| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
What's this? |
Mycology and Plant Pathology |
210300 Baltimore Avenue, Room 304, Building 011A, Beltsville, Maryland 20705 USA; 3Box 90338, Department of Biology, Duke University, Durham, North Carolina 27708 USA
Received for publication November 1, 2006. Accepted for publication June 27, 2007.
ABSTRACT
In the Pucciniomycetes, a class of fungi that includes the plant pathogenic rust fungi, insect parasitism is restricted to a single family, the Septobasidiaceae. The Septobasidiaceae form a variety of symbioses with scale insects and have remained largely unstudied since the 1930s. Transitions between plant and animal parasitism and between mutualism and parasitism cannot be fully addressed in the Basidiomycota without a clear phylogenetic hypothesis for the Septobasidiales. Here, molecular phylogenetic methods were applied to understand the origin of scale insect parasitism, test the monophyly of the order Septobasidiales, and evaluate the infrageneric concepts in the largest genus of scale insect parasites, Septobasidium. DNA sequence data from rRNA genes were used to infer higher-level relationships within the Pucciniomycetes, and data from translation elongation factor 1-alpha (tef1) were added for phylogenetic inference within the Septobasidiaceae. Data from tef1 revealed different intron arrangements within Septobasidium, but the molecule did not provide much additional phylogenetically informative data. Likelihood-model-based phylogenetic analyses of 44 Pucciniomycotina taxa provided moderate support for a single origin of insect parasitism. Within the Septobasidiaceae, there was little or no support for a monophyletic Septobasidium, and well-resolved subclades of Septobasidium species contradict previous morphological delimitations of groups within the genus.
Key Words: Auriculoscypha • Coccoidea • mutualism • parasitism • Pucciniomycetes • rDNA • scale insect • tef1 • Uredinella
Mutualistic and parasitic symbioses between fungi and plants are widely acknowledged to have profound influences on the evolution and ecology of terrestrial life, but less well known are the symbioses between fungi and insects. Basidiomycete fungi have evolved many symbiotic associations with plants and animals, but fungi in the Septobasidiales are the only large group of basidiomycetes that are obligately parasitic on insects. Understanding the evolution of insect parasitism and switches from plant parasitism in the Basidiomyocta requires a phylogeny to place the Septobasidiales within the Pucciniomycotina and to determine whether the different forms of insect parasitism in the Septobasidiales and Septobasidium have a single origin. Although Septobasidium sterilize the individuals they parasitize, the fungi may protect uninfected individuals and thereby benefit the population of scale insects (Couch, 1938
). All fungi in the Septobasidiales do not display this type of symbiosis. Some may be wholly parasitic because they do not form substantial protective structures. This fungus–insect symbiosis is important because of its unique altruistic and parasitic characteristics and because of its phylogenetic position within the Basidiomycota, but few studies since the 1930s have focused on elucidating either the ecology or evolution of the Septobasidiales.
The order Septobasidiales, which contains two families and approximately 180 described species, is currently placed taxonomically within the otherwise plant pathogenic Pucciniomycetes, the subclass that includes the well-known rust fungi (Swann et al., 2001
; Bauer et al., 2006
). All fungi in the family Septobasidiaceae are parasitic on scale insects (Coccoidea), while Pachnocybe, the only genus in the family Pachnocybaceae, is found in wood. Few morphological characters unite the Septobasidiales as distinct within the Pucciniomycetes. Pachnocybe has been linked to the Septobasidiaceae based on weakly supported phylogenetic inference (Frieders, 1997
) and the reported presence of microscala, cross-linking membranes between mitochondria and sometimes endoplasmic reticulum, in the few species examined (Kleven and McLaughlin, 1989
; Bauer and Oberwinkler, 1990
). The presumed monophyly of the Septobasidiaceae is based entirely on its association with scale insects. However, dependence on host association as a phylogenetic character is known to be misleading with regard to phylogenetic relationships in other fungal groups such as the Hypocreales (Rossman, 2000
; Artjariyasripong et al., 2001
), and inter-kingdom host switching has been documented in the life cycle of at least one Pucciniomycete genus, Helicobasidium (Lutz et al., 2004
).
Within the Septobasidiaceae, segregation of sections and genera has been problematic because the fungi are morphologically reduced but highly variable (Figs. 1–12). Five genera have been included in this family, but almost all of the known species (
175) are placed in the genus Septobasidium despite differences among species in gross morphology (Couch, 1938
). The genus Auriculoscypha was first described as related to Auricularia (Reid and Manimohan, 1985
), but upon discovery of a coccid associate, it was moved to the Septobasidiales (Lalitha and Leelavathy, 1990
). Uredinella was described by Couch as an intermediate fungus between Septobasidium and the rust fungi (Couch, 1941
). Oberwinkler (1989)
suggested that Septobasidium is an overly broad generic concept, and he erected a new genus Coccidiodictyon and argued for the resurrection of the genus Ordonia. Coccidiodictyon and Ordonia were described as closely related to Septobasidium and Uredinella, respectively (Oberwinkler, 1989
).
|
). To our knowledge, only three new species have been described since Couch's last descriptions of Septobasidium species in 1946 (Gomez and Kisimova, 2001
; Henk, 2005
). Most knowledge of the biology, morphology, and systematics of Septobasidium originates from Couch's (1938)
authoritative publication, The Genus Septobasidium.
