| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
MUCOSAL BIOLOGY
and endodermal transcription factors cooperatively activate Fabpl: MODY3 mutations abrogate cooperativity
1Division of Biology and Biomedical Sciences and 2Department of Pediatrics, Washington University School of Medicine, St. Louis, Missouri 63110
Submitted 14 February 2003 ; accepted in final form 11 March 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
plays a central role in intestinal
and hepatic gene regulation and is required for hepatic expression of the
liver fatty acid binding protein gene (Fabpl). An Fabpl
transgene was directly activated through cognate sites by HNF-1 and
HNF-1
, as well as five other endodermal factors: CDX-1, C/EBP,
GATA-4, FoxA2, and HNF-4. HNF-1 activated the Fabpl
transgene by as much as 60-fold greater in the presence of the other five
endodermal factors than in their absence, accounting for up to one-half the
total transgene activation by the group of six factors. This degree of
synergistic interaction suggests that multifactor cooperativity is a critical
determinant of endodermal gene activation by HNF-1. Mutations in
HNF-1 that result in maturity onset diabetes of the young (MODY3)
provide evidence for the in vivo significance of these synergistic
interactions. An R131Q HNF-1 MODY3 mutant exhibits complete loss of
synergistic activation in concert with the other endodermal transcription
factors despite wild-type transactivation ability in their absence.
Furthermore, whereas wild-type HNF-1 exhibited pairwise cooperative
synergy with each of the other five factors, the R131Q mutant could synergize
only with GATA-4 and C/EBP. Selective loss of synergy with other
endodermal transcription factors accompanied by retention of native
transactivation ability in an HNF-1 MODY mutant suggests in vivo
significance for cooperative synergy.
and HNF-1, which share highly homologous
homeodomain DNA-binding motifs that recognize the same DNA sequence
(34,
45). The HNF-1 factors bind to
DNA as heterodimers or homodimers and are found in the liver, kidney, small
intestine, pancreas, and genitourinary tissues
(13). Many genes in these
tissues are activated by HNF-1
(50). Genes activated by
HNF-1 may also be activated by HNF-1, or HNF-1 may inhibit
HNF-1 activation (6).
Mice with targeted disruption of both HNF-1 alleles exhibit Fanconi
syndrome, enlarged fatty liver, defects in bile acid metabolism, and diabetes
(31,
43,
44). Altered expression of
numerous genes occurs in the livers of these animals
(50). Loss of one HNF-1
allele in humans results in renal dysfunction
(26,
35) and pancreatic -cell
defects that give rise to diabetes, termed maturity onset diabetes of the
young (MODY) (66).
MODY is characterized by onset between 10 and 60 years of age, with a
defect in insulin secretion (7,
17). Mutations in HNF-1,
HNF-4, and PDX-1 also result in MODY with phenotype similar to that of
HNF-1 mutations (40).
The exact targets of these transcription factors that result in MODY are
unknown. These transcription factors along with factors of the FTF, FOXA
(formerly HNF-3), GATA, and HNF-6 families comprise a genetic network critical
in endodermal development (16,
42,
70). MODY due to
haploinsufficiency of HNF-1 is termed MODY3, and complete loss of
HNF-1 activity at one allele causes the disease
(64). However, two-thirds of
the 
80 defined HNF-1 gene MODY3 mutations are missense mutations
that result in a full-length protein containing a single amino acid change
(17,
46). Other MODY3 HNF-1
mutations result in early protein truncation, a loss of transactivation
potential, or dominant-negative activity against HNF-1 target genes in
cellular transfection assays
(46). However, some MODY3
missense mutants retain significant ability to transactivate HNF-1 target
genes (63). The defects in
gene regulation resulting from HNF-1 MODY mutations are largely
unknown. In addition to MODY3 HNF-1 mutations that are autosomal
dominant for a severe phenotype, a non-MODY HNF-1 mutation has been
described that is a risk factor for type 2 diabetes but does not result a
dominant phenotype (59).
