| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
Department of Biochemistry, University of Texas Health Science Center, San Antonio, Texas 78284-7760
Address all correspondence and requests for reprints to: Dr. Martin L. Adamo, Department of Biochemistry, University of Texas Health Science Center, 7703 Floyd Curl Drive, San Antonio, Texas 78284-7760. E-mail: adamo{at}bioc02.uthscsa.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Our data suggest that the sequence surrounding and including start site 3 in exon 1 functions as a minimal independent promoter. The minimal exon 2 promoter is contained within the 73 bp upstream and 44 bp downstream of the transcription start site cluster. These minimal promoters contain similar and distinct elements that are important for basal transcription. Upstream sequences may contain cell type-specific silencer elements.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The tissue-specific distribution of exon 1 and exon 2 transcripts suggests that the two promoters are subject to different regulation. Moreover, exon 1 and exon 2 transcripts appear at different times during development (11, 12, 13, 14, 15). Exon 1 and exon 2 transcripts may be differentially regulated by GH status in a tissue-specific manner (10). However, there are examples of similar fold changes in the levels of exon 1 and exon 2 transcripts in response to GH, fasting and refeeding, and diabetes (5, 13), suggesting the probability of similar regulation of the exon 1 and exon 2 promoters as well.
Nuclear run-on assays and assays of nuclear transcript levels indicate that in many cases, changes in the levels of exon 1 and exon 2 transcripts are indeed due to changes in transcription (reviewed in Ref.16). In transient transfection assays in two cell lines, exon 1 and exon 2 promoter activities were positively correlated with endogenous expression of exon 1 and exon 2 transcripts (6). Increased transcription of the liver IGF-I gene during postnatal development and in response to GH is associated with alterations in chromatin structure in different regions of the IGF-I gene (15, 17, 18, 19). Specific binding of nuclear proteins has been observed in the IGF-I promoter regions (18, 20, 21, 22, 23). Regions or elements that are responsive to ectopically expressed transcription factors and humoral regulators of IGF-I gene transcription have been identified by transfection and DNA-protein binding assays (20, 23, 24, 25, 26, 27, 28, 29, 30). The mechanisms by which exon 1 and exon 2 promoter activities are regulated have not been firmly established, however. Such studies are complicated by the fact that the exon 1 promoter, in particular, presents some unusual structural features.
Exon 1 transcription initiation occurs at multiple sites dispersed over 350 nucleotides of the genomic sequence (1, 3, 5, 9, 13, 14). Studies of endogenous mRNA levels suggest that transcription start sites 2 and 3, located at approximately 345 and 245 nucleotides (nt), respectively, upstream of the 3'-end of exon 1, are major sites of transcription initiation (1, 5, 13, 14). In some tissues, such as liver, there is an approximately equivalent distribution of exon 1 transcripts resulting from use of these sites, whereas in many other tissues, there is a preponderance of the start site 3 transcript (5, 14). Regions upstream and downstream of the exon 1 transcription start sites contribute to promoter activity (5, 15, 17, 18, 19, 22, 23, 24, 25, 26, 27, 28, 29, 30). However, the precise location and nature of the core promoter have not been characterized. There are no TATA or CCAAT homologies, and the region is not GC rich. Although an initiator-like sequence is near transcription start site 2 (2, 4, 5), it has been suggested that this initiator is not functional, as multiple dispersed transcription start sites are used.
