| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
,*
Departments of
* Microbiology and Immunology,
Pathology and Laboratory Medicine, and
Biological Sciences, University of Kentucky, Lexington, KY 40536
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
mRNA encoding pIgR is transcribed from a single gene (locus
PIGR) that has been mapped to human chromosome 1q31–32
(14, 15, 16), mouse chromosome 1E3 (17, 18), and
bovine chromosome 16q13 (19). Initial characterizations of
the promoter regions of the human (20, 21, 22), rat
(23), and mouse (24) PIGR genes
indicate that a region including approximately 100 nt upstream of the
transcription start site is required for basal transcription. Within
this region, a binding site for the ubiquitous transcription factor
upstream stimulatory factor (USF) has been shown to regulate basal
transcription of the PIGR gene (22, 23, 24).
Expression of pIgR mRNA is up-regulated by the proinflammatory
cytokines IFN-
, TNF-
, and IL-1
and the Th2 cytokine IL-4
(reviewed in Refs. 2 and 3). A binding site
for IFN regulatory factor-1 (IRF1) in the first exon of the
PIGR promoter has been implicated in the response to IFN-
(20, 25) and TNF- (26, 27). Other TNF-
responsive elements include binding sites for NF-
B in the
5'-flanking region and the first intron of the human PIGR
gene (27, 28). The signaling pathways for IL-1
regulation of the PIGR gene have not yet been identified,
although they are likely to share elements with the TNF- pathway. A
binding site for STAT6 has recently been implicated in the
up-regulation of PIGR gene transcription by IL-4
(29)
Here we present a detailed analysis of elements in the promoter of the human PIGR gene that regulate basal transcription. Our analysis indicates that the E box motif at position -71 is the most important positive regulatory element of the human PIGR promoter. The major E box binding protein in human intestinal CaCo2 cells was identified as a heterodimer of USF1 and USF2. Other potential regulatory elements include a binding site at position -46 for the transcription factor AP2 and a binding site for an unidentified factor between the USF and AP2 sites. Two inverted repeat motifs in the PIGR promoter may have the potential to down-regulate the PIGR promoter by forming an inhibitory stem-loop structure. Our data suggest a model by which the human PIGR promoter is regulated by cooperative interactions among multiple transcription factors.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The CaCo2 human colon carcinoma cell line, the HepG2 human hepatoma cell line, and the Hepa1–6 mouse hepatoma cell line were obtained from American Type Culture Collection (Manassas, VA). Transimmortalized mouse intestinal cells (m-ICcl2) that maintain a crypt phenotype (30) were a gift from Dr. J.-P. Kraehenbuhl (University of Lausanne, Lausanne, Switzerland). CaCo2 cells were maintained in RPMI 1640 medium with a final concentration of 25 mM HEPES, 2 mM L-glutamine, 10% FCS, and antibiotics (100 U/ml penicillin, 100 µg/ml streptomycin, and 250 ng/ml fungizone). HepG2 cells were maintained in DMEM/Ham’s F-12 (1/1) medium with a final concentration of 15 mM HEPES, 2.5 mM L-glutamine, 10% FCS, and antibiotics. Hepa1–6 cells were maintained in DMEM with a final concentration of 4.5 g/L glucose, 20 mM HEPES, 4 mM L-glutamine, 10 µg/ml insulin, 10% FCS, and antibiotics. The m-ICCl2 cells were maintained in DMEM/Ham’s F-12 (1/1) medium with a final concentration of 20 mM HEPES, 2.5 mM L-glutamine, 10 µg/ml insulin, 5.5 µg/ml transferrin, 6.7 ng/ml sodium selenite, 2% FCS, and antibiotics. All cell culture reagents were obtained from BioWhittaker (Walkersville, MD), except for insulin, transferrin, and sodium selenite (Life Technologies, Gaithersburg, MD). Cells were maintained at 37°C in a humidified 5% CO2 atmosphere.
Construction of chimeric pIgR-luciferase reporter plasmids
Chimeric reporter plasmids were constructed in which sequences from the PIGR gene were inserted upstream of the firefly luciferase gene in the pGL2-Basic plasmid (Promega, Madison, WI) as previously described (20). Internal deletions and mutations in the human PIGR promoter were generated by the two-step megaprimer PCR method (31). In the first PCR step, mutagenic primers were paired with a lower primer corresponding to the reverse complement of nt -1 to +29, and a 12-nt extension at the 5' end that generated a HindIII site. All primers were obtained from Life Technologies. The wild-type (WT) PIGR promoter was used as a template. PCR was performed using the AmpliWax PCR gem-facilitated hot start method (PerkinElmer, Norwalk, CT) with ELONGASE Enzyme Mix (Life Technologies). The conditions for PCR were 30 cycles of 94°C for 2 min, 55°C for 2.5 min, and 72°C for 4 min, followed by a final extension at 72°C for 10 min. The products of these reactions were gel-purified and used as the lower megaprimers in the second PCR reaction. The upper primer corresponded to nt -280 to -251 and a 12-nt extension at the 5' end that generated an XhoI site. The amplified products from the second PCR reactions were gel-purified, digested with XhoI and HindIII, and cloned into the promoterless pGL2-Basic plasmid. The sequences of all plasmids were verified by automated fluorescence sequencing.
