Implications of cnidarian gene expression patterns ... - Oxford Academic

701

Implications of cnidarian gene expression patterns for the origins of bilaterality—is the glass half full or half empty? Eldon E. Ball, Danielle M. de Jong,,† Bernd Schierwater,‡ Chuya Shinzato,† David C. Hayward and David J. Miller†,1 *ARC Centre for the Molecular Genetics of Development, Research School of Biological Sciences, Australian National University, PO Box 475, Canberra, ACT 2601, Australia; †ARC Centre of Excellence in Coral Reef Studies and Comparative Genomics Centre, James Cook University, Townsville, Queensland 4811, Australia; ‡ITZ, Ecology & Evolution, TiHo Hannover, Bu¨nteweg 17d, D-30559 Hannover, Germany

Synopsis The past two years have seen a dramatic increase in the available data on gene sequence and gene expression for cnidarians and other ‘‘lower’’ Metazoa, and a flurry of recent papers has drawn on these to address the origins of bilaterality. Cnidarian homologs of many genes that play key roles in the specification of both the A/P and D/V axes of bilaterians have been characterized, and their patterns of expression determined. Some of these expression patterns are consistent with the possibility of conservation of function between Cnidaria and Bilateria, but others clearly differ. Moreover, in some cases very different interpretations have been made on the basis of the same, or similar, data. In part, these differences reflect the inevitable uncertainties associated with the depth of the divergence between cnidarians and bilaterians. In this article, we briefly summarize the cnidarian data on gene expression and organization relevant to axis formation, the varying interpretations of these data, and where they conflict. Our conclusion is that the presently available data do not allow us to unequivocally homologize the single overt axis of cnidarians with either of the bilaterian axes.

The axes and assumptions concerning their origins

The fossil record and the anatomy of the ancestral cnidarian

Although still frequently described as radially symmetrical and diploblastic, there is a longstanding debate over the axes and the body layers of cnidarians and how it/they relate to those of Bilateria. Do some modern cnidarians illustrate an evolutionary stage through which bilaterians passed on the way to becoming overtly bilateral (Boero et al. 1998; Seipel and Schmid 2005), or are they an evolutionary offshoot that evolved from an already bilateral common ancestor? The standard textbook view has been that the Cnidaria are separated from the Bilateria by two major dichotomies. Firstly, they are considered to be radial, while Bilateria are bilateral. Secondly, they are considered to be diploblastic (have two body layers) in contrast to the triploblastic Bilateria. Further, it is commonly assumed that the oral end of Cnidaria, where the mouth lies, corresponds to the anterior end of Bilateria. Although these are widely prevalent views, none of them has gone undebated.

Before evaluating the textbook generalizations outlined above, there are several matters that should be considered. Firstly, the nature of the common ancestor of Cnidaria and Bilateria is still a matter of active debate; the possibilities being a medusoid form (see Seipel and Schmid 2005 for a summary of arguments), a polypoid form or a planuloid form. The fossil record does not provide any guidance as to the nature of the common ancestor since the earliest known fossil assemblages, from the Neoproterozoic Ediacaran deposits, ‘‘included a mixture of stem- and crown-group radial animals, stem-group bilaterian animals, ‘failed experiments’ in animal evolution, and perhaps representatives of other eukaryotic kingdoms’’ (Narbonne 2005). Secondly, there is a hypothesis, traced by Willmer (1990) from Haeckel (1874, 1875), through Hyman (1940) to Salvini-Plawen and Splechtna (1978), which proposes that cnidarians arose from a settling biradial planula with a single pair of tentacles. This hypothesis and related scenarios

From the symposium ‘‘Key Transitions in Animal Evolution’’ presented at the annual meeting of the Society for Integrative and Comparative Biology, January 3–7, 2007, at Phoenix, Arizona. 1 E-mail: [email protected] Integrative and Comparative Biology, volume 47, number 5, pp. 701–711 doi:10.1093/icb/icm028 Advanced Access publication July 9, 2007 ß The Author 2007. Published by Oxford University Press on behalf of the Society for Integrative and Comparative Biology. All rights reserved. For permissions please email: [email protected].

702

E. E. Ball et al.

Fig. 1 Many members of the Class Anthozoa exhibit bilaterality. (A) Mesentery pattern in an ancient rugose coral, a now extinct group which flourished during the Paleozoic (redrawn from Willmer 1990). (B) Skeleton of a modern solitary fungiid coral (Anthozoa; Scleractinia) showing extreme elongation of the mouth. (C and D) Cross sections of an Alcyonium (Anthozoa; Octocorallia) polyp cut at the level of the pharynx (C) and further proximally (D), showing several aspects of bilaterality (redrawn from Russell-Hunter 1979).

(including the placula hypothesis; Bu¨tschli 1884) sit fairly comfortably with the molecular evidence presented subsequently. If the common ancestor of Cnidaria and Bilateria was indeed such an animal, this would help explain the lack of candidate fossils since preservation of such an organism would require highly specialized conditions. All extant cnidarians are likely to have some derived characteristics, so any conclusions about ancestral states based solely on living cnidarians are open to question. Nevertheless, with that caveat, the Anthozoa, which for reasons summarized below, is now thought to best reflect the ancestral condition, sits least comfortably within the textbook radial— diploblastic scheme outlined above. In contrast, the Hydrozoa, on whose characteristics the scheme is mostly based, is now considered to be the most derived. Today, the most commonly cited arguments supporting the proposition that the Anthozoa comprises the least derived of modern forms are molecular, and include mitochondrial genome structure, which is circular in Anthozoa and Bilateria, but linear in Medusozoa (the Classes Hydrozoa, Cubozoa, and Scyphozoa), and phylogenetic analyses of sequence data from several classes of genes. Even in the premolecular era, however, many scholars took this point of view. For example, Pantin (1966) argued that both differentiation of nematocysts and, above all, the simplicity of the anthozoan nervous system, which in contrast to that of hydrozoans, really is a nerve net showing few signs of centralization except an increased density of neurons around the mouth, supported this point of view. Willmer (1990) has succinctly summarized the nonmolecular evidence for the primitive nature of the Anthozoa as follows: simpler life cycles, lesser ability to cope with physiologically difficult environments, and less elaborate and diverse cnidoblasts. Hydrozoans, by contrast, have more complex musculature and