Exon 2 transcription occurs from a cluster of sites located from
52–68 nt upstream of the 3'-end of exon 2 (1, 3, 5, 9, 13, 14). The
region upstream and surrounding these start sites also has functional
promoter activity, which contributes to basal and cell type-specific
exon 2 transcription (6, 8). The more focal exon 2 transcription
initiation may reflect the presence of the putative promoter sequence
motifs TTAA at 27 nt and CCAAAT at 80 nt upstream of the first start
site (1, 5), although this hypothesis has not been tested. To
understand how the unusual core IGF-I promoters interact with
regulatory regions, we have at present characterized transcriptional
activity and DNA-protein binding in the proximal exon 1 and exon 2
promoters.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
], and 362 bp of exon 1. The BglII site was
filled in with Klenow enzyme, and the fragment was ligated into the
SmaI site of pGEM 4Z. The insert was removed from pGEM by
using SmaI to cut the 5'-end and BamHI (which
occurs in the 4Z polylinker immediately adjacent to the SmaI
site) to cut the 3'-end and was ligated into the SmaI and
BglII sites of the promoterless pGL2-Basic luciferase
expression vector (Promega, Madison, WI). This resulted in the
-1122/+362 exon 1 promoter-luciferase construct. A construct, in which
861 bp of sequence from -783 to +75 was removed, was prepared
by digestion of the -1122/+362 construct with PvuII, gel
purification of the vector away from the insert, and religation of the
vector. This resulted in the -1122/-783-+75/+362 promoter construct.
This construct lacks transcription start sites 1 and 2. The constructs
-783/+362 and +75/+362 were prepared by cutting the -1122/+362
construct with SmaI, followed by partial digestion with
PvuII. The respective partial digests were gel purified and
religated. The -783/+362 construct contains all four transcription
start sites mapped in rat liver and 783 bp of 5'-flanking sequence. The
+75/+362 construct contains transcription start sites 3 and 4, but
lacks start sites 1 and 2. A -133/+362 promoter construct was prepared
by digesting an IGF-I subclone in pGEM 3 with SmaI and
BglII. The SmaI site was in the polylinker
immediately adjacent to the BamHI site into which the
Sau3A site at -133 had originally been ligated. The
BglII site is, as in the constructs described above, at the
-362 position. This fragment was ligated into the SmaI and
BglII sites of pGL2-Basic and contains all exon 1
transcription start sites. To produce a promoter fragment that
contained 133 bp of 5'-flanking sequence and start sites 1 and 2, but
lacked start sites 3 and 4, the -133/+362 plasmid was digested with
PvuII and BglII. The BglII site was
filled in with Klenow enzyme, and the vector was gel purified away from
the insert and religated to produce the -133/+74 IGF-I
promoter-luciferase construct. Finally, a fragment containing 133 bp of
5'-flanking sequence and 192 bp of exon 1 was prepared by ligating a
Sau3A-FspI fragment into the BglII and
filled in HindIII sites of pGL2-Basic. This fragment lacks
transcription start site 4, as mapped in rat liver (32).
|
], and the remainder is 5'-flanking sequence. The
SmaI-SmaI fragment was ligated into the
SmaI site of pGL2-Basic to produce the -1500/+44 exon 2
promoter-luciferase construct. A PstI-SmaI
fragment extending from -485 to +44 was removed from pGEM 3Z by
digesting with HindIII (which is 11 bp 5' to the
PstI site) and KpnI (which is immediately
adjacent to the SmaI site at the 3'-end). The
HindIII site was filled in with Klenow enzyme, and the
fragment was ligated into the SmaI and KpnI sites
of pGL2-Basic. A HincII-SmaI fragment (-362 to
+44) was removed from pGEM 3Z with HincII and
KpnI (which is immediately adjacent to the Sma I
site at the 3'-end) and was ligated into the SmaI and
KpnI sites of pGL2-Basic.
|
Control plasmids
The promoterless pGL2-Basic vector (Promega) was used for
preparation of IGF-I promoter-luciferase constructs and as a negative
control in transfection assays. The positive control was the
pGL2-Control (Promega), which uses the simian virus 40 promoter and
enhancer to drive transcription of the luciferase structural gene. In
some experiments, the pCMV-ß-gal (obtained from Clontech, Palo Alto,
CA) plasmid was cotransfected to assess transfection efficiency. This
plasmid uses the cytomegalovirus (CMV) promoter to drive transcription
of the ß-galactosidase (ß-gal) structural gene.
Plasmid preparation
Plasmids were transformed into competent Escherichia
coli cells (HB-101 or DH5-
) for amplification, and plasmids
were purified using either the Qiagen column system (Chatsworth, CA) or
the column system from 5 Prime-3 Prime (Boulder, CO). Plasmid DNA was
quantified on the basis of absorbance measurement at 260 nm (A260).
