|
|
|
|||||||
|
|||||||||
(Received for publication, August 9,
1995; and in revised form, December 15, 1995) From the
The significance of conserved cysteines at positions 288, 337,
and 369 in the hormone binding domain of the human vitamin D receptor
was evaluated by individual site-directed mutagenesis to glycine.
Neither nuclear localization nor heterodimerization with retinoid X
receptors in binding to the vitamin D-responsive element was
appreciably affected by altering these cysteines, but vitamin D hormone
(1,25-(OH)
The 1,25-dihydroxyvitamin D Cysteine residues are known to play a vital role in the formation
and maintenance of protein conformation. Eight cysteine residues in the
DNA binding domain are absolutely conserved among the
steroid/retinoid/thyroid hormone receptors and form two zinc fingers
which are involved in binding to the cognate-responsive elements of
these receptors(14) . Furthermore, in keeping with an important
structural role for cysteines, several cysteinyl residues in the HBD of
steroid hormone receptors have been proposed to be involved in ligand
binding. Previous studies revealed that cysteine residues in the
glucocorticoid receptor (GR)(15, 16, 17, 18, 19, 20) and
estrogen receptor (ER) (21, 22, 23, 24) play important
roles in high affinity ligand binding. In the HBD of VDR, cysteines at
positions 288, 337, and 369 are conserved in the human(25) ,
rat(26) , and avian (27, 28) receptors. Little
information is available on the importance of HBD cysteines in the
mechanism of VDR action, although in an early biochemical experiment,
Coty et al. (29) showed that treatment of
hormone-occupied avian VDR with mercurial reagents causes a
dissociation of the 1,25-(OH)
Figure 1:
Schematic representation and
expression of mutant hVDRs. A, mutant hVDR cDNAs were
generated by site-directed mutagenesis within the expression vector
pSG5hVDR to replace Cys-288, -337, and -369 with glycine. The resulting
mutant hVDR proteins are depicted schematically in the context of the
hormone binding domain. B, wild-type and mutant hVDRs were
expressed in COS-7 cells, and cell extracts were prepared as described
under ``Materials and Methods.'' These lysates were subjected
to immunoblotting with a specific monoclonal antibody to VDR
(9A7
Figure 2:
Heterodimerization of the wild-type or
mutant hVDRs with retinoid X receptors on the VDRE. Cellular extracts
from COS-7 cells expressing wild-type or mutant hVDRs were incubated
with
Figure 3:
Cellular distribution and stability of
expressed wild-type and mutant hVDRs in COS-7 cells. COS-7 cells
transfected with wild-type or mutant hVDR expression plasmids were
incubated in the absence (upper panel) or presence (lower
panel) of 10
Figure 4:
Transcriptional activation by
1,25-(OH)
To amplify this analysis of transactivation, we performed a
series of dose-response experiments in co-transfected COS-7 cells
treated with 1,25-(OH)
Figure 5:
Specific binding of
1,25-(OH)
Because of the striking differences between ligand binding kinetics
with receptor extracts at 4 °C and in intact cells at 37 °C (Fig. 5A), we performed a final
1,25-(OH) Evidence for the involvement of cysteine residues in steroid
hormone receptor-ligand binding was originally reported when it was
observed that mercurial reagents which interfere with protein thiol
residue interaction reduce ligand binding to the ER (21) and
GR(15, 41) . Coty et al. (29) also
found that treatment of occupied avian VDR with mercurial reagents
causes a dissociation of 1,25-(OH) The
present data advance our understanding of the ligand binding domain of
hVDR and provide insight into the potential role of the three conserved
cysteine residues in this region. Cysteine 288 is clearly essential for
normal high affinity hormone binding (Fig. 5, A and B) and stimulation of transcription at physiologic doses of
1,25-(OH) The recent crystal
structure elucidation of the HBD of human RXR The two
mutants reported here for hVDR which have substantial effects on
hormone binding and hormone-dependent transactivation, namely Cys-288
and Cys-337, occur in areas corresponding to the 1st Covalent modification of specific residues in hVDR with ligands
using affinity labeling techniques will be required to extend the
present conclusions. Further studies of the type carried out in this
report could involve altering Cys-288 and Cys-337 in hVDR to serine or
alanine residues, since they may better preserve the size of the
R-group and possibly the protein conformation. Ultimately, a physical
examination of the molecular structure of the
1,25-(OH)
Volume 271,
Number 9,
Issue of March 1, 1996 pp. 5143-5149
©1996 by The American Society for Biochemistry and Molecular Biology, Inc.
