SSU Rona Sequence Data 1

EXPERIMENTAL MYCOL()(;Y 19,7-15 (1995)

Phylogeny of Discomycetes and Early Radiations of the Apothecial Ascomycotina Inferred from -SSU rONA Sequence Data 1

ANDREA GARGAS",·2 AND JOHN W. T A YLORt

·Department of BOlany, NHB-I66, National Museum of Natural His/ory, Smithsonian Ins/itution, Washington, DC l056fJ; and tDeparlment of Plant Biology, I I J Koshland Hail, University of CaUfornia, Berkeley, California 94720

Accepted for publication October 16. 1994

G ARGAS, A" AND TAYLOR, J. W. 1995. Phylogeny of discomycetes and earl y radiati ons of the apothecial Ascomycotina inferred from SSU rDNA sequence data. Experimen/al Mycology 19, 7- 15. We used nucleotide sequences of the small subunit ribosomal genes (SSU rONA) to examine evolutionary relationsh ip s of apothecial ascomycetes (division Ascomycota; class Discomycetes sensu), commo nl y known as the cup fungi. The apothecial ascomycetes include both lichen fonning and free-living fungi. We sequenced the SSU rONA from representatives of 10 funga l genera from four orders: Pe zizale~ (Ascobolus Uneola/us, Morchel/a elata agg" Peziw badia); Leotiales (Leo/ia lubrica, Sclero/inia sclerotiorum); Caliciales (CaUdum tricolor, Myc()Calicium albonigrum, Sphaerophorus globosus); and Lecanorales (Lecanora dispersa, Porpidia crus/ula/a). Of these, C. Iricolor, S. globosus, L. dispersa, and P. cruslillata are lichen-forming fungi. Based on parsimony analyses of approximately 1750 aligned nucleotides of their SSU rO NA, we deter mined a most parsimonious tree (MPT). Thi s hypothesis suggests that the apothecial ascomycetes are a paraph yletic assemblage, basal to other groups of filamentous ascomycetes including repre sentati ves of the perithecial fungi and cleistothecial fungi. The most parsimonious tree produced using thi s dataset suppor1ed the monophyly of the orders Pezizales, Leotiales, and Lecanorales. However, there was no suppor1 for monophyly of the representative Caliciales: S. g/oboslls had affi nities with member5 of the Lecanorale~. This ph ylogenetic hypothesis recognizes Pezizales as basal and suppor1 S Nannfeldt's hypothesis (1932) of a primitive apothecial ascomata with su b ~e quent evolution of perithecial and cleistothecial forms. Thi~ MPT provide~ a foundation for un derstanding evolution of the ascomycetous fungi. c 1 99~ Acadcm;c I'r< ... II><, I NDEX DESC RIPTORS: fungi: Ascomycetes; discomycetes; lichen-forming; apothecia; 18S rONA; SSU rDNA; small subunit ribosomal DNA; molecular evolution; phylogeny.

Fungi that produce cup-shaped ascomata mycetes, nor on their relationships to other (apothecial ascomycetes) are commonl y fungi . Most of the apothecial ascomycetes known as the cup fungi or Discomycetes. (around 6000 spp.) are lichen-forming fungi The placement of this group within the fil in symbiotic re lationships with green algae amentous ascomycetes is important since or cyanobacteria, The 13,000 species of li the apothecial fungi represent almost 9000 chen-forming ascomycetes, including mem of the 28 ,000 described ascomycete species bers of the apothecial ascomycetes as we ll (Hawksworth et al., 1983). There is no hi s as members of other classes, constitute torical consensus on the evolutionary rela nearly half of the descri bed species of as tionships within the apothecial asco- comycetes. These fungi have been treated separately from pathogenic and free-living 1 Sequence data from this article have been depos fungi, but must be considered to produce a ited with the GenBank Data Library under Accession Nos. U5732- US74 I, comprehensive phylogeny of the asco 2 To whom correspondence shou ld be addres~ed. mycetes. To address the relationships of Fax: (:ro2) 786-2563. E- mai l: garga~@onyx . si.edu. the apothecial ascomycetes. we sequenced