Purity was verified by observing the ethidium bromide-stained,
linearized plasmid DNA on agarose gels.
Cell culture and transient transfection assays
Rat C6 glial tumor cells, rat GH3
pituitary tumor cells, and Chinese hamster ovary (CHO) cells were
obtained from American Type Culture Collection (Rockville, MD). Human
OVCAR-3 ovarian adenocarcinoma cells were kindly provided by Dr.
Douglas Yee, University of Texas Health Science Center at San Antonio,
Division of Medical Oncology, or from American Type Culture Collection.
C6 and GH3 cells were grown in DMEM (4.5 g/liter
glucose) containing 10% newborn calf serum or 10% FBS. OVCAR-3 cells
were grown in RPMI 1640 medium containing 10% FBS. CHO cells were
grown in Ham’s F-12 medium containing 10% FBS. All cells were
maintained in a humidified 5% CO2-95% air environment at
37 C. Before transfection, about 6–8 x 105 cells
were plated onto 60-mm tissue culture plates (Falcon 3004, Becton
Dickinson, Mountain View, CA) and were grown for 48 h until
reaching confluence. C6, OVCAR-3, and CHO cells were transfected using
calcium phosphate precipitation with reagents supplied by 5 Prime-3
Prime. GH3 cells were transfected with Lipofectin
reagent (Life Technologies, Gaithersburg, MD). Typically, 10 µg
promoter-luciferase plasmid DNA were transfected with or without 0.5
µg pCMV-ß-gal plasmid DNA according to the manufacturer’s
instructions. After 24 h, the luciferase (luc) and ß-gal
activities were measured using Luciferase Assay System reagents
(Promega) and the Galacto-Light kit (Tropix, Bedford, MA),
respectively. Chemiluminescence measurements were made over 10-sec
intervals in the model ILA911 semiautomatic luminometer (Tropix).
Luciferase activity was normalized to either ß-gal activity or
absorbance at 562 nm, as determined using the BCA protein assay reagent
(Pierce Chemical Co., Rockford, IL). Generally, we found that
normalization to either protein or to ß-gal activity gave similar
results (data not shown). However, the reduction in activity of the
shorter exon 1 and exon 2 promoter constructs in cells transfected with
the ß-gal plasmid suggested the possibility of squelching the IGF-I
promoter by the CMV promoter. Thus, we have chosen to show in
Results data from transfections performed without
ß-gal.
Nuclear extract preparation and electrophoretic mobility shift
assays (EMSA)
DNA fragments were isolated by appropriate restriction enzyme
digestion followed by gel purification and were end-labeled with the
appropriate [32P]deoxy (d)-NTPs using the Klenow enzyme
fill-in reaction. The labeled DNA fragments were purified using
Elutip-D columns (Schleicher and Schuell, Keene, NH). In some
experiments, probes were labeled by PCR (33). A typical reaction
contained 50 ng template DNA (the 1.5-kb
SmaI-SmaI fragment from the exon 2 promoter); 0.5
µM sense and antisense primers; 50 µM dATP;
200 µM dCTP, dTTP, and dGTP; 5 µl of
32P-dATP (3000 Ci/mmol); 10 µl 10 x reaction buffer
(10X reaction buffer contains: 500 mM KCl, 100
mM Tris-HCl, pH 8.3, 15 mM MgCl2,
and 0.01% (wt/vol) gelatin); and 0.5 U native Taq DNA
polymerase in a total volume of 100 µl. All reagents for PCR were
obtained from Perkin-Elmer/Cetus. After an initial denaturation at 94
C, 20 cycles of PCR were performed as follows: denaturation for 1 min
at 94 C, annealing for 1 min at 50 C, and elongation for 1 min at 72 C.