nervous systems, as well as often having specialized sense organs, all features pointing to a higher degree of derivation. Not only do the lines of evidence cited above point to anthozoan morphology as best reflecting ancestral characteristics, but also this is in agreement with the fossil record. Conway-Morris’ (2006) recent summary of Cambrian fossils lists numerous anthozoan groups as probable but fails to list any hydrozoans and states that there are no convincing Cambrian scyphozoans or cubozoans (but see Hagadorn et al. 2002; Pickerill 1982). This presents a challenge to the commonly accepted idea of a radially symmetrical ancestral cnidarian since most extant anthozoans are either biradial or bilateral. Obvious examples of bilaterality can be found among corals, both ancient and modern (Fig. 1A and B) but most other anthozoans show similar tendencies, having an elongate mouth with specializations at one or both ends (Fig. 1C and D) and corresponding asymmetries in septal musculature. As might be expected, this bilaterality is the result of asymmetical gene activity, as will be discussed subsequently, following consideration of patterning of the oral/ aboral axis and the possible role of Hox genes in this process.

Does oral–aboral correspond to anterior–posterior and, if so, what is the correspondence? It has commonly been assumed that oral–aboral does correspond to anterior–posterior. Leaving aside the correctness of this assumption, if this is the case then how do the two axes match up (i.e., is oral anterior or posterior)? The case becomes particularly complicated when it is realized that some of the answers will depend on whether the planula or the polyp stage is considered and that there is no clear answer as to whether the ancestral cnidarian was planuloid,

Axis determination in Cnidaria

polypoid, or even medusoid. Considering the planula first, some of the arguments for the aboral end being anterior are that it is anterior as the animal swims, it frequently has a concentration of sensory neurons, and in planulae of Podocoryne (Class Hydrozoa) the nervous system develops sequentially from that end (Gro¨ger and Schmid 2001), as is the case in bilaterians. The main counter argument appears to be that the blastopore becomes the oral pore, which in turn becomes the mouth after settlement, and the mouth is generally toward the anterior end in bilaterians. However, as Rieger et al. (2005) pointed out, the mouth is not actually AT the anterior pole. In fact, in acoel flatworms, which may be the most similar living form to the ancestral bilaterian, the mouth is located ventrally on the posterior third of the animal. If one considers the polyp to be the ancestral life form then the argument that oral corresponds to anterior becomes stronger since in many cnidarians there is a concentration of neurons associated with the mouth (Grimmelikhuijzen and Westfall 1995). One way of assessing axial correspondence is to look at the patterns of gene expression of cnidarian homologs of genes that are characteristically expressed either anteriorly or posteriorly in bilaterians. This apparently straight-forward approach, however, has a potential flaw in that expression is sometimes found to switch ends at settlement. Thus, both Podocoryne Cnox2-pc and Acropora Pax-Dam switch from aboral to oral expression at the time of settlement (Fig. 2). In the absence of functional studies, it is unclear whether this change is associated with a change in gene function.

What is a ‘‘Hox’’ gene? The cnidarian ‘‘Hox’’ data have been subject to very different interpretations (e.g., compare Finnerty et al. 2004; Kamm et al. 2006; Chourrout et al. 2006; Ryan et al. 2007). A contributing factor has been the shifting definition of ‘‘Hox’’, which continues to plague the field today. The definition originally given to the word by a distinguished group of investigators was ‘‘any homeobox containing gene or genomic fragment may be given the designation Hox (Homeobox) providing that a significant fraction of the amino acids it encodes are identical to those of the homeobox in the Drosophila Antennapedia gene’’ (Martin 1987). By the mid-1990s, however, following the realization that flies and mice had orthologous clusters of genes performing similar functions (Scott 1992), the definition of ‘‘Hox gene’’ became more restricted. This new approach is encapsulated by the

703

Fig. 2 A source of potential confusion in using the position at which a gene is expressed along the oral-aboral axis to homologize this to the A/P axis is that this position is sometimes reversed at the time of settlement. Two genes exhibiting this phenomenon are: (A) Podocoryne Cnox2-Pc (drawn from Masuda-Nakagawa et al. 2000), and (B) Acropora Pax-Dam (drawn from de Jong et al. 2006). Expression is shown in black.

following two quotes: ‘‘Hox genes are characterized by three criteria: (1) sequence similarity to genes of the Drosophila HOM-C complex; (2) the serial position in a cluster; and (3) expression along the anteroposterior axis’’ (Ruddle et al. 1994) and ‘‘The epithet Hox should be used only for the clustered genes homologous to the insect and vertebrate homeotic complexes, and not for other classes of homeobox genes’’ (Akam 1995). This view was dominant during the late 1990s and into the first few years of this century. Thus, much of the cnidarian ‘‘Hox’’ literature of this period dealt with classification of Hox-like genes as anterior, posterior, intermediate, or ParaHox, and with the search for clusters. However, as sequencing of ‘‘Hox’’ genes from diverse organisms continued, more and more broken clusters were discovered in organisms that were clearly bilaterians. The paper by Seo et al. (2004) on disintegration of the Hox cluster in Oikopleura and other tunicates, raises the thorny issue of using clustering in any definition of Hox since tunicates are unquestionably advanced bilaterians. However, despite extensive gene loss and cluster disintegration, Oikopleura has some remaining genes that unquestionably fall into the classes known from Drosophila and vertebrates, and expresses these in patterns that are characteristic of Hox genes. Thus, although the Hox genes in Oikopleura are not clustered, it had an ancestor in which they were. Hox cluster integrity should no longer be mandatory for membership of the club, as it seems only to be essential for maintenance of temporal colinearity (Kmita and Duboule 2003) – in fact the

