Genomic, phylogenetic, and cell biological insights into metazoan origins

Over 600 million years ago (Ma), the first multicellular metazoans evolved from their single- celled ancestors. Although not recorded in the fos- sil record, the earliest events in metazoan evolution can be inferred by integrating findings from phylo- genetics, genomics, and cell biology. Comparisons of choanoflagellates (microeukaryote relatives of metazoans) with sponges (the earliest known metazoans) reveal genetic innovations associated with metazoan origins. Among these are the evolu- tion of the gene families required for cell adhesion and cell signalling, the presence of which catalysed the evolution of multicellularity and the functions of which have since been elaborated to regulate cell differentiation, developmental patterning, morphogenesis, and the functional integration of tissues. The most ancient tissues—differentiated epithelia—are found in sponges and evolved before the origin and diversification of modern phyla.


3.1 Introduction

Metazoans are one of evolution’s most dramatic experiments with multicellularity, and yet we know surprisingly little about their origins. The fossil record provides no insight into the biology of the unicellular ancestors of metazoans. Indeed, the relatively abrupt appearance of fossils attributable to modern metazoan phyla over the c. 80 million year span of the Cambrian radiation obscures the sequence of metazoan phylogenesis. Nonetheless, by merging phylogenetics and comparative gen- omics with comparative cell biology, we can infer some of the earliest events in metazoan evolution.
Metazoan origins required at least two innov- ations: the evolution of simple colonies of equipotent cells followed by the organization and integration of cell function and behaviour within an ‘individ- ualized’ organism (Pfeiffer and Bonhoeffer, 2003; King, 2004; Michod, 2007). Both of these phenom- ena required regulated cell signalling, adhesion, and differentiation mechanisms, the origins of which directly address fundamental questions about the evolution of multicellularity.
Metazoan multicellularity evolved independ- ently from that of all other macroscopic lineages. In fact, although unicellular life predominates in all considerations of total biomass and biodiversity, at least 16 separate transitions to multicellularity have occurred during the history of eukaryotic life (King, 2004). The imprint of these separate origins can be seen at the level of phylogenetics, compara- tive genomics, and comparative cell biology. In the following discussion, we review how insights from choanoflagellates and sponges have begun to illuminate some of the earliest events in metazoan history, the origin of multicellularity, and the dif- ferentiation of epithelial tissues.


3.2 Phylogenetics: are there any
‘living models’ of early metazoan ancestors?

3.2.1 The case for choanoflagellates

Choanoflagellates and sponges have classically been thought to straddle the evolutionary divide between metazoans and their unicellular ances- tors. Choanoflagellates, a group of heterotrophic

24

ORIGIN S OF MUL T IC E L L UL A RIT Y 25



microeukaryotes, originally captured the atten- tion of cell biologists for their striking similar- ity to the ‘feeding cells’ (choanocytes) of sponges (James-Clark, 1866; Saville-Kent, 1880–82; see Figure 3.1). This resemblance was first noted by Henry James-Clark in 1866, prompting one of two interpretations: either that sponges and choanoflagellates are derived from an ancestral species that used choanoflagellate-like cells to capture bacterial prey, or that these cell types are only superficially similar and have evolved independently. Subsequent molecular phylogen- etic analyses and comparative genomic data have
firmly established that sponges are metazoans, that metazoans are monophyletic, and that cho- anoflagellates are sister to metazoans (Burger et al., 2003; Medina et al., 2003; Steenkamp et al.,
2006; Moreira et al., 2007; King et al., 2008; Ruiz- Trillo et al., 2008; see Figure 3.2). Furthermore, mitochondrial genome data and species-rich phylogenetic analyses demonstrate that choano- flagellates are not derived from metazoans, but instead represent a distinct lineage that evolved before the origin and diversification of metazoans (Lavrov et al., 2005; Steenkamp et al., 2006; Rokas et al., 2005; Jimenez-Guri et al., 2007).

