleome and plastome of the entire genus (Denk et al., 2002; Okaura & Harada, 2002; Grimm et al., 2007; Gömöry & Paule, 2010; Hatziskakis et al., 2009; Lei et al., 2012; Zhang et al., 2013). This is not surprising; ancient and reiterated hybridization events can be assumed for the evolution of Fagus based on all assembled morphological, fossil and molecular data, and with respect to its biogeographic history and ecology (Denk et al., 2005; Denk & Grimm, 2009; Oh et al., 2016; Renner et al., 2016). Accordingly, plastid DNA variation has shown to be not sorted by speciation but rather by geography in all broadly sampled groups of Fagales studied so far (Fagaceae: Simeone et al., 2016; Yan et al., 2019; Nothofagaceae: Acosta & Premoli, 2010; Premoli et al., 2012).
Nuclear ribosomal DNA spacers, organized in multi-copy arrays, have an unsurpassed potential to detect past and recent reticulation events (e.g. Grimm & Denk, 2008, 2010; Pilotti et al., 2009; de Castro et al., 2013 for Platanus ). In two subfamilies of Fagaceae, Quercoideae (oaks, Quercus ) and Fagoideae (Fagus ), the most widely used nuclear marker for intra-generic studies in plants, the ITS region of the 35S rDNA cistron, has shown only limited discrimination capacity while requiring extensive cloning because of substantial intra-genomic, length- and sequence-polymorphism (Denk et al., 2002, 2005; Denk & Grimm, 2010). Cloning and special methodological frameworks to extract useful phylogenetic signals and infer evolutionary patterns (e.g. Göker & Grimm, 2008; Potts et al., 2014) are also required for the shorter but typically more divergent and phylogenetically informative non-transcribed intergenic spacers of the 5S rDNA (see for instance Forest et al., 2005; Denk & Grimm, 2010; Simeone et al., 2018). This marker has never been used for large-scale genetic studies inFagus but it is known to consist of two paralogous loci inF. sylvatica (Ribeiro et al., 2011) which may lead to substantial intra-genomic variation.
The increasing availability of inexpensive High-Throughput Sequencing (HTS) approaches is now boosting new efforts into research questions that were previously considered too expensive, labour-intensive, and/or inefficient (Babik et al., 2009; Glenn, 2011). Amplicon sequencing of loci with high information content (target sequencing) substantially reduces costs, as it makes cloning unnecessary and many individuals can be combined in the same sequencing run (Ekblom & Galindo, 2010). In a recent study, Piredda et al. (2020) analysed 5S-IGS HTS amplicon data generated for Quercus (Fagaceae) and demonstrated the great potential of this technique for inspecting range-wide diversity patterns and for assessing inter- and intra-species relationships. In the present study, we explored the applicability of the same experimental pathway in six geographic samples of the core group of ‘subgenus Fagus’: theF. crenata – F. sylvatica s.l. lineage, a dominant element of temperate mesophytic forests of Eurasia, climaxing in cooler versions ofCfb climates (i.e. warm temperate, fully humid climates with warm summer; Peters, 1997; Grimm & Denk, 2012; Kottek et al., 2016; Peel et al., 2017). Our objectives were to test the utility of 5S-IGS HTS data for delineating beech species, assessing intra- and inter-specific diversity, and gaining deeper insights into an evolutionary history that likely involved complex reticulations, facilitating or obscuring speciation processes during the different phases of beech evolution (Gömöry et al., 2007, 2018; Gömöry & Paule, 2010; Müller et al., 2019).