Most species within the genus Septobasidium grow as mats of hyphae covering and embedding scale insects on branches and leaves of trees. A few species form only a minimal web of hyphae external to infected insects, and some species form large structures that protrude from the underlying branch or leaf. Because the hyphal mat covers both infected and uninfected insects, Couch (1931)
postulated that the fungi actually protect uninfected insects from parasitoid wasps and desiccation, resulting in a mutualism between colonies of insects and fungi. However, no published literature since has corroborated that assertion. Some species form special structures under which insects may settle. These structures are useful for species identification and are termed "insect houses" or "insect traps" (Figs. 5, 6), depending on the interpretation of the relationship (Couch, 1938
). Specialized hyphae called haustoria grow inside the infected scale insects. These are categorized by shape as coiled, regularly coiled, or spindle shaped.
In the only available circumscription of Septobasidium, Couch (1938)
divided it primarily based upon morphological characters related to the structure of the basidium and probasidium. Basidia may be curved or straight and may have one, two, three, or four cells. The probasidium either remains as an empty cell at the base of the mature basidium (persistent) or becomes a spore-bearing cell in the mature basidium (not persistent). Other characters used as indicators of major groups in Septobasidium include the layered nature of the thallus and the presence of pillar structures. Despite this wide range in macromorphology, classification beyond species identification has remained problematic because homology assessment is difficult for these characters and they are often considered misleading in higher-level classification (Oberwinkler, 1989
).
Molecular phylogenetic work in the Pucciniomycotina has proven to be informative in morphologically reduced groups such as yeasts and has shown that many morphologically based genera are polyphyletic (Fell et al., 2001
). To date, only a few molecular studies have attempted to assess relationships within the important basidiomycete group Pucciniomycetes (Swann and Taylor, 1995
; Frieders, 1997
; Maier et al., 2003
), and none have attempted to test the monophyly of the Septobasidiales or the Septobasidiaceae. Generic, subgeneric, and species concepts have been tested in some Pucciniomycetes using molecular phylogenetic techniques (Maier et al., 2003
; Chung et al., 2004
; Lutz et al., 2004
; Tian et al., 2004
; Chatasiri et al., 2006
; Liang et al., 2006
; Szabo, 2006
), but no studies have focused on the Septobasidiales.
To address the monophyly of the Septobasidiales and the origin of insect association in the Pucciniomycotina, we used sequence data from large and small subunits of the nuclear ribosomal RNA genes (nLSU-rDNA and nSSU-rDNA, respectively). To address the phylogenetic relationships within the Septobasidiaceae and the genus Septobasidium, we added sequence data from the internal transcribed spacer region of the nuclear rRNA (ITS) and the protein-coding gene, translation elongation factor 1 alpha (tef1). Our objectives were to (1) assess the monophyly of the Septobasidiales, Septobasidiaceae, and Septobasidium; (2) establish the sister-group relationships of the monophyletic entities within the Septobasidiales; and (3) evaluate the efficacy of morphological criteria used to define groups within the Septobasidiales.
MATERIALS AND METHODS
Specimen collection and taxon sampling
We obtained all type specimens and many other specimens of Septobasidium and Uredinella housed at the UNC herbarium where Couch had assembled almost all of the material used for his monograph. Although we were not able to obtain DNA suitable for PCR from the herbarium specimens, they were referenced for identification of subsequent collections. We collected specimens primarily from the southeastern United States and Costa Rica to represent the major groups of Septobasidium informally defined by Couch and Uredinella coccidiodphaga from Florida (Table 1; Appendix 1). Collections were made by removing infected plant parts from host trees and generally included sufficient material to enable DNA extraction from the fungus, host insects, and host plant. Dr. P. Manimohan kindly provided us with specimens of Auriculoscypha anacardiicola from its type locality, and Dr. F. Porcelli and Ms. J. Hoy provided Septobasidium specimens from Italy and New Jersey. All collections have been deposited in the Duke cryptogamic herbarium (DUKE). Dr. D. McLaughlin kindly provided us with cultures of Pachnocybe, Eocronartium, Jola, and Helicobasidium.
|
and the orders within the Urediniomycetidae, later modified to Pucciniomycetes by Bauer et al. (2006)
. We then used a narrower range of taxa to determine generic and subgeneric relationships within the Septobasidiales. We chose Helicobasidium to use as a single outgroup taxon from among the potential Pucciniomycetes outgroups based on branch length in the larger analysis and because alignments suggested that there we could include more unambiguously alignable characters if we did not include taxa from the Pucciniales or Platygloeales.
DNA extraction, PCR amplification, DNA sequencing
DNA from fungal tissue was extracted following extraction protocols using CTAB extraction buffer (Zolan and Pukkila, 1986
). Mycelia were placed in extraction buffer and ground with a plastic pestle, and yeast cells were scraped from the surface of culture media and lysed in extraction buffer by vortexing with sterile sand for 2 min. Proteins were removed with 24 : 1 isoamylalcohol : chloroform, and DNA precipitated with isopropanol. DNA extracts were eluted in 100 µL of sterile H2O for use directly in PCR. PCR followed a standard cycling protocol –95°C for 4 min followed by a loop repeating 35 times of 95°C for 30 s, 50°C (48°C with SEF1a primers) for 30 s, 72°C for 60 s; finishing with 72°C for 420 s. We used standard primers ITS1, ITS4, NS1, NS4 (White et al., 1990
), LR0r, LR5, or LR3 (Vilgalys and Hester, 1990
) for PCR amplification and sequencing of rRNA genes. We also designed the primers UredinioSSUf (5'-ATCAATTGGANGGCAAGTNTGG-3') and Septo5.8sf (5'-TTTGAACGCACCTTGCAC-3') for rRNA genes to exclude some nontarget fungal sequences and to amplify and sequence regions where other primers could not anneal because of group I intron insertions at priming sites. We used primers SEF1a1fI (5'-CTYGGIAAGGGITCNTTCAAG-3'), matching the amino acid sequence LGKGSFK, and SEF1a1r2 (5'-CATICCGGCCTTGATNGTNCC-3'), matching the reverse complement of the amino acid sequence IKAGMVVT, for the amplification of tef1 genes. Automated sequencing was performed on an Applied Biosystems Incorporated (Foster City, California, USA) 3700 using ABI BigDye terminator mix V2.1 or V3.0.
