Cloning and characterization of a functional promoter of the rat pp120 gene, encoding a substrate of the insulin receptor tyrosine kinase.

Cloning of the 5′-flanking region of the rat pp120 gene has indicated that it is a housekeeping gene: it lacks a functional TATA box and contains several Sp1 binding sites and multiple transcription initiation sites at nucleotides −101, −71, −41, and −27 spread over a GC-rich area. A fragment between nucleotides −21 and −1609 exhibited promoter activity when ligated in a sense orientation into a promoterless luciferase reporter plasmid and transiently transfected into rat H4-II-E hepatoma cells. 5′ progressive deletion and block substitution analyses revealed that the three proximal Sp1 boxes (boxes 3, 5, and 6) are required for basal transcription of the pp120 gene. Promoter activity was stimulated 2-3-fold in response to insulin, dexamethasone, insulin plus dexamethasone, and cAMP. Although unaltered by phorbol esters alone, promoter activity was stimulated 4-5-fold in response to phorbol esters plus cAMP. Several motifs resembling response elements for insulin (in the rat phosphoenolpyruvate carboxykinase gene), glucocorticoids, cAMP, and phorbol esters as well as a number of putative binding sites for activating proteins-1 (Jun/Fos) and −2, and liver-specific factors were detected. The role of these sites in tissue-specific expression of pp120 remains to be investigated.

Molecular cloning revealed that the rat pp120 gene consists of nine exons, the 7th of which (53 bp) is alternatively spliced during mRNA processing, generating two alternatively spliced variants that differ in the intracellular cytoplasmic domain (11). The truncated isoform lacks 61 of the 71 amino acid cytoplasmic domain, including basal (Ser 503 ) and insulin-stimulated (Tyr 488 ) phosphorylation sites (4).
The function of pp120 remains unclear. We have observed that the rate of insulin-stimulated internalization and degradation of the ligand-receptor complex is 2-3-fold higher in NIH 3T3 cells coexpressing insulin receptors and pp120 than in cells expressing insulin receptors alone, suggesting a role for pp120 in the hepatic clearance of insulin from circulation (12). Since pp120 may play a physiological role in insulin metabolism, it has become essential to understand the mechanism of its in vivo expression, both basal and regulated.
Glucocorticoid treatment of rats increased pp120 protein level 2-3-fold in liver plasma membrane (1). To understand the molecular basis for the regulation by glucocorticoids, we have cloned the 5Ј-flanking region and demonstrated a functional promoter activity of the pp120 gene. This region lacks a TATA box and contains multiple transcription initiation sites as well as potential binding sites for basal and regulatory transcriptional elements. Additionally, we have observed that Sp1 plays an important role in determining basal promoter activity of the gene.

EXPERIMENTAL PROCEDURES
Screening of Genomic Library-Detection of a -2 positive plaque upon screening of a rat genomic DNA library (Clontech, Palo Alto, CA) by hybridization with a 32 P-labeled full-length rat liver ecto-ATPase cDNA was described previously (11). The genomic DNA insert that contains exon 1-4 and ϳ20 kb of the 5Ј-flanking region (11) was subcloned into pBluescript II KS ϩ plasmid vector (Promega, Madison, WI) at the XhoI sites using T4 DNA ligase (TaKaRa, Berkeley, CA). A plasmid clone of Ͼ10 kb of genomic DNA (p-2(7)) containing exon 1 and its 5Ј-flanking region was identified by hybridization with a 32 P-labeled exon 1-specific primer (␣-60, see Table I). Similarly, a plasmid clone p-2(12) of ϳ2.2 kb containing exon 2 and parts of intervening introns 1 and 2 was identified by hybridization with a 32 P-labeled exon 2-specific primer (␣-100, see Table I). Hybridization was carried out as described previously (11).
