| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Internal Medicine (N.J.K.) Division of
Nephrology, Bone and Mineral Metabolism University of Kentucky
Medical Center Lexington, Kentucky 40536-0084
Department
of Medicine (S.A., J.R.) Albert Einstein College of Medicine
Bronx, New York 10021
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
A mutant produced an attenuated negative
transcriptional response while the GC mutant resulted in a
reproducibly weak positive transcriptional outcome. The double mutant,
however, yielded a 4-fold increase in transcription, similar to the
7-fold increase observed from an analogous construct using the human
osteocalcin VDRE. UV light cross-linking to gapped oligonucleotides
assessed the polarity of heterodimer binding to the wild-type and
double mutant sequences and was consistent with the vitamin D receptor
preferentially binding to the 5'-half of both elements. Finally, DNA
affinity chromatography was used to immobilize heterodimer complexes
bound to the wild-type and double mutant sequences as bait to identify
proteins that may preferentially interact with these DNA-bound
heterodimers. This analysis revealed the presence of a p160 protein
that specifically interacted with the heterodimer bound to the
wild-type VDRE, but was absent from complexes bound to response
elements associated with positive transcriptional activity. Thus, the
sequence of the individual VDRE appears to play an active role in
dictating transcriptional responses that may be mediated by
altering the ability of a vitamin D receptor heterodimer to interact
with accessory factor proteins. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
The hormone, however, also negatively regulates expression of several gene products (23, 24, 25, 26, 27), including direct transcriptional effects on PTH gene expression (28, 29, 30, 31, 32). Response elements were identified that were in close proximity to the promoters of the avian and human PTH genes (28, 29). The avian PTH (aPTH) VDRE followed the classical DR+3 format while the hPTH VDRE appeared to function as a single half-site and may involve RXR-independent binding (33). In vitro DNA binding to the aPTH VDRE was shown to require heterodimerization of recombinant VDR with RXR-containing extracts, although interference footprinting results revealed only weak interactions over the two half-sites (29). As part of the same study, transient transfection experiments utilizing the aPTH VDRE demonstrated the element’s ability to down-regulate gene expression in response to hormone, either within the context of its natural position in the aPTH promoter or when linked to a heterologous promoter.
Negative gene regulation by other members of the steroid hormone receptor superfamily has also been described. Negative regulation of the bovine PRL gene by glucocorticoids was shown to occur by direct interaction of the glucocorticoid receptor with a negative glucocorticoid response element (GRE) that bore only a modest resemblance to a consensus GRE derived from positive regulatory elements (34, 35). Interestingly, when two mutations were introduced into this negative element, the transcriptional activity was transformed from negative to positive in response to hormone administration. That data, combined with additional evidence that GREs may act as allosteric regulators of glucocorticoid-mediated gene transcription (36, 37), give credence to the notion that individual DNA sequences may play an active role in dictating transcriptional responses.
As part of continuing studies on the aPTH gene, we noted that
replacement of the wild-type negative VDRE in the aPTH promoter with
the positively responding human osteocalcin (hOC) VDRE resulted in a
strong positive transcriptional response to
1,25-(OH)2D3 administration in transient
transfection studies (see Table 1
). This
indicated that the aPTH promoter could support both negative or
positive transcriptional outcomes and suggested that the DNA sequence
of the VDRE may actively participate in dictating the observed
response. Given these preliminary observations, we systematically
investigated the capacity to alter the negative transcriptional
potential of the aPTH VDRE by directed mutagenesis of the wild-type
element within the context of its natural promoter.
