urchin species O/A polarity is not established until after cleavage has begun (Cameron et al, 1989; Davidson et al, 1998). The cyllla cis-regulatory system is relatively well understood (Kirchhamer and Davidson, 1996; Coffman etal., 1997). Figure 2.2A illustrates the two cis-regulatory modules that control cyllla expression in the embryo. The proximal module initiates function early on, during the initial processes of ectoderm specification, and a more distal module drives expression in the blastula-gastrula stage as the aboral ectoderm differentiates into squamous epithelium. The endogenous pattern of cyllla expression, as it appears in the late gastrula, is shown in Fig. 2.2B.
Four different positive regulators work together to activate the early module, all of maternal origin, while expression of the late module depends on a largely zygotic Runt family transcription factor, which is transcribed actively following the blastula stage (Coffman et al, 1996). But the activity of none of these positive regulators is confined to the aboral ectoderm cell lineages. Although the spatial details are not known for some, they all appear to be functional in multiple embryonic territories, including both oral and aboral ectoderm. It will no longer surprise (though it did on discovery, Hough-Evans et al, 1990) that expression is confined to the aboral ectoderm by repressive cis-regulatory interactions. Each of the two modules includes target sites for its own spatial repressors. The two sites labeled "P" in the early module of Fig. 2.2A bind a repressor called "P3a2." If either site is deleted, the reporter is found to be expressed ectopically in the oral ectoderm [Kirchhamer and Davidson, 1996; in this case a gene encoding chloramphenicol acetyltransferase (CAT) is the reporter]. Results from such an experiment are illustrated in Fig. 2.2D and its control, Fig. 2.2C. Exactly the same effect is observed if instead of removing the target sites for P3a2, the P3a2 factor is inactivated in vivo. This has been accomplished by introduction of a vector encoding a single chain antibody against the P3a2 DNA-binding site (Bogarad etal, 1998). But there are other repressive interactions required as well. When the late module becomes dominant, i.e., when the Runt factor is prevalent, two different repressors come into play: a 12-Zn finger repressor binding at the site labeled "2" in Fig. 2.2A prevents ectopic expression in skeletogenic cells (Wang etal, 1995; Fig. 2.2E); and a Myb-class repressor binding at site "M" prevents expression in oral ectoderm and also gut (Coffman et al., 1997; Fig. 2.2F). By late in embryogenesis the Myb repressor is localized, as shown rather dramatically in Fig. 2.2G, to those tissues where it functions as a cyllla repressor, viz, oral ectoderm, gut, and mesenchyme cells. In keeping with our theme of intramodular repression functions, it is an important experimental result that if the whole of the late module is deleted, repressor sites and all, no ectopic expression is observed, only a low level of spatially correct expression. Therefore the late module repressors are just needed to control expression driven by the late module activator (i.e., the Runt factor). In other words, ectopic expression is seen only if the Runt site is present, and the Myb or Zn finger repressor sites are also mutated or absent from the construct.
The role of the P3a2 repressor is additionally interesting because its asymmetric activation appears to play a role in the initial process by which O/A (oral/aboral)
polarity is set up early in cleavage. The activity of P3a2 depends on ambient redox state (Coffman and Davidson, 2001). As has long been known, the sea urchin egg displays an early O/A asymmetry in redox potential (Czihak, 1963; Coffman and Davidson, 2001) possibly caused by cytoskeletal rearrangements that concentrate mitochondria on one side of the egg following fertilization. If artificial means of altering the redox gradient are imposed, O/A specification is correspondingly constrained. The redox gradient may explain how the maternal P3a2 factor acts as a repressor on the oral side of the early embryo, wherever the oral pole comes to lie. It is the job of the early czs-regulatory module of cyllla to interpret the spatial distribution of regulatory activity; and thus to transduce the polarized cline of P3a2 activity into an asymmetric and ultimately lineage-specific pattern of zygotic gene activity.
cis -Regulatory Design for Autonomous Modular Function
In the two czs-regulatory modules we have so far discussed, the repressors can be thought of as acting by canceling the output of the positive regulators that bind within the same module, when and where these repressors are present. These modules are independent devices with self-contained positive and negative spatial control functions. The further examples summarized in Figs. 2.3 and 2.4 illustrate their autonomy, and provide some idea of the constraints on this kind of design.