Figure 3.1 Similarities between choanoflagellates and sponge choanocytes. Choanoflagellates are heterotrophic microeukaryotes that use an apical flagellum to swim and to generate water flow, thus trapping bacterial prey on an actin-filled microvillar collar. Some choanoflagellates, like the species of Proterospongia shown here, have both unicellular (a) and colonial (b) life-history stages. The ultrastructural and functional characteristics of choanoflagellates are conserved in the feeding cells of sponges, choanocytes (c, adapted from Leys and Eerkes-Medrano, 2006), despite vast differences in overall organismal morphology (d). Arrows indicate flagellum and braces indicate the collar of individual cells.

26 AN I M AL EV O L UTI O N




Bilateria Ctenophora Cnidaria Trichoplax
Sponges Choanoflagellates Filasterea
Ichthyosporea

Fungi

Dictyostelium

Figure 3.2 Phylogenetic relationships among metazoans and their close relatives. The preponderance of available evidence supports sponges as the earliest branching metazoan lineage and choanoflagellates as the closest living relatives of the Metazoa.
As such, comparisons with these lineages can uniquely inform us about the nature of the last common metazoan ancestor (white circle) and the last unicellular ancestor of Metazoa (black circle). Other close unicellular relatives of Metazoa, such as Filasterea and Ichthyosporea are poorly understood, but ongoing genome projects for members of these lineages promise to feature prominently in future studies of metazoan origins.



Increasing numbers of molecular phylogen- etic analyses of diverse microeukaryotes have recently revealed a collection of taxa (including Filasterea, Ministeria, Capsaspora owczarzaki, and Ichthyosporea) in the internode between meta- zoans and fungi (Ruiz-Trillo et al., 2004, 2008; Steenkamp et al., 2006; Shalchian-Tabrizi et al.,
2008). Choanoflagellates remain the closest known relatives of Metazoa, and the cell morph- ology of these other diverse microeukaryotes does not provide an obvious link to choanoflag- ellates and metazoans. Nonetheless, molecular phylogenetic analyses reveal that metazoans count diverse single-celled and colony-forming lineages, in addition to choanoflagellates, among their close relatives (Medina et al., 2001; Ruiz- Trillo et al., 2004, 2007, 2008; Steenkamp et al.
2006). A phylogenetically informed comparison of genomes from diverse microeukaryotes with those of metazoans and choanoflagellates prom- ises to further refine our understanding of pre- metazoan genome evolution.
3.2.2 The case for sponges

Arguments in support of sponges as useful ‘liv- ing models’ of the last common metazoan ances- tor (LCMA) stem from their cytological similarities with choanoflagellates, their phylogenetic position, and the antiquity of their fossil record. Of these arguments the evolutionary link between cho- anoflagellates and sponge choanocytes is perhaps the most compelling. The strength of this argu- ment lies in its proven predictive power; it was the hypothesized homology of sponge choanocytes with choanoflagellates that first suggested an evo- lutionary relationship between choanoflagellates and metazoans (to the exclusion of countless other eukaryotes). As discussed above, this predicted relationship has since been independently borne out in phylogenetic analyses.
In addition to the observation that the sponge body plan is organized around the most ancient metazoan cell type, a preponderance of phylogen- etic analyses based upon both morphological and molecular data sets place sponges as the ‘earliest branching’ metazoans (i.e. all other metazoans are more closely related to each other than to sponges: Collins, 1998; Borchiellini et al., 2001; Medina et al.,
2001, 2003; Eernisse and Peterson, 2004; Peterson and Butterfield, 2005; da Silva et al., 2007; Jimenez- Guri et al., 2007; Sperling et al., 2007; Ruiz-Trillo et al., 2008). With this perspective, we can begin to reconcile the ‘primitive’ nature of the modern sponge body plan with the fact that they, like most metazoans, are the product of at least 600 million years of independent evolution. Chance, key inno- vations, or (more likely) both, resulted in drastic- ally different evolutionary outcomes in sponges compared with other metazoans. Only after other metazoans diverged from sponges did traits such as nerves, muscles, tissues, and a digestive gut arise.
The fossil record is consistent with the hypoth- esis that the sponge body plan has remained nearly unchanged since the late Neoproterozoic (reviewed in Carrera and Botting, 2008). Specifically, sponge fossils from between 750 Ma (Reitner and Wörheide,
2002) and 580 Ma (Li et al., 1998) represent the earli- est known metazoan body fossils. By the time of the Cambrian, sponge diversity was high, with

