Contributions of rpb2 and tef1 to the phylogeny of mushrooms and allies (Basidiomycota, Fungi) P. Brandon Matheny a,*, Zheng Wang a, Manfred Binder a, Judd M. Curtis a, Young Woon Lim b, R. Henrik Nilsson c, Karen W. Hughes d, Valerie �� Hofstetter e, Joseph F. Ammirati f, Conrad L. Schoch g, Ewald Langer h, Gitta Langer h, David J. McLaughlin i, Andrew W. Wilson a, Tobias Fr��slev j, Zai-Wei Ge k, Richard W. Kerrigan l, Jason C. Slot a, Zhu-Liang Yang k, Timothy J. Baroni m, Michael Fischer n, Kentaro Hosaka o, Kenji Matsuura p, Michelle T. Seidl q, Jukka Vauras r, David S. Hibbett a a Biology Department, Clark University, 950 Main St., Worcester, MA 01610, USA b Department of Wood Science, University of British Columbia, 2900-2424 Main Mall, Vancouver, British Columbia V6T 1Z4 Canada c Department of Plant and Environmental Sciences, Goteborg �� University, Box 461, SE-405 30, Goteborg, �� Sweden d Botany Department, 437 Hesler Biology Building, University of Tennessee, Knoxville, TN 37996-1100, USA e Department of Biology, Box 90338, Duke University, Durham, NC 27708, USA f Biology Department, Box 351800, University of Washington, Seattle, WA 98195, USA g Botany and Plant Pathology, Oregon State University, Corvallis, OR 97331, USA h FB 18, Naturwissenschaften, Institut fur �� Biologie, Universitat �� Kassel, FG Okologie, �� Heinrich-Plett Str. 40, D-34132, Kassel, Germany i Department of Plant Biology, University of Minnesota, 250 Biological Science Center, 1445 Gortner Ave., St. Paul, MN 55108, USA j Department of Microbiology, Institute of Biology, University of Copenhagen, Oster Farimagsgade 2D, DK-1353 Copenhagen K, Denmark k Kunming Institute of Botany, Chinese Academy of Sciences, Heilongtan, Kunming 650204, PR China l Sylvan Research, 198 Nolte Drive, Kittanning, PA 16201, USA m Department of Biological Sciences, SUNY Cortland, Box 2000, Cortland, NY 13045-0900, USA n Staatliches Weinbauinstitut Freiburg, Merzhauser Stra��e 119, D-79100 Freiburg, Germany o Department of Botany, The Field Museum, 1400 S. Lake Shore Drive, Chicago, IL 60605-2496, USA p Faculty of Agriculture, Okayama University, 1-1-1 Tsushima-naka, Okayama 700-8530, Japan q Environmental Microbiology Laboratory, Inc., 1400 12th Ave SE, Bellevue, WA 98004, USA r Herbarium, University of Turku, FI-20014 Turku, Finland Received 8 April 2006 revised 19 August 2006 accepted 24 August 2006 Available online 23 September 2006 Abstract A phylogeny of the fungal phylum Basidiomycota is presented based on a survey of 160 taxa and five nuclear genes. Two genes, rpb2, and tef1, are presented in detail. The rpb2 gene is more variable than tef1 and recovers well-supported clades at shallow and deep tax- onomic levels. The tef1 gene recovers some deep and ordinal-level relationships but with greater branch support from nucleotides com- pared to amino acids. Intron placement is dynamic in tef1, often lineage-specific, and diagnostic for many clades. Introns are fewer in rpb2 and tend to be highly conserved by position. When both protein-coding loci are combined with sequences of nuclear ribosomal RNA genes, 18 inclusive clades of Basidiomycota are strongly supported by Bayesian posterior probabilities and 16 by parsimony boot- strapping. These numbers are greater than produced by single genes and combined ribosomal RNA gene regions. Combination of nrDNA with amino acid sequences, or exons with third codon positions removed, produces strong measures of support, particularly for deep internodes of Basidiomycota, which have been di���cult to resolve with confidence using nrDNA data alone. This study produces 1055-7903/$ - see front matter �� 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.ympev.2006.08.024 * Corresponding author. Fax +1 508 793 7174. E-mail address: pmatheny@clarku.edu (P. Brandon Matheny). www.elsevier.com/locate/ympev Molecular Phylogenetics and Evolution 43 (2007) 430���451

