Mycol. Res. 106 (5) : 541–548 (May 2002). # The British Mycological Society 541 DOI : 10.1017}S0953756202006019 Printed in the United Kingdom. Phylogenetic relationships among smut fungi parasitizing dicotyledons based on ITS sequence analysis Teresa ALMARAZ1, Christophe ROUX2*, Stephan MAUMONT3 and Guy DURRIEU4 " Real JardıU n BotaU nico (CSIC ), Plaza de Murillo 2, 28014 Madrid, Spain. # Equipe de Mycologie VeU geU tale, UMR CNRS 5546, PoW le de biotechnologie veU geU tale, chemin de Borde Rouge, BP 17, 31326, Castanet-Tolosan, France. $ Laboratoire d’Ecologie Terrestre, UMR CNRS 5552, UniversiteU Paul Sabatier, 118 route de Narbonne, bat. 4R3, 31062 Toulouse Cedex 4, France. % Chemin Flou de Riou, 31000 Pouvourville, Toulouse, France. E-mail : roux!smcv.ups-tlse.fr Received 8 October 1999 ; accepted 30 April 2002. The phylogenetic relationships of several smut fungi parasitic on dicotyledons were analysed. Parsimony analysis was performed based on the sequences of the ITS regions of the rDNA genes. Three genera were considered : Microbotryum, Sphacelotheca and Ustilago. The cladogram showed a dichotomy : the species of Microbotryum and Ustilago parasitic on dicotyledons (dicot Ustilago) were found to be divided into two independent taxa among the Microbotryaceae. A divergent mechanism of evolution of the species of each of these two clades with their respective hosts could be involved in this dichotomy. According to our results, Microbotryum appears monophyletic and restricted to the anthericolous smuts on Caryophyllaceae. However, no morphological characters have yet been found to support this distinction, and we refute the denomination Bauhinus as defined by Moore to describe the group of ‘ dicot Ustilago ’ as it leads to controversial determinations. Sphacelotheca belongs to Microbotryales, but could not be synonymised with Microbotryum as S. polygoni-persicariae is in an independent clade. Lastly, U. duriaeana is independent of the ‘ dicot Ustilago ’ clade ; the position of this species among Microbotryales is still uncertain. INTRODUCTION Many studies have tried to clarify the systematic and taxonomic position of smuts parasitic on dicotyledons in the Basidiomycota. Biochemical (Prillinger et al. 1991), ultrastructural (Bauer, Oberwinkler & Va! nky 1997) and molecular approaches (Blanz & Gottschalk 1984, Swann & Taylor 1995, Begerow, Bauer & Oberwinkler 1997, Roux, Almaraz & Durrieu 1998) have established that the smut fungi, formerly grouped in the Ustilaginales, are polyphyletic and should be split into different groups at different taxonomic levels. As discussed by Begerow et al. (1997), three lineages are considered in Basidiomycetes : Hymenomycetes, Ustilaginomycetes, and Urediniomycetes. This last lineage is thought to include Uredinales and Microbotryales (Bauer et al. 1997), or Microbotryomycetidae (Swann, Frieders & McLaughlin 1999). One of the taxa excluded from Ustilaginales and placed in the Microbotryales, the Microbotryaceae * Corresponding author. (Moore 1996), includes smuts on dicotyledons. However, the delimitation of this family and its genera is still uncertain. Different authors have considered whether the species belong to Microbotryum and whether to add genera other than Microbotryum to Microbotryaceae. Some have proposed the inclusion of all Ustilago species parasitic on dicotyledons together in a single genus, Microbotryum (Prillinger et al. 1990, 1991, Denchev 1994, Va! nky 1998a, Begerow et al. 1997, Bauer et al. 1997), while Moore (1992, 1996) and Denchev (1997) proposed the separation of these species into two genera : species infecting Caryophyllaceae (genus Microbotryum) and species parasitic on dicotyledons other than Caryophyllaceae (genus Bauhinus). Bauer & Oberwinkler (1997) pointed out that up till now no morphological characters clearly support this distinction. To avoid ambiguities, and before making any firm statements, below we will call ‘ dicot Ustilago ’ the species of Ustilago parasitic on non-caryophyllaceous dicotyledons. Lastly, based on ultrastructural