DNA Sequencing-Plasmid clone p-2 (7), which contains pp120 5Јflanking region, was treated with either EcoRI or HindIII, and the * This work was supported in part by an American Diabetes Association Research Award (to S. M. N.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EMBL Data Bank with accession number(s) U27207 and U27208 for the 5Ј-flanking region and intron 1 sequences of the pp120/ecto-ATPase gene, respectively, and HUMCEA01 for the 5Ј-flanking region of the human CEA gene (16). ¶  1 pp120 has been designated pp120/ecto-ATPase, pp120/HA4, pp120/ C-CAM. We have used the designation pp120/ecto-ATPase in this report to be consistent with our original publication on the cloning of the gene encoding the translated region of pp120/ecto-ATPase (11). For brevity, we have used pp120 for the subsequent part of the manuscript. 2 The abbreviations used are: CAM, cell adhesion molecule; CEA, carcinoembryonic antigen; bp, base pair(s); kb, kilobase(s); nt, nucleotide(s); PCR, polymerase chain reaction; GH, growth hormone; hGH, human growth hormone; PMA, phorbol 12-myristate 13-acetate; TAP, tobacco acid pyrophosphatase; HNF, hepatic nuclear factor; LF, liver factor; AP, activating protein; PEPCK, phosphoenolpyruvate carboxykinase.
Genomic DNA Cloning-Double-stranded genomic DNA corresponding to nt Ϫ1609 to Ϫ21 was synthesized and amplified by Vent polymerase (New England Biolabs) in a polymerase chain reaction (PCR) using 1 g of p-2(7-HindIII) as template and 100 ng of each of sense s(Ϫ1609) and antisense ␣(-21) primers. Following initial DNA denaturation at 94°C for 5 min, 30 cycles of PCR were carried out as described previously (11).
An aliquot of amplified DNA (50 ng) was ligated into pCRII plasmid (Invitrogen, San Diego). Constructs containing the genomic DNA insert in either sense or reverse orientation were treated with XhoI and HindIII, and the DNA was analyzed on 1% agarose gel and electroeluted in 0.2 ϫ TBE (Tris-borate-EDTA) buffer at 60 V for 3 h. Purified genomic DNA was then ligated at the XhoI and HindIII sites of pGL3-BASIC promoterless firefly luciferase reporter plasmid (Promega).
The same PCR technique was employed to synthesize the 5Ј deletion products using the same ␣(-21) antisense primer and different sense 21-mer primers whose 5Ј ends were at nt Ϫ439, Ϫ249, Ϫ194, Ϫ147, Ϫ131, Ϫ124, and Ϫ112, respectively. Amplified genomic DNA was subsequently subcloned at the XhoI and HindIII sites of pGL3-BASIC plasmid.
Scanning mutants between nt Ϫ209 and Ϫ128 were created in the Ϫ249/Ϫ21 promoter fragment by replacing 20 bp of the native sequence with 20 bp of a heterologous sequence that includes an EcoRI restriction site (5Ј-CTCgaattcGATGGCGGTAT-3Ј) in a PCR-based reaction. Mutant promoters were prepared by EcoRI restriction and ligation of the appropriate pairs of the 5Ј and 3Ј fragments. The mutant promoter fragments were cloned into the XhoI/HindIII sites of pGL3-BASIC plasmid. Using this procedure, three mutant constructs were synthesized and designated as Ϫ249Mut3pLuc, Ϫ249Mut5pLuc, and Ϫ249Mut6pLuc containing block substitution mutation from nt Ϫ209 to Ϫ189, Ϫ169 to Ϫ149, and Ϫ148 to Ϫ128, respectively.
Cell Culture and Transfection of H4-II-E Cells-H4-II-E cells, derived from the well differentiated Reuber H35 rat hepatoma (13) and expressing endogenous pp120 (3), were maintained in RPMI 1640 (Biofluids Inc., Rockville, MD) supplemented with 10% fetal calf serum (Upstate Biotechnology, Inc., Lake Placid, NY), 100 units/ml penicillin (Sigma), and 10 g/ml streptomycin (Sigma) at 37°C, 5% CO 2 . Cells were adapted to Dulbecco's modified Eagle's medium (Biofluids Inc.) two passages before transfection, plated in 60-mm dishes at 50 -70% confluence, and then transfected with 5 g of the pGL3-pp120 construct using the DEAE-dextran method. To correct for transfection efficiency, cells were cotransfected with 1.4 g of pXGH5 human growth hormone plasmid driven by the mouse metallothionein-1 promoter (Nichols Institute, San Juan Capistrano, CA). 100 l of DEAE-dextran (1 mg/ml) in Tris-buffered saline, pH 7.5, was added to the DNA construct for 10 min at room temperature before addition to the cells at room temperature for 15 min. After addition of 3 ml of serum-supplemented Dulbecco's modified Eagle's medium, cells were allowed to incubate at 37°C, 5% CO 2 for 4 h. Medium was then replaced with fresh complete Dulbecco's modified Eagle's medium and the cells incubated at 37°C, 5% CO 2 for 24 h. One ml of medium was then collected to assay secreted hGH as a measure for transfection efficiency (14), and cells were incubated at 37°C, 5% CO 2 for additional 24 h in complete medium supplemented with either 1 g/ml insulin (Sigma), 10 Ϫ6 M dexamethasone (Sigma), 0.5 mM phorbol 12-myristate 13-acetate (PMA, Sigma) or 1 mM cAMP (Sigma). At the end of the incubation period, cells were lysed, and the luciferase activity was measured (14) using an automated luminometer (Lumat, LB 9501, Berthold, Gaithersburg, MD).
