| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Articles |
Departments of Biochemistry (C.J.T., G.L.D.) and Medicine (S.J.U.), East Carolina University School of Medicine, Greenville, North Carolina 27858; and the Department of Physiology and Biophysics, University of Iowa (J.E.P.), Iowa City, Iowa 52242
Address all correspondence and requests for reprints to: Christopher J. Torrance, Ph.D., Department of Oncology, Johns Hopkins University, Baltimore, Maryland 21231.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In conclusion, these data demonstrated the presence of a specific TR-binding site within a T3-responsive region of the rat GLUT4 promoter and thus support the supposition that thyroid hormone acts directly to stimulate GLUT4 transcription in rat skeletal muscle. Moreover, characterization of a novel TR-binding site with low affinity suggests an additional mechanism by which the intrinsic activity and responsiveness of thyroid hormone regulated genes may be modulated.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The molecular defect in glucose disposal leading to NIDDM has been suggested to reside within the insulin signaling pathway (5, 6, 7), although cause and effect have not been established in this disease. Nevertheless, elucidating positive regulators of GLUT4 as well as their mechanism(s) of action are of considerable interest as a possible treatment in NIDDM. Overexpression of GLUT4 in transgenic diabetic mice has been shown to be highly effective in ameliorating postprandial hyperglycemia, primarily by stimulating basal (noninsulin-mediated) glucose disposal (8). Previous studies from our laboratory (9) and that of Weinstein et al. (10) have demonstrated that thyroid hormone (T3) stimulates basal and, to some extent, insulin-mediated glucose uptake in rat skeletal muscle. The mechanism for this stimulatory effect of T3 was determined to be due predominantly to an induction of GLUT4 protein expression (9, 10). In the preceding manuscript, GLUT4 induction by T3 was further defined to be primarily via transcriptional induction in red muscle, and a separate translational and/or posttranslational mechanism in white muscle fiber types (11). Moreover, in a previous study, constructs containing various regions of the GLUT4 upstream of a reporter gene, when transfected into C2C12 myotubes and treated with T3, isolated a 281-bp region responsive to thyroid hormone (12). However, further experiments are required to determine whether T3 acts directly on GLUT4 transcription and whether transcriptional induction of skeletal muscle GLUT4 by T3 or a previously suggested (11) muscle-selective T3-receptor agonist represents a viable therapeutic strategy in NIDDM.
The effects of T3 on gene transcription are
mediated directly via a family of nuclear receptor/transcription
factors: c-erbA
1, -ß1, and -ß2 (13, 14, 15, 16, 17, 18, 19). These
thyroid hormone receptors (TRs) bind to specific thyroid hormone
response elements (TREs) consisting of hexameric half-sites
[consensus: AGGT(C/A)A], orientated in singlet or multiplex
configurations within the promoter regions of target genes. A limited
number of TREs have been elucidated; many consist of directly repeated
half-sites separated by 4 bp (DR+4) (13), whereas, an everted half-site
TRE separated by 6 bp (F2) has been reported to function as a silencer
in the chick lysozyme promoter (20, 21). TRs bind with high affinity to
their cognate DNA elements in both the presence and absence of
T3; however, TRs only function to stimulate (on
positive TREs) or repress (on negative TREs) gene transcription in
response to binding thyroid hormone (13, 14, 15, 17, 18, 19). In contrast,
unoccupied TRs on positive TREs (e.g. DR+4 elements)
actively repress basal transcription (22, 23, 24, 25, 26, 27).
TRs bind to composite TREs as monomers; however, cooperative interactions favor the formation of homodimers. Moreover, TRs interact with coregulatory proteins or thyroid hormone auxiliary proteins in vivo, some of which have been characterized [e.g. the cis-retinoic acid receptors (RXRs)] (28, 29, 30, 31, 32, 33, 34). TR/RXR heterodimers demonstrate higher binding affinities than TR homodimers, are favored in the presence of T3, and produce larger increases in T3-induced gene expression (13, 14, 32, 33). Indeed, heterodimers are considered to be the primary complexes mediating T3-regulated gene expression in vivo.
