Although many of the genes that pattern the segmented body plan of the Drosophila embryo are known, there remains much to learn in terms of how these genes and their products interact with one another. Like many of these gene products, the protein encoded by the pair-rule gene odd-skipped (Odd) is a DNA-binding transcription factor. Genetic experiments have suggested several candidate target genes for Odd, all of which appear to be negatively regulated. Here we use pulses of ectopic Odd expression to test the response of these and other segmentation genes. The results are complex, indicating that Odd is capable of repressing some genes wherever and whenever Odd is expressed, while the ability to repress others is temporally or spatially restricted. Moreover, one target gene, fushi tarazu, is both repressed and activated by Odd, the outcome depending upon the stage of development. These results indicate that the activity of Odd is highly dependent upon the presence of cofactors and/or overriding inhibitors. Based on these results, and the segmental phenotypes generated by ectopic Odd, we suggest a number of new roles for Odd in the patterning of embryonic segments. These include gap-, pair-rule- and segment polarity-type functions.

Early Drosophila embryogenesis is optimized for rapid subdivision of the embryo into segmentally repeated units. This process is controlled by a hierarchy of maternally and zygotically expressed segmentation genes. The zygotic segmentation genes have been classified into three groups based on their mutant phenotypes, gap genes, pair-rule genes and segment polarity genes (Nusslein-Volhard and Wieschaus, 1980). Gap genes define regions that span several contiguous segments, pair-rule genes control homologous regions within every other segment and segment polarity genes control homologous regions within every segment (for a review, see Ingham, 1988).

The majority of these genes encode transcription factors. They interact combinatorially (Gergen and Wieschuas, 1985; DiNardo and O’Farrell, 1987; Ingham et al., 1988a; Manoukian and Krause, 1992, 1993), generating expression patterns of increasing complexity. The outcome is a segmental blueprint in the form of segment polarity gene expression patterns. Shortly thereafter, the first physical signs of segmentation, epidermal indentations referred to as parasegmental grooves, are observed (Martinez-Arias and Lawrence, 1985). These subdivide the embryo along the anterior-posterior axis into segment-wide metameres called parasegments (Martinez-Arias and Lawrence, 1985; Lawrence et al., 1987). Parasegments are shifted anteriorly relative to the segmental units that will form later.

odd-skipped (odd) is a member of the pair-rule class of segmentation genes. It encodes a zinc-finger-containing transcription factor (Odd), which is expressed first in stripes that span the even-numbered parasegments, and later in narrow stripes in the middle of both even- and odd-numbered parasegments (Coulter et al., 1990; Manoukian and Krause, 1993). Mutations in the odd gene cause loss of portions of even-numbered parasegments and partial replacement by mirror-image duplications of flanking regions (Coulter and Wieschaus, 1988). These patterning defects have been explained, in part, by alterations in the expression patterns of three putative target genes, fushi tarazu (ftz), engrailed (en) and wingless (wg) (DiNardo and O’Farrell, 1987; Mullen and DiNardo, 1995). Similar pattern rearrangements have been noted in HSeve and HSrun embryos (Manoukian and Krause, 1992, 1993), where ectopic expression of eve and run led to repression of odd. Taken together, these studies suggest that the main function of Odd may be to repress expression of the ftz, en and wg genes in the middle regions of even-numbered parasegments. The paired gene may also be negatively regulated by odd in these regions (Baumgartner and Noll, 1990).

Genetic analyses such as those just described are important as they determine the positive or negative relationships between genes. However, they do not distinguish between direct and indirect interactions. For example, Odd may bind to the promoters or promoter-binding proteins of the ftz and en genes to directly repress their transcription. Alternatively, Odd may act indirectly by regulating the expression of intermediary genes whose products then act on the ftz and en promoters. In the case of en, for example, Odd could act as a direct repressor of the en promoter, or alternatively, could act indirectly by repressing ftz, since ftz is required for en activation (Mullen and DiNardo, 1995).

A further complication in the interpretation of these studies is that segmentation gene expression patterns are extremely dynamic, resulting in protein combinations that change rapidly within a particular cell. Because the segmentation proteins are highly dependent upon combinatorial interactions, their effects on target genes may change within the span of a few minutes. At one time, a protein could function as an activator of a particular target gene and then, with the addition or removal of another protein, could be neutralized, or could even become a repressor. An example of this is the pair-rule protein Even-skipped, which changes from an activator to a repressor of the ftz gene within a period of 10-20 minutes (Manoukian and Krause, 1992). Such changes in activity are difficult to elucidate solely by the observation of mutant embryos, since these would exhibit the cumulative effects of both regulatory interactions.

For these reasons, we have been using an alternative approach to help distinguish between direct and indirect gene interactions, and to dissect changing gene interactions that are temporally or spatially regulated (Manoukian and Krause, 1992, 1993). These studies make use of heat-inducible segmentation gene constructs to enable activation of the genes in developing embryos. Expression is induced at the desired stage using short (2-6 minutes) heat pulses and changes in putative target gene expression patterns are monitored at 5 minute intervals thereafter. Typically, we find that there are two temporal windows of response within 40 minutes of the heat pulse. Genes that we presume to be direct targets begin to respond within 15 minutes of the heat pulse, while those that require the expression and function of intermediary gene products show delayed responses beginning at 30 and then 45 minutes after trans-gene induction. Our definition of direct is that the interaction occurs in the absence of intermediary gene regulation and the subsequent synthesis or degradation of these intermediary gene products. However, they need not involve direct molecular contact between the induced regulator and target gene promoters.

Using this kinetic approach, we confirm interactions suggested by previous genetic analyses as well as identifying several new targets of Odd. The regulation of these targets changes rapidly in both a temporal and spatial fashion. For one target, Odd switches from an activator to a repressor. Taken together, our data show that Odd plays major roles in determining both the size and polarity of even-numbered as well as odd-numbered parasegments. Studies such as this indicate that current models of the segmentation gene hierarchy are somewhat oversimplified and need further investigation.

pHSodd construction and fly transformation

The vector pHSodd was constructed as follows: the odd full-length cDNA (1.95 kb) was excised as a NotI-EcoRV fragment from the clone podd 7.4 II-F (Coulter et al., 1990), obtained from D. Coulter. This fragment was inserted between the NotI and StuI cloning sites of the P-element transformation vector, pCaSpeR-HS (Thummel and Pirrotta, 1992). This vector has a white selectable marker and allows expression of the cloned cDNA under control of the hsp70 promoter. HSodd transgenic flies were obtained by P-element-mediated germline transformation (Spradling and Rubin 1982). The pHSodd vector was injected into w; P[Δ2-3]/TM6 embryos, which contain a stable integrated source of transposase (Robertson et al., 1988). Two independent transformant lines were obtained, HSodd1 and HSodd2, both carrying a single copy of the pHSodd vector on the second chromosome. HSodd1 is homozygous lethal and balanced over SM5. HSodd2 is homozygous viable. Two more lines, HSodd3 and HSodd9, each with a single P-element insertion on the third chromosome, were obtained from the HSodd1/SM5 line by jump-start transposition (Engels et al., 1987). HSodd3 is homozygous viable and HSodd9 is kept as a balanced stock over TM3. HSodd; odd lines were generated using the third chromosome HSodd3 line crossed to the odd mutant lines oddIIId and odd7L (Coulter and Wieschaus, 1988). odd mutant chromosomes were balanced over a CyO balancer marked with a hb-lacZ reporter gene.