Tobacco Acid Pyrophosphatase (TAP) Reverse Ligation PCR-The TAP (Epicenter Technologies, Madison, WI) reverse ligation PCRs were carried out as described previously (15). Total rat liver RNA (10 g) was treated with DNase I (10 units, Boehringer Mannheim) at 37°C for 30 min to remove residual DNAs. The phosphate groups at the 5Ј end of partially degraded RNAs were then removed by calf intestinal alkaline phosphatase (50 units Table I) as exon 1-specific antisense primers. The second PCR was carried out using 10% volume of the first PCR as template. After initial DNA denaturation at 94°C for 5 min, 20 (first reaction) or 25-30 (second reaction) cycles of PCR were carried out as described previously (15).
The nucleotide sequence of intron 1 (769 bp) was similarly determined using plasmid clone p-2(12) as template (data not shown) and beginning with ␣-100, an exon 2-specific primer, and plasmid-specific T3 and T7 primers.
Sequence analysis by MacVector 4.5 program (IBI) revealed the presence of several putative DNA-binding elements that may play a role in basal and regulated pp120 transcriptional activity. These include multiple putative Sp1 binding sites in the proximal promoter (Table II) and a number of motifs conforming to the consensus binding site for HNF-1/HP-1 liverspecific transcription factor (nt Ϫ1624 to Ϫ1612) and for LF-A1-RS (Table III), a trans-acting factor that is required for the expression of several genes such as human apolipoprotein A1, FIG. 1. Map of the 5-flanking region and intron 1 of the pp120 gene. Genomic DNA of -2, a bacteriophage containing the 5Ј-flanking region of the pp120 gene (11), was cleaved at the intronic XhoI sites (X) to yield different DNA fragments that were subcloned into the pBluescript II KS ϩ plasmid vector. p-2(12) subclone contains exons 2 and most of the intervening introns 1 and 2 sequences, whereas p-2(7) contains exon 1 and greater than 10 kb of the 5Ј-flanking region of the gene. p-2(7) was further cleaved with either HindIII (H) or EcoRI (E) and self-religated to yield p-2(7-HindIII) and p-2(7-EcoRI), respectively. haptoglobin-related genes, and human ␣ 1 -antitrypsin in hepatocytes (17). Overlapping sequences matching perfectly the consensus binding sequence for HNF-5 (nt Ϫ1359 to Ϫ1373), a liver-specific factor in the rat tyrosine aminotransferase gene (18), were characteristically found in proximity to other liverspecific factor binding sites such as CAAT enhancer-binding protein (nt Ϫ1376 to Ϫ1346) binding sites in the liver-specific rat tyrosine aminotransferase gene (TTTGTTTT, Ref. 18).
Potential sites for activating proteins AP-1 (Jun/Fos) and AP-2 were also found (Table IV). Nucleotide sequences matching 9 out of 10 bp in the insulin response element of the rat phosphoenolpyruvate carboxykinase gene (PEPCK, Ref. 19), 12 out of 15 bp in the loose consensus sequence of glucocorticoid response element (20), 8 out of 8 bp in TPA response element (21), and 6 out of 6 bp in cAMP response element (22) were also identified ( Table IV).
