The DNA Replication-related Element (DRE)/DRE-binding Factor System Is a Transcriptional Regulator of the Drosophila E2FGene*

Two mRNA species were observed for theDrosophila E2F (dE2F) gene, differing with regard to the first exons (exon 1-a and exon 1-b), which were expressed differently during development. A single transcription initiation site for mRNA containing exon 1-b was mapped by primer extension analysis and numbered +1. We found three tandemly aligned sequences, similar to the DNA replication-related element (DRE; 5′-TATCGATA), which is commonly required for transcription of genes related to DNA replication and cell proliferation, in the region upstream of this site. Band mobility shift analyses using oligonucleotides containing the DRE-related sequences with or without various base substitutions revealed that two out of three DRE-related sequences are especially important for binding to the DRE-binding factor (DREF). On footprinting analysis with Kc cell nuclear extracts and a glutathioneS-transferase fusion protein with the N-terminal fragment (1–125 amino acid residues) of DREF, all three DRE-related sequences were found to be protected. Transient luciferase expression assays in Kc cells demonstrated that the region containing the three DRE-related sequences is required for high promoter activity. We have established transgenic lines of Drosophila in which ectopic expression of DREF was targeted to the eye imaginal disc cells. Overexpression of DREF in eye imaginal disc cells enhanced the promoter activity ofdE2F. The obtained results indicate that the DRE/DREF system activates transcription of the dE2F gene.

sites in their promoters. Transcription of these genes is induced by E2F family members during progression from G 0 to S phase (2).
In general, the levels of DP1, DP2, and E2F4 proteins are constant throughout the cell cycle (5,6), whereas dramatic increase in expression of E2F1 and E2F2 genes during progression from G 0 to S phase has been reported (7,8). The mechanisms of transcriptional regulation have been studied (8 -10), and E2F1 and E2F2 genes contain E2F-binding sites in their promoters, their expression being repressed by E2F-mediated negative control through these sites during G 0 to G 1 phase (8 -11). The repression is likely mediated by an E2F-Rb family complex. In late G 1 phase, E2F-mediated repression is relieved by G 1 cyclin-dependent kinase-mediated phosphorylation of a Rb family protein (12). Then, transcription factors like CCAATbinding protein and YY-1 appear to activate the E2F1 gene promoter (11,13), while transcription factor Myc activates the E2F2 gene promoter (8).
In Drosophila, a single set of E2F and DP has been identified and characterized (14,15), but two different E2F cDNAs have been reported (14,15). Although these two cDNAs encode an identical protein, the nucleotide sequences in their 5Ј-untranslated regions differ.
Comparison of the newly determined genomic sequences of the Drosophila E2F gene (dE2F) and its 5Ј-flanking region 2 with the cDNAs nucleotide sequences allowed us to map two genomic regions, corresponding to the diverged sequences of two E2F cDNAs. These two genomic regions likely represent parts of two different first exons, named 1-a and 1-b, that are, respectively, located 10.1 and 5.4 kb upstream of the common second exon containing the translation initiation codon. Northern hybridization analyses using specific probes confirmed the existence of two transcripts, named transcripts-a and -b. Although transcript-a was detected exclusively in embryos, transcript-b could be identified in other stages of development. Fluctuation of transcript-b level was roughly similar to those of other DNA replication-related genes (16,17). Thus, we were interested in knowing whether synthesis of transcript-b is regulated by the same factors involved in expression of DNA replication-related genes.
Like their mammalian counterparts, Drosophila E2F and DP form a heterodimer to generate a sequence-specific DNA binding factor that appears to activate transcription of DNA replication-related genes through E2F-binding sites (14, 15, 18 -22).
In addition to E2F-binding sites, a common 8-base pair (bp) palindromic sequence named the DNA replication-related element (DRE: 5Ј-TATCGATA) is frequently found in the promoters of growth-related genes (23)(24)(25)(26)(27). A transcription factor DREF (DRE-binding factor) that specifically binds to the DRE sequence has been isolated as an 80-kDa polypeptide homodimer (23), and its cDNA has been cloned (28). Our previous studies performed in vitro and in vivo suggested that E2Fbinding sites and DRE function synergistically to activate promoters of the PCNA and DNA polymerase ␣ 180-kDa subunit genes (23,29,30).
