Acoel Flatworms: Earliest Extant Bilaterian Metazoans, Not Members of Platyhelminthes

See allHide authors and affiliations

Science  19 Mar 1999:
Vol. 283, Issue 5409, pp. 1919-1923
DOI: 10.1126/science.283.5409.1919

This article has a correction. Please see:


Because of their simple organization the Acoela have been considered to be either primitive bilaterians or descendants of coelomates through secondary loss of derived features. Sequence data of 18S ribosomal DNA genes from non–fast evolving species of acoels and other metazoans reveal that this group does not belong to the Platyhelminthes but represents the extant members of the earliest divergent Bilateria, an interpretation that is supported by recent studies on the embryonic cleavage pattern and nervous system of acoels. This study has implications for understanding the evolution of major body plans, and for perceptions of the Cambrian evolutionary explosion.

“Since the first Metazoa were almost certainly radial animals, the Bilateria must have sprung from a radial ancestor, and there must have been an alteration from radial to bilateral symmetry. This change constitutes a most difficult gap for phylogeneticists to bridge, and various highly speculative conjectures have been made” (1, p. 5). So began Libbie Hyman's discussion on the origin of bilaterian Metazoa, and despite a century of morphological studies and a decade of intensive molecular work, the nature of the simplest bilaterian animal remains elusive (1, 2). Paleontological and molecular data indicate that most bilaterian phyla appeared and diversified during the Cambrian explosion (3, 4). Three main clades emerged—the Deuterostomia, the Ecdysozoa, and the Lophotrochozoa (5), although their branching order is unresolved. The acoel flatworms, traditionally classified as an order of the Platyhelminthes, are perhaps the simplest extant members of the Bilateria and have been viewed as either basal metazoans that evolved from ciliate protozoans (“syncytial or ciliate-acoel theory”) (6) or a direct link between diploblasts and triploblasts (“planuloid-acoeloid theory”) (1, 7). However, the lack of complexity has also been interpreted as a loss of derived features of more complex ancestors (“archicoelomate theory”) (8).

The proposed metazoan phylogenetic trees that include acoels have shown them to branch after the diploblasts, indicating that they are considered primitive triploblastic animals (9–11). However, all 18Sribosomal DNAs (rDNAs) from acoels that have been sequenced so far show rates of nucleotide substitution that are three to five times the rates of most other Metazoa (10), resulting in the long-branch attraction effect in which rapidly evolving taxa cluster and branch together artifactually at the deepest base of the trees (12). We examined the relationship of the Acoela to other metazoan taxa by sequencing complete 18S rDNAs (13) from 18 acoel species (14). In addition, we sequenced the 18S rDNA of the catenulid Suominasp. and the nemertodermatid Meara sp. as additional representatives of basal orders of Platyhelminthes thought to be closely related to acoels. The 18S ribosomal gene was chosen because of the large number of sequences available in the molecular data banks (GenBank and European Molecular Biology Laboratory) for representatives of the entire animal kingdom.

To avoid the long-branch effect, we broadly sampled the Acoela to find species that have normal rates of nucleotide substitution (non–fast-clock species). As representatives of most animal phyla a wide range of metazoan species were selected from the data banks (Table 1) and their sequences aligned and compared with those of acoels (15). A preliminary phylogenetic analysis (by the neighbor-joining method) showed that all 18 acoels form a very clear monophyletic group that branches at the base of the triploblasts. As expected (10), inclusion of the long-branch acoels leads to several inconsistencies in tree topology. Therefore, to select those taxa with uniform rates of change, we first performed a relative rate test (16) comparing all the species by pairs with the diploblast species as reference taxa. Because extremely long branches characterize most acoel species, only the four with shortest branches were included in the test. Only one acoel species (Paratomella rubra) passed the test, and one other (Simplicomorpha gigantorhabditis) came very close (Table 1). Although the latter was not included in subsequent analyses reported here, very similar results were obtained when both species or a single one (Paratomella rubra) was used. Of 74 bilaterian species tested (including the four acoels), 57 passed the test. Subsequent analyses were performed with only these 57 species that demonstrated uniform and comparable rates of evolution (representing 21 phyla) plus the four diploblasts representing three phyla. The second step in the analysis was to determine the phylogenetic content of the data resulting from this selection. Two tests were carried out. A plot of the observed (total, transitions or transversions) versus inferred number of substitutions (4,17) showed that, although the curves tend to level off (Fig. 1A), they do not reach a plateau, meaning that the sequences studied are only moderately mutationally saturated. From a likelihood-mapping analysis (18) 81.5% of quartets had resolved phylogenies and only 10.7% of all quartet points were in the star-tree region (Fig. 1B), indicating that the rDNA data contain a reasonably high degree of phylogenetic information.