The Rat pp120 Gene Is Transcribed from Four Initiation Sites in the 5Ј-Flanking Region-Initial attempts to map the cap sites by primer extension and nuclease protection assays failed, perhaps because of the low abundance of mRNA encoding pp120 and the potential formation of mRNA secondary structures due to the high GC content of the proximal promoter. To overcome this problem, we used a more sensitive PCR-based assay, TAP reverse ligation PCR (15,23). In these experiments, an RNA linker was ligated to the exposed 5Ј end of mRNA, and the linked mRNA was used as template to FIG. 2. Nucleotide sequence of the 5-flanking region of the rat pp120 gene. Numbers on the right are nucleotide number relative to ϩ1 at the A ϩ1 TG translation initiation codon (indicated by an arrow). Nucleotides upstream of this codon are labeled with negative numbers. Potential regulatory cis-elements are indicated either by boxes or underlining. They include binding sites for HNF-1 and HNF-5, CAAT enhancer-binding protein (C/EBP), liver factor-A1 (LF-A1-RS), activating proteins-1 and -2 (AP-1 and AP-2), glucocorticoid response element (GRE), insulin response element (IRE), and Sp1. Sp1 boxes are in bold and are designated Sp1 box 1-6. with two nested exon 1-specific antisense primers (␣-38 and ␣-19, Table I). The size of the PCR products was measured by comparison with a sequencing ladder using p-2(7-HindIII) as template and ␣-19 as primer (Fig. 3). Four specific PCR prod-TABLE II Nucleotide sequences matching to the consensus binding sites for Sp1 in the 5Ј-flanking region of the pp120 gene The nucleotide sequence of the 5Ј-flanking region of the pp120 gene was analyzed on the MacVector 4.5 software program. Several sequences matching with different degrees of identity (78 -100%) to the consensus binding sites for Sp1 were found. Nucleotide C in the GC box of the consensus sequence is underlined.

Consensus sequence
Sequence in the pp120 gene Matches Box no.

5Ј-G/T G/A GGC G/T G/A G/A G/T-3Ј
Indicates that the nucleotide sequence that matches to a consensus sequence in the 5Ј-flanking region of the pp120 gene is found in the reverse orientation.
b Unmatched nucleotides are represented by lowercase letters.  13/13 (Ref. 55) a Indicates that the nucleotide sequence that matches to a consensus sequence in the 5Ј-flanking region of the pp120 gene is found in the reverse orientation.
b The consensus sequence of CAAT enhancer-binding protein (C/EBP) binding sites is more loose than that of other liver-specific factors. This consensus sequence was compiled by combining results in both reports.  The consensus sequence for glucocorticoid response elements is loose, but in general, the symmetrical G and C (underlined) around the triple N box (twice underlined) is characteristic of cis-elements that mediate the regulatory effect of glucocorticoids on genes expression. c Indicates that the nucleotide sequence that matches to a consensus sequence in the 5Ј-flanking region of the pp120 gene is found in the reverse orientation. d cAMP response element-binding protein.
ucts of 145 (band A), 115 (band B), 85 (band C), and 71 bp (band D) were thus obtained at an annealing temperature of 54°C, 5°C below the melting temperature of ␣-19 (Fig. 3). These bands persisted at annealing temperatures of up to 70°C (data not shown). However, none of these bands was detected when either TAP or reverse transcriptase was omitted from the reaction (Fig. 3). After subtracting the size (25 bp) of the DNA sense primer (DNA Pr-1) that corresponds to the RNA linker (Table I), transcription start sites were mapped to nt Ϫ101, Ϫ71, Ϫ41, and Ϫ27. Transcription start sites were mapped to these nucleotides even when another reverse transcriptase primer or different combinations of nested DNA primers were used (data not shown). The bands corresponding to the four transcription start sites were excised and the DNA eluted and reamplified in a PCR using DNA Pr-1 and ␣-19 as sense and antisense primers, respectively. Hybridization with a pp120 exon 1-specific primer (␣-10, Table I) confirmed specificity of all bands to the pp120 gene (data not shown). Subsequent sequencing of each of amplified bands A, B, C, and D by the thermostable DNA polymerase in the f-mole cycle sequencing kit (Promega) using ␣-19 as primer confirmed the mapping of distinct mRNA transcription initiation sites to nt Ϫ101, Ϫ71, Ϫ41, and Ϫ27 (data not shown). Since no TATA box was found properly positioned in proximity to the start sites, we conclude that the rat pp120 gene does not contain a functional TATA box. The region mapping the transcription start sites is GC-rich and contains a number of GC boxes (Table II) that may be recognized by Sp1 transcription factor (24). Demonstration of a Functional Promoter Activity in the 5Ј-Flanking Sequence-To determine whether the 5Ј-flanking sequence contains functional promoter activity, we directionally subcloned an amplified DNA fragment (nt Ϫ21 to Ϫ1609) at the XhoI/HindIII sites of the pGL3 promoterless luciferase plasmid. Lysates from H4-II-E rat hepatoma cells transfected with the sense construct showed greater than 30-fold increase in luciferase activity compared with lysates from cells transfected with a construct containing the promoter ligated in the reverse orientation relative to the luciferase gene ( Fig. 4; sense 1,812.3 Ϯ 535.49 versus reverse: 43.4 Ϯ 2.33, p Ͻ 0.05). There was no difference (p Ͼ 0.05) in the luciferase activity in lysates made from untransfected cells ( Fig. 4; untransfected 40.6 Ϯ 3.54) and cells transfected with the reverse construct ( Fig. 4; reverse 43.4 Ϯ 2.33). This suggests that the pp120 promoter functions in an orientation-specific manner.
