| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Zoological Museum, The Natural History Museum of Denmark, University of Copenhagen, Universitetsparken 15, DK-2100 Copenhagen, Denmark
E-mail: cnielsen{at}snm.ku.dk
 |
Abstract |
|---|
 TOP Abstract IntroductionMorphology and Feeding Biology...Early Eumetazoan RadiationTypes of Information About...The Origin of Indirect...Literature Cited |
|---|
We have irrefutable proof in support of the pelagic larval life as an original feature... . This is so well documented that it would be superfluous to enter into the matter more closely, were it not for its great importance and the fact that there exist authors, who still assert that direct development is the original condition (Jägersten, 1972, p. 3).
|
Introduction |
|---|
|
TOPAbstractIntroductionMorphology and Feeding Biology...Early Eumetazoan RadiationTypes of Information About...The Origin of Indirect...Literature Cited |
|---|
The newly advanced theory for the origin of the eumetazoans (Nielsen, 2008b) proposes that the ancestral eumetazoan which Haeckel (1874) called gastraea did not evolve directly from a blastaea, but from sexually mature larvae of a homoscleromorph-like sponge with a pelago-benthic life cycle. This is compatible with the majority of the morphological information (from embryology, ultrastructure studies, and molecular biology), and it provides a function-based explanation of the origin of the gut with extracellular digestion. It is in accordance with some of the results from molecular phylogenetic studies (e.g., Sperling et al., 2007; Srivastava et al., 2008) and with results from studies of, for example, microRNAs (Grimson et al., 2008). One recent molecular study places the ctenophores at the base of the Metazoa (Dunn et al., 2008), and other studies show a monophyletic clade of sponges and cnidarians (and sometimes ctenophores) as the sister group of the Bilateria (e.g., Wang and Lavrov, 2007; Dunn et al., 2008) These phylogenies imply several major homoplasies or losses of important morphological characters such as nervous system and basal membrane, and therefore they appear less plausible. The new support for a modified gastraea theory with a holopelagic ancestor of the eumetazoans invites a reassessment of the various theories for the origin of biphasic life cycles comprising pelagic, planktotrophic larvae and benthic adults. In this connection it should be stressed that holopelagic just means free-swimming as opposed to benthic (vagile or sessile). It does not imply that the larvae were swimming in the open ocean: early planktotrophic organisms may well have lived in the near-bottom layers, which were probably richer in food particles than the open oceans; in fact the difference between the two types of life style is small.
At the outset, I want to emphasize that "tree thinking" (Baum et al., 2005) is essential for all discussions on deep animal phylogeny, but I further find that three principles should be adhered to in all speculations: (1) every proposed ancestor should have been able to feed, and its feeding mechanism should be considered; (2) every evolutionary step between the proposed ancestors should likewise have been able to feed; and (3) the adaptational value(s) of the proposed evolutionary changes should be considered and, if possible, explained.
The invertebrates show an immense variety of developmental types. Not unexpectedly, this has resulted in considerable inconsistencies in the terminology used for the developmental types—see, for example, McEdward and Janies (1993). There seem to be almost as many definitions of "larva" and "metamorphosis" as there are authors (see, for example, Bishop et al., 2006), so in accordance with the very sound views of Hadfield (in the same paper) I here give my definitions.
Some animals develop directly from the late embryonic stage to the adult via a juvenile. Examples are nematodes, planarians, and fishes, which all hatch from the fertilization membrane almost as miniature adults. However, most animals have indirect development, with an ontogenetic stage between the embryo and the juvenile, which deviates from the adult not only in size but also in structure and feeding biology and often in habitat. This stage is called a larva. The structural difference lies usually in the presence of special larval organs, such as locomotory or feeding structures, that are lost in the transition from the larva to the juvenile. This transition is known as metamorphosis (Nielsen, 2000). The larval organs may be inconspicuous, as in the apparently directly developing nemertine Carinoma (Maslakova et al., 2004), whose larva has trochoblasts, as indicated by the cell lineage, but with unspecialized cilia. These large cells are shed at metamorphosis just like the trochoblasts of planktotrophic trochophores. Thus Carinoma has larval organs and should be classified as an indirect developer. The other extreme, with highly specialized larval types with larval organs shed during a cataclysmic metamorphosis, is well known—for example, from the nemertine pilidium larva and the larva of the annelid Polygordius. The differences in feeding biology may be, for instance, between planktotrophy or macrophagy and between lecithotrophy or active feeding. In this paper I discuss only the "primary" ciliated larvae (Jägersten, 1972); the secondary larvae, such as nauplii and caterpillars, seem to be of no interest for the present discussion.
|
Morphology and Feeding Biology of the Eumetazoan Ancestor |
|---|
|
TOPAbstractIntroductionMorphology and Feeding Biology...Early Eumetazoan RadiationTypes of Information About...The Origin of Indirect...Literature Cited |
|---|
It appears that the ancestor must have been feeding by the extracellular digestion of particles in a gut lined by a digestive epithelium (Nielsen, 2008b). Osmotrophy—that is, feeding on dissolved organic matter—can probably be left out of the discussion; it is known to be the sole means of feeding only in specialized parasites.