The rat liver fatty acid binding protein gene (Fabpl) has been
utilized as an experimental model to study gene regulation in endoderm-derived
tissues (10,
52,
56). Rat Fabpl is
highly expressed in hepatocytes and enterocytes, and expression is primarily
regulated at the transcriptional level
(3). A transgene constructed of
Fabpl nucleotides –596 to +21 relative to the start site of
transcription is active in murine hepatocytes, all small intestinal epithelial
cells, renal proximal tubular epithelial cells, and the urinary tract
(47,
52,
56). Two HNF-1 binding sites
were noted in the Fabpl promoter, and mice with targeted disruption
of HNF-1 exhibit complete loss of Fabpl expression in the
liver (1). We found functional
binding sites for five additional endodermal transcription factor families in
the proximal Fabpl promoter and examined the interaction between
HNF-1 and these other endodermal transcription factors to determine why
HNF-1 is essential for Fabpl expression. Experiments
demonstrated that multifactor cooperativity is a critical determinant of
Fabpl activation by HNF-1 and that HNF-1 MODY3
mutations result in loss of multifactor cooperativity but not individual
Fabpl activating ability.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Plasmids. An Fabpl transgene was constructed from Fabpl nucleotides –596 to +21, relative to the start site of transcription, linked to the entire human growth hormone (hGH) gene lacking regulatory sequences. The Fapbl promoter was released from pEPLFABP (56) by cleaving with EcoRI and BamHI and inserted into pBluescript II SK+ cut with the same enzymes to produce pTS9. The entire hGH gene was released from pBShGH (52) by BamHI digestion and ligated into the BamHI site of pTS9 to create pTS10. pTS10 was digested with XbaI and religated, deleting an 18-nucleotide fragment containing the BamHI site distal to the hGH gene and creating pTS154. A glucocorticoid receptor site in the first intron of the hGH gene (39) was destroyed by site-directed mutagenesis to create pTS245. Site-directed mutagenesis was performed with a commercial kit (QuikChange; Stratagene, La Jolla, CA). The mutation changed hGH nucleotides 5268–5269 (GenBank accession no. J03071 [GenBank] ) from TG to GT, using complimentary oligonucleotides with sense strand sequence 5'-CTAAAATCCCTTTGGGCACAATGgtTCCTGAGGGGAGAGGCAGCG-3'. The presence of this mutation and the absence of other mutations were confirmed between the AvrII and BamHI sites of the mutated pTS154. This fragment was released by endonuclease digestion and ligated into pTS154 digested with the same enzymes, yielding PTS245.
Potential transcription factor binding sites in Fabpl were destroyed by site-directed mutagenesis of pTS10 as described above. These sites are indicated in Fig. 1, and the sense sequence from one of each complementary oligonucleotide pair with changed bases after mutagenesis in lower case is: HNF-4 –55 5'-ATCGACAATCACTGAaaTATGGaaTATATTTGAGGAGGAA-3'; Cdx –78/–82 and C/EBP –78 overlapping sites, 5'-GGAGTTAATGTTTGATCCTGGCCATggAGggATCGACAATCACTGACCTATGGCC-3'; FoxA –94, 5'-GACCATTGCTCTCAGGAGTTAATGaTcGAcCCTGGCCATA-3'; HNF-1 –95, 5'-GACCATTGCTCTCAGGAGggccTGTTTGATCCTGGCCATA-3'; GATA –128/–130, 5'-CTTCTGCCTTGCCCATTCTacTTTTTAgtGTTGACCATTGC-3', FOXA –155/–169; 5'-CCTTGATTGGACTCACTAAgGcTTtCTGAATTAGAACAggCTTCTGCC-3'; GATA –229, 5'-ACTCTTATTTCATGAGCGGTacTAAGACACCAAAAATGC-3'; and GATA –557, 5'ACAGCTTTAGGGACTacTAAAATATATGTAAAATTATGT-3'.
|
Targeted mutations were confirmed by sequencing the entire Fabpl sequence and a functional hGH reporter verified by protein production in cultured cells (see below). Promoters with multiple mutations were created by sequential rounds of mutagenesis. Plasmids created by site-directed mutagenesis were termed: pTS146, HNF-1 –95 site mutated; pTS147, HNF-4 –55 site mutated; pTS179, Cdx –78/–82 and C/EBP –78 overlapping sites mutated; pTS187, GATA –128/–130/–229/–557 sites mutated; and pTS211, FoxA –94/–155/–169 sites mutated. The hGH reporter with mutagenized glucocorticoid receptor binding site was released from pTS245 by digestion with BamHI and NotI and was ligated into pTS146, pTS147, pTS179, pTS187, or pTS211 digested with the same enzymes to form pTS255, pTS247, pTS256, pTS246, and pTS248, respectively.
Transcription factor expression plasmids were produced by inserting the
transcription factor coding sequences into pSG5 (Stratagene), a mammalian
expression vector containing the early SV40 promoter. The murine C/EBP
open reading frame was released from MSV/EBP (provided by Steve
McKnight) by digestion with EcoRI and BamHI and inserted
into the EcoRI/BamHI sites in pSG5 to form pTS142. The
murine HNF-1 coding sequence was released from pBJ5-HNF-1
[provided by Peter Traber, Baylor University Medical School, Houston, TX
(67)] with
EcoRI/NotI and ligated into the BamHI site in pSG5
to form pTS156 after blunting the ends of both fragments with DNA
bacteriophage T4 DNA polymerase. The murine HNF-1 coding sequence was
released from pBJ-HNF-1 (provided by Peter Traber) with
EcoRI/EcoRV and ligated into pSG5 digested with
BglII (blunted) then EcoRI to form pTS258. The murine GATA-4
was released from pMT2615A [provided by David Wilson, Washington University,
St. Louis, MO (4)] by
EcoRI digestion and inserted into the EcoRI site of pSG5 to
form pTS186. The rat FoxA2 open reading frame was released from pHNF3
[provided by Robert Costa, University of Illinois, Chicago, IL
(41)] with EcoRI
digestion, and this sequence was inserted into the pSG5 EcoRI site to
form pTS190. The human CDX-1 protein coding sequence was released from pCDX1
[provided by Beatrice Levy-Wilson, Palo Alto Research Foundation, Palo Alto,
CA (30)] with EcoRI
and cloned into the pSG5 EcoRI site to form pTS197. Sequences
containing the coding sequence for human HNF-42 from pHNF-42
[provided by Gerhart Ryffel, Institut für Zellbiologie, Essen, Germany
(15)] was isolated with
HindIII/NotI digestion and cloned into the pSG5
EcoRI site to create pTS276 after blunting both fragments.
