Significance of molecular phylogenetic analyses for paleobotanical investigations on the origin of angiosperms

Abstract

Molecular phylogenetic analyses have provided increasing evidence that angiosperms are not related to Gnetales, thus contradicting the anthophyte hypothesis based on morphological cladistic analyses and throwing the question of angiosperm relatives back to paleobotanists. Previous analyses of gene sequences based on a molecular clock conflicted with the fossil record in indicating a Late Palaeozoic or Triassic origin of the angiosperms, but closer examination suggests that these dates were biased by the use of herbaceous taxa with accelerated rates of molecular evolution. Despite uncertainty on angiosperm relatives, analyses of many genes consistently place Amborella, Nymphaeales, Austrobaileya, Trimenia and Illiciales (the ‘ANITA grade') at the base of extant angiosperms, possibly followed by Chloranthaceae. Molecular phylogenies imply that the first crown-group angiosperms had columellar exine structure, suggesting that Hauterivian-Barremian reticulate-columellar monosulcates may be closer to the origin of angiosperms than was thought when granular Magnoliales were believed to be basal. Hauterivian pollen with a verrucate tectum and microspinules is especially similar to Amborella. The ANITA lines and Chloranthaceae have ascidiate carpels sealed by secretion and often exotestal seeds, fitting the abundance of such carpels and seeds in Barremian-Aptian mesofloras. Similarities between Aptian angiosperm leaves and ANITA taxa, such as chloranthoid teeth and variable stomatal structure, also suggest that Early Cretaceous angiosperms were more primitive than previously appreciated. Molecular results may help refine search images for extinct angiosperm relatives, away from Gnetales and toward groups such as Caytonia, Glossopterids, Bennettitales and Corystosperms. Since molecular data place the vesselless taxa Amborella and Nymphaeales at the base of the angiosperms, the presence of vessels is not evidence that gigantopterids are related to angiosperms. The conclusion that columellar structure is ancestral reaffirms the potential of Triassic reticulate-columellar Crinopolles pollen as angiosperm relatives.