A cis-acting DNA element located between TATA box and transcription initiation site is critical in response to regulatory sequences in human angiotensinogen gene.

The promoter of the human angiotensinogen (hAG) gene functioned in its own core promoter context but not when replaced with simian virus 40 (SV40) core promoter, suggesting the presence of a transcriptionally important cis-acting sequence. Electrophoretic mobility shift assays demonstrated that a ubiquitously expressed nuclear factor, AGCF1, bound to AGCE1 (hore promoter lement 1; positions −25 to −1) located between the TATA box and transcription initiation site. Substitution mutation in AGCE1 which disrupted AGCF1 binding affected the promoter activity more severely than a nonsense mutation of the hAG TATA sequences did. When AGCE1 was placed at the downstream of SV40 core promoter, the responsiveness to hAG upstream region was significantly restored. Furthermore, mutation and in vivo competition experiments suggested that AGCF1 acts as a critical regulator of hAG transcription by mediating the activity of the hAG upstream and downstream enhancer elements. DNase I footprinting and UV cross-linking analyses showed that AGCF1 with apparent molecular masses of 31, 33, and 43 kDa as the components protected the region from −26 to −9 which partially overlapped with the TATA box consensus sequences. These findings indicate that AGCE1 in addition to the TATA box plays a key role in mediating the hAG regulatory elements.

A number of transcriptional DNA control elements, such as upstream and core promoter elements, have been identified as being important in affecting promoter activity of eukaryotic genes transcribed by RNA polymerase II. These elements direct the action of two classes of transcription factors, sequencespecific regulatory factors and general transcription factors, the former regulates the rate of transcription initiation and the latter is essential for initiating a basal level of transcription (1)(2)(3). Although core promoter elements have been considered to be important for general reaction of transcription initiation, several studies have indicated that these elements play a key role in determining the characteristic pattern of gene expression. For example, a muscle-specific enhancer activates transcription when fused to the myoglobin core promoter but has no influence upon the transcription from SV40 early core pro-moter (4). A 15-bp 1 core promoter region that includes the TATA box of the growth hormone gene has been shown to regulate its expression through the pituitary-specific transcription factor GHF-1 (5). Moreover, the degree of transcriptional stimulation by activating transcription factor is dependent upon the context of core promoter elements (6). Therefore, these observations provide evidence that each core promoter element would have intrinsic differences dependent upon its promoter context.
Angiotensinogen is the precursor of angiotensin II that acts as a physiologically important regulator of blood pressure and electrolyte homeostasis as well as a growth factor of cardiac myocytes (7)(8)(9)(10). Genetic linkage analyses proposed association between essential hypertension and molecular variants of the human angiotensinogen (hAG) gene (11)(12)(13). Our previous studies have shown that cis-acting elements located at nucleotide positions Ϫ1222 to ϩ44 are sufficient for the hAG gene expression in transiently transfected human hepatoma HepG2 cells and in the liver and the neuroectodermal tumor of the transgenic mice (14 -17). In particular, the removal of the DNA element located at Ϫ16 to ϩ44 of the transcription start site led to the dramatic reduction of its transcriptional activity, suggesting the presence of a core promoter with a functional importance for transcription (14). Furthermore, we recently identified several regulatory elements of the hAG gene including the downstream enhancer elements (18 -20), but it is still unknown how these cis-acting sequences activate the hAG promoter. In the present study, to examine the putative hAG core promoter function and the mechanism of action of hAG transcriptional elements, we have analyzed the core promoter element and identified a ubiquitously expressed nuclear factor, AGCF1 (hAG core promoter binding factor 1) that bound to AGCE1 (hAG core promoter element 1; nucleotide positions Ϫ25 to Ϫ1), which plays a major role in mediating the hAG enhancer function.
