|
Figure 1. Northern blot analysis of discordant patterns of mRNA expression by related Egr genes. |
|
|
|
|
|
|
|
JUN LIU1, MARION NAU2, LIAM GROGAN2, POWELL BROWN2, CARMEN ALLEGRA2, EDWARD CHU1, and JOHN WRIGHT2
1Department of
Medicine and Pharmacology
Yale Cancer Center
Yale University School of Medicine and
VA Connecticut Healthcare System
New Haven, CT 06520; and
2Medicine
Branch
Division of Clinical Science
National Cancer Institute
National Institutes of Health
Bethesda, MD 20889
Address correspondence to:
Jun Liu, M.D.
VA CT Healthcare System, Cancer Center-111D
950 Campbell Avenue
West Haven, CT 06516
Tel: (203) 932-5711 Ext. 4033
Fax: (203) 937-4869
Email: JunLiuj@cs.com
Key words: Egr-3, Promoter, Transcription regulation
|
Abstract |
Egr-3 is an immediate-early primary response gene encoding a zinc finger transcription factor. We cloned the human Egr-3 genomic locus including greater than 1100 bp of the 5' flanking region and analyzed this region for putative cis-acting elements. The GC-rich promoter forms part of a representative CpG island that extends into the genomic locus. The Egr-3 promoter contains a region of TATA homology located 25bp upstream from a major transcriptional start site. One serum response element and two variant Egr consensus sequences were identified. Features that distinguished Egr-3 from other human Egr gene promoters included the presence of at least five E-box motifs and a retinoblastoma response element. In addition, an overlapping tandem repeat of 16 GC-rich nonamers was identified in the flanking region that may represent a novel regulatory region for this primary response gene. Reporter constructs coupled with Egr-3 5' flanking sequences in sense and antisense orientation were tested in transient transfection assays. The functional activity of the Egr-3 regulatory region was position-specific. Deletional analysis in serum stimuIated embryonic lung fibroblasts identified that the major elements responsible for growth-induced Egr-3 expression are located within the first 378 bp upstream of the major transcription start site. Analysis of the human Egr-3 genomic locus revealed a complex regulatory organization with significant differences from other Egr genes. These findings may provide insights into the expression of Egr-3 in normal and neoplastic tissues.
|
Introduction |
The primary response genes encode a spectrum of structural and regulatory proteins including several families of nuclear transcription factors that presumably regulate a select group of target genes involved in tissue- and signal-specific responses (Herschman, 1991). Among them, the Egr (early growth response) gene family is a structurally related group consisting, to date, of four zinc finger transcription factors, Egr-1, Egr-2, Egr-3, and pAT 13 (Sukhatme et al., 1988; Joseph et al., 1988; Muller et al. 1991). The Egr genes encode proteins with three tandem zinc finger motifs of the Cys2- His2 subclass that are highly homologous and mediate sequence-specific DNA binding. The prototypic member of this group, Egr-1, has been most extensively studied in this regard. A GC-rich nonameric consensus sequence (GCGGGGGCG) was initially identified as an Egr-1 binding site (Christy and Nathans, 1989; Cao et al., 1993), and the interaction of murine Egr-1 with the GCGTGGGCG motif was characterized by X-ray crystallography studies (Pavletich and Pabo, 1991). Binding to the nonameric consensus sequence has been demonstrated for all the Egr family members, which complex to this domain with different levels of affinity. Other putative Egr response elements include a homopurine/homopyrimidine domain (TCCTCCTCCTCCTCTCC) (Wang and Deuel, 1992) and variations of the consensus sequence (Swirnoff and Milbrandt, 1995; Nakagama et al., 1995). Some of the Egr-binding sequences are present in the promoter region of various genes involved in cell proliferation, thus linking this family of regulatory proteins to transcriptional control of cellular growth processes.
A potential relationship of some Egr genes and related zinc finger proteins to the development of the malignant phenotype has also been suggested. Egr-1 is located on chromosome 5q31, a region commonly deleted in therapy-related myeloid leukemias (Nucifora et al., 1993). Dys-regulated expression of Egr-1 and Egr-2 by human retrovirus-transformed cells (Wright et al., 1990) and soft-tissue sarcomas has been identified. Putative tumor suppressor activity of Egr-1 has been described based on studies showing inhibition of v-sis transformation in murine fibroblasts co-transfected with an expression vector containing Egr or Egr gene fragments (Huang wt al., 1994). Finally, the WT1 gene, the loss of which results in the development of Wilms tumors, encodes a transcription factor with zinc finger regions that share a high level of sequence homology to the corresponding region of Egr proteins and binds the Egr consensus sequence as well (Nakagama et al., 1995).
