Repression of Drosophila pair-rule segmentation genes by ectopic expression of tramtrack

The establishment of the segmental body pattern of Drosophila is governed by a hierarchy of maternal and zygotic genes. The zygotically expressed segmentation genes have been classified into three categories according to their temporal and spatial expression patterns. The gap genes are expressed as broad domains along the antero-posterior axis of the embryo. The pair-rule genes are expressed with double-segment periodicity in the early embryos and the segment polarity genes are expressed in a part of every segment of the developing embryo (for reviews see Akam, 1987; Nusslein-Volhard et al., 1987; Scott and Carroll, 1987; Ingham, 1988).

The fushi tarazu (ftz) gene is a member of the pair-rule class of segmentation genes. Embryos homozygous for the ftz mutation are missing the even-numbered parasegments and die late in embryogenesis (Wakimoto and Kaufman, 1981). Ftz RNA is first detected weakly throughout the embryo during nuclear cycle 10 (Hafen et al., 1984; Weir and Kornberg, 1985). During the later division cycles, expression becomes restricted to 15%-65% egg length (0% is the posterior pole). By late nuclear cycle 14 (the cellu-lar blastoderm stage), ftz RNA and protein expression evolves into 7 evenly spaced stripes encircling the embryo (Carroll and Scott, 1985; Krause et al., 1988; Karr and Kornberg, 1989). These stripes of ftz expression correspond approximately to the regions that are missing in a ftz mutant embryo (Wakimoto and Kaufman, 1981). Ftz is expressed in specific cells of every segment of the developing ner-vous system at 5-12 hours of embryogenesis (Carroll and Scott, 1985; Doe et al., 1988), and in a restricted section of the developing hindgut around 12-15 hours of development (Krause et al., 1988).

The expression of ftz is largely controlled at the level of transcription (Hiromi et al., 1985; Edgar et al., 1986). Three cis-acting elements have been mapped by promoter fusion analysis (Hiromi et al., 1985): the zebra element (740 bp) confers a weak striped pattern of expression in the early embryonic mesoderm; the upstream element (6.1 kb to 3.4 kb upstream of the translational start) has enhancer-like properties that ensure high levels of ftz expression in the ectoderm and mesoderm, and the neurogenic element (2.45 kb upstream from the translational start) is necessary for ftz expression in the developing nervous system.

A number of trans-regulators of ftz have been identified by genetic studies. The pair-rule segmentation gene hairy plays a critical role in repressing ftz in the interstripe regions (Carroll and Scott, 1986; Howard and Ingham, 1986; Frasch and Levine, 1987; Hiromi and Gehring, 1987; Ish-Horow-icz and Pinchin, 1987; Carroll et al., 1988; Hooper et al., 1989). Although it has been shown that repression by hairy is mediated through the zebra element, it remains to be established whether hairy protein binds DNA directly, or acts indirectly through the association with another factor. In addition, ftz protein positively autoregulates ftz expression through the binding of its homeodomain to mul-tiple sites in the upstream element (Pick et al., 1990; Schier and Gehring, 1992).

Biochemical approaches have led to the identification of many protein binding sites within the regulatory sequences of ftz (Harrison and Travers, 1988, 1990; Dearolf et al. 1989a, 1989b; Ueda et al., 1990; Pick et al., 1990; Brown et al., 1991; Topol et al., 1991). A number of the DNA-binding factors that interact with these sequences have now been identified, including the gap gene caudal and two previously unidentified genes FTZ-F1 and FTZ-F2 / tramtrack. The gap gene caudal acts as an activator of ftz in the pos-terior region of the embryo (Dearolf et al., 1989b). FTZ-F1 is a new member of the steroid hormone receptor super-family and is implicated in the overall activation of ftz, par-ticularly in stripes 1, 2, 3 and 6 (Ueda et al., 1990; Lavorgna et al., 1991). Tramtrack (ttk) / FTZ-F2 has been identified as a zinc finger protein that is capable of binding to several sites in the zebra element and to sites in the enhancer ele-ment of ftz (Harrison and Travers, 1988, 1990; Brown et al., 1991). The ttk protein has also been shown to bind in vitro to the pair-rule gene even skipped (eve) (Read et al., 1990; Jiang et al., 1991; Read and Manley, 1992a,b).