Site-directed Mutagenesis-13cat, DM10cat, and DM12cat were used as templates to construct mutations in TATA box and AGCE1 by oligonucleotides-directed mutagenesis (21). Once the site-directed mutations were obtained and confirmed by sequencing, the altered 1266-bp (position Ϫ1222 to ϩ44), 150-bp (Ϫ106 to ϩ44), or 76-bp (Ϫ32 to ϩ44) fragments were used for various constructions.
Cell Culture and Transient Expression Assays-HepG2 cells were maintained in minimum essential medium containing 10% fetal bovine serum and nonessential amino acids. The cells were plated at a density of 5 ϫ 10 5 cells/60-mm dish and transfected 24 h later by calcium phosphate co-precipitation (22) with reporter plasmids (3 g) and a ␤-galactosidase expression plasmid, pCH110 (1 g), to normalize transfection efficiency. After 48 h of culture, ␤-galactosidase activities were measured, and cell extracts containing equivalent amounts of ␤-galactosidase activity were used for CAT assay (23). In the competition assay, 2 g of reporter plasmids were transfected with 1, 2, or 4 g of competitive plasmids. Total amounts of DNA were adjusted 6 g by pUC119. An aliquot of cell extracts containing equal amounts of total protein (40 g) was used in CAT assay. The extent of conversion of chloramphenicol to its acetylated form was measured with a Bio-Imaging analyzer (model BAS2000; Fujix, Tokyo, Japan). All experiments were performed at least four times for each construct.
Preparation of Nuclear Extracts-Nuclear extracts from HepG2 cells were prepared using the protocol of Dignam et al. (24). The final protein concentration was about 5 mg/ml.
Electrophoretic Mobility Shift Assays (EMSA)-Single-stranded oligonucleotides were annealed, and the double-stranded DNA probe (100 ng) was end-labeled using [␥-32 P]ATP and T4 polynucleotide kinase. Five micrograms of nuclear extracts were incubated with 1 g of poly[d(I-C)] (Boehringer Mannheim) and end-labeled oligonucleotide (0.5 ng, approximately 15,000 cpm) at 20°C for 15 min in the presence or absence of the unlabeled oligonucleotides. The binding reaction was carried out in a solution containing 12 mM Hepes (pH 7.9), 60 mM KCl, 4 mM MgCl 2 , 1 mM EDTA, 12% glycerol, 1 mM dithiothreitol, and 0.5 mM phenylmethylsulfonyl fluoride. The reaction mixtures, final volume of 20 l, were directly loaded onto a 4.5% nondenaturing polyacrylamide gel (9:1, acrylamide/bisacrylamide) containing 4% glycerol made in 1 ϫ TBE (90 mM Tris-HCl (pH 8.0), 89 mM boric acid, and 2 mM EDTA) that had been pre-electrophoresed for 20 min. After electrophoresis was performed at 130 V for 2.5 h at 4°C, the gels were dried and autoradiographed with an intensifying screen.
DNase I Footprinting-On the basis of EMSA, ammonium sulfate precipitation at 40% saturation was a necessary step to remove the nonspecific DNA binding activity. Only nonspecific DNA binding activity remained in the ammonium sulfate supernatant (data not shown). The BglII site of DM10cat was end-labeled with T4 polynucleotide kinase and [␥-32 P]ATP, followed by digestion with HindIII to generate a probe for DNase I footprinting. After gel purification, the probe (approximately 20,000 cpm) was incubated with 2 g of poly[d(I-C)] and 125 g of 40% saturation ammonium sulfate precipitation in the 50 l of EMSA-binding buffer. The mixture was incubated for 20 min on ice, followed by 2 min at 25°C by the addition of 5 l of DNase I footprinting solution (4, 10, or 20 ng/l DNase I, 5 mM CaCl 2 , and 10 mMMgCl 2 ). The reaction was stopped by the addition of 100 l of 20 mM EDTA, 0.2% SDS, 300 mM NaCl, and 25 ng/l salmon sperm DNA. The digested DNA was extracted with phenol/chloroform (1:1, v/v) and precipitated with 2.5 volumes of ethanol prior to electrophoresis on an 8% polyacrylamide, 8 M urea sequencing gel. In order to define the position of the protected region, G ϩ A sequence ladders were prepared (data not shown).