Sequence alignment and phylogenetic analyses
Sequences were inspected and edited using Sequencher 4.1 (Gene Codes, Ann Arbor, Michigan, USA), then deposited in GenBank (accession nos. DQ241405–DQ241504 and DQ648057–DQ648073). Edited sequences and other pucciniomycete sequences from GenBank (accession nos. AB021697, AB055193, AB178480, AF189985, AF522164, AF522166, AF522182, AY123286, AY123307, AY512887, AY629314, AY634278, AY665775, AY745723, L20282, M94338, U78043—see Appendix 1) were then aligned manually using MacClade 4.05 (Maddison and Maddison, 2002
). Ambiguously aligned regions including all introns were excluded from phylogenetic analysis. Each gene was analyzed separately and then compared to determine whether there was significant conflict between their inferred phylogenies. We estimated the phylogeny using maximum parsimony (MP) analyses with PAUP* 4.10b (Swofford, 2003
), and Bayesian analyses with MrBayes v3.1.2 (Huelsenbeck and Ronquist, 2001
; Ronquist and Huelsenbeck, 2003
). For Bayesian analyses, we allowed each gene region to follow a separate model of evolution chosen using Modeltest 3.06 (Posada and Crandall, 1998
). The GTR + I + 
model was used in all rDNA analyses with default settings for the priors and separate models for each region. The following models were used for tef1 codon positions 1, 2, and 3, respectively: a GTR + I + model, a one substitution type (F81) + I + model with default priors, and GTR + model with default priors. Trees saved every 100 generations from four iterated runs of 1 000 000 generations were checked for convergence and assembled into a single distribution composed of 20 000 sampled trees after discarding burn in trees. The proportion of a bipartition's occurrence in the distribution was considered its posterior probability (PP). We also assessed statistical support for nodes in the phylogeny using MP bootstrapping (MPB) and Bayesian bootstrapping (BB). For MPB analyses all constant characters were excluded, and each of 100 bootstrap replicates was analyzed using heuristic searches with TBR with a maximum of 500 trees saved from each replicate. For BB analyses, we used the bootstrap method of the P4 package (Foster, 2003 http://www.nhm.ac.uk/zoology/external/p4.htm) to generate bootstrapped data sets and a short Python script to format those data sets for use in MrBayes. We used 100 bootstrap replicates each run in MrBayes with the same parameters as those used for the entire data set. However, we used only one run for each bootstrapped data set rather than four iterated replicate runs, and we discarded all but the final 1000 trees from each run. We concatenated the trees from each bootstrap replicate. The proportion of a bipartition's occurrence in the 100 000 concatenated trees was its BB support value.
Morphological data
To determine the efficacy of traditional morphological characters for defining natural groups within the Septobasidiales, we first scored six characters for the species sampled using light microscopy (Table 1). We did not have the original specimen for the S. carestianum Bres. culture that we sequenced and instead used data from herbarium specimens and published data (Couch, 1938
) for this species. We focused on characters used in the only available key for Septobasidium and characters previously identified as potentially phylogenetically useful (Couch, 1938
; Oberwinkler, 1989
). We scored three characters related to the basidium: (1) persistence or not of the probasidium (Figs. 9, 10), (2) coiled or straight basidium (Figs. 10, 11), and (3) the number of cells in the mature basidium (Fig. 9). We also noted the presence or absence of vertically arranged hyphae below the fertile layer and distinctly separated from each other and other contextual elements (pillars—Fig. 4), and three distinct horizontal layers of mycelial growth. We grouped the type of haustoria into two categories, spindle-shaped or coiled (Figs. 7, 8). Once we scored the characters, we examined their distribution on the phylogenetic trees. Characters that were consistent with the phylogeny were considered most useful for defining natural groups.
RESULTS
Analysis of SSUr DNA and LSUr DNA from Pucciniomycetes
Our broad taxonomic sampling included 44 taxa representing major lineages within the Pucciniomycotina and the major groups within the Pucciniomycetes. For the SSUr DNA, we included 998 unambiguously aligned base pairs, and for the LSUr DNA, we included 797 unambiguously aligned base pairs. Group I introns lacking homing endonucleases similar to those found in other fungi (Perotto et al., 2000
; Lickey et al., 2003
) were detected and sequenced within both the nSSU-rDNA and nLSU-rDNA. Most analyses converged in less than 30 000 generations and all reached stationarity by 50 000 generations. Independent runs converged on the same likelihood plateau in each analysis. Independent analyses of the nLSU-rDNA and nSSU-rDNA data sets did not reveal conflict between the regions. Therefore, we proceeded with analysis of the combined genes data set (Fig. 13). Support from MPB, PP, and BB increased in the combined analysis relative to the independent analyses. The accepted major groups within the Pucciniomycotina including the Pucciniomycetes were recovered with moderate to high support. For the Bayesian analyses, the major groups within the Pucciniomycetes were recovered including a monophyletic Pucciniales (rust fungi) clade, a Platygloeales clade, and the Septobasidiales, including Pachnocybe and the Septobasidiaceae. However, there was conflict between MPB and PP in the placement of Uredinella. In the MPB analysis, Uredinella was strongly supported as a member of the Pucciniales, but PP showed support for Uredinella being within the Septobasidiales. When 10 nucleotide matrices generated from the 95% PP tree under a GTR + I + model using Mesquite 1.0 (Maddison and Maddison, 2003, http://mesquiteproject.org) were reanalyzed with parsimony, MPB consistently misplaced Uredinella in 10 of 10 simulations, most often in the same or similar phylogenetic position as the MPB analysis of the actual nSSU-rDNA and nLSU-rDNA data (results not shown). However, Bayesian analyses consistently recovered the correct topology in 10 of 10 simulations whether matrices were simulated based on the PP or MPB topology or under a simple Jukes-Cantor or GTR + I + model.
