sign in sign up
<<
Home , Acrasidae, Fonticula, Phytomyxea, Rhizaria, Rosculus, Schizopyrenida

Current Biology 22, 1123–1127, June 19, 2012 ª2012 Elsevier Ltd All rights reserved DOI 10.1016/j.cub.2012.04.021 Report Aggregative Multicellularity Evolved Independently in the Eukaryotic Supergroup

Matthew W. Brown,1,* Martin Kolisko,1,2 called a sorocarp (Figure 1A). Their life cycles are socially Jeffrey D. Silberman,3 and Andrew J. Roger1,* intricate because each cell retains its individuality yet works 1Centre for Comparative Genomics and Evolutionary with other like-cells to form a variably complex multicellular Bioinformatics, Department of Biochemistry sorocarp. The ‘‘patchy’’ phylogenetic distribution of soro- and Molecular Biology carpic across the tree of suggests that 2Centre for Comparative Genomics and Evolutionary they have independently converged upon this analogous Bioinformatics, Department of Biology mode of propagule dispersal, which is also similar to that of Dalhousie University, Halifax, Nova Scotia, B3H 4R2, Canada the analogous life cycle in myxobacteria [9]. Here we conclu- 3Department of Biological Sciences, University of Arkansas, sively determine the evolutionary position of Guttulinopsis Fayetteville, AR, 72701, USA vulgaris, a ubiquitous but under-studied sorocarpic , using a phylogenomic approach with one of the largest protein supermatrices assembled to date that encompasses all major Summary groups of eukaryotes. The genus Guttulinopsis, composed of four , was Multicellular forms of life have evolved many times, indepen- described over a century ago [10] and has been variously dently giving rise to a diversity of organisms such as classified [11, 12]. Guttulinopsis vulgaris is the most common , , and fungi that together comprise the visible species and can be found globally on the dung of various biosphere. Yet multicellular life is far more widespread herbivores [11, 13–15]. Despite morphological and ultrastruc- among eukaryotes than just these three lineages. A particu- tural studies [11, 12, 14, 16] of sorocarps (Figures 1A–1C) larly common form of multicellularity is a social aggregative and trophozoites (Figures 1D and 1E), the phylogenetic fruiting lifestyle whereby individual cells associate to form position of Guttulinopsis has remained unclear. The lobose a ‘‘-like’’ sorocarp. This complex developmental pro- morphology of the amoebae suggests an affinity to either cess that requires the interaction of thousands of cells the supergroup or the phylum Heterolo- working in concert was made famous by the ‘‘cellular slime bosea (Figures 1D and 1E). The mitochondria of G. vulgaris mold’’ discoideum, which became an impor- have flattened discoid cristae [2, 16] and the cells lack a tant model organism [1]. Although sorocarpic protistan discrete Golgi apparatus [17], both characteristics of Heterolo- lineages have been identified in five of the major bosea [2, 18]. On the other hand, G. vulgaris lacks rough endo- groups [2–8], the ubiquitous and globally distributed plasmic reticulum surrounding its mitochondria, a character species Guttulinopsis vulgaris has eluded proper classifica- present in most heteroloboseans [16, 18]. The balance of tion. Here we demonstrate, by phylogenomic analyses of morphological data hinted that G. vulgaris might be related a 159-protein data set, that G. vulgaris is a member of Rhiza- to the putative heterolobosean amoeba genus, , ria and is thus the first member of this eukaryote supergroup which shares all of the above morphological characteristics known to be capable of aggregative multicellularity. [2, 19, 20] but does not form sorocarps. However, to date, no sequence data from Guttulinopsis was available to test Results and Discussion these hypotheses, and the only sequence currently available from Rosculus spp. is a nuclear encoded mitochondrial The origins of multicellular life forms from unicellular ancestors single-subunit RNA polymerase gene (mtSNAP) (GenBank are among the most profound innovations in the . DQ388527) that, to our knowledge, has not been phylogeneti- Although multicellularity has originated multiple times in a cally analyzed. great diversity of eukaryotic lineages and, more controver- The standard approach to investigate phylogenetic affinity sially, in some prokaryote lineages [9], many of these indepen- of protistan taxa is to utilize an 18S ribosomal RNA gene dent evolutionary transitions have been overlooked. Molecular phylogeny. However, attempts to amplify G. vulgaris 18S