Transfection of cells and analysis of PIGR promoter activity
Cells were transiently cotransfected with reporter-luciferase
plasmids and a control vector in which the CMV immediate early promoter
controls expression of the bacterial chloramphenicol acyltransferase
(CAT) gene (pcDNA3.1(-)CAT; Invitrogen, San Diego, CA; Figs. 2
and 3)
or Renilla luciferase gene (pRL-CMV; Promega; Figs. 4 and 6). Cells were plated in 35-mm culture wells at 
70–80% confluence
24 h before transfection. Two micrograms of PIGR
promoter-luciferase plasmid and either 1 µg pcDNA3.1(-)CAT plasmid
or 0.4 µg pRL-CMV plasmid were mixed with Tfx-50 reagent (Promega;
2:1 Tfx-50:DNA charge ratio) in a final volume of 1 ml serum-free
medium and incubated with cells for 4 h at 37°C. Cells were then
overlaid with 2-ml complete medium with serum, without removing the
Tfx-50-DNA mixture. Culture media were changed daily until the cells
were harvested at 72 h (
Figs. 2–4) or 24 h (Fig. 5) after
transfection. For the experiments shown in Figs. 2 and 3, preparation
of cell lysates and analysis of luciferase and CAT activities were
performed as previously described (20). For the
experiments shown in Figs. 4 and 5, firefly and Renilla
luciferase activities were analyzed using the Dual-Luciferase Reporter
Assay System according to the manufacturer’s protocol (Promega).
Promoter activity was calculated as [firefly luciferase
activity/control enzyme activity (CAT or Renilla
luciferase)] x 100%.
|
|
|
|
|
Nuclear protein extracts were prepared as previously described (32), and protein concentrations were determined by the Bradford assay (33) (Bio-Rad, Hercules, CA). EMSAs were performed as previously described (20) with the following modification to reduce the formation of nonspecific protein/DNA complexes. Nuclear extracts (10 µg protein) were preincubated for 15 min at room temperature with 2 µg poly d(I-C) plus unlabeled homologous or heterologous double-stranded oligodeoxynucleotides. Next, 10-14 mol 32P-labeled double-stranded oligodeoxynucleotide was added, and incubation was performed for an additional 15 min at room temperature. Where indicated, 2 µg Abs to the transcription factors USF1, USF2, c-Myc, Max, or IRF1 (Santa Cruz Biotechnology, Santa Cruz, CA) were added and incubated for 1 h at 5°C. Protein-bound DNA was separated from free DNA by nondenaturing PAGE as previously described (20).
DNase I footprinting analysis
A nonradioactive detection method was developed by end-labeling dsDNA probes with digoxigenin (DIG), and detection with anti-DIG Abs. DNA probes, spanning nt -280/-47 or -280/+29 of the PIGR gene, were generated by PCR as described above, except that the upper primer was labeled at its 5' end with DIG according to the manufacturer’s protocol (Roche, Indianapolis, IN). Purified DIG-labeled DNA was incubated for 10 min on ice with or without added protein (50 µg nuclear protein from CaCo2 or HepG2 cells, or increasing volumes of bacterial extract containing 0.5, 1.0, or 2.0 µg recombinant AP2 protein). DNase I digestions were conducted with the Core Footprinting System (Promega). The concentrations of DNase I and the incubation times were optimized for each sample and are listed in the figure legends. We determined in preliminary experiments that it was necessary to add higher concentrations of DNase I in the presence of nuclear extracts from CaCo2 and HepG2 cells to achieve a degree of digestion of the DNA probe equivalent to that seen in the absence of extract. DNA sequencing ladders were generated using the T7 DNA polymerase chain termination method (Sequenase version 2.0; Amersham Life Science, Cleveland, OH) and the same 5'-DIG-labeled oligodeoxynucleotide that was used to generate the labeled footprinting probes. The DNase I-digested samples and the sequencing ladders were separated by electrophoresis in a 6% polyacrylamide gel containing 100 mM Tris-borate (pH 8.3), 1 mM EDTA, and 7.0 M urea. The gel was capillary blotted overnight onto a charged nylon membrane (Osmonics Laboratory Products, Minnetonka, MN) and baked for 1 h at 80°C. DIG-labeled DNA was detected using anti-DIG Ab and an alkaline phosphatase-based fluorescent detection system (Roche). Fluorescent membranes were exposed at room temperature to XAR AF 1417 film (Action Scientific, Carolina Beach, NC).