|
Independent analyses of the rDNA regions yielded trees that were not significantly in conflict with each other regardless of the method of analysis. The rDNA data were combined into a single matrix and analyzed together to infer phylogenetic relationships (Fig. 14). The combined analysis of rDNA regions resolved a monophyletic statistically supported Septobasidiaceae but a paraphyletic Septobasidium. Auriculoscypha was recovered as sister group to the rest of the Septobasidiaceae with only weak support by all three methods. Uredinella along with S. burtii, S. canescens Burt, S. castaneum Burt, S. ramorum, S. michelianum Pat., and S. pinicola Snell were unresolved with respect to the clades I–IV within Septobasidium.
|
Clade II includes taxa from Costa Rica and the United States and was recovered with strong support. All species in clade II have spindle- or sausage-shaped haustoria. Within this clade, S. apiculatum is unresolved with respect to subclades IIa and IIb. It is the only member of clade II without a persistent probasidium or a third horizontal layer and with three cells per basidium. Subclade IIa contains three species. Septobasidium fumigatum is strongly supported as sister to an undescribed species from Costa Rica that is morphologically quite distinct (Henk, 2005
). The two specimens differ dramatically in gross morphology, having different contextual arrangements and strikingly different hues. However, they have some similar microscopic characters, including the spindle type of haustoria. Although the Costa Rican specimen may not be formally described, it is clearly a distinct species closely related to S. fumigatum Burt. Septobasidium grandisporum Couch is similar to S. fumigatum in macroscopic characters but is one of only two described species of Septobasidium with basidia that give rise to a single large spore.The species in clade IIb overlap in their morphological characteristics, all of them having two cells per basidium, similar-sized spores, similar hymenia, grey and black colorings, and similar margins. Differentiating these species can be particularly difficult in the tropics where the few distinguishing characteristics such as pillars and the presence of rhizomorphs may be variable within a colony or completely disrupted or obscured by the abundance of other epiphytes. In clade IIb the two S. griseum collections were recovered as sister taxa, but the S. sinuosum collection from Costa Rica was sister to the other Costa Rican material rather than the S. sinuosum specimen from the USA.
Clade III and the relationships within it were weakly supported by PP, supported by MPB, but not supported by BB and. This group includes specimens collected from Costa Rica and the United States. All the taxa in this clade lack pillars and have four-celled basidia with persistent probasidia, except S. pilosum, which has not been observed with fully developed basidia. Septobasidium arachnoideum and S. taxodii Couch have coiled basidia, while S. wilsonianum and S. meredithiae have straight club-like basidia. Septobasidium arachnoideum and S. pilosum have spindle-shaped haustoria, while the other species have haustoria of the irregularly coiled type. Septobasidium taxodii, S. pilosum, and S. wilsonianum each forms a thin subiculum with distinct spike-like upright hyphae. Septobasidium arachnoideum and S. meredithiae each forms a more distinct third layer supported by thinner upright hyphae.
Clade IV was weakly supported by PP, BB, and MPB analyses, but the taxa are strongly linked by morphology. The three members of this clade have pillars, lack persistent probasidia, and have four-celled basidia. Septobasidium septobasidioides Lloyd, which has straight to somewhat curved basidia, was strongly supported as sister to S. westonii Couch, which has distinctly coiled basidia. Septobasidium gomezii Henk shares a distinctive insect house type, basidial type, and haustorial type with S. septobasidioides. However, S. septobasidioides has a papery, grey hymenial layer and a taller less dense context than the similar taxon with a nonpapery, salmon-colored hymenial layer.
Data from tef1 resolved few supported clades (Fig. 14). Like the rDNA analysis, the PP of the tef1 data supported S. pseudopedicellatum and S. cokeri as sister taxa as were S. grandisporum and S. fumigatum. Unlike the rDNA analysis, S. mariani was recovered as sister to S. alni, and S. apiculatum was sister to S. sinuosum, each with strong PP support in the tef1 analysis. The relationships of S. apiculatum to S. sinuosum and S. mariani to S. alni are unresolved in the rDNA tree (both a reduced sampling that matches the tef1 data set and the complete 33 taxon sampling) and therefore not in conflict with tef1 phylogeny. Neither BB nor MPB analyses supported any nodes in the tef1 phylogeny. Because there was no topological incongruence, we combined the rDNA and tef1 data for the 19 taxa. Most of the nodes supported by the rDNA alone were recovered in the combined analysis (results not shown). In the combined tree, S. sinuosum was sister to S. apiculatum with MPB support above 70% but without support by PP or BB, and S. mariani and S. alni were recovered as sister taxa with high PP and MPB support. Septobasidium taxodii was not recovered as sister to the other representatives of clade III—S. meredithiae and S. pilosum. When all the taxa were combined into a single matrix with missing data for those taxa that lacked tef1 sequences, some of the rDNA phylogeny clades were recovered. However, many clades were no longer significantly supported, overall support values declined, and no clades were resolved that were not present in the rDNA phylogeny.