Receptor (*)
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
D
) activated transcription was
compromised significantly in the C288G and C337G mutants. Only the
C288G mutant displayed depressed (3-fold) 1,25-(OH)D ligand binding affinity at 4 °C, in vitro, although
at elevated temperatures (23-37 °C), ligand binding was
attenuated severely in C288G, moderately in C337G and very mildly in
C369G. The degree of impairment of ligand binding at physiologic
temperatures correlated with the requirement for increased
concentrations of 1,25-(OH)D ligand to
maximally stimulate transcriptional activity in co-transfected COS-7
cells. Thus cysteine 288 and, to a lesser extent, cysteine 337 are
important for high affinity hormone binding to the vitamin D receptor,
which ultimately leads to ligand-dependent transcriptional activation.
receptor (VDR) (
)is a nuclear protein that mediates many of the biological
actions of the 1,25-dihydroxyvitamin D (1,25-(OH)D) hormone, such as regulating
calcium/phosphorus metabolism and cellular differentiation (1) . The VDR belongs to the steroid/retinoid/thyroid hormone
receptor superfamily and, as with other members of this receptor
family, consists of a highly conserved domain, which contains two zinc
finger motifs required for DNA binding, and a carboxyl-terminal hormone
binding domain (HBD) responsible for the specific, high affinity
binding of 1,25-(OH)D, the active form of
vitamin D(2) . The binding of ligand presumably initiates a
conformational change in the VDR protein whereupon the hormone-receptor
complex binds to distinct sequences of nucleotides, termed vitamin
D-responsive elements (VDREs), located upstream of target genes and
thereby modulates transcription(3) . VDRE sequences have been
identified in the promoter regions of the human (4) and rat (5) osteocalcin genes as well as in the mouse osteopontin
gene(6) . Generally, VDREs consist of an imperfect direct
repeat of six nucleotide bases, GGGTGA, separated by a three-base pair
spacer. Gel mobility shift analysis using these VDREs has revealed that
an additional nuclear factor, the family of retinoid X receptors
(RXRs), forms heterodimers with the VDR (7, 8, 9) and facilitates its binding to DNA.
VDR-RXR heterodimerization on VDREs is postulated to play an essential
role in transcriptional modulation of target genes through the VDR.
Mutational analysis has revealed that a conserved region between
residues 244 and 263(10, 11) , as well as the fourth
and ninth heptad repeats (12) in the HBD of human VDR (hVDR),
are essential for heterodimerization on the VDRE. We have also reported
that the region between amino acids 403 and 427 in the HBD of hVDR may
be involved in transcriptional regulation(12) . Therefore, the
HBD of hVDR must be considered as a multifunctional domain important
not only for binding to the 1,25-(OH)D ligand,
but also for forming a heterodimer with RXR and likely for interacting
with the transcriptional machinery. These functions are closely
related, since the 1,25-(OH)D hormone enhances
the formation of VDR heterodimers with RXR on the VDRE (9, 13) and co-expression of RXRs enhances
ligand-dependent transactivation mediated by the VDR(9) . D ligand. This
result suggested that VDR amino acids with sulfhydryl-containing side
chains may play a crucial role in ligand binding, perhaps by
maintaining the proper conformation of the hormone binding pocket.
Detailed involvement of specific cysteines in this and other functions
of VDR remains to be elucidated. In the present study, we constructed
several site-specific mutant hVDRs to examine the precise roles of each
of these conserved cysteine residues in the HBD of hVDR.
Preparation of Wild-type and Mutant VDRs
The
hVDR expression vector, pSG5hVDR(30) , was utilized to create
point-mutant hVDR plasmids by in vitro site-directed
mutagenesis(31) . Three cysteine residues in the HBD of hVDR at
positions 288, 337, and 369, which are conserved among species, were
replaced by glycines (designated C288G, C337G, and C369G,
respectively). COS-7 monkey kidney cells (5 
10
cells/150-mm culture dish) were transfected with 50 µg of
each plasmid/plate by the calcium-phosphate DNA co-precipitation
method, as described previously(32) , and cultured in
Dulbecco's modified Eagle's medium (DMEM) (Life
Technologies, Inc.) with 10% dextran-coated charcoal-stripped fetal
bovine serum (Gemini Bioproducts, Calabasas, CA) for 48 h. The cells
were then scraped, washed three times with phosphate-buffered saline
(136 mM NaCl, 26 mM KCl, 8 mM NaHPO
, 1.5 mM KHPO, pH 7.2), and resuspended in KETD-0.3
buffer (10 mM Tris-HCl, pH 7.6, 1 mM EDTA, 300 mM KCl, 10% glycerol, 1 mM dithiothreitol, 0.1 mM phenylmethylsulfonyl fluoride, 15 µg/ml aprotinin, 1 µg/ml
leupeptin, 1 µg/ml pepstatin A). After sonication, samples were
centrifuged at 215,000 g for 30 min at 2 °C. The
supernatant was collected, divided into small aliquots, and stored at
-70 °C. Western blotting of VDR was performed with anti-VDR
monoclonal antibody, 9A7, as described previously(33) .
