mycological research 113 (2009) 1322–1334 journal homepage: www.elsevier.com/locate/mycres Evolutionary relationships between aquatic anamorphs and teleomorphs: Tricladium and Varicosporium Jinx CAMPBELLa,*, Ludmila MARVANOVÁb, Vladislav GULISc a Department of Coastal Sciences, University of Southern Mississippi, Ocean Springs, MS 39564, USA Czech Collection of Microorganisms, Institute of Experimental Biology, Faculty of Science, Masaryk University, Tvrdého 14, 602 00 Brno, Czech Republic c Department of Biology, Coastal Carolina University, Conway, SC 29528-6054, USA b article info abstract Article history: Tricladium, with 21 accepted species, is the largest genus of aquatic hyphomycetes. It encom- Received 5 May 2009 passes species with dematiaceous as well as mucedinaceous colonies. Conidiogenesis is thal- Received in revised form loblastic; conidiogenous cells proliferate percurrently or sympodially. Conidia have typically 28 August 2009 two alternate primary lateral branches. Fontanospora and Variocladium are segregates of Tricla- Accepted 3 September 2009 dium, differing by conidial branching. Varicosporium comprises nine species, one not well Available online 10 September 2009 known. Conidiogenesis is blastic or thalloblastic, conidiogenous cells proliferate sympodially Corresponding Editor: or are determinate; conidia regularly produce primary and secondary branches and often frag- Marc Stadler ment into part conidia. Molecular analyses on the 28S rDNA of 86 isolates, including 16 species of Tricladium, five species of Varicosporium, two species of Fontanospora and one species of Var- Keywords: iocladium, place these hyphomycetes within Helotiales. Tricladium is polyphyletic and placed in Aquatic hyphomycetes six clades; Varicosporium is polyphyletic and placed in three clades; Fontanospora is polyphyletic Fontanospora within a single clade. Variocladium is placed with poor support as a sister taxon to Varicosporium Molecular taxonomy giganteum, Hymenoscyphus scutula and Torrendiella eucalypti. Morphology ª 2009 The British Mycological Society. Published by Elsevier Ltd. All rights reserved. Variocladium Introduction Previously published results, based on teleomorph–anamorph relationships and on molecular analyses of various genes of rDNA, place the ascomycetous aquatic hyphomycetes in the subphylum Pezizomycotina in five classes, viz Dothideomycetes, Leotiomycetes, Orbiliomycetes, Pezizomycetes and Sordariomycetes, and the basidiomycetous aquatic hyphomycetes in two classes of Basidiomycota, viz Urediniomycetes and Hymenomycetes (Marvanová 2002, 2007). In the proposed new classification of fungi (Hibbett et al. 2007) they appear in the subkingdom Dicarya; the aquatic ascomycetous members (66 taxa) are distributed in the same classes as mentioned above. Most of them (42) show affinity to Leotiomycetes, eight to Orbiliomycetes, seven to Sordariomycetes, eight to Dothideomycetes and one to Pezizomycetes. This contribution continues our study of phylogenetic relationships of aquatic hyphomycetes. Previously (Campbell et al. 2006), using molecular analyses of 28S rDNA, we have shown that species of Lemonniera, Margaritispora and Goniopila support the hypothesis of convergent evolution of the asexual propagules in aquatic hyphomycetes (Ingold 1975), very probably due to their specific adaptation to the life in a lotic water environment. Species of Lemonniera, a morphologically relatively homogeneous holoanamorphic taxon with phialidic conidiogenesis, appeared to belong to two quite distinct clades (Leotiomycetes and Dothideomycetes) of ascomycetes. * Corresponding author. Tel.: þ1 228 818 8878; fax: þ1 228 872 4264. E-mail address: jinx.campbell@usm.edu 0953-7562/$ – see front matter ª 2009 The British Mycological Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.mycres.2009.09.003 Evolutionary relationships of Tricladium and Varicosporium This time we have chosen two morphologically heterogeneous genera with thalloblastic (Hennebert & Sutton 1994) or blastic conidiogenesis: Tricladium (with some segregates) and Varicosporium. Conidia of both are ‘‘open branching systems’’ in the sense of Kendrick (2003), and can be understood as modified hyphae with the function of propagules (Descals 1985). The inter-generic differences are not quite clear; there are species whose classification in one or the other genera is debatable (e.g. Tricladium indicum, Tricladium terrestre, Varicosporium tricladiiforme, and the anamorph of Hymenoscyphus varicosporoides). Tricladium, with 21 recognized species, is the largest anamorph genus among aquatic hyphomycetes. It is based on Tricladium splendens (Ingold 1942). The generic characters are branched conidia, typically consisting of an axis and two broadly divergent primary alternate branches, developing in acropetal succession. Conidiogenesis is thalloblastic, i.e. the conidial axis during its development is integrated with the conidiogenous cell. The species are delineated on: (1) the degree of conidiophore branching (from simple to profusely branched); (2) the conidiogenous cell proliferation (percurrent, sympodial or absent); (3) the shape of the conidial axis and branches (parallel-walled, tapering distally, nearly straight, curved or geniculate, apices rounded or acute, branch insertion constricted to varying extent or unconstricted); and (4) the presence of a parabasal axial appendage before conidial secession. Colonies may be mucedinaceous or dematiaceous. Some species of Tricladium produce occasional secondary branches in culture, but it was described as a specific character for T. terrestre as this character also occurs in specimens in nature (Park 1974). Tricladium is morphologically heterogeneous. With an increasing number of species, some also with more than two primaries and with secondary branches, it became difficult to define clearly its scope and to delimit them against similar taxa. To reduce the morphological heterogeneity there were several attempts to segregate some species from Tricladium into new genera. Iqbal (1974) erected Scorpiosporium, based on Scorpiosporium minutum, where he also included Tricladium