phylogenetic data have recently revealed that the sorocarpic rDNA were unsuccessful when employing a comprehensive life style, commonly known as the cellular habit set of ‘‘universal’’ eukaryotic 18S primers (data not shown). [1], is the most widely distributed form of multicellularity, Instead, we obtained multiple genes for phylogenetic analyses rivaled only by simple colonial pluricellularity. Examples of by conducting massively parallel 454-pyrosequencing (Roche, this form are found in at least five major eukaryote groups Branford, CT, USA) of complementary DNA (cDNA) to sample [2]; these are the and Copromyxa in Amoebozoa, the transcriptome of G. vulgaris and another sorocarpic the acrasids in Excavata, alba in Opisthokonta, amoeba, F. alba, that our previous analyses indicate is closely Sorogena in the Alveolata, and Sorodiplophrys in the Strame- related to the fungi [6]. We integrated these protein-coding nopiles [3–8]. The phylogenetic position of Copromyxella has gene sequences into a pre-existing data set united from yet to be confirmed by molecular analyses, and it may two prior phylogenomic studies [21, 22]. The final 67 represent yet another independent derivation of this lifestyle. taxa, 159-protein phylogenomic supermatrix (43,615 amino In all of these organisms, cell aggregates are capable of devel- acid positions), included representatives from each of the oping into a ‘‘fungus-like’’ propagule-bearing fruiting body, proposed eukaryotic supergroups and the , , , and telonemids [2]. Unexpectedly, phylogenetic analyses using both Bayesian inference (BI) and *Correspondence: @live.com (M.W.B.), [email protected] (A.J.R.) maximum-likelihood (ML) methods revealed that G. vulgaris Current Biology Vol 22 No 12 1124

Figure 1. The Life Stages of Guttulinopsis vulgaris Sorocarps appear as white to pale-yellow stalked ‘‘balls’’ projecting from the surface of 1–7 day old dung. They are macroscopic (200–500 mm in height) and consist of >20,000 cells [1] with a distinct stalk and one to several spore masses (sorus[i]) atop the stalk (Figures 1A–1C). Mature sorocarps consist of distinct cell types, irregu- larly shaped walled spores (Figure 1C) along with stalk compartments containing degenerated, amoeboid, and encysted stalk cells [14]. (A) Multicellular fruiting body with a single sorus fruiting on a piece of cow dung substrate. Scale bar represents 250 mm. (B) Multicellular fruiting body with three sori projecting from a single stalk. Scale bar represents 250 mm. (C) Whole mount of the fruiting body shown in (A). Photo- graphed using differential interference contrast micros- copy. Scale bar represents 50 mm. (D) Typical irregularly shaped spores. Scale bar repre- sents 10 mm. (E) Lobose amoeba and round spore isolated from sorus. Scale bar represents 10 mm. The lobose amoebae (shown in E) of G. vulgaris are very similar to those in the genus Rosculus. The phylogenetic relationship of the two taxa is confirmed through analyses of the mtSNAP gene (Figure S1). robustly groups within the supergroup Rhizaria (1.0/100%, contains organisms such as the taxon-rich [26], BI posterior probability and ML bootstrap, Figure 2). This , testate amoeboid gromiids, and marine plank- affiliation was unanticipated considering the morphological tonic , along with and parasites (Haplo- and ultrastructural characters of Guttulinopsis, which are not sporidia, ). Despite this diversity, there are two especially similar to any known rhizarian. unifying morphological themes in Rhizaria: possession of The exact position of G. vulgaris within Rhizaria is less tubular mitochondrial cristae and the ability of cells to form definitively resolved. Our phylogenies demonstrate that fine that are either reticulose (network with G. vulgaris consistently branches within Filosea (the ‘‘core- thin thread-like connections) or filose (fine thread-like) [25]. Cercozoa’’) (0.99/99%) (Figure 2), here represented by the However, both of these likely represent symplesiomorphies taxa Limnofila sp., natans, and because tubular mitochondrial cristae are common outside marina. Guttulinopsis vulgaris specifically branches sister to of Rhizaria and fine pseudopodia are not restricted to Rhizaria P. marina with moderately high topological support in our (i.e., nucleariid amoebae [Opisthokonta] and diplophoriid ML tree (1.0/94%). Topology tests employing our phyloge- amoebae [stramenopiles]). Guttulinopsis is a very unusual nomic data set reject: (1) the placement of G. vulgaris outside rhizarian, even considering