The basic characteristics of embryonic process throughout Metazoa are considered with focus on those aspects that provide insight into how cell specification occurs in the initial stages of development. There appear to be three major types of embryogenesis: Type 1, a general form characteristic of most invertebrate taxa of today, in which lineage plays an important role in the spatial organization of the early embryo, and cell specification occurs in situ, by both autonomous and conditional mechanisms; Type 2, the vertebrate form of embryogenesis, which proceeds by mechanisms that are essentially independent of cell lineage, in which diffusible morphogens and extensive early cell migration are particularly important; Type 3, the form exemplified by long germ band insects in which several different regulatory mechanisms are used to generate precise patterns of nuclear gene expression prior to cellularization. Evolutionary implications of the phylogenetic distribution of these types of embryogenesis are considered. Regionally expressed homeodomain regulators are utilized in all three types of embryo, in similar ways in later and postembryonic development, but in different ways in early embryonic development. A specific downstream molecular function for this class of regulator is proposed, based on evidence obtained in vertebrate systems. This provides a route by which to approach the comparative regulatory strategies underlying the three major types of embryogenesis.

As we learn how embryos work, we begin to see the relations between their external cellular and morphological processes and the underlying regulatory mechanisms. How embryos of diverse forms utilize their cell lineage, how they position differentially functioning cells with respect to their three-dimensional body plans, and how and when they initially carry out cell specification, reveal functions that must be served by their developmental gene regulatory systems. The primary role of the regulatory systems that drive the initial construction of the embryo is of course to establish differential patterns of gene expression in cells located in different regions. However, an embryo is not simply equivalent to a set of differentiating cells, even a spatially organized set. A particular function of embryonic cells is to interact in specific ways, in order to generate morphological structure. The nature of the developmental process provides information both about how an embryo arranges the presentation of different sets of active transcriptional regulators in its different cells, and about the location and role of specific intercellular ligand-receptor interactions, that in turn will further affect the internal regulatory milieu of these cells.

While the relation between external embryonic character and underlying regulatory process is in principle evident, it is also intellectually disturbing, because metazoan embryos differ so greatly in their external character. Caenorhabditis elegans embryos have an invariant cell lineage while the cell lineage of a chicken or mouse or fish embryo is always different from that of another of the same species; portions of sea urchin or jellyfish embryos can regulate to generate whole new embryos, while equivalent portions of ascidian or annelid embryos cannot; Drosophila embryos specify elegant spatial patterns of gene expression before there are any cells to interact, while in Xenopus or sea urchin embryos the initial spatial diversification of gene expression depends causally and extensively on intercellular interaction. The two fundamental issues that I shall try to address here are: what are the common regulatory organizations that underlie the diversity of observed processes in metazoan embryos, and how might these organizations have arisen in evolution; that is, is there a simple phylogenetic sense that one might make out of their comparative distribution. These are not necessarily intertwined issues, but I shall deal with them as such. As our predecessors in embryology were well aware a century ago, comparative considerations can be immensely illuminating, and, even at the advanced state to which we have been carried by our superb molecular technology, it might still be said that we need all the help we can get in trying to understand embryonic development.

I shall approach these problems by the following route. There appears to be an associated set of characters that define a canonical type of embryonic process found in many, though not all, invertebrate groups. One can reasonably infer, on present knowledge, the outlines of the molecular modus operand! of these embryos. However, I shall argue that their regulatory organization is clearly not directly applicable either to advanced insect or to vertebrate embryos, nor are most of the mechanisms by which insect and vertebrate embryos initially organize themselves directly applicable to one another. The phylogenetic occurrence of these three embryo types leads to the strong conclusion that advanced insect and vertebrate embryogenesis are derived processes that differ, amongst other things, in their initial use of homeo-domain regulators (i.e. the DNA-binding regulatory protein products of homeobox genes). I suggest a specific downstream developmental function for homeodomain regulators, and this can be utilized to explore further the general mechanisms that in each of these three basic forms direct the transformation of egg into organism.

Those characters that are most useful and most important to consider reveal the processes of cell specification as embryogenesis begins, i.e. the processes by which are initially established the identity, and the fates, of the progeny of individual blastomeres. An analysis of such characters in a recent comparative review (Davidson, 1990) showed that similar regulatory strategies apparently underlie embryogenesis in four entirely unrelated organisms, viz the nematode Caenor-habditis, the sea urchin Strongylocentrotus, the neogastropod mollusc Ilyanassa, and the ascidian Ciona. By most criteria this is of course an impossibly diverse group of organisms, and of embryonic forms. Nematodes are typically direct developing terrestrial animals of distant protestóme affinity, that lack a true (i.e. mesoderm-lined) coelom, and the embryos of which display a unique non-spiral form of cleavage; while neogastropods such as Ilyanassa are advanced marine coelomate protostomes that utilize classic spiral cleavage in embryogenesis and give rise to a special form of free-living larva. The ascidian Ciona and the sea urchin Strongylocentrotus are both indirectly developing lower deuterostomes, but in traditional classifications the embryos of these groups present an archetypical contrast: sea urchin embryos are renowned for their amazing regulative abilities, while ascidian embryos are famous for the opposite quality. Yet, if the focus is set upon certain essential features of early development, there emerges a distinctive suite of key developmental characters that is shared by all four of these disparate forms, as follows, (i) Within each species, cleavage is more or less invariant, meaning that the cleavage planes are in. approximately the same positions in each individual relative to the axes of the egg, and in the undisturbed embryo the lineages descended from specific blastomeres display generally predictable, invariant fates, (ii) One axis of the egg is preformed, while the second is set up after fertilization, by processes that are various and specific to each phyletic group (the mechanism of second axis formation is thus not a general or fundamental character), (iii) Where studied with molecular markers, histotypic specification and the appearance of an initial set of differentially functioning cell types is found to occur before gastrulation in all regions of the embryo, and to precede any large-scale embryonic cell migration (though when cells do become migratory they may add to or change their differentiated properties further), (iv) Some founder cells of the lineages of which the early embryo is composed are specified autonomously, while others, within the context of the invariant lineage, are specified conditionally. Since this last character leads us directly to the underlying embryonic processes, it is important that at the outset of this discussion we explore the implications of these fundamental modes of initial blastomere specification.

Modes of blastomere specification, and the basis of regulative development

Automonous specification signifies a process in which the identity of a lineage founder cell depends only on its internal constituents, and is independent of contact with other coexistent cells. Thus if a blastomere inherits maternal factors that are necessary and sufficient for its specification, it will exhibit the same functions irrespective of what other cells are next to it. If isolated, its progeny will do so even in culture, given that conditions are acceptable. In contrast, the identity of a conditionally specified founder cell depends at least in part on local intercellular interactions. The state of differentiation displayed by lineages descendant from a conditionally specified founder blastomere thus depends on the location in the embryo of this blastomere. In embryos with invariant lineage, such interactions are generally short range (i.e. compared to blastomere diameter). Thus changing the immediate neighbors of a prospective founder blastomere can result in a change in its own identity, as it may receive different signals from the blastomeres to which it is now apposed than it would have in its original position. To the extent that an embryo can select among a number of different potential cell fates by specific intercellular interactions, it does not have to specify all of these fates autonomously, but to position pluripotential cells in the right place that can give the right response. Thus, ab initio, founder cells that are to be specified conditionally possess more potentialities than they will display in any given embryo, normal or chimeric, while, on the other hand, autonomously specified founder cells have no functional need for embryonic plasticity. There are two corollaries that are important for what follows. First, in embryos with invariant cell lineage, founder cell specification, including conditional specification, always begins during cleavage, and it follows that the selective intercellular interactions also must occur during cleavage. Second, because conditionally specified blastomeres are initially pluripotential, they can generate what embryologists term regulative development, as perceived in ectopic or chimeric blastomere recombinations, or in partial embryos. In these situations, blastomeres normally subject to conditional specification can give rise to cell types that would be missing were the lineages normally generating these cell types their sole possible source (see Davidson, 1989, 1990 for discussion and examples). The observation of regulative behavior reliably indicates a conditional specification process. This is useful, since it is unfortunately often the case that no detailed experimental examination of blastomere potentialities has been carried out on embryos belonging to many of the less well studied groups.

Autonomous specification and lineage specific gene expression

We now consider briefly how embryos that display all of the external characters indicated above are likely to work, with illustrative examples drawn in each case from several of the four embryos so far mentioned. For convenience in the following I shall refer to embryos that display these characters as Type 1 embryos. In Type 1 embryos the means by which the primordial cytoarchitectural polarity of the egg is transmuted into an axial feature of the body plan of the embryo, is the autonomous specification of lineage founder cells at (at least) one pole of the original egg axis. All four of our initial Type 1 embryos provide examples: in C. elegans, in which the A/P axis is primordial, the autonomously specified germline (P) and body wall muscle (D, Cap and Cpp) lineages arise at or near the posterior end of the egg (Strome, 1989; Wood, 1991; reviewed in Davidson, 1986, 198-212, 436-445); in ascidians the endodermal lineages (A6.1, A6.3, B6.1 founder cells) arise at the vegetal pole of the primordial D/V (=An/Vg) axis (Jeffery, 1990; Whittaker, 1990); in molluscan eggs endoderm lineages also arise at the vegetal pole of the primordial An/Vg axis; and in sea urchin embryos the autonomously specified skeleto-genic founder cells (VAMk, VOMk and VLMk cells) arise at the vegetal pole of the primordial An/Vg axis. Autonomously specified founder cells may also be formed at (at least) one pole of the second axis of the egg. In ascidian eggs, for example, one of the most intensely studied cases of an autonomous embryonic cell lineage is that giving rise to the larval tail muscles, the founder cells for which arise within the lineages of the bilateral B4.1 cells formed at the posterior pole of the second, or A/P axis of the embryo. Extensive evidence (Meedel et al. 1987; reviewed in Davidson, 1986, 430-436, 489-491; Davidson, 1990) indicates that specification of the founder cells depends on maternal factors, localized in the particular region of egg cytoplasm that is inherited by the B4.1 cells. The localization of these factors -with respect to the invariant cleavage planes -occurs through a cytoskeletal cortical reorganization of the egg, following fertilization (reviewed by Davidson, 1986, 489-493; Bates and Jeffery, 1988; Jeffery, 1990). The Ilyanassa embryo provides a second example of an autonomous lineage arising at a pole of the second axis, viz the specification by maternal factors of the mesoderm germ band founder cells, as these factors are sequestered into the D macromere and then certain of its progeny (reviewed in Davidson, 1986, 453-473).

Studies with molecular markers demonstrate the early expression of cell-type-specific genes in the autonomous lineages of C. elegans and sea urchin embryos. For example in C. elegans the terminal muscle differentiation regulator MyoD begins to be expressed at mid-cleavage in the immediate daughters of the body wall muscle founder cells, i.e. within one cell cycle following the segregation of the D, Cap and Cpp cells from sister cells of other lineages (Krause et al. 1990). In sea urchin (S. purpuratus) embryos, mRNA for SM50, a skeletal matrix protein, begins to accumulate within 2-3 cleavage cycles after completion of segregation of the skeletogenic founder cells at 5th cleavage (Benson et al. 1986; Killian and Wilt, 1989). Furthermore, as might be expected if autonomously specified blastomeres contain localized regulatory factors that control the activation of their lineage-specific genes, exogenous fusion genes under the control of the relevant marker gene regulatory sequences are activated appropriately in autonomous lineages. This was demonstrated with a MyoD-lacZ gene fusion in C. elegans (Krause et al. 1990), and with an SM50-CAT gene fusion in S. purpuratus (Sucov et al. 1988). Gene regulatory factors that are sequestered into the cytoplasmic domains of autonomous founder cells would be divided up amongst the clonal descendants of their lineages as cleavage progresses. As I discussed earlier (Davidson, 1990), cytochalasin experiments that have been carried out on ascidian and nematode eggs support this mechanism of autonomous founder cell specification. These experiments show in essence that at whatever cleavage stage cytokinesis is blocked by cytochalasin treatment, those blastomeres that then contain cytoplasmic regulatory factors destined for autonomous lineages ultimately express their lineage-specific markers (for C. elegans, see Cowan and McIntosh, 1985; Edgar and McGhee, 1986; for ascidians, Whittaker, 1973, 1980, 1990; Crowther and Whittaker, 1986; Meedel et al. 1987; Nishikata et al. 1988; Jeffery, 1989). That is the regulatory factors function in whatever nuclei are included within the domains bounded by the cell membranes present when the drug was added. Thus in the undisturbed Type 1 embryo, the correct positional location of these gene regulatory domains depends directly on the location of the cleavage planes bounding the autonomously specified founder cells, with respect to the axes of the egg.

Conditional specification and the spatial regulation of gene expression in Type 1 embryos

Founder cells of Type 1 embryos that are specified conditionally may constitute a minor fraction of the embryo, as in ascidian embryos, or far more commonly, a large fraction as in C. elegans or sea urchin embryos. Lineages of the ascidian embryo that are conditionally specified include those responsible for brain, sensory pigment cells and terminal caudal tail muscle formation (see, e.g. Meedel et al. 1987; Nishida and Satoh, 1989; other references cited in Davidson, 1990). In C. elegans probably the entire set of anterior (AB) and likely other lineages as well are conditionally specified (it is revealing that in contrast to the autonomous D lineage discussed above, the AB and a large part of the MS lineages are not clonal, in that entirely different cell types arise from sister cells throughout cleavage; Sulston et al. 1983). In sea urchin embryos all of the oral (including neurogenic), and aboral ectoderm lineages, and probably elements of the gut lineages as well are conditionally specified. In ectopic recombinations virtually all embryonic cell fates can be elicited from pluripotential cells that in the unfertilized embryo normally undergo conditional specification (Hbrs-tadius, 1939, 1973; for interpretation see Davidson, 1986, 505-509; Wilt, 1987; Davidson, 1989). Since conditional specification implies regulative ability, and vice versa, differences in the extent of conditional specification account for the large differences amongst embryos in their regulative capacity. For example, ascidian embryos are unable to compensate for missing or killed cleavage stage blastomeres, or to regenerate missing right or left halves (Conklin, 1905; Reverberi et al. 1960; Ortolani, 1987), whereas C. elegans embryos can respecify their asymmetric right/left lineage fates if the positions of the descendants of the AB blastomere are shifted experimentally early in cleavage (Wood, 1991). Similarly, exchange of the positions of certain anterior blastomeres of C. elegans causes reversal of the second (D/V) axis (Priess and Thomson, 1987).

Like autonomously specified lineages in Type 1 embryos, conditionally specified lineages transcriptionally activate particular cell-type-specific marker genes very early in development. Examples include aboral ectoderm markers in the sea urchin embryo such as Cyllla cytoskeletal actin, or the Sped Ca2+-binding protein, the mRNAs for which appear before cleavage is complete (Lynn et al. 1983; Shott et al. 1984; Lee, 1986; Lee et al. 1986; Hickey et al. 1987). Ascidian embryos produce tyrosinase, a marker of the conditionally specified pigment cell lineage, within a few hours of the end of cleavage, at about the same time as markers of autonomously specified lineages appear (Nishida and Satoh, 1989; development in Halocynthia roretzi, as in other ascidians, is so rapid that this is already neural plate stage). The first point is thus that conditional specification in Type 1 embryos results in immediate activation of downstream batteries of histotypic genes, in blastomere lineages that remain located in situ in more or less the same positions in which their founder cells were born (just as in autonomously specified lineages). Immediate conditional activation of such lineage markers has been studied most extensively in cultured combinations of blastomeres in sea urchin embryos. Thus for example blastomeres from the animal pole tiers, that in undisturbed embryos produce only oral and aboral ectodermal derivatives, are made to express a gut alkaline phosphatase marker if combined with skeleto-genic micromeres (Khaner and Wilt, 1990). The same result, and also the expression of SM50 mRNA, can be elicited by treatment of animal blastomeres with Li+ (Livingston and Wilt, 1989).

A second point is that the conditional specification functions monitored by these markers may be mediated by known signal transduction pathways such as the phosphatidyl inositide cycle, and/or tyrosine kinase-linked signal receptors. A role for the phosphatidyl inositol pathway is probably required to explain the frequently observed developmental effects of Li+ (Berridge et al. 1989). Signal transduction effects on gene regulatory systems, mediated by specific ligandreceptor interactions, provide an obvious and general starting point in considering conditional specification in embryos. Recently many examples of molecules belonging to known families of ligands that activate signal transduction systems via specific receptors have been identified in amphibian, mammalian and Drosophila embryos. For Type 1 embryos, a good example is provided by the glp-1 locus of C. elegans (Austin and Kimble, 1989). This gene, the product of which encodes a receptor with EGF-like repeats, is required for the conditional specification of pharyngeal muscle cells descendant from the AB blastomere (Priess et al. 1987), but not for the apparently autonomously specified pharyngeal muscle cells descendant from the MS blastomere. In addition, though there are yet no functional data, mRNAs encoding apparent ligands of the EGF family have been cloned from very early sea urchin embryos (Delgadillo-Reynoso et al. 1989; Grim-wade et al. 1991) as has a member of the TGFβ family that resembles Drosophila dpp (J. L. Micol and E. H. Davidson, unpublished).

Since it is so essential and widespread an aspect of early Type 1 development, one of our most important objectives must be to erect a working hypothesis that provides a general and testable molecular model for conditional specification, and I recently suggested a possible mechanism (Davidson, 1989, 1990). Given a range of receptors, the effect of a particular ligand interaction in the early Type 1 embryo can be accounted for as follows (Davidson, 1989, 1990): prior to conditional specification blastomeres can be envisioned to contain key maternal gene regulatory factors (or their maternal mRNAs) but these factors initially exist in a cryptic or inactive state. Selection of the appropriate gene expression program would thus depend on which set of such factors a specific inductive signal transduction will affect, i.e. by activating them by means of covalent modification, mobilization of their access to the nucleus from a site of sequestration in the cytoplasm, or the presentation of an active cofactor. The initial degree of pluripotentiality of such a blastomere might be defined directly by the range of such factors or factor sets that it contains. The selection of the particular pattern of differential gene expression would then depend on the particular intercellular interaction in which a blastomere at a given position in the canonical cleavage pattern engages (this of course includes interactions with regulatory factors that function negatively, in that they repress gene expression, as well as positive interactions). Were a blastomere moved to another position it might encounter a different ligand, presented by a different neighboring blastomere, and it would be specified differently, which we might notice as a remarkable example of regulative behavior.

The organization of the initial body plan in Type 1 embryos may thus begin with the presentation of particular ligands on the axially positioned, autonomously specified founder cells. These will affect the conditional specification of the adjacent blastomeres, thus selecting their identity. Once again, examples are at hand from disparate Type 1 embryos, some of which have already been touched upon. For instance, the ability of the autonomously specified skeletogenic precursors of cleavage stage sea urchin embryos to specify gut lineage identity in recombinations with pluripotential animal cap blastomeres probably reveals a normal function. Thus the skeletogenic founder cells are in contact with the prospective gut founders in the undisturbed embryo (Davidson, 1989). The experiments of Wood (1991) cited above in which right/left handedness is shifted in early C. elegans embryos by changing blastomere positions can be interpreted similarly. In essence this operation will cause the lefthand AB derivative (ABpl) to contact the autonomously specified P3 cell instead of the right-hand derivative (ABpr), and contacts with the E blastomere may be reversed as well; the prediction might be that the P3 (or its descendants) provide the ligands for the conditional specification of these AB descendants. Similarly, Priess and Thompson (1987) infer that the interactions required to specify anterior pharyngeal muscle precursors in this embryo occur with contiguous derivatives of posterior lineages that by their test, produce the more caudal pharyngeal muscle cells autonomously. As a third example, cells of the autonomously specified D quadrant lineage of Ilyanassa are required for the correct conditional specification of the micromere lineages that give rise to eye, foot and shell gland (reviewed by Davidson, 1986, 455-462). Autonomous specification in Type 1 embryos may always involve specific ligand presentation, as well as the internal institution of a cell-type-specific pattern of gene expression.