Protein concentrations were determined according to the method of
Bradford (34) .
Ligand Binding Assays
COS-7 cell extracts
containing wild-type or mutant VDR proteins were incubated with the
indicated concentrations of
1,25-(OH)[H]D (Amersham
Corp.) for 10 h at 4 °C or for 2 h at 23 °C along with 20
µg of rat liver nuclear extract in the presence or absence of a
400-600-fold molar excess of unlabeled
1,25-(OH)D. Addition of rat liver nuclear
extract, which alone does not possess 1,25-(OH)D binding activity, to samples containing overexpressed VDR is
required for efficient high affinity, saturable
1,25-(OH)D hormone binding(35) . Bound
and free ligand were separated with dextran-coated charcoal for
analysis of hormone binding(36) .Cellular Distribution of VDR
COS-7 cells
transfected with wild-type or mutant VDR plasmids were incubated in the
absence or presence of 10M 1,25-(OH)
D for 12 h at 37 °C. The
nuclear versus cytosolic partitioning of VDR was determined as
described previously(33) .Nuclear Uptake Assay
Assessment of the nuclear
uptake of 1,25-(OH)[H]D by COS-7 cells was performed essentially as described previously (37) with slight modifications. COS-7 cells overexpressing
hVDRs were harvested with 0.025% trypsin, 0.02% EDTA and washed twice
with 5 ml of DMEM. Cells were then resuspended in DMEM containing 1%
fetal bovine serum at a concentration of 2.5 10 cells/ml and incubated with the indicated concentrations of
1,25-(OH)[H]D in the
presence or absence of a 600-fold molar excess of unlabeled
1,25-(OH)D at 37 °C for 2 h with shaking
(100 rpm). Nuclear fractions were obtained as described
elsewhere(33) , and the radioactivity was measured by liquid
scintillation counting. Specific binding was calculated by subtracting
nonspecific binding from total binding.Gel Mobility Shift Assay
Gel mobility shift assays
were performed in the absence of 1,25-(OH)D ligand as described previously(38) . Briefly, cellular
extracts from COS-7 cells overexpressing wild-type or mutant hVDR were
incubated with 0.5 ng of P-labeled rat osteocalcin VDRE
(5`-AGCTGCACTGGGTGAATGAGGACATTACA-3`) in the presence of RXR
or
RXR
for 30 min at 22 °C. The reaction mixtures were loaded
onto 4% nondenaturing polyacrylamide gels in 0.25 TBE (11
mM Tris, 11 mM borate, 0.5 mM EDTA). Gels
were run at 10 mA for 70 min, dried, and subjected to autoradiography.Transcriptional Activation Assay
Four copies of
the CT4 oligonucleotide, which contains the rat osteocalcin VDRE, were
subcloned into the upstream region of the viral thymidine kinase
promoter-growth hormone reporter gene construct (Nichols Institute, San
Juan Capistrano, CA) creating (CT4)-TKGH(5) . COS-7
cells (5 10
cells/60-mm culture plate) were
co-transfected with 5 µg of (CT4)-TKGH and wild-type or
mutant hVDR plasmids and then incubated with a final concentration of
10 to 10
M 1,25-(OH)
D or ethanol vehicle for 12 h at
37 °C. Growth hormone concentrations in the culture medium were
determined by radioimmunoassay (Nichols Institute).
Expression of Wild-type and Mutant
hVDRs
Cysteine residues in the HBD of hVDR at positions 288,
337, and 369, which are conserved in VDRs from various species, were
individually replaced with glycine by site-directed mutagenesis
(designated C288G, C337G, and C369G, respectively). The resulting
mutated hVDR proteins are depicted schematically in Fig. 1A. Wild-type and mutant hVDR expression plasmids
were transfected into COS-7 cells, and immunoblotting was employed to
assess the levels and stability of these proteins (Fig. 1B). The mutations introduced do not appear to
dramatically affect the expression or stability of the hVDR protein in
COS-7 cells.