Mapping of the Region Regulating Basal Promoter Activity-The proximal region of the pp120 promoter contains six potential Sp1 boxes ( Fig. 2 and Table II). Three of these potential Sp1 binding sites (boxes 4 -6) lie adjacent to one another in a 30-bp region between nt Ϫ143 and Ϫ114, with boxes 4 and 5 overlapping by one nucleotide. Boxes 1, 2, 4, 5, and 6 are variant isoforms in which the C of the GGCG core is replaced by A. Box 3 (nt Ϫ157 and Ϫ149) showed perfect homology, with 9 out of 9 FIG. 3. Identification of the transcription start sites for the rat pp120 gene. The TAP reverse ligation PCR method was applied to total rat livers. RNA was sequentially treated with DNase I, CIP, and TAP before ligation to an RNA linker (Table I) as described under "Experimental Procedures." Reactions were conducted in the absence (Ϫ) or presence (ϩ) of TAP and avian myeloblastosis virus reverse transcriptase as indicated. 100 ng of ligated RNA was reverse transcribed using exon 1-specific ␣-45 antisense primer and amplified with nested exon 1-specific ␣-38 and ␣-19 antisense primers (Table I). The PCR products were analyzed on a 6% polyacrylamide-urea sequencing gel. The size of the PCR products (bands A-D) was determined by comparison with a sequencing ladder obtained by the dideoxy sequencing of p-2(7-HindIII) plasmid using ␣-19 primer.

FIG. 4. The 5-flanking region of the rat pp120 gene contains functional promoter activity in H4-II-E rat hepatoma cells.
A ϳ1.6-kb fragment (nt Ϫ1609 to Ϫ21) from the 5Ј-flanking region was subcloned into a promoterless luciferase reporter plasmid in both sense and reverse orientation. These constructs were transiently transfected into H4-II-E rat hepatoma cells, and the luciferase activity was determined in cell lysates 48 h after transfection. A hGH reporter plasmid (pXGH5) was cotransfected to monitor transfection efficiency. For comparison, luciferase activity in lysates from untransfected cells was included (Untrsft). Results were normalized for GH secretion and expressed as mean Ϯ S.D. of triplicate transfections in relative light units. The graph represents typical results from four separate experiments. bp matches in the consensus site for Sp1-CS4 binding (Table  II). Interestingly, the Sp1 cluster (boxes 4 -6) is embedded in a region containing direct and inverted repeats (the CTCTGG-GAGG sequence is repeated between nt Ϫ143 and Ϫ134, and Ϫ126 and Ϫ117, and the CCCTCCTCT sequence at nt Ϫ131 to Ϫ123 is directly fused in direct tandem with its inverted complement, GGGAGGAGA, at nt Ϫ122 to Ϫ114). To determine whether any of these putative Sp1 boxes is important for basal promoter activity, we measured luciferase activity in H4-II-E cells transfected with pGL3 constructs containing promoter fragments in which individual Sp1 boxes were progressively removed from the 5Ј end (Fig. 5). As Fig. 5 reveals, deletion of the region between nt Ϫ1609 and Ϫ439 induced a slight increase in the basal promoter activity (Ϫ1609pLuc 2,444.9 Ϯ 820.81 versus Ϫ439pLuc 5,198.1 Ϯ 875.48; p Ͻ 0.05), suggesting that this region may contain potential down-regulators and confirming our initial observations that the putative TATA boxes located distally at nt Ϫ939, Ϫ1117, Ϫ1484, and Ϫ1532 are not functional. Removal of the sequence between nt Ϫ439 and Ϫ194 (which deletes boxes 1 and 2) did not lead to a significant change in promoter activity relative to the In addition to the 5Ј deletion analysis described above, a 