In a previous study, we have proposed that ttk / FTZ-F2 (hereafter referred to as ttk) functions as a negative regu-lator of ftz, since point mutations that eliminate ttk binding to the zebra element cause derepression of ftz-lacZ con-structs in transformed embryos (Brown et al., 1991). In addition to abberant expression in the germband-extended-stage embryos, global derepression of ftz was observed in preblastoderm-stage embryos, as early as the third nuclear division cycle. With the exception of the detection of engrailed protein in pre-cycle 9 embryos (Karr et al. 1989), transcriptional activity at such an early stage of embryoge-nesis has not been observed. The earliest transcription by RNA polymerase II from the zygotic genome is known to occur around nuclear division cycles 9 or 10 (Lamb and Laird, 1976; McKnight and Miller, 1976; Anderson and Lengyel, 1979; Edgar and Schubiger 1986). It is widely assumed that preblastoderm-stage embryos are transcrip-tionally silent due to the absence of specific activators or to the inactivity of the general transcriptional machinery. Our findings with the mutated ftz-lacZ constructs suggest that, in the case of ftz, there is specific repression by ttk in pre-cycle 10 embryos.

In this paper, we have examined the distribution of ttk protein during embryogenesis and the effects of ectopic ttk expression on the expression of ftz and on subsequent embryonic development. Our results provide further evi-dence in support of the hypothesis that ttk acts as a direct, maternal repressor of ftz, and perhaps of other pair-rule seg-mentation genes.

An NdeI site was engineered at the transcription start site of the ttk gene by site-directed mutagenesis. The full-length clone was then isolated as a 2.5kb NdeI-BamHI fragment (the BamHI restric-tion was a partial digest since there is a BamHI site near the 3′ end of the gene), and cloned into NdeI-BamHI restricted JC10 (Clos et al., 1990), a derivative of the pET3C vector (Studier and Moffatt, 1986). The resulting plasmid (pJLBG), has ttk under the control of the T7 promoter.

400 ml Luria-Bertani broth with 100 mg/ml ampicillin was inoc-ulated with a fresh transformant colony of pJLBG in BL21(DE3) and grown with shaking at 37°C until the culture reached OD_600nm 0.75. The expression of ttk was induced by adding IPTG to 0.4 mM and shaking at 37°C for 1 hour. Cells were harvested by spin-ning 10 minutes at 6000 g, resuspended in 4 ml lysis buffer (50 mM Tris pH 8.0, 1 mM EDTA, 100 mM NaCl, 0.75 mg apro-tinin, 5 mg/ml leupeptin, 2 mg/ml pepstatin, 0.5 mM phenyl-methylsulphonylfluoride (PMSF), 0.5 mM DTT) and frozen at − 70°C.

The cell suspension was subjected to three rapid freeze-thaw cycles, sonicated with 6×30 second bursts at 100 mW and incubated for 5 minutes at 37°C with 2 mg/ml deoxycholic acid. DNAase I was added to a final concentration of 10 mg/ml and incubated 30 minutes at 25°C. The lysate was centrifuged for 10 minutes at 6000 g, at 4°C. The supernatant was discarded and the pelleted inclusion bodies were resuspended in a Dounce homogeniser in 9 ml lysis buffer containing 0.5% Triton X-100, 10 mM EDTA, incubated at room temperature for 10 minutes, and centrifuged for 10 minutes at 6000 g at 4°C. The resuspension and centrifugation steps were repeated twice and the final pellet was resuspended in 4 ml Laemmli gel loading buffer (50 mM Tris-HCl pH 6.8, 2 mM EDTA, 1% SDS, 1% mercaptoethanol, 8% glycerol, 0.025% bromphenol blue). The sample was divided into 750 ml aliquots and either frozen at −70°C, or loaded on an SDS gel for electrophoretic purification of the protein.

750 μl aliquots of ttk protein in loading buffer were elec-trophoresed on 7.5% SDS gels. A small vertical slice was excised and stained with Coomassie Blue to locate ttk protein. The slice was realigned with the remainder of the gel. The gel region con-taining ttk was excised and homogenised with 4 ml gel running buffer. The slurry was incubated at 25°C for 48 hours and passed over glass wool to remove residual acrylamide. The protein was precipitated by adding KCl to 1 M, and incubating for 20 min-utes on ice, and pelleted by centrifugation in a microfuge for 10 minutes. The pellet was dissolved in 500 ml 10 mM Tris pH 8.0, 0.1% SDS and dialysed for 4 hours against the same buffer. Pro-tein concentration was measured with the Bradford assay and purity was assayed by applying an aliquot to an SDS gel and stain-ing with Coomassie Blue after electrophoresis.

Polyclonal antibodies against ttk were prepared by subcuta-neously injecting rats with 50 μg of purified ttk protein followed by two injections of 50 μg each at 3 week intervals, (Hazleton Laboratories).