). Molecular weight standards are shown in the first and last
lanes.
Heterodimerization on the VDRE
Because the
replaced cysteines are near heptad repeat regions of hVDR (Fig. 1A) known to be required for dimerization with
RXR, we first performed gel mobility shift assays with cellular
extracts from transfected COS-7 cells overexpressing wild-type or
mutant hVDRs in order to compare the abilities of the receptors to form
heterodimers with RXR on the VDRE. The amount of COS-7 cellular extract
utilized in these VDRE binding assays was adjusted based on the level
of hVDR expression as determined by quantitative immunoblotting.
Wild-type and all mutant hVDRs tested formed hVDR-dependent, shifted
complexes with both RXR (Fig. 2, lanes 1-4),
and RXR (Fig. 2, lanes 6-9). C288G appeared
to heterodimerize with the RXRs on the VDRE with a capacity identical
to that of the wild-type receptor (compare lanes 1 and 2 or 6 and 7 of Fig. 2). However, both
C337G and C369G were slightly diminished in their ability to form
shifted complexes in relation to the native hVDR (see lanes 3 and 4 or 8 and 9, Fig. 2). This
analysis utilizing enriched preparations of RXR and RXR
revealed that all mutant hVDRs examined formed qualitatively normal
heterocomplexes with the RXRs, with only minor quantitative attenuation
of C337G and C369G, the latter effect best detected by careful
examination of the relative amounts of unshifted probe in Fig. 2. Extracts of COS-7 cells transfected with a control
vector lacking the hVDR cDNA, even in the presence of RXRs (Fig. 2, lanes 5 and 10), did not form a
complex on the VDRE. This observation, plus the fact that shifted
complexes could be inhibited by VDR-specific monoclonal antibody
9A7 (data not shown), argues that the retarded bands are indeed
comprised of complexes of hVDR. Finally, identical conclusions
regarding the mutant hVDRs expressed in COS-7 cells were reached when
heterodimerization was assessed in combination with endogenous nuclear
factors and RXRs present in a rat liver nuclear extract (data not
shown).
P-labeled rat osteocalcin VDRE in the presence of
RXRs. Detailed procedures are described under ``Materials and
Methods.'' Extracts from COS-7 cells transfected with expression
plasmids for wild-type hVDR (lanes 1 and 6), C288G (lanes 2 and 7), C337G (lanes 3 and 8), C369G (lanes 4 and 9), or expression
vector pSG5 without the hVDR cDNA insert (lanes 5 and 10) were incubated with
P-labeled VDRE in
combination with human RXR
(lanes 1-5) or mouse
RXR (lanes 6-10). RXR and RXR were
expressed and partially purified as described
elsewhere(9) .
Nuclear Translocation of Expressed
Receptors
Steroid/retinoid/thyroid hormone receptors are
localized to, or translocated into, the cell nucleus upon ligand
binding. In the case of VDR, like the retinoid and thyroid hormone
receptors, the unoccupied receptor is already predominantly localized
in the nucleus(39, 40) . Although the
``hinge'' region between the DNA binding and HBD of many
nuclear receptors has been proposed to be responsible for nuclear
translocation(14) , it is possible that conformational changes
resulting from the introduction of point mutations in the hormone
binding domain may attenuate nuclear translocation and therefore
diminish DNA binding and transcriptional activity of mutant hVDRs. Fig. 3illustrates VDR immunoblotting utilizing equivalent
amounts of total protein of nuclear (N) and cytosolic (C) fractions of COS-7 cells transfected with the indicated
receptor and incubated in the absence or presence of
1,25-(OH)D. As expected, VDR is enriched in the
nuclear fraction and the subcellular partitioning is not markedly
affected by treatment with the hormonal ligand. Most relevant, all of
the mutant hVDRs exhibit a similar apparent N/C ratio when compared
with the wild-type receptor as determined by quantitative densitometric
scanning of these immunoblots (data not shown). In this experiment, we
noted that C337G was either expressed in lower amounts or was more
unstable (especially in the absence of ligand) during the fractionation
procedure than wild-type receptor and the other two hVDR mutants, but
nuclear localization was still evident. These results strongly suggest
that mutations at cysteine 288, 337, or 369 do not significantly alter
the ability of the receptor to translocate into the nucleus.