7 0147-5975195 $6.00 Copyright c 1 \l9~ by A ca.d~mk I'r< ... In<:. All righ,. 0( ~producti"" in any form ~.., .... ed. 8 GARG ..... S AND TAYLOR the small subunit rONA gene (SSU rO NA») (Berbee and Taylor, 1992) from asco from representatives of 10 species and llsed mycetes with closed ascomata (cleistothe these nucleotide sequences to produce a cia) and fl ask-shaped ascomata (perithecia) phylogenetic hypothesis. have suggested that these are useful phylo Classification schemes for apolhecial as genetic characters. and that the classes comycetes have been reviewed by Kim based on these characters, Plectomycetes brough (1970) and for all ascomycetes by and Pyrenomycetes, respectively, are sup Hawksworth (1985), Based on reproductive ported as monophyletic lineages. Landvik structures of open, simple cups in contrast et af. (1993) included members of the Leo to the closed structurcs of the cleistothecial tiales and Pezizales, but found no clear ascomycetes or the fl ask-shaped structures monophyly for either group. Saenz el af. of the perithecial ascomycetes, mycologists (1994) showed the apothecial ascomycetes have variously proposed that the cup fungi to be a basal assemblage which included the constitute a monophyletic group of recent powdery mildew Blumeria graminis (Erysi origin, an ancestral group. or several inde phales). The fi rst analysis to include SSU pe ndent groups. Karf and Eriksson have rONA from lichen-forming fungi (Gargas, diffe red as to whether or not the Disco 1992) also found the Calic iales, Lecano mycetes are monophyletic. Korf (1973) rales, Leotiales, and Pezizales to be a basal placed the apothecial ascomycetes in the assemblage within the filamentous asco single class .. Discomycetes." Eriksson mycetes, although monophyly of each of (1981), in an outline of the ascomycetes, these groups was not rejected by maximum classified the apothecial ascomycetes into li kelihood analysis of the SSU rDNA data groups of one to several fami lies. Eriksson set. concluded that without a comprehensive in In the present study we have used SSU vestigation of their morphology and ontog rONA sequences to test the monophyly of eny, the ascomycetes could not be arranged the class Discomycctes, those fungi which into classes. Even when morphology and possess apothecia. We selected taxa of ontogeny were well known, interpretation apothecial ascomycetes that would address was difficult. Eriksson (1981) foc used atten the fo llowing questions: (I) with in all asco tion on the orders of ascomycetes and dis mycetes, are the fila mentous forms sup couraged the use of supraordinal taxa; the ported as a single lineage? (2) which type of Dictionary of Fungi (Hawksworth et 01., ascoma should be considered ancestral 1983) and Outline of Ascomycetes (Eri ks within the ascomycetes? (3) within the non son and Hawksworth, 1993) present the as yeast ascomycetes. arc the apothecial asco comycete fungi without class distinctions. mycetes supported as monophyletic and To examine monoph yly of the morpho separate from the c1cistothecial and peri logically defined class di scomycetes, as thccial ascomycetes? and (4) within the well as orders within this group, we have apothecial ascomycetes. are re presenta sought a source of characters independent tives of the fou r orders Pezizales. Cali from their morphology. An independent dales. Leoliales, and Lecanorales each test of phylogenetic hypotheses based on supported as monophyletic? morphological characters can be provided by molecular data. For example, recent MATERIALS AND METHODS analyses using S$U rONA sequence data Isolares ) Abbreviati ons used: PC R, polymerase chai n reac tion: MPT, most pa rsimonious tree: SSU. small sub We sequenced the SSU rONA from 10 unit. fungal species representing four orders of PHYLOGENY OF ASCOMYCETES FROM SSU rONA 9 ascomycete apothecial fungi . One species man) to sequence both the coding and the from a represe ntative genus for each order noncod ing strands. was sequenced, as we ll as one or two other To confirm the identity of the fungal species from ge nera considered to be within DNA being sequenced, sequence data for a the order. The species sequenced from diagnostic region was