- Top of page
- Materials and Methods
- Supporting Information
The grass family (Poaceae) consists of c. 10 000 species, most of which belong to two major clades: BEP (Bambusoideae, Ehrhartoideae and Pooideae) (Soreng et al., 2000) and PACCMAD (Panicoideae, Arundinoideae, Centothecoideae, Chloridoideae, Micrairoideae, Aristidoideae and Danthonioideae) (Gabriel Sánchez-Ken et al., 2007). Poaceae evolved > 70 million yr ago in an ecosystem characterized by a warm climate (Grass Phylogeny Working Group, 2001; Edwards & Smith, 2010; Strömberg, 2011), but have successively diversified outside their ecological zone of origin, and today we find grasses adapted to a wide range of climatic regimes, from tropical forests to freezing Arctic and Antarctic ecosystems.
Pooideae, one of the most species-rich grass subfamilies, have successfully adapted to and diversified in cool climate ecosystems (Hartley, 1973; Edwards & Smith, 2010). However, it is unknown if cold and freezing tolerance evolved before, coincidentally with or during Pooideae evolution. What is clear is that Pooideae presently occupy much colder temperatures during the coldest month than other BEP lineages (see fig. S1 in Edwards & Smith, 2010), suggesting a shift in climate adaptation some time between the BEP divergence (Fig. 1) and early Pooideae diversification. Pooideae consist of 14 tribes (Soreng et al., 2000). The 10 earliest diverging, referred to hereafter as basal Pooideae (BP), represent a paraphyletic set of lineages with variable chromosome numbers (e.g. five, nine, 11 or 13) and significant morphological and ecological diversity, but only moderate species diversity, accounting for c. 20–30% of the total c. 3000 Pooideae species (http://delta-intkey.com) (Grass Phylogeny Working Group, 2001; Hilu, 2004). The earliest diverging BP tribe is Brachyelytreae, while the most recently diverging (i.e. sister to core Pooideae (CP)) is Brachypodieae. All four remaining tribes belong to a species-rich clade in which the basal chromosome number is seven, referred to as the core Pooideae (CP) (Fig. 1) (Soreng et al., 2000; Hilu, 2004). The CP encompass Hordeeae, Bromeae, and Poeae, in which all agriculturally important Pooideae crops belong.
Figure 1. Study species and climate adaptation. Branch colors reflect subfamily differences in the mean temperature of the coldest month as shown in fig. S1 in Edwards & Smith (2010). Hordeeae (Triticum aestivum and Hordeum vulgare), Poeae (Festuca pratensis and Lolium perenne), and PACMAD outgroup (Sorghum bicolor and Zea mays) clades are collapsed into single branches.
Download figure to PowerPoint
How Pooideae evolved to become a highly successful cold-adapted lineage is not well understood. Comparative genomics has identified some intriguing examples of Pooideae-specific evolution of low-temperature-induced (LTI) responses involved in freezing tolerance. For example, the important C-repeat binding factor (CBF) gene family have diversified extensively in Pooideae (Skinner et al., 2005; Li et al., 2012) possibly through some selection-driven mechanism (Sandve & Fjellheim, 2010). Two other intriguing examples are the Pooideae lineage gains of fructosyl transferase enzymes (Francki et al., 2006; Li et al., 2012) and the ice-interacting ice recrystallization inhibition proteins (Sidebottom et al., 2000; Sandve et al., 2008), which have both been shown to increase plant freezing tolerance (Hisano et al., 2004; Zhang et al., 2010). Although such fixations of lineage-specific molecular features may be driven by adaptive evolution, the existence of lineage-specific LTI stress responses per se reveals nothing about ancestral selection pressures. Moreover, these comparative genomics case studies do not provide a framework for testing hypotheses about the mechanisms involved in Pooideae cold climate adaptation.
One way to specifically test hypotheses concerning adaptive evolution is to reconstruct ancestral selection pressures by estimating the relationship between synonymous (dS) and nonsynonymous (dN) substitution rates (dN/dS) at individual codons during gene evolution (Zhang et al., 2005). Branches with a dN/dS ratio > 1 are considered to reflect positive selection pressure (i.e. adaptive evolution), while dN/dS = 1 and dN/dS < 1 are signatures of neutral and purifying selection, respectively. If adaptive evolution was important for Pooideae cold climate adaptation, we predict stronger signatures of positive selection (elevated dN/dS ratio) in LTI genes compared with the genome-wide background level.
A simple model of Pooideae climate adaptation can be hypothesized on the basis of the evolution of habitat temperature preference in the BEP clade (Edwards & Smith, 2010). According to this hypothesis, some change in the environment triggered adaptive evolution of cold-stress tolerance, and was followed by niche expansion and ecological diversification in cooler habitats, either before Pooideae divergence or during the evolution and diversification of Pooideae. This study is the first to reconstruct the evolution of substitution rates and ancestral selection pressure between early BEP diversification (the Ehrhartoideae–Pooideae split) and the diversification of the CP lineages, and specifically investigate whether LTI genes were key targets for adaptive evolution. Our results suggest that LTI genes were under strong adaptive selection before the divergence of the CP clade and reveal new insights into a poorly understood chapter in the evolutionary history of some of the world's most important crops.