Chloroplast phylogeny of Triticum/Aegilops species is not incongruent with an ancient homoploid hybrid origin of the ancestor of the bread wheat D‐genome

Abstract

In a recent Letter to New Phytologist Li et al. (2015) presented a re-evaluation of our recent finding from a phylogenomics study that the bread wheat D-genome lineage is probably of homoploid hybrid origin (Marcussen et al., 2014). The authors claim to present evidence from chloroplast phylogeny and gene tree topologies that necessitate additional hybridization events compared to the phylogeny proposed by us. Unfortunately Li et al.'s (2015) conclusions appear to be confounded by differences in nomenclatures used to describe the major Triticum/Aegilops clades (Fig. 1) and a misunderstanding concerning the temporal dimension of our phylogenetic model (i.e. when hybridization happened) (Fig. 1). In our original paper we presented phylogenomic data suggesting that one or more basal hybridization(s) between the ancestor of the A-genome and the ancestor of the B-genome in modern bread wheat gave rise to an ancestor of the bread wheat D-genome lineage 5–6 million years ago (Fig. 1). We used the terms ‘A-genome lineage’, ‘B-genome lineage’, and ‘D-genome lineage’ to refer to the three deep lineages encompassing all extant species in the Triticum/Aegilops clade (Fig. 1 in this Letter; Figs 2, 3, and S6 and their captions in Marcussen et al., 2014). This informal nomenclature was chosen because Marcussen et al. (2014) was part of a series of papers on the hexaploid bread wheat genome constituents (i.e. A, B and D) (International Wheat Genome Sequencing Consortium (IWGSC), 2014; Pfeifer et al., 2014) in addition to the lack of a phylogenetic infrageneric classification of Triticum/Aegilops. Thus, our use of ‘D-genome lineage’ in Marcussen et al. (2014) refers to the entire lineage containing T. aestivum-D, Ae. tauschii, and Ae. sharonensis (Fig. 1). This is not synonymous with how Li et al. (2015) define the ‘D-genome’ according to the haplome nomenclature of wheats (Fig. 1 and summarized in Table 1 of Li et al., 2015). Hence, in their re-evaluation of our hypothesis Li et al. (2015) consider the D-lineage as encompassing Aegilops tauschii only, while Ae. sharonensis is considered to belong to an S*+M clade (Fig. 1). One sentence from the Li et al. (2015) Letter that highlights this misunderstanding is: ‘Marcussen et al. (2014) proposed that A. tauschii originated from homoploid hybridization between A. speltoides and the ancestor of modern T. urartu/T. monococcum.’ This is a factual error. Marcussen et al. (2014) never claimed that Ae. tauschii evolved through a homoploid hybridization. What we proposed was that a homoploid hybridization gave rise to an ancestor of all D+S*+M clade species (including Ae. tauschii and Ae. sharonensis) (Fig. 1). In addition, Li et al. (2015) seem to have overlooked the timescale of our phylogenetic analyses. We report on an ancient hybridization (illustrated clearly in Fig. 2b in Marcussen et al., 2014) several million years before Aegilops tauschii and Ae. sharonensis existed as species. The hybrid history of the D-genome lineage (as Marcussen et al., 2014 refers to) is therefore not restricted to Ae. tauschii but encompasses the entire D+S*+M clade. The very same chloroplast phylogeny that Li et al. (2015) use to re-evaluate our conclusions is therefore completely congruent with our proposed phylogenetic network for wheats. Finally, we fully agree with Li et al. (2015) that the evolutionary history of Tritium/Aegilops is likely to be even more complex than we currently understand it to be. Judging from the young age of the Triticeae and the numerous reports of suspected reticulations, sequencing and analysis of additional diploid and polyploid Triticum/Aegilops genomes will most likely provide evidence for additional hybridization events not yet detected.