UV Cross-linking-The DNA probe was generated by primed synthesis after annealing of a 10-mer primer 5Ј-GATCTAGGGC-3Ј to a 33-mer oligonucleotide containing AGCE1 5Ј-GATCCCCCTGGCCGGGTCAC-GAGGCCCTAGATC-3Ј using [␣-32 P]dCTP, dATP, dGTP, and 5-bromodeoxyuridine triphosphate. Ten micrograms of nuclear extracts were incubated with 4 ng (approximately 250,000 cpm) of probe and 1 g of poly[d(I-C)] in the 50 l of EMSA-binding buffer. After incubation for 20 min at 20°C, the samples were exposed to a UV source (254 nm) at a distance of 5 cm for 30 min at 4°C. One microliter of UV-DNase I solution (10 g/l DNase I, 100 mM CaCl 2 , and 100 mM MgCl 2 ) was added, and the complex was digested for 30 min at 37°C. The reaction was terminated with 1 l of 250 mM EGTA. After addition of 15 l of 3 ϫ loading buffer, the samples were boiled for 3 min and electrophoresed on a 12.5% SDS-polyacrylamide gel. The gel was fixed in 10% acetic acid, dried, and exposed for autoradiography.

RESULTS
In order to examine the functional significance of putative hAG core promoter (nucleotide positions Ϫ32 to ϩ44) responsible for its upstream DNA region (positions Ϫ1222 to Ϫ36), we first constructed a series of mutant promoters linked to the CAT reporter gene and analyzed their promoter activities in transiently transfected HepG2 cells (Fig. 1). Substitution of SV40 core promoter for hAG core promoter 13STcat did not produce the CAT activity at all, as observed in 13-36cat that lacks hAG core promoter. In contrast, the addition of hAG core promoter to 13-36cat, 13ATcat, restored the CAT activity to the levels similar to those found in the wild-type promoter, 13cat, although hAG core promoter alone represented little activity (data not shown). These two core promoters had TATA box and CAP site but no consensus initiator sequence (25) and apparently differed in the sequences of their TATA boxes. 13Tmcat with substitution mutation of hAG TATA box sequences (TATAAAT) to SV40-type sequences (TATTTAT) in the wild-type promoter context moderately reduced the promoter activity but still retained the efficient activity. These results indicated that hAG and SV40 core promoters have functional differences in response to the hAG upstream DNA region and suggested that the effective responsiveness of hAG core promoter is partially attributable to the TATA box element but largely dependent upon the other potential constituents of hAG core promoter.
To identify a functional element in the hAG core promoter region, we tried to find nuclear factors that could bind to this region by EMSA using the dissected DNA fragments that cover with the core promoter region. Incubation of AGCE1 fragment with the nuclear extracts prepared from HepG2 cells produced retarded complexes, AGCF1 (Fig. 2B), which represented a sequence-specific interaction between AGCE1 fragment and a FIG. 1. Evidence for a functional distinction between the hAG core promoter and heterologous SV40 core promoter. Left, reconstituted constructs of the hAG promoter-CAT hybrid genes. Thick lines, oval, open boxes, and hatched box represent wild-type promoter sequences, SV40 core promoter sequences (114-bp NcoI/HindIII fragment containing TATA box and initiation site), wild-type TATA box, and SV40-type TATA box whose base substitutions are indicated below the box, respectively. Right, HepG2 cells were transfected with 3 g of the indicated CAT vectors and 1 g of the ␤-galactosidase expression plasmid (pCH110) as an internal control for transfection efficiency. After a 48-h culture, ␤-galactosidase activities were measured and extracts containing equivalent amounts of ␤-galactosidase activities were used for CAT assays. The CAT activity of 13cat is designated as 100 and each value of CAT activity represents the mean Ϯ S.E. for at least four independent experiments. N.D., not detected. nuclear factor, since the formation of this complex was specifically reduced with molar excess of unlabeled competitors (Fig.  2B, lanes 1-6). Moreover, double-stranded oligonucleotides containing the consensus binding sequences for C/EBP, ATF, Sp1, AP-1, AP-2, AP-3, and NF-I/CTF failed to compete with AGCF1-binding activity (Fig. 2B, lane 7, and data not shown). The DNA-protein complex formed by AGCF1 binding to AGCE1 was, however, inhibited by molar excess of nonlabeled AGCE1, Am2, and Am3 (Fig. 2, A and B). Although Am1 could partially prevent this complex formation, Am4 did not compete for this binding at all (Fig. 2B, lanes 8-14).
