The Secret of the Bilaterians Abstract Regulatory Design in Building Adult Body Parts

Law Of Attraction For Kids

Smart Parenting Guide

Get Instant Access

The Evolutionary Significance of "Pattern Formation" 103 The First Step: Transcriptional Definition of the Domain of the Body Part 105

Morphological Pieces and Regulatory Subpatterns 110

Heart Parts 110

Forelimb and Hindlimb Buds 111

Transcriptional Domains in the Gut Endoderm 115

Patterns in the Developing Hindbrain 117 Appendage Parts and Transcriptional Patterns in Drosophila

Imaginal Discs 121

Glimpses of How It Works 122 Transcriptional Domains and the Pattern Program for the

Drosophila Wing Disc: Modularity and as-Regulatory Inputs 123

Patterning the Heart Progenitor Field in Drosophila 129

Encoding Hindbrain Regulatory Patterns 131

The Role of Signaling 140

The Last Routines: Calling in Differentiation Programs 146 Specification of Peripheral Nervous System Elements in the

Drosophila Wing 147 Installation of Cell Type-specific Differentiation Programs in the

Pituitary 152

Concluding Remark 153


The bilaterians share principles of developmental regulatory design that are uncannily similar in the most diverse examples, for instance flies and ourselves.

Yet at the DNA level these designs also hold the specific solutions to the general question with which this book begins: what accounts causally for morphological diversity in the bilaterians? The networks of gene regulatory interactions that direct the developmental construction of body parts in bilaterian development vary enormously in detail and outcome, but nonetheless they have common organizational properties; or one might say, they have a common sort of architecture. This is now to be our subject. From either an evolutionist's or a developmental biologist's point of view there may be no more generally important subject. For in their developmental regulatory architecture lies the real "secret of the bilaterians." That is so in two senses, one developmental and one evolutionary. The regulatory architecture explains how bilaterian genomes encode their body parts; but it follows also that the evolutionary appearance of the bilaterians had to have depended on the genomic assembly of such regulatory architecture in their common ancestors (a theory of bilaterian origins based in part on this concept is discussed elsewhere: Davidson et al., 1995; Peterson et al., 2000b; Peterson and Davidson, 2000). When conditions permitted, bilaterian diversification followed, powered not only by the capacities of their large complement of genes or gene families (Chapter 1), but by a shared heritage of regulatory "know how," that enables these genes to be deployed in an endless variety of deep regulatory formats.

"Deep" is a useful word here, because it connotes the stepwise and sequential nature of the necessary regulatory transactions, resulting in an architecture of multiple layers of control. At each level the functional object is to specify sets of transient spatial domains, or regions, and to control their growth, until all the subparts of an element of the body plan have been laid out. Within these subparts the batteries of differentiation genes that endow each with its tissue functions are finally activated. This is the process of "pattern formation" touched on in Chapter 1. The key mechanism in regional specification is the installation of new transcriptional regulatory states within each element of space. As indicated in Chapter 1, a transcriptional regulatory state denotes a region of a developing system in the cells of which given transcription factors are present and active. In the following we take up where this initial description leaves off: among other things we need to know what these transcriptional states accomplish; the nature of the genomic regulatory wiring needed to set them up; and how they are linked into a progressive series.

The outlines of the regulatory mechanism for body part formation seem clear, or are at least rapidly clarifying, but with respect to any particular body part, e.g., appendages or heads or glands, we can as yet only see through the glass darkly. Often we know relatively much about genomic mechanisms for the early stages, but almost nothing about the final steps of cell type diversification, and sometimes it is just the reverse. So the treatment that follows is synthetic. By use of many examples, drawn either from amniotes or from flies, we can see the elements of mechanism, the common principles of the bilaterian way of regulating adult body plan formation. Through these examples the elegance, the richness, and the variety of these mechanisms will hopefully be appreciated. But this is not the place to learn, from A to Z, how any particular piece of any particular animal develops: the focus of this chapter, like the object of its focus, is basically abstract. The mechanisms we want understood are abstract in that they utilize complex programs of genetic interactions to specify spatial domains, one after another, which do not in themselves resemble or function like parts of the structure that will emerge, until the final stages. To the observer taking a snapshot in developmental time, these domains appear as patterns, abstract patterns of regulatory gene expression.


