Local selection and adaptation
The A. viridiflora complex was already present in Asia by 1.19 Mya is evident from the age model of Fior et al. in accordance with molecular clock estimates of the flora of Altai (A. glandulosa ,A. sibirica , and A. viridiflora ) (Fior et al., 2013). It is clear that present lineages must have diverged much more recently through our demographic history simulations. The model indicates that a very young divergence time in the A. viridiflora complex: c. 239 Kya, c. 211 Kya, and c. 168 Kya (Figure 3B). Although these ages must be explained with caution, the most important climate transition occurred in the Middle Pleistocene, which might have resulted in lineage divergence by changing the suitable habitat of species. The role of Middle Pleistocene climate transitions in speciation processes has long been advised, such as the Ranunculus auricomus complex (Tomasello, Karbstein, Hodač, Paetzold, & Hörandl, 2020), Populus rotundifolia (J. L. Li et al., 2021) and Cerapanorpa brevicornis(Gao, Hua, Xing, & Hua, 2022). Here, we add to these “species on the speciation way” examples in the recent Middle Pleistocene speciation and indicate that the very high rates of speciation associated with Aquilegia adaptive radiation might be driven by the Middle Pleistocene climate transition. Based on our observation, the CN lineage representing A. hebeica and the NW lineage representing A. viridiflora , which have experienced heterogeneous environments and different environmental variables, exert differential selection pressure on the different lineages. We found that 83 genes related to environmental factors that may play a key role in the continuously adaptive process (Figure 4C). Some genes associated with the abscisic acid (ABA) signaling pathway (AIN1 (Dong et al., 2021) and AAO4 (Seo et al., 2004)) can regulate numerous ABA responses and may induce the abiotic stress responses for defense in different environments.
Apart from the genetic divergence revealed by phylogenetic and population structure analyses, clear differentiation in phenotypic traits was also exhibited based on common garden data (Figure 1). Divergence in corolla diameter, petal length, spur length, pistil length, inflorescence number and leaf area was probably driven by selection. Among these traits, floral characteristic differentiation might act as a prezygotic isolation mechanism between the four lineages. Additionally, we found 487 genes under positive selection and exhibiting high divergence between lineages (Figure 4A). These genes were significantly enriched on ABC-2 type transport family protein, a gene family involved in a wide range of metabolism in plants and playing import roles in seed germination, stomatal movement, lateral root development and responses to various environmental stresses (Liu, Li, & Liu, 2013; Matsuda et al., 2012; X. D. Zhang, Zhao, & Yang, 2018). Among these selected genes, KNU (Bollier et al., 2018) andCKX5 (Bartrina et al., 2011) are involved in the regulation of flower morphology and development, of which a major difference exists between the CN and NW lineages. Similarly, GRF2 (Beltramino et al., 2018) and PHT4;2 (Irigoyen et al., 2011) initiate fixation in those lineages due to their relation to leaf development and morphology. NRPD1B is related to panicle branches (L. Xu et al., 2020), and high allelic divergence and fixation of this gene in CN lineages may contribute to producing more inflorescence relative to NW lineages. Moreover, several genes (e.g., PAD2 (Parisy et al., 2007), DRB3 (Mehdi et al., 2022) and EDM2 (Eulgem et al., 2007)) are involved in disease resistance in response to biotic stresses. Plants can adjust growth and defense based on different environments to survive and reproduce in the natural world, which might contribute to lineage divergence (He, Webster, & He, 2022). Therefore, our study illuminated that geographic isolation and local selection drove the lineage divergence of the A. viridiflora complex and created geographic distributions of phenotypic variations. Further work is needed to acquire more accurate functions of genes in the whole genome under the selective pressure and clarify potential adaptation patterns.