DISCUSSION
Monophyly of Septobasidiales and Septobasidiaceae
Understanding the evolution of insect parasitism, its origin and possible transition to mutualism in the Basidiomycota, requires a phylogenetic hypothesis for the members of the Septobasidiales. Little morphological evidence supports the monophyly of the Septobasidiales. In our molecular analyses we detected not only some support for a monophyletic Septobasidiales, but also some conflict between phylogenetic methods. While PP supports a monophyletic Septobasidiales and Septobasidiaceae, BB is equivocal, and MPB supports an alternative hypothesis. However, the conflict is not in the placement of Pachnocybe, the only Septobasidialean fungus not parasitic on scale insects and linked morphologically only by the unpublished report of microscala (Swann et al., 2001
). Pachnocybe is well supported as sister to the Septobasidiaceae by PP and unresolved within the Pucciniomycetes by MPB or BB (Fig. 13). The conflict occurs in the placement of Uredinella. The alternative hypothesis supported by MPB places Uredinella sister to the Pucciniales rather than the Septobasidiales. This phylogenetic position was implicitly considered a competing hypothesis to a monophyletic Septobasidiaceae by Couch (1937)
, who placed Uredinella as intermediate between Septobasidium and the Pucciniales. Unlike Septobasidium or Auriculoscypha, Uredinella infects only a single insect and forms a thin mycelial mat covering only the infected insect. It also differs from Septobasidium and Auriculaoscypha in microscopic characters, particularly in that Uredinella forms a very thick-walled probasidium that has a distinct apical pore (Fig. 1) and gives rise to either a detached four-celled basidium or a binucleate cylindrical spore. It was the presence of these characters that led Couch to link Uredinella to the rusts. He considered the Uredinella probasidium similar to a teliospore, and the binucleate spore reminiscent of a urediniospore or aeciospore. However, Uredinella is on a very long branch, and its placement is suspect because of long-branch attraction, particularly in parsimony analyses (Anderson and Swofford, 2004
). Our simulation studies confirm that long-branch attraction probably explains the placement of Uredinella in MPB analyses. We tentatively conclude that the Septobasidiales is monophyletic and that there is a single origin of scale insect parasitism in the Pucciniomycetes defining a monophyletic Septobasidiaceae. Because PP is the only measure that supports these hypotheses, we must remain cautious. Although PP is known to better recover correct short internodes that cannot easily receive high bootstrap support, PP may be sensitive to branch lengths as well (Alfaro et al., 2003
; Yang and Rannala, 2005
). Here we favor the PP support in part because, in our simulation analyses based on both topologies, PP consistently recovered only the simulated topology.
The monophyly of the Septobasidiaceae confirms another large clade of morphologically diverse Pucciniomycotina that is completely restricted to a parasitic life style. The Pucciniales represent the largest monophyletic assemblage of obligate plant parasitic Basidiomycota, and the Septobasidiaceae likely represent the largest monophyletic group of obligate insect parasitic Basidiomycota. That these lineages are closely related suggests that a feature of the Pucciniomycetes may encourage phylogenetic fidelity to a host-restricted parasitic life style. However, the recent connection of the anamorphic Tuberculina, which is a parasite on rust fungi, to its teleomorph Helicobasidium, a plant pathogen, shows that inter-kingdom host switching can occur even within the lifecycle of a fungus (Lutz et al., 2004
). Although the Septobasidiaceae are monophyletic, the close association of scale insects to plants easily provides an opportunity for pathogenic fungi to shift hosts, as in the Hypocreales. Other urediniomycete parasites of scale insects may yet be discovered as the study of molecular ecology expands to include more of the phyllosphere and its inhabitants.
Monophyly of Septobasidium and phylogenetic relationships within the Septobasidiaceae
Within the Septobasidiaceae, natural groups are difficult to define based on morphological characters, and our analyses of rDNA did not resolve a monophyletic Septobasidium. Although Uredinella and Auriculoscypha are quite distinct from Septobasidium species, they are also the only members of their respective genera, making the matter of their respective monophyly trivial and their positions in the rDNA phylogeny along the basal unresolved portion of the Septobasidiaceae uninformative. Within Septobasidium, several well-supported clades exemplify the difficulty of using previously defined morphological characters to determine natural groups. Even the most conservative approach, that is, accepting only clades strongly supported by all three phylogenetic methods, reveals that any one of the morphological characters is not sufficient to define a natural group (Fig. 15). Clade I includes the most common morphology and the type species of the genus. Clade I is rather uniform and unites taxa that all have at least three layers, pillared growth, persistent probasidia, and irregularly coiled haustoria. However, not all taxa with these characteristics are included in this clade. Interestingly S. ramorum and S. castaneum share clade I characteristics, are common, and are frequently collected along with clade I representatives S. mariani, S. pseudopedicellatum, and S. cokeri. Clade IV is similarly homogeneous, with each species having irregularly coiled haustoria, pillars, and four-celled basidia lacking persistent probasidia. Consistency begins to break down in clade III. Although all members of clade III lack pillars and have four-celled basidia with persistent probasidia, clade III includes both spindle-shaped and irregularly coiled haustoria. Virtually every character breaks down in clade II, the best-supported major clade in every analysis. Multiple basidial morphologies occur in clade II, including one-, two-, three-, and four-celled basidia, most of which have persistent probasidia, although the three-celled basidium of S. apiculatum lacks a persistent probadium. Clade II includes some species that form pillars, some that do not, and also S. apiculatum, which does not even form a three-layered structure. Only the spindle-shaped haustorial type is shared by all members of clade II. Overall, the key characters of pillar formation, haustorial type, and a persistent probasidium are not completely consistent with the rDNA phylogeny but require only one reversal or parallelism to fit the well-supported portions of the rDNA phylogeny. Only the number of basidial cells, which is not necessarily a simple two-state character, would require more than a single homoplasious event to fit the rDNA phylogeny. At least one key character unifies each of the major clades, and in combination they define natural monophyletic groups. Focusing on the combination of these key characters is promising as a way forward in defining groups within the Septobasidiaceae, but further studies are needed to confirm the efficacy of any particular combination to define natural groups.
