The three core phyla of the Deuterostomia are the echinoderms, the hemichordates (enteropneusts and pterobranchs), and the chordates (urochordates, cephalochordates, and craniates plus vertebrates; Figure 3). In addition, recent molecular phyloge-netic research has added the enigmatic and rather simple worm X. bocki to the deuterostomes as well (Bourlat et al., 2003). Of the three main bilaterian clades, the high-level phylogeny of the deuteros-tomes is currently best understood. This section will describe deuterostome phylogeny in detail, including an account of high-level vertebrate phylogeny.
1.02.3.3.1 Echinodermata, Hemichordata, and Xenoturbella Currently, a sister group relationship of the echinoderms and hemichordates in a clade Ambulacraria is robustly supported by both molecular and morphological evidence (Smith et al., 2004; Figure 3). Previous hypotheses often considered Hemichordata not to be monophyletic, with the enteropneusts being the sister group to the chor-dates, and the pterobranchs diverging more basally within the deuterostomes. Recent molecular and morphological phylogenetic analyses established the monophyly of the Hemichordata and its placement as a sister group to the Echinodermata (Cameron, 2005; Zeng and Swalla, 2005). A reinterpretation of morphological characters in these groups, as well as in the lophophorates that were previously considered to be deuterostomes, supports this hypothesis.
The exclusively marine echinoderms contain some of the most familiar invertebrates, such as sea stars and sea urchins. Their pentaradial symmetry, calcareous endoskeletal elements, and their elaborate coelomic water vascular system are diagnostic features of the phylum. Although the living species are classified into five well-demarcated groups (asteroids or sea stars, echinoids or sea urchins, ophiuroids or brittle stars, crinoids, feather stars or sea lilies, and holothuroids or sea cucumbers), the morphological diversity of extinct echinoderms yields a further 18 or so distinct echinoderm taxa.
The phylogeny of the extant echinoderms is far from established, but some provisional conclusions may be drawn on the basis of currently available morphological and molecular data sets. First, the crinoids are the sister group to the remaining taxa, which are grouped in a clade Eleutherozoa. Within the Eleutherozoa, the asteroids are the most likely sister group to a clade comprising the ophiuroids, echinoids, and holothuroids. Within this clade there is robust support for a sister group relationship between the echinoids and the holothuroids in a clade Echinozoa. In contrast, support for a sister group relationship between the clade Echinozoa and the ophiuroids is not very strong.
The Hemichordates comprise the enteropneusts, or acorn worms, and the pterobranchs. Enteropneusts and pterobranchs are very easily distinguishable. Enteropneusts are large (typically between 20 and 25 cm), free-living, bottom-dwelling worms equipped with a bulbous anterior proboscis (prosome) that is clearly separated from the long posterior trunk (metasome) by a collar (mesosome) at the anterior end of which the mouth opens. Pterobranchs, on the other hand, are tiny sessile, colonial animals that carry a prominent crown of ciliated tentacles. Like the enteropneusts the ptero-branchs possess a tripartite body, with coelomic cavities in each of the three subdivisions of the body, with an anterior part called the oral shield (prosome that secretes the tubes in which they live), a short middle collar (mesosome bearing the tentacles), and a sac-like trunk (metasome).
In several early morphological phylogenetic analyses Hemichordata was assumed to be monophyletic, but this assumption was not confirmed by most morphological phylogenetic analyses that separated the enteropneusts and pter-obranchs as individual taxa. Hemichordate monophyly is confirmed, however, by phylogenetic analyses of SSU and LSU sequences, and recent reinterpretation of hemichordate morphology confirms two potential hemichordate synapomorphies that had previously been hypothesized: possession of a stomochord (anterior dorsal extension of the pharynx wall into the prosome) and presence of mesocoelomic ducts that connect the coelomic cavities of the mesosome with the outside (Cameron, 2005; Ruppert, 2005).
The monophyly of the pterobranchs is not in doubt, as they share such unique features as the secretion of tubes, or coenecia, with their prosome, or cephalic shield, and the possession of tentacles on the mesosome. However, SSU and LSU evidence and phylogenetic analysis of combined molecular and morphological data suggest the possibility that the enteropneusts are paraphyletic with respect to the pterobranchs. Alternatively, pterobranchs may be the sister group to a monophyletic Enteropneusta. If pterobranchs have evolved from within the enter-opneusts, this implies several rather drastic changes in body architecture probably associated with miniaturization. These changes would include simplification of the pterobranch gill skeleton, and simplification of the nervous system from possession of the well-developed dorsal nerve cord or tube in the collar of enteropneusts, to a simpler neural ganglion in the collar region of pterobranchs.