Haeckel (1874) suggested that the eumetazoan ancestor was a gastraea with ectoderm, endoderm, and archenteron with blastopore, essentially having the same structure as that of larval and adult cnidarians and of ontogenetic stages of many bilaterians. The cnidarians feed by ingesting particles into a gut with extracellular digestion, and this is the case in almost all the bilaterians. The syncytial endoderm found in acoels develops through fusion of blastomeres at a late stage of embryogenesis (Ramachandra et al., 2002) and must be considered an acoel apomorphy. Most morphology-based studies and some studies on molecular phylogeny place the cnidarians as the sister group of the remaining eumetazoans, the bilaterians (see, for example, Sperling et al., 2007; Srivastava et al., 2008), and their larval development has therefore been important in several discussions on life-cycle evolution.
Planula theories of various types (see classical papers like that of Hyman, 1951, and more recent papers like those of Salvini-Plawen, 1978, and Willmer, 1990) propose a free-living adult ancestor, planula or planaea, which should have resembled a compact cnidarian larva, but its feeding biology is either mentioned in a brief sentence or not discussed at all. Willmer (1990, p. 169) mentioned the problem, saying "it is not clear how the initial planula stage was supposed to feed, having no mouth or gut," and Salvini-Plawen (1978, p. 58) proposed the evolution of "an occasional mouth opening." It should be noted that all free-living organisms of the compact planula type are lecithotrophic larvae with so much yolk that feeding is not necessary until after metamorphosis and development of a gut.
Many authors state that all cnidarian planula larvae are lecithotrophic, although several examples of planktotrophic anthozoan larvae are known (see below). The anthozoans are now believed to represent the ancestral cnidarian life cycle (Collins, 2002; Collins et al., 2006), so phylogenetic inference about the ancestral eumetazoan larva indicates that it could have been planktotrophic as represented in Fig. 1 rather than as in diagrams such as those of Sly et al. (2003) and Raff (2008), which indicate that the cnidarians have only lecithotrophic larvae.
|
The new version of the gastraea theory, which proposes that the eumetazoan ancestor was a gastraea evolved from a "neotenic" larva of a homoscleromorph-like ancestor (Nielsen, 2008b), explains how a gastraea-like "homoscleromorph" larva could have evolved into the eumetazoan ancestor with sealed epithelia and a gut with extracellular digestion. Choanoflagellates and adult sponges are filter-feeders that rely mainly on particles in the size range of 1–5 µm. With their undulating cilia, they create water currents that transport water away from the cell body. Captured particles are ingested individually. However, the sponge larvae have "effective-stroke cilia" that propel the water parallel to the cell surface, and their beat is coordinated so that metachronal waves are formed. This is also seen in several lecithotrophic eumetazoan larvae, which can actually be quite difficult to identify even to phylum. The evolution of the eumetazoan gut and blastopore made the swallowing and digestion of larger eukaryotes possible (Peterson et al., 2005), and several anthozoan larvae are now known to feed on suspended material either in the plankton or in the detritus at the upper layer of the sea floor (Martin and Koss, 2002). Their method of particle collection has not been studied in detail (see below), but the locomotory cilia may well be engaged in wafting particles to the mouth. Filter-feeding with ciliary bands like those of trochophora and dipleurula larvae (see below) has not been observed. Ciliated epithelia of cnidarians are known to be able to reverse the direction of the effective stroke (Holley and Shelton, 1984), so the ciliated gastraea could well have transported particles in and out of the archenteron. The adaptational value of the evolution of extracellular digestion, enabling the digestion of larger particles, seems obvious.
In the gallertoid hypothesis (Bonik et al., 1976) the metazoans are derived from a cellularized ciliate and the epithelial eumetazoan guts from ciliated canals in sponge-like ancestors; the theories of Rieger (1994) and Dewel (2000) similarly describe the eumetazoan gut as derived from an (unspecified) sponge-like organization. Neither of these suggestions finds support from morphological or embryological observation.
|
Early Eumetazoan Radiation |
|---|
|
TOPAbstractIntroductionMorphology and Feeding Biology...Early Eumetazoan RadiationTypes of Information About...The Origin of Indirect...Literature Cited |
|---|
The eumetazoan ancestor, gastraea, has usually been described as radially symmetrical. Anthozoans, which are now believed to represent the ancestral cnidarian life cycle, show some bilateral traits, but cnidarians have no head and their concentration of nerve cells around the mouth cannot be described as a brain, so morphologically the anthozoan bilaterality cannot be related to that of the bilaterians. Some genes related to bilaterian symmetries are expressed in anthozoans, but there is no agreement about the interpretation of these results; for example, see the opposing views of Matus et al. (2006) and Rentsch et al. (2006).