704

Hox ‘‘cluster’’ of Drosophila is fragmented and some of its members clearly no longer function as Hox genes. Thus ‘‘Hox’’ is now a very imprecise term meaning different things to different people and in the case of Cnidaria perhaps generating more heat than light. The presence of a hexapeptide motif in two of the Nematostella Hox-like proteins, anthox7 and anthox8, has been put forward as one line of evidence that these are indeed true Hox proteins (Finnerty et al. 2004). However, although sequence coding for this protein motif is present in most true Hox gene classes, it is also present in other members of the extended Hox family and hence predates the divergence of (for example) emx, msx and Hox genes. It is interesting, however, that although it is clearly present in the two Nematostella genes, the motif has not been reported thus far in any other cnidarian homeodomain protein—neither Hox-like genes, nor any of the cnidarian Msx or Emx genes characterized to date encode hexapeptides. It is also remarkable that anthox6, the one gene that all workers agree IS orthologous to a true Hox gene class, lacks sequence coding for a hexapeptide. Thus, the presence of a hexapeptide has no diagnostic value—most likely this motif was present in the ancestor of a large suite of ANTP-class genes, and has been lost from most of the cnidarian genes. One group of potential regulators of Hox genes (reviewed by Pearson et al. 2005) which, judging from the Nematostella genome, appears to be limited to four members of three gene families from cnidarians, is the microRNAs (Prochnik et al. 2007). This is in contrast to at least 30 gene families conserved across the Bilateria (Prochnik et al. 2007).

Animals predating the canonical Hox system? Although our focus is on the Cnidaria, recent findings on the paucity of true Hox genes in closely related lower Metazoa is relevant here. Turning first to sponges, and paraphrasing Larroux et al. (2006) no Hox, ParaHox, or extended Hox genes have been conclusively recovered from any sponges despite attempts on several species (Seimiya et al. 1994; Manuel and Le Parco 2000). Yet the larva of Amphimedon (then known as Reniera) is clearly axially organized (Degnan et al. 2005) and bears a remarkable resemblance to the planula larva of Acropora. Indeed, Le´vi (1963) has characterized sponge larvae as having axial, or even tetraradial, symmetry. Thus, in sponges it appears that there is a clear axial (oral–aboral) organization in the absence

E. E. Ball et al.

of Hox genes. This suggestion should soon be proven or disproven on completion of the sequencing of the Amphimedon genome. In Placozoa, which are morphologically polarized relative to the substratum but otherwise show no sign of axial specialization (summarized by Miller and Ball 2005; Schierwater 2005), five ANTP-class genes have been recovered (summarized by Monteiro et al. 2006). However, the only one closely related to the Hox/ParaHox classes is Trox-2 (Schierwater and Kuhn 1998), which groups with the cnidarian cnox2 and bilaterian Gsx-type sequences in phylogenetic analyses (Gauchat et al. 2000; Hayward et al. 2001). The acoel flatworms, which may have diverged from the bilaterian stem after cnidarians but before the protostome/deuterostome split (Ruiz-Trillo et al. 2002; Telford et al. 2003), are likely to be highly significant in terms of understanding the origins of the bilaterian Hox cluster. Cook et al. (2004) and Baguna and Riutort (2004) have investigated the Hox complement of phylogenetically divergent acoels and have found representatives of four paralogy groups of Hox genes. Although only partial sequences are available at present, these appear to correspond to anterior, group 3, intermediate and posterior Hox types. Unfortunately, although it is likely to be of great importance for our understanding of Hox genes, no information appears to be yet available concerning possible linkages between these genes or on their expression.

Cnidarian Hox and Hox-like Genes Finally we return to the question of Hox genes in Cnidaria, which is by far the most investigated lower metazoan phylum in this respect. In the early 1990s, most of us were influenced by one of the central proposals of the seminal paper by Slack et al. (1993)—that the Hox cluster is a defining characteristic of animals, and to some extent this influenced interpretation of the cnidarian data. Some of the present authors were as susceptible as anyone to this suggestion and interpreted the linkage of an evenskipped gene and the Hox-like gene AntpC (which corresponds to Nematostella anthox6) as a remnant of an ancestral Hox cluster (Miller and Miles 1993), and set out in search of linked Hox genes in the coral, Acropora. Unfortunately, the search proved fruitless. Hox-like genes have been identified in a range of cnidarians (summarized by Gauchat et al. 2000; Ryan et al. 2006) but recent analyses (Kamm et al. 2006; Chourrout et al. 2006) cast serious doubts on the idea that a Hox cluster existed at the time of the

705


Fig. 3 Orthologous Hox-like genes are expressed in highly divergent patterns across the Cnidaria. The diversity of these expression patterns contrasts with the high degree of similarity in the expression patterns of true Hox genes across the Bilateria. (A) Podocoryne Cnox-1pc is expressed in ectoderm and endoderm at the aboral end the planula and is not expressed in the polyp. (B) Nematostella anthox6 expression is restricted to the oral endoderm of the planula and not expressed in the polyp. (C) Eleutheria Cnox-5ed is expressed in the aboral ectoderm of the planula and in ectoderm and endoderm at either end of the polyp. Expression is shown in black. Drawn from Kamm et al. (2006).