Data analysis
Analyses of free energy changes associated with DNA stem-loop
formation were performed using OLIGO 4.0 Primer Analysis Software
(National Biosciences, Plymouth, MN). Two-way ANOVA was used to test
for statistical differences in promoter activity, controlling for
multiple independent experiments (StatView; SAS Institute, Cary, NC).
For Figs. 2
and 5, individual comparisons were made using Student’s
t test. For Figs. 3 and 4, pairwise multiple comparisons
were made using Fisher’s protected least significant difference
test.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We previously reported that the region from -280/+29 of the human
PIGR gene was sufficient to promote transcription of a
heterologous reporter gene in a human colon carcinoma cell line
(20). Fig. 1 illustrates an
alignment of this region of the human PIGR gene with the
corresponding regions of the rat and mouse PIGR genes. The
region from -206/+29 of the human PIGR gene has
overall similarities of 63 and 66% to the region from
-188/+25 of the rat and mouse PIGR genes,
respectively. Upstream of position -206, however, the sequence of the
human PIGR gene diverges completely from the rodent
PIGR genes. To compare directly the activities of the human
and mouse PIGR promoters, the homologous regions were cloned
into reporter vectors containing the firefly luciferase gene, which
were transiently transfected into human and mouse intestinal and liver
epithelial cell lines (Fig. 2). We chose
intestine and liver because the endogenous expression of pIgR mRNA
varies between these tissues in a species-specific pattern (see
Discussion). The human PIGR promoter was
significantly more active than the mouse homolog in all cell lines
tested regardless of the tissue of origin: 4-fold higher in CaCo2 human
colon carcinoma cells, 6-fold higher in HepG2 human hepatoma cells,
4-fold higher in m-ICCl2 mouse intestinal
epithelial cells, and 9-fold higher in Hepa1–6 mouse hepatoma cells.
In addition, PIGR promoter activity was consistently higher
in intestinal than in liver cell lines. The human PIGR
promoter was 11-fold more active in human intestinal cells than in
human liver cells and 7-fold more active in mouse intestinal cells than
in mouse liver cells. Similarly, the activity of the mouse
PIGR promoter was 16-fold higher in human intestinal cells
than in human liver cells and 15-fold higher in mouse intestinal cells
than in mouse liver cells. Although we observed a consistent pattern,
we caution that our conclusions are based on results from only four
cell lines. It should also be noted that PIGR promoter
activity was expressed relative to the activity of a cotransfected CMV
promoter. We cannot rule out the possibility that tissue differences in
CMV promoter activity contributed to the observed differences in
PIGR promoter activity between intestinal and liver cell
lines.
|
Comparison of the sequences of the human PIGR promoter
within the region from -206/+1 with the corresponding regions of the
rodent PIGR promoters revealed subregions of significant
similarity or divergence (Fig. 1). We designated four subregions (sites
A–D) in which notable differences between the human and rodent
PIGR genes were observed. Sites A and B differ significantly
in sequence between the human and rodent PIGR genes, and
site C contains an extra 18-bp sequence with no homolog in the rodent
PIGR genes. Homology between the human and rodent
PIGR genes is higher in site D, except for a critical A-G
difference within the E box motif and differences in the nucleotides
flanking the 3' end of the E box. At positions -120 (between sites B
and C) and -63 (downstream of site D) are two 10-bp motifs that are
complementary in sequence (i.e., inverted repeats). The rodent
PIGR genes also contain 10-bp inverted repeats at this
position, which are 90% identical in sequence to the inverted repeats
in the human PIGR gene. A consensus binding site for the
transcription factor AP2 was observed at position -48 of the human,
but not the rodent, PIGR promoters. To determine which
subregions of the human PIGR promoter bind nuclear proteins,
we performed a DNase I footprint analysis of the region spanning nt
-280/-47 (Fig. 3A). Binding
of nuclear proteins from CaCo2 cells completely protected the region
spanning site D and the second inverted repeat from DNase I digestion.
Binding of nuclear proteins from HepG2 cells caused a similar, but less
striking, footprint over site D, consistent with the lower promoter
activity in HepG2 vs CaCo2 cells (Fig. 2). Binding of nuclear proteins
from CaCo2 cells, but not HepG2 cells, created a DNase I-hypersensitive
site at the junction of sites C and D. This hypersensitive site may
have resulted from bending of the DNA strand induced by binding of
proteins to site D. While binding of nuclear proteins caused some minor
changes in DNase I sensitivity in sites A, B and C, they were not as
striking as the changes in site D. These results suggest that
recruitment of nuclear proteins to site D may be important for
transcription of the PIGR gene.
