Molecular phylogeny of the genus Russula in Europe with a comparison of modern infrageneric classifications

Species in the large mushroom genus Russula are important ecologically as ectomycorrhizal fungi and economically as comestibles. Most infrageneric classification schemes of this genus have originated in Europe, but because of nomenclatural history and an evolving suite of characters these systems remain largely incongruent. Using ribosomal DNA sequences for 87 species representing all infrageneric taxa described from Europe, the phylogenetic position and relationships among these species were examined. Cladistic analysis of the ITS1, 5.8S, and ITS2 regions showed a cluster of five to six small to large clades basal in the topology and one large apical clade arising from the deeper nodes, none of which has been previously recognized in toto at the subgeneric level. Two of these groups, the Compactae and Pallidosporinae, which have been previously recognized as subsections of section Compacta, did not appear to be closely related. Bootstrap support and Bremer decay values indicated that collapse of the tree into monophyly at the deeper nodes would result in two large groups which are consistent with the classical subgeneric concept of the Eurussulae and a narrowed Compactae. The topology confirmed some previously described infrageneric taxa at the section level including the Tenellae and Heterophyllae and at the subsection level including the Cupreinae, Laricinae, Lilaceinae, Integroidinae, Violaceinae, Sphagnophilinae, Viridantinae, Emeticinae, Subvelatae, Pallidosporinae, and portions of the Polychromae and Sardoninae. The molecular analysis also indicated many interesting new combinations or relationships not previously conceived. Mapping of characters such as spore print colour, taste, and presence of acid resistant incrustations, which have been used to define infrageneric taxa in Russula, onto the phylogeny identified interesting patterns consistent with hypotheses regarding plesiomorphic and apomorphic characters. However, because of potential loss or reversal of character states, this analysis did not support their unequivocal use in infrageneric classification.