cnidarian/bilaterian divergence. Phylogenetic analyses indicate that cnidarians have anterior Hox-like genes and that several other cnidarians, including the hydrozoan Eleutheria, have a second type that is related to both the posterior Hox and Cdx classes (Kamm et al. 2006). All other presently characterized cnidarian Hox-like genes postdate the bilaterian divergence. In Nematostella, three paralogous Hoxlike genes are linked to the eve/anthox6 pair, but this ‘‘cluster’’ arose since the bilaterian divergence, and is therefore unrelated to the Hox clusters of bilaterians. Moreover, the expression patterns of the cnidarian Hox-like genes are highly divergent and do not fit the expectation, based on the true Hox genes of bilaterians, of relatively conserved expression across the phylum (Fig. 3). The only Nematostella gene (anthox6) that all investigators appear to accept as a bona-fide member of a Hox orthology group, is expressed in a narrow stripe in the pharynx (Finnerty et al. 2004). Presently available data are consistent

with the hypothesis that the cnidarian/bilaterian common ancestor (Ureumetazoa) is likely to have had two genes (Hox precursors; see above) that may have later given rise to the Hox cluster in Urbilateria. The cnidarian Hox debate has recently been taken up again, with yet another analysis (Ryan et al. 2007) directly challenging the conclusions of the papers by both Kamm et al. (2006) and Chourrout et al. (2006). Surprisingly, the results presented in all of these recent papers do not differ radically—what differs is their interpretation. Thus, Kamm et al. (2006) found a ‘‘true’’ anterior Hox gene and a gene similar to a posterior Hox/Cdx gene in Nematostella and Eleutheria, with no evidence of linkage of orthologous Hox genes. Chourrout et al. (2006) found only anterior Hox genes in Nematostella and Hydra, and Ryan et al. (2007) found Hox1 and Hox2 (i.e., anterior Hox) and a Hox 9þ (i.e., posterior Hox) gene. The two views are essentially that either a Hox system was present in the urmetazoan and it has subsequently diverged almost beyond recognition (Ryan et al. 2007) or such a system evolved in Urbilateria after the divergence of the Cnidaria, so that what we see today in the latter reflects a pre-Hox state (Kamm et al. 2006; Chourrout et al. 2006). Arguments can be made for either scenario, but in any case it is clear that the definition of a ‘‘Hox’’ gene has itself degenerated, almost beyond recognition, when applied in the context of the lower Metazoa. Until there is general agreement on such a definition the debate cannot be taken much farther in the absence of new data.

Cnidarian homologs of Drosophila head-gap genes In the absence of a canonical Hox cluster, what other genes might be good markers for the cnidarian equivalent of the anterior pole? Two of the most universal anterior markers are homeobox genes corresponding to the Drosophila head-gap genes empty-spiracles (ems) and orthodenticle (otd), which have conserved roles in anterior patterning from insects to mammals (reviewed by Lichtneckert and Reichert 2005). Whereas the Hox data are equivocal, unambiguous homologs of ems and otd have been cloned from several cnidarians (Mokady et al. 1998; Mu¨ller et al. 1999; Smith et al. 1999; de Jong et al. 2006). In Acropora, the ems gene emx-Am is expressed in the aboral two thirds of the planula larva (Fig. 4A), most obviously in presumed neurons (de Jong et al. 2006), but also more generally throughout the ectoderm; this is reminiscent of empty-spiracles

706

E. E. Ball et al.

The ‘‘directive’’ axis of cnidarians

Fig. 4 Apparently contradictory expression patterns of ‘‘anterior’’ genes in the coral Acropora. emx-Am and otx-Am, which are members of gene families widely considered to be diagnostic anterior genes, are expressed at opposite ends of the Acropora planula larva. Direct comparisons are complicated by the fact that both genes have been independently duplicated within the Cnidaria. Even if paralogs of one or both of these genes are expressed at opposite ends of the axis to those shown here, however, the ancestral pattern will be equivocal. Externally visible expression is shown for emx-Am, while otxA-Am is shown in section. Drawn from de Jong et al. (2006).

function in Drosophila, which has roles in patterning both the anterior ectoderm and the brain (Dalton et al. 1989; Hartmann et al. 2001). While expression of this head-gap gene at the aboral end of the primary axis (i.e. the ‘anterior’ end with respect to swimming direction) suggests correspondence between the aboral and anterior, an Acropora homolog of another ‘definitive’ anterior marker, otd/Otx, is expressed at the opposite end of the primary axis (Fig. 4B). The otxA-Am is expressed initially around the blastopore (which later becomes the oral pore), and later in an undefined cell type at the base of the ectoderm; these cells are distributed along the axis but only those in the oral half of the planula express otxA-Am (de Jong et al. 2006). A second otx gene is present in Acropora but is first expressed in presumptive endoderm at the start of gastrulation and later throughout the planula endoderm (Hayward et al., unpublished data). Although the first gene is more widely expressed at the oral end of Acropora, it is expressed in the pharynx in a position corresponding to that shown for Nematostella otx (Matus et al. 2006a; Fig. 7, compare de Jong et al., 2006; Fig. 1D), and thus may turn out to correspond to one of the Nematostella otx genes described there, once data on sequence and expression are available for Nematostella. Nevertheless, it is difficult to justify using these ‘‘head genes’’ as anterior markers when they indicate that opposite ends of the planula are anterior.