M 1,25-(OH)
D for 12 h at 37 °C. Nuclear (N) and cytosolic (C) fractions were prepared as
described elsewhere(33) . Western blotting to detect the
specific activity of VDR in each fraction was performed with 15 µg
of total protein from each preparation.
1,25-(OH)
Considering that mercurial
reagents had been found to dissociate 1,25-(OH)DLigand
Binding at 4 °C, in VitroD from VDR(29) , we next examined the ligand binding
activity of the cysteine-mutant hVDRs utilizing an assay performed in vitro. Extracts from transfected COS-7 cells were incubated
at 4 °C for 10 h with
1,25-(OH)[H]D at
concentrations of 0.23, 0.45, 0.90, 1.35, and 1.8 nM.
Scatchard analyses of saturable, specific hormone binding by wild-type
and mutant hVDRs yielded evidence for a single class of high affinity
binding sites in each case (data not shown). A summary of the average
results (±S.E.) from four separate experiments is as follows.
The dissociation constant of C288G (293 ± 40 pM) was
3.4 times higher than the wild-type K
of 86
± 15 pM (p < 0.005). In contrast, C337G (K = 35 ± 7 pM)
demonstrated approximately one-half of the dissociation constant of the
wild-type (p < 0.02), while C369G (K = 77 ± 2 pM) was essentially equivalent to
the wild-type receptor with respect to this parameter. Thus, only
mutation of cysteine 288 appears to significantly diminish the affinity
of hVDR for the 1,25-(OH)D ligand, at least as
assessed by assay of binding at 4 °C, in vitro.Ligand-dependent Transcriptional Activation by 10 nM 1,25-(OH)
Even
though alteration of the three cysteines in question did not
dramatically affect ligand or DNA binding, in vitro, we
proceeded to investigate 1,25-(OH)DD-mediated
transcriptional activation by wild-type or mutant hVDRs via
co-transfection of individual hVDR plasmids and a rat osteocalcin
VDRE-containing growth hormone reporter gene construct in COS-7 cells. Fig. 4A shows the effects of a slightly supraphysiologic
concentration of 10M 1,25-(OH)
D
on VDRE-mediated
transactivation. The hormone generated a 33-fold increase in the
transcription of the reporter in the presence of wild-type hVDR. C337G
demonstrated transcriptional activation by 1,25-(OH)
D
at levels approximately 60% of the wild-type hVDR, while the
activity of C369G was not significantly different from the wild-type
receptor. Surprisingly, C288G did not show detectable transcriptional
activation (Fig. 4A) by the
1,25-(OH)
D
ligand, even though this mutant
exhibited reduced but still relatively high binding affinity for the
ligand under the conditions of the in vitro binding assay at 4
°C.
D via wild-type and cysteine point
mutant hVDRs. COS-7 cells were co-transfected with wild-type or mutant
pSG5hVDR expression plasmids along with the (CT4)-TKGH
reporter construct. The cells were then incubated with
1,25-(OH)D or ethanol vehicle for 12 h, and
growth hormone secretion into the medium was assessed. A,
transcriptional activation in the presence of 10M 1,25-(OH)
D. The amounts of
growth hormone secreted into the media were compared with the amount in
cultures receiving wild-type hVDR +
1,25-(OH)D, which was set at 100 arbitrary
units. Means (±S.E.) from four separate experiments are
depicted. Expression of mutant hVDR proteins in the cells used in this
experiment was similar to that of wild-type hVDR as determined by
immunoblotting (data not shown). B, dose-response curves.
COS-7 cells were transfected with wild-type hVDR (), C369G
(
), C337G (
), or C288G (
) expression plasmids along
with (CT4)-TKGH reporter and then treated with various
concentrations of 1,25-(OH)D as indicated. Each
point represents the average of assays on duplicate plates of cells,
and the data were normalized to the maximal transcriptional activation
effect of 1,25-(OH)D with the wild-type
receptor. Virtually identical dose-response curves were obtained in two
repeats of the experiment shown (see ``Results'' for
compilation of EC values). The typical maximal stimulation
of transcription by 1,25-(OH)
D with the C337G
and C288G mutant receptors was 15-fold, comparable with wild-type and
C369G in these experiments, but less than the 33-fold effect reported
for the experiment in A.
