Amoebae that make fruiting bodies consisting of a stalk and spores and classified as closely related to the myxogastrids have classically been placed in the taxon Eumycetozoa. Traditionally, there are three groups comprising Eumycetozoa: myxogastrids, dictyostelids, and the so-called protostelids. Dictyostelids and myxogastrids both make multicellular fruiting bodies that may contain hundreds of spores. Protostelids are those amoebae that make simple fruiting bodies consisting of a stalk and one or a few spores. Protostelid-like organisms have been suggested as the progenitors of the myxogastrids and dictyostelids, and they have been used to formulate hypotheses on the evolution of fruiting within the group. Molecular phylogenies have been published for both myxogastrids and dictyostelids, but little molecular phylogenetic work has been done on the protostelids. Here we provide phylogenetic trees based on the small subunit ribosomal RNA gene (SSU) that include 21 protostelids along with publicly available sequences from a wide variety of amoebae and other eukaryotes. SSU trees recover seven well supported clades that contain protostelids but do not appear to be specifically related to one another and are often interspersed among established groups of amoebae that have never been reported to fruit. In fact, we show that at least two taxa unambiguously belong to amoebozoan lineages where fruiting has never been reported. These analyses indicate that we can reject a monophyletic Eumycetozoa, s.l. For this reason, we will hereafter refer to those slime molds with simple fruiting as protosteloid amoebae and/or protosteloid slime molds, not as protostelids. These results add to our understanding of amoebozoan biodiversity, and demonstrate that the paradigms for understanding both nonfruiting and sporulating amoebae must be integrated. Finally, we suggest strategies for future research on protosteloid amoebae and nonfruiting amoebae, and discuss the impact of this work for taxonomists and phylogenomicists.
Citation: Shadwick LL, Spiegel FW, Shadwick JDL, Brown MW, Silberman JD (2009) Eumycetozoa = Amoebozoa?: SSUrDNA Phylogeny of Protosteloid Slime Molds and Its Significance for the Amoebozoan Supergroup. PLoS ONE 4(8): e6754. doi:10.1371/journal.pone.0006754
Editor: Simon Joly, McGill University, Canada
Received: May 8, 2009; Accepted: July 17, 2009; Published: August 25, 2009
Copyright: © 2009 Shadwick et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work has been supported in part by the National Science Foundation Grants DEB 039102 and DEB 0316284 http://www.nsf.gov/. LLS was partially supported by a distinguished doctoral fellowship from the cell and molecular biology program (CEMB) provided by the Walton Family Charity Support Foundation's endowment to the University of Arkansas. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
A microscopic drop of water resting upon the tip of a fine hair, this is the search image for organisms historically called protostelids. When researchers see this, they know they might be looking at a protostelid fruiting body. The so-called protostelids are amoebae that make simple fruiting bodies consisting of a delicate stalk that supports one or a few spores (Figure 1) –. Other fruiting amoebae, the dictyostelids and myxogastrids (also referred to as myxomycetes), make relatively complex fruiting bodies with many cells: the dictyostelids by aggregative fruiting and the myxogastrids by division of large, multinucleate cells into uninucleate spores , , . Olive , [see also 2],  thought that the simplicity of protostelid fruiting bodies suggested that the ancestors of dictyostelids and myxogastrid amoebae might have made protostelid-like fruiting bodies. Olive called this group the taxon Eumycetozoa and envisioned the monophyletic taxa Myxogastria, and Dictyostelia arising from a paraphyletic taxon Protostelia (Figure 251 of ) . Different interpretations of morphology by both Olive and Spiegel were used to call this idea into question , , , . However, early molecular phylogenies that included species from all three groups suggested that there might be a clade of eukaryotes that includes myxogastrids, dictyostelids, and protostelids –. Was this support for the taxon Eumycetozoa?
Brightfield light micrographs of standing protosteloid fruiting bodies. A) Protostelium mycophaga, B) Nematostelium ovatum, C) Ceratiomyxella tahitiensis, D) Soliformovum expulsum, E) Soliformovum irregularis, F) Cavostelium apophysatum, G) Schizoplasmodiopsis amoeboidea, H) Tychosporium acutostipes, I) Clastostelium recurvatum, J) Protosporangium articulatum, K) Protosteliopsis fimicola, L) isolate LHI05, M) Endostelium zonatum. Scale bar is 10 µm.
Since 1654 when the first record of a myxogastrid was purported to be a fungus, the insidious perception that fruiting body formation has phylogenetic relevance has perpetuated a divide between those biologists who study amoebae and those biologists who study amoebae that fruit (for reviews see , , ). Fruiting amoebae are those amoebae that make spores, usually supported by stalks, at some point during their life-cycle, and are typically studied by classically trained mycologists. These amoebae are identified, isolated, and described beginning with their fruiting bodies (for reviews see –, ). Amoebae that are not known to fruit are typically studied by classically trained protistologists. Such amoebae are identified, isolated, and described by their amoebal morphology and sometimes by their cysts (for review see , ). The reasons for this scientific divide are historical and methodological, not biological.
Until the last decade, when molecular phylogenies began to show otherwise, amoebae were thought of as a polyphyletic assemblage of eukaryotes. Baldauf et al.  were the first to show that some classical amoebae and some dictyostelids and myxogastrids grouped together. There has been a flush of recent molecular phylogenetic evidence showing that some fruiting and some nonfruiting amoebae belong to the supergroup Amoebozoa , –. Taxonomic sampling of protostelids has been a major limitation in all of these studies. In fact, those protostelids that have been included in phylogenies do not span the breadth of the purported morphological groups of protostelids , . In fact, the ribosomal small subunit RNA gene (SSU) sequences from only two very closely related species have been used as exemplars in all the above studies that include any protostelids . We think that the term protostelid has led to confusion in the literature because it implies an evolutionarily cohesive taxonomic unit , , while at the same time, ,  the term protostelid is used to describe a morphology . Therefore, to avoid this double meaning we will hereafter refer to these organisms in a descriptive sense as protosteloid amoebae, not as protostelids.
Protosteloid amoebae have simple fruiting structures (Figure 1), and a range of highly diverse amoeboid trophic cells (Figure 2 and –). Analysis by Spiegel of amoebal morphology as well as fruiting led to five proposed, morphologically identifiable groups of protosteloid amoebae that he thought were good candidates for being closely related to myxogastrids and dictyostelids , . These five groups include 28 of the 36 species described as protosteloid amoebae. Of the other eight species, one, Echinostelium bisporum, is clearly a myxogastrid , , and the rest are of doubtful affinity . Only by including protosteloid amoebae that span this known diversity in analyses with an appropriately broad set of outgroups will it be possible to determine whether there is a clade that corresponds to Olive's  hypothesis that there is robust phylogenetic support for the taxon Eumycetozoa. If the Eumycetozoa hypothesis is correct, then fruiting amoebae should form a monophyletic (or natural) group that includes some protosteloid amoebae, the myxogastrids, and the dictyostelids to the exclusion of nonfruiting amoebae.
Light micrographs of protosteloid amoebae. A) Protostelium mycophaga differential interference contrast microscopy (DIC), B) Nematostelium ovatum phase contrast microscopy (PC), C) Soliformovum expulsum DIC, D) Cavostelium apophysatum DIC, E) Schizoplasmodiopsis amoeboidea PC, F) Protosporangium articulatum PC, G) Protosteliopsis fimicola DIC, H) isolate LHI05 PC, I) Endostelium zonatum DIC. Scale bars are 10 µm except B which is 50 µm.