Heat-shock protocol and cuticle preparation

For cuticle preparations, embryos were collected on apple juice/agar plates for 20 minutes. After aging at 25°C for the indicated times (see figure legends), embryos were washed off the plates into small plastic cylinders containing a nylon mesh fastened to the bottom. Heat shocks were performed by immersion of the cylinder in a 36.5°C water bath. For the precise timing of different cuticular phenotypes, five collections of 20 minutes each were made and aged prior to heat shock such that each collection represented a different 20 minute interval, each separated by 10 minutes, between 2:10 (2 hours, 10 minutes) and 3:10 AEL (i.e. 2:10-2:30 AEL; 2:20-2:40; 2:30-2:50; etc.). After heat shock, embryos were rinsed with 25°C water, blotted dry, transferred to a microscope slide and covered with a thin layer of halocarbon oil. Properly staged embryos (approximately 100) were then selected under a dissecting microscope and transferred to an apple juice plate for further aging. After 24 to 30 hours at 25°C, hatched larvae were transferred into methanol. Unhatched larvae were dechorionated in bleach, the vitelline membranes removed by vigorous agitation in methanol/heptane (1:1) and then washed once with methanol. Hatched and unhatched larvae were pooled and washed several times in methanol prior to being deposited on a slide. After evaporation of the methanol, larvae were cleared in Hoyer’s medium/lactic acid (4:1) for 24 hours at 65°C (Weischaus and Nusslein-Volhard, 1986).

In situ hybridization

For in situ hybridization, embryos were collected on apple juice/agar plates for 30 minutes or 1 hour as indicated. After appropriate aging at 25°C, the embryos were heat shocked as described above. After heat shock, embryos were allowed to recover for different periods of time and then fixed in 4% formaldehyde as described by Lehmann and Tautz (1994) in protocol 1A.

The detection of transcripts in embryos was achieved by whole-mount in situ hybridization using digoxigenin-11-dUTP-labeled probes (Tautz and Pfeifle, 1989), essentially as described in protocol 2 by Lehmann and Tautz (1994), except that the hybridization was performed at 48°C. Embryos were mounted in 70% glycerol, 1× PBS. In experiments involving odd mutant embryos, homozygous mutants were identified by double labeling (Manoukian and Krause, 1992) with odd mRNA probes and anti-β-galactosidase antibodies. Expression of β-galactosidase indicates the presence of the wild-type chromosome (marked with hb-lacZ).

Images were captured using a Sony XC275 CCD camera and Northern Exposure software. Figures were compiled using Adobe Photoshop.

Ectopic expression of Odd in transgenic embryos

A heat-inducible odd vector (pHSodd) was generated by placing the odd cDNA under control of the hsp 70 heat-shock gene promoter in the P-element vector pCaSpeR (Thummel and Pirrotta, 1992). Two transformant lines harboring this construct were isolated, HSodd1 and HSodd2, each with single insertions mapped to the second chromosome. Additional lines were generated by jumping the P-element construct in HSodd1 to new chromosomal locations (see Materials and Methods). All lines gave similar results. The majority of results presented here were obtained with the HSodd2 line.

HSodd2 embryos were heat pulsed at 36°C for different lengths of time (2-10 minutes) in order to find conditions that would induce physiological levels of odd expression. As shown in Fig. 1B, a 4 minute heat shock was sufficient to induce ectopic transcript levels similar to those normally expressed in odd stripes. Note that the underlying pattern of six stripes normally seen at this stage is still visible (although darker) and that the intensity of ectopic expression is similar to or lower than the levels seen in wild-type embryo stripes (Fig. 1A). With a 6 minute heat shock (Fig. 1C), levels of transgene expression were substantially higher, making the endogenous striped expression pattern difficult to make out. Since these levels are probably higher than endogenous levels of expression, the remaining data were obtained using 4 minute inductions. Oregon R embryos heat shocked for the same duration were used as controls.

Fig. 1.

odd-skipped expression in HSodd embryos. odd-skipped transcripts were detected by in situ hybridization in wild-type (A) and HSodd embryos heat shocked for 4 minutes (B) or 6 minutes (C) between 2:30 and 2:50 AEL. Embryos in B and C were fixed 15 minutes after the beginning of heat shock. The wild-type embryo (A) was fixed at a similar stage. Note that endogenous odd stripes are still visible in B and that the intensity of interstripe staining in B is similar to or less than the intensity of normal odd stripes in the WT embryo shown in A.

Fig. 1.

odd-skipped expression in HSodd embryos. odd-skipped transcripts were detected by in situ hybridization in wild-type (A) and HSodd embryos heat shocked for 4 minutes (B) or 6 minutes (C) between 2:30 and 2:50 AEL. Embryos in B and C were fixed 15 minutes after the beginning of heat shock. The wild-type embryo (A) was fixed at a similar stage. Note that endogenous odd stripes are still visible in B and that the intensity of interstripe staining in B is similar to or less than the intensity of normal odd stripes in the WT embryo shown in A.

Ubiquitous expression of odd causes three cuticular phenotypes

HSodd embryos were heat pulsed at different stages of embryogenesis in order to determine when segmental patterning was most susceptible to ectopic Odd expression and to determine how segmental patterning would be affected. Segmental patterning was assessed by examination of larval cuticles prepared at the completion of embryogenesis (Fig. 2). Segmental defects were found to occur at the highest penetrance when heat pulses were administered between 2:10 and 3:10 (stages 5-7) after egg laying (AEL). Factoring in the 10-20 minute delay required for heat-inducible transcripts to accumulate, this interval closely approximates the window of endogenous odd expression. Three major segmental phenotypes were observed. These include a phenotype with head defects only (Fig. 2B), a pair-rule phenotype (Fig. 2C) and a pair-rule phenotype restricted to the dorsal half of the embryo (Fig. 2D). The stage at which each phenotype was induced was determined by administering heat shocks to collections of embryos that were collected over 20 minute intervals and aged appropriately (i.e., aged 2:10-2:30, 2:20-2:40 etc at the time of heat shock). The correlation between when heat shocks were administered and the frequency with which each phenotype was observed is summarized in Table 1. The earliest phenotype observed (‘Head’) is a specific loss of head regions beginning just anterior to the T1 ‘beard’ up to the maxillary sense organs (Fig. 2B). A cone-shaped indentation replaces the missing structures. Central portions of the mouth hooks such as the ectostomal sclerite are also malformed or missing. This phenotype prevails when Odd is induced between 2:10 and 2:30 AEL. Head defects, usually more severe, are also observed following later Odd inductions, but are not indicated in Table 1 since they are also associated with other defects. Also seen at the 2:10-2:30 interval, but at very low frequency (6%: grouped within ‘Other’ category), is a deletion of even-numbered thoracic and odd-numbered abdominal denticle belts, as seen in ftz mutant cuticles. However, this phenotype was also observed in heat-shocked Oregon R controls, albeit at a lower frequency (∼2%).

Table 1.

Correlation between cuticular phenotypes and times of odd induction

Correlation between cuticular phenotypes and times of odd induction
Correlation between cuticular phenotypes and times of odd induction
Fig. 2.

