help button home button Endocrine Society Molecular Endocrinology ENDO 08 Sessions Library
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Molecular Endocrinology, doi:10.1210/me.2003-0289
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow NURSA Molecule Pages Link
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Mengeling, B. J.
Right arrow Articles by Privalsky, M. L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Mengeling, B. J.
Right arrow Articles by Privalsky, M. L.
Molecular Endocrinology 19 (1): 35-51
Copyright © 2005 by The Endocrine Society

Novel Mode of Deoxyribonucleic Acid Recognition by Thyroid Hormone Receptors: Thyroid Hormone Receptor ß-Isoforms Can Bind as Trimers to Natural Response Elements Comprised of Reiterated Half-Sites

Brenda J. Mengeling, Fan Pan and Martin L. Privalsky

Section of Microbiology, Division of Biological Sciences, University of California, Davis, California 95616

Address all correspondence and requests for reprints to: Martin L. Privalsky, Section of Microbiology, Division of Biological Sciences, One Shields Avenue, University of California at Davis, Davis, California 95616. E-mail: mlprivalsky{at}ucdavis.edu.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
Thyroid hormone receptors (TRs) regulate gene expression by binding to specific DNA sequences, denoted thyroid hormone response elements (TREs). The accepted paradigm for TRs proposes that they bind as homo- or heterodimers to TREs comprised of two AGGTCA half-site sequences. In the prototypic TRE, these half-sites are arranged as direct repeats separated by a four-base spacer. This dimeric model of TR binding, derived from analysis of artificial DNA sequences, fails to explain why many natural TREs contain more than two half-sites. Therefore, we investigated the ability of different TR isoforms to bind to TREs possessing three or more half-sites. We report that the TRß isoforms (TRß0, TRß1, TRß2), but not TR{alpha}1, can bind to reiterated DNA elements, such as the rat GH-TRE, as complexes trimeric or greater in size. The TRß0 isoform, in particular, formed homo- and heterotrimers (with the retinoid X receptor) with high efficiency and cooperativity, and TRß0 preferentially used reporters containing these reiterated elements to drive gene expression in vivo. Our data demonstrate that TRß isoforms can form multimeric receptor complexes on appropriately reiterated DNA response elements, providing a functional distinction between the TR isoforms and an explanation for TREs possessing three or more half-sites.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
THYROID HORMONE RECEPTORS (TRs) are members of the nuclear receptor family of ligand-regulated transcription factors (1). TRs play multiple roles in vertebrate homeostasis and differentiation, including control of general metabolic rate, glucose utilization, thermogenesis, cardiac output, lipid metabolism, central nervous system development, and amphibian metamorphosis (2, 3, 4, 5, 6, 7). Genetic disruptions have demonstrated the importance of TRs for proper skeletal, cochlea, cardiac, and intestinal development (8, 9, 10, 11, 12, 13). Close nuclear receptor relatives of TR include the retinoic acid receptors (RARs) and vitamin D3 receptor (VDR). Each of these receptors regulates different genes in vivo, yet share very similar DNA recognition domains (14, 15, 16). Studies in vitro have shown that all three of these receptor classes are able to bind to the same six-nucleotide consensus DNA sequence (half-site) of AGGTCA. These receptors bind to DNA primarily as protein dimers (either as homodimers or heterodimers with the retinoid X receptor, RXR), and therefore a functional hormone response element is considered to consist of two AGGTCA half-sites. Both the orientation and spacing of these half-sites are believed to determine which response elements are recognized by which receptor. Two AGGTCA half-sites oriented as a direct repeat with a three-base spacer [a direct repeat (DR) 3] selects for VDRs, a DR4 of the same half-sites for TRs (Fig. 1AGo), and a DR5 for RARs (17, 18). This 3-4-5 rule is not absolute, however; VDRs, TRs, and RARs can bind to, and regulate gene expression from DNA elements containing nonconsensus half-sites or comprised of half-site spacings and orientations that differ significantly from these model elements (19, 20, 21, 22). The existence of additional members of the nuclear receptor family with DNA recognition properties that overlap these DR-3, -4, -5 spacings further questions how these different nuclear receptors display the specificity necessary to regulate only their appropriate target genes in vivo.



View larger version (28K):
[in this window]
[in a new window]
  		Fig. 1. Structure of the DR4 and rGH TRE and Schematic of the Different TR Isoforms

A, The nucleotide sequence of a consensus DR4 and the native rGH TRE are shown, with the three hexanucleotide half-site sequences (A–C) highlighted; arrows beneath indicate relative orientation. B, The four predominant TR isoforms are illustrated from N to C terminus. The locations of the DNA binding and hormone binding domains within these receptors are indicated. The TRß0 and TRß2 isoforms (TRß1 and TRß2 in mammals) are encoded from a single genetic locus by alternative mRNA splicing, and differ only in their amino-terminal A/B domains as indicated. The TR{alpha}1 isoform is encoded by a distinct locus and differs from the ß isoforms at multiple positions throughout the open reading frame.

 
Many of the target genes regulated by TRs have thyroid hormone response elements (TREs) that contain more than two half-sites in close proximity (23, 24). The rat GH gene (rGH) promoter has served as a prototype of this type of reiterated TRE. At the core of the rGH element is a TRE consisting of two half-sites oriented as a DR4, followed in turn by a third half-site in an inverted repeat (INV) orientation with a one base spacer (an INV1; Fig. 1AGo) (25, 26, 27, 28, 29). Brent and colleagues (25) found that all three of the half-sites in the rGH core element contribute to reporter gene regulation by TR receptors in vivo, but that paradoxically TR{alpha}1 binds to the rGH element primarily as a receptor dimer when tested in vitro. Receptor occupancy at all three half-sites could only be demonstrated on the rGH element at artificially high TR{alpha}1 concentrations, was noncooperative (i.e. occurred only after saturation of monomer and dimer binding sites) and was disrupted by addition of RXR, resulting in the preferential formation of TR/RXR heterodimers. Based on these results, most subsequent investigations have focused on the DR4 as the principal operative component of the rGH element, and the role of the third half site has remained a mystery.