obtained from at least each order were: Pezizales-Ascobolus Ii one related species, or the fungal sequence neoimus Brumm. , Morcheffa elaw agg., was compared to that obtained by another Peziza badia Pers.; Leoli ale s-Leotia 111- researcher (5. Landvik, personal communi brica Pers.:Fr., Sclerotinia scJerotiorum cation). The S5U rONA sequences from (lib . DeBary); Ca li ciales-Calicillm tricolor the 10 fungi have been archived in Gen F . Wilson, Mycocalicium a/bonigrllm Bank. (Ny!.) TibelJ , Sphaerophorus globosus (Huds.) Vain. ; and Lecanorales-Lecanora Sequence Afignment and Construction 0/ dispersa (Pers.) Sommcrf., Porpidia crus Phylogenetic Trees tulata (Ach.) Hertel and Knoph (syn: Lecidea crustulata). The sequences were aligned with the fol lowi ng fungi (with GenBank accession DNA Extraction and Amplification Nos.): Saccharomyces cerevisiae Meyen ex Hansen (101353, M27607) (Rubtsov et We used standard fungal miniprep proto al., 1980); Neurospora crassa Shear and cols (Lee and Taylor, 1990; White et al., Dodge (X04971) (Chambers et al., 1986); 1990) to extract the tolal DNA from fungal Ascosphoeru apis (Maasen ~x Claussen) cultures, fre shly collected material, or her 1 Olive ex Spiltoir (M83264) , Byssochlamys barium specimens. From dilutions (10 - , 2 niveo Westling (M83256), Chaetomium ela 10 - ) of this total DNA we specificall y am tum Kunze (M83257), Eremascus albus Ei plified the fungal nuclear SSU rO NA using dam (M83258), Leucostoma persoonii polymerase chain reaction (PCR) primers Hoh n (M83259) , Ophiostoma 111m; (Buism.) designed to favor the amplification of fun Nannf. (M83258), and Thremoascus crusta gal over green-algal DNA (NSI7UCB ceus (Apinis and Chesters) A. C. Stolk NS24UCB; Gargas and Taylor, 1992). The (M83263) (Berbee and Taylor, 1992); Euro initial PCR am plification ran for 30 cycles tfllm rubrum Konig, Spieckermann. and W. (each cycle was for 2 min at 97"C, 1 min at Bremer (U00970), Pleosporo rudis Rabenh. 48°C, 45 s at 72°C with a 4 s/cycle extension ex Ces. and de Not. (U00975) (Berbee and at 7rC). A second PCR amplification of 30 Taylor, 1993) ; Cudonia con/lisa Bres. cycles was used 10 obtain either single (Z30240), Gryromitra esculenta (Pers.) Fr. stranded (Gy lJensten and Erlich, 1988) or (Z30238), In ermisia aggregata Berk. and double-stranded DNA products (Kusukawa Broome (Z30241) , Neolecta vitellino et al., 1990) for sequencing by the dideoxy (Bres.) Korf and 1. K. Rogers (Z27393), labeling method (TAQucnce kit ; V.S. Bio Plectania nigrella (Pers.:Fr.) P. Karst chemical Corp ., Cleveland. OH). (Z27408), and Spath lila ria flavida Pers. Sequencing Reactions (Z30239) (Landvik et at.. 1993); Blumeria gramit/is (DC.) Speer f. sp. hordei (L26253) We used the peR primers NS2-NS7 (Saenz et af., 1994). (White et al., 1990), UCBNSI7-UCBNS24 We ali gned the sequences with the (Gargas and Taylor, 1992), MB2 (the com Pile V p computer program (Genetics Com plement of UCBNS23; M. Berbee, persona) pu ter Group, Madison, WI) and hand communicati on), and CNS26 (TCGAA corrected this alignment. S. cerevisiae and AGTTG ATAGG GCAG; gift of B. Bow- N. vitellino were chosen as the outgroups 10 GARGAS AND TAYLOR based on the results of Landvik et ai., 1993 . and Oamberger. manuscript In prepara We used maximum-parsimony analysis tion). (Camin and Sakal, 1965 ; PAUP 3.1, Swof ford, 1991) to determine a single most par Parsim'ony Analysis Supports the Orders simonious tree (MPT) from this dataset. We Pezizales, Leotiales, and Lecanorales resampled the alignment 500 times using To test for monophyly of various groups bootstrap (Felsenstein, 1985) to derive 500 within the ascomycetes, our alignment in replicate datasets. Each of these datasets cluded the apothecial ascomycetes with was subjected to maximum-parsimony other re presentati ve ascomycetes. This analysis to determine a MPT. From the alignment of 1750 nuc1eotides had a total of MPTs of these replicate datasets, we calcu 4% variable sites and 307 informative sites. lated the