Insights & Perspectives hn again Think

Multicellularity arose several times in the

evolution of (Response to DOI 10.1002/bies.201100187)

Laura Wegener Parfrey1) and Daniel J. G. Lahr2)

The cellular slime mold has cell-cell connections similar in struc- strate that these intermingle, and the ture, function, and underlying molecular mechanisms to animal epithelial cells. actual diversity within this ancient clade is much greater [4, 6]. These similarities form the basis for the proposal that multicellularity is ancestral amoebae spend much of their life cycle to the clade containing animals, fungi, and (including Dictyostelium): foraging in soils and leaf litter for Amorphea (formerly ‘‘unikonts’’). This hypothesis is intriguing and if true could bacteria as single celled amoebae. In precipitate a paradigm shift. However, phylogenetic analyses of two key genes response to stress or starvation hun- reveal patterns inconsistent with a single origin of multicellularity. A single origin dreds of thousands of amoebae aggre- in Amorphea would also require loss of multicellularity in each of the many uni- gate to form a motile multicellular slug that already contains differentiated popu- cellular lineages within this clade. Further, there are numerous other origins of lations of cells [1]. The slug migrates until multicellularity within eukaryotes, including three within Amorphea, that are not it reaches a suitable place, such as the top characterized by these structural and mechanistic similarities. Instead, conver- layer of soil, and then a multicellular gent evolution resulting from similar selective pressures for forming multicellular fruiting body develops in a process called structures with motile and differentiated cells is the most likely explanation for culmination. Culmination produces a dif- ferentiated multicellular fruiting body the observed similarities between animal and dictyostelid cell-cell connections. (sorocarp) composed of a mass of thou- sands of resistant spores supported by a Keywords: cellulosic stalk filled with stalk cells [1]. .Amoebozoa; catenin; Dictyostelium; metazoa; microbial ; This process is organized by stalk cells Opisthokonta; vinculin that form the tip of the slug, and later, the tip of the fruiting body. These tip cells are also the location of the cell-cell connec- Introduction: Dictyostelium development, and cell biology [1–4]. tions akin to epithelia in animals [7]. multicellularity has structural Dictyostelium is a member of the dic- Dictyostelium is used as model for tyostelid family of amoebae, which fall animal multicellularity because many and molecular similarities within the major eukaryotic clade key features are shared, such as cell to animal multicellularity Amoebozoa [5]. The are adhesion, communication and signal- alternatively referred to as cellular slime ing, differentiation, and development. Dictyostelium discoideum is a model molds or social amoebae. There are Despite deep evolutionary divergence organism that has provided insight into three described genera of dictyostelids, [8], many genes crucial to these processes the origins of multicellularity, sociality, however molecular analyses demon- in animals are found in Dictyostelium [9]. The shared processes related to multicel- lularity have been shown to be similar DOI 10.1002/bies.201200143 in structure, function, and underlying molecular machinery [3, 4, 10, 11]. 1) Department of Chemistry and Biochemistry, Both authors contributed equally to this work. These parallels are surprising because University of Colorado, Boulder, CO, USA Dictyostelium multicellularity occurs by 2) Department of Zoology, University of Sa˜ o aggregation of solitary amoebae during Paulo, Sa˜ o Paulo, Brazil Abbreviations: AM, aggregative multicellularity; DM, multicellularity one stage of the life cycle, which is by division; ML, maximum likelihood. fundamentally different from repeated *Corresponding author: Daniel J. G. Lahr division of a zygote that leads to E-mail: [email protected] multicellularity in animals. Further, the

Bioessays 00: 000–000,ß 2013 WILEY Periodicals, Inc. www.bioessays-journal.com 1 L. W. Parfrey and D. J. G. Lahr Insights & Perspectives .....

last common ancestor of dictyostelids tive explanations for the similarity convergence, meaning that the two and animals lived more than one billion between animals and Dictyostelium: groups arrived at similar methods for years ago and they are separated by homology, i.e. the common ancestor evolving multicellularity independently. numerous unicellular and multicellular of both organismal groups already pos- Multicellularity in different clades is lineages (Fig. 1). There are two alterna- sessed many of these characteristics; or commonly attributed to convergent evol- ution across eukaryotes [10, 12, 13]. Convergence is also presumed between Type Proportion animals and Dictyostelium [9]. In both Rhizaria Cercozoa (Guttulinopsis) A taxa the multicellular organism is com- posed of motile cells that must commu- Polycystina

Think again nicate and adhere to one another. In

Foraminifera Stramenopiles contrast, multicellular plants and algae Brown algae (kelp) D contain rigid, non-motile cells [10, 13]. Diatoms Based on the striking similarities between metazoan and dictyostelid cell Labyrinthulids (Sorodiplophrys) A adhesion, Dickinson and colleagues Alveolates Apicomplexa hypothesize that homology is a more Dinoflagellates likely explanation and that the last com- mon ancestor of dictyostelids and Ciliates (Sorogena) A animals was multicellular [7]. Haptophytes

Centroheliozoa Archaeplastida Green algae (incl. plants) D* Cell-cell junctions in Glaucophytes Dictyostelium resemble Red algae D* animal epithelia Cryptomonads Central to the arguments of Dickinson Euglenozoa (Trypanosoma) et al. [7] are the similarities in the organ-

Excavata ization and cell-cell junctions between Heterolobosea (Acrasids) A animal epithelium and the tip cells that Preaxostyla organize the D. discoideum slug and then Fornicata (Giardia) fruiting body. Epithelium is one of the major tissue types in animals and forms Parabasalia (Trichomonas) the lining of all free body surfaces. Epithelial cells are connected through cell (Entamoeba) junctions formed by protein complexes. (Copromyxa) A Adherens junctions are one type of junc- Protosporangids Amoebozoa tion and are found throughout animals Dictyostellids A [14, 15]. In animals, adherens junctions mechanically link the actin cytoskeleton Fractovitellida of adjacent epithelial cells to provide Myxogastrea structural integrity [16]. Adherens junc- Centramoebida tions are formed by a protein complex Amorphea that consists of transmembrane cadherin Vannelidae proteins that bind cadherins of other cells Breviates extracellularly. Inside the cell cadherins Nucleariids + Fonticula are connected to the actin cytoskeleton