After a final 7-min elongation, PCR products were purified using
Elutip-D columns. PCR reactions were conducted using the Crocodile II
thermocycler from Appligene (Rohnert Park, CA). Nuclear extracts from
cell lines were prepared using the high salt method (34). Protein
concentrations of the extracts were determined using the method of
Bradford (35). EMSA was carried out using the BandShift kit (Pharmacia,
Piscataway, NJ). Briefly, 1 ng labeled DNA (5000 cpm) was
incubated with nuclear protein and other unlabeled DNA fragments or
oligonucleotides as indicated in Results. Binding assays
were carried out in 10 mM Tris-HCl (pH 7.5), 50
mM NaCl, 0.5 mM dithiothreitol, 10% glycerol,
and 0.05% Nonidet P-40 at 25 C for 20 min. The reactions were then
electrophoresed through nondenaturing 5% polyacrylamide gels. The
complexes formed were visualized by autoradiography of the dried
gels.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). These constructs
were transiently transfected into C6, OVCAR-3,
GH3, and CHO cells.
C6 cells express only exon 1 mRNAs (7). The construct containing the
longest amount of 5'-flanking sequence (1122 bp) and almost all of exon
1 (i.e. 362 of 381 bp; fragment 1 in Fig. 1) stimulated
luciferase activity by 2.4-fold over the control, promoterless plasmid
(pGL-2 basic). Deletion of 339 bp (fragment 3) or 989 bp (fragment 4)
of 5'-flanking sequence led to increased luciferase activity. Highest
luciferase activity was observed in fragment 4 (Fig. 1), which
contained 133 bp of 5'-flanking sequence and 362 bp of exon 1,
including all four transcription start sites. This fragment was
arbitrarily divided into -133/+74 and +75/+362 fragments to determine
the contributions of the different start sites and exon 1 sequence to
promoter activity. The +75/+362 fragment (fragment 5), which contains
only start sites 3 and 4 as well as 287 bp of exon 1 sequence,
stimulated luciferase activity by 6.2-fold over the promoterless
luciferase plasmid. In contrast, the -133/+74 fragment, which contains
133 bp of 5'-flanking sequence and 74 bp of 5' exon 1 sequence,
including start sites 1 and 2, stimulated luciferase activity less than
2-fold. Deletion of 170 bp from the 3'-end of fragment 4 to produce
fragment 6 (Fig. 1) resulted in a fragment that stimulated luciferase
activity by 7-fold. This fragment retained 133 bp of 5'-flanking
sequence and transcription start sites 1, 2, and 3, but transcription
start site 4 was deleted. When the -1122 to -783 sequence was placed
upstream of the minimal promoter containing only start sites 3 and 4, a
4-fold stimulation of luciferase activity was observed (Fig. 1, fragment 2).
OVCAR-3 cells are reported to express much lower levels of exon 1
transcripts than exon 2 transcripts (6, 36, 37). The longest exon 1
promoter fragment (-1122/+362) was inactive in transient transfection
assays in OVCAR-3 cells (Fig. 1). The exon 1 promoter fragments in
which deletions of 5'-sequence occurred were, however, active in
OVCAR-3 cells. Similar patterns of expression of luciferase activity
were seen in the OVCAR-3 cells as in C6 cells using these deletion
constructs (Fig. 1).
GH3 cells express both exon 1 and exon 2 mRNAs
(38, 39). The longest exon 1 promoter fragment (fragment 1) did not
stimulate luciferase activity (Fig. 1). Moreover, deletion of 322 bp of
5'-flanking sequence (fragment 3), which was associated with increased
luciferase activity in C6 and OVCAR-3 cells (i.e. 3- to
6-fold over pGL2-Basic), resulted in a luciferase activity that was
only 1.8-fold over that seen with the promoterless control plasmid. As
with C6 and OVCAR-3 cells, the -133/+362 fragment had highest promoter
activity (6.5-fold stimulation of luciferase activity). Deletion of
transcription start sites 3 and 4 along with 287 bp of sequence at the
3'-end of exon 1 (fragment 7) led to a loss of promoter activity in
GH3 cells.
To date, endogenous expression of IGF-I mRNA in CHO cells has not been
reported. However, in confirmation and extension of previous
observations in CHO cells (8), the exon 1 promoter fragments were
active in transient transfection assays in CHO cells. The pattern of
activity among the different fragments was similar to that observed in
the other cells lines (Fig. 1).