The Ambulacraria hypothesis, according to which echinoderms and hemichordates are sister taxa, has significant implications for understanding deuterostome evolution in general, and the origin of the chordate body plan in particular, because the origin of the chordate nervous system, the notochord, and pharyngeal gill slits has frequently been discussed with respect to either enteropneust anatomy or hemichordate and echinoderm larval morphology.
For example, the classic hypothesis proposed for the origin of the chordate nervous system is Walter Garstang's auricularia or dipleurula hypothesis, later amended and elaborated by other authors, most recently Thurston Lacalli. At the core of this hypothesis is the derivation of the chordate dorsal nerve tube from the fused ciliary bands of a dipleur-ula-type larva, as is commonly found in the life cycles of both echinoderms and enteropneusts. However, monophyly of the Ambulacraria implies that such a larval form is a possible synapomorphy of the echinoderms and hemichordates, and is not a plesiomorphy for the chordates (Lacalli, 2005).
The recent addition of the small, morphologically simple worm X. bocki to the morphologically much more complex coelomate deuterostomes also has potentially important implications for understanding deuterostome character evolution. Molecular evidence indicates a possible phylogenetic position for Xenoturbella as the sister group to the Ambulacraria (Bourlat et al., 2003; Figure 3). Since Xenoturbella lacks complex morphology shared by the other deuterostomes, including coelo-mic cavities and a gut with both mouth and anus (it has a blind-ending gut), it raises the interesting question of whether Xenoturbella may have become secondarily simplified, or whether it has retained its morphological simplicity from the root of the Bilateria, and as indicated by the acoelomorphs discussed above.
1.02.3.3.2 Chordata: Urochordata and Cephalo-chordata The monophyly of Chordata (Urochordata, Cephalochordata, and Craniata) is widely accepted (Figure 3). The urochordates, or tunicates, have long been considered to be the sister group of the clade Cephalochordata plus Craniata, with the latter two taxa also being sister groups. Chordate monophyly is principally supported by the presence of a notochord, or chorda, a dorsal neural tube, longitudinal muscles along the noto-chord, or its derivatives, the presence of an endostyle that secretes a mucus filter used in feeding, and the presence of special cerebral sensory organs in the form of optic and otic receptors. It should be noted that some of these features are only present in the tadpole larvae of the tunicates, not in the sessile adults (with the possible exception of the larva-ceans, or appendicularians, which are likely derived from more typical ascidian ancestors). The body plan of adult tunicates is entirely different from those of the cephalochordates and craniates.
Interestingly, until recently, available molecular phylogenetic evidence did not yield any convincing support for the monophyly of the chordates. The precise placement of the urochordates turned out to be a particularly vexing issue. SSU and LSU sequences have variously suggested urochordates to be the sister group to a monophyletic clade of Cephalochordata plus Craniata, sister group to Ambulacraria, or sister group to all of the deuteros-tomes. More recent phylogenetic studies based on the simultaneous analysis of a larger number of different genes have instead started to provide tentative support for an unexpected sister group relationship between vertebrates and tunicates
(Blair and Hedges, 2005; Delsuc et al., 2006). If this turns out to be correct, it will have important consequences for understanding the origin of the vertebrates, for which cephalochordates are commonly interpreted as a stand-in of the last common vertebrate ancestor. However, straightforward comparisons between urochordates and vertebrates are very difficult, especially because of the strong modifications of the many forms that are sessile as adults.
The vast majority of tunicates are represented by the familiar sessile ascidians or sea squirts. The remaining species comprise the pelagic thaliaceans (salps, pyrosomids, and doliolids), and the appendi-cularians or larvaceans. A highly complex and unique cuticular exoskeleton known as the tunic, which includes free cells, has given the phylum its name, although it is not present in all tunicates. A large pharynx or branchial basket is another conspicuous organizational feature of urochordates. A similar pharynx can also be identified in the cepha-lochordates and the ammocoetes larva of lampreys. The characteristic tadpole larva with a bulbous body and slender muscular tail is also unique for urochordates. In fact, larvaceans look somewhat like urochordate tadpole larvae during their entire life cycle. It has been hypothesized that larvaceans have evolved by truncating development so that the original tadpole larva has now become the definitive adult (a phenomenon known as pedomorphosis). Recent phylogenetic evidence deriving both the pelagic larvaceans and thaliacians from within a paraphyletic clade of ascidians provides some support for this hypothesis (Zeng and Swalla, 2005).