strong boostrap support and significant posterior probabilities for the first time for the following monophyletic groups: (1) Ustilagino- mycetes plus Hymenomycetes, (2) an inclusive cluster of hymenochaetoid, corticioid, polyporoid, Thelephorales, russuloid, athelioid, Boletales, and euagarics clades, (3) Thelephorales plus the polyporoid clade, (4) the polyporoid clade, and (5) the cantharelloid clade. Strong support is also recovered for the basal position of the Dacrymycetales in the Hymenomycetidae and paraphyly of the Exobasidiomycetidae. �� 2006 Elsevier Inc. All rights reserved. Keywords: Basidiomycota EF1-a Fungi Introns RNA polymerase II Systematics 1. Introduction Although the number of multi-gene phylogenetic studies of fungi has increased over the past dozen years, the propor- tion of such studies that utilize multiple loci has not grown (Lutzoni et al., 2004). PCR primers, however, have been developed for single or low-copy nuclear protein-coding genes for fungal phylogenetic research across a broad spec- trum of major fungal groups including the phyla Chytridi- omycota, Zygomycota, Ascomycota, and Basidiomycota (Fr��slev et al., 2005 Helgason et al., 2003 James et al., 2006 Kretzer and Bruns, 1999 Liu et al., 1999 Liu and Hall, 2004 Matheny et al., 2002 Matheny, 2005 O���Donnell et al., 1998 O���Donnell et al., 2001 Thon and Royse, 1999 Voigt and Wostemeyer, 2000, 2001). Of these, several genes have arisen as candidates to complement or rival the suite of nuclear and mitochondrial ribosomal RNA genes (rDNA) that have served as the platform for fungal molecular sys- tematics for almost the past 20 years (Blanz and Unseld, 1987 Bruns et al., 1991 Bruns et al., 1998 Hibbett, 1992). Examples include atp6, a mitochondrial protein-cod- ing gene (Kretzer and Bruns, 1999 Robison et al., 2001) gapdh, the glyceraldehyde 3-phosphate dehydrogenase gene ( Den Bakker et al., 2004 Berbee et al., 1999 Smith, 1989) rpb2, which encodes the second largest subunit of RNA polymerase II (Liu et al., 1999 Liu and Hall, 2004 Mathe- ny, 2005 Reeb et al., 2004 Wang et al., 2004 Zhang and Blackwell, 2002) rpb1, the gene that encodes the largest subunit of RNA polymerase II (Kropp and Matheny, 2004 Matheny et al., 2002 Matheny, 2005 Tanabe et al., 2002) b-tubulin encoding genes (Ayliffe et al., 2001 Begerow et al., 2004 Einax and Voigt, 2003 Keeling et al., 2000 Thon and Royse, 1999) actin encoding genes (Cox et al., 1995 Helgason et al., 2003 Tarka et al., 2000 Voigt and Wostemeyer, 2000) and tef1, which codes for the translation elongation factor 1-a (Baldauf and Palmer, 1993 O���Donnell et al., 2001 Rehner and Buckley, 2005 Roger et al., 1999). The development of alternative markers has been critical in the field of molecular systematics due to inherent limitations of taxon-rich single gene studies (Lips- comb et al., 1998 Rokas and Carroll, 2005 Rosenberg and Kumar, 2001) and limitations of rDNA regions in general (Aanen et al., 2001 Alvarez and Wendel, 2003 Buckler et al., 1997 Hasegawa and Hashimoto, 1993 Kauserud and Schumacher, 2003 Moncalvo et al., 2002 O���Donnell and Cigelnik, 1997 Okabe and Matsumoto, 2003 Sang, 2002 Stiller and Hall, 1999). Here, we present a multi-gene analysis of the Basidiomycota using rpb2, tef1, and nuclear 18S, 25S, and 5.8S rRNA genes. The Basidiomycota (basidiomycetes) includes about 30,000 described species (Kirk et al., 2001) distributed across three classes: the Ustilaginomycetes (true smuts and allies and yeast forms), Urediniomycetes (rusts, anther smuts, and diverse yeasts), and Hymenomycetes (mush- rooms and allies and other molds). A fourth class, the Wallemiomycetes (Matheny et al., 2006b Zalar et al., 2005), was recently proposed to accommodate several unusual xerophilic molds. The phylum is important not just because of its taxonomic diversity but also due to var- ious ecological roles as mutualists with vascular plants, bryophytes, green algae, and cyanobacteria as decompos- ers primarily in terrestrial ecosystems and as pathogens of plants, animals, and other