angulatum. The main distinguishing characters were the ‘scorpioid’ (geniculate) conidial axis and unconstricted branch insertion. However, these features are not always linked (cf. also Ando & Kawamoto 1985); there are species with geniculate axis and constricted branch insertion (e.g. Tricladium patulum) and, moreover, the proliferations of the conidiogenous cells are percurrent in S. minutum, but sympodial in T. angulatum. Therefore Marvanová & Descals (1996) recombined S. minutum in Tricladium. Dyko (1978) segregated Tricladium eccentricum into a new genus, Fontanospora, together with a new species Fontanospora alternibrachiata. The main distinguishing features from Tricladium are the median constriction of the conidial axis and two branches, inserted closely to each other, one below and one above this constriction. Variocladium was published by Descals et al. (1998) to accommodate two similar species, Tricladium giganteum and Scorpiosporium rangiferinum, with large conidia of variable branching pattern (alternate and opposite primaries, and also secondaries appearing in situ) and acute distal ends. The conidiogenous cells are percurrently proliferating in the former and determinate in the latter. A spermatial state resembling Phialocephala dimorphospora was reported in 1323 Variocladium giganteum by Willoughby & Minshall (1975) as well as in Variocladium rangiferinum (Descals & Webster 1982). Three Tricladium species have known teleomorphs, all classified in Helotiales: Hymenoscyphus splendens (Helotiaceae) is the teleomorph of T. splendens (Abdullah et al. 1981), Hydrocina chaetocladia (Hyaloscyphaceae) is the teleomorph of Tricladium chaetocladium (Webster et al. 1991), and Cudoniella indica (Helotiaceae) was described as the teleomorph of Tricladium indicum isolate from South Africa (Webster et al. 1995), but the conspecificity of Webster’s isolate with the type of this species described from Himalaya (India) was doubted by Sivichai et al. (2003, see below). Phialidic andromorphs have been described in T. chaetocladium, Tricladium curvisporum, Tricladium minutum, Tricladium obesum, Tricladium robustum, T. splendens and T. terrestre. Varicosporium is based on Varicosporium elodeae (Kegel 1906). Nine species are accepted at present, one of them not well known and by some authors classified in Tricladium (Sivichai et al. 2003). Varicosporium is characterized by branched conidia with rarely nearly straight, but often variously curved conidial axis and by regularly formed, diverging primary and secondary (in some species also of higher order) conidial branches. Unlike in Tricladium, the conidia have a strong tendency to fragment into simpler part conidia, often similar to those of a Tricladium. Conidiogenesis in V. elodeae is blastic, whereas it is thalloblastic in the rest of Varicosporium species. There is no species of Varicosporium with percurrent proliferation of conidiogenous cells known at present. In V. elodeae the conidiogenous cells do not elongate after having produced the first conidium and the subsequent conidia arise retrogressively (or randomly) down the conidiophore. In the rest of the species there are sympodial elongations or no growth of the conidiogenous cells after release of the first conidium. The species are distinguished by the curvature of the conidial elements (nearly straight, arcuate, helical or geniculate) by the morphology of conidial elements (parallel-walled, subulate, slightly clavate), by the shapes of apices (rounded or acute) and by branch insertion (constricted or unconstricted). The only published teleomorph in Varicosporium is H. varicosporoides, but the classification of the anamorph in Varicosporium was doubted by Sivichai et al. (2003) and Boonyuen et al. (2006). Marvanová (unpubl.) has obtained minute apothecia of a leotiomycetan discomycete in a mating experiment with two isolates of V. giganteum from Canada. A phialidic andromorph is known in this species (Crane 1968) and was discovered in Varicosporium scoparium (Marvanová unpubl.). The aims of this study were: 1) to test the hypothesis that the morphologically defined taxa are supported by molecular analyses of LSU rDNA nucleotide sequences; 2) to establish phylogenetic relationships of the anamorphic species to ascomycete clades; 3) to establish the relationships of the pleomorphic aquatic taxa with their terrestrial relatives. Materials and methods Collection, isolation and characterisation Collections were made from streams in Europe and North America. A loopful of fresh foam was streaked onto a microscope slide coated with a thin layer of malt extract agar 1324 (0.1–2 % MEA plus 100 mg L1 chloramphenicol or penicillin/ streptomycin [Sigma, St. Louis, MO]) and kept at 10–15  C. When isolated from submerged plant litter (USA isolates), material was first incubated in sterile distilled water in Petri dishes to induce sporulation and then suspended conidia were transferred onto 0.1 % MEA with 200 mg L1 of streptomycin and 200 000 units/L of penicillin G. After 24–48 h germinating conidia were transferred to MEA plates supplemented with antibiotics and incubated at 10–18  C (Marvanová & Bärlocher 1998; Marvanová & Gulis 2000). Cultures were checked for bacterial and fungal contaminants, subcultured to MEA plates and incubated at 15  C. All isolates were monoconidial. DNA extraction, sequencing and cladistic analysis Mycelia were harvested directly from MEA plates, incubated with 200 units of lyticase at 30  C for 4 h, and ground at hourly intervals during incubation using a micropestle. RNaseA (40 mg) was added and incubated for 20 min at 65  C and genomic DNA extracted using Qiagen’s DNeasy Plant Mini Kit (Qiagen. 2004) following the procedure by Raja et al. (2003). The 50 end of the 28S ribosomal gene was amplified with Taq PCR Master Mix Kit (Qiagen. 2002a) using fungal primers LROR (Bunyard et al. 1994) and LR6 (Vilgalys & Hester 1990). The PCR products were cleaned up with Qiaquick PCR Purification Kit (Qiagen. 