Core Promoter Functions in the Regulation of Gene Expression of Drosophila Dorsal Target Genes*

Background: The Dorsal transcription factor regulates the dorsal-ventral developmental gene regulatory network by binding to enhancers. Results: The core promoters of multiple Dorsal target genes are evolutionarily conserved and functionally dependent on specific elements. Conclusion: The core promoter composition is an important determinant of the transcriptional outcome. Significance: Transcriptional regulation results from an interplay between enhancers and core promoter composition. Developmental processes are highly dependent on transcriptional regulation by RNA polymerase II. The RNA polymerase II core promoter is the ultimate target of a multitude of transcription factors that control transcription initiation. Core promoters consist of core promoter motifs, e.g. the initiator, TATA box, and the downstream core promoter element (DPE), which confer specific properties to the core promoter. Here, we explored the importance of core promoter functions in the dorsal-ventral developmental gene regulatory network. This network includes multiple genes that are activated by different nuclear concentrations of Dorsal, an NFκB homolog transcription factor, along the dorsal-ventral axis. We show that over two-thirds of Dorsal target genes contain DPE sequence motifs, which is significantly higher than the proportion of DPE-containing promoters in Drosophila genes. We demonstrate that multiple Dorsal target genes are evolutionarily conserved and functionally dependent on the DPE. Furthermore, we have analyzed the activation of key Dorsal target genes by Dorsal, as well as by another Rel family transcription factor, Relish, and the dependence of their activation on the DPE motif. Using hybrid enhancer-promoter constructs in Drosophila cells and embryo extracts, we have demonstrated that the core promoter composition is an important determinant of transcriptional activity of Dorsal target genes. Taken together, our results provide evidence for the importance of core promoter composition in the regulation of Dorsal target genes.

Developmental processes are highly dependent on transcriptional regulation by RNA polymerase II. The RNA polymerase II core promoter is the ultimate target of a multitude of transcription factors that control transcription initiation. Core promoters consist of core promoter motifs, e.g. the initiator, TATA box, and the downstream core promoter element (DPE), which confer specific properties to the core promoter. Here, we explored the importance of core promoter functions in the dorsal-ventral developmental gene regulatory network. This network includes multiple genes that are activated by different nuclear concentrations of Dorsal, an NFB homolog transcription factor, along the dorsal-ventral axis. We show that over two-thirds of Dorsal target genes contain DPE sequence motifs, which is significantly higher than the proportion of DPE-containing promoters in Drosophila genes. We demonstrate that multiple Dorsal target genes are evolutionarily conserved and functionally dependent on the DPE. Furthermore, we have analyzed the activation of key Dorsal target genes by Dorsal, as well as by another Rel family transcription factor, Relish, and the dependence of their activation on the DPE motif. Using hybrid enhancer-promoter constructs in Drosophila cells and embryo extracts, we have demonstrated that the core promoter composition is an important determinant of transcriptional activity of Dorsal target genes. Taken together, our results provide evidence for the importance of core promoter composition in the regulation of Dorsal target genes.
Focused core promoters, i.e. promoters in which transcription initiates at a single nucleotide or within a narrow region of several nucleotides, exist in all eukaryotes and are predominant in simple organisms, such as Drosophila (13,17,18). Focused core promoters are typically ϳ80 nucleotides in length and encompass the RNA start site (13,17,18).
Core promoters may contain one or more functional DNA sequence elements, termed core promoter elements or motifs, such as the TATA box, TFIIB recognition elements (BRE u and BRE d ), initiator (Inr), 2 TCT motif, motif 10 element (MTE), and downstream core promoter element (DPE), which confer specific properties to the core promoter (13, 19 -21). The TATA box, the Inr, the MTE, and the DPE motifs are recognized and bound by subunits of TFIID, the first complex that recognizes and binds the core promoter in the process of RNA polymerase II recruitment to the core promoter.
