614 Mycologia, 96(3), 2004, pp. 614–626.  2004 by The Mycological Society of America, Lawrence, KS 66044-8897 Tuberculina: rust relatives attack rusts 1 Matthias Lutz 2 Robert Bauer Dominik Begerow Franz Oberwinkler Universita ¨t Tu ¨bingen, Botanisches Institut, Lehrstuhl Spezielle Botanik und Mykologie, Auf der Morgenstelle 1, 72076 Tu ¨bingen, Germany Dagmar Triebel Botanische Staatssammlung Mu ¨nchen, Menzinger Straße 67, 80638 Mu ¨nchen, Germany Abstract: Molecular sequence data together with ul- trastructural features were used to infer the phylo- genetic position of Tuberculina species. Additional ul- trastructural characteristics were used to determine their mode of nutrition. We investigated ultrastruc- tural morphology of the type species Tuberculina per- sicina and determined base sequences from the D1/ D2 region of the nuclear large-subunit ribosomal DNA of the three commonly distinguished Tubercu- lina species, T. maxima, T. persicina and T. sbrozzii. Analyses of sequence data by means of a Bayesian method of phylogenetic inference using a Markov Chain Monte Carlo technique reveal the basidiomy- cetous nature of Tuberculina. Within the Uredini- omycetes, Tuberculina clusters as a sister group of Hel- icobasidium, closely related to the rusts (Uredinales). This phylogenetic position is supported by the ure- dinalean architecture of septal pores in Tuberculina. In addition, we present aspects of the ultrastructural morphology of the cellular interaction of Tuberculina and rusts showing a unique interaction with large fu- sion pores, revealing the mycoparasitic nature of Tub- erculina on its close relatives, the rusts. Key words: cellular interaction, molecular phy- logeny, mycoparasitism, nuc-LSU rDNA, septal pore morphology, systematics, ultrastructure, Urediniomy- cetes INTRODUCTION In 1817, a member of the genus Tuberculina Sacc. first was described accurately by Ditmar (1817) as Accepted for publication September 29, 2003. 1 Part 212 in the series Studies in Heterobasidiomycetes from the Bo- tanical Institute, University of Tu ¨ bingen. 2 Corresponding author. E-mail: matthias.lutz@uni-tuebingen.de Tubercularia persicina. His morphological character- ization, with some additions, remains valid: Members of the genus Tuberculina are characterized by the for- mation of hemispherical lilac to violet sporodochia. They consist of palisade-like arranged, short, mod- erately thick conidiogenous cells, each of which bears one globose, smooth conidium at the tip. The spo- rodochia break through the surface of higher plants and emit a powdery mass of conidia. Sometimes spherical sclerotium-like structures are formed. In addition, Tuberculina is known to exist only in asso- ciation with rusts as first postulated by Saccardo (1880) and later elaborated by Tubeuf (1901) and others. Saccardo (1880) excluded Tubercularia persi- cina Ditmar from the genus Tubercularia Tode : Fr., which should have pleurogenous conidiogenesis and thread-like conidiogenous cells, and described the new genus Tuberculina Sacc. for species with acroge- nous conidiogenesis, short and broad conidiogenous cells which are parasitic on the aecial stage of rust fungi. The three Tuberculina species, T. maxima, T. per- sicina and T. sbrozzii, commonly are recognized (e.g., von Arx 1981, Ellis and Ellis 1988). They are distrib- uted worldwide, living in association with more than 150 rust species from at least 15 genera. However, up to 45 species were described with the authors follow- ing strikingly different species concepts. Adopting a concept based on morphological characters, plant parasites (e.g., T. solanicola Ellis parasitic on fruits of Solanum melongena L. [Ellis 1893]) and parasites of non-rust fungi (e.g., T. ovalispora Pat. parasitic on Darluca filum [Biv.] Castagne [Patouillard and Gail- lard 1888]) were included in the genus. Other au- thors used a species concept based on host specific- ities, distinguishing Tuberculina species on different rust hosts (Spegazzini 1880, 1884) or even plant hosts (Gobi 1885). After controversial discussions whether Tuberculi- na-like fungi should be treated as smuts, rusts, asco- mycetes or hymenomycetes, the genus presently is as- signed mostly to the Fungi Imperfecti because no stages of sexual reproduction are known. Research on Tuberculina was motivated by two main factors: -taxonomy (e.g., Cooke 1888, Patouil- lard and Gaillard 1888, Spegazzini 1880, 1884, 1911) and the use of Tuberculina as a biological agent in