SDS-PAGE (7.5% polyacrylamide) was performed according to standard procedures. Proteins were electrophoretically transferred to nitrocellulose membrane in 50 mM Tris, 380 mM glycine, 0.1% (w/v) SDS, 20% methanol. Membranes were processed with pri-mary antibody (1:250 dilution rat polyclonal anti-ttk, pre-absorbed overnight at 1:5 dilution against fixed embryos), followed by sec-ondary antibody (1:500 dilution alkaline phosphatase-conjugated goat anti-rat IgG). The alkaline phosphatase II kit (Vector Labo-ratories) was used for color development according to the manu-facturer’s instructions.

Nuclear extracts used were as described in Brown et al. (1991). Bacterial extracts were prepared as described above from BL21(DE3) cells carrying pJLBG or from XL1 Blue (Strategene) cells carrying plasmids with ttk p69 (pHD) or p88 (λ409) fused to β-galactosidase at the aminoterminus (Read and Manley, 1992a). The p69 and p88 proteins were induced with IPTG.

2 μl of nuclear embryonic extract (see Brown et al., 1991) was incubated in a total volume of 10 μl containing 20 fmoles labelled DNA-binding site, 10 μg tRNA, 4 μg poly(dI-dC).poly(dI-dC), 1 μl 5% BSA, 3.5 μl PBS and 1 μl of antibody or prebleed serum (diluted in PBS). Samples were incubated for 30 minutes at 25°C, supplemented with 2 μl gel loading buffer (2.5% Ficoll 400, 0.5× TBE and tracking dyes) and electrophoresed on a 1% agarose, 0.5× TBE gel.

pLB20, a plasmid containing the full-length ttk cDNA in pBlue-script with an NdeI site engineered at the transcription start site, was cut with NdeI and filled in with Klenow. The cDNA insert was excised with XbaI. The NdeI-XbaI fragment was ligated to SacII (cut and filled in)/XbaI cut pCaSpeRhs vector (gift of C. Thummel and V. Pirotta), thereby putting the ttk cDNA under the control of the heat-shock promoter. The filling in reaction had some exonuclease activity which nibbled the fragment ends but which left the ATG codon and the downstream sequences intact and fortuitously maintained the necessary translation start signals. The resulting plasmid is named phs-ttk and the organisation is illustrated in Fig. 3A. Our initial experience with a heat shock-ttk construct with 252 nt of 5′ untranslated leader sequence showed that these sequences interfered with the translation of the p69 polypeptide (unpublished observations).

P element-mediated germline transformations were performed as described in Rubin and Spradling (1982). The host strain, Df(1)w67c2,y was injected with 300 μg/ml phs-ttk and 150 μg/ml pd2-3wc helper DNA. Transformed G₁ progeny were selected on the basis of partial rescue of the w⁻ phenotype. Homozygous flies were generated by crossing with the appropriate balancer strains; two homozygous lines were obtained. In the transformant line 4C, the phs-ttk construct is inserted on chromosome 2 and, in line 4B, on chromosome 3.

For antibody staining and in situ hybridization, untransformed embryos and hs-ttk embryos were processed in parallel. The embryos were collected at 25°C for 45 minutes on grape juice agar plates, washed, dried, transferred onto a coverslip and cov-ered with Voltaleff 3S halocarbon oil to clear the chorion. Embryos were allowed to develop (at 25°C) until most of the embryos were in late syncytial blastoderm or cellular blastoderm formation. Unfertilised eggs and postblastoderm stages were dis-carded before heat shock. Embryos were heat shocked at 36°C for 15 minutes in a humidified chamber on a constant temperature block. After heat shocking, the embryos were allowed to recover for 30 minutes at 25°C and were then rinsed with heptane to remove the halocarbon oil and dechorionated in 50% Chlorox.

Fixation, devitellinisation and antibody staining of embryos was as described in DiNardo and O’Farrell (1987) using the Vectas-tain Elite ABC kit. The polyclonal ttk serum and the anti-ftz anti-body (a gift from H. Krause), were used at a concentration of 1:500 after preabsorbing at a 1:5 dilution against Oregon R embryos. The color reaction was developed using the peroxidase substrate kit DAB (Vector Laboratories). Embryos were mounted in Permount (Fisher Biotech.) and examined using Nomarski optics.