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
). These positions are largely
restricted to cytosine (+6) and adenine (+7) residues in the majority
of positive VDREs, but are relegated to guanine and thymidine residues
in the wild-type aPTH element. Thus, in the present study, mutations
were introduced into these positions comprised of single conversions of
GC at +6 (cPTH) and TA at the +7 position (mPTH) because of the
aforementioned high conservation of those nucleotides in positive DR+3
VDREs. A double mutant (dmPTH) was also created by changing the GT at
positions +6/+7 to CA to generate a perfect direct repeat of the GGGTCA
hexanucleotide half-site, similar to a sequence previously shown to be
an enhancer of vitamin D actions (7). Notably, this represents a
conversion from a purine/pyrimidine to a pyrimidine/purine dinucleotide
ending.
|
(rhRXR)
extracts (Fig. 2). No binding was
observed by either extract alone (lanes 1 and 2), and while binding was
observed in the absence of hormone, addition of
1,25-(OH)2D3 resulted in a 2-fold stimulation
of binding under these conditions (lanes 3 and 4). The anti-VDR
antibody, 9A7
, was able to block the appearance of the bound band
while generating a minor supershifted complex, and inclusion of an
anti-RXR antibody was able to supershift the complex indicating that
both proteins were part of the bound complex (lanes 5 and 11).
Competition with an excess of unlabeled aPTH or hOC double-stranded
oligonucleotides prevented observation of the bound radiolabeled
complex, while the perfect estrogen response element from the chicken
vitellogenin II (cvitERE) gene had no effect on the complex (lanes
7–9). Analogous results were obtained using radiolabeled probes for
each of the other mutant aPTH VDRE sequences (data not shown),
indicating that the mixture of rhVDR/rhRXR was required and bound to
the these sequences in a specific manner.
|
). Based on an analysis of the
half-maximal competition curves, the data indicate that the dmPTH
oligonucleotide was approximately 3 to 4 times more effective at
competing for heterodimer binding to the aPTH probe than either the
wild-type sequence or the cPTH mutant oligonucleotide. The mPTH
sequence was the least effective competitor of these four
oligonucleotides in this analysis. Furthermore, additional mutant
oligonucleotides were evaluated that replaced highly conserved guanine
dinucleotide positions in the 5'-half of the wild-type element (5'Mut,
positions -5 and -6, GGCC) or in the 3'-half of the element
(3'Mut, positions +3 and +4, GGCC). Neither of these latter mutants
was effective at competing for binding by the heterodimer, although the
3'Mut, which retained the intact GGGTCA 5'-half-site, produced a modest
reduction in the bound band at the highest tested concentration. This
indicates the importance of both half-sites in binding by the
heterodimeric complex.
|
). Weak interference was observed over
the modified guanines at positions -7 to -5 and +2 to +4. On the
opposite strand (data not shown), weak interference was again seen over
the TGA sequence at -2' to -4', with even less pronounced
interactions over the ACA at +5' to +7'.
|
).
Notably, interactions were again evident at the +2 to +4 guanine
positions, but they now displayed a measurable increase in the degree
of interference, with the strongest occurring at the +2 and +3
positions. The opposite strand exhibited a slight gain in interference
over the +5' to +7' positions (data not shown). The strands of the
other single mutant, cPTH, produced a unique footprint pattern that was
now significantly altered from the previous observations (Fig. 4c).
Interference over the guanine triplet at +2 to +4 was again evident;
however, very strong interference was now observed over the modified
guanine at +3, with a more modest gain in interference evident at the
+4 guanine. Significant increases in interference were also observed
over the -7 to -5 guanine triplet in the 5'-half of the element as
well as additional degrees of contact over the corresponding AGA at +5'
to +7' on the opposite strand (data not shown).
The interference footprint of the double mutant deviated significantly
from the previously examined wild-type and single-mutant sequences
(Fig. 4d). Strong, nearly complete interference was observed over the
guanine residues at +2/+3, and as noted for the cPTH mutant, there was
significant interference over the guanines at -5/-6. In addition,
interference in the opposite strand was now strongly evident over the
pair of TGA sequences at +5' to +7' and -2' to -4' (data not shown).
The overall pattern of interference for the dmPTH was similar in nature
to the strong interference seen for the osteopontin (11) and
osteocalcin VDREs (N. J. Koszewski, unpublished observations),
both of which are known to be strong positive regulators of vitamin
D-stimulated gene transcription (2, 3, 5). A densitometric analysis is
presented in Fig. 4e to depict the changes that occurred in the
footprint patterns over the 3'-halves of each of the sequences. These
range from a very modest degree of interference observed as a result of
the heterodimer complex binding to the wild-type aPTH VDRE to the very
strong, almost complete, interference observed with the dmPTH sequence.
