The segmented body pattern along the longitudinal axis of the Drosophila embryo is established by a cascade of specific transcription factor activities. This cascade is initiated by maternal gene products that are localized at the polar regions of the egg. The initial long-range positional information of the maternal factors, which are transcription factors (or are factors which activate or localize transcription factors), is transferred through the activity of the zygotic segmentation genes. The gap genes act at the top of this regulatory hierarchy. Expression of the gap genes occurs in discrete domains along the longitudinal axis of the preblastoderm and defines specific, overlapping sets of segment primordia. Their protein products, which are DNA-binding transcription factors mostly of the zinc finger type, form broad and overlapping concentration gradients which are controlled by maternal factors and by mutual inter-actions between the gap genes themselves. Once established, these overlapping gap protein gradients provide spatial cues which generate the repeated pattern of the subordinate pair-rule gene expression, thereby blue-printing the pattern of segmental units in the blastoderm embryo. Our results show different strategies by which maternal gene products, in combination with various gap gene proteins, provide position-dependent sets of transcriptional activator/repressor systems which regulate the spatial pattern of specific gap gene expression. Region-specific combinations of different transcription factors that derive from localized gap gene expression eventually generate the periodic pattern of pair-rule gene expression by the direct interaction with individual cis-acting “stripe elements” of particular pair-rule gene promotors. Thus, the developmental fate of blastoderm cells is programmed according to their position within the anterior-posterior axis of the embryo: maternal transcription factors regulate the region-specific expression of first zygotic transcription factors which, by their specific and unique combinations, control subordinate zygotic transcription factors, thereby subdividing the embryo into increasingly smaller units later seen in the larva.

The initial establishment of polarity along the longitudinal axis of the Drosophila embryo is carried out by three fairly independent genetic pathways of maternal origin. They are required for the development of the anterior, the posterior and the terminal regions of the embryo (Nüsslein-Volhard et al., 1987; Nüsslein-Volhard, 1991; St. Johnston and Nüsslein-Volhard, 1992; for reviews). The first of the zygotic genes to interpret the positional information provided by the maternal genes are members of the gap class of segmentation genes (Nüsslein-Volhard and Wieschaus, 1980). They are responsible for the development of specific, overlapping sets of segment primordia. For example, hunchback (hb) is needed to establish the gnathal and thoracic segments (Lehmann and Nüsslein-Volhard, 1987), Krüppel (Kr) for thoracic and anterior-abdominal segments (Gloor, 1954; Nüsslein-Volhard and Wieschaus, 1980; Wieschaus et al., 1984; Preiss et al., 1985), knirps (kni) and giant (gt) for abdominal segments (Nüsslein-Volhard and Wieschaus, 1980; Petschek et al., 1987) and tailless (tll) and huckebein (hkb) for pattern elements in the head and tail regions (Strecker et al., 1988; Weigel et al., 1990). The correct spatial regulation of the gap genes is critical for normal development since ectopic presence of the gap gene products results in severe alterations of the embryonic body pattern (Jackie et al., 1986; Gaul et al., 1987; Klingler et al., 1988; Eldon and Pirrotta, 1991; Kraut and Levine, 1991a; Steingrimsson et al., 1991). All of the gap genes studied to date encode proteins containing DNA-binding motifs (Pankratz et al., 1990b; for review). With the exception of the gt protein (GT), the above mentioned gap genes encode multiple zinc finger motifs (Rosenberg et al., 1986; Tautz et al., 1987; Nauber et al., 1988; Pignoni et al., 1990; G. Brbnner and H. Jackie, unpublished results), and they are expressed in regions of the preblas-toderm embryo that are affected in the larval segmentation pattern seen later in the corresponding mutants (Knipple et al., 1985; Gaul et al., 1987; Tautz et al., 1987; Mohler et al., 1989; Rothe et al., 1989; Pignoni et al., 1990; Brönner and Jackie, 1991).

The mechanism by which the gap genes are regulated has been studied in detail for one of the gap genes, hb, which is expressed in the anterior half of the early embryo (Tautz et al., 1987; Schroder et al., 1988). These studies indicated that hb is directly activated in response to the protein product encoded by the anterior maternal organizer gene bicoid (bed). After transcription in the nurse cells, bed mRNA is transferred into the growing oocyte, and it eventually becomes localized in the anterior pole region of the egg. After translation, the bed protein (BCD) forms a concentration gradient along the longitudinal axis of the embryo and acts as the anterior morphogen (Frohnhbfer and Nüsslein-Volhard, 1986; Driever and Nüsslein-Volhard, 1988a,b, 1989; Driever et al., 1989; Struhl et al., 1989). The spatial restriction of zygotic hb expression to the anterior region of the embryo is due in most part to the limiting amount of the activator BCD in the posterior half. It has been suggested that other zygotic segmentation genes, in particular in the prospective head region, might be activated through a similar mechanism by BCD (Driever et al., 1989; Struhl et al.. 1989). The detailed mechanisms by which the gap genes acting in the central and posterior regions, such as Kr, kni and gt, and the gap genes in the posterior terminal region, tll and hkb, are spatially regulated are the subject of current studies reported here. These studies also address the question of how the maternal morphogenetic input on gap gene expression is then interpreted by the pair-rule genes by making use of the several overlapping gap protein gradients along the anterior-posterior axis.

Combined genetic and molecular analyses indicate that gap genes most likely carry out segmentation function by directing the expression of the pair-rule genes (Akam, 1987; Ingham, 1988; Pankratz and Jackie, 1990; for reviews). Members of this class of segmentation genes are expressed in tandemly repeating stripes in the blastoderm embryo. The patterning of all pair-rule genes examined to date is altered in embryos mutant for any of the gap genes which act in the trunk region of the embryo, indicating that those gap genes are involved in generating the stripe pattern of pairrule gene expression. Because a regulatory hierarchy exists even among the pair-rule genes, i.e. primary pair-rule genes such as hairy (h) or even-skipped (eve) are required for the normal expression of secondary pair-rule genes such as fushi tarazu (ftz), the gap genes act most likely at the level of the primary pair-rule genes. In this paper we summarize results indicating that the molecular mechanism underlying gap gene expression in different regions of the Drosophila embryo are fundamentally different from and more complex than the activation of hb by BCD. We show that Kr and kni expression patterns arise from superimposition of distinct transcriptional activators and repressors of both maternal and gap gene origin, and that the combinations of distinct gap gene activities provide the regulatory input to the individual stripe elements of the pair-rule gene promotors. Finally, we show that the Kr protein (KR) is capable of acting as a concentration-dependent transcriptional activator and repressor, providing the two opposing activities through different portions of the protein.

Anterior-posterior prepattern formed in response to spatially regulated transcription factors

During the early stages of Drosophila embryogenesis, the zygotic nuclei divide through a series of twelve mitotic divisions without forming cells. Towards the end of the syncytial phase of development, the nuclei migrate to the periphery to form monolayers of nuclei which then become surrounded by membranes to form the cellular blastoderm (for a review see Campos-Ortega and Hartenstein, 1985). The basic prepattern of the anterior-posterior axis is established prior to the cellularization by the localized expression of the zygotic gap genes (Fig. 1). As one moves from anterior to posterior, the embryo shows transcripts reflecting the expression domains of hkb, tll, hb (and gt), Kr, kni, gt, tll and hkb. All of these genes encode DNA-binding proteins, the majority of which have been shown to act as transcription factors (see below). The spatial domains of gap gene expression require the input of maternal gene activities and regulatory interactions among themselves, leading to a genetic circuitry required for the formation of trunk and tail segments in the embryo (summarized in Fig. 2).

Fig. 1.

Expression patterns of gap genes in the blastoderm embryo as visualized by whole mount in situ hybridization. In the text we focus on the anterior domain of hb expression (posterior is right; dorsal is up) and the posterior expression domains of the other gap genes. The spatial domains of expression require the genetic input of maternal gene activities and regulatory interactions among the gap genes, as summarized in Fig. 2.

Fig. 1.

Expression patterns of gap genes in the blastoderm embryo as visualized by whole mount in situ hybridization. In the text we focus on the anterior domain of hb expression (posterior is right; dorsal is up) and the posterior expression domains of the other gap genes. The spatial domains of expression require the genetic input of maternal gene activities and regulatory interactions among the gap genes, as summarized in Fig. 2.

Fig. 2.