Since recruitment of nuclear proteins to the PIGR promoter
was striking in CaCo2 cells, we used that cell line to test the effects
of specific mutations on promoter activity. We began by creating
internal deletions of sites A, B, C, and D (Fig. 3B). All
deletions were 22 bp in length, representing approximately two turns of
the DNA helix, to minimize potential effects of disrupting spacing
between regulatory elements. This approach has been used to preserve
the stereospecific alignment of regulatory elements in the promoter of
the HLA-DRA gene (34). Deletion of site A caused a modest,
but highly reproducible, decrease in promoter activity. Deletion of
site B had no effect, while deletion of site C caused a significant
increase in promoter activity. Deletion of site D caused a dramatic
decrease in PIGR promoter activity to a level that was not
significantly different from that of the promoterless luciferase
plasmid. These results suggested that sites A and D may contain
positive regulatory elements, and that an element(s) in site D may be
required for basal transcription of the PIGR gene. In
contrast, an element(s) in site C may act as a negative regulator(s) of
PIGR promoter activity.
Effects of point mutations in potential regulatory motifs on activity of the human PIGR promoter
The identification of site D as a major focus of protein binding
and promoter activity led us to investigate the E box motif at position
-71. The canonical E box sequence CACGTG is a target for transcription
factors of the basic helix-loop-helix leucine zipper (bHLH-zip) family,
which includes USF, members of the Myc/Max/Mad family, TFE3, and TFEB
(35). A study of base preferences for DNA binding
demonstrated that bHLH-zip proteins can also bind the core sequence
CACATG, although with lower affinity (36). To test the
hypothesis that the E box motif is responsible for the transcriptional
activity of site D, we observed the effects of two mutations within the
E box on activity of the human PIGR promoter (Fig. 4A). For the E box up
mutation, the core sequence CACATG was changed to the
canonical E box sequence CACGTG (nucleotides
differing from the WT human PIGR sequence are underlined).
For the E box down mutation, the core sequence CACATG
was changed to CACAGA, which has been reported to abolish
binding of bHLH-zip proteins (36). We also included the
mouse PIGR promoter in this analysis, since its E box has
the canonical CACGTG sequence (see Fig. 1). The E box up
mutation caused a significant increase in PIGR promoter
activity, presumably due to enhanced binding of bHLH-zip proteins. In
contrast, the E box down mutation caused a significant decrease in
PIGR promoter activity to a level that was statistically
indistinguishable from the promoterless control plasmid. This effect
was comparable to deletion of the entire site D (Fig. 3B),
confirming our hypothesis that the E box is responsible for the
transcriptional activity of site D. As shown in Fig. 2, the activity of
the mouse PIGR promoter was significantly less than that of
the human PIGR promoter despite the fact that the mouse
PIGR promoter contains the canonical E box sequence (based
on the multiple group comparison test used to analyze the data in Fig. 4A, the activity of the mouse PIGR promoter
appeared to be indistinguishable from the promoterless control plasmid;
however, an individual comparison by t test did detect a
significant difference between the mouse PIGR promoter and
the control plasmid). This result suggests that sequences outside the E
box are responsible for the enhanced activity of the human
PIGR promoter relative to its murine counterpart.
We next tested the effects of mutations in the 10-bp inverted repeats
at positions -120 and -63 of the human PIGR promoter. The
sequence of repeat 2 is the reverse complement of the sequence of
repeat 1, raising the possibility that stem-loop structures could form
by base-pairing of the upper and lower strands of repeat 1 with the
corresponding strands of repeat 2 (Fig. 4B). Base-pairing of
these 10-nt sequences would theoretically be associated with a 
G of
-11.9 kcal/mol. Formation of stem-loops could result in
transcriptional repression, since the important E box motif between
repeat 1 and repeat 2 would be disrupted. Conversely, binding of
bHLH-zip proteins such as USF to the E box motif should stabilize the
DNA in a double-stranded conformation, inhibiting stem-loop formation.