Consider body parts from an evolutionary perspective. The Cambrian lobopods lacked appendages which had the pattern elements of modern arthropod legs; the modern cephalochordates lack the organ level glands, and have at best only the simplest rudiments of the anterior brain organization of vertebrates (see Lacalli, 1996); modern agnathans and fossil conodonts lack the paired limbs of the jawed vertebrates, and so forth. Specific genomic regulatory apparatus is of course required for developmental formulation of every individual body part, and in each case the genomic apparatus must have been assembled stepwise during evolution, building upon the apparatus required to generate the relevant region of the body plan of the latest ancestor lacking that body part. Whatever its stage of evolutionary elaboration, a given body part is produced by a regulatory system that has a unit quality: it is a subprogram that starts with a particular developmental regulatory step, and proceeds from there. The following is about this first step.

The first regulatory step in body part formation is so general that it can be considered a law of bilaterian developmental biology. It is to set up a discrete regulatory state by expression of a gene or genes encoding transcription factors, specifically within a growing field of cells from which the body part will derive. A convenient term for this is the "progenitor field" for that part (Davidson, 1993). The os-regulatory system activating each such gene in the progenitor field must read and integrate spatial inputs that define this field in terms of the preexisting regulatory coordinates of the body plan (here preexisting is used in the developmental sense, but it may often be true in an evolutionary sense as well). These coordinates specify position with respect to the midline, the anterior/posterior axis, the dorsal/ventral axis, the right/left location, or whatever. The following steps, by which the growing progenitor field is divided into subregions that will form the subparts of the future organ or structure, depend mechanistically on this first step. The observable earmarks of the first-step process are the appearance at a certain point in development of a pattern of regulatory gene expression such as can be detected by in situ hybridization or immunocytology, and which conforms to the cellular progenitor field for the body part; the conservation of this pattern within the clade of animals for which this body part is a specific shared character (i.e., the pattern is a synapomorphy); failure of the body part to develop if expression of the first-step regulatory genes can be knocked out or altered; and sometimes ectopic formation of elements of the body part if these genes are expressed ectopically, at least within those regions of the body plan wherein the additional necessary inputs are to be found. Of course this does not mean that expression of these first-step regulatory genes is both necessary and sufficient to make the whole body part; no one gene makes a structure (a strangely simple-minded view which for some reason has died hard). What it does mean is that the first step is necessary and sometimes sufficient to entrain all the subsequent processes, which do control the developmental assembly of the body part.

As an initial illustration of generality with respect to body part, Fig. 4.1 includes localized patterns of transcription of regulatory genes for four different organs of the mouse. In each case these patterns are evident in advance of morphogenesis of these structures. For example, the first known step in the transcriptional specification of the progenitor cells for the heart is activation of the nkx2.5 gene (Fig. 4.1 A, B; Harvey, 1996). There are several closely related members of the same gene family which have overlapping functions (reviewed by Evans, 1999), and orthologous genes are expressed at the beginning of heart development in the heart forming regions of all vertebrate classes (Harvey, 1996). As discussed below, nkx2.5 expression is known to lie upstream of other regulatory genes that execute subsequent steps in patterning the heart, and much evidence exists for the role of this gene and its close relatives in the initial specification of the heart progenitor field (for reviews, Lin et al., 1997; Evans, 1999; cf. Tanaka etal., 1999). For instance, the introduction into Xenopus eggs of trans-acting negative versions of Nkx2.5 and the related Nkx2.3 factors that bear the engrailed repressor domain can result in total loss of heart specification, heart marker gene expression, heart morphogenesis, and cardiocyte differentiation (Fu et al., 1998; Grow and Krieg, 1998). The point is that the pattern illustrated in Fig. 4.1A, B has a purpose: the first-step regulators set in train a sequential genetic machine, which first subdivides and specifies the parts of the heart, and then builds it.