As an initial point of reference, a control cleavage pattern generated
from a portion of the chemically modified aPTH VDRE probe not used in
the separation analysis is also included in each graph (dashed
line).
Ribonuclease protection assays were next used to assess the
transcriptional response to administration of
1,25-(OH)2D3 to OK cells transfected with
reporter constructs containing wild-type or mutant VDREs upstream from
the aPTH promoter linked to the chloramphenicol acetyl transferase
(CAT) gene (Table 1). In transfection studies using plasmid constructs
containing the wild-type aPTH VDRE, treatment with
1,25-(OH)2D3 inhibited CAT mRNA transcripts by
54%, similar to earlier observations using CAT enzyme activity itself
to assess gene expression (29). Substitution of the wild-type sequence
with the mPTH VDRE resulted in approximately 50% attenuation of the
inhibitory effect of 1,25-(OH)2D3 (-54%
vs. -28%). More interestingly, the single mutation of
GC at position +6 (cPTH) now resulted in a modest, but highly
reproducible, increase (+21%) of CAT gene transcripts in response to
hormone. Finally, transfection with the plasmid construct containing
the dmPTH sequence resulted in a dramatic increase (
4-fold) in CAT
gene transcription, similar to the 7-fold increase observed with the
prototypical positive hOC VDRE used as a positive control in these
experiments.
Selected mutations within the wild-type aPTH response element were able to alter the points of DNA contact as delineated by the interference assay in conjunction with a graded response from negative to positive in hormone-stimulated transcriptional activity. However, heterodimeric interactions by the RXR/VDR complex with a direct repeat response element implies an order or polarity of the individual proteins with respect to DNA binding. Previous studies demonstrated that RXR occupies the 5'-half and VDR the 3'- half of VDREs isolated from genes that are up-regulated by the vitamin (20, 21, 22, 38). In light of the negative regulation imparted by the wild-type aPTH sequence, and the clearly positive response obtained with the dmPTH element, we sought to ascertain whether this might not be the result of an altered polarity of the proteins binding to these elements. Using a strategy of gapped oligonucleotides and UV light cross-linking (39), proteins in contact with the 5'- and 3'-halves of both of these VDREs were analyzed.
As noted in Fig. 5, cross-linking to
either half of both the aPTH and dmPTH sequences gave similar results.
Control cross-linking to either 5' sequence yielded two prominent bands
at approximately 59 and 82 kDa, which was consistent with linked
oligonucleotides to rhVDR and rhRXRß, respectively. Specific
competition with an excess amount of hOC VDRE resulted in the
disappearance of both bands; however, competition with an excess of the
nonspecific cvitERE had a dramatic effect on either eliminating or
strongly reducing the intensity of the 82-kDa band, while a more modest
effect on the 59-kDa band was observed. This suggests that the 82-kDa
band is not specific and these results would be consistent with a
polarity of binding that places rhVDR in the 5'-position. Using the
same type of analysis, cross-linking to the 3'-halves of both elements
resulted in the most prominent, specific band at 82 kDa, indicative
of hRXRß linked to the respective 3'-oligonucleotides. Thus, while
the polarity appears reversed for the wild-type negative element from
other known positive VDREs, this alignment is maintained in the dmPTH
mutant sequence that elicits a positive transcriptional response.
Therefore, an altered polarity by itself cannot account for the
negative to positive transcriptional response observed with these two
sequences.
|
in
the presence of 1,25-(OH)2D3 and washed with a
moderately high-salt (200 mM KCl) buffer to limit binding
by nonspecific proteins. These matrices were then equilibrated in a
low-salt buffer and incubated with extracts prepared from MG-63 cells,
a human osteoblastic cell line, that permitted having human recombinant
receptors interacting with human-derived cofactors. After this, the
matrices were washed extensively with low-salt buffer, and all of the
proteins were recovered in a high-salt wash and analyzed by SDS-PAGE
and silver staining. As a further set of controls, the hOC VDRE and a
DR+1 known to bind RXR homodimers (40, 41) were similarly prepared. In
the latter case, the column was prebound with rhRXR extracts alone
in the presence of 9-cis-retinoic acid.