Expression constructs for HNF-1 MODY3 mutations were derived by
site-directed mutagenesis of pTS158 using primers as previously described
(63). The entire open reading
frame of each HNF-1 mutant was sequenced to ensure that no additional
mutations were introduced.
Cell culture and transfections. Caco-2 and HepG2 cells were from
American Type Culture Collection (Manassas, VA) and were maintained as
recommended, and HeLa cells were a kind gift from Alan Schwartz. Transient
transfections were performed with calcium phosphate precipitation as follows.
All plasmids utilized in transfection assays were purified with a commercial
kit that yields reduced endotoxin contamination (Qiagen, Valencia, CA). Each
assay contained an Fabpl reporter plasmid, transcription factor
expression plasmids, and plasmids to control for expression efficiency.
Fabpl reporter gene plasmids pTS10 or pTS245 and their mutagenized
derivatives were used interchangeably with equivalent results. Transfection
efficiency was monitored by including identical amounts in each assay of
pSV40-galactosidase (Promega, Madison, WI) or pGL3 (Promega), which
constitutively express bacterial -galactosidase or Photinus
pyralis luciferase, respectively. The amount of DNA (5–9 µg per
well) was kept constant in each experiment by the addition of pSG5 plasmid.
Enough DNA for three wells was diluted with water to a volume of 157.5 µl.
An equal volume of 0.5 M CaCl2 was added, then 315 µl
BES-buffered saline (50 mM BES, 280 mM NaCl, 1.5 mM
Na2HPO4) was added. A precipitate was allowed to form
for 20 min at room temperature before adding the DNA solution to the cells.
Cells were in six-well plates at 30–50% confluence at the time of
transfection, and one-tenth volume (200 µl) of each transfection solution
was added to three separate wells. Cells were washed twice with 2 ml
phosphate-buffered saline after overnight incubation and then covered with the
appropriate culture medium. Culture media were renewed on the following day,
and media and cells were harvested 24 h later. hGH was detected in the media
using a specific radioimmunoassay (Nichols Institute). Dilutions with media
were utilized when necessary to remain in the linear assay range. Transfection
efficiency was assayed by using either a -galactosidase assay kit
(Promega) or luciferase assay kit (Promega). Values were calculated as the
average of the three wells for each DNA solution, and error was calculated as
SD or propagated error for calculated values. Values are reported as fold
activation over the activity of the native Fabpl reporter with no
added transcription factor expression plasmids. All experiments were repeated
at least twice with similar results.
Nuclear extract preparation and electrophoretic mobility shift
assays. Nuclear extracts were prepared from Caco-2 cells after
transfection with transcription factor expression plasmids. Caco-2 cells at
40% confluence in a single 75-cm2 flask were transfected by calcium
phosphate precipitation exactly as for the expression studies, except that
precipitates of 48 µg of either pTS158 (expressing HNF-1), pTS156
(expressing HNF-1), or pSG5 in a larger volume were utilized.
Transfection efficiency was monitored by inclusion of 16 µg pXGH5 (Nichols
Institute). pXGH5 expresses hGH from the metallothionine promoter. Nuclear
extracts were prepared by using a commercial kit (NE-PER kit; Pierce,
Rockford, IL). Approximately 50 µl of packed cells were obtained from each
flask. Nuclear extract protein concentration was determined with a commercial
protein assay kit (Bio-Rad, Hercules, CA).
EMSAs were performed as previously described
(51), except 1.5 µg nuclear
extract, 1 µg herring sperm DNA, and no dIdC were included in each 20-µl
reaction. Reactions were incubated 18 min at room temperature before
electrophoresis, except that supershift assays contained 2 µg antibody
(Santa Cruz Biotechnology, Santa Cruz, CA) and were incubated for 32 min. The
radiolabeled probe was a double-stranded oligonucleotide derived from the
putative Fabpl HNF-1 recognition sequence shown in
Fig. 3A. Competitors
were either a double-stranded oligonucleotide with the mutagenized
Fabpl sequence noted in Fig.