To evaluate the role of AGCE1 in the native promoter context, the Am4 mutations that abolished AGCF1-binding activity were introduced into the hAG promoter. The DM10cat, which could supply a sufficient CAT activity in our previous experiments (16), was used as the wild-type promoter for the convenience of constructions. As shown in Fig. 3A, DM10Am4cat with the Am4 mutation decreased the transcription activity to 10% of the wild-type construct, whereas DM10TGmcat with a nonsense TATA box (TATGGAT) (26) reduced the activity by 50%. To examine whether AGCE1 provides the functional differences between hAG and SV40 core promoters in addition to TATA box, we reconstituted the chimeric core promoters by fusing SV40 core promoter with AGCE1 and assayed their promoter activities (Fig. 3B). As expected, AGCE1 could significantly restore the core promoter activity under the control of hAG upstream DNA region. Taken together, these findings suggested that AGCE1 in addition to TATA element is critical for the efficient hAG promoter activity.
To define the contribution of AGCE1 to hAG transcription in the whole gene context, we reconstituted the hAG gene containing its 1.3-kilobase promoter (13 fragment) and the downstream enhancer (B2 fragment) that is composed of the three enhancer core elements, d61-2, GM, and ME (18,19) (Fig. 4A). When the Am4 mutation was introduced into the 13 fragment, the mutated promoter activity decreased to 40% as compared with that of the wild-type sequence (Fig. 4B, 13cat and  13Am4cat). Furthermore, the transcriptional activity dramat-ically dropped by the Am4 mutations in the reconstituted hAG gene even when the three sets of enhancer core elements are included (Fig. 4B). These results indicated that AGCE1 is an authentic regulatory element for hAG transcription.
We have previously identified hASR in the upstream region (20) and the three enhancer core elements in the downstream region (18,19) as hAG regulatory elements. Recently, we found  Fig. 2A, and mutated TATA box whose base substitutions are indicated below the box, respectively. Right, HepG2 cells were transfected with the indicated CAT vectors, and CAT assays were performed as described in Fig. 1. The CAT activity of DM10cat is designated as 100, and each value of CAT activity represents the mean Ϯ S.E. for at least four independent experiments. B, reconstitution of the hAG core promoter. Left, the reconstituted constructs of the hAG promoter-CAT hybrid genes. Oval and hatched box represent SV40 core promoter sequences and SV40type TATA box whose base substitutions are indicated in Fig. 1, respectively. Hatched box marked AGCE1 represents the AGCF1-binding site (Fig. 2). Right, the CAT activity of 13Tmcat is designated as 100 and each value of CAT activity represents the mean Ϯ S.E. for at least four independent experiments. N.D., not detected. another upstream element, ALE (ATF-Like Element), of which deletion reduced the hAG promoter activity by 50% and the binding of multiple factors including CREB/ATF family and novel ones to ALE. 2 Therefore, we next examined the mechanism of action of AGCE1 in response to these hAG regulatory elements (Figs. 4A and 5). When hASR was placed in front of a native or Am4-mutated hAG core promoter fragment, it functioned efficiently in both the core promoter contexts. In contrast, the degrees of stimulation by ALE, d61-2, GM, and ME dramatically decreased upon the Am4 mutations. To confirm the functional importance of AGCF1, we performed in vivo competition experiments (Fig. 6). The various amounts of pUC-AGCE1, which included eight tandem copies of AGCE1 in pUC119, were cotransfected with DM12d61-2(Ϫ) or SV3cat. The CAT activity derived from DM12d61-2(Ϫ) decreased with sequential titration of AGCF1 binding, although the activity driven from a positive control, SV3cat, was little influenced. These results suggested that the responsiveness of hAG core promoter to the activation by factors bound to ALE and the hAG downstream enhancer core elements are largely dependent upon AGCF1, whereas the factors bound to hASR function independently of AGCF1.