In situ hybridisation to detect mRNA was based on the method of Tautze and Pfeifle (1989) with modifications according to Rick Garber (personal communication). After fixation and devitellinisation, the embryos were transferred to EtOH and were stored (for up to 6 months) at −20°C. For staining, the embryos were rinsed in 50% EtOH, 50% xylene and soaked in 100% xylene for several hours to reduce the background; they were then rinsed again with 50% EtOH, 50% xylene followed by several rinses in 100% EtOH (rinses are for 5 minutes each unless otherwise stated). The embryos were then rinsed in MeOH, followed by 50% MeOH, 50% PBT (PBS [130 mM NaCl, 7 mM Na₂HPO₄, 3 mM NaH₂PO₄] with 0.1% Tween 20) plus 5% formaldehyde and fixed for 20 minutes in PBT plus 5% formaldehyde. After fixing, the embryos were rinsed three times in PBT (2 minute rinses) and then digested with 50 μg/ml proteinase K (Boehringer Mannheim) for 4 minutes at 37°C. Digestion was terminated by rinsing 2×2 minutes with 2 mg/ml glycine in PBT followed by two rinses in PBT. The embryos were postfixed for 20 minutes in PBT plus 5% formaldehyde. The fixative was removed and the embryos were washed 5×5 minutes with PBT. The embryos were then rinsed in 50% PBT, 50% hybridization solution (50% deionised formamide, 5× SSC, 100 μg/ml sonicated, boiled salmon sperm DNA, 100 μg/ml tRNA, 50 μg/ml heparin, 0.1% Tween 20) and incubated in 100% hybridization solution at 48°C for 2 hours. Hybridisation was at 48°C overnight in 100 μl of hybridization solution containing 0.1 μg/ml heatdenatured DNA probe.

DNA probes were prepared from gelpurified restriction fragments of pairrule gene cDNA clones. The ftz probe was 1.7 kb EcoRI fragment (gift of A. Laughon and M. Scott); the even skipped probe was a 1.1 kb HindIII-EcoRI fragment (gift of M. Levine and M. Scott); the runt probe was a 2.2 kb BamHI fragment (gift of A. Manoukian, H. Krause, and P. Gergen); the hairy probe was a 2.0 kb EcoRI fragment (gift of M. Horner and C. Rushlow); the odd skipped probe was 400 bp EcoRI fragment (gift of D. Coulter); and the Kruppel and hunchback probes were 2.1 kb PstI/EcoRI and 2.3 kb Xba I fragments, respectively (obtained through D-H. Hwang). 100 ng of DNA fragment was incubated in a total volume of 15 μl with 100 μg random primers, pd(N)₆ (Pharmacia), denatured by boiling 5 minutes and quick chilled on ice. 2 μl of 10× Vogel buffer (1 M Pipes (pH 6.6), 50 mM MgCl₂, 0.6 mM β-mercaptoethanol), 2 μl deoxynucleotide mix containing digoxigenin-UTP (Genius kit, Boehringer Mannheim), and 5-10 units of Klenow enzyme were added. The labelling reaction was incubated overnight at 15°C and then for 4 hours at room temperature after the addition of a further 5-10 units of Klenow. The reaction was stopped by adding 0.5 M Na₂EDTA to 50 mM and incubating at 65°C for 10 minutes. The probe was precipitated at −70°C after the addition of 80 μl of H₂O, 50 μg tRNA, 5 M LiCl to 0.4 M and 300 μl ethanol.

After hybridization, excess probe was removed and the embryos were washed for 20 minutes at 48°C with hybridization solution, 50% hybridization solution/50% PBT, and 5 washes with PBT. The embryos were then incubated for 1 hour at room temperature with 1:2000 dilution of antidigoxigenin alkalinephosphateconjugated antibodies (Genius kit) in PBT with 0.1% BSA (the antibodies were preabsorbed at a 1:5 dilution against an equal volume of embryos). The embryos were washed four times for 20 minutes with PBT and then rinsed 2× 20 minutes in 100 mM NaCl, 50 mM MgCl₂, 100 mM Tris pH 9.5, 0.1% Tween 20. To the second wash (1 ml), 4.5 μl nitroblue tetrazolium (NTB) and 3.5 μl of X-phosphate (Genius kit) were added for the color reaction and were incubated for several hours in the dark until the color developed. The reaction was stopped by rinsing several times with PBT. The embryos were stored overnight at 4°C, mounted in Permount (Fisher Biotech) and examined using Nomarski optics.

Embryos were collected and heat shocked as described above. After heat shock the embryos were returned to a humid chamber at 25°C for at least 24 hours to allow development to proceed. The number of hatched versus unhatched embryos was scored under the microscope for each class. The embryos were then rinsed with heptane to remove the halocarbon oil, dechorionated in 50% chlorox for 3 minutes, rinsed with embryo wash, dried, transferred to a microscope slide with a drop of 1:1 acetic acid/Hoyers reagent (as described in Wieschaus and Nusslein-Volhard, 1986) and incubated overnight at 65°C. The cuticle preparations were then examined by phasecontrast microscopy.