20-bp nucleotide block substitution (which selectively mutated boxes 2, 3, 4/5) in the Ϫ249/Ϫ21 promoter fragment was also made to assess the functional significance of these respective Sp1 boxes in basal promoter activity (Fig. 6). Mutation of Sp1 box 2 did not alter promoter activity of the Ϫ249/Ϫ21 promoter fragment (Ϫ249Mut3pLuc 5,965.5 Ϯ 381.69 versus Ϫ249pLuc 4,422.4 Ϯ 971.42), confirming the previous observation that box 2 did not play a significant role in the overall pp120 promoter activity. However, mutation of the region spanning Sp1 box 3 (Ϫ249Mut5) or that spanning Sp1 box 4 and box 5 together (Ϫ249Mut6) in the Ϫ249/Ϫ21 promoter fragment led to a significant decrease (p Ͻ 0.05) in luciferase activity relative to the intact Ϫ249pLuc construct (Ϫ249Mut5pLuc 1,501.9 Ϯ 579.37 and Ϫ249Mut6pLuc 1,173.1 Ϯ 293.07 versus Ϫ249pLuc 4,422.4 Ϯ 971.42). Thus, selective block mutation of either box 3 or boxes 4/5 led to a comparable reduction in the overall activity of the pp120 promoter, indicating that these Sp1 binding sites are equally important. It is unclear whether boxes 4 and 5 are both essential for activity, since mutations in Ϫ249Mut6pLuc did target both of these boxes. However, 5Ј deletion analysis suggests that box 4 did not contribute to basal activity since deletion of box 4 did not further decrease basal activity of the pp120 promoter (Ϫ147pLuc versus Ϫ131pLuc). Taken together, 5Ј deletion and block mutation analyses suggest that the collective presence of Sp1 boxes 3, 5, and 6 is needed for full basal activity of the pp120 promoter in H4-II-E cells.
Regulation of the pp120 Promoter by Insulin, Glucocorticoids, PMA, and cAMP-The 5Ј-flanking region of the pp120 gene contains putative response elements for insulin, glucocorticoids, phorbol esters, and cAMP (Table IV). To determine whether the promoter activity is regulated by these agents, H4-II-E cells transfected with the full-length sense construct were treated with either insulin, dexamethasone, cAMP, PMA, insulin and dexamethasone, or with PMA and cAMP for 24 h, and luciferase activity was assayed in cellular lysates (Fig. 7). Insulin or dexamethasone treatment increased promoter activity 2-3-fold compared with nontreated cells, and this increase is additive in cells cotreated with both insulin and dexamethasone (insulin-treated, 4,040.2 Ϯ 287.44; dexamethasonetreated, 3,109.8 Ϯ 498.09; insulin-and dexamethasone-treated, 5,644.7 Ϯ 515.38 versus untreated, 1,812.3 Ϯ 535.49, p Ͻ 0.05). DISCUSSION pp120, a 120-kDa rat liver plasma membrane glycoprotein, is an endogenous substrate of the insulin receptor kinase (1,3,4). It is identical to a Ca 2ϩ /Mg 2ϩ ecto-ATPase (cell-CAM-105), possessing two ATP binding sites and three immunoglobulinlike loops in its extracellular domain (6,10). In this report, we have cloned the 5Ј-flanking region of the rat pp120 gene and showed that it contains functional promoter activity when transfected into H4-II-E rat hepatoma cells.
The 5Ј-flanking region of the pp120 gene shares ϳ60% homology with that of the human tumor marker CEA (Fig. 8), a member of the immunoglobulin superfamily gene (25). Since CEA has been implicated in neoplasia and cell adhesion (26) and pp120 has been proposed to play a role in insulin clearance from blood (12), identification of basal and regulatory DNA elements may therefore provide important insight into the molecular mechanisms underlying regulation of expression of these gene products.
