The Regulatory Demands of Development

Development is the execution of the genetic program for construction of a given species of organism. The program itself is our concern, and those aspects of its execution that regulate its own readout from the DNA. The most cursory consideration of the developmental process produces the realization that the program must have remarkable capacities, for development imposes extreme regulatory demands. Of the thousands of genes indicated in the diagram of Fig. 1.2, most are utilized at some time during development, and all must be controlled accurately in space and time. But this is really the least of the problem: the essence of development is the progressive increase in complexity which it invariably entails. In informational terms the increase in developmental complexity is measured by the generation of new populations of cells, each of which reads out a defined genetic subprogram; and each of which arises in a particular spatial domain of the embryo. All the while these populations are being instructed to expand to given extents, by cell growth (and sometimes to contract, by cell death). Thus the spatial components of morphology are laid out.

Metaphors often have undesirable lives of their own, but a useful one here is to consider the regulatory demands of building a large and complex edifice, the way this is done by modern construction firms. All of the structural characters of the edifice, from its overall form to minute aspects that determine its local functionalities such as placement of wiring and windows, must be specified in the architect's blueprints. The blueprints determine the activities of the construction crews from beginning to end. At first glance the blueprints for a complex building might seem to provide a good metaphoric image for the developmental regulatory program that is encoded in the DNA. For example, just as in considering organismal diversity, it can be said that all the specificity is in the blueprints: a railway station and a cathedral can be built of the same stone, and what makes the difference in form is the architectural plan. But there is an interesting problem with this metaphor, in the way the regulatory program is used in development, compared to how the blueprint is used in construction. In development it is as if the wall, once erected, must then turn around and talk to the ceiling in order to place the windows in the right positions, and the ceiling must use the joint with the wall to decide where its wires will go, etc. Thus much of the genomic regulatory program for development is devoted to the progressive organization of spatial interactions between fields of cells of differing character. The morphological outcome is the completed organism, and within a narrow anatomical range, for each species this outcome is the same, almost 100% of the time.

With the exception of unusual cases such as the syncytial early embryo of Drosophila, specification of cell fate in early development usually involves signaling between cells, and the same is true of adult body plan formation. Here developmental "specification" is defined as the process by which cells acquire the identities or fates that they and their progeny will adopt. Specification results in differential expression of genes, the readout of particular genetic subprograms. For specification to occur, genes have to make decisions, depending on the inputs they receive. This brings us back to the information processing capacities of bilaterian c¿s-regulatory systems. As illustrated in the next chapter, genes operating at every level of the developmental process have to carry out information processing functions in the course of developmental specification. The point cannot be overemphasized: were it not for the ability of czs-regulatory systems to integrate spatial signaling inputs together with multiple temporal, and other inputs, bilaterian development could never occur. This is because development depends on creating new spatial and temporal domains of gene expression from preexisting information.

Until terminal form is achieved in each morphological element of a developing organism the genetic subprograms expressed in its progenitor cells are transient; they are way stations along the pathway to the final state of spatial organization. Transience implies temporal control, and in development temporal regulation is often exquisitely precise. There are many examples of czs-regulatory systems that operate for only a few hours of the whole life cycle, since the functions mediated by the genes they control are not useful at any other time. For instance, the establishment of transcriptional dorso/ventral (D/V) asymmetry in the frog embryo must occur within the crucial hour or so when the embryo genomes are activated at the end of the blastula stage. Genes encoding particular homeodo-

main transcriptional regulators, viz, siamois, twin, and goosecoid, are then transiently activated, specifically on the dorsal side of the embryo. The cell signaling and other events regulated by these factors are required even before gastrulation commences (for review, see Harland and Gerhart, 1997; Moon and Kimelman, 1998). Similarly, specialized gene products of individual differentiated cell types are often required only at specific stages, and only in those cells. A striking example is provided by the "glue proteins," which are produced by the salivary glands of Drosophila larvae just prior to the pupal stage (Beckendorf and Kafatos, 1976). These adhesive proteins, encoded by a small family of coordinately expressed genes, are necessary to fix the larva to a dry surface. There, immobile and enclosed within the pupal case, it will undergo metamorphosis. The glue proteins are of use only for this specific purpose, and are synthesized only during the last several hours of the 4-day larval period, out of the whole life cycle. Needless to say, the nature of the specific inputs which are utilized to effect temporal regulation in transiently expressed genes is determined explicitly by the target sites that are hardwired into the respective civ-regulatory systems of these genes.