Transcription of regulatory genes in the anterior neural tube of mouse embryos, from which the forebrain forms (Fig. 4.1C), probably plays similar roles. It is not easy to ferret out these roles because single gene knock-out phenotypes are not simply related to the early gene expression domains illustrated (e.g., for the emx and otx genes, see Boncinelli, 1999; Mallamaci et al., 2000). Interpretation is complicated by the existence of multiple gene family members with partially overlapping functions; by prior embryonic lethality when the gene is knocked out; and more generally, by the density and parallel wiring of the regulatory systems that in mammals typically surround such genes. For example, activation of emxl and emx2 in the domains indicated in Fig. 4.1C4 requires prior expression in the dorsal neural tube of a gene encoding another transcription factor, gli3; but expression of neither otxl, otx2 (Fig. 4.1C4), nkx2.1 (Fig. 4.1C1), bfl (Fig. 4.1C2), nor shh (Fig. 4.1C3) requires gli3 (Theil et al., 1999). These regulators are evidently "wired" in parallel and are initially activated by factors other than gli3-A powerful index of the importance of the patterns of gene expression in brain progenitor fields such as shown in Fig. 4.1C, though of a different sort than a gene knock-out phenotype, is their very remarkable conservation: throughout the vertebrates similar relative spatial patterns of genes encoding certain transcription factors occur early in brain development (e.g., for forebrain see Bally-Cuif and Boncinelli, 1997; Smith Fernandez et al., 1998; for the midbrain-hindbrain region Joyner, 1996; and for hindbrain see below).

The role of the pax6 gene in the development of the eye is taken up in the next chapter, but the first stage of that process illustrates the present point too well to omit here (Fig. 4.ID). Bilateral expression of pax6 in the neuroepithelium of the amniote head region long before there is an eye, or any discrete structure relating thereto, is required for development of eyes (pax6 is also utilized in development of brain regions and nasal placodes). In a pax6mouse mutant (sey, small eye) the earliest morphological manifestations of the progenitor field for the eye lens, the lens placode, fails to form, as shown in Fig. 4.1D2, 3); (Grindley et al., 1995).

Figure 4. IE is included here because it provides a beautiful visual display of the kind of process by which the initial transcriptional regulatory states for body parts are established. They are located in the body plan by signaling. The spatial specification of these regulatory states demands integrative processing of signal inputs within individual cz's-regulatory elements of genes that encode transcription factors (cf. Chapter 2). In the case considered in Fig. 4.IE, the body parts that will arise are the wings, halteres, and legs of Drosophila, which derive from the thoracic imaginal discs; the observation in the figure concerns the origins of these discs. The dorsal discs which give rise to wings and halteres and the ventral discs from which the legs emerge have a common origin in three patches of cells on each side in the postgastrular thoracic epithelium (reviewed by Cohen, 1993). These imaginal progenitor cells are marked by expression of the distal-less (dll) gene, or of a lacz gene insertion in the enhancer that directs dll expression in the early (5 h) postgastrular embryo. The expression of the dll gene is required for the developmental outcome of the ventral discs of the thorax, i.e., for formation of the legs (and also for the antennal discs of the head). In the absence of dll parts of the leg extending out from the body fail to form, while ectopic dll expression induces ectopic leg structures (Cohen andjurgens, 1989; Gorfinkiel et al., 1997). Figure 4.IE shows that the imaginal primordia are located precisely at the intersections between the parasegmental D/V stripes of wg gene expression and the bilateral longitudinal stripes of dpp gene expression. Expression of the dll Lacz marker is shown here superimposed on the ladder-like pattern of expression formed by these two genes, both of which encode signaling ligands. The pattern shown in Fig. 4.IE is causal as well as pretty, Wg acting as inducer of dll, and Dpp limiting its expression dorsally (Cohen et al., 1993; Goto and Hayashi, 1997).