markers, Bauer et al. (1997) included Sphacelotheca in Microbotryales, but Delimitation of Microbotryum 542 Table 1. List of the fungi used in this study. Microbotrym dianthorum M. dianthorum M. lychnidis-dioicae M. violaceum M. violaceum M. silenes-inflatae M. violaceo-verrucosum M. violaceum Sphacelotheca polygoni-serrulati Ustilago cordae U. duriaeana U. duriaeana U. kuehneana U. pinguiculae U. scabiosae U. scolymi U. scorzonerae U. vinosa Host (family) Spore morphology} sorus location Location (date collected) Dianthus hyssopifolius (Caryophyllaceae) Dianthuscarthusianorum (Caryophyllaceae) Silene latifolia (Caryophyllaceae) Saponaria ocymoides (Caryophyllaceae) Saponaria officinalis (Caryophyllaceae) Silene inflata (Caryophyllaceae) Silene mellifera (Caryophyllaceae) Silene ciliata (Caryophyllaceae) Polygonum serrulatum (Polygonaceae) Polygonum spp. (Polygonaceae) Cerastium spp. (Caryophyllaceae) Arenaria spp. (Caryophyllaceae) Rumex acetosella (Polygonaceae) Pinguicula spp. (Lentibulariaceae) Knautia dipsacifolia (Dipsacaceae) Scolymus hispanicus (Asteraceae) Scorzonera spp. (Asteraceae) Oxyria digyna (Polygonaceae) Reticulated} anthers Reticulated} anthers Reticulated} anthers Reticulated} anthers Reticulated} anthers Reticulated} anthers Echinulated} anthers Reticulated} anthers Reticulated} ovaries Reticulated} flowers Reticulated} ovaries Reticulated} ovaries Reticulated}flowers, stems, leaves Reticulated} anthers Reticulated} anthers Reticulated} inflorescence Reticulated} inflorescence Reticulated} flowers SP : Huesca (VIII-95) SP : Le! rida (VI-96) SP : Cuenca (V-95) SP : Burgos (4-VI-1996) FRA : Pyre! ne! es-Or. (VII-96) its relationships with the other members of this order is not well established. Most of the works on the molecular phylogeny of smuts have treated the question at the suprageneric level by using small and large subunits of the rDNA genes (Swann & Taylor 1995, Begerow et al. 1997). Our purpose was to investigate the limits of Microbotryum and Microbotryaceae by analysing the ITS (Internal Trancribed Spacer) regions of the rDNA genes. The ITS regions are known to be useful for the analysis of closely related taxa of fungi (Bruns, White & Taylor 1991). These sequences are rather constant inside a species and are highly variable between species, leading to a sufficient number of informative sites to define phylogenetic relationships of genera. The utility of the ITS in studying the phylogeny of smut fungi has been reported previously (Roux et al. 1998). In order to compare the molecular data with morphological characters used in the recent classifications of smuts, the species included in our analysis were chosen according to family host range, sorus position, and spore morphology. A parsimony analysis was performed on the ITS sequences of these species. SP : Albacete (VI-96) SP : Palencia (VII-95) SP : Albacete (VI-96) SP : Albacete (VI-96) SP : A; vila (VI-96) SP : Le! rida (VII-96) SP : Guadalajara (I-97) Collection GenBank accession no. MA-Fungi 34534 (TAl281) MA-Fungi 35419 (TAL 367) MA-Fungi 33572 (TAl171) MA-Fungi 37708 (TAl321) CR5±3, 96 AF038834 AF045872 ex VKM J2974 AF045875 MA-Fungi 37728 (TAl329) MA-Fungi 33613 (TAl238) ex VKM Y2691 AF045874 AF038835 ex VKM J2967 AF045878 MA-Fungi 37798 (TAl326) MA-Fungi 37799 (TAl328) MA-Fungi 37803 (TAl337) ex CBS 184±42 AF287152 AF045881 CR3±9, 96 AF045880 MA-Fungi 37554 (TAl401) ex CBS 365±33 AF038830 ex VKM J973 AF045876 Not submitted AF045873 AF038832 AF038833 Not submitted AF045879 AF045877 MATERIALS AND METHODS Fungal isolates and sequences The samples used in our assays, their host and geographic origin, some of their morphological characters and the GenBank accession numbers are listed in Table 1. DNA extraction, PCR amplification and sequencing A sample of 10 mg of yeast cells or teliospores was taken for each species considered in this analysis. The samples were ground in sand with 1 ml of extraction buffer (0±2  Tris–HCl pH 8, 0±02  EDTA, 1±4  NaCl, 2 % CTAB, 0±2 % β mercaptoethanol) (Gardes & Bruns 1993) and then