Interestingly, we found three tandemly aligned DRE-related sequences, each of which matched 6 bp out of the 8-bp DRE consensus sequence, located at positions Ϫ540 to Ϫ533 (site I), Ϫ532 to Ϫ525 (site II), and Ϫ521 to Ϫ514 (site III) with respect to the initiation site of transcript-b. The region from Ϫ2184 to ϩ203 showed promoter activity in Drosophila Kc cells with the three DRE-related sequences playing an important role. In addition, DREF specifically bound to these DRE-related sequences, and overexpression of DREF in Drosophila eye imaginal disc cells activated the expression of the lacZ reporter gene, which was under the control of the dE2F gene promoter. The obtained results indicate that transcription of the dE2F gene is regulated by the DRE/DREF system.

EXPERIMENTAL PROCEDURES
Cell Culture-Kc cells derived from Drosophila melanogaster embryos were grown at 25°C in M3(BF) medium (31) supplemented with 2% fetal calf serum in the presence of 5% CO 2 .
Isolation of Drosophila Genomic Clones-A genomic library was constructed by Sau3AI partial digestion of D. melanogaster Oregon-R DNA and ligation into the BamHI site of EMBL3. The details of the procedures for screening and nucleotide sequencing of the genomic clones for dE2F will be described elsewhere 2 (accession number AB011813).
The sequences of double-stranded oligonucleotides containing the DRE of the PCNA gene promoter (DRE-P), or a 3 base-deleted derivative (DRE-P⌬3), were as described previously (29).
Plasmid Constructions-All nucleotide positions of the dE2F gene in the following part of this paper are numbered with the transcription initiation site for transcript-b set as ϩ1. Plasmid DPA124SE4.5 contains the region corresponding to exon 1-b and at least 4 kb upstream region in the vector pBluescript II SK(Ϫ). Plasmid DPA124SS5.2 contains the region corresponding to exon 1-a and at least 3.5 kb upstream in the vector pBluescript II SK(Ϫ). 2 The expression plasmid Act-dE2F (15) contains a dE2F full-length cDNA placed under the control of the Drosophila actin 5C gene promoter (32).
To construct the plasmid pRL-actin5C for the internal control of the luciferase-transient expression assay, pActCAT was digested with SacI and HindIII, and the DNA fragment that contained the actin 5C gene promoter was isolated and blunt-ended using T4 DNA polymerase. Then, this DNA fragment was inserted into EheI site of the plasmid pRL-null vector (Promega).
To construct the plasmid p-3x6/8-pBluescript II SK(Ϫ) for the footprinting analysis, plasmid DPA124SE4.5 was digested with PvuII, and the DNA fragment containing the region from Ϫ843 to Ϫ440 was isolated. Then, this fragment was inserted into the EcoRV site of the plasmid pBluescript II SK(Ϫ).
To construct the plasmid p-2184E2Fbluc for the luciferase-transient expression assay, plasmid DPA124SE4.5 was digested with NdeI and EcoRV, and the DNA fragment containing the dE2F gene fragment was isolated and inserted into blunt-ended BglII site of the plasmid PGVB. To construct the plasmid p-1159E2Fbluc, PCR was performed using plasmid DPA124SE4.5 as a template, and primers ϩ8230X and Ϫ9610HS2 in combination. The PCR products were digested with XhoI and HindIII and inserted between the XhoI and HindIII sites of plasmid PGVB. To construct the plasmid p-913E2Fbluc, plasmid DPA124SE4.5 was digested with EheI and EcoRV, and the DNA fragment containing the dE2F gene fragment was isolated and inserted into the blunt-ended BglII site of the plasmid PGVB. To construct the plasmid p-691E2Fbluc, plasmid DPA124SE4.5 was digested with DraI and EcoRV, and the DNA fragment containing the dE2F gene fragment was isolated and inserted into the blunt-ended BglII site of the plasmid PGVB. To construct the plasmid p-613E2Fbluc, PCR was performed using plasmid DPA124SE4.5 as a template, and primers ϩ8771 and Ϫ9610HS2 in combination. The PCR products were digested with XhoI and HindIII, and inserted between the XhoI and HindIII sites of plasmid PGVB. To construct the plasmid p-440E2Fbluc, plasmid DPA124SE4.5 was digested with PvuII and EcoRV, and the DNA fragment containing the dE2F gene fragment was isolated and inserted into the blunt-ended BglII site of the plasmid PGVB. To construct the plasmid p-34E2Fbluc, plasmid-DPA124SE4.5 was digested with NaeI and EcoRV, and the DNA fragment containing the dE2F gene fragment was isolated and inserted into the blunt-ended BglII site of the plasmid PGVB.