A Opisthokonta via a protein complex containing a-and Fungi D b-catenin [16]. b-Catenin binds to an Mesomycetozoa (Capsaspora) intracellular domain of cadherin, and Animals D a-catenin binds to b-catenin and actin, and both a-catenin and b-catenin are Choanoflagellates required for this process [16]. Catenins are essential in adherens junctions and Figure 1. Tree of eukaryotes based on Parfrey et al. [39] and with Amoebozoa relationships key to animal multicellularity, hence the based on Lahr et al. [5] and the placement of key multicellular lineages derived from [42, 45, significance of the discovery of catenin 53, 58]. Bold lineages contain multicellular representatives. Type: A, Lineage contains at least one taxon with aggregative multicellularity. D, Contains at least one taxon that is multicellular homologs in Dictyostelium. by division. Multicellularity has arisen multiple times independently in these clades. Tip cells of D. discoideum are intercon- Proportion: Black represents the proportion of taxa within the clade that are multicellular. nected by structures similar to metazoan

2 Bioessays 00: 000–000,ß 2013 WILEY Periodicals, Inc...... Insights & Perspectives L. W. Parfrey and D. J. G. Lahr

adherens junctions [11, 17]. D. discoideum logs play a major role in cell adhesion in were performed on the CIPRES Science has homologs of both a-catenin and adherens junctions and in cell signaling Gateway version 3.2 [30]. Further, we b-catenin called Dd a-catenin and pathways, especially Wnt [24]. It has been tested whether a history of a-catenin hn again Think Aardvark, respectively [11, 18], as well as suggested that the ancestral metazoan consistent with the hypothesis of an additional member of the vinculin fam- b-catenin was a single protein that Dickinson and colleagues could be ily [6, 9]. Dd a-catenin localizes to the cell carried out both of these roles [21]. rejected using the Approximately surface during the multicellular phase of Unbiased test [31, 32]. Briefly, we gener- thelifecycle(butcruciallynotwhen ated maximum likelihood reconstruc- Dictyostelium is unicellular [11]) and shares Reconstructing the tions with RAxML using the same biochemical properties with metazoan evolutionary history of parameters as described above, but with a-catenin, but not vinculin [11]. The other the tree topology constrained to place vinculin family member was not analyzed. epithelial molecular dictyostelid a-catenins sister to the met- The functional similarities of Dd a-catenin machinery azoan a-catenins (and not nested within and metazoan a-catenin include the bind- the vinculin clades). The constrained ing of Dd a-catenin to Aardvark as well as We assess the evolutionary history of a- results were then compared to the best to mouse b-catenin and the ability of Dd and b-catenin by constructing broadly tree found in the unconstrained RAxML a-catenin to bundle actin filaments [11]. sampled phylogenetic trees of diverse analysis to calculate per-site likeli- Further, normal multicellular develop- catenin sequences and use these to hoods. The per-site likelihoods were ment of the fruiting body and the polarized evaluate the hypothesis that multicellu- then analyzed in CONSEL with standard epithelia-like tissue within the tip are dis- larity is ancestral to the Amorphea parameters (including 100,000 repli- rupted when Dd a-catenin is knocked clade. We used the online database cates) to obtain p-values [33]. down [11]. Aardvark is also required for OrthoMCL (www.orthomcl.org) to We find that the history of both normal fruiting body development and tip identify homologs of a-catenin and b- a-andb-catenin supports a scenario of organization [11, 19]. Aardvark knockouts catenin and then performed additional independent co-option for their current arereportedtodisrupttheadherensjunc- BLASTp and tBLASTn searches against roles in epithelium-like connections in tions in D. discoideum in some studies [19], each major clade and specific taxa of Dictyostelium and epithelia in animals, though not others [11]. The b-catenin interest using cutoff of e10 to increase respectively (Figs. 2 and 3, and see Box 1). homolog in D. discoideum, Aardvark,also taxon sampling. Sequences were then exhibits the dual role of cell-cell adhesion aligned with MAFFT version 6.882 [25] and signaling [17]. and filtered using GUIDANCE [26] in the Vinculin and b-catenin were Thus while the work of Dickinson online server (http://guidance.tau.ac.il/). independently co-opted for et al. does support the idea of ancestral The best fitting substitution matrices multicellularity, elucidating the evol- were determined to be LG for both a-cat- similar functions in cell-cell utionary history of catenins is key to enin and b-catenin by the ProtTest junctions evaluating the hypothesis of ancestral 2.4 online server (http://darwin.uvigo. multicellularity. However, this history is es/software/prottest2_server.html) [27], Dickinson and colleagues show through complicated by numerous gene dupli- and these were used in subsequent biochemical analysis that Dd a-catenin cations followed by subfunctionaliza- analyses. Gene trees were estimated functions like metazoan a-catenin and tion [20–22]. a- and b-catenins are not using (i) maximum likelihood (ML) with not like vinculin [11]. However, our homologous as their names would RAxML HPC version 7.2.8 [28] and (ii) analyses demonstrate that Dd a-catenin suggest, instead a-catenin is a member Bayesian analysis with MrBayes version is not an ortholog of metazoan a-catenin, of the vinculin gene family while b-cat- 3.1.2 [29]. We performed a thorough ML but rather originated by a duplication of enin is a member of the broader catenin search in RAxML by generating 100 vinculin (Fig. 2). We find that dictyostel- family [20]. Vinculins are actin-binding independent maximum parsimony trees ids, including D. discoideum,havetwo proteins that are involved in cell-cell as start points for optimization. The best copies of vinculin genes that appear to adhesion at adherens junctions and scoring ML tree was then annotated with have duplicated at the base of dictyostel- integrin-mediated junctions in animals the results of 1,000 bootstraps per- ids (Fig. 2). This duplication may have [22]. There have been numerous gene formed independently with the same taken place earlier in the history of duplications and subfunctionalization parameters. All analyses were per- Amoebozoa (indicated by an intermedi- of a-catenin within animals, and especi- formed using PROTGAMMA search ate branching Acanthamoeba castellanii ally vertebrates [20, 22]. Vinculins are algorithm. The Bayesian analysis was vinculin), but the taxonomic sampling is present in the unicellular relatives of run for ten million generations, saving not sufficient to distinguish this possib- animals, including Capsaspora [23] trees every 10,000 generations, with two ility. Additionally, while the low support and Dictyostelium [6, 11], but are independent MCMC runs with four values for the backbone of the tree not found in most eukaryotes [23]. chains each, and a heating parameter decreases our power to distinguish b-Catenin, on the other hand, has of 0.05. We obtained convergence after among alternative topologies, we can experienced many gene duplications, four million generations, the 400 trees marginally reject a single origin especially within animals [20, 21], and before convergence were discarded as of animal a-catenin and dictyostelid homologs of b-catenin are widely distrib- burnin and analyses used the remaining ‘‘a-catenin’’ (AU-test p ¼ 0.034 for uted across eukaryotes. b-Catenin homo- trees. Bayesian phylogenetic analyses monophyly of ‘‘a-catenin’’ and 0.031