In addition to TR{alpha}1, a distinct genetic locus, denoted TRß, also encodes TRs (30, 31). Further TRß isoform diversity arises as a result of alternative mRNA splicing (32, 33, 34). In mammals, two predominant TRß isoforms are produced in a tissue-specific manner: TRß1 and TRß2 (the latter is found primarily in the pituitary and hypothalamus). These isoforms differ only in their amino-terminal (A/B) domains; their DNA binding and ligand binding domains are identical (Fig. 1BGo). In birds, reptiles, and amphibians, a TRß0 isoform, bearing a more truncated A/B domain, appears to substitute for the TRß1 form.

We previously noted that some, but not all, RXR isoforms can form receptor homotrimers and homotetramers on suitably reiterated response elements (35, 36). Because prior studies of the rGH TRE have focused principally on the TR{alpha}1 isoform, we examined whether the TRß isoforms might differ from TR{alpha}1 and recognize the reiterated rGH element as a receptor multimer. We report here that, in contrast to TR{alpha}1, the TRß isoforms are able to form homo- and heterotrimeric complexes cooperatively on the rGH TRE. The extent of oligomer binding depended upon the specific TRß isoform used and on the orientation and spacing of the half-sites within the element. This ability to form oligomeric complexes was reflected in vivo as a greater ability to activate transcription from the rGH TRE compared with a DR4 element. Our results help explain the existence of natural TREs with multiple half-sites, and shed light on the importance of the different TR isoforms in gene specific regulation. The ability to form receptor multimers also provides an additional level of DNA recognition that may contribute to target gene discrimination by different members of the nuclear receptor family.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
All TR Isoforms Can Bind as Receptor Dimers to DNA Elements Containing Two Half-Sites
To verify our assays and to identify the positions of receptor monomers, homodimers, and heterodimers, we first performed a series of EMSAs on a panel of prototypic TREs containing one or two AGGTCA half-sites: a DR4, a divergent-repeat with a six-base spacer (DIV6), an inverted repeat with no spacer (INV0), and a DR4 element in which the second half-site has been replaced with 6 Ts (1-HS). Equal amounts of each TR isoform were used for each EMSA as assessed by SDS-PAGE and quantitative Sypro Ruby staining; all receptors were expressed as full-length proteins in a recombinant baculovirus/Sf9 system, and all were of avian origin except for TRß1, which was derived from a human cDNA clone. All EMSAs were performed in probe excess and all experiments were performed multiple times with comparable results; representative electrophoretograms are presented.

No protein/DNA complex was observed on any of the DNA elements when using non-recombinant baculovirus/Sf9 extracts (Fig. 2Go, A and B, lanes 1 and 2, and data not shown). TR{alpha}1 binds to single half-site DNA elements as a receptor monomer and to DR4 elements as a mix of receptor monomers and homodimers (37, 38, 39, 40); we assigned the receptor/DNA complexes displaying these properties in our own EMSA accordingly (Fig. 2AGo, lanes 3 and 4). Further paralleling prior studies (37, 38, 39, 40), the complexes we denote as TR{alpha} homodimers were destabilized by addition of T3 hormone, whereas the complexes denoted monomers were stable to T3 (Fig. 2AGo, compare lanes 3 and 4 with lanes 9 and 10). TRs heterodimerize with RXRs, and adding RXR{alpha} to our TR{alpha} preparation resulted in the appearance of the anticipated heterodimer complexes migrating at the appropriate mobilities (intermediate to that of the TR{alpha} homodimers and RXR{alpha} homodimers; Fig. 2AGo, lanes 4–6 and 10–12). Analogous results were seen with TRß0, with the exception that TRß isoforms do not detectably bind DNA as monomers; as a result, TRß0 formed primarily homodimers and (together with RXR{alpha}) heterodimers on the DR4 element (Fig. 2BGo). The composition of the TRß0 and RXR complexes were also confirmed by titration and antibody supershift experiments (see Figs. 3Go and 4CGo).



View larger version (63K):
[in this window]
[in a new window]
  		Fig. 2. All Four TR Isoforms Are Able to Bind as Receptor Dimers to Dual Half-Site TREs

The different TR isoforms were obtained from a recombinant baculovirus/Sf9 cell expression system and were tested for the ability to bind to radiolabeled DNA probes using an EMSA protocol. Phosphorimager scans of representative electrophoretograms are presented. A, TR{alpha}1 binding to a single half-site (1-HS) vs. a DR4 element was compared in the absence and presence of RXR{alpha}, and in the absence or presence of T3 hormone, as indicated above the panel. The positions of unbound DNA (free probe, fp), TR{alpha}1 monomer/DNA complexes (TR{alpha}1), TR{alpha}1 homodimer/DNA complexes [2(TR{alpha}1)], TR{alpha}1/RXR heterodimer/DNA complexes (TR{alpha}1/RXR), and RXR homodimer/DNA complexes [2(RXR{alpha})] are indicated on the left. Preparations isolated from nonrecombinant baculovirus/SF9 cells were used as negative controls (Non-Recomb). B, TRß0 binding to the 1-HS and DR4 DNA probes was analyzed by the methods in panel A. C, The ability of the different TR isoforms to bind to a series of two half-site DNA response elements was characterized. See Table 1Go for the nucleotide sequences of the probes. The position of monomeric, homodimeric, and heterodimeric protein/DNA complexes are indicated to the left of the panel for each receptor as in panel A. To conserve space, only the relevant portion of each phosphorimager scan, containing the receptor/DNA complexes, is shown.