Expression data for genes of the Dpp/BMP2/4 type in Acropora (Hayward et al. 2002 and unpublished) and Nematostella (Finnerty et al. 2004) provides one of the most convincing lines of evidence that cnidarians (or, at least, anthozoans) have more than one axis. In flies and vertebrates, Dpp/BMP4 plays a key role in D/V patterning; on one side of the D/V axis the activity of this antineurogenic protein is antagonized by sog/chordin, allowing the nerve cord to develop in both groups. During gastrulation of both Acropora (Hayward et al. 2002) and Nematostella (Finnerty et al. 2004), Dpp/BMP4 is expressed asymmetrically in one quadrant of the blastopore lip. In planulae of Nematostella (Finnerty et al. 2004) Dpp/BMP4 is also asymmetrically expressed in the pharyngeal ectoderm. Combined with expression data for a number of Nematostella ‘‘Hox’’ genes, which show at least three distinct domains of expression along the primary axis (Finnerty et al. 2004), these intriguing similarities have been used to argue for a direct correspondence of cnidarian ‘‘bilaterality’’ with that of members of the Bilateria, underlain by common molecular mechanisms—so that, as in bilaterians, cnidarian ‘‘Hox’’ genes pattern along the primary axis (therefore, O/A ¼ A/P) and Dpp is involved in patterning the second axis (therefore, the directive axis ¼ D/V). More recently many of the same authors have established the 3D patterns of expression of these genes in more detail (Matus et al. 2006a) and have proposed that they may be involved in the patterning of the nervous system. Although expression data for Dpp/BMP4 (see above) provided much cause for speculation concerning correspondence of cnidarian and bilaterian axes, more extensive investigation of the expression of TGFb ligands, antagonists and receptors (Rentzsch et al. 2006; Matus et al. 2006a, 2006b) considerably complicates the situation and appears to rule out any simple correspondence between the O/A cnidarian axis and either axis of bilaterians. As outlined above, in bilaterians Dpp/BMP4 activity is antagonized on the opposite side of the D/V axis by short gastrulation (sog) in Drosophila or its homolog chordin in vertebrates, allowing neuroectoderm to develop on the sog/chordin side. In Nematostella, however, chordin is expressed on the same side of the directive axis as dpp, in a partially overlapping manner (Rentzsch et al. 2006; Matus et al. 2006b). One of the two Nematostella homologs of another TGFb antagonist, noggin, is also expressed specifically on the dpp/chordin side of the directive axis


707

Fig. 5 Expression of Nematostella homologs of key axial patterning genes. Because gene expression patterns are constantly changing, static diagrams are inadequate but do serve to illustrate overall patterns of message distribution. (A) In contrast to the situation in bilaterians, dpp and chordin are expressed on the same side of the directive axis and in gradients both along the oral-aboral axis and at right angles to it, as shown in this diagram redrawn from Rentzsch et al. (2006). (B) Expression domains of some genes potentially involved in axis determination or providing axial markers (redrawn from Matus et al. 2006a). The expression of many additional genes, belonging to a broader Hox-like class, has recently been described (Ryan et al. 2007), but has been omitted in the interests of clarity. (C) Transverse section of (B) at the level of the pharynx showing the distribution of differentially expressed messages in the directive axis (redrawn from Matus et al. 2006a with gremlin distribution estimated from Rentzsch et al. 2006). (D) Expression patterns of four Nematostella Hox-like genes along the directive (left column) and oral/aboral axes (right column). The position of the oral pore is shown by an asterisk. These expression patterns suggest the possibility of functions in patterning in the directive axis, rather than, or in addition to, the oral/aboral axis (drawn from Ryan et al. 2007).

(Fig. 5B) (Matus et al. 2006a). On the opposite side, the Nematostella homolog of yet another class of TGFb antagonist, gremlin, and another TGFb ligand, gdf5-like are again expressed in complex and semioverlapping patterns (Rentzsch et al. 2006). It should be noted here that the differing expression patterns reported by Matus et al. (2006a) and Rentzsch et al.

(2006) for Nematostella gremlin are the result of working on different paralogs of the gene. The gene described by the former authors is expressed at a time and in a manner that precludes a role in axis formation and therefore is not discussed here. Analyses in heterologous systems imply that these molecules function in Nematostella as do their

708

orthologs in flies and vertebrates—for example, both NvChordin and NvGremlin can antagonize the ability of NvDpp mRNA to ventralize zebrafish (Rentzsch et al. 2006). What this means is that, ‘‘Taken together, the expression pattern of NvGrm is compatible with the generation of two endodermal BMP-like activity gradients: as an antagonist of NvGDF-5-like along the oral-aboral axis and as an antagonist of NvDpp along the directive axis (Rentzsch et al. 2006).’’ In addition to components of TGFb signaling systems, the Martindale laboratory has catalogued a number of other genes that are differentially expressed in the so-called directive axis in Nematostella. As outlined earlier, the most obvious restriction of the expression domains of the Hox-like genes anthox1a, anthox7 and anthox8 is in the directive axis (Finnerty et al. 2004; Ryan et al. 2007)) and these genes are expressed on the gdf5like þ gremlin side (Rentzsch et al. 2006; Matus et al. 2006a). In addition, homologs of netrin and goosecoid (gsc) are clearly differentially expressed in the second axis (Matus et al. 2006a). The significance of the netrin data is unclear, as this is an extensive family of axon-guidance molecules in bilaterians, but the gsc data are intriguing given that this is a key player in the vertebrate organizer—the small group of cells on the blastopore lip that directs the development of the embryonic body plan. The Nematostella goosecoid gene NvGsc is not expressed during gastrulation, but in the planula is expressed asymmetrically with respect to the directive axis–two bands of expression are seen, the stronger being on the same side as anthox1a/7/8, gdf5-like and gremlin, but a less extensive stripe is seen on the opposite side (i.e., where dpp, chordin, and noggin expression is seen).

Cnidarian homologs of Drosophila ‘‘columnar’’ genes Another suite of genes that are potentially informative in terms of the correspondence of axes are those that in Drosophila are known as the ‘‘columnar’’ genes, which include three interacting homeobox genes that subdivide the neuroectoderm into three columns in the D/V axis. In vertebrates, orthologous genes are expressed in strikingly similar patterns, leading to this being considered an evolutionarily conserved system (reviewed in Chan and Jan 1999; Cornell and von Ohlen 2000). As in the case of the head-gap genes, unambiguous members of each of the columnar homeobox gene classes are also present in cnidarians. If the directive axis of cnidarians does correspond to the D/V axis of bilaterians, then one might expect

E. E. Ball et al.

differential expression of these genes in the directive axis, but this is not the case. Instead, in Acropora expression of each of these is restricted along the primary (O/A) axis. A member of one of these gene classes, the ind/Gsh gene cnox2-Am, is expressed in a subset of ectodermal cells resembling neurons in an O/A restricted manner in developing planulae (Hayward et al. 2001). O/A-restricted expression zones with expression in all ectodermal cells have also been demonstrated for members of the other two columnar gene classes, the msh/Msx gene msx3-Am, and for three vnd/NKX2.1 genes (de Jong et al. 2006). Thus, although expression of these genes, which have restricted dorsal/ventral expression in bilaterians, is limited along the oral/aboral axis, the differences are so marked that it is difficult to use their expression to support an O/A ¼ dorsal/ventral correspondence.