Genetic circuitry establishing gap gene expression domains in the trunk and tail region of the blastoderm embryo. Arrows indicate positive regulation, lines ending with a vertical bar designate negative inputs. Note that bed activates both hb and Kr, hb activates and represses Kr in a (likely) concentrationdependent manner and represses the anterior limits of both kni and gt expression, Kr activates kni and represses gt, and that the terminal gap genes til and hkb act as repressors of Kr, kni and gt. Note that til is also required to activate posterior hb gene expression. For a detailed description of the terminal signal transduction pathway in which a torso-like (tV)-dependent signal (from outside the germ cells) is transmitted through activation (*) of the transmembrane receptor encoded by torso (tor) and the Drosophila raf oncogene-homologous Draf gene, see a recent review by St Johnston and Niisslein-Volhard (1992). Note that this signal transduction pathway is likely to activate an unknown transcription factor (TF) for reasons described in the text. Note also that hb activity is both of maternal and zygotic origin, and that the maternal activity is post-transcriptionally controlled by nanos (nos) in the posterior half of the embryo (see text). Finally note that kni (and possibly gt) are activated by a global activator system (GTS) for which no genetic basis is established. A detailed description of the formal interactions depicted here is given in the text. For the position of the expression domains in the embryo see Fig. 1.

Fig. 2.

Genetic circuitry establishing gap gene expression domains in the trunk and tail region of the blastoderm embryo. Arrows indicate positive regulation, lines ending with a vertical bar designate negative inputs. Note that bed activates both hb and Kr, hb activates and represses Kr in a (likely) concentrationdependent manner and represses the anterior limits of both kni and gt expression, Kr activates kni and represses gt, and that the terminal gap genes til and hkb act as repressors of Kr, kni and gt. Note that til is also required to activate posterior hb gene expression. For a detailed description of the terminal signal transduction pathway in which a torso-like (tV)-dependent signal (from outside the germ cells) is transmitted through activation (*) of the transmembrane receptor encoded by torso (tor) and the Drosophila raf oncogene-homologous Draf gene, see a recent review by St Johnston and Niisslein-Volhard (1992). Note that this signal transduction pathway is likely to activate an unknown transcription factor (TF) for reasons described in the text. Note also that hb activity is both of maternal and zygotic origin, and that the maternal activity is post-transcriptionally controlled by nanos (nos) in the posterior half of the embryo (see text). Finally note that kni (and possibly gt) are activated by a global activator system (GTS) for which no genetic basis is established. A detailed description of the formal interactions depicted here is given in the text. For the position of the expression domains in the embryo see Fig. 1.

Expression o/hb in the anterior region

The early zygotic expression of hb is dependent on the BCD gradient as it does not form in bed mutant embryos and it expands posteriorly when the dose of the bed gene is increased in the egg-laying females (Driever and Nüsslein-Volhard, 1988a,b). bed encodes a homeodomain protein (Frigerio et al.. 1986: Berieth et al., 1988). Its RNA is localized in the cytoplasm at the anterior pole of the egg (Frigerio et al., 1986; Berieth et al.. 1988; St. Johnston et al., 1989). After translation, the BCD transcription factor forms an anterior-to-posterior concentration gradient which determines the domain of zygotic hb gene expression extending from the anterior tip to about 50% of the egg length. The hb czs-acting region that can direct bed-dependent expression of a reporter gene in the anterior half of the embryo (Schroder et al., 1988) contains six BCD-binding sites; three of them are of high affinity and three of low affinity (Driever and Nüsslein-Volhard, 1989). Four tandem copies of a high-affinity BCD-binding site direct reporter gene expression in the anterior half of the embryo, while the corresponding number of low-affinity sites direct expression in the anterior third only (Driever et al., 1989; Struhl et al., 1989). Addition of extra BCD-binding sites of equal or lower affinity to either construct leads to an increase in the level of expression but it has no significant effect on the spatial expression domains (Driever et al., 1989). Therefore, Driever and Nüsslein-Volhard have concluded that promotors with low-affinity BCD-binding sites require high BCD concentrations to be activated, while the high-affinity binding sites require lower concentrations of BCD to direct expression in more posterior regions of the embryo. There is some doubt whether a threshold concentration of BCD is read solely by the binding site with the highest BCD affinity (Struhl et al.. 1989); it appears likely, however, that a given threshold within the BCD gradient is sufficient to activate the hb gene above a defined BCD concentration, giving rise to the hb anterior domain of expression (Driever et al., 1989; Struhl et al., 1989).

Expression o/kni and gt in the abdominal region

The products of kni and gt are required for the establishment of the abdominal region of the embryo (Nüsslein-Volhard and Wieschaus, 1980; Pctschek et al., 1987). The corresponding expression domains of these “abdominal gap genes” are adjacent to each other, and they are both regulated in response to the maternal posterior organizer gene nanos (Rothe et al., 1989; Eldon and Pirrotta, 1991; Kraut and Levine, 1991b). In contrast to bed which is thought to regulate all anteriorly acting gap genes directly, nanos provides its activity through regulation of a posterior repressor encoded by maternal hb transcripts. This means that hb is transcribed not only zygotically (see above), but also during oogenesis, and gives rise to uniformly distributed hb transcripts in the mature egg. This maternal transcript is degraded in response to nanos activity in the posterior half of the embryo (Wang and Lehmann, 1991), thereby restricting the translation of maternal hb mRNA to the anterior half (Tautz, 1988). Thus, hb protein (HB) is already present in the anterior half of the embryo before its zygotic complement is expressed in response to the BCD concentration gradient. In nanos mutants, both the hb transcript and HB are present throughout the embryo (Tautz, 1988). In this case, posterior expression of both kni and gt are blocked and abdominal segments fail to be established (Nüsslein-Volhard et al., 1987; Rothe et al., 1989; Eldon and Pirrotta, 1991; Kraut and Levine, 1991b). Since HB contains two putative DNA-binding domains, one with four and the other with two zinc finger motifs (Tautz et al., 1987), it is likely that hb prevents abdomen formation through repression of the abdominal gap genes kni and gt. Thus, the role of the posterior maternal organizer system is to prevent posterior expression of the repressor HB (Hiilskamp et al., 1989; Irish et al., 1989; Struhl, 1989). In the anterior region, HB serves as a repressor of kni and gt, setting the anterior borders of their posterior expression domains (Hiilskamp et al., 1990).

This interpretation is also consistent with recent genetic studies of Struhl et al. (1992) showing that HB acts as a concentration-dependent morphogen which prevents kni and gt expression at different concentration values; low levels suppress kni expression, and more posteriorly, even lower levels (which do not interfere with kni expression) also suppress gt expression. Thus, the anterior limits of the kni and gt expression domains can be set in response to HB. In addition, there is also evidence from genetic studies that the two genes function as weak, mutual repressors (Eldon and Pirrotta, 1991). Thus, KNI is part of the repressor system which delimits the gt expression domain (from anterior), and GT may serve similar function on kni expression (from posterior; see below).

The posterior limits of the kni and gt expression domains are mainly set in response to the terminal gap genes, which by genetic means have been shown to be strong repressors of the gap genes Kr, kni and gt (see below). The activators) of kni and gt have not yet been identified, but there is increasing evidence that it might be a factor which is capable of activating KNI throughout the embryo (Pankratz et al., 1992).

Expression of the terminal gap genes in the tail region

Expression of the terminal gap genes tll and hkb in the posterior pole region in the embryo is almost entirely controlled by the maternal terminal organizer system, and the two genes are likely to encode the zygotic transcription factors which are the functional zygotic response to the activity of a signal transduction pathway (Weigel et al., 1990; Brönner and Jackie, 1991). In this pathway, the product of the gene torso probably acts as a receptor for an extracellular signal that is produced at the two poles of the egg. torso encodes a putative transmembrane receptor tyrosine kinase which, in response to a localized extracellular signal (Sprenger et al., 1989; Stevens et al., 1990), activates downstream gene products including the as yet unknown transcription factor which eventually activates tll and hkb. The activation of tll and hkb in response to torso activity, as well as the effect of ectopic tll and hkb expression on other gap genes, has been demonstrated in torso gain-of-function alleles which encode mutant receptors with constitutive tyrosine kinase activity in the absence of activating ligands. Embryos laid by females containing these gain-of-function alleles develop normal terminal structures but they have defects in the segmented region of the pattern likely to be caused by the correlated ectopic expression of tll and hkb in these embryos. The segmental pattern defects are suppressed in embryos that are also mutant for tll and hkb, indicating that the gain-of-function phenotypes are likely due to the ectopic expression of the two terminal gap genes in the central region of the embryos. In fact, their ectopic expression leads to the repression of central gap genes such as Kr and kni (Brönner and Jackie, 1991; and references therein).