ORIGIN S OF MUL T IC E L L UL A RIT Y 27



spicules from most major sponge groups forming an abundant component of the Cambrian fossil record globally (Gehling and Rigby, 1996).
A second, and less well established, phylo- genetic result that has emerged is the possibility of sponge paraphyly. Under this scenario, some sponge lineages (e.g. calcareous and homoscle- romorph sponges) might be more closely related to eumetazoans than to other, earlier branching sponge lineages (Collins, 1998; Borchiellini et al.,
2001; Medina et al., 2001; Peterson and Butterfield,
2005; Sperling et al., 2007). The evolutionary impli-
cations of sponge paraphyly have been thoroughly
explored and can be distilled into an argument
that all extant metazoans are derived sponges
(Sperling et al., 2007; Nielsen, 2008). However, the
proposition of sponge paraphyly remains tenuous,
in part because analyses that include expressed
sequence tag (EST) and mitochondrial genome
data from the homoscleromorph species Oscarella
carmela strongly support sponge monophyly
(Jimenez-Guri et al., 2007; Lartillot and Philippe,
2008; Lavrov et al., 2008; Ruiz-Trillo et al., 2008;
Wang and Lavrov, 2008).


3.2.3 The controversy

Despite the weight of evidence supporting the placement of sponges at the base of the metazoan tree, placozoans (Dellaporta et al., 2006) and, more recently, ctenophores have also been posited as the earliest branching metazoan phylum (Dunn et al.,
2008). The case for placozoans derives from an ana- lysis of the mitochondrial genome from the only characterized species, Trichoplax adhaerens. This ana- lysis can be distilled into three arguments: (1) like choanoflagellates and unlike most metazoans the mitochondrial genome of T. adhaerens is large (c.
43 kb compared with the 15–24 kb genomes typ- ical of metazoans); (2) it contains an assortment of introns, intergenic spacers, and genes that are lack- ing from all other sequenced metazoan mitochon- drial genomes (albeit, also without orthologues in choanoflagellates or other non-metazoans); and (3) phylogenetic analyses of predicted mitochondrial proteins support T. adhaerens as the earliest branch- ing lineage in an unprecedented clade that also contains sponges, ctenophores, and cnidarians. The
existence of this clade is contradicted by numer- ous independent analyses and can be explained by accelerated rates of evolution within Bilateria (Dellaporta et al., 2006; Wang and Lavrov, 2008). More recently, a genome-scale analysis of predicted proteins from single-copy loci in the draft genomes of the sponge Amphimedon queenslandica and T. adhaerens strongly supported placozoans as an independent lineage that branches after sponges, and before cnidarians (Srivastava et al., 2008). This result and others (Collins, 1998; da Silva et al., 2007) cast doubt on the hypothesis that T. adhaerens is the earliest branching metazoan.
Recently, Dunn et al. (2008) published a phyl- ogeny based upon 150 EST-derived genes that supports ctenophores as branching before two sampled sponge species. This finding would imply one of two unlikely evolutionary scenarios: that the LCMA was much more complex than previously predicted (e.g. it had nerves, muscles, and a digest- ive gut) or that the ctenophore lineage and other eumetazoans underwent extensive convergent evo- lution (Giribet et al., 2007). The former scenario is not supported by the fossil record—sponges would have had to undergo morphological simplification before their appearance as the first recognizable metazoan fossils—and the latter explanation would require the improbable, independent evolution of nerves, muscles, and a gut in the ctenophore and cnidarian/bilaterian lineages. Instead, the weight of evidence places choanoflagellates as the closest living metazoan outgroup, sponges as the earliest branching metazoan phylum, and argues that the choanocyte-based feeding strategy of sponges is ancestral to all Metazoa.