percentage of MP'fs which had a Initial analyses showed that the sequence particular branch as an assessment of sup. of C. tricolor had man y apomorphic port for each internal branch. changes in its SSU rONA sequence which caused " Iong-branch-attractioo" problems (leading to spurious convergences): for this RESULTS reason C. tricolor was excluded from the We sequenced over 1600 nucleotides. or determination of the MPT. The MPT (Fig. over 90%, of SSU rONA from each of 10 1) required 1095 steps; adding one step pro taxa currently recognized as members of duced 9 equally parsimonious trees; adding the four orders Caliciales. Lecanorales, Le two steps added 28 equally parsimonious otiales. and Pezizales. These four orders re trees. Each branch supported by bootstT'dP flect the diversity of morphologies and nu at greater than 50% (derived from 500 boot tritional modes (saprobic, parasitic, and li strap replications) are indicated in Fig. I. chen-forming) found within the apothecial This analysis suggests that the apothecial ascomycetes. The nucleotide seque nces ascomycetes radiated early in the evolution were obtained by direct sequencing of PCR of the filamentous ascomycetes and that products from specific amplifications of cleistothecial and perithecial forms were the SSU rONA. To assure the accuracy of derived from within this lineage. The group thi s data, both strands of the SSU rONA which includes apothecial ascomycetes is were sequenced (except for a short region separated by 35 synapomorphies from the near the 5' and 3' ends, close to primers outgroups N. vitellina and the yeast S. ce UCBNSI7 and UCBNS24). These nucle revisiae. The order Pezizalcs is supported otide sequences were comparable to those by 17 synapomorphies. the order Leotiales of the SSU rONA reported for other fungi is supported by 10 synapomorphies. includ from the division Ascomycota. Although ing the Erysiphales (powdery mildew) rep most eukaryotic SSU rONA sequences resentative B. graminis. The cleistothecial have been approximately 1800 nucleotides and perithecial ascomycetes are each sup long, the PeR products from the isolates ported by 41 and 43 synapomorphies. re M. albonigrum, L . dispersa and P. crus/u speclively, a result consistent with the con lata were longer, with L. dispersa nearly clusions of Berbee and Taylor (1992). double the expected length. Upon sequenc ing , we found that the increase in length DISCUSSION was due to discrete insertions of 78-388 nu Analysis for Monophyly of Each of c1eotides. A total of 15 insertions were the Orders found with between I and 8 insertions per isolate (Gargas et al., 1995) , several of these Phylogenetic analysis of the SSU rONA insertions being group I inlrons (Gargas sequences from representative asco- PHYLOGENY OF ASCOMYCETES FROM SSU rONA 11

ChaelOmium datum '00 _~.~oo:.~:~~==-~Neurospora crasS-----< Leucostoma persoonii By.ssochlamys nivea EurO( ium rubrum

~-,~00;'''2!~=--99 Ascosphaera apis Eremascus albus Mycocalicium albonigrum Porpidia cruslulata Sphaerophorus globosus Lecanora di spersa LeOlia lubrica Cudonia confusa Spathularia naviJa Sclerotinia sclerotiorum " Blumeria graminis Gyromitra esculenta Morchella elata Inennis aggregata Plectania ni grella Peziza bOOia Ascobolus lineolatus Neolec ta vi tellina Saccharomyces ccrtvisiae FI(;. I. The single most pa rsimonious tree based on SSU rONA se quence data ( 1750 al igned nu cleotides. 1097 steps, consistency index - 0.6(4). Boolstrolp percentages >50% based on 500 repli cations are shown over thickened branches. Distance is rel ative to line length (scale on figure). mycetes suggests that the class " Disco· consistently recognize each order as mono mycetes" is paraphyletic and basal to the phyletic. The basal taxa within each order cleistothecial and perithecial fungi. How· (i.e .