Figure 1

Phylogenetic content of the data. (A) Substitution saturation curve. The y axis shows the frequency of observed differences between pairs of species sequences determined with MUST (4, 17), and the xaxis shows the inferred distance between the same two sequences determined by maximum likelihood (ML) with PUZZLE v. 4.0 (38). Each dot thus defines the observed compared with the inferred number of substitutions for a given pair of sequences. The resulting curve lies between the diagonal line (no saturation) and a horizontal plateau line (full saturation), which means that the data set is only moderately saturated [for further information see (4, 17)]. (B) Likelihood mapping analysis (18) of the data set, represented as a triangle. Values at the corners indicate the percentages of well-resolved phylogenies for all possible quartets (18), and values at the central and lateral regions are percentages of unresolved phylogenies. The cumulatively high percentage (81.5%) from the corner values indicates the data set is phylogenetically informative.

Table 1

List of species included in this study, GenBank accession numbers, and result of the relative rate test (rrt). The names of the 61 species finally selected for analysis are in bold. Ph., phylum; O., order; Cl., class.

View this table:

We next built a tree using maximum likelihood (19). The best tree that we found is shown in Fig. 2. In this tree, Deuterostomia, Ecdysozoa, and Lophotrochozoa (5) form monophyletic groups. Interestingly, the acoelomate and pseudocoelomate groups cluster at the base of the Ecdysozoa and Lophotrochozoa. Most importantly, the tree shows the acoels as the first offshoot after the diploblasts. However, the nemertodermatids, an order of Platyhelminthes usually classified as the sister group of the Acoela and forming the Acoelomorpha (20,21), and here represented by the single species that passed the relative rate test, group within the bulk of the Platyhelminthes rather than with the acoels. On the other hand, both catenulids cluster at the base of the Platyhelminthes. All alternative hypotheses concerning the relationships between acoels and other platyhelminths (20,22–24) and their position within the Bilateria (10) were also compared by the Kishino-Hasegawa test and all were significantly poorer than the phylogeny obtained originally. The robustness of the internal branch separating acoels from the rest of bilaterians was further evaluated by the four-cluster likelihood mapping method (25) and resulted in 100% support for this branch.

Figure 2

Diagrammatic representation of the best 18S rDNA-based maximum-likelihood tree of 61 metazoan species (bold names are in Table 1) with homogeneous rates of nucleotide substitution. The final matrix included 1181 sites (584 variable and 383 informative under parsimony); log ln = 11,862. The number 100 on the branch separating acoels from the rest of triploblasts represents the percentage of support to that branch obtained by the four-cluster likelihood mapping (25). The tree was obtained with fast DNAml (19). It illustrates the relationships of the Acoela (bold, upper case) and the rest of the Platyhelminthes (bold, lower case) to the rest of the Metazoa. The general topology of the tree defines three main bilaterian phylogenetic groups: Deuterostomia, Lophotrochozoa (including Platyhelminthes and Gastrotricha as basal phyla), and Ecdysozoa (with Priapulida and Kinorhynchia as basal phyla). The position of the Acoela renders the Platyhelminthes polyphyletic, whereas the Nemertodermatida (underlined) appears buried within the bulk of Platyhelminthes. For taxa and species names, see Table 1, the complete tree with all the species names is available in the supplementary material