To gain insights into the relationships among protosteloid amoebae and where they fit among other amoebae, we have sequenced the SSU of 21 isolates representing 17 species of protosteloid amoebae including multiple representatives of each of the five “eumycetozoan” groups of Spiegel  and three other species, Endostelium zonatum, Protosteliopsis fimicola, and undescribed protosteloid isolate LHI05, whose morphologies suggest questionable affinity to the other purported eumycetozoans (Figures 1,2). These were included in phylogenetic analyses along with the SSU sequences from a broad range of amoebozoans (for recent reviews of Amoebozoa see , , ), and from a diverse assemblage of outgroup eukaryotes. Several cercozoans and stramenopiles were included, because Spiegel , ,  had suggested members of both groups as possible close relatives to protosteloid amoebae. The SSU gene was chosen because it is the most widely sequenced among amoebozoans and because it has been used to support the phylogenies of a number of clearly monophyletic lineages within Amoebozoa , , , , , , – including dictyostelids  and myxogastrids . We included multiple representatives from well supported amoebal lineages in our analysis to (a) look for congruence between our results and other amoebozoan phylogenies and (b) test whether any or all organisms described as protosteloid amoebae fell into a clade of amoebozoans that also included the myxogastrids and dictyostelids, i.e. Eumycetozoa sensu Olive . Further, we wanted to know if protosteloid amoebae were indeed a grade of Eumycetozoa sensu Olive . We show here that protosteloid amoebae are all members of the supergroup Amoebozoa and that there are several discrete lineages that include protosteloid species. There is no evidence for a group that corresponds to Eumycetozoa sensu Olive; rather, stalked fruiting is widespread among the supergroup.
The SSU rRNA genes of 21 isolates, representing 17 species of protosteloid amoebae were sequenced to assess their phylogenetic affinities (Table 1). The sequences ranged from 1,786 bp in Schizoplasmodiopsis pseudoendospora to 2,493 bp in Endostelium zonatum (Table 1). No group 1 introns were observed, and nearly all variation in length was contained within hypervariable regions of the SSU rRNA gene. Those seven isolates with especially short SSU genes<1,850 bp had some truncations in regions that are generally conserved across a diverse array of eukaryotes. Most SSU genes of protosteloid amoebae were AT rich with GC contents ranging from 38% in Schizoplasmodiopsis amoeboidea to 50% in isolate LHI05 (Table 1). Within-isolate sequence heterogeneity was detected in nine isolates, and was most extensive in E. zonatum, unnamed isolate LHI05, Protosporangium articulatum, and all isolates of Protosteliopsis fimicola (Table 1).
All of the organisms with protosteloid types of fruiting group within Amoebozoa in our maximum likelihood tree (Figure 3). There are several well supported clades that contain protosteloid species. Five of these clades with more than one species include protosteloid amoebae exclusively. We refer to these by the following informal designations: the protosporangiid clade, the protosteliid clade, the soliformoviid clade, the cavosteliid clade, and the schizoplasmodiid clade. Two species of protosteloid amoebae branch with high support within established, species rich amoebal lineages where no fruiting members have previously been reported. Protosteliopsis fimicola is a vannellid (Figures 1K, 2G, 3, 4A), and the undescribed isolate LHI05, is an acanthamoebid (Figures 3, 4B). So far, only one sequenced protosteloid species has no obvious close relatives, E. zonatum (Figures 1M, 2I, 3). We had originally included LHI05 in the analysis because its amoeba and mode of fruiting are reminiscent of E. zonatum (Figures 1I,M, 2H,I). However, these two taxa do not appear to be specifically related to each other. While all of the protosteloid species branched within a monophyletic Amoebozoa in our ML tree, the bootstrap support for a monophyletic Amoebozoa is lacking (Figure 3).
Colored branches indicate lineages in which protosteloid fruiting occurs. Black branches highlight amoebozoan lineages and gray branches show other eukaryotes used as outgroups. Red, black, and gray fonts indicate species of amoebae with protosteloid fruiting, nonfruiting amoebozoans, and other eukaryotes used as outgroups, respectively. To allow the figure to fit legibly on a single page, and to conserve the long branch length, the long branch leading to Lindbladia has been broken and shifted above and left. One hundred twenty nine taxa and 1,169 aligned positions were used to infer the optimal maximum likelihood (ML) tree in RAxML 7.0.4 using the following model (GTR + Γ, α = 0.513834, 20 discrete rate categories). ML bootstrap values from analyses of 1,000 RAxML datasets and 1,000 GARLI 0.96 datasets are shown above and below the node respectively. ML bootstrap values: black oval = 90–100, white oval with black outline = 80–90, gray oval = 70–80, unmarked<20. Black circle highlights the support values for monophyly of Amoebozoa. For the GARLI analyses, the following model was used (GTR + Γ + I, α = 0.71950104, 4 discrete rate categories). The scale bar represents evolutionary distance in changes per site.
The monophyly of myxogastrids and the monophyly of dictyostelids are maintained in our analyses. The myxogastrids form a clade that is divided into the dark spored and light spored lineages , and the dictyostelids show the four clades of Schaap et al.  (Figure 3). Our highest likelihood tree has protosteloid clades as sister to the myxogastrids and dictyostelids. The soliformoviid clade is a poorly supported sister to the myxogastrids and the cavosteliid clade appears as a sister to the dictyostelids, again with weak support (Figure 3).
There is no discrete clade of Amoebozoa that exclusively contains all the fruiting species we included in our taxon sample in our highest likelihood tree (Figure 3). That is, we recovered no monophyletic taxon Eumycetozoa sensu Olive . There is no clade that exclusively includes the myxogastrids, the dictyostelids, and some subset of the protosteloid species and no nonfruiting amoebae i.e. an exclusively fruiting clade that could be consistent with Olive's  Eumycetozoa hypothesis in a more limited sense.
There is an essentially unsupported clade that occurs in our highest likelihood tree that includes most of the protosteloid species thought to be Eumycetozoa by Spiegel . This unsupported clade includes the protosporangiids, protosteliids, soliformoviids, cavosteliids, schizoplasmodiids, myxogastrids and dictyostelids with a number of nonfruiting amoebozoans including archamoebids, Arachnula, both Filamoeba spp., Acramoeba, and the amoebozoan flagellates Multicilia and Phalansterium (Figure 3).
Three clades that contain both protosteloid amoebae and amoebozoans that have never been reported to fruit were examined in more detail.
Within the poorly supported group that contains the myxogastrids, dictyostelids, many protosteloid amoebae and several nonfruiting amoebozoans, there is an interesting sister group relationship recovered between the schizoplasmodiid clade and the amoebozoan flagellate Phalansterium solitarium. This group appears with low support in the large tree that is restricted to 1,169 alignable postitions (Figure 3). However, P. solitarium is alignable with all of the schizoplasmodiids across nearly their entire SSU rRNA genes, including hypervariable regions, such that 1,735 unambiguously aligned positions are amenable to phylogenetic analyses (Figure 4C). More detailed analyses of this region of the tree using P. solitarium as an outgroup to the schizoplasmodiids shows that Nematostelium ovatum and Schizoplasmodium cavostelioides are sister to each other with Ceratiomyxella tahitiensis branching basally to them (Figure 4C).