The cephalochordates, or lancelets (also known as amphioxus), are widely accepted to be the sister group to the craniates. Lancelets are the most vertebrate-like of all invertebrates, and they resemble little fish in their morphology. The segmented nature of their body musculature can readily be distinguished, and shows a striking similarity to the segmentally arranged muscle blocks or myo-meres found in vertebrates. These muscle blocks are derived from coelomic (somitic) pouches. Moreover, vertebrates and cephalochordates share the attainment of a certain degree of brain complexity reflected in both its ultrastructure and the expression of developmental regulatory genes, and the elaboration of special sense organs, principally an olfactory organ (corpuscles of Quatrefages in lancelets) marked by the expression of a developmental regulatory gene also expressed in craniate ectodermal placodes. However, the evolutionary significance of these features remains unclear when, as discussed above, urochordates indeed prove to be more closely related to vertebrates.
1.02.3.3.3 Craniata and Vertebrata The last decade has witnessed remarkable progress in resolving the major phylogenetic relationships within the extant craniates and vertebrates (Rowe, 2004; Figures 4-6). Especially the application of molecular evidence from a variety of sources has contributed significantly to the emergence of the current consensus.
The basal-most extant groups within the Craniata are the hagfishes (Myxinoida) and lampreys (Petromyzontida), which together are referred to as Cyclostomata, representing the jawless agnathans. Whether Cyclostomata is monophyletic or paraphy-letic with hagfishes as the sister group to the Vertebrata, including the lampreys, is currently unclear. Different analyses contradict each other. For the purpose of this article I accept cyclostoma-tan paraphyly as a working hypothesis, with lampreys as the sister group to the Gnathostomata, which comprises all living jawed vertebrates (Figure 4).
The basal-most split within Gnathostomata is between the sister groups Chondrichthyes (cartilaginous fishes including sharks and rays) and Osteichthyes (bony fishes and tetrapods). Within the Osteichthyes there is a basal split between the Actinopterygii (ray-finned fishes, comprising all living fishes except the coelacanths and lungfishes) and the Sarcopterygii (lobe-finned fishes: coelacanths, lungfishes, and tetrapods). Although it has commonly been thought that the coelacanths are the nearest relatives of the Tetrapoda, accumulating molecular evidence instead supports a basal split within the Sarcopterygii between the sister groups Actinistia (coelacanths) and Choanata (the lung-fishes and tetrapods). The lungfishes (Dipnoi) are the sister group to the Tetrapoda, or terrestrial vertebrates (Figure 4).
An important caveat may obtain here. Although the basal vertebrate relationships described above are widely accepted, a recently published phyloge-netic analysis of complete mitochondrial genomes and 18S and 28S ribosomal sequences instead suggested that a fish is a fish, and a tetrapod is a tetrapod; the tetrapods were a sister clade to a clade comprising all gnathostome fishes (Arnason et al., 2004). This study provided suggestive evidence that the traditional phylogeny in which tetrapods evolve from within a paraphyletic group of fishes is an artifact of rooting the phylogeny within the gnathostomes, either with bony fishes or cartilaginous fishes. This procedure may already assume paraphyly of gnathostome fishes. Only when phylogenies are rooted instead with a non-gnathostome outgroup, such as hagfish or lamprey, can the hypothesis of the monophyly of gnathos-tome fishes be really tested. In this case a monophyletic clade of gnathostome fishes, including lungfishes and coelacanths, is recovered. Yet, nothing is ever unambiguous in phylogenetics, and this result was again contradicted by more recent studies published between first submission and revision of this article, which on the basis of phylogenetic analysis of a large number of nuclear protein-coding genes showed that gnathostome fishes are paraphyletic with respect to tetrapods (Blair and Hedges, 2005).
Extant tetrapods comprise three clades: Amphibia, Reptilia, and Mammalia (Figure 4). The amphibians (anurans, such as frogs; caudatans, such as salamanders; and caecilians) are the sister group to the Amniota, comprising the extant sister groups Reptilia and Mammalia (Figures 4 and 5). Although long considered to be a paraphyletic group, the reptiles are currently recognized as the monophy-letic sister clade of the Synapsida, which includes the modern mammals plus extinct stem taxa. Reptilia comprises two major clades of extant reptiles: Archosauria and Lepidosauria (Lee et al., 2004; Figure 5). Monophyly of Archosauria is well established, and the clade comprises the extant crocodilians and birds, as well as the extinct pterosaurs and dinosaurs.