The TATA box is the first core promoter motif identified and is conserved from archaebacteria to humans (22,23). The upstream T is typically located at Ϫ30 or Ϫ31 relative to the transcription start site (24). The TATA box is bound by the TATA box-binding protein (TBP) subunit of TFIID. The Inr encompasses the transcription start site and is the most common core promoter element (11,19,25). The Inr is bound by the TAF1 and TAF2 subunits of TFIID (26). The DPE was originally discovered as a TFIID recognition site that is located downstream of the initiator element (precisely from ϩ28 to ϩ33 relative to the Aϩ1 of the Inr) and is conserved from Drosophila to humans (27,28). It is bound by the TAF6 and TAF9 subunits of TFIID (28). The MTE is located immediately upstream of the DPE at precisely ϩ18 to ϩ27 relative to the Aϩ1 in the Inr and is also conserved from Drosophila to humans (29,30). Both the DPE and the MTE motifs are dependent on the Inr and function cooperatively with it (28, 30 -33). Moreover, we have previously demonstrated that gene expression levels can be modulated via the core promoter (31).
The existence of different types of core promoters implies that core promoter elements play regulatory roles beyond the specification of transcription initiation. Transcription of TATAdependent genes differs from transcription of DPE-dependent genes in many respects. First, the set of basal transcription factors that is necessary to transcribe TATA-dependent promoters in vitro is insufficient to transcribe DPE-dependent promoters (34,35). Second, enhancers with preference for DPEcontaining promoters or TATA-containing promoters have been identified, supporting the existence of enhancer-promoter specificity (16, 36 -38). Third, TBP is necessary for TATA-dependent transcription. However, TBP down-regulates DPE-dependent transcription (39). The two transcriptional regulators, NC2 and MOT1, which have been shown to be positive regulators of DPE-dependent transcription, do so by counteracting TBP, thus relieving its inhibition of DPE transcription (39 -41).
In this study, we explored the contribution of the DPE motif to the regulation of gene expression in the dorsal-ventral developmental gene network that is governed by the Rel family transcription factor Dorsal. We show that the promoters of many Dorsal target genes contain DPE sequence motifs. We demonstrate that core promoters of multiple Dorsal target genes are evolutionarily conserved and are functionally dependent on the DPE motif. Moreover, we have analyzed the activation of key Dorsal target genes by Dorsal and another Rel family transcription factor, Relish, and the dependence on the DPE motif. Using Drosophila S2Rϩ cells, we further demonstrate that the TATA box cannot compensate for the loss of DPE in the Dorsal target brinker and can only partially compensate for the loss of DPE in the Dorsal target twist. The addition of a TATA box to the brinker and twist promoter can, however, compensate for the loss of DPE using nuclear extracts derived from Drosophila embryos. We show that Dorsal is able to discriminate between a TATA box-containing promoter and a DPE-containing promoter. Transcription of enhancer-promoter constructs in both Drosophila S2Rϩ cells and embryo extracts manifests the importance of the core promoter in transcriptional regulation of Dorsal target genes. Collectively, we demonstrate that the DPE motif plays an important role in the expression of Dorsal target genes regulating dorsal-ventral patterning and provides evidence for the key contribution of the core promoter composition to gene regulatory networks.

EXPERIMENTAL PROCEDURES
Sequence Conservation Analysis-Sequence conservation analysis was performed using the UCSC genome browser.
Calculation of Core Promoter Element Frequency-Drosophila transcripts that initiate at different chromosomal positions (based on the RefSeq database) were used to calculate the frequency of the core promoter elements (see Tables 1 and 2). Inr elements were identified in the region Ϫ10 to ϩ10 relative to the transcription start site if there was an A at ϩ1 and at least three of five additional matches to the consensus (TCAKTY, where the A is the ϩ1 of the transcript). DPE motifs that were precisely located at ϩ28 relative to the A ϩ 1 of the Inr, were identified based on at least five of six matches to the DPE functional range set (DSWYVY). Putative TATA box elements were identified based on a match to a TATA sequence located from Ϫ45 to Ϫ19 relative to the RefSeq transcription start site. The significance of differences in core promoter composition between Drosophila transcripts and Dorsal target genes was calculated using the chi-square test.