Bioessays 00: 000–000,ß 2013 WILEY Periodicals, Inc. 3 L. W. Parfrey and D. J. G. Lahr Insights & Perspectives .....

between Dictyostelium Aardvark and Box 1 metazoan b-catenins has arisen through Innovation often arises through gene duplication convergence and co-option of the same eukaryotic machinery. Proposed by Ohno [66] in the 1970s as a major driver of biological evolution, gene duplication is likely responsible for providing most of the basic materials for evolution to work with. Gene duplication is pervasive across all domains of life: the Multicellularity has arisen percentage of genes that are duplicated in organismal genomes varies from 17% in many times from unicellular the bacterium Helicobacter pylori up to 65% in the plant Arabidopsis thaliana [67]. The rate at which genes duplicate and achieve fixation is comparable to the rate of ancestors across the

Think again nucleotide substitutions in mammal genomes [68]. Gene duplication can arise eukaryotic tree of life through multiple mechanisms, mainly unequal crossing over, retroposition, and whole-genome duplication. These have been reviewed elsewhere [67]. When evaluating the origin and evol- Afterduplication,thetwocopiespresentinthesamegenomearecalledparalogs ution of traits in Dictyostelium and and many rounds of duplication can result in gene families. Paralogous genes may animals, it is useful to consider these undergo different evolutionary fates: because one copy of the gene is performing lineages within the broader context of the original function, the other copy is free to diverge. The most common fate of eukaryotic diversity. The breadth of duplicated genes is pseudogenization, because there are no functional constraints eukaryotic diversity is comprised by the gene receives mutations that render the gene non-functional [68, 69]. However, more than 70 lineages that are predom- in a small percentage of cases, duplicate genes do not become pseudogenes. Some inantly microbial and have largely been do not even diverge, due to either concerted evolution by gene conversion or very defined by unique patterns of sub-cel- strong purifying selection for genes that produce high amounts of an essential lular structure by electron microscopy – product, like histones [70]. Those genes which do diverge may acquire new Patterson’s ‘‘ultra-structural identity’’ functions either by subfunctionalization or neofunctionalization, and these are [38]. These lineages have been grouped an abundant source of evolutionary novelty [71]. into a handful of major clades in recent Duplicated genes have an important consequence for phylogenetic recon- years by evolutionary hypotheses, classi- struction: ancient paralogy. This issue has been extensively explored by system- fications and molecular phylogenetic atists and is still a source of error and concern in deep level phylogenies [72]. analyses (Fig. 1; [39, 40, 41]). Deep This is because genes that duplicated before the diversification of a lineage relationships within eukaryotes are present a sampling challenge: if they have been lost in some of the derived beginning to stabilize as taxonomic lineages or if the experimental sampling was inadequate, the historical recon- sampling of the microbial lineages has struction will inevitably present artifactual relationships [73]. For instance, if we increased in large-scale phylogenetic fail to sample the vinculins in Fig. 2, then the dictyostelid actin binding proteins analyses (e.g. [39, 41, 42]). The currently will artifactually fall sister to the metazoan a-catenins. recognized major clades of eukaryotes are Opisthokonta, SAR, Amoebozoa, Excavata, and Archaeplastida (Fig. 1). for reciprocal monophyly of ‘‘a-catenin’’ varies across eukaryotes, and is Multicellularity has arisen more and vinculin). Thus, while the ability of unknown in many taxa. For instance, than 25 times across the eukaryotic tree Dd a-catenin to bind Aardvark and the b-catenin homolog in A. thaliana, of life and in all of the major clades mouse b-catenin was interpreted as con- named Arabidillo after the Drosophila (Fig. 1; [12, 13]), though the majority served function, our analysis suggests b-catenin called Armadillo,promoteslat- of eukaryotic lineages are unicellular that this ability is an example of mol- eral root branching [36]. However, this in nature [38]. Multicellularity is accom- ecular convergence. There is precedence role is not conserved within the plant plished by two major strategies: aggrega- for this degree of sequence level conver- lineage because early land plants such tion of individual cells as in Dictyostelium gence, see for example protein level as Selaginella and Physcomitrella that do and division of a single cell (zygote or convergence of Prestin in echolocating not exhibit elaborate root systems also spore) as in animals, plants and fungi bats and dolphins [34] and lysozyme in have Armadillo homologues, indicating [43]. Aggregative multicellularity (AM) foregut fermenters [35]. a more complicated history of shifting occurs as part of the life cycle when Our phylogenetic reconstruction of function. These b-catenin homologs single-celled organisms come together b-catenin demonstrates that homologs (Selagidillo and Physcodillo)havebeen to produce a fruiting body [4, 44]. AM of this protein are present in most major suggested to act instead in cell division arose in all of the major eukaryotic clades eukaryotic groups (Fig. 3). Consistent and/or cell elongation in these mosses with the exception of the Archaeplastida with previous analyses, we also recover [37]. Given the broad distribution of (A, Fig. 1; [45]). Multicellularity by a history rife with duplications, with homologs, it is likely that b-catenin division (DM) is also widespread, much taxon-specific duplication in was present in the common ancestor of animals, fungi, and plants being the most animals and vertebrates in particular all eukaryotes (Fig. 3). Because b-catenin recognizable examples (D, Fig. 1). DM is [20, 21]. However, all of the duplications homologs in most eukaryotes are not quite common in algal lineages, there occur within taxa and none predate involved in adherens junctions or cell being numerous origins in both red and the origin of major eukaryotic clades adhesion more generally, it appears likely green algae, and once at the base of the (Fig. 3). The function of b-catenins that the functional similarity observed brown algae (kelps) [12, 43]. Both types of

4 Bioessays 00: 000–000,ß 2013 WILEY Periodicals, Inc...... Insights & Perspectives L. W. Parfrey and D. J. G. Lahr

Macaca mulatta XP001113555 Canis lupus ENSCAFP00000001898 Pan troglodytes ENSPTRP00000043016 α-catenin Rattus norvegicus ENSRNOP00000008041 Homo sapiens ENSP00000347190 Equus caballus XP001504306