Cnidaria
Cnidarian life cycles show an immense variation. Anthozoans are pelago-benthic with small, ciliated larvae, which are feeding in some species, and sessile adults. Larval feeding of some species is through secretion of mucous "nets" that are subsequently ingested (Fadlallah, 1983; Schwarz et al., 2002). Ciliary filter-feeding has not been observed. The ancestral medusozoan added the free-swimming, sexually mature medusa stage to the ancestral pelago-benthic life cycle (Collins et al., 2006).
Since some of the anthozoan larvae depend on food from the plankton for normal development (Goreau et al., 1981), the ancestral cnidarian may well have had a feeding gastrula larva.
Ctenophora
The comb jellies are at the same organizational stage as the cnidarians, with only one gut opening. The two small anal pores can transport small undigested particles, but larger undigested particles are egested through the mouth, so functionally the gut must be described as sac-shaped (Reeve and Walter, 1978). They are biradial, generally holopelagic, but a few types have added a benthic adult stage that may be creeping (e.g., in Coeloplana; Fricke, 1970) or sessile (in Tjalfiella; Mortensen, 1912). Their phylogenetic position is much debated and they will not be considered here, but it should be noted that they show how a benthic adult stage could be added to an ancestral holopelagic life cycle.
Bilateria
Bilaterian monophyly now seems unquestioned. It is supported by many morphological characters, such as the bilateral architecture with a brain and longitudinal nerve cords, and the molecular phylogenies are almost unanimous in agreement (e.g., Dunn et al., 2008; Peterson et al., 2008).
For more than a century, the Bilateria have traditionally been divided into Protostomia and Deuterostomia (or through a misinterpretation of Hyman, 1951, into Acoelomata, Pseudocoelomata, and Eucoelomata, a concept that unfortunately seems to survive in some American textbooks). However, new information, especially from molecular phylogeny, has demonstrated that the Acoela and Nemertodermatida, which traditionally have been treated as flatworms, are actually basal bilaterians. This is supported by several studies on RNA sequences (see, for example, Giribet et al., 2007) and also by observations on the distribution of Hox genes (see below) and microRNAs (Sempere et al., 2007). The two groups are interpreted either as one clade, the Acoelomorpha (e.g., Ruiz-Trillo et al., 2002), or as two separate clades with the Acoela as the most basal bilaterians (e.g., Baguñà et al., 2008a). They will be discussed together below. The modern concept of Platyhelminthes comprises only Catenulida and Rhabditophora and places the group in the Lophotrochozoa or Spiralia (Giribet et al., 2007; Dunn et al., 2008).
Acoelomorpha
Acoels and nemertodermatids are small creeping organisms that share a number of characters. They have only one gut opening, which is usually midventral. There is nothing to indicate that their ancestors had an anus. Only a few Hox genes have been found, one to two anterior genes, one to two central class genes, and a posterior class gene (Cook et al., 2004; Baguñà et al., 2008b; Hejnol and Martindale, 2008). This supports the idea that the acoelomorphs have not reached the more complicated evolutionary stage with an anterior-posterior regionation organized by the longer Hox cluster found in the eubilaterians (see below) and a tubular gut. The nervous system comprises longitudinal nerve cords and a brain, which is not associated with the mouth.
The acoelomorph organization resembles the early "bilaterogastraea" proposed by Jägersten (1959)—that is, a creeping, bilateral gastraea-like form with mesoderm. The development is direct, but this could well be a small specialization from a holopelagic ancestor or from an indirect type. The acoelomorphs show many characters that could be inferred as ancestral in the Bilateria, but also, as is to be expected, a number of specializations both in the genome and in the morphology and embryology (Hejnol and Martindale, 2008). Nevertheless they must be regarded as the living animals that are morphologically closest to the bilaterian ancestor.
Eubilateria
The evolution of a tubular gut and a new anterior-posterior regionation apparently associated with a Hox cluster with at least seven genes comprising anterior, group 3, median, and posterior genes expressed along the longitudinal body axis (Garcia-Fernàndez, 2005; Lemons and McGinnis, 2006) gives obvious ecological advantages. This is reflected in the enormous radiation and diversification of the "higher" animals in the Ediacaran (Peterson et al., 2008), so the "Cambrian explosion" could perhaps better be described as an Ediacaran explosion of the Eubilateria.