Flies in the homology ointment—mere details or critical differences? It is clear that the gene complement of Ureumetazoa was surprisingly complex and did not differ greatly from those of modern chordates (Kortschak et al. 2003; Technau et al. 2005; Nichols et al. 2006). Anthozoan cnidarians appear to have maintained much of this ancestral complexity, and we should not therefore be surprised that Nematostella and Acropora have clear homologs of many of the genes that play key roles in patterning both major axes of bilaterians. Many cnidarians are not radially symmetrical, and among anthozoans a kind of bilaterality is most clearly and convincingly seen. The big question is whether the axes of cnidarians and bilaterians are homologous and, while the assumption has often been that this is the case, the presently available data are equivocal. While it has been argued that the expression data for Hox-like genes and dpp support a direct correspondence of the cnidarian O/A with the bilaterian A/P axis, and the bilaterian D/V axis with a second cnidarian axis most often known as the ‘‘directive’’ axis (Finnerty et al. 2004), this simple interpretation is undermined by more detailed analyses of these and other key genes. At the present time, one can essentially pick gene expression data to support any favoured evolutionary scenario—for example, anterior Hox expression data are consistent with the oral end of the planula being equivalent to the bilaterian anterior pole, Emx expression data with the aboral end being anterior and Vnd data with the oral end being either dorsal or ventral. Significantly, the Nematostella dpp and ‘‘Hox’’ genes are clearly restricted not just in one axis, but in

709


both (Fig. 5). Dpp is expressed on one side of the directive axis, but is also restricted to the oral region during gastrulation, and the combined body of data for TGFb antagonists and other TGFb genes (Rentzsch et al., 2006; Matus et al. 2006a, 2006b) is also inconsistent with any simple correspondence of axes. Moreover, Nematostella anthox1a, anthox7 and anthox8 are not genuine ‘‘Hox’’ genes in the sense of being orthologs of bilaterian genes and their most obvious restriction of expression is in the directive axis (Fig. 5D). While apparent similarities in expression patterns are generally interpreted as reflecting conservation of function, discrepancies between cnidarians and bilaterians are often ignored because of the assumption that these are consequences of the depth of the divergence. However, if preconceptions are discarded, then the presently available data suggest a more intriguing possibility—that cnidarians might reflect a separate experiment (or, perhaps, a separate set of experiments) in axial patterning. Starting with much the same molecular toolbox as Urbilateria but lacking a canonical Hox cluster, they have independently achieved a tremendous diversity of body plans—witness, for example, the spectacular morphologies of siphonophores and octocorals. Independent solutions to the problem of how to pattern along the primary axis may include deployment of Wnts (Kusserow et al. 2005, Miller et al. 2005, Guder et al. 2006) and homologs of both the head gap and columnar genes of bilaterians (de Jong et al. 2006). Martindale (2005) presented an attractive general theory in which cnidarians fit neatly into a general animal plan. In the long run, this may turn out to be true when we have more data on the genes of various cnidarian groups and their expression and better understand how these have changed in the course of cnidarian evolution. Nevertheless, at present some of the data discussed here appear inconsistent with such a theory. We are only beginning to uncover the complexity of gene duplication and gene loss within the Cnidaria and some of the present inconsistencies may well be due to comparisons of paralogs (e.g., the two Nematostella gremlins cited earlier). Further uncertainty is added by the lack of information on protein distribution, rather than message distribution, for the genes considered here. Thus, apparent inconsistencies may disappear as more work is carried out on cnidarian genes and genomes. Nevertheless, in our present state of knowledge it seems premature to try to force cnidarians into the same mould as all other animals.

Acknowledgments The coral research presented here was supported by Grants from the Australian Research Council, both directly to D. M. and E. B. and via both the Centre for the Molecular Genetics of Development and the Centre of Excellence for Coral Reef Studies. We also acknowledge support from the DFG (Schi 277/20) and are especially grateful to Sharyn Wragg for her skilful and patient help with the illustrations.