Early transcription in Type 1 embryos

This discussion of the regulatory mechanisms underlying Type 1 early development indicates that transcription must be required during cleavage, for expression of ligands, receptors, and gene regulatory factors or cofactors, as well as of the characteristic early sets of downstream, cell-type-specific lineage markers. In fact, in sea urchin, ascidian, molluscan (reviewed in Davidson, 1986, 126-140) and C. elegans embryos (Schauer and Wood, 1990), transcriptional activity has been demonstrated at early cleavage stages. Indeed, no period of transcriptional quiescence preceding a ‘midblastula transition’ has ever been observed in a Type 1 embryo.

To summarize, Type 1 embryos generate spatial patterns of differential, cell-type-specific gene expression by processes of both conditional and autonomous specification that take place in situ during cleavage. These processes produce a three-dimensional array of functional cytoplasmic gene regulatory domains, that are positioned in space by the canonical cleavage planes that separate lineage founder cells of differing fate. A set of criteria that in capsulized form define Type 1 embryogenesis is shown in Table 1.

Table 1.

External embryo characters used for type classification

External embryo characters used for type classification
External embryo characters used for type classification

Variable lineage, cell migration and late gene expression in vertebrate embryos

The suite of mechanisms by which Type 1 embryos operate is not universally found. Thus the key biological characters that provide the conditions for the process of early development just outlined are in some forms of embryo absent, or are replaced by other characters. For example, no vertebrate embryo displays an invariant cell lineage. In each individual of a vertebrate species, the cleavage pattern may differ, at least after the first few divisions. More importantly, given structures are always composed of cells of different exact lineage. Even far into cleavage, cells labeled with a lineage tracer are found to give rise to a variety of different cell types, that in quantitative and qualitative detail is unique to each embryo, although majority fates can be correlated with given regions of the blastula (reviewed by Davidson, 1986, 246-268; 1990; examples of recent results obtained with cells labeled for lineage tracing by various methods include for zebrafish, Kimmel and Warga, 1986; Warga and Kimmel, 1990; Kimmel, 1989; Kimmel et al. 1990; for Xenopus, Keller, 1976; Wetts and Fraser, 1989; Dale and Slack, 1987; for chick, Stern and Canning, 1990; Stern et al. 1988; Stern, 1990; for mouse, Gardner and Rossant, 1979; Lawson et al. 1986; Turner et al. 1990; Walsh and Cepko, 1990). Two of the most important sources of internal structures, tissues and organs in vertebrate embryos, the mesoderm and the neural crest, are formed from pluripotential migratory cell layers or mesenchymal cells. The neural crest provides a paradigm of a vertebrate embryological process in which the fate of individual cells is in large part determined environmentally during their migration or at their final destination (see, e.g. Anderson, 1989; Bronner-Fraser and Fraser, 1989; Le Douarin, 1990). In large measure, cell-type-specification throughout the vertebrate embryo occurs after the general migratory dispersion and reorganization of sheets of cells into multilayered structures that is the result of gastrulation, and that continues into neurulation and beyond. Some earlier, more general specification processes as of whole germ layers or regional anlagen can be identified, e.g. the inductive specification of dorsal axial mesoderm in Xenopus, which occurs between 5th cleavage and the midblastula stage (i.e. 12th cleavage; see Gimlich and Gerhart, 1984; Gurdon et al. 1985; Jones and Woodland, 1987; Godsave and Slack, 1991). In the chick the initial decision of which cells are to generate the embryonic mesoderm and endoderm vs the ectoderm is also apparently made before gastrulation (Stern, 1990; Stern and Canning, 1990). However, cell-type-specific molecular markers of differentiated function can be detected in vertebrate embryos only at early gastrula stages, or thereafter. Here most available data are from Xenopus. The earliest example is an epidermal cyto-keratin, zygotic mRNA for which appears at late blastula-early gastrula stage (stage 9-10; Jamrich et al. 1987). Significantly, activation of this gene is an autonomous function, in that it occurs in isolated cultured cells (Sargent et al. 1986). MyoD also begins to be expressed at early gastrula stage in presumptive skeletal muscle (Hopwood et al. 1989), followed later by muscle actins (Mohun et al. 1984). Most known histotypic markers first appear at the end of gastulation or in the neurula stage, e.g. several CNS neural markers (Sharpe and Gurdon, 1990) or a notochord keratin marker (La Flamme et al. 1988).

Vertebrate embryos differ from Type 1 embryos in utilizing specification processes that are essentially independent of lineage and that follow migratory relocation, but they do not differ in all respects. As in Type 1 embryos, the axis of the egg is prespecified in teleost, avian, and amphibian eggs (for references see Davidson, 1990). Some autonomously specified lineages may exist in amphibian eggs, e.g. ectoderm precursors at the animal pole of the primordial A/V axis, or at the vegetal pole, the lineages that give rise clonally to endoderm. However, unlike autonomous Type 1 lineages, these endoderm precursors remain pluripotent and plastic in fate, and they may not represent truly autonomous specifications (Davidson, 1986, 250-257; Sargent et al. 1986; Snape et al. 1987; Wylie et al. 1987). Cytoplasm of the dorsal posterior cells of Xenopus embryos may also contain determinants for dorsal specification (or at least for inductive interactions that lead to dorsal specification) since this cytoplasm, if transferred to ventral blastomeres, may cause a second dorsal axis to form (Yuge et al. 1990). However, most cell types in all vertebrate embryos are certainly conditionally specified, and in neither zebrafish nor mammals is there any evidence whatsoever for any autonomously specified embryonic cell type, sensu strictu. Nor is there any axial prespecification in the mammalian egg, in which all early blastomeres up to the 8- to 16-cell stage remain totipotent (reviewed in Davidson, 1986, 515-524).

Long-range diffusible morphogens

Vertebrate embryos rely extensively on an additional mechanism that is a fundamental aspect of their method of generating their initial body plans. They appear to operate spatially crude regional identification functions that initially specify major parts, and thus inform immigrant and resident cells that they are to be part of a future head, posterior region, trunk, or dorsal or ventral part of the body. At this level, regional identification is distinct from cell type specification, and must be, since the same cell types appear in all regions. Muscle, or CNS neuroblasts, are formed in head, trunk and tail, for instance, and originate from very many different variable lineages. In early Type 1 embryos, this distinction does not apply, since the initial embryo plan is directly a consequence of the relative positions occupied by the given differentiated cell types specified in cleavage, e.g. muscle or gut in C. elegans, and aboral ectoderm or skeletogenic mesenchyme in sea urchins. Classical and modern studies show that following the establishment of the initial biaxial anisotropies in the early embryo, the regional identification processes of postgastrular vertebrate development are inductive. Regional specification is mediated by diffusible intercellular morphogens, and at least some of the morphogens have been identified in amphibian, avian, and mammalian embryos. These are members of the FGF, TGFβ, and Wnt growth factor families, and retinoic acid (e.g. Kimelman and Kirschner, 1987; Durston et al. 1989; Mitrani et al. 1990; Ruberte et al. 1990; Smith et al. 1990; Green and Smith, 1990; Thomsen et al. 1990; van den Eijnden-Van Raaij et al. 1990; Cooke and Wong, 1991; Ruiz i Altaba and Jessell, 1991; McMahon and Moon, 1989). Within the inductively specified regions such as neural plate, head, and dorsal and ventral mesoderm, all of which are canonical features of neurula stage vertebrate embryos, there then begin to operate the postgastrular morphogenetic programs by which specific structures and tissues are progressively generated, as we discuss below.

The characters general to the vertebrate embryonic process are a distinct set, and this brief sketch suggests that at the level of mechanism vertebrate embryos ‘work’ in a distinct manner, with respect to Type 1 embryos. For ease of communication, in the following I refer to the vertebrate embryonic process as Type 2; a canonical set of characters that identify this form of embryogenesis is listed in Table 1.

Forms of insect embryogenesis

The insects display (to varying extents) a third form of basic embryonic process by which the initial spatial patterns of differential gene expression may be organized. The eggs of all orders of winged (pterygote) insects undergo intralecithal cleavage, forming a yolky syncytium. Outward migration of some of the nuclei late in cleavage gives rise to a syncytial blastoderm (reviewed by Anderson, 1973, 214-217; 1972a,b; this process also occurs in all wingless insect orders except Collembola [springtails], the eggs of which undergo several holoblastic cleavages, but then form the usual blastoderm secondarily, surrounding a yolky syncytium). Cellularization of the pterygote blastoderm occurs as membranes wall off the nuclei from one another and from the yolk mass, which is ultimately incorporated in the midgut. There is no possibility of cellular interaction while the embryo is syncytial, and after cellularization the subsequent cell lineage is variable (e.g. see Beer et al. 1987; Technau et al. 1988; Technau and Campos-Ortega, 1986). The body plan becomes morphologically visible after gastrulation, with the formation of the segmented germ band along the future ventral surface of the larva, in which neuroblasts and mesoderm lie within the epidermal precursors that will give rise to the larval cuticular pattern. The fate maps of all pterygote embryos at the completed germ band stage are remarkably homologous, from primitive orders such as dragonflies to the Díptera (Anderson, 1973, 209-225, 249-253). However, the manner in which the germ band is formed differs greatly amongst the different pterygote orders (reviewed by Sander, 1976). It is largely this variable initial process with which we must be concerned, for it is this process that specifies the body plan of the organism; that results directly in the spatial disposition of differentially functioning cells; and that is comparable with the stages of Type 1 and Type 2 development that we considered above. In ‘long germ band’ insects, e.g. the Díptera, the whole metameric body plan is established more or less simultaneously by subdivision and diversification of the blastoderm. We know that in Drosophila the major head, tail, thoracic and abdominal regions, the future metameric pattern, and dorso-ventral diversification, are already foreshadowed in differential patterns of regulatory gene expression by the time cellularization is complete (reviewed by Akam, 1987; Scott and Carroll, 1987; Ingham, 1988). ‘Intermediate germ band’ insects, examples of which include dragonflies and some beetles, also form the head and thoracic segments of the germ band by direct subdivision of the blastoderm, but most of the abdominal segments are elaborated sequentially from a posterior ‘growth zone’ that contains the anlagen for the posterior ectoderm and mesoderm. In further contrast, insects from several lower hemimetabolous orders (e.g. some beetles, grasshoppers, termites, mantids) initially form a ‘short germ band,’ in which all segments posterior to the antennal region of the head are generated from a posterior growth zone. Among the best known short germ band examples is the grasshopper (Orthoptera). Development from a growth zone is a very different process from specification of the blastoderm prior to cellularization, in that it appears to depend explicitly on intercellular interaction (Sander, 1976; Sander and Lehman, 1988; Tear et al. 1988). It is interesting that the eggs of long germ band insects are formed by meroistic oogenesis (Sander, 1976), i.e. in an ovariole in which most of the oocyte transcripts derive from polyploid nurse cells of germ line origin (reviewed in Davidson, 1986, 307-326), while the ovarioles of short germ band insects lack nurse cells, and at a much slower rate, synthesize their own transcripts. Recent molecular data provide a striking demonstration of the argument that despite the basic contrast in their modes of formation, once completed, the embryos of long and short germ band forms are functionally homologous. Thus, for example the spatial patterns of expression of the AbdA gene is the same in the abdominal segments of a grasshopper, Schistocerca, as in those of Drosophila (Tear et al. 1990); and the metameric expression of the engrailed gene is also the same in the embryos of these species (Patel et al. 1989b).

Conditional specification in Drosophila

In Drosophila many of the morphogenetic processes that follow cellularization depend on intercellular interactions, as in all embryos. For example, the differentiation of individual epidermal cells that generate cuticular pattern within each metamere is under the control of an intercellular signalling system mediated by ‘segment polarity gene’ products (Botas et al. 1988; Martinez-Arias et al. 1988; Hidalgo and Ingham, 1990). Among these genes, for example, wingless encodes a secreted protein that probably functions as a signalling ligand (Cabrera et al. 1987; Rijsewijk et al. 1987); patched encodes a transmembrane protein that might interact with the receptor for the wingless protein (Hooper and Scott, 1989; Nakano et al. 1989; Hidalgo and Ingham, 1990); and others encode transcription factors that likely control the presentation of relevant ligands and receptors. As another example, the product of the decapentaplegic (dpp) gene is a member of the TGFβ family of developmentally active ligands (Padgett et al. 1987), and is required for cells to express dorsal embryonic fates (Irish and Gelbart, 1987; St Johnston and Gelbart, 1987), as well as for inductive functions in midgut morphogenesis (Reuter et al. 1990). A third example is the set of genes required for neuroblast formation within the ventral neurogenic region of the cellular blastoderm, which has been demonstrated to be a conditional specification process that also depends on intercellular signalling (reviewed in Davidson, 1986, 271-280). Among the required ‘neurogenic’ genes are Notch and Delta, which both encode apparent transmembrane proteins with extracellular domains displaying homology to mammalian EGF, and which may mediate intercell interactions by interacting with one another (Artavanis-Tsakonas, 1988; Campos-Ortega, 1990; Johansen et al. 1989; Fehon et al. 1990; Hartenstein and Posakony, 1990). Though genetic evidence provides a somewhat different slant on these processes than did, say, the experiments of Hbrstadius and Spemann and their schools on sea urchin and amphibian embryos, molecular biology suggests that embryonic conditional specification is essentially similar throughout Metazoa, except that sometimes the ligands may be cell bound and sometimes diffusible (see Davidson, 1990). It is the processes of regional nuclear specification that occur before cellularization that distinguish embryogenesis in Drosophila, and by implication in at least other long and intermediate germ band insects (the same processes could be used in the head regions of short germ band insects as well).

Precellular processes by which spatial patterns of gene expression are established

Networks of regulatory gene interaction, which have different features, and which of course interlock (e.g. Carroll et al. 1987; Strecker et al. 1991), specify the anteroposterior metameric organization, the head and tail regions, and dorsoventral organization of the Drosophila egg, all to at least some extent prior to the cellularization of the blastoderm. The primordial A/P polarity of the embryo is reflected in a declining A to P gradient of the maternal bicoid transcription factor (Nüsslein-Vollhard et al. 1987); this is interpreted directly at the level of regulatory DNA-protein interactions by the zygotically expressed ‘gap genes’ so as to set up a series of broad overlapping domains of the transcription factors produced by the gap genes; and their local concentration clines and ratios are in turn ‘read’ in the syncytial blastoderm nuclei to produce transverse premetameric ‘stripes’ of expression of ‘pairrule’ genes such as eve, runt and hairy. These genes then control one another, as well as downstream pair-rule genes such as ftz and the segment polarity signalling system (see reviews of Scott and Carroll, 1987; Ingham, 1988; Reinitz and Levine, 1990). The homeotic genes are also initially controlled by gap gene products, by pair-rule gene products, and later by segment polarity gene products (Scott and Carroll, 1987; Carroll et al. 1988; Harding and Levine, 1988; Ingham, 1988; Reinitz and Levine, 1990). A revealing mechanistic insight derives from analyses of the ≥8 kb eve regulatory domain (Goto et al. 1989; Stanojevic et al. 1989; Jiang et al. 1991; Small et al. 1991), which show that there are separate regulatory subdomains responsible for generating different stripes, and that each subdomain consists of clusters of multiple positively and negatively acting target sequences for gap gene and other transcription factors, including that encoded by eve itself. This is an integrating regulatory system that resolves multiple inputs into a single scalar function, transcription. Since it evidently depends on cooperative factor-DNA interactions it is exquisitely sensitive to factor concentrations in each nucleus, and thus to the shape of the gap gene product concentration profiles and ratios along the A/P axis (Small et al. 1991). The hairy gene is evidently regulated similarly (Howard and Struhl, 1990). The relevant point here is that these interactions take place in a syncytium, and the nuclei communicate by the concentration clines of their diffusible macromolecular regulatory products, rather than by intercellular interactions (Akam, 1987). In the head and tail are found other variations. At the anterior end, different gap genes encoding transcriptional regulators are activated by the bicoid factor before cellularization and may directly control segment polarity genes (viz, ems, otd, and btd;Dalton et al. 1989; Cohen and Jurgens, 1990; Finkelstein and Perrimon, 1990). In the head and the tail, furthermore, a uniformly distributed maternal receptor tyrosine kinase, the product of the torso gene, is regionally activated, apparently by a ligand presented on the terminal follicle cells. The signal transduction system that it controls regulates nuclear expression, again before cellularization, of various downstream genes required for terminal differentiation (including the gap genes tailless and otd’, Casanova and Struhl, 1990; Sprenger et al. 1989; Finkelstein and Perrimon, 1990; Pignoni et al. 1990; Strecker and Lipshitz, 1990; Strecker et al. 1991). Still a different precellular regulatory system controls dorsoventral pattern formation. Here the key transcription regulator is dorsal (Ip et al. 1991). The presentation of dorsal is mediated along the D/V axis by a localized cytoplasmic machinery, generated by the products of a number of maternal genes required for dorsoventral specification, that acts to modify the cryptic dorsal regulator so that it is transported into the cell nuclei only on the ventral side of the blastoderm (Levine, 1988; Steward, 1989; Roth et al. 1990; Rushlow et al. 1989). The downstream targets of the dorsal regulator are again genes encoding other transcription factors, as well as the dpp gene (Ip et al. 1990).

Thus there is a variety of distinct mechanisms utilized to achieve precellular spatial pattern of expression. By means of gene interactions, following the prelocalized biaxial maternal spatial cues, these mechanisms essentially accomplish what Type 1 and Type 2 processes accomplish by entirely different mechanisms, which are not available to the syncytial stage Drosophila embryo. For convenience the processes exemplified in the early Drosophila embryo are referred to as Type 3 embryogenesis in the following, and can be summarized in outline as in Table 1. Their end result is of course the same as of Types 1 and 2 processes, viz, generation of an assembly of spatially organized cells, according to the body plan, that can now mount the intercellular functions required for cell-type-specification and morphogenesis.

A speculative extension of typological classification, and ubiquity of Type 1 process

Thus far we have been concerned with ‘model systems,’ i.e. with relatively well studied embryos, which we assume are archetypical for their respective taxa. In considering other metazoan groups, the available information is rarely adequate. Relatively little molecular biology, developmental genetics, or experimental embryological manipulations such as might directly reveal the nature of specification processes are to be encountered outside the ‘model systems.’ In as much as each type of embryonic process can be defined by an easily observed set of external characters, however, it is worth trying to assess the available descriptive and experimental data so as to arrive at a more inclusive idea of the phylogenetic distribution of these Types. Table 1 summarizes in brief form the sets of external criteria that I have used in this speculative exercise, and Fig. 1 shows its result. The embryological evidence, and references used for these characterizations are very briefly indicated in Notes to Fig. 1, except for the groups whose relatively well-known representatives were considered above, and in Davidson (1990).

Fig. 1.

Phylogenetic chart of Metazoa, with embryo type indicated in color (blue, Type 1; red, Type 2; yellow, Type 3)

Fig. 1.