The anus has been lost in a number of eubilaterians; undoubtedly so in the articulate brachiopods and the ophiuroids, for example. Also the Platyhelminthes sensu stricto are now generally believed to have lost the anus. This loss is indicated not only by their phylogenetic position within the Eubilateria, as shown by several molecular analyses (Dunn et al., 2008; Lartillot and Philippe, 2008), but also by the embryology, which shows a unique programmed cell death of the blastomeres 3A–C and 4D—that is, almost all of the cells that give rise to the gut in other spiralians (Boyer et al., 1998).
Protostomia
The most characteristic morphological trait of the protostomes is the shape of the central nervous system, which consists of a circumoral brain and a pair of ventral longitudinal nerve cords (fused in nematodes, among others) (Hatschek, 1888: Zygoneura). This type of CNS is not found in Phoronida and Brachiopoda, which cluster with nemertines in almost all molecular analyses (e.g., Baguñà et al., 2008a; Dunn et al., 2008). However, almost all morphological and embryological characters indicate deuterostome affinities (Nielsen, 2001, 2005). The two groups will not be discussed here.
The modified trochaea theory, which considers the trochaea as the ancestor of the protostomes (Nielsen, 2001, 2009), proposes that the protostomian ancestor, trochaea, was a gastraea with a periblastoporal ring of compound cilia functioning as a downstream-collecting system using the catch-up principle (as seen today in Symbion, which has a ring-shaped downstream-collecting system around the mouth; see Riisgård et al., 2000). Adults of this holopelagic ancestor took up a creeping habit, feeding on deposited particles; they accordingly lost the compound cilia, but retained them in the planktotrophic larval stage. The creeping habit favoured the establishment of a new anterior-posterior axis, with collected food particles entering the blastopore at its anterior side and undigested particles leaving at the posterior side. Fusion of the lateral blastopore lips created a tubular gut, which made digestion more efficient and opened the way for specializations of various parts of the gut. This part of the theory is in accordance with observed blastopore closure or fusion of longitudinal neurogenic areas giving rise to the ventral nerve cords in, for example, annelids (Shankland and Savage, 1997), arthropods (Schwalm, 1997), nematodes (Sulston et al., 1983), and chaetognaths (Doncaster, 1902).
Although the larvae remained planktotrophic, they soon began to show development of the tubular gut, through the process aptly called "adultation" by Jägersten (1972). The band of compound cilia around the mouth could be pulled out into a pair of lateral loops, just like the loops of the prototroch on the entoproct tentacles (Nielsen, 2001). The anterior part of the loop became the larger prototroch, also involved in locomotion, and the two bands together formed the downstream collecting system so characteristic of planktotrophic trochophores. The perianal part of the circumblastoporal band became the telotroch. Planktotrophic trochophores are found in a number of spiralians, and the trochophore is believed to be the ancestral larval type of the spiralians (lophotrochozoans) (Nielsen, 2004, 2005; Raff, 2008). Cilia are generally lacking in the ecdysozoans, where only the development of the CNS reveals the relationships with the spiralians.
It should be emphasized that the modified version of the trochaea theory, with trochaea as the ancestor of the protostomes only, does not suggest that trochaea was "an adult holopelagic resembling a particular type of living feeding larva" as stated by Sly et al. (2003) and Raff (2008, p. 1474), among others. Trochaea is a hypothetical early holopelagic ancestor of the protostomes. The trochophora was originally believed to be "close to" the ancestor of the Protostomia (Zygoneura; Hatschek, 1891). The trochaea theory interprets it as the larva of the somewhat later evolutionary stage, the "gastroneuron," which had added a benthic deposit-feeding adult stage to become pelago-benthic.
The trochaea theory proposes that the protostomian ancestor was radially symmetrical. This seems problematic, because bilaterality is generally assumed to have become established in the ancestor of the Bilateria. However, the short Hox cluster of the acoelomorphs invites the question of homology of "bilaterality" in cnidarians, acoelomorphs, and eubilaterians, but so far this can only be speculation. It may furthermore indicate that the tubular gut evolved independently in Protostomia and Deuterostomia, but the uncertainty about the early evolution of the deuterostomes (see below) makes this difficult to ascertain.
Deuterostomia
The deuterostomes comprise two quite separate lineages, Ambulacraria (+ Xenoturbella: Bourlat et al., 2003, with unknown development) and Chordata. The sister-group relationship is strongly supported by almost all morphological and molecular studies. The fate of the blastopore, which has given name to the group, is clearly observed in the ambulacrarians where it directly becomes the anus, but the chordate embryology has become much more complicated by the neurulation. Most of the classical and newer theories derive the chordates from a dipleurula-like ancestor (Garstang, 1928; Nielsen, 1999), but there is no consensus about this. However, similarities in the morphology of the branchial sac of enteropneusts and chordates (Ruppert, 1997; Nielsen, 2001) could indicate that the common ancestor was enteropneust-like, perhaps with a dipleurula larva. This larval type is considered ancestral by most authors (e.g., Nielsen, 2001; Raff, 2008).