In summary, Fig. 4.1 shows us that development of very different body parts begins in just the same way, with the establishment of a patch of progenitor cells expressing certain genes encoding transcription factors. For each case we know

FIGURE 4.1 Initial stages of transcriptional progenitor field specification. (A), (B) Expression of nkx2.5 gene marks the cardiac crescent and heart progenitor field, as visualized by in situ hybridization. This gene encodes a home-odomain transcription factor. (A) Mouse: (AI) 7.5 dpc, anterior aspect; (A2) 8.75 dpc, ventral view. (B) For comparison, expression in chick: (Bl) stage 6, (B2) stage 10, ventral views. [(A, B) From Harvey (1996) Dev. Biol. 178, 203-216.] (C) Expression of various transcription factors visualized by in situ hybridization in anterior neural tube of mouse. The expression domains demarcate future regions of the brain. (CI) nkx2.l expression in central region of developing forebrain, dorsoanterior view (he, heart; f, forebrain; m, midbrain). (C2) Expression of gene encoding a forkhead transcription factor, Bfl, at anterior end of neural plate; anterior view. (C3, C4) Two diagrammatic views of early gene expression domains: (C3) Expression of bfl, otxl, nkx2.2, shh, and emx2 genes in anterior neural tube at 7-somite stage, dorsal view, (oc, optic cup; pr, prosencephalon; me, mesencephalon; rh, rhombencephalon; ap, alar plate; bp, basal plate; fp, floor plate; pos, postotic sulcus.) [(CI-C3) From Shimamura et al. (1995) Development 121, 3923-3933 and The Company of Biologists Ltd.] (C4) Mouse embryo, 10 dpc, showing domains of expression of otxl, otx2, emx2, and emxl genes; telencephalon (Te), diencephalon (Di), mesencephalon (Mes). These genes are not expressed in the midbrain or hindbrain, i.e., metencephalon (Met), and myelencephalon (My). The hox genes are expressed in the hindbrain and more caudally. The domains of expression of the four genes indicated appear to be nested within one another such that otx2 is expressed in the most broad pattern and emxl in the most confined pattern. [(C4) From Simeone et al. (1992) Nature 258, 687-690, copyright Macmillan Magazines Ltd.] (D) Expression of pax6 gene establishes the eye progenitor field in amniotes. (DI) Expression of pax6 gene in chicken embryo, viewed in section after whole mount in situ hybridization. The embryo is at 2-somite stage (stage 7); the domain of expression shown is lateral to the V-shaped neural plate at the forebrain level, and includes the progenitors of eye lens and corneal epithelium. [(DI) From Li et al. (1994) Dev. Biol. 162, 181-194.] (D2, D3) Consequences of mutation in pax6 gene (sey, small eye mutation) in mouse: (D2) normal histology of early optic vesicle, in 7.5 dpc embryo. The lens placode (Ip) arises as a local thickening of the surface epithelium (se). (D3) In sey homozygotes the lens placode fails to form. [(D2, D3) From Grindley et al. (1995) Development 121, 1433-1442 and The Company of Biologists Ltd.] (E) Location of thoracic imaginal disc progenitors in Drosophila, with respect to intersection of stripes of wg and dpp expressing cells. The embryo has been dissected and spread out, anterior left, ventral surface toward viewer; (tl), first thoracic segment. Longitudinal stripes of dpp expression, and metameric stripes of wg expression, both visualized by in situ hybridization, form a ladder-like pattern. This is shown together with six Lacz expression domains from a transgene driven by the early expression c/s-regulatory module of the distal-less gene (dark blue). The cells expressing the transgene indicate the common progenitors of wing and leg imaginal discs [(E) From Cohen et al. (1993) Development 117, 597-608 and The Company of Biologists Ltd.]

that these patterns of expression are an important first step: what happens next depends upon it. But the following steps are much more complicated. We first consider the ensuing process from the outside, i.e., we shall look at examples of more regulatory gene expression patterns, the function of which is to subdivide the initial field into the future subparts of that piece of the body plan. Then in the following section we turn to the underlying regulatory circuitry.