To test the hypothesis that stem-loop formation may regulate the
activity of the PIGR promoter, we created a series of
mutations in the two inverted repeats, as described in Fig. 1. For
MutRpt1 and MutRpt2, multiple point mutations were made in repeats 1
and 2, respectively, while the opposite repeat was maintained in the WT
configuration. Either of these mutations would prevent stem-loop
formation. The mutations in repeats 1 and 2 were designed to be
complementary, so that stem-loop formation could be restored in the
double mutant (predicted G, -9.2 kcal/mol). However, if the actual
sequences of the repeats were important (for example, for DNA-protein
interactions), then the double mutant promoter might have altered
activity relative to the WT promoter. Mutations in repeat 1 caused a
50% increase in PIGR promoter activity, consistent with the
hypothesis that disruption of stem-loop formation could enhance
promoter function (Fig. 4A). The relatively small difference
in activity between the MutRpt1 and WT promoters could have been
because binding of protein(s) to the E box constitutively inhibited
stem-loop formation in the WT promoter. If this were the case, then the
reduced activity of the E box down mutant could have partly been due to
enhanced stem-loop formation secondary to loss of protein binding at
the E box. To test this hypothesis we created a new reporter plasmid
that contained both the MutRpt1 and E box down mutations (MR1/EB Dn),
in which disruption of stem-loop formation might compensate for loss of
E box function. However, we found that the promoter activity of the
combined MR1/EB Dn mutant was just as low as that of the single E box
down mutant (Fig. 4A). We subsequently discovered that while
the E box down mutation specifically inhibited binding of the
transcription factor USF to the E box, other unidentified proteins
remained bound (see Fig. 6). We therefore favor the hypothesis that the
reduced promoter activity of the E box down mutant is due to changes in
the composition of proteins bound to the E box and not to enhanced
stem-loop formation.
In contrast to repeat 1, mutations in repeat 2 caused a significant
decrease in PIGR promoter activity to a level
indistinguishable from that of the promoterless control vector.
Activity of the double mutant promoter was the same as that of the
MutRpt2 promoter, suggesting that decreases in PIGR promoter
activity were due to mutations in the sequence of repeat 2 and not to
changes in stem-loop formation. In this context it is significant that
nuclear extracts from CaCo2 cells protected the region around repeat 2,
but not repeat 1, from DNase I digestion (Fig. 3A). Repeat 2
may therefore act as a positive regulatory motif by associating with a
specific transcription factor(s). Because of the strategic location of
Repeat 2 between the E box and a potential binding site for AP2 (see
Fig. 1), it was important to determine whether the mutations we created
in repeat 2 had disrupted the activity of adjacent regulatory elements.
For example, the mutations in repeat 2 included an A-T transition at
the nt immediately flanking the 3' end of the E box, a position that
can affect binding of transcription factors to the E box
(36). To isolate the effect of this mutation, we created a
single point mutation at this position (E box flank A-T; see Fig. 1).
This mutation had no effect on PIGR promoter activity (Fig. 4A), indicating that the suppressive effect of the mutations
in repeat 2 was not due to disruption of E box function. We next used
DNase I footprinting analysis to demonstrate that the mutations in
repeat 2 did not inhibit binding of recombinant AP2 to the adjacent
motif (Fig. 5A). To determine
whether the AP2 site was active in CaCo2 cells, we tested the effects
of mutations in the AP2 site on PIGR promoter function. The
CC dinucleotide at position -46 was the location of the major AP2
footprint (see arrows in Figs. 1 and 5A). We changed the C
at nt 66 to an A to match the sequence of the rodent PIGR
genes, which do not have a consensus AP2 motif at this
location. We also mutated the GG dinucleotide flanking the 5' end of
the AP2 site to CT, as found in the rodent PIGR genes. These
mutations disrupted binding of recombinant AP2 to the PIGR
promoter (Fig. 5A) and significantly reduced PIGR
promoter activity (Fig. 5B). Taken together, these results
demonstrate that mutations in the repeat 2 motif did not affect the
function of the adjacent E box and AP2 elements and, alternatively, may
have inhibited binding of a unique protein(s) to this region of the
PIGR promoter. We have not been able to identify protein(s)
binding to the either the repeat 1 or 2 motifs by EMSA (data not
shown), but this does not rule out the possibility that binding of
transcription factor(s) to repeat 2 may require complex interactions
with factors bound to the adjacent E box and AP2 motifs.
Identification of the major E box binding protein in nuclear extracts of CaCo2 cells
To determine the sequence requirements for binding of nuclear
proteins to site D of the human PIGR promoter, we performed
EMSAs using as probe a WT site D oligo or an oligo containing the E box
up mutation. Three specific DNA-protein complexes were observed in
nuclear extracts from CaCo2 cells (Fig. 6B). The pattern of
DNA-protein complexes was similar for the site D and E box up probes,
although the intensities of the bands were stronger with the E box up
probe. The pattern of competition by unlabeled DNA revealed a higher
binding affinity for the E box up oligo. Weak displacement was seen
with a 200-fold molar excess of unlabeled site D oligo, while the same
concentration of E box up oligo completely displaced complexes 1 and 2
(but not complex 3). Surprisingly, the E box down oligo caused some
displacement of complex 1 and complete displacement of complexes 2 and
3 from the WT site D oligo. However, this oligo did not compete with
the higher affinity E box up oligo. The simplest explanation for the
partial competition by the E box down oligo is that this mutation
lowered the affinity, but did not eliminate binding of transcription
factors to the E box motif. Some competition was therefore observed
with the low affinity WT E box probe, but not with the high affinity E
box up probe, in an in vitro binding assay. In living cells, however,
the lower affinity of the E box down motif was insufficient to support
PIGR promoter activity (Fig. 4A). To test whether
binding of proteins to site D required the E box motif, we used a
competitor oligo (Myc) that was homologous to site D only within the E
box and which contained the higher affinity CACGTG motif. At a 200-fold
molar excess, the Myc oligo displaced all three specific protein
complexes from the WT site D probe, confirming that binding of all
three complexes required the E box motif. These results were confirmed
by testing a range of concentrations of site D and Myc oligo
competitors (data not shown).