Cuticular phenotypes caused by ectopic odd. Dark-field photomicrographs show cuticular phenotypes that arise from HSodd2 embryos that received a 4 minute heat shock between 2:10 and 3:10 AEL. (A) Wild-type cuticle: thoracic segments are numbered T1 to T3, and abdominal segments A1 to A8. (B) ‘Head’ phenotype: this phenotype is most prevalent when Odd is induced between 2:10 and 2:30 AEL. Defects are limited to a region (indicated with arrow) just anterior to the ‘beard’ in T1. Portions of the mouth hooks are missing or malformed. Missing regions are often replaced with a funnel-shaped hole. (C) Pair-rule phenotype:this phenotype is most frequently observed when HSodd embryos are heat shocked between 2:40 and 2:50 AEL. Posteriorly, the filzkorper (fz) and spiracles (sp) are present, although not always well formed, but the stigmatophore is not. Fused denticle belts are indicated. (D) Bean-shape cuticle: this phenotype arises most frequently when heat shocks are administered between 2:50 and 3:10 AEL. Pair-rule deletions (indicated with arrows) are observed in the dorsal half of the embryo. (E) Higher magnification of a wild-type cuticle showing T3 and A1 ventral denticle belts. Note that the T3 denticle belt contains two to three rows of tiny denticles, while the denticles in A1 are much larger. (F) The same region as in E in a HSodd cuticle. The denticle belts of T2 and A1 are fused, with the intervening regions deleted. (G) Schematic representation of a wild-type larval cuticle (top) showing regions missing in HSodd pair-rule mutants as gray boxes. The resulting phenotype is shown below. Also indicated are posterior defects. Head defects are not shown.

Fig. 2.

Cuticular phenotypes caused by ectopic odd. Dark-field photomicrographs show cuticular phenotypes that arise from HSodd2 embryos that received a 4 minute heat shock between 2:10 and 3:10 AEL. (A) Wild-type cuticle: thoracic segments are numbered T1 to T3, and abdominal segments A1 to A8. (B) ‘Head’ phenotype: this phenotype is most prevalent when Odd is induced between 2:10 and 2:30 AEL. Defects are limited to a region (indicated with arrow) just anterior to the ‘beard’ in T1. Portions of the mouth hooks are missing or malformed. Missing regions are often replaced with a funnel-shaped hole. (C) Pair-rule phenotype:this phenotype is most frequently observed when HSodd embryos are heat shocked between 2:40 and 2:50 AEL. Posteriorly, the filzkorper (fz) and spiracles (sp) are present, although not always well formed, but the stigmatophore is not. Fused denticle belts are indicated. (D) Bean-shape cuticle: this phenotype arises most frequently when heat shocks are administered between 2:50 and 3:10 AEL. Pair-rule deletions (indicated with arrows) are observed in the dorsal half of the embryo. (E) Higher magnification of a wild-type cuticle showing T3 and A1 ventral denticle belts. Note that the T3 denticle belt contains two to three rows of tiny denticles, while the denticles in A1 are much larger. (F) The same region as in E in a HSodd cuticle. The denticle belts of T2 and A1 are fused, with the intervening regions deleted. (G) Schematic representation of a wild-type larval cuticle (top) showing regions missing in HSodd pair-rule mutants as gray boxes. The resulting phenotype is shown below. Also indicated are posterior defects. Head defects are not shown.

The second phenotype observed is a classic pair-rule phenotype, with alternate segmental regions deleted (Fig. 2C).

This phenotype prevails when heat pulses are provided between 2:30 and 2:50 AEL. Closer examination of the fused and partially fused segments reveals a phasing of the deleted regions that is neither segmental nor parasegmental (Fig. 2C,F,G). On the ventral side, composite denticle belts are observed (T1/2, T3/A1, A2/3, A4/5 and A6/7), with the regions in between deleted. Most of the remaining regions are composed of even-numbered parasegments, which is where odd is normally expressed. This phenotype is complementary to the odd mutant phenotype, in which the regions remaining are primarily derived from odd-numbered parasegments (Coulter et al., 1990). Also complementary to the odd mutant phenotype (Coulter and Wieschaus, 1988) is the frequency at which the different regions are deleted. The most frequently fused segments in HSodd embryos are A4 and A5, which are fused in 100% of the intermediate pair-rule cuticles examined. Sensitivity then decreases away from A4/A5 towards the extremities of the embryo (A6/A7>A3/A4>T3/A1>T1/T2).

The third phenotype observed (‘bean-shape’) is a pair-wise deletion of segmental regions identical to the pair-rule phenotype just described, but limited to the dorsal half of the embryo (Fig. 2D). This phenotype prevails when heat shocks are administered between 2:50 and 3:10 AEL.

Effects of ectopic Odd on late en and wg expression

Expression patterns of the segment polarity genes en and wg were monitored about a third of the way through embryogenesis (stage 11) in order to help elucidate the mechanisms underlying the HSodd pair-rule phenotypes (Fig. 3). These two genes are normally expressed in 14 equally spaced stripes flanking either the posterior (en, Fig. 3A) or anterior (wg, Fig. 3D) edges of parasegmental grooves. In Hsodd adjacent stripes is no longer uniform. For example, en stripes 1 and 2, 3 and 4, and so on are closer together than normal, giving a coupled appearance. The same is true for wg stripes 0 and 1, 2 and 3, and so on. In approximately 30% of these embryos, the even-numbered en stripes and odd-numbered wg stripes are partially missing. The most severely affected are en stripe 10 and wg stripe 9. This is consistent with the observation that denticle belt fusions occur with the highest frequency between A4 and A5 (parasegments 9 and 10). In embryos fixed at progressively later stages, these stripes were missing at increasingly higher frequencies. Loss of wg expression appeared to precede that of en. Based on these observations and the cuticular phenotypes, we surmise that the deleted segmental regions include each set of unstable en and wg stripes, and extend approximately half way to the en embryos, the spacing and/or intensity of these stripes are altered. Two different patterns that show pair-wise alterations in stripe spacing or intensity are seen for each gene. These expression patterns (described below) correlate with the pair-rule and bean-shape cuticular phenotypes based upon the timing of the heat shocks that produced them.

Fig. 3.

Late expression of the segment polarity genes engrailed and wingless. Embryos were stained for en (A-C) or wg (D-F) transcripts by in situ hybridization. (A,D) Wild-type embryos fixed at 5:00-5:30 AEL (stage 10). (B,C,E,F) HSodd embryos, heat shocked for 4 minutes, and fixed at the same stage as the embryos in A and D. (B,E) Embryos heat shocked at 2:40-2:50 AEL: even-numbered en stripes and odd-numbered wg stripes appear to be shifted anteriorly and are often weaker or partially missing. (C,F) Embryos heat shocked at 2:50-3:10: spacing of stripes is more normal, but even-numbered en stripes and odd-numbered wg stripes are weak or missing, particularly dorsolaterally. The arrows highlight en stripe 10 and the adjacent wg stripe 9, which are generally the first to be lost.

Fig. 3.

Late expression of the segment polarity genes engrailed and wingless. Embryos were stained for en (A-C) or wg (D-F) transcripts by in situ hybridization. (A,D) Wild-type embryos fixed at 5:00-5:30 AEL (stage 10). (B,C,E,F) HSodd embryos, heat shocked for 4 minutes, and fixed at the same stage as the embryos in A and D. (B,E) Embryos heat shocked at 2:40-2:50 AEL: even-numbered en stripes and odd-numbered wg stripes appear to be shifted anteriorly and are often weaker or partially missing. (C,F) Embryos heat shocked at 2:50-3:10: spacing of stripes is more normal, but even-numbered en stripes and odd-numbered wg stripes are weak or missing, particularly dorsolaterally. The arrows highlight en stripe 10 and the adjacent wg stripe 9, which are generally the first to be lost.