The tramtrack gene encodes two related proteins of predicted molecular masses 69×10³ (p69) and 88×10³ (p88), with different DNA-binding specificities generated through alternative splicing of related zincfinger motifs (Read and Manley, 1992a). Complementary DNAs for ttk p69 have been previously cloned and shown to interact specifically with ftz regulatory elements (Harrison and Travers, 1988, 1990; Brown et al., 1991). We have raised rat polyclonal antibodies to fulllength recombinant ttk p69 and used them to analyze the distribution of ttk protein during embryogenesis by immunohistochemical staining.

Antibodies to ttk p69 react with one major and three minor ttk protein-DNA complexes that are formed when embryo extracts are incubated with a segment of the ftz zebra element and analyzed by native gel electrophoresis (Fig. 1A). The putative ttk proteins (previously named FTZ-F2 A-D and defined by their specific binding to TGC_AGGAC_T sequences on the zebra element [Brown et al. 1991]) are all supershifted by treatment with immune serum but not with preimmune serum. Hence, all four gel shift complexes are composed of multiple forms of ttk p69 (or a highly related antigen) that could result from covalent modification, proteolytic degradation or association as homomeric or heteromeric complexes.

The specificity of the antibodies to ttk p69 was determined by western blot analysis of embryonic nuclear extracts after separation on a SDS-polyacrylamide gel. The antibodies showed reaction to a major protein species of M_r 92×10³, which corresponds to the anomalously migrating, fulllength p69 protein expressed in E. coli (Fig. 1B; Read and Manley, 1992a). Anomalous migration of other developmental transcription factors on SDS gels, e.g. ftz (Krause et al. 1988) and Kruppel (Ollo and Maniatis, 1987) has been observed previously. Since fulllength ttk p69 shares significant sequence identity with the ttk p88 isoform, we expected substantial crossreaction of our antiserum with the latter isoform. However, the antiserum only faintly stained a band of M_r ∼150×10³ on the western blot (Fig. 1B), corresponding to the anomalous migration of the p88 isoform (Read and Manley, 1992a). The preferential staining of ttk p69 over p88 was observed with nuclear extracts prepared from embryos staged throughout embryonic development, and with the bacterially expressed ttk isoforms (Fig. 1C). Hence, the results indicate that the antibodies are primarily directed against epitopes unique to p69, the ttk isoform that has been implicated in the repression of ftz.

We employed the antibodies to ttk p69 to determine the pattern of p69 expression at various stages throughout embryogenesis (Fig. 2). At the earliest stages observed (prestage 3 embryos; stages according to Campos-Ortega and Hartenstein, 1985) staining for ttk protein, presumably of maternal origin, can be detected throughout the embryo (Fig. 2A). This early appearance of ttk does not seem to accumulate exclusively in the cleavage nuclei. The general distribution of ttk persists through the syncytial blastoderm stage and staining for ttk is also observed in the pole cells (Fig. 2B). During formation of the cellular blastoderm, staining for ttk diminishes, and staining is completely absent by the completion of the cellular blastoderm (Fig. 2C), a period when ftz expression in stripes is at its height.

The zygotic expression of ttk is first evident around late embryonic stage 9-early stage 10, in the anterior midgut primordium (Fig. 2D). This is followed in stage 10 by the staining of isolated cells along the mesodermal surface of the yolk space (Fig. 2E). These cells may correspond to the precursors of macrophages, which have been identified morphologically by Campos-Ortega and Hartenstein (1985) in stage 11 embryos. Weak staining of the pole cells is also detectable at this stage. Subsequently, ttk expression is observed in the posterior midgut primordium and in patches of cells along the embryo that correspond to staining of the tracheal placodes. The salivary gland placode and the head region also show staining at this stage (Fig. 2F). By the time of germband retraction, the pattern becomes harder to discern. There is strong staining of the visceral mesoderm and weaker staining of the ectoderm (Fig. 2G). At full germband retraction, the staining is more general, although the staining is stronger in the visceral mesoderm and the ectoderm of the embryo. There is also weak staining of the amnioserosa (Fig. 2H).

The distribution of ttk protein that we have observed during embryogenesis is similar to the distribution of ttk RNA (Harrison and Travers, 1990; Read and Manley, 1992a,b; Brown and Wu; unpublished observations), although the appearance of ttk protein lags behind that of ttk RNA by a short interval (about 30 minutes). The presence of ttk protein in preblastodermstage embryos is consistent with its proposed role as a maternal repressor of ftz. The complex zygotic expression of ttk in late embryos suggests that ttk has multiple regulatory functions separate from the repression of ftz at specific stages throughout embryogenesis.