3.3 Reconstructing the genetic toolkit for cell–cell interactions

Choanoflagellates and sponges, by virtue of their positions on the tree of life, bracket metazoan ori- gins and are well situated to help us understand the genetic innovations associated with the transi- tion to multicellularity. Indeed, a wealth of genomic data have begun to pour out from representatives of both of these groups. The single-celled choanoflag- ellate Monosiga brevicollis is the subject of a recently completed genome project (King et al., 2008), and

28 AN I M AL EV O L UTI O N



genome sequencing projects are under way for the freshwater choanoflagellate Monosiga ovata and a colony-forming choanoflagellate, Proterospongia sp. (Ruiz-Trillo et al., 2007). Likewise, sponges have received increasing attention from a genomics per- spective. Pilot EST projects have been completed for the sponges O. carmela and Suberites domuncula, and genome-scale data are available for the sponge A. queenslandica (Nichols et al., 2006; Perina et al., 2006; Adamska et al., 2007a). The juxtaposition of sponge and eumetazoan sequences with those from cho- anoflagellates is beginning to reveal the catalogue of genes present in their common ancestor, thus per- mitting the construction of hypotheses about gen- omic innovations underlying metazoan origins.
A prediction from the field of evo-devo is that genes involved in regulating development play important roles in morphological evolution. One such class of genes includes those involved in the conserved signalling pathways that transduce extracellular cues in diverse metazoans. Although all cellular organisms engage in cell signalling, the pathways required for metazoan development are more elaborate than those of unicellular organisms and distinct from those found in other multicellu- lar lineages (e.g. fungi and plants). Traditionally, seven intercellular signaling pathways are consid- ered unique to and abundant in Metazoa: nuclear hormone receptor, WNT, TGF-E, Jak/STAT, Notch/ Delta, Hedgehog, and RTK (Gerhart, 1998; Barolo and Posakony, 2002; Pires-daSilva and Sommer,
2003). At least six of these seven pathways (Wnt, TGF-E, RTK, Notch, Hedgehog, and Jak-STAT) have conserved components that are expressed in sponges and thus were present in the LCMA (Adell et al., 2003; Nichols et al., 2006; Adamska et al.,
2007a,b; Adell et al., 2007).
In contrast with sponges, only two of the major
metazoan signalling pathways, the RTK pathway
and components of the Hedgehog signalling path-
way, are present in the genome of the choanoflag-
ellate M. brevicollis (King et al., 2008). Despite early
suggestions that RTK signalling might represent
a key innovation in the evolution of metazoans
from their single-celled ancestors (Hunter, 2000),
components of the pathway are abundant in the
choanoflagellate genome (including c. 120 tyrosine
kinase domains, c. 30 tyrosine phosphatases, and
c. 80 SH2 domains: King et al., 2008; Manning et al., 2008, Pincus et al., 2008). In addition, two choanoflagellates (M. brevicollis and M. ovata) con- tain homologues of the proto-oncogene Src and biochemical analyses reveal these homologues to conserve most of the regulatory interactions asso- ciated with metazoan Srcs (Segawa et al., 2006; Li et al., 2008). These observations establish the pres- ence of bona fide tyrosine kinase signalling during the pre-metazoan era.
With the accumulation of genome-scale data from early branching metazoans and their close out- groups, an emerging theme is that the functional protein domains found in developmentally import- ant metazoan signalling and adhesion genes have histories and, presumably, ancestral functions inde- pendent of their roles in metazoan proteins. In other words, these protein domains evolved prior to meta- zoan origins and only later, as a product of domain or exon shuffling (see Patthy, 1999), were linked in the combinations found in the canonical signalling and adhesion proteins of modern metazoans.
One example of this is the case of the secreted ligand Hedgehog. In bilaterians, the Hedgehog signalling pathway is involved in developmental patterning events as diverse as segment polarity in Drosophila and brain, bone, muscle, and gut pattern- ing in vertebrates. The canonical Hedgehog ligand is composed of two protein domains, an N-terminal signalling domain that is released through auto- proteolytic cleavage by a linked C-terminal intein domain (reviewed in Perler, 1998). Analyses of the choanoflagellate, sponge, and cnidarian genomes reveal that the two functional domains known from the bilaterian Hedgehog family evolved independently and were subsequently coupled through domain shuffling early in metazoan evo- lution (Figure 3.3). Specifically, the genomes of the choanoflagellate M. brevicollis and the sponge A. queenslandica encode the Hedgehog N-terminal and C-terminal domains on separate, unrelated proteins, whereas the cnidarian Nematostella vect- ensis has orthologues of these proteins in addition to true Hedgehog proteins typical of bilaterians (Adamska et al., 2007a; King et al., 2008; Matus et al.,
2008). This pattern suggests that the Hedgehog gene family evolved through domain shuffling after the divergence of sponges from other metazoans and


