Tribal phylogeny of the Asteraceae based on two non-coding chloroplast sequences, the trnL intron and trnL/trnF intergenic spacer

Asteraceae are the largest family of dicotyledonous plants and have long been known for their taxonomic complexity. The ubiquitous parallelisms in morphology within the family have made phylogenetic reconstruction and tribal circumscription an area of long debate. In this study we explored the utility of using two relatively short non-coding chloroplast DNA sequences, the trnL intron and trnL/trnF intergenic spacer, to resolve phylogenetic relationships among the tribes. The results of the phylogenetic analysis produced trees that are topologically congruent with prior phylogenetic hypotheses based on both morphological and molecular data sets. The Asteroideae are a monophyletic group, but the Cichorioideae are paraphyletic. The primary clades of the Cichorioideae are the Mutisieae-Cardueae, Liabeae-Vernonieae, and of the Asteroideae, the Inuleae-Plucheeae, Astereae-Anthemideae, Senecioneae-Gnaphalieae, and the helianthoid clade (Helenieae, Heliantheae s. str., and Eupatorieae). The Inuleae-Plucheeae clade is sister to the remainder of the Asteroideae, and the paraphyly of the Inuleae s.l. (Gnaphalieae, Inuleae s. str., and Plucheeae) is firmly supported by our analysis. Our study illustrates the utility of the trnL intron and trnL/F intergenic spacer for resolving relationships among tribes of the Asteraceae. Using approximately 874 bp, we were able to produce a phylogeny of comparable resolution to phylogenies based on well-known coding regions such as rbcL and ndhF. For phylogenetic inference at the family level the trnL intron and trnL/F spacer provide similar levels of resolution to longer coding sequences (e.g., rbcL, ndhF), while having the advantage of being much easier io amplify and sequence due to their short lengths and universal primers. The numerous insertions and deletions commonly found in this region are easily aligned and are phylogenetically informative, thus adding considerably to the information content per base pair sequenced.