 


View larger version (64K):
[in this window]
[in a new window]
  		Fig. 3. The TRß Isoforms, But Not TR{alpha}, Bind to the rGH TRE as Protein Trimers

The ability of different TR isoforms to bind to a DR4 element vs. an rGH response element was analyzed. A, DNA binding by TR{alpha}1 and TRß0 was compared. EMSAs were performed using 50 nM of 32P-labeled oligonucleotide probes representing a DR4 or rGH TRE, and identical ranges of TR{alpha}1 or TRß0 concentrations, as indicated above the panels and as described in Materials and Methods. A phosphorimager scan of a representative electrophoretogram is shown. The positions of receptor/DNA complexes representing TR{alpha}1 homodimers, TRß0 homodimers, and TRß0 homotrimers are indicated on the sides of the panel (see legend), as is the position of free DNA probe not bound to protein (fp). B, DNA binding by the three different TRß isoforms was compared. The EMSA was performed as in panel A except 100 nM of each oligonucleotide probe was used. See Table 1Go for the nucleotide sequences of the probes. A phosphorimager scan of a representative electrophoretogram is provided. For the TRß2 experiment in panel B, the gain in the phosphorimager was elevated 4-fold compared with TRß0 and TRß1 to permit the weak formation of trimeric complexes by this isoform to be detected. C, The ability of the TRß isoforms to bind to the DR4 element as dimers or to the consensus rGH element as trimers was compared as in panel B and quantified. Bars represent the mean of four experiments with standard errors. TRß0 binding to the DR4 element was defined as 100. D, TRß0 binding to a DR4, native rGH, consensus rGH, and mutated versions of the native rGH element was compared by EMSA with phosphorimager analysis and quantified. Reactions contained 100 nM of each oligonucleotide probe and a range of TRß0 concentrations. Only the highest two TRß0 concentrations were used for the A–C site mutants. The amount of receptor/DNA complexes migrating as a receptor homodimer (top panel), or as a receptor homotrimer (bottom panel) are presented; maximal TRß0 binding to the DR4 and con rGH elements was defined as 100. The average of four independent experiments is presented; error bars represent SE.

 


View larger version (58K):
[in this window]
[in a new window]
  		Fig. 4. TRß0 and TRß1, But Not TRß2, Can Form Heterotrimers with All Three RXR Isoforms on the Tripartite rGH TRE

The ability of the different TRß isoforms to heterodimerize with RXRs was examined. A, TRß0 and ß1 can form both heterodimers and heterotrimers on suitably reiterated response elements. EMSA reactions were performed as in Fig. 2Go, and contained equal amounts of each corresponding TRß isoform, a series of RXR concentrations (indicated as triangles above the panel), and 100 nM of the 32P-labeled oligonucleotide probe indicated, as described in Materials and Methods. The TR isoform used in each analysis is indicated to the right of each panel; lanes 18–20 contain the maximum amount of RXR used in the dilution series, but in the absence of TRs. B, Quantification of TRß0 binding to the DR4 (left panel) and consensus rGH (right panels) elements in the absence or presence of RXR{alpha}; the amount of probe migrating as a receptor homodimer or heterodimer is presented; bars represent the mean of four experiments with standard errors. C, The identities of the receptors in the dimeric and trimeric complexes were confirmed by use of antibody supershifts. EMSA antibody supershift experiments were performed as described in Materials and Methods. Antibodies to the different receptors were either omitted from (–) or included in the EMSA reactions (T = anti-TRß; X = anti-RXR).

 
We next extended these studies to the DIV6 and INV0 elements and to the TRß1 and TRß2 isoforms. Notably, all four TR isoforms formed receptor homodimers on the DR4 element, and less efficiently, on the DIV6 and INV0 elements (Fig. 2CGo, lanes 1–3); TR{alpha}1 was also able to form receptor monomers on many of these elements, whereas TRß0, 1, and 2 did not detectably bind any of these elements as receptor monomers (Fig. 2CGo and data not shown). As anticipated, the addition of RXR{alpha} to the binding assay resulted in the formation of TR/RXR heterodimer complexes that migrated at a slower electrophoretic mobility (Fig. 2CGo, lanes 4–6). All three TR isoforms formed heterodimers with RXR{alpha} efficiently on all three DNA elements tested (Fig. 2CGo); comparable results were obtained with RXRß and with RXR{gamma} (see Fig. 4Go). Our results indicate that all four TR isoforms can bind as receptor homo- and heterodimers to DNA elements composed of two consensus half-sites, whereas TR{alpha}1 can also bind to many elements as a receptor monomer. It should be also noted that for each receptor isoform, the spacing and orientation of the half-sites in the TRE had only minimal effects on the electrophoretic mobility of the resulting homo- or heterodimers.

TRß Isoforms, But Not TR{alpha}1, Bind as Receptor Trimers to the rGH Element
We next investigated the binding of TRs to response elements that contain three or more half-sites. The rGH TRE served as our prototype, and consists of two half-sites oriented as a DR4, followed by a third half-site in an INV1 topology [e.g. A(DR4)B(INV1)C; Fig. 1AGo] (25, 27). Despite the presence of three half-sites in this element, TR{alpha}1 has been reported to bind to the rGH element primarily as receptor monomers or dimers, with 3-fold occupancy observed only noncooperatively at extreme receptor excess and in the absence of RXR (25, 27). In agreement with these prior results, TR{alpha}1 bound to the rGH element principally as a dimeric protein/DNA complex displaying the same mobility as that observed on a DR4 element (Fig. 3AGo, compare lanes 1–5 with 6–10). TRß0, in contrast, bound to the rGH element as a protein/DNA complex with an electrophoretic mobility distinctly slower than that observed for TRß0 homodimers on a DR4, DIV6, or INV0 element (compare Fig. 3AGo, lanes 11–15 to lanes 16–20 and Fig. 2CGo). This novel TRß0 complex on the rGH element migrated at a position consistent with that of a receptor trimer (see Ferguson Analysis at the end of this section) and formed in a cooperative manner (Fig. 3AGo and quantified from repeated experiments in Fig. 3DGo).

The half-sites in the native rGH element differ from the idealized AGGTCA sequence (Fig. 1AGo). To determine whether these nonconsensus half-sites were required for trimer formation by TRß0, we tested a modified DNA element in which all three half-sites in the rGH element were converted to the consensus AGGTCA sequence (denoted as con rGH). TRß0 bound to this idealized element with a mobility again characteristic of a receptor trimer and with greater efficiency than to the native rGH element (Fig. 3BGo and quantified in Fig. 3DGo). Similarly, although TRß1 and TRß2 bound to the native rGH element very weakly and primarily as receptor homodimers, the ß1 and ß2 isoforms bound more strongly to the consensus rGH element, and as receptor trimers (Fig. 3BGo and, compared with homodimer formation on a DR4 in Fig. 3CGo). Our results suggest that the ability to form receptor trimers is shared by all three TRß isoforms tested, but that the TRß1 and ß2 isoforms require consensus half-site sequences to fully unmask this activity, whereas TRß0 is able to form receptor trimers on suitably reiterated elements bearing either consensus or nonconsensus half-sites.