D concentrations ranging
from 10 to 10
M. A
representative result is pictured in Fig. 4B, showing
that the C369G hVDR mediates transcriptional activation nearly as
effectively as wild-type receptor, with only a slight defect apparent
at the lower doses of 10
and 10
M 1,25-(OH)
D. In the case of
C337G, the ligand response curve is appreciably shifted to the right,
with normalization of transactivation only occurring at the relatively
high level of 10M 1,25-(OH)
D and no significant effect at
10M 1,25-(OH)
D (Fig. 4B). C288G is clearly the most severely
affected mutant, with a minuscule but statistically significant
response at 10M 1,25-(OH)
D and the requirement for a
concentration of 10M 1,25-(OH)
D to restore maximal
transactivation. From the experiment depicted in Fig. 4B and two independent repeats, the following EC values
(in nM 1,25-(OH)
D ± S.D., n = 3) were calculated: wild-type VDR = 1.0 ±
0.2; C369G = 1.7 ± 0.3; C337G = 4.3 ± 1.5;
and C288G = 187 ± 23. These data provide indirect
evidence for a mild hormone binding suppression in C369G and more
substantial reductions in ligand binding affinities at 37 °C for
C337G and especially C288G. The observation that transactivation can be
effectively rescued in all mutant hVDRs by increasing ligand
concentrations argues against any of the three cysteines in question
participating in subsequent interaction with the transcription
machinery.1,25-(OH)
Although ligand binding studies
with cytosolic extracts at 0-4 °C are traditional for steroid
hormone receptors, the dissociation constants for the three cysteine
mutant hVDRs obtained in this fashion (see above) are inconsistent with
apparent ligand binding affinities inferred from the transcriptional
dose-response patterns shown in Fig. 4B. For example,
C288G exhibits only a 3.4-fold increase over wild-type in apparent KDBinding
at Elevated Temperatures as assessed at 4 °C, while the EC for 1,25-(OH)
D action via this mutant is
187 times that of the wild-type receptor. To resolve this discrepancy,
we more directly examined the ligand binding activities of wild-type
and mutant hVDRs at 37 °C, in vivo, by measuring the
nuclear uptake of tritiated 1,25-(OH)D in
intact COS-7 cells overexpressing the various hVDRs. Fig. 5A
illustrates the specific binding of ligand in nuclear fractions of
transfected COS-7 cells at five different concentrations of
1,25-(OH)[
H]D
. In studies
such as the one pictured in Fig. 5A, we have found that
saturation of wild-type hVDR even at low, endogenous receptor levels
occurs only at relatively high concentrations of
1,25-(OH)
D
ligand, i.e.
5
nM(37) . All three mutant hVDRs showed reduced nuclear
uptake of the ligand in the intact COS-7 cells. Significantly, C288G
demonstrated the lowest uptake, which was approximately 5% of wild-type
levels; this finding correlates with the requirement for very high
concentration of hormonal ligand (10M) to
stimulate receptor-mediated transcription for this mutant (Fig. 4B). C337G, which paradoxically exhibited higher
ligand binding activity, in vitro, at 4 °C as compared
with the wild-type receptor, but a shift to the right in the
transcriptional activity dose-response curve with
1,25-(OH)
D
at 37 °C (Fig. 4B), displayed markedly reduced nuclear uptake of
1,25-(OH)
D
. Because nuclear partitioning of all
mutant hVDR proteins appears normal (Fig. 3), these data suggest
that 1,25-(OH)
D
ligand binding activity, in
vivo, at 37 °C is attenuated by the mutation at cysteine 337
and to a greater degree by altering cysteine 288. The results
illustrated in Fig. 5A also reveal a small but
significant suppression of nuclear uptake of ligand at 37 °C for
the C369G receptor. Again, this slight defect appears to be in concert
with the 70% increase in EC
for
1,25-(OH)
D
-stimulated transcription with this
mutant as noted in Fig. 4B and described in the
accompanying text. Therefore, the level of nuclear uptake of the
1,25-(OH)
D
ligand by COS-7 cells transfected
with the various mutant hVDRs (Fig. 5A) correlates with
the EC
values for
1,25-(OH)
D
-enhanced transcriptional activity of
these mutant receptors (Fig. 4B), thereby demonstrating
that reduced hormone binding is the major source of attenuation in
transcriptional activation exhibited by the C288G and C337G mutants.