Observations of similar gene expressions in the mouth region of annelid and enteropneust larvae have been interpreted as indication of homology (Arendt et al., 2001), but gene expression alone cannot be taken as a proof of homology (Nielsen and Martinez, 2003). The structure and function of the ciliary bands of trochophora and dipleurula larvae are completely different, and it seems impossible to derive one from the other. Both protostomes and deuterostomes originated in the Ediacaran, and the split between the two groups has been dated to around the Earliest Ediacaran (Peterson et al., 2008).
The evolution of the Deuterostomia from the eubilaterian ancestor appears completely obscure.
|
Types of Information About Life Cycle Evolution |
|---|
|
TOPAbstractIntroductionMorphology and Feeding Biology...Early Eumetazoan RadiationTypes of Information About...The Origin of Indirect...Literature Cited |
|---|
Information from Precambrian/Cambrian fossils
Various types of microfossils in the size range of 600–1100 µm from the Neoproterozoic Doushanto Formation (about 600 mya) have been interpreted as metazoan embryos indicating direct development (Xiao and Knoll, 2000). However, the animal nature of some of these fossils has been questioned, for example by Bailey et al. (2007), who suggested that some of them may have been sulphur bacteria. This has again been questioned by Donoghue (2007), but it was nevertheless concluded that the classification of these fossils as animal embryos is far from certain. Smaller fossils, in the range of 150–250 µm, from the same deposits have been interpreted as early cleavage stages and embryos/larvae with indirect development or as tiny adults (Chen et al., 2000, 2004a, b), but this also has been questioned (Bengtson and Budd, 2004). It appears that the information from these fossils does not contribute to the discussion.
The development of the Early Cambrian Olivooides indicated that it is a direct-developing cnidarian, perhaps related to scyphozoans (Bengtson and Zhao, 1997). Other putative embryos of Middle Cambrian age are also rather large, which indicates direct development (Lin et al., 2006). The presence of Cambrian planktotrophic mollusc larvae has been discussed vigorously, but apparently no firm conclusion can be drawn (Freeman and Lundelius, 2007; Runnegar, 2007). However, it appears that a number of Cambrian brachiopods had planktotrophic larvae (Freeman and Lundelius, 2005).
Unquestionable fossils of small eggs that would indicate the presence of a planktotrophic larval stage and ciliated larvae have not been found in the Precambrian or Cambrian strata. However, as often stated: "absence of proof is not proof of absence." So, as also pointed out, for example by Gostling et al. (2008), the apparent absence in the Cambrian strata of fossilized small eggs/embryos is not a proof of the plesiomorphic character of direct development.
Could the Precambrian-Cambrian metazoans have had planktotrophic larvae?
Most of the Precambrian was characterized by the absence of eukaryotes, but the Ediacaran is a transitory period where both planktonic and benthic eukaryotes became abundant (Butterfield, 2007). The Precambrian planktonic prokaryotes were obviously sufficient for the filter-feeding of the choanoflagellates and sponges, which both rely on minute particles for intracellular digestion. The early cnidarian larvae may have captured prokaryotes with a mucous net like that of some living species. The ancestral protostomes and ambulacrarians, which probably had larvae with particle-collecting ciliary bands that generally capture particles larger than 2 µm in diameter (Riisgård et al., 2000; Strathmann, 2007), originated in the Ediacaran (Peterson et al., 2008), so this would fit well with the evolution of the ecosystems.
Can small animals have planktotrophic larvae?
Olive (1985) stated that extremely small organisms rarely, if ever, have planktotrophic development. This general trend is very obvious in many phyla, but it has sometimes been interpreted as a "law" that small species cannot have planktotrophic larvae (Sly et al., 2003, p. 626): "These [small] animals cannot produce enough gametes to afford planktotrophic larvae with a distinct body plan." First of all, it should be pointed out that the planktotrophic anthozoan larvae have a body plan that is not very different from that of small solitary adults, so that the difference between larva and adult is more in the life style than in the morphology. Two examples from spiralians demonstrate that tiny species may have planktotrophic larvae: One of the smallest living gastropods, Caecum glabrum (soft-part volume about 0.025 mm3; see Nielsen, 2004), deposits small egg cocoons each containing a few eggs with a diameter of about 50–60 µm, as is characteristic of species with planktotrophic development. All the eggs develop into minute planktotrophic veliger larvae (Götze, 1938). The solitary entoproct Loxosomella elegans broods its eggs, which have a diameter of about 50 µm, and tiny feeding trochophores are released (Nielsen, 1971). Although the examples from living species are few, they demonstrate that minute organisms can have planktotrophic larvae.
Developmental types in clades with well-known phylogeny
A number of interesting studies of smaller or larger clades with information about developmental types have been published during the last decades.