The human and murine Egr-3 genes were isolated from a serum-activated cDNA library by low-stringency hybridization with an Egr-1 probe containing the zinc finger (Patwardhan et al., 1991). The putative Egr-3 protein featured three tandem zinc finger motifs that were 90% homologous to the Egr-1 zinc finger domains as well as a significant degree (~35%) of similarity in the N-terminus to the corresponding region of Egr-1 and Egr-2. Analysis of an Egr-3 genomic clone identified a gene structure similar to Egr-1 and Egr-2 with one intron approximately 1.3 kb in size located between two exonic regions beginning at nt 435 of the known Egr-3 cDNA sequence. The Egr-3 genomic locus mapped to human chromosome 8p21-23. Although Egr-3 RNA expression in response to serum- and phytohemagglutinin-induction resembled the activation pattern of Egr-1 and Egr-2, several features of Egr-3 regulation apparently are in contrast to these other Egr family members. Egr-3 expression was not identified in any adult rat tissues despite high levels of Egr-1 expression in rat brain, lung, and heart. In addition, although Egr-1 is highly expressed in rat PCJ 2 cells following nerve growth factor stimulation, Egr-3 is not activated under similar conditions.
In this report, we provide the results of cloning and characterization of the human Egr-3 promoter as the first step in identifying mechanisms responsible for Egr-3 regulation. Our findings suggest significant differences in Egr-3 regulation that may be mediated at the transcriptional level.
|
Materials and Methods |
Isolation and Sequence Analysis of the Egr-3 Promoter. Recombinant phage clones containing the Egr-3 gene were isolated by plaque hybridization from a Mbo I-generated human placental genomic DNA library in the Lambda Fix II vector (Stratagene, La Jolla, Calif.). The phage library was screened with an equal mixture of two Egr-3 probes generated by PCR amplification based on the previously isolated Egr-3 cDNA and were random prime-labeled with [a -32P] dCTP (Amersham 3000Ci/mmol). The first probe was a 242 bp fragment of exon 1 extending from nt 253 to 495 and the second probe was 662 bp in length, spanning nt 511 to 1173 of Egr-3 exon 2. Approximately 106 clones were screened from the library. Filter membranes were hybridized at 420C in Rapid Hyb Buffer (Amersham) and washed four times at 650C for 15 minutes. A 3 kb genomic phage isolate positive on quaternary screening was subcloned into plasmid vector Bluescript SK (-) (Stratagene). DNA restriction fragments from this clone were analyzed by restriction endonuclease mapping and Southern blotting using standard techniques.
Dideoxy sequencing of double-stranded templates with vector- and sequence-specific oligonucleotide primers was performed using Sequenase II (U.S. Biochemical Corp) or "hot tub" sequencing kit (Amersham) as described by the manufacturers. Computer-based sequence analyses were performed with the Xgrail program (Uberbacher and Mural, 1991; Guigo et al., 1992).
Construction of Luciferase Reporter Vectors. Segments of the Egr-3 promoter region were amplified by PCR using sequence-specific oligonucleotide primers designed with a terminal Bgl II enzyme site. The 3' end of all Egr-3 promoter fragments began at nt +33. Egr-3 promoter fragments of 378 (nt -345), 568 (nt -535) and 998 (nt -965) base pairs in length were subcloned in the sense and antisense orientations into the PGL2 Genelight vector (Promega; Madison, WI) which contains the luciferase gene without a promoter. Full sequence analysis of the constructs was performed to verify orientation and fidelity with the genomic template sequence. The reporter plasmids were purified by CsCI density gradient centrifugation.
Transfection and Luciferase Assay. Transient transfections of the Egr-3 promoter constructs were performed at 75%-80% confluence in 6 well plates. The cell lines used in these studies were MRC-9 human fetal lung cells, lung cancer cell line H-23, bladder carcinoma cell line SAOS-2, and immortalized mammary epithelial cell line MCF10A. The cell culture conditions were modified according to the individual cell lines. In general, equal volumes of test plasmid (2-3 µg/well) and 15-20 µl of Lipofectin (Bethesda Research Lab.) diluted in 0.1 ml of RPMI 1640 medium (Bethesda Research Lab) with 0-5% serum were mixed and incubated for 15 minutes at room temperature before addition to cells. The pSV-ß-gal plasmid (1µg/well) was co-transfected to normalize for transfection efficiency. After overnight incubation at 370C, the transfection media was replaced by media with 10% serum and incubated for another 24-48 hr before being harvested for luciferase and ß-Gal activity assays. To test for the effects of serum stimulation, medium with 0.25% serum was added after overnight transfection and incubated for an additional 48 hr. Cells were then serum-stimulated by addition of medium containing 20% FCS. At 1-7 hr after serum addition, cells were harvested for luciferase and ß-galactosidase assays. In brief, following lysis with Tropix's Galacto Lysis Buffer and removal of cell debris by centrifugation, 40 µl of cellular extract was used to determine luciferase and ß-gal activity. Reactions were initiated with injection of 100 µl of 1 mM luciferin (Promega) and Galacton Substrate (Tropix), respectively, and enzyme activity was then determined using a luminometer (Berthold L8955).
|
Results |
Discordant Expression Patterns by Related Egr Genes. A rapid and coordinate up-regulation of primary response genes Egr-1 and Egr-3 was observed following serum stimulation of quiescent MRC-9 fetal lung fibroblast cells, synchronized in G0 by serum deprivation. Egr-3 gene expression was detectable after 30 minutes following serum stimulation and peaked at approximately 1 hr post-serum exposure. In fact, the induction was so transient that Egr-3 mRNA level was undetectable by 3 hr. (Figure 1A, left panel). Activation of Egr-1 was also time-dependent but significantly stronger and of longer duration. The Egr-3 induction following serum induction indicated that a common regulatory motif, specifically one or multiple serum response elements, might be present in the Egr-3 promoter region.