Transmembrane domain
Immunoglobulin/EGF
cassette
SH2 domain EC domain vWA domain
Hh N-terminal domain
Hh C-terminal Hint domain
\
Figure 3.3 Evolution of the Hedgehog ligand by domain shuffling. The two functional domains of the signalling protein Hedgehog, the
N-terminal signal domain (black), and the C-terminal Hint domain (white), evolved on separate proteins in the ancestors of choanoflagellates and animals. One of these ancient proteins, Hedgeling (that links the N-terminal signal peptide to extracellular cadherin domains on a transmembrane protein), has homologues in sponges and cnidarians but was lost in the ancestors of bilaterians. A second ancestral protein, Hoglet, containing only the Hint domain, has been conserved in choanoflagellates, sponges, cnidarians, and bilaterians. Hedgehog, a ubiquitous signalling ligand among bilaterians, is also found in cnidarians and evolved by domain shuffling after the divergence of sponges and eumetazoans (Snell et al., 2006; Adamska et al., 2007a; King et al., 2008; Matus et al., 2008).




that a heretofore unrecognized gene family con- taining the Hedgehog signal peptide linked to extracellular cadherin domains and a transmem- brane domain was in place in the last common ancestor of choanoflagellates and metazoans.
In addition to the Hedgehog protein itself, other components of the Hedgehog signalling pathway can be traced back to the last common ancestor of choanoflagellates and metazoans. For example, sponges and choanoflagellates encode orthologues of the upstream protein that once dispatched releases the Hedgehog protein from signalling cells, and the receptor patched that localizes to the surface of downstream Hedgehog signalling targets (Nichols et al., 2006; King et al., 2008). Additionally,
the sponge O. carmela expresses a gene with simi- larity to the negative regulator, suppressor of fused (Nichols et al., 2006).
The case of hedgehog evolution is illustrative of the evolution of metazoan signalling systems because it demonstrates how seemingly emer- gent metazoan cell signalling machinery might have been assembled piecemeal through domain shuffling and the co-option of genes with differ- ent (if related) ancestral functions. The most excit- ing work lies ahead and will entail the exploration of how these genes function in choanoflagellates, sponges, and cnidarians, and how their functions were altered as they were recruited for their roles in regulating bilaterian development.