To construct the luciferase expression plasmid having three DRErelated sequences placed upstream of the heterologous TATA promoter (Fig. 9), the oligonucleotides DRE-Eb or DRE-Eb (mut I II III) were ligated in a head-to-tail manner as described previously (23). The ligated oligonucleotides were isolated and digested with BglII and BamHI, and inserted into the BamHI site of the p-TATA-SK. The fragments containing DRE and TATA promoter were isolated by digestion with SmaI and SacI, and inserted between SmaI and SacI sites of PGVB. For p-TATA-SK, TATA-CAT (23) was digested with XbaI, after which the DNA fragment containing metallothionein basal promoter was isolated, and inserted into the XbaI site of pBluescript II SK(Ϫ). To construct the control plasmid TATA-PGVB carrying the metallothionein gene basal promoter alone, TATA-SK was digested with KpnI and SacI, after which the DNA fragment containing metallothionein gene basal promoter was isolated and inserted into the KpnI and SacI sites of PGVB.
To construct the plasmid pGMR-GAL4, PCR was performed using plasmid pGaTB as a template, and primers GAL4ϩ and GAL4Ϫ in combination. The PCR products were digested with EcoRI and BglII, and inserted between the EcoRI and BglII sites of plasmid pGMR. To construct the pUAS-DREF, PCR was performed using plasmid DC-DREF2.2 (23) as a template, and primers DREF-ATG and T7 primer (Toyobo) in combination. The PCR products were digested with BamHI and XhoI, and then inserted between BamHI and XhoI sites of pUAST (33).
All plasmids were propagated in Escherichia coli XL-1 Blue, isolated by standard procedures (34) and further purified through two cycles of ethidium bromide/CsCl density-gradient centrifugation.
Northern Blot Hybridization Analysis-Total cellular RNA was isolated from bodies of Drosophila at various developmental stages by the acid guanidium thiocyanate-phenol-chloroform extraction method (35). Twenty micrograms of total RNA were separated on a 1% agarose gel containing formaldehyde and blotted onto a sheet of GeneScreen Plus membrane (DuPont). Probes were radiolabeled using the random primer method (36). Hybridization and washing conditions were the same as described elsewhere (16). Blots were exposed to Kodak X-Omat XAR films or quantified with a BAS2000 (Fuji Film) imaging analyzer. The exon 1-a specific probe ( Fig. 1) was prepared by means of PCR using plasmid DPA124SS5.2 as a template and primers 5.2ss3761ϩ and 5.2ss4777Ϫ in combination. The exon 1-b specific probe ( Fig. 1) was prepared by means of PCR using plasmid DPA124se4.5 as a template and primers 4.5se9007ϩ and 4.5se10049Ϫ. Each PCR product was gel-purified and radiolabeled. The coding region specific probe was obtained by digestion of Act-dE2F with HindIII and BamHI. 1.2 kb of fragment was gel-purified and radiolabeled. Full-length cDNA of ribosomal protein 49 (Rp-49) gene was used as a control.
Primer Extension-Primer extension analysis was performed as detailed earlier (25). Total cellular RNA was extracted from Drosophila 8 -12-h embryos, unfertilized eggs, third instar larvae, and Kc cells by the methods described previously (35). A 34-mer primer (4.5pet), which was complementary to the region downstream of the 5Ј-end of the cDNA reported by Dynlacht et al. (15), was chemically synthesized. The reaction product was analyzed by gel electrophoresis under denaturing conditions, followed by autoradiography. 35 S-Labeled DNA fragments, produced in the dideoxy-sequencing reaction with the plasmid DPA124SE4.5 as a template using 4.5pet as a primer, were run in parallel, allowing precise mapping of the transcription initiation site.
Band Mobility Shift Analysis-Band mobility shift analysis was performed as detailed previously (23) with minor modifications. An expression plasmid for the glutathione S-transferase (GST)-DNA binding domain of a DREF (DREF1-125) fusion protein was constructed, and the fusion protein was expressed in E. coli as described previously (28). Kc cell nuclear extracts were prepared as described previously (23) and used for the band mobility shift analysis. The oligonucleotides DRE-Eb or DRE-P were end-labeled with 32 P, and DNA-protein complexes were electrophoretically resolved on a 6% polyacrylamide gel in 50 mM Tris borate, pH 8.3, 1 mM EDTA containing 2.5% glycerol at 25°C. The gels were dried and autoradiographed or quantified with a BAS2000 (Fuji Film) imaging analyzer.