centrifuged for 10 min at 5000 g. Proteinase K (20 µg was added to the supernatant and left for 5 min at 62 °C. After centrifugation for 10 min at 5000 g, 1 ml of phenol}chloroform was gently mixed with the supernatant. Proteins were pelleted by centrifugation for 15 min at 5000 g. The supernatant was gently mixed with half a volume of chloroform and centrifuged for 15 min at low speed. The supernatant T. Almaraz and others 543 Puccinia caricina Puccinia consimilis Ustilago duriaeana strain 328 Ustilago duriaeana strain 326 Sphacelotheca polygoni persicariae Ustilago scabiosae Ustilago vinosae Ustilago kuehneana Ustilago cordae Ustilago pinguiculae Ustilago scorzonerae Ustilago scolymi Microbotryum violaceum Microbotryum lychnidis-dioicae Microbotryum silenes-inflatae Microbotryum violaceum* Microbotryum violaceum* Microbotryum violaceo-verrucosum Microbotryum dianthorum Microbotryum dianthorum Fig. 1. Cladogram obtained with 20 aligned ITS1-5±8S-ITS2 sequences of different species of smut fungi. The ITS sequences of Puccinia caricina and P. consimilis were added as outgroup. One most parsimonious tree, resulting from 1000 bootstrap replicates on 325 sites (137 informative), and 388 steps were required for maximum parsimony. Microbotryum violaceum* was on Saponaria officinalis. Values below branches correspond to the number of changes per branch (nc, value non-calculated in case of trichotomy). Numbers above branches correspond to bootstrap values. Tree length ¯ 912, consistency index (CI) ¯ 0±787, retention index (RI) ¯ 0±764. was then mixed with 0±7 vol. of isopropanol. DNA was pelleted by centrifugation (for 15 min, 5000 g) and washed twice with 70 % ethanol. The air-dried pellet was resuspended in 100 µl of TE buffer (10 m Tris–HCl pH 8 and 1 m EDTA). The quality and quantity of DNA were checked by spectrophotometry at 260, 280 and 310 nm. The ITS1 and ITS4 primers used in PCR to amplify the ITS sequence were described by Gardes & Bruns (1993). The amplifications used 25 µl DNA template solution in 50 µl of reaction mixture. The final solution contained 25 µ each of dATP, dCTP, dGTP and dTTP (Euromedex, Belgium), 0±5 µ of each primer and 1 unit of pfu-polymerase (Appligen-Oncor, France). The cycling conditions consisted of an initial denaturation step at 94 °C for 5 min, followed by 30 cycles of 1 min at 94 °, 1 min at 52 °, and 2 min at 72 °. The amplified DNA was visualised by electrophoresis on 1±5 % agarose gels. The PCR product (680 to 700 bp, except for U. duriaeana 550 bp) was purified using the QIAquickTM Kit (QIAGEN, France) according to the manufacturer’s protocol. The DNA was sequenced directly using the ABI PRISMTM Dye terminator Cycle Sequencing Kit (Perkin Elmer, USA) and an automated sequencer (ABI 373, Perkin Elmer, USA). Sequences alignment The sequences ITS1-5±8S-ITS2 were analysed on a UNIX station using the software SeaView for alignment procedures (Galtier, Gouy & Gautier 1996). This software uses the ClustalW algorithm (Thompson, Delimitation of Microbotryum 544 Sphacelotheca polygonipersicariae Microbotryum silenes-inflatae Microbotryum lychnidisdioicae Microbotryum violaceum boostrap analysis. The following settings were used : heuristic, 1000 replicates with a 50 % majority rule. The numbers of changes per branch were obtained using MacClade v. 3±03. The DNA sequence alignment used in these analyses and a tree file for Figs 1–2 have been deposited in TreeBASE under accession no. SN934. Microbotryum violaceum* Microbotryum violaceum* Microbotryum violaceoverrucosum Microbotryum dianthorum Microbotryum dianthorum Ustilago scabiosae Ustilago kuehneana Ustilago vinosae Ustilago scorzonerae Ustilago scolymi Ustilago cordae Ustilago pinguiculae Fig. 2. Cladogram obtained with 16 aligned ITS1-5±8S-ITS2 sequences of different species of Microbotryales (outgroup : Sphacelotheca polygoni-serrulati). One most parsimonious tree, resulting from 1000 bootstrap replicates on 448 sites (62 informative), and 174 steps were required for maximum parsimony. Microbotryum violaceum* was collected on