Ectopic expression of tll protein (TLL) under the control of an inducible promotor results in a torso gain-of-function, mutant-like phenotype. This is due to repression of Kr and kni (Klingler et al., 1988; Steingrimsson et al., 1991). Furthermore, in tll mutant embryos, the posterior band of gt expression (Eldon and Pirrotta, 1991) is broader and extends posteriorly as compared to wild-type expression. In hkb/tll double-mutant embryos or embryos which lack the torso gene product, gt extends to the posterior tip. Similarly, the domain of kni expression expands posteriorly in tll mutants or in hkb/tll double-mutant embryos. These results argue that the activity of the terminal gap genes is capable of repressing the activation of the central gap genes. By contrast, the expression domains of the terminal gap genes are not affected by the absence of central gap gene activity (Brönner and Jackie, 1991). Since the terminal gap genes do not receive regulatory inputs from the other gap genes in the posterior region of the embryo, the spatial limits of tll and hkb expression set a stringent posterior boundary in the embryo which separates trunk from tail primordia. In view of the ectopic activation of both tll and hkb in response to ectopic torso activity and the lack of control by other gap genes or by mutual interactions, one must postulate that terminal gap genes are solely dependent on a transcription factor which is locally activated in response to torso activity. By analogy to the known long-range morphogen gradient provided by BCD in the anterior region of the embryo, both hkb and tll may be activated above different concentration values of the active form of this transcription factor, in a manner similar to hb activation by BCD.

Expression of Yet in the central region

Expression of Kr fills the region between the anterior hb and posterior kni expression domains (Fig. 1). In the absence of zygotic activity of each of these two genes or of gt, the Kr expression domain expands, to various degrees, towards the wild-type expression domain of the gene whose activity is missing. This suggests that TLL (see above), HB, kni protein (KNI) and GT act as potential repressors of Kr gene expression, although their ability to repress could be of different strengths (see below). This leaves the question of which factor(s) activate Kr in a central region of the wild-type embryo.

In the absence of the anterior morphogen BCD or zygotic HB, Kr is expressed in a more anterior region than normal (Gaul and Jackie, 1987, 1989). In the absence of both BCD and HB (maternal as well as zygotic), however, no Kr expression can be observed (Hülskamp et al., 1990). These findings suggest that BCD and HB can each activate Kr gene expression. Since Kr expression expands anteriorly in the absence of zygotic HB, it appears likely that different HB concentrations would have different effects on Kr expression in the anterior region of the embryo: low levels of HB present in the center of the embryo would activate it (Hülskamp et al., 1990) while high levels of HB in a more anterior position would repress Kr expression (Gaul and Jackie, 1987, 1989). This model would explain why Kr expression, although dependent on BCD, can expand in the posterior-most region of the embryo: the expansion occurs under the control of maternally supplied HB when posterior repressors such as KNI, GT, TLL and hkb protein (HKB) are absent from embryos which lack both nanos and torso activity (see above).

Once activated, KR influences the expression of its neighbouring genes by acting as a repressor, delimiting secondary zygotic expression of hb in the anterior, and gt expression in the anterior and posterior region. In addition, it enhances kni expression by two means: KR acts directly through interaction with a kni cis-acting region and also through repression of gt which itself is a repressor of kni. Evidence for this proposal comes from reduced levels of kni gene expression in Kr mutant embryos. In these embryos, gt expression extends towards the anterior. However, wild-type levels of kni expression are observed in the absence of both Kr and gt gene activities. The finding of functional KR binding sites in the Kr-dependent kni cis-acting region (Pankratz et al., 1989; F. Sauer, unpublished results) suggests that Kr and gt carry antagonizing functions for the control of the level of kni gene expression.

The genetic interactions between maternal and gap genes described here (Fig.2) and elsewhere (Gaul and Jackie, 1990; Nüsslein-Volhard, 1991; Hiilskamp and Tautz, 1991; St. Johnston and Nüsslein-Volhard, 1992; for reviews) allow each of the gap genes to be placed in a specific scenario of putative transacting factors required for the normal spatial control of their expression. For comparative reasons and to point out the similarity and differences in the mode of gap-gene spatial regulation with respect to the known regulation of hb, we focus here on Kr and kni. The genetic circuitry leading to Kr and kni gene expression and the known transacting gene activities capable of spatially controlling the expression of Kr and kni as deduced from the genetic studies described above are summarized in Fig. 2. In the following we focus on initial studies which, in part, reveal different molecular strategies by which the spatial domains of the gap genes Kr and kni are set in response to maternal factors and other gap gene products.

Spatial control of Kr expression by competitive binding of the BCD activator and repressors to overlapping binding sites

By deletion analysis of the Kr civ-acting sequences, we defined a 730 bp minimal Kr DNA sequence which is sufficient to mediate gene expression in the authentic early Kr expression domain (Hoch et al., 1990, 1991). The corresponding DNA fragment, termed Kr730-element, mediates gene expression in bed, hb, kni, gt and tll mutant embryos in a manner indistinguishable from the alterations observed for endogenous Kr gene expression in these mutants (see above). This suggests that the Kr730-element contains the cis-acting requirement for the activation of gene expression in response to BCD, and the target sites for the interactions with the gap gene proteins HB, KNI, GT and TLL. The ability of these proteins to bind to Kr730-DNA was analysed by DNase I in vitro footprinting using E. coli-derived proteins (Hoch et al., 1991, 1992). As summarized in Fig. 3, K/-730-DNA contains six binding sites for BCD, seven TLL binding sites, five GT binding sites and one region protected by KNI. In addition, ten HB binding sites were found.

Fig. 3.

BCD, HB, KNI, GT and TLL in vitro binding sites in the Á7730-C lenient which directs reporter gene expression in place of the Kr central domain during blastoderm stage.

Fig. 3.

BCD, HB, KNI, GT and TLL in vitro binding sites in the Á7730-C lenient which directs reporter gene expression in place of the Kr central domain during blastoderm stage.

The common feature of BCD and HB binding sites is that they occur in clusters. From 5 ′ towards 3 ′ of Kr730-DNA, there is a pair of HB binding sites followed by a single site and another single site embedded in a cluster of five BCD binding sites, followed by two HB binding sites, followed by a cluster of four HB binding sites and a single BCD binding site at the end (Fig. 3). One common feature of TLL, GT and KNI in vitro binding sites on Kr730-DNA is the overlap with binding sites of BCD, an activator of Kr and Kr730-mediated gene expression. In order to determine whether the overlap of the BCD binding sites with potential Kr repressor protein binding sites might be of functional significance for gene expression in vivo, we placed the Kr730-lacZ reporter gene into transgenic embryos where KNI, GT and TLL can be expressed under the control of a promotor element responsive to heat-shock.

When TLL was induced throughout the embryo by heat-shock treatment, lacZ expression directed by the Kr730-element was eliminated (Fig. 4) (Hoch et al., 1992). Similarly, when the reporter gene expression was examined in embryos where GT was ectopically expressed, lacZ expression was greatly reduced or even absent in some of these embryos (Fig. 4). In contrast, KNI ectopic expression reduced KrT3Q expression only slightly (Hoch et al., 1992). These observations correlate well with genetic data indicating that KNI only weakly represses Kr (Gaul and Jackie, 1987), whereas GT is a strong repressor (Eldon and Pir-rotta, 1991; Kraut and Levine, 1991a) and TLL can even abolish Kr expression completely (Steingrimsson et al., 1991). Since none of these gap genes interfere with bicoid activity or the formation of the BCD gradient, it appeared likely that these repressors prevent BCD-dependent activation of Kr730-mediated gene expression by competitive binding to ÀT730-DNA, thereby preventing the activator from binding.

Fig. 4.

Expression of lacZ directed by the Kr730-element in wild-type and experimentally manipulated embryos, (a) KrJ30-mediated gene expression in early blastoderm embryos. Staining of the even-skipped protein (stripe pattern) served as an internal standard for the staining intensity of β -galactosidase expression, (b) Kr730-mediated gene expression in transgenic embryos containing the kni gene under the control of the heat-inducible hsp70 promotor. After heat shock prior to blastoderm formation resulting in ectopic KNI expression, the amount of Kfl3G-mediated gene expression was reduced in comparison to the amount of even-skipped protein staining. The altered stripe pattern is indicative of ectopic kni activity, (c) Ær730-mediated gene expression in transgenic embryos containing the hsp70tll gene; after heat shock, gene expression was absent. The distribution of even-skipped protein was abnormal, indicative of ectopic tll activity.

Fig. 4.

Expression of lacZ directed by the Kr730-element in wild-type and experimentally manipulated embryos, (a) KrJ30-mediated gene expression in early blastoderm embryos. Staining of the even-skipped protein (stripe pattern) served as an internal standard for the staining intensity of β -galactosidase expression, (b) Kr730-mediated gene expression in transgenic embryos containing the kni gene under the control of the heat-inducible hsp70 promotor. After heat shock prior to blastoderm formation resulting in ectopic KNI expression, the amount of Kfl3G-mediated gene expression was reduced in comparison to the amount of even-skipped protein staining. The altered stripe pattern is indicative of ectopic kni activity, (c) Ær730-mediated gene expression in transgenic embryos containing the hsp70tll gene; after heat shock, gene expression was absent. The distribution of even-skipped protein was abnormal, indicative of ectopic tll activity.