Figure
1. Northern
blot analysis of discordant patterns of mRNA expression by
related Egr genes.
Kinetics
of Egr-1 and Egr-3 expression following serum
stimulation. Left panel: Egr-3 expression and right
panel: Egr-1 expression. MRC-9 cells were
synchronized in G0 by serum deprivation, and total RNAs were
extracted at 0, 0.5, 1, 2, 3, and 4 hr following 20% serum
stimulation. Ten µg of total RNA was electrophoresed
and hybridized with 32P-labeled Egr-3 cDNA
probe.
Comparison
of Egr-1, Egr-2 and Egr-3 expression by
human HTLV-1 transformed. T-lymphocyte cell line 702 and
fetal lung cell line MRC-9. Cells were harvested at 60 min
following serum stimulation of quiescent cells. Ten µg
of total RNA was electrophoresed per lane and hybridized
with the respective 32P-labeled cDNA
probes.
In contrast to the co-regulated induction of Egr-1, Egr-2, and Egr-3 in serum-activated cells, high levels of Egr-1 were expressed in human fetal and adult organs such as lung, mammary gland but fetal brain and liver as well as placenta had very low or undetectable levels. Under similar assay conditions, Egr-3 expression was present in all fetal and adult primary tissues (data not shown). In addition to this discrepancy in tissue patterns of expression, alternative regulation of these genes was evident in T lymphocytes transformed by the human retrovirus HTLV-1. The HTLV-l transformed cell line 702 constitutively expresses high levels of Egr-1 and Egr-2 but Egr-3 expression is undetectable (Figure 1B). Since HTLV-l activation of the Egr-1 and Egr-2 promoter has been attributed to the HTLV-l Tax protein, corresponding promoter motifs regulated by Tax may be inaccessible or not present in the Egr-3 promoter.
Molecular Cloning of the 5' Promoter Region of Egr-3. The Egr-3 genomic locus was subcloned from a human placenta genomic library by screening 1 x 106 recombinant phage clones under high stringency conditions with two PCR-generated probes from the Egr-3 5' coding region. One clone was studied that contained a 12 kb Not I fragment spanning the entire length of the Egr-3 gene including 5' and 3' flanking sequences. Following analysis by restriction endonuclease digestion and Southern blotting, an approximately 3.0 kb Pst I fragment was identified that encompassed about 1.5 kb of the 5' regulatory region of Egr-3. This fragment was subcloned into Bluescript SK (-) plasmid and M13 vectors and over 1.2 kb of the upstream region was sequenced on both strands. Southern blot analysis of human genomic DNA revealed restriction fragments identical to those in the clone, suggesting that gene rearrangement did not occur during subcloning (Figure 2).
Figure
2. Isolation of Egr-3 genomic locus
Top: Schematic representation of Egr-3 gene with 5'
and 3' flanking regions including the 3 kb Pst I
fragment subcloned from a human placenta library.
Restriction endonuclease sites identified are Pst I
(Ps), Ava II (A), BstX I (Bx), EcoR I
(R), BamH I (B), Sph I (S), Pvu II (V),
and Bsm I (M) Filled boxes represent Egr-3
intron (small diameter) and exonic (large diameter) regions
described by Patwardhan et al. Open boxes represent putative
5' cDNA terminus (small diameter) and 5' flanking regulatory
region (large diameter). A cross-hatched region upstream of
the promoter was not fully sequenced. Bottom: Southern blot
analysis of genomic DNA from human breast cancer cell line
MCF-7 (lanes 1, 3, 5) and a human placenta genomic clone
(lanes 2, 4, 6).
Structural Features of the 5' Flanking Region. The sequence of 1368 bp of a previously unreported genomic segment flanking the Egr-3 gene is shown in Figure 3.
Figure
3. Sequence of Egr-3 5' flanking
region
Nucleotides are numbered from the previously published 5'
end (nt-1) (Patwardhan et al., 1991). CAAT and TATA sites
are boxed. CArG box is underlined. Filled arrowhead
indicates the major transcription start site. For a detailed
listing of the putative regulatory motifs, please refer to
Table 1. The sequence was assigned the genebank accession
number AY026865.
The nucleotide composition of the region was GC-rich (61.3% GC in composition) and contained expected numbers of CpG and TpG dinucleotides. A 200 bp segment at the downstream end was exceptionally GC rich (75% GC). Analysis of the region with the Xgrail computer program identified this segment of the promoter as part of a CpG island that extended into the coding region of the Egr-3 gene (Figure 4).
Figure
4. Sequence of Egr-3 5' flanking
region
The lower bar represents overlap Egr-3
promoter/S40832 1371:2210 and the upper bar represents
S40832 Coding region 1729:2892.