The processes which occur next in the formation of adult body parts reflect the shape of the genomic regulatory networks that control bilaterian morphogenesis. This is our real objective, and though it is just becoming possible to perceive the architecture of such networks, the process itself is of tremendous interest. The initial progenitor field is transformed into a mosaic of regulatory subdomains, and, remarkably, these prefigure the morphological pieces of the body part. Many sequential steps are sometimes required before the process is complete. Setting up the developmental components of the morphology is to be distinguished from setting up particular differentiation programs, for the need for differentiation programs is frequently overlapping among the diverse pieces of which the body part is composed: cartilage, osteoclasts, and the various skeletal muscle fiber types are required within all the morphologically distinct components of the vertebrate limb; and the cell types of peripheral nervous system elements, veins, and inter-vein regions are generated alike in the diverse pattern elements to which the Drosophila wing disc gives rise. The last chapter dealt with regulatory processes that are aimed at direct cell type specification, but not so these. Furthermore, and most importantly, because we are large animals (i.e., compared to eggs or Type 1 embryos), the spatial components of the forming body part must also function during development as growth control units.

Here is where the term "abstract," as in the title of this chapter, is most needed. Genetic regulatory units hardwired in the genome produce as their output spatial transcriptional domains, which after many further patterning steps produce, as their output, the morphological elements of the body part. Five briefly discussed examples follow, to illustrate these causal, though abstract relations. For each case development of a particular morphological piece of the finished product, the body part, is seen to depend on an identifiable piece of the regulatory apparatus, that operates far upstream of the development of the final form.

Heart Parts

Upon specification of its progenitor field morphogenesis of the heart begins with the formation of a tube. Regions of this tube give rise to the different morphological and functional components of the heart (reviewed by Fishman and Chien,

1997; Christoffels et al, 2000): in amniotes, these are the left and the right ventricles, and the left and right atria, plus the inflow and outflow tracts. Specification of the ventricular domains provides a particular case of the general process by which the subparts of an organ begin their development. The forming ventricles express two genes encoding bHLH transcription factors, the dhand and ehand genes (Srivastava et al., 1997; Biben and Harvey, 1997), the expression patterns for which are shown in Fig. 4.2A and B. These patterns are quite remarkable: while the dhand gene is expressed strongly in the future right ventricle and more weakly in the future left ventricle (Fig. 4.2A2), ehand is expressed not at all in the future right ventricle but strongly in the future left ventricle (Fig. 4.2B2). If the dhand gene is knocked out, ventricular morphogenesis fails in the developing heart (Fig. 4.2C). The phenotype of the knockout includes failure to mobilize neural crest components that produce the aortic arches (in which the dhand gene, is also expressed; Cserjesi et al., 1995). The hand genes are downstream of the nkx2.5 gene (Biben and Harvey, 1997; Srivastava et al., 1997), and regional dhand expression also requires the mef2c gene, which together with the nkx2.5 gene is expressed throughout the early heart tube and is required for its further morphogenesis (Lin et al., 1997). Activation of dhand therefore depends on initial specification of the heart progenitor field. But obviously regulation of the dhand and ehand genes depends as well on additional spatial cues in the early embryo.

The dhand gene in turn controls other genes needed for ventricular development. Among these are gata4 (Srivastava et al., 1997) which is essential for further heart morphogenesis (Molkentin et al., 1997). Another regulatory gene, irx4, encodes a homeodomain factor which in turn is needed for expression of the genes which encode the ventricle-specific myosin heavy chain isoform; but ventricular expression of irx4 is not maintained in the absence of dhand expression (Bao et al., 1999; Bruneau et al., 2000). We can anticipate a czs-regulatory network that links the initial specification of the heart progenitor field (Fig. 4.1A, B) to the regionalization functions which specify the parts of the heart, and thence to the morphogenetic and functional development of these parts. A prominent aspect of the format of this network is that the genes which define each successive spatial and morphogenetic domain continue to be used to provide inputs for the following regulatory phase, together with additional new, regionally confined inputs. In this way the spatial achievement of each regulatory stage is integrated together with new cues so that specification is progressive. Some detailed examples of this principle appear in the following section.