The echinoderms are probably the most thoroughly studied group. It has almost universally been concluded that the various types of feeding echinoderm larvae—echinopluteus, ophiopluteus, bipinnaria, and auricularia—and the enteropneust tornaria are all evolved from a planktotrophic dipleurula larva, which was the larva of the common ambulacrarian ancestor (Byrne et al., 2007; Raff, 2008). The many examples of lecithotrophic larvae or direct development in echinoderms are generally interpreted as specializations within a number of smaller or larger lineages (Nielsen, 1998; Peterson et al., 2000). Lecithotrophic "schmoo" larvae have evolved several times within separate echinoderm lineages of species with planktotrophic larvae (Wray, 1995, 1996). The most thoroughly studied echinoid genus is Heliocidaris, where the two sister species H. tuberculata with a feeding echinopluteus larva and H. erythrogramma with a schmoo larva are shown to have evolved from their latest common ancestor, with a planktotrophic larva, about 10 mya (Smith et al., 1990; Wray, 1995).
Large echinoderm clades, such as most of the holothurians and possibly all of the living crinoids, have lecithotrophic larvae. However, most species in the holothurian families Holothuriidae and Stichopodidae (order Aspidochirota) and Synaptidae (order Apoda) develop through the characteristic planktotrophic auricularia larvae (Sewell and McEuen, 2002). These larvae subsequently develop into nonfeeding doliolaria larvae, resembling the larvae of most crinoids (Lacalli, 1993). This clearly indicates that the doliolaria is a secondary larval type. Furthermore, it is now known that the feather star Metacrinus develops through a nonfeeding stage resembling a young auricularia before it develops into a doliolaria like that of the holothurians (Nakano et al., 2003).
Thus, it can hardly be doubted that the ancestral ambulacrarian had an indirect development with a planktotrophic dipleurula larva.
A well-documented example of multiple losses of larval structures in chordates is seen in ascidians. Most species have so-called tadpole larvae, which have a tail with notochord and muscles used in swimming. However, a number of species of the family Molgulidae have "anuran" larvae that lack the tail (but elements of both neural tube and mesoderm are found in the dorsal side of the embryo). The phylogeny of these families indicates that the larval tail has been lost independently a number times (Hadfield et al., 1995; Huber et al., 2000) (see also below).
Among the protostomes, the ecdysozoans lack ciliated primary larvae, but many spiralians have ciliated planktotrophic larvae.
Two studies of gastropod phylogeny and development should be mentioned: Reid's (1990, 1996) studies on littorinids and Collin's (2004) study of calyptraeids. Indirect development with planktotrophic veligers and "direct" development with crawling juveniles escaping from the cocoons was observed in various lineages, but the fact that the "direct" developers go through a veliger stage within the cocoons clearly indicates that the ancestor of both groups had a feeding veliger larva.
A very comprehensive phylogeny of the annelid families Oweniidae, Sabellariidae, Serpulidae, and Sabellidae has been presented by Rouse and Pleijel (2001). Planktotrophic larvae with downstream-collecting ciliary bands are well known from members of the three first-mentioned families, whereas planktotrophic larvae have not been observed in any of the sabellids. Surprisingly, Pernet (2003) observed complete downstream-collecting ciliary systems with prototroch, adoral ciliary zone, and metatroch in larvae of the sabellid polychaetes Schizobranchia, Demonax, Myxicola, and Pseudopotamilla. These larvae are lecithotrophic without a functioning gut at this stage. The larvae of Schizobranchia were observed to collect particles with the ciliary bands and transport them toward the mouth, where they were rejected and transported away along the gastrotroch. Pernet (2003, p. 295) concluded that the species "have lost larval feeding very recently, that opposed bands are inexpensive to construct and operate, or that opposed bands have some alternative function."
Extrapolation from studies of smaller spiralian clades to the spiralian ancestor is a controversial issue—see, for example, Rouse (1999, 2000a, b) versus Nielsen (2004, 2005). However, it appears that all observations on structures and feeding mechanisms of planktotrophic spiralian larvae with downstream-collecting ciliary bands using the catch-up principle are consistent with the evolution outlined by the trochaea-trochophora theory, which implies that the compound cilia of prototroch, metatroch, and telotroch have a ciliary beat in the direction of the ancestral blastopore. It would indeed be a very remarkable example of convergence if these structures should all have evolved convergently from an ancestral uniform ciliation (see below). It should be stressed that the loss of a character, such as the loss of the planktotrophic larval stage, is just as likely to have evolved in a speciation event giving rise to a large clade of direct developers as in one of a pair of sister species with different larval types, as seen in the echinoid genus Heliocidaris (Wray, 1995).
Could the downstream-collecting ciliary bands of various trochophores have evolved convergently?