• A. lineo/atus or Le. lubrica) having ever, the same analysis supported most of many apomorphic changes, were subject to the traditional orders, recognized by char long branch attraction to taxa in other or· acteristic ascomata and ascal tip structures, ders. The inclusion of several diverse taxa as monophyletic, The order Lecanorales is withi n each order reduced these long monophyletic if the taxon S, globosus (for branches to a series of hi.erarchical inter· merly Caliciales) is included ; the order Le· nodes and resolved the clades. With proper otiales is supported as paraphyletic. includ taxon sampling we may be able to include ing the powdery mildew B, graminis (for· problem taxa such as the long-branched C. merly Erysiphales). Taxon sampling is tricolor. critical to the resolution of monophyly of The order Pezizales or true cup fungi in each order in parsimony analyses; analyses cludes species with minute, simple cups as which sampled fewer taxa, or a less diverse in A.lineolatus, larger cups such as those in range of taxa within each order, did not Pe. badia, as well as species with large, 12 GARGAS AND TAYLOR convoluted , ascocarps such as Mo. eta/a apomorphic c hanges resulted in long and G. esculenta. Their asci are usually branch attraction between Le. lubrica and operculate, though this is not evident in all other nonLeotialean taxa. members. Typicall y the Pezizales are The Caliciales are a group of less than 400 saprobic, though some form mycorrhizal species of mostly lichen-forming fungi char associations. The MPT shows support for acterized by stalked ascomata or mazaedia. the Pezizales representati ves as a mono Their asci are simple. or prototunicate. and ph yletic group. Bootstrapping supports the the ascospores are released passively in a sister taxa relationships of G. eSC/J. lenta and powdery mass. We included these fungi to Mo. elata, of I. aggregata and PI. nigre/la, test whether mazaedia are a derived char as well as a group of all four of those taxa; acter that can delineate this group. Repre relationships predicted by ultrastructural sentatives of the lichen-forming S. globo characters (J. Kimbrough, personal com sus, C. tricolor, and the saprobic M. albo munication), The MPT places A.lineo/atu$ nigrum were sequenced from this group. basal to the other Pezizales, its branch in The MPT was determined without C. tri cludes 63 apomorphic changes. As the most color, as long branch attraction made the basal of the Pezizales, A . lineola/us should position of C. tricolor ambiguous. The MPT be examined for characters it may share placed S. gloho.~us as a sister taxon to L. with the basal members of other apothecial dispersa within the Lecanorales; M. albo orders, such as simple ascomata. The nigrum was placed on a lineage between the branch leading to the Pezizales including A. Lecanorales and c1eistothecial and perithe lineolatus was not supported by bootstrap cial ascomycetes. The MPT does not sup ping, as is typical for clades including taxa port recognition of the Caliciales including with many apomorphic changes. S. globosus. When we used the MPT as a The Leotiales produce small , fleshy backbone constraint (PAUP 3. 1.1) to deter apothecia whose asci typically open with a mine the most parsimonious placement of simple pore . This group includes saprobic C. tricotor, C. tricolor was a sister taxon to fungi and also many important plant patho Le. tubrica with a long branch of 87 autapo gens such as species of Sclerorinia. The morphies. We consider the placement of C. branch leading to this assemblage of Leo tricolor to be ambiguous as the result of its tialean fungi was supported in the MPT , yet many autapomorphies. Careful taxon sam bootstrapping supported only the branch pling of more representati ves of the Cali leading to Cu. confusa and Sp.flavida, con ciales is critical to further testing of these sistent with the results of Landvik et al. hypotheses. The MPT suggests that mazae (1993). Interestingly, the powdery mildew dia presence is not sufficient to delimit this (formerly Erysiphales) B. graminis grouped group. and may have arisen multiple times with Sc. sclerotiorum , a result consistent within the