References Akam M. 1995. Hox genes and the evolution of diverse body plans. Philos Trans R Soc Lond B Biol Sci 349:313–9. Baguna J, Riutort M. 2004. The dawn of bilaterian animals: the case of acoelomorph flatworms. Bioessays 26:1046–57. Boero F, Gravili C, Pagliara P, Pariano S, Bouillon J, Schmid V. 1998. The cnidarian premises of metazoan evolution: from triploblasty, to coelom formation, to metamery. Ital J Zool 65:5–9. Bu¨tschli O. 1884. Bemerkungen zur gastraeatheorie. Morph Jahrb 9:415–427. Chan YM, Jan YN. 1999. Conservation of neurogenic genes and mechanisms. Curr Opin Neurobiol 9:582–8. Chevalier S, Martin A, Leclere L, Amiel A, Houliston E. 2006. Polarised expression of FoxB and FoxQ2 genes during development of the hydrozoan Clytia hemisphaerica. Dev Genes Evol 216:709–20. Chourrout D, et al. 2006. Minimal ProtoHox cluster inferred from bilaterian and cnidarian Hox complements. Nature 442:684–7. Conway Morris S. 2006. Darwin’s dilemma: the realities of the Cambrian ‘explosion’. Philos Trans R Soc Lond B Biol Sci 361:1069–83. Cook CE, Jimenez E, Akam M, Salo E. 2004. The Hox gene complement of acoel flatworms, a basal bilaterian clade. Evol Dev 6:154–63. Cornell RA, Ohlen TV. 2000. Vnd/nkx, ind/gsh, and msh/msx: conserved regulators of dorsoventral neural patterning? Curr Opin Neurobiol 10:63–71. Dalton D, Chadwick R, McGinnis W. 1989. Expression and embryonic function of empty spiracles: a Drosophila homeo box gene with two patterning functions on the anteriorposterior axis of the embryo. Genes Dev 3:1940–56. Degnan BM, Leys SP, Larroux C. 2005. Sponge development and antiquity of animal pattern formation. Integr Comp Biol 45:335–41. de Jong DM, Hislop NR, Hayward DC, Reece-Hoyes JS, Pontynen PC, Ball EE, Miller DJ. 2006. Components of both major axial patterning systems of the Bilateria are differentially expressed along the primary axis of a ‘radiate’ animal, the anthozoan cnidarian Acropora millepora. Dev Biol 298:632–43. Finnerty JR, Pang K, Burton P, Paulson D, Martindale MQ. 2004. Origins of bilateral symmetry: Hox and dpp expression in a sea anemone. Science 304:1335–7.

710 Gauchat D, Mazet F, Berney C, Schummer M, Kreger S, Pawlowski J, Galliot B. 2000. Evolution of Antp-class genes and differential expression of Hydra Hox/paraHox genes in anterior patterning. Proc Natl Acad Sci USA 97:4493–8. Grimmelikhuijzen C, Westfall JA. 1995. The nervous systems of cnidarians. In: Breidback O, Kutsch W, editors. The nervous systems of invertebrates: an evolutionary and comparative approach. Basel: Birkhauser. p 7–24. Gro¨ger H, Schmid V. 2001. Larval development in Cnidaria: a connection to Bilateria? Genesis 29:110–4. Guder C, Philipp I, Lengfeld T, Watanabe H, Hobmayer B, Holstein TW. 2006. The Wnt code: cnidarians signal the way. Oncogene 25:7450–60. Haeckel E. 1874. The gastraea-theory, the phylogenetic classification of the animal kindgdom and the homology of the germ-lamellae. Quart J Microsc Sci 14:142–65. Haeckel E. 1875. Die Gastrula und die Entfuchung der Thiere. Jena Z Naturwiss 9:402–508.

E. E. Ball et al. biology and phylogeny. Berkeley, Los Angeles: University of California Press. p 375–382. Lichtneckert R, Reichert H. 2005. Insights into the urbilaterian brain: conserved genetic patterning mechanisms in insect and vertebrate brain development. Heredity 94:465–77. Manuel M, Le Parco Y. 2000. Homeobox gene diversification in the calcareous sponge, Sycon raphanus. Mol Phylogenet Evol 17:97–107. Martin GR. 1987. Nomenclature for homeobox-containing genes. Nature 325:21–2. Martindale MQ. 2005. The evolution of metazoan axial properties. Nat Rev Genet 6:917–27. Masuda-Nakagawa LM, Groer H, Aerne BL, Schmid V. 2000. The HOX-like gene Cnox2-Pc is expressed at the anterior region in all life cycle stages of the jellyfish Podocoryne carnea. Dev Genes Evol 210:151–6.

Hagadorn JW, Dott RH Jr., Damrow D. 2002. Stranded on a late Cambrian shoreline: medusae from central Wisconsin. Geology 30:147–50.

Matus DQ, Pang K, Marlow H, Dunn CW, Thomsen GH, Martindale MQ. 2006a. Molecular evidence for deep evolutionary roots of bilaterality in animal development. Proc Natl Acad Sci USA 103:11195–200.

Hartmann B, Reichert H, Walldorf U. 2001. Interaction of gap genes in the Drosophila head: tailless regulates expression of empty spiracles in early embryonic patterning and brain development. Mech Dev 109:161–72.

Matus DQ, Thomsen GH, Martindale MQ. 2006b. Dorso/ ventral genes are asymmetrically expressed and involved in germ-layer demarcation during cnidarian gastrulation. Curr Biol 16:499–505.

Hayward DC, Catmull J, Reece-Hoyes JS, Berghammer H, Dodd H, Hann SJ, Miller DJ, Ball EE. 2001. Gene structure and larval expression of cnox-2Am from the coral Acropora millepora. Dev Genes Evol 211:10–9.

Miller DJ, Miles A. 1993. Homeobox genes and the zootype. Nature 365:215–6.

Hayward DC, Samuel G, Pontynen PC, Catmull J, Saint R, Miller DJ, Ball EE. 2002. Localized expression of a dpp/BMP2/4 ortholog in a coral embryo. Proc Natl Acad Sci U S A 99:8106–11.

Miller DJ, Ball EE, Technau U. 2005. Cnidarians and ancestral genetic complexity in the animal kingdom. Trends Genet 21:536–9.

Miller DJ, Ball EE. 2005. Animal evolution: the enigmatic phylum placozoa revisited. Curr Biol 15:R26–8.

Hyman LH. 1940. The Invertebrates: Protozoa through Ctenophora. Metazoa of the tissue grade of construction– the Radiate Phyla - Phylum Cnidaria. New York: McGrawHill Book Company Inc. p 365–641.

Mokady O, Dick MH, Lackschewitz D, Schierwater B, Buss LW. 1998. Over one-half billion years of head conservation? Expression of an ems class gene in Hydractinia symbiolongicarpus (Cnidaria: Hydrozoa). Proc Natl Acad Sci USA 95:3673–8.

Kamm K, Schierwater B, Jakob W, Dellaporta SL, Miller DJ. 2006. Axial patterning and diversification in the cnidaria predate the Hox system. Curr Biol 16:920–6.