As expected, rhVDR and rhRXR were readily evident in the elutions
from the matrices in comparable amounts (Fig. 6). Several proteins, including a cluster
in the range of 45–47 kDa and a 240-kDa band, appeared in all four
eluted samples, although the 45-kDa band was more prominently observed
in the hOC elution with little of the 47-kDa band evident. In addition,
a protein at 50 kDa appeared more intensely in the elutions from the
columns bound with VDR/RXR heterodimer complexes than in the
rhRXR homodimer experiment. A pair of intense bands at approximately
60 and 54 kDa from the MG-63 cell extract, above and below the position
of rhRXR, appeared to specifically associate with the heterodimer
complex bound to the hOC VDRE and were not evident in the other
elutions. A 98-kDa protein appeared in elutions from both the hOC and
aPTH VDRE affinity columns, but did not appear in the other columns.
Interestingly, the profile from the aPTH column yielded one unique band
at approximately 160 kDa that was not observed in any of the other
elutions. Analogous profiles from the four columns were obtained from
complexes formed in the absence of ligands in the binding mixtures
(data not shown).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
antibody
directed against the hinge region of the VDR generated a minor, but
significant, supershifted band in the mobility shift assay (Fig. 2).
Previous work with this antibody under identical conditions using
positively responding VDREs demonstrated that binding of the VDR was
blocked, and no bound or supershifted products were observed (11, 29).
Collectively, these data would be consistent with an altered
conformational change in the VDR and/or an altered polarity of binding
by the heterodimer to these different types of VDREs. Interference footprinting is predicated on the concept that limited chemical modifications to singular positions in a DNA binding site may preclude the ability of a protein to bind to that site, either through steric, structural, or electrostatic perturbations to the DNA, and thus be excluded from the bound fraction during the separation procedure. Interactions of the heterodimer with the wild-type and mutant sequences appeared highly specific using recombinant sources of receptor proteins, and there were only minor differences between these sequences in their ability to compete for binding in the relative competitive binding electrophoretic mobility shift assay (EMSA). Yet distinct footprints were evident for each DR+3, suggesting different modes of interaction by the heterodimeric partners. This implies that the heterodimer complex may be making additional contacts within the various sequences with individual nucleotides not examined in this study or that the bound complex is able to somehow accommodate these uniquely modified positions. In the case of the aPTH VDRE then, this latter possibility suggests that any single modification by itself may not be sufficient to entirely block the complex from still binding to that sequence.
Surprisingly, the mPTH sequence that placed an adenine residue at
position +7, which is highly conserved in the majority of positive DR+3
VDREs (1), caused only a minor shift in the footprint analysis and
demonstrated only a moderate attenuation of the negative response in
the transcription assay. Conversely, the significant increase in the
degree of interference observed by the GC swap at position +6 in the
cPTH mutant, together with the observed weak, yet clearly positive
transcriptional response, revealed the relative importance this
position may exert in acting as a ‘switch’ to distinguish negative
and positive activity. Finally, the dmPTH sequence that exhibited the
strongest interference footprint pattern also displayed the strongest,
positive transcriptional response. This sequence is a perfect direct
repeat of the GGGTCA half-site, highly similar to DNA elements
associated with positive transcriptional responses (1, 7). In addition,
binding site selection analyses indicated that the preferred
DNA-binding sequence of the heterodimer ended in the CA dinucleotide
(positions -3/-2 and +6/+7) for both half-sites (12, 42). However,
other than the dmPTH sequence, there was no strong association between
the order of competition in the EMSA, dmPTH aPTH cPTH
mPTH, and the rank order of transcriptional outcomes from strongly
positive to strongly negative: dmPTH cPTH mPTH aPTH. This
suggests that factors other than binding avidity may be playing a role
in determining transcriptional activity.