Mus musculus ENSMUSP00000049007 again Think Monodelphis domestica ENSMODP00000000518 Monodelphis domestica ENSMODP00000015093 89/1 Ornithorhyncus anatinus ENSOANP00000005566 Gallus gallus ENSGALP00000003847 Takifugu rubripes ENSTRUP00000029393 Tetraodon nigrovirides ENSTNIP00000011862 Danio rerio ENSDARP00000052440 Homo sapiens ENSP00000389714 Pan troglodytes ENSPTRP00000004385 Macaca mulatta XP001088422 Macaca mulatta XP001088875 Monodelphis domestica ENSMODP00000021498 100/1 Equus caballus XP001503610 Vertebrates Mus musculus ENSMUSP00000101081 Gallus gallus ENSGALP00000004478 Homo sapiens ENSP00000384638 54/0.53 Macaca mulatta XP001113131 Equus caballus XP001499507 Mus musculus XP001112948 Mus musculus ENSMUSP00000078277 Monodelphis domestica ENSMODP00000019286 Pan troglodytes ENSPTRP00000039080 Macaca mulatta XP001113096 Ratus norvegicus ENSRNOP00000007972 Gallus gallus ENSGALP00000025681 Canis lupus ENSCAFP00000011929 Tetraodon nigrovirides ENSTNIP00000007542 Takifugu rubripes ENSTRUP00000006660 Danio rerio ENSDARP00000092080 O. anatinus ENSOANP00000022446 Culex pipiens CPIJ005773 72/0.66 Aedes aegypti AAEL003754-PB Anopheles gambiae AGAP003424-PA Bombix mori BGIBMGA011888 Drosophila melanogaster FBpp0070037 Arthropods 100/1 Apis mellifera XP625229 64/1 Acyrtosiphon pisum XP001944375 Pediculus humanus PHUM450820 Ixodes scapularis EEC17127 99/1 C. elegans WBGene00001978 C. briggsae WBGene00041635 Brugia malayi 13185.m00099 Nematodes Trichoplax adherens e_gw1.8.263.1 Ciona intestinalis ENSCINP00000018199 Nematostella vectensis e_gw.6.323.1 Pan troglodytes ENSPTRP00000004598 Equus caballus XP001918196 Mus musculus ENSMUSP00000022369 Homo sapiens ENSP00000211998 Equus caballus XP001918197 Vinculins Ornithorhyncus anatinus ENSOANP00000016509 Macaca mulatta XP001118118 Canis lupus ENSCAFP00000022529 Rattus norvegicus ENSRNOP00000015179 Danio rerio ENSDARP00000100647 Tetraodon nigrovirides ENSTNIP00000013215 Vertebrates Macaca mulatta XP001103990 Monodelphis domestica ENSMODP00000000758 Gallus gallus ENSGALP00000008131 94/1 Takifugu rubripes ENSTRUP00000033393 Danio rerio ENSDARP00000039827 Tetraodon nigrovirides ENSTNIP00000019644 Takifugu rubripes ENSTRUP00000037161 Allomyces Allomyces Nematostella vectensis Fungi Trichoplax adherens Ciona intestinalis ENSCINP00000001738 Aedes aegypti AAEL005879 Culex pipiens CPIJ013873 Anopheles gambiae AGAP007532 Drosophyla melanogaster FBpp0070404 Bombyx mori BGIBMGA013369 Acyrtosiphon pisum XP001944078 Arthropods Pediculus humanus PHUM397210 68/1 Ixodes scapularis EEC14249 Caenorhabditis elegans WBGene00000942 Caenorhabditis briggsae WBGene00028154 Brugia malayi 14972.m07282 Nematodes 100/1 Proterospongia Monosiga brevicollis Choanoflagellates Capsaspora owczarzaki EFW47179 Thecamonas trahens Dictyostelium purpureum XP003288732 95/1 D. discoideum G0281499 88/0.99 Polysphondilium pallidum EFA84300 Dictyostelium fasciculatum EGG23992 94/1 Acanthamoeba castellanii Dyctiostelium fasciculatum EGG20897 Amoebozoa Polysphondilium pallidum EFA83885 Actin binding 100/1 Dictyostelium purpureum XP003291429 ”catenin” D. discoideum “Ddα-catenin” XP638018 0.4

Figure 2. Phylogenetic analysis of the vinculin gene family demonstrates that a-catenin came and together with two nested unicellular into being by gene duplication at the base of the metazoa. The most likely phylogenetic tree lineages Breviata and Apusomonadida, shows that Dictyostelium a-catenin (Dd a-catenin) arose through a separate duplication is now formally recognized as clade within the dictyostelid lineage and is not orthologous to metazoan a-catenin. The tree is Amorphea [42]. When evaluating char- shown unrooted because the position of the root of the eukaryotic tree remains uncertain. acter evolution within the Amorphea Support values are maximum likelihood bootstrap followed by Bayesian posterior probabil- ities. The sequence for Thecamonas trahens was obtained from the supplement of Sebe- clade, one must consider the position Pedros et al. [23] (where it was listed as Amastigamonas sp.). of the root of eukaryotes, which has been hypothesized to fall either within Amorphea [49, 50] or between Amorphea multicellularity are thus widespread including dictyostelids. These clades and the remaining eukaryotes [51, 52]. across eukaryotes and in the sister groups were originally joined together under The placement of the root on the eukary- of animals and Dictyostelium. the ‘‘unikont’’ hypothesis [46]. While otic tree of life remains an outstanding Two major clades of eukaryotes of the original conception of ‘‘unikonts’’ question in biology and an active area of particular interest here: the Opisthokonta, has since been refuted [47], the research [47, 50]. If the root falls between comprising animals, fungi and their Amoebozoa þ Opisthokonta relation- Amoebozoa and Opisthokonta the pos- microbial relatives; and Amoebozoa, ship is widely recovered in molecular ition of ancestral multicellularity postu- home to 20 or so lineages of amoebae, phylogenetic analyses [39, 41, 45, 48], lated by Dickinson et al. [7] would move

Bioessays 00: 000–000,ß 2013 WILEY Periodicals, Inc. 5 L. W. Parfrey and D. J. G. Lahr Insights & Perspectives .....