The trochaea theory proposes that the proto-, meta-, and telotroch of all trochophora larvae are specializations of parts of the circumblastoporal band of compound cilia at the ancestral trochaea (see above). In contrast, the intercalation theories propose that these larval ciliary bands have evolved independently a number of times (e.g., Rouse, 1999, 2000b; Salvini-Plawen, 1980).
Through experiments with various ciliated larvae, Emlet (1991) showed that ciliary bands placed on ridges or at the posterior end of a larva enhanced its swimming effectiveness, and a number of rings of compound locomotory cilia are found on larvae such as those of protobranch bivalves (Buckland-Nicks et al., 2002), of the pteropod Pneumodermopsis (Nielsen, 1979), and of the annelid family Chaetopteridae (Pernet et al., 2002). This could have been the basis for the origin of prototroch and telotroch of the lecithotrophic trochophores, and the circumblastoporal archaeotroch of the trochaea (the origin of the downstream-collecting mechanism has not been studied). However, the evolution of a metatroch and a particle-collecting ciliary mechanism in a lecithotrophic larva is very difficult to visualize. These could not have been of positive adaptational value before the complicated structure was fully formed and functioning. In fact, the developing metatroch could initially be disadvantageous because its beating would counteract the locomotory current of the prototroch.
In more general terms, Sly et al. (2003) proposed the evolution of a larva in a direct-developing life cycle through the introduction of a "facultative novel feature" into the development, followed by its transformation to a required "larval" feature and finally, through co-option with other features, to the fully formed larva. This is completely without consideration of possible adaptational values of the several evolutionary steps. To paraphrase Raff (2008, p. 1474), one could say that the intercalation requires a selective role for the evolving "larval genes" before a new planktotrophic larval stage has evolved, which requires selection for novel developmental elements prior to need.
The trochaea theory proposes a continuous evolution of living protostomes with a tubular gut from a gastraea-like ancestor with a circumblastoporal ring of compound cilia. Structure and function of the ciliary bands of the trochophore, including the orientation of the effective stroke of the bands of compound cilia against the ancestral blastopore, remain unmodified, and all evolutionary stages must have been able to feed. It implies the loss of planktotrophy, and of larvae in general, in a number of lineages reaching from single species to phyla, but the loss of planktotrophy in the ancestor of a phylum is just as likely as in any other species. Loss of planktotrophy has been well documented in a number of clades (see above), whereas a de novo evolution of a downstream-collecting system has never been demonstrated in any extant phylum.
Convergent evolution of the trochophore larvae therefore appears improbable.
Information from gene expression
The only well-documented demonstration of the genetic background for the loss of a larval structure is the loss of the larval tail in several species of the ascidian family Molgulidae (see above). The genetic mechanisms responsible for tail development have been studied in the species pair Molgula oculata, which has a tadpole larva, and M. occulta, which has an anuran larva (Swalla and Jeffery, 1996). The experiments showed that the zinc finger gene Manx is needed for the differentiation of the notochord from the gut and consequently for the following normal development of the tail. The development of a normal tail depends on the interplay of numerous genes, but the silencing of Manx is enough to suppress the organization. This shows that the loss of a larval character or structure that is not needed for further normal development can be caused by the silencing of just one gene, and this evolutionary mechanism may have been at work in many other clades at any level of the phylogenetic tree.
Love et al. (2008) have tried to infer whether direct or indirect development "originated first" by comparing gene expression patterns in the gut rudiment in larvae of the indirect-developing sea urchin Heliocidaris tuberculata and its direct-developing sister species H. erythrogramma. They found temporally changing patterns of gene expression during the early, almost totally bilaterally symmetrical, developmental stages of the pluteus larvae of H. tuberculata: expression was weak in the stomodaeum and oral field, which degenerate at metamorphosis, and stronger in the stomach and rectum, which are retained in the adults. The later stages with the adult structures developing at the left side were not studied. Larvae of H. erythrogramma with developing gut and developing adult structures including primary podia showed strong expression of the gene mandrillin in the developing gut, just as in the developing gut of H. tuberculata; however, other gene expressions differed between the larvae of the two species. That different stages show different gene expressions is clearly demonstrated in the early stages of H. tuberculata, so comparisons between different stages of two species cannot give any information about the evolution of developmental types in the genus Heliocidaris, and the data seem perfectly compatible with the interpretation of indirect development as ancestral. Thus, the further conclusions about the ancestral developmental type of deuterostome larvae appear unfounded.
Information from hybridization experiments with the two Heliocidaris species and with a more distantly related direct developer likewise indicate that direct development is the derived developmental type (Raff et al., 1999, 2003) just as in the ascidians discussed above.