In contrast to previous reports on other VDREs (20, 21, 22, 43), the cross-link data place RXR preferentially over the 3'-half of both the aPTH and dmPTH elements. Preliminary results would indicate that this arrangement may also hold for a VDRE recently identified in the rat PTH gene that exhibits striking sequence similarity to the avian element (J. Russell, S. Ashok, and N. J. Koszewski, manuscript submitted). Previous data demonstrated that the VDR can bind as a homodimer to a DR+3 element (9, 11, 12), indicating that the VDR is capable of occupying the 5' half-site of a VDRE. The retinoid receptors have also been reported to display an altered polarity in binding to selected response elements that resulted in enhanced or repressed transcriptional activity (44). However, an altered polarity by itself did not restrict the VDR heterodimer complex to a unique transcriptional outcome in the present study, as exhibited by the wild-type aPTH and dmPTH sequences that maintained the same polarity, but with opposing transcriptional responses.
The VDR, like other members of the nuclear receptor gene family, has also been shown to interact with a variety of accessory proteins (45, 46, 47, 48, 49, 50). In the studies with the transcriptionally negative aPTH element, binding to that sequence by the heterodimer generated a weak interference footprint and promoted unique interaction with a p160 protein. In contrast, heterodimer binding to the positive hOC and dmPTH sequences, the latter that differed by only two nucleotides from the aPTH element, failed to exhibit an interaction with this protein. The elution from the hOC affinity column also revealed the significant presence of additional proteins at approximately 60 and 54 kDa that appeared specific for this sequence. This may reflect proteins from human osteoblast-like cells associating with the recombinant human heterodimer complex bound to a regulatory sequence from a gene that is naturally expressed in these cells (51). Additional studies are ongoing, but the relative presence or absence of these proteins may be involved in the observed negative or positive transcriptional responses from these specific response elements.
It was noted that essentially identical elution profiles from the affinity chromatography columns have been obtained from complexes bound in the presence or absence of added vitamin D hormone. A similar lack of hormone dependence was also observed in analogous experiments using HeLa cell nuclear extracts (data not shown). There may be several reasons to account for this observation, the first of which may be an inherent lack of sensitivity in this experimental system. Either in the presence or absence of hormone, a relatively large amount of affinity-purified hVDR/hRXR complex was immobilized on these VDRE columns relative to the interacting factors present in the diluted cell extracts. Also, the binding conditions and washes for the affinity columns were performed under low-salt conditions. Thus, if the effect of hormone is to alter the affinity for an interacting protein, then having such an excess amount of complex in either a hormone-bound or unbound state under low-salt conditions may simply transcend those affinity differences and result in an inability to visualize hormone-dependent interactions. Alternatively, this experiment may be revealing only those basal, hormone-independent interactions that occurred in an element-specific fashion. That is, these factors may associate with the VDR complex as it binds to these VDREs to assist in achieving the proscribed regulatory response, while the function of the hormone is to recruit yet additional factors not revealed by this analysis.
A general working model has appeared in the literature that incorporates many of the current concepts in explaining activation of gene expression by the thyroid/retinoid subfamily of nuclear receptors, to which the VDR belongs (52). The chromatin remodeling scheme proposes that the unliganded receptor binds to its response element as part of a multiprotein complex that also includes corepressor or silencer molecules and histone deacetylases. In this state, transcription of the target gene is repressed or silenced. Hormone binding is proposed to alter the conformation of the receptor complex, and transcriptional coactivators are recruited to the scene along with histone acetylases, and gene transcription is turned on. While accounting for enhancement of gene transcription, a question arises as to what factors are involved that would