ascomycetes. Kimbrough (1970) with Saenz et al. (1994). Further research classified the caliciaceous fungi as the fam should address whether other Erysiphales ily Caliciaceae within the Helotiales (here have their closest relatives within the Leo considered to be Leotiales), placement of tiales. The most basal taxon in Leotiales C. tricolor in the Leotiales could not be ex was the type-genus Leotia . Previous analy cluded based on its SSU rDNA sequence. ses, including fewer Leotialean taxa, have The Lecanorales is a large and diverse had difficulties resolving Leolia as mono order containing most of the lichen-forming phyletic with other members of this order fungi. Lichen thalli from this group may be (Gargas, 1992; Landvik et at., 1993 ; Saenz crustose, foliose. or fruticose. Lecanorales et at., 1994). Probably the large number of has the greatest diversity of ascal types of PHYLOGENY OF ASCOMYCETES FROM SSU rONA Il the four orders examined; some elaborate the hypothesis that cleistothecia were basal asci have been described as "nearl y tritu to perithecia and that the open apothecia of ni cate" (Hawks worth, 1985). The MPT the discomycetes arose from within perith shows that L. dispersa and P. crusflliata ecial forms (Luttrell, 19.5.5), but has been are monophyletic, if one includes the cali supported by other recent molecular analy ciaceous S. globosus. Although S. gfobosus ses (G. Haase, personal communication). produces mazaedia, it has asexual charac The MPT suggests that within the apo ters similar to those found in the Lecano thecial ascomycetes th e Pezizales are rales (J. Hafellner and L. Tibell , personal basal , and this group may have retained communi cation). Other members of the traits whi ch are primi tive for the asco Sphaerophoraceae should be examined to mycetes. The Pezizales are distinguished see if they should be recognized as mem by operculate asci and ascospores, which bers of the Lecanorales. are single-celled and often highl y orna mented. Developmentall y, the Pezizales lack spermatia and the vegetative primordia Evolution of Ascomata within the within which ascogonia develop, further Filamentous Ascomycetes supporting a separation of he Pezizales from the inoperculate di scomycetes and from the perithecial ascomycetes (Kim The MPT suggests that the apothecial as brough, 1981). Within thi s phylogenetic comycetes diverged after the outgroups context evolution of lifestyle may be con represented by N. vitellina and the yeast sidered. 1t is interesting that the Pezizales Sa. cerevisiae. The distinction of represen contain rew plant parasites. 1f the habit of tatives of N. vitellina as separate from the parasitism arose, perhaps many times, in other ascomycetes is supported by their other lineages, it is reasonable that basaJ taxa lack of paraphyses and the unusual staining would retain an ancestral saprobic lifestyle. of their asci. This analysis suggests that Based on the MPT the lichen habit arose paraphyses evolved wit hin th e ascomy in the ascomycetes after the divergence of cetes before divergence of the apothecial Pezizales and Leotiales. It is likely that the form s and brings into question whether the lichen habit was acquired separately in ancestral ascomycetes were yeast-like or other groups such as the Arthoniales (A. filamentous. The position of the apothecial Tehler, personal communication). U ascomycetes as paraphyletic and ancestral chenologists have recently used ph yloge to the c1eistothecial and the perilhecial netic methods to examine relationships of fungi supports the hypothesis that fungi the lichen-forming fungi . Hafellner's (1988) with relatively undifferentiated apothecia classification incorporated both free-living gave rise to derived cleistothecia and per and lichen-forming fungi . Using ascal char ithecia as proposed by Nannfeldt (1932). acters Hafellner grouped the Leotiales (as Nannfe ldt suggested a gradual transition Heloliales), Caliciales, and Pezizales (with between apothecia and peri