Exon 2 promoter activity in transient transfection assays
C6 cells do not express detectable exon 2 transcripts (7). The
longest exon 2 promoter fragment (Fig. 2, fragment 1), which included
1500 bp of 5'-flanking sequence and 44 bp of exon 2 (including all
transcription initiation sites), was essentially inactive in C6 cells.
Serial deletions of 5'-flanking sequence were made to generate
constructs containing 485, 362, 79, and 73 bp (fragments 2, 3, 4, and
5, respectively). These promoter fragments stimulated luciferase
activity by 4- to 9-fold in C6 cells (Fig. 2). However, a promoter
construct that contained only 36 bp of 5'-flanking sequence (fragment
6) did not stimulate luciferase activity.
OVCAR-3 cells express high levels of exon 2 transcripts (6, 36, 37).
The construct containing 1500 bp of 5'-flanking sequence and 44 bp of
exon 2 stimulated luciferase activity by 13-fold. Serial deletion of
5'-flanking sequence resulted in promoter constructs that stimulated
luciferase activity by 13- to 23-fold (fragments 2–5; Fig. 2). The
fragment containing only 36 bp of 5'-flanking sequence (fragment 6) was
inactive.
The longest exon 2 promoter construct (-1500/+44) stimulated
luciferase activity by 2.6-fold in GH3 cells and,
as previously reported (8), by 4-fold in CHO cells (Fig. 2). Serial
deletion of 5'-flanking sequences resulted in similar changes in
promoter activity in GH3 and CHO cells as in C6
and OVCAR-3 cells.
IGF-I promoter regions specifically bound by nuclear proteins in
EMSA
IGF-I promoter regions were examined for their ability to bind
nuclear proteins from the different cell lines. A single band was
observed in EMSA when 1 ng 32P-labeled exon 1 promoter
fragment 5 (+75/+362) was incubated with 1 µg nuclear protein from C6
and GH3 cells (Fig. 3A). This band was not visible with 1
µg protein from CHO cell nuclear extracts in the EMSA shown in Fig. 3A. However, in other EMSA using 2 or 4 µg protein from CHO cell
nuclear extracts or upon longer exposure using 1 µg protein, this
band was observed (data not shown). The appearance of this band in EMSA
using C6 and GH3 nuclear extracts was completely
abolished by coincubation with a 100- or 300-fold molar excess of the
unlabeled +75/+362 fragment, but its appearance was not affected by
coincubation of a 300-fold molar excess of DNA containing an Oct-1
transcription factor-binding site (Fig. 3A). When the -133/+74 exon 1
fragment was used, no bands were observed in EMSA under the conditions
used (data not shown).
|
, B and
C). As the same CHO cell nuclear extract was used for exon 1 and exon 2
binding studies, the positive results obtained with the exon 2 promoter
fragment suggest that the lack of exon 1 promoter binding by 1 µg CHO
cell nuclear protein was not due to a gross protein degradation in the
CHO extract. Exon 2 binding was inhibited by excess unlabeled -236/+44
DNA fragment, but not by excess unlabeled DNA containing an Oct-1
consensus binding site. The upstream fragment (-485/-237) did not
produce any bands on EMSA under the conditions used (data not
shown).
To determine whether the binding of the exon 1 and exon 2 promoter
fragments could be due to common cis-acting elements,
competition EMSAs were conducted. As shown in Fig. 4A, the unlabeled exon 2 fragment (-236/+44) inhibited the binding of the
labeled exon 1 +75/+362 fragment by nuclear extract from
GH3 cells. Similarly, excess unlabeled exon 1
fragment inhibited the binding of the labeled exon 2 fragment by
OVCAR-3 cell nuclear extracts (Fig. 4B).