2002b) and sequenced directly. Sequences (Table 1) were aligned with published sequence data (Table 2) using Clustal X (Thompson et al. 1997), and then refined manually in Se-Al (Rambaut 1996). The rationale for the taxon sampling followed Campbell et al. (2006). Additionally, the sequences derived in this study were entered into BLAST, NCBI (Altschul et al. 1997) to help identify sequence similarity to other taxa and to determine which orders and families should be included in the phylogenetic analyses. Bayesian Metropolis coupled Markov chain Monte Carlo (B-MCMCMC) analyses were performed with the general time reversible model of substitution model (Rodriguez et al. 1990) with invariant sites and gamma distribution (GTR þ I þ G) using MrBayes 3.0 (Huelsenbeck & Ronquist 2001): searches were conducted for a total of 2 000 000 generations with phylogenetic trees sampled every 100 generations. Three independent B-MCMCMC analyses were conducted to verify likelihood convergence and burn-in parameter. Out of 20 001 resulting trees from each analysis, the initial 3111 trees were identified as burn-in prior to the convergence of likelihoods and were excluded from post-run analyses. The majority rule consensus tree from 16 890 trees was generated with average branch lengths (Fig 1). Maximum parsimony (MP) analyses were performed using PAUP*4.0b 10 (Swofford 2002). A two-step search approach was employed in an attempt to avoid local optima (Olmstead et al. 1993). Step-one consisted of 100 heuristic replicates with random starting trees, random stepwise addition and tree-bisection–reconnection branch swapping with MulTrees off and a maximum of two trees saved per replicate. All of the shortest trees from these initial runs were saved and then used as starting trees for the second step, which consisted of searches with MulTrees on and the maximum number of trees set to 10 000. Gaps were treated as missing data. J. Campbell et al. Nodal support was assessed by non-parametric bootstrapping and Bayesian posterior probabilities. Bootstrap values (Felsenstein 1985) were calculated from 1000 replications using a heuristic search on 100 replicates with random starting trees, random stepwise addition and MulTrees off. Posterior probabilities were calculated from the majority rule consensus tree of the 20 000 trees from the Bayesian analyses, with the initial 3111 trees excluded as burn in prior to the convergence of likelihoods. Results and discussion Initial analyses were performed on w250 species with representatives from 17 orders of ascomycetes, and with basidiomycetes as outgroup taxa (data not shown). All species of Tricladium (16), Fontanospora (2), Variocladium (1) and Varicosporium (5) treated in this study are placed among the Leotiomycetes. These placements were also supported in the BLAST searches. Further analyses were run with 86 taxa, including our 45 isolates; members of the Pezizomycetes were used as outgroup taxa. The Bayesian tree (Fig 1) indicates that Tricladium is polyphyletic and placed in six clades; Varicosporium is polyphyletic and placed in three clades; Fontanospora is polyphyletic within a single clade; Variocladium is placed with poor support as a sister taxon to Varicosporium giganteum, Hymenoscyphus scutula and Torrendiella eucalypti. The first group (Figs 1, 2) includes Tricladium species with dark colonies (Tricladium clade 1). These five species (Tricladium castaneicola, Tricladium indicum, Tricladium obesum, Tricladium splendens and Tricladium terrestre) form a relatively homogeneous group from the point of colony color and of conidial morphology. They all have dark colonies; conidiophores in culture are mostly lateral, relatively short, often unbranched (except in T. splendens, where they typically consist of a stipe with acrotonous branches); conidiogenous cell proliferation is percurrent (in T. splendens the type of the genus, and in T. obesum), sympodial (in T. castaneicola and T. terrestre) or they are determinate (in T. indicum); conidia are medium sized to large (span ca. 100–200 (400) mm, smaller in T. obesum), and the conidial elements are mostly 5–11 mm wide (except in T. castaneicola); secondary branches are occasional or regularly present (in T. terrestre), branch insertion is constricted to varying extent; phialidic andromorph is known in T. obesum (Marvanová 2004), T. splendens and T. terrestre (Descals & Webster 1982). However, four Tricladium species with dark colonies appear in at least three other clades (see below). The two strains of T. splendens treated here were also sequenced by Baschien et al. (2006, Fig 3). Neighbor joining analyses on the 18S rDNA placed T. splendens in a well supported clade with the helotialean aquatic anamorphs Anguillospora crassa, Anguillospora furtiva, Anguillospora fustiformis, Anguillospora mediocris and Geniculospora grandis. The same grouping was found with another isolate of T. splendens by Belliveau & Bärlocher (2005, Fig 2) using maximum parsimony and weighted parsimony on the 18S rDNA. Baschien et al. (2006, Fig 15, ITS1–5.8S–ITS2 rDNA sequences, maximum likelihood) found a close relationship to Zalerion varium (which according to Bills et al. (1999) may be a member of Helotiales), and to Evolutionary relationships of Tricladium and Varicosporium 1325 Table 1 – Fungal isolates used in this study. Species Culture no.a Country of origin Fontanospora eccentrica Fontanospora eccentrica Fontanospora fusiramosa Fontanospora fusiramosa Fontanospora fusiramosa Hydrocina chaetocladiab Tricladium angulatum Tricladium angulatum Tricladium attenuatum Tricladium attenuatum Tricladium biappendiculatum Tricladium biappendiculatum Tricladium castaneicola Tricladium castaneicola Tricladium caudatum Tricladium caudatum Tricladium chaetocladium Tricladium chaetocladium Tricladium curvisporum Tricladium curvisporum Tricladium indicum Tricladium indicum Tricladium minutum Tricladium obesum Tricladium obesumb Tricladium patulum Tricladium patulum Tricladium patulum Tricladium procerumb Tricladium splendens Tricladium splendens Tricladium terrestre Tricladium terrestre Tricladium sp.1 Tricladium sp.2 Varicosporium delicatum Varicosporium delicatum Varicosporium elodeae Varicosporium elodeae Varicosporium giganteum Varicosporium giganteum Varicosporium scopariumb