In an earlier study, we proposed that ttk functions as a negative regulator of ftz, since point mutations that eliminate ttk binding to the zebra element cause derepression of ftzlacZ constructs in transformed embryos (Brown et al., 1991). To test this hypothesis, we generated flies carrying ttk under the control of the heatshock promoter and examined the effects of ectopic expression of ttk on ftz expression. The fulllength ttk p69 cDNA clone (with the entire 5′ leader sequence removed; see Methods) was inserted into the pCaSpeRhs vector as shown in Fig. 3A. The ability of the ttk construct lacking 5′ leader sequences to be translated was confirmed by in vitro transcriptiontranslation assays before proceeding with P elementmediated germline transformation. Two independent transformant lines were established as homozygous stocks.

Since blastodermstage embryos represent the earliest developmental period when the heatshock response can be elicited, embryos were harvested and aged at 25°C until the late syncytial blastoderm or early cellular blastoderm stages. Embryos were then heat shocked for 15 minutes at 36°C and allowed to recover for 30 minutes at 25°C before fixation and analysis by wholemount in situ hybridisation or immunohistochemical staining. The lack of ttk protein expression at the cellular blastoderm stage of unshocked embryos carrying phsttk is confirmed in Fig. 3B, panel (a). Upon heat shock, these embryos showed extensive ttk protein staining with clear nuclear localization essentially throughout the length of the embryo. In some cases, phsttk expression seems to be slightly lower in the polar regions and is not evident in the pole cells (Fig. 3B, panel (b). The halflife of ectopic ttk protein was found to be approximately 90 minutes.

Embryos collected from flies lacking or carrying the phsttk construct were heat shocked in parallel and stained for ftz RNA expression by in situ hybridization (Fig. 4). As shown in Fig. 4A,B, the normal expression of ftz in seven stripes is not affected by heat shock of embryos derived from the strain lacking phsttk. However, heatshockinduced expression of ttk in embryos carrying phsttk caused a drastic reduction in the expression of ftz in the entire population of shocked embryos. Out of a cohort of 150 embryos examined, 60% showed no detectable ftz RNA expression (Fig. 4D) and the remaining 40% showed only weak, residual staining in stripes 2,3 and 7 (see Fig. 4E). 100% of nonheatshocked embryos carrying phsttk showed the normal staining for ftz in 7 stripes (Fig. 4C). The above observations were similar for both transformant lines analyzed.

When heatshocked embryos carrying phsttk were stained for ftz protein expression by immunohistochemistry, the same trend was observed, although the appearance of ftz protein lagged slightly behind that of ftz RNA, so that most embryos had residual ftz expression in at least some of the seven stripes (Fig. 5). Nonetheless, the residual expression in the stripes was limited to small patches in embryos expressing phsttk and in no case was the threeto fourcell width of normal ftz expression retained (compare Fig. 5E and F).

In order to analyze the phenotype of embryos after ectopic expression of ttk, embryos carrying phsttk were collected and allowed to develop at 25°C to late syncytial blastoderm or early cellular blastoderm stages. The embryos were then heat shocked for 15 minutes at 36°C, returned to 25°C and permitted to develop for another 24 hours. The numbers of hatched larvae and unhatched larvae or embryos were scored by microscopic analysis. Typical results of such an experiment are shown in Table 1. No embryos carrying phsttk survived after the heat shock to hatch from the eggshell (100% lethality) while the lethality was 20-50% for the control embryos. Cuticle preparations from the unhatched embryos carrying phsttk were analyzed. All of the heatshocked embryos examined showed a range of cuticular abnormalities that were not seen with the nonheatshocked controls or with the heatshocked Df(1)w67c2,y line. The most predominant class of defects observed (about 50%) had naked cuticles completely devoid of denticle belts (Fig. 6D). Other cuticular abnormalities included defects in the number, spacing or appearance of the denticle belts (Fig. 6B,C); some embryos did not secrete any cuticle at all. In addition, many of the embryos also showed severe abnormalities of the mouthparts. This highly nonspecific phenotype indicates that the ectopic expression of ttk throughout the blastodermstage embryo must interfere with the normal expression or function of a number of genes besides ftz that are important for embryogenesis. The wide range of phenotypes observed could be due to the precise developmental age of the individual embryo when phsttk was expressed, leading to a different or overlapping set of affected genes or functions.