To identify the proteins binding to site D, we performed an EMSA in
which specific DNA-protein complexes were supershifted by Abs to known
members of the bHLH-zip family of transcription factors. We first
tested Abs to USF1 and USF2, protein subunits that combine as hetero-
and homodimers to form the transcription factor, USF (37).
Using the WT site D probe, Abs to USF1 and USF2 strongly inhibited the
formation of DNA-protein complex 1 and also caused a weak supershift
(Fig. 6C). Using the higher affinity E box up probe, a
stronger supershift was observed. Abs to USF1 completely supershifted
complex 1, while Abs to USF2 supershifted most, but not all, of complex
1. No supershift was observed with Abs to IRF1, an unrelated
transcription factor. These results are consistent with the
identification of USF as the protein constituent of complex 1. The USF
appears to be comprised predominantly of heterodimers of USF1 and USF2,
which are supershifted by Abs to either USF1 and USF2, and a lesser
amount of USF1 homodimers, which are supershifted by Abs to USF1, but
not USF2. These results suggest that the increased activity of the E
box up and decreased activity of the E box down promoters was due to
increases and decreases in binding of USF. Abs to c-Myc and its binding
partner Max did not supershift any of the site D complexes (data not
shown). However, since Myc/Max dimers are inherently unstable and
difficult to isolate from cells (38), these results of in
vitro binding assays do not necessarily preclude the binding of Myc/Max
to site D of the human PIGR promoter in vivo.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A direct comparison of the homologous regions of the human and mouse
PIGR promoters demonstrated that the transcriptional
activity of the human PIGR promoter was significantly higher
in both human and mouse cell lines (Fig. 2). Subsequent analyses did
not define a specific region unique to the human PIGR gene
that would explain the species differences in PIGR promoter
activity. We therefore favor a model by which complex interactions
among multiple regulatory sites contribute to the superior activity of
the human PIGR promoter. The tissue specificity of pIgR mRNA
expression varies between humans and rodents (3). Northern
blot analysis of human tissues demonstrated that pIgR mRNA was
expressed at high levels throughout the small and large bowels, but was
undetectable in the liver (41). By contrast, Northern blot
analysis of mouse tissues demonstrated high expression of pIgR mRNA in
both intestine and liver (42). A quantitative RT-PCR
analysis of mouse tissues indicated that expression of pIgR mRNA
(relative to -actin mRNA) was 2- to 3-fold higher in intestine than
in liver (25). In the present study the activities of both
the human and mouse PIGR promoters were 5- to 10-fold higher
in intestinal than in liver cell lines, suggesting that tissue-specific
mechanisms may contribute to the regulation of PIGR
transcription. The higher activity of the mouse PIGR
promoter in intestinal than in liver cell lines is consistent with the
higher endogenous expression of pIgR mRNA in mouse intestine. Although
the activity of the human PIGR promoter was higher in
intestinal than in liver cell lines, our finding of significant human
PIGR promoter activity in liver cell lines is inconsistent
with the lack of expression of pIgR mRNA in human hepatocytes. It is
possible that additional mechanisms may down-regulate steady state pIgR
mRNA levels in human liver, such as the effects of more distal
regulatory regions on PIGR gene transcription,
liver-specific control of pIgR mRNA stability, etc.
The regions designated sites A, B, and C in Fig. 1 represent the most
divergent regions between the human and rodent PIGR
promoters. Deletion of site A caused a modest, but reproducible,
decrease in PIGR promoter activity in both intestinal and
nonintestinal cells (Fig. 3), suggesting that it may act as a positive
regulatory element. By EMSA, we have detected ubiquitous and
intestine-specific proteins that bind to the unique sequence of the
human PIGR site A region (unpublished observations). We have
further observed that point mutations in site A that reduce protein
binding also reduce PIGR promoter activity (unpublished
observations). We have not yet confirmed the identities of the site A
binding proteins, and the sequence of site A does not contain binding
sites for any known transcription factors in the TRANSFAC database
(43). In contrast to site A, site C appears to act as a
negative regulatory element (Fig. 3). The sequences of sites B and C
are highly related, suggesting that site C may have evolved by
duplication and divergence of site B. Both sites contain the motif
AAANGAAANNAAA, which represents a potential binding site for proteins
of the IRF family of transcription factors (20). Despite
the similarity to IFN-stimulated response elements, sites B and C
appear to bind the same set of proteins in the presence and absence of
IFN- stimulation (20 ; in Fig. 5 of that reference, site
B is designated upstream ISRE-1, site C is designated upstream ISRE-2,
and site A is the irrelevant oligo). We are currently characterizing
interactions among proteins that bind to sites A, B, and C.