The most frequently observed en and wg patterns originating from the heat-shock interval that yields the pair-rule phenotype are shown in Fig. 3B (en) and E (wg). For both genes, the spacing between and wg stripes in front and behind. These deletions are out of register with both segmental and parasegmental boundaries.

The temporal window giving rise to the bean-shaped phenotype generates the en and wg patterns shown in Fig. 3C,F. In these embryos, the spacing between en and wg stripes is usually normal. However, even-numbered stripes of en and odd-numbered stripes of wg are weaker or missing at their dorsal edges, consistent with the dorsal deletions seen in the bean-shaped cuticles. As with the pair-rule patterns, en stripe 10 and wg stripe 9 often show defects that are more severe than the others.

Effects of Odd on pair-rule gene expression

To address the nature of the genetic alterations underlying the HSodd phenotype, we analyzed the expression of segmentation genes expressed during the Odd-sensitive period (2:10-3:10 AEL). In previous studies, we found that genes responding directly to ectopic gene expression show peak changes in expression patterns within 25 minutes of the start of heat shock (Manoukian and Krause, 1992, 1993). Genes that require an intermediary gene product to be transcribed and translated, or alternatively repressed and degraded, in order to respond to the induced protein only begin to respond 30-35 minutes post-heat shock, and peak at 40-45 minutes post-heat shock. Fig. 4 shows pair-rule gene expression patterns in embryos fixed 25 minutes after the beginning of a 4 minute heat shock. Hence, these effects likely reflect direct transcriptional responses to ectopic Odd expression. These Odd-induced patterns (center column) are compared to the corresponding pair-rule gene expression patterns in wild-type (left) and odd (right) embryos. The changes in expression patterns shown in HSodd embryos were not observed in Oregon R control embryos heat shocked under identical conditions (not shown).

Fig. 4.

Effect of ectopic odd-skipped on pair-rule genes. RNA expression patterns are shown, from top to bottom, for the pair-rule genes odd-paired (opa; A-C), fushi tarazu (ftz; D-F), hairy (h; G-I), runt (run; J-L), even-skipped (eve; M-O), and paired (prd; P-R) in stage 5-7 embryos. Embryos on the left (A,D,G,J,M,P) are wild-type, embryos in the middle (B,E,H,K,N,Q) are Hsodd embryos fixed 25 minutes after a 4 minute heat shock and embryos on the right (C,F,I,L,O,R) are similarly staged odd embryos. Changes in expression patterns are described in the text.

Fig. 4.

Effect of ectopic odd-skipped on pair-rule genes. RNA expression patterns are shown, from top to bottom, for the pair-rule genes odd-paired (opa; A-C), fushi tarazu (ftz; D-F), hairy (h; G-I), runt (run; J-L), even-skipped (eve; M-O), and paired (prd; P-R) in stage 5-7 embryos. Embryos on the left (A,D,G,J,M,P) are wild-type, embryos in the middle (B,E,H,K,N,Q) are Hsodd embryos fixed 25 minutes after a 4 minute heat shock and embryos on the right (C,F,I,L,O,R) are similarly staged odd embryos. Changes in expression patterns are described in the text.

Two of the seven pair-rule genes tested do not show significant changes in expression at the stages shown. These include the genes odd-paired (opa, Fig. 4B) and, to our surprise, ftz (Fig. 4E). In odd embryos, ftz stripes do not resolve properly (Fig. 4F), remaining about 3 cells wide until well into the process of germ band extension (Mullen and DiNardo, 1995). This suggested that Odd may be a repressor of ftz. However, as shown in Fig. 4E, ectopic Odd does not repress ftz expression. Also unexpected was the fact that ectopic Odd has effects on all three of the ‘primary’ pair-rule genes. These were previously thought not to be regulated by Odd (Carroll and Scott, 1986; Carroll et al., 1988; Klingler and Gergen, 1988). In stage 5 embryos, stripe 1 of h is efficiently repressed by ectopic Odd (Fig. 4H). The first stripe of eve was also repressed at this stage (not shown). Repression of h stripe 1 continues in older embryos and is accompanied by weaker repression of stripes 2-6. These effects of Odd on h correlate with what appears to be a modest broadening of h stripes in odd embryos, particularly stripe 1 (Fig. 4I). Early repression of the first stripes of h and eve likely accounts for the cuticular head defects that arise from early pulses of ectopic Odd expression (Fig. 2B). Interestingly, in odd embryos, the entire 7-stripe pattern of h appears to expand, both anteriorly and posteriorly. This is also true of eve and run stripes (see below). Our data provide no explanation for this, but it may explain the fairly consistent spacing of h stripes, despite their apparent broadening.

In stage 6 embryos, all seven stripes of run are moderately repressed by ectopic Odd (Fig. 4K). This correlates with what appears to be a slight broadening and strengthening of run stripes in odd embryos (Fig. 4L). eve is the most dramatically affected of the so called primary pair-rule genes at this stage. All 7 eve stripes are strongly repressed by ectopic Odd (Fig. 4N). In odd embryos, however, eve stripes are only marginally wider (Fig. 4O), but at later stages (not shown), widening is more obvious, with both primary and secondary eve stripes widening from one to two or three cells in width. The response of prd is unusual in that only the anterior portions of each stripe are repressed (Fig. 4Q). This is the stage at which each prd stripe normally begins to split into two narrower stripes, beginning at the anterior end of the embryo (Fig. 4P). In odd embryos, stripe splitting is substantially delayed (Fig. 4R).

Effects of Odd on late pair-rule gene and early segment polarity gene expression

Odd induction affects segmentation when induced as late as 3:10 AEL, suggesting that Odd may also be affecting late pair-rule and early segment polarity gene expression. Therefore, we heat pulsed embryos at these later stages and fixed them 25 minutes after heat shock to test for direct effects on these genes in stage 6/7 embryos. Except for prd and slp, the effects of ectopic Odd on the pair-rule genes are similar to those described in the previous section. For prd and sloppy-paired (slp), expression normally changes at gastrulation from 7-stripe to 14-stripe patterns of expression (Fig. 5A,C). Following induction of ectopic Odd, expression of every second prd and slp stripe is repressed (Fig. 5B,D). The repressed stripes are the secondary stripes, which lie in the posterior regions of odd-numbered (eve-expressing) parasegments and, in the case of prd, also overlap with the anterior edges of even-numbered (ftz-expressing) parasegments. In odd embryos, these same stripes of prd (Baumgartner and Noll, 1990) and slp (data not shown) become wider.

Fig. 5.

Effects of ectopic odd on late pair-rule and early segment polarity gene expression. RNA expression patterns are shown, from top to bottom, for paired (prd; A,B), sloppy-paired (slp; C,D) engrailed (en; E,F) and wingless (wg; G,H) in stage 6 (gastrulating) embryos. Embryos on the left are wild-type (A,C,E,G) and those on the right are HSodd embryos (B,D,F,H) fixed 25 minutes after a 4 minute heat shock. Secondary stripes of paired and sloppy-paired are differentially repressed by ectopic Odd, whereas all 14 en and wingless stripes are repressed.