Saponaria officinalis ; M. violaceum was species collected on Silene ciliata. Values below branches correspond to the number of changes per branch (nc, value non-calculated in case of trichotomy). Numbers above branches correspond to bootstrap values. Tree length ¯ 442, consistency index (CI) ¯ 0±785, retention index (RI) ¯ 0±769. Higgins & Gibson 1994) and allows to align partial regions of the ITS sequences with different parameters of alignment. This procedure is particularly interesting to align alternative variable and conserved regions like ITS1-5±8S-ITS2 sequences. Minor adjustments of the alignments were made manually with the Seaview Software. Phylogenetic inferences Maximum parsimony analysis was first performed on Phylojwin software (Galtier et al. 1996) using the DNApars algorithm from PHYLIP (Felsenstein 1985). Bootstrap analyses were made of 1000 replicates, and the tree was built with the consensus option of Phylojwin. We used the ITS1-5±8S-ITS2 sequences of Puccinia consimilis and Puccinia caricina to root the tree (GenBank accession nos U88215 and U88234 respectively). Further Cladistic analysis was made using the PAUP 3±1.1 program (Swofford 1993) with the branch and bound method. Confidence in specific clades of the resulting topologies was estimated by RESULTS The general phylogeny of the species analysed is given in Fig. 1. One most parsimonious tree was obtained, and the robustness of the major clades was assessed by high bootstrap values. No differences occurred in the topology of the tree when phylogenetic inferences were performed with different subsets of species (not shown). Five lineages were obtained with the species analysed. The species infecting Caryophyllaceae (M. silenesinflatae, M. lychnidis-dioicae, M. violaceo-verrucosum, M. dianthorum, M. violaceum) were well characterized in a single clade, although divided into two groups. Except for U. duriaeana, the species of ‘ dicot Ustilago ’ were in a single lineage, separated from the Microbotryum species, although the clade of U. scabiosae is weakly supported. This dichotomy was observed whatever the set of species used for intermediate analyses. In the set of species presented in the strict consensus tree (Fig. 1), no clear distinction occurred among the species of ‘ dicot Ustilago ’, and they were presented as an undefined clade. Sphacelotheca polygoni-persicariae was found to be separated from the sister groups Microbotryum and ‘ dicot Ustilago ’. However, the ITS sequence of Sphacelotheca shows higher similarity with Microbotryum and ‘ dicot Ustilago ’ species than with the species of the upper clades (Table 2). As indicated by the bootstrap values, these three lineages (Microbotryum, ‘ dicot Ustilago ’, and Sphacelotheca) are distinct but closer to each other than to the upper clades. The position of Ustilago duriaeana is intermediate between the outgroup and the three previously discussed lineages. To avoid a sampling error, two isolates of U. duriaeana collected on Cerastium sp. and Arenaria sp. from two different stations were used, and shared slightly divergent ITS sequences. In Fig. 1, the addition of divergent ITS sequences (Puccinia sp.) led us to choose sites unambiguously aligned to enable the phylogenetic inferences to be determined. As the ITS sequences of the Microbotryum and ‘ dicot Ustilago ’ species were more homologous (see Table 2), the phylogenetic relationships of these species were defined in a separate analysis to avoid any bias simply due the choice of the sites. One most parsimonious tree, rooted with S. polygoni-persicariae, is shown in Fig. 2. In this tree, sublineages can be separated within Microbotryum. A group containing M. violaceum (on Silene ciliata), M. silenes-inflatae, and M. lychnidisdioicae can be distinguished from M. dianthorum and M. violaceo-verrucosum. An additional lineage is formed T. Almaraz and others Table 2. Homology of the ITS1­ITS2 sequences (without 5±8S region) for some of the species analysed (in percentage). M. silenes-inflatae M. silenes-inflatae