We decided to delineate the AGCF1 binding site more thoroughly, since this location between the TATA box and transcription initiation site is very unique. DNase I footprinting using 40% saturation ammonium sulfate precipitation of HepG2 nuclear extracts showed a protected region from Ϫ26 to Ϫ9 (Fig. 7A, denoted by bracket). Interestingly, the protected region partially overlapped with the consensus sequences of the TATA box (TATAAAT). Finally, UV cross-linking was performed to determine the molecular masses of the nuclear factors involved in AGCF1 (Fig. 7B). Ten micrograms of HepG2 nuclear extracts were incubated with 5-bromodeoxyuridine-substituted AGCE1 probe in the absence or presence of the unlabeled oligonucleotides. The mixtures were UV crosslinked, digested with DNase I, and analyzed on a 12.5% SDSpolyacrylamide gel. Factors corresponding to apparent 31, 33, and 43 kDa were identified as the major components of AGCF1 (Fig. 7B, lane 1), since these factors were specifically eliminated by the molar excess of unlabeled AGCE1 fragment (lane 2) but not by that of unlabeled C/EBP fragment (lane 3).

DISCUSSION
In the present study, we have analyzed hAG core promoter region and found that hAG core promoter but not SV40 early core promoter responds to hAG upstream region (Fig. 1). This functional difference was likely due to the major action of hAG core promoter element. We identified a potential DNA element, AGCE1 (nucleotide positions Ϫ25 to Ϫ1), that interacted with a nuclear factor, AGCF1 (Fig. 2). Mutation analyses and reconstitution experiments combined with CAT assays suggested that AGCE1 in addition to TATA element plays a key role in maintaining the hAG promoter activity (Fig. 3). Furthermore, AGCF1 acts as an authentic regulator of hAG transcription (Fig. 4) by mediating the responsiveness to one of the upstream elements, ALE, and the downstream enhancer core elements (Figs. 5 and 6). DNase I footprinting and UV cross-linking experiments indicated that AGCF1 with apparent molecular masses of 31, 33, and 43 kDa protected the region from Ϫ26 to Ϫ9 that partially overlapped with the consensus sequences of the TATA box (Fig. 7).
It has been reported that the functional diversity of multiple classes of TATA sequences plays an important role in responding to the regulatory elements. For example, the sequence differences between myoglobin TATA motif and SV40 early TATA element are responsible for the differential response to muscle-specific enhancer (4). In another instance, heat shock protein 70 and EIIa TATA elements confer high activity on the ATF site whereas the SV40 early TATA motif hardly responds to it (6). In contrast to these observations, a conversion of hAG TATAAAT sequence to SV40 TATTTAT in the hAG promoter context resulted only in the moderate reduction of the promoter activity, although hAG core promoter but not SV40 early core promoter responds to hAG upstream region (Fig. 1).