Phylogenetic chart of Metazoa, with embryo type indicated in color (blue, Type 1; red, Type 2; yellow, Type 3)

Phyletic Aspects: It is impossible to cite a phylogenetic tree, or a cladogram, that is not in some measure and in some regions controversial (see, e.g., Willmer, 1990). The chart is divided horizontally into three main regions. As indicated on the right margin all groups above the bottom dashed line have nerves and mouth; above the second dashed line all groups have bilateral symmetry and mesoderm; above the third dashed line all groups have mesodermal coeloms or coelomic compartments. These are body cavities lined with mesodermal sheaths (e.g., peritoneum) so that internal organs are not situated directly in the body cavities, but rather behind mesodermal boundaries (that are often of major developmental importance). The overall geometrical design of the chart, the angles at which the various lines intersect, and the lengths of the vertical and horizontal stems that connect these lines to the current phyla shown, are entirely arbitrary design features, and have no biological significance. The diagram shown here is particularly based on the following sources: the division of arthropods into crustacean and uniramian phyla reflects the conclusions of Anderson (1973) with respect to arthropod embryological forms; of Manton (1977, 236-291) regarding arthropod structure; and the paleontological assessment of Valentine (1989). The chelicerates (e.g., horseshoe crabs, spiders, ticks) are omitted, since there is little relevant experimental embryological information. The molluscs are not considered fundamentally metameric (e.g., see Barnes, 1980, 462), and like the annelids and the arthropod phyla, can be regarded as descendant from an ancestral assemblage of spirally cleaving, protostomial ancestors. However, neither group is regarded as derived from annelid ancestors. The tentaculates (phoronids, bryozoa, brachiopods) possess some deuterostome characteristics, e.g., their trimerous coeloms, the mesoderm of which arises by a similar process of delamination and outpocketing from the archenteron as in echinoderms and other deuterostomes. However, they also display protostome characteristics, e.g., formation of mouth from the site of blastoporal invagination. Thus the shaded areas indicating the regions occupied by the Deuterostomia and Protostomia, respectively, overlap in the tentaculate area of the chart. The relation of nematodes (and related minor phyla) to the contiguous groups, flatworms and nemerteans, is uncertain, since while flatworms and nemerteans both display obvious spiral cleavage, nematodes have a different characteristic cleavage pattern, but on the other hand they share with nemerteans a pseudocoel body organization (mesodermal body wall but no mesodermal layer surrounding the gut or other visera, so that the body cavity is not enclosed in mesoderm). Flatworms (acoel grade) have mesoderm but neither coelomic nor pseudocoelic cavities. The ctenophores (e.g., comb jellies) and cnidaria (jellyfish, anemones) are radially symmetric animals, and have no true mesoderm. For the deuterostome branch I have followed Jeffries (1986) in placing hemichordates as the most primitive living true deuterostome representatives, and as such the sister group of all remaining deuterostomes, though this view remains controversial. Echinoderms, tunicates (e.g.; ascidians), acraniates (e.g., Amphioxus), and vertebrates can all be considered ‘dexiothetes’; the term refers to the development of organ systems from coelomic domains that are homologous to the left coeloms of hemichordates, as if in ancestral deuterostomes of hemichordate grade the right side had become ventral and the right coeloms and other structures had disappeared (Jeffries, 1986, 52; Jeffries, 1990; Paul, 1990; Willmer, 1990, 319-326); note that it would be of little import for the main purposes of this Figure were the positions of hemichordates and echinoderms reversed. Acraniates have been placed as the sister group of vertebrates rather than tunicates, on the basis of the distribution of a large number of shared characters amongst these groups, as enumerated by Maisey (1986) and Schaeffer (1987), for example. Maisey (1986) and Jeffries (1986, Ch. 5-6) and the earlier authorities whose works they review, have assembled a large number of characteristics shared by all true vertebrates that cannot be enumerated here, e.g., appearance of genuine neural crest, olfactory and optic capsules of brain, somitic structures, etc. Examples of prominent characteristics important for the organization of the diagram that appear in particular phyletic branches, are indicated across the stems leading to the modern groups; of course these are merely indicative, in each case, of a large number of definitive characters. Proceeding upwards along the protostome and deuterostome branches the characters indicated by the black thin lines parallel to the heavy green main branches are shown cumulatively; thus for example, on the protostome line, from the ancestors of the nemerteans up, all groups descended from ancestors with hemocoels (fluid filled body cavities not lined with coelomic mesoderm), and from the ancestors of annelids up, all groups descended from metameric ancestors (i.e., animals composed of repetitive homologous body segments). These are selected from among the nested sets of shared characters that support the order shown on the phyletic stems.

Embryological Aspects: The classifications indicated in color are based on the criteria summarized in Table 1 and discussed in text. Adequate experimental data are in some cases not available, particularly with respect to the key parameter, the initial mode of embryonic cell fate specification. In these cases the phyla are indicated by striped rather than solid color patterns. For relatively well studied nematode, dipteran, gastropod and other molluscan, echinoderm, ascidian, amphibian, teleost, and mammalian embryos the assignments of embryo type are based on data summarized in Davidson (1990), and review of relevant information in Davidson (1986, Ch. 4 and 6). Evidence supporting the Type 1 assignments shown for cnidarian, ctenophore, nemertean, crustacean, annelid, tentaculate, hemichordate, and acraniate embryos is very briefly cited in Note to Fig. 1 . The classification obviously rests on observations on a few type species of each group that have served as objects of investigation, and there may be many deviant embryological modes, particularly among direct developing species, or species with very yolky eggs (see discussion in text). These, however, are generally regarded as derivative forms. For the Uniramia, Anderson (1973) provides a key comparative treatment of hexapod, myriapod (i.e., millipedes, centipedes, etc.), and onchyphoran embryos. Very little detailed experimental information exists that in fact supports the generalization of Type 3 mechanisms outside the long and intermediate germ band orders of insects (see text). Myriapod embryos could be described as intermediate germ band forms in recombinations no respecification of blastomere fates along the animal-vegetal axis can be detected (Horstadius, 1937). The egg is polarized in this axis when shed. However, isolated 1/2 eggs, including, e.g., both 1/2 blastomeres of one egg, generate normal bilaterally symmetrical larvae, as do at least some 1/4 eggs (Wilson, 1903; Horstadius, 1937). This clearly requires regulation and hence in normal embryos, conditional specification, at least along the second (A/P) axis of the egg.

Crustacea: These are basically spirally cleaving eggs, as can be seen, for example, in barnacles (reviewed by Anderson, 1973, 268-280 and Green, 1971, 320-333). Yolky crustacean eggs display adaptive modifications, including radial cleavage, sometimes involving secondary features such as generation of a ‘blastoderm’ covering the yolk. In contrast to insects, canonical invariant lineages have been described for many crustacean species (Kumé and Dan, 1968, 341-352). The fate map of the spirally cleaving forms is decisively different from that of annelids, but is homologous in all crustacean forms (Anderson, 1973, Chapter 8). There is little detailed modern evidence on the nature of blastomere specification processes. However, germ cells are probably autonomously specified by maternal determinants in copepods and other groups (reviewed by Wilson, 1925, 314). The egg is evidently polarized in the initial An/Vg axis; and mesoderm derives from specific founder cells (3A, 3B, 3C), as does the gut, though the lineage differs amongst orders (Anderson, 1973, op cit). These founder cells are likely to be autonomously specified, as in other spirally cleaving forms (though the fate of individual lineage elements is strikingly different from that in the annelid spiral cleavage map). On the other hand, regulative capability has been reported (Green, 1971, op cit). Furthermore in crayfish posterior segments develop progressively from a ‘growth zone,’ as in intermediate or short germ band insects, and these segments express engrailed at their posterior boundaries (Patel et al. 1989a), suggesting a homologous process that depends on intercell interaction for posterior segment formation. The assumption that crustaceans have Type 1 eggs is tentative, though it seems unlikely to be incorrect.

Annelids: Polychaete annelid embryos exemplify canonical invariant spiral cleavage, and have a canonical fate map (for recent review see Anderson, 1973). Oligochaetes that have yolky eggs display certain modifications in cleavage pattern but homologous fate maps. Many classical studies demonstrate autonomous specification of both oligochaete and polychaete annelid lineages. Among these are mesoderm and the ectodermal neuroblast germ band lineages, apical tuft, specific bristle and protocyte lineages. Such results have been obtained in experiments in which individual blastomeres or specific groups of blastomeres are killed, raised in isolation, or transplanted to ectopic positions; see experiments of Penners on Tubifex, (reviewed by Morgan, 1927, 370-380), of Novikoff (1938) and of Render (1983) on Sabellaria; of Henry (1986, 1989) and Henry and Martindale (1987) on Chaetopterus. Furthermore, partial double embryos that do not regulate to form a single larva are produced in situ by equalization of cleavage in Tubifex (reviewed by Morgan, 1927, 412-417), and partial twin embryos are generated by equalization of cleavage in Chaetopterus (Henry and Martindale, 1987). All of these results imply autonomous specification by maternal determinants localized in given lineage founder cells (Davidson, 1990). However, in the twin studies on Chaetopterus of Henry and Martindale (1987), it is clear that, to account for the morphological results, the D blastomere must exercise an inductive effect on C blastomere derivatives as well as giving rise to autonomously specified structures. In direct developing annelid forms exemplified by the leech Holabdella (Class Hirudinea) autonomous specification of the ‘mesoblast’ stem cell (in fact this also gives rise to a few neuroblasts; Weisblat et al. 1984), and of the neuroectoteloblast, is demonstrated by centrifugation experiments. These show that supernumerary founder cells develop when a visible maternal cytoplasm (teloplasm) is distributed to the C as well as D blastomere (Astrow et al. 1987). However, several studies also show that in this and other leeches the founder cells for several neuroblast lineages are conditionally specified, so that their identity depends on that of their neighbors (Weisblat and Blair, 1984; Keleher and Stent, 1990). Annelid embryos thus have all the characteristics of Type 1 embryos; a monopolar egg organization ab initio; invariant cleavage; and utilization of both autonomous and conditional founder cell specification processes.

Phoronida: Little is known of the cellular embryology of tentaculates outside Phoronida. The egg is apparently preorganized along the A/V axis. The animal pole is clearly marked by polar bodies and by egg shape, which is flattened at A and V poles. The first two planes of cleavage are orthogonal in this axis, the third transverse, just as in hemichordates and echinoderms (Kumé and Dan, 1968, 242-245; Zimmer, 1964; Emig, 1977). In cleavage two radially segmented tiers of eight cells each are formed, and then four tiers (as also in other tentaculate phyla, viz ectoprocts and brachiopods). There is no good evidence for localization of determinants along the A/V axis, except that like sea urchin embryos, the animal blastomere quartet isolated after the transverse 3rd cleavage fails to regulate, forming only a ciliated ball of cells that express none of the neuronal markers normally produced by the progeny of these cells (Freeman, 1991). However, vegetal half 3rd cleavage embryos regulate to produce whole new embryos (Freeman, 1991), while 1st and 2nd cleavage blastomeres form almost complete embryos which, however, lack normal tentacles (Emig, 1977). These characteristics are typical of Type 1 embryos, in which there is normally some autonomous specification along the A/V axis, though it is clear that neurectodermal differentiation in animal cap cells requires influences from ‘ vegetal cells. Furthermore the second axis is specified after fertilization, as in Type 1 eggs. On the other hand, anteroposterior blastomere fate specification occurs only late in development, in gastrulation, and the lineage and cleavage patterns are somewhat variable (Freeman, 1991). These are significant deviations from Type 1 characteristics. Hence the Figure displays a striped color pattern for this phylum indicating that the assignment of Type 1 embryonic process to this group is not unequivocal.

Hemichordates: Significant information exists only for some species of Class Enteropneusta (acorn worms). Eggs of Saccoglossus display radial cleavage of a form remarkably similar to that of echinoids; thus 4th cleavage is radial in the animal half of the embryo and horizontal in the vegetal half, etc. The egg is polarized ab initio in the A/V axis, which corresponds to the future A/P axis of the embryo (Colwin and Colwin, 1953). Cleavage is canonical and was followed to the 128-cell stage, but Colwin and Colwin (1953) observed some variation in the positioning of the 4th cleavage planes, without any disturbance of subsequent development. The initial fate map along the A/P axis is similar to that of echinoderms. Thus the lowest tier of blastomeres at the 16-cell stage gives rise to gut and coeloms (like veg2 in sea urchins) and the animal half plus the vegi equivalent produces ectoderm (Colwin and Colwin, 1953). Both members of pairs of isolated 1/2, and some isolated 1/4 blastomeres, form complete embryos, again as in sea urchins. Though this indicates conditional specification (at least along the A/V axis), there is no direct evidence of any autonomous specification in these embryos. However, the detailed similarities (very briefly noted here) to early sea urchin embryos suggest that early developmental processes are generally similar as well; hence their (somewhat tentative) Type 1 classification.

Acraniates (or Cephalochordates): As described by classical workers (reviewed by Reverberi, 1971), the canonical cleavage gives rise to a four-tiered, 32-cell embryo, with a more or less invariant fate map (Tung et al. 1962; Reverberi, 1971). Extensive blastomere tier recombinations, and other experiments in which one or another tier is deleted, demonstrate that most of the lineage fates are plastic, i.e., are in normal embryos conditionally specified (reviewed by Reverberi, 1971). The neural tube depends for its formation on contact with vegetal anterior blastomeres, and can be formed from cells that normally give rise only to ectoderm in ectopic recombinations. Similarly, ectopic founder cells for notochord are generated in chimeric recombinations. However, at least some autonomous specification capacity is elicited in cultures of isolated vegetal tiers which in some cases give rise to gut, a normal product of this tier, and embryos lacking the veg2 tier give rise to larvae lacking the gut. In addition, pairs of isolated 1/2, 1/4 or 2/4 blastomeres form complete larvae, though animal half 4/8cell embryos produce only ciliated balls (Wilson, 1893; Conklin, 1932; Tung et al. 1958). The embryo thus in many respects resembles that of the sea urchin (Horstadius, 1939; Davidson, 1989), relying largely on conditional founder cell specification, but with an initial A/V organization reflected in some autonomous vegetal (gut) specification (in the sea urchin the vegetal-most lineage is skeletogenic, and it is this, rather than gut, which is specified autonomously).

Fig. 1 is organized as a phylogenetic chart of Metazoa, omitting several minor phyla of which relatively little is known. Superimposed on the chart is a color code indicating embryo type for each phylum; blue for Type 1, red for Type 2, and yellow for Type 3. Phylogenetic aspects of the chart are described in the legend. There is no universally or perhaps even generally accepted phylogenetic organization for such a diagram, because of the uncertainties surrounding the adult and embryonic morphological homology arguments, and the evolutionary and paleontological interpretations, on which it must be based. However, the particular arrangement shown in Fig. 1 is not very important for our present purposes, given its rather simple and direct import. This is that Type 1 embryogenesis is the basic, typical mode of early development in almost every invertebrate phylum. In many of these phyla certain orders or families or even particular genera display significant deviations from the canonical Type 1 process (see below), but only embryo types that appear to possess Order, Class or Phylum levels of generality are indicated by the color code in Fig. 1. At this level Type 2 embryonic process is characteristic only of vertebrates, and possibly of the cephalopod molluscs and the hydroid cnidarians. Type 3 embryonic processes occur in long germ band orders of winged insects, and most probably in intermediate germ band orders. This includes wingless as well as winged insects. It is important to note that uniramian embryos of all classes, e.g. myriapods as well as insects, display some version of intralecithal or syncytial cleavage (see legend to Fig. 1 for references and further details).

Fig. 1 implies a fascinating evolutionary argument, viz that the original form of embryogenesis, which must have appeared before the Cambrian, was and is Type 1 embryogenesis. Other forms of embryonic process might therefore be regarded as derivations of Type 1 embryogenesis. Note in this connection that the oldest known fossil true vertebrate, a jawless fish, appears in the Upper Cambrian; an acraniate is known from the Middle Cambrian and the oldest uniramian ancestors also appear to have arisen in the course of the Cambrian, though the earliest myriapods and wingless insects appear only in the Devonian. The earliest true coelomate protostome forms are already present at the beginning of the Cambrian, however, and animals similar to the acoel and pseudocoel protostomes of today are thought to have existed in the pre-Cambrian (see, e.g. Manton, 1977, 26-27; Richards and Davies, 1977, 421-427; Valentine, 1989; Jeffries, 1986, 333-335; Kuhn-Schnyder and Rieber, 1986). It thus becomes instructive to consider what regulatory changes might we imagine by which Type 2 and Type 3 embryonic process might have derived from their respective deuterostome and protostome ancestors.

Direct developing Type 1 variants

A clue perhaps emerges from minor phyletic variants in embryonic process that are not indicated in Fig. 1. The Type 1 groups shown in this Figure characteristically develop indirectly, in that the business of the embryo is to produce a feeding larva, which in turn provides a life support system for more complex morphogenetic processes required to generate the juvenile form of the adult. However, in many of these phyla there occur variants that develop the juvenile form directly. These are undoubtedly derivative variants of the canonical, indirectly developing forms in each phylum (e.g. for annelids, see Anderson, 1973, 88-90; for echinoids Raff, 1987; for ascidians Jeffery and Swalla, 1990). For instance, in the sea urchins, about 20% of the approximately 200 studied species show some variation from the canonical form of indirect development, and direct development has arisen independently in six of the ten echinoid orders. Sometimes, as in both echinoids and ascidians, the same genus will include species that develop directly, and species that develop indirectly. This shows that the types of evolutionary change leading to direct development can occur relatively easily, in evolutionary, and in regulatory terms. The kinds of regulatory change that have occurred include deletion of regulatory programs; heterochronic changes, in which developmental processes that are used in larval morphogenesis in indirectly developing forms are instead used in embryonic development; and other qualitative inventions. For instance, in directly developing ascidians, there is an alteration in the localization of maternal determinants required to specify gene expression in the larval tail muscle, and the expression of embryonic muscle genes is deleted (Jeffery and Swalla, 1990; Swalla et al. 1991). Changes in the timing and order of internal organ system morphogenesis, in stage of founder cell specification, and in timing of expression of lineage-specific marker genes (Parks et al. 1988; Raff et al. 1990; Wray and Raff, 1990) are among the differences that distinguish direct development in sea urchins from typical indirect development. Indeed heterochronic (and other) changes in lineage specific gene expression can be observed even in comparing various indirectly developing sea urchin species (Wray and McClay, 1989). For the most part, species belonging to typically Type 1 phyla, which display direct development, nonetheless retain a Type 1 form of embryonic process, by the external criteria of Tablel. Examples include the direct developing ascidian species cited above; and the direct developing oligochaete and leech classes of annelids. These embryos fulfill Type 1 criteria as well as do the archetypical marine polychaete annelid embryos, despite their often enormous, yolky eggs, and altered forms of spiral cleavage (Anderson, 1973, 51-90; references and summary in Notes to Fig. 1, where examples from other phyla are also indicated). However, what is of particular interest here is that in some of the independent explorations of direct development that have occurred frequently in evolution, variations appear that seem to foreshadow Type 2 or Type 3 embryonic process. For instance, early development in the order Hydroida, of the cnidarian class Hydrozoa, has many features that are amazingly reminiscent of Type 2 development (reviewed by Freeman, 1983), though other hydrozoan orders display Type 1 development (as do ctenophore embryos; see Fig. 1, and Notes to Fig. 1). Thus, in Hydroida cleavage is variable and there is no canonical lineage; regional organization is evidently mediated in part by diffusible morphogens, as well as shorter range interactions; all early blastomeres are totipotent; and the whole early body plan can be regenerated from pieces that normally would have formed only ectoderm or endoderm (reviewed by Freeman, 1983, 1990). In a direct developing sea urchin, as another example, the lineages of vegetal cells have become variable, while the animal cell lineages remain invariant and similar to those of typical indirect development (Wray and Raff, 1989). Similarly, in the large, yolky eggs of some decapod crustácea (i.e. shrimps, lobsters and crabs), spiral cleavage is lost or obscured; some syncytial cleavage is observed in the presumptive midgut region of the egg; and a blastoderm-like layer of cells forms around the outside of the yolk mass (Anderson, 1973, 278-282); these features are reminiscent of Type 3 development. Adaptive advantages of various forms of direct development have evidently led to the repeated derivation of these kinds of changes. Thus the major incidences of Type 2 and Type 3 embryogenesis shown in Fig. 1 may represent only the phyletically more important fixations, and extensions, of processes that have started to occur many times over in diverse evolutionary branches.