To better define the molecular mass of the TRß0 complexes on the rGH elements, we applied a Ferguson analysis, which compares the relative electrophoretic mobilities of the relevant receptor/DNA complexes to a set of protein molecular weight markers in a series of native gels of increasing polyacrylamide concentration (41); this approach is more accurate than use of a single gel porosity and helps compensate for potential differences in receptor and DNA conformation. The apparent molecular weight of the TRß0 complex on either the natural or consensus rGH element was 1.51 (±0.03, n = 4) times the molecular weight of the homodimer TRß0 complex on the DR4 element (electrophoretograms not shown), fully consistent with our identification of the protein/DNA complex on the rGH element as a receptor homotrimer. To confirm the calibration of our Ferguson calculation, we artificially increased the size of TRß0 by fusion of GFP sequences; as expected, the GFP-TRß0 constructs formed complexes on the rGH element of slower mobility than did native TRß0, but which were again 1.52 times the calculated molecular weight of the GFP-TRß0 complexes on the DR4 element (data not shown).

TRß0 and ß1 Form Heterotrimers with RXR on DNA Elements Bearing Suitably Reiterated Half-Sites
RXRs are important dimer partners for TRs, and RXR/TR heterodimers assemble in preference to TR/TR homodimers on classical, bivalent TREs (42). Therefore, we examined how the different TRß isoforms bound to the rGH TRE in the presence of an RXR partner. Consistent with prior reports, all three TRß isoforms formed heterodimers with RXRs on a DR4 element in preference to TRß/TRß homodimers (Fig. 4AGo, compare lane 1 with lanes 2–4). Most of the RXR/TR heterodimers migrated at a slower mobility than did the corresponding TRß/TRß homodimers, although the position of the RXRß/TRß2 and RXR{gamma}/TRß2 heterodimers partially overlapped that of the corresponding TRß2/TRß2 homodimers (Fig. 4AGo, lanes 1–4). The RXR isoforms bound minimally to the DR4 element in the absence of a TR partner under these conditions (Fig. 4BGo, and data not shown).

In the absence of RXRs, TRß0 formed a homotrimer on the consensus rGH element (Fig. 4AGo, top panel, lane 5). Adding increasing amount of RXR{alpha} to TRß0 resulted in formation of a new complex that migrated more slowly than did either the TRß0 homotrimer on the rGH element or the TRß0/RXR heterodimer on the DR4, suggestive of the formation of a heterotrimer (Fig. 4AGo, top panel, lanes 6–9, and quantified in repeated experiments in Fig. 4BGo). This same phenomenon was observed using RXRß or RXR{gamma} (Fig. 4AGo, lanes 10–13 and 14–17) or when using the natural rGH element (data not shown). None of the RXR isoforms bound significantly to the rGH element in the absence of TRs (Fig. 4AGo, top panel, lanes 18–20). Of interest, TRß0 homo- or heterotrimers were the predominant species formed on the rGH element at all RXR concentrations tested, and little or no RXR/TRß0 heterodimer formation was observed (i.e. Figure 4AGo, top panel, compare lane 2 with lanes 5–9).

To verify the composition of these different complexes, we performed EMSA antibody supershift experiments. Antibodies to TRß supershifted the TRß0 complexes on all three TREs tested (DR4, rGH, and consensus rGH) in both the presence and absence of RXR{alpha} (Fig. 4CGo, lanes 2, 5, 8, 11, 14, and 17). Antibodies to RXRs, as expected, were only able to supershift complexes to which RXR had been added; the anti-RXR antibody shifted both the heterodimers formed on the DR4 element, and the presumptive heterotrimers that formed on the rGH elements in the combined presence of TRß0 and RXR{alpha} (Fig. 4CGo, compare lanes 12 with 10 and 18 with 16). We conclude that the latter complexes contain both RXR{alpha} and TRß0, and based on their relative mobility, are indeed receptor heterotrimers. The EMSA technique did not permit us to determine the relative stoichiometry of RXR to TRß0 in these complexes: formation of either a 2(TRß0):1(RXR), or a 1(TRß0):2(RXR) complex would be consistent with our results.

The TRß1 isoform also formed a heterotrimer with all three RXR isoforms, but less efficiently than did TRß0 (Fig. 4AGo, middle panel, compare lanes 5, 6–9, 10–13, and 14–17). Unlike TRß0, however, substantial formation of a RXR/TRß1 heterodimer was observed at higher RXRß or RXR{gamma} concentrations (Fig. 4AGo, middle panel, lanes 10 and 14), suggesting that heterodimers, not heterotrimers, are the most stable form of TRß1/RXRß and TRß1/RXR{gamma} complex on the consensus rGH element. TRß2 failed to form detectable heterotrimers with any of the RXR isoforms, and instead formed TRß2/RXR heterodimers on both the rGH and the DR4 elements (Fig. 3AGo, bottom panel). We conclude that although all three TRß isoforms can form receptor homotrimers on appropriately reiterated DNA elements, TRß0 forms these homotrimers with the highest efficiency, particularly on the natural rGH element. TRß0 is also the most efficient isoform of TR at forming heterotrimers with RXRs, whereas TRß1 forms heterotrimers with less efficiency, and TRß2 does not exhibit such ability.

The Ability of TRß0 to Bind DNA as a Receptor Trimer Is Determined by Both the Spacing and the Orientation of the Half-Sites in the Element
Mutagenesis of any one of the three half-sites in the native or consensus rGH disrupted homotrimer binding by TRß0 and by TRß1, confirming that all three half-sites participate in receptor recruitment on the rGH element (Fig. 5AGo, quantified in Fig. 3CGo, and data not shown). TRß0 and TRß1 were able to bind weakly as receptor dimers to the nonconsensus DR4 remaining in the native rGH element when the C half-site was disrupted (Figs. 3CGo and 5AGo). The significantly stronger binding of TRß0 and ß1 trimers when the C half-site is present (i.e. in the native rGH), compared with when the C site is mutated, is indicative of a significant gain in DNA binding stability resulting from receptor trimerization.