Varicosporium trimosum Varicosporium trimosum Variocladium giganteum CCM F-11402 CCM F-46394 CCM F-03680b CCM F-12900 VG 66-6 CCM F-10890 CCM F-00282 CCM F-14186 CCM F-06485 CCM F-10103 CCM F-13000 CCM F-19794 CCM F-11296 CCM F-13005 CCM F-13498 CCM F-21299 CCM F-03485 VG 23-1 CCM F-23387 VG 67-5w VG 112-1 VG 113-4 CCM F-10203 CCM F-13798 CCM F-14598 CCM F-15199 CCM F-17199 VG 8-1w CCM F-16786 CCM F-16599 CCM F-19087 CCM F-10101 CCM F-10201 VG 68-1 VG 69-2 CCM F-03977 CCM F-18499 CCM F-13589 CCM F-20087 CCM F-10987 CCM F-11287 CCM F-10303 CCM F-14398 CCM F-32694 CCM F-16686 UK Canada Slovak Republic Czech Republic USA UK Czech Republic Czech Republic Switzerland Canada Czech Republic Canada Czech Republic Portugal Czech Republic Czech Republic UK USA Canada USA USA USA UK Czech Republic Czech Republic Czech Republic Czech Republic USA Slovak Republic Czech Republic Canada Portugal Portugal USA USA Czech Republic Czech Republic UK Canada Canada Canada Spain Czech Republic Canada Slovak Republic GenBank accession no. GQ477304 GQ477305 GQ477306 GQ477307 GQ477308 GQ477309 GQ477310 GQ477311 GQ477312 GQ477313 GQ477314 GQ477315 GQ477316 GQ477317 GQ477318 GQ477319 GQ477320 GQ477321 GQ477322 GQ477323 GQ477324 GQ477325 GQ477326 GQ477327 GQ477328 GQ477329 GQ477330 GQ477331 GQ477332 GQ477333 GQ477334 GQ477335 GQ477336 GQ477337 GQ477338 GQ477339 GQ477340 GQ477341 GQ477342 GQ477343 GQ477344 GQ477345 GQ477346 GQ477347 GQ477348 a CCM ¼ Czech Collection of Microorganisms; VG ¼ Culture Collection of Vladislav Gulis. b Ex-isotype cultures. A. furtiva, A. crassa and Cudoniella sp. AY89371, confirming partly the results of the SSU rDNA analyses (see above). The teleomorphs Hymenoscyphus splendens and Cudoniella indica were respectively connected to T. splendens (Abdullah et al. 1981) and T. indicum, an isolate from S. Africa (Webster et al. 1995). We did not have sequences for H. splendens or C. indica, but we did have sequences for H. scutula, Cudoniella clavus and Cudoniella sp., however these were not placed in the clade with T. splendens and T. indicum. H. scutula was placed as sister taxon to V. giganteum in the second clade (see below), and Cudoniella sp. and C. clavus were placed in a clade with no anamorphic taxa (Fig 1). However, as pointed out in the Introduction, the anamorph–teleomorph connection between the type of T. indicum described from Himalaya by Sati & Tiwari (1992) and C. indica from South Africa (Webster et al. 1995) was questioned by Sivichai et al. (2003). They obtained an anamorph, similar to that isolated by Webster et al. (1995) from apothecia on wood baits in Thailand, which they identified as Hymenoscyphus varicosporoides. They suggested the Varicosporium anamorph of H. varicosporoides should be reclassified in Tricladium, but have not renamed it formally (cf. also Baschien et al. 2006). Later, Boonyuen et al. (2006) in their study based on analyses of the ITS rDNA of the ex-type cultures CBS 651.66 (H. varicosporoides from Japan derived from ascospores) and CBS 430.94 (T. indicum – here as C. indica – from Indian Himalaya derived from conidia), reported a high percentage 1326 J. Campbell et al. Table 2 – Sequences obtained from GenBank. Species Baeomyces placophyllus Brasiliomyces trina Bulgaria inquinans Chlorencoelia sp. Cudoniella clavus Cudoniella sp. Cyathicula coronata Cystotheca wrightii Dermea acerina Dibaeis baeomyces Erysiphe betae Fabrella tsugae Gyromitra esculenta Helicodendron conglomeratum Helvella lacunosa Heyderia abietis Hyaloscypha daedaleae Hydrocina chaetocladia Hymenoscyphus scutula Hymenoscyphus sp. Icmadophila ericetorum Lachnum cf. bicolor Lachnum virgineum Lambertella tubulosa Leveillula taurica Microsphaera pulchra Morchella esculenta Neofabraea malicorticis Ombrophila violacea Ostropa barbara Pezicula carpinea Phyllactinia moricola Podosphaera longiseta Potebniamyces pyri Rhytisma acerinum Siphula ceratites Spirosphaera floriformis Torrendiella eucalypti Typhulochaeta japonica Uncinula septata GenBank accession no. AF356658 AB022350 DQ470960 AY789351 DQ470944 AY789375 AF222491 AB022355 DQ247801 AF279385 AB079684 AF356694 FJ176906 AY856900 U42681 AY789289 AY789415 AY789412 AY789431 AF430278 DQ883694 AY544674 AY544646 AY616237 AB022387 AB022389 AF279398 AY544662 AY789365 AY584642 DQ470967 AB022401 AB022423 DQ470949 AF356696 DQ986775 AY616238 DQ195800 AB022415 AB183532 (99.5 %) of bp similarity between these specimens (see their Table 2), and suggested this indicated that H. varicosporoides had a Tricladium anamorph. Moreover, there is a third large-spored (holoanamorphic) species of Tricladium, Tricladium marylandicum, described from USA (Crane 1968) whose relationships with T. indicum and H. varicosporoides are also unresolved. All these three taxa are characterized by dark colonies and large conidia with 1–4 primary and (in culture) also secondary branches. Multi-gene molecular analyses employing all these species may help solve this problem. This first group also contains Lambertella tubulosa, Helicodendron conglomeratum and Spirosphaera floriformis. They also have dark colonies, but with their helicoform conidia are morphologically and ecologically quite different. They are called aero-aquatic hyphomycetes and unlike aquatic hyphomycetes, they thrive in standing or slow-moving waters, with low dissolved oxygen, and their conidia have air-trapping shapes, resistant against submerging, and adapted to dispersal on the water surface. Spirosphaera appears to be polyphyletic however, as Voglmayr (2004, partial nuc 28S and ITS1–5.8S–ITS2 rDNA) and Baschien et al. (2006, the same gene regions) placed Spirosphaera cupreorufescens in a well supported clade with Anguillospora longissima in the Dothideomycetes. The second group (Figs 1, 3, 4) contains Tricladium p.p.-Varicosporium p.p.