|
1200-fold molar excess) containing activating
protein-1 (AP-1) and AP-3 binding sites, respectively, partially
inhibited the binding of the +75/+362 exon 1 promoter fragment to
GH3 cell nuclear extracts (Fig. 5A). Unlabeled oligonucleotides containing AP-2 and AP-3
consensus binding sequences inhibited binding of the -236/+44 labeled
exon 2 promoter fragment to GH3 nuclear extracts
(Fig. 5B). An oligonucleotide containing a nuclear factor-1 binding
site had little effect on the binding of either the exon 1 or the exon
2 promoter fragments by GH3 extracts. The AP-2
oligonucleotide did not inhibit exon 1 promoter binding (Fig. 5A), and
the AP-1 oligonucleotide did not inhibit exon 2 binding (Fig. 5B). An
oligonucleotide containing a stimulatory protein 1 binding site
appeared to slightly inhibit exon 1 binding (Fig. 5A).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Deletion of 170 bp from the 3'-end of the +75/+362 fragment, including
transcription start site 4, did not lead to a major loss of promoter
activity. Previous studies, in which deletion of start site 3 and
downstream exon 1 sequence resulted in a large reduction of promoter
activity despite the presence of the first 107 bp of exon 1, including
start sites 1 and 2, had first suggested the importance of start site 3
to exon 1 promoter activity (5, 28) (Fig. 6). The
current results strongly support the hypothesis that a core exon 1
promoter is located around transcription start site 3. This site is
contained within site HS3C, which is footprinted in vitro by
rat liver extracts (18) (Fig. 6). However, it should also be noted that
the HS3C footprint is not produced by rat osteoblast nuclear extracts
(28). We have observed a terminal deoxynucleotidyl transferase gene
(TdT)-like initiator motif around start site 3. The TdT core initiator
is 5'-CCCTCATTCT-3', where the A residue is the single start
site (40, 41). The putative start site 3 initiator element would be
5'-CCCTCTTCT-3', where transcription initiation
sites are in boldface. The lack of a purine residue in the
putative start site 3 initiator may explain why multiple start sites
are used.
|
) by C6 cell nuclear extracts (22). Mutations
in this sequence decreased or abolished promoter activity despite the
presence of transcription start site 3 and downstream sequence (22).
Thus, in the studies of An and Lowe (22), the -18/+75 fragment
contained major basal promoter elements, whereas in our studies the
+75/+362 fragment appears to contain a second basal promoter activity
(Fig. 6). An and Lowe (22) obtained generally higher luciferase
activity than that observed in the current study. The reason for this
is not known. The reason for our inability to observe protein binding
to the -133/+74 fragment could be due to our use of a lower amount of
nuclear protein (1 µg) in EMSA. However, in qualitative terms, the
results from the two studies are, in fact, consistent. An and Lowe (22)
observed that addition of the 5'-flanking sequence from -18 to -412
resulted in a large decrease in promoter activity in C6 cells and in
fibroblasts when only the first 75 bp of exon 1 (i.e. start
sites 1 and 2) were present. Extension of the 3'-end of this construct
to +183 (i.e. to include start site 3) slightly relieved the
inhibition in C6 cells and fully relieved the inhibition in fibroblasts
(22). Extension of the exon 1 sequence to +302 resulted in maximal
promoter activity in C6 cells (22) (Fig. 6). The studies of An and Lowe (22) and our current work suggest the hypothesis that there are two potentially independent promoters directing IGF-I exon 1 transcription. The promoter determined by An and Lowe (22) to be regulated by the -10 to +9 region may be more sensitive to inhibition by upstream sequences between -18 and -412. This inhibition may be relieved by the presence of a start site 3 promoter and downstream exon 1 sequences. In addition, the IGF-I-FP1 promoter may be activated in the presence of start site 3 and downstream exon 1 sequence. The presence of two promoters directing transcription in a single exon in other genes and their physiological significance have been recently discussed (42). Elucidation of the mechanisms by which the two hypothetical exon 1 promoters may be coordinately and differentially regulated awaits further study.
Location of minimal exon 2 promoter
Deletion analysis in transfection experiments indicated that exon
2 promoter activity was retained when as little as 73 bp of 5'-flanking
sequence and 44 bp of exon 2 were placed upstream of the luciferase
structural gene. However, exon 2 promoter activity was completely
abolished when only 36 bp of 5'-flanking sequence were present. Thus,
the sequences required for minimal exon 2 promoter activity are
contained between -73 and -36. The critical region between -73 and
-36 falls within two in vitro footprints produced by rat
liver nuclear extracts (21) (Fig. 7). A putative
CACCC-box (5'-CCCCACCC-3') is at the -53 position (43, 44, 45) (Fig. 7).