View larger version (55K):
[in this window]
[in a new window]
  		Fig. 5. Trimeric TRß Binding Requires the Presence of Three Half-Sites Displayed with Appropriate Spacings and Orientations

The ability of TRß0 and TRß1 to bind to altered versions of the rGH element was tested. A, Disruption of any one half-site in the rGH element prevents trimer binding by TRß0 or ß1. Each reaction contained 50 nM of 32P-labeled oligonucleotide probe and equal mass amounts of TR{alpha}1 and TRß0. The identity of the DNA response element used in each experiment is depicted above each lane; mut A, mut B, or mut C denote rGH elements with the indicated half-site mutated as to be dysfunctional (see Fig. 1AGo and Table 1Go). B, The ability of TR{alpha}1 and ß0 to bind to a series of different artificial response elements was tested; the effect of altering the orientation of the B and C half-sites was examined. EMSAs were performed as in panel A except the oligonucleotide probes were used at 100 nM. See Table 1Go for the nucleotide sequences of the probes. C and D, The effect of different half-site spacings on trimer formation by TRß0 was examined. EMSAs were performed as in panel B. The receptors used in each analysis are indicated to the left of the panel. E, The ability of TRß0 and TR{alpha}1 to bind to a selection of natural tripartite DNA elements was examined by EMSA as in panel A. See Table 1Go for the sequences of the probes.

 
TR dimers can form on response elements composed of pairs of half-sites displayed in a variety of orientations and spacings. We therefore examined the effect of changing the half-site orientation and spacing within the consensus rGH element on the ability of TRß0 to bind as a receptor trimer. Reorienting the C half-site into a DR1 configuration (DR4/DR1, Table 1Go) or a DR4 configuration (DR4/DR4, Table 1Go) abolished trimer binding by TRß0 and resulted in the exclusive formation of receptor homodimers (Fig. 5BGo, lanes 2 and 3). Inversion of the A half-site to create a DIV6/INV1 element (Table 1AGo) significantly destabilized, but did not fully abrogate, trimer binding by TRß0, resulting in formation of both dimeric and trimeric protein/DNA complexes (Fig. 5BGo, lane 4). Combining the DIV6 orientation of the A and B sites with a DR4 orientation for the C site (DIV6/DR4, Table 1AGo) abolished trimer binding by TRß0 and resulted in the exclusive formation of homodimers (Fig. 5BGo, lane 5). Notably the TR{alpha}1 isoform bound to all of these modified response elements as a receptor dimer or monomer, with no evidence of TR{alpha}1 trimer formation on any element tested (Fig. 5BGo, lower panel, lanes 1–5).


View this table:
[in this window]
[in a new window]
  		Table 1. Oligonucleotide Probes Used in these Experiments

 
We next altered the spacing between the B and C half-sites of the consensus rGH element while preserving the INV topology. Changing the spacing between the B and C half sites from the native 1 nucleotide to 0, 3, 5, 7, or 9 bases destabilized or eliminated homotrimer formation although permitting weak homodimer binding (Fig. 5CGo, top panel). Heterotrimer binding was slightly more permissive, with an inverted repeat spacing of nine nucleotides resulting in the formation of detectable, if modest, levels of TRß0/RXR{alpha} heterotrimers (Fig. 5CGo, bottom panel); spacings of 0, 3, 5, and 7 nucleotides resulted primarily in the formation of heterodimers (Fig. 5CGo).

TR dimers are able to bind to AGGTCA half-sites in DR3 and DR5 orientations, although less efficiently than to the prototypic DR4 spacing (20, 21, 37). Therefore, we tested whether changing the spacing between the A and B half-sites in the rGH TRE to a DR3 or DR5 permitted TRß0 to bind as a receptor trimer. A DR3 spacing between the A and B sites retained detectable homotrimeric and heterotrimeric binding, with the latter relatively robust (Fig. 5DGo). The DR5 spacing, however, abrogated virtually all homotrimer and heterotrimer formation, although permitting some residual formation of heterodimers (Fig. 5DGo). We conclude that the DR4/INV1 orientation and spacing of half-sites in the native rGH element is optimal for recruitment of TRß0 trimers; nonetheless, a limited series of additional spacings and orientations is capable of recruiting TRß0 as a homotrimer or as a heterotrimer with RXRs.

Other natural TREs have been identified that contain three half-sites, but displayed in different topologies from the DR4/INV1 in the rGH TRE (see Table 1Go). These include response elements mapped within the rat myosin heavy chain gene, the mouse myoD gene, and the human skeletal {alpha}-actin gene (23, 24). We tested these additional natural TREs for their ability to bind TR{alpha}1 and TRß0. TR{alpha}1 recognized all elements as a receptor dimer, or in the case of the MHC element, as a receptor monomer; no higher order complexes were detected (Fig. 5EGo, lower panel). TRß0 also bound to the MHC, myoD and skeletal {alpha}-actin elements primarily as a receptor dimer, although some evidence of a higher order species could be found for the myoD element (Fig. 5BGo, upper panel, lanes 3–5); these results are consistent with the preference of TRß0 trimers for a DR4/INV1 topology as determined in our mutagenesis studies.

TRß Trimers Can Recruit Both Corepressors and Coactivators
Once bound to a response element, nuclear receptors modulate target gene transcription by recruiting auxiliary proteins, denoted corepressors and coactivators (1, 2, 3). TR homodimers bind to both SMRT (silencing mediator of retinoic acid and thyroid hormone action) and nuclear receptor corepressor (N-CoR) corepressors in the absence of thyroid hormone, but they release these auxiliary factors and bind to coactivators, such as ACTR (activator of thyroid and retinoic acid receptors), in the presence of hormone. We employed an EMSA supershift protocol to determine whether TRß trimers also are able to recruit corepressors and coactivators. As anticipated, TRß0 homodimers, assembled on a DR4, interacted strongly with SMRT and with N-CoR in the absence of hormone; these interactions were observed as a further reduction in electrophoretic mobility of the receptor/DNA complex when the corepressor was included in the binding reaction (Fig. 6Go, upper panel, compare lane 1 with lanes 3 and 5). The same phenomenon was observed for TRß0 homotrimers assembled on the consensus rGH element (Fig. 6Go, upper panel, compare lane 9 with lanes 11 and 13). Addition of T3 resulted in the release of corepressor from both the dimeric and trimeric receptor complexes (Fig. 6Go, upper panel, lanes 4, 6, 12, and 14). Heterodimers of TR and RXR interact only weakly with SMRT and N-CoR, and the same proved true of the TR/RXR heterotrimers that assembled on the rGH element (Fig. 6Go, compare lanes 1, 3, and 5 in top and bottom panels); note in each case the preferred supershift of the residual TRß0 homomeric complexes, and the more minimal corepressor supershift for the corresponding TRß0/RXR heteromeric complexes.