DNA Transfection and Luciferase Assay-Kc cells (2 ϫ 10 6 /dish) were grown in 60-mm plastic dishes for 24 h and cotransfected with 5 g of the reporter plasmid DNA and 10 ng of pRL-actin5C DNA using Cell-Fectin reagent (Life Technologies, Inc.) (37). Cells were harvested at 48 h after DNA transfection. The luciferase assay was carried out by means of a Dual-Luciferase reporter assay system (Promega). Firefly luciferase activities were normalized to Renilla luciferase activities. Transfections were performed several times with at least two independent plasmid preparations.
DNase I Footprinting Analysis-DNase I footprinting analysis was performed essentially as described previously (23). The DNA fragment (Ϫ440 to Ϫ843) was obtained by digestion of p-3x6/8-pBluescript II SK(Ϫ) with BamHI (lower) and HindIII (upper), labeled at 5Ј-end of the upper or lower strand. After electrophoresis, gels were dried and autoradiographed.
Fly Strains-Fly stocks were maintained at 25°C on standard food. The Canton S fly was used as the wild type strain. The dE2F 729 allele, described previously (19), was kindly supplied by Drs. A. Brook and N. Dyson.
Establishment of Transgenic Flies-P-element-mediated germ line transformation was carried out as described previously (38). F1 transformants were selected on the basis of white eye color rescue (39). Four independent lines were obtained with pGMR-Gal4 constructs and line 16 carrying pGMR-GAL4 on X chromosome was used. Five independent lines were obtained for pUAS-DREF constructs. The line carrying pUAS-DREF in the second chromosome was used in this study. The details of the procedures for establishment of lines with the UAS-DREF transgene will be described elsewhere. 3 Immunohistochemistry-The line carrying GMR-GAL4 in the X chromosome was crossed with the line carrying UAS-DREF in the second chromosome, and the resultant hybrids were crossed with dE2F 729 . The progenies were analyzed as detailed below. Third instar larvae were dissected in Drosophila Ringer solution, and imaginal discs were fixed in 4% paraformaldehyde/PBS for 20 min at room temperature. After washing with PBS/0.3% Triton X-100 (PBS-T), the samples were blocked with PBS-T containing 10% normal goat serum for 30 min at room temperature. Samples were incubated with rabbit anti-DREF IgG at a 1:2000 dilution or with mouse anti-␤-galactosidase monoclonal antibody at a 1:1000 dilution at 4°C for 16 h. After extensive washing with PBS-T, the imaginal discs were incubated with alkaline phos-phatase-conjugated goat anti-rabbit IgG (Promega) at a 1: 1000 dilution or with peroxidase-conjugated goat anti-mouse IgG (E-Y Laboratory) at a 1:500 dilution as second antibody. After extensive washing with PBS-T, color was developed in a solution containing 100 mM Tris-HCl, pH 9.5, 100 mM NaCl, 5 mM MgCl 2 , 0.34 mg/ml nitro blue tetrazolium salt, and 0.175 mg/ml 5-bromo-4-chloro-3-indolyl phosphate toluidinium salt for the alkaline phosphatase reaction and in a solution containing 0.5 mg/ml diaminobenzidine, 2.5 mM CoCl 2 , and 0.003% H 2 O 2 for the peroxidase reaction. The tissues were washed with PBS and mounted in 90% glycerol/PBS for microscopic observation.