Pattern Formation

"Pattern formation" is a term of many uses, because spatial patterns appear at every level of biological organization from subcellular to organismal. In this book it is applied specifically to the developmental process by which spatial domains of unspecified cells are assigned "regulatory states," thus ultimately creating fields of cells that will give rise to the diverse parts of a structure, an organ, an organism. The early events of pattern formation establish the basic elements of the body plan, e.g., the metameric segmentation which arises early in the embryogenesis of a fly; or the anterior/posterior (A/P) axis, and left-right asymmetry in a mouse. Later pattern formation events define the spatial organization of the main parts of the body plan, and still later ones define more detailed and smaller elements, e.g., the arrangement of the sensory bristles in the fly or of the limb digits in the mouse. Pattern formation processes are progressive in that as development proceeds, spatial elements are specified at increasingly fine resolution. Regardless of the scale on which it occurs, or at what stage of development, pattern formation requires the partitioning of existing domains of cells into novel subdomains. Until terminal differentiation processes are called into play, these developmental spatial patterns are therefore all transient. Their only significance is to enable the next stages of spatial subdivision to occur.

At the heart of pattern formation mechanisms is the process of regional specification. New subdivisions of an existing cellular domain are created by the expression of one or more particular transcription factors within a bounded region. This establishes a new regulatory state in this region, distinguishing it from the remaining cells, and setting up the program for subsequent developmental events. To view the process of pattern formation in this way is to orient our quest for causality in the pathways by which morphogenesis is organized; we now have an algorithm for tracking the genetic program for pattern formation to its lairs in the genome. The track leads to the cfe-regulatory elements controlling spatial expression of genes encoding those transcription factors which execute regional specification processes. It is these c/s-regulatory systems that perform the crucial function of integrating the signaling, lineage, or other spatial inputs which define the regions to undergo regional specification. In turn, the transcription factors which they control provide specific downstream inputs into the transcriptional activities of other genes, at the next level of the regulatory network. These are often genes encoding other spatially expressed transcription factors, as well as genes encoding elements of intercellular signaling systems; hence the progressive nature of the pattern formation process.

A very clear example of the way in which a pattern formation process is encoded in cis-regulatory DNA sequence is afforded by the process of metamer-ization early in the embryonic development of Drosophila. Metameric organization is a basic feature of the dipteran body plan, and the process by which it is created begins even before cellularization of the blastoderm. Among the genes which initiate this pattern formation process are the "pair-rule" genes which begin to be expressed in seven repeating "stripes" (alternating on and off) along the A/P axis of the embryo. These stripes define two-segment units from which the full complement of body segments is ultimately generated (see for review Rivera-Pomar and Jackie, 1996). The primary pair-rule genes encode transcription factors. Each stripe of pair-rule gene expression constitutes a distinct regulatory subdivision, which is positioned specifically along the A/P axis within a much larger cellular domain at the cellular blastoderm-stage, i.e., the future "trunk" region of the body plan. The location of each stripe of expression of course depends on the cis-regulatory systems controlling these genes. In some well-studied cases the necessary control sequences for given stripes are contained within discrete cz's-regulatory modules (Small etal., 1991; 1992), and a demonstration of this is shown in Fig. 1.4.

Modules controlling expression in different stripes contain sites for different collections of transcription factors, and together they define all the regions of A/P cellular space in which the respective pair-rule genes are expressed. Each of the individual stripe modules reads and integrates the irans-regulatory inputs with which it is confronted, in each cell. As we discuss in detail in the following chapter, each of the modules contains target sites for positively acting transcription factors which cause it to be expressed; and for negatively acting transcription factors which set the anterior and posterior boundaries within which expression is allowed. That is, a given stripe module directs expression of the gene it controls when two external conditions are met: the transcriptional activators for which it includes binding sites are present and active, and the transcriptional repressors for which it includes binding sites are not present at significant levels. This is a paradigmatic case, since in the stripe modules of the pair-rule genes we are brought face to face with czs-regulatory elements of which the sole function is to organize spatial patterns of transcription factor expression.

FIGURE 1.4 Accurate expression of eve stripe 2 and eve stripes 3+7 mediated by individual ci's-regulatory elements. The evenskipped (eve) gene generates a metameric pattern consisting of seven stripes arrayed along the A/P axis, which appears in the precellular embryo. The brown stain displays the location of the seven eve protein stripes, as revealed by anti-Eve antibodies. (A) Transgenic embryo carrying a minimal 480 base pairs (bp) stripe 2 c/s-regulatory construct with the lacz gene as a reporter. The mRNA generated by the lacz reporter is visualized by in situ hybridization (purple stain), and the construct can be seen to recreate the endogenous stripe 2 expression domain perfectly. [(A) From Small et al. (1992) EMBOJ. 11, 40474057 and by permission of the European Molecular Biology Organization.] (B) Transgenic embryo carrying a 500 bp stripe 3 + 7 c/s-regulatory construct, lacz mRNA detected as in (A). The expression domain again corresponds exactly with the endogenous eve 3 + 7 stripes. (C) RNA in situ hybridization using a probe that detects eve RNA reveals the same seven stripes as displayed by immunostaining in (A) and (B). [(B, C) From Small et al. (1996) Dev. Biol. 175, 314-324.]