fungi. When reproducing sexual- ly, basidiomycetes typically produce four meiotic products (basidiospores) on the exterior of their meiosporangia (bas- idia). Not all basidiomycetes reproduce sexually, and some are better known by their asexual or anamorphic state (e.g., polyphyletic yeasts), including the medically important Malassezia, Trichosporon, Pseudozyma, and Cryptococcus species (Fell et al., 2000 Fell et al., 2001 Gueho �� et al., 1998 Oberwinkler, 1987 Sampaio, 2004 Scorzetti et al., 2002 Sugita et al., 2003 Xu et al., 2000). The rusts alone constitute an order of about 7000 species that are obligate pathogens of ferns, conifers, and flowering plants (Cum- mins and Hiratsuka, 2003). One thousand five hundred species of smuts, bunts, and allies make up another diverse assemblage of pathogens of mostly flowering plants (Vanky, �� 2002). The mushroom-forming fungi include roughly 20,000 species (Kirk et al., 2001) that decompose lignin and/or cellulose, in addition to a diverse group of root symbionts that form mycorrhizas with trees, shrubs, and grasses. In addition, some basidiomycetes form diverse and unusual symbioses with insects (Aanen et al., 2002 Chapela et al., 1994 Henk and Vilgalys, 2006). Few multi-gene phylogenies have addressed higher-level relationships in the Basidiomycota. Though some recent papers have utilized multiple regions of rDNA at a relative- ly broad taxonomic level (Binder and Hibbett, 2002 Binder et al., 2005 Larsson et al., 2004 Lutzoni et al., 2004), only one study (Begerow et al., 2004) has incorporated a pro- tein-coding gene (b-tubulin) for exemplars across the three major lineages (classes) of the phylum. Previous phyloge- netic assessments of the Basidiomycota relied principally on nuclear rDNA gene analyses, which have contributed P. Brandon Matheny et al. / Molecular Phylogenetics and Evolution 43 (2007) 430���451 431

considerable insight. Nevertheless, many aspects of the Basidiomycota phylogeny have not been resolved with rDNA, especially at deeper nodes. Many 18S or 25S studies suggest that the Hymenomycetes and Ustilaginomycetes are sister groups but with weak or no bootstrap support (Berbee and Taylor, 1993 Gargas et al., 1995b Swann and Taylor, 1995a Swann and Taylor, 1995b Prillinger et al., 2002 Tehler et al., 2003 Weiss et al., 2004a). A recent analysis of ultrastructural characters also supports this view (Lutzoni et al., 2004). However, other rDNA studies suggest alternative arrangements such as a sister relation- ship between Hymenomycetes and Urediniomycetes or an unresolved trichotomy between these three classes (Gueho �� et al., 1989 Swann and Taylor, 1993 Berres et al., 1995 McLaughlin et al., 1995a Nishida et al., 1995 Taylor, 1995 Begerow et al., 1997 Swann et al., 1999 Tehler et al., 2000 Bauer et al., 2001 Swann et al., 2001 Begerow et al., 2004). Gross relationships among major lineages within each class of Basidiomycota are also poorly resolved or not strongly supported by taxon-rich studies using nucle- ar and/or mitochondrial rDNA regions (Begerow et al., 2000 Binder and Hibbett, 2002 Binder et al., 2005 Hibbett and Thorn, 2001 Lutzoni et al., 2004 Moncalvo et al., 2002 Weiss et al., 2004a). The development of protein-cod- ing genes as phylogenetic markers is thus required to resolve the phylogeny of the Basidiomycota. Availability of primers for orthologous and phylogenetically informa- tive genomic regions should facilitate development of mul- ti-gene genealogies, which provides another impetus for their use (Taylor et al., 2000 Taylor and Fisher, 2003 Xu et al., 2000). We present a study of 160 species of Basidiomycota for which we have sequenced a 950���1200 bp fragment of tef1 between exons 4 and 8 (Wendland and Kothe, 1997) and, in most instances, 2200 bp of rpb2 between conserved domains 5 and 11 (Liu et al., 1999). Both genes are useful for analyses at high taxonomic levels because they are con- served at the amino acid level and easy to align, and because