The functional organization of the neuroectodermal control module of the Drosophila gene rhomboid (rho) is described in Fig. 2.3. This gene encodes a cell surface component of a signaling pathway, which begins to be expressed during specification of the neuroectodermal territory from which the ventral CNS will derive (about one fourth of the cells will become neuroblasts; the remainder ectoderm). The embryo is at this stage still syncytial, and the neuroectodermal regions consist of two longitudinal stripes of nuclei in which rho is expressed, overlying the ventral mesodermal domain on either side. The relevant cz's-regulat-ory module of the rho gene contains binding sites for several activators. Of these the most important is Dorsal (Dl), a Rel domain regulator which is the key early transcription factor in dorsoventral (D/V) patterning. As discussed below in more detail, by late cleavage Dl is present in the syncytial nuclei in a graded concentration cline from ventral to dorsal. Other positive regulators which function syner-gistically with Dl in the rho neuroectoderm czs-regulatory element are Twist, a bHLH activator, and some additional bHLH factors (Ip et al., 1992a; Jiang and Levine, 1993). Figure 2.3A shows that this rho cis-regulatory element generates the two lateral bands of expression that constitute the neuroectodermal territories if incorporated in a lacz fusion construct (Gray et al., 1994). Exclusion of expression from the ventral mesodermal domain is due to the Snail (Sna) repressor, which binds at the indicated sites within the rho as-regulatory module. The sna gene is itself activated in the ventralmost regions of the embryo of the future mesodermal domain by high levels of nuclear Dl protein (Ip et al., 1992b; Jiang and Levine, 1993). If the Sna target sites in the rho regulatory element are mutated, lacz
50 bp 50 bp
FIGURE 2.3 Modes of c/s-regulatory expression in Drosophila. (A-D) Short-range repression; (E, F), long-range repression. (A) rhomboid (rho) gene neuroectodermal module, and its expression revealed by a lacz transgene, displayed by in situ hybridization; ventral view of an early cellularization-stage embryo. The construct is expressed in two broad lateral stripes directly abutting the ventral mesodermal domain, in which it is silent. Factors that bind the target sites shown are Dorsal (red); a bHLH factor (green); Twist (yellow); (all activators are shown below the line representing the DNA); and Snail (black boxes), a repressor. (B) The Snail sites are responsible for repression in the mesoderm, since ectopic expression occurs there if
50 bp 50 bp
FIGURE 2.3 Modes of c/s-regulatory expression in Drosophila. (A-D) Short-range repression; (E, F), long-range repression. (A) rhomboid (rho) gene neuroectodermal module, and its expression revealed by a lacz transgene, displayed by in situ hybridization; ventral view of an early cellularization-stage embryo. The construct is expressed in two broad lateral stripes directly abutting the ventral mesodermal domain, in which it is silent. Factors that bind the target sites shown are Dorsal (red); a bHLH factor (green); Twist (yellow); (all activators are shown below the line representing the DNA); and Snail (black boxes), a repressor. (B) The Snail sites are responsible for repression in the mesoderm, since ectopic expression occurs there if expression spreads throughout the ventral mesoderm (Fig. 2.3B), a dramatic effect. The experiment reproduced in Fig. 2.3C provides insight into the structural organization of czs-regulatory elements of this sort. Here two Sna target sites have been restored to the mutated construct in Fig. 2.3B, but they are positioned 50 bp outside all the positive regulatory sites on either end of the sequence. Nonetheless, ventral repression is restored, and in fact, a single Sna site suffices. However, if the Sna sites are removed to distances of 150 and 120 bp, the spatial repression that they exert is greatly weakened (Gray et al., 1994). The Sna sites do not have to be in any particular position, e.g., overlapping with, or immediately adjacent to, the sites for positive regulators, but they have to be within about 100 bp of at least some of those sites: there is considerable design flexibility, but also a crucial design constraint in the sequence organization of this regulatory module.
The term "autonomy" here means that the package of target sites within a module suffice for it to execute its function, irrespective of its regulatory environs. as-Regulatory autonomy is demonstrated precisely in the experiment shown in Fig. 2.3D (Gray and Levine, 1996). The eve stripe 2 module and the rho neuroectoderm module are here combined in a single lacz expression vector: they both operate, independently, producing a summed pattern of lacz mRNA.