In Vitro Transcription Assays-Double-stranded oligonucleotides comprising core promoter sequences from Ϫ10 to ϩ40 (see Fig. 1 and supplemental Fig. S1) of the tested Dorsal target genes core promoters were inserted into the PstI and XbaI sites of pUC119. Mutation of the DPE in the core promoters was identical to that used previously (30), where the mutant DPE contains CATA at ϩ30 to ϩ33 relative to A ϩ 1. Natural tinman (tin), brinker (brk), twist (twi), and leak (lea) enhancerpromoter constructs driving the luciferase reporter gene were used in Figs. 3-5, and tin-twi hybrid enhancer-promoter constructs driving the luciferase reporter gene were used in Fig. 6. In vitro transcription reactions were carried out as described previously (42) using 250 ng of supercoiled DNA templates with Drosophila high salt nuclear extracts (43). The resulting transcripts were subjected to primer extension analysis with an M13 reverse sequencing primer (AGCGGATAACAATTTCA-CACAGGA; see Fig. 2) or with a reverse luciferase primer (TCTTCCAGCGGATAGAATGGCGCC; see Figs. 5 and 6). Quantitation of reverse transcription products was carried out using ImageQuant, ImageJ, and GelQuantNET. All experiments were carried out a minimum of three independent times to ensure reproducibility of the data.
Expression Plasmids and Luciferase Reporter Genes-pAc-Dorsal expression plasmid was kindly provided by Dr. Albert Courey (UCLA) For the construction of a Relish expression vector, the active N terminus of Relish (amino acids 1-532; not including the ankyrin repeat-containing inhibitory domain) (44) was cloned into a pAc expression vector. An N-terminally FLAG-tagged full-length Caudal was cloned into a pAc expression vector. A C-terminally V5-His-tagged full-length Bicoid was cloned into a pAc expression vector. The natural enhancerpromoter tin reporter construct encompasses a genomic fragment from Ϫ2089 to ϩ80 relative to the Aϩ1 of the Inr (containing either a wt, or a mDPE), upstream of a firefly luciferase reporter gene. The natural enhancer-promoter brk reporter construct encompasses a genomic fragment from Ϫ3531 to ϩ60 relative to the Aϩ1 of the Inr (containing either a wt, a mDPE, or a mDPE with a TATA box), upstream of a firefly luciferase reporter gene. The natural enhancer-promoter twi reporter constructs encompasses a genomic fragment from Ϫ2041 to ϩ50 relative to the A ϩ 1 of the Inr (containing either a wt, a mDPE or a mDPE with a TATA box), upstream of a firefly luciferase reporter gene. The natural enhancer-promoter lea reporter construct encompasses a genomic fragment from Ϫ3931 to ϩ69 relative to the Aϩ1 of the Inr (containing either a wt, a mDPE, or a mDPE with a TATA box), upstream of a firefly luciferase reporter gene. The tin enhancer-twi core promoter hybrid promoter was constructed using the tin enhancer from Ϫ2089 to Ϫ138 followed by the twi promoter from Ϫ134 to ϩ50 (with a 3-nucleotide spacer in between), upstream of a firefly luciferase reporter gene.
Luciferase reporter constructs driven by either the tin, brk, lea, or twi genomic region with a mutated DPE (in which nucleotides at positions ϩ28 to ϩ34 relative to Aϩ1 in the Inr were mutated to CTCATGT) or an additional TATA box (TATAAAAA, where the upstream T is at position Ϫ31 relative to the A ϩ 1 of the Inr) were generated using a QuikChange II site-directed mutagenesis kit (Stratagene). DNA sequence-verified fragments that encompass the mutated nucleotides were subcloned into their corresponding locations in the wild-type vectors. Pol III-Renilla luciferase reporter was kindly provided by Dr. Norbert Perrimon (Harvard Medical School).
Transient Transfection and Reporter Gene Assays-Drosophila melanogaster Schneider S2Rϩ adherent cells were cultured in Schneider's Drosophila medium (Biological Industries) that was supplemented with 10% heat-inactivated FBS. Cells were transfected in 24-well plates by using the Escort IV reagent (Sigma). For Dorsal dose-response dual luciferase assays, cells were transfected with the firefly luciferase reporter constructs (80 ng) and the indicated amounts of a Dorsal expression vector that was supplemented, where necessary, with pAc control expression vector to give a total of 1.5 g of DNA of expression vector. For assaying the tin, brk, twi, or lea reporter constructs with mDPE and an added TATA box, cells were co-transfected with either 0.5 g of Dorsal, Relish, Bicoid, or Caudal expression vector that was supplemented, where necessary, with pAc control expression vector to give a total of 1.0 g of DNA of expression vector, in addition to either tin (80 ng), brk (80 ng), lea (80 ng), or twi reporter (20 ng).