In order to see whether such a molecular mechanism could account for the results obtained with Kr73O-DNA, we analysed the single BCD-binding site overlapped by a KNI-binding site in detail. For this, we have chosen the 16-bp sequence (5 ′)-ACTGAACTAAATCCGG-(3 ′) at the 3 ′ -end of the Kr730-DNA. KNI and BCD competed for binding to this sequence in vitro. Increasing amounts of BCD competed for the binding of KNI and vice versa (Hoch et al., 1992). These experiments demonstrated that each of the two proteins can bind to the 16-bp sequence, but their binding is mutually exclusive. To examine the possible function of interaction of KNI and BCD with the 16-bp element in vivo, we transfected Drosophila Schneider tissue culture cells with reporter gene constructs containing the bacterial chloramphenicol acetyltransferase (CAT) gene under the control of this element (Gorman et al., 1982). Cotransfection experiments with plasmid DNA containing the bed or the kni gene showed that BCD can activate 16-bp-mediated gene expression. When KNI was introduced along with BCD, BCD-dependent CAT gene activation was suppressed in a manner responsive to the dosage of KNI. KNI did not make the reporter gene inactivable, but it increased the amount of BCD necessary for activation. No repression was observed when the DNA-binding domain of KNI was mutated such that the first Cis in the second zinc finger motif was replaced by a Leu residue, thereby impairing DNA-binding function. These results indicate that KNI can act as an antagonist of BCD-dependent gene activation, suggesting that this function of KNI is provided by competitive binding to the overlapping BCD binding site. In order to see whether KNI could also act as a suppressor of BCD-dependent gene activation in the embryo, we placed the 16-bp element upstream of a lacZ reporter gene construct.

The 16-bp lacZ reporter gene construct was inserted into the Drosophila genome and lacZ expression patterns were monitored in the transgenic embryos. No lacZ expression was observed in embryos derived from >cí/-deficient females. Embryos laid by females containing the two normal copies of the bed gene show lacZ expression in the anterior third of the embryo, where BCD is present at its highest concentration. When BCD concentration within the embryo was increased by addition of wild-type copies of the bed gene in the females, the extent of lacZ expression correspondingly expanded towards the posterior, indicating that the 16-bp element responds to alterations of the BCD concentration along the longitudinal axis. In order to determine whether KNI is also able to interfere with this BCD-dependent gene activation in the embryo, we placed the 16-bp lacZ reporter gene into transgenic embryos where KNI was ectopically expressed under the control of the heat-shock responsive promotor element. In cases when kni gene expression was induced throughout the embryo, lacZ expression directed by the 16-bp element was eliminated (Hoch et al., 1992). Taken together, these findings indicate that the 16-bp element is a responsive site for the action of the activator BCD and the repressor KNI, suggesting a mechanism in which KNI represses by competitive DNA-binding (Ptashne, 1986, 1988; Levine and Manley, 1989) thereby preventing the action of the BCD activator.

In the Kr73()-element TLL overlaps all BCD-binding sites, and GT overlaps three of them. This observation and the finding that Kr730-mediated lacZ gene expression is abolished or greatly reduced upon ectopic expression of TLL and GT, respectively, suggest that gene activation or repression by competition of opposing transcriptional regulators at the same binding sites, as established here for the 16-bp element, may provide a molecular basis for gap gene products to contribute to the spatial control of Kr gene expression in response to the morphogen BCD. The concentration of a gap protein at a particular point in its gradient may effectively block gene activation by the amount of BCD found at that location, thus restricting the initial gene activation in réponse to the gradient of the BCD morphogen. This interpretation of our data is consistent with a finding that the initially broad Kr expression domain narrows, from both ends, during blastoderm formation when the concentrations of the surrounding gap gene products increase significantly. It leaves open, however, how Kr is initially activated in a more posterior position of the embryo than hb (see above). In view of the mechanism established for hb activation, i.e. the affinity of the BCD-binding sites determines the threshold concentration above which a gene becomes activated in response to BCD, we looked for high-affinity binding sites within ÀT730-DNA. By direct comparison with the high-affinity binding sites of the hb promotor, none of the six BCD-binding sites bound BCD with higher affinity than the sites within the hb promotor (M. Hoch, unpublished). This finding suggests that the mechanism by which BCD sets the posterior limit of the expression domain in response to BCD may differ from the one observed with hb. In the case of hb, the BCD-binding sites are separated by about 140 nucleotides corresponding to one nucleosome equivalent (Driever et al., 1989), whereas the clustered BCD-binding sites within the Krl30-DNA are separated by maximum of 40 nucleotides. Thus, BCD-binding to these sites might be susceptible to cooperative interactions leading to a crowding effect and subsequently to positive interaction with the transcription machinery in more posterior regions of the BCD gradient than in the case of hb. In this model, each of the repressors provided by TLL, GT and KNI may interfere with the BCD cooperative interactions, and thus may interfere more severely with BCD-dependent activation than by blocking of individual activator sites. However, since we do not know yet how HB activates through the Kr730-element, cooperative interactions between BCD and HB cannot be excluded to account for the activation of Kr in a more posterior region of the embryo than hb.

Spatial control of kni expression by general activation and local repression through distinct c/s-acting elements

As with the terminal gap genes there are no genetically identified factors that could activate kni and gt expression. However, these factors are capable of activating kni (and possibly also gt) throughout the embryo. The spatial domain of kni expression, in the prospective abdominal region, is set by localized repressors which act on top of the activator. HB appears to be responsible for repression in the anterior half of the embryo whereas TLL acts as a repressor in the pole region. Deletion analysis indicates that the kni cis-acting region consists of at least three functional elements which respond to these factors (Figs 5, 6). One is needed for activation throughout the embryo, another is required to prevent expression in the anterior half in response to HB, while the third is required to prevent expression in the posterior pole region in response to TLL. The element required for setting the anterior border of kni expression contains HB-binding sites while the element required for setting the posterior border contains TLL-binding sites. Deletion of each of the two kni c/s-acting elements results in an expansion of reporter gene expression towards the site of the repressor gene product which is now unable to interfere with the m-acling element in transgenic wild-type embryos (Pankratz et al., 1992). In the absence of both repressor elements, gene expression can be obtained throughout the embryo. On top of this basic regulatory scheme, additional factors are required to further regulate kni wild-type gene expression. For example, genetic studies indicate that the genes Kr and gt which act in the kni adjacent domains have opposing effects on kni expression: KR may act as a co-activator of kni expression (Pankratz et al., 1989), while GT may provide a repressor function. The latter effect is relatively weak since no significant extension occurs in gt mutants, and ectopic expression of GT under the control of a heat-shock element does not result in the complete suppression of kni. In Kr;gt double mutant embryos kni expression is expanded and restored to wild-type levels.

Fig. 5.

Transformation constructs derived from the kni genomic region. Top line shows the positions of the various restriction sites utllized for making the DNA constructs used for fly transformation experiments (see Pankratz et al., 1992; for details). The hatched bar in the 9-kb kniSH kni rescue fragment denotes the transcribed region. This construct, which contains the coding region and approximately 6 kb of flanking sequences, fully rescues the kni mutant phenotype The thick lines on the left indicate the fragments used to make the lacZ indicator gene fusion constructs. The first three constructs contain the endogenous kni promoter. All others are fused to the basal heat shock promoter. The schematized embryos on the right illustrate the 13-gal expression patterns directed by the corresponding fusion constructs. Striped areas denote weaker staining. S, SalI; X, Xbal; B. BamHI; K, KpnI; P, PstV. N, Aral; D. Oral; Bg, BglI; H, HindIII; T, Taql;C, Clal.

Fig. 5.

Transformation constructs derived from the kni genomic region. Top line shows the positions of the various restriction sites utllized for making the DNA constructs used for fly transformation experiments (see Pankratz et al., 1992; for details). The hatched bar in the 9-kb kniSH kni rescue fragment denotes the transcribed region. This construct, which contains the coding region and approximately 6 kb of flanking sequences, fully rescues the kni mutant phenotype The thick lines on the left indicate the fragments used to make the lacZ indicator gene fusion constructs. The first three constructs contain the endogenous kni promoter. All others are fused to the basal heat shock promoter. The schematized embryos on the right illustrate the 13-gal expression patterns directed by the corresponding fusion constructs. Striped areas denote weaker staining. S, SalI; X, Xbal; B. BamHI; K, KpnI; P, PstV. N, Aral; D. Oral; Bg, BglI; H, HindIII; T, Taql;C, Clal.

Fig. 6.