Rhizaria’s extreme morphological but sister to Filosea and (2) as outside but sister to Rhizaria diversity. The amoebae of Guttulinopsis do not produce fine (see Table S1 available online). However, determination of pseudopodia; rather, their pseudopodia are broad-lobose the deep phylogenetic branching patterns within Rhizaria and form eruptively along the cell’s anterior surface (Figure 1E; using this protein supermatrix should be seen as tentative until see also Movies S1 and S2) in a similar fashion to members of a much broader taxon sampling from each major rhizarian the Heterolobosea [14, 19]. The life cycle of G. vulgaris lacks taxonomic group is achieved. a flagellated stage, although flagellated cells (often flagellated Regardless of its position within the group, the placement gametes) are present in nearly all described rhizarian taxa. of Guttulinopsis within Rhizaria does not appear to be an arti- Discoid mitochondrial cristae like that of G. vulgaris are also factual consequence of long-branch attraction [23] between rare in Rhizaria—known only in a few members such as the its long terminal branch and the long branches of other Sainouroidea [27, 28]. Protein-coding data from taxa in the rhizarians. We showed this by inferring trees from a reduced Sainouroidea are needed to examine whether this ultrastruc- taxon data set devoid of all long-branched rhizarians except tural character indicates a specific phylogenetic affinity to for G. vulgaris (i.e., removing Phyllostaurus, Astrolonche, G. vulgaris. , and filosa). The resulting trees To assess the morphology-based hypothesis of a Guttuli- continue to recover a very strong rhizarian affiliation of nopsis/Rosculus relationship [2, 19, 20] as noted above, we G. vulgaris (see Supplemental Information). We also applied recovered an expressed sequence tag (EST) cluster from several different phylogenetic methods that accounted for G. vulgaris that encodes a mtSNAP protein and constructed heterotachy (i.e., site-rate shifts across the tree), and all of a phylogeny. The G. vulgaris and Rosculus sp. mtSNAP homo- them also recovered a similar phylogenetic position for logs are clearly closely related, sharing 91% sequence identity G. vulgaris as shown in Figure 2, indicating that heterotachy- at the amino acid level (n = 153) and robustly grouping together induced artifacts were not responsible for the result (see on a phylogenetic tree (Figure S1). This confirms their close Supplemental Information). relationship and by extension, draws Rosculus into Rhizaria. Rhizaria was originally established solely on the basis of Unfortunately, mtSNAP sequences are not currently available molecular phylogenetic evidence [24] and represents one of from other rhizarians (including the recently released genome the most morphologically diverse groups of eukaryotes with of B. natans), so the placement of the G. vulgaris/Rosculus sp. no clearly defined morphological synapomorphy [25]. It within Rhizaria could not be further addressed. Aggregative Multicellularity in Rhizaria 1125

Figure 2. Unrooted Phylogenetic Tree Estimated from the 67-Taxon 159-Protein Data Set Inferred by Phylobayes under the CAT-Poisson Model The tree shown is the consensus tree of seven converged Phylobayes Markov chain Monte Carlo chains run under the CAT-Poisson model run for 2,500 generations with a burnin of 250 generations. Taxa from each representative taxonomic group are color-coded accordingly. Topological support for each branch was estimated using 1,000 ML bootstrap replicates using the LG + G + F model implemented in RAxML and is shown as the upper value on each branch. The lower values are Bayesian posterior probabilities. Black dots indicate ML bootstrap values of 100% and Bayesian posterior probabilities of 1.0. ML bootstrap support values <50% are denoted by ‘‘–.’’ The histogram to the left of the tree illustrates the gene sampling in the supermatrix for each taxon. Guttulinopsis vulgaris clearly groups within the Rhizaria in both Bayesian and ML analyses (see also the Supplemental Information). Analyses of individual genes from G. vulgaris show consistency of base composition, codon usage, and phylogenetic signal (see Figure S2 and the Supplemental Infor- mation). As with any phylogenomic analysis of eukaryotic diversity, computational constraints prevents comprehensive sampling of taxa in all major groups and the exact branching order within groups can change somewhat depending on the precise subsampling of taxa (e.g. see [21] for example).