Several lines of evidence have demonstrated that the first step in preinitiation complex assembly involves binding of a multiprotein complex TFIID to the promoter via sequencespecific interactions between the TATA box and TFIID at TATA-containing promoters (27)(28)(29)(30). The TFIID complex is proposed to be recruited by a "tethering factor," such as Sp1, which could function as a substitute for TATA box via proteinprotein interactions between the Sp1 and TFIID at TATA-less promoters (31)(32)(33). Aird et al. (34) reported that the rat platelet factor 4 gene contained a GATA motif in place of the TATA sequence, and this GATA motif functioned as a repressor by competition between GATA-binding proteins and basal factors for the core promoter. The similar repression mechanism was observed in several genes (35,36). In the case of AGCF1 that binds to AGCE1 (Fig. 2), it is necessary for the hAG core promoter activity (Fig. 3), although the protected region from DNase I was partially overlapped with the TATA sequences (Fig. 7A). Moreover, the mutations that abolished AGCF1-binding activities (Fig. 2) affected the hAG promoter activity more severely than a nonsense mutation of the TATA element did (Fig. 3). It is conceivable that the efficient activity even in the mutated hAG core promoter along with a nonsense TATA element may be maintained by a tethering activity of AGCF1 in the presence of TFIID.
A positive regulatory element located around or downstream of the TATA box has recently been identified and discovered to possess an intrinsic function for cell type-specific regulation. The human gastrin gene fragment from Ϫ17 to ϩ57 shows a considerable cell type specificity in transfection experiments (37). The cell type specificity determinant of the peripherin gene is localized to a region overlapping the TATA box, whereas  lines represent the wild-type promoter sequences, wild-type TATA box, and AGCE1 mutations whose base substitutions are indicated in Fig. 2A, respectively. Boxes marked hASR, ALE, d61-2, GM, and ME represent the upstream elements, hASR and ALE, and downstream enhancer core elements, d61-2, GM, and ME, respectively. Right, HepG2 cells were transfected with the indicated CAT vectors, and CAT assays were performed as described in Fig. 1. The CAT activity of DM12cat is designated as 1.0, and each value of CAT activity represents the mean Ϯ S.E. for at least four independent experiments.
FIG. 6. In vivo competition analysis of the AGCE1. A competitive plasmid, pUC-AGCE1, containing eight head-to-tail tandem repeats of the AGCE1 fragment was used in this experiment. Two micrograms of reporter plasmids (SV3cat or DM12d61-2(Ϫ)) were cotransfected with 1, 2, or 4 g of pUC-AGCE1. Total amounts of DNA were adjusted to 6 g by pUC119. After a 48-h culture, the protein concentration was measured, and aliquots of cell extract containing equal amounts of total protein (40 g) were used in CAT assay. The CAT activity of each reporter plasmid, cotransfected with 4 g of pUC119, is designated as 100, and each value of CAT activity represents the mean Ϯ S.E. for at least four independent experiments. elements determining the strength of the promoter are localized upstream (38). Interestingly, the core promoter (Ϫ36 to ϩ12) of the myelin basic protein gene is shown to direct brainspecific in vitro transcription (39). In addition to the above elements that interact with cell-specific nuclear factors, other examples of core promoter regions have been reported including the pro-opiomelanocortin gene (Ϫ18 to ϩ6) that bind to widely distributed nuclear factors (40). Despite the positions between the TATA box and transcriptional initiation site, there was no extensive homology between AGCE1 and the DNA sequences of those previously identified core promoter elements (data not shown).
Our striking finding in the present study is that AGCE1 plays a critical role in activating hAG transcription by one of the upstream elements, ALE, or the downstream enhancer core elements, but not by another upstream element, hASR (Figs. [3][4][5]. In other words, AGCF1 can mediate the transcriptional activating function of ALE-and downstream core elementbinding factors. In the case of the latter factors, at least in part, this notion is supported by the sequential titration of AGCF1 binding by means of the in vivo competition (Fig. 6). On the other hand, other core promoter elements, for example TATAbox sequences, may be needed for hASR function. Taken together, we now hypothesize that the cis-acting sequences of hAG gene selectively activate its transcription through the diversity of core promoter elements. Further study will be necessary to define, by means of purifying AGCF1 and cloning its cDNA, the molecular mechanism by which it can respond to the upstream or downstream regulators of the hAG gene. Such investigations are now in progress in our laboratory.