The published Egr-3 cDNA sequence includes 282 bp of 5'-untranslated region (Patwardhan et al., 1991). Sequence analysis of the 5'-flanking region identified a canonical TATA box approximately 200 bp upstream from the published 5' end. We performed anchored PCR cloning of cDNAs from the human breast carcinoma cell line MCF-7 as part of a more comprehensive approach to identify the Egr-3 transcription start site. More than 20% of these 5' RACE' clones began 25 bp downstream from the TATA box indicating that it serves as a major site of TATA directed transcription initiation (data not shown). The potential regulatory elements present in the promoter region are listed in Table 1.
Table
1
Potential regulatory elements present in the promoter region
of Egr-3.
The conserved palindromic CRE (cyclic AMP response element) sequence has been identified in all other Egr promoter regions consistent with the demonstrated regulation of these genes by the cyclic AMP signal transduction pathway (Irving et al., 1989). Recently, it was shown that the CRE of the Egr-1 promoter is essential for the transcriptional response to the cytokines GM-CSF and IL3 (Sakamoto et al., 1994). The CRE site may be a focal regulatory point for integration of additional signals as well. The CRE of the Egr-1 promoter is involved in the regulation of this gene by TAX-1 and TAX-2, the transactivating proteins of human retroviruses HTLV-1 and HTLV-2 respectively .
The similar kinetics of Egr-1, Egr-2, and Egr-3 expression in response to serum activation suggested that either one or multiple serum response elements (SRE) or CArG boxes, the core of an SRE, were likely to be present in the Egr-3 promoter since these motifs mediate serum induction of the Egr-1 and Egr-2 genes (Christy and Nathans, 1989; Rangnekar et al., 1990). In support of this hypothesis, one serum response element was identified that was present in the Egr-3 proximal promoter region located within 35 bp 5' of the TATA box. The core consensus sequence (GGAAG) of an ETS motif, the recognition site for the ETS family of transcription factors, is frequently located in the 5' flanking arm of SRE in the promoter region of primary response genes such as c-fos, Egr-1 and Egr-2 (Treisman, 1990). Formation of a ternary binding complex at the SRE by serum response factor (SRF) and ETS proteins is essential for transcriptional activation through the SRE by serum growth factors. An ETS-binding site was not identified adjacent to the Egr-3 SRE, suggesting an alternative mechanism of SRF-mediated induction at this response element. The Egr-3 promoter contained two ETS motifs and a third ETS-binding site was present approximately 50 bp 3' of the TATA box. None of these segments, however, matched the CGGAAG sequence characteristic of the subset of high affinity ETS sites involved in SRE-ETS protein interactions.
Five F box elements (CANNTG) were scattered throughout the Egr-3 promoter region. These motifs are potential binding sites for helix-loop-helix transcription factors with leucine zippers (bHLH-LZ) that regulate growth and differentiation, such as c-myc and myoD, the later being considered a "master" regulator of myogenic differentiation (Desprez et al., 1995; Kim et al., 1995). The E box sequence CACGTG, located at nt-844 is a high-affinity binding site for myc, max, and several other bHLH-LZ transcription factors (Miltenberger et al., 1995).
In addition to these cis-acting sequences, there are two other putative binding sites in the Egr-3 promoter region for various cellular proteins associated with malignant transformation. One such protein includes the retinoblastoma (Rb) tumor suppressor protein that recognizes the CCACCC sequence located in the flanking region. Of interest, this sequence is also a frequent erythroid gene promoter motif (Pan and McEver, 1993). Moreover, two variant IGF-2 motifs, which serves as response elements for several interferon-induced regulatory factors includingIRF-1 (Tanaka et al., 1993) were also present.
Egr-binding sites, including the nonameric consensus sequence (GCGG/TGGGCG), the TC-rich tract (TCCTCCTCCTCCTCTCC) or the iterative PCR-selected site (TGCGTG/AGGCGGT) were not present in the Egr-3 promoter region. However, two identical one nucleotide variants of the Egr nonameric consensus sequence (GCAGGGGCG) were located in this regulatory region at nt-486 and nt-666. These variant motifs are altered at the same nucleotide as an Egr-binding segment in the human neurofilament light promoter and may therefore bind Egr (Pospelov et al., 1994). A SP-1 site overlaps the variant Egr-binding segment in the distal Egr-3 promoter. This same structural relationship has been noted in the promoter region of several other genes (Patwardhan et al., 1991; Ackerman et al., 1991; Harrington et al., 1993; Shingu and Bornstein, 1994) and is believed to have important consequences for transcriptional regulation by these Egr and SP-1 binding sites.
An unusual structural feature of the Egr-3 flanking region, not present in any of the other Egr genomic loci, was an overlapping tandem array of GC-rich nonameric segments located 106 to 159 bp downstream of the TATA box. In a region that spans 45 bp, a total of 16 overlapping nonameric segments can be defined that include eight identical nonamers and eleven nonameric units that vary from the Egr consensus sequence at an identical single nucleotide (Figure 5). The remaining five nonamers in the region vary from the canonical Egr binding site by two or three base pairs.
Figure
5. Unusual overlapping tandem array of GC-rich nonameric
segments in the 5' untranslated region of Egr-3
downstream of the TATA box
The nucleotides that are different from those of the
consensus Egr binding site are
underlined.