In addition, Sly et al. (2003) used the expression patterns of Hox genes as the basis for a discussion of the molecular background for possible intercalation of a larval stage into a direct development. However, it appears to me that "tree thinking" about this subject gives a different picture. The sister group of the eubilaterians, the acoelomorphs, has only a short Hox cluster that is not related to the anterior-posterior regionation along a tubular gut (see above). This is an indication of the structure of the Hox cluster in the bilaterian ancestor. In protostomes, Hox genes are generally not expressed in the episphere region (Nielsen, 2008a), which indicates that the long Hox cluster is associated with the somewhat later developing structures of the larvae, many of which are adult characters that have already begun development in the larva.
So far, the intercalation hypothesis has not found any support from gene expression.
|
The Origin of Indirect Development |
|---|
|
TOPAbstractIntroductionMorphology and Feeding Biology...Early Eumetazoan RadiationTypes of Information About...The Origin of Indirect...Literature Cited |
|---|
A generalized picture of early metazoan radiation indicates three or four independent evolutions of benthic stages from holoplanktonic ancestors (Fig. 2).
|
A true gastrula with sealed epithelia and an archenteron is seen in all eumetazoans, both in the larval and adult organization of cnidarians and ctenophores (and acoelomorphs) and as an ontogenetic stage in eubilaterians. Only the ctenophores have remained holopelagic.
According to my views expressed above, a benthic adult stage has been added to the ancestral holopelagic eumetazoan life cycle two or three times.
It seems generally accepted that the ancestral cnidarian was pelago-benthic with a sessile adult, and it is more or less tacitly agreed that the cnidarians evolved from a planula-like ancestor (Werner, 1984; Ruppert et al., 2004). I find it impossible to imagine a nonfeeding ancestor, such as a compact planula larva, so the ancestor must have been a feeding organism of the general structure of a gastraea. There is no problem with "transfer of reproductive capacity from the larva to the adult" as suggested by Degnan and Degnan (2006); it is simply the adult holopelagic ancestor that becomes sessile. There has been very little specialization of the cnidarian larvae; only the zoanthina and zoanthella larvae of certain zoanthid anthozoans show special ciliary bands (Martin and Koss, 2002).
The acoelomorphs can be interpreted as creeping bilateral gastrulae with mesoderm.
The life cycle of the latest common eubilaterian ancestor is difficult to make out, because the life cycle of the ancestral deuterostome is so difficult to deduce. However, the evolution of the protostomes appears well explained through the trochaea-trochophora theory, which is supported by very many morphological and embryological observations. As shown above, most arguments against the theory can be refuted.
There are numerous larval specializations (caenogenesis, or "detours" from the ancestral ontogeny) in the protostomes. Good examples are the well-known veliger larvae of several gastropods and bivalves with the particle-collecting ciliary system situated at the edge of a pair of wing-like expansions; the "blown-up" trochophores of certain polygordiid, phyllodocid, and oweniid polychaetes (Pernet et al., 2002); and the various types of pericalymma larvae (Nielsen, 2001). Other types are of course the lecithotrophic larvae, which span the whole spectrum from those resembling planktotrophic larvae in most details to developmental types almost without larval traits.
The deuterostomes are much more difficult to interpret. As stressed above, any attempt at visualizing deep evolutionary processes should include continuous structural and functional changes, and I know of no attempt at explaining the origin of the tubular gut in the deuterostomes or of the upstream-collecting ciliary system of the dipleurula larva. Is the deuterostome mouth really new, which would indicate that the tubular gut evolved independently in protostomes and deuterostomes?
In conclusion, I find strong support for the addition of a benthic stage into the ancestral holopelagic life cycle of sponges, cnidarians, and protostomes; only the deuterostomes remain enigmatic. The eumetazoan ancestor was a direct-developing, swimming or creeping gastraea or planula from which a pelago-benthic life cycle with a feeding larva evolved three times: (1) in the Cnidaria, where the adult sessile phase was added and the pelagic phase retained with very little modification; (2) in the Protostomia, perhaps via the evolution suggested by the trochaea-trochophora theory; and (3) in the Deuterostomia through a still-unknown evolution. Ctenophores and acoelomorphs remained at the gastrula stage of organization, being respectively holopelagic (mainly) and holobenthic. Small undigested particles may exit via the two small anal pores of the ctenophores, but the larger particles are egested through the mouth (Reeve and Walter, 1978). So the ctenophore gut is functionally like that of the cnidarians.
In contrast, I find no support for the intercalation hypothesis in relation to the origin of the indirect development in several groups of the just-mentioned clades.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
Literature Cited |
|---|
|
TOPAbstractIntroductionMorphology and Feeding Biology...Early Eumetazoan RadiationTypes of Information About...The Origin of Indirect...Literature Cited |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. B. Emlet, S. A. Maslakova, A. L. Shanks, and C. M. Young Biological Bulletin Virtual Symposium: Biology of Marine Invertebrate Larvae Biol. Bull., June 1, 2009; 216(3): 201 - 202. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