-Fontanospora. It consists of two subclades: The first subclade (Figs 1, 3) contains Tricladium attenuatum, Tricladium biappendiculatum, Tricladium minutum, Tricladium patulum, Varicosporium elodeae, Varicosporium trimosum, Fontanospora eccentrica, Fontanospora fusiramosa. If we exclude T. minutum, which is on a poorly supported branch within this subclade, then the unifying features here are pale colonies, conidial branch insertion constricted to various extent, relatively narrow (usually up to 4 mm wide) conidial elements, absence of spermatial synanamorph and unknown teleomorph. Except for T. biappendiculatum, the conidiophore usually consists of a stipe and a fertile head which, if fully developed, is a multibranched structure with numerous conidia in lateral as well as terminal position. However, the branching pattern of the conidia is different: in Fontanospora there are two nearly opposite laterals inserted near the middle of a bent axis; V. elodeae and Varicosporium trimosum have conidia with primary and secondary branches (in the latter the conidia are relatively simple, consisting of an axis bearing one primary and one secondary branch, both inserted at right angle to their parent structures); the two Tricladium species have typical conidia for this genus differing by geniculate axis and blunt conidial apices in T. patulum against nearly straight axis and acute conidial apices in T. attenuatum. T. biappendiculatum also has acute conidial ends, but its conidiophores are short, often simple, and conidiogenous cell proliferation is percurrent. T. minutum is on a poorly supported branch within the above subclade. This may support its former segregation in Scorpiosporium by Iqbal (1974), but in morphology, there are no characters diagnostic enough on the generic level to justify its separation. T. minutum has brown colonies, percurrent conidiogenous cells, small to medium large conidia with geniculate axis, and typically two primary branches with unconstricted branch insertion. A phialidic andromorph was seen in one of the British isolates by E. Descals (Marvanová & Descals 1996). V. elodeae was sequenced by Baschien et al. (2006, Fig 3, 18S rDNA, neighbor joining, maximum likelihood and parsimony analyses) and was placed in a poorly supported clade with T. patulum, which is congruent with our results. Both strains of V. elodeae sequenced by Baschien et al. (2006), as well as those sequenced in our study, produce extracellular green pigments in agar cultures, and the conidia have parallel-walled axis and branches, typical for isolates from water. The second subclade differs from the species of the first subclade by having dark colonies. V. giganteum has large conidia (spanning over 200 mm), with geniculate axis produced on sympodial conidiogenous cells and primary and secondary branches with unconstricted insertion. Variocladium giganteum is placed with poor support as a sister taxon to the species within the second subclade. It is relatively isolated from other Tricladium species, indicating that its segregation might have been justified. It is unique in the large dimensions of conidia Evolutionary relationships of Tricladium and Varicosporium 1327 Fig 1 – B-MCMCMC majority rule consensus tree generated with average branch lengths from 16 890 trees inferred from Bayesian analyses on the 28S rDNA data. Maximum parsimony bootstrap supports and Bayesian posterior probability values greater than 50 % are given, respectively, at the corresponding nodes. and variable branching pattern (alternate or opposite). Colonies are dark, conidiogenous cells percurrent and conidial branch insertion is unconstricted. The third group (Figs 1, 5) contains Tricladium chaetocladium (including its teleomorph Hydrocina chaetocladia), Tricladium curvisporum, Tricladium sp.1 and sp.2 and Varicosporium delicatum. In our study, these species form a poorly supported clade within Helotiales. There are only two morphological characters which the four Tricladium species have in common: a sigmoid conidial axis and a phialidic andromorph in two of them. Otherwise they are culturally and morphologically quite different. T. chaetocladium, represented by two isolates from ascospores of the teleomorph and two from the anamorphic state, has V. delicatum as a sister taxon. However, their morphological Fig 2 – Group 1. Conidia from Tricladium clade 1. (A) Tricladium terrestre CCM F-10101. (B–C) Tricladium indicum VG 112-1. (D) Tricladium castaneicola CCM F-11296. (E) T. castaneicola CCM F-10605. (F) Tricladium splendens CCM F-16599. (G) T. splendens CCM F-19087. (H–J) Tricladium obesum CCM F-14598. (K–L) T. obesum CCM F-13798. Scale bar a (A–C) [ 100 mm, scale bar b (D–L) [ 50 mm. Evolutionary relationships of Tricladium and Varicosporium 1329 Fig 3 – Group 2, subclade 1. Conidia from Varicosporium clade 1. (A–B) Varicosporium elodeae CCM F-20087. (C) Varicosporium trimosum CCM F-14398. (D) V. trimosum CCM F-32694. Conidia from Fontanospora clade 1. (E–F) Fontanospora fusiramosa VG 66-6. (G) F. fusiramosa CCM F-03680b. (H) Fontanospora eccentrica CCM F-11402. (I) F. eccentrica CCM F-46394. Conidia from Tricladium clade 2. (J) Tricladium attenuatum CCM F-06485. (K) T. attenuatum CCM F-10103. (L) Tricladium patulum CCM F-17199. (M–N) Tricladium biappendiculatum CCM F-13000. (O–R) Tricladium minutum CCM F-10203. Scale bar [ 50 mm. 1330 J. Campbell et al. Fig 4 – Group 2, subclade 2. Conidia from Varicosporium clade 1. (A–B) Varicosporium giganteum CCM F-11287. Conidium from Variocladium clade 1. (C) Variocladium giganteum CCM F-16686. Scale bar [ 100 mm. similarity is low and can be seen only in the relatively large conidia with constricted or subconstricted branch insertion. Another small subclade linked to T. chaetocladium, consists of two isolates of T. curvisporum and two problematic taxa Tricladium sp.1 VG 68-1 and Tricladium sp.2 VG 69-2, both isolated from sedges submerged in fresh water and representing undescribed species. Tricladium sp.1 has conidia somewhat similar to those of T. angulatum or T. minutum, but it has black colonies, the conidial elements are wider, the conidial axis is sigmoid and axis and branches have conspicuously acute apices; elongation of conidiogenous cell is sympodial. Tricladium sp.2 is a similar isolate with dark colonies and sympodial elongation of conidiogenous cells, but conidia are much more delicate with 0–2(3) primary branches and an occasional secondary branch produced after release. Conidia may bear some resemblance to