For all trees, the scale bars represent evolutionary distance in changes per site. Red font indicates protosteloid amoebae. ML bootstrap values from analyses of 1,000 datasets and Bayesian posterior probabilities are shown above and below the nodes respectively. A) Placement of Protosteliopsis fimicola among vannellids. ML tree of 9 SSU genes and 1,837 aligned positions inferred with a GTR + Γ + I (α = 0.5042, 4 discrete rate categories, and I = 0.3421) model of nucleotide substitution. For the Bayesian analyses two runs, each consisting of 4 MCMC chains, were run for 2,000,000 generations, sampling every 100th tree. The first 100 trees were discarded as burnin after assessing for convergence of parameters. Purple branches highlight Protosteliopsis fimicola clade. B) Placement of protosteloid Isolate LHI05 among acanthamoebids. ML tree of 18 ssu genes and 1,476 aligned positions inferred with a TrN + Γ, α = 0.2228, 4 discrete rate categories model of nucleotide substitution. For the Bayesian analyses two runs each consisting of 4 MCMC chains were run for 2,000,000 generations, sampling every 100th tree. The first 4,000 trees were discarded as burnin after assessing for convergence of parameters. C) Branching order of Schizoplasmodiids rooted with Phalansterium. Four taxa and 1,735 aligned positions were used to infer the optimal ML tree with a TrN + Γ, α = 0.3693, 4 discrete rate categories model. For the Bayesian analyses two runs each consisting of 4 MCMC chains were run for 5,000,000 generations, sampling every 100th tree. The first 5,000 trees were discarded as burnin after assessing for convergence of parameters.
Isolate LHI05 groups with high support within the acanthamoebids with specific and robust affinity to Protacanthamoeba bohemica (Figure 4B).
Protosteliopsis is clearly a vannellid with Vannella placida as its sister species (Figures 3, 4A). The sister group relationship between Protosteliopsis fimicola and Vannella placida is upheld when multiple isolates of P. fimicola and additional vannellids are included in a fine-scale analysis (Figure 4A).
Some previous hypotheses about the relationships among the organisms traditionally considered to be eumycetozoans  and their relationships with other eukaryotes were not compatible with branching patterns recovered in our maximum likelihood tree. To test these hypotheses we built topologically constrained trees and compared their likelihoods to our maximum likelihood trees and bootstrap trees using the Approximately Unbiased (AU) test. Table 2 lists some important hypotheses of relationships among the purported eumycetozoans and other organisms that have been listed in the literature. Rejection of constrained trees was established at an AU test p-value of 0.05.
Brief descriptions of some of the more interesting AU test results follow. A group that exclusively contains all the protosteloid amoebae is rejected (Table 2). Eumycetozoa in the strictest sense, i.e., a group that includes all the protosteloid species, myxogastrids, and dictyostelids to the exclusion of other groups is also rejected (Table 2). However, Eumycetozoa cannot be rejected if it is defined to include only protosteliids, schizoplasmodiids, soliformoviids, cavosteliids, protosporangiids, dictyostelids, and myxogastrids, where, E. zonatum, P. fimicola, isolate LHI05, and all nonfruiting amoebozoans are excluded from the constraint (Table 2).
Within the well supported, species rich clades that contain protosteloid amoebae, i.e., the protosteliid clade, the cavosteliid clade, the vannellids, and the acanthamoebids, some hypotheses can be rejected and others cannot. In the protosteliid clade, while Planoprotostelium aurantium is nestled within the protosteliid clade in all of our highest likelihood trees, it cannot be rejected as sister to Protostelium (Table 2). Likewise in the cavosteliid clade, Cavostelium apophysatum cannot be rejected as the sister to all other cavosteliids (Table 2), nor can the protosporangiids be rejected as sister to Myxogastria (Table 2). Trees in which LHI05 was excluded from the acanthamoebids were soundly rejected (Table 2). Protosteliopsis fimicola branched as the sister taxon to the remaining vannellids when it was excluded from that group, and this was a relationship that could not be rejected (Table 2).
Our findings show that the organisms formerly called protostelids are scattered among Amoebozoa. Our trees clearly show: 1) as expected, that protosteloid amoebae are not monophyletic , , , and 2) contrary to predictions, they are not a grade within Eumycetozoa, sensu Olive , . Therefore, our results justify our decision to reject the term protostelids in favor of the strictly descriptive term protosteloid amoebae. While we recognize that the SSU gene presents a problem in resolving deep structure , , , it is ideal for delimiting well supported groups of clearly related organisms.
When we look past the obvious trait of fruiting, we do find that there are five such groups containing only protosteloid amoebae and that their morphological identity is clear when all other detailed stages of the life-cycle are considered. These correspond to Spiegel's groups I, II, III, IV, and Va (see Figure 3 and Table 2 in ). In brief these groups are described below. Monographic treatments with formal taxonomic revisions are being prepared separately.
The Protosteliid Clade - Group I (100% Bootstrap Support)
This group includes the first described protosteloid amoeba, Protostelium mycophaga . Its taxa have amoebae with orange pigment and acutely pointed subpseudopodia (Figure 1A, 2A). There are three points of interest within this group: 1) the branch lengths within the species Protostelium mycophaga are relatively long, 2) the species which forcibly discharge their spores, Protostelium nocturnum  and Protostelium okumukumu , branch basally, and 3) one member of this clade Planoprotostelium aurantium makes an amoeboflagellate cell . It was supposed that Planoprotostelium was sister to Protostelium because of its ability to make flagella and that this ability was lost once, ancestrally to other members of the clade , , , –. Our optimal trees do not support that hypothesis because the genus Planoprotostelium is embedded within the protosteliid clade (Figure 3). However, when Planoprotostelium was constrained outside of Protostelium it branched as sister to Protostelium, and this relationship was not rejected by the AU test (Table 2). While the AU test does not let us reject P. aurantium as sister to the rest of the protosteliid clade, the similarity of its fruiting body  to that of Protostelium mycophaga  compared to Protostelium okumukumu  and Protostelium nocturnum  is quite clear, and we predict that further analysis of a broader taxon sampling of the protosteliid clade will further support its position within the group rather than as a basal lineage.
The Schizoplasmodiid Clade - Group II (100% Bootstrap Support)
The first schizoplasmodiids were described together under the genus name Schizoplasmodium  based on the plasmodial trophic state that gives rise to the fruiting bodies, and their shared characteristic of a stalk-spore junction with an annular hilum on the spore that articulates with a knob-like apophysis on the stalk (Figure 1B, C, 2B) , –. The plasmodial amoeba has both filose and anastomosing subpseudopodia (Figure 2B), and similar “bead on a string” plasmodial mitosis , , –, , . Subsequently, schizoplasmodiids were divided into three genera based on variations in fruiting-body stalk length, presence or absence of forcible spore discharge, and presence of an amoeboflagellate in the life-cycle of Ceratiomyxella tahitiensis , , . While our 129 taxa tree groups all three species, C. tahitiensis, Nematostelium ovatum, and Schizoplasmodium cavostelioides together with 100% bootstrap support (Figure 3), a more inclusive mask was required in order to recover the branching order among the three species (Figure 4C). The fine-scale analysis resolved that the two non-flagellates, N. ovatum and S. cavostelioides, are sister to one another with 96% bootstrap support and a Bayesian posterior probability of 0.95 (Figure 4C), which suggests that the flagellate state may have been lost once in this group as previously supposed . However more taxon sampling of the non-flagellate schizoplasmodiid species, including Schizoplasmodium obovatum, Schizoplasmodium seychellarum, and Nematostelium gracile, will be necessary to resolve this group completely. Given the large number of morphological synapomorphies and the nearly identical SSU rRNA gene sequences, maintaining three genera within the schizoplasmodiid clade may not be well justified. For instance, the separate genus names may well have served to confuse researchers with little firsthand knowledge of protosteloid amoebae, leading them to misclassify members of this clade .