Since deletion of site D completely abolished the activity of the human
PIGR promoter, we hypothesize that proteins bound to this
region are critical for initiation of transcription. Within this region
is an E box motif, a consensus binding site for transcription factors
of the bHLH-zip family (35). It was previously reported
that point mutations in the E box motif caused a significant decrease
in activity of the human (22) and mouse (24)
PIGR promoters. Furthermore, overexpression of USF1, a
ubiquitously expressed protein of the bHLH-zip family, resulted in a
modest enhancement of the activity of the mouse PIGR
promoter (24). Here we demonstrated that an abundant
nuclear protein complex associated with the PIGR E box in
CaCo2 cells was a heterodimer of USF1 and USF2, with a minor
contribution by USF1 homodimers. Given the critical role of USF in
regulation of PIGR promoter activity, it is informative to
examine the sequence of the PIGR E box in the context of
known requirements for USF binding. Three approaches have been used to
determine the optimal DNA binding site for USF: analysis of USF binding
to E box motifs of varying sequence within active promoters
(35), analysis of USF binding to oligonucleotides of
varying sequence by EMSA (36), and structural analysis of
purified USF bound to its cognate DNA element (44). USF
shares with all members of the bHLH-zip family an affinity for the core
palindromic sequence CACGTG. The DNA-binding surface of USF extends to
contact 2 additional bp on either side of the core E box, with a
preference for maintaining the palindromic structure across the entire
10 bp. While the prototype sequence for bHLH-zip proteins is
RYCACGTGRY, USF shows a clear preference for a T residue in the second
position, thus having an optimal binding site of RTCACGTGAY
(36). The sequence of the human PIGR E box is a
near-perfect match for the USF consensus, with the single exception of
the A in position 4 of the core hexanucleotide. Significantly, the DNA
binding domain of USF has been shown to contact the core sequence
CACATG with a small reduction in binding affinity compared with the
palindromic sequence CACGTG (36). We demonstrated that
mutation of the human PIGR E box to the canonical sequence
CACGTG significantly increased promoter activity (Fig. 4). We
hypothesize that the increase in promoter activity was due to enhanced
binding of USF, as supported by our EMSA data (Fig. 6).
In addition to regulating the affinity of USF binding, the sequence of the E box motif in the human PIGR promoter may influence the availability of this site for binding other members of the bHLH-zip family of transcription factors. Substitution of an A in position 4 of the core E box motif (CACATG) has been shown to significantly diminish the affinity of Myc/Max binding in vitro while causing only a minimal decrease in the affinity of USF binding (36). Similarly, the presence of a T immediately 5' to the core E box (TCACATGA) favored USF binding while diminishing Myc/Max binding. Similar sequence variations have been shown to affect USF vs Myc/Max binding to the E box motif in the cad gene promoter in living cells (45, 46). Significantly, the human PIGR E box contains the sequence TCACATGA, which would be predicted to favor USF over Myc/Max binding. We have recently shown that overexpression of USF1 or USF2 enhances and c-Myc inhibits PIGR promoter activity in intestinal epithelial cells (M. E. C. Bruno and C. S. Kaetzel, manuscript in preparation). This situation is reversed in the case of the cad gene promoter, which is enhanced by c-Myc and inhibited by USF (45, 46). The significance of the opposing effects of USF and c-Myc may relate to their biological roles. USF1 and USF2 have been shown to enhance transcription of genes associated with differentiated cellular function (37), while c-Myc has been shown to enhance the transcription of genes associated with cellular proliferation and oncogenesis, such as the cad gene (35). In this context it is significant that pIgR expression is correlated with differentiation of intestinal epithelial cells and is diminished in disease states associated with proliferation of intestinal epithelial cells, such as colon cancer and inflammatory bowel disease (3).
We identified a consensus binding site for the AP2 transcription
factor that is unique to the human PIGR promoter (Figs. 1 and 5). This element bound recombinant AP2, and mutations that
disrupted AP2 binding caused a significant decrease in PIGR
promoter activity. AP2 expression is known to be induced by several
signaling pathways, including protein kinase C, cAMP, and retinoic acid
(47, 48, 49). AP2 may therefore play a role in both basal and
inducible transcription of the human PIGR gene.
Positive regulation of the PIGR gene by USF may be relevant
to the function of the inverted repeat elements that are highly
conserved among the human and rodent PIGR genes (Fig. 1). We
hypothesized that the inverted repeats could base pair to form a double
stem-loop conformation of the DNA in the proximal promoter region (Fig.