3 or an authentic -fibrinogen HNF-1 site
(11).
|
|
RESULTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Endodermal transcription factors directly activate Fabpl through
interactions with cognate sites. Transient transfection assays were
utilized to determine the potential function of the transcription factor
binding sites identified through Fabpl promoter sequence analysis. A
transgene was constructed from rat Fabpl nucleotides –596 to
+21 linked to a reporter consisting of the entire hGH gene minus its
regulatory regions. The Fabpl transgene was active when transfected
into Caco-2 cells or HepG2 cells (Fig.
2, "native + control" in all panels). These cell lines
were chosen to resemble enterocytes and hepatocytes, respectively. The
potential of transcription factor families with cognate binding sites in the
Fabpl promoter to transactivate the Fabpl transgene was
assessed by cotransfection with an expression plasmid for one transcription
factor from each family. CDX-1, C/EBP, FoxA2, GATA-4, HNF-1, and
HNF-4 all activate the native Fabpl transgene
(Fig. 2). To demonstrate that
activation by these factors was direct and through their cognate sites,
transgenes were created with all sites for each factor family mutagenized to
destroy binding. For example, to test for indirect activation by FoxA2, a
transgene was created with all three FoxA sites mutagenized
(Fig. 1). This mutagenized
transgene was active in both cell lines but was not stimulated by FoxA2
(compare mutant control and mutant + factor in
Fig. 2). This result indicates
the Fabpl transgene activation by FoxA2 is mediated by interaction
with these three sites. Similar transgenes were created to test for indirect
activation by each other factor family. Mutagenesis of potential binding sites
essentially eliminated activation by every factor except C/EBP, which
displayed reduced activity. A second potential HNF-1 binding site in the
Fabpl promoter has been described at –355
(1), but this site did not
mediate Fabpl transgene transactivation (data not shown). These
results indicate that the six factor families with binding site(s) in the
Fabpl promoter activate Fabpl through interaction with their
cognate site(s).
|
HNF-1 and HNF-1 interact with the
cognate Fabpl binding sequence in vitro. HNF-1 has been proposed
as a critical regulator of intestinal epithelial gene regulation
(57) and to directly regulate
Fabpl in hepatocytes
(1). However, because the
proposed Fabpl HNF-1 binding site differs from the consensus at one
nucleotide (Fabpl –97 is G not A;
Fig. 1), interaction of
HNF-1 and HNF-1 with the Fabpl site was tested in vitro.
Nuclear extracts were prepared from Caco-2 cells transfected with either an
expression construct for HNF-1, an expression construct for
HNF-1, or a control construct. EMSAs were performed with these nuclear
extracts and a radiolabeled double-stranded oligonucleotide probe derived from
the putative Fabpl HNF-1 recognition site
(Fig. 3A). Three
specific complexes formed with extracts from Caco-2 cells transfected with the
control vector (Fig.
3B). Formation of these complexes was prevented by
competition with a 128-fold molar excess of unlabeled probe or a 128-fold
molar excess of an oligonucleotide with the sequence of an authentic HNF-1
binding site from the -fibrinogen promoter
(11). Complex formation was
not affected by inclusion of a 128-fold molar excess of the mutagenized
Fabpl HNF-1 binding site that lacked activity in the transient
transfection assay. The three specific complexes that form between Caco-2
extracts and the Fabpl binding site have mobilities similar to those
identified for HNF-1/HNF-1 homo- and heterodimers
(45). The slowest-moving
complex is the HNF-1 homodimer, the fastest-moving complex is the
HNF-1 homodimer, and the middle complex is the heterodimer. These EMSA
with nuclear extracts from cells transfected with HNF-1 or HNF-1
expression constructs demonstrated that abundant binding protein is produced
with each expression construct in Caco-2 cells and is consistent with the
complex identification for extracts from cells transfected with the control
vector. Supershift EMSA confirmed the identity of the complexes
(Fig. 3C). These
experiments demonstrate that the Fabpl HNF-1 site readily forms
complexes with HNF-1 and HNF-1 despite differing from the
consensus sequence.
HNF-1 and HNF-1 transactivate Fabpl
individually and together without interference or cooperation. Because
HNF-1 and HNF-1 both bind to the cognate Fabpl site in
vitro, the transactivation potential of both factors was determined in
transient transfection assays (Fig. 4,
A and B). Both HNF-1 and HNF-1
transactivated the Fabpl transgene in both cell lines, and this
activation was eliminated by specific mutagenesis of the HNF-1 binding site.
Transfection with 2 µg of expression vector for both factors resulted in
greater activation by HNF-1 than HNF-1 in both cell lines. Adding
4 µg expression plasmid resulted in significant activation by HNF-1
in both cell lines (Fig. 4, C and
D). Because HNF-1 has been reported to interfere
with HNF-1 transactivation of other genes
(6,
22,
54), the interaction of
HNF-1 and HNF-1 in Fabpl transgene activation was
determined. Transient transfections were performed with various ratios of
expression plasmids for each factor (Fig.