The Soliformoviid Clade - Group III (100% Bootstrap Support)
The genus Soliformovum includes two species with identical fan-shaped amoebae with acutely pointed subpseudopodia and indistinct, diffuse nucleoli (Figure 2C) , , . Both species make a characteristic prespore cell that resembles a “sunny-side-up” fried egg, the character for which the genus is named , .
The Cavosteliid Clade - Group IV (65/69% Bootstrap Support)
This is by far the most morphologically diverse clade of protosteloid amoebae. The cavosteliids all have relatively thin amoebae (Figure 2D,E), with filose subpseudopodia, although flagellates and plasmodia also occur as additional stages within some species of the group (see Figure 3 of ). Most have round, centrally located nucleoli, except Schizoplasmodiopsis amoeboidea, which has indistinct, diffuse nucleoli similar to those seen in Soliformovum , , . They all have sculpturing on their spore walls, and the spores are not deciduous –. Many of these species are common. We were surprised that Tychosporium acutostipes branched within this group. Spiegel et al.  had placed Tychosporium as a basal Protostelium noting similar prespore cells and some aspects of amoebal morphology, but Tychosporium lacked the orange pigment . The cavosteliids are a highly diverse and fascinating group that requires more work. For instance, this clade has the lowest bootstrap support of any of the morphological groups, but support for the group and for nearly every node within the group jumps to nearly 100% if Cavostelium apophysatum is removed from the analysis (data not shown). Removal of Cavostelium apophysatum from the group was not rejected by the AU test (Table 1). However, we still tentatively accept this clade because of its morphological identity .
The Protosporangiid Clade - Group Va in part (100% Bootstrap Support)
The two species of Group Va that we included, Protosporangium articulatum and Clastostelium recurvatum, have essentially identical life cycles, essentially identical amoeboflagellates (Figure 2F) and non-flagellated amoebae, and fruiting bodies with 2–4 spores (Figure 1I, J) (see , , –). Spiegel's  group Va also includes Ceratiomyxa, for which we do not have sequence data; it has a similar life-cycle and amoeboflagellates , . Group Va was thought to be sister to the myxogastrids (Spiegel's group Vb) , ,  on the basis of amoeboflagellate ultrastructure. Although the AU test does not allow us to reject this relationship to myxogastrids, we prefer to be skeptical about this hypothesis until further work either supports or fully rejects it. For the same reason, we also remain skeptical about the sister group relationship with Protostelium that we recovered with low support in our highest likelihood tree (Figure 3).
Groups VI & VII
Spiegel suggested that Endostelium zonatum and Protosteliopsis fimicola might be members of amoeboid groups unrelated to other protosteloid amoebae . Our results support this hypothesis. The placement of Endostelium zonatum in the SSU tree is equivocal since it has no strong affinities towards any particular amoebozoan taxon (Figure 3, Table S2) –. Based on similar amoebal morphology and fruiting body development, we thought that our new protosteloid amoeba, isolate LHI05, and E. zonatum might be specifically related. However, E. zonatum and isolate LHI05 show no close relationship in our highest likelihood trees, though a possible sister group relationship of LHI05 and E. zonatum, embedded within the acanthamoebids was not rejected by the AU test (Table 1). If E. zonatum were closely related to LHI05, then it would be an acanthamoebid according to the maximum likelihood constrained tree (data not shown). LHI05 is clearly closely related to Protacanthamoeba bohemica among the acanthamoebids (Figures 3, 4B) . In fact constraining LHI05 away from Protoacanthamoeba bohemica results in a maximum likelihood tree that is strongly rejected by the AU test (Table 1). Two of us, L.L. Shadwick and F.W. Spiegel are currently in the process of describing this new species. Protosteliopsis fimicola was originally described as a Protostelium, then moved to the monotypic genus Protosteliopsis because it was so different (lacking both orange pigment and filose subpseudopodia) from the other species of Protostelium , . Protosteliopsis fimicola robustly groups within the vannellids (Figure 3). In fact, P. fimicola displaces Vannella epipetala as the sister to Vannella placida (Figure 4A) . In addition, P. fimicola shares similar patterns of SSU sequence microheterogeneity with both V. placida and V. epipetala , . These molecular phylogenetic findings are consistent with published light and electron-microscopy images , –. For instance, Protosteliopsis fimicola has the typically conspicuous contractile vacuole, the anterior hyaline veil, the floating form, and the complete lack of uropodia – all typical characters of vannellids , , . It is clear from Olive's drawings that he was sometimes observing an amoebal state intermediate between the locomotive form and the floating form, which partially explains his constant assertion that Protosteliopsis fimicola makes filose pseudopodia, which both P. fimicola and vannellids lack , , . In light of the description of P. fimicola and the recent clarification of the genus Vannella, it is astounding that no one recognized that P. fimicola was a vannellid prior to this study. The misclassification of P. fimicola is a testament to both the bias that fruiting induces in the minds of researchers, and the lack of clear, published morphological guidelines for classification of vannellids at the time P. fimicola was described , –.
Protosteloid amoebae are at the crux of the Eumycetozoa hypothesis sensu Olive . Stalked fruiting body formation was thought to be a synapomorphy of Eumycetozoa , , . Purported eumycetozoans, such as sessile myxogastrids, whose fruiting bodies lacked a stalk were thought to be derived from a stalked ancestor . Other morphological characters were used to support the monophyly of Eumycetozoa, e.g., morphology of stalk-producing cells, amoebal morphology and ultrastructure of amoeboflagellates , , , , , , . These characters were also used to delineate the major groups of protosteloid amoebae as discussed above . However, these characters were not considered in the absence of fruiting for comparison to other morphologically similar nonfruiting amoebae; in fact no amoeboid outgroups were considered at all , , , . These additional characters were used, instead, to support fruiting as a character. Just as morphological characters were used to support stalked fruiting as a character, so were early molecular markers –. Here, we have shown with the most extensive taxon sampling that we are aware of for both fruiting and nonfruiting amoebae that: 1) protosteloid ameobae are not monophyletic, and 2) amoebozoans with stalked fruiting are not monophyletic. We have recovered the same well supported clades that others have recovered for myxogastrids , dictyostelids , and most well to moderately supported clades of nonfruiting amoebae found consistently in the literature , , , , , , , , , . However, support for the deeper relationships among these groups is lacking as in all SSU trees of Amoebozoa , , , –, –, , , , , .
Rigor demands that we reject the insidious lure of the fruiting body. Our results show, that stalked fruiting, s.l., by amoebozoans, though taken as significant, has no a priori phylogenetic significance , , . Rather, it is our view that fruiting has to be taken in context with all the characters of an organism's morphological traits and life history before its significance can be understood.