[H]D in intact
cells at 37 °C and in cellular extracts at 23 °C, in
vitro. A, nuclear uptake of
1,25-(OH)[H]D by COS-7
cells expressing wild-type or mutant hVDRs. COS-7 cells expressing
hVDRs were harvested, resuspended in culture medium containing 1% fetal
bovine serum, and incubated with five concentrations of
1,25-(OH)[H]D in the
presence or absence of a 600-fold molar excess of unlabeled
1,25-(OH)D at 37 °C for 2 h. B,
specific binding of
1,25-(OH)[H]D to extracts
of COS-7 cells transfected with hVDR and cysteine point mutants.
Incubations were carried out for 2 h at 23 °C in the presence of
4.3 nM labeled ligand ± a 400-fold molar excess of
radioinert 1,25-(OH)D to obtain specific
binding. The binding shown is corrected for level of
expression/degradation as determined by Western
blotting.
D binding experiment with cellular
extracts at the intermediate temperature of 23 °C. Incubation of
extracted VDR at this elevated temperature was found to elicit
degradation (data not shown), so we were limited to the relatively
short incubation time of 2 h to preserve the receptor. Under these
conditions, saturation kinetics were not achieved, precluding the
determination of K values. However, specific
binding levels at 23 °C for each mutant at a
1,25-(OH)D ligand concentration of 4.3 nM (Fig. 5B), corrected for receptor expression by
normalizing the results to the signals from a Western blot performed
after a 23 °C incubation (data not shown), strongly support the
conclusion that C369G binds 1,25-(OH)D reasonably well at elevated temperatures while C337G and
especially C288G hVDRs are defective in ligand binding.
D. With
active analogs utilized for affinity labeling of the receptor, such as
dexamethasone 21-mesylate, Cys-656 of the rat GR (corresponding to
Cys-644 of the mouse GR) was identified as a critical residue for
ligand binding(16, 17, 18) . Utilizing
arsenite as a thiol bridging reagent, it has been demonstrated that
Cys-656 and -661 of rat GR are important for hormone
binding(20) . In the case of ER, Harlow et al. (22) reported that both an estrogen agonist and an estrogen
antagonist covalently bind to Cys-530 in the HBD of human ER.D ligand (Fig. 4, A and B), but is not required for heterodimeric association
of hVDR with RXR on the VDRE (Fig. 2). Because significant
transcriptional activation can be generated when cells expressing C288G
hVDR are treated with the high dose of 10M 1,25-(OH)
D (Fig. 4B),
cysteine 288 does not seem to be as crucial for transactivation, per se, as are residues 403-427(12) . In
contrast, cysteine 369 is not critical for high affinity hormone
binding at 4 °C, and mutation of this residue results in only minor
suppressions of 1,25-(OH)D nuclear uptake (Fig. 5A), ligand binding at 23 °C (Fig. 5B), and possibly of heterodimerization with RXRs (Fig. 2); these effects are manifest as a small but significant
shift in the dose-response curve with respect to transactivation (Fig. 4B). Finally, alteration of cysteine 337 to
glycine elicits a paradoxical enhancement of ligand binding at 4 °C
and a minor attenuation of RXR heterodimerization capacity (Fig. 2), but results in a significant diminution in hVDR
transactivation function (Fig. 4, A and B),
the latter finding most likely being explained by relatively
ineffective 1,25-(OH)D ligand binding at
23-37 °C (Fig. 5, A and B). Strong
evidence supporting this conclusion is provided by the fact that
transactivation by C337G is restored to normal in the presence of
10M 1,25-(OH)
D (Fig. 4B). Mutations such as C337G, therefore,
appear to result in a temperature-dependent phenotype for as yet
unexplained reasons and reveal a necessity for functional testing at or
near physiological temperatures. This concept is further illustrated in
the case of C288G. When cell extracts were assayed by traditional
ligand binding, at 4 °C, in vitro, the Cys-288 mutant
exhibited approximately one-third of the binding affinity of the
wild-type receptor. Yet this mutant VDR displayed only 5% of wild-type
nuclear uptake of the hormonal ligand in vivo, at 37 °C (Fig. 5A). Although there are several possibilities to
explain these results, including: i) instability and degradation of the
receptor protein at physiologic temperatures, ii) attenuated nuclear
translocation of the receptor, and iii) weaker binding to the ligand in vivo; the first two possibilities are not likely because
nuclear fractions from COS-7 cells expressing the C288G mutant receptor
contained a similar amount of intact hVDR protein compared with the
cells expressing wild-type hVDR (Fig. 3). Thus, we again
conclude that there is a temperature-sensitive defect, in this case in
the hormone binding activity of the C288G mutant. (42) appears to provide a prototype for this region in nuclear
receptors. The RXR ligand binding domain consists of an
antiparallel -helical sandwich containing 11 -helices
surrounding two -strands(42) , and the proposed ligand
binding pocket is a hydrophobic cavity bordered by helix 5, both
-strands, helix 7, the COOH-terminal portion of helix 10 and the
NH-terminal part of helix 11. That a similar ligand binding
pocket may exist in the other members of the nuclear receptor
superfamily is suggested not only by the homologies seen in this region
(typically 20-30% across the superfamily), but also by
mutagenesis studies with ER and GR. Katzenellenbogen et al. (43) have previously suggested that Cys-381 and Cys-530 lie at
the ``mouth'' of a putative ligand binding pocket; these two
residues in fact correspond to positions in human RXR within helix
5 and the NH-terminal portion of helix 11, respectively.