4, C and D). Fabpl activation with any
combination of HNF-1 and HNF-1 in both cell lines exhibited no
interference or cooperativity.
|
Endodermal transcription factors exhibit strong cooperative
synergy. HNF-1 or HNF-1 Fabpl transgene activation
was assayed in concert with a mixture of five transcription factors,
consisting of one member from each of the other five endodermal transcription
factor families with functional binding sites in the Fabpl promoter
(Fig. 5). HNF-1 by
itself stimulated the Fabpl transgene eightfold in Caco-2 cells. A
mixture of HNF-1 plus the other five endodermal transcription factors
stimulated the transgene 157-fold, whereas the five-factor mix alone
stimulated the transgene 68-fold. Thus addition of HNF-1 to the five
factors resulted in an 89-fold increase in transgene activity relative to the
unstimulated activity. These results reveal that HNF-1 activated the
transgene in Caco-2 cells 11 times better in the presence of the other factors
than by itself (89- vs. 8-fold). The result of cooperation between
HNF-1 with the five-factor group resulted in transgene activation
2.1-fold compared with activation by the five factors together plus activation
of HNF-1 by itself. Similar results were obtained in HepG2 cells in
which HNF-1 activated the Fabpl transgene 187-fold in the
presence of the five factors vs. threefold by itself. Cooperative contribution
of HNF-1 to the factor mix in HepG2 cells is 1.3-fold. Activation with
the mixture of five factors plus HNF-1 activated the transgene 157-fold
in Caco-2 cells, compared with a calculated additive value of 23-fold.
Calculated additive values were derived from the sum of the values obtained
for transgene activation by the individual factors in the mix. In HepG2 cells,
five factors plus HNF-1 activated the transgene 740-fold vs. a
calculated additive value of 67-fold relative to the unstimulated transgene.
These results demonstrate that cooperative synergy among all the factors is
quantitatively more important for Fabpl gene expression in these
cells than activation by any single factor. This degree of synergistic
interaction suggests that multifactor cooperativity is a critical determinant
of endodermal gene activation by HNF-1. In contrast to HNF-1,
HNF-1 obstructed Fabpl activation by the other five endodermal
factors in HepG2 cells and to a lesser extent in Caco-2 cells
(Fig. 5).
|
Pairwise cooperative interaction between the HNF-1 factors and each of the
other five endodermal transcription factors were evaluated
(Fig. 6). The actual activation
of the Fabpl transgene by each factor pair together was compared with
the calculated additive value for transgene stimulation by each factor
separately. Cooperative interactions of twofold or greater were observed
between HNF-1 and CDX-1, C/EBP, GATA-4, FoxA2, and HNF-4
in Caco-2 cells relative to the unstimulated transgene, consistent with the
extensive cooperative interaction between HNF-1 and these factors as a
mixture (Fig. 5). In HepG2
cells, HNF-1 exhibited significant cooperative activation with all
factors except FoxA2. In contrast, HNF-1 did not exhibit significant
synergy with any factor in Caco-2 cells
(Fig. 6). In HepG2 cells,
HNF-1 had greater than twofold cooperative activation with C/EBP
and GATA-4 but significant anergy with HNF-4. Lack of pairwise
interactions between HNF-1 and other factors is consistent with the lack
of cooperative interaction between HNF-1 and the factors as a group
(Fig. 5).
|
HNF-1 MODY3 mutants exhibit a selective defect in
cooperative activation with other endodermal transcription factors.
HNF-1 MODY3 point mutations have been reported that result in a
full-length protein with significant transactivation ability and no
dominant-negative activity
(63). Because cooperative
multifactor interactions are more important in target gene activation than the
action of any one factor (Fig.
5), HNF-1 MODY mutants with significant individual
activation ability were examined for defects in cooperative interactions. Five
of 10 HNF-1 MODY3 mutations examined in the original report retained
significant transactivation ability for a synthetic target gene consisting of
HNF-1 binding sites upstream of a minimal promoter
(63). Each of these five
mutants (Y122C, R131Q, R159Q, K205Q, R272H) was able to transactivate the
Fabpl transgene to varying degrees in Caco-2 and HepG2 cells (data
not shown). In contrast, these five HNF-1 MODY3 mutants exhibited loss
of cooperative synergy with the group of five endodermal transcription factors
in both cell lines (data not shown). Two MODY3 mutants, R131Q and Y122C, were
particularly informative. Both of these mutants are in the HNF-1 DNA
binding domain, and both proteins localize to the nucleus and form complexes
with a canonical HNF-1 binding site in EMSA despite a reduced halflife
(63). The R131Q MODY3 mutant
retained wild-type transactivation ability for the Fabpl transgene in
HeLa cells, and the Y122C mutant retained 63% of the wild-type activity
(Fig. 7A). In HepG2
cells, both mutants showed significantly less ability to transactivate the
Fabpl transgene (Fig.