Reporter gene expression directed by various kni upstream sequences in early embryos. Embryos are oriented with anterior to the left and dorsal up, and were stained with anti-B-gal antibody to visualize reporter gene expression (Pankratz et al., 1992). (a) KBg construct; the spatial pattern of the posterior expression resembles that of the kni transcripts, (b) KD construct; the spatial pattern is similar to the above KBg construct, (c) KR construct; the posterior domain is expanded to the posterior tip of the embryo. At late blastoderm (not shown) the posterior domain retracts from the posterior tip of the embryo; this is also seen with endogenous kni expression in embryos derived from torso females, (d) KP/RBg construct; the posterior domain is expanded to the anterior region of the embryo. As kni is also expressed at the anterior tip (see Fig.1), we do not know to what extent the strong signal in the very anterior region is due to the increased expression of the anterior domain, the anteriorly expanded posterior domain, or both. However, the ectopic expression between approximately 50%-75% egg length (posterior pole is 0%) is due to the anteriorly expanded posterior domain because that expression is stlll present in embryos derived from bed mutant mothers. The endogenous kni expression as well as the reporter gene expression of the kni 4.4 lac Z construct (M. J. Pankratz, unpublished result) in the anterior cap domain are absent in bed embryos, (e) KFn construct; reporter gene is expressed throughout the embryo. As in the KP/RBg construct above, this construct also directs strong expression in the anterior kni domain. Constructs PN and PX, which both delete a common 58-bp Kpn/Pst fragment, did not direct any reporter gene expression (not shown).

Fig. 6.

Reporter gene expression directed by various kni upstream sequences in early embryos. Embryos are oriented with anterior to the left and dorsal up, and were stained with anti-B-gal antibody to visualize reporter gene expression (Pankratz et al., 1992). (a) KBg construct; the spatial pattern of the posterior expression resembles that of the kni transcripts, (b) KD construct; the spatial pattern is similar to the above KBg construct, (c) KR construct; the posterior domain is expanded to the posterior tip of the embryo. At late blastoderm (not shown) the posterior domain retracts from the posterior tip of the embryo; this is also seen with endogenous kni expression in embryos derived from torso females, (d) KP/RBg construct; the posterior domain is expanded to the anterior region of the embryo. As kni is also expressed at the anterior tip (see Fig.1), we do not know to what extent the strong signal in the very anterior region is due to the increased expression of the anterior domain, the anteriorly expanded posterior domain, or both. However, the ectopic expression between approximately 50%-75% egg length (posterior pole is 0%) is due to the anteriorly expanded posterior domain because that expression is stlll present in embryos derived from bed mutant mothers. The endogenous kni expression as well as the reporter gene expression of the kni 4.4 lac Z construct (M. J. Pankratz, unpublished result) in the anterior cap domain are absent in bed embryos, (e) KFn construct; reporter gene is expressed throughout the embryo. As in the KP/RBg construct above, this construct also directs strong expression in the anterior kni domain. Constructs PN and PX, which both delete a common 58-bp Kpn/Pst fragment, did not direct any reporter gene expression (not shown).

As outlined in the formal genetic circuitry that regulates kni expression in the prospective abdominal region (see Fig. 2), there are two possibilities as to how KR and GT may work on kni: KR may act as a repressor of gt, thereby preventing GT repressor function on kni. In this case, Kr function on kni would be indirect and mediated by GT, as pro posed by Capovilla et al. (1992). We favour a model in which KR compensates in the wild-type situation the repressive effect of gt, and both interactions are of direct nature. In the absence of gt, KR will have no detectable regulatory effect on kni as the level of kni expression is highest due to the absence of its repressor. Furthermore, the level of kni expression will be unchanged in double mutant embryos since both the co-activator and the repressor are absent. This model is based on the finding of several KR binding sites in the cis-acting region of kni. Deletion of these binding sites results in the decrease of reporter gene expression (Pankratz et al., 1989) mediated by the deleted kni cis-acting sequences. In tissue culture, the fragment containing the KR sites is able to mediate reporter gene activation in response to co-transfected Kr gene DNA (F. Sauer, unpublished). Thus, in the wild-type situation Kr expression is likely to compensate for gt function on kni expression by two means. Firstly, it spatially limits gt expression and thereby inhibits its repressor function. Secondly, it restores wild-type expression levels, compensating for GT-dependent repression of kni.

Concentration-dependent activation or repression by KR

Based on genetic data, both HB and KR may act as positive and negative regulators of transcription on other genes of the zygotic segmentation hierarchy (see above). Both genes encode zinc finger-type proteins which form steep concentration gradients which extend significantly beyond the domains of transcript accumulation. We have examined the regulatory potential of such gradients by a series of co-transfection experiments in the Drosophila Schneider cell-line-system using the Kr gene as the source of a specific transcription factor. Different doses of Kr expression plasmid were tested for their ability to drive the already mentioned CAT reporter gene expression mediated by a single 11-bp Kr in vitro binding site, 5 ′ AAAAGGGTTAA-3 ′, common to several putative Kr target genes (Pankratz et al., 1989; Stanojevic et al., 1989; Treisman and Desplan, 1989) including kni (see above). When the KR expression plasmid was used in co-transfection experiments, both activation and repression of the reporter gene were observed (Fig. 7). Low amounts of the KR expression plasmid led to low levels of KR and to transcriptional activation, whereas high amounts of the expression plasmid led to high levels of KR which resulted in reporter gene repression (Licht et al., 1990; Zuo et al., 1990). The basal amount of reporter gene expression was left unchanged by co-transfected Kr DNA when the 11-bp-element was absent or replaced in the reporter gene construct, and no activation or repression occurs when plasmids containing the DNA-encoding transcription factors other than KR were used in the co-transfection experiments. Thus, KR is able to provide opposing regulatory effects in a concentration-dependent manner.

Fig. 7.

Concentration-dependent activation and repression by the Kr gene require different regions of the protein. Co-transfection experiments were performed using a constant amount of reporter gene plasmid DNA (1 μ g) and various concentrations of effector gene plasmid DNA, carrying either the entire Kr gene (a) or truncated (b, c) versions of it. The reporter gene DNA contains 5 copies of a KR in vitro binding site in front of a heterologous promoter which drives the expression of the basal level of the reporter gene CAT. (a) Reporter gene activity is dependent on KR: low dosages of Kr gene mediate activation while high dosages repress reporter gene expression, (b) Reporter gene expression in response to different dosages of a C-terminal truncated Kr protein. Note that this gene only mediates activation, (c) Deletion construct which leads to a N-terminal truncated KR only represses reporter gene expression (Sauer and Jackie, 1991).

Fig. 7.

Concentration-dependent activation and repression by the Kr gene require different regions of the protein. Co-transfection experiments were performed using a constant amount of reporter gene plasmid DNA (1 μ g) and various concentrations of effector gene plasmid DNA, carrying either the entire Kr gene (a) or truncated (b, c) versions of it. The reporter gene DNA contains 5 copies of a KR in vitro binding site in front of a heterologous promoter which drives the expression of the basal level of the reporter gene CAT. (a) Reporter gene activity is dependent on KR: low dosages of Kr gene mediate activation while high dosages repress reporter gene expression, (b) Reporter gene expression in response to different dosages of a C-terminal truncated Kr protein. Note that this gene only mediates activation, (c) Deletion construct which leads to a N-terminal truncated KR only represses reporter gene expression (Sauer and Jackie, 1991).

In order to determine whether transcriptional activation and repression by KR require specific portions of the protein, we examined truncated versions of the protein in the tissue culture assay (Fig. 7). High and low concentrations of KR lacking the 64 carboxy-terminal amino acids could not repress, but they were able to activate gene expression in a dose-dependent manner. Conversely, KR lacking the 116 amino-terminal aminoacids was not able to activate gene expression, but repression of gene expression was still observed. These results demonstrate that distinct portions of KR other than the DNA-binding domain are essential for gene activation and repression. Preliminary data suggest that KR is able to form a concentration-dependent homodimer, involving the C-terminal region of the protein as a dimerization domain, and that KR can activate transcription as a monomer, but loses this ability when forming homodimers at high concentrations of the protein (F. Sauer, unpublished). If these results obtained with tissue culture and in vitro experiments apply to the wild-type situation in the embryo, Kr could only act as a repressor in regions of high protein concentrations (roughly corresponding to the domain of transcript accumulation) and it could activate target genes only in regions which correspond to the domains of adjacent gap gene expression. However, Kr does not always act as an activator, as low concentrations of KR can act as transcriptional repressors within the kni expression domain (see below).

