The Essential Role of "Minor" Phyla in Molecular Studies of Animal Evolution(1).

INTRODUCTION

What are "minor" phyla? Minor phyla are often referred to as enigmatic or problematic, are usually of uncertain affinity, and generally are treated superficially in invertebrate texts. Minor phyla are considered to be of little consequence to mainstream animal evolution, usually because they are not well represented in present day macrofauna (see e.g., Simonetta and Conway Morris, 1991). This is a major error, since the modern day or paleontological prominence of a taxon does not necessarily reflect its role, or the role of its ancestors in the metazoan radiation. If we use the questionable definition of a phylum as a taxon with a distinctly unique body plan and leave aside the requirement of monophyly, then minor phyla represent the majority of nature's experimentation with animal body plans. In contrast, the "major" phyla are a small number of groups that are prominent among modern macrofauna and are composed of annelids, arthropods, chordates, cnidarians, echinoderms, molluscs, and perhaps platyhelminths.

Several kinds of minor taxa are important to this discussion. Some groups such as mesozoans and placozoans have been considered to be representatives of the stem lineage leading to triploblastic animals (see Ax, 1996). Other groups have uncertain affinities, appear to have simple body plans, and are generally small in size. Many of these historically have been lumped together into the "Aschelminthes" (e.g., Rotifera, Acanthocephala, Nematoda, Nematomorpha, Priapulida, Kinorhyncha, Gastrotricha), based on the dubious assessment that each possesses a pseudocoelom. Other minor phyla appear to be sister taxa to larger and more well defined groups. For example, echiurans, sipunculids, pogonophorans and vestimentiferans have long been considered to be protostomes, possibly allied with annelids or molluscs. Then there are the lophophorates, a group of three phyla (Phoronida, Brachiopoda, Bryozoa) that have been placed either intermediate between protostomes and deuterostomes (see Willmer, 1990), as deuterostomes (Brusca and Brusca, 1990) or have been proposed to be polyphyletic with some being protostomes and others deuterostomes (Nielsen, 1995). Entoprocts have been associated with molluscs (Bartolomaeus, 1993), with aschelminths (Brusca and Brusca, 1990), or allied with ectoprocts (Nielsen, 1995). Onychophorans are usually allied with annelids or arthropods, while tardigrades have been linked to both "aschelminths" (Ruppert and Barnes, 1994) and arthropods (Brusca and Brusca, 1990) at various times. Some groups such as chaetognaths or the newly discovered Cycliophora (Funch and Kristensen, 1995; Funch, 1996) have only vaguely defined relationships to other taxa.

In general, the perception of basic evolutionary relationships of the major phyla have remained similar to that shown in Figure 1A (acoelomate/coelomate tree) with tribloblastic animals branching from diploblastic animals, leading to an acoelomate ancestor similar to modern day flatworms which then split into the pseudocoelomate ("Aschelminthes") and eucoelomate lineages. The eucoelomates branched from a common eucoelomate ancestor into protostomes and deuterostomes. This succession of acoelomates--pseudocoelomates--coelomates goes back to Hyman (1951, see p. 23), although it was originally not intended to precisely reflect phylogeny. A more modern variation on that theme is shown in Figure 1B (protostome/deuterostome tree), with a protostome/deuterostome split early on, and flatworms as basal protostomes. In this scheme, "aschelminth" taxa are usually ignored.

[Figure 1 ILLUSTRATION OMITTED]

The phylogeny within the protostomes has been locked into place by the "obvious" relationship between arthropods and annelids as segmented Articulata, sometimes extended to include the molluscs as a third group. A few other phyla (tardigrades and onychophorans) have been invoked as modern day representatives of ancestral forms that were transitional between annelids and arthropods. All the other protostome phyla except perhaps molluscs have essentially played a secondary role to these "mainstream" protostomes in phylogeny for over a century because it has been difficult to integrate most minor groups into overall phylogenetic hypotheses. Part of the reason for this is the difficulty in finding appropriate characters in many of the minor phyla whose members are often tiny, with simple bauplans, of little economic importance and therefore under-studied.

The advent of ultrastructural studies, coupled with cladistic analysis and molecular phylogenetic methods, have dramatically improved our ability to incorporate more minor taxa into phylogenetic hypotheses. Ultrastructural studies suggested that body cavities are more plastic than previously thought and perhaps not a good character for phylogenetic studies (Ruppert, 1991; Kristensen, 1995). They also suggested that aschelminths are polyphyletic, but could not relate them to other phyla because they lacked appropriate character sets (Ruppert, 1991; Kristensen, 1995; Wallace et al., 1996). In a sense, molecular phylogenetic methods democratized phylogeny, providing the theoretical and practical framework to integrate nearly any taxon into phylogenies, regardless of how understudied or unknown it was previously.

The 18S rRNA gene and unequal rate effects

Although there are likely to be other genes better suited for metazoan phylogenetic studies at the level of the phylum, the 18S rRNA gene has been the gene of choice because of the large number of sequences available and because its properties are well known (Hillis and Dixon, 1991; Dixon and Hillis, 1993). The major problems with molecular phylogeny have been the development of adequate methods, the ability to acquire sufficient data and the recognition of the limitations of molecular analysis (e.g., see Maley and Marshall, 1998). Unequal rate effects are one important source of error in the phylogenetic analysis of molecular data (Felsenstein, 1978; Hillis et al., 1994; Aguinaldo et al., 1997) that is often ignored. In unequal rate effects, genes of some taxa (fast evolving) have sequences that have much higher substitution rates than in other taxa (slow evolving). A combination of alignment errors and problems with tree making algorithms cause taxa with long branches to be…