If we wished to argue for a clade to call Eumycetozoa that includes myxogastrids, dictyostelids, and some or all of the protosteliid, schizoplasmodiid, soliformoviid, cavosteliid, and protosporangiid clades, then there are several nodes we could select on the tree as basal to such a group (Figure 3). We think this is unwise for two reasons. First, while it is interesting that our trees show that there may be protosteloid sister groups to both dictyostelids and myxogastrids, the affinities are very poorly supported. Second, it must be recognized that almost all of the groups in the part of the tree that includes most of the fruiting organisms have extremely long branches; thus, some of the deep structure in our tree could simply be a result of long branch attraction . We have attempted to alleviate some long branch effects by using a conservative inclusion set of unambiguously alignable sequence and by including multiple representatives of each lineage where possible, but we cannot confidently rule out long branch attraction. Therefore, we think formal taxonomic revision should be restricted only to well supported clades. Further work, such as comparative genomics, will be necessary to resolve the uncertain deeper relationships. We are inclined to be very conservative when using our results to revise the higher level taxonomy of these organisms. In fact, we are strongly disinclined to even propose informal names for poorly supported groups that happen to occur in our highest likelihood tree unless subsequent research provides more support for them. Taxonomic revisions based on poorly resolved phylogenetic nodes only clutter the literature with names that can lead to confusion. Therefore we think it is best, for now, to relinquish the concept of Eumycetozoa.
We would like to make our point as strongly as possible. Our results show that stalked fruiting is widespread among the Amoebozoa. Thus, if we were presumptuous enough to accept 1) that our tree of this amoeboid supergroup (Figure 3) is true, 2) that stalked fruiting in the supergroup has only one origin, and 3) that fruiting is important as a defining character, then the name Eumycetozoa Zopf 1885 would be correct for the whole supergroup since the name has taxonomic priority over the name Amoebozoa Lühe, 1913 , –. However, at least two issues need to be resolved before we would consider formally renaming Amoebozoa. First, the evolutionary and developmental origin of fruiting must be understood. Second, the higher order relationships among fruiting and nonfruiting lineages must be resolved.
Materials and Methods
At least one protosteloid amoeba from each major morphological group of Spiegel  was sampled. Type cultures were used where available, as were multiple isolates. When necessary we isolated organisms from nature into monoeukaryotic or dieukaryotic culture as previously described , ,  and other cultures were acquired from culture collections (Table 3).
Protosteloid amoebae were grown on weak malt yeast extract agar plates (wMY) (0.002 g malt extract, 0.002 g yeast extract, 0.75 g K2HP04, 15.0 g Difco Bacto Agar, 1.0L deionized [DI] H20) with appropriate food organisms (Table 3) in the laboratory at ambient temperatures (approx 21–25°C) , . Protosporangium articulatum could not be established in culture.
All cultures were vouchered by rigorous microscopical examinination on a Zeiss Axioskop 2 Plus under the 10×dry differential interference contrast (DIC) and bright field (BF), 40×dry (DIC/Phase contrast (PC)), and 63×oil (DIC) to verify the proper taxon identification of each organism, and to check for possible contaminants. All cultured organisms were observed to form fruiting bodies in culture. All cultures were digitally photographed using AutoMontage (Syncroscopy, Frederick, MD). All cultures, even those obtained from ATCC and CCAP except Schizoplasmodiopsis pseudoendospora (which later succumbed to a bacterial contamination) were put into a viable frozen stasis in liquid nitrogen and are stored at the University of Arkansas. Many were also depostited at the ATCC (Table 3).
DNA from cultures was made available for PCR by using a chelex method that requires little starting material and provides few chances for cross contamination of reagents . Chelex solution was 6% (w/v) chelex100 resin (Bio-Rad Laboratories, Hercules, CA) in double distilled, diethylpyrocarbonate treated H20. Under a dissecting microscope (Nikon SMZ 1500), organisms were scraped off their media using an ethanol-flamed spear-point needle and placed into 150 µl of chelex solution. Negative controls include 1) chelex solution only 2) chelex solution + food organism. All were placed into a thermal cycler for 4 hours at 56°C followed by 30 min. at 98°C, then stored at −20°C until needed. Protosporangium articulatum was treated differently because we were not able to grow it in culture. Fruiting bodies on their natural substrate were vouchered through photomicroscopy (Figure 1J), then approximately 30 spores were collected with an insect pin (see ). DNA was extracted from the collected spores using the MasterPure Complete DNA and RNA Purification Kit (Epicentre, Madison, WI) according to manufacturer's protocols.
Two additional sequences were included to increase phylogenetic signal from the Myxogastria. Lindbladia tubulina fruiting bodies were a kind gift from Sergey Karpov, Herzen State Pedagogical University, St. Petersburg, Russia. Trichia decipiens fruiting bodies were collected from their natural substrate in Halifax Nova Scotia, Canada during the summer of 2001. DNA from each myxogastrid was isolated from spores or maturing plasmodium using PureGene tissue lysis kit, per manufacturer's recommendations.
PCR and DNA sequencing
The SSU from protosteloid amoebae were amplified with either “universal” eukaryote SSU primer pairs Medlin A, Medlin B , 30F, 1492R, 5′SSU17!, or specifically designed biased primers PmycF1, PmycR2, Myxo3′ (Table 4). The SSU of Protosporangium articulatum was obtained by a nested PCR protocol using Medlin A : B in the first PCR reaction that served as template for a second PCR using 30F : 1492R. For the two myxogastrids, L. tubulina and T. decipiens, SSU rDNA was amplified with the “universal” eukaryotic primers Medlin A : B , cloned and sequenced as previously described .
For the SSU from protosteloid species, a stepdown thermal cycling program was used with Platinum Blue RTS PCR Super Mix (Invitrogen, Carlsbad, CA) in 20 µl reactions with the following cycling parameters: preheat lid 105°C, initial denaturation 94°C 1 min., followed by 5 cycles of 94°C 30 sec., 1 min. for primer annealing at 60°C followed by 3 min. 5 sec. elongation at 72°C. Then 9 cycles were done with denaturation 94°C for 30 sec., with an initial annealing temperature of 59°C that decreased 1°C/cycle to a final 50°C annealing, elongation for 3 min. 5 sec. at 72°C, followed by 20 cycles as above with a 50°C annealing temperature and an elongation time of 2 min. 5 sec. PCR products were either sequenced directly after removal of unincorporated nucleotides and primers using QIAquick Gel-Extraction kit (Qiagen, Valencia, CA), or they were T/A cloned into TOPO vector pCR4 and transformed into TOP 10 Escherichia coli cells per manufacturer's instruction (Invitrogen). For cultures grown on yeast (where no size difference between the SSU of the yeast and amoeba was seen), amplified and cloned SSU inserts were PCR amplified directly from transformed bacterial colonies and screened by TaqI restriction fragment length polymorphism with TaqαI restriction endonuclease (New England Biolabs, Ipswich, MA) to distinguish between yeast SSU clones and protosteloid SSU clones.
DNA sequencing reactions were performed using big-dye chemistry and resolved on an Applied Biosystems 3100 Genetic Analyzer (Foster City, CA). PCR products were sequenced directly, where possible in both orientations. Otherwise, one or two clones and/or a pool of 6–10 clones were sequenced fully in both orientations. Partial sequence was often obtained from additional clones.
For protosteloid amoebae that grew with yeast as food organisms, we partially sequenced one fungal clone (identified by restriction fragment length polymorphism), and fully sequenced all available (1 to 10) protosteloid clones. In several cases, within-isolate SSU sequence microheterogeneity was observed. In these cases pooled clones and/or PCR products were sequenced plus one or two individual clones through regions of heterogeneity (Protosteliopsis fimicola, Endostelium zonatum, isolate LHI05). In cases of within-isolate microheterogeneity that inhibited sequencing through certain regions of pooled clones/PCR products, all individual cloned SSUs were sequenced and included in preliminary trees. In all cases those partial sequences clustered tightly together with other sequences from the same isolate. An individual clone from each of these isolates was used for subsequent analyses. All new sequences were accessioned in GenBank (Table S1).