Another repetitive DNA segment was noted in the Egr-3 genomic clone isolated during this study. A tetranucleotide tandem repeat consisting of eleven ATGG sequences was present in the 5' region of the intron. In addition to representing a microsatellite repeat sequence that potentially may be useful as a marker of genetic heterogeneity, the tetranucleotide repeat unit is a binding site for transcription factor YY1 (Shi et al., 1991) which regulates several immediate early genes such as c-fos and c-myc. Therefore, this intronic repeat sequence may identify a region involved in transcriptional regulation of Egr-3 expression.
Transient Expression Analysis of the 5' Flanking Region of Egr-3. We wished to determine whether Egr-3 genomic flanking sequences could act as a functional promoter and whether this region was serum-inducible. To test for this activity in transient cell transfection assays, heterologous constructs were generated using a promoterless plasmid with a luciferase reporter gene by inserting Egr-3 flanking segments beginning at nt+33bp downstream of TATA in the sense and antisense orientation. Luciferase expression by reporter constructs with 998 bp of Egr-3 promoter sequence in either orientation, was measured following transfection of three cell lines, lung cancer cell line H-23, bladder carcinoma cell line SAOS-2, and an immortalized but non-malignant mammary epithelial cell line, MCF10A (Figure 6A). The Egr-3 promoter sense constructs exhibited varying levels of total luciferase activity that differed by approximately 10- to 100-fold with the highest levels of luciferase expression in the H-23 cell line. These variances were independent of transfection efficiency indicating significant differences in basal Egr-3 activation in the individual cell lines. In all cases, however, activation of the sense construct was substantially higher (15- to 100-fold) than following transfection with antisense constructs or background levels with the pGL-2 Basic vector alone. The orientation-dependent increase in luciferase gene activation indicated that sequences supporting promoter activity were present in these regions of Egr-3 flanking sequence.
Figure
6. Transient expression analysis of Egr-3 5' flanking
region. Three
constructs were made by subcloning into the pLG-2 reporter
the promoter region of Egr-3 from the same position
from 3' end (nt+33) to &emdash;345, -545, and &emdash;965,
respectively.
Promoter
activity of the Egr-3 5' flanking region in different
cell lines. Hatched bars represent the 998 bp 5' flanking
fragment of Egr-3 in the report vector in sense
orientation, while the dotted bars represent the antisense
construct and clear bars show the basic activity of control
pGL-2 vector. The luciferase measurements were scaled up by
100-fold for Saos-2 and 10-fold for MCF10A over their actual
readings. Error bars represent standard deviations
calculated from a duplicate set of experiments.
Identification
of the serum responsive region of the Egr-3 promoter.
Clear bars show the 0 time point value while black bars show
the activity that tested at 7 hr point after serum
stimulation.
Serum responsiveness of the Egr-3 promoter was tested next using reporter constructs with serially deleted Egr-3 promoter fragments spanning 378, 568, and 998 bp. The MRC-9 cell line was used since our previous studies showed Egr-3 induction in these cells by serum (Figure 1A). Luciferase expression by serum-deprived MRC-9 cells (0 time) transfected with each of the three promoter fragments was uniformly low and equivalent to background expression levels in pGL-2 Basic transfectants. Following serum activation, a 50- to 100-fold increase in promoter activity above 0 time was measured in multiple assays confirming that the Egr-3 promoter was indeed serum-inducible. However, there was not a significant difference in normalized luciferase expression levels between the three constructs with serially truncated fragments, i.e., the extent of activation was not correlated to the length of the promoter region. Luciferase measurement in serum-activated cells transfected with the 378 was equivalent to the 568 and the 998 bp construct (Figure 6B). These results, reproducible in multiple assays, also were independent of transfection efficiency and thus suggested that the elements that are responsible for the induced Egr-3 expression is located within the first 378 bp region of the Egr-3 promoter.
|
Discussion |
Egr-3 is an immediate early gene product with transcriptional regulatory activity that is expressed during extracellular ligand-activated transition from G0 to G1 stages of the cell cycle, and presumably plays a critical intermediary role in regulating the cascade of genetic events leading ultimately to cell division. Activation of Egr-3 and other Egr genes occurs in response to a number of common mitogenic stimuli. But like many structurally related families of transcription factors, there is evidence that individual family members are differentially expressed in similar biological processes, suggesting alternative regulatory influences that may be controlled at the level of transcription (Morgenbesser et al., 1995). Thus, activation of Egr-3 by nerve growth factor or basal Egr-3 expression in rat primary tissues is undetectable despite induction of Egr-1 under similar conditions (Patwardhan et al., 1991). We also have identified differential expression of Egr-1 and Egr-3 at the RNA level in several human fetal primary tissues. In addition, we have shown that Egr-3 is not constitutively expressed by a human HTLV-1 transformed cell line in contrast to Egr-1 and Egr-2 that are upregulated in these cells.