RESULTS
Two Transcripts of Drosophila E2F Are Possibly Synthesized by Different Mechanisms during Development-Two different cDNAs of Drosophila E2F have been reported (14,15). Although these two cDNAs encode an identical protein, the nucleotide sequences in their 5Ј-untranslated regions are divergent. Comparison of the newly determined genomic sequence of dE2F gene and its 5Ј-flanking region 2 with those of cDNAs allowed us to map two genomic regions corresponding to the divergent sequences of the two E2F cDNAs. These two genomic regions likely represent parts of two different first exons, named 1-a and 1-b, that are, respectively, located 10.1 and 5.4 kb upstream of the common second exon containing the translation initiation codon (Fig. 1). We examined the dE2F mRNA level using total RNA from Drosophila bodies at various developmental stages (Fig. 2). Northern hybridization analysis using the exon 1-b specific probe detected on an approximately 4.2-kb transcript (transcript-b) at a relatively low level in unfertilized eggs and 0 -2-h embryos (Fig. 2). Transcript-b was detectable throughout developmental stages, although its size appeared slightly larger in 4 -24-h embryos, larvae, pupae, and adult flies (Fig. 2 and The primer was extended using reverse transcriptase as described under "Experimental Procedures." To align the extended products with the genomic sequence, a parallel dideoxysequencing reaction was carried out using the same 34-mer primer (lanes 1-4). The numbers on the right indicate the nucleotide positions from the transcription initiation site, defined as ϩ1. data not shown). In contrast, the exon 1-a specific probe detected a transcript of 4.7 kb (transcript -a) at highest levels in 4 -8-h embryos, and only at low levels in 0 -4-h and 8 -12-h embryos. Transcript-a was not detectable in unfertilized eggs, late stages of embryogenesis, larvae, and pupae. Since both the regions corresponding to the first exons are short, the signals detected by both probes were rather weak. The coding region specific probe detected both transcripts-a and -b (data not shown). A ribosomal protein 49 (Rp-49) cDNA was used as a reference probe to monitor the mRNA integrity. The results suggest that transcripts-a and -b are likely to be under the control of different regulatory mechanisms during development.
Determination of the Initiation Site of Transcript-b-Although the initiation site of transcript-a has been reported (40), that of transcript-b has hitherto not been determined. We therefore performed a primer extension analysis using a primer complementary to the 5Ј region of exon 1-b (Fig. 3). Using total RNA from unfertilized eggs, 8 -12-h embryos, third larvae, and Kc cells, a single transcription initiation site was mapped to 78 bp upstream from the 5Јend of the reported cDNA (15), and defined as nucleotide position ϩ1. It should be pointed out that the same transcription initiation site was used for both the maternally transcribed mRNA (Fig. 3, lane 5) and the zygotically transcribed mRNA (Fig. 3, lanes 6 -8).
DRE-related sequences, with a 6 out of 8 bp DRE consensus match, located at positions Ϫ540 to Ϫ533 (site I), Ϫ532 to Ϫ525 (site II), and Ϫ521 to Ϫ514 (site III) with respect to the transcription initiation site. A potential E2F-binding site (TTTGC-CGG) was found at positions Ϫ41 to Ϫ34. Although a DRErelated sequence was found at Ϫ213 to Ϫ206, band mobility shift assays revealed that this DRE-related sequence was not recognized by DREF (data not shown).

DREF Recognizes the Three DRE-related Sequences Located Upstream of the Initiation Site for Transcript-b-To determine
whether DREF can recognize the three DRE-related sequences located upstream of the initiation site of transcript-b, we performed band mobility shift assays using Kc cell nuclear extracts (Fig. 5, A-C). DRE-Eb is the oligonucleotide containing three DRE-related sequences located upstream of the initiation site of transcript-b of the dE2F gene. DRE-P is the oligonucleotide containing DRE from the promoter region of the Drosophila PCNA gene (23), used as a control. DRE-P⌬3 is a 3-basedeletion derivative of the DRE-P (29).
To examine whether the shift bands observed with the DRE-Eb probe represent a DRE/DREF complex, we examined effect of the anti-DREF monoclonal antibodies (mAbs 1 and 4) on the binding reaction with Kc cell nuclear extracts. mAb 1 inhibited the binding of DREF to DRE-P, and mAb 4 supershifted the band with DRE-P (20, 28) (Fig. 5C, lanes 10 -12). As shown in Fig. 5C, one of the shift bands with DRE-Eb probe was diminished by the addition of mAb 1 (lanes 2 and 3) and was supershifted by mAb 4 (lane 6). These results indicate that one of the DNA-protein complexes formed between DRE-Eb and Kc cell nuclear extracts contains DREF.
