Ventral

dorsal pattern

wnt7a

Imx1 Imx1/en1 RCAS

type

Imx1 / wnt7a RCAS

Imx1 RCAS / Infected

FIGURE 4.11 A typical relation: transcriptional control of a gene encoding a signal ligand determines the dorsal transcriptional domain in chick limb bud. (A) Diagram of limb bud and early regulatory interactions required for D/V specification. The engrailed (enl) gene is expressed in ventral ectoderm; a Lim class homeodomain gene (Imxl) is expressed in dorsal mesoderm; and the wnt7a gene is expressed in dorsal ectoderm. [(A) From Johnson and Tabin (1997) Cell 90, 979-990;

(Continues)

wnt7a

wnt7a

wnt7a/en1 RCAS

Imx1 Imx1/en1 RCAS

type

Imx1 RCAS / Infected

FIGURE 4.11 A typical relation: transcriptional control of a gene encoding a signal ligand determines the dorsal transcriptional domain in chick limb bud. (A) Diagram of limb bud and early regulatory interactions required for D/V specification. The engrailed (enl) gene is expressed in ventral ectoderm; a Lim class homeodomain gene (Imxl) is expressed in dorsal mesoderm; and the wnt7a gene is expressed in dorsal ectoderm. [(A) From Johnson and Tabin (1997) Cell 90, 979-990;

(Continues)

dorsal pattern

Lmx1 lmx1

1 /en 1 RCAS

Imx1 / wnt7a RCAS

state upstream of the dorsal program of morphogenesis, and the example shows how its spatial localization is controlled by repression of the gene encoding the Wnt7a signaling ligand.

We have now briefly explored some of the forms of regulatory design that underlie bilaterian pattern formation, and the morphogenesis of body parts. There is not nearly complete enough knowledge of any patterning system, and so we have been forced to extract elements of mechanism where they can be found, from a variety of different developmental processes. But there is an advantage. The complete generality of the mechanisms underlying the pattern formation process emerges strongly from these comparisons, disparate as they are. cis-Regulatory modularity, the requirement for spatial cis-regulatory information processing, the depth of the regulatory networks that control first the definition and then the subdivision of future body parts and subparts, and the intimate sequential linkage between installation of regional transcriptional states and signaling events: these are all entirely general features.

THE LAST ROUTINES: CALLING IN DIFFERENTIATION PROGRAMS

There remains the culmination of the process of building body parts. After the successive transcriptional patterns have been generated, and the right morphogen-etic processes set in train in the right places, differentiated effector cells must be produced. Regional specification continues until the final fine-scale pattern is achieved and the transcriptional mechanisms which determine the spacing of copyright Cell Press.] (B) Experimental demonstrations supporting genetic interactions indicated in (A); visualization of gene products in sections (BI-B6) or whole mounts (B7, B8) by in situ hybridization. (BI) Normal ventral expression of enl gene in hindlimb of stage 21 chick embryo; (B2) Transcription of enl gene extended to dorsal ectoderm by introduction of a retroviral vector expressing the mouse enl gene (en/RCAS); stage 10; (B3) normal expression of wnt7a gene, stage 21; (B4) loss of wnt7a expression in hindlimb bud injected with en/RCAS; (B5) Imxl expression in dorsal mesoderm, stage 21; (B6) loss of almost all Imxl expression in limb bud injected with en/RCAS. [(BI-B6) From Logan et al. (1997) Development 124, 2317-2324 and The Company of Biologists Ltd.] (B7) Normal expression of Imxl, stage 24; (B8) ectopic expression of Imxl in limb bud injected with wnt7aRCAS. (B9) Histological cross-sections of normal hindlimb, day II, dorsal up, at two successively more distal levels in foot. In the diagrams green indicates tendons; red, dorsal muscles; pink, ventral muscles; yellow, ventral tendons; light blue, bones. (BIO) Similar sections and diagrams of foot deriving from limb bud that had been injected with ImxlRCAS at stage 8-10. Ventral structures are missing and the morphology approaches mirror image dorsal symmetry. Gray, tendons appearing in an abnormal position. [(B7-BI0) From Riddle et al. (1995) Cell 83, 631-640; copyright Cell Press.]

terminal morphological elements are in place. Terminal differentiation typically follows mitotic expansion of already specified precursors, and this process also has to be locked in to the underlying morphological pattern. From a programmatic point of view a lot of special regulatory functions are required to set in place the right patches, arrays, or masses of differentiated cells, and many or most of the differences that distinguish closely related species are controlled at this level.

A revealing aspect of the processes that control final events of cell type differentiation is how separate they are from what has gone before. For example, return for a moment to Fig. 4.7E: there we saw that a unique combination of patterning functions is required to specify each of veins L2-L5 of the Drosophila wing, but we notice that at the end all the vein territories are identically divided into initial EGFR and flanking Notch territories (Sturtevant et al, 1993; Baonza et al., 2000). That is, the diverse vein patterning processes terminate with the installation of the same differentiation programs. Of course this is a commonplace, true as well of the diverse patterning processes that end by calling in the same differentiation programs for muscle cell types, or bone forming cells, given types of neuron, and so forth, both within and amongst organisms.