Fig. 5.

Effects of ectopic odd on late pair-rule and early segment polarity gene expression. RNA expression patterns are shown, from top to bottom, for paired (prd; A,B), sloppy-paired (slp; C,D) engrailed (en; E,F) and wingless (wg; G,H) in stage 6 (gastrulating) embryos. Embryos on the left are wild-type (A,C,E,G) and those on the right are HSodd embryos (B,D,F,H) fixed 25 minutes after a 4 minute heat shock. Secondary stripes of paired and sloppy-paired are differentially repressed by ectopic Odd, whereas all 14 en and wingless stripes are repressed.

The effects of ectopic Odd were also examined on expression of the segment polarity genes engrailed (en) and wingless (wg), since previous studies suggested that Odd may be a repressor of both genes (Manoukian and Krause, 1992, 1993; Mullen and DiNardo, 1995). Consistent with these studies, all 14 stripes of en and wg were rapidly repressed in HSodd embryos (Fig. 5F,H). This repression is short-lived, as stripes of both en and wg were observed once again in stage 11 embryos (Fig. 3). Indeed, expression returned about 45-60 minutes post-heat shock (data not shown).

Early and late effects of Odd on ftz

Given the results of previous genetic studies, the lack of an effect of ectopic Odd on ftz was somewhat surprising. Hence, we looked more carefully for possible effects at other stages of ftz expression. Fig. 6 shows that ectopic Odd does indeed affect ftz expression in earlier stage 4/5 embryos. However, the result was once again unexpected. Rather than negatively regulating ftz, ectopic Odd causes a rapid expansion of all 7 ftz stripes (Fig. 6B). In some embryos, interstripe regions are difficult to discern (not shown). Since this activation is observed within 20-25 minutes of Odd induction, it likely reflects a direct interaction between the two genes. Consistent with this positive relationship, initiating ftz stripes are irregular in width and intensity in odd mutant embryos (Fig. 6C). Stripes 3-6 are the most strongly affected, particularly stripe 4.

Fig. 6.

Odd activates ftz in syncytial blastoderm embryos. Expression of ftz mRNA is shown in stage 5 (cellularization 50% complete) embryos. (A) Wild-type embryo: ftz stripes are on average 2-3 cells wide. (B) HSodd embryo fixed 25 minutes after a 4 minute heat shock: ectopic Odd causes a rapid broadening and intensification of all seven ftz stripes. (C) odd embryo: stripes 2-6 are diminished in intensity and width.

Fig. 6.

Odd activates ftz in syncytial blastoderm embryos. Expression of ftz mRNA is shown in stage 5 (cellularization 50% complete) embryos. (A) Wild-type embryo: ftz stripes are on average 2-3 cells wide. (B) HSodd embryo fixed 25 minutes after a 4 minute heat shock: ectopic Odd causes a rapid broadening and intensification of all seven ftz stripes. (C) odd embryo: stripes 2-6 are diminished in intensity and width.

In germband-extending (stage 7) embryos, ectopic Odd, once again, does not appear to have an effect on ftz expression (Fig. 7B). At this stage, endogenous odd is no longer expressed in the same cells as ftz, but rather immediately posterior to each narrowing ftz stripe (Manoukian and Krause, 1993). In odd embryos, ftz continues to be expressed in the cells that normally would express odd, generating stripes that are 2 to 3 cells wide, rather than 1-2 cells wide, as normally observed at this stage (Mullen and DiNardo, 1995; Fig. 7C). Taken together, these observations suggest the possibility that Odd may function as a repressor of ftz, but only in the cells where Odd is normally expressed. To test this possibility, we induced Odd expression in odd embryos. Fig. 7D shows that ectopic expression of Odd in odd embryos does indeed repress the broadened 2-to 3-cell-wide ftz stripes, returning their width to a normal 1-2 cell width. Thus, Odd does appear to be able to repress the ftz gene, but only in the middle region of the ftz parasegment, and only after gastrulation.

Fig. 7.

Odd represses ftz in gastrulating embryos. In order to test whether Odd could repress ftz (A-D) within the normal domains of odd expression, ectopic Odd was expressed in odd embryos. In a wild-type stage 7 embryo, stripes 1-6 of ftz are on average 1-2 cells wide (A). No effect is seen on these stripes when Odd is expressed ectopically (B). In odd embryos, stripes 1-6 are 2-3 cells wide (C). These posteriorly expanded stripes are repressed by ectopic Odd, reverting the stripes to wild-type width (D). (E-H) Corresponding en patterns later on in stage 10 (5:00-5:30 AEL) embryos. (E) Wild-type embryo with stripes 1-14 indicated. (F) HSodd embryo: all 14 stripes are present. (G) odd embryo: 7 additional ftz-dependent en stripes appear posterior to the normal even-numbered en stripes. (H) HSodd; odd embryo: repression of the broadened ftz stripes in HSodd; odd embryos (D) prevents ectopic en stripes from forming.

Fig. 7.

Odd represses ftz in gastrulating embryos. In order to test whether Odd could repress ftz (A-D) within the normal domains of odd expression, ectopic Odd was expressed in odd embryos. In a wild-type stage 7 embryo, stripes 1-6 of ftz are on average 1-2 cells wide (A). No effect is seen on these stripes when Odd is expressed ectopically (B). In odd embryos, stripes 1-6 are 2-3 cells wide (C). These posteriorly expanded stripes are repressed by ectopic Odd, reverting the stripes to wild-type width (D). (E-H) Corresponding en patterns later on in stage 10 (5:00-5:30 AEL) embryos. (E) Wild-type embryo with stripes 1-14 indicated. (F) HSodd embryo: all 14 stripes are present. (G) odd embryo: 7 additional ftz-dependent en stripes appear posterior to the normal even-numbered en stripes. (H) HSodd; odd embryo: repression of the broadened ftz stripes in HSodd; odd embryos (D) prevents ectopic en stripes from forming.

This effect on ftz is paralleled by the response of en. When examined immediately after heat shock, all 14 en stripes are effectively repressed, as shown in Fig. 5. However, all 14 stripes reinitiate within an hour of heat shock, and are relatively normal in stage 9 embryos (Fig. 7F). In odd embryos, ectopic en stripes appear along the posterior edges of the widened ftz stripes (Fig. 7G) due to activation by ectopic Ftz (DiNardo and O’Farrell, 1987; Mullen and DiNardo, 1995). Like the other 14 en stripes, these are also repressed by ectopic Odd when induced in stage embryos (not shown), but unlike the other stripes, these fail to reinitiate (Fig. 7H). The failure of en to reinitiate in these cells is likely due to lack of the en activator Ftz, which was repressed earlier by Odd (Fig. 7D).