M. violaceum** M. violaceum* M. violaceo-verrucosum U. vinosae U. cordae U. scolymi U. scorzonerae Sp. polygoni-serrulati U. duriaeana P. caricina M. violaceum** M. violaceum* 91 92 89 M. violaceoverrucosum 87 86 90 U. vinosae U. cordae U. scolymi U. scorzonerae 74 76 73 74 79 77 81 77 81 74 69 76 73 79 80 79 77 82 77 78 81 85 Sp. polygoniserrulati 68 65 66 65 73 73 70 68 U. duriaeana P. caricina 28 27 27 28 30 30 29 28 44 19 20 20 19 24 27 24 25 33 32 The values were calculated by SEAVIEW using the pairwise alignments of the ClustalW algorithm. M., Microbotryum ; U., Ustilago ; Sp., Sphacelotheca ; P., Puccinia. ** On Silene ciliata ; * on Saponaria officinalis. 545 Delimitation of Microbotryum by two strains of M. violaceum. The sequences of these two isolates of M. violaceum, collected on Saponaria ocymoides and S. officinalis from different geographic regions (Table 1), differ significantly from the ITS of the isolate from Silene ciliata. In Fig. 2, the ‘ dicot Ustilago ’ species tested in our analysis are in a different lineage than the Microbotryum group (see also Fig. 1). These species always grouped together regardless which outgroup was used for the analysis (not shown). The relationships among ‘ dicot Ustilago ’ species is less clear than in the Microbotryum taxon. As for Microbotryum species, different groups appear : U. kuehneana and U. vinosae, U. cordae and U. pinguiculae, U. scorzonerae and U. scolymi. Although this last group is formed by species parasitic on Asteraceae, no strict relation according to the host plant family was observed as U. cordae, parasitic on Polygonaceae, is not grouped with U. kuehneana and U. vinosae, but with U. pinguiculae, parasitic on Lentibulariaceae. DISCUSSION As defined by Moore (1996), the Microbotryaceae forms a natural group. In the present study, we investigated the phylogenetic relationships of some members of this family, and their relationships with species with an uncertain position. We used the ITS sequences to compare these closely related species and to define the delimitation of each genus. Phylogenetic relationships and taxonomical implications among Urediniomycetes Comparison of the 18S (Swann & Taylor 1993, Swann & Taylor 1995), 28S (Begerow et al. 1997) and ITS15±8S-ITS2 rDNA sequences (Roux et al. 1998) leads to the conclusion that the Microbotryales are related to the Uredinales, and grouped in the Urediniomycetes (Swann & Taylor 1995). Our results showed that Sphacelotheca polygoni-persicariae and Ustilago duriaeana were intermediate between Uredinales and the other species analyzed. Parsimony analysis of the ITS region of Sphacelotheca polygoni-persicariae showed that this species was distinct from the clades Microbotryum and ‘ dicot Ustilago ’ (Microbotryaceae), although the homology of the sequences indicated that it is related to them. This analysis supports the position of the genus Sphacelotheca in Microbotryales, as proposed by Prillinger et al. (1993) and Bauer et al. (1997) based on biochemical and ultrastructural data. In a comparison of the 5S sequences, Blanz & Gottschalk (1984) also found that M. violaceum, U. scabiosae, and a species related to S. hydropiperis formed a monophyletic group. However, the distinction we observed between the clades Microbotryum and ‘ dicot Ustilago ’ vs Sphacelotheca suggests that Sphacelotheca is different from Microbotryum, and could belong to a family distinct from Microbotryaceae. 546 It would be interesting to add the ITS sequences of Ustilentyloma species to verify whether the Microbotryales are only formed by two families, one parasiting dicotyledons, the other parasiting monocotyledons (Bauer et al. 1997, Begerow et al. 1997), or whether Sphacelotheca species form a third family. The case of U. duriaeana is complex. The morphology of this species is different, its spores being larger than in other ‘ dicot Ustilago ’ species, and its sorus being formed in the ovary instead of the anthers as for other caryophyllaceaous species studied. Compared to the other smuts on dicots in Fig. 1, U. duriaeana is intermediate