Irrespective of the deep distinctions amongst various types of biological embryonic process, at the level of the molecular biology of gene regulation they all seem at first glance the same. All metazoan embryos probably use zinc finger regulators, various homotypic and heterotypic pairs of helix-loop-helix regulators, and regionally expressed homeodomain regulators during development. These could be regarded merely as common building blocks of which diverse regulatory structures are constructed, but homologies between distant developmental processes extend to higher levels of organization as well. Our field has recently been absorbed for example, by the evidence suggesting that homologous, evolutionarily conserved clusters of homeobox genes might be utilized for anteroposterior developmental processes throughout the bilateral metazoans. The regional spatial expression of homeodomain regulators in all forms of developing animals, and the homeotic morphological effects of gain-of-function regulatory mutations in some of the genes encoding these regulators in Drosophila, make the functional role of these genes a particularly interesting issue. We now have a relatively good understanding in Drosophila of how their mutual cross-regulatory interactions are utilized in setting up their own spatial domains of expression within the organism. The general nature of the downstream targets of regionally expressed homeodomain transcriptional regulators has so far proved an elusive experimental objective. It is clear, however, that the general problem of how spatial patterns of gene expression are coordinated in embryogenesis intersects with the problem of homeodomain regulatory function. In the following I explore a general interpretation of the downstream biological functions of these elements of the developmental regulatory system, and then briefly reexamine the differences - and similarities -in the regulation of spatial patterns of gene expression in the different forms of embryogenesis.

What do regionally expressed vertebrate homeodomain regulators actually control?

A great deal of descriptive evidence has also accumulated regarding the patterns of expression of homeobox genes in vertebrate development. Some of this evidence, considered in conjunction with the biological character of vertebrate developmental processes, suggests a specific functional role for homeodomain regulators, at least in vertebrate ontogeny. We are here concerned only with those homeodomain regulators that are expressed regionally during vertebrate development, in the specific sense that they appear in a given spatial domain of the body plan, but in a variety of different cell types within that domain. This excludes the cell-type-specific homeodomain regulators of vertebrates, of which the downstream targets are indeed genes encoding cell-type-specific products, e.g. Pitl in rat pituitary cells (Ingraham et al. 1988); HNF-1 in the rat liver (Baumhueter et al. 1990); Oct-2 in human B cells (Clerc et al. 1988); and perhaps the product of the mouse Cdx-1 gene, which is expressed only in differentiating intestinal epithelium cells late in embryogenesis and in adults (Duprey et al. 1988). Note that the same homeobox genes can be expressed regionally at one stage in development, while at a later stage they may have a cell-type-specific role elsewhere in the organism (e.g. Vogels et al. 1990); we are here concerned only with the circuitry and function of such genes when they are being expressed regionally. In amphibian embryos regionally expressed homeodomain regulators first appear during mid-late gastrulation, after the initial broad specification of the major regions of the embryonic body plan has already occurred, or as it is in process. Many such genes then continue to be expressed in a changing spatial pattern throughout postembryonic development, particularly during morphogenetic and organogenic processes.

A simple interpretation of the developmental function of this class of regionally expressed regulatory gene, for which some evidence is discussed below, is as follows:

(i) Downstream functional role

The biological function of regionally expressed homeodomain regulators might be specifically to promote spatially organized expression of genes encoding morphogenetic ligands, receptors, particular signal transduction machinery, and extracellular components that mediate ligand-receptor interactions. The premise is that organogenesis and local morphogenesis of structure proceeds by means of short-range ligandreceptor interactions amongst different cell types, which specify the location and function of the constituent cells with respect to one another. The specific downstream targets of homeodomain regulators would thus be regulatory binding sites in these ‘cell communication’ genes, or perhaps in genes encoding transcription factors that specifically control ‘cell communication’ functions. The function of regional homeodomain regulator expression would therefore be to control the coordinate expression of the ligands and the receptors required for given morphogenetic developmental modules. This could be accomplished by control of both the ligand and the receptor system by the same homeodomain regulator (in different cells), or by different homeodomain regulators, as illustrated in a simplified way in Fig. 2 (see legend).

Fig. 2.

Abstract depiction of simplest versions of proposed function of regionally expressed homeodomain regulators in vertebrate development. (A) Intranuclear regulatory requirements if the homeodomain regulator directly activates (or represses) both the ligand and receptor genes of locally interacting cells (the same functional relationship would obtain if the homeodomain regulator directly activates a gene encoding a transcription factor, that in turn activates both ligand and receptor genes in the respective cell types). In 1, cells of two different layers that are destined to interact in order to produce a morphogenetic event are indicated in different colors. These cell types have been established by a prior process of specification; e.g., they could be mesoderm and ectoderm or mesenchyme and epithelium, neural crest derivatives and cells with which they will react, etc. Cell type factors (CTF) are each specific to one cell type; the ligand (lig) gene has a site for one of these factors and the receptor (rec) gene for the other. At 2, in the following stage of development the homeodomain regulator (HDR) is shown bound to a second regulatory target site in the gene coding for the inductive ligand and also in the gene coding for the receptor. That is, both genes have sites that can be productively occupied by the homeodomain regulators. Solid arrow indicates transcription of .these genes, which occurs when both the CTF and HDR sites are bound, presumably by a mechanism involving factor interactions (double headed arrow). (B) Morphological scale. In 1, a relatively large region of the two layers (pink and tan) is indicated, separated by an extracellular matrix (gray). Later in development this region will give rise to two different morphological structures, but until regional expression of the homeodomain regulators occurs, these tissue layers are developmentally equivalent in the horizontal direction. In 2, three different regulatory circuits are shown in the boxes, representing different ways in which coordination of regional expression of ligands and their receptors might be organized by homeobox (HBX) gene expression. On the left, as in part (A) of the Figure, a homeobox (HBX) gene regulates its ligand in one of the early cell types and its receptor in another; this gene, and its HDR product are shown in red. The gene might have been turned on in both layers by certain levels of a diffusible morphogen, for example. An alternative is shown in the middle panel, in which the homeobox gene controlling the expression of the ligand in the inducing cell also makes use of a preexistent signalling system in the responding cell type to induce its own expression. This then results in the production of the specific downstream interaction system symbolized by the ligand and receptor, that will mediate morphogenetic development (were the preexistent signalling system adequate to mediate morphogenetic development itself, homeobox gene expression in the responding cell would not be required, nor presumably, observed, unless the same molecule acts as ligand and receptor). In the right box two different homeobox genes are active (genes and products in blue and turquoise). The top gene controls the ligand and the bottom the receptor, as shown. The genes are coordinately activated in the different cell types either by an internal regulatory network that depends in part on cell-type-specific factors as in A; or by an inductive activation as in the middle box. Different homeodomain regulators might be required because their signal and receptor functions are utilized separately elsewhere in the organism. Of course innumerable more complex (and no doubt more realistic) variants can be imagined than those shown in this cartoon: e.g., there may usually be needed not one but several homeodomain regulators in certain particular concentrations; or the receptors could be constitutive in some cell types and not be regulated; or the cell-type-specific transcription factors shown in A could be several factors, some controlling temporal expression, some responding to signals from other cells; etc. Below, the presentation of two different sets of morphogenetic ligands and receptors is indicated, resulting in two different morphological processes. The different homeodomain proteins expressed in the red, and in the blue and turquoise, region mediate different modular ligandreceptor interactions, producing spatially separate, diverse structures.

Fig. 2.

Abstract depiction of simplest versions of proposed function of regionally expressed homeodomain regulators in vertebrate development. (A) Intranuclear regulatory requirements if the homeodomain regulator directly activates (or represses) both the ligand and receptor genes of locally interacting cells (the same functional relationship would obtain if the homeodomain regulator directly activates a gene encoding a transcription factor, that in turn activates both ligand and receptor genes in the respective cell types). In 1, cells of two different layers that are destined to interact in order to produce a morphogenetic event are indicated in different colors. These cell types have been established by a prior process of specification; e.g., they could be mesoderm and ectoderm or mesenchyme and epithelium, neural crest derivatives and cells with which they will react, etc. Cell type factors (CTF) are each specific to one cell type; the ligand (lig) gene has a site for one of these factors and the receptor (rec) gene for the other. At 2, in the following stage of development the homeodomain regulator (HDR) is shown bound to a second regulatory target site in the gene coding for the inductive ligand and also in the gene coding for the receptor. That is, both genes have sites that can be productively occupied by the homeodomain regulators. Solid arrow indicates transcription of .these genes, which occurs when both the CTF and HDR sites are bound, presumably by a mechanism involving factor interactions (double headed arrow). (B) Morphological scale. In 1, a relatively large region of the two layers (pink and tan) is indicated, separated by an extracellular matrix (gray). Later in development this region will give rise to two different morphological structures, but until regional expression of the homeodomain regulators occurs, these tissue layers are developmentally equivalent in the horizontal direction. In 2, three different regulatory circuits are shown in the boxes, representing different ways in which coordination of regional expression of ligands and their receptors might be organized by homeobox (HBX) gene expression. On the left, as in part (A) of the Figure, a homeobox (HBX) gene regulates its ligand in one of the early cell types and its receptor in another; this gene, and its HDR product are shown in red. The gene might have been turned on in both layers by certain levels of a diffusible morphogen, for example. An alternative is shown in the middle panel, in which the homeobox gene controlling the expression of the ligand in the inducing cell also makes use of a preexistent signalling system in the responding cell type to induce its own expression. This then results in the production of the specific downstream interaction system symbolized by the ligand and receptor, that will mediate morphogenetic development (were the preexistent signalling system adequate to mediate morphogenetic development itself, homeobox gene expression in the responding cell would not be required, nor presumably, observed, unless the same molecule acts as ligand and receptor). In the right box two different homeobox genes are active (genes and products in blue and turquoise). The top gene controls the ligand and the bottom the receptor, as shown. The genes are coordinately activated in the different cell types either by an internal regulatory network that depends in part on cell-type-specific factors as in A; or by an inductive activation as in the middle box. Different homeodomain regulators might be required because their signal and receptor functions are utilized separately elsewhere in the organism. Of course innumerable more complex (and no doubt more realistic) variants can be imagined than those shown in this cartoon: e.g., there may usually be needed not one but several homeodomain regulators in certain particular concentrations; or the receptors could be constitutive in some cell types and not be regulated; or the cell-type-specific transcription factors shown in A could be several factors, some controlling temporal expression, some responding to signals from other cells; etc. Below, the presentation of two different sets of morphogenetic ligands and receptors is indicated, resulting in two different morphological processes. The different homeodomain proteins expressed in the red, and in the blue and turquoise, region mediate different modular ligandreceptor interactions, producing spatially separate, diverse structures.

(ii) Upstream regulation of regionally expressed homeobox genes

Spatial expression of this class of homeobox genes in vertebrates will depend on the prior general axial specification processes, and will be regulated in part by diffusible morphogens, since these may be utilized in vertebrate embryos to specify domains in which particular morphogeneses will occur, i.e. to specify what are sometimes called ‘embryonic fields’ (this is discussed in the following section). In part they will be regulated by homeobox gene interactions, including interactions at the .intranuclear network level, and at intercellular boundaries, by short-range inductive interactions (cf. Fig. 2). ‘

(iii) Conditional specification

Morphogenesis proceeds by conditional specification, in the particular sense considered in the initial section of this paper, which deals with Type 1 embryos. That is, the number of inductively competent cells always exceed those actually utilized in a given response, and certain cells other than those actually responding to any given signal could always have responded, had they been in the right place (or were they put in the right place). The interpretation of conditional specification (or regulative capacity) given above implies that celltype-specific genes that function as a result of morphogenetic induction, e.g. in organogenesis, are likely to be activated by selective modification of preexisting transcription factors, mediated by inductive signal transduction. The perhaps surprising interpretation emerges that genes encoding cell-type-specific products of terminal differentiation are unlikely to be the direct downstream targets of homeodomain regulators.

(iv) Evolutionary conservation of ligand, receptor and homeobox gene control circuits

Homeodomain regulators and the downstream ligand and receptor genes that they control are likely to constitute conserved genetic regulatory modules (rather than just the homeobox genes per se), since change in the DNA-binding elements of these regulators would require corresponding changes in the regulatory domains of the target genes, and vice versa. Thus given homeobox genes and their downstream targets can be thought of as ‘regulatory kernels,’ that are likely to be found in many organisms, and are likely to be used over and over within the same organism, and amongst organisms, to mediate different inductive morphogeneses. The number of different downstream target genes of regionally expressed homeodomain regulators thus may be quite limited.

In a local morphogenetic process proceeding by way of short-range inductive intercellular interactions, the inducing and the induced tissues have in common only the rather teleological property of ‘similar positional identity’, and the mechanistic property of coordinate expression, respectively, of the ligands and the receptors that mediate their interaction. They do not in general express other downstream structural genes in common. Thus the experimental observation that homeodomain regulators are regionally expressed in both inducing and induced cells is significant. Many examples that may illustrate the principles supposed in Fig. 2B are to be found in the recent literature, obtained by in situ hybridization or immunocytology. These show for various stages of mouse, Xenopus, or chicken development the tissue, and sometimes the cellular locations of homeobox gene transcripts, or their homeodomain protein products. A very brief account of a few pertinent examples follows. The first of these examples concerns regional, late embryonic morphogenesis of the CNS. Formation of the anteroposterior elements of the brain and spinal cord is probably programmed by interactions with the immediately contiguous mesodermal tissues (see Holtfreter and Hamburger, 1955; Keynes and Stern, 1988). Expression in both mesoderm and CNS neuroectoderm early in CNS morphogenesis is observed for many homeodomain genes in the mouse (e.g. see Gaunt, 1987, 1988; Kessel and Gruss, 1990; Murphy and Hill, 1991; see Fig. 2B, left-most box). The early expression patterns of the Hox-2.9 gene (labial related) are particularly interesting. In the late gastrula of the mouse embryo, transcripts of the Hox-2.9 homeobox gene first appear in migrating mesodermal cells that will become associated with the CNS in the posterior half of the embryo (Frohman et al. 1990). The anterior boundary of the position that these cells will occupy is the posterior hindbrain. By the early neurula stage that portion of the neuroectoderm overlying the expressing mesodermal cells also transcribes Hox-2.9, sharing the same anterior boundary of Hox-2.9 transcription with the expressing mesoderm. Later the pattern sharpens and intense Hox-2.9 expression is limited to rhombomere 4 as that structure differentiates, but the transcripts continue to be expressed in both mesoderm and neuroectoderm (Murphy et al. 1989). The initial registration of expression may be the direct result of induction of Hox-2.9 expression in neuroectoderm by Hox-2.9 expressing mesoderm (Frohman et al. 1990; cf. Fig. 2B, middle box). In chick the paralogous homeobox gene, whose expression is also ultimately confined to rhombomere4, is called Ghox-lab, and the mRNA, and the nuclear homeodomain protein are again found in the neuroectoderm and underlying mesoderm at the gastrula stage, even before neural plate formation (Sundin and Eichele, 1990; Sundin et al. 1990). A second example also emphasized by Frohman et al. (1990) concerns the expression of Hox-2.9 not only in the branchial arch mesoderm, which derives from neural crest, but also in the contiguous endoderm of the pharyngeal pouch, and the overlying ectoderm, amongst which the anterior boundaries of expression are again all in register. The mesoderm is thought to provide short-range inductive signals that direct foregut morphogenesis (Copenhaver, 1955; Okata, 1957; Wessels, 1977, 123-127). In morphogenesis of head structures, a similar pattern is observed, as shown by Hunt et al. (1991). Each of the anterior branchial arches is populated by neural crest cells that express a particular, unique set of homeobox genes of the Hox-2 cluster, determined initially by the region of the neural plate from which they derive; and the overlying ectoderm comes to express the same homeobox genes as the apposed neural crest cells: A third example concerns homeobox gene expression in the differentiating kidney, in which various homeobox genes are expressed (e.g. Graham et al. 1988; Erselius et al. 1990; Kress et al. 1990; Vogels et al. 1990). Hox-3.2 is expressed in cortex and also in mesenchymal cells of the developing kidney, for instance, and Hox-2.3 is expressed in both ducts and tubules in mesonephros and metanephros. Formation of the kidney is regarded as a metameric, anteroposterior process, and the expression of homeobox genes in kidney preserves regionality, in that only Hox genes expressed at certain anteroposterior axial positions are also expressed in the developing kidney. Furthermore, expression of the same homeodomain regulators clearly occurs in the various interacting cell types that give rise to the differentiated organ, i.e. duct, mesenchyme, etc. (Holtfreter, 1944; Burns, 1955). An additional striking feature is that the mesenchyme cells that express Hox-3.2 also express TGFβ1 at the same time (Schmid et al. 1991). In fact, many similar parallels appear to exist between organ- and tissue-specific patterns of expression of growth factors (e.g. see Pelton et al. 1990; Millan et al. 1991; Jones et al. 1991), and of retinoic acid receptors and binding proteins (Dollé et al. 1990), and the patterns of expression of various homeobox genes, for example in developing heart, lung, stomach, prevertebrae, limb buds, etc. Similarly, in the mouse Wnt5a, a mammalian homologue of the Drosophila wingless gene, is expressed throughout the posterior region of the late gastrula, and in a graded distribution in the developing limb buds that is independent of cell type or tissue (Gavin et al. 1990). Parallels between morphogen distribution and homeodomain proteins or their mRNAs have been noted (e.g. Robert et al. 1989; Ruberte et al. 1991; Smith and Eichele, 1991; Jones et al. 1991) and are likely to become far more obvious when we have higher resolution knowledge of which specific ligands (and their receptors) and which specific homeodomain regulators are present in the key cell populations at the time of formation of each organ system. A fourth example is afforded by developing limb buds in Xenopus (Oliver et al. 1988a) and feather buds in chicken (Chuong et al. 1990). A homeodomain protein called XlHbox-1 (recognized by an antibody) is expressed in both morphogenetic processes. Expression in the Xenopus limb buds is confined to the forelimbs, and when it first appears it defines the region or ‘field’ of lateral plate mesoderm where the buds will form. Early in limb bud morphogenesis expression is high anterior-proximal and low posterior-distal; i.e. it is of opposite polarity to a morphogenetic inductive process initiated at the posterior ‘zone of polarizing activity’ (see Wanek et al. 1991; Noji et al. 1991; for recent evidence that this inductive morphogenesis could be mediated by shortrange cell- to-cell interactions, or by an unknown morphogen, but not by a retinoic acid gradient). Many intercellular interactions clearly must occur within the limb bud mesenchyme, and it is significant that both it and the overlying ectoderm are positive for XlHbox-1. In feather bud morphogenesis, the mesoderm provides inductive direction and, through its immediate interaction with adjacent ectoderm, it generates the feather epithelial placode, and then the feather follicle. In the mesodermal component, the XlHbox-1 nuclear protein is again graded in distribution, high in the anterior-proximal region of the bud, and all the interacting ectoderm cells also express the protein relatively strongly.

It is to be stressed that in many cases homeobox genes are not expressed in both inducing and responding tissues, contrary to these examples. In the mouse limb bud, for instance, transcripts of the Hox-5 gene cluster form an elegantly precise, nested set of spatial zones of expression, centered at the posterior-distal site named the ‘zone of polarizing activity’; but these genes are expressed only in the mesodermal component (Dollé et al. 1989). Similarly, in the lung, kidney, stomach and trachea, e.g. the paralogous Hox-1.4, Hox-2.6 and Hox-5.1 genes are expressed in mesoderm but not in endodermal endothelial layers (Gaunt et al. 1989); in Xenopus, the XlHbox-8 gene is expressed just before, and during differentiation of a region of the endoderm that includes the presumptive duodenum, the pancreas and its duct system, i.e. in a number of different local endodermal cell types (Wright et al. 1988). Different homeobox genes are expressed in the overlying gut mesoderm (Oliver et al. 19886). Supposing that the downstream targets of homeodomain regulators are indeed genes encoding ligands and receptors, there are many reasons why only one of the interacting cell types might be observed to be expressing a given homeobox gene: the ligand might be regulated by one homeodomain protein and the receptor by another, as in Fig. 2B, right box (see legend); the receptor, or the ligand, could be very broadly expressed and thus only the other would require spatial regulation; the cell ‘type’ to which expression is observed to be confined, e.g. an organogenic mesenchyme, might in fact be heterogeneous, and the cells may be engaged in interactions with one another; or these cells might be required to interact with another cell type not noticed at the time of the observation. For example, as the above references show, many homeobox genes that are expressed in the mesoderm of given organs are also expressed in neural crest, derivatives of which ultimately contact and innervate these organs.