Guttulinopsis vulgaris exhibits a unique form of multicellu- some rhizarians form propagule dispersal structures that are larity that is not associated with the life cycle of any other the result of plasmodial cleavage (i.e., some chlorarachnio- known rhizarian. Until now, the extent of multicellularity or phytes, plasmodiophrids, and haplosporidians), and some pluricellularity in Rhizaria only included simple colonial foraminifera can form complex multilocular skeletons with members that form as the result of clonal cell division within different compartments having different functions. It is likely a gelatinous matrix, for example, the filosean spongomonads that aggregative cooperative multicellularity is molecularly (Spongomonas spp. and Rhipidodendron spp.). There are and developmentally more complex than these examples. In also several groups that have multinucleated life-cycle stages any case, our results now indicate that aggregative fruiting is that could be interpreted as ‘‘multicellular.’’ For example, the most broadly distributed form of multicellularity among Current Biology Vol 22 No 12 1126

eukaryotes. As far as we know, the (green Bigna1), the Broad Institute for access to the genome data of Capsaspora + land plants, , and ) is the only re- owczarzaki, Salpingoeca rosetta, and Thecamonas trahens (http://www. maining supergroup that lacks aggregative fruiting members. broadinstitute.org/annotation/genome/multicellularity_project), and Baylor College of Medicine for access to the genome data of cas- The evolutionary pressures that led to the convergent evolu- tellanii (http://www.hgsc.bcm.tmc.edu/microbial-detail.xsp?project_id=163). tion of aggregative fruiting lifestyles at least seven times We thank Frederick W. Spiegel for discussion during the early stages of across the tree of life have not been examined, but represent this project. We thank the late Robert K. Beeler and Donald R. Beeler for an intriguing and outstanding problem in evolutionary biology. providing access to their cattle. We thank Daniel Gaston and Javier Alfaro These organisms are on the boundary between uni- and multi- for bioinformatic help. The authors would like to thank the Canadian consor- cellularity, having a life cycle with both lifestyles and emergent tium Protist Expressed Sequence Tag Project (PEP) that has made publicly available several EST projects through its database (http://tbestdb.bcm. social behavior. Sorocarpic protists like the dictyostelids umontreal.ca). Computing resources were provided by the University of Copromyxa, Fonticula, acrasids, and Guttulinopsis are excel- Toronto’s SciNet supercomputing facility, of Compute/Calcul Canada, lent model organisms to examine the evolution of cellular, a national infrastructure for supercomputing-powered innovation. This developmental, and regulatory components that result in a work was made possible by the Discovery grant 227085-11 from multicellular state with attributes that must include cell-cell the Natural Sciences and Engineering Research Council of Canada and communication, some degree of differentiation, and simple the Canada Research Chair awarded to A.J.R. This work was partly sup- ported through National Science Foundation grant DEB 0329102 and the programmed cell death [1]. This suite of fruiting organisms is Arkansas Biosciences Institute. M.W.B. is supported by a fellowship from broadly distributed across the eukaryotic tree of life and the Tula Foundation. A.J.R. is partly supported by the Canadian Institute provides an excellent opportunity for rigorous comparative for Advanced Research’s program in Integrated Microbial Biodiversity. developmental genomics. Received: February 13, 2012 Experimental Procedures Revised: March 23, 2012 Accepted: April 12, 2012 The Supplemental Information details complete experimental procedures. Published online: May 17, 2012 In summary, total RNA from G. vulgaris and F. alba cultures was extracted. From the total RNA, cDNA libraries were constructed and sequenced using References high throughput massively parallelized 454 sequencing. The raw 454 sequence data from these organisms were assembled, and then 159 protein 1. Bonner, J.T. (2003). On the origin of differentiation. J. Biosci. 