Forelimb and Hindlimb Buds

The two bilateral pairs of limb buds arise in the lateral plate mesoderm of the amniote embryo. Their position is thought to depend on the prior A/P expression domains of box genes (e.g., Gibson-Brown et al., 1998; Cohn et al., 1997), though the molecular linkages by which this is accomplished are unknown. At the beginning of limb bud outgrowth hoxd9 is expressed in both leg and wing buds in

FIGURE 4.2 Transcriptional specification domains for the ventricular components of the heart in the mouse. (A, B) Expression of dhand and ehand genes visualized by in situ hybridization. (A) Expression of dhand gene; (A I) 8.0 dpc; (A2) 9.5 dpc. (B) Expression of ehand gene: (BI) and (B2) same stages as (A I) and (A2). Abbreviations: ht, heart; hf, head fold; ba, branchial arch; Iv, left ventricle; rv, proper ventricular, Im, lateral plate mesoderm; ct, conotrunchus. (C) dhand is required for right ventricle formation: (CI) Normal morphology, transverse section 9.5 dpc. (C2) morphology in embryo of same age in which the dhand gene has been knocked out: a, v indicate regions where atria and ventricles would have formed. [(C) Adapted from Srivastava et al. (1997) Nature Genetics 16, 154-160,410; copyright Macmillan Magazines Ltd.]

FIGURE 4.2 Transcriptional specification domains for the ventricular components of the heart in the mouse. (A, B) Expression of dhand and ehand genes visualized by in situ hybridization. (A) Expression of dhand gene; (A I) 8.0 dpc; (A2) 9.5 dpc. (B) Expression of ehand gene: (BI) and (B2) same stages as (A I) and (A2). Abbreviations: ht, heart; hf, head fold; ba, branchial arch; Iv, left ventricle; rv, proper ventricular, Im, lateral plate mesoderm; ct, conotrunchus. (C) dhand is required for right ventricle formation: (CI) Normal morphology, transverse section 9.5 dpc. (C2) morphology in embryo of same age in which the dhand gene has been knocked out: a, v indicate regions where atria and ventricles would have formed. [(C) Adapted from Srivastava et al. (1997) Nature Genetics 16, 154-160,410; copyright Macmillan Magazines Ltd.]

the chick (Nelson et al., 1996; Cohn et al., 1997), but not in the intervening lateral plate mesoderm. Application of beads which emit certain FGFs to this intervening region suffices to induce ectopic limb buds and limbs (Isaac et al., 2000; whether or how these FGF family members function in the normal process of limb bud specification is not clear). If the beads are implanted in the anterior portion of this region, wing-like structures are formed, and if in the posterior region legs are formed (Cohn et al., 1995). The extant positional inputs thus superimpose fore-limb or hindlimb identity on the ectopic limb outgrowth. The process by which "identity" is established with respect to the future morphogenetic character of a body part is now familiar: it requires institution of a specific transcriptional regulatory domain, as in all the cases considered in this chapter. Some of the regulatory players responsible for forelimb vs. hindlimb fate are shown in Fig. 4.3. The pitxl gene and the tbx4 gene are expressed specifically in the hindlimb (Fig. 4.3A, C) in chicken (Logan et al., 1998) and mouse (Szeto et al., 1996); and the tbx5gene is expressed specifically in the forelimb (Fig. 4.3B; ibid.). The tbx4 gene responds to expression of the pitxl gene, though the latter is activated in a broader area than the limb bud alone, while tbx4 transcripts are here confined to the limb bud: pitxl is activated earlier than is tbx4 in either normal or ectopi-cally induced limbs; and if pitxl expression is induced in the wing bud by introduction of a retroviral expression vector, the consequence is to activate tbx4 expression there (Fig. 4.3D; Takeuchi et al., 1999; Logan and Tabin, 1999). Furthermore, tbx4 expression is exclusive of tbx5 expression. A limb bud induced with an FGF bead in the anterior region of the flank expresses tbx5 (Fig. 4.3E1) and as indicated above it produces an ectopic wing, illustrated in Fig. 4.3E3. But if expression of tbx4 is forced within the ectopic limb by introduction of a retroviral expression vector, tbx5 expression is turned off (Fig. 4.3E2), and the resulting ectopic limb now often displays morphological features of the hindlimb; e.g., it forms a claw (Fig. 4.3E4; Rodriguez-Esteban et al., 1999; Takeuchi et al., 1999).