The cloning and transfection analysis of Egr-3 upstream flanking sequences has identified a number of response motifs that may regulate Egr-3 transcription and govern the differential co-expression of Egr-3 and other Egr family members. TATA and CCAAT motifs that normally promote accurate initiation of transcription were evident in the Egr-3 promoter. Preliminary studies indicate that a major transcriptional initiation site exists approximately 25 bp downstream of the TATA box. The Egr-3 promoter contains several regulatory motifs in common with previously characterized flanking segments of Egr-1, Egr-2 and NGFI-C, the rat homologue of the fourth Egr family member (Crosby et al., 1991) including binding sites for cyclic AMP responsive element binding protein (CREB) and transcription factors Sp-1 and NF-k B. One serum response element (SRE) was identified in the Egr-3 promoter. In contrast, five SRE are located in the Egr-1 promoter (Sakamoto et al., 1991) while 2 CArG elements, the core sequence of an SRE, are identified in the Egr-2 upstream regulatory region (Rangnekar et al., 1990). This may partially explain why the Egr-3 induction level was lower than that of Egr-1. The promoter region of NGFI-C contains no SRE or CArG elements, and activation of this gene by serum is markedly reduced compared to other Egr genes. An ETS motif was not located adjacent to the Egr-3 SRE as described for SRE in c-fos, Egr-1 and Egr-2. Egr-3 thus falls into a class of primary response genes, including cyrGi, SRF, and Xenopus laevis type 5 actin (Williams and Lau, 1993) with promoters containing CArG boxes without associated ETS motifs. Regulation of these genes may involve recruiting a more distal ETS motif to the SRE or activation through an entirely different pathway.
The presence of multiple E-box motifs, the binding site for bHLH-LZ proteins may signal a role for Egr-3 in myogenesis. A total of five E-box core elements were present in the Egr-3 promoter, while not one E-box is identified in greater than 690bp of Egr-1 promoter and only one E-box is present in the Egr-2 promoter. Tissue-specific class B bHLH15-LZ proteins have important regulatory functions in several differentiation pathways including skeletal myogenesis, hematopoiesis, and neurogenesis. The E-box at nt-844 of the Egr-3 promoter is a common E-box sequence (CACGTG) that in various systems is a high-affinity binding site for bHLH-LZ transcription factors including Myc, max, USF, TFE3, CBF-1 and TFEB. The prominent representation of E-box binding motifs in the Egr-3 promoter may indicate that the gene is a target for lineage-specific or stage-restricted regulation by bHLH-LZ transcription factors during differentiation
The sequence CCACCC may serve as a retinoblastoma control element (RCE) for mediating p105Rb (Rb) protein regulation of gene expression (Kim et al., 1991). An RCE is located at nt-367 of the Egr-3 promoter, the only known human Egr gene promoter with this response element. The transcriptional regulatory potential of Rb extends to a number of other primary response genes such as c-fos and c-jun. Rb-mediated regulation may, in some instances, occur through cell-lineage specific interactions with other transcription factors that center around the RCE or other response elements (e.g. Sp-1, ATF2) and may include positive or repressive effect (Bremner et al., 1995). The CCACC sequence has also been identified in many erythroid-specific gene promoters including ß-globin and the porphobilinogen carboxylase genes and has been implicated in a number of cell-specific signaling pathways presumably by serving as a binding site for several regulatory proteins (Frampton et al., 1990).
Egr-3 is not constitutively activated in the HTLV-1 transformed cell line 702 in contrast to Egr-1 and Egr-2. Expression of these and other primary response genes by HTLV-transformed cells results from transactivation by the HTLV regulatory gene TAX. Although TAX does not bind directly to DNA, TAX-responsive regulatory domains have been identified in the HTLV long terminal repeat (LTR) and in the promoter regions of TAX target genes suggesting that TAX-mediated stimulation requires other specific cellular factors. The TAX gene product of HTLV-l and HTLV-ll transformed cells may activate Egr-1 expression through different response elements based on work by Sakamoto who identified the Egr canonical response element and the cyclic AMP response element.
Transfection of the approximately 1 kb of 5'-flanking region of Egr-3 linked to a luciferase reporter vector indicated that the essential elements for Egr-3 promoter activity is present in the region between nt-965 and nt+33, which support high-level gene expression in H-23 but only basal expression in MCF-10A cells. These significant differences in basal Egr-3 activation in individual cells suggest that the 998 base pair flanking sequence contains sufficient information for promoter-specific expression in cultured cell lines. Moreover, the fact that activation of the sense construct was substantially higher than that observed with transfection of antisense construct indicated that sequences that support promoter activity are also orientation-dependent. Serum stimulation experiments using serially truncated segments of Egr-3 promoter revealed that the positive regulatory region responsible for serum stimulated activation of Egr-3 expression exists within the first 378 bp of the 5' flanking sequence. In this region, there is a SRE element that has been shown to mediate the expression of mitogen inducible genes such as c-fos (Treisman, 1987), EGR-1 (Christy et al., 1988) and Egr-2 (Rangnekar et al., 1990). Mutagenesis of this CArG box in Egr-3 promoter is needed to determine whether this motif is solely essential for serum reponsiveness.