To normalize for transfection efficiency, cells were co-transfected with a Pol III-Renilla luciferase control plasmid (10 ng). Cells were harvested 36 -48 h post-transfection and assayed for dual luciferase activities, as specified by the manufacturer (Promega). The graphs represent averages of three to four independent experiments. Tar-get Genes-To discover distinct gene regulatory networks (GRNs) and pathways that are regulated via the core promoter, we have examined the core promoter composition of genes that are involved in early embryonic fly development. This analysis led us to focus on the dorsal-ventral GRN and in particular on Dorsal target genes. The dorsal-ventral GRN is critical for early embryonic development. The network includes multiple genes that are activated by different nuclear concentrations of the Dorsal transcription factor along the dorsal-ventral axis (45)(46)(47)(48)(49)(50)(51)(52)(53)(54). Activation of Dorsal target genes is achieved by the recruitment of Dorsal to the enhancers of these genes, which contain Dorsal-binding sites hundreds or even thousands of base pairs upstream of the transcription start site.

Identification of DPE Motifs in Core Promoters of Dorsal
To examine whether core promoter elements play a role in the regulation of Dorsal-target gene expression, we first analyzed the core promoter sequences of Dorsal-target genes. Using the RefSeq database, we have carried out a comprehensive analysis of the core promoter composition of previously identified Dorsal target genes (Refs. 47 and 51 and references therein) (Table 1). Strikingly, the majority of known Dorsal target genes appear to contain Inr and DPE motifs. Specifically, in the mesoderm, where Dorsal nuclear concentration is highest, of a total 30 mesodermal genes analyzed, 17 genes contain potential Inr and DPE motifs with high probability (Table 1, Inr and DPE ϩϩϩ and Inr and DPE ϩϩ), including Dorsal target genes critical for mesodermal development such as tin, twi, Hbr, and Mef2, and 3 genes contain TATA box, Inr, and DPE motifs. In the neuroectoderm, where Dorsal nuclear concentration is intermediate, 17 genes contain Inr and DPE motifs with high probability, including well characterized Dorsal target genes such as brk, ths, and pyr, and 7 genes contain TATA box, Inr, and DPE motifs of a total 35 neuroectodermal genes analyzed. In the dorsal ectoderm, where Dorsal nuclear concentration is lowest, 10 genes contain Inr and DPE motifs with high probability, including well characterized dorsoectodermal targets such as dpp and tup, and 9 genes contain TATA box, Inr, and DPE motifs of a total 26 dorsal ectoderm genes analyzed. Overall, 63 of 91 genes analyzed (over 69%) are likely to contain functional DPE motifs.
Computational analyses of the frequencies of Drosophila core promoter elements had previously indicated that the DPE is present in ϳ2.1-22% of Drosophila core promoters (19,25,29). We have calculated the frequencies of core promoter elements among Drosophila transcripts that initiate at different chromosomal positions using strict Inr-DPE spacing criteria. We have discovered that 23% of Drosophila promoters contain Inr and DPE motifs, whereas 4% contain Inr, DPE, and TATA box motifs ( Table 2). The frequency of DPE-containing promoters among Dorsal target genes (69%) is significantly higher than the calculated (23%) and the previously reported frequencies (2.1% to 22%) of DPE in Drosophila core promoters ( Table 2).
The overall frequency of a TATA box among Drosophila transcripts is 18% (Table 2). Interestingly, only a small subset of the Dorsal target genes appear to contain a TATA box and no DPE. In the mesoderm there are no genes that contain only a TATA box, 2 genes that contain a TATA box and an Inr motif, and 3 more genes that contain TATA box, Inr, and DPE motifs. In the neuroectoderm, 1 gene contains a TATA box, there are no genes that contain a TATA box and Inr, and 7 genes contain TATA box, Inr, and DPE motifs. In the dorsal ectoderm, 1 gene contains a TATA box, 3 genes contain a TATA box and Inr, and 9 genes contain TATA box, Inr, and DPE. The overall frequency of TATA box without a DPE motif among Dorsal target genes is less than 8% (7 of 91), and only a few (3 of 65) of the Dorsal targets that are regulated by high or intermediate nuclear concentrations of Dorsal contain a TATA box without a DPE motif, which is significantly lower than the overall frequency of Drosophila core promoters containing a TATA box without a DPE motif (13%; Table 2). Taken together, the Inr and DPE motifs appear to be overrepresented in the core promoters of multiple Dorsal target genes.