part conidia of V. delicatum. H. chaetocladia (HME 4375, ex-isotype, teleomorph of T. chaetocladium) was found to be placed within a poorly supported, so called ‘‘bsa’’ clade (Wang et al. 2005, LSU, SSU, 5.8S rDNA) containing ecologically similar saprotrophic fungi, mostly with bright apothecia, preferring aquatic or at least humid habitats. In their parsimony analysis the closest neighbors are two species of Mitrula and two of Vibrissea (Vibrisseaceae), but in their maximum likelihood tree topology Hydrocina appears closest only to Vibrissea spp. (Wang et al. 2005). In a later study, however, (Wang et al. 2006a, parsimony analysis of LSU, SSU and 5.8S rDNA regions) they place this fungus into Hyaloscyphaceae, but with no bootstrap support. Finally, in an extensive study based on SSU, LSU and 5.8S rDNA sequences of 99 taxa (Wang et al. 2006b), it is placed in the so-called Mitrula clade among Helotiaceae, again without bootstrap support (cf. also Raja et al. 2008, LSU, maximum parsimony). T. chaetocladium in the study of Belliveau & Bärlocher (2005, SSU rDNA) formed a poorly supported clade with T. angulatum and with Monilinia fructicola (Sclerotiniaceae) a fruit parasite on Rosaceae. The classification in Hyaloscyphaceae, suggested tentatively by Webster et al. (1991) remains questionable. The only Hyaloscypha hit in our blast search grouped with Tricladium procerum. The remaining species treated in our study form rather small clades within Helotiales. The fourth group (Figs 1, 6) consists of T. angulatum and Varicosporium scoparium, both of which have pale colonies and sympodial conidiogenous cell proliferation, but otherwise there is very little morphological similarity between them. V. scoparium is a rare species, which differs from other Varicosporium species by ‘‘scopiform’’ (broom-like) conidia. This clade also contains the teleomorphs Chlorencoelia sp., Heyderia abietis and Fabrella tsugae, all placed in Dermateaceae, but this alignment is only poorly supported. The only other aquatic hyphomycetes having teleomorphs in that family are Anguillospora crassa, Casaresia sphagnorum and Filosporella sp. (Marvanová 2007); none of them are morphologically similar to Tricladium or Varicosporium. Our results corroborate the distant placement of T. angulatum and T. splendens, the type of the genus, revealed by Belliveau & Bärlocher (2005) and by Baschien et al. (2006). Although in T. splendens the teleomorph was described in Hymenoscyphus, Helotiaceae (Abdullah et al. 1981), T. angulatum seems to have affinity to Hyaloscyphaceae (Belliveau & Bärlocher 2005, SSU rDNA; Baschien et al. 2006, SSU, ITS, rDNA). Both taxa differ considerably also in phenotypic characters: T. splendens has dark colonies, percurrent conidiogenous cells and constricted conidial branch insertion, whereas T. angulatum has pale colonies, polyblastic sympodial conidiogenesis and unconstricted conidial branch insertion. The fifth group (Figs 1, 6) contains T. procerum, forming a small, but highly supported clade with members of two families of Helotiales: Hyaloscypha daedaleae and Hymenoscyphus sp. (Hyaloscyphaceae and Helotiaceae, respectively). It does not seem to be related to other species of Tricladium. T. procerum seems rare, known so far only from the type locality in the Slovak Republic and from a few streams in the Czech Republic, all Fig 5 – Group 3. Conidia from Tricladium clade 3. (A) Hydrocina chaetocladia conidial state, CCM F-10890. (B) Tricladium chaetocladium CCM F-03485. (C) Tricladium sp. 1, VG 68-1. (D) Tricladium sp. 2, VG 69-2. (E) Tricladium curvisporum CCM F-23387. Conidia from Varicosporium clade 2. (F) Varicosporium delicatum CCM F-03977. (G) V. delicatum CCM F-18499. Scale bar [ 50 mm. Fig 6 – Groups 4–6. Conidium from Tricladium clade 4. (A) Tricladium angulatum CCM F-00282. Conidium from Tricladium clade 5. (B) Tricladium procerum CCM F-16786. Conidia from Tricladium clade 6. (C–D) Tricladium caudatum CCM F-21299 (this culture was isolated from a conidium without parabasal extension, but it appears in some conidia). Conidium from Varicosporium clade 3. (E) Varicosporium scoparium CCM F-10303. Scale bar [ 50 mm. 1332 with soft water (Marvanová 1988). Affinity to Hyaloscyphaceae based on rDNA sequence data (Sokolski et al. 2006, ITS1– 5.8S–1TS2) was found also in Dwayaangam colodena, an aquatic hyphomycete collected in foam on streams in the Canadian boreal forests and isolated from Picea mariana needles. Conidia of D. colodena have dichotomous and trichotomous branches on a stalk and are quite different from those of T. procerum or T. angulatum, mentioned above as possible hyaloscyphaceaean members. The sixth group (Figs 1, 6) contains Tricladium caudatum, which appears with 88 % support in a small clade with Rhytisma acerinum (Rhytismatales). The only other aquatic hyphomycete with such affinity is Tricladiopsis flagelliformis which has quite different conidia. Tricladium caudatum is also the only species in our group of taxa with affinity outside Helotiales. It is unique within this genus by typical in situ production of an excentric caudal extension on conidia. CCM F-21299, isolated from a conidium lacking this extension, produces extensions only in a small percentage of conidia. The high similarity of nucleotide sequences with isolate CCM F-13498, collected in the same area and cultured from a conidium bearing this extension supports the conspecificity of both isolates. Conclusions Using molecular analyses of the partial 28S rDNA, we were not able to unequivocally support the classification of taxa based on conidiogenesis or on morphological similarity of conidia. The ambiguity in inter-generic classification of Tricladium and Varicosporium evident on the morphological level is not resolved on the basis of the LSU rDNA gene analysis. Neither the branching pattern nor the conidiogenesis seem to be diagnostic for generic separation. Molecular phylogenetic studies in aquatic hyphomycetes have one limitation, which is not always fully recognized and in some individual cases may lead to misinterpretations: there are very few ex-type cultures (Marvanová 2007) and therefore we