Following the specification of forelimb or hindlimb identity for which the tbx genes are necessary and sufficient, other genes initiate the morphogenesis of the limb. Among those are sets of box cluster genes, different members of which are expressed in specific patterns in the developing wing and leg (Nelson et al., 1996; see Chapter 5). The boxd9 gene is upregulated in the wing bud and hoxc9 down-regulated, while the opposite occurs in the hindlimb bud. Introduction of retroviral vectors encoding tbx4 into wing bud causes the leg bud pattern of hoxc9 and hoxd9 expression; and introduction of a tbx5 vector in the leg bud causes installation of the wing bud pattern of activity of the box genes (Takeuchi et al., 1999).

All these results show that the tbx genes lie rather far upstream in the regulatory patterning network that establishes the limbs. Their regional activation follows initial specification of the limb buds in respect to the A/P axis of the embryo, while they themselves provide inputs that directly and/or indirectly set up the forelimband hindlimb-specific patterns of box gene expression, and no doubt as well activate (or repress) many other target genes (in addition to affecting one another). But this is only an early stage of the abstract process of regional

ectopic wing

specification in the limb, which has many steps and stages and which is not yet very well known. Prominent elements of the control apparatus are the regulatory circuits that establish the A/P, D/V, and proximo-distal domains from which the different subparts of the limb will develop, and which later direct the actual morphogenesis of the wing.

Transcriptional Domains in the Gut Endoderm

In the two examples so far considered (i.e., in Figs. 4.2 and 4.3) we have looked at patterning within the progenitor fields for particular structures. But these kinds of mechanisms also apply to the overall body plan, and for this reason Fig. 4.4 is included. It serves as a reminder of the scope of the phenomenon we are discussing, before proceeding with better known and more detailed examples. Figure 4.4A shows in diagrammatic form the whole gut tube of a chick embryo: on the left is shown a series of transcriptional regulatory domains that have been imposed within the gut endoderm at the 15 somite stage; and in the center the organ that will form from each such domain (Grapin-Botton and Melton, 2000). The diagram shows that each organ arises from a region distinguished by a unique regulatory state at this stage. These patterns appear to be established by signaling from the surrounding mesoderm and ectoderm (Wells and Melton, 2000). Two examples of these patterns are also shown, vizpdxl expression in the dorsal and ventral pancreatic buds (Fig. 4.4B); and ┬┐ex expression in the endoderm that will

FIGURE 4.3 Early transcriptional specification of hindlimb and forelimb identity in chick. Domains of expression are visualized by in situ hybridization. (A) Expression of pitxl homeodomain regulatory gene in mouse hindlimb bud (this gene is also expressed in branchial arches and in the pituitary anlage). (B) Expression of tbxJ gene in forelimb (wing) of chick; and (C) expression of tbx4 in hindlimb. [(A-C) From Logan et al. (1998) Development 125, 2825-2835 and The Company of Biologists Ltd.] (D) Demonstration that pitxl expression is upstream of tbx4 expression: pitxl expression was forced in the wing bud by introduction of a retroviral expression vector, and tbx4 expression can now be detected by in situ hybridization; (DI) control, (D2) ectopic pitxl wing bud, in which tbx4 is expressed (red arrow). [(D) From Logan and Tabin (1999) Science 283, 1736-1739; copyright American Association for the Advancement of Science.] (E) Fate of induced limb bud altered by forced tbx4 gene expression. (El) Induction of ectopic limb bud (arrow) in chick by implantation of bead emitting Fgf2 in lateral plate mesoderm at the level of somite 21; the induced ectopic bud expresses the tbx5 gene. (E2) Expression of tbx5 in induced ectopic limb bud (arrow) is nearly extinguished on introduction of a retroviral vector (RCAS) expressing tbx4. (E3) Limb bud induced as in (El) produces an ectopic limb that displays skeletal character of wing (70% of cases; arrow). (E4) Induced limb bud in which tbx4 expression is forced instead produces mainly ectopic legs (56% of cases; arrow). [(E) From Rodriguez-Esteban et al. (1999) Nature 398, 814-818; copyright Macmillan Magazines Ltd.]