Living lepidosaurs are the tuataras, lizards, and snakes. Lizards and snakes are grouped together in a clade Squamata, the monophyly of which is well supported, but the internal relationships of which remain uncertain (Lee, 2005; Vidal and Hedges, 2005). Monophyly of the Lepidosauria is generally believed to be well supported as well, but several recent molecular phylogenetic analyses of nuclear genes contradict morphological and mitochondrial sequence support for a monophyletic Lepidosauria, suggesting instead a paraphyletic Lepidosauria with squamates as sister group to all remaining extant reptiles.
The perennially problematic phylogenetic position of turtles has not yet been convincingly resolved (Figure 5). On the basis of morphological and paleontological evidence turtles were either allied with a group of extinct marine lepidosaurian reptiles, or were placed outside all other living reptiles. However, the application of molecular sequence data has yielded no support for any of these hypotheses. Instead, these data surprisingly suggest that turtles are closely related to archosaur-ian reptiles (Rest et al., 2003; Lee et al., 2004). So far no anatomical support for this relationship has been recovered, and a recent morphological phylo-genetic study places turtles as a sister group to Lepidosauria (Hill, 2005).
The congruence of diverse molecular evidence has recently generated a new consensus of mammalian phylogeny that is significantly at odds with traditional ideas established on the basis of morphological and paleontological evidence (Murphy et al., 2004; Springer et al., 2004). The basal divergences among the mammals are in agreement with traditional ideas. The Monotremata (duck-billed platypus, echidnas) and their fossil stem group are together called Protheria, and Protheria is the sister group to the marsupial and placental mammals and their stem groups. The marsupials and their stem group are called Metatheria, while the Placentalia and their stem group are called Eutheria. Metatheria and Eutheria are sister groups within a clade Theria.
The emerging molecular picture of placental mammal phylogeny is strikingly at odds with traditional ideas (Figure 6). Four main clades are recognized: Xenarthra, Afrotheria, Euarchontoglires, and Laurasiatheria. The latter three clades were not recognized previously. Euarchontoglires and Laurasiatheria are sister clades with northern hemisphere origins. Xenarthra is sister group to this superclade, and the divergence between Afrotheria and the remaining pla-cental mammals represents the basal-most split. The xenarthrans and afrotherians had a southern hemisphere origin.
Afrotheria includes morphologically very disparate forms such as elephants, hyraxes, manatees, golden moles, elephant shrews, tenrecs, and aard-varks. Xenarthra includes sloths, armadillos, and anteaters. Euarchontoglires unites rodents, lago-morphs (rabbits, pikas) together with flying lemurs, tree shrews, and primates. Laurasiatheria includes bats, certartiodactyls (artiodactyls and cetaceans), perissodactyls, carnivores, pangolins, and insectivores such as moles, hedgehogs, and shrews. One of the most striking implications of the new placental mammal phylogeny is the extent of parallel morphological evolution between afrotherians and laurasiatherians. Both clades show independent radiations of similar adaptive forms, including mole-like animals (golden mole and mole), hedgehog-like animals (tenrec and hedgehog), shrew-like animals (elephant shrew and shrew), anteaters (aardvarks and pangolin), and fully aquatic forms (manatees and dolphins).
The implications of the new mammal phylogeny for understanding phenotypic evolution are enormous, and many received wisdoms need to be freshly scrutinized. Humans turn out to be more closely related to mice than to cows or dogs, ungulates or hoofed mammals do not form a clade, and the traditional order Insectivora is equally spread out across the mammalian tree.
The enormous range of morphological variation characterizing clades such as Afrotheria also poses a real challenge to those who wish to reconstruct their ancestral phenotype. For example, so far just a single morphological synapomorphy has been proposed for the clade Afrotheria (Carter et al., 2006), a situation reminiscent of the difficulty of establishing morphological synapomorphies for the large invertebrate clade Lophotrochozoa (see below). Moreover, current evidence suggests that morphologically very distinct taxa, such as hyraxes, elephants, and sirenians, and which form a clade within Afrotheria, may have radiated very rapidly, establishing their distinctive body plans in a short amount of time (Nishihara et al., 2005). Evidently, a satisfactory synthesis of the molecular phylogeny with morphology and fossil evidence is the daunting challenge still before us.
Was this article helpful?