View larger version (82K):
[in this window]
[in a new window]
  		Fig. 6. Trimeric TRß0 Complexes Assembled on the rGH Element Can Interact with Corepressors and Coactivators

The ability of TRß0 dimeric and trimeric complexes (formed on the DR4 or consensus rGH element, respectively) to interact with corepressors and coactivators was analyzed. GST-fusion proteins containing the receptor interaction domains of SMRT, of N-CoR, or of ACTR were added to the EMSA binding reactions in the presence or absence of 1 µM T3 as indicated above the panels. Interaction of the corepressor or coactivator with the receptor/DNA complex leads to formation of a complex with slower mobility than that of receptor/DNA alone (supershift). All lanes contained 100 nM 32P-labeled oligonucleotide probe and either TRß0 (top panel) or TRß0/RXR{alpha} (bottom panel); receptor concentrations were identical in all lanes of each panel.

 
The p160 family of coactivators, such as ACTR, operates reciprocally to corepressors, interacting with TRs in the presence, but not in the absence, of T3 hormone. Consistent with these previous findings, both TRß0 homodimers and TRß0/RXR heterodimers interacted with ACTR when bound to a DR4 element, and this interaction required T3 hormone (Fig. 6Go, compare lanes 1, 7, and 8). Notably, TRß0 homotrimers and TRß0/RXR heterotrimers also interacted with ACTR when bound to the consensus rGH element, and this interaction was again T3-dependent (Fig. 6Go, lanes 15 and 16). The presence of two supershifted complexes in some contexts (e.g. Fig. 6Go, top panel, lane 8) but not others (e.g. lane 16) may reflect changes in the stoichiometry of coactivator recruitment, but the precise significance of this observation remains unclear. We conclude that TRß0 trimers, in common with TRß0 dimers, release from SMRT corepressor and recruit p160 coactivators in response to T3 hormone.

The Ability of TRß Isoforms to Utilize the rGH Element Correlates with the Ability to Form Trimers
To determine whether the ability of TRß0 to form receptor trimers in vitro results in differences in transcriptional activation in vivo, we created luciferase reporter vectors containing single copies of either the consensus DR4 or the rGH TRE. We then introduced these reporter constructs by transient transfection into CV-1 cells together with an expression vector for either TRß0 or TRß2. In cells, where RXRs are present at high levels, TRß0 would be expected to form heterotrimeric complexes on the rGH TRE, whereas TRß2 is likely to form only heterodimers. A constitutive lacZ expression vector was used in all experiments as an internal control for transfection efficiency. We first compared the ability of the two different TR isoforms to activate transcription at a saturating hormone concentration, over a range of receptor vector inputs (Fig. 7AGo). Both TRß0 and TRß2 activated the consensus DR4 equally well, and Mann-Whitney analysis showed no significant difference between TRß0 and TRß2 on the DR4 element at every plasmid concentration, when four separate transfections were analyzed (95% confidence interval). In contrast, on the wt rGH element, TRß0 activated transcription significantly better than TRß2 at every plasmid input (Mann-Whitney, 95% CI, P < 0.03 for each concentration, n = 4). Next, we compared transcriptional activation with different versions of the rGH element, using a single concentration of the TR expression vector. As in panel A, TRß0 (black bars) and TRß2 (open bars) activated the DR4 reporter to a comparable extent in response to T3 (P = 0.89, n =4), whereas TRß0 activated the native rGH element reporter significantly better than did TRß2 (P = 0.029, n =4) (Fig. 7BGo). The A–C site mutations in the rGH element each severely inhibited reporter activation by either TRß0 or TRß2, as expected given the poor ability of these elements to bind either TR isoform in vitro (Fig. 7BGo). Interestingly, the consensus rGH element was efficiently activated by both TRß0 and by TRß2, likely a consequence of the ability of TRß2 to efficiently form heterodimers on the consensus DR4 in this element. We also transfected COS1 cells (a derivative of CV-1 cells) with the same expression vectors and isolated nuclear extracts to test in an EMSA reaction. Both the TRß0 transfected and TRß2 transfected COS nuclear extracts were able to form heterodimers on the DR4 oligonucleotide in the presence of RXR, whereas only TRß0 transfected COS nuclear extracts, but not those transfected by TRß2, also formed heterotrimers on the consensus rGH TRE (data not shown).



View larger version (24K):
[in this window]
[in a new window]
  		Fig. 7. TRß0 Activates Transcription from an rGH Element More Strongly than Does TRß2, Whereas Both Receptors Activate From a DR4 Element Equally

The ability of TRß0 and of TRß2 to activate transcription through DR4 or rGH response elements was tested by a cell transfection protocol, as described in Materials and Methods. A, The ability of TRß0 and of ß2 to activate transcription was tested over a range of receptor concentrations. CV-1 cells were transfected with a luciferase reporter gene bearing either a DR4 or native rGH TRE element, and a range of concentrations of pSG5 vectors expressing either TRß0 or TRß2, as indicated. A pSG5 empty construct was added to maintain total DNA concentration constant. All transfections were analyzed in the presence of 100 nM T3, and a ß-galactosidase expression vector was included as an internal control for transcription efficiency. Fold relative activation was calculated as the ratio of reporter gene expression in the presence vs. the absence of a TR allele in the pSG5 expression vector; the mean fold relative luciferase values of four independent experiments are shown. Error bars represent the standard error. The results for TRß0 (square symbols) and TRß2 (triangles), either minus (open symbols) or plus (solid symbols) 100 nM T3 are shown. B, Transient transfections were performed as in panel A, except 5 ng of the pSG5-TR vector was tested with the different TRE luciferase reporters indicated. The A, B, and C site mutations were in the context of the native rGH element. The results for TRß0, plus T3 (black bars) and TRß2 plus T3 (open bars) are shown. Results represent the mean of four independent experiments; error bars are the standard error. C, The same experiments as in panel B are shown, but in the absence of hormone.