7B). In the presence of the five-factor mix, wild-type
HNF-1 shows a dramatic cooperative synergy (Figs.
5 and
7C). Neither MODY3
mutant exhibited cooperative synergy with the five other factors, but both
actually inhibited activation by the five factors
(Fig. 7C). The G319S
HNF-1 mutant does not result in MODY or a dominant phenotype
(59) and has wild-type ability
to transactivate the Fabpl transgene in HeLa or HepG2 cells
(Fig. 7, A and
B). Furthermore, this mutant did not display a defect in
cooperative synergistic activation with the other endodermal factors
(Fig. 7C). HepG2 cells
are known to endogenously express many endodermal transcription factors,
whereas HeLa cells do not
(14). The decrease in
Fabpl activation in HepG2 cells compared with HeLa cells only in
those factors deficient in cooperative synergy may reflect a contribution of
synergy by the endogenous factors. These HNF-1 MODY3 mutants exhibited
defects in specific pairwise interactions with the other endodermal
transcription factors (Fig.
7D), whereas the non-MODY G319S mutant did not. The
wild-type HNF-1 exhibited synergy with CDX-1, C/EBP, GATA-4, and
HNF-4 in HepG2 cells, whereas the R131Q MODY3 HNF-1 mutant lost
cooperative activation with CDX-1 and HNF-4 and exhibited anergy with
FoxA2. The Y122C MODY3 HNF-1 lost cooperative activation with CDX-1,
C/EBP, and HNF-4 and exhibited anergy with FoxA2. The G319S
displayed the same pairwise cooperative interactions as the wild-type
HNF-1. Similar results were obtained for the R131Q and Y122C mutants
with Caco-2 cells and HeLa cells with the five-factor mix (data not
shown).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
exhibits extensive cooperative synergy with other
endodermal transcription factors. The target gene is activated up to 60 times
greater by HNF-1 in the presence of the other factors than in their
absence, with HNF-1 contributing up to one-half the total activation.
Hepatocytes contain numerous transcription factors capable of transactivating
Fabpl, including C/EBP, FOXA, GATA, and HNF-4 family factors
(12), but mice with
HNF-1 gene null mutations exhibit complete loss of Fabpl
expression in the liver (1).
The lack of Fabpl expression in the liver of HNF-1-null mice
despite the presence of the other activating factors is consistent with a
central role for HNF-1 in multifactor synergy.
Pairwise cooperative interactions occur between HNF-1 and each of
the other endodermal transcription factors that activate the Fabpl
transgene. These multiple cooperative interactions between pairs of factors
may combine to yield the observed multifactor cooperative synergy.
HNF-1 has been reported to interact pairwise to cooperatively activate
genes besides Fabpl with these same factors: FoxA2
(9,
61), C/EBP
(8,
68,
69), and HNF-4
(23). Cooperative activation
is dependent, at least in part, on the target gene, because no synergy in
activation between HNF-1 and HNF-4 is observed with some targets
(33) and actual anergy occurs
with other targets (28,
29). HNF-1
cooperatively interacts with GATA-4 and Cdx-2 to activate the
sucrase-isomaltase gene (5).
HNF-1 has also been reported to exhibit pairwise synergy to activate
genes besides Fabpl with other factor family members, including FoxA3
(61), GATA-4
(62), GATA-5
(27), and Cdx2
(38), plus additional factors
important in endoderm: DBP (2),
HNF-6 (20), Oct-1
(24), COUP-TF
(32), and HOXC11
(37). The mechanism for most
of these synergies is unknown, but protein-protein interaction has been
described between HNF-1 and Cdx-2
(21) or GATA-5
(27). Synergistic gene
activation through recruitment of multiple factors has been hypothesized to
result from increased efficiency of assembly of a competent RNA polymerase II
initiation complex through multiple mechanisms
(36). We observed results
similar to those shown for HepG2 cells in
Fig. 7C in HeLa cells,
indicating that the mechanism of cooperative synergy is not unique to
endodermal cells.
Direct relevance for the significance of cooperative synergy in vivo is
obtained from experiments with the MODY3 HNF-1 mutations. MODY3 is an
autosomal dominant disease in which loss of one copy of the HNF-1 gene
is sufficient to disturb pancreatic gene expression
(17). It is perhaps surprising
that haploin-sufficiency of one transcription factor results in disease when
genes are typically activated by numerous factors. Haploinsufficiency of
transcription factors frequently results in disease, and it has been suggested
this may be due to loss of transcriptional synergy
(65). The R131Q MODY3 mutation
exhibits wild-type target gene activation alone but results in a severe MODY
phenotype with average age of onset at 14 years
(7,
18,
25). This phenotype compares
with an average age of onset of 24 yr for all MODY3 mutations
(49), indicating that
selective loss of synergy results in a disease at least as severe as that
caused by other mutations. The R131Q mutant activates the Fabpl
target gene at least as well as wild type in HeLa cells but only 47% in HepG2
cells. This difference can be explained by cooperative synergy between
HNF-1 and endogenous endodermal transcription factors present in the
HepG2 but not HeLa cells
(14).