Macaca mulatta XP002802884 [53]. Within Amoebozoa the majority of Pan troglodytes ENSPTRP00000058940 Rattus norvegicus ENSRNOP00000026016 lineages are unicellular [5, 57], with two Mus musculus ENSMUSP00000007130 Macaca mulatta XP002802887 origins of AM in the dictyostelids and Macaca mulatta XP002802886 Homo sapiens ENSP00000385604 the genus Copromyxa [58]. Copromyxa Macaca mulatta XP002802885 Equus caballus XP001501760 falls within the Tubulinea, a diverse Monodelphis domestica ENSMODP00000014004 Ornitorhynchus anatinus ENSOANP00000010865 clade that includes Amoeba proteus Canis lupus ENSCAFP00000032261 Gallus gallus ENSGALP00000019402 of high school biology fame, and the Macaca mulatta XP001115474 Takifugu rubripes ENSTRUP00000025607 arcellinid testate amoebae [5, 58]. Tetraodon nigrovirides ENSTNIP00000009199 100/1 Danio rerio ENSDARP00000076115 Vertebrates The Myxogastrea, or plasmodial slime Danio rerio ENSDARP00000032935 Takifugu rubripes ENSTRUP00000008263 molds, are sometimes discussed as a Tetraodon nigrovirides ENSTNIP00000022288

Think again Macaca mulatta XP001107394 further example of multicellularity Homo sapiens ENSP00000377508 90/1 Pan troglodytes ENSPTRP00000015634 within the Amoebozoa because they Mus musculus ENSMUSP00000001592 Rattus norvegicus ENSRNOP00000043950 also form reproductive fruiting bodies 100/1 Canis lupus ENSCAFP00000023403 Equus caballus XP001917840 with acellular stalks. However, myxo- Monodelphis domestica ENSMODP00000018353 60/na 100/1 Takifugu rubripes ENSTRUP00000001389 gastrids are not an example of AM Tetraodon nigrovirides ENSTNIP00000007525 Danio rerio ENSDARP00000075461 because the fruiting body does not Danio rerio ENSDARP00000026915 100/1 Takifugu rubripes ENSTRUP00000034058 form from aggregation of single cells. Tetraodon nigrovirides ENSTNIP00000000864 Danio rerio ENSDARP00000076517 Instead, myxogastrids are giant multi- Ciona intestinalis ENSCINP00000014417 Trichoplax adherens e_gw1.3.898.1 nucleated cells, a syncytial amoeba, Schistosoma mansoni Smp_023550 that undergo cleavage during the Culex pipiens CPIJ006700 43/0.83 Aedes aegypti AAEL002887-PA Anopheles gambiae AGAP001043-PA development of the fruiting body Drosophila melanogaster “Armadillo” FBpp0089035 Bombyx mori BGIBMGA006959 giving rise to a mature fruiting body Apis mellifera XP623374 consisting of many cells [59]. This phy- Apis mellifera XP391947 Arthropods 98/1 Pediculus humanus PHUM000580 logenetic perspective highlights prepon- Pediculus humanus PHUM000560 Pediculus humanus PHUM280840 derance of unicellular lineages within 87/1 Acyrtosiphon pisum XP001943912 Acyrtosiphon pisum XP001946088 the Amorphea and the differences in Ixodes scapularis EEC18089 57/0.84 Nematostella vectensis e_gw.183.1.1 the nature of multicellularity that has Phytophtora XP002998437 92/1 Phytophtora EEY69790 arisen across this clade (Fig. 1), both of Phytophtora EGZ25524 Albugo CCA23071 which make a multicellular ancestor 100/1 Capsaspora owczarzacki EFW41665 SAR Plasmodium AAN35980 unlikely. Ectocarpus CBJ28776 62/0.58 Entamoeba histolytica XP65416 Entamoeba dispar XP001739153 Amoebozoa Penicillium ABO31326 100/1 Cryptococcus XP00319217 S. cerevisae NP010903 Fungi Ogataea ABO31070 Multicellularity evolves D. discoideum XP636508 D. discoideum “Aardvark” AAG17931 readily through convergent 100/1 D. purpureum XP003294817 Amoebozoa P. pallidum EFA76493 D. fasciculatum EGG15434 evolution Salpingoeca EGD79747 100/1 Salpingoeca EGD79779 98/1 Salpingoeca EGD79791 Choanoflagellates The broad distribution of multicellular Salpingoeca EGD76725 Selaginella “Selagidillo” AFI41878 taxa across the eukaryotic tree suggests Selaginella “Selagidillo” AFI41877 Arabidopsis thaliana “Arabidillo” AEC10482 100/1 that multicellularity is ‘‘easy’’ to evolve Arabidopsis thaliana “Arabidillo” AEE80050 Green Algae Physcomytrella “Physcodillo” ACQ55238 [13], and the varied mechanisms for Physcomytrella “Physcodillo” ACQ55239 0.5 achieving the requisites of multicellular- ity (e.g. cell adhesion, communication, Figure 3. Phylogenetic analyses show that there are numerous unicellular lineages and differentiation) found across taxa Dictyostelium Aardvark is not sister to the nested within. Animals and fungi are demonstrate that there are many routes animal clade and demonstrates the antiquity the most diverse groups within the to its evolution [10]. The frequent occur- of b-catenin. Other notes as in Fig. 2. Opisthokonta and both achieve multi- rence of multicellularity is likely due to cellularity by division, though the strong selective pressures favoring multi- to the base of extant eukaryotes, neces- mechanisms are very different [53, 54]. cellularity, few genetic changes necessary sitating a large number of losses. Animals are entirely multicellular, to enable the switch, or a combination of but have many unicellular relatives, these factors [13]. Experimental evidence including the Choanoflagellates and from several taxa support this view as Dictyostelium and animals Capsaspora [55]. Fungi are predomi- multicellularity can be induced rapidly represent two of the five nately multicellular, though early diverg- in the face of favorable selective pressure. ing lineages are largely unicellular and Experiments in several diverse lineages of multicellular lineages in there have been numerous reversions green algae show that predation pressure Amorphea to unicellularity across fungi [55, 56]. can repeatedly induce multicellularity The sister group of fungi is a clade that in a few generations [13], and subjecting Multicellularity is broadly distributed contains the nucleariid amoebae plus yeast to selective pressure for larger cells across the Amorphea clade, although Fonticula, an amoeboid taxon with AM produces aggregates repeatedly [60].