Odd functions primarily as a repressor

Previous genetic studies suggested that the product of the odd gene would function as a repressor. Four putative target genes were identified; the pair-rule genes ftz (Manoukian and Krause 1992, 1993; Mullen and DiNardo, 1995) and prd (Baumgartner and Noll, 1990), and the segment polarity genes en (Manoukian and Krause, 1992, 1993; Mullen and DiNardo, 1995; Florence et al., 1997) and wg (Mullen and DiNardo, 1995). Our results are consistent with Odd acting predominantly as a repressor of target gene expression. All four genes, ftz, prd, en and wg were repressed in the same short time frame after ectopic Odd induction, suggesting that they are indeed direct targets of Odd. In addition, we identified four more genes that appear to be directly repressed by Odd; the genes slp, eve, h and run. Although we classify these interactions as direct, in that they most likely do not require the activation or repression of intermediary genes, they need not involve direct interactions between Odd and each of the target gene promoters. These regulatory relationships, summarized in Fig. 8, are consistent with the temporal and spatial expression patterns of these genes relative to odd in wild-type embryos, and with the changes observed in these patterns in odd embryos. They are also consistent with the segmental phenotypes produced by odd and HSodd embryos (discussed below).

Fig. 8.

Spatial and regulatory relationships between Odd and putative target genes. This schematic depicts segmentation gene expression patterns and relevant regulatory interactions within a 3-parasegment-wide interval (indicated by text and solid black lines at the top of A and vertical dashed gray lines in A and B). (A) Expression patterns and interactions that occur near the completion of cellularization (stage 5/6). (B) Patterns and interactions that occur in gastrulating embryos (stage 6/7). Segmentation gene stripes are indicated by boxes. The width of each box correlates with the relative width of the stripe indicated. Sloped sides indicate regions of stripes in the process of narrowing. Boxes indicating secondary stripes of eve in B are half the heights of others, indicating lower levels of expression. Positive regulatory interactions between genes are indicated as green arrows, while negative interactions are indicated as red, blunt-ended arrows. Dashed arrows indicate a weak interaction. Purple brackets at the bottom of B indicate the regions missing in HSodd cuticles. Earlier non-periodic interactions are not shown.

Fig. 8.

Spatial and regulatory relationships between Odd and putative target genes. This schematic depicts segmentation gene expression patterns and relevant regulatory interactions within a 3-parasegment-wide interval (indicated by text and solid black lines at the top of A and vertical dashed gray lines in A and B). (A) Expression patterns and interactions that occur near the completion of cellularization (stage 5/6). (B) Patterns and interactions that occur in gastrulating embryos (stage 6/7). Segmentation gene stripes are indicated by boxes. The width of each box correlates with the relative width of the stripe indicated. Sloped sides indicate regions of stripes in the process of narrowing. Boxes indicating secondary stripes of eve in B are half the heights of others, indicating lower levels of expression. Positive regulatory interactions between genes are indicated as green arrows, while negative interactions are indicated as red, blunt-ended arrows. Dashed arrows indicate a weak interaction. Purple brackets at the bottom of B indicate the regions missing in HSodd cuticles. Earlier non-periodic interactions are not shown.

Odd can also function as an activator

An advantage of using a heat-inducible transgene is that expression can be induced at virtually any stage of development. This allowed dissection of early, mid and late roles of Odd on other segmentation genes. A surprising discovery was that Odd functions not only as a repressor of the ftz gene, but also as an activator. Activation of ftz occurs when ectopic Odd is expressed prior to the completion of cellularization and is most pronounced at the beginning of cellularization. It occurs with the same rapid kinetics as observed for genes that are repressed by Odd, indicating that this also is likely to occur in the absence of intermediary gene regulation. This positive relationship between Odd and ftz is consistent with the expression patterns of the two genes at this stage: odd and ftz stripes overlap perfectly, except for stripe 7 of odd which is missing at this stage (Manoukian and Krause, 1993). Beginning at stage 6, odd and ftz stripes begin to resolve into non-overlapping patterns and it is at this later time that Odd becomes a repressor of ftz.

Since ftz is the only gene amongst those studied here that is activated by Odd, and is only activated within a short temporal window, the primary activity of Odd is most likely as a repressor. Its function as an activator probably requires the participation of a cofactor that is expressed early and exhibits specificity for the ftz promoter. A similar observation was made in a previous study (Manoukian and Krause, 1992), where it was noted that the even-skipped protein (Eve) could also function as both an activator and a repressor of ftz. The switch in Eve activities, from an activator to a repressor, occurs at about the same time as the switch in Odd activities. Moreover, the switch also occurs as the stripes of eve and ftz resolve into non-overlapping patterns. These similarities in Odd and Eve target gene specificity and regulation may be coincidental, or may be due to the participation of a common cofactor (or cofactor complex).

Odd activity is controlled temporally and spatially

Our ability to express Odd uniformly and at precise developmental stages also revealed a number of temporal and spatial limitations in Odd activities. For example, when Odd switches from an activator to a repressor of ftz, its ability to repress ftz is excluded from the anterior-most cells of each ftz stripe. The inability of Odd to repress ftz in these cells indicates that, either a necessary cofactor for Odd is missing in these cells, or that an overriding factor is present. Another possibility is that the levels of Odd required to repress ftz are higher than those that we induced.

Another example of spatial and temporal limitations in Odd activities is in the ability of Odd to repress h and eve. In early cellularizing embryos, repression of these genes is limited to their anterior-most stripes. This suggests a possible interaction at this stage with anteriorly localized gap or terminal group proteins. Shortly thereafter, Odd is capable of repressing eve throughout the embryo while h continues to be repressed predominantly in anterior regions. Yet another example of spatial and temporal limitations of Odd activity is the ability of Odd to repress only the secondary stripes of prd and slp, which are normally initiated in gastrulating embryos. This indicates differential interactions with the enhancers and proteins that regulate these stripes at this time.

A final example of spatially restricted Odd activity is its inability to alter segmentation in the ventral half of the embryo during later stages of the Odd-sensitive period. Induction of Odd between 2:50 and 3:10 AEL causes pair-rule fusions, but only in the dorsal half of the embryo. This suggests some form of interaction with components of the dorsoventral controlling pathways. Links between anteroposterior and dorsoventral controlling genes have been noted previously (see for example: DiNardo and O’Farrell, 1987; Manoukian and Krause, 1992; Akimura et al., 1997; Nibu et al., 1998; Fisher and Caudy, 1998), and likely occur at a variety of levels.

Revision of pair-rule gene paradigms

Three of the pair-rule genes that we show are repressed by Odd were previously classified as a subgroup of pair-rule genes called ‘primary’ pair-rule genes. These were thought to function as intermediary genes between the earlier functioning gap genes and the other pair-rule genes (Carroll and Scott, 1986; Howard and Ingham, 1986; Pankratz and Jackle, 1990). This deduction was based on three observations. First, primary pair-rule gene stripes were thought to form prior to those of most other pair-rule genes. Second, these patterns were thought to be unaffected by mutations in the other pair-rule genes. Third, all three genes were thought to be regulated by large, complex promoters that possess a unique ability to decode non-periodic gap gene cues using stripe-specific enhancers.

Our results here, as well as those of others (Gutjahr et al., 1993; Tsai and Gergen, 1994; Yu and Pick, 1995; Klingler et al., 1996), suggest that the primary/secondary pair-rule gene paradigm requires revision. For example, odd expression initiates first in a single stripe and then in broad patterns that suggest direct responses to gap gene cues (Coulter et al., 1990). These patterns are initiated early enough (late stage 4/early stage 5) to regulate eve, h, run and ftz. Indeed, we show that ectopic Odd is capable of regulating all four genes and that the expression patterns of these genes are altered in odd mutant embryos. These changes may have been missed in earlier studies due to their subtle nature in stage 6 embryos. The full effect of these actions may also be masked by the redundant actions of other segmentation genes. Indeed, redundancy may be masking a number of important gene interactions that can only be revealed by approaches such as double mutant and ectopic expression analyses.