The R131Q mutant and to a greater extent the Y122C mutant inhibited target
gene activation by the other five transcription factors
(Fig. 7C). R131Q has
wild-type ability to transactivate the Fabpl transgene, and Y122C
retains 63% of the wild-type target gene transactivation
(Fig. 7A). The
decrease in Y122C target gene transactivation may be a result of the reported
decreased stability of these mutants compared with the wild-type protein
(63). However, the significant
inhibition of Fabpl activation by the five factors in the presence of
Y122C is difficult to attribute solely to a decrease in stability. It is
interesting that a loss of cooperation with some but not all of the five
factors tested with the HNF-1 mutants results in a complete loss of
synergy with the entire group (Fig. 7,
C and D). These results could explain the in
vivo MODY phenotype in which the numerous interactions that might occur on the
target gene promoters could be disrupted by loss of a few critical
interactions. The G319S does not have any loss of cooperative interactions
compared with wild-type HNF-1 and does not lead to the autosomal
dominant phenotype observed with the MODY mutants but rather to a more subtle
phenotype that manifests as a risk factor for type 2 diabetes
(59).
We describe a MODY mutation that results in loss of interaction with
multiple endodermal transcription factors. Another MODY mutation has been
described that results in loss of pairwise synergy between HNF-4 E276Q
MODY1 and COUP-TFII to activate the HNF-1 promoter
(55). This finding suggests
that other MODY mutations may also function through loss of synergy and that
this effect may be amplified through the genetic network of endodermal
transcription factors. All the MODY transcription factors are constituents of
a genetic network for transcriptional activation in endoderm that also
encompasses FTF, FOXA, GATA, and HNF-6
(16,
19,
42,
70). Each transcription factor
may activate the gene for another factor and/or autoactivate its own gene.
Thus loss of one transcription factor may lead to endodermal defects through
failure to activate other factors. Synergistic activation of the transcription
factor genes would amplify the loss of any one factor and contribute to the
loss of direct activation of target genes in differentiated tissue.
|
ACKNOWLEDGMENTS |
|---|
This work was supported by grants from the March of Dimes Foundation (to T. C. Simon) and National Institute of Diabetes and Digestive and Kidney Diseases Grants DK-56361 (to T. C. Simon) and P30-DK-52574 (to the Washington University Digestive Disease Research Core Center).
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(HNF-1). Alterations
in fatty acid homeostasis in HNF-1-deficient mice. J Biol
Chem 275:
27117–27122, 2000., GATA-4, and
caudal related homeodomain protein Cdx2 interact functionally to modulate
intestinal gene transcription. Implication for the developmental regulation of
the sucrase-isomaltase gene. J Biol Chem
277: 31909–31917,
2002. to HNF-1. J Biol Chem
276: 32122–32128,
2001.
and 3. J Biol Chem 275:
18358–18365, 2000.1-antitrypsin promoters. Science
238: 688–692,
1987.) mutations in maturity-onset
diabetes of the young. Hum Mutat
16: 377–385,
2000.[Web of Science][Medline]
-cell genes and diabetes:
molecular and clinical characterization of mutations in transcription factors.
Diabetes 50:
S94–S100, 2001. gene expression. Mol Cell Biol
21: 7320–7330,
2001. regulates transcription of the guanylin gene. Am J
Physiol Gastrointest Liver Physiol 273:
G833–G841, 1997.1-antitrypsin gene
expression in human intestinal epithelial cell line Caco-2 by HNF-1 and
HNF-4. Am J Physiol Gastrointest Liver Physiol
276: G1181–G1194,
1999.-mediated activation of the human UDP
glucuronosyltransferase 2B7 promoter. Mol Pharmacol
57: 940–947,
2000./MODY3 gene in Japanese subjects with early- and late-onset
NIDDM. Diabetes 46:
1504–1508, 1997.[Abstract]
. Am J Physiol Gastrointest Liver Physiol
281: G69–G84,
2001. knockout mouse. Mol Cell
Biol 18:
3059–3068, 1998. and
HNF-1 (vHNF-1) share dimerization and homeo domains, but not activation
domains, and form heterodimers in vitro. Genes Dev
5: 1042–1056,
1991. (HNF-1) gene. Diabet
Med 15:
816–820, 1998.[Web of Science][Medline]
-dependent transcription. J Biol Chem
273: 13297–13306,
1998. mediates expression of the
lactase-phlorizin hydrolase gene. Biochem J
346: 529–535,
2000.[Web of Science][Medline]
contains two transcriptional activation domains, one of which
is novel and conserved with the Drosophila fork head protein.