6 Bioessays 00: 000–000,ß 2013 WILEY Periodicals, Inc...... Insights & Perspectives L. W. Parfrey and D. J. G. Lahr

Table 1. Characteristics of aggregative multicellular organisms (slime molds) base of the Metazoa. However, this view- point has been overturned by recent genomic sequencing projects that have

Taxon Description of fruiting body Refs. again Think identified many genes involved in these Fonticula Spores inside a mucus matrix within stalk, and these are forced [44] processes in the unicellular relatives of upward to form the sorocarp. Sorocarp surrounded by fibrillar animals, especially choanoflagellates, material Capsaspora, and Mesomycetozoa [10, Copromyxa Stalk and sorocarp formed by amoebae crawling to end of [58] 23, 61–63]. As more microbial opistho- fruiting body and encysting. Fruiting body and sorocarp formed konts are sequenced it is becoming clear by encysted amoebae that the genetic toolkit for multicellular- Dictyostelids Stalk formed by excretion of cellulose matrix that may or may [1] ity was already present in unicellular not contain cells. Developmental process organized by tip cells organisms, but in case after case was Acrasids Amoeboid cells within a sheath covering form the stalk and [44] co-opted to function in a multicellular projections. Sorocarp elongates and branches so that cells are context within the Metazoa [10, 61, 62, 64]. single file, and then they encyst The widespread presence of genes Sorogena Fibrous material surrounds the spores. Acellular stalk consists [74] involved in multicellularity in uni- of a mucilaginous matrix covered by a sheath cellular organisms supports the view of Sorodiplophrys Spores of the sorocarp embedded in firm gelatinous matrix and [75] Dickinson and colleagues that the mol- covered with mucus ecular machinery necessary for multicel- Guttulinopsis Spores suspended in a matrix of slime. Stalk consists of amoeboid [44] lular organisms is ancestral. However, it cells, cysts, and disintegrated cells within a mucus matrix is also consistent with independent ori- gins of multicellularity and co-option of machinery used for signaling, communi- There are diverse strategies for achiev- different mechanisms both within cation, and other processes in microbial ing both major types of multicellularity Amorpheaandacrosseukaryotesasa lineages for new purposes in multicellu- both within Amorphea and across result of similar selective pressures [45]. lar organisms [10, 23, 62, 65]. eukaryotes, making the view of deep This survey of multicellularity in ancestral multicellularity less likely. eukaryotes highlights the diverse modes AM arose independently in at least of becoming multicellular and also the Conclusions seven eukaryotic lineages, each with unique similarities between Dictyostelium similar ecological habits [45]. These and animals. In both cases, the multi- The majority of evidence supports are terrestrial organisms that generally cellular structure consists of motile cells multiple origins of multicellularity within fruit on dung or plant material, and that are differentiated and have a com- Amorphea and eukaryotes as a whole. their similar habits suggest that similar plex pattern of development. Though Overall, this analysis highlights the selective pressures resulted in conver- dictyostelid multicellularity is much remarkable convergence in the cell-cell gence of this life history strategy [45]. different than that achieved by division adhesion in Dictyostelium and animals, In each instance AM is achieved by differ- of a single cell in animals there are demonstrated by Dickinson and col- ent mechanisms (Table 1), although many parallels because in both systems leagues [7, 11]. This likely reflects the the molecular mechanisms have only motile cells must communicate and similar selective pressures faced by been assessed in Dictyostelium and cell join together to orchestrate coordinated these two lineages that form multi- biology is much less studied in other AM movement and cellular differentiation. cellular structures from non-walled cells. lineages. The epithelia-like cell-cell con- We suggest that it is this similarity that Understanding the duplication events nections described in Dictyostelium has driven the remarkable convergence and co-option of the same molecular appear to be unique: the structure of in cell adhesion as well as other aspects machinery to perform similar functions the sorocarp does not suggest the exist- of cell biology. will illuminate the requirements for, ence of epithelia-like cell junctions in and evolution of, animal multicellularity. any other AM lineage. In most cases Moving forward, it will be crucial to the sorocarp is held together by fibrillar Molecular mechanisms incorporate an array of dictyostelids that material or a mass of mucus (Table 1). underlying animal encompass the breadth of morphological There are no apparent similarities in the multicellularity have been formsfoundintheclade[3,4],aswellas structure or formation of the sorocarp other amoebozoans, in analyses aimed at in the Amorphea lineages with AM: co-opted from unicellular elucidating the origin and evolution of Copromyxa, Fonticula, and dictyostelids ancestors dictyostelid multicellularity. These stud- (Table 1; [1, 44]). Under the scenario of a ies will in turn provide a deeper under- multicellular Amorphea ancestor, one The origin of animal multicellularity has standing of our own multicellularity. Most would expect the same mechanisms to been a topic of interest for comparative importantly, because metazoans are but a have been utilized in the evolution of biologists. For a long time the prevailing recent twig in an ancient tree, the diver- multicellularity in these taxa. All evi- view was that key features of animal sity and complexity of unicellular eukar- dence supports the widely held view that multicellularity, such as cell adhesion yotes must be acknowledged when deep AM arose independently and through and signaling, must have arisen at the evolutionary inferences are made.

Bioessays 00: 000–000,ß 2013 WILEY Periodicals, Inc. 7 L. W. Parfrey and D. J. G. Lahr Insights & Perspectives .....