Repetitive patterns of pair-rule gene expression in response to gap gene-encoded local activators and repressors

The general scheme that emerged from the genetic and molecular analysis of pair-rule genes is that they are required for the establishment of alternating segmental units in the larva, and that they are expressed, apart from a few minor differences, in a basic outward appearance of seven evenly spaced stripes in the blastoderm embryo (Akam, 1987; Ingham, 1988; for reviews). The stripe pattern of expression of pair-rule genes such as fushi taraza (ftz) is controlled by a small upstream fragment that directs expression of all the stripes (Hiromi et al., 1985), and their formation depends on the preceding activity of other “primary” pair-rule genes. The work of Howard et al. (1988) suggested that this may not hold true for all the pair-rule genes. They studied mutations in the pair-rule gene hairy (h) that deleted the expression of specific stripes in the early embryo. Furthermore, studies on the regulatory region of another pair-rule gene, even-skipped {eve), showed a similar effect: DNA fragments derived from the eve upstream region could drive expression of a reporter gene in regions corresponding to particular eve stripes (Goto et al., 1989; Harding et al., 1989). These results strongly supported the view that h and eve are not responding to a preset periodic pattern, but that they respond instead to unique spatial cues that have been established along the axis of the embryo. In addition, proper ftz expression requires the preceding activities of eve and h whereas these two genes do not require ftz activity for their normal spatial control. Thus, despite similar expression patterns, the ways in which these patterns emerge are quite different. Both h and eve are primary pair-rule genes which can be placed at the receiving end of a long range of maternal positional information that is transmitted through localized activity of different gap genes.

We have used the h cis-regulatory upstream region (Pankratz et al., 1990a) to ask whether or not individual gap gene proteins are able to interact with h DNA. Different gap gene proteins were incubated with small subfragments of the h target DNA, and the protein/DNA complexes were then precipitated with antibodies directed against the different proteins. We found that the gap proteins bind, in various combinations and to different degrees, to a number of different fragments. The DNA fragments precipitated by this assay were then used as prospective control elements in front of a cis reporter gene construct. The precipitated DNA fragments acted as separate control elements giving rise to individual stripes in the transgenic embryos which are localized almost precisely in the position of the stripe of endogeneous h gene expression. Only the elements giving rise to stripe 3 and 4 of h expression could not be separated. Here we focus on the regulation of stripe number 6.

Genetic studies suggest that the two gap genes Kr and kni are mainly responsible for the control of h stripe 6 expression in the wild-type embryo, which occurs in an area of very low KR and high KNI concentrations (Pankratz et al., 1990a; Riddihough and Ish-Horowicz, 1991). In the absence of kni, h stripe 6 fails to be expressed. In the absence of Kr, reporter gene expression under the control of the h stripe 6 element is shifted towards anterior, and the area of expression expands. This suggests that in the wild-type situation, KR would negatively interact with the h stripe 6 element to set the anterior border of expression. However, there is a second indirect interaction of KR with the stripe 6 element, provided through KNI. As outlined above, KR co-activates kni gene expression. Thus, the absence of KR leads to a lower level of KNI expression, and therefore activation mediated by the h stripe 6 element occurs in a more anterior position of the embryo, where the threshold level of KNI would now be sufficient for activation. Due to the concomitant absence of the KR repressor, h stripe 6-mediated gene activation would continue throughout the region where KNI is high enough for activation. This mechanism implies that KR can repress h stripe 6-mediated gene expression even at low concentrations, and repression overrides activation. Stripe 6 expression would then be observed only in an area with a high enough KNI concentration to activate, but not enough KR to repress transcription. This model is consistent with a finding that ectopic Kr expression (via heat-shock control; see above) does not allow for gene expression mediated by the h stripe 6 element. At present, we do not know whether activation and repression by KNI and KR, respectively, is provided by competition for overlapping activator/repressor binding sites, as described above for the regulation of Kr expression in response to BCD and gap gene proteins, or whether it follows the mode of kni gene expression by individual activator/repressor elements. We note, however, that a direct link has been established between binding sites for gap regulatory proteins in the stripe 2 promotor element of eve and the expression of this stripe in the embryo (Small et al., 1991; Stanojevic et al., 1991). In this case, regulation of stripe gene expression occurs by overlapping activators (such as the HB and BCD) and repressors (such as GT and KR). Interplay between activators and repressors, as observed for Kr expression, is therefore used at the different levels of the segmentation gene hierarchy.

The metameric prepattern in the Drosophila blastoderm is generated through a series of DNA-protein interactions that is initiated by at least three transcription factor gradients in the egg. In the terminal regions, one (or several) unknown transcription factor(s) appear to be activated through a signal transduction cascade. Most likely, this results in a local and steep gradient of the activated form of the transcription factor which, at different threshold levels of activity, controls the expression of the terminal gap genes tll and hkb. Once activated, their gene products delimit the region of the embryo which gives rise to segmental patterns. This occurs by repression of the other known gap gene activities which cause the periodic pattern of pair-rule gene expression. In the anterior region of the preblastoderm embryo, two morphogen gradients are formed, either by diffusion of the protein from a local source (BCD), or by post-transcriptional regulation of evenly distributed mRNA under the control of nanos, to restrict the expression of maternal HB to the anterior half of the embryo. While BCD is likely to provide only an activating function that is limited to the anterior portion of the egg (with the central gap gene Kr likely to represent the posterior-most gene controlled by BCD activity), hb activity is likely to both activate and repress gene expression and reaches into the prospective abdominal region, as it marks the anterior border of gt expression within the posterior region of the embryo.

The mechanisms by which the spatially localized gap gene expression patterns are regulated are fundamentally different, hb expression occurs in response to the BCD gradient, likely to be regulated through the affinity of the BCD binding sites within the hb promotor: in regions where the BCD concentration is high enough to bind (through the site with the highest affinity within the promotor), hb becomes activated. The activation then leads to a pattern which extends from the anterior pole region, where bed concentration is highest, to the position in the gradient where BCD fails to bind. This mechanism may also account for other genes which are expressed in regions of the embryo where the slope of the BCD gradient is steep. In more posterior regions of the embryo, i.e. in the more shallow portions of the BCD gradient, this mechanism for gene activation might not be sufficient to provide sharp boundaries of the expression domain. For this reason, a different mechanism for BCD-dependent gene activation might be used, as in the case Kr. In view of the fact that none of the BCD binding sites of the Kr730-element shows a higher affinity binding site for BCD than in the hb promotor region, we favour a model in which cooperative interactions may lead to a crowding effect to activate Kr by BCD. The use of such a mechanism would be consistent with the finding that gap gene proteins such KNI, GT and TLL (which act as repressors to delimit the Kr expression domain) have target sequences that overlap the BCD binding sites within Kr73Q DNA: binding of either TLL, GT or KNI would interfere with BCD binding and thus would lead to repression. This effect would also interfere with cooperative interactions facilitating the filling of the additional BCD binding sites in the Kr regulatory region, and the binding of the repressor instead of the activator would shift Kr gene expression towards regions of higher BCD concentrations. In the extreme situation, the binding of several repressors should knock out BCD-dependent activation even if high BCD concentrations were present, as for example in the anterior pole position. The way in which HB activates and represses Kr gene expression in a concentration-dependent manner and how it may interfere with BCD-dependent activation has not yet been analysed. It appears, however, that the zinc finger-type proteins encoded by the gap genes are capable of providing these opposing functions, as exemplified in the case of KR.

An interplay between activators and repressors as observed for Kr expression is also observed for kni. However, the mechanism by which kni is spatially regulated in the posterior region again differs significantly from the one that regulates hb and Kr. Our results suggest that kni activation occurs throughout the entire embryo through the activity of a general, as yet unidentified activator. The spatial borders are then actively set through repression by factors that are regionally localized: HB, GT and TLL repress through separate modular ".s-acting elements that are different from the activation element within the kni regulatory region. Another feature, common to both Kr and kni regulation, is interesting to note. Regulatory inputs of the same kind are provided by more than one gene product. For example, Kr can be activated by HB when BCD is absent, and vice versa. Furthermore, TLL, HB and GT in the anterior and KNI, GT, TLL and possibly also HKB (Weigel et al.,1990) in the posterior, can provide local repression of Kr. Currently, we do not know whether these multiple activation and repressor functions represent a fine tuning system to determine and specify the precise position in the embryo where the target gene is active (or repressed), whether the regulatory inputs are redundant, or whether some of the regulatory pathways represent just evolutionary relics which are not decisive any longer in the Drosophila wild-type embryo.

We would like to thank our colleagues in the lab for comments on the manuscript. The work was supported by the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie.