The aim of the present study was to determine and characterize a putative TRE within the rat GLUT4 promoter and thus perhaps establish whether T3 acts directly to stimulate GLUT4 transcription. Sequential electrophoretic mobility shift assay (EMSA) analyses using a previously described 281-bp T3-responsive region of the GLUT4 promoter (12) identified a TR-binding site, orientated in the classical DR+4 motif, between bases -450 and -434. However, in contrast to other established TREs, the GLUT4 TR-binding element demonstrated a significantly lower affinity, and as a likely consequence only bound TRs appreciatively in the presence of RXR. Therefore, these data are also suggestive of an additional level of complexity that may govern the inducibility and responsiveness of T3-regulated genes.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1 in pCMV (35) and
for RXR and retinoic acid receptor (RAR) complementary DNAs
in pSKXR3–1 (36) were used to program TNT T7-coupled reticulocyte
lysates (Promega, Madison, WI). Unlabeled and
[35S]methionine-labeled receptors were synthesized in
parallel using in vitro translation reactions, and the
protein concentrations were determined from trichloroacetic
acid-precipitable counts according to the manufacturer’s instructions.
The synthesis of the human 1 (h1) TR, however, is not as
efficient as those for ß1 and RXR and thus could not be used in
sufficient amounts to give optimal results in EMSAs. In some
experiments purified chick 1 TR was used to visualize 1 complexes
more readily. This receptor was the gift of Dr. H. Samuels (New York
University Medical Center, New York, NY). The isolation procedure for
chick 1 has been described previously (37) and was used here in
approximately equimolar ratios with hRXR. A plasmid containing a 281-bp EcoNI-BstXI fragment of the GLUT4 promoter (bases -517 to -238), which accounts for full thyroid hormone responsiveness of 2212 bp from the transcriptional start site of GLUT4, was obtained from Dr. J. Pessin (University of Iowa, Iowa City, IA). Various double stranded DNA oligomers spanning this region were also designed to further localize the putative GLUT4 TRE. These were synthesized at the East Carolina University School of Medicine DNA-Core Facility.
Antibodies
The N-terminal polyclonal antibody specific for rat ß1
(rß1PAb, amino acids 62–92 in rat ß1) and the C-terminal 1/ß
cross-specific polyclonal antibody (r1/ßPAb, amino acids 447–461
and 393–407 in rat ß1 and 1, respectively) were provided by Drs.
J. Oppenheimer and H. Schwartz (University of Minnesota, Minneapolis,
MN). The hRXR polyclonal antibody (RXRPAb) was the generous gift
of Dr. R. Evans (The Salk Institute, San Diego, CA).
Isolation and extraction of nuclei from rat skeletal muscle
Nuclei were isolated from rat skeletal muscle in the presence of
protease inhibitors (0.5 mM phenylmethylsulfonylfluoride,
10 µg/ml leupeptin, 1 µg/ml aprotinin, and 10 mM EDTA)
by the previously described method of Neufer et al. (38).
Approximately 500 µg purified rat muscle nuclei were extracted using
100 µl cold (4 C) M4 extraction buffer [20 mM HEPES (pH
7.8), 0.4 M KCl, 2 mM dithiothreitol, and 20%
glycerol] for 30 min (plus protease inhibitors), and the nuclear
debris and DNA were pelleted at 12,000 rpm in a microfuge for 15 min.
Nuclear extracts were assayed for protein concentration by the method
of Bradford (39) and were used either immediately for analysis or
stored at -80 C until required.
EMSAs
EMSAs were performed as previously described (40). Briefly,
approximately 100 ng (60,000 cpm) of a 32P-labeled double
stranded DNA oligomer were incubated with either 1 µl (10 fmol)
reticulocyte lysate synthesized receptor (1 µl hß1 and/or hRXR)
or 5 µg nuclear extract for 30 min at room temperature in a total
volume of 20 µl reaction buffer [25 mM Tris (pH 7.8),
0.5 mM EDTA, 88 mM KCl, 1 mM
dithiothreitol, 150 µg/ml poly(dI·dC)-poly(dI·dC), 0.05% Triton
X-100, and 12.5 µg/ml sonicated salmon sperm DNA]. For supershift
experiments, 1 µl polyclonal antisera was incubated with in
vitro translated receptors for 30 min at room temperature before
addition of the DNA oligomer. Reaction mixtures were subjected to
electrophoresis on a 5% nondenaturing polyacrylamide gel at 4 C and 40
mA (200 V) for approximately 2.5 h, and the dried gels were
exposed with intensifying screens to x-ray film for 16–48 h at -80
C.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
),
previously delineated in transient transfection studies to be
T3 responsive (12). EMSA using this fragment was
sufficient to first establish the presence of a TR-binding site within
the GLUT4 promoter. The binding of reticulocyte lysate-synthesized 1
and ß1 TR/RXR heterodimers to this 281-bp fragment is shown in the
four left lanes of Fig. 2 (complex
composition confirmed in Fig. 6). In addition, complexes were formed
using rat skeletal muscle nuclear extracts, in particular one that
migrates closely with the human in vitro translated
1/RXR heterodimer (Fig. 2, left lanes 4 and 2,
respectively) and one highly abundant species with very low mobility.