Acknowledgments transition? Annu Rev Ecol Evol Syst 38: 33. Shimodaira H, Hasegawa M. 2001. 621–54. CONSEL: for assessing the confidence of Thanks to Mark Q. Martindale, Antonio 14. Leys SP, Nichols SA, Adams ED. 2009. phylogenetic tree selection. Bioinformatics C. Marques, and Mariana C. de Oliveira Epithelia and integration in sponges. Integr 17: 1246–7. for insightful discussions during the Comp Biol 49: 167–77. 34. Liu Y, Cotton JA, Shen B, Han X, et al. 2010. development of ideas within this manu- 15. Magie CR, Martindale MQ. 2008. Cell-cell Convergent sequence evolution between adhesion in the cnidaria: insights into the evol- echolocating bats and dolphins. Curr Biol script. We thank Jessica Metcalf, Se Jin ution of tissue morphogenesis. Biol Bull 214: 20: R53–4. Song, and Marcos Barbeitos for com- 218–32. 35. Kornegay JR, Schilling JW, Wilson AC. ments on the manuscript. The manu- 16. Niessen CM. 2007. Tight junctions/adherens 1994. Molecular adaptation of a leaf-eating junctions: basic structure and function. bird: stomach lysozyme of the hoatzin. Mol script was improved following the J Invest Dermatol 127: 2525–32. Biol Evol 11: 921–8. comments of two anonymous reviewers 17. Grimson MJ, Coates JC, Reynolds JP, 36. Coates JC, Laplaze L, Haseloff J. 2006. Shipman M, et al. 2000. Adherens junctions Armadillo-related proteins promote lateral Think again and the editor. This work was conceived and partly written during the Sa˜o Paulo and [beta]-catenin-mediated cell signalling root development in Arabidopsis. Proc Natl in a non-metazoan organism. Nature 408: Acad Sci USA 103: 1621–6. School of Advanced Sciences—evol- 727–31. 37. Moody L, Saidi Y, Smiles E, Bradshaw S,et ution, funded by FAPESP Grant #2011/ 18. Coates JC, Harwood AJ. 2001. Cell-cell al. 2012. ARABIDILLO gene homologues in 51916-0. D.J.G.L. is funded by CNPq adhesion and signal transduction during basal land plants: species-specific gene Dictyostelium development. J Cell Sci 114: duplication and likely functional redundancy. (501089/2011-0). 4349–58. Planta 236: 1927–41. 19. Coates JC, Grimson MJ, Williams RS, 38. Patterson DJ. 1999. The diversity of eukar- Bergman W, et al. 2002. Loss of the beta- yotes. Am Nat 154: S96–124. catenin homologue aardvark causes ectopic 39. Parfrey LW, Grant J, Tekle YI, Lasek- References stalk formation in Dictyostelium. Mech Dev Nesselquist E, et al. 2010. Broadly sampled 116: 117–27. multigene analyses yield a well-resolved 1. Kessin R. 2001. Dictyostelium: Evolution, 20. Zhao Z-M, Reynolds A, Gaucher E. 2011. eukaryotic tree of life. Syst Biol 59: 518–33. Cell Biology, and the Development of The evolutionary history of the catenin gene 40. Adl SM, Simpson AGB, Farmer MA, Multicellularity. Cambridge: Cambridge family during metazoan evolution. BMC Evol Andersen RA, et al. 2005. The new higher University Press. 284 p. Biol 11: 198. level classification of eukaryotes with empha- 2. Williams RS, Boeckeler K, Graf R, Muller- 21. Schneider SQ, Finnerty JR, Martindale MQ. sis on the of protists. J Eukaryot Taubenberger A, et al. 2006. Towards a mol- 2003. Protein evolution: structure-function Microbiol 52: 399–451. ecular understanding of human diseases relationships of the oncogene beta-catenin 41. Burki F, Okamoto N, Pombert JF, Keeling using Dictyostelium discoideum. Trends Mol in the evolution of multicellular animals. PJ. 2012. The evolutionary history of hapto- Med 12: 415–24. J Exp Zool B Mol Dev Evol 295: 25–44. phytes and cryptophytes: phylogenomic evi- 3. Urushihara H. 2008. Developmental biology 22. Kobielak A, Fuchs E. 2004. Alpha-catenin: at dence for separate origins. Proc Biol Sci 279: of the social amoeba: history, current knowl- the junction of intercellular adhesion and actin 2246–54. edge and prospects. Dev Growth Differ 50: dynamics. Nat Rev Mol Cell Biol 5: 614–25. 42. Adl SM, Simpson AGB, Lane CE, Lukesˇ J, S277–81. 23. Sebe-Pedros A, Roger AJ, Lang FB, King N, et al. 2012. The revised classification of eukar- 4. Romeralo M, Escalante R, Baldauf SL. et al. 2010. Ancient origin of the integrin-medi- yotes. J Euk Microbiol 59: 429–514. 2012. Evolution and diversity of dictyostelid ated adhesion and signaling machinery. Proc 43. Bonner JT. 1998. The origins of multicellular- social amoebae. Protist 163: 327–43. Natl Acad Sci USA 107: 10142–7. ity. Integrative Biol 1: 27–36. 5. Lahr DJ, Grant J, Nguyen T, Lin JH,etal. 24. Wikramanayake AH, Hong M, Lee PN, Pang 44. Blanton RL. 1990. Phylum acrasea. In 2011. Comprehensive phylogenetic recon- K, et al. 2003. An ancient role for nuclear Margulis L, Corliss JO, Melkonian M, struction of amoebozoa based on concaten- [beta]-catenin in the evolution of axial polarity Chapman DJ, eds; Handbook of Protoctista. ated analyses of SSU-rDNA and actin genes. and germ layer segregation. Nature 426: 446– Boston, MA: Jones and Bartlett. p. 75–87. PLoS ONE 6: e22780. 50. 45. Brown MW, Kolisko M, Silberman JD, 6. Heidel AJ, Lawal HM, Felder M, Schilde C, 25. Katoh K, Toh H. 2008. Recent developments Roger AJ. 2012. Aggregative multicellularity et al. 2011. Phylogeny-wide analysis of social in the MAFFT multiple sequence alignment evolved independently in the eukaryotic amoeba genomes highlights ancient origins program. Brief Bioinform 9: 286–98. supergroup Rhizaria. Curr Biol 22: 1123–7. for complex intercellular communication. 26. Penn O, Privman E, Ashkenazy H, Landan 46. Cavalier-Smith T. 2003. Protist phylogeny Genome Res 21: 1882–91. G, et al. 2010. GUIDANCE: a web server for and the high-level classification of Protozoa. 7. Dickinson DJ, Nelson WJ, Weis WI. 2012. assessing alignment confidence scores. Euro J Protistol 39: 338–48. An epithelial tissue in Dictyostelium chal- Nucleic Acids Res 38: W23–8. 47. Roger AJ, Simpson AGB. 2009. Evolution: lenges the traditional origin of metazoan multi- 27. Abascal F, Zardoya R, Posada D. 2005. revisiting the root of the eukaryote tree. Curr cellularity. BioEssays 34: 833–40. ProtTest: selection of best-fit models of Biol 19: R165–7. 8. Parfrey LW, Lahr DJ, Knoll AH, Katz LA. protein evolution. Bioinformatics 21: 2104–5. 48. Katz LA, Grant J, Parfrey LW, Gant A,etal. 2011. Estimating the timing of early eukaryotic 28. Stamatakis A, Hoover P, Rougemont J. 2011. Subulatomonas tetraspora nov. gen. diversification with multigene molecular 2008. A rapid bootstrap algorithm for the nov. sp. is a member of a previously unrecog- clocks. Proc Natl Acad Sci USA 108: RAxML Web-servers. Syst Biol 57: 758–71. nized major clade of eukaryotes. Protist 162: 13624–9. 29. Ronquist F, Huelsenbeck JP. 2003. 762–73. 9. Eichinger L, Pachebat JA, Glockner G, MrBayes 3: Bayesian phylogenetic inference 49. Stechmann A, Cavalier-Smith T. 2002. Rajandream MA, et al. 2005. The genome under mixed models. Bioinformatics 19: Rooting the eukaryote tree by using a derived of the social amoeba Dictyostelium discoi- 1572–4. gene fusion. Science 297: 89–91. deum. Nature 435: 43–57. 30. Miller MA, Pfeiffer W, Schwartz T. 2010. 50. Katz LA, Grant JR, Parfrey LW, Burleigh JG. 10. Abedin M, King N. 2010. Diverse evolutionary Creating the CIPRES Science Gateway for 2012. Turning the crown upside down: gene paths to cell adhesion. Trends Cell Biol 20: inference of large phylogenetic trees. tree parsimony roots the eukaryotic tree of 734–42. Proceedings of the Gateway Computing life. Syst Biol 61: 653–60. 11. Dickinson DJ, Nelson WJ, Weis WI. 2011. Environments Workshop (GCE), 14 Nov 51. Stechmann A, Cavalier-Smith T. 2003. The A polarized epithelium organized by beta- and 2010, New Orleans, LA pp. 1–8. root of the eukaryote tree pinpointed. Curr alpha-catenin predates cadherin and meta- 31. Shimodaira H. 2002. An approximately Biol 13: R665–6. zoan origins. Science 331: 1336–9. unbiased test of phylogenetic tree selection. 52. Derelle R, Lang F. 2012. Rooting the eukary- 12. Knoll AH. 2011. The multiple origins of com- Syst Biol 51: 492–508. otic tree with mitochondrial and bacterial plex multicellularity. Annu Rev Earth Pl Sci 39: 32. Shimodaira H. 2004. Approximately proteins. Mol Biol Evol 29: 1277–89. 217–39. unbiased tests of regions using multistep- 53. Brown MW, Spiegel FW, Silberman JD. 13. Grosberg RK, Strathmann RR. 2007. The multiscale bootstrap resampling. Ann Stat 2009. Phylogeny of the ‘‘forgotten’’ cellular evolution of multicellularity: a minor major 32: 2616–41. slime mold, Fonticula alba, reveals a key evol-