|
Molecular phylogenetic analyses
We encountered three interesting challenges in our phylogenetic analyses. First, we did not recover similar values of support from our different phylogenetic measures. Not only does this require that we remain cautious about assigning much confidence to our phylogenetic inference, but it requires some explanation of the differences between the methods. The overall trend was for PP to be much higher than either BB or MPB, as with many studies using these methods. One explanation is that PP recovers nodes that are just too short to be supported by the resampling methods such as bootstrap or jackknife (Alfaro et al., 2003
). Our data seem consistent with this explanation because most of the nodes with high PP and little or no bootstrap support are quite short. Only the sister relationship of S. apiculatum and clade IIb was supported by MPB and not supported by PP. Because MPB is the most unreliable of the methods and PP is the only method that can correctly recover relationships with short internodes, we proceed by assuming that nodes recovered with high PP support are correct as long as they are not contradicted by the other methods. The second challenge arose when our different measures of support actually conflicted in one case (discussed in detail later). Because this conflict involved two lineages on long branches, we suspected that MPB might be incorrect because of long-branch attraction. This was confirmed as a likely explanation for the observed conflict by our simulations. It is not surprising that model misspecification and long branches led to an incorrect inference of phylogeny (Buckley, 2002
). However, MPB is still common currency in much of the phylogenetic literature, and consistency across analytical methods can provide confidence beyond any one statistical measure. Our third problem was that the analysis of the tef1 data that supported few nodes. Complicating the analysis of tef1 data was our inability to recover the DNA sequence via PCR for any but the best-preserved and recently collected material. A PCR strategy might have been developed to obtain more complete results, but we considered this unlikely because of the otherwise high functionality of the given PCR conditions, even on distantly related taxa such as the Sporidiobolales (results not shown), and the consistency of failure on older or poorly prepared material. Regardless, the current analyses were left with either a reduced taxon sampling for which we could combine the rDNA with tef1 data or a larger taxon sampling with an abundance of missing data. We used both approaches for combining the data. For the reduced taxon sampling with all the data included, we recovered some support for two relationships not recovered by the rDNA alone, but we did not recover all the nodes originally recovered by the rDNA alone. However, we discovered that the combined data set that included large amounts of missing data had greatly reduced support values as compared to the data set without the tef1 data added. Reduction in support has been noted before as an effect of adding missing data (Wilkinson, 1995
, but see Wiens, 2003
, 2005
). Further sequencing of tef1 may improve phylogenetic inference in the Septobasidiales.
APPENDIX
Taxon—GenBank accessions: 18S, 25S, ITS, tef1; Source; Voucher specimen.
Agaricostilbum hyphaens—AY665775, AY634278, —, —; —; —. Auriculoscypha anacardiicola—DQ241434, DQ241503, DQ241470, —; P. Manimohan; DUKE 40144. Coleosporium asterum—AY123286, AF522164, —, —; —; —. Cronartium ribicola—M94338, AF522166, —, —; —; —. Eocronartium musicola—DQ241438, AF014825, —, —; —; —.
Helicobasidium mompa—DQ241439, AY254178, DQ241472, DQ648074; ATCC 56070; —. Helicogloea variabilis—U78043, L20282, —, —; —; —. Jola cf. javensis—DQ241437, DQ241505, —, —; E. Fieders/D. Mclaughlin; —. Occultifur externus—AB055193, AY745723, —, —; —; —. Pachnocybe ferruginea—DQ241440, L20284, DQ241473, DQ648075; ATCC 66701; —. Platygloea disciformis—DQ241436, AY629314, —, —; E. Fieders/D. Mclaughlin; —. Septobasidium alni—DQ241405, DQ241474, DQ241441, DQ648057; D.A. Henk; DUKE 40149. Septobasidium apiculatum—DQ241406, DQ241475, DQ241442, DQ648058; D.A. Henk; DUKE 40143. Septobasidium arachnoideum—DQ241407, DQ241476, DQ241443, —; D.A. Henk; DUKE 40136. Septobasidium burtii—DQ241408, DQ241477, DQ241444, DQ648059; D.A. Henk; DUKE 40129. Septobasidium canescens—DQ241410, DQ241479, DQ241446, DQ648061; D.A. Henk; DUKE 40120. Septobasidium carestianum—DQ241412, DQ241481, DQ241448, DQ648063; D.A. Henk; —. Septobasidium castaneum—DQ241411, DQ241480, DQ241447, DQ648062; D.A. Henk; DUKE 40134. Septobasidium cavarae—DQ241409, DQ241478, DQ241445, DQ648060; F. Porcelli/D.A.Henk; DUKE 40125. Septobasidium cokeri—DQ241413, DQ241482, DQ241449, DQ648064; D.A. Henk; DUKE 40148. Septobasidium fumigatum—DQ241415, DQ241484, DQ241451, DQ648066; D.A. Henk; DUKE 40131. Septobasidium gomezii—DQ241426, DQ241495, DQ241462, —; D.A. Henk; DUKE 40135. Septobasidium grandisporum—DQ241417, DQ241486, DQ241453, DQ648067; D.A. Henk; DUKE 40139. Septobasidium griseum—DQ241418, DQ241487, DQ241454, —; D.A. Henk; DUKE 40145. Septobasidium griseum—DQ241419, DQ241488, DQ241455, —; D.A. Henk; DUKE 40128. Septobasidium mariani—DQ241420, DQ241489, DQ241456, DQ648068; D.A. Henk; DUKE 40126. Septobasidium meredithae—DQ241429, DQ241498, DQ241465, DQ648072; D.A. Henk; DUKE 40146. Septobasidium michelianum—DQ241421, DQ241490, DQ241457, DQ648069; F. Porcelli/D.A. Henk; DUKE 40123. Septobasidium pilosum—DQ241422, DQ241491, DQ241458, —; D.A. Henk; DUKE 40124.