these sites are mutated (open black boxes). (C) Snail still functions to repress expression in the mesodermal domain when its target sites are moved to positions 50 bp proximal and distal to the nearest activation sites; the natural Snail target sites mutated as in (B). [(A-C) Adapted from Gray et al. (1994) Genes Dev. 8, 1829-1838.] (D) Autonomy of short-range repression. Two different modules, each including short-range repressors, are physically linked in a single lacz expression construct. This contains a version of the rho element of (A), distal, and the eve stripe 2 element, proximal (see Fig. 2.1 for structure; blue, Kr repressor; black, Bed activator). The two modules function independently of one another, generating both the A/P rho pattern and the D/V eve stripe 2, so as to produce a crossed pattern. The Sna repressor does not prevent eve stripe 2 expression in the mesoderm, and the Kr repressor does not prevent rho expression, though it has a mild weakening effect (bracket). [(D) Adapted from Gray and Levine (1996) Genes Dev. 10,700-710.] (E) Expression of a vector containing two copies of the ventral longrange repression element of the zen gene, distal to the same eve stripe 2 module (eve2) as above, "evellacz" denotes the promoter/lacz reporter fusion. The zen element contains sites for Dorsal (red circles) and three sites (I, II, III) where other factors bind (colored boxes). Though the zen elements are positioned hundreds of bp upstream of the BTA, and outside of the eve stripe 2 element, lacz mRNA is seen only in the dorsal region of the eve stripe 2, and expression is silenced elsewhere. (F) Expression of eve stripe 2 is restored throughout if the second of the three sites in the zen repression element that are shown as colored boxes (I, II, III) is changed to be the same as the first; this is the site at which Deadringer and Cut corepressors interact (Valentine et al., 1998). [(E, F) Adapted from Cai et al. (1996) Proc. Natl. Acad. Sci. USA 93, 9309-9314, copyright National Academy of Sciences, USA.]
The experiment proves that the repression functions that set the anterior and posterior boundaries of eve stripe 2 (i.e., those mediated by Kr and Gt regulators), and the repressor setting the ventral boundaries of the rho stripes as well (i.e., Sna), are dedicated to the control of their own respective modules. There is no gap in the eve stripe where Sna, the rho repressor, is active, and only a very slight attenuation in the rho stripes where the Kr repressor is active. With respect to the disposition of their czs-regulatory target sites, all of these repressors work over a short range, in that their effect is confined to the modules within which they bind.
A similar example from the sea urchin is shown in Fig. 2.4. Here again two different czs-regulatory modules are combined. One of these, which includes the only basal transcription apparatus in the construct, is from the sm.50 skeletal matrix protein gene, and this regulatory module confers expression of the CAT reporter exclusively to skeletogenic mesenchyme cells (Fig. 2.4A). The second is from the endo 16 gene, which at the stage shown functions only in the gut (Fig. 2.4B). The workings of the endol6system are discussed below, and for now the only important point is that it includes target sites for repressors required to prevent expression of endol 6 in skeletogenic mesenchyme cells. But again this repression function is pointed at its own czs-regulatory system, and it does not interfere at all with skeletogenic expression driven by the sm50 module. When combined, each module operates as if the other were not there: the output is just the sum of the two expression patterns (Fig. 2.4C).
Other czs-regulatory repressors function in a different way, silencing all transcriptional activity mediated by the basal transcription apparatus (BTA) which
FIGURE 2.4 Autonomy of repression in a synthetic Strongylocentrotus cis-regulatory vector. (A) Expression of a vector in which a CAT gene serves as reporter, driven by a cis-regulatory module from the skeletogenic gene sm50. CAT mRNA is visualized in a late gastrula-stage embryo viewed from the side by in situ hybridization. Expression is confined to skeletogenic mesenchyme cells (white arrowheads). The sm50 module contains sites for positive mesenchyme specific factors but no negatively acting elements (Makabe et al., 1995). [(A) From Lee et al. (1999) Dev. Growth Differ. 41, 303-312.] (B) Expression of endo 16 c/s-regulatory system in gut (white arrowheads). This system specifically represses endo 16 expression in skeletogenic mesenchyme cells, as well as in ectoderm (see text). [(B) From Yuh et al. (1994) Mech. Dev. 47, 165-186; copyright Elsevier Science.] (C) Synthetic fusion of regulatory systems shown in (A) and (B). The endo 16 system is distal and the sm50 module including the BTA are proximal. Expression is additive, and there is no repression of the transcriptional activity in skeletogenic mesenchyme which is mediated by the smSO module. Black arrowheads indicate staining in skeletogenic cells; white arrowheads indicate expression in gut cells. Similarly skeletogenic expression driven by the sm50 module remains unaffected when an endo 16 DNA fragment containing only the skeletogenic repressor sites is directly linked to it. [(C) Adapted from Kirchhamer et al. (1996) Proc. Natl. Acad. Sci. USA 93, 13849-13854; copyright National Academy of Sciences, USA.]
Was this article helpful?