give rise to the liver (Fig. 4AC). A pdxl knockout shows that this gene is indeed essential for proliferation and differentiation of pancreatic cell types within the bud (Offield et al., 1996). We can be sure that Fig. 4.4A illustrates only a fraction of the regulatory transactions required to specify downstream morphogenetic fates in the endoderm. Even so, it shows that spatially confined transcriptional patterning is utilized throughout the whole extent of the endodermal germ layer.

Patterns in the Developing Hindbrain

The embryonic hindbrain of vertebrates transiently displays a segmental organization. In molecular and developmental terms the metameric units, or rhombo-meres, are genetic regulatory units. That is, transcriptional regulatory states specific to each rhombomere are established early in hindbrain development, while each rhombomere can be considered the subunit of the hindbrain from which will ultimately derive specific facial and lower head ganglia, as well as specific populations of neural crest (Nieto et al., 1995). The developmental outcome depends on both autonomous and nonautonomous (i.e., signaling) functions of the individual rhombomeres (reviewed by Lumsden and Krumlauf, 1996; Marin and Charnay, 2000). Rhombomere specification affords a clear causal link connecting early gene expression patterns with the later morphological output of the regulatory system.

Regional expression of many different genes has been recorded in the developing hindbrain. The domains of transcription of some genes which encode transcription factors, and others which encode components of signaling systems are shown against the metameric register of rhombomeres (r) 2-6 in Fig. 4.5A.

FIGURE 4.4 Regional expression of transcription factors foreshadows endodermal fate. (A) Summary of gene expression domains (left) mapped onto completed gut tube of chick (center) to illustrate the correlation between early transcriptional domains in the 15-somite embryo (right) and regional organogenesis in the endoderm (center). Pink triangles indicate hex gene expression in thyroid (top left) and in liver bud (center left); blue triangle (center left) indicates expression of pdxl gene in pancreatic bud; BA, positions of branchial arches. [(A) From Grapin-Botton and Melton (2000) Trends Genet 16, 124-130.] For additional hox cluster expression patterns in the endoderm see Beck et al. (2000). (B) Example of pattern shown in (A): expression of pdxl in pancreatic buds in mouse, visualized by immunocytology. Top domain is dorsal pancreatic epithelium, lower is ventral. [(B) From Ahlgren et al. (1996) Development 122, 1409-1416 and The Company of Biologists Ltd.] (C) Expression of hex gene in endoderm that will give rise to liver bud (L) adjacent to heart epithelium in a stage 12 (18-somite) chick embryo, visualized by in situ hybridization. (CI) Whole mount; (C2) sagittal section showing that expression is confined to endoderm (end); myo, myocardium; som, somite. [(C) From Yatskievych et al. (1999) Mech. Dev. 80, 107-109; copyright Elsevier Science.]

Was this article helpful?

0 0
Confident Kids

Confident Kids

Although nobody gets a parenting manual or bible in the delivery room, it is our duty as parents to try to make our kids as well rounded, happy and confident as possible. It is a lot easier to bring up great kids than it is to try and fix problems caused by bad parenting, when our kids have become adults. Our children are all individuals - they are not our property but people in their own right.

Get My Free Ebook

Post a comment