between Uredinales and Microbotryales. The ultrastructure, such as nucleus-associated characters (Swann et al. 1999), and the biology of this species will be further investigated to assess its particular phylogenetic position, and its interest in evolution as an intermediate species between the two orders of the Urediniomycetes. From our results, although this species could remain inside the Microbotryales, it does not seem appropriate to include it in Microbotryum as proposed by Va! nky (1998b). Phylogenetic relationships among Microbotryaceae and taxonomic implications Our results indicate that the anther smuts on Caryophyllaceae are monophyletic. We found that the ITS sequences of the species of Microbotryum were highly homologous (Table 2). These species could have a common ancestor whose host was probably a Caryophyllaceae, or was related to this family. This could be the result of a single phenomenon of parasitism followed by coevolution of the hosts and the parasites. Lepage et al. (1994) reported that this coevolution is rather ancient as Microbotryum probably already existed inside the anthers of plants during the Eocene. Moreover, different isolates of M. violaceum show a very strict host specificity (Baker 1947). The existence of special forms of M. violaceum has been supported by different molecular approaches (Perlin 1996, Perlin et al. 1997). However, the comparison of the ITS sequences of different strains of M. violaceum suggests that some of these could be considered as distinct species. For instance, two strains of M. violaceum from Saponaria sp. formed a separate clade from M. violaceum isolated on Silene ciliata (Fig. 2). The position of the clade M. violaceo-verrucosum compared to the clades M. dianthorum and M. violaceum}Saponaria sp. is uncertain as their bootstrap support is weak (48). Inside the ‘ dicot Ustilago ’ clade, the results showed a lesser monophyly based on the host plant family. U. scorzonerae and U. scolymi, both parasitic on Asteraceae, are grouped, but not the species which are pathogenic to Polygonaceae. Although the number of samples was low for each host family, these results invalidate the hypothesis of a strict coevolution of all ‘ dicot Ustilago ’ species with their hosts. As proposed for coevolution of Uredinales with their hosts (Durrieu T. Almaraz and others 1980, Savile 1990), different mechanisms seem to have been involved in the evolution of the ‘ dicot Ustilago ’ group, like jumps (biogenic radiation) to host plants belonging to different families, without direct phylogenetical relationships, i.e. from Polygonaceae to Lentibulariaceae (Lamiidae), Dipsacaceae, and Compositae (Asteridae). In a study based on ultrastructural characters, Bauer et al. (1997) established a classification of Ustilaginales and related taxa different to that adopted in Hawksworth et al. (1995). They consider the Microbotryales to include Microbotryaceae and Ustilentylomataceae. They proposed the Microbotryaceae be used for Microbotryum including the species already known as Microbotryum together with ‘ dicot Ustilago ’ species. Va! nky (1999) also included three other genera (viz. Sphacelotheca, Liroa, Zundeliomyces), all parasitising dicotyledonous hosts. However, our results showed a clear dichotomy within species placed in Microbotryum and ‘ dicot Ustilago ’, the clade Microbotryum being in a basal position of the clade ‘ dicot Ustilago ’. The Microbotryum species form a monophyletic lineage and could have evolved from an ancestor common to ‘ dicot Ustilago ’ species, but independently from them. Although our results support that Microbotryum and ‘ dicot Ustilago ’ belong to a unique family, the taxonomic significance of the dichotomy of these two groups has to be discussed. Two contradictory hypotheses can be put forward : (1) Microbotryum species and ‘ dicot Ustilago ’ could be considered as a unique taxon, Microbotryum sensu Bauer & Oberwinkler (1997). The anthericolous species on Caryophyllaceae constitute an infraspecific lineage with an independent mechanism of