28, alignments were generated as described in Supplemental Information. 523–528. Single protein alignments were used to estimate trees and the data from 2. Adl, S.M., Simpson, A.G.B., Farmer, M.A., Andersen, R.A., Anderson, G. vulgaris, and other taxa were screened for significant incongruence O.R., Barta, J.R., Bowser, S.S., Brugerolle, G., Fensome, R.A., with a conservative reference tree of accepted eukaryotic relationships. Fredericq, S., et al. (2005). The new higher level classification of eukary- Individual gene data from G. vulgaris was tested for congruence in codon otes with emphasis on the of protists. J. Eukaryot. Microbiol. usage, base composition, and phylogenetic signal (see Figure S2 and 52, 399–451. Supplemental Information). All new sequence data used in this study from 3. Baldauf, S.L., Roger, A.J., Wenk-Siefert, I., and Doolittle, W.F. (2000). A G. vulgaris and F. alba have been deposited in GenBank (www.ncbi.nlm. -level phylogeny of eukaryotes based on combined protein nih.gov) under Bioproject PRJNA82891 and PRJNA82893, respectively. data. Science 290, 972–977. Suspected contaminating sequences or suspected paralogs were removed 4. Brown, M.W., Silberman, J.D., and Spiegel, F.W. (2011). ‘‘Slime molds’’ as described in Supplemental Information. The ‘‘cleaned’’ 159 protein align- among the (Amoebozoa): molecular systematics and ments were concatenated into a supermatrix of 43,615 amino acid charac- taxonomy of Copromyxa. Protist 162, 277–287. ters. After appropriate ‘‘best’’ models were selected for ML and Bayesian 5. Brown, M.W., Silberman, J.D., and Spiegel, F.W. (2012). A contemporary analyses, trees were inferred using PhyloBayes v3.3b [29] employing the evaluation of the acrasids (, Heterolobosea, Excavata). Eur. J. CAT-Possion model and RAxML [30] employing the the Le and Gascuel Protistol. 48, 103–123. G (LG) model + distribution with four rate categories and empirical amino 6. Brown, M.W., Spiegel, F.W., and Silberman, J.D. (2009). Phylogeny of acid frequencies for ML analyses. Topological support for the phylogenetic the ‘‘forgotten’’ cellular slime mold, Fonticula alba, reveals a key evolu- trees included 1,000 individual bootstrap ML replicates for the 159-protein tionary branch within Opisthokonta. Mol. Biol. Evol. 26, 2699–2709. alignments and the mtSNAP data set. For details of further phylogenetic 7. Lasek-Nesselquist, E., and Katz, L.A. (2001). Phylogenetic position of analyses including the congruence, topology testing, heterotachy analyses, Sorogena stoianovitchae and relationships within the class Colpodea and partitioned analyses, see the Supplemental Information. (Ciliophora) based on SSU rDNA sequences. J. Eukaryot. Microbiol. 48, 604–607. Accession Numbers 8. Schaap, P., Winckler, T., Nelson, M., Alvarez-Curto, E., Elgie, B., Hagiwara, H., Cavender, J., Milano-Curto, A., Rozen, D.E., All new sequence data used in this study from G. vulgaris and F. alba Dingermann, T., et al. (2006). Molecular phylogeny and evolution of have been deposited in GenBank (www.ncbi.nlm.nih.gov) under BioProject morphology in the social . Science 314, 661–663. PRJNA82891 (GenBank accession numbers JT844884–JT845030) 9. Bonner, J.T. (1998). The origins of multicellularity. Integr. Biol. 1, 27–36. and PRJNA82893 (GenBank accession numbers JT844813–JT844883), 10. Olive, E.W. (1901). A preliminary enumeration of the Sorophoreae. Proc. respectively. Am. Acad. Arts Sci. 37, 333–334. 