Genomic differentiation programs are of a different, because more ancient, evolutionary origin than pattern formation processes (Davidson et al., 1995; Peterson and Davidson, 2000). Bilaterian clades are distinguished by the specific design of the pattern formation processes by which their definitive morphologies are constructed. But many of the differentiation programs they use are completely panbilaterian, and others are shared by the very diverse forms that make up huge clades, such as those required for formation of the endoskeletal and dermal bone of all vertebrates or for the exoskeleton of ecdysozoans. That is, genomic programs for many forms of cell type differentiation were already present in the common ancestors that preceded the morphological diversification of the Bila-teria, and some even preceded the bilaterian-cnidarian split. Differentiation programs have remained as separate subelements of the developmental regulatory network, that can be called into play and discretely "wired" into diverse regional specification systems. To link them into body part development means that their genomic controls must respond to a transcriptional output of the preceding pattern formation process. Two examples follow which illustrate the nature of the interfaces immediately upstream of differentiation gene batteries.

Specification of Peripheral Nervous System Elements in the Drosophila Wing

The wing disc control systems we last looked at determine the position of the proneural clusters. As precise and elegant as these systems are, further regulatory apparatus is still needed in order to refine the position at which SOPs are actually specified. Only thereafter does the terminal differentiation of peripheral nervous system cell types occur. The function of the additional positioning apparatus is to limit proneural fate both without and within the proneural clusters of AS-C

expressing cells. For example, the extramacrochaetae (emc) gene provides an additional negative spatial input in the 3rd instar wing disc, the effect of which is to increase the accuracy of the proneural positioning system (Modolell, 1997; Garrell and Modolell, 1990; Van Doren et al., 1992). Another transcriptional regulator which spatially controls AS-C expression is Hairy (Van Doren et al., 1994), which in certain regions acts as a DNA-binding repressor that is required to prevent ectopic expression. Within the clusters proneural fate is ultimately confined to the cells that will actually become SOPs. In the case of the large macro-chaete bristles of the notum the SOP which emerges is one of a patch of very few cells precisely positioned within the field of 20-30 cells that make up the proneural cluster. The SOP accumulates high levels of AS-C gene products, expresses downstream neurogenic genes, and gives rise to a single bristle. In large bristle clusters such as on the wing margin, the positions of the nervous system elements are less precise, except for their characteristic spacing (Simpson, 1990; reviewed by Garcia-Bellido, 1981; Modolell, 1997). SOP cells actively express Delta, the Notch (N) ligand (under direct AS-C control; Kunisch et al., 1994), and in consequence of activation of their N receptors the surrounding cells of the cluster are diverted to epidermal fate (reviewed by Jan and Jan, 1994).

Many of the N effects on gene expression are mediated in the proneural cluster cells by genes of the Enhancer of split [E(spl)] cluster. A direct link between N and its target genes is a transcription factor, Suppressor of Hairless [Su(H)], that forms a transcriptionally active nuclear complex with the intracellular domain of N, which is released upon N activation. Figure 4.12A shows diagrams of the cis-regulatory elements for a number of E(spl) genes (Nellesen et al., 1999). Some of these genes are critical elements of the spatial refinement system of the wing proneural clusters, in that they encode bHLH repressors of the proneural genes (see legend). The key point here is how the activity of the E(spl) genes is focused to the appropriate cells. The answer is to be seen in the target sites of the E(spl) cis-regulatory elements: Fig. 4.12A shows that each of these genes includes sites both for Su(H) and AS-C factors, i.e., these genes will run in prospective proneural cells, when these cells receive N signals. The result of their activation is suppression of neural differentiation potential, and execution of epidermal fate. Some E(spl) genes respond directly to ectopic Su(H) and ^45-Cexpression in vivo (Cooper et al., 2000) and in Fig. 4.12B two of the genes, m4 and my, are shown to increase their expression dramatically on stimulation of N signaling under heat shock control (Nellesen et al., 1999). But note that the expression still occurs in proneural clusters, in the case of my in clusters in which it is not normally expressed.

Once specified, the SOP undergoes the stereotypic divisions shown in Fig. 4.12C (Posakony, 1994; Gho et al., 1999). The four cells of the bristle assemblage, i.e., shaft, sheath, socket, and neuron, express four distinct differentiated fates, and at each division the asymmetry in fate depends on N signaling (see legend). We are now finally at the level at which cell type-specific differentiation processes must be called into play. A transcriptional regulatory factor which executes that function is Pax2. The pax2 gene is expressed initially in all the SOP

Was this article helpful?

0 0

Post a comment