In conclusion, cloning and analysis of the human Egr-3 5'-untranslated region revealed a unique complex regulatory organization. The biological activity of the Egr-3 promoter has been demonstrated by means of differential transcription activation and serum-induced transactivation. Considering the large number of putative regulatory elements present in this region, it is conceivable that Egr-3 may play a significant role in transcriptional control of cellular proliferation, differentiation, or other growth processes.
|
Acknowledgements |
This work was supported in part by grants from the National Cancer Institute (CA 16359 and CA 75712 to E.C.) and the Department of Veterans Affairs (VA Merit Review to E.C.)
|
References |
Ackerman, S. L., Minden, A. G., Williams, G. T., Bobonis, C., and Yeung, C. Y. (1991) Functional siginificance of an overlapping consensus binding motif for Sp1 and Zif268 in the murine adenosine deaminase gene promoter. Proc. Natl. Acad. Sci. USA. 88:7523-7.
Bremner, R., Cohen, B. l., Sopta, M., Hamel, P. A., Ingeles, C. J., Gallie, B. L., and Phillips, R. A. (1995) Direct transcriptional repression by pRB and its reversal by specific cyclins. Mol. Cell. Biol. 15:3256-3265.
Cao, X., Mahendran, R., Guy, G. R., and Tan, Y. H. (1993) Detection and characterization of cellular EGR-1 binding to its recognition site. J. Biol. Chem. 268:16949-16957.
Christy, B. A., Lau, L. F., and Nathans, D. (1988) A gene activated in mouse 3T3 cells by serum growth factors encodes a protein with "zinc finger" sequences. Proc. Natl. Acad. Sci. USA. 85:7857-7861.
Christy, B., and Nathans, D. (1989) DNA binding site of the growth factor-inducible protein Zif268. Proc. Natl. Acad. Sci. USA. 86:8737-8741.
Crosby, S. D., Puets, J. J., Simburger, K. S., Fahrner, T. J., and Milbrandt, J. (1991) The early response gene NGFI-C encodes a zinc finger transcriptional actiator and is a member of the GCGGGGGCG (GSG) element-binding protein family. Mol. Cell. Biol. 11: 3835-3841.
Desprez, P. Y., Hara, E., Bissell, M. J., and Campisi, J. (1995) Supression of mammary epithelial cell differentiation by the helix-loop-helix protein Id-1. Mol. Cell. Biol. 15:3398-3404.
Frampton, Walker, J., M., Plumb, M., and Harrison, P. R. (1990) Synergy between the NF-E1 erythroid-specific transcription factor and the CACCC factor in the erythroid-specific promoter of the human porphobilinogen deaminase gene. Mol. Cell. Biol. 10:3838-3842.
Guigo, R., Knudsen, S., Drake, N., and Smith, T. (1992) Prediction of gene structure. J. Mol. Biol. 226: 141-157.
Harrington, M. A., Konicek, B., Song, A., Xia, X. L., Fredericks, W. J., and Rauscher, F. J. I. (1993) Inhibition of colony-stimulating factor-1 promoter activity by the product of the Wilm's tumor locus. J. Biol. Chem. 268:21271-21275.
Huang, R. P., Darland, T., Okamura, D., Mercola, D., and Adamson, E. D. (1994) Supression of v-sis-dependent transformation by the transcription factor, Egr-1. Oncogene. 9:1367-1377.
Irving, S. G., June, C. H., Zipfel, P. F., Siebenlist, U., and Kelly, K. (1989) Mitogen-induced genes are subject to multiple pathways of regulation in the initial stages of T-cell activation. Mol. Cell. Biol. 9:1034-1040.
Joseph, L. J., Le Beau, M. M., Jamieson, G. A. J., Acharya, S., Shows, T. B., Rowley, J. D., and Sukhatme, V. P. (1988) Molecular cloning, sequencing, and mapping of EGR2, a human early growth response gene encoding a protein with "zinc-binding finger" structure. Proc. Natl. Acad. Sci. USA. 85:7164-7168.
Kim, S. J., Lee, H. D., Robbins, P. D., Busam, K., Sporn, M. B., and Roberts, A. B. (1991) Regulation of transforming growth factor beta 1 gene expression by the product of the retinoblastoma-susceptibility gene. Proc. Natl. Acad. Sci. USA. 88:3052-3056.
Kim, Y. I., Christman, J. K., Fleet, J. C., Cravo, M. L., Salomon, R. N., Smith, D., Ordovas, J., Selhub, J., and Mason, J. B. (1995) Moderate folate deficiency does not cause global hypomethylation of hepatic and colonic DNA in rats. Am. J. Clin. Nutr. 61:1083-1090.
Miltenberger, R. J., Farnham, P. J., Smith, D. E., Stommel, J. M., and Cornwell, M. M. (1995) v-Raf activates transcription of growth-reponsive promoters via GC-rich sequences that bind the transcription factor Sp1. Cell Growth Differ. 6:549-556.
Morgenbesser, S. D., Schreiber-Agus, N., Bidder, M., Mahon, K. A., Overbeek, P. A., Horner, J., and DePinho, R. A. (1995) Contrasting roles for c-myc and L-myc in the regulation of cellular growth and differentiation in vivo. EMBO. J. 14:743-756.
Muller, H. J., Skerka, C., Bialonski, A., and Zipfel, P. F. (1991) Clone pAT 133 identifies a gene that encodes another human member of a class of growth factor-induced genes with almost identical zinc-finger domains. Proc. Natl. Acad. Sci. USA. 88:10079-10083.