Similar to immunoglobulin genes such as neural cell adhesion molecule and Thy-1 (27,28) and to many genes that are involved in signal transduction (29 -37), the rat pp120 appears to belong to the GC-rich promoter class of TATA-less housekeeping genes. This class usually contains several transcription start sites in close proximity to clustered potential Sp1 binding sites spread over a GC-rich area (24, 38 -41). The region between nt Ϫ165 and Ϫ25 in the proximal pp120 promoter is GC-rich (62%), and it contains four transcription start sites (nt Ϫ101, Ϫ71, Ϫ41, and Ϫ27) as well as four potential binding sites for Sp1 (Table II and Fig. 5) which are clustered in a 72% GC-rich domain (nt Ϫ165 to Ϫ114). Although further studies such as specific binding of purified Sp1 to Sp1 boxes in DNase I footprinting and protection assays are required to determine whether Sp1 binding to boxes 3, 5, and 6 is essential for efficient transcription of the TATA-less pp120 gene, our observations suggest that these boxes are essential for basal functional promoter activity of the gene. At least two of these boxes (boxes 3 and 5) may act independently as was reported for the Sp1 boxes of the TATA-less gene of the rat insulin-like growth factor-binding protein-2 (42). It is interesting that box 3 contains elements that fully match the consensus sequence for Sp1 binding site, and boxes 6 and 5 contain either a single C to A mutation at position 5 (box 6) or in addition to another mutation at position 1 of the Sp1 canonical motif (box 5) ( Table  II). The minimal effect exhibited by these mutations on the pp120 promoter activity is in agreement with the notion that Sp1 binding sites containing the C to A mutation at position 5 are fully functional and bind Sp1 with high affinity in many FIG. 6. Effect of block substitution of specific Sp1 boxes in the ؊249/؊21 promoter fragment on basal promoter activity in H4-II-E cells. A series of 20-bp (5Ј-CTCGAATTCGATGGCGGTAT-3Ј) block mutations was prepared in the Ϫ249/Ϫ21 promoter fragment (Ϫ249pLuc), which overlapped and mutated specific potential Sp1 boxes. Ϫ249Mut3pLuc, Ϫ249Mut5pLuc, and Ϫ249Mut6pLuc constructs containing block mutations spanning Sp1 box 2 in nt Ϫ209 to Ϫ189, Sp1 box 3 in nt Ϫ169 to Ϫ149, and Sp1 boxes 4/5 in nt Ϫ148 to Ϫ128, respectively, were thus obtained. Luciferase activity in lysates derived from cells transfected with these constructs is graphically shown on the right. Luciferase activity was normalized for GH secreted in the media and calculated as mean Ϯ S.D. in relative light units of triplicate transfections. The graph represents typical results of three different experiments.
FIG. 7. Regulation of the pp120 promoter activity in H4-II-E cells by insulin, dexamethasone, cAMP, and phorbol esters. A ϳ1.6-kb fragment (nt Ϫ1609 to Ϫ21) of the 5Ј-flanking region was subcloned into pGL3 and transfected into H4-II-E cells using DEAEdextran as described in the legend to Fig. 4. Transfected cells were incubated overnight in serum-containing medium followed by a 24-h incubation in serum-free medium. Insulin (Ins, 1 mg/ml), dexamethasone (Dex, 1 mM), cAMP (1 mM), and PMA (1 mM) were added and the incubation continued for an additional 24 h. Untreated transfected cells (plain) were included as control. Luciferase activity was measured in the cell lysates, normalized for hGH secretion, and calculated as mean Ϯ S.D. in relative light units of triplicate transfections. The graph represents typical results of four different experiments. natural promoters (43)(44)(45) and that mutation at position 1 is the least important (43). In addition to the C to A mutation at position 5, box 4 contains mutation at position 8. Although this mutation may be responsible for the apparent minimal role of box 4, we cannot exclude in these studies the possibility of a functional role for box 4 in the transcription of the pp120 gene. Individual block mutation of boxes 4 and 5 is required to address this question in more detail. The apparent nonfunctionality of Sp1 boxes 1 and 2 is not unexpected since they are spatially separated from box 3 by 250 and 42 nt, respectively. Additionally, these boxes are located outside the GC-rich area (nt Ϫ25 to Ϫ165) of proximal pp120 promoter.