The  (59 -63), and genes that are activated by low levels of Dorsal: thisbe (ths), pyramus (pyr) in the neuroectoderm, and tailup (tup) in the dorsalectoderm (64,65). In particular, we have searched these genes for both an Inr motif and a DPE motif that matches the functional DPE range set (DSWYVY) (33). Additionally, we have verified that the DPE motif is precisely located at ϩ28 relative to the A ϩ 1 of the Inr. Such spacing was previously shown to be critical for the function of DPE-dependent promoters (28,33). Importantly, none of these genes contain a TATA box.
To assess the potential biological significance of the DPE motifs in the core promoters of these Dorsal target genes, we   Fig. 1 and supplemental Fig. S1). vn is highly conserved throughout the core promoter sequence but in only five Drosophila species (supplemental Fig. S1). Hence, the core promoters of multiple Dorsal target genes appear to contain conserved Inr and DPE motifs that match the DPE functional range set (33).
The DPE Motifs in Multiple Dorsal Target Genes Are Functionally Important for Transcription-To examine whether the conserved DPE motifs are functionally important for transcription of Dorsal target genes, we have cloned both wild-type (wt) and mutant DPE (mDPE) core promoters (Ϫ10 to ϩ40 relative to the A ϩ 1 of the Inr) of 12 putative DPE-containing Dorsal target genes (namely, twi, brk, tin, pyr, ths, Mdr49, phm, RhoL, tup, lea, soxN, and vn) into the pUC119 vector. The activity of wt and mDPE core promoters was tested by in vitro transcription with Drosophila embryo nuclear extracts followed by primer extension analysis. Mutations in the DPE motif of each of the core promoters of the aforementioned Dorsal target genes resulted in reduced transcription levels (Fig. 2). Hence, transcription of the core promoters of these 12 Dorsal target genes is highly dependent on the DPE.
Activation of Dorsal Target Genes via the DPE-To examine the effect of core promoter composition on activation of Dorsal target genes in the context of their natural enhancers, we mapped the putative Dorsal DNA binding sites within the  To test the activation of the tin, brk, and twi genes by Dorsal and their dependence on the DPE motif, genomic fragments encompassing either 2.2 kb of the natural tin enhancer and promoter, 3.6 kb of the natural brk enhancer and promoter (each containing 6 putative Dorsal binding sites), or 2.1 kb of the natural twi enhancer and promoter (which contains 15 putative Dorsal binding sites) were subcloned upstream of the firefly luciferase reporter gene (illustrated in Fig. 3). D. melanogaster Schneider S2Rϩ cells were co-transfected, and the transcriptional activity of wt and mDPE reporter genes in the pres-ence of varying amounts of transfected Dorsal expression plasmid was assayed. The transcriptional activity of both tin and brk is higher in the presence of increasing amounts of transfected Dorsal expression plasmid, as expected (Fig. 3, A and B). The levels of transcription of the mDPE tin and brk do not increase in response to increasing amounts of transfected Dorsal, suggesting that the transcription of tin and brk is highly dependent on the presence of the DPE motif.
The transcriptional activity of twi is augmented with increasing amounts of transfected Dorsal expression plasmid, as expected (Fig. 3C). The levels of transcription are dependent on the presence of the DPE motif, because the transcriptional activity of mDPE twi is significantly lower. There is, however, an increase in the mDPE twi reporter activity with increasing amounts of transfected Dorsal expression vector. The basal transcription levels (in the absence of transfected Dorsal) of the twi reporter constructs are also dependent on the DPE motif (Fig. 3D). The basal activity of the wt twi promoter was 19-fold higher than that of the mDPE twi, whereas the basal activities of both wt tin and wt brk promoters were 6-fold higher than that of their corresponding mDPE reporters (data not shown).

The TATA Box Motif Cannot Substitute for the Loss of a DPE Motif in brk Transcription and Can Only Partially Substitute for the Loss of a DPE Motif in twi in Drosophila S2Rϩ
Cells-Transcription of TATA-dependent genes differs from transcription of DPE-dependent genes in many respects. We demonstrated that mutation of the DPE reduces activation of brk and twi (Fig. 3). Thus, to examine whether the activation of these genes by Dorsal could occur via a TATA box motif, we have generated reporter constructs containing either brk or twi natural enhancer-promoter that, in addition to a mutation in the DPE motif, contain a TATA box. Addition of a TATA box to the brk natural enhancer-promoter with a mutation in the DPE motif could not restore transcription levels in the absence or in the presence of transfected Dorsal (Fig. 4A).