The lesser degree of h1 TR binding to the GLUT4 sequence as well as
other TREs used throughout this report is due to the lower efficiency
of the in vitro translation reaction for h1, such that
equimolar ratios with hß1 and RXR could not be used in this
assay.
|
|
|
). In addition, consistent with the presence of a MyoD and/or
myocyte enhancer factor-2 (MEF-2) consensus binding site within the
5'-restriction fragment (Fig. 1), both the full-length and 5'-AlwNI
sequences bound low mobility complex produced with muscle nuclear
extracts (left and right four lanes of Fig. 2, respectively). Based on these data, three oligomers were synthesized to
span the 5'-AlwN1 restriction fragment: oligo 1 (bases -526/-485),
oligo 2 (bases -491/-452), and oligo 3 (bases -460/-418) (Fig. 1),
to further define the position of the GLUT4 TRE as well as establish
the putative binding of MyoD and/or MEF activities within this thyroid
hormone-responsive region of the GLUT4 promoter. Specifically, oligo 1
encompassed the three upstream putative TRE sequences, oligo 2 covered
the MyoD and MEF consensus binding sites, and oligo 3 spanned the
remaining near-consensus TRE sequence within this AlwN1 restriction
fragment (Fig. 1). Oligo 3 was not synthesized to cover the residual
downstream sequences in the 5'-AlwN1 fragment, because this region is
simply composed of a stretch of AC microsatellite dinucleotide repeats
and thus is highly unlikely to contain a specific TR-binding
element.
The bindings of in vitro translated TRs and skeletal muscle
nuclear extracts to oligo 1 and a previously characterized TRE (the
chicken lysozyme F2 TRE) (20, 21) were compared in Fig. 3. These data clearly demonstrated that bases
-526/-485 of the GLUT4 promoter do not contain a TR-binding element,
although the putative TRE half-sites within this region (highlighted in
Fig. 1) are close to the consensus AGGT(C/A)A. The experiment shown in
Fig. 4, left panel, directly compared the
binding of TRs and skeletal muscle nuclear extracts to oligos 2 and 3.
Consistent with the respective putative consensus elements within these
oligos (Fig. 1), binding of in vitro translated TRs was
localized to oligo 3 (bases -460/-418), whereas the postulated MEF-2
and/or MyoD binding complex was defined within oligo 2. Moreover,
competition of the oligo 2 binding species with a large excess of an
unlabeled MEF-2, but not a MyoD consensus binding site, confirmed this
DNA-binding species to be composed solely of MEF-2 (Fig. 4, right
panel).
|
|
, solid and dashed arrows,
and Fig. 5, top panel). Therefore, to further
define the boundaries and orientation of the GLUT4 TRE, two additional
oligomers were synthesized with mutations designed to abrogate TR
binding dependent upon which sequence represented the true TR-binding
site (Fig. 5). Mutations within the 5'- to 3'-orientated putative TRE
were shown to abolish TR binding, whereas mutations within the
oppositely orientated sequence did not (Fig. 5). Therefore, these data
unambiguously established the location of the GLUT4 TR-binding site
between bases -457/-426, composed in a classical DR+4 arrangement
orientated in the 5' to 3' direction.
|
, lanes 1–6) first demonstrated that the TR complexes
formed using in vitro translated TRs and hRXR on the
GLUT4 TRE were composed of TR/RXR heterodimers. Moreover, the
successful competition of these species with a large excess of an
unlabeled oligo containing the F2 TRE, but not with a mutated
(non-TR-binding) version of this element (M2; Fig. 6, lanes 7–12),
demonstrated the specificity of the GLUT 4 sequence for TRs.
All EMSA experiments described thus far were performed for the most
part by coincubating in vitro translated TRs with hRXR,
principally with the intent of visualizing the maximum possible number
of TR/TRE interactions. However, only a single species has consistently
been observed in these assays, i.e. DNA binding was unable
to be demonstrated by TRs in the absence of hRXR. These data,
therefore, suggest that binding of TRs to the GLUT4 TRE may be
dependent upon heterodimerization. Hence, the GLUT4 TRE was next
compared to a canonical DR+4 TRE sequence composed of perfect consensus
half-sites to determine the relative ability of these sequences to bind
TRs (Fig. 7; both elements labeled to identical specific activities).
Complexes on both elements demonstrated the characteristic downshift of
heterodimeric TR/TRE complexes and dissociation of TR homodimers in the
presence of T3 (Fig. 7).