In order to test the possibility that ttk might be involved in the regulation of genes other than ftz, we analyzed the expression of four other members of the pairrule class of segmentation genes. As shown in Fig. 7, embryos expressing phsttk cause significant repression of runt and eve and also showed effects on hairy and odd skipped. Like ftz, the runt and eve RNA expression patterns are severely repressed in the heatshocked embryos, although some residual expression is observed in stripes 1,2 and 7 (see Fig. 7A-D). More than 90% of the heatshocked embryos show these effects whereas the nonheatshocked line carrying phsttk, and the nonheatshocked or heatshocked Df(1)w67c2,y embryos showed the normal expression patterns for runt and eve, as described by Gergen and Butler (1988), and MacDonald et al. (1986), respectively.

In the case of hairy and odd skipped the repression is more localized. For hairy, 92% of the heatshocked embryos carrying phsttk showed a severe repression of stripes 3 and 4 but virtually no effect on the other stripes (Fig. 7F). This effect was not observed with any of the heatshocked Df(1)w67c2,y embryos, or the unshocked line 4C carrying phsttk (Fig. 7E). For odd skipped, repression is seen in stripes 1, 5, 6 and 7, but not stripes 2, 3 and 4 (Fig. 7H); there also appears to be increased odd skipped expression in the regions between stripes 2, 3 and 4. We have investigated the possibility that the ectopic expression of ttk also leads to repression of the gap genes. As shown in Fig. 8, we failed to observe any effects of ectopic ttk on the expression of the gap genes Kruppel and hunchback.

In this paper, we have tested the hypothesis that ttk p69 acts as a direct, maternal repressor whose decay and dilution in relation to the exponentially dividing cleavage nuclei leads to the onset of ftz transcription in blastodermstage embryos (Brown et al., 1991). We show that the endogenous distribution of ttk p69 protein as revealed by immunohistochemical staining of preblastodermstage embryos is consistent with its proposed role as a maternal repressor of ftz. The absence of ttk p69 at the cellular blastoderm stage is also consistent with the peak of ftz expression at that stage. Most importantly, the ectopic production of ttk p69 in blastodermstage embryos carrying a heat shockttk construct causes complete or nearcomplete repression of the endogenous ftz gene, and a significant repression of even skipped, odd skipped, hairy and runt. No effects were observed on the expression of the gap genes Kruppel and hunchback. In an independent study, Read and Manley (1992b) have also found that expression of ttk p69, but not ttk p88, leads to repression of evenskipped and ftz, but not hunchback, Kruppel and giant. The combined results offer strong support for the hypothesis that maternal ttk p69 functions as a specific repressor of ftz and suggest further that this repression may extend to at least several other pairrule segmentation genes.

That there should be a requirement for specific repression of ftz in preblastodermstage embryos is unusual, since it has been generally accepted that there is no zygotic transcription prior to nuclear division cycle 9 or 10 (Lamb and Laird, 1976; McKnight and Miller, 1976; Anderson and Lengyel, 1979; Edgar and Schubiger, 1986). A priori, a lack of transcription could be caused either by a lack of activators (general or specific) or by the presence of repressors. To our knowledge, current evidence on transcriptional controls at the preblastoderm stage do not preclude specific repression as one mechanism (among others) for the overall lack of preblastoderm transcription. Nonetheless, it is generally assumed that the extreme rapidity with which cleavage nuclei divide during the first nine nuclear division cycles could be sufficient to prevent significant transcription in preblastoderm embryos. Indeed, early cleavage nuclei divide approximately every nine minutes, with an interphase or S phase of about 4 minutes (Rabinowitz, 1941; McKnight and Miller, 1977; Foe and Alberts, 1983). Despite this narrow time window available for transcription, it is possible that small genes (less than 5 kb) could be fully transcribed in under 4 minutes, assuming a rate of 1.1 to 1.4 kb per minute for RNA polymerase II [Thummel et al. (1990); Shermoen and O’Farrell, (1991)]. While transcripts from large developmental genes such as Ubx would be aborted by the interference of rapid mitotic cycles (Shermoen and O’Farrell, 1991), other repressive mechanisms must prevail for small genes, which constitute a significant fraction of the Drosophila genome.