thecia, and tran functionall y unitunicate asci), as separate sitory forms may be identified when more from the Lecanorales (wh ich have ascal taxa of intermediate positions are included tips with internal apical beaks). The MPT is in molecular studies. The Eurotiales are a consistent with Hafellner's grouping. Both derived group within the filamentous asco Hafellner (1988) and Tehler (1988, 1990) mycetes even though they have many re treated the lichen-forming fungi along with ductions in morphology such as naked as saprobic and parasitic fungi and concurred cogonia and simple asci. The derived state that the lichen-forming habit was wide of perithecia from apothecia differs fro m spread within the ascomycetes and could 14 GAROAS AND T AYLOR

not be used to separate these fungi into a evolution. Building on this siable phylog monophyletic group. eny for the ascomycetes. which recognizes morphologicall y supported orders. we can CONCLUSION exlend our analyses to include problemalic The most parsimonious tree based on taxa such as those with highly derived and SSU rONA gene sequences implies an reduced morphology or those thai lack te early radiation of apothecial fungi within leomorphic states. The inclusion of more the fil amentous ascomycetes (Fig. 2). The taxa from other di verse groups of asco c1eistothecial and pcrithecial ascomycetes mycetes will reveal other detai ls of fungal have undergone more recent radiations. N. evolution. lIiteiUna represents a basal lineage whic h is distinct from the apothecial ascomycetes. AC KNOWLEDGMENTS The MPT supports monophyly for the or We thank P. T . DePriest for providing support with ders Pezizales and Leotiales and the Lecan Ihe analyses and manuscript. M. L. Berbee sequenced orales (including S. globoslIs). The Cali many of the pe rithetial and cleistOlhetial fungi used dales as currently recognized (Tibell. 1984) for tomparisons and provided technical advice and support . Cu ltures or herbarium material we re ge ner. could not be supported as monophyletic ; ously provided by O. E. Eriksson, L. M. Kahn, L we propose that this group may need redef Tavares, and L. Tibell and prepUblication phyloge inition . Wit h this view of the apothecial netic information by J. Haase. Helpful comments were fungi as an unnatural assemblage. we are provided by O. E. Eriksson, D. L. Hawksworth, free to form new hypotheses of ascomycete J. w. Kimbrough . and K. L. O·Oonnell. P. T . Spieth, I. Tavares, M. Grube. and two anonymous reviewers provided manustript suggestions. This work was par. tiall y supported by a Smithsonian Institution Postdoc --{~== SOll&AII)A~. IIRI:SAl.f.~ toral Fellowship to A.O. (Scholarly Studies Grant) and r NIH RO I AI 2S343 and NSF BSR 9007 141 to J.W.T.

EuMOTI~U .~ &Au . II .~ REFERENCES BERBEE , M. L., ANOTAV L.OR , J. W. I992. Two Asco mycele classes based on fruiting-body characters and ribosomal DNA sequence. Mol. Bioi. E,·ol. 9: 27S-284. BERB EE , M. L., ANO TAYLO R. J. W. 1993 . Dating the ~volutionary radiations of the true fungi . Call. J. L EOT I ~lf.S BOI. &ERI$YPlIAI.fS 71: 1114-1127. CA MIN , J. H. , ANO SOKAl, R. R. 1963. A method ror deducing b'-dnching sequences in phylogeny. Evolu tion 19: 311 -326. CHAMBERS, C., DUlTA, S. K., AND CROUCH, R. J. 1986. N,uraspora crassa ribosomal DNA : sequence of internal transcribed spacer and comparison with NlOUCTALES N. illUrmt Jia and N. silophila. Gtllt 44: 159-164. S ACOIARO .... Vct.-r AUS ERI KSSON. O. E. 1981 . Origin and evolution of the As FIG . 2. Phylogeny of the subdivision Ascomytotina comycetes. Op,," BOI . 60: 175-209. suggested by the most parsimonious tree from SSU ERlItSSON. O. E., AND HAW KSWOIlTH, D. L. 1993 . rONA sequentes. This phylogenetic hypothesis sup Outline of the Ascomycetes 1993. Sysuma Asco ports the traditional orders PezizaJes. uotiales. and mycttum 12: 31-257. Utanorales (intluding S. globosus) and suggests that FEL.Sf.NSTEIN, J. 1985. Confid ence li mits on phyloge Erysiphales is wit hin the Leotiales. The apothetial as· nies: An approach using the bootstrap. Evolution 39: comycetes are paraphyletit and basal to the perithetial 783-791. and cleistothetial ascomycetes which are of compara GA RGAS, A. 1992. PhylogtllY of Discomyctles tIIIJ tively recent divergence. Early Rodiuliolls of ,h, Filamt nlous Ascomycttts PHYLOGENY OF ASCOMYCETES FROM SSU rONA 15