Sp1-activating proteins acting either alone (46,47) or in conjunction with a transcriptional initiator located downstream from the Sp1 binding site (48,49) attract the transcription factor IID complex to direct proper transcription. pp120 proximal promoter sequence contains a 7-bp fragment (CCAAATC, nt Ϫ45 to Ϫ38) that includes the transcription start site at nt Ϫ41, thus conforming to the loose consensus sequence for a transcriptional initiator (PyPyA*N T/A PyPy) at the start site (*) (49). In the pp120 gene, this putative initiator is part of an AGCCAAAT sequence (nt Ϫ47 to Ϫ39) that matches 6 out of 8 bp in the consensus octamer binding site found in immunoglobulin genes (ATGCAAAT, Ref. 28). Whether this potential initiator-octamer binding site complex participates in the transcriptional activation of the pp120 gene remains to be investigated.
A number of motifs for liver-specific transcription factors are present in the 5Ј-flanking region of the pp120 gene. Some of these binding sites, such as those for HNF-5, are characteristically located in proximity to other liver-specific factors, such as the potential CAAT enhancer-binding protein binding sites in the pp120 gene (nt Ϫ1373 to Ϫ1359). Although many of these factors are neither restricted to nor sufficient to confer hepatocyte specificity, they are commonly classified as liver-specific transcription factors (50). Nonetheless, they have been implicated in polarized epithelium expression (50). Hence, it is of future interest to investigate whether these potential liverspecific factors play a role in pp120 expression on the bile canalicular domain of hepatocytes. Nucleotide sequences matching response elements for cAMP and TPA were found in the 5Ј-flanking region of the pp120 gene. TPA and cAMP may exert their effects by binding either to their respective DNA binding sites (TPA response element and cAMP response element, respectively) or through the activation of AP-1 (TPA, Ref. 51) or AP-2 (TPA and cAMP, Ref. 52). pp120 promoter activity was not affected by PMA alone, whereas it was elevated 4 -5-fold in response to both cAMP and PMA. This suggests a synergistic effect of PMA and cAMP on the promoter activity of pp120. cAMP and PMA may coordinate their concerted effect by binding either to the same site, such as AP-2, or to different sites where the binding of one is required to mediate regulation by the other. Although nucleoside monophosphates are not substrates of the enzymatic activity that is believed to be associated with the protein gene product (53), our observations of the up-regulatory effect of cAMP on pp120 FIG. 8. Comparison of the nucleotide sequence of the 5-flanking region of the rat liver pp120 and the human CEA genes. Approximately 1 kb of the 5Ј-flanking region of each of the rat pp120 (GenBank U27207) and the human CEA (GenBank HUMCEA01) genes was compared using the GeneWorks software program (Intelligenetics). Sequence homologies are shown as boxes.
promoter activity warrant further investigations.
Dexamethasone treatment of rats increased pp120 protein expression 2-3-fold in hepatocytes (1). Hence, the 2-fold increase in pp120 promoter activity by glucocorticoids suggests that the up-regulatory effect of glucocorticoids on the pp120 protein occurs at the transcriptional level. These observations are in marked contrast to the post-transcriptional down-regulatory effect of glucocorticoids on the insulin receptor substrate-1 mRNA levels, the occurrence of which is not accompanied by any effect on the insulin receptor substrate-1 promoter activity in transfected 3T3-F442A adipocytes (37).
The up-regulatory effect of insulin on the pp120 promoter activity suggests that insulin may act at the transcriptional level to regulate pp120 expression, an insulin receptor substrate. In view of the proposed role of pp120 to regulate hepatic insulin clearance from blood (12), modulation of pp120 gene expression may represent an important feedback loop in insulin action in liver. The promoter activity of insulin receptor substrate-1 was not altered by insulin treatment of transfected 3T3-F442A adipocytes (37). Thus, pp120 constitutes a first example of an insulin receptor substrate whose promoter activity is up-regulated by insulin. The cis-regulatory elements required to regulate transcription of the pp120 gene by insulin have not been determined. Insulin may act through nuclear factors binding to the sequence that matches the insulin response element in the PEPCK gene (Table IV) or to a novel site. Insulin down-regulates the transcriptional rate and counteracts the up-regulatory effect of glucocorticoids on PEPCK, the enzyme that catalyzes the rate-limiting step in gluconeogenesis (19). Unlike the PEPCK gene, there is no apparent interaction between insulin and glucocorticoids to regulate transcription of the pp120 gene.
In the present report, we have identified the presence of potential cis-elements that constitute the binding sites of various trans-acting factors on the pp120 promoter. Further studies of the interaction between these putative binding sites with specific trans-acting nuclear factors will provide important insight into the elements required for tissue-specific expression of the rat pp120 gene.