When expression of a homeodomain regulator is prevented in an amphibian embryo, or ectopic expression is forced, there are morphological consequences of the sort one might imagine were a local morphogenetic inductive program interrupted by failure of, or ectopic, ligand-receptor presentation. In general the effects are confined to those regions of the body plan to which expression is normally confined, indicating that the normal downstream function of homeodomain regulators is not mediated by long-range diffusible morphogens. Several informative cases have been reported from Xenopus. XlHbox-1 is expressed in two transcripts, ‘long’ and ‘short’, which may function antagonistically. The short protein is expressed more anteriorly, in CNS, neural crest, mesoderm and internal organs (Oliver et al. 1988b). Wright et al. (1989) showed that injection into the fertilized egg of an antibody recognizing the more posterior ‘long’ protein causes the upper spinal cord to begin development as a hindbrain, normally the next most anterior CNS structure, while a similar effect is produced by injection of excess ‘short’ protein mRNA. This result suggests modular, local control of morphogenetic fate, involving reorganization of similar cell types; thus motor nerves are generated by both normal and transformed CNS elements but they are organized differently. Synthetic transcripts of a homeobox gene called Xhox-1A, that is normally expressed in axial structures of the trunk, were injected into early blastomeres by Harvey and Melton (1988), with the primary result that somitic morphogenesis was disorganized. Nonetheless, molecular muscle-specific markers were expressed as in normal somites, emphasizing that excess Xhox-1A expression affects tissue organization, i.e. intercellular arrangement and communication, rather than differentiation per se. The effects of ectopic expression of Xhox-3, a homeobox gene normally expressed in a declining posterior-to-anterior gradient, were reported by Ruiz i Altaba and Melton (1989a). Injection of this transcript into early anterior blastomeres permits normal gastrulation, and the formation of differentiated notochord, muscle and neuronal cell types, but it then results essentially in the deletion of head formation. An inductive process appears to have been affected, since the anterior neural plate is normally specified inductively by underlying anterior mesoderm and, though the mesodermal cells and the neural plate are in place in the experimental animals, they evidently fail to interact properly. The effects of ectopic expression of homeodomain regulators during development have also been reported in the mouse (reviewed by Kessel and Gruss, 1990). In one interesting case, overexpression of a germline Hox-1.4 construct under the control of an SV40 promoter resulted in apparent failure of innervation of the large intestine (Wolgemuth et al. 1989) either because of an effect on the innervating neural crest derivatives, or on the gut mesenchyme during development. In either case, the result could again be interpreted as interference with specific processes of intercellular interaction. Germline deletion of the Hox-1.5 gene (Chisaka and Capecchi, 1991) in the mouse causes a series of defects in structures that derive from the pharyngeal arches and pouches, including thymus, parathyroid, thyroid, arteries and lower facial structures. These defects can be interpreted as the direct downstream consequences of failures of inductive interactions between cephalic neural crest and pharyngeal endoderm, mesoderm and ectoderm (Chisaka and Capecchi, 1991). Thus the Hox-1.5 gene might be specifically required for the regional manifestation of the ligands and receptors, signalling machinery, etc. that modulate these interactions. Another striking and relevant observation is that, as shown by Thomas and Capecchi (1990) and McMahon and Bradley (1990), germline disruption of the Wnt-1 gene results in what appear to be modular regional defects in the ongoing inductive processes by which morphogenesis of the cerebellum and mesencephalon occurs. The mouse Wnt-1 gene apparently encodes a secreted ligand homologous to that produced by the wingless gene of Drosophila (van den Heuvel et al. 1989), and it is a reasonable speculation that there is a relation between Wnt-1 expression, and the expression of at least one homeodomain regulator, engrailed, which may be required for this aspect of CNS development (McMahon and Bradley, 1990).

The arguments in this section suggest a mechanism by which homeodomain regulators may work in the late embryonic and postembryonic vertebrate development. Using this idea as a lens, it is interesting now to examine again the three major types of early embryonic process, and their evolutionary interrelationships.

Different modes of activation of regionally expressed homeobox genes in early development

The later developmental processes by which organs are formed and by which morphogenesis of three-dimensional multicellular structures occurs are probably common to all animal forms above sponges. Assuming the downstream functions of regionally expressed homeobox genes envisioned in respect to vertebrate development, we might therefore expect to encounter these same functions universally, wherever short-range inductive morphogenetic functions are required. This should, for example, include the regional inductive interactions that mediate anteroposterior diversification in bilateral animals. On the other hand, the initial mechanisms by which an egg turns into an embryo capable of later generating subsequent regional morphogenesis, are clearly different. If regionally expressed homeobox genes control local inductive morphogenesis, then in the types of embryo that we have discussed the mechanistic meaning of these differences will be revealed, in part, by the ways that homeobox genes are initially activated. ‘

In Type 1 embryos cell specification occurs in a fixed spatial geometry with respect to lineage, by conditional and autonomous processes. There is no obvious requirement for early regional homeobox gene expression (i.e. expression in cells that will give rise to a variety of different cell types), since the positions of the initial embryonic tissues with respect to the larval body plan are already established by lineage, cleavage pattern and the geometry of gastrular invagination. Of course such genes might indeed control conditional specification by intercellular interaction in particular lineages, or be used for other functions in early Type 1 development. The arguments given above would suggest that regional control functions will only be required in postembryonic Type 1 development, as lineage relations blur, cells multiply and migrate, and are organized de novo into organs and structures interrelated in position (except perhaps in heterochronic direct developing species). Unfortunately, not much is known about spatial patterns of homeobox gene expression in Type 1 embryos. A homeobox gene called AHox-1 is expressed in ascidian embryos but not until larval and juvenile stages (Saiga et al. 1991). Transcripts appear in gut, at the time of the functional differentiation of this organ, and in juvenile coelomocytes. The ascidian endoderm derives from a fixed embryonic lineage, as described earlier, and juvenile coelomocytes are probably also of endodermal lineage (Whittaker, 1990). In the sea urchin embryo, transcripts of a gene called TgHbox-1 accumulate in a particular region of the aboral ectoderm, in mid-late embryogenesis (Angerer et al. 1989). However, cytotypic marker gene expression in aboral ectoderm is initiated in cleavage, long prior to TgHbox-1 expression, so the conditional process of initial aboral ectoderm specification cannot depend on the activity of this homeobox gene. Expression in the late embryo appears to be confined to a single clonal lineage, which is the descendant of the VA founder cell, located at the posterior vertex of the larva. The postembryonic role of this regulator could be to mediate intercellular interactions in larval skeletoge-nesis, which will be active in the vertex region. In C. elegans a homeobox gene called mab-5 is required for postembryonic morphogenesis in the posterior region, where it is expressed after completion of embryonic founder cell specification in many different cell types (Kenyon, 1986; Costa et al. 1988). Among a variety of morphogenetic functions mab-5 is required for development of certain male sensory structures, and in this process mab-5 expression is under the control of a cell ‘ signalling system (Waring and Kenyon, 1990, 1991). Thus this gene both responds to, and probably regulates (to account for its morphogenetic effects), specific intercellular interactions, just as is typical for regionally expressed homeobox genes in postembryonic vertebrate development.

In the Type 2 embryos of vertebrates the initial regional expression of homeobox genes is apparently activated by long-range diffusible morphogens that are responsible for the broad organization of the body plan. Regional expression of specific homeobox genes has been related to the activity of known morphogens in Xenopus. Ruiz i Altaba and Melton (1989b) showed, for example, that Xhox-3 is transcribed in isolated animal caps if these are exposed to growth factors of the TGFβ and bFGF families that are known to be required for the initial axial specification of mesoderm. As we recall Xhox-3 transcripts are normally concentrated in the dorsal posterior mesoderm at late gastrula-neurula stages, and the accumulation of these transcripts is stimulated by high levels of bFGF, a known morphogen for posterior-dorsal mesoderm. High levels of a TGFβ relative (XTC-MIF), that induce anterior-dorsal mesoderm, correspondingly depress Xhox-3 expression. Similarly Cho and DeRobertis (1990) showed that bFGF activates XlHbox-6, which is expressed in the posterior CNS (Sharpe et al. 1987; Wright et al. 1990), while XTC-MIF also activates XlHbox-1, which is expressed in the anterior trunk (see above). The causal linkage between early embryonic processes that initially specify anteroposterior fates and later regional homeobox gene expression is further demonstrated by the effects of UV treatment of zygotes, and Li+ treatment of early embryos (Ruiz i Altaba and Melton, 1989c). Ectopic respecification of anterior-dorsal fates toward posterior-ventral fates by UV irradiation results in elevation of Xhox-3 expression, and the ectopic respecification toward anterior-dorsal fates caused by exposure to Li+ results in depression of Xhox-3 expression. Clearly Xhox-3 is not autonomously specified with respect to its domain of expression. Another diffusible morphogen that effects anteroposterior specification in the CNS and mesoderm of Xenopus is retinoic acid (Durston et al. 1989; Sive et al. 1990). This agent promotes posterior specification, and it also promotes expression of XIHbox-6. Like the transcripts of the homeobox genes whose activity they initially control, in gastrula-neurula stage embryos, the effects of retinoic acid are distributed regionally, in that developmental-sensitivity to this morphogen is observed at the organ or body plan level. Thus retinoic acid treatment alters development of anterior CNS, cement gland and other anterior structures that include contributions from both ectodermal and mesodermal derivatives (Sive et al. 1990). These early effects of retinoic acid in Xenopus embryos can at least in part be accounted for as modulation of the anterior mesoderm specification initially induced by growth factors (Ruiz i Altaba and Jessell, 1991). Thus retinoic acid apparently acts downstream of Xhox-3 expression in the mesoderm but upstream of other homeobox gene expressions involved in regional morphogenesis. Retinoic acid probably also exercises important developmental functions in mammalian and avian embryos. Thus it activates homeobox genes in cultured EC cells (e.g. Deschamps et al. 1987; Simeone et al. 1990). Forced ectopic expression of the Hox-1.1 gene in transgenic mice causes a distinct subset of the morphological phenotypes caused by high doses of retinoic acid, and that is probably due to failure of morphogenetic neural crest cell functions in the head and face (Balling et al. 1989). This interpretation implies that in later development as well, the morphogenetic effects of retinoic acid may be directly exercised through homeobox gene activation.

In Type 3 embryonic process homeobox genes are initially activated in an entirely different way, dependent neither on diffusible growth factors, nor on shortrange cell interactions or cell lineage. As discussed above, in these embryos homeobox genes are used to generate the embryonic body plan ab initio, rather than being used to carry morphogenesis onward following prior generation of the body plan by other means. The properties of pair-rule homeobox gene expression that enable them to regulate one another are utilized to set up precellular spatial patterns of regulatory gene expression. Among these properties are regulatory domains that include positively and negatively acting target sites for the homeodomain protein products of other such genes, as well as for other transcription factors (e.g. see Dearolf et al. 1989; Biggin and Tjian, 1988; Hayashi and Scott, 1990); and also their clustered genomic organization, as discussed in an interesting way by Peifer et al. (1987). As stressed earlier, homeobox gene interactions are only one of several unique regulatory devices utilized in precellular Drosophila embryos for spatial patterning, but their mode of initial utilization highlights the distinction between Type 3 and other forms of early embryonic processes. After the initial spatial patterning is established, however, i.e. from the beginning of gastrulation on, Drosophila embryos utilize their homeobox genes in ways more similar to those seen in postembryonic vertebrate development. As noted above, homeodomain regulators of the segment polarity group such as engrailed, evidently control expression of ligands and receptors (directly or indirectly). A very interesting example recently described is the relation between dpp and wingless expression, and homeotic gene expression, in the midgut of Drosophila (Panganiban et al. 1990; Immerglück et al. 1990; Reuter et al. 1990; Reuter and Scott, 1990). The products of the dpp gene, a TGFβ family member, and of the wg gene, apparently a diffusible ligand, are required for the inductive morphogenesis of the midgut. These factors are produced in the visceral mesoderm, and the genes encoding them are evidently downstream targets of homeotic genes that regulate midgut development. In parasegment8 expression of the AbdA gene causes wg expression and blocks dpp expression, and, in parasegment 7, Ubx gene expression causes dpp expression. The effects observed in various ectopic expression experiments in which these homeotic selector genes were driven by heat-shock promoters, suggest that the dpp gene may be a direct target of the homeodomain regulatory factors (Reuter et al. 1990). The dpp factor in turn affects Ubx expression in the visceral mesoderm, and induces labial gene expression in the adjacent endodermal epithelium. Mutations that prevent dpp expression in the midgut have modular effects on midgut morphogenesis, here failure of given local structures to be formed. In the midgut the exclusive regional patterns of homeobox gene expression are thus mediated by intercellular inductive effects on homeobox gene expression, by the negative interactions of homeodomain regulators with other homeobox genes within the same nucleus, and by the activation of intercellular signalling systems by homeodomain regulators that in turn effect induction of homeobox genes (Tremml and Bienz, 1989; Reuter et al. 1990). Similar sets of regulatory interactions are likely to be involved in control of homeotic gene function in other regions of the post-cellularization Drosophila embryo, and maybe used to define the boundaries of the multicellular regions of homeodomain regulator expression in all forms of embryo. That is, intercellular signalling that regulates homeobox gene expression’ is likely to be required at the lateral boundaries of regions that are defined by the patterns of expression of these genes.

The modular nature of the downstream effects of homeodomain regulators depends of course on their regional expression. This suggests that true homeotic effects, i.e. exchange of one modular morphogenetic program for another, might operate by simply causing a switch of one set of downstream ligand-receptor interactions that promote a certain morphogenesis for a different one.

Evolutionary derivation of embryonic processes

The implication of Fig. 1 is that Type 2 and 3 embryogenesis are, in evolutionary terms, derived. The view adduced here is that, while all Metazoa probably utilize the same kinds of intercellular, conditional specification processes for organogenesis and morphogenesis, there are several quite different processes by which gastrula-stage embryos are generated from eggs. Such differences can be seen starting to occur in direct developing forms of various Type 1 taxa, and are extensively amplified in what I have called Type 2 and Type 3 embryonic processes. One way of imagining how this might have come about is to regard both of these as heterochronic derivatives of Type 1 processes. Type 3 embryos exemplify a revolutionary, early use of regional homeobox gene expression, together with other special regulatory adaptations. Furthermore, in these embryos the autonomous, precellular development of spatial regulatory gene expression patterns is, until the end of this stage of the process, separated from direct control of downstream target genes that mediate intercell interaction. In Type 2 embryonic process, regional homeodomain regulators are used extensively in mid-late embryonic development, because even at these relatively early stages these embryos utilize the same kinds of morphogenetic mechanisms that continue in postembryonic development. It is a fascinating thought that the progenitors of the advanced insects, and the different progenitors of the vertebrates, both found in the available library of postembryonic genetic regulatory devices ways to replace the lineage-based processes of early embryogenesis, which from the beginning of the Cambrian to the present probably provided the universal basic pathways for embryogenesis.

Needless to say, I have relied heavily on the thoughtful critical advice, and the expert insights of many colleagues who generously expended their valuable time in reviewing drafts of this manuscript. I cannot too strongly express my appreciation to the following people: Frank Calzone, UC Irvine; Rick Firtel, UC San Diego; Gary Freeman, University of Texas; Antonio Garcia-Bellido, Centro de Biología Molecular, Madrid; Mike Levine, UC San Diego; Andrew McMahon, Roche Institute; Jim Posakony, UC San Diego; Richard Whittaker, Marine Biological Laboratory; and Roy Britten, Andy Cameron, Scott Fraser, Barbara Hough-Evans, Howard Lipshitz, Ellen Rothenberg, and Paul Sternberg, all of Caltech. This work was partially supported by a grant from the National Institutes of Health (HD-05753).

