A few-gene plastid phylogenetic framework for mycoheterotrophic monocots

Few-gene studies with broad taxon sampling have provided major insights into phylogeny and underpin plant classification. However, they have typically excluded heterotrophic plants because of loss, pseudogenization, or rapid evolution of plastid genes. Here we performed a phylogenetic survey of three commonly retained plastid genes to assess their utility in placing mycoheterotrophs.

We surveyed accD, clpP, and matK for 34 taxa from seven monocot families that include full mycoheterotrophs and a broad sampling of photosynthetic taxa. After screening for weak contaminants, we conducted phylogenetic analyses and characterized among-lineage rate variation.

Likelihood analyses strongly supported local placements of fully mycoheterotrophic taxa for Corsiaceae, Iridaceae, Orchidaceae, and Petrosaviaceae, in positions consistent with other studies. Depression of likelihood bootstrap support values near mycoheterotrophic clades was alleviated when each mycoheterotrophic family was considered separately. Triuridaceae (Sciaphila) monophyly was recovered in a partitioned likelihood analysis, and the family then placed as sister to Cyclanthaceae-Pandanaceae. Burmanniaceae placed in Dioscoreales with weak to strong support depending on analysis details, and we inferred a plastid-based phylogeny for the family. Thismiaceae species may retain a plastid genome, based on accD retention. The inferred position of Thismiaceae is unstable, but was close to Taccaceae (Dioscoreales) in some analyses.

Long branches/elevated substitution rates, missing genes, and occasional contaminants are challenges for plastid-based phylogenetic inference with full mycoheterotrophs. However, most mycoheterotrophs can be readily integrated into the broad picture of plant phylogeny using several plastid genes and broad taxonomic sampling.