conserved primers sites have been identified (Liu et al., 1999 Rehner and Buckley, 2005). We explore the evolutionary dynamics of these genes at the level of coding sequences and gene structure with respect to spliceosomal intron placement and evaluate each gene phylogenetically. We also combine the protein-coding data with nuclear rDNA sequences (18S, 25S, and 5.8S) in an attempt to resolve nodes that have been unresolved or weakly sup- ported by previous studies. The strongly supported, highly resolved phylogenetic trees resulting from these multi-gene analyses will be useful for revising the classification of Basidiomycota, but that is not the purpose of the present study. Instead, the taxon names used here are derived from prior works, and they include a mixture of formal Linnae- an taxa and informal clade names. A revised classification of Basidiomycota (and other fungi) that draws heavily on the present study is in preparation and will be published elsewhere (see http://www.clarku.edu/faculty/dhibbett/ AFTOL/AFTOL.htm). 2. Materials and methods 2.1. Taxon sampling One hundred and forty sequences were produced for tef1 and one hundred and thirty-two for rpb2. We sampled 146 taxa for the nuclear ribosomal DNA (nrDNA) tandem repeat, which includes the 18S, 25S, and 5.8S genes and the two internal transcribed spacers (S1). Taxa from the tef1 and rpb2 datasets were merged with those from a matrix of nrDNA for a final total of 146 taxa in the combined data set, including 12 Urediniomycetes, 6 Ustilaginomycetes, 125 Hymenomycetes, and 3 Ascomycota that were used for outgroup purposes. Exemplars were sampled from every subclass of Basidiomycota (Kirk et al., 2001) except for the Entorrhizomycetidae(Bauer etal.,2001), whichiscomposed of a single genus, Entorrhiza, in the Ustilaginomycetes, and the Wallemiomycetes, which includes the sole genus, Wall- emia (Zalar et al., 2005). Taxon sampling was extensive with- in the Hymenomycetidae or Homobasidiomycetes and allies. Exemplars have been selected to represent every major lineage of Binder et al. (2005), the most exhaustive sampling of Homobasidiomycetes to date, with exception of the Glo- eophyllum clade (Thorn et al., 2000). The intent of our sam- pling strategy was to avoid missing data. However, of the taxa included in the combined data set, 11 lack rpb2 sequenc- es and 18 lack tef1 sequences. Some rpb2 sequences from domains 5to7or7to11weremergedinthecombined matrix but omitted from the rpb2 only analysis. Despite these char- acter gaps, almost 90% of the taxa are represented by sequences from the five genesin the combined dataset. Meth- ods of phylogenetic inference generally do not appear sensitive to moderate amounts of missing data for large alignments, thus, we include taxa for which a gene region is missing (Philippe et al., 2004 Wiens, 2006). For outgroup purposes we used the following sequences of Asco- mycota: Schizosaccharomyces pombe (Z19136, AY046272, AY046223, D13337, NC_003421), Saccharomyces cerevi- seae (J01355, AF548094, AJ544253, M15693, M10992), Ale- uria aurantia AY544654, AY544698, AF072090, DQ247785, DQ466085, and Neurospora crassa (AY681158, AY046271, AY681193, AF107789, XM_329192). 2.2. DNA extraction Sources of genomic DNA included fresh fruit bodies or spores, dried herbarium specimens, and live or lyophilized cultures. Most DNA extractions were performed at Clark University following the procedures detailed at http:// www.clarku.edu/faculty/dhibbett/HibbettLab.protocols.htm, as well as those outlined in Hughes et al. (1998) and Hof- stetter et al. (2002). Extractions at Clark University relied principally on the E.Z.N.A Fungal Miniprep Kit (Omega Bio-tek, Doraville, Georgia) or, in the case of rust and smut spores, the E.Z.N.A. Forensic DNA Extraction Kit (Omega Bio-tek). Genomic DNA was also isolated from several lyophilized cultures at the Centraalbureau voor 432 P. Brandon Matheny et al. / Molecular Phylogenetics and Evolution 43 (2007) 430���451