have to rely upon our ability to correctly identify specimens used for sequencing in accordance with the protologue. This may be sometimes difficult due to the phenotypic plasticity and genetic variability (Hebert et al. 2003), an insufficient protologue or a lack of experience in distinguishing morphological characters. The molecular taxonomy is still in the stage of accumulating knowledge and search for suitable genes for sequencing. Presently, especially concerning the staurosporous and scolecosporous hyphomycetes, the amount of treated taxa and sequenced genes is small. The inconsistencies between classifications based on morphology and phylogenetic analysis are especially striking in staurosporous anamorphs because our classification in this group is, besides conidiogenesis, predominantly based on conidial morphology. We face a disquieting problem: our visual approach to classification of objects forces us to group those with similar shapes together, but the phylogenetic analyses sometimes link taxa with markedly dissimilar conidia. Some recent results in other groups with conspicuous conidial configuration, e.g. Spirosphaera cupreorufescens, an aero-aquatic hyphomycete with complex conidia consisting of intertwined J. Campbell et al. branched filament and closely related to scolecosporous Anguillospora longissima (Voglmayr 2004) or Aquaphila albicans with sickle-shaped conidia and close affinity to some helicosporous hyphomycetes (Tsui et al. 2007), raise a similar question. Close phylogenetic relationships between aquatic hyphomycete taxa with branched and with unbranched conidia indicate that conidial branching might have arisen repeatedly and as the response to microhabitat conditions influencing distribution of species within a stream. A stream is a hydrologically complex habitat and offers various niches in its rapid and slow or deep and shallow parts. Branch number as well as their arrangement might represent advantages or disadvantages in leaf colonization within a particular microhabitat. The majority of anamorphic aquatic hyphomycetes treated here show relationships to Helotiales, contributing so to the 23 species whose affinity to that order is already known (Marvanová unpubl.). Only one species is placed in Rhytismatales, an order housing among others the leaf parasite Rhytisma acerinum. However, Helotiales in the contemporary interpretation is considered paraphyletic (Hibbett et al. 2007) and not well resolved (Wang et al. 2006b). Its future refinement may link at least some aquatic hyphomycetes with possible smaller monophyletic lineages. As in previous studies (Belliveau & Bärlocher 2005; Baschien et al. 2006), the teleomorphs described in some pleomorphic taxa on the basis of cultural studies, have not shown clear relationships to terrestrial representatives of the same genera. One of the reasons may be that not enough terrestrial teleomorphs were included in the study. The LSU gene region sequences are not frequently represented in GenBank. Acknowledgements This article is part of the project MSM00216222416 (LM). LM is also grateful to C. Baschien for valuable comments during preparation of the manuscript. references Abdullah SK, Descals E, Webster J, 1981. Teleomorphs of three aquatic hyphomycetes. Transactions of the British Mycological Society 77: 475–483. Altschul SF, Madden TL, Schäffer AA, Zhang J, Zhang Z, Miller W, Lipman DJ, 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Research 25: 3389–3402. Ando K, Kawamoto I, 1985. New and interesting microfungi from fallen leaves in Japan: two Tricladium species. Transactions of the Mycological Society of Japan 26: 39–47. Baschien C, Marvanová L, Szewzyk U, 2006. Phylogeny of selected aquatic hyphomycetes based on morphological and molecular data. Nova Hedwigia 83: 311–352. Belliveau MJ-R, Bärlocher F, 2005. Molecular evidence confirms multiple origins of aquatic hyphomycetes. Mycological Research 109: 1407–1412. Bills GF, Platas G, Peláez F, Masurekar P, 1999. Reclassification of a pneumocandin-producing anamorph Glarea lozoyensis gen et Evolutionary relationships of Tricladium and Varicosporium sp. nov., previously identified as Zalerion arboricola. Mycological Research 103: 179–192. Boonyuen N, Sivichai S, Worapong J, Hywel-Jones N, 2006. A molecular phylogenetic study of selected in goldian (sic) species. BRT Research Reports 2006: 56–63. Bunyard BA, Nicholson MS, Royse DJ, 1994. A systematic assessment of Morchella using RFLP analysis of the 28S ribosomal RNA gene. Mycologia 86: 762–772. Campbell J, Shearer C, Marvanová L, 2006. Evolutionary relationships among aquatic anamorphs and teleomorphs: Lemonniera, Margaritispora, and Goniopila. Mycological Research 110: 1025–1033. Crane JL, 1968. Freshwater hyphomycetes of the Northern Appalachian Highland including New England and three Coastal Plain states. American Journal of Botany 55: 996–1002. Descals E, 1985. Conidia as modified hyphae. Proceedings of the Indian Academy of Sciences (Plant Sciences) 94: 209–227. Descals E, Webster J, 1982. Taxonomic studies on aquatic hyphomycetes III. Some new species and a new combination. Transactions of the British Mycological Society 78: 405–437. Descals E, Marvanová L, Webster J, 1998. New taxa and combinations of aquatic hyphomycetes. Canadian Journal of Botany 76: 1647–1659. Dyko BJ, 1978. New aquatic and waterborne hyphomycetes from the Southern Appalachian Mountains in the United States. Transactions of the British Mycological Society 70: 409–416. Felsenstein J, 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783–791. Hebert PDN, Cywinska A, Ball SL, deWaard JR, 2003. Biological identifications through DNA barcodes. Proceedings of the Royal Society of London 270: 313–321. Hennebert GL, Sutton BC, 1994. Unitary parameters in conidiogenesis. In: Hawksworth DL (ed), Ascomycete Systematics. Problems and Perspectives in the Nineties. Plenum Press, New York and London, pp. 65–76. Hibbett DS, Binder M, Bischoff JF, Blackwell M, Cannon PF, Eriksson OE, Huhndorf S, James T, Kirk PM, Lücking R, Thorsten Lumbsch H, Lutzoni F, Matheny PB, McLaughlin DJ, Powell MJ, Redhead S, Schoch CL, Spatafora