Interestingly, transfection of Relish, another Rel family transcription factor, which binds the same or nearly identical DNA sequence motifs as Dorsal (66,67), was able to activate the wt brk, albeit to lower levels as compared with Dorsal. Addition of a TATA box to the brk natural enhancer-promoter with a mutation in the DPE motif could not restore transcription levels in the presence of transfected Relish (Fig. 4A).
Addition of a TATA box to the mDPE twi reporter was only able to partially restore activation by transfected Dorsal in the context of the natural enhancer (less than 30%) (Fig. 4B). Similarly, addition of a TATA box to the mDPE twi reporter was only able to partially restore activation by transfected Relish in the context of the natural enhancer. Interestingly, addition of a TATA box to a twi reporter containing a mDPE motif was unable to restore the basal transcription activity in the absence of transfected Dorsal (Fig. 4C). The twi enhancer contains 15 putative Dorsal binding sites, whereas the brk enhancer contains 6 putative Dorsal binding sites. To examine whether twi is less dependent on the DPE because it contains a higher number of putative Dorsal binding sites, we have cloned a genomic fragment encompassing 4.0 kb of the natural leak (lea) enhancer and promoter, which contains 13 putative Dorsal DNA binding sites. We have also generated lea reporter constructs containing either a mDPE motif or a mDPE with a TATA box and have tested their activation by Dorsal and Relish (Fig. 4D). A mutation of the DPE reduces activation of lea by both Dorsal and Relish. Interestingly, addition of a TATA box to the lea reporter containing a mDPE motif was able to restore the basal transcription activity, as well as the activation by transfected Dorsal or Relish (Fig. 4D).
Taken together, in Drosophila S2Rϩ cells, brk, twi, and lea promoters are dependent on the DPE motif. Dorsal activates transcription of its targets brk and twi via the DPE, whereas it activates lea via the DPE as well as the TATA box. Furthermore, Relish, which binds the same or nearly identical DNA sequence motifs as Dorsal, is able to activate transcription of Dorsal targets with similar core promoter preference as Dorsal.
The DPE Motif, Which Is Important for the Activity of twi, lea, and brk in Drosophila Nuclear Extracts, Can Be Replaced by a TATA Box-We next sought to examine the dependence of the twi, lea, and brk on the DPE motif in Drosophila embryo nuclear extracts, which provide a different experimental system. To this end we have performed in vitro transcription reactions of the enhancer-promoter reporters containing wt, mDPE, or mDPE ϩ TATA with Drosophila embryo nuclear extracts followed by primer extension analysis. Mutations in the DPE motifs of the core promoters of twi, lea, and brk result in reduced transcription levels of the enhancer-promoter constructs (Fig. 5A). Hence, transcription of these Dorsal target genes is dependent on the DPE.
The addition of a TATA box to mDPE-containing twi and brk promoters cannot fully restore the reporter activity of the mDPE constructs in transfected S2Rϩ cells (Fig. 4). The addition of a TATA box to mDPE-containing twi, lea, and brk promoters can, however, restore in vitro transcription using Drosophila embryo nuclear extracts (Fig. 5A). This demonstrates both the strength of the TATA box in general as well as the apparent absence of transcription factors from Drosophila S2Rϩ cells. We have analyzed the enhancer of brk for putative binding sites of transcription factors that are expressed in 0 -12-h embryos but are not expressed in S2Rϩ cells. The brk enhancer contains 19 putative Bicoid binding sites and 60 putative Caudal bindings sites (JASPAR). To examine whether Caudal or Bicoid could activate transcription of a mDPE ϩ TATAcontaining brk reporter, we co-transfected S2Rϩ cells with natural brk reporter constructs containing wt, mDPE, or mDPE ϩTATA box promoter, as well as expression vectors of either Dorsal, Bicoid, or Caudal (Fig. 5B). As expected, Caudal, which has previously been shown to be a preferential DPE-specific transcription factor (68), was able to activate the wt brk promoter but was unable to activate transcription of mDPE or mDPE ϩ TATA-containing promoters. Interestingly, Bicoid was able to activate transcription of the wt brk reporter. Unlike Dorsal and Caudal, Bicoid does not have such core promoter preference because it activates the mDPE reporter and to some extent, also the mDPE ϩ TATA brk reporter. It is of note that Bicoid activates its natural target gene giant, which contains functional TATA box, Inr, and DPE motifs, regardless of its core promoter composition (data not shown). Hence, a factor other than Caudal or Bicoid accounts for the transcription of the brk reporter containing a mDPE ϩ TATA box using embryo nuclear extracts.