FIGURE 1.4 Accurate expression of eve stripe 2 and eve stripes 3+7 mediated by individual ci's-regulatory elements. The evenskipped (eve) gene generates a metameric pattern consisting of seven stripes arrayed along the A/P axis, which appears in the precellular embryo. The brown stain displays the location of the seven eve protein stripes, as revealed by anti-Eve antibodies. (A) Transgenic embryo carrying a minimal 480 base pairs (bp) stripe 2 c/s-regulatory construct with the lacz gene as a reporter. The mRNA generated by the lacz reporter is visualized by in situ hybridization (purple stain), and the construct can be seen to recreate the endogenous stripe 2 expression domain perfectly. [(A) From Small et al. (1992) EMBOJ. 11, 40474057 and by permission of the European Molecular Biology Organization.] (B) Transgenic embryo carrying a 500 bp stripe 3 + 7 c/s-regulatory construct, lacz mRNA detected as in (A). The expression domain again corresponds exactly with the endogenous eve 3 + 7 stripes. (C) RNA in situ hybridization using a probe that detects eve RNA reveals the same seven stripes as displayed by immunostaining in (A) and (B). [(B, C) From Small et al. (1996) Dev. Biol. 175, 314-324.]

All morphological features of adult bilaterian body plans are created by means of pattern formation process. The cz's-regulatory systems that mediate regional specification are thereby the keys to understanding how the genome encodes development of the body plan. These systems also provide the most fundamental and powerful approach to understanding the evolution of bilaterian forms.

Terminal Differentiation

With the completion of development, populations of permanently specialized cells have been put in place throughout the organism. These express sets of genes encoding the definitive attributes of each component of the body plan. Terminally differentiated cells include familiar cell types: muscle cells, many kinds of hematopoietic cells, nerve cells, skeletogenic cells, gut cells, liver cells, light receptor cells, gland cells, and so forth. All metazoans produce some differentiated cell types, but only the bilaterians generate the three-dimensional morphological arrays of terminally differentiated cell types that we term "organs," and only in the adult body forms of bilaterians do layers or tubes of differentiated mesodermal cell types constitute major components of all the body parts. Under normal conditions, in bilaterians terminal differentiation is indeed just that: additional functions will never be expressed by these cells, and if they retain replication capacity they will only produce more of their own sort. There are some examples of highly differentiated cells which in the course of development dedifferentiate and then re-differentiate in a second direction, but these cases are unusual ("trans-differentiation"; see review in Davidson, 1986). Terminally differentiated cells rely on diverse stabilization mechanisms to lock down their transcriptional patterns (i.e., what embryologists have always called "commitment"). These include stabilizing feedback circuits in the gene networks that regulate differentiation; changes in the state of chromatin structural domains that include the regulatory systems of the relevant genes; local DNA methylation or demethy-lation; decrease in mRNA turnover rates; in some cell types even loss or total inactivation of the whole nuclear transcriptional apparatus once the cell has been sufficiently loaded with the necessary specific mRNAs. But lockdown mechanisms are not our first concern; our focus is rather on the genomic regulatory apparatus that institutes given states of differentiation. Terminal differentiation can usually be defined in terms of a precise set of differentially expressed protein coding genes, such as the hemoglobins and enzymes of the vertebrate red blood cell, or the cell surface receptors and signaling molecules of immune effector cells. Our question is what features of genomic regulatory organization account for the coordinate expression of such dedicated sets of genes, in the same cells at about the same time.

A useful concept here (and in other contexts as well) is that of the "gene battery," defined as a set of functionally linked genes expressed in concert. The term "gene battery" was originally Morgan's (1934), and was later adapted by Britten and Davidson (1969) to denote sets of genes expressed together for the specific reason that their cis-regulatory systems respond to common trans-regulatory inputs. A good bit is now known of the c/s-regulatory organization underlying expression of gene batteries in terminally differentiating cells, particularly in vertebrates (Arnone and Davidson, 1997). As a heuristic example, relevant cis-regulatory elements from four amniote muscle genes are diagrammed in Fig. 1.5. We see that several types of DNA-binding transcription factor are utilized by these regulatory systems, viz, a muscle specific bHLH factor of the MyoD family (MDF);

Was this article helpful?

0 0

Post a comment