8 Bioessays 00: 000–000,ß 2013 WILEY Periodicals, Inc...... Insights & Perspectives L. W. Parfrey and D. J. G. Lahr

utionary branch within Opisthokonta. Mol Biol 60. Ratcliff WC, Denison RF, Borrello M, 66. Ohno S. 1970. Evolution by Gene Duplication. Evol 26: 2699–709. Travisano M. 2012. Experimental evolution New York: Spinger-Verlag. 54. Medina M, Collins A, Taylor J, Valentine J, of multicellularity. Proc Natl Acad Sci USA 67. Zhang J. 2003. Evolution by gene duplication: et al. 2003. Phylogeny of Opisthokonta and 109: 1595–600. an update. Trends Ecol Evol 18: 292–8. hn again Think the evolution of multicellularity and complex- 61. Nichols SA, Roberts BW, Richter DJ, 68. Lynch M, Conery JS. 2000. The evolutionary ity in Fungi and Metazoa. Int J Astrobiol 2: Fairclough SR, et al. 2012. Origin of meta- fate and consequences of duplicate genes. 203–11. zoan cadherin diversity and the antiquity of Science 290: 1151–5. 55. Torruella G, Derelle R, Paps J, Lang BF, the classical cadherin/beta-catenin complex. 69. Walsh JB. 1995. How often do duplicated genes et al. 2012. Phylogenetic relationships within Proc Natl Acad Sci USA 109: 13046–51. evolve new functions? Genetics 139: 421–8. the Opisthokonta based on phylogenomic 62. Suga H, Dacre M, de Mendoza A, Shalchian- 70. Nei M, Rogozin IB, Piontkivska H. 2000. analyses of conserved single-copy protein Tabrizi K, et al. 2012. Genomic survey of Purifying selection and birth-and-death evol- domains. Mol Biol Evol 29: 531–44. Premetazoans shows deep conservation of ution in the ubiquitin gene family. Proc Natl 56. Hibbett DS, Binder M, Bischoff JF, cytoplasmic tyrosine kinases and multiple radi- Acad Sci USA 97: 10866–71. Blackwell M, et al. 2007. A higher-level phylo- ations of receptor tyrosine kinases. Sci Signal 5: 71. De Smet R, Van de Peer Y. 2012. genetic classification of the Fungi. Mycol Res ra35. Redundancy and rewiring of genetic networks 111: 509–47. 63. Sebe-Pedros A, de Mendoza A, Lang BF, following genome-wide duplication events. 57. Smirnov AV, Chao E, Nassonova ES, Degnan BM, et al. 2011. Unexpected reper- Curr Opin Plant Biol 15: 168–76. Cavalier-Smith T. 2011. A revised classification toire of metazoan transcription factors in the 72. Qiu H, Yang EC, Bhattacharya D, Yoon HS. of naked lobose amoebae (Amoebozoa: unicellular holozoan Capsaspora owczarzaki. 2012. Ancient gene paralogy may mislead ). Protist 162: 545–70. Mol Biol Evol 28: 1241–54. inference of plastid phylogeny. Mol Biol Evol 58. Brown MW, Silberman JD, Spiegel FW. 64. Sebe-Pedros A, Ruiz-Trillo I. 2010. Integrin- 29: 3333–43. 2011. ‘‘Slime molds’’ among the Tubulinea mediated adhesion complex: cooption of 73. Maddison WP. 1997. Gene trees in species (Amoebozoa): molecular systematics and signaling systems at the dawn of Metazoa. trees. Syst Biol 46: 523–36. taxonomy of Copromyxa. Protist 162: 277–87. Commun Integr Biol 3: 475–7. 74. Olive LS, Blanton RL. 1980. Aerial sorocarp 59. Frederick L. 1990. Phylum plasmodial slime 65. Mikhailov KV, Konstantinova AV, Nikitin development by the aggregative ciliate molds, class Myxomycota. In Margulis L, MA, Troshin PV, et al. 2009. The origin of Sorogena stoianovitchae. JProtozool27: 293–9. Corliss JO, Melkonian M, Chapman DJ, Metazoa: a transition from temporal to 75. Dykstra MJ, Olive LS. 1975. Sorodiplo- eds; Handbook of Protoctista. Boston: spatial cell differentiation. BioEssays 31: phrys – unusual sorocarp producing protist. Jones and Bartlett. p. 467–83. 758–68. Mycologia 67: 873–9.

Bioessays 00: 000–000,ß 2013 WILEY Periodicals, Inc. 9