The Core Promoter Composition Is an Important Contributor to the Transcriptional Output-To determine the contribution of the core promoter to the transcriptional output, we have generated hybrid enhancer-promoter constructs with the tin enhancer and twi core promoter and have tested their activities. Co-transfection of a tin enhancer-twi core promoter hybrid reporter and increasing amounts of Dorsal expression vector to wt (1) mDPE (2) mDPE+TATA (3) -Dorsal Bicoid Caudal S2Rϩ cells resulted in an overall Dorsal activation that is typical for the tin enhancer (compare Figs. 3A and 6A). The tin enhancer-twi core promoter hybrid reporters containing mDPE or mDPE ϩ TATA can be somewhat activated by Dorsal, as we have observed for twi enhancer-promoter reporters (Figs. 3, 4, and 6A). Hence, the ability of Dorsal to activate the mDPE or mDPE ϩ TATA hybrid promoters in S2Rϩ cells relies on the core promoter in addition to the enhancer. We next compared the activity of the hybrid reporters using Drosophila embryo nuclear extracts. As can be seen in Fig. 6B, the transcription activities of the hybrid tin enhancer-twi core promoter reporter constructs using embryo nuclear extracts are very similar to the activities of the natural twi enhancerpromoter constructs (Fig. 5A), suggesting that even though the enhancers are present in these constructs, the activity of the constructs in the embryo nuclear extracts is dictated by the core promoters. Collectively, the analysis of the hybrid enhancerpromoter constructs in both S2Rϩ cells and nuclear extracts derived from Drosophila embryos highlights the contribution of the core promoter to the transcriptional output.

DISCUSSION
The DPE Is a Transcriptional Element Shared by Many Dorsal Target Genes-In this study we demonstrate that the DPE is an important, conserved transcription element shared by multiple Dorsal target genes, which comprise the dorsal-ventral gene regulatory network. Specifically, we have shown that over two-thirds of the known Dorsal target genes contain DPE motifs, which is significantly higher than the percentage of DPE promoters in Drosophila genes (19,25,29). Remarkably, only less than 8% of the Dorsal target genes contain TATA box elements without DPE motifs. The DPE is most prevalent in mesodermal Dorsal targets. The number of DPE containing genes decreases in the neuroectoderm and further decreases in the dorsal ectoderm, where the Dorsal nuclear concentration is lowest. On the other hand, the number of genes containing a TATA box is higher in regions where Dorsal nuclear concentration is decreased. The occurrence of the DPE in many developmentally regulated genes (68), as well as our analysis of the frequencies of core promoter elements in Drosophila genes ( Table 2) and the identification of DPE motifs in the majority of Dorsal targets, imply that the DPE is not randomly distributed in ϳ23% of the genes; rather it is enriched in specific GRNs and pathways.
Regulation of twi, lea, and brk via the Core Promoter-We have examined the transcription of the natural enhancers and promoters of twi, lea, tin, and brk in Drosophila Schneider S2Rϩ cells and have discovered that the basal transcription levels of twi, lea, tin, and brk in the absence of ectopically expressed Dorsal are highly dependent on the DPE motif ( Figs.  3 and 4).
We show that the DPE core promoter motif is an important regulatory component in Dorsal target genes in S2Rϩ cells. The brk and twi core promoters are dependent on the DPE motif and could not be fully activated in S2Rϩ cells by an added TATA box, whereas the lea core promoter is functionally dependent on the DPE, but its activity could be restored via an added TATA box.
Unlike the results obtained using Drosophila S2Rϩ cells, in vitro transcription analysis using nuclear extracts derived from Drosophila embryos has demonstrated that the DPE motif, which is important for the transcriptional activity of twi, lea, and brk, can be replaced by a TATA box. This demonstrates the strength of the TATA box, which can restore transcription of some mDPE-containing promoters. Nevertheless, the TATA box is only naturally used in a minority of Dorsal target genes.