The participation of the two -strands and helix 7 in a generalized
hormone binding site is confirmed by the findings of Chakraborti et
al.(20) , who implicate Cys-640 (1st -strand), as
well as Cys-656 and Cys-661 (both in helix 7), as being important for
ligand binding by rat GR. In addition, a previous report describing the
effect of an artifactual mutation in the cloned human ER from MCF-7
cells (44) indicates that Gly-400 raises the dose of
estradiol-17 required for maximal transactivation by
10-100-fold; this residue is also located in the region of ER
that corresponds to the second -strand in RXR. More recent
site-directed mutagenesis of the mouse GR (45) points to the
significance of Cys-742 (COOH-terminal portion of helix 10) in
ligand-dependent transcriptional activation. The location of all of
these residues implicated in hormone-binding or hormone-dependent
functions of the respective receptors seems in complete agreement with
the proposed prototypical hydrophobic binding pocket.-strand in
RXR and in helix 8, respectively. Cys-288 would therefore take its
place along with Cys-640 in rat GR and Gly-400 in human ER as
confirming the general importance of the -strand region in ligand
binding. Furthermore, recently reported natural mutations of hVDR which
display impaired hormone binding lie in helix 5 (Arg-274 (46) )
and helix 7 (Ile-314(47) ), both critical regions in RXR ligand
association. In contrast, Cys-337 resides in an area of hVDR
corresponding to helix 8, which places it outside the proposed ligand
binding pocket. However, mutations at analogous positions in the human
ER at Cys-447 (48) and in the mouse GR at Cys-671 (45) result in impaired ligand-induced transcriptional
activation at physiological temperatures. Because these mutations have
marked effects on hormone binding and transactivation, it is suggested
that helix 8, which lies adjacent and parallel to helix 5 in the
structure of RXR, might somehow be important in stabilizing the
conformation of the ligand binding cavity. Thus, the fact that RXR, ER,
GR, and VDR represent widely diverse members of the nuclear receptor
superfamily argues strongly that many features of the proposed ligand
binding pocket may be shared across the nuclear receptor superfamily. D-occupied, hormone binding domain of
VDR, such as through x-ray crystallography, will be necessary to
elucidate the mechanism of 1,25-(OH)D ligand
binding and how this can influence the control of gene transcription.
) receptor; hVDR, human VDR;
1,25-(OH)D, 1,25-dihydroxyvitamin
D; HBD, hormone binding domain; VDRE, vitamin D responsive
element; RXR, retinoid X receptor; GR, glucocorticoid receptor; ER,
estrogen receptor; DMEM, Dulbecco's modified Eagle's
medium.
We thank Drs. Ronald M. Evans and David J. Mangelsdorf
for the generous gift of the human RXR and mouse RXR cDNAs.
We also gratefully acknowledge Mina C. Reeder, Michelle Thatcher, and
Sanford H. Selznick for their technical assistance.
©1996 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  G. JONES, S. A. STRUGNELL, and H. F. DeLUCA Current Understanding of the Molecular Actions of Vitamin D Physiol Rev, October 1, 1998; 78(4): 1193 - 1231. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. C. Li, C. Bergwitz, H. Juppner, and M. B. Demay Cloning and Characterization of the Vitamin D Receptor from Xenopus laevis Endocrinology, June 1, 1997; 138(6): 2347 - 2353. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| All ASBMB Journals | Molecular and Cellular Proteomics |
| Journal of Lipid Research | ASBMB Today |