Akam
,
M.
(
1987
).
The molecular basis for metameric pattern in the Drosophila embryo
.
Development
101
,
1
22
.
Anderson
,
D. J.
(
1989
).
The neural crest cell lineage problem: neuropoiesis?
Neuron
3
,
1
12
.
Anderson
,
D. T.
(
1972a
).
The development of hemimetabolous insects
.
In Developmental Systems: Insects vol. 1 (eds, S. J. Counce and C. H. Waddington)
, pp.
95
163
.
London
:
Academic Press
.
Anderson
,
D. T.
(
1972b
).
The development of holometabolous insects
.
In Developmental Systems: Insects vol. 1 (eds, S. J. Counce and C. H. Waddington)
, pp.
165
242
.
London
:
Academic Press
.
Anderson
,
D. T.
(
1973
).
Embryology and Phylogeny in Annelids and Arthropods
.
Oxford
:
Pergammon
.
Angerer
,
L. M.
,
Dolecki
,
G. J.
,
Gagnon
,
M. L.
,
Lum
,
R.
,
Wang
,
G.
,
Yang
,
Q.
,
Humphreys
,
T.
and
Angerer
,
R. C.
(
1989
)
Progressively restricted expression of a horneo box gene within the aboral ectoderm of developing sea urchin embryos
.
Genes Dev
.
3
,
370
383
.
Artavanis-Tsakonas
,
S.
(
1988
).
The molecular biology of the Notch locus and the fine tuning of differentiation in Drosophila
.
Trends Genet
.
4
,
95
100
.
Astrow
,
S.
,
Holton
,
B.
and
Weisblat
,
D.
(
1987
).
Centrifugation redistributes factors determining cleavage patterns in leech embryos
.
Devi Biol
.
120
,
270
283
.
Austin
,
J.
and
Kimble
,
J.
(
1989
).
Transcript analysis of glp-1 and lin-12, homologous genes required for cell interactions during development of C. elegans
.
Cell
58
,
565
571
.
Balling
,
R.
,
Mutter
,
G.
,
Gruss
,
P.
and
Kessel
,
M.
(
1989
).
Craniofacial abnormalities induced by ectopic expression of the homeobox gene Hox-1.1 in transgenic mice
.
Cell
58
,
337
347
.
Barnes
,
R. D.
(
1980
).
Invertebrate Zoology, Fourth Edition
.
Philadelphia
:
Saunders College
.
Bates
,
W. R.
and
Jeffery
,
W. R.
(
1988
).
Polarization of ooplasmic segregation and dorsal-ventral axis determination in ascidian embryos
.
Devi Biol
.
130
,
98
107
.
Baumhueter
,
S.
,
Mendel
,
D. B.
,
Conley
,
P. B.
,
Kuo
,
C. J.
,
Turk
,
C.
,
Graves
,
M. K.
,
Edwards
,
C. A.
,
Courtois
,
G.
and
Crabtree
,
G. R.
(
1990
).
HNF-1 shares three sequence motifs with the POU domain proteins and is identical to LF-B1 and APF
.
Genes Dev
.
4
,
372
379
.
Beer
,
J.
,
Technau
,
G. M.
and
Campos-Ortega
,
J. A.
(
1987
).
Lineage analysis of transplanted individual cells in embryos of Drosophila melanogaster. IV. Commitment and proliferative capabilities of mesodermal cells
.
Roux’ Arch, devl Biol
.
196
,
222
230
.
Benson
,
S. C.
,
Crise Benson
,
N.
and
Wilt
,
F.
(
1986
).
The organic matrix of the skeletal spicule of sea urchin embryos
.
J. Cell Biol
.
102
,
1878
1886
.
Berridge
,
M. J.
,
Downes
,
C. P.
and
Hanley
,
M. R.
(
1989
).
Neural and developmental actions of lithium: a unifying hypothesis
.
Cell
59
,
411
419
.
Biggin
,
M. D.
and
Than
,
R.
(
1988
).
Transcriptional factors that activate the Ultrabithorax promoter in developmentally staged extracts
.
Cell
53
,
699
711
.
Botas
,
J.
,
Cabrera
,
C. V.
and
Garcia-Bellido
,
A.
(
1988
).
The reinforcement-extinction process of selector gene activity: a positive feed-back loop and cell-cell interactions in Ultrabithorax patterning
.
Roux’ Arch, devl Biol
.
197
,
424
434
.
Boyer
,
B. C.
(
1986
).
Determinative development in the polyclad turbellarian Hoploplana inquilina
.
Intn. J. Invert. Reprod. Dev
.
9
,
243
251
.
Boyer
,
B. C.
(
1987
).
Development of in vitro fertilized embryos of the polyclad flatworm, Hoploplana inquilina, following blastomere separation and deletion
.
Roux’ Arch, devl Biol
.
196
,
158
164
.
Boyer
,
B. C.
(
1989
).
The role of the first quartet micromeres in the development of the polyclad Hoploplana inquilina
.
Biol. Bull. mar. biol. Lab., Woods Hole
177
,
338
343
.
Bronner-Fraser
,
M.
and
Fraser
,
S.
(
1989
).
Developmental potential of avian trunk neural crest cells in situ
.
Neuron
3
,
755
766
.
Burns
,
R. K.
(
1955
).
Urinogenital System
.
In Analysis of Development
(ed.
B. H.
Willier
,
P. A.
Weiss
, and
V.
Hamburger
), pp.
462
491
.
Philadelphia
:
W. B. Saunders Company
.
Cabrera
,
C. V.
,
Alonso
,
M. C.
,
Johnston
,
P.
,
Phillips
,
R. G.
and
Lawrence
,
P. A.
(
1987
).
Phenocopies induced with antisense RNA identify the wingless gene
.
Cell
50
,
659
663
.
Campos-Ortega
,
J. A.
(
1990
).
Mechanisms of a cellular decision during embryonic development of Drosophila: epidermogenesis or neurogenesis
.
Adv. Gen
.
27
,
403
453
.
Carroll
,
S. B.
,
Dinardos
,
S.
,
O’Farrell
,
P. H.
,
White
,
R. A. H.
and
Scott
,
M. P.
(
1988
).
Temporal and spatial relationships between segmentation and homeotic gene expression in Drosophila embryos: distributions of the fushi taraza, engrailed, Sex combs reduced, Antennapedia, and Ultrabithorax proteins
.
Genes Dev
.
2
,
350
360
.
Carroll
,
S. B.
,
Winslow
,
G. M.
,
Twombly
,
V. J.
and
Scott
,
M. P.
(
1987
).
Genes that control dorsoventral polarity affect gene expression along the anteroposterior axis of the Drosophila embryo
.
Development
99
,
327
332
.
Casanova
,
J.
and
Struhl
,
G.
(
1990
).
Localized surface activity of torso, a receptor tyrosine kinase, specifies terminal body pattern in Drosophila
.
Genes Dev
.
3
,
2025
2038
.
Chisaka
,
O.
and
Capecchi
,
M. R.
(
1991
).
Regionally restricted developmental defects resulting from targeted disruption of the mouse homeobox gene hox-1.5
.
Nature
350
,
473
479
.
Cho
,
K. W. Y.
and
De Robertis
,
E. M.
(
1990
).
Differential activation of Xenopus horneo box genes by mesoderm-inducing growth factors and retinoic acid
.
Genes Dev
.
4
,
1910
1916
.
Chuong
,
C.-M.
,
Oliver
,
G.
,
Ting
,
S. A.
,
Jegalian
,
B. G.
,
Chen
,
H. M.
and
De Robertis
,
E. M.
(
1990
).
Gradients of homeoproteins in developing feather buds
.
Development
110
,
1021
1030
.
Clerc
,
R. G.
,
Corcoran
,
L. M.
,
Lebowitz
,
J. H.
,
Baltimore
,
D.
and
Sharp
,
P. A.
(
1988
).
The B-cell-specific Oct-2 protein contains POU box- and horneo box-type domains
.
Genes Dev
.
2
,
1570
1581
.
Cohen
,
S. M.
and
Jurgens
,
G.
(
1990
).
Mediation of Drosophila head development by gap-like segmentation genes
.
Nature
346
,
482
485
.
Colwin
,
A. L.
and
Colwin
,
L. H.
(
1953
).
The normal embryology of Saccoglossus kowalevskii (Enteropneusta)
.
J. Morph
.
92
,
401
453
.
Conklin
,
E. G.
(
1905
).
The organization and cell lineage of the ascidian egg
.
J. Acad. natn. Sci. Phila
.
13
,
1
119
.
Conklin
,
E. G.
(
1932
).
The embryology of Amphioxus
.
J. Morph
.
54
,
69
120
.
Cooke
,
J.
and
Wong
,
A.
(
1991
).
Growth-factor-related proteins that are inducers in early amphibian development may mediate similar steps in amniote (bird) embryogenesis
.
Development
111
,
197
212
.
Copenhaver
,
W. M.
(
1955
).
Heart, blood vessels, blood, and entodermal derivatives
.
In Analysis of Development
(ed.
B. H.
Willier
,
P. A.
Weiss
, and
V.
Hamburger
), pp.
440
461
.
Philadelphia
:
W. B. Saunders Company
.
Costa
,
M.
,
Weir
,
M.
,
Coulson
,
A.
,
Sulson
,
J.
and
Kenyon
,
C.
(
1988
)
Posterior pattern formation in C. elegans involves position-specific expression of a gene containing a homeobox
.
Cell
55
,
747
756
.
Cowan
,
A. E.
and
Mcintosh
,
J. R.
(
1985
).
Mapping the distribution of differentiation potential for intestine, muscle and hypodermis during early development in Caenorhabditis elegans
.
Cell
41
,
923
932
.
Crowther
,
R. J.
and
Whittaker
,
J. R.
(
1986
).
Differentiation without cleavage: multiple cytospecific ultrastructural expressions in individual one-celled ascidian embryos
.
Devi Biol
.
117
,
114
116
.
Dale
,
L.
and
Slack
,
J. M. W.
(
1987
).
Fate map for the 32-cell stage of Xenopus laevis
.
Development
99
,
527
551
.
Dalton
,
D.
,
Chadwick
,
R.
and
Mcginnis
,
W.
(
1989
).
Expression and embryonic function of empty spiracles: a Drosophila homeobox gene with two patterning functions on the anterior-posterior axis of the embryo
.
Genes Dev
.
3
,
1940
1956
.
Davidson
,
E. H.
(
1986
).
Gene Activity in Early Development
, Third Edition.
Orlando, Florida
:
Academic Press
.
Davidson
,
E. H.
(
1989
).
Lineage-specific gene expression and the regulative capacities of the sea urchin embryo: a proposed mechanism
.
Development
105
,
421
445
.
Davidson
,
E. H.
(
1990
).
How embryos work: a comparative view of diverse modes of cell fate specification
.
Development
108
,
365
389
.
Dearolf
,
C. R.
,
Topol
,
J.
and
Parker
,
C. S.
(
1989
).
Transcriptional control of Drosophila fushi tarazu zebra stripe expression
.
Genes Dev
.
3
,
384
389
.
De Leo
,
G.
(
1972
).
Data on the cell-lineage and the early stages of development of Sepiola rondelettii
.
Acta Embryol. exp
.
1972
,
25
44
.
Delgadillo-Reynoso
,
M. G.
,
Rollo
,
D. R.
,
Hursh
,
D. A.
and
Raff
,
R. A.
(
1989
).
Structural analysis of the uEGF gene in the sea urchin Strongylocentrotus purpurtus reveals more similarity to vertébrate thart to invertebrate genes with EGF-like repeats
.
J. molec. Evol
.
29
,
314
327
.
Deschamps
,
J.
,
De Laaf
,
R.
,
Joosen
,
L.
,
Meijlink
,
F.
and
Destree
,
O.
(
1987
).
Abundant expression of homeobox genes in mouse embryonal carcinoma cells correlates with chemically induced differentiation
.
Proc. natn. Acad. Sci. U.S.A
.
84
,
1304
1308
.
Dollé
,
P.
,
Izpisúa-Belmonte
,
J.-C.
,
Falkenstein
,
H.
,
Renucci
,
A
and
Duboule
,
D.
(
1989
).
Coordinate expression of the murine Hox-5 complex homeobox-containing genes during limb pattern formation
.
Nature
342
,
767
772
.
Dollé
,
P.
,
Ruberte
,
E.
,
Leroy
,
P.
,
Morriss-Kay
,
G.
and
Chambón
,
P.
(
1990
).
Retinoic acid receptors and cellular retinoid binding proteins. I. A systematic study of their differential pattern of transcription, during mouse organogenesis
.
Development
110
,
1113
1151
.
Duprey
,
P.
,
Chowdhury
,
K.
,
Dressler
,
G. R.
,
Balling
,
R.
,
Simon
,
D.
,
Guenet
,
J.-L.
and
Gruss
,
P.
(
1988
).
A mouse gene homologous to the Drosophila gene caudal is expressed in epithelial cells from the embryonic intestine
.
Genes Dev
.
2
,
1647
1654
.
Durston
,
A. J.
,
Timmermans
,
J. P. M.
,
Hage
,
W. J.
,
Hendriks
,
H. F. J.
,
De Vries
,
N. J.
,
Heideveld
,
M.
and
Nieuwkoop
, (
1989
)
P D Retinoic acid causes an anteroposterior transformation in the developing central nervous system
.
Nature
340
,
140
144
.
Edgar
,
L. G.
and
Mcghee
,
J. D.
(
1986
).
Embryonic expression of a gut-specific esterase in Caenorhabditis elegans
.
Devi Biol
.
114
,
109
118
.
Emig
,
C. C.
(
1977
).
Embryology of Phoronida
.
Am. Zool
.
17
,
21
37
.
Erselius
,
J. R.
,
Goulding
,
M. D.
and
Gruss
,
P.
(
1990
).
Structure and expression pattern of the murine Hox-3.2 gene
.
Development
110
,
629
642
.
Fehon
,
R. C.
,
Kooh
,
P. J.
,
Rebay
,
L
,
Regan
,
C. L.
,
Xu
,
T.
,
Muskavitch
,
M. A. T.
and
Artavanis-Tsakonas
,
S.
(
1990
).
Molecular interactions between the protein products of the neurogenic loci notch and delta, two EGF-homologous genes in Drosophila
.
Cell
61
,
523
534
.
Finkelstein
,
R.
and
Perrimon
,
N.
(
1990
).
The orthodenticle gene is regulated by bicoid and torso and specifies Drosophila head development
.
Nature
346
,
485
488
.
Freeman
,
G.
(
1976a
).
The role of cleavage in the localization of developmental potential in the ctenophore Mnemiopsis leidyi
.
Devi Biol
.
49
,
143
177
.
Freeman
,
G.
(
1976b
).
The effects of altering the position of cleavage planes on the process of localization of developmental potential in ctenophores
.
Devi Biol
.
51
,
332
337
.
Freeman
,
G.
(
1977
).
The establishment of the oral-aboral axis in the ctenophore embryo
.
J. Embryol. exp. Morph
.
42
,
237
260
.
Freeman
,
G.
(
1978
).
The role of asters in the localization of the factors that specify the apical tuft and the gut of the nemertine Cerebratulus lacteus
.
J. exp. Zool
.
206
,
81
108
.
Freeman
,
G.
(
1981
).
The role of polarity in the development of the hydrozoan planula larva
.
Wilhelm Roux’ Arch, devl Biol
.
190
,
168
184
.
Freeman
,
G.
(
1983
).
Experimental studies on embryogenesis in hydrozoans (Trachylina and Siphonophora) with direct development
.
Biol. Bull. mar. biol. Lab., Woods Hole
165
,
591
618
.
Freeman
,
G.
(
1990
).
The establishment and role of polarity during embryogenesis in hydrozoans
.
In The Cellular and Molecular Biology of Pattern Formation
(ed.
D. L.
Stocum
and
T. L.
Karr
).
Oxford
:
Oxford University Press
.
Freeman
,
G.
(
1991
).
The bases for and timing of regional specification during larval development in Phoronis
.
Submitted for publication
.
Freeman
,
G.
and
Miller
,
R. L.
(
1982
).
Hydrozoan eggs can only be fertilized at the site of polar body formation
.
Devi Biol
.
94
,
142
152
.
Freeman
,
G.
and
Reynolds
,
G. T.
(
1973
).
The development of bioluminescence in the ctenophore Mnemiopsis leidyi
.
Devi Biol
.
31
,
61
100
.
Frohman
,
M. A.
,
Boyle
,
M.
and
Martin
,
G. R.
(
1990
).
Isolation of the mouse Hox-2.9 gene; analysis of embryonic expression suggests that positional information along the anterior-posterior axis is specified by mesoderm
.
Development
110
,
589
607
.
Gardner
,
R. L.
and
Rossant
,
J.
(
1979
).
Investigation of the fate of 4-5 day post coitum mouse inner cell mass cells by blastocyst injection
.
J. Embryol. exp. Morph
.
52
,
141
152
.
Gaunt
,
S. J.
(
1987
).
Homeobox gene Hox-1.5 expression in mouse embryos: earliest detection by in situ hybridization is during gastrulation
.
Development
101
,
51
60
.
Gaunt
,
S. J
: (
1988
).
Mouse homeobox gene transcripts occupy different but overlapping domains in embryonic germ layers and organs: a comparison of Hox-3.1 and hox-1.5
.
Development
103
,
135
144
.
Gaunt
,
S. J.
,
Krumlauf
,
R.
and
Duboule
,
D.
(
1989
).
Mouse homeo-genes within a subfamily, Hox-1.4, -2.6 and -5.1, display similar anteroposterior domains of expression in the embryo, but show stage- and tissue-dependent differences in their regulation
.
Development
107
,
131
141
.
Gavin
,
B. J.
,
Mcmahon
,
J. A.
and
Mcmahon
,
A. P.
(
1990
).
Expression of multiple novel Wnt-/bir-7-related genes during fetal and adult mouse development
.
Genes and Dev
.
4
,
2319
2332
.
Gimlich
,
R.
and
Gerhart
,
J.
(
1984
).
Early cellular interactions promote embryonic axis formation in Xenopus
.
Devi Biol
.
104
,
117
130
.
Godsave
,
S. F.
and
Slack
,
J. M. W.
(
1991
).
Single cell analysis of mesoderm formation in the Xenopus embryo
.
Development
111
,
523
530
.
Goto
,
T.
,
Macdonald
,
P.
and
Maniatis
,
T.
(
1989
).
Early and late periodic patterns of even skipped expression are controlled by distinct regulatory elements that respond to different spatial cues
.
Cell
57
,
413
422
.
Graham
,
A.
,
Papalopulu
,
N.
,
Lorimer
,
J.
,
Mcvey
,
J. H.
,
Tuddenham
,
E. G. D.
and
Krumlauf
,
R.
(
1988
).
Characterization of a murine homeobox gene, Hox-2.6, related to the Drosophila Deformed gene
.
Genes Dev
.
2
,
1424
1438
.
Green
,
J.
(
1971
).
Crustaceans
.
In Experimental Embryology of Marine and Fresh-water Invertebrates
(ed.
G.
Reverberi
), pp.
312
362
.
Amsterdam
:
North-Holland
.
Green
,
J. B. A.
and
Smith
,
J. C.
(
1990
).
Graded changes in dose of Xenopus activin-A homolog elicit stepwise transitions in embryonic-cell fate
.
Nature
347
,
391
394
.
Grimwade
,
J. E.
,
Gagnon
,
M. L.
,
Yang
,
Q.
,
Angerer
,
R. C.
and
Angerer
,
L. M.
(
1991
).
Expression of two RNAs encoding EGF-related proteins identifies subregions of sea urchin embryonic ectoderm
.
Devi Biol
.
143
,
44
57
.
Gurdon
,
J. B.
,
Fairman
,
S.
,
Mohun
,
T. J.
and
Brennan
,
S.
(
1985
).
Activation of muscle-specific actin genes in Xenopus development by an induction between animal and vegetal cells of a blastula
.
Cell
41
,
913
922
.
Harding
,
K.
and
Levine
,
M.
(
1988
).
Gap genes define the limits of antennapedia and bithorax gene expression during early development in Drosophila
.
EMBO J
.
7
,
205
214
.
Hartenstein
,
V.
and
Posakony
,
J. W.
(
1990
).
A dual function of the Notch gene in Drosophila sensillum development
.
Devi Biol
.
142
,
13
30
.
Harvey
,
R. P.
and
Melton
,
D. A.
(
1988
).
Microinjection of synthetic Xhox-1A homeobox mRNA disrupts somite formation in developing Xenopus embryos
.
Cell
53
,
687
697
.
Hayashi
,
S.
and
Scott
,
M. P.
(
1990
).
What determines the specificity of action of Drosophila homeodomain proteins?
Cell
63
,
883
894
.
Henry
,
J. J.
(
1986
).
The role of unequal cleavage and the polar lobe in the segregation of developmental potential during first cleavage in the embryo of Chaetopterus variopedatus
.
Roux Arch, devl Biol