Akam
,
M.
(
1987
).
The molecular basis for metameric pattern in the Drosophila embryo
.
Development
101
,
1
22
.
Berleth
,
T.
,
Burri
,
M.
,
Thoma
,
G.
,
Bopp
,
D.
,
Richstein
,
S.
,
Frigerio
,
G.
,
Noll
,
M.
and
Nüsslein-Volhard
,
C.
(
1988
).
The role of localization of bicoid RNA in organizing the anterior pattern of the Drosophila embryo
.
EMBOJ
.
1
,
1749
1756
.
Bronner
,
G.
and
Jackie
,
H.
(
1991
).
Control and function of terminal gap gene activity in the posterior pole region of the Drosophila embryo
.
Meeh. Dev
.
35
,
205
211
.
Campos-Ortega
,
J. A.
and
Hartenstein
,
V.
(
1985
).
The Embryonic Development of Drosophila melanogaster
.
Berlin
:
Springer
.
Capovilla
,
M.
,
Eldon
,
E. D.
and
Pirrotta
,
V.
(
1992
).
The giant gene of Drosophila encodes a b-ZIP DNA-binding protein that regulates the expression of other segmentation gap genes
.
Development
114
,
99112
.
Driever
,
W.
and
Nüsslein-Volhard
,
C.
(
1988a
).
A gradient of bicoid protein in Drosophila embryos
.
Cell
54
,
83
93
.
Driever
,
W.
and
Nüsslein-Volhard
,
C.
(
1988b
).
The bicoid protein determines postion in the Drosophila embryo in a concentration dependent manner
.
Cell
54
,
95
104
.
Driever
,
W.
and
Nüsslein-Volhard
,
C.
(
1989
).
The bicoid protein is a positive regulator of hunchback transcription in the early Drosophila embryo
.
Nature
331
,
138
143
.
Driever
,
W.
,
Thoma
,
G.
and
Nüsslein-Volhard
,
C.
(
1989
).
Determination of zygotic domains of gene expression in the Drosophila embryo by the affinity of binding sites for the bicoid morphogen
.
Nature
340
,
363
367
.
Eldon
,
E. D.
and
Pirrotta
,
V.
(
1991
).
Interactions of the Drosophila gap gene giant with maternal and zygotic pattern forming genes
.
Development
111
,
367
378
.
Frigerio
,
G.
,
Burri
,
M.
,
Bopp
,
D.
,
Baumgartner
,
S.
and
Noll
,
M.
(
1986
).
Structure of the segmentation gene paired and the Drosophila PRD gene set as part of a gene network
.
Cell
47
,
735
746
.
Frohnhôfer
,
H.
and
Nüsslein-Volhard
,
C.
(
1986
).
Organization of anterior pattern in the Drosophila embryo by the maternal gene bicoid
.
Nature
.
324
,
120
125
.
Gaul
,
U.
,
Seifert
,
E.
,
Schuh
,
R.
and
Jackie
,
H.
(
1987
).
Analysis of Krüppel protein distibution during early Drosophila development reveals post-transcriptional regulation
.
Cell
50
,
639
647
.
Gaul
,
U.
and
Jackie
,
H.
(
1987
).
Pole region-dependent repression of the Drosophila gap gene Krüppel by maternal gene products
.
Cell
51
,
549555
.
Gaul
,
U.
and
Jackie
,
H.
(
1989
).
Analysis of maternal effect mutant combination elucidates regulation and function of the overlap of Krüppel and hunchback gene expression in the blastoderm Drosophila embryo
.
Development
107
,
651
662
.
Gaul
,
U.
and
Jackie
,
H.
(
1990
).
Role of gap genes in early Drosophila development
.
Adv. Genetics
27
,
239
272
.
Gloor
,
H.
(
1954
).
Phânotypen der Heterozygoten bei der unvollstandig dominanten, homozygot lethalen Mutante Kr (-Krüppel) von Drosophila melanogaster
.
Arch. Jul. Klaus Stiftung
29
,
277
287
.
Gorman
,
C. M.
,
Moffat
,
L. F.
and
Howard
,
B. H.
(
1982
).
Recombinant genomes which express chloramphenicol acetyltransferase in mammalian cells
.
Mol. Cell. Biol
.
2
,
1044
1051
.
Goto
,
T.
,
Macdonald
,
P.
and
Maniatis
,
T.
(
1989
).
Early and late periodic patterns of even-skipped expression are controlled by distinct regulatory elements that respond to different spatial cues
.
Cell
57
,
413
422
.
Harding
,
K.
,
Hoey
,
T.
,
Warrior
,
R.
and
Levine
,
M.
(
1989
).
Auto-regulatory and gap response elements of the even-skipped promoter of Drosophila
.
EMBO J
.
8
,
1205
1212
.
Hiromi
,
Y.
,
Kuroiwa
,
A.
and
Gehring
,
W. J.
(
1985
).
Control elements of the Drosophila segmentation genefushi tara-u
.
Cell
43
,
603
613
.
Hoch
,
M.
,
Schroder
,
C.
,
Seifert
,
E.
and
Jackie
,
H.
(
1990
).
Cis-acting control elements for Krüppel expression
.
EMBO J
.
9
,
2587
2595
.
Hoch
,
M.
,
Seifert
,
E.
and
Jackie
,
H.
(
1991
).
Gene expression mediated by cis-acting sequences of the Krüppel gene in response to the Drosophila morphogens bicoid and hunchback
.
EMBO J
.
10
,
2267
2278
.
Hoch
,
M.
,
Gerwin
,
N.
,
Taubert
,
H.
and
Jackie
,
H.
(
1992
).
Competition for overlapping sites in the regulatory region of the Drosophila gene Krüppel
.
Science
256
,
94
97
.
Howard
,
K.
,
Ingham
,
P.
and
Rushlow
,
C.
(
1988
).
Region-specific alleles of the Drosophila segmentation gene hairy
.
Genes Dev
.
2
,
1037
1046
.
Hülskamp
,
M.
,
Schoder
,
C.
,
Pfeifle
,
C.
,
Jackie
,
H.
and
Tautz
,
D.
(
1989
).
Posterior segmentation of the Drosophila embryo in the absence of a maternal posterior organizer gene
.
Nature
338
,
629
632
.
Hülskamp
,
M.
,
Pfeifle
,
C.
and
Tautz
,
D.
(
1990
).
A morphogenetic gradient of hunchback protein organizes the expression of the gap genes Krüppel and knirps in the early Drosophila embryo
.
Nature
346
,
577580
.
Hülskamp
,
M.
and
Tautz
,
D.
(
1991
).
Gap genes and gradients - the logic behind the gaps
.
BioEssays
13
,
261
268
.
Irish
,
V. F.
,
Lehmann
,
R.
and
Akam
,
M.
(
1989
).
The Drosophila posterior-group gene nanos functions by repressing hunchback activity
.
Nature
338
,
646
648
.
Ingham
,
P. W.
(
1988
).
The molecular genetics of embryonic pattern formation in Drosophila
.
Nature
335
,
25
34
.
Jackie
,
H.
,
Tautz
,
D.
,
Schuh
,
R.
,
Seifert
,
E.
and
Lehmann
,
R.
(
1986
).
Cross-regulatory interactions among the gap genes of Drosophila
.
Nature
324
,
668
670
.
Klingler
,
M.
,
Erdélyi
,
M.
Szabad
,
J.
and
Nüsslein-Volhard
,
C.
(
1988
).
Function of torso in determining the terminal anlagen of the Drosophila embryo
.
Nature
335
,
275
277
.
Knipple
,
D. C.
,
Seifert
,
E.
,
Rosenberg
,
U.
,
Preiss
,
A.
and
Jackie
,
H.
(
1985
).
Spatial and temporal pattern of Krüppel gene expression in early Drosophila embryos
.
Nature
317
,
40
44
.
Kraut
,
R.
and
Levine
,
M.
(
1991a
).
Mutually repressive interactions between the gap genes giant and Krüppel define middle body regions of the Drosophila embryo
.
Development
111
,
611
622
.
Kraut
,
R.
and
Levine
,
M.
(
1991b
).
Spatial regulation of the gap gene giant during Drosophila development
.
Development
111
,
601
609
.
Lehmann
,
R.
and
Nüsslein-Volhard
,
C.
(
1987
).
hunchback, a gene required for segmentation of an anterior and posterior region of the Drosophila embryo
.
Dev. Biol
.
119
,
402
417
.
Levine
,
M.
and
Manley
,
J. L.
(
1989
).
Transcriptional repression of eukaryotic promotors
.
Cell
59
,
405
408
.
Licht
,
J. D.
,
Grossel
,
M. J.
,
Figge
,
J.
and
Hansen
,
U.
(
1990
).
Krüppel is a transcriptional repressor
.
Nature
346
,
76
79
.
Mohler
,
J.
,
Eldon
,
E. D.
and
Pirrotta
,
V.
(
1989
).
A novel spatial transcription pattern associated with the segmentation gene, giant, of Drosophila
.
EMBO J
.
8
,
1539
1548
.
Nauber
,
U.
,
Pankratz
,
M.
,
Kienlin
,
A.
,
Seifert