Protosteloid amoeba SSU sequences were hand aligned into an existing SSU rDNA multiple sequence alignment in MacClade (Sinauer Associates, Sunderland, MA) . Ambiguous regions in the alignment were excluded from phylogenetic analyses. The largest data set consisted of 129 taxa that span the known diversity of Amoebozoa plus a wide array of outgroup taxa that included at least two members from most other major eukaryotic lineages.
Multiple representatives from each sequenced amoebozoan lineage were included to deeply sample the molecular diversity available within Amoebozoa and to assess congruence with other amoebozoan phylogenies. Phylogenies were inferred using maximum likelihood (ML) as implemented in PAUP* 4.0 , RAxML 7.0.4 , and GARLI 0.96  using a GTR + Γ + I model of nucleotide substitution (except RAxML which implemented a GTR CAT model for topology search and a GTR + Γ model for tree optimization, both with 20 rate categories one of which essentially corresponds to invariant sites). Specific model parameters were determined using ModelTest 3.7  selected using the Akaike information criterion (AIC)  for analyses run in PAUP*, while RAxML and GARLI were allowed to estimate models during their respective analyses. A single optimum ML tree was inferred in PAUP* while the highest likelihood tree was identified from 300 RAxML and 300 GARLI runs each starting from a different parsimony tree. The optimum ML tree was inferred in RaxML with branch lengths optimized in PAUP* is shown in Figure 3. Topological support for branches was assessed from the consensus of 1,000 ML bootstrap trees inferred in RAxML and GARLI. For the 129 taxa data set Bayesian parameters failed to converge even after 30 million generations.
For all other finer-scale phylogenetic analyses, both ML and Bayesian analyses were performed. ML analyses were performed in PAUP* with the nucleotide substitution model and specific parameters selected for each dataset using ModelTest as implemented in PAUP* . For Bayesian analyses we used Mr.Bayes 3.1.2  with 4 Markov chain Monte Carlo (MCMC) chains in each of two independent runs with nst = 6 and rates set to invgamma (corresponding to a GTR + Γ +I model of nucleotide substitution). Trees generated prior to convergence of parameters were discarded as ‘burn-in’. Convergence was detected only after the standard deviation of split frequencies dropped below 0.01 and the sump function provided in Mr.Bayes and the program Tracer (part of the BEAST package)  all indicated convergence.
Phylogenetic tree topologies conforming to a variety of specific hypotheses were tested in a likelihood framework. Hypothesis testing was done on the 129 taxa tree (Figure 3) using the Approximately Unbiased (AU) test in the program Consel 0.1i , . Consel compares likelihoods, but the likelihoods calculated by the programs RAxML and GARLI, which were used to generate topologies and estimate models, are not directly comparable , . Also, each topology generated by RAxML and GARLI has unique model parameters estimated with it. A single set of model parameters was needed for input into PAUP* so that comparable likelihood values could be calculated for every topology no matter how it was generated. RAxML and GARLI each produce a highest likelihood topology (as calculated within the program). We considered the model parameters associated with that highest likelihood topology to be the optimal model parameters produced by that program. Thus PAUP* was used to calculate likelihoods using the optimal model parameters estimated by RAxML for all of the tree topologies generated in RAxML and GARLI. Then PAUP* was used to calculate the same likelihoods using the optimal model parameters estimated by GARLI. The likelihoods of tree topologies produced by the two methods produced distributions that overlapped almost entirely. Paup* was used to calculate likelihoods for all topologies using the model estimated by RAxML for all subsequent analyses. Six hundred ten constraint topologies were created by manually constraining specific taxa to branch together followed by reoptimization of the branching among the remaining taxa. Reoptimization was performed by inferring three ML trees in RAxML with a GTR + Γ model (as specified in RAxML) and keeping the highest likelihood of these trees. Thus, 610 constraint tree topologies (hypotheses) were generated. For a list of constraints see table S2. Site likelihoods, required for import into Consel, were calculated from the optimal ML topologies, the 610 specific constraint topologies and a set of plausible topologies consisting of the 1,000 RAxML and 1,000 GARLI bootstrap topologies in PAUP* using the RAxML substitution model. Significant differences in the likelihood among all trees were tested by the AU, Shimodaira-Hasegawa (SH), and Kishino-Hasegawa (KH) tests as implemented in Consel 0.1i , .
Genbank Accession Numbers for Additional Sequences Used in Phylogenetic Analyses. Bold font highlights protosteloid species. * = organisms sequenced in this study.
(0.05 MB XLS)
Constrained Taxa Abbreviations and Constraints Used for Statistical Tests. Each well supported amoebozoan lineage is constrained with all other amoebozoan and outgroup lineages. The first two columns refer to the taxon and its abbreviation in the constraint. The taxa constrained to branch together are within the parentheses. For more complex constraints, where some taxa are constrained away from a group the notation used is ((taxa constrained together) taxa constrained away);. All other taxa (from 129 taxa dataset see figure 3) that are not constrained in or out of a group are ommitted from this table for ease of reading. Constrained trees (not shown) were built in RAxML and likelihood scores (-lnL) were estimated in PAUP* using the constraints shown. Significant differences in the likelihood among all trees were tested by the AU, SH, KH tests as implemented in Consel .
(0.22 MB XLS)
Tychosporium acutostipes sequences were generated by Ronan Bakker as part of his undergraduate honor's thesis. Thanks to Bobbie Okimoto and Sha Jin at the Unversity of Arkansas DNA sequencing lab for sequencing and troubleshooting. Thanks to George Ndiritu, Maria Aguilar, Carlos Rojas, Sally Edwards, Katherine Winsett, and Satyendra Rajguru for helpful comments and discussions. Thanks to Ed Haskins for providing culture advice and Cryptococcus laurentii cultures.
Conceived and designed the experiments: LLS FWS JDS. Performed the experiments: LLS JDLS MWB. Analyzed the data: LLS JDLS JDS. Contributed reagents/materials/analysis tools: FWS JDS. Wrote the paper: LLS FWS JDLS JDS.
- 1. Olive LS (1975) The Mycetozoans. New York: Academic Press. 293 p.
- 2. Spiegel FW (1990) Phylum plasmodial slime molds: Class Protostelids. In: Margulis L, Corliss JO, Melkonian MJ, Chapman D, editors. Handbook of Protoctista. Boston: Jones and Bartlett Publishers, Inc.. pp. 484–497.
- 3. Spiegel FW, Haskins EF, Cavender JC, Landolt JC, Lindley-Settlemyre LA, et al. (2007) A beginner's guide to isolating and culturing eumycetozoans. Available: http://slimemold.uark.edu/educationframe.htm. Accessed 25 January 2008.
- 4. Raper KB (1984) The Dictyostelids. New Jersey: Princeton Univeristy Press.. 453 p.
- 5. Martin GW, Alexopoulos CJ (1969) The Myxomycetes. Iowa: University of Iowa Press.. 561 p.
- 6. Spiegel FW (1991) A proposed phylogeny of the flagellated protostelids. BioSystems 25: 113–120.
- 7. Adl SM, Simpson AGB, Farmer MA, Andersen RA, Anderson OR, et al. (2005) The new higher level classification of Eukaryotes with emphasis on the taxonomy of protists. J Euk Microbiol 52: 399–451.
- 8. Spiegel FW, Lee SB, Rusk SA (1995) Eumycetozoans and molecular systematics. Can J Bot 73: 738–746.