Role of the Three DRE-related Sequences in DREF Binding-To determine the nucleotide sequences in the DRE-Eb required for binding to DREF, a set of oligonucleotides having mutations in and around the three DRE-related sequences were chemically synthesized (Fig. 7A), and added to the binding reaction as competitors (Fig. 7, B and C). With the 32 Plabeled DRE-Eb, mutant oligonucleotide DRE-Eb (mut I) or DRE-Eb (mut IV) competed as effectively as wild type DRE-Eb (Fig. 6B, lanes 5-8 and 33-36). Mutant oligonucleotide DRE-Eb (mut II III) or DRE-Eb (mut I II III) did not compete for the binding (Fig. 6B, lanes 21-24 and 29 -32). These results indicate that sites II and III in DRE-Eb are important for the binding to GST-DREF1-125.
Footprinting analyses with the Kc cell nuclear extracts and GST-DREF1-125 fusion protein demonstrated protection of the region from Ϫ540 to Ϫ514, which covers not only sites II and III but also site I (Fig. 8, A and B), and the Kc cell nuclear extract also protected a similar region (Fig. 8, A and C). The data provide evidence that the DREF-binding to site I is dependent on sites II and III.
In addition, it should be noted that another protection was observed with the Kc cell nuclear extract in the regions from Ϫ635 to Ϫ615 and Ϫ709 to Ϫ674, both containing a homeodomain protein-binding consensus, TAAT (41) (Figs. 4 and 8A).
The Three DRE-related Sequences Are Required for High Promoter Activity of the dE2F Gene-Genomic fragments from Ϫ2184 to ϩ203 of the dE2F gene containing the initiation site of transcript-b and its 5Ј-deletion derivatives were ligated with the luciferase reporter gene, and then the constructs were transfected into Kc cells to determine promoter activities (Fig.  9). Deletion from positions Ϫ2184 to Ϫ1159 resulted in 40% reduction of luciferase expression. A CFDD (common regulatory factor for DNA replication and DREF genes) recognition site, which is important for the promoter activity of the PCNA gene (42), was found at Ϫ1303 to Ϫ1299, although we have not yet confirmed whether this is responsible for the promoter activity. Deletion from Ϫ1159 to Ϫ913 decreased luciferase expression to 16%. Further deletions up to Ϫ613 did not show any additional significant change in luciferase expression. However, a deletion from Ϫ613 to Ϫ440 decreased luciferase expression to 1%, indicating the existence of strong positive regulatory elements within this region. As noted above, it contains the three DRE-related sequences (Ϫ540 to Ϫ514), sug- gesting that these are important for the promoter activity. A further deletion to position Ϫ34 almost completely abolished the promoter activity.
For further confirmation of the stimulatory effects of the three DRE-related sequences on the promoter, we prepared luciferase expression constructs having the wild type DRE-Eb or DRE-Eb (mut I II III) with mutations in all three DRErelated sequences ligated in a head-to-tail manner with the basal promoter of the Drosophila metallothionein gene, and transfected them into Kc cells (Fig. 10). When one, two, or three copies of DRE-Eb were ligated to the heterologous TATA-containing promoter, the promoter activity was progressively increased by 3.0-, 8.5-, or 20.0-fold, respectively, while the addition of three copies of DRE-Eb (mut I II III) did not activate the promoter activity. Thus, the three DRE-related sequences in DRE-Eb are important for high promoter activity and function as positive regulatory elements.
DREF Activates the dE2F Gene Promoter in Vivo-The results obtained in vitro and in cultured cells clearly demonstrate important roles for the three DRE-related sequences in dE2F gene promoter activity. Although we tried co-transfection of the DREF expression plasmid and the p-2184E2Fbluc plasmid into Kc cells, significant stimulatory effects on the promoter activity were not observed (data not shown), probably due to the high level of endogenous DREF in Kc cells (28). Therefore, we designed another type of experiment using transgenic flies to examine the effects of DREF overexpression on the dE2F gene promoter activity.
We thus established transgenic flies in which ectopic expression of DREF was targeted to the eye imaginal disc. Overexpression of DREF in cells posterior to the morphogenetic furrow in the larval eye imaginal disc was confirmed by immunostaining with the anti-DREF antibody (Fig. 11B). These transgenic flies were crossed with a dE2F mutant, dE2F 729, in which the lacZ gene had been inserted near the translation initiation site of the dE2F gene in the same orientation as the dE2F gene (Fig. 1A, and Ref. 19). The lacZ expression was therefore directed by the dE2F gene promoter and its pattern was indistinguishable from expression of the dE2F gene (43).