Mol Cell Biol 12:
3723–3732, 1992., HNF-4, and
HNF-1 gene promoters. J Biol Chem
276: 13136–13144,
2001.-deficient
mice. J Clin Invest 101:
2215–2222, 1998.[Web of Science][Medline]
is an essential regulator of bile acid and plasma cholesterol
metabolism. Nat Genet 27:
375–382, 2001.[Web of Science][Medline]
and 1. J Biol Chem 273:
28408–28418, 1998. found in type 1 maturity-onset
diabetes of the young: impaired synergy with chicken ovalbumin upstream
promoter transcription factor II on the hepatocyte nuclear factor-1 promoter.
Diabetes 48:
1162–1167, 1999.[Abstract]
G319S, a transactivation-deficient mutant, is associated with
altered dynamics of diabetes onset in an Oji-Cree community. Proc
Natl Acad Sci USA 99:
4614–4619, 2002.,
HNF-3, and NF-I contributes to the activation of the liver-specific protein C
gene. DNA Cell Biol 16:
569–577, 1997.[Web of Science][Medline]
results in
synergistic activation of the human lactase-phlorizin hydrolase promoter.
J Biol Chem 277:
27659–27667, 2002.
(HNF-1) and HNF-1 regulate transcription via two elements in an
intestine-specific promoter. J Biol Chem
269: 17080–17085,
1994.This article has been cited by other articles:
|
|
 |
|
  M. Yu, J. Wang, W. Li, Y. Z. Yuan, C. Y. Li, X. H. Qian, W. X. Xu, Y. Q. Zhan, and X. M. Yang Proteomic screen defines the hepatocyte nuclear factor 1{alpha}-binding partners and identifies HMGB1 as a new cofactor of HNF1{alpha} Nucleic Acids Res., March 27, 2008; 36(4): 1209 - 1219. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. J. Jeong, H.-S. Lee, K.-S. Kim, Y.-K. Kim, D. Yoon, and S. W. Park Sterol-dependent regulation of proprotein convertase subtilisin/kexin type 9 expression by sterol-regulatory element binding protein-2 J. Lipid Res., February 1, 2008; 49(2): 399 - 409. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. S. Belaguli, M. Zhang, M. Rigi, M. Aftab, and D. H. Berger Cooperation between GATA4 and TGF-beta signaling regulates intestinal epithelial gene expression Am J Physiol Gastrointest Liver Physiol, June 1, 2007; 292(6): G1520 - G1533. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Bosse, J. J. Fialkovich, C. M. Piaseckyj, E. Beuling, H. Broekman, R. J. Grand, R. K. Montgomery, and S. D. Krasinski Gata4 and Hnf1{alpha} are partially required for the expression of specific intestinal genes during development Am J Physiol Gastrointest Liver Physiol, May 1, 2007; 292(5): G1302 - G1314. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. K. Divine, L. J. Staloch, H. Haveri, C. W. Rowley, M. Heikinheimo, and T. C. Simon Cooperative interactions among intestinal GATA factors in activating the rat liver fatty acid binding protein gene Am J Physiol Gastrointest Liver Physiol, August 1, 2006; 291(2): G297 - G306. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Bosse, H. M. van Wering, M. Gielen, L. N. Dowling, J. J. Fialkovich, C. M. Piaseckyj, F. J. Gonzalez, T. E. Akiyama, R. K. Montgomery, R. J. Grand, et al. Hepatocyte nuclear factor-1{alpha} is required for expression but dispensable for histone acetylation of the lactase-phlorizin hydrolase gene in vivo Am J Physiol Gastrointest Liver Physiol, May 1, 2006; 290(5): G1016 - G1024. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. W. Rowley, L. J. Staloch, J. K. Divine, S. P. McCaul, and T. C. Simon Mechanisms of mutual functional interactions between HNF-4{alpha} and HNF-1{alpha} revealed by mutations that cause maturity onset diabetes of the young Am J Physiol Gastrointest Liver Physiol, March 1, 2006; 290(3): G466 - G475. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. K. Divine, L. J. Staloch, H. Haveri, C. M. Jacobsen, D. B. Wilson, M. Heikinheimo, and T. C. Simon GATA-4, GATA-5, and GATA-6 activate the rat liver fatty acid binding protein gene in concert with HNF-1{alpha} Am J Physiol Gastrointest Liver Physiol, November 1, 2004; 287(5): G1086 - G1099. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. M. van Wering, T. Bosse, A. Musters, E. de Jong, N. de Jong, C. E. Hogen Esch, F. Boudreau, G. P. Swain, L. N. Dowling, R. K. Montgomery, et al. Complex regulation of the lactase-phlorizin hydrolase promoter by GATA-4 Am J Physiol Gastrointest Liver Physiol, October 1, 2004; 287(4): G899 - G909. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. J. Oesterreicher and S. J. Henning Rapid induction of GATA transcription factors in developing mouse intestine following glucocorticoid administration Am J Physiol Gastrointest Liver Physiol, June 1, 2004; 286(6): G947 - G953. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