Septobasidium pinicola—DQ241423, DQ241492, DQ241459, —; D.A. Henk; DUKE 40147. Septobasidium pseudopedicellatum—DQ241424, DQ241493, DQ241460, DQ648070; D.A. Henk; DUKE 40118. Septobasidium ramorum—DQ241414, DQ241483, DQ241450, DQ648065; D.A. Henk; DUKE 40116. Septobasidium septobasidioides—DQ241425, DQ241494, DQ241461, —; D.A. Henk; DUKE 40132. Septobasidium sinuosum—DQ241427, DQ241496, DQ241463, DQ648071; D.A. Henk; DUKE 40150. Septobasidium sinuosum—DQ241428, DQ241497, DQ241464, —; D.A. Henk; DUKE 40142. Septobasidium sp.—DQ241416, DQ241485, DQ241452, —; D.A. Henk; DUKE 40140. Septobasidium taxodii—DQ241430, DQ241499, DQ241466, DQ648073; D.A. Henk; DUKE 40113. Septobasidium velutinum—DQ241431, DQ241500, DQ241467, —; D.A. Henk; DUKE 40137. Septobasidium westonii—DQ241432, DQ241501, DQ241468, —; F. Porcelli/D.A. Henk; DUKE 40130. Septobasidium wilsonianum—DQ241433, DQ241502, DQ241469, —; D.A. Henk; DUKE 40138. Sporidiobolus salmonicolor—AB021697, AY512887, —, —; —; —. Sporobolomyces gracilis—AB178480, AF189985, —, —; —; —. Uredinella coccidiophaga—DQ241435, DQ241504, DQ241471, —; D.A. Henk; DUKE 40114. Uromyces appendiculatus—AY123307, AF522182, —, —; —; —.
FOOTNOTES
1 The authors thank D. J. McLaughlin, E. M. Frieders, F. Porcelli, J. Hoy, L. D. Gomez, and P. Manimohan for providing valuable discussion as well as specimens and/or DNA sequence data. They especially thank T. Y. James for assistance in the development of primers, L. Bukovnik for sequencing, and C. Cox for valuable input regarding phylogenetic analysis. They also thank anonymous reviewers for their valuable comments. This research was supported in part by NSF grant DEB-0408011. 
4 Author for correspondence (e-mail: dan{at}nt.ars-grin.gov
)
LITERATURE CITED
Alfaro M. E. Zoller S. Lutzoni F.. 2003. Bayes or bootstrap? A simulation study comparing the performance of Bayesian Markov chain Monte Carlo sampling and bootstrapping in assessing phylogenetic confidence. Molecular Biology and Evolution 20: 255-266..
Anderson F. E. Swofford D. L.. 2004. Should we be worried about long-branch attraction in real data sets? Investigations using metazoan 18S rDNA. Molecular Phylogenetics and Evolution 33: 440-451..[CrossRef][Web of Science][Medline]
Artjariyasripong S. Mitchell J. I. Hywel-Jones N. L. Gareth-Jones E. B.. 2001. Relationship of the genus Cordyceps and related genera, based on parsimony and spectral analysis of partial 18S and 28S ribosomal gene sequences. Mycoscience 42: 503-517..[CrossRef]
Bauer R. Begerow D. Sampaio J. P. Weiss M. Oberwinkler F.. 2006. The simple-septate basidiomycetes: a synopsis. Mycological Progress 5: 41-66..[CrossRef][Web of Science]
Bauer R. Oberwinkler F.. 1990. Meiosis, spindle pole body cycle, and taxonomy of the heterobasidiomycete Pachnocybe ferruginea. Plant Systematics and Evolution 172: 241-261..[CrossRef][Web of Science]
Buckley T. R.. 2002. Model misspecification and probabilistic tests of topology: evidence from empirical data sets. Systematic Biology 51: 509-523..[CrossRef][Web of Science][Medline]
Chatasiri S. Kitade O. Ono Y.. 2006. Phylogenetic relationships among Puccinia hemerocallidis, P. funkiae, and P. patriniae (Uredinales) inferred from ITS sequence data. Mycoscience 47: 123-129..[CrossRef]
Chung W. H. Tsukiboshi T. Ono Y. Kakishima M.. 2004. Morphological and phylogenetic analyses of Uromyces appendiculatus and U. vignae on legumes in Japan. Mycoscience 45: 233-244..
Couch J. N.. 1931. The biological relationship between Septobasidium retiforme (B. & C.) Pat. and Aspidiotus osborni New. and Ckll. Quarterly Journal of Microscopical Science 74: 383-437..
Couch J. N.. 1937. A new fungus intermediate between the rusts and Septobasidium. Mycologia 29: 665-673..[CrossRef]
Couch J. N.. 1938. The genus Septobasidium. University of North Carolina Press, Chapel Hill, North Carolina, USA..
Couch J. N.. 1941. A new Uredinella from Ceylon. Mycologia 33: 405-410..[CrossRef]
Couch J. N.. 1946. Two species of Septobasidium from Mexico with unusual insect houses. Journal of the Elisha Mitchell Scientific Society 62: 87-94..
Fell J. W. Boekout T. Fonseca A. Sampaio J. P.. 2001. Basidiomycetous yeasts. In D. J. McLaughlin, E. G. McLaughlin, P. A. Lemke, editors Systematics and evolution: the Mycota VII, 1-36. Springer-Verlag, Berlin, Germany..
Frieders E. M.. 1997. An integrated approach to understanding the moss parasites and their role in basidiomycete evolution. Ph.D. dissertation, University of Minnesota, St. Paul, Minnesota, USA..