evolution. Such a strategy of coevolution is not unique among Microbotryum as so circumscribed as it has also been observed on species parasitic of Asteraceae. (2) Microbotryum species and ‘ dicot Ustilago ’ can be considered as different taxa. This is supported by the observation, in our cladogram, that the two groups are not nested, and present a dichotomy. This divergence was also observed by comparison of karyotypes of different isolates of Microbotryum violaceum and several ‘ dicot Ustilago ’, showing chromosome size polymorphism (Perlin 1996). The genus Microbotryum should then be restricted to the anthericolous smuts parasitic on Caryophyllaceae. Meanwhile, neither the anthericolous behaviour, nor structural (spore and basidium morphology) and ultrastructural (septa without pores) characters at present described allow two natural groups to be distinguished. Many species of ‘ dicot Ustilago ’ on dicotyledons are anthericolous and have reticulated spores (Va! nky 1994). Ustilago vinosa has spores identical in shape and size to those of M. violaceum, while U. scabiosae is anthericolous too. For these reasons, we also do not accept the generic name Bauhinus (Moore 1992) described and used confusedly for Ustilago on dicotyledons in some recent works. We therefore agree with Bauer & Oberwinkler (1997) that 547 Bauhinus also includes in its circumscription the genera Microbotryum s. str. and Sphacelotheca. In conclusion, our results indicate that Sphacelotheca and an ovariicolous species on Caryophyllaceae, U. duriaeana, are generically distinct from Microbotryum. The addition of ITS sequences of other ovariicolous species on Caryophyllaceae (i.e. U. nivalae and U. holostei) would allow the analysis of the autapomorphy of the character ‘ ovariicolous ’ and the phylogenetic position of this group. Moreover, a clear dichotomy is revealed between the Microbotryum s. str. clade and the ‘ dicot Ustilago ’. However, although Microbotryum species are morphologically homogeneous, consistent morphological differences have not been observed to differentiate the genus from ‘ dicot Ustilago ’. As yet, this leads to the two clades being included in Microbotryum sensu Bauer & Oberwinkler (1997). While micromorphological characters (teliospore or basidium morphology) are often inadequate for separations in this group of fungi (Swan et al. 1999), further investigations of the ultrastructure of these species could reveal characters supporting the distinction of Bauhinus, corresponding to ‘ dicot Ustilago ’, from Microbotryum. Lastly, it could be inferred from our results that Microbotryales may have been originally parasitic on ancestors of the phylum Caryophyllidae (orders Polygonales and Caryophyllales). Later separation of host plant species in independent natural groups could have been involved in the individualisation of fungal parasitic populations. A C K N O W L E D G E M E N TS T. A. was supported by the program DGES ‘ Flora Micologica Iberica III ’, PB95-0129-C03-01 of the Spanish government. REFERENCES Baker, H. G. (1947) Infection of species of Melandrium by Ustilago violaceae (Pers.) Fuckel and the transmission of the resultant disease. Annals of Botany 11 : 333–348. Bauer, R. & Oberwinkler, F. (1997) The Ustomycota : an inventory. Mycotaxon 64 : 303–319. Bauer, R., Oberwinkler, F. & Va! nky, K. (1997) Ultrastructural markers and systematics in smut fungi and allied taxa. Canadian Journal of Botany 75 : 1273–1314. Begerow, D., Bauer, R. & Oberwinkler, F. (1997) Phylogenetic studies on nuclear large subunit ribosomal DNA sequences of smut fungi and related taxa. Canadian Journal of Botany 75 : 2045–2056. Blanz, P. A. & Gottschalk, M. (1984) A comparison of 5S rRNA nucleotide sequences from smut fungi. Systematic and Applied Microbiology 5 : 518–526. Bruns, T. D., White, T. J. & Taylor, J. W. (1991) Fungal molecular systematics. Annual Review of Ecology and Systematics 22 : 525–564. Denchev, C. M. (1994) Validation of the name Microbotryum vinosum (Ustilaginales). Mycotaxon 50 : 331. Denchev, C. M. (1997) New