As shown in Fig. 11D, strong staining signals with the anti-␤-galactosidase antibody were observed in cells posterior to the morphogenetic furrow on compared with the control fly case (Fig. 11C). No significant difference in the lacZ expression pattern in flies with and without DREF overexpression was observed in cells anterior to the furrow, a region where DREF is not overexpressed. Thus, we conclude that DREF can activate the dE2F gene promoter in vivo.

DISCUSSION
Promoters of Drosophila DNA replication-related genes contain common transcriptional regulatory elements such as E2Fbinding sites and DRE (20,23,25,29). Our previous studies using a transgenic fly system indicated that both E2F-binding sites and DRE are required for activity of the Drosophila PCNA gene promoter throughout all stages of development (30). dE2F is essential for transcription of Drosophila DNA replicationrelated genes, since transcripts of the latter are completely Open boxes indicate the homeodomain protein-binding consensus TAAT, protected in the Kc cell nuclear extract from DNase I digestion (Fig. 8A). Luciferase activities are expressed as percentages of the p-2184E2Fbluc value on the right.
FIG . 10. Effects of the three DRE-related sequences on the heterologous promoter. Luciferase expression plasmids having the three DRE-related sequences placed upstream of the heterologous TATA promoter, derived from the metallothionein gene basal promoter, are illustrated. These reporter plasmids were transfected into Kc cells, and luciferase activities were determined and expressed on the right as values relative to that with the construct carrying the metallothionein gene basal promoter alone. lacking in the dE2F Ϫ mutant embryos (18,19,21,22). Cyclin E, which in the form of a complex with cyclin dependent kinase a critical regulator of G 1 to S phase transition, activates dE2F in central nervous system cells (18). This regulation appears to be mediated by a homologue of the retinoblastoma protein, RBF (44).
In addition to this physiological regulation of dE2F activity, the level of dE2F can be increased through a transcription step during the onset of proliferation. In mammals, auto-regulation through E2F-binding sites in the E2F gene promoters and several transcription factors such as CCAAT-binding protein, YY-1 and c-Myc, have been suggested to play important roles in transcriptional activation of E2F genes (8,11,13). However, in the Drosophila case, little is known about transcriptional regulators. In the present study, we have shown that the DRE/ DREF system activates transcription of the dE2F gene. Probably during the onset of proliferation, increased dE2F activity in combination with DREF could coordinately activate transcription of DNA replication-related genes to prepare for S phase.
Adult eyes of transgenic flies overexpressing DREF in the eye imaginal discs shows a severe rough eye phenotype. 3 A half-reduction of the dE2F gene copy number suppresses this rough eye phenotype. However, half-reductions of other DREFtarget genes such as those for PCNA, DNA polymerase ␣, and cyclin A apparently exerted no effect on the rough eye phenotype. 4 Thus, the dE2F gene appears to be one of the most critical target genes of DREF, at least in eye imaginal disc cells.
We have found that overexpression of DREF can stimulate the expression of the lacZ gene that is placed under the control of the dE2F gene promoters in the eye imaginal disc. Since the lacZ transgene is located downstream of two different dE2F gene promoters, one for transcript-a and the other for transcript-b, we do not know which promoter is actually activated by DREF. The present studies indicate that the DRE-related sequence located upstream of the initiation site of transcript-b is required for the promoter activity and specifically bind DREF. Although we also found DRE-related sequences in the upstream region of the initiation site of transcript-a, these were not recognized by DREF on band mobility shift analysis (data not shown). Thus, the overexpression of DREF likely activates the promoter for transcript-b rather than that for transcripta. It should be noted that the promoter for transcript-b appears to function throughout all stages of development, while that for transcript-a may only act at limited stages during embryogenesis. FIG. 11. Overexpression of DREF activated lacZ under the control of the dE2F gene promoter. Eye imaginal discs from third instar larvae are shown. The morphogenetic furrow is marked by arrows. A, GMR-GAL4; UAS-DREF/ϩ; ϩ eye disc stained with control IgG. B, GMR-GAL4; UAS-DREF/ϩ; ϩ eye disc stained with anti-DREF antibody. DREF is present in the posterior portion to the morphogenetic furrow in the disc. C, GMR-GAL4; ϩ/ϩ; dE2F 729 /ϩ eye disc stained with anti-␤-galactosidase antibody. D, GMR-GAL4; UAS-DREF/ϩ; dE2F 729 /ϩ eye disc stained with anti-␤galactosidase antibody. Enhancement of lacZ expression under the control of the dE2F gene promoter is apparent.