Ifl/erred from 18S rDNA Data. Ph.D. dissertation, ferred from 18S rDNA sequences. Syslema Asco University of California, Berkeley. 99 pp. mycetum II: 107-11 5. GA RGAS, A ., AND TAYLOR , I . W. 1992. Polymerase LEE, S. B., AN D TAYLOR, J. W. 1990. Isolation of chain reaction (PCR) primers for amplifying and se DNA from fungal mycelia and single spores. In PCR quencing 18S rONA from licheni zed fungi. Mycafo PrO/ocols: A Guide to Methods and Applications. gia M: 589-592. (M. A. Innis, D. H. Gelfand, I . I . Sninsky, and GARGAS , A., DEPRIEST, P. T., AN D TAYLOR, J. W. T . J. White, Eds.), pp. 282-287. Academic Press, 1995. Positions of multiple insertions in the SSU New York. rONA of lichen-forming fungi . Mol. BioI. EYQI., in LUTTRELL, E. S. 1955. The ascostromatic Asco press. mycetes. Mycologia 47: 511-533. GYlLENSTEN, U. B., AND ERLICH, H. A. 1988. Gen NANN FELDT, J. A. 1932. Studien uber die Morpholo eration of single-stranded DNA by the polymerase gie und SY$tematik der Nicht-licheni sierten inoper chain real;tion and its application to direct sequenc culaten Discomyceten. Nova Acla Regia~ Soc. Sci. ing of the HLA-DQA locus. Proc. Natl. Acad. Sci. Upsali~nsis (S~r. IV) 8: 1-368. USA 85: 1652- 1656. RUBTsov, P. M ., MU SA KHANOV, M. M., ZA HAFELL NER, I. 1988. Principles of classification and KHARYEY, V. M. , KRAYEV, A. S., SK RYAB1 N, main taxonomic groups. In CRC Handlmok of Li K. G., AND BAYEY, A. A. 1980. The $tructureofthe chenofogy (M. Galun, Ed.), Vol. 3, pp. 41-52. CRe yeast ribosomal RNA genes. I. The complete nucle P-ress, Boca Raton, FL. otide sequence ofthe 18S ribosomal RNA gene from HA WkSWORT H, D . L., SUTTON, B . C. , ANO Saccharomyces cerevisiae. Nucleic Acids Res. 8: AINSWORTH, G. C. 1983. Ainsworth and Bisby's 5119-5794. Dictianary afthe Fungi. CAB, Kew. SAENZ, G. S., T AY LOR, J. W., AND GA RGAS, A. 1994. ISS rRNA gene sequences and supraordinal classi HAW KSWORTH, D. L. 1985. Problems and prospects in the systematics of the Ascomycotina. Proc. In fication of the Erysiphales. Mycologia 86: 212-216. SWOFFORD. D. L 1991. PAUP: Phylogen~l jc Analy dian Acad. Sci. 94: 319-339. sis Using Parsjmony, Versjon 3.1 . Computer pro KIMBROUGH, J. W. 1970. Current trends in the classi· gram distributed by the Illinois Natural History Sur fication of the discomycetes. Bot. Rev. 36: 91-161 . vey, Champaign, IL. KIMBROUGH, J. W . 1981. The Discomycete centrum. TEHLER , A. 1988. A cladistic outline ofthe Eum ycota. In Ascomycete Systematics, the Lulfrellian Concept Cladistjcs 4: 221-271. (D. R. Reynold s, Ed.), pp. 72-101. Springe r-Verlag, TEHLER, A. 1990. A new approach to the phylogeny of New York. Euascomycetes with a cladistic outline of Artho KORF, R. P. 1973. Discomycetes and Tuberales. In niales focusing on Roccellaceae. Can. J . Bot. 68: The Fungi, an Advanced Treatise (G. C. Ainsworth, 2458-2492. F. K . Sparrow, and A. S. Sussman, Eds.), Vol. 4A, TIBELL. L. 1984. A reappraisal of the taxonomy of pp. 249-319. Academic Press, New York. Caliciales. Bejh. Nova Hedwigia 79: 597-713. KU SU KAWA, N., UENDitl, T., ASA DA, K., AND KATO, WHITE , T . J., BRU NS, T., LEE, S. B., AND TAY LOR, I. 1990. Rapid, reliable protocol for direct sequenc J. W. 199()' Amplification and direct sequencing of ing of material amplified by the PCR. Biol/'chniques fungal ribosomal RNA genes for phylogenelics. In 9(1): 66-12. PCR Protocols: A Guide to Methods and Applica LANDVIK , 5., ERIKSSON, 0., GARGAS , A., AND tions. (M. A. Innis, D. H. Gelfand, I. J. Sninsky, GUSTAFSSON , P. 1993. Relationships of the genus and T. J. White, Eds.), pp. 315-322. Academic Neo/ecta (Neolectales ordo. nov., Ascomycotina in- Press. New York.