In addition to ftz, the four pairrule genes affected by ectopic ttk expression are all under 5 kilobases in size (Coulter et al., 1990; Kania et al., 1990; MacDonald et al., 1986, Laughon and Scott, 1984; Rushlow et al., 1989). Hence, they could in principle undergo complete transcription in preblastoderm embryos and therefore require specific repression by ttk until the blastoderm stage has been reached. For hairy and odd skipped, which are affected by ectopic ttk expression in only a subset of their stripe patterns, it would be necessary to invoke negative regulators other than ttk that would ensure the complete repression of those genes in the preblastoderm stage. It should be emphasized, however, that a direct interaction between ttk p69 and DNA has only been shown for ftz and even skipped (Brown et al. 1991; Read and Manley, 1992a). For even skipped, the location of negative cis-elements that interact directly with ttk has not been determined. Further studies will be required to ascertain whether the repression of hairy, runt, even skipped and odd skipped that we have observed is actually dependent on direct interactions with ttk p69 or is the result of crossregulatory interactions between pairrule genes. The results could also be caused by artefactual interactions between ttk p69 and stripespecific activators that are unrelated to the normal function of ttk. It is interesting to note that hairy stripes 3 and 4, which are repressed by ectopic ttk expression, are located in the domain of action of the Kruppel gene, raising the possibility of competition between these two zinc finger proteins. The lack of ttk repression of gap genes indicates that the subjects of ttk repression are somewhat restricted, and leave open the question whether other early embryonic genes are under repression by a different mechanism in the preblastoderm stage.

Our previous observation of the expression in preblastoderm embryos of a zebra element-lacZ construct with mutated ttk binding sites suggests that general or specific activators must be present in preblasotoderm embryos that interact with the 740 bp zebra element and drive expression of β-galactosidase. Two maternally provided activators that bind to the zebra element have been identified so far: caudal (Dearolf et al., 1989b) and FTZ-F1 (Ueda et al 1990). The orphan hormone receptor FTZ-F1 is uniformly distributed throughout the preblastoderm embryo (G. Lavorgna and C. Wu, unpublished results) and behaves as an activator of all the ftz stripes, with greater effects on stripes 1,2,3 and 6 (Ueda et al., 1990). The caudal protein is expressed in the posterior of the embryo (MacDonald and Struhl, 1986) and has an effect on ftz activation in this region of the embryo (Dearolf et al., 1989b). It is possible that ttk p69 may act to antagonize the positive influences of FTZ-F1 and caudal on RNA polymerase II by blocking the binding of these factors to DNA, to other components of the transcriptional machinery, or indirectly by forming nonproductive interactions with general transcription factors. Whatever the mechanism of preblastoderm repression by ttk, such repression should be necessary to prevent premature initiation of the ftz-dependent, positive autoregulatory loop until the necessary positional cues are in place (Hiromi and Gehring, 1987; Pick et al., 1990).

The reemergence of ttk p69 in a highly complex pattern of expression subsequent to the cellular blastoderm stage indicates that the zygotically expressed p69 protein has additional regulatory functions aside from ftz repression. The pleiotropic nature of ttk is also indicated by the wide range of cuticular abnormalities observed for embryos that have undergone ectopic expression of ttk p69. Such embryos die before hatching and do not exhibit a specific pairrule phenotype. In addition, a ttk null mutation has recently been isolated and reveals an embryoniclethal phenotype (Xiong and Montell, personal communication). In order to test genetically the regulation of ftz by ttk, it will be necessary to eliminate the maternal contribution by the creation of germline clones homozygous for the ttk null mutation.

The model that we have proposed for preblastoderm repression of ftz by maternal ttk protein is not the only system of repression for this pairrule segmentation gene. Previous microinjection experiments using the protein synthesis inhibitor cycloheximide have revealed two systems by which ftz transcription is repressed in early embryos. A polar repression system has been proposed to repress ftz expression in the anterior 35% and posterior 15% of the embryo, and a periodic system of repression is responsible for repression in the interstripe regions (Edgar et al., 1986). Our studies have elucidated a previously unforseen requirement for repression at a time when a need for repression was not generally anticipated. We suggest that the number of genes in the Drosophila genome that may be subject to preblastoderm repression by ttk or by other maternal repressors should not be underestimated.

We thank Judy Kassis for invaluable assistance and advice in setting up the germline transformation and wholemount embryo staining techniques, and Jim Manley and Craig Montell for sharing their data prior to publication. We also thank Carl Thummel and Vince Pirrotta for pCaSpeRhs, Henry Krause for antibodies to ftz; Matthew Scott, Mike Horner, Chris Rushlow, Mike Levine, Allan Laughon, Sean Carroll, Peter Gergen, Armen Manoukian, Henry Krause, Doug Coulter, Doug Read, John Colgan, and Jim Manley for gifts of cDNA clones and hybridization probes; Rick Garber and Henry Krause for sharing in situ hybridization protocols; and Jim Kennison and Judy Kassis for helpful suggestions and comments on the manuscript.