Because the position of the acoels might be due to the most variable sites of the alignment, we removed them from the whole data set (26); the acoels still appeared at the base of the trees, although the phylogenetic signal within the triploblasts almost faded away. Alternatively, the sequence regions that show the highest variation among acoels might represent noisy data that separate them from the rest of the Bilateria. To test this idea, we aligned the 18 acoel sequences, found their most variable positions, and removed the latter from the 61-species alignment (27). Again, this resulted in the acoels on a shorter branch at the base of the bilaterian tree (the three trees obtained in both tests are available as supplementary material Finally, because some important phyla such as Chaetognatha, Acanthocephala, Gnathostomulida, Mesozoa, and Nematoda did not pass the relative rate test and were not included in the maximum-likelihood analyses, the four-cluster likelihood mapping was used again to test the position of these groups against acoels and the rest of the triploblasts. In all cases, acoels and diploblasts cluster together (Table 2). Importantly, of all the phyla tested, some of those previously proposed as “primitive” bilaterians (Mesozoa, Nematoda, and Gnathostomulida), always cluster with the triploblasts.

Table 2

Four-cluster likelihood mappings to test acoel position against fast-clock phyla. Four-cluster likelihood mappings (18) were performed arranging species into three groups: diploblasts (D), acoels (A), triploblasts (T), and a fourth group (X) taken from each of the phyla to be tested. If the phylum under test is more basal than the acoels, it should cluster with high support with the diploblasts. Conversely, if acoels are more basal the phylum under test should cluster with the triploblasts. Results show that acoels cluster more closely to the diploblasts than all other triploblasts.

View this table:

Our analyses clearly indicate that acoels are not members of the phylum Platyhelminthes, but occupy a key position in the Metazoan tree of life, most likely as the earliest branch within the bilaterian clade that left extant descendants. The monophyly of Platyhelminthes has been criticized (23, 28, 29) because of the weakness of the synapomorphies on which it is based (22, 24): multiciliation of epidermal cells, the biciliary condition of the protonephridia, and the lack of mitosis in somatic cells. In contrast, the Acoela have a characteristic set of well-defined apomorphies: a network of ciliary roots of epidermal cells, tips of the cilia with a distinct step, lack of extracellular matrix, absence of protonephridia, and, most importantly, the duet spiral cleavage. The first three features are usually considered to be derived (22, 24); the 18S molecular data, however, suggest a different interpretation for the other two. The lack of protonephridia in acoels may be a plesiomorphic feature (29). The Acoela exhibit duet spiral cleavage, in contrast to the quartet pattern that characterizes the Spiralia and some turbellarian Platyhelminthes. However, acoel cleavage is actually more bilateral than spiral (30), suggesting that duet cleavage and typical quartet cleavage are not related. Moreover, all spiralian embryos have both ecto- and endomesoderm and exhibit determinative development, whereas acoel embryos generate only endomesoderm (30) and are highly regulative (31); the latter two features are considered to be ancestral. Most diploblastic and several triploblastic phyla exhibit a radial cleavage pattern; thus, it is more parsimonious to assume that the first bilaterian also had radial cleavage (32). This evidence supports our proposed phylogenetic tree in which the acoels branch before the Cambrian radiation from unknown bilaterian ancestors with radial cleavage and suggests that duet cleavage and quartet spiral cleavage arose independently from an ancestral radial pattern. The structure of the nervous system also indicates that the acoels are not related to the other platyhelminths. Most Platyhelminthes have a bilobed brain with neuropile surrounded by nerve cells and two main longitudinal nerve cords with commisures making an orthogon (33). In contrast, the nervous system of acoels comprises a simple brain formed by clusters of nerve cells that lack a neuropile, and a variable number of longitudinal nerve cords that do not make an orthogon (34).

The 18S rDNA sequences, embryonic cleavage patterns and mesodermal origins (30), and nervous system structure data (34) support the position of the Acoela as the earliest branching Bilateria (Fig. 2) and the polyphyly of the Platyhelminthes. This argues for an extended period before the Cambrian within which different bilaterian lineages may have originated, with the acoels being the descendants of one of these lineages. This interpretation is supported by recent data on protein sequence divergence (35). Direct development, which characterizes all extant acoels, as opposed to the biphasic life cycle with a larval stage and a benthic adult (36), probably represents the ancestral bilaterian condition [see (37), for a recent discussion]. Our findings suggest that the Acoela (or Acoelomorpha if the Nemertodermatida are shown to remain as their sister group) should be placed in their own phylum.

  • * To whom correspondence should be addressed. E-mail: bagunya{at}


View Abstract

Navigate This Article