Gap-like functions of odd

The earliest observed effects of ectopic Odd (late stage 4/early stage 5) are on the anterior-most stripes of eve and h. Repression of these stripes results in larval head defects in structures primarily derived from maxillary segment primordia, which is where the first stripes of h and eve are centered. These defects are similar to those previously reported in h (Howard et al., 1988; Ingham et al., 1988b) and eve (Nusslein-Volhard et al., 1985) mutant cuticles. At this time, endogenous odd is expressed in a single stripe immediately posterior to the first stripes of eve and h (Coulter et al., 1990), suggesting that it acts at this time to prevent posterior expansion of the anterior-most eve and h stripes. Indeed, in odd mutant embryos, the first stripes of eve and h appear to expand (Fig. 4 and data not shown), and head defects occur in structures normally derived from adjacent regions (Nusslein-Volhard et al., 1985; Coulter and Wieschaus, 1988). Thus, odd appears to have an early gap-like function which is exerted on the primary pair-rule genes h and eve.

Early pair-rule functions of odd

When ectopic Odd is induced just prior to gastrulation, all seven stripes of eve are repressed while ftz stripes are broadened. Later, en stripes that normally mark the boundaries of eve-dependent parasegments are closer together than in wild-type embryos, and conversely, those bordering ftz-dependent parasegments are farther apart (note, for example, the space between en stripes 5 and 6 in Fig. 3B). The same effect is seen in eveID19 embryos grown at 25°C (Frasch et al., 1988) and, although not previously noted, the opposite is true in odd mutant embryos. Stripes of en that border the eve-dependent parasegments are farther apart and those that border ftz-dependent parasegments are closer together (Fig. 7G). These data suggest that Odd functions in early embryos to help position the parasegmental borders. This is achieved by activation of ftz and repression of eve.

Late pair-rule and segment polarity functions of odd

Previous studies could not establish unambiguously whether Odd acts as a direct or indirect repressor of the en and wg genes (Manoukian and Krause, 1992, 1993; Mullen and DiNardo, 1995; Florence et al., 1997). The data presented here show that, during gastrulation, Odd appears to regulate both genes, not only directly, but indirectly as well. Indirect repression is mediated by selective repression of the en and wg activators, ftz, prd, eve and slp. The result of these interactions in HSodd embryos is first the loss of all fourteen en and wg stripes due to direct repression and then failure of certain stripes to reinitiate.

Stripes of en reappear because their activators, ftz, prd and opa are not repressed where required for en activation. An exception is the set of ectopic en stripes that initiate later in odd mutant embryos. In HSodd; odd embryos, these stripes failed to initiate due to selective repression of Ftz in these cells. Interestingly, initiation of odd-numbered en stripes normally requires expression of eve as well as prd, and yet these stripes recover well despite near complete repression of eve by ectopic Odd. It is possible that low levels of eve are sufficient to activate en, or that eve acts earlier in an indirect fashion, as has previously been suggested (Manoukian and Krause 1992, 1993).

Odd-numbered wg stripes fail to recover fully, most likely due to repression of the prd and slp stripes that are required for full wg expression. In turn, loss of these wg stripes eventually leads to loss of the adjacent (even-numbered) en stripes in older (stage 11) embryos. This is because en expression is dependent upon signaling from adjacent wg-expressing cells at this stage (Martinez-Arias et al., 1988; DiNardo and Heemskerk, 1990; Heemskerk et al., 1991). Loss of these en and wg stripes partially explains the HSodd pair-rule phenotype: the segmental regions that are missing include these cells (see Fig. 8). Thus, a major function of Odd in gastrulating embryos is to establish polarity within the middle of even-numbered parasegments by preventing expression of en and wg, and their respective activators. The expansion of en and wg stripes in odd mutant embryos (Mullen and DiNardo, 1995) is consistent with this function. The secondary stripes of odd likely play a similar role in odd-numbered parasegments, although such a role was not noted in the analysis of odd mutant embryos (Mullen and DiNardo, 1995) and may be functionally redundant. Odd may also regulate other segmentation genes at this stage, since the boundaries of regions deleted in HSodd embryos extend anteriorly and posteriorly beyond the missing en- and wg-expressing cells (see Fig. 8). Repression of odd-numbered slp stripes, which extend one cell anterior to wg stripes, may partially explain this observation.

Further elucidation of the segmentation gene hierarchy

The experimental approach described here, generating brief pulses of uniformly expressed protein, has confirmed findings of previous genetic studies, as well as revealing a number of unsuspected Odd activities. This approach is now well tested (Manoukian and Krause, 1992, 1993; A. N. and H. M. K., unpublished data), and has a great deal of potential for testing the regulatory relationships between other genes in an in vivo setting. Once these relationships are better understood, the intensive efforts required to identify and characterize relevant regulatory elements and trans-acting factors can be approached more confidently.

The results that we have obtained thus far emphasize the complexity and highly dynamic nature of the segmentation gene hierarchy. Previous genetic analyses of the circuitry involved have only provided a rough basis for the numerous and rapidly changing gene and protein interactions that occur. On occasions, interpretations have even been misleading. A great deal of work is still required to fully delineate and verify all of the interactions that occur between and within the various classes of segmentation genes. In the case of Odd, one of the next important steps will be to identify cofactors that modify its function in a temporally and spatially restricted fashion, permitting it to act as both an activator and a repressor of target gene expression.

We would like to thank Doug Coulter for providing us with odd cDNA and for helpful discussions. We would also like to thank Howard Lipshitz, Carol Schwartz and Mark Perry for comments on the manuscript. This work was supported by a grant from the National Cancer Institute of Canada, with funds from the Canadian Cancer Society. B. S.-Le D. was supported by a fellowship awarded by the Association pour la Recherche sur le Cancer (ARC France).