.
195
,
103
116
.
Henry
,
J. J.
(
1989
).
Removal of the polar lobe leads to the formation of functionally deficient photocytes in the annelid Chaetopterus variopedatus
.
Roux Arch, devl Biol
.
198
,
129
136
.
Henry
,
J. J.
and
Martindale
,
M. Q.
(
1987
).
The organizing role of the D quadrant as revealed through the phenomenon of twinning in the polycheate Chaetopterus variopedatus
.
Roux Arch, devl Biol
.
196
,
499
510
.
Hickey
,
R. J.
,
Boshar
,
M. F.
and
Crain
,
W. R.
, JR
(
1987
).
Transcription of three actin genes and a repeated sequence in isolated nuclei of sea urchin embryos
.
Devl Biol
.
124
,
215
227
.
Hidalgo
,
A.
and
Ingham
,
P.
(
1990
).
Cell patterning in the Drosophila segment: spatial regulation of the segment polarity gene patched
.
Development
110
,
291
301
.
Holtfreter
,
J.
(
1944
).
Experimental studies on the development of the pronephros
.
Revue Canadienne de Biologie
3
,
220
250
.
Holtfreter
,
J.
and
Hamburger
,
V.
(
1955
).
Amphibians
.
In Analysis of Development
(ed.
B. H.
Willier
,
P. A.
Weiss
, and
V.
Hamburger
), pp.
230
296
.
Philadelphia
:
W. B. Saunders Company
.
Hooper
,
J. E.
and
Scott
,
M. P.
(
1989
).
The Drosophila patched gene encodes a putative membrane protein required for segmental patterning
.
Cell
59
,
751
785
.
Hopwood
,
N. D.
,
Pluck
,
A.
and
Gurdon
,
J. B.
(
1989
).
MyoD expression in the forming somites is an early response to mesoderm induction in Xenopus embryos
.
EMBO J
.
8
,
3409
3417
.
Horstadius
,
S.
(
1937
).
Experiments on determination in the early development of Cerebratulus lacteus
.
Biol. Bull. mar. biol. Lab., Woods Hole
73
,
317
342
..
Horstadius
,
S.
(
1939
).
The mechanics of sea urchin development studied by operative methods
.
Biol. Rev. Cambridge Phil. Soc
.
14
,
132
179
.
Horstadius
,
S.
(
1973
).
Experimental Embryology of Echinoderms
.
London
:
Clarendon Press
.
Howard
,
K. R.
and
Struhl
,
G.
(
1990
).
Decoding positional information: regulation of the pair-rule gene hairy
.
Development
110
,
1223
1231
.
Hunt
,
P.
,
Wilkinson
,
D.
and
Krumlauf
,
R.
(
1991
).
Patterning the vertebrate head: murine HOX-2 genes mark distinct subpopulations of premigratory and migrating cranial neural crest
.
Development
,
112
,
43
50
.
Immerglück
,
K.
,
Lawrence
,
P. A.
and
Bienz
,
M.
(
1990
).
Induction across germ layers in Drosophila mediated by a genetic cascade
.
Cell
62
,
261
268
.
Ingham
,
P. W.
(
1988
).
The molecular genetics of embryonic pattern formation in Drosophila
.
Nature
335
,
25
34
.
Ingraham
,
H. A.
,
Chen
,
R.
,
Mangalam
,
H. J.
,
Elsholtz
,
H. P.
,
Flynn
,
S. E.
,
Lin
,
C. R.
,
Simmons
,
D. M.
,
Swanson
,
L.
and
Rosenfeld
,
M. G.
(
1988
).
A tissue-specific transcription factor containing a homeodomain specifies a pituitary phenotype
.
Cell
55
,
519
529
.
Ip
,
Y. T.
,
Kraut
,
R.
,
Levine
,
M.
and
Rushlow
,
C. A.
(
1991
).
The dorsal morphogen is a sequence-specific DNA-binding protein that interacts with a long-range repression element in Drosophila
.
Cell
64
,
439
446
.
Irish
,
V. F.
and
Gelbart
,
W. M.
(
1987
).
The decapentaplegic gene is required for dorsal-ventral patterning of the Drosophila embryo
.
Genes Dev
.
1
,
868
879
.
Jamrich
,
M.
,
Sargent
,
T. D.
and
Dawid
,
I. B.
(
1987
).
Cell typespecific expression of epidermal cytokeratin genes during gastrulation of Xenopus laevis
.
Genes Dev
.
1
,
124
132
.
Jeffery
,
W. R.
(
1989
).
Requirement of cell division for muscle actin expression in the primary muscle cell lineage of ascidian embryos-
.
Development
105
,
75
84
.
Jeffery
,
W. R.
(
1990
).
Ultraviolet irradiation during ooplasmic segregation prevents gastrulation, sensory cell induction, and axis formation in the ascidian embryo
.
Devl Biol
.
140
,
388
400
.
Jeffery
,
W. R.
and
Swalla
,
B. J.
(
1990
).
Anural development in ascidians: Evolutionary modification and elimination of the tadpole larva
.
Seminars Devl Biol
.
1
,
253
261
.
Jeffries
,
R. P. S.
(
1986
).
The Ancestry of the Vertebrates
.
London
:
British Museum (Natural History
).
Jeffries
,
R. P. S.
(
1990
).
The solute Dendrocystoides scoticus from the Upper Ordovician of Scotland and the ancestry of chordates and echinoderms
.
Palaeontology
33
,
631
679
.
Jiang
,
J.
,
Hoey
,
T.
and
Levine
,
M.
(
1991
).
Autoregulation of a segmentation gene in Drosophila, combinatorial interaction of the even-skipped homeobox protein with a distal enhancer element
.
Genes Dev
.
5
,
265
277
.
Johansen
,
K. M.
,
Fehon
,
R. G.
and
Artavanis-Tsakonas
,
S.
(
1989
)
The Notch gene product is a glycoprotein expressed on the cell surface of both epidermal and neuronal precursor cells during Drosophila development
.
J. Cell Biol
.
109
,
2427
2440
.
Jones
,
C. M.
,
Lyons
,
K. M.
and
Hogan
,
B. L. M.
(
1991
).
Involvement of Bone Morphogenetic Protein-4 (BMP-4) and Vgr-1 in morphogenesis and neurogenesis in the mouse
.
Development
111
,
531
542
.
Jones
,
E. A.
and
Woodland
,
H. R.
(
1987
).
The development of animal cap cells in Xenopus. the effects of environment on the differentiation and the migration of grafted ectodermal cells
.
Development
101
,
23
32
.
Keleher
,
G. P.
and
Stent
,
G. S.
(
1990
).
Cell position and developmental fate in leech embryogenesis
.
Proc. natn. Acad. Sci. U.S.A
.
87
,
8457
8461
.
Keller
,
R. E.
(
1976
).
Vital dye mapping of the gastrula and neurula of Xenopus laevis. II. Prospective areas and morphogenetic movements of the deep layer
.
Devl Biol
.
51
,
118
137
.
Kenyon
,
C.
(
1986
).
A gene involved in the development of the posterior body region of C. elegans
.
Cell
48
,
477
487
.
Kessell
,
M.
and
Gruss
,
P.
(
1990
).
Murine developmental control genes
.
Science
249
,
374
379
.
Keynes
,
R. J.
and
Stern
,
C. D.
(
1988
).
Mechanisms of vertebrate segmentation
.
Development
103
,
413
429
.
Khaner
,
O.
and
Wilt
,
F.
(
1990
).
The influence of cell interactions and tissue mass on differentiation of sea urchin mesomeres
.
Development
109
,
625
634
.
Killian
,
C. E.
and
Wilt
,
F. H.
(
1989
).
The accumulation and translation of a spicule matrix protein mRNA during sea urchin embryo development
.
Devl Biol
.
133
,
148
156
.
Kimelman
,
D.
and
Kirschner
,
M.
(
1987
).
Synergisyic induction of mesoderm by FGF and TGF-/5 and the identification of an mRNA coding for FGF in the early Xenopus embryo
.
Cell
51
,
869
877
.
Kimmel
,
C. B.
(
1989
).
Genetics and early development of zebrafish
.
Trends Genet
.
5
,
283
288
.
Kimmel
,
C. B.
and
Warga
,
R. M.
(
1986
).
Tissue-specific cell lineages originate in the gastrula of the zebrafish
.
Science
231
,
365
368
.
Kimmel
,
C. B.
,
Warga
,
R. M.
and
Schilling
,
T. F.
(
1990
).
Origin and organization of the zebrafish fate map
.
Development
108
,
581
594
.
Krause
,
M.
,
Fire
,
A.
,
Harrison
,
S. W.
,
Priess
,
J.
and
Weintraub
,
H.
(
1990
).
CeMyoD accumulation defines the body wall muscle cell fate during C. legans embryogenesis
.
Cell
63
,
907
919
.
Kress
,
C.
,
Vogels
,
R.
,
De Graaff
,
W.
,
Bonnerot
,
C.
,
Majlink
,
F.
,
Nicolas
,
J.-F.
and
Deschamps
,
J.
(
1990
).
Hox-2.3 upstream sequences mediate lacZ expression in intermediate mesoderm derivatives of transgenic mice
.
Development
109
,
775
786
.
Kuhn-Schnyder
,
E.
and
Rieber
,
H.
(
1986
).
Handbook of Paleozoology
(
E.
Kucera
, trans.).
Baltimore
:
The Johns Hopkins University Press
.
Kumé
,
M.
and
Dan
,
K
(eds.) (
1968
).
Invertebrate Embryology
(
J. C.
Dan
, trans.).
NOLIT, Belgrade, Yugoslavia
.
Laflamme
,
S. E.
;
Jamrich
,
M.
,
Richter
,
K.
,
Sargent
,
T. D.
and
Dawid
,
I. B.
(
1988
).
Xenopus endo B is a keratin preferentially expressed in the embryonic notochord
.
Genes Dev
.
2
,
853
862
.
Lawson
,
K. A.
,
Meneses
,
J. J.
and
Pedersen
,
R. A.
(
1986
).
Cell fate and cell lineage in the endoderm of the presomite mouse embryo, studied with an intracellular tracer
.
Devl Biol
.
115
,
325
339
.
Le Douarin
,
N. M.
(
1990
).
Cell lineage segregation during neural crest ontogeny
.
Annals New York Acad. Sciences
559
,
131
140
.
Lee
,
J. J.
(
1986
).
Activation of sea urchin actin genes during embryogenesis: nuclear synthesis and decay rate measurements of transcripts from five different genes
.
Ph.D. Thesis, California Institute of Technology
.
Lee
,
J. J.
,
Calzone
,
F. J.
,
Britten
,
R. J.
,
Angerer
,
R. C.
and
Davidson
,
E. H.
(
1986
).
Activation of sea urchin actin genes during embryogenesis. Measurement of transcript accumulation from five different genes in Strongylocentrotus purpuratus
.
J. molec. Biol
.
188
,
173
183
.
Levine
,
M.
(
1988
).
Molecular analysis of dorsal-ventral polarity in Drosophila
.
Cell
52
,
785
786
.
Livingston
,
B. T.
and
Wilt
,
F. H.
(
1989
).
Lithium evokes expression of vegetal-specific molecules in the animal blastomeres of sea urchin embryos
.
Proc. natn. Acad. Sci. U.S.A
.
86
,
3669
3673
.
Lynn
,
D. A.
,
Angerer
,
L. M.
,
Bruskin
,
A. M.
,
Klein
,
W. H.
and
Angerer
,
R. C.
(
1983
).
Localization of a family of messenger-RNAs in a single cell type and its precursors in sea urchin embryos
.
Proc. natn. Acad. Sci. U.S.A
.
80
,
2656
2660
.
Maisey
,
J. G.
(
1986
).
Heads and tails: a chordate phylogeny
.
Cladistics
2
,
201
256
.
Manton
,
S. M.
(
1977
).
The Arthropoda: Habits, Functional Morphology, and Evolution
.
Oxford
:
Clarendon Press
.
Marthy
,
H. J.
(
1973
).
An experimental study of eye development in the cephalopod Loligo vulgaris: Determination and regulation during formation of the primary optic vesicle
.
J. Embryol. exp. Morph
.
29
,
347
361
.
Marthy
,
H. J.
(
1982
).
The cephalopod egg, a suitable material for cell and tissue interaction studies
.
In Embryonic Development, Part B: Cellular Aspects
(ed.
M. M.
Burger
and
R.
Weber
), pp.
223
233
.
New York
:
Alan R. Liss
.
Marthy
,
H. J.
and
Dale
,
B.
(
1989
).
Dye-coupling in the early squid embryo
.
Roux Arch, devl Biol
.
198
,
211
218
.
Martindale
,
M. Q.
(
1986
).
The ontogeny and maintenance of adult symmetry properties in the ctenophore, Mnemiopsis mccradyi
.
Devl Biol
.
118
,
556
576
.
Martinez-Arias
,
A.
,
Baker
,
N. E.
and
Ingham
,
P. W.
(
1988
).
Role of segment polarity genes in the definition and maintenance of cell states in the Drosophila embryo
.
Development
103
,
157
170
.
Mcmahon
,
A. P.
and
Bradley
,
A.
(
1990
).
The Wnt-1 (int-1) proto-oncogene is required for development of a large region of the mouse brain
.
Cell
62
,
1073
1085
.
Mcmahon
,
A. P.
and
Moon
,
R. T.
(
1989
).
Ectopic expression of the proto-oncogene int-1 in Xenopus embryos leads to duplication of the embryonic axis
.
Cell
58
,
1075
1084
.
Meedel
,
T. H.
,
Crowther
,
R. J.
and
Whittaker
,
J. R.
(
1987
).
Determinative properties of muscle lineages in ascidian embryos
.
Development
100
,
245
260
.
Millan
,
F. A.
,
Denhez
,
F.
,
Kondaiah
,
P.
and
Akhurst
,
R. J.
(
1991
).
Embryonic gene expression patterns of TGF /JI, /32 and 33 suggests different developmental functions in vivo
.
Development
111
,
131
143
.
Mitrani
,
E.
,
Ziv
,
T.
,
Thomsen
,
G.
,
Shimoni
,
Y.
,
Melton
,
D. A.
and
Bril
,
A.
(
1990
).
Activin can induce the formation of axial structures and is expressed in the hypoblast of the chick
.
Cell
63
,
495
501
.
Mohun
,
T.
,
Brennan
,
S.
,
Dathan
,
N.
,
Fairman
,
S.
and
Gurdon
,
J.
(
1984
).
Cell type-specific activation of actin genes in the early amphibian embryo
.
Nature
311
,
716
720
.
Morgan
,
T. H.
(
1927
).
Experimental Embryology
.
New York
:
Columbia University Press
..
Murphy
,
P.
,
Davidson
,
D. R.
and
Hill
,
R. E.
(
1989
).
Segmentspecific expression of a homeobox-containing gene in the mouse hindbrain
.
Nature
341
,
156
159
.
Murphy
,
P.
and
Hill
,
R. E.
(
1991
).
Expression of the mouse labial-like homeobox-containing genes, Hox-2.9 and Hox-1.6, during segmentation of the hindbrain
.
Development
111
,
61
74
.
Nakano
,
Y.
,
Guerrero
,
L
,
Hidalgo
,
A.
,
Taylor
,
A. M.
,
Whittle
,
J. R. S.
and
Ingham
,
P. W
, (
1989
).
The Drosophila segment polarity gene patched encodes a protein with multiple potential membrane spanning domains
.
Nature
341
,
508
513
.
Nishida
,
H.
and
Satoh
,
N.
(
1989
).
Determination and regulation in the pigment cell lineage of the ascidian embryo
.
Devi Biol
.
132
,
355
367
.
Nishikata
,
T.
,
Mita-Miyazawa
,
I.
and
Satoh
,
N.
(
1988
).
Differentiation expression in blastomeres of cleavage-arrested embryos of the ascidian Halocynthia roretzi
.
Dev. Growth Differ
.
30
,
371
381
.
Noh
,
S.
,
Nohno
,
T.
,
Koyama
,
E.
,
Muto
,
K.
,
Ohyama
,
K.
,
Aoki
,
Y.
,
Tamura
,
K.
,
Ohsugi
,
K.
,
Ide
,
H.
,
Taniguchi
,
S.
and
Saito
,
T.
(
1991
).
Retinoic acid induces polarizing activity but is unlikely to be a morphogen in the chick limb bud
.
Nature
350
,
83
86
.
Novikoff
,
A. B.
(
1938
).
Embryonic determination in the annelid, Sabellaria vulgaris. II. Transplantation of polar lobes and blastomeres as a test of their inducing capacities
.
Biol. Bull, mar. biol. Lab., Woods Hole
74
,
211
234
.
Nüsslein-Volhard
,
C.
,
Ffrohnhofer
,
H. G.
and
Lehmann
,
R.
(
1987
).
Determination of anteroposterior polarity in Drosophila
.
Science
238
,
1675
1681
.
Okata
,
T. S.
(
1957
).
The pluripotency of the pharyngeal primordium in urodelan neurulae
.
J. Embryol. exp. Morph
.
5
,
438
448
.
Oliver
,
G.
,
Wright
,
G. V. E.
,
Hardwicke
,
J.
and
De Robertis
,
E M.
(
1988a
).
A gradient of homeodomain protein in developing forelimbs of Xenopus and mouse embryos
.
Cell
55
,
1017
1024
.
Oliver
,
G.
,
Wright
,
C. V. E.
,
Hardwicke
,
J.
and
De Robertis
,
E M.
(
1988b
).
Differential antero-posterior expression of two proteins encoded by a homeobox gene in Xenopus and mouse embryos
.
EMBO J
.
7
,
3199
3209
.
Ortolani
,
G.
(
1987
).
The developmental capacity of the animal blastomeres of the 8-cell ascidian embryo
.
Acta Embryol. Morph, exp
.
8
,
353
360
.
Padgett
,
R. W.
,
St Johnston
,
R. D.
and
Gelbart
,
W. M.
(
1987
).
A transcript from a Drosophila pattern gene predicts a protein homologous to the transforming growth factor-beta family
.
Nature
325
,
81
84
.
Panganiban
,
G. E. F.
,
Reuter
,
R.
,
Scott
,
M. P.
and
Hoffman
,
F M.
(
1990
).
A Drosophila growth factor homolog, decapentaplegic, regulates homeotic gene expression within and across germ layers during midgut morphogenesis
.
Development
110
,
1041
1050
.
Parks
,
A. L.
,
Parr
,
B. A.
,
Chin
,
J.-E.
,
Leaf
,
D. S.
and
Raff
,
R. A.
(
1988
).
Molecular analysis of heterochronic changes in the evolution of direct developing sea urchins
.
J. evol. Biol
.
1
,
27
44
.
Patel
,
N. H.
,
Kornberg
,
T. B.
and
Goodman
,
C. S.
(
1989a
).
Expression of engrailed during segmentation in grasshopper and crayfish
.
Development
107
,
201
212
.
Patel
,
N. H.
,
Martin-Blanco
,
E.
,
Coleman
,
K. G.
,
Poole
,
S. J.
,
Ellis
,
M. C.
,
Kornberg
,
T. B.
and
Goodman
,
C. S.
(
1989b
).
Expression of engrailed proteins in arthropods, annelids, and chordates
.
Cell
58
,
955
968
.
Paul
,
C.
(
1990
).
Thereby hangs a tail
.
Nature
348
,
680
681
.
Peifer
,
M.
,
Karch
,
F.
and
Bender
,
W.
(
1987
).
The bithorax complex: control of segmental identity
.
Genes Dev
.
1
,
891
898
.
Pelton
,
R. W.
,
Dickinson
,
M. E.
,
Moses
,
H. L.
and
Hogan
,
B. L M.
(
1990
).
In situ hybridization analysis of TGF/33 RNA expression during mouse development: comparative studies with TGF/31 and J2
.
Development
110
,
609
620
.
Pignoni
,
F.
,
Baldarelli
,
R. M.
,
Steingrimsson
,
E.
,
Diaz
,
R. J.
,
Patapoutian
,
A.
,
Merriam
,
J. R.
and
Lengyel
,
J. A.
(
1990
).
The Drosophila gene tailless is expressed at the embryonic termini and is a member of the steroid receptor superfamily
.
Cell
62
,
151
163
.
Priess
,
J. R.
,
Schnabel
,
H.
and
Schnabel
,
R.
(
1987
).
The glp-1 locus and cellular interactions in early C. elegans embryos
.
Cell
51
,
601
611
.
Priess
,
J. R.
and
Thomson
,
J. N.
(
1987
).
Cellular interactions in early C. elegans embryos
.
Cell
48
,
241
250
.
Raff
,
R. A.
(
1987
).
Constraint, flexibility, and phylogenetic history in the evolution of direct development in sea urchins
.
Devi Biol
.
119
,
6
19
.
Raff
,
R. A.
,
Parr
,
B. A.
,
Parks
,
A. L.
and
Wray
,
G. A.
(
1990
)
Heterochrony and other mechanisms of radical evolutionary change in early development
.
In Evolutionary Innovations
(eds.
M. H.
Nitecki
and
D. V.
Nitecki
), pp.
71
98
.
Chicago
:
University of Chicago Press
.
Reinitz
,
J.
and
Levine
,
M.
(
1990
).
Control of the initiation of homeotic gene expression by the gap genes giant and tailless in Drosophila
.
Devi Biol