explain hormone-dependent repression and activation in the same cell. The present data suggest that an additional detail to consider is the specific sequence of the DNA-binding site. That is, beyond the potential role in binding affinity, the subtle and not-so-subtle differences in DNA sequences capable of binding the VDR specifically may be acting as allosteric effectors of receptor function (36, 37, 53, 54). Binding of the VDR to different types of response elements, therefore, is proposed to elicit conformational changes in the receptor complex and/or the exposure of different regions of the molecule to the external environment, similar in concept to agonist or antagonist binding to the ligand-binding domains of hormone receptors. Collectively, this suggests that plasticity exists in the VDR/RXR heterodimer arrangement and corresponding transcriptional response that is inextricably associated with the exact DNA sequence (55, 56). These changes in the VDR complex invoked by binding to distinct DNA response elements may then facilitate the interaction of available cell type-specific coactivators or corepressors to bring about the desired transcriptional response.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
32P-ATP (6000 Ci/mmol) and 32P-dATP
(3000 Ci/mmol) were purchased from Dupont NEN Research Products
(Wilmington, DE). Oligonucleotides were synthesized at the University
of Kentucky Molecular Structure Analysis Facility with
BamHI- and XbaI-compatible ends and cloned into
pGEM11zf (Promega Corp, Madison, WI). Sequences were confirmed by
dideoxysequencing. Sequences of the top strands are as follows: aPTH,
5'-CTAGAATGAGGGTCAGGAGGGTGTGCTGG; mPTH,
CTAGAATGA-GGGTCAGGAGGGTGAGCTGG; cPTH, CTAGAGAGGGTCAGGAGGGTCTGCG;
dmPTH, CTAGAGAGGGTCAGGA-GGGTCAGCG; hOC,
CTAGATTGGTGACTCACCGGGTGAACGGGGGCATTGCG. The aPTH promoter sequence was
synthesized by PCR and ligated into a commercially available pCAT
expression vector using the SalI and XbaI
restriction sites present in the multiple cloning site. Synthetic
oligonucleotides containing the sequences for wild-type aPTH, the
various aPTH mutants, and wild-type hOC VDREs were excised from
pGEM11zf with the combination of SalI and HindIII
and placed immediately upstream from the aPTH promoter/pCAT construct.
The 9A7 antibody was generously provided by Dr. J. Wesley Pike
(University of Cincinnati), while the anti-RXR and anti-RXRß
antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz,
CA).
Mobility Shift Assays
Recombinant hVDR- and hRXR-containing extracts were prepared
from baculovirus-infected Sf9 insect cells as described previously
(11). The cell pellets were homogenized in 6 volumes of buffer [20
mM Tris (pH 7.5), 1 mM EDTA, 2 mM
dithiothreitol (DTT), 350 mM KCl, 10 mM NaF,
100 µM Na3VO4, 0.1 mM
leupeptin, and 10% glycerol] on ice. The cytosols were clarified by
centrifugation at 30,000 x g for 60 min at 4 C,
snap-frozen, and stored at -70 C before use. For radiolabeling, DNA
fragments were excised from the pGEM11zf-based plasmids using the
combination of EcoRI and HindIII restriction
enzymes. Radioactive probes were generated by fill-in labeling
reactions using Klenow fragment and -32P-dATP. Cytosols
of recombinant hVDR and hRXR were diluted 1:25 in ice-cold KTEDG
buffer [400 mM KCl, 20 mM Tris (pH 7.5), 1
mM EDTA, 2 mM DTT, and 10% glycerol] before
use. Samples were kept cold, and 1 µl each of diluted extracts was
used in a 20 µl final volume in a buffer that included 100
mM KCl, 20 mM Tris (pH 7.5), 1.5 mM
EDTA, 2 mM DTT, 5% glycerol, 0.5%
3-(3-cholamidopropyl)dimethylammonio-1-propanesulfonate (CHAPS),
10 mM NaF, 100 µM
Na3VO4, 1.0 µg dIdC, and 100 nM
1,25-(OH)2D3. After incubation on ice, the
radiolabeled probes were added and incubation continued for 1 h.
The samples were then applied to cooled, prerun 5% polyacrylamide gels
(29:1) in 0.5x Tris-borate-EDTA buffer, and electrophoresis was
initiated at 14 V/cm for 3 h. Gels were transferred and dried
followed by autoradiography.
Antibodies (1 µl each of undiluted sample) were allowed to incubate with the gel shift samples for 1 h at 4 C before the addition of the radiolabeled probes. Cold competition binding reactions were carried out using the cvitERE (5'-GATCCCTGGTCAGCGTGACCGGAG) and hOC VDRE (5'-CTAGATTGGTGACTCACCGGGTGAACGGGGGCATTGCG). For the relative competitive binding experiments, recombinant cytosols were diluted 1:50, and unlabeled competitor oligonucleotide samples were mixed with radiolabeled wild-type aPTH probe and this combination was added to the binding reactions.