JW, Stalpers JA, Vilgalys R, Aime MC, Aptroot A, Bauer R, Begerow D, Benny GL, Castlebury LA, Crous PW, Dai Y-C, Gams W, Geiser DM, Griffith DM, Gueidan C, Hawksworth DL, Hestmark G, Hosaka K, Humber RA, Hyde KD, Ironside JE, Kõljalg U, Kurtzman CP, Larsson K-H, Lichtwardt R, Longcore J, Mia x dlikowska J, Miller A, Moncalvo J-M, Mozley-Standridge S, Oberwinkler F, Parmasto E, Reeb V, Rogers JD, Roux C, Ryvarden L, Sampaio JP, Schüssler A, Sugyiama J, Thorn RG, Tibell L, Untereiner W, Walker C, Wang Z, Weir A, Weiss M, White MM, Winka K, Yao Y-J, Zhang N, 2007. A higher-level phylogenetic classification of the Fungi. Mycological Research 111: 509–547. Huelsenbeck JP, Ronquist F, 2001. MR BAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17: 754–755. Ingold CT, 1942. Aquatic hyphomycetes of decaying alder leaves. Transactions of the British Mycological Society 25: 339–417. Ingold CT, 1975. Hooker Lecture 1974. Convergent evolution in aquatic fungi: the tetraradiate spore. Biological Journal of the Linnean Society 7: 1–25. Iqbal SH, 1974. Scorpiosporium minutum gen. nov., sp. nov., an aquatic hyphomycete. Biologia (Lahore) 20: 17–21. Kegel W, 1906. Varicosporium Elodeae, ein Wasserpilz mit auffallender Konidienbildung. Berichte der Deutschen Botanischen Gesellschaft 24: 213–216. Kendrick WB, 2003. Analysis of morphogenesis in hyphomycetes: new characters derived from considering some conidiophores and conidia as condensed hyphal systems. Canadian Journal of Botany 81: 75–100. Marvanová L, 1988. New hyphomycetes from aquatic environments in Czechoslovakia. Transactions of the British Mycological Society 90: 607–617. 1333 Marvanová L, 2002. Aquatic hyphomycetes – emerging outlines of their classification in the fungal system. In: Third International Meeting on ‘‘Plant Litter Processing in Freshwater’’, September 7–12, 2002, Szentendre, Hungary. (Abstract), p. 11. Marvanová L, 2004. New hyphomycete species from streams in the Šumava National Park (Bohemian Forest, Czech Republic). Czech Mycology 56: 193–202. Marvanová L, 2007. Aquatic hyphomycetes and their meiosporic relatives: slow and laborious solving of a jig-saw puzzle. In: Ganguli BN, Deshmukh SK (eds), Fungi Multifaceted Microbes. Anamaya Publishers, New Delhi, pp. 128–152. Marvanová L, Bärlocher F, 1998. Hyphomycetes from Canadian streams III. Arcispora bisagittaria anam. gen. et sp. nov. Mycologia 90: 531–536. Marvanová L, Descals E, 1996. Hyphomycetes from streams: new taxa and new combination. Mycotaxon 60: 455–469. Marvanová L, Gulis V, 2000. Notes on aquatic hyphomycetes and streamborne spora from Austria. Österreichische Zeitschrift für Pilzkunde 9: 125–140. Olmstead RG, Bremer B, Scott KM, Palmer JD, 1993. A molecular systematic analysis of the Asteridae sensu lato based on rbcL sequences. Annals of the Missouri Botanical Garden 80: 700–722. Park D, 1974. Tricladium terrestre sp. nov. Transactions of the British Mycological Society 63: 179–183. Qiagen, 2004. DNeasyTM Plant Mini Handbook. QIAGEN Inc., USA. Qiagen, 2002a. Taq PCR Handbook for Taq PCR Master Mix Kit. QIAGEN Inc., USA. Qiagen, 2002b. QIAquick Spin Handbook for QIAquick PCR Purification Kit. QIAGEN Inc., USA. Raja H, Campbell J, Shearer CA, 2003. Freshwater ascomycetes: Cyanoannulus petersenii a new genus and species from submerged wood. Mycotaxon 88: 1–17. Raja HA, Miller AN, Shearer CA, 2008. Freshwater ascomycetes: Aquapoterium pinicola, a new genus and species of Helotiales (Leotiomycetes) from Florida. Mycologia 100: 141–148. Rambaut A, 1996. Se-Al: sequence alignment editor. vers. 1.0 alpha 1.0. University of Oxford, UK. Rodriguez F, Oliver J, Marin A, Medina J, 1990. The general stochastic model of nucleotide substitution. Journal of Theoretical Biology 142: 485–502. Sati SC, Tiwari N, 1992. A new species of Tricladium from Kumaun Himalaya, India. Mycological Research 96: 229–232. Sivichai S, Jones EBG, Hywel-Jones NL, 2003. Lignicolous freshwater Ascomycota from Thailand: Hymenoscyphus varicosporoides and its Tricladium anamorph. Mycologia 95: 340–346. Sokolski S, Piché Y, Laitung B, Bérubé JA, 2006. Streams in Quebec boreal and mixed-wood forests reveal a new aquatic hyphomycete species, Dwayaangam colodena sp. nov. Mycologia 98: 628–636. Swofford DL, 2002. PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods). ver. 4. Sinauer Associates, Sunderland, Massachusetts. Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG, 1997. The Clustal-X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Research 25: 4876–4882. Tsui KM, Sivichai S, Rossman AY, Berbee ML, 2007. Tubeufia asiana, the teleomorph of Aquaphila albicans in the Tubeufiaceae, Pleosporales, based on cultural and molecular data. Mycologia 99: 884–894. Vilgalys R, Hester M, 1990. Rapid genetic identification mapping of enzymatically amplified ribosomal DNA from several Cryptococcus species. Journal of Bacteriology 172: 4238–4246. Voglmayr H, 2004. Spirosphaera cupreorufescens sp. nov., a rare aeroaquatic fungus. Studies in Mycology 50: 221–228. Wang Z, Binder M, Hibbett DS, 2005. Life history and systematics of the aquatic discomycete Mitrula (Helotiales, Ascomycota) 1334 based on cultural, morphological, and molecular studies. American Journal of Botany 92: 1565–1574. Wang Z, Binder M, Schoch CL, Johnson PR, Spatafora JW, Hibbett DS, 2006a. Evolution of Helotialean fungi (Leotiomycetes, Pezizomycotina): a nuclear rDNA phylogeny. Molecular Phylogenetics and Evolution 41: 295–313. Wang Z, Johnson PR, Takamatsu S, Spatafora JW, Hibbett DS, 2006b. Toward a phylogenetic classification of the Leotiomycetes based on rDNA data. Mycologia 98: 1065–1075. J. Campbell et al. Webster J, Eicker A, Spooner BM, 1995. Cudoniella indica sp. nov. (Ascomycetes, Leotiales), the teleomorph of Tricladium indicum, an aquatic fungus isolated from a South African river. Nova Hedwigia 60: 493–498. Webster J, Scheuer Ch, Om-Kalthoum Khattab S, 1991. Hydrocina chaetocladia gen. et sp. nov., the teleomorph of Tricladium chaetocladium. Nova Hedwigia 52: 65–72. Willoughby LG, Minshall GW, 1975. Further observations on Tricladium giganteum Iqbal. Transactions of the British Mycological Society 65: 77–82.