However, TR binding to the GLUT4 TRE was overall considerably lower,
and as a likely consequence, homodimer and even monomeric binding
evident for ß1 TRs on the canonical DR+4 element, was not detectable
on the GLUT4 TR-binding site.
|
to the GLUT4 and canonical DR+4 TR-binding sites. The results of this
experiment (Fig. 8) demonstrated that unlike the
canonical DR+4, only at the highest ß1TR concentration was TR
homodimer marginally apparent on the GLUT4 TRE.
|
). Specifically, these mutations were
designed to replace nonconsensus half-site bases with consensus
nucleotides. In addition, various substitutions were introduced within
the intervening and flanking regions of the GLUT 4 TR-binding element,
e.g. mutant oligo 7 was synthesized, replacing all the
intervening and flanking bases with those present in the canonical DR+4
TRE.
|
receptors to these mutant GLUT4 TREs are shown in Fig. 9.
Interestingly, bases within the intervening and/or flanking sequences
were also shown to be crucial for TR binding; in particular intervening
bases 2 and 3. In contrast, the mutations in oligo 5 and oligo 6 (which
converts the first base of the 3'-TRE half-site to the consensus)
apparently imparted an additional affinity to these sequences, the
degree of which and perhaps complex formation on would seem to be
comparable with those of the canonical DR+4 TRE. Therefore, in
combination with the previous titration experiment, these data suggest
that it is the overall lower affinity of the GLUT4 sequence that
accounts for the visualization of only the more avid heterodimer in
these experiments, i.e. a quantitative, rather than
qualitative, phenomenon. This conclusion would also be supported by the
reduced binding and formation of only heterodimer observed by mutant 7,
wherein all the flanking and intervening bases were substituted with
those from the canonical TR-binding element. In conclusion, these data demonstrated that the newly identified GLUT4 TRE has a specific affinity for TRs. However, in contrast to other established TREs, GLUT4 TRE was shown to have a much lower affinity, and this property appears to be responsible for the observation that only high affinity TR/heterodimeric receptors bind appreciatively on this TRE. Moreover, combined with the results of a previous study that demonstrated this region of the GLUT4 promoter to confer thyroid hormone responsiveness in transfection studies (12), these data indicate that the effects of T3 on GLUT 4 transcription are likely to be mediated directly by this TR-binding element. However, formal proof of this hypothesis requires additional functional analyses using mutations within this TR-binding element.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The exquisite specificity of base sequences directing protein-DNA
interaction(s) was similarly demonstrated in the mutation experiment,
which revealed that intervening and flanking sequences surrounding the
GLUT4 half-sites were also crucial for mediating TR binding. Indeed, we
may have been fortunate in that the mutations introduced within the
intervening and flanking regions to discern the orientation of the
GLUT4 TRE did not also abrogate TR binding. However, none of the
mutations analyzed in Fig. 9 directly altered the
propensity of the GLUT4 TRE to bind heterodimers. Moreover, certain
substitutions (in particular the conversion of the first base in the 3'
half-site to the consensus adenine; mutant oligo 6) that enhanced the
affinity of the GLUT4 TRE, resulted in binding properties similar to
those of the canonical DR+4 TRE. Of note, increased binding to the
mutation oligo 6 would be consistent with the fact that TRs in
heterodimeric complexes are considered to bind this particular
half-site. In contrast, mutant oligo 7, in which all bases other than
those within the GLUT4 TRE half-sites were replaced with the
corresponding sequences in the canonical DR+4, demonstrated properties
comparable to those of the wild-type GLUT4 TRE. These data, therefore,
suggest that the apparent heterodimer specificity observed for the
GLUT4 TR-binding site is simply a function of the overall lower
affinity of the half-sites within this element. Indeed, the titration
experiment demonstrated the GLUT4 sequence to be equally deficient in
binding both homodimers and heterodimers compared to a canonical DR+4
TR-binding element (estimated to be approximately 5-fold from
densitometric scanning).
Finally, characterization of a low affinity TR-binding site has a
number of implications for thyroid hormone-regulated genes, the first
of which would presumably be the inducibility of responsive genes.
Indeed, the degree of transcriptional induction observed for the GLUT4
gene (
2.5 under chronic hyperthyroid vs. hypothyroid
conditions) (11) compared to genes containing high affinity TRE(s)
(e.g. malic enzyme; transcription stimulated 3- to 4-fold in
the liver of euthyroid rats) (43) is consistent with this hypothesis.