Preliminary results indicate that mutation of this sequence completely
abolishes exon 2 promoter activity (Wang, X., and M. L. Adamo,
unpublished data). The current deletion analysis indicates that the
CCAAAT element at -80 is not required for exon 2 promoter activity.
Indeed, it is not clear what sequence constitutes the core exon 2
promoter. The exon 2 start site cluster may contain two TdT-like (40, 41) (Fig. 7) initiator sequences:
5'-TTCGGCCTCATAATA
CCCACTCT-3' (transcription start sites are in
boldface, and the two putative initiators are in
italics).
|
Cell type specificity of promoter activity
The inhibition of proximal exon 1 and exon 2 promoter activity by
upstream sequence appears to be cell type specific and may contribute
to actual patterns of endogenous gene expression. For example, the
-1122/+362 exon 1 promoter was essentially inactive in OVCAR-3,
GH3, and CHO cells, whereas it stimulated
luciferase activity by 2.4-fold in C6 cells (Fig. 2). Deletion of
upstream sequences resulted in promoter activity in all cell lines. C6
cells express IGF-I mRNA, resulting primarily from the use of exon 1
start site 3 (Yang, H., and M. L. Adamo, manuscript in preparation).
Exon 1 mRNA is expressed in OVCAR-3 cells at extremely low levels, if
at all (35, 36). In GH3 cells, initiation of exon
1 transcription from start sites 2 and 3 is very low compared to
initiation from a site near the 3'-end of exon 1 (Yang, H., and M. L.
Adamo, manuscript in preparation). This site is not contained in our
promoter constructs. This may explain why the -1122/+362 construct was
not active in GH3 cells. Thus, the degree to
which the proximal promoter is inhibited by upstream sequences may
contribute to the actual level of endogenous exon 1 gene expression in
the cell lines studied. Upstream sequence markedly stimulated proximal
promoter activity in SK-N-MC cells, had little stimulatory effect in
osteoblast cells, and appeared to be inhibitory in C6 cells (4, 5, 6, 7, 22, 28) (Figs. 1 and 6). The mechanism of cell type-specific inhibition of
the proximal promoters may be due to the presence of cell type-specific
inhibitor proteins that bind to upstream silencer elements.
Our results in OVCAR-3 cells confirm a previous report in the literature that the exon 2 promoter is more active than the exon 1 promoter in this predominantly exon 2-expressing cell line (6). These researchers also noted that the exon 2 promoter was more active in the OVCAR-3 cells than it was in exon 1-producing SK-N-MC cells (6). In the current study, upstream sequences, especially between -1500 and -362, inhibited the activity of the promoter. Inhibition of exon 2 promoter activity by upstream sequences was also reported previously for the human promoter (6), although in that study, the effects of 5'-deletions beyond approximately -1200 were not reported. As the inhibition by upstream sequences was less in the exon 2-expressing OVCAR-3 cells and GH3 cells than in the C6 cells, it is possible that the sequence between -1500 and -362 interacts with cell type-specific inhibitors. The presence of silencer elements upstream of both the exon 1 and exon 2 minimal promoters suggests the possibility that derepression is a potential mechanism for IGF-I gene activation.
Common and distinct features of the exon 1 and exon 2 promoter
regions
Our EMSA data clearly show that the exon 1 and the exon 2 proximal
promoter regions have different nuclear protein binding
characteristics. The differences in the exon 1 and exon 2 promoter
binding patterns may reflect differences in the regulation of the two
promoter activities. In cross-competition EMSA, however, the exon 1 and
exon 2 promoter fragments inhibited one another’s binding. These data
suggest that there are both common and distinct elements in the exon 1
and exon 2 promoter regions, which may explain common and differential
regulation of exon 1 and exon 2 mRNA levels (10, 11, 12, 13, 14, 15). The
cis-acting elements reported or suggested to be important
for exon 1 and exon 2 promoter activity have been summarized in Figs. 6 and 7, respectively. It is evident that sequences that are important
for both basal and regulated transcription are located in the exon 1
proximal promoter region. This may also hold true for the exon 2
proximal promoter region.