Akimaru
,
H.
,
Hou
,
D. X.
and
Ishii
,
S.
(
1997
).
Drosophila CBP is required for dorsal-dependent twist gene expression
.
Nat. Genet.
17
,
211
214
.
Baumgartner
,
S.
and
Noll
,
M.
(
1990
).
Network of interactions among pairrule genes regulating paired expression during primordial segmentation of Drosophila
.
Mech. Dev.
33
,
1
18
.
Carroll
,
S. B.
and
Scott
,
M. P.
(
1986
).
Zygotically active genes that affect the spatial expression of the fushi tarazu segmentation gene during early Drosophila embryogenesis
.
Cell
45
,
113
26
.
Coulter
,
D. E.
and
Wieschaus
,
E.
(
1988
).
Gene activities and segmental patterning in Drosophila: analysis of odd-skipped and pair-rule double mutants
.
Genes Dev.
2
,
1812
23
.
Coulter
,
D. E.
,
Swaykus
,
E. A.
,
Beran-Koehn
,
M. A.
,
Goldberg
,
D.
,
Wieschaus
,
E.
and
Schedl
,
P.
(
1990
).
Molecular analysis of odd-skipped, a zinc finger encoding segmentation gene with a novel pair-rule expression pattern
.
EMBO J.
9
,
3795
804
.
DiNardo
,
S.
and
O’Farrell
,
P. H.
(
1987
).
Establishment and refinement of segmental pattern in the Drosophila embryo: spatial control of engrailed expression by pair-rule genes
.
Genes Dev.
1
,
1212
25
.
DiNardo
,
S.
and
Heemskerk
,
J.
(
1990
).
Molecular and cellular interactions responsible for intrasegmental patterning during Drosophila embryogenesis
.
Sem. in Cell Biol.
1
,
173
83
.
Engels
,
W. R.
,
Benz
,
W. K.
,
Preston
,
C. R.
,
Graham
,
P. L.
,
Phillis
,
R. W.
and
Robertson
,
H. M.
(
1987
).
Somatic effects of P element activity in Drosophila melanogaster: pupal lethality
.
Genetics
117
,
745
57
.
Fisher
,
A. L.
and
Caudy
,
M.
(
1998
).
Groucho proteins: transcriptional corepressors for specific subsets of DNA-binding transcription factors in vertebrates and invertebrates
.
Genes Dev.
13
,
1931
1940
.
Florence
,
B.
,
Guichet
,
A.
,
Ephrussi
,
A.
and
Laughon
,
A.
(
1997
).
Ftz-F1 is a cofactor in Ftz activation of the Drosophila engrailed gene
.
Development
124
,
839
47
.
Frasch
,
M.
,
Warrior
,
R.
,
Tugwood
,
J.
and
Levine
,
M.
(
1988
).
Molecular analysis of even-skipped mutants in Drosophila development
.
Genes Dev.
2
,
1824
1838
.
Gergen
,
P. J.
and
Wieschaus
,
E.
(
1985
).
The localized requirements for a gene affecting segmentation in Drosophila: analysis of larvae mosaic for runt
.
Dev. Biol.
109
,
321
335
.
Gutjahr
,
T.
,
Frei
,
E.
and
Noll
,
M.
(
1993
).
Complex regulation of early paired expression: initial activation by gap genes and pattern modulation by pairrule genes
.
Development
117
,
609
23
.
Heemskerk
,
J.
,
DiNardo
,
S.
,
Kostriken
,
R.
and
O’Farrell
,
P. H.
(
1991
).
Multiple modes of engrailed regulation in the progression towards cell fate determination
.
Nature
352
,
404
10
.
Howard
,
K.
and
Ingham
,
P.
(
1986
).
Regulatory interactions between the segmentation genes fushi tarazu, hairy, and engrailed in the Drosophila blastoderm
.
Cell
44
,
949
57
.
Howard
,
K.
,
Ingham
,
P.
and
Rushlow
,
C.
(
1988
).
Region-specific alleles of the segmentaion gene hairy
.
Genes Dev.
2
,
1037
1046
.
Ingham
,
P. W.
(
1988
).
The molecular genetics of embryonic pattern formation in Drosophila
.
Nature
335
,
25
34
.
Ingham
,
P. W.
,
Baker
,
N. E.
and
Marinez-Arias
,
A.
(
1988a
).
Regulation of segment polarity genes in the Drosophila blastoderm by fushi tarazu and even-skipped
.
Nature
331
,
73
75
.
Ingham
,
P. W.
,
Howard
,
K. R.
and
Ish-Horowicz
,
D.
(
1988b
).
Transcription pattern of the Drosophila segmentation gene hairy
.
Nature
318
,
439
445
.
Klingler
,
M.
,
Soong
,
J.
,
Butler
,
B.
and
Gergen
,
J. P.
(
1996
).
Disperse versus compact elements for the regulation of runt stripes in Drosophila
.
Dev. Biol.
177
,
73
84
.
Lawrence
,
P. A.
,
Johnston
,
P.
,
Macdonald
,
P.
and
Struhl
,
G.
(
1987
).
Borders of parasegments in Drosophila embryos are delimited by the fushi tarazu and even-skipped genes
.
Nature
328
,
440
2
.
Lehmann
,
R.
and
Tautz
,
D.
(
1994
).
In situ hybridization to RNA
.
Methods Cell Biol.
44
,
575
98
.
Manoukian
,
A. S.
and
Krause
,
H. M.
(
1992
).
Concentration-dependent activities of the even-skipped protein in Drosophila embryos
.
Genes Dev.
6
,
1740
51
.
Manoukian
,
A. S.
and
Krause
,
H. M.
(
1993
).
Control of segmental asymmetry in Drosophila embryos
.
Development
118
,
785
96
.
Martinez-Arias
,
A.
and
Lawrence
,
P. A.
(
1985
).
Parasegments and compartments in the Drosophila embryo
.
Nature
313
,
639
42
.
Martinez-Arias
,
A.
,
Baker
,
N. E.
and
Ingham
,
P. W.
(
1988
).
Role of segment polarity genes in the definition and maintenance of cell states in the Drosophila embryo
.
Development
103
,
157
70
.
Mullen
,
J. R.
and
DiNardo
,
S.
(
1995
).
Establishing parasegments in Drosophila embryos: roles of the odd- skipped and naked genes
.
Dev. Biol.
169
,
295
308
.
Nibu
,
Y.
,
Zhang
,
H.
and
Levine
,
M.
(
1998
).
Interaction of short range repressors with drosohila CtBP in the embryo
.
Science
280
,
101
104
.
Nusslein-Volhard
,
C.
and
Wieschaus
,
E.
(
1980
).
Mutations affecting segment number and polarity in Drosophila
.
Nature
287
,
795
801
.
Nusslein-Volhard
,
C.
,
Kluding
,
H.
and
Jurgens
,
G.
(
1985
).
Genes affecting the segmental subdivision of the Drosophila embryo
.
Cold Spring Harbor Symp. Quant. Biol.
50
,
145
154
.
Pankratz
,
M. J.
and
Jackle
,
H.
(
1990
).
Making stripes in the Drosophila embryo
.
Trends Genet.
6
,
287
92
.
Robertson
,
H. M.
,
Preston
,
C. R.
,
Phillis
,
R. W.
,
Johnson-Schlitz
,
D. M.
,
Benz
,
W. K.
and
Engels
,
W. R.
(
1988
).
A stable genomic source of P element transposase in Drosophila melanogaster
.
Genetics
118
,
461
70
.
Spradling
,
A. C.
and
Rubin
,
G. M.
(
1982
).
Transposition of cloned P elements into Drosophila germ line chromosomes
.
Science
218
,
341
7
.
Tautz
,
D.
and
Pfeifle
,
C.
(
1989
).
A non-radioactive in situ hybridization method for the localization of specific RNAs in Drosophila embryos reveals translational control of the segmentation gene hunchback
.
Chromosoma
98
,
81
5
.
Thummel
,
C. S.
and
Pirrotta
,
V.
(
1992
)
New pCaSpeR P element vectors
.
Dros. Info. Serv.
71
,
150
.
Tsai
,
C.
and
Gergen
,
J. P.
(
1994
).
Gap gene properties of the pair-rule gene runt during Drosophila segmentation
.
Development
120
,
1671
83
.
Weischaus
,
E.
and
Nusslein-Volhard
,
C.
(
1986
). Looking at embryos. In
Drosophila: A Practical Approach
(ed.
D. B.
Roberts
) pp.
199
-
228
. Oxford, UK: IRL Press.
Yu
,
Y.
and
Pick
,
L.
(
1995
).
Non-periodic cues generate seven ftz stripes in the Drosophila embryo
.
Mech. Dev.
50
,
163
75
.