Importantly, the mDPE and mDPE ϩ TATA reporter constructs of twi, brk, and lea have comparable strength in vitro, yet tin enhancer twi core promoter mDPE+ TATA FIGURE 6. The core promoter is a key contributor to the transcriptional output. Drosophila S2Rϩ cells were transfected with firefly luciferase reporter constructs containing hybrid enhancer-promoters with either wt, mDPE, or MDPE ϩ TATA core promoter, as well as varying amounts of a Dorsal expression plasmid, as indicated. To normalize for transfection efficiency, cells were co-transfected with a Pol III-Renilla luciferase control plasmid and assayed for dual luciferase activity. The activities are reported relative to the wild-type promoter in the absence of co-transfected Dorsal expression plasmid, which was defined to be 1. A schematic diagram of the hybrid genomic fragments is shown on top. A, transcriptional activation of the hybrid tinman enhancer-twist core promoter constructs by Dorsal (n ϭ 3). B, the hybrid tinman enhancer-twist core promoter constructs containing either wt, mDPE, or mDPE ϩ TATA motifs, were subjected to in vitro transcription analysis with a Drosophila embryo nuclear extract. The resulting transcripts were detected by primer extension-reverse transcription analysis.
they do not have the same strength in S2Rϩ cells. This suggests that there are additional DPE specificity factors in the cells. The identification of such DPE specificity factor(s) that act in the activation of twi and brk, but not in lea, awaits future investigation. Taken together, our data indicate that the core promoter contributes to the overall transcription levels and adds an important regulatory dimension to the complex dorsal-ventral gene network. Not All Sequence-specific Transcription Factors Activate Their Target Genes in a Core Promoter Preferential Manner-In this study we have demonstrated that multiple Dorsal target genes are dependent on the DPE and that Dorsal has the ability to preferentially activate some of its targets via the DPE. It is interesting to note that Relish, another Rel family transcription factor, which is important for the activation of the Imd (Immune deficiency) pathway of innate immunity and binds the same or nearly identical DNA sequence motifs as Dorsal, activates transcription in a similar manner to Dorsal, albeit to a lower extent, presumably because of the lack of additional factors that might be missing in the S2Rϩ cells in which the Imd pathway has not been activated (Fig. 4).
In a separate study, we have observed that the maternal sequence-specific transcription factor Bicoid activates its natural target gene giant, which contains functional TATA box, Inr, and DPE motifs, regardless of its core promoter composition (data not shown). In this study we show that Bicoid activates the mDPE brk reporter (Fig. 5B), highlighting the fact that the ability of Dorsal, Relish, and Caudal to activate transcription of some of their target genes with a preference for the core promoter composition is not a general property of sequencespecific transcription factors; rather, it is a unique feature of these specific transcription factors.
DPE Transcription May Provide a Specialized Transcription System Directed toward Development-The present study demonstrates that the DPE is broadly used in the regulation of genes that mediate the formation of the dorsal-ventral axis during early embryonic development. Furthermore, our findings highlight the importance of the core promoter, in addition to sequence-specific DNA binding motifs in enhancers, in the regulation of gene expression. We have previously demonstrated that the core promoters of the majority of the Hox genes, which are key regulators of the development of the embryonic body plan, contain functional DPE motifs (68). We have also shown that Caudal, a key regulator of the Hox gene network and a sequence-specific enhancer binding transcription factor, activates transcription with a distinct preference for the DPE over the TATA box (68). We now propose that the prevalence of the DPE in developmentally regulated genes provides these genes with certain advantages in the complex regulation of gene expression (e.g. kinetics of responsiveness to specific signals).
The concept of a specialized transcription system has been articulated with regards to the TCT core promoter element, which has been shown to play a key role in a system that is directed toward the synthesis of ribosomal proteins (20). Transcription of TATA-dependent genes is different from transcription of DPE-dependent genes in many respects, e.g. the necessary basal transcription factors, the existence of core promoter-specific enhancers, and the necessity for TBP for TATA transcription as opposed to DPE transcription, where TBP exerts inhibitory effects. Taken together, DPE transcription may be regarded as a specialized transcription system that is directed toward development.