8A). A similar double stem-loop structure has been reported
to inhibit transcription of the human platelet thromboxane receptor
gene (50). In that system binding of the Sp1 transcription
factor to the region between the inverted repeats enhanced
transcription by stabilizing the DNA in a double-stranded configuration
and inhibiting stem-loop formation. In the case of the PIGR
gene we hypothesize that binding of USF to the E box motif, which is
located between the inverted repeats, would inhibit stem-loop
formation. Given the high level of USF binding activity in CaCo2 cells
(Fig. 6), we predict that stem-loop formation would be inhibited
strongly. This model explains the modest effect of mutations in repeat
1 on PIGR transcription (Fig. 8), since stem-loop formation
would already be inhibited by USF binding to the region between the
inverted repeats. Thus, USF may play two roles in regulating the
PIGR promoter: first, by acting as a general
trans-activating factor; and second, by disrupting the
formation of a potentially inhibitory stem-loop structure. The
possibility remains that the inverted repeats may play a role in
inhibiting PIGR promoter activity in cells where USF
activity is low. An alternative role for repeat 2 could involve binding
of transcription factor(s). Mutation of the second inverted repeat
caused a dramatic decrease in PIGR promoter activity, which
was not due to disruptions in the adjacent E box and AP2 motifs (Figs. 4 and 5). DNase I footprinting analysis indicated that nuclear proteins
from CaCo2 cells bound to the repeat 2, but not the repeat 1 motif
(Fig. 3). Because the sequences of repeats 1 and 2 are the same (in
opposite orientations), we conclude that binding of proteins to these
motifs is dependent on context, i.e., interactions with adjacent
elements. The repeat 2 motif is strategically placed between the E box
and AP2 motifs and may bind an as yet unidentified protein in
cooperation with USF and AP2. Significantly, Johansen et al.
(22) identified several protein complexes from epithelial
cell-lines that bound to a DNA probe spanning the E box and repeat 2
motifs of the human PIGR promoter. One of these complexes
was specifically competed by a shorter DNA probe including repeat 2,
but not the E box, and may represent a unique factor that binds to
repeat 2 in cooperation with USF or other E box binding proteins.
In summary, our data suggest a model for transcriptional regulation of
the human PIGR gene that involves cooperative binding among
multiple transcription factors. Factors that regulate basal
transcription may include USF, AP2, a protein(s) that binds the repeat
2 element, and proteins that associate with the regions we have
designated sites A and C. During immune and inflammatory responses,
these basal regulatory factors may interact with inducible factors such
as IRF-1, NF-B, and STAT6 to up-regulate PIGR gene
transcription.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Current address: Department of Oncology-Cancer Biology, The Johns Hopkins University, 464 Cancer Research Building, 1650 Orleans Street, Baltimore, MD 21231.
3 Address correspondence and reprint requests to Dr. Charlotte S. Kaetzel, Department of Pathology and Laboratory Medicine, University of Kentucky, MS117 Medical Center, Lexington, KY 40536. E-mail address: cskaet{at}pop.uky.edu
4 Abbreviations used in this paper: pIgR, polymeric Ig receptor; bHLH-zip, basic helix-loop-helix leucine zipper; CAT, chloramphenicol acyltransferase; DIG, digoxigenin; IRF, IFN regulatory factor; pIg, polymeric Ig; SC, secretory component; S-IgA, secretory IgA; USF, upstream stimulatory factor; WT, wild type.
Received for publication April 2, 2002. Accepted for publication June 11, 2002.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. Mol. Immunol. 34:75.[Medline]
B/Rel site in intron 1 cooperates with proximal promoter elements to mediate TNF--induced transcription of the human polymeric Ig receptor. J. Immunol. 167:6412.B in the expression by tumor necrosis factor- of the human polymeric immunoglobulin receptor (plgR) gene. Immunogenetics 51:289.[Medline]
This article has been cited by other articles:
|
|
 |
|
  L. V. DeSouza, J. Grigull, S. Ghanny, V. Dube, A. D. Romaschin, T. J. Colgan, and K. W. M. Siu Endometrial Carcinoma Biomarker Discovery and Verification Using Differentially Tagged Clinical Samples with Multidimensional Liquid Chromatography and Tandem Mass Spectrometry Mol. Cell. Proteomics, July 1, 2007; 6(7): 1170 - 1182. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Schjerven, P. Brandtzaeg, and F.-E. Johansen Hepatocyte NF-1 and STAT6 Cooperate with Additional DNA-Binding Factors to Activate Transcription of the Human Polymeric Ig Receptor Gene in Response to IL-4 J. Immunol., June 15, 2003; 170(12): 6048 - 6056. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