30 AN I M AL EV O L UTI O N



3.4 From single cells to epithelia

The general architecture of choanoflagellates and sponge choanocytes extends to eumetazoan planar epithelial cells, with the central flagellum of cho- anoflagellates and the primary cilium of epithelial cells probably sharing a common ancestry (Singla and Reiter, 2006). Neither sponges nor choanoflag- ellates have epithelial tissues as classically defined, but they do form structures that exhibit the rudi- ments of epithelial tissue architecture and epithe- lial cellular machinery. The absence from sponges of abundant intercellular junctions and a basement membrane, two features that contribute to the mechanical and absorptive/transport properties of eumetazoan epithelial cells, has led to their char- acterization as lacking epithelia. Instead they are considered to be a somewhat loose association of cells in which the internal environment (i.e. meso- hyl) is, at least in terms of ionic homeostasis, undif- ferentiated from the external environment (e.g. seawater: de Ceccatty, 1974; Cereijido et al., 2004). Nonetheless, sponges have cell layers such as those formed by choanocytes (i.e. the choanoderm) that are specialized for filter-feeding. In addition, these cell layers are packed closely together (particularly in embryos) with highly regular paracellular spaces suggesting that they have the same kind of direct intercellular interactions that characterize eumeta- zoan epithelia and are not simply embedded in a
common extracellular matrix (sensu Harwood and
Coates, 2004; Figure 3.4).
Contrary to the dogma that sponges lack func-
tional epithelia is the observation that the cell
layers that line their various body cavities are dif-
ferentiated. For example, in addition to the cho-
anoderm, the body surface, internal water canals,
and spermatic cysts are lined by T-shaped pinaco-
cyte cells, whereas oocytes and embryos are often
encased in a layer of large, cuboidal follicle cells.
Furthermore, in some species the basal epithelium
is uniquely differentiated and larvae develop an
outer presumptive epithelium composed of col-
umnar cells more than 15 µm high and c. 2 µm in
diameter (e.g. O. carmela; SAN, personal observa-
tion). The morphological differences between as
many as five presumptive sponge epithelial tissues
plausibly reflect functional differences.
Amongst choanoflagellates that form multicelled
colonies, colony architecture varies between spe-
cies, with cells typically connected at their lateral
surfaces to form two-dimensional ‘chains’ or by
the bundling of secreted extracellular pedicels to
form rosettes (Figure 3.5). Additionally, colonies of
the genus Proterospongia form spherical clusters of
polarized cells in direct contact with each other.
Cell contacts in colonies from Codosiga botrytis
and Desmarella moniliformis comprise cytoplasmic
bridges (Hibberd, 1975; Karpov and Coupe, 1998)
and therefore differ from the protein-plaque-based

Figure 3.4 Transmission electron micrographs of larval and adult epithelial tissues in the sponge, Oscarella carmela. Like other sponges, the larval (a) and adult (b) epithelial tissues of O. carmela are characterized by closely apposed membranes that have very small, uniform paracellular spaces (arrows). However, in contrast to other sponges, only homoscleromorphs are reported to have a loose, ladder-like basement membrane composed of type IV collagen (arrowhead; Boute et al., 1996). Scale bars are shown; n, nucleus.

ORIGIN S OF MUL T IC E L L UL A RIT Y 31



intercellular junctions typical of eumetazoan epi- thelial tissues. Furthermore, with the exception of an uncorroborated account by William Saville-Kent (1880–82), no choanoflagellate colony is known to display cell differentiation.
Epithelial tissues therefore represent a metazoan innovation that evolved de novo after the evolution of multicellularity. Aspects of choanoflagellate biology hint at the types of cell biological phenom- ena that might have laid the foundation for epithe- lial origins. The capacity of most choanoflagellates to adhere to surfaces suggests the presence of a ubiquitous adhesion mechanism that might have been co-opted to support intercellular adhesion in diverse choanoflagellate lineages and in the lin- eage leading to Metazoa. This is consistent with the discovery of more than 23 cadherin genes and a diversity of predicted proteins with C-type lectin,
immunoglobulin, D-integrin, collagen, fibronec- tin, and laminin adhesion domains encoded by the genome of the exclusively unicellular species, M. brevicollis (Abedin and King, 2008; King et al.,
2008).
Only one group of sponges, the homosclero-
morphs, has been argued to have a bona fide epi-
thelium, complete with intercellular junctions in
the larva and a basement membrane underlying
larval and adult tissues (Boute et al., 1996; Boury-
Esnault et al., 2003). However, due to uncertainty
about the phylogenetic position of this group (see
Section 3.2), it is unclear if other sponges have
lost these epithelial features or if homosclero-
morphs are more closely allied with eumetazoans.
Another possibility is that the molecular machin-
ery characteristic of intercellular junctions and
the basement membrane in eumetazoan epithelia

Figure 3.5 Diversity of colony morphology in choanoflagellates. All known species of choanoflagellates have unicellular life-history stages, but some species are also capable of forming simple evidently undifferentiated colonies. Among these species, colony morphology is diverse and suggestive of independent evolutionary origins. For example, colony morphology can vary from ‘chains’ (a) or ‘balls’ (b) of athecate cells
in direct contact with each other to thecate cells embedding in a common gelatinous matrix (c) or sharing a common stalk (d).