 
In the absence of hormone, TRß0 represses DR4 reporter gene expression (Ref. 43 and Fig. 7CGo). Notably, TRß0 also displayed an enhanced ability to repress on the native rGH element relative to the DR4 element (Fig. 7CGo, closed bars; P = 0.029, n = 4), and still stronger repression was seen on the consensus rGH element (Fig. 7CGo, closed bars; P = 0.029, n = 4). Mutations in any of the three rGH half-sites that abolished TRß0 trimer binding also abolished reporter gene repression (Fig. 7CGo). TRß2 does not repress on a DR4 (Ref. 43 and Fig. 7CGo), and did not repress any of the rGH-derived reporter constructs tested here (Fig. 7CGo, open bars).

The Ability of TRß0 to Form Receptor Oligomers Extends to High Order Complexes Greater than Trimer in Size
Although many studies have focused on the tripartite nature of the rGH element, the promoter region of the native rGH gene actually contains two additional half-sites, denoted D and E (29). These additional half-sites are oriented as a DR5 just downstream of the ABC half-sites (Fig. 6AGo and Ref. 29), raising the possibility that all five half-sites together might be able to recruit a pentameric form of TRß0. Indeed, TRß0 alone, or TRß0 together with RXRß, was able to bind to an oligonucleotide probe containing all five half-sites (an ABCDE element) as a receptor complex displaying a slower mobility than that of the previously characterized receptor trimer on the ABC element (Fig. 8Go). Although this multimeric receptor complex was most evident at high receptor concentrations, it formed under conditions of probe excess and in preference to trimeric or dimeric complexes (Fig. 8Go and data not shown). We suggest that these TRß0 complexes are pentamers; consistent with this hypothesis, an oligonucleotide containing four half-sites (oriented as two DR4 repeats) produced a TRß0 complex intermediate in electrophoretic mobility to the tripartite rGH (ABC) and pentapartite rGH(ABCDE) TREs (data not shown). These results indicate that TRß has the capability to form dimers, trimers, tetramers, and pentamers, depending on the number and the topology of the half-sites within the DNA binding site.



View larger version (46K):
[in this window]
[in a new window]
  		Fig. 8. TRß0 Can Bind an Expanded Segment of the rGH Promoter as a Presumptive Receptor Pentamer

A, The nucleotide sequence and arrangement of possible half-sites in the regions encompassing and flanking the core rGH element is presented (29 ). Each presumptive half-site is underlined. Half-sites A–C comprise the trimeric rGH TRE previously noted in Fig. 1Go. B, TRß0 can form both homo and heteropentameric receptor complexes on the extended rGH promoter element. For each binding reaction, TRß0, or a mix of TRß0 and RXRß, was incubated with 100 nM radiolabeled oligonucleotide probe (representing a DR4, rGH, or rGH 5-half-site TRE) and the resulting complexes were resolved by EMSA as in Fig. 2Go and Materials and Methods. A phosphorimager scan of a representative electrophoretogram is presented.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
MATERIALS AND METHODS
 REFERENCES
 
The TRß Isoforms Can Bind as Trimers to the Tripartite rGH TRE
Analysis of artificial DNA sequences has permitted elucidation of a prototypic TRE comprised of two AGGTCA half-sites arrayed as a DR4 (17, 18). This two half-site element efficiently binds both TR homodimers and TR/RXR heterodimers; the former have been implicated in transcriptional repression and the latter in transcriptional activation (42, 44). Examination of the native T3-regulated genes present in vertebrate genomes, however, reveals an apparent paradox: many native response elements contain three or more half-sites in close proximity to one another. The tripartite TRE found in the rGH promoter has served as a prototype for these reiterated response elements, and disruption of any one of the three half-sites in the rGH element significantly decreases T3-driven reporter gene activation in vivo (25). Notably, many of these prior studies on the rGH TRE employed the TR{alpha}1 and RXR{alpha} isoforms. All three half-sites in the rGH element can be occupied by TR{alpha}1 at very high receptor concentrations; however, these weak trimeric complexes assemble noncooperatively and are readily disrupted by addition of RXRs, resulting in a preferential formation of TR{alpha}1/RXR heterodimers. Therefore, despite its tripartite structure, it has been assumed that the rGH element operates by recruiting receptor dimers.

Prior studies have shown that RXRß, but not RXR{alpha} or {gamma}, forms receptor trimers and tetramers on RXR response elements possessing appropriately reiterated half-sites (35, 36). Therefore, we investigated if TRs might display similar, if unrecognized, isoform-specific oligomerization properties on the tripartite rGH TRE. We report here that the TRß isoforms can assemble as trimeric receptor complexes on the rGH element with high cooperativity, whereas the TR{alpha}1 isoform does not. All three ß isoforms (TRß0, TRß1, and TRß2) tested were able to form homotrimeric complexes, if with distinct efficiencies, whereas only TRß0 and TRß1 were also able to form heterotrimeric complexes with RXRs. TRß0/RXR{alpha}, TRß0/RXRß, TRß0/RXR{gamma}, and TRß1/RXR{alpha} heterotrimers assemble on the rGH element in clear preference to the corresponding heterodimers; in contrast, TRß1/RXRß and TRß1/RXR{gamma} heterotrimers can be detected on the rGH element, but are converted into heterodimers in the presence of high levels of the RXR partner.

Our identification of these TRß complexes on the rGH element as authentic trimers was confirmed by a number of criteria. Our EMSAs were calibrated by using response elements containing different numbers of half-sites and by employing gel matrices of differing porosity; quantification by Ferguson analysis verified that the TRß complexes on the rGH element matched closely the predicted mobility of a receptor trimer relative to a dimer. All the DNA probes were of identical overall length and contained identical flanking sequences outside of the response elements themselves. Disrupting any one of the three half-sites in the rGH element disrupted trimer assembly; conversely converting all three rGH half-sites into AGGTCA consensus sequences further stabilized trimer formation by the TRß isoforms. Notably, the same TRß preparations formed only conventional receptor dimer complexes when incubated with DNAs composed of two half-sites, and changes in the spacing or orientations of these two half-site elements had little or no effect on the mobility of these dimeric complexes. No orientation or spacing of two half-sites produced a TRß complex that mimicked the mobility characteristic of the trimeric complex on the three half-site rGH elements. None of the response elements tested in our studies allowed trimer formation by the TR{alpha}1 isoform. Antibody supershifts confirmed the compositions of the various receptor complexes analyzed here. Taken together, our results strongly indicate that the TRß complexes on the rGH element are true receptor trimers, and not receptor dimers migrating at an aberrant mobility due to an unusual protein or DNA conformation.