,
E.
,
Klemm
,
U.
and
Jackie
,
H.
(
1988
).
Abdominal segmentation of the Drosophila embryo requires a hormone receptor-like protein encoded by the gap gene knirps
.
Nature
336
,
489
492
.
Nüsslein-Volhard
,
C.
and
Wieschaus
,
E.
(
1980
).
Mutations affecting segment number and polarity in Drosophila
.
Nature
287
,
795
801
.
Nüsslein-Volhard
,
C.
,
Frohnhofer
,
H. G.
and
Lehmann
,
R
, (
1987
).
Determination of anterioposterior polarity in Drosophila
.
Science
238
,
1675
1681
.
Nüsslein-Volhard
,
C.
(
1991
).
Determination of the embryonic axes of Drosophila
.
Development Suppl
.
1
,
1
10
.
Pankratz
,
M. J.
,
Hoch
,
M.
,
Seifert
,
E.
and
Jackie
,
H.
(
1989
).
Krüppel requirement for knirps enhancement reflects overlapping gap gene activities in the Drosophila embryo
.
Nature
341
,
337
340
.
Pankratz
,
M. J.
,
Seifert
,
E.
,
Gerwin
,
N.
,
Billi
,
B.
,
Nauber
,
U.
and
Jackie
,
H.
(
1990a
).
Gradients of Krüppel and knirps gene products direct pairrule stripe patterning in the posterior region of the Drosophila embryo
.
Cell
61
,
309
317
.
Pankratz
,
M. J.
,
Hoch
,
M.
,
Rothe
,
M.
,
Nauber
,
U.
,
Gerwin
,
N.
,
Brönner
,
G.
,
Gôrlich
,
K.
and
Jackie
,
H.
(
1990b
).
DNA-binding proteins in Drosophila development
.
Nucleic Acids and Molecular Biology
, Vol
4
(ed.
F.
Eckstein
and
D. M. J.
Lilley
), pp.
185
204
.
Berlin, Heidelberg
:
Springer-Verlag
.
Pankratz
,
M. J.
and
Jackie
,
H.
(
1990
).
Making stripes in the Drosophila embryo
.
Trends Genet
.
6
,
287
292
.
Pankratz
,
M. J.
,
Busch
,
M.
,
Hoch
,
M.
,
Seifert
,
E.
and
Jackie
,
H.
(
1992
).
Spatial control of the gap gene knirps in the Drosophila embryo by the posterior morphogen system
.
Science
255
,
986
989
.
Petschek
,
J.
,
Perrinton
,
N.
and
Mahowald
,
A. P.
(
1987
).
Region specific effects in 1(1) giant embryos of Drosophila
.
Dev. Biol
.
119
,
177
189
.
Pignoni
,
F.
,
BaldarelU
,
R. M.
,
Steingrimsson
,
E.
,
Diaz
,
R. J.
,
Patapoutian
,
A.
,
Merriam
,
J. R.
and
Lengyel
,
J. A.
(
1990
).
The Drosophila gene tailless is expressed at the embryonic termini and is a member of the steroid receptor superfamily
.
Cell
62
,
151
163
.
Preiss
,
A.
,
Rosenberg
,
U. B.
,
Kienlin
,
A.
,
Seifert
,
E.
and
Jackie
,
H.
(
1985
).
Molecular genetics of Krüppel, a gene required for segmentation of the Drosophila embryo
.
Nature
313
,
27
33
.
Ptashne
,
M.
(
1986
).
Gene regulation by proteins acting nearby and at a distance
.
Nature
322
,
697
701
.
Ptashne
,
M.
(
1988
).
How eukaryotic transcriptional activators work
.
Nature
335
,
683
689
.
Riddibough
,
G.
and
Ish-Horowicz
,
D.
(
1991
)
Individual stripe regulatory elements in the Drosophila hairy promoter respond to maternal, gap and pair-rule genes
.
Genes Dev
.
5
,
840
854
.
Rosenberg
,
U.
,
Schroder
,
C.
,
Preiss
,
A.
,
Kienlin
,
A.
,
Coté
,
S.
,
Riede
,
I.
and
Jackie
,
H.
(
1986
).
Structural homology of the product of the Drosophila Krüppel gene with Xenopus transcription factor IIIA
.
Nature
319
,
336
339
.
Rothe
,
M.
,
Nauber
,
U.
and
Jackie
,
H.
(
1989
).
Three hormone receptor-like Drosophila genes encode an identical DNA-binding finger
.
EMBO J
.
8
,
3087
3094
.
Sauer
,
F.
and
Jackie
,
H.
(
1991
).
Concentration-dependent transcriptional activation or repression by Krüppel from a single binding site
.
Nature
353
,
563
566
.
Schroder
,
C.
,
Tautz
,
D.
,
Seifert
,
E.
and
Jackie
,
H.
(
1988
).
Differential regulation of the two transcripts from the Drosophila gap segmentation gene hunchback
.
EMBO J
.
7
,
2882
2887
.
Small
,
S.
,
Kraut
,
R.
,
Hoey
,
T.
,
Warrior
,
R.
and
Levine
,
M.
(
1991
).
Transcriptional regulation of a pair-rule stripe in Drosophila
.
Genes Dev
.
5
,
827
839
.
Sprenger
,
F.
,
Stevens
,
L. M.
and
Nüsslein-Volhard
,
C.
(
1989
).
The Drosophila gene torso encodes a putative receptor tyrosine kinase
.
Nature
338
,
478
483
.
Stanojevic
,
D.
,
Hoey
,
T.
and
Levine
M.
(
1989
).
Sequence-specific DNA-binding activities of the gap proteins encoded by hunchback and Krüppel in Drosophila
.
Nature
341
,
331
335
.
Stanojevic
,
D.
,
Small
,
S.
and
Levine
,
M.
(
1991
).
Regulation of a segmentation stripe by overlapping activators and repressors in the Drosophila embryo
.
Science
254
,
1385
1387
.
Steingrimsson
,
E.
,
Pignoni
,
F.
,
Liaw
,
G. J.
and
Lengyel
,
J. A.
(
1991
).
Dual role of the Drosophila pattern gene tailless in embryonic termini
.
Science
254
,
418
420
.
Stevens
,
L. M.
,
Frohnhofer
,
H. G.
,
Klingler
,
M.
and
Nüsslein-Volhard
,
C.
(
1990
).
Localized requirement for torsolike expression in follicle cells for the development of terminal anlagen of the Drosophila embryo
.
Nature
346
,
660
663
.
St. Johnston
,
D.
,
Driever
,
W.
,
Berieth
,
T.
,
Richstein
,
S.
and
Nüsslein-Volhard
,
C.
(
1989
).
Multiple steps in the localization of bicoid RNA to the anterior pole of the Drosophila oocyte
.
Development (Suppl.)
107
,
1319
.
St. Johnston
,
D.
and
Nüsslein-Volhard
,
C.
(
1992
).
The origin of pattern and polarity in the Drosophila embryo
.
Cell
68
,
201
219
.
Strecker
,
T. R.
,
Merriam
,
J. R.
and
Lengyel
,
J. A.
(
1988
).
Graded requirement for the zygotic terminal gene, tailless, in the brain and tail region of the Drosophila embryo
.
Development
102
,
721
734
.
Struhl
,
G.
,
Struhl
,
K.
and
Macdonald
,
P. M.
(
1989
).
The gradient morphogen bicoid is a concentration-dependent transcriptional activator
.
Cell
57
,
1259
1273
.
Struhl
,
G.
,
Johnston
,
P.
and
Lawrence
,
P. A.
(
1992
).
Control of Drosophila body pattern by the hunchback morphogen gradient
.
Cell
69
,
237
249
.
Tautz
,
D.
,
Lehmann
,
R.
,
Schniirch
,
H.
,
Schuh
,
R.
,
Seifert
,
E.
,
Kienlin
,
A.
,
Jones
,
K.
and
Jackie
,
H.
(
1987
).
Finger protein of novel structure encoded by hunchback, a second member of the gap class of Drosophila segmentation genes
.
Nature
327
,
383
389
.
Tautz
,
D.
(
1988
).
Regulation of the Drosophila segmentation gene hunchback by two maternal morphogenetic centres
.
Nature
332
,
281
284
.
Treisman
,
J.
and
Desplan
,
C.
(
1989
).
The products of the Drosophila gap gene hunchback and Kriippel bind to the hunchback promoters
.
Nature
341
,
335
337
.
Wang
,
C.
and
Lehmann
,
R.
(
1991
).
Nanos is the localized posterior determinant in Drosophila
.
Cell
66
,
637
647
.
Weigel
,
D.
,
Jiirgens
,
G.
,
Klingler
,
M.
and
Jackie
,
H.
(
1990
).
Two gap genes mediate maternal terminal pattern formation in Drosophila
.
Science
248
,
495
498
.
Wieschaus
,
E.
,
Nüsslein-Volhard
,
C.
and
Kluding
,
H.
(
1984
).
Kriippel, a gene whose activity is required early in the zygotic genome for normal embryonic segmentation
.
Dev. Biol
.
104
,
172
186
.
Zuo
,
P.
,
Stanojevic
,
D.
,
Colgan
,
J.
,
Han
,
K.
,
Levine
,
M.
and
Manley
,
J. L.
(
1990
).
Activation and repression of transcription by the gap proteins hunchback and Kriippel in cultured Drosophila cells
.
Genes Dev
.
5
,
254264
.