Interference Assay
The ethylation interference footprints were obtained as
previously described (11). The DNA probes were generated from the
pGEM11z-based plasmids containing the inserted oligonucleotides by
end-labeling linearized, CIP-treated plasmids with polynucleotide
kinase and -32P-ATP using the combination of
EcoRI and HindIII restriction enzymes. Briefly,
32P-end-labeled DNA probes in 50 mM sodium
cacodylate buffer (pH 8.0) were treated with ethylnitrosourea-saturated
ethanol for 20 min at 55 C. After precipitation with sodium
acetate/ethanol and reprecipitation (three times), the pellets were
washed with 70% ethanol, dried, and resuspended in water. Ethylated
probes were then used in the gel mobility shift assay as above, except
the amounts of probe were increased to 10–15 fmol. Acrylamide sections
corresponding to bound and free DNA were excised, and the DNA was
recovered by electrochemical elution and precipitation. Cleavage of the
modified DNA was accomplished by treating with 100 mM
NaOH/0.1 mM EDTA in 10 mM phosphate buffer at
95 C for 30 min followed by neutralization with 3 M sodium
acetate (pH 5.2) and precipitation with ethanol. Samples were separated
by 8% sequencing gels and dried, and autoradiography was performed
followed by densitometric analysis.
Transfection Studies
OK cells were cotransfected by lipofectin as previously
described with the appropriate VDRE-CAT construct (2.5 µg, Fig. 1)
and cytomegalovirus (CMV)-ß-galactosidase (2.5 µg) plasmids (29).
Analysis of CAT gene transcripts after treatment with
1,25-(OH)2D3 for 24 h was quantified by
ribonuclease protection assay within the same assay. Total RNA from OK
cells transfected with the various PTH-CAT and CMV-ß-galactosidase
plasmid constructs was prepared by extraction with phenol-guanidinium
thiocyanate. The RNA extracts were incubated with DNase I for 5 min at
room temperature and ethanol precipitated. The RNA pellets were
incubated with radiolabeled RNA probes (158 bases for CAT and 370 bases
for ß-galactosidase) in 20 µl of buffer containing 80% formamide,
100 mM sodium citrate, pH 6.4, 300 mM sodium
acetate, pH 6.4, 1 mM EDTA, overnight at 43 C. The
hybridization mixtures were digested with a combination of ribonuclease
A and ribonuclease T1 at 37 C for 30 min. The protected RNA hybrids
were precipitated with propanol-guanidinium thiocyanate and
separated on 6% nondenaturing polyacrylamide gels. The gels were dried
and exposed to x-ray film for 6 h at -80 C. The resulting
autoradiograms were quantified by densitometric scanning, and values
for the CAT gene transcripts were normalized with respect to the values
for cotransfected ß-galactosidase gene transcripts.
UV Cross-Linking
The 5'- or 3'-oligonucleotides (5':
GCATCTAGAATGAGGGBrUCAGG; 3':
AGGGBrUGBrUGCTGGATCCCTCGA) of the aPTH were
end-labeled with -32P-ATP and T4 polynucleotide kinase.
The labeled oligonucleotides were mixed with an excess of the other
unlabeled oligonucleotide and annealed to the complementary sequence.
After hybridization, the double-stranded oligonucleotides were purified
by 8% PAGE and recovered by electroelution. The hybridized DNA
fragments were then included in incubations as described above for the
mobility shift assay except the amount of recombinant cytosols was
increased 2.5-fold, and 100,000 cpm of gapped, radiolabeled probes were
added. The sample tubes were maintained on ice and irradiated for 45
min at a wavelength of 254 nm. After irradiation, the binding reactions
were mixed with 2x SDS-PAGE loading buffer and denatured at 95 C for 5
min, and the proteins were separated by 10% SDS-PAGE. After
electrophoresis the gels were fixed and dried, and autoradiography was
performed. Cross-linking probes for the dmPTH (5':
CCGCATCTAGAGAGGGBrUCAGG; 3':
AGGGBrUCAGCGGATCCCTCGAG) were prepared and treated in an
analogous fashion.