combinations in Bauhinus (Microbotryaceae). Mycotaxon 65 : 419–426. Durrieu, G. (1980) Les me! canismes e! volutifs chez les Melampsora (Basidiomyce' tes, Ure! dinales). Comptes Rendus de l ’AcadeU mie des Sciences de Paris, seU rie 3, Sciences de la Vie 291 (se! rie D) : 849–852. Felsenstein, J. (1985) Confidence intervals on phylogenies : an approach using the bootstrap. Evolution 39 : 783–791. Delimitation of Microbotryum Galtier, N., Gouy, M. & Gautier, C. (1996) SEAVIEW and PHYLOjWIN : two graphic tools for sequence alignment and molecular phylogeny. Computer Applications in Bioscience 12 : 543–548. Gardes, M. & Bruns, T. D. (1993) ITS primers with enhanced specificity for basidiomycetes – application to the identification of mycorrhizae and rusts. Molecular Ecology 2 : 113–118. Hawksworth, D. L., Kirk, P. M., Sutton, B. C. & Pegler, D. N. (1995) Ainsworth & Bisby’s Dictionary of Fungi. 8th edn. CAB International, Wallingford. Lepage, B. A., Currah, R. S. & Stockey, R. A. (1994) The fossil fungi of the Princeton Chert. International Journal of Plant Science 155 : 828–836. Moore, R. T. (1992) The genus Bauhinus gen. nov. : for species of Ustilago on dicot hosts. Mycotaxon 45 : 97–100. Moore, R. T. (1996) An inventory of the phylum Ustomycota. Mycotaxon 59 : 1–31. Perlin, M. H. (1996) Pathovars or formae speciales of Microbotryum violaceum differ in electrophoretic karyotype. International Journal of Plant Sciences 157 : 447–452. Perlin, M. H., Hughes, C., Welch, J., Akkaraju, S., Steinecker, D., Kumar, A., Smith, B., Garr, S. S., Brown, S. A. & Andom, T. (1997) Molecular approaches to differentiate subpopulations or formae speciales of the fungal phytopathogen Microbotryum violaceum. International Journal of Plant Sciences 158 : 568–574. Prillinger, H., Doerfler, C., Laaser, G. & Hauska, G. (1990) A contribution to the systematics and evolution of higher fungi : yeast types in the basidiomycetes. Part III. Ustilago-type. Zeitschrift Fuer Mykologie 56 : 251–278. Prillinger, H., Deml, G., Do$ rfler, C., Laaser, G. & Lockau, W. (1991) Ein Beitrag zur Systematik und Entwicklungsbiologie ho$ herer Pilze : Hefe-Typen der Basidiomyceten. Teil II : MicrobotryumTyp. Botanica Acta 104 : 5–17. Prillinger, H., Oberwinkler, F., Umile, C., Tlachac, K., Bauer, R., Do$ rfler, C. & Taufratzhofer, E. (1993) Analysis of cell wall 548 carbohydrates (neutral sugars) from ascomycetous and basidiomycetous yeasts with and without derivitization. Journal of General and Applied Microbiology 39 : 1–34. Roux, C., Almaraz, T. & Durrieu, G. (1998) Phyloge! nie de champignons responsables des charbons des ve! ge! taux a' partir de l’analyse des se! quences ITS. Comptes Rendus de l ’AcadeU mie de Sciences de Paris, SeU rie 3, Sciences de la Vie 321 : 603–609. Savile, D. B. O. (1990) Coevolution of Uredinales and Ustilaginales with vascular plants. Reports of the Tottori Mycological Institute 28 : 15–24. Swann, E. C., Frieders, E. M. & McLaughlin, D. J. (1999) Microbotryum, Kriegeria and the changing paradigm in basidiomycete classification. Mycologia 91 : 51–66. Swann, E. C. & Taylor, J. W. (1993) Higher taxa of basidiomycetes : an 18S rRNA gene perspective. Mycologia 85 : 923–936. Swann, E. C. & Taylor, J. W. (1995) Phylogenetic diversity of yeastproducing basidiomycetes. Mycological Research 99 : 1205–1210. Swofford, D. L. (1993) PAUP : phylogenetic analysis using parsimony. Version 3±1.1. Center for Biodiversity, Illinois Natural History Survey, Champaign, IL. Thompson, J. D., Higgins, D. G. & Gibson, T. J. (1994) ClustalW improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Research 2 : 4673–4680. Va! nky, K. (1994) European Smut Fungi. Gustav Fisher, Stuttgart. Va! nky, K. (1998a) The genus Microbotryum (smut fungi). Mycotaxon 67 : 33–60. Va! nky, K. (1998b) Proposal to conserve the generic name Thecaphora against Sorosporium (Fungi, Ustilaginales). Taxon 47 : 153. Va! nky, K. (1999) The new classificatory system for smut fungi, and two new genera. Mycotaxon 70 : 35–49. Corresponding Editors : R. J. Vilgalys & D. L. Hawksworth