Secondly, as 1) TRs bind DNA in the presence and absence of thyroid
hormone, 2) unoccupied TRs actively repress basal transcription, and 3)
heterodimers demonstrate higher affinities over homodimers and are
favored in the presence of T3 (13), a low
affinity TRE would have special relevance for thyroid hormone-regulated
genes. For example, one can speculate that although a low affinity TRE
would not illicit as large a T3 induction
compared to a high affinity TRE, by the same token, in the absence of
T3 a low affinity TRE would not be subjected to
such a great degree of basal repression.
A previous study suggested a criterion by which different combinations of weak and strong artificial TRE half-sites modulate TR complex formation and T3 responsiveness of thyroid hormone-regulated genes (44). In this classification, the GLUT4 TRE presumably falls into the category of either a weak/weak (w/w) or, more likely from the high affinity imparted to the mutant 6, a strong/weak half-site composition. Interestingly, consistent with our observations, both of these previously described TREs were weakly T3 responsive (44). Moreover, distinct from all other TRE combinations, these elements were also dependent upon RXR to mediate T3 induction in transfection assays. Unfortunately, however, the basal repression properties of a w/w or strong/weak TRE were not investigated in this study (44).
Interestingly, both the half-site sequences, the immediately flanking
bases, and three of the four intervening sequences of the GLUT4 TRE are
absolutely conserved among humans, mice, and rats (Fig. 10). Their
conservation would, therefore, further suggest the probable
functionality of this low affinity TRE as well as the importance of
sequences other than those within the TR half-sites for mediating TR
binding. Indeed, one can envisage that during the evolution of genes,
the properties of promoter regulatory elements, e.g.
affinity, may be similarly selected so that they compliment the
function of the particular gene product, i.e. for a TRE; a
low affinity TRE(s) may predominate in genes that need to be
constitutively expressed, but under certain conditions need to be
up-regulated to some extent, e.g. GLUT4 (11). In contrast, a
higher affinity TREs would be highly responsive to thyroid hormone, but
would also be subjected to a greater degree of active repression by
unoccupied heterodimers and/or homodimers. Such a TRE would, therefore,
be highly sensitive to changes in T3
status/availability and may be advantageous for imparting high
inducibility on, for instance, temporally expressed genes.
|
|
Acknowledgments |
|---|
Received July 1, 1996.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  P. de Lange, R. Senese, F. Cioffi, M. Moreno, A. Lombardi, E. Silvestri, F. Goglia, and A. Lanni Rapid Activation by 3,5,3'-L-Triiodothyronine of Adenosine 5'-Monophosphate-Activated Protein Kinase/Acetyl-Coenzyme A Carboxylase and Akt/Protein Kinase B Signaling Pathways: Relation to Changes in Fuel Metabolism and Myosin Heavy-Chain Protein Content in Rat Gastrocnemius Muscle in Vivo Endocrinology, December 1, 2008; 149(12): 6462 - 6470. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Karnieli and M. Armoni Transcriptional regulation of the insulin-responsive glucose transporter GLUT4 gene: from physiology to pathology Am J Physiol Endocrinol Metab, July 1, 2008; 295(1): E38 - E45. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. S. Malo, W. Zhang, F. Alkhoury, P. Pushpakaran, M. A. Abedrapo, M. Mozumder, E. Fleming, A. Siddique, J. W. Henderson, and R. A. Hodin Thyroid Hormone Positively Regulates the Enterocyte Differentiation Marker Intestinal Alkaline Phosphatase Gene via an Atypical Response Element Mol. Endocrinol., August 1, 2004; 18(8): 1941 - 1962. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Moreno, A. L. Serrano, T. Santalucia, A. Guma, C. Canto, N. J. Brand, M. Palacin, S. Schiaffino, and A. Zorzano Differential Regulation of the Muscle-specific GLUT4 Enhancer in Regenerating and Adult Skeletal Muscle J. Biol. Chem., October 17, 2003; 278(42): 40557 - 40564. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. P. Schroder-van der Elst, D. van der Heide, J. Kastelijn, B. Rousset, and M. Jesus Obregon The Expression of the Sodium/Iodide Symporter Is Up-Regulated in the Thyroid of Fetuses of Iodine-Deficient Rats Endocrinology, September 1, 2001; 142(9): 3736 - 3741. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Santalucia, K. R. Boheler, N. J. Brand, U. Sahye, C. Fandos, F. Vinals, J. Ferre, X. Testar, M. Palacin, and A. Zorzano Factors Involved in GLUT-1 Glucose Transporter Gene Transcription in Cardiac Muscle J. Biol. Chem., June 18, 1999; 274(25): 17626 - 17634. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
