Hierarchies of Regulatory Circuits Batteries and Networks

Some TFs are known as "selectors", meaning that they activate a hierarchical cascade of change in the expression of other, developmentally "downstream" genes. This is comparable to the notion of an embryological "organizer" that we will see in later chapters. Another term for the idea is "master control" genes, but it can be misleading to think of genes so metaphorically in terms of our own culture's social structure; rarely if ever is one gene the master of a whole organ in a very meaningful sense. Selectors cause a major differentiation commitment or branch point. They activate a series of other TFs that in turn activate still more TFs that ultimately alter the expression of structural genes, enzymes, and the like. Hox, Pax, and MyoD genes are examples. Different selectors work at different levels or stages of embryological development (Carroll, Grenier et al. 2001). But they don't act alone, and, typically, other factors (perhaps controlling their own downstream effects) are also involved.

We will elaborate further on this in Chapter 9, but here we can indicate the uses of linkage relationships among selector genes. The members of the Hox class of homeobox TF genes have become an archetype of high-level selectors in which linkage arrangements are important. The Hox genes are linked in chromosomal clusters (shown in Figure 4-4), whose individual genes are expressed in a colinear, sequential, cumulative fashion in cells from anterior to posterior regions of the early embryo, respectively. That is, the gene at the 3' end of a Hox cluster is expressed earlier in development, in cells in the most anterior parts of the embryo. A bit later, as the next most posterior region of the embryo begins to develop, the next most 5' Hox gene in the cluster also becomes expressed (that is, along with the Hox gene that is already "on").This continues down the line of genes in the cluster to the most 5' gene, when the most posterior anatomic segments develop, which is when most or all of the genes are being expressed. Because the latest gene to be activated as more posterior segments differentiate has its effect on those segments, even though earlier (more "anterior") genes may still be expressed, the effect is known as "posterior prevalence."

Figure 7-4A repeats part of Figure 4-4 to show the gene clusters in Drosophila and in mammals, and is another oft-used figure because of its great importance in the recent history of developmental biology and genetics. The colinear summed expression pattern sets up combinatorial sets of Hox transcripts that effect gene expression cascades that differ locally along an anatomic axis. The figure shows the similarity in this aspect of the expression pattern that is shared between vertebrates and invertebrates, first documented in a famous paper by Lewis in flies (Lewis 1978) and later shown in vertebrates.

Essentially, the same summed colinear expression system is used to specify the primary AP axis in species across much of the animal world, although the details of what the various combinations achieve vary. However, the Hox cluster genes in vertebrates are also used in the development of many other structures, including the segmentation of the gut and limbs. Thus, a Hox gene is not a "vertebra" gene, but can be viewed as a selector gene, or in some contexts a patterning selector gene in several structures. The use is functionally arbitrary as noted above and not inherent in the genes themselves, and can only work because of the contingent, hierarchical, sequestered nature of development (e.g., limb buds are not the same kinds of cells as are early vertebral precursors).

Not all pathways are as deeply conserved as the Hox system, nor rely on conserved linkage relationships. Even this classic system is subject to all the sources of variation that one might expect; in fact, the more intensely it is studied, the less completely conserved it seems to be (e.g., Levine et al. 2002), including the variety of ways the genes in the cluster may be regulated to coordinate their spatial and temporal domains of expression (Kmita and Duboule 2003). Nonetheless, much is conserved, and similar situations apply to a number of other differentiation hierarchies, such as the role of Pax6 in eye development (Chapter 14), or MyoD in muscle development (Gerhart and Kirschner 1997). Figure 7-4B shows the general logic of a selector gene, and the way that REs are used to start and proliferate a cascade of downstream effects on gene expression.

A regulatory gene that acts early in helping trigger differentiation cascades may be widely expressed in an organism; some TFs seem to help regulate a high fraction of all the genes in the genome, as seems to be the case with the Ftz and Eve genes in Drosophila (Liang and Biggin 1998). However, the spatiotemporal distribution of most regulatory genes is restricted at least somewhat, and like a Venn diagram of the embryo, the set of regulatory signals activated in a given cell is assumed to be correlated with the functional genes activated in that cell. It is assumed that there is a unique combination for each function, or at least for each context in an organism; although this is far from proven, it is central to the present view of differentiation by sequestered combinations of cis-regulating factors.

A critical aspect of this that enables so much combinatorial pleiotropy to work is that gene activation states can be mitotically heritable. This is a most important sequestration factor. Once a differentiation cascade has been initiated, a cell and its descendants in the organism can become strikingly autonomous, either self-



Chr 6


Chr 11


Chr 15


Chr 2

vi* Ubx Antp Scr Dfd Zen pb lab

A13 A11 A10 A9 A7 A6 A5 A4 A3 A2 A1


D13 D12 D11 D10 D9 D8 D4 D3 D1


0 0

Post a comment