.
140
,
57
72
.
Render
,
J. A.
(
1983
).
The second polar lobe of the Sabellaria cementarium embryo plays an inhibitory role in apical tuft formation
.
Wilhelm Roux Arch, devl Biol
.
192
,
120
129
.
Reuter
,
R.
,
Panganiban
,
G. E. F.
,
Hoffman
,
F. M.
and
Scott
,
M. P.
(
1990
).
Homeotic genes regulate the spatial expression of putative growth factors in the visceral mesoderm of Drosophila embryos
.
Development
110
,
1031
1040
.
Reuter
,
R.
and
Scott
,
M. P.
(
1990
).
Expression and function of the homeotic genes Antennapedia and Sex combs reduced in the embryonic midgut of Drosophila
.
Development
109
,
289
303
.
Reverberi
,
G.
(
1971
).
Amphioxus
.
In Experimental Embryology of Marine and Fresh-water Invertebrates
(ed.
G.
Reverberi
), pp.
551
572
.
Amstermam
:
North-Holland Publ. Co
.
Reverberi
,
G.
,
Ortolani
,
G.
and
Ferruzza
,
N. F.
(
1960
).
The causal formation of the brain in the ascidian larva
.
Acta Embryol. Morph, exp
.
3
,
296
336
.
Richards
,
O. W.
and
Davies
,
R. G.
(
1977
).
Imms General Textbook of Entomology
, Tenth Edition, Vol. 2.
London
:
Chapman and Hall
.
Rijsewijk
,
F.
,
Schuermann
,
M.
,
Wagenaar
,
E.
,
Parren
,
P.
,
Weigel
,
D.
and
Nusse
,
R.
(
1987
).
The Drosophila homolog of the mouse mammary oncogene int-1 is identical to the segment polarity gene wingless
.
Cell
50
,
649
657
.
Robert
,
B.
,
Sassoon
,
D.
,
Jacq
,
B.
,
Gehring
,
W.
and
Buckingham
,
M.
(
1989
).
Hox-7, a mouse homeobox gene with a novel pattern of expression during embryogenesis
.
EMBO J
.
8
,
91
100
.
Roth
,
S.
,
Stein
,
D.
and
Nusslein-Volhard
,
C.
(
1989
).
A gradient of nuclear localization of the dorsal protein determines dorsoventral pattern in the Drosophila embryo
.
Cel!
59
,
1189
1202
.
Ruberte
,
E.
,
Dollé
,
P.
,
Krust
,
A.
,
Zelent
,
A.
,
Morriss-Kay
,
G
and
Chambón
,
P.
(
1990
).
Specific spatial and temporal distribution of retinoic acid receptor gamma transcripts during mouse embryogenesis
.
Development
108
,
213
222
.
Ruberte
,
E.
,
Dollé
,
P.
,
Chambón
,
P.
and
Morriss-Kay
,
G.
(
1991
)
Retinoic acid receptors and cellular retinoid binding proteins. II. Their differential pattern of transcription during early morphogenesis in mouse embryos
.
Development
111
,
45
60
.
Ruiz I Altaba
,
A.
and
Jessell
,
T.
(
1991
).
Retinoic acid modifies mesodermal patterning in early Xenopus embryos
.
Genes Dev
.
5
,
175
187
.
Ruiz I Altaba
,
A.
and
Melton
,
D. A.
(
1989a
).
Involvement of the Xenopus homeobox gene Xhox3 in pattern formation along the anterior-posterior axis
.
Cell
57
,
317
326
.
Ruiz I Altaba
,
A.
and
Melton
,
D. A.
(
1989b
).
Interaction between peptide growth factors and homeobox genes in the establishment of antero-posterior polarity in frog embryos
.
Nature
341
,
33
38
.
Ruiz I Altaba
,
A.
and
Melton
,
D. A.
(
1989c
).
Bimodal and graded expression of the Xenopus homeobox gene Xhox3 during embryonic development
.
Development
106
,
173
183
.
Rushlow
,
C. A.
,
Han
,
K.
,
Manley
,
J. L.
and
Levine
,
M.
(
1989
).
The graded distribution of the dorsal morphogen involves selective nuclear transport in Drosophila
.
Cell
59
.
1165
1177
.
Saiga
,
H.
,
Mizokami
,
A.
,
Makabe
,
K. W.
,
Satoh
,
N.
and
Mita
,
T.
(
1991
).
Molecular cloning and expression of a novel homeobox gene AHox-1 of the ascidians, Halocynthia roretzi
.
Development
111
,
821
828
.
St Johnston
,
R. D.
and
Gelbart
,
W. M.
(
1987
).
Decapentaplegic transcripts are localized along the dorsal-ventral axis of the Drosophila embryo
.
EMBO J
.
6
,
2785
2791
.
Sander
,
K.
(
1976
).
Specification of the basic body pattern in insect embryogenesis
.
Adv. Insect Physiol
.
12
,
125
258
.
Sander
,
K.
and
Lehman
,
R.
(
1988
).
Drosophila nurse cells produce a posterior signal required for embryonic segmentation and polarity
.
Nature
335
,
68
70
.
Sargent
,
T. D.
,
Jamrich
,
M.
and
Dawid
,
I. B.
(
1986
).
Cell interactions and the control of gene activity during early development of Xenopus laevis
.
Devl Biol
.
114
,
238
246
.
Schaeffer
,
B.
(
1987
).
Deuterostome monophyly and phylogeny
.
In Evolutionary Biology, Vol. 21
(ed.
M. K.
Hecht
,
B.
Wallace
and
G. T.
Prance
), pp.
179
235
.
New York
:
Plenum Press
.
Schauer
,
I. E.
and
Wood
,
W. B.
(
1990
).
Early C. elegans embryos are transcriptionally active
.
Development
110
,
1303
1317
.
Schmid
,
P.
,
Cox
,
D.
,
Bilbe
,
G.
,
Maier
,
R.
and
Mcmaster
,
G. K.
(
1991
).
Differential expression of TGF /31, ft! and p3 genes during mouse embryogenesis
.
Development
111
,
117
130
.
Scott
,
M. P.
and
Carroll
,
S. B.
(
1987
).
The segmentation and homeotic gene network in early Drosophila development
.
Cell
51
,
689
698
.
Sharpe
,
C. R.
,
Fritz
,
A.
,
De Robertis
,
E. M.
and
Gurdon
,
J. B
(
1987
).
A homeobox-containing marker of posterior neural differentiation shows the importance of predetermination in neural induction
.
Cell
50
,
749
758
.
Sharpe
,
C. R.
and
Gurdon
,
J. B.
(
1990
).
The induction of anterior and posterior neural genes in Xenopus laevis
.
Development
109
,
765
774
.
Shott
,
R. J.
,
Lee
,
J. J.
,
Britten
,
R. J.
and
Davidson
,
E. H.
(
1984
).
Differential expression of the actin gene family of Strongylocentrotus purpuratus
.
Devl Biol
.
101
,
295
306
.
Simeone
,
A.
,
Acampora
,
D.
,
Arcioni
,
L.
,
Andrews
,
P. W.
,
Boncinelli
,
E.
and
Mavillo
,
F.
(
1990
).
Sequential activation of Hox-2 homeobox genes by retinoic acid in human embryonal carcinoma cells
.
Nature
346
,
763
766
.
Sive
,
H. L.
,
Draper
,
B. W.
,
Harland
,
R. M.
and
Weintraub
,
H
(
1990
).
Identification of a retinoic acid-sensitive period during primary axis formation in Xenopus laevis
.
Genes Dev
.
4
,
932
942
.
Small
,
S.
,
Kraut
,
R.
,
Hoey
,
T.
,
Warrior
,
R.
and
Levine
,
M.
(
1991
).
Transcriptional regulation of a pair-rule stripe in Drosophila
.
Genes Dev
.
5
,
827
839
.
Smith
,
J. C.
,
Price
,
B. M. J.
,
Van Nimmen
,
K.
and
Hhuylebroeck
,
D.
(
1990
).
Identification of a potent Xenopus mesoderm-inducing factor as a homologue of activin A
.
Nature
345
,
729
731
.
Smith
,
S. M.
and
Eichele
,
G.
(
1991
).
Temporal and regional differences in the expression pattern of distinct retinoic acid receptor-/? transcripts in the chick embryo
.
Development
111
,
245
252
.
Snape
,
A.
,
Wylie
,
C. C.
,
Smith
,
J. C.
and
Heasman
,
J.
(
1987
).
Changes in states of commitment of single animal pole blastomeres of Xenopus laevis
.
Devl Biol
.
119
,
503
510
.
Sprenger
,
F.
,
Stevens
,
L. M.
and
Nusslein-Volhard
,
C.
(
1989
).
The Drosophila gene torso encodes a putative receptor tyrosine kinase
.
Nature
338
,
478
483
.
Stanojevií
,
D.
,
Hoey
,
T.
and
Levine
,
M.
(
1989
).
Sequencespecific DNA-binding activities of the gap proteins encoded by hunchback and Krilppel in Drosophila
.
Nature
341
,
331
335
.
Stern
,
C. D.
(
1990
).
The marginal zone and its contribution to the hypoblast and primitive streak of the chick embryo
.
Development
109
,
667
682
.
Stern
,
C. D.
and
Canning
,
D. R.
(
1990
).
Origin of cells giving rise to mesoderm and endoderm in chick embryo
.
Nature
343
,
273
275
.
Stern
,
C. D.
,
Fraser
,
S. E.
,
Keynes
,
R. J.
and
Primmett
,
D. R. N.
(
1988
).
A cell lineage analysis of segmentation in the chick embryo
.
Development
104
Supplement
,
231
244
.
Steward
,
R.
(
1989
).
Relocalization of the dorsal protein from the cytoplasm to the nucleus correlates with its function
.
Cell
59
,
1179
1188
.
Strecker
,
T. R.
and
Lipshitz
,
H. D.
(
1990
).
Functions of the Drosophila terminal genes in establishing embryonic pattern
.
In Developmental Biology
(ed.
E. H.
Davidson
,
J. V.
Ruderman
and
J. W.
Posakony
), pp.
85
94
.
New York
:
Wiley-Liss
.
Strecker
,
T. R.
,
Yip
,
M. L. R.
and
Lipshitz
,
H. D.
(
1991
).
Zygotic genes that mediate torso receptor tyrosine kinase functions in the Drosophila melanogaster embryo
.
Proc. natn. Acad. Sci. U.S.A., submitted
.
Strome
,
S.
(
1989
).
Generation of cell diversity during early embryogenesis in the nematode Caenorhabditis elegans
.
Int. Rev. Cytol
.
114
,
81
123
.
Sucov
,
H. M.
,
Hough-Evans
,
B. R.
,
Franks
,
R. R.
,
Britten
,
R. J.
and
Davidson
,
E. H.
(
1988
).
A regulatory domain that directs lineage-specific expression of a skeletal matrix protein gene in the sea urchin embryo
.
Genes Dev
.
2
,
1238
1250
.
Sulston
,
J. E.
,
Schierenberg
,
E.
,
White
,
J. G.
and
Thomson
,
J. N.
(
1983
).
The embryonic cell lineage of the nematode Caenorhabditis elegans
.
Devl Biol
.
100
,
64
119
.
Sundin
,
O. H.
,
Busse
,
H. G.
,
Rodgers
,
M. B.
,
Gudas
,
L. J.
and
Eichele
,
G.
(
1990
).
Region-specific expression in early chick and mouse embryos of Ghox-lab and Hox-1.6, vertebrate homeobox-containing genes relating to Drosophila labial
.
Development
108
,
47
58
.
Sundín
,
O. H.
and
Eichele
,
G.
(
1990
).
A horneo domain protein reveals the metameric nature of the developing chick hindbrain
.
Genes Dev
.
4
,
1267
1276
.
Swalla
,
B. J.
,
Badgett
,
M. R.
and
Jeffery
,
W. R.
(
1991
).
Identification of a cytoskeletal protein localized in the myoplasm of ascidian eggs: localization is modified during anural development
.
Development
111
,
425
436
.
Tear
,
G.
,
Akam
,
M.
and
Martinez-Arias
,
A.
(
1990
).
Isolation of an abdominal-A gene from the locust Schistocerca gregaria and its expression during early embryogenesis
.
Development
110
,
915
925
.
Tear
,
G.
,
Bate
,
C. M.
and
Martinez-Arias
,
A.
(
1988
).
A phylogenetic interpretation of the patterns of gene expression in Drosophila embryos
.
Development
104
Supplement
,
135
145
.
Technau
,
G. M.
,
Becker
,
T.
and
Campos-Ortega
,
J. A.
(
1988
).
Reversible commitment of neural and epidermal progenitor cells during embryogenesis of Drosophila melanogaster
.
Rouxs Arch, devl Biol
.
197
,
413
418
.
Technau
,
G. M.
and
Campos-Ortega
,
J. A.
(
1986
).
Lineage analysis of transplanted individual cells in embryos of Drosophila melanogaster. II. Commitment and proliferative capabilities of neural and epidermal cell progenitors
.
Roux Arch, devl Biol
.
195
,
445
454
.
Thomas
,
K. R.
and
Capecchi
,
M. R.
(
1990
).
Targeted disruption of the murine int-1 proto-oncogene resulting in severe abnormalities in midbrain and cerebellar development
.
Nature
346
,
847
850
.
Thomsen
,
G.
,
Woolf
,
T.
,
Whitman
,
M.
,
Sokol
,
S.
,
Vaughan
,
J.
,
Vale
,
W.
and
Melton
,
D. A.
(
1990
).
Activins are expressed early in Xenopus embryogenesis and can induce axial mesoderm and anterior structures
.
Cell
63
,
485
493
.
Tremml
,
G.
and
Bienz
,
M.
(
1989
).
Homeotic gene expression in the visceral mesoderm of Drosophila embryos
.
EMBO J
.
8
,
2677
2685
.
Tung
,
T. C.
,
Wu
,
S. C.
and
Tung
,
Y. F. Y.
(
1958
).
The development of isolated blastomeres of Amphioxus
.
Scientia Sinica
7
,
1280
1320
.
Tung
,
T. C.
,
Wu
,
S. C.
and
Tung
,
Y. F. Y.
(
1962
).
The presumptive areas of the egg of Amphioxus
.
Scientia Sinica
11
,.
629
644
.
Turner
,
D. L.
,
Snyder
,
E. Y.
and
Cepko
,
C. L.
(
1990
).
Lineageindependent determination of cell type in the embryonic mouse retina
.
Neuron
4
,
833
845
.
Valentine
,
J. W.
(
1989
).
Bilaterians of the Precambrian-Cambrian transition and the annelid-arthropod relationship
.
Proc. natn. Acad. Sci. U.S.A
.
86
,
2272
2275
.
Van Den Eunden-Van Raau
,
A. J. M.
,
Van Zoelent
,
E. J. J.
,
Van Nimmen
,
K.
,
Koster
,
C. H.
,
Snoek
,
G. T.
,
Durston
,
A. J.
and
Huylebroeck
,
D.
(
1990
).
Activin-like factor from a Xenopus laevis cell line responsible for mesoderm induction
.
Nature
345
,
732
734
.
Van Den Heuvel
,
M.
,
Nusse
,
R.
,
Johnston
,
P.
and
Lawrence
,
P. A.
(
1989
).
Distribution of the wingless gene product in Drosophila embryos: a protein involved in cell-cell communication
.
Cell
59
,
739
749
.
Vogels
,.
R.
,
De Graaff
,
W.
and
Deschamps
,
J.
(
1990
).
Expression of the murine homeobox-containing gene Hox-2.3 suggests multiple time-dependent and tissue-specific roles during development
.
Development
110
,
1159
1168
.
Walsh
,
C.
and
Cepko
,
C. L.
(
1990
).
Cell lineage and cell migration in the developing cerebral cortex
.
Experientia
46
,
940
947
.
Wanek
,
N.
,
Gardiner
,
D. M.
,
Muneoka
,
K.
and
Bryant
,
S. V.
(
1991
).
Conversion by retinoic acid of anterior cells into ZPA cells in the chick wing bud
.
Nature
350
,
81
83
.
Warga
,
R. M.
and
Kimmel
,
C. B.
(
1990
).
Cell movements during epiboly and gastrulation in zebrafish
.
Development
108
,
569
580
.
Waring
,
D. A.
and
Kenyon
,
C.
(
1990
).
Selective silencing of cell communication influences anteroposterior pattern formation in C elegans
.
Cell
60
,
123
131
.
Waring
,
D.
and
Kenyon
,
C.
(
1991
).
Regulation of cellular responsiveness to inductive signals in the developing C. elegans nervous system
.
Nature
350
,
712
715
.
Weisblat
,
D. A.
and
Blair
,
S. S.
(
1984
).
Developmental interdeterminacy in embryos of the leech Helobdella triserialis
.
Devl Biol
.
101
,
326
335
.
Weisblat
,
D. A.
,
Kim
,
S. Y.
and
Stent
,
G. S.
(
1984
).
Embryonic origins of cells in the leech Helobdella triserialis
.
Devl Biol
.
104
,
65
85
.
Wessells
,
N. K.
(
1977
).
Tissue Interactions and Development
.
Menlo Park, CA
:
W. A. Benjamin
.
Wetts
,
R.
and
Fraser
,
S. E.
(
1989
).
Slow intermixing of cells during Xenopus embryogenesis contributes to the consistency of the blastomere fate map
.
Development
105
,
9
15
.
Whittaker
,
J. R.
(
1973
).
Segregation during ascidian embryogenesis of egg cytoplasmic information for tissue-specific enzyme development
.
Proc. natn. Acad. Sci. U.S.A
.
70
,
2096
2100
.
Whittaker
,
J. R.
(
1980
).
Acetylcholinesterase development in extra cells caused by changing the distribution of myoplasm in ascidian embryos
.
J. Embryol. exp. Morph.
55
,
343
354
.
Whittaker
,
J. R.
(
1990
).
Determination of alkaline phosphatase expression in endodermal cell lineages of an ascidian embryo
.
Biol. Bull. mar. biol. Lab., Woods Hole
178
,
220
230
.
Willmer
,
P.
(
1990
).
Invertebrate Relationships. Patterns in Animal Evolution
.
Cambridge
:
Cambridge University Press
.
Wilson
,
E. B.
(
1893
).
Amphioxus and the mosaic theory of development
.
J. Morph
.
8
,
579
638
.
Wilson
,
E. B.
(
1903
).
Experiments on cleavage and localization in the nemertine egg
.
Arch. f. Entw
.
16
,
411
460
.
Wilson
,
E. B.
(
1925
).
The Cell in Development and Heredity
, Third Edition.
New York
:
Macmillan
.
Wilt
,
F.
(
1987
).
Determination and morphogenesis in the sea urchin embryo
.
Development
100
,
559
575
.
Wolgemuth
,
D. J.
,
Behringer
,
R. R.
,
Mosteller
,
M. P.
,
Brinster
,
R. L.
and
Palmiter
,
R. D.
(
1989
).
Transgenic mice overexpressing the mouse homeobox-containing gene Hox-1.4 exhibit abnormal gut development
.
Nature
337
,
464
467
.
Wood
,
W. B.
(
1991
).
Evidence for reversal of handedness in C. elegans embryos for early cell interactions determining cell fates
.
Nature
349
,
536
538
.
Wray
,
G. A.
and
Mcclay
,
D. R.
(
1989
).
Molecular heterochronies and heterotopies in early echinoid development
.
Evolution
43
,
803
813
.
Wray
,
G. A.
and
Raff
,
R. A.
(
1989
).
Evolutionary modification of cell lineage in the direct-developing sea urchin Heliocidaris erythrogramma
.
Devl Biol
.
132
,
458
470
.
Wray
,
G. A.
and
Raff
,
R. A.
(
1990
).
Novel origins of lineage founder cells in the direct-developing sea urchin Heliocidaris erythrogramma
.
Devl Biol
.
141
,
41
54
.
Wright
,
C. V. E.
,
Cho
,
K. W. Y.
,
Hardwicke
,
J.
,
Collins
,
R. H.
and
De Robertis
,
E. M.
(
1989
).
Interference with function of a homeobox gene in Xenopus embryos produces malformations of the anterior spinal cord
.
Cell
59
,
81
93
.
Wright
,
C. V. E.
,
Morita
,
E. A.
,
Wilkin
,
D. J.
and
De Robertis
,
E. M.
(
1990
).
The Xenopus XlHbox-6 horneo protein, a marker of posterior neural induction, is expressed in proliferating neurons
.
Development
109
,
225
234
.
Wright
,
C. V. E.
,
Schnegelsberg
,
P.
and
De Robertis
,
E. M.
(
1988
).
XlHbox-8, a novel Xenopus horneo protein restricted to a narrow band of endoderm
.
Development
104
,
787
794
.
Wylie
,
C. C.
,
Snape
,
A.
,
Heasman
,
J.
and
Smith
,
J. C.
(
1987
).
Vegetal pole cells and commitment to form endoderm in Xenopus laevis
.
Devl Biol
.
119
,
496
502
.
Yuge
,
M.
,
Kobayakawa
,
Y.
,
Fujisue
,
M.
and
Yamana
,
K.
(
1990
)
A cytoplasmic determinant for dorsal axis formation in early embryo of Xenopus laevis
.
Development
110
,
1051
1056
.
Zimmer
,
R. L.
(
1964
).
Reproductive biology and development of Phoronida
.
Ann Arbor, Michigan
:
University Microfilm
.