- 9. Baldauf SL, Doolittle WF (1997) Origin and evolution of the slime molds (Mycetozoa). Proc Natl Acad Sci U S A 94: 12007–12012.
- 10. Baldauf SL, Roger AJ, Wenk-Siefert I, Doolittle WF (2000) A kingdom-level phylogeny of eukaryotes based on combined protein data. Science 290: 972–977.
- 11. Lister A (1925) A monograph of the Mycetozoa. London: British Museum of Natural History.. 296 p.
- 12. Swanson AR, Spiegel FW, Cavender JC (2002) Taxonomy, slime molds, and the questions we ask. Mycologia 94: 968–979.
- 13. Page FC (1988) A new key to freshwater and soil gymnamoebae. Ambleside: Freshwater Biological Association.. 122 p.
- 14. Smirnov AV, Brown S (2004) Guide to the methods of study and identification of soil gymnamoebae. Protistology 3: 148–190.
- 15. Bapteste E, Brinkmann H, Lee JA, Moore DV, Sensen CW, et al. (2002) The analysis of 100 genes support the grouping of three highly divergent amoebae: Dictyostelium, Entamoeba, and Mastigamoeba. Proc Natl Acad Sci U S A 99: 1414–1419.
- 16. Fahrni JF, Bolivar I, Berney C, Nassonova E, Smirnov A, et al. (2003) Phylogeny of lobose amoebae based on actin and small-subunit ribosomal RNA genes. Mol Biol Evol 20: 1881–1886.
- 17. Nikolaev SI, Mitchell EAD, Petrov NB, Berney C, Fahrni J, et al. (2005) The testate lobose amoebae (order Arcellinida Kent, 1880) finally find their home within Amoebozoa. Protist 156: 191–202.
- 18. Nikolaev SI, Berney C, Petrov NB, Mylnikov AP, Fahrni JF, et al. (2006) Phylogenetic position of Multicilia marina and the evolution of Amoebozoa. Int J Syst Evol Microbiol 56: 191–202.
- 19. Parfrey WL, Barbero E, Lasser E, Dunthorn M, Bhattacharya D, et al. (2006) Evaluating support for the current classification of Eukaryotic diversity. PloS Genetics 2: 2062–2071.
- 20. Michel R, Wylezich C, Hauroder B, Smirnov AV (2006) Phylogenetic position and notes on the ultrastructure of Sappinia diploidea (Thecamoebidae). Protistology 4: 319–325.
- 21. Brown MW, Spiegel FW, Silberman JD (2007) Amoeba at attention: phylogenetic affinity of Sappinia pedata. J Euk Microbiol 54: 511–519.
- 22. Smirnov AV, Nassonova ES, Cavalier-Smith T (2008) Correct identification of species makes the amoebozoan rRNA tree congruent with morphology for the order Leptomyxida Page 1987; with description of Acramoeba dendroida n.g., n. sp., originally misidentified as ‘Gephyramoeba sp.’. Eur J Protistol 44: 35–44.
- 23. Tekle YI, Grant J, Anderson OR, Nerad TA, Cole JC (2008) Phylogenetic placement of diverse amoebae inferred from multigene analyses and assessment of clade stability within ‘Amoebozoa’ upon removal of varying rate classes of SSU-rDNA. Mol Phylogenet Evol 47: 339–352.
- 24. Yoon HS, Grant J, Tekle Y, Wu M, Chaon B, et al. (2008) Broadly sampled multigene trees of eukaryotes. BMC Evol Biol 8: 14.
- 25. Kudryavtsev A, Bernhard D, Schlegel M, Chao EE, Cavalier-Smith T (2005) 18S ribosomal RNA gene sequences of Cochliopodium (Himatismenida) and the phylogeny of Amoebozoa. Protist 156: 215–224.
- 26. Kudryavtsev A, Wylezich C, Schlegel M, Walochnik J, Michel R (2009) Ultrastructure, SSU rRNA gene sequences and phylogenetic relationships of Flamella Schaeffer, 1926 (Amoebozoa), with description of three new species. Protist 160: 21–40.
- 27. Pawlowski J, Burki F (2009) Untangling the Phylogeny of Amoeboid Protists. J Euk Microbiol 56: 16–25.
- 28. Whitney KD, Bennet WE, Olive LS (1982) Observations on Echinostelium bisporum. Mycologia 74: 677–680.
- 29. Spiegel FW, Feldman J (1989) Fruiting body development in the mycetozoan Echinostelium bisporum. Canad J Bot 67: 1285–1293.
- 30. Pawlowski J (2008) The twilight of Sarcodina: a molecular perspective on the polyphyletic origin of amoeboid protists. Protistology 5: 281–302.
- 31. Bolivar I, Fahrini JF, Smirnov A, Pawloski J (2001) SSU rRNA-based phylogenetic position of the genera Amoeba and Chaos (Lobosea, Gymnamoebia): the origin of gymnamoebae revisited. Mol Biol Evol 18: 2306–2314.
- 32. Dykova I, Veverkova-fialova M, Fiala I, Dvorakova H (2005) Protacanthamoeba bohemica sp. n., Isolated from the liver of tench, Tinca tinca (Linnaeus, 1758). Acta Protozool 44: 369–376.
- 33. Fiore-Donno A-M, Berney C, Pawlowski J, Baldauf SL (2005) Higher-order phylogeny of plasmodial slime molds (Myxogastria) based on elongation factor 1A and small subunit rRNA gene sequences. J Euk Microbiol 52: 201–210.
- 34. Schaap P, Winckler T, Nelson M, Alvarez-Curto E, Elgie B (2006) Molecular phylogeny and evolution of morphology in the social amoebas. Science 314: 661.
- 35. Silberman JD, Clark CG, Diamond LS, Sogin ML (1999) Phylogeny of the genera Entamoeba and Endolimax as deduced from the small-subunit ribosomal RNA sequences. Mol Biol Evol 16: 1740–1751.
- 36. Smirnov AV, Nassonova ES, Chao E, Cavalier-Smith T (2007) Phylogeny, evolution, and taxonomy of vannellid amoebae. Protist 158: 295–324.
- 37. Olive LS, Stoianovitch C (1960) Two new members of the Acrasiales. Bull Torrey Bot Club 87: 1–20.
- 38. Spiegel FW (1984) Protostelium nocturnum, a new, minute, ballistosporus protostelid. Mycologia 76: 443–447.
- 39. Spiegel FW, Shadwick JD, Hemmes JD (2006) A new ballistosporous species of Protostelium. Mycologia 98: 144–148.
- 40. Olive LS, Stoianovitch C (1971) Planoprotostelium: a new genus of protostelids. The Journal of the Mitchell Society 4: 115–119.
- 41. Spiegel FW (1982) Mitosis in the protostelid Planoprotostelium aurantium. Protoplasma 113: 178–188.
- 42. Spiegel FW (1982) The ultrastructure of the trophic cells of the protostelid Planoprotostelium aurantium. Protoplasma 113: 165–177.
- 43. Olive LS, Stoianovitch C (1966) Schizoplasmodium, a Mycetozoan genus intermediate between Cavostelium and Protostelium, a new order of mycetozoa. J Protozool 13: 164–171.
- 44. Olive LS, Stoianovitch C (1966) A simple new mycetozoan with ballistospores. Am J Bot 53: 344–349.
- 45. Olive LS, Stoianovitch C (1971) A new genus of protostelids showing affinities with Ceratiomyxa. Am J Bot 58: 32–40.