Gómez L. D. Henk D. A.. 2004. Validation of the species of Septobasidium (Basidiomycetes) described by John N. Couch. Lankesteriana 4: 75-96..
Gómez L. D. Kisimova-Horovitz L.. 2001. A new species of Septobasidium from Costa Rica. Mycotaxon 80: 255-259..[Web of Science]
Henk D. A.. 2005. New species of Septobasidium from southern Costa Rica and the southeastern United States. Mycologia 97: 908-913..
Huelsenbeck J. P. Ronquist F.. 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17: 754-755..
Kleven N. J. McLaughlin D. J.. 1989. A light and electron microscopic study of the developmental cycle in the basidiomycete Pachnocybe ferruginea. Canadian Journal of Botany 67: 1336-1348..[CrossRef]
Lalitha C. R. Leelavathy K. M.. 1990. A coccid association in Auriculoscypha and its taxonomic significance. Mycological Research 94: 571-572..[Web of Science]
Liang Y. Tian C. Kakishima M.. 2006. Phylogenetic relationships on 14 morphologically similar species of Pucciniastrum in Japan based on rDNA sequence data. Mycoscience 47: 137-144..[CrossRef]
Lickey E. B. Hughes K. W. Peterson R. H.. 2003. Variability and phylogenetic incongruence of an SSU nrDNA group I intron in Artomyces, Auriscalpium, and Lentinellus (Auriscalpiaceae: Homobasidiomycetes). Molecular Biology and Evolution 20: 1909-1916..
Lutz M. Bauer R. Begerow D. Oberwinkler F.. 2004. Tuberculina–Thanatophytum/Rhizoctonia crocorum–Helicobasidium: a unique mycoparasitic–phytoparasitic life strategy. Mycological Research 108: 227-238..[CrossRef][Web of Science][Medline]
Maddison D. R. Maddison W. P.. 2002. MacClade 4 analysis of phylogeny and character evolution. Sinauer, Sunderland, Massachusetts, USA..
Maier W. Begerow D. Weiss M. Oberwinkler F.. 2003. Phylogeny of the rust fungi: an approach using nuclear large subunit ribosomal DNA sequences. Canadian Journal of Botany 81: 12-23..[CrossRef]
Oberwinkler F.. 1989. Coccidiodictyon, new genus and Ordonia, two genera in the Septobasidiales. Opera Botanica 100: 185-192..
Perotto S. Nepote-Fus P. Saletta L. Bandi C. Young J. P. W.. 2000. A diverse population of introns in the nuclear ribosomal genes of ericoid mycorrhizal fungi includes elements with sequence similarity to endonuclease-coding genes. Molecular Biology and Evolution 17: 44-59..
Posada D. Crandall K. A.. 1998. MODELTEST: testing the model of DNA substitution. Bioinformatics 14: 817-818..
Reid D. A. Manimohan P.. 1985. Auriculoscypha, a new genus of Auriculariales (Basidiomycetes) from India. Transactions of the British Mycological Society 85: 532-535..[Web of Science]
Ronquist F. Huelsenbeck J. P.. 2003. MRBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 1572-1574..
Rossman A. Y.. 2000. Towards monophyletic genera in the holomorphic Hypocreales. Studies in Mycology 45: 27-34..
Swann E. C. Frieders E. M. McLaughlin D. J.. 2001. Pucciniomycotina. In D. J. McLaughlin, E. G. McLaughlin, P. A. Lemke, editors Systematics and evolution: the ycota VII, 37-56. Springer-Verlag, Berlin, Germany..
Swann E. C. Taylor J. W.. 1995. Toward a phylogenetic systematics of the basidiomycota: integrating yeasts and filamentous basidiomycetes using 18S rRNA gene sequences. Studies in Mycology 38: 147-161..
Swofford D. L.. 2003. PAUP*: phylogenetic analysis using parsimony (*and other methods), version 4.10b. Sinauer, Sunderland, Massachusetts, USA..
Szabo L. J.. 2006. Deciphering species complexes: Puccinia andropogonis and Puccinia coronata, examples of differing modes of speciation. Mycoscience 47: 130-136..[CrossRef]
Tian C. Shang Y. Zhuang J. Wang Q. Kakishima M.. 2004. Morphological and molecular phylogenetic analysis of Melampsora species on poplars in China. Mycoscience 45: 55-66..
Vilgalys R. V. Hester M.. 1990. Rapid genetic identification and mapping of enzymatically amplified ribosomal DNA from several Cryptococcus species. Journal of Bacteriology 172: 4238-4246..
White T. J. Bruns T. Lee S. Taylor J. W.. 1990. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. In M. A. Innis, D. H. Gelfand, J. J. Sninsky, T. J. White, editors PCR protocols: a guide to methods and applications, 315-322. Academic Press, New York, New York, USA..
Wiens J. J.. 2003. Missing data, incomplete taxa, and phylogenetic accuracy. Systematic Biology 52: 528-538..
Wiens J. J.. 2005. Can incomplete taxa rescue phylogenetic analyses from long-branch attraction?. Systematic Biology 54: 731-742..
Wilkinson M.. 1995. Coping with abundant missing entries in phylogenetic inference using parsimony. Systematic Biology 44: 501-514..
Yang Z. Rannala B.. 2005. Branch-length prior influences Bayesian posterior probability of phylogeny. Systematic Biology 54: 455-470..
Zolan M. E. Pukkila P. J.. 1986. Inheritance of DNA methylation in Coprinus cinereus. Molecular and Cellular Biology 6: 195-200..
CiteULike Complore Connotea Del.icio.us Digg Facebook Reddit Technorati Twitter What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