11. Olive, L.S. (1975). The Mycetozoans (New York: Academic Press). Supplemental Information 12. Raper, K.B. (1984). The Dictyostelids (Princeton, New Jersey: Princeton University Press). Supplemental Information includes two figures, two tables, Supplemental 13. Olive, E.W. (1902). Monograph of the Acrasieae. Proc. Boston Soc. Nat. Experimental Procedures, and two movies and can be found with this article Hist. 30, 451–510. online at doi:10.1016/j.cub.2012.04.021. 14. Olive, L.S. (1965). A Developmental Study of Guttulinopsis vulgaris (Acrasiales). Am. J. Bot. 52, 513–519. Acknowledgments 15. Raper, K.B., Worley, A.C., and Kessler, D. (1977). Observations on Guttulinopsis vulgaris and Guttulinopsis nivea. Mycologia 69, 1016– We wish to thank Fabien Burki for assembled EST clusters from the 1030. Rhizarian taxa Phyllostaurus sp., Astrolonche sp., sphaerica, 16. Dykstra, M.J. (1977). The possible phylogenetic significance of Spongospora subterranea, and brassicae. We wish to mitochondrial configurations in the acrasid cellular slime molds with thank the Department of Energy’s Joint Genome Institute (JGI) for access reference to members of the Eumycetozoa and fungi. Mycologia 69, to the genome data of (http://genome.jgi-psf.org/ 579–591. Aggregative Multicellularity in Rhizaria 1127

17. Erdos, G.W., and Raper, K.B. (1978). Ultrastructural Aspects of 2 Species of Guttulinopsis. Am. J. Bot. 65, 552–561. 18. Page, F.C., and Blanton, R.L. (1985). The Heterolobosea (Sarcodina, Rhizopoda), a New Class Uniting the and the Acrasidae (Acrasida). Protistologica (Paris) 21, 121–132. 19. Page, F.C. (1988). A New Key to Freshwater and Soil Gymnamoebae (Ambleside, Cumbria, UK: Freshwater Biological Association). 20. Patterson, D.J., Simpson, A.G.B., and Rogerson, A. (2000). Amoebae of uncertain affinities. In An Illustrated Guide to the , Volume 2, J.J. Lee, G.F. Leedale, and P. Bradbury, eds. (Lawrence, Kansas: Society of Protozoologists), pp. 804–826. 21. Hampl, V., Hug, L., Leigh, J.W., Dacks, J.B., Lang, B.F., Simpson, A.G.B., and Roger, A.J. (2009). Phylogenomic analyses support the monophyly of Excavata and resolve relationships among eukaryotic ‘‘supergroups’’. Proc. Natl. Acad. Sci. USA 106, 3859–3864. 22. Burki, F., Shalchian-Tabrizi, K., Minge, M., Skjaeveland, A., Nikolaev, S.I., Jakobsen, K.S., and Pawlowski, J. (2007). Phylogenomics reshuf- fles the eukaryotic supergroups. PLoS ONE 2, e790. 23. Philippe, H. (2000). Opinion: long branch attraction and protist phylogeny. Protist 151, 307–316. 24. Cavalier-Smith, T. (2002). The phagotrophic origin of eukaryotes and phylogenetic classification of Protozoa. Int. J. Syst. Evol. Microbiol. 52, 297–354. 25. Pawlowski, J. (2008). The twilight of Sarcodina: a molecular perspective on the polyphyletic origin of amoeboid protists. 5, 281–302. 26. Cavalier-Smith, T., and Chao, E.E.Y. (2003). Phylogeny and classifica- tion of phylum Cercozoa (Protozoa). Protist 154, 341–358. 27. Cavalier-Smith, T., Lewis, R., Chao, E.E., Oates, B., and Bass, D. (2008). Morphology and phylogeny of Sainouron acronematica sp. n. and the ultrastructural unity of Cercozoa. Protist 159, 591–620. 28. Cavalier-Smith, T., Lewis, R., Chao, E.E., Oates, B., and Bass, D. (2009). Helkesimastix marina n. sp. (Cercozoa: Sainouroidea superfam. n.) a gliding zooflagellate of novel ultrastructure and unusual ciliary behaviour. Protist 160, 452–479. 29. Lartillot, N., Lepage, T., and Blanquart, S. (2009). PhyloBayes 3: a Bayesian software package for phylogenetic reconstruction and molecular dating. Bioinformatics 25, 2286–2288. 30. Stamatakis, A. (2006). RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22, 2688–2690.