Nakagama, H., Heinrich, G., Pelletier, J., and Housman, D. E. (1995) Sequence and structural requirements for high-affinity DNA binding by the WT1 gene product. Mol. Cell. Biol. 15:1489-1498.
Nucifora, G., Birn, D. J., Espinosa, R. D., Erickson, P., LeBeau, M. M., Roulston, D., McKeithan, T. W., Drabkin, H., and Rowley, J. D. (1993) Involvement of the AML1 gene in the t(3;21) in therapy-related leukemia and in chronic myeloid leukemia in blast crisis. Blood. 81:2728-2734.
Pan, J., and McEver, R. P. (1993) Characterization of the promoter for the human P-selectin gene. J. Biol. Chem. 268:22600-22608.
Patwardhan, S., Gashler, A., Siegel, M. G., Chang, L. C., Joseph, L. J., Shows, T. B., Le Beau, M. M., and Sukhatme, V. P. (1991) EGR3, a noval member of the Egr family of genes encoding immediate-early transcription factors. Oncogene. 6:917-928.
Pavletich, N. P., and Pabo, C. O. (1991) Zinc finger-DNA recognition: crystal structure of a Zif268-DNA complex at 2.1A. Science. 252:809-817.
Pospelov, V. A., Pospelova, T. V., and Julien, J. P. (1994) AP-1 and Krox-24 transcription factors activate the neurofilament light gene promoter in P19 embryonal carcinoma cells. Cell Growth Differ. 5:187-196.
Rangnekar, V. M., Aplin, A. C., and Sukhatme, V. P. (1990) The serum and TPA responsive promoter and intron-exon structure of EGR2, human early growth response gene encoding a zinc finger protein. Nucleic Acids Res. 18:2749-2757.
Sakamoto, K. M., Nimer, S. D., Rosenblatt, J. D., and Gasson, J. C. (1992) HTLV-1 and HTLV-II tax trans-activate the human EGR-1 promoter through different cis-acting sequences. Oncogene. 7:2125-2130.
Sakamoto, K. Bardeleben, M., C., Yates, K. E., Raines, M. A., Golde, D. W., and Gasson, J. C. (1991) 5' upstream sequence and genomic stucture of the human primary response gene, EGR-1/TIS8. Oncogene. 6:867-871.
Sakamoto, K. M., Fraser, J. K., Lee, H. J., Lehman, E., and Gasson, J. C. (1994) Granulocyte-macrophage colony-stimulating factor and interlukin-3 signaling pathways converge on the CREB-binding site in the human egr-1 promoter. Mol. Cell. Biol. 14:5975-5985.
Shi, Y., Seto, E., Chang, L. S., and Shenk, T. (1991) Transcriptional repression by YY1, a human GLI-Kruppel-related protein, and relief of repression by adenovirus E1A protein. Cell. 67:377-388.
Shingu, T., and Bornstein, P. (1994) Overlapping Egr-1 and Sp1 sites function in the regulation of transcription of the mouse thrombospondin 1 gene. J. Biol. Chem. 269:32551-32557.
Swirnoff, A. H., and Milbrandt, J. (1995) DNA-binding specificity of NGFI-A and related zinc finger transcription factors. Mol. Cell. Biol 15:2275-2287.
Sukhatme, V. P., Cao, X. M. L., Chang, C., Tsai-Morris, C. H., Stamenkovich, D., Ferreira, P. C., Cohen, D. R., Edwards, S. A., Shows, T. B., Curran, T., and e. al. (1988) A zinc finger-encoding gene coregulated with c-fos during growth and differentiation, and after cellular depolarization. Cell. 53:37-43.
Tanaka, N., Kawakami, T., and Taniguchi, T. (1993) Recognition DNA sequence of interferon regulatory factor 1 (IRF-1) and IRF-2, regulators of cell growth and the interferon system. Mol. Cell. Biol 13:4531-4538.
Treisman, R. (1987) Indentification and purification of a polypeptide that binds to the c-fos serum response element. EMBO. J 6:2711-2717.
Treisman, R. (1990) The SRE: a growth factor resposive transcriptional regulator. Semin. Cancer Biol 1:47-58.
Uberbacher, E. C., and Mural, R. J. (1991) Locating protein-coding regions in human DNA sequences by a multiple sensor-neural network approach. Proc. Natl. Acad. Sci. USA. 88: 11261-11265.
Wang, Z. Y., and Deuel,T. F. (1992) An S1 nuclease-sensitive homopurine/homopyrimidine domain in the PDGF A -chain promoter contains a novel binding site for the growth factor-induciable protein EGR-1. Biochem. Biophys. Res. Commun 188:433-439.
Williams, G. T., and Lau, L. F. (1993) Activation of the inducible orphan receptor gene nur77 by serum growth factors: dissociation of immediate-early and delayed-early response. Mol. Cell. Biol 13:6124-6136.
Wright, J. J., Gunter, K. C., Mitsuya, H., Irving, S. G., Kelly, K., and Siebenlist, U. (1990) Expression of a zinc finger gene in HTL-I- and HTL-II-tranformed cells. Science. 248: 588-591. Herschman, H. R. (1991) Primary response genes induced by growth factors and tumor promoter. Annu. Rev. Biochem. 60:281-319.