In summary, 1) we have located a potential minimal exon 1 promoter around transcription start site 3; 2) we have determined the 5'-limit of the minimal exon 2 promoter; 3) we have presented data suggesting that there are common and distinct cis-acting elements in both the exon 1 and exon 2 proximal promoters; and 4) we have presented data suggesting that sequences upstream of the proximal promoters may contain cell type-specific silencer elements. Future studies will focus on characterization of the core promoters in the IGF-I gene and how they interact with cell type-, developmental stage-, and stimulus-specific factors to achieve integrated transcriptional control of IGF-I gene expression.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received September 16, 1996.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and
LAP. Mol Endocrinol 8:1636–1654[Abstract]
activates promoter 1 of the human
insulin-like growth factor-I gene via two distinct sites. Mol
Endocrinol 9:1488–1499[Abstract]
This article has been cited by other articles:
 |
|
 |
|
  C. M. Alfieri, H. J. Evans-Anderson, and K. E. Yutzey Developmental regulation of the mouse IGF-I exon 1 promoter region by calcineurin activation of NFAT in skeletal muscle Am J Physiol Cell Physiol, May 1, 2007; 292(5): C1887 - C1894. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. S. McLellan, T. Kealey, and K. Langlands An E box in the exon 1 promoter regulates insulin-like growth factor-I expression in differentiating muscle cells Am J Physiol Cell Physiol, August 1, 2006; 291(2): C300 - C307. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. W. Wynes and D. W. H. Riches Transcription of macrophage IGF-I exon 1 is positively regulated by the 5'-untranslated region and negatively regulated by the 5'-flanking region Am J Physiol Lung Cell Mol Physiol, June 1, 2005; 288(6): L1089 - L1098. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. E. McCall, D. L. Allen, F. Haddad, and K. M. Baldwin Transcriptional regulation of IGF-I expression in skeletal muscle Am J Physiol Cell Physiol, October 1, 2003; 285(4): C831 - C839. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. F. Kuemmerle and H. Zhou Insulin-like Growth Factor-binding Protein-5 (IGFBP-5) Stimulates Growth and IGF-I Secretion in Human Intestinal Smooth Muscle by Ras-dependent Activation of p38 MAP Kinase and Erk1/2 Pathways J. Biol. Chem., May 31, 2002; 277(23): 20563 - 20571. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Wang, X. Wang, and M. L. Adamo Two Putative GATA Motifs in the Proximal Exon 1 Promoter of the Rat Insulin-Like Growth Factor I Gene Regulate Basal Promoter Activity Endocrinology, March 1, 2000; 141(3): 1118 - 1126. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J.-L. Zhu, C.-I Pao, E. Hunter Jr., K.-w. M. Lin, G.-j. Wu, and L. S. Phillips Identification of Core Sequences Involved in Metabolism-Dependent Nuclear Protein Binding to the Rat Insulin-Like Growth Factor I Gene Endocrinology, October 1, 1999; 140(10): 4761 - 4771. [Abstract] [Full Text]
|
|
|
|
|
|
C. Benbassat, L. N. N. Shoba, M. Newman, M. L. Adamo, S. J. Frank, and W. L. Lowe Jr. Growth Hormone-Mediated Regulation of Insulin-Like Growth Factor I Promoter Activity in C6 Glioma Cells Endocrinology, July 1, 1999; 140(7): 3073 - 3081. [Abstract] [Full Text]
|
|
|
|
|
|
X. Wang, J. L. Talamantez, and M. L. Adamo A CACCC Box in the Proximal Exon 2 Promoter of the Rat Insulin-Like Growth Factor I Gene Is Required for Basal Promoter Activity Endocrinology, March 1, 1998; 139(3): 1054 - 1066. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||