32 AN I M AL EV O L UTI O N



is in place in all sponges, but is not sufficiently concentrated to be detected as an electron dense structure under transmission electron microscopy. Indeed, diverse adhesion gene families, and other cell junction components, are widely conserved in sponges (reviewed in Bowers-Morrow et al., 2004). Furthermore, in Ephydatia muelleri—a species that lacks electron-dense epithelial cell junctions and a basement membrane—immunofluorescent stain- ing reveals that actin plaques characteristic of adherens junctions (Yap et al., 1997) are present at points of cell contact in the pinacoderm (Elliot and Leys, 2007).
Epithelial tissues were the first metazoan tis- sue type to evolve and are thought to have been required for body plan diversification by allowing early metazoans to compartmentalize and regu- late physiological homeostasis within and between body compartments (Tyler, 2003). Sponges and cho- anoflagellates exclusively share the ancient charac- teristic of collared cells, yet when viewed at the cell biological level the gulf between their morphology and that of other metazoans is not as wide as it may seem. For example, it seems that the fundamental characteristics of epithelial tissues can be traced to the rudimentary epithelia of sponges and, to some extent, to the simple, undifferentiated colonies of choanoflagellates.


3.5 The biology of the earliest metazoan ancestors

The study of metazoan origins is still in its infancy, despite recent advances in our understanding catalysed by genome-scale data from choanoflag- ellates, sponges, and other early branching meta- zoan phyla. It is premature to assume that any one (or few) species is(are) representative of each phylum—the diversity within these groups is high and phylogenetic divergences are deep—so an immediate goal is to acquire genomic data from a more representative sampling. Nevertheless, from the available data we can begin to reconstruct the minimal genomic, cell, and developmental charac- teristics of the first metazoans.
The cell biology of the last common ancestors of choanoflagellates and metazoans probably resem- bled that of modern choanoflagellates. However,
we do not know whether the last common ances- tor of choanoflagellates and metazoans was cap- able of forming simple multicelled colonies like those formed by some choanoflagellate species and other microeukaryote relatives of metazoans (Leadbeater, 1983; Ruiz-Trillo et al., 2007). A key to addressing this question may be to determine the molecular mechanisms underlying cell interac- tions in diverse colony-forming choanoflagellates species.
If choanoflagellate-like cells are the most ancient metazoan characteristic and the LCMA was a bacterivorous filter-feeder, was the LCMA more like a choanoflagellate or a sponge? Sponges and other metazoans share many developmental and genomic characteristics that choanoflagellates lack. Specifically, sexual reproduction in sponges is typical of other metazoans in that it involves the fusion of a large, nutritive egg with a small, motile sperm to produce an embryo that under- goes programmed patterns of cleavage, cell dif- ferentiation, and morphogenetic patterning. Also, many components of the molecular machinery that regulate development in other metazoans are con- served in sponges and, in some cases, expressed during development (Nichols et al., 2006; Adamska et al., 2007a,b). In contrast, sex is undocumented (though likely) in choanoflagellates, and there is no record of gametic differentiation or of develop- ment beyond the formation of simple colonies in some species. The genome of the choanoflagellate M. brevicollis also encodes few genes with hom- ology to those that regulate development in meta- zoans. Integrating these data provides an early impression of a sponge-like LCMA, and suggests a series of specific developmental and cell biological innovations that separate modern metazoans from their single-celled ancestors


3.6 Acknowledgements

We are grateful to R. Howson for contributing to the artwork used in Figure 3.5. Financial support was provided by the American Cancer Society (SAN), the Miller Institute for Basic Research in Science (MJD), Richard Melmon and the Gordon and Betty Moore Foundation Marine Microbiology Initiative (NK).