TRß0 Trimers Assemble Preferentially on Response Elements Bearing a DR4/INV1 Topology
We examined the features of the rGH element necessary for trimeric binding by the TRß isoforms. Notably, trimer formation is not dependent on the nonconsensus character of the native rGH half-sites, and substituting these nonconsensus half-sites with AGGTCA sequences enhances, rather than diminishes, trimer formation by TRß. In contrast, the DR4 orientation of half-sites A and B and the INV1 orientation of half-sites B and C do play a pivotal role in receptor trimerization; most other spacings and orientations tested here are either less permissive or completely dysfunctional for TRß trimer formation.

What might the trimeric TRß complex look like once assembled on the rGH element? The topology of a DR4 element necessitates a head-to-tail arrangement of the DNA binding domains of any bound receptor dimer (14). This head-to-tail arrangement is stabilized through a protein-protein dimerization surface present within the DNA binding domain of the receptor (14). The addition of a third half-site, inverted to the second half-site, as in the rGH TRE, would place the DNA binding domain of the third TRß in a tail-to-tail conformation with that of the TRß on the second half-site (i.e. the same orientation as on the dimeric INV0 element). What stabilizes the tail-to-tail binding of this additional receptor molecule? Most nuclear receptors possess a second, strong dimerization domain in their hormone binding domains, and it is thought that a "swivel" in the receptor permits a multiplicity of orientations and spacings of half-sites to be recognized by the receptor DNA binding domain while preserving the dimerization interaction mediated by the receptor hormone binding domain (15, 45). We propose that receptor trimers are stabilized by an extension of this basic model through which a mix of protein-protein interaction domains within both the DNA binding and hormone binding domains act together to form a multimeric receptor complex. Of course, there may also be additional interaction surfaces on the TRßs that are exploited specifically for trimer formation.

The Capacity to Form Receptor Oligomers Depends on the Nature of the TR N-Terminal Domain
The TRß isoforms, but not TR{alpha}1, are able to form receptor trimers. Notably, among the ß isoforms, TRß0 exhibited the greatest ability to form homotrimers and the broadest ability to form heterotrimers with different RXRs. Although truncated at its N terminus, TRß0 is otherwise identical in sequence to the TRß2 isoform, which formed trimers relatively poorly. Presumably the extended N terminus in TRß2 interferes with trimer formation, rather than TRß0 possessing a trimerization domain absent in TRß2. A similar phenomenon has been observed among the different isoforms of RXR: RXRß can form receptor trimers and tetramers, whereas RXR{alpha} and {gamma} cannot, and these properties correlate with the highly truncated N terminus of RXRß (35, 36). We propose that the receptor N-terminal receptor domain can restrict DNA binding to dimeric complexes, and that the loss of these sequences in specific isoforms can permit an oligomeric mode of DNA recognition.

The TRß0 isoform is expressed in avian, amphibian, and reptilian lineages, whereas mammals encode a related TRß1 isoform possessing a 32-amino acid N-terminal sequence found neither in TRß0 nor TRß2. In our assays, mammalian TRß1 displays a somewhat more limited ability to form trimers than does avian TRß0. Although muted relative to TRß0, the trimerization capabilities of TRß1 may nonetheless be sufficient to function in mammalian cells on tripartite TREs such as the rGH element. Alternatively, two novel isoforms of TRß, denoted TRß3 and TRß{Delta}3, have been described recently in rat (46); these isoforms possess truncated N-terminal domains and conceivably may possess stronger trimerization properties than TRß1.

TRß0 Can Form Receptor Complexes Greater than Trimer in Size on Elements Possessing Suitably Reiterated Half-Sites
RXRß can form trimeric, tetrameric, and pentameric complexes depending upon the multiplicity of the half-sites in the response element (35, 36). We report here that TRß0 can similarly form complexes greater than trimer in size. Although the rGH TRE is usually considered to be composed of three half-sites, a broader examination of the rGH promoter demonstrates that two additional half-sites flank this rGH core (29). When the pentameric rGH sequence was included in our analysis, it was able to recruit what appears to be a pentameric complex of TRß0; these pentamers formed cooperatively, under conditions of probe excess, and both TRß0 homopentamers and TRß0/RXR{alpha} heteropentamers were observed. TRß0 therefore appears capable of the cooperative formation of relatively large receptor complexes on suitably reiterated response elements. Given that the TRß0 isoform is expressed in amphibia, it is interesting that a natural TRE involved in T3-driven metamorphosis in Xenopus, in the TH/bZIP promoter (47), contains four half-sites and recruits a TRß0 tetrameric complex in our EMSA experiments (our unpublished results).

The Ability to Bind the rGH Element as a Heterotrimer in Vitro Correlates with Enhanced Transcriptional Regulation by TRß0 in Transfected Cells
Both the TRß0 and TRß2 isoforms activate reporter gene expression equally well from a DR4 response element in transfected cells. On the native rGH element, however, TRß0 activates gene expression up to 3-fold more strongly than on the DR4, whereas TRß2 does not. This suggests that the ability of the TRß0 isoform to bind as a trimer to the reiterated rGH element in vitro is paralleled by a preferential ability to utilize this element to activate transcription in vivo. This is likely to be due, in part, to the greater ability of TRß0 trimers to bind to the rGH element compared with TRß2 dimers. It is also possible that, once bound to a target gene, trimeric TR complexes might recruit coactivators more efficiently than do dimeric complexes, resulting in increases in transcriptional efficiency as well as increases in promoter occupancy. Although the ACTR coactivator interacted with TR dimers and trimers with approximately equal efficiencies in vitro, circumstances may differ in vivo and