Affinity Chromatography Studies
Oligonucleotide probes biotinlyated at the 5'-end for aPTH and
dmPTH were synthesized as above, annealed to their complementary
strand, and bound to streptavidin-agarose beads. In addition, a DR+1
matrix was similarly prepared (5'-GATCCGGGTAGGGGTCAGAGGTCACTCGT).
Binding reactions of a mixture of rhVDR and rhRXR in the presence of
1,25-(OH)2D3 or rhRXR alone in the presence
of 9-cis-retinoic acid (10 µM) were prepared
as outlined above, but in amounts to completely saturate binding to the
response elements present in each column. After incubation of the
recombinant proteins with the agarose matrices for 60 min at 4 C, the
matrices were washed with approximately 50 volumes of KTEDG-200 buffer.
After this, the columns were reequilibrated in 10 volumes of KTEDG-60
buffer. Whole-cell extracts were prepared from human osteosarcoma MG-63
cells as outlined above and diluted to 60 mM KCl
concentration. The MG-63 extracts were then mixed with the agarose
matrices at 4 C for 60 min and then washed with approximately 50
volumes of KTEDG-60 buffer. The bound proteins were subsequently eluted
by incubation with KTEDG-400 buffer, denatured, separated by 10%
SDS-PAGE, and silver stained.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
This work was supported by NIH Grants DK-47883 (N.J.K.), DK-38422 (J.R.), and Dialysis Clinic Incorporated (N.J.K.).
Received for publication May 14, 1998. Revision received November 20, 1998. Accepted for publication November 30, 1998.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
heterodimers are specified by dinucleotide differences in the vitamin
D-responsive elements of the osteocalcin and osteopontin genes. Mol
Endocrinol 10:1444–1456[Abstract]
This article has been cited by other articles:
 |
|
 |
|
  H. Du, G. S. Daftary, S. I. Lalwani, and H. S. Taylor Direct Regulation of HOXA10 by 1,25-(OH)2D3 in Human Myelomonocytic Cells and Human Endometrial Stromal Cells Mol. Endocrinol., September 1, 2005; 19(9): 2222 - 2233. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. R McGaffin and S. A Chrysogelos Identification and characterization of a response element in the EGFR promoter that mediates transcriptional repression by 1,25-dihydroxyvitamin D3 in breast cancer cells J. Mol. Endocrinol., August 1, 2005; 35(1): 117 - 133. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J.-N. Bastie, N. Balitrand, F. Guidez, I. Guillemot, J. Larghero, C. Calabresse, C. Chomienne, and L. Delva 1{alpha},25-Dihydroxyvitamin D3 Transrepresses Retinoic Acid Transcriptional Activity via Vitamin D Receptor in Myeloid Cells Mol. Endocrinol., November 1, 2004; 18(11): 2685 - 2699. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. P. Alimov, M. C. Langub, H. H. Malluche, and N. J. Koszewski Sp3/Sp1 in the Parathyroid Gland: Identification of an Sp1 Deoxyribonucleic Acid Element in the Parathyroid Hormone Promoter Endocrinology, July 1, 2003; 144(7): 3138 - 3147. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G.-J. C. M. van den Bemd, M. Jhamai, A. Staal, A. J. van Wijnen, J. B. Lian, G. S. Stein, H. A. P. Pols, and J. P. T. M. van Leeuwen A Central Dinucleotide within Vitamin D Response Elements Modulates DNA Binding and Transactivation by the Vitamin D Receptor in Cellular Response to Natural and Synthetic Ligands J. Biol. Chem., April 19, 2002; 277(17): 14539 - 14546. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. POLLY, M. HERDICK, U. MOEHREN, A. BANIAHMAD, T. HEINZEL, and C. CARLBERG VDR-Alien: a novel, DNA-selective vitamin D3 receptor-corepressor partnership FASEB J, July 1, 2000; 14(10): 1455 - 1463. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||