- 46. Olive LS, Stoianovitch C (1976) The ballistosporic protostelid genus Schizoplasmodium. Am J Bot 63: 1385–1389.
- 47. Whitney KD, Bennett WE (1984) An ultrastructural study of feeding techniques in three protostelids. Can J Bot 62: 1750–1755.
- 48. Olive LS, Stoianovitch C (1970) The Mycetozoa: a revised calssification. Bot Rev 36: 59–89.
- 49. Spiegel FW, Feldman J (1986) Mitosis in the protostleid Ceratiomyxella tahitiensis (Eumycetozoa). Can J Bot 64: 932–942.
- 50. Furtado JS, Olive LS (1971) Ultrastructure of the protostelid Ceratiomyxella tahitiensis including scale formation. Nova Hedwigia 21: 537–576.
- 51. Cavalier-Smith T, Chao E-Y, Oates B (2004) Molecular phylogeny of Amoebozoa and evolutionary significance of the unikont Phalansterium. Eur J Protistol 40: 21–48.
- 52. Spiegel FW, Gecks S, Feldman J (1994) Revision of the genus Protostelium (Eumycetozoa) I: The Protostelium mycophaga group and the P. irregularis Group. J Euk Micorbiol 41: 511–518.
- 53. Lindley LA, Edwards SM, Spiegel FW (2006) Variations in nucleolar morphology in Eumycetozoans. Revista Mex Micol 23: 75–81.
- 54. Olive LS, Whitney K (1982) A new species of the protostelid genus Schizoplasmodiopsis. Mycologia 74: 655–661.
- 55. Dykstra MJ (1978) Ultrastructure of the genus Schizoplasmodiopsis (Protostelia). J Protozool 25: 44–49.
- 56. Whitney KD (1984) Sporocarp ultrastructure and development in the protostelid Cavostelium apophysatum. J Protozool 31: 253–259.
- 57. Spiegel FW, Moore D, Feldman J (1995) Tychosporium acutostipes, a new protostelid which modifies the concept of the Protosteliidae. Mycologia 87: 265–270.
- 58. Olive LS, Stoianovitch C (1972) Protosporangium: a new genus of protostelids. J Protozool 19: 563–571.
- 59. Spiegel FW, Feldman J, Bennett WE (1986) Ultrastructure and development of the amoeboflagellate cells of the prototstelid Protosporangium articulatum. Protoplasma 132: 115–128.
- 60. Olive LS, Stoianovich C (1977) Clastostelium, a new ballistosporous protostelid (Mycetozoa) with flagellate cells. Trans Br Mycol Soc 69: 83–85.
- 61. Spiegel FW, Feldman J (1988) The trophic cells of Clastostelium recurvatum, a third member of the myxomycete-like protostelids. Mycologia 80: 525–535.
- 62. Bennett WE (1986) A new species of Protosporangium (Protostelida, Eumycetozoa) Mycologia 78: 857–860.
- 63. Spiegel FW (1981) Phylogenetic significance of the flagellar apparatus of Ceratiomyxa fruticulosa. Journal of the Elisha Mitchell Society 97: 183–189.
- 64. Olive LS, Stoianovitch C (1969) Monograph of the genus Protostelium. Am J Bot 56: 978–988.
- 65. Olive LS, Bennett WE, Deasey MC (1984) The new protostelid genus Endostelium. Mycologia 76: 884–891.
- 66. Bennett WE (1986) Fine structure of the trophic stage of Endostelium Olive, Bennett and Deasey 1984 (Eumycetozoea, Protostelia). Protistologica 22: 205–212.
- 67. Olive LS (1962) The genus Protostelium. Am J Bot 49: 297–303.
- 68. Olive LS, Stoianovitch C (1966) Protosteliopsis, a new genus of the Protostelida. Mycologia 58: 452–455.
- 69. Amaral-zettler LA, Cole J, Laatsch AD, Nerad TA, Anderson OR, et al. (2006) Vannella epipetala n. sp. isolated from the leaf surface of Spondias mombin (Anacardiaceae) growing in the dry forest of Costa Rica. J Euk Microbiol 53: 522–530.
- 70. Hung CY, Olive LS (1972) Ultrastructure of the amoeboid cell and its vacuolar system in Protosteliopsis fimicola. Mycologia 64: 1312–1327.
- 71. Page FC (1968) Generic criteria for Flabellula, Rugipes and Hyalodiscus, with descriptions of species. J Protozool 15: 9–26.
- 72. Bovee EC (1965) An emendation of the ameba genus Flabellula and a description of Vannella Gen. Nov. Trans Am Micro Soc 84: 217–227.
- 73. Page FC (1969) Platyamoeba stenopodia n. g., n. sp., of a freshwater amoeba. J Protozool 16: 437–441.
- 74. Spiegel FW, Feldman J (1985) Obligate amoebae of the protostelids: significance for the concept of Eumycetozoa. BioSystems 18: 377–386.
- 75. Spiegel FW (1981) Phylogenetic significance of the flagellar apparatus in protostelids (Eumycetozoa). BioSystems 14: 491–499.
- 76. Stiller JW, Hall BD (1999) Long-branch attraction and the rDNA model of early eukaryotic evolution. Mol Biol Evol 16: 1270–1279.
- 77. Cavalier-Smith T (1998) A revised system of life. Biol Rev 73: 203–266.
- 78. Cavalier-Smith T (2003) Protist phylogeny and the high-level classification of Protozoa. Eur J Protistol 39: 338–348.
- 79. Zopf W (1885) Die pilzthiere oder schleimpilze: nach dem neuesten Standpunkte bearbeitet. Breslau: Verlag von Eduard Trewendt.. 174 p.
- 80. Singer-Sam J, Tanguay RC, Riggs AD (1989) Use of Chelex to improve the PCR Signal from a small number of cells. Amplifications: A forum for PCR users 3: 11.
- 81. Medlin L, Elwood HJ, Stickel S, Sogin ML (1988) the characterization of enzymatically amplified eukaryotes 16S like ribosomal RNA coding regions. Gene 71: 491–500.
- 82. Maddison DR, Maddison WP (2002) MacClade Version 4.05: Analysis of phylogeny and character evolution. Sunderland: Sinauer Associates.
- 83. Swofford DL (2003) PAUP*, Phylogenetic Analysis Using Parsimony (*and Other Methods). Version 4. Sunderland: Sinauer Associates.
- 84. Stamatakis A (2006) RAxML-VI-HPC: Maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22: 2688–2690.
- 85. Zwickl DJ (2006) Genetic algorithm approaches for the phylogenetic analysis of large biological sequence datasets under the maximum likelihood criterion. Ph.D. dissertation, The University of Texas at Austin.
- 86. Posada D, Crandall KA (1998) Modeltest: testing the model of DNA substitution. Bioinformatics 14: 817–818.
- 87. Posada D, Buckley TR (2004) Model selection and model averaging in phylogenetics: advantages of the AIC and Bayesian approaches over likelihood ratio tests. Syst Biol 53: 79.
- 88. Ronquist F, Huelsenbeck JP (2003) MRBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 1572–1574.
- 89. Drummond AJ, Rambaut A (2007) BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evol Biol 7: 21.
- 90. Shimodaira H, Hasegawa M (2001) CONSEL: for assessing the confidence of phylogenetic tree selection. Bioinformatics 17: 1246–1247.
- 91. Shimodaira H (2002) An approximately unbiased test of phylogenetic tree selection. Syst Biol 51: 492–508.