Monteiro AS, Schierwater B, Dellaporta SL, Holland PW. 2006. A low diversity of ANTP class homeobox genes in Placozoa. Evol Dev 8:174–82.

Kmita M, Duboule D. 2003. Organizing axes in time and space; 25 years of colinear tinkering. Science 301:331–3.

Mu¨ller P, Yanze N, Schmid V, Spring J. 1999. The homeobox gene Otx of the jellyfish Podocoryne carnea: role of a head gene in striated muscle and evolution. Dev Biol 216:582–94.

Kortschak RD, Samuel G, Saint R, Miller DJ. 2003. EST analysis of the cnidarian Acropora millepora reveals extensive gene loss and rapid sequence divergence in the model invertebrates. Curr Biol 13:2190–5. Kusserow A, et al. 2005. Unexpected complexity of the Wnt gene family in a sea anemone. Nature 433:156–60. Larroux C, Fahey B, Liubicich D, Hinman VF, Gauthier M, Gongora M, Green K, Worheide G, Leys SP, Degnan BM. 2006. Developmental expression of transcription factor genes in a demosponge: insights into the origin of metazoan multicellularity. Evol Dev 8:150–73. Le´vi C. 1963. Gastrulation and larval phylogeny in sponges. In: Dougherty EC, editor. The lower Metazoa: comparative

Narbonne GM. 2005. The Ediacara biota: Neoproterozoic origin of animals and their ecosystems. Annu Rev Earth Planet Sci 33:421–42. Nichols SA, Dirks W, Pearse JS, King N. 2006. Early evolution of animal cell signaling and adhesion genes. Proc Natl Acad Sci USA 103:12451–6. Pantin CFA. 1966. Homology, analogy and chemical identity in the Cnidaria. In: Rees WJ, editor. The Cnidaria and their Evolution. New York: Academic Press. p 1–15. Pearson JC, Lemons D, McGinnis W. 2005. Modulating Hox gene functions during animal body patterning. Nat Rev Genet 6:893–904.

711

Axis determination in Cnidaria Pickerill RK. 1982. Cambrian medusoids from the St John Group, southern New Brunswick: Canadian Geological Survey Paper. 82-1B:71–6.

Salvini-Plawen L, Splechtna H. 1978. On the origin and evolution of the lower metazoa. Z. f. Zool. Systematik Evolutionsforschung 16:40–88.

Prochnik SE, Rokhsar DS, Aboobaker AA. 2007. Evidence for a microRNA expansion in the bilaterian ancestor. Dev Genes Evol 217:73–7.

Schierwater B, Kuhn K. 1998. Homology of Hox genes and the zootype concept in early metazoan evolution. Mol Phylogenet Evol 9:375–81.

Rentzsch F, Anton R, Saina M, Hammerschmidt M, Holstein TW, Technau U. 2006. Asymmetric expression of the BMP antagonists chordin and gremlin in the sea anemone Nematostella vectensis: implications for the evolution of axial patterning. Dev Biol 296:375–87.

Schierwater B. 2005. My favorite adhaerens. Bioessays 27:1294–302.

animal,

Trichoplax

Scott MP. 1992. Vertebrate homeobox gene nomenclature. Cell 71:551–3.

Rieger RM, Ladurner P, Hobmayer B. 2005. A clue to the origin of the Bilateria? Science 307:353–5.

Seimiya M, Ishiguro H, Miura K, Watanabe Y, Kurosawa Y. 1994. Homeobox-containing genes in the most primitive metazoa, the sponges. Eur J Biochem 221:219–25.

Ruddle FH, Bartels JL, Bentley KL, Kappen C, Murtha MT, Pendleton JW. 1994. Evolution of Hox genes. Annu Rev Genet 28:423–42.

Seipel K, Schmid V. 2005. Evolution of striated muscle: jellyfish and the origin of triploblasty. Dev Biol 282:14–26.

Ruiz-Trillo I, Paps J, Loukota M, Ribera C, Jondelius U, Baguna J, Riutort M. 2002. A phylogenetic analysis of myosin heavy chain type II sequences corroborates that Acoela and Nemertodermatida are basal bilaterians. Proc Natl Acad Sci USA 99:11246–51. Russell-Hunter WD. 1979. A life of invertebrates. New York: Macmillan. Ryan JF, Burton PM, Mazza ME, Kwong GK, Mullikin JC, Finnerty JR. 2006. The cnidarian-bilaterian ancestor possessed at least 56 homeoboxes. Evidence from the starlet sea anemone, Nematostella vectensis. Genome Biol 7:R64. Ryan JF, Mazza ME, Pang K, Matus DQ, Baxevanis AD, Martindale MQ, Finnerty JR. 2007. Pre-Bilaterian Origins of the Hox Cluster and the Hox Code: evidence from the Sea Anemone, Nematostella vectensis. PLoS ONE 2:e153.

Seo HC, et al. 2004. Hox cluster disintegration with persistent anteroposterior order of expression in Oikopleura dioica. Nature 431:67–71. Slack JM, Holland PW, Graham CF. 1993. The zootype and the phylotypic stage. Nature 361:490–2. Smith KM, Gee L, Blitz IL, Bode HR. 1999. CnOtx, a member of the Otx gene family, has a role in cell movement in hydra. Dev Biol 212:392–404. Technau, et al. 2005. Maintenance of ancestral complexity and non-metazoan genes in two basal cnidarians. Trends Genet 21:633–9. Telford MJ, Lockyer AE, Cartwright-Finch C, Littlewood DT. 2003. Combined large and small subunit ribosomal RNA phylogenies support a basal position of the acoelomorph flatworms. Proc Biol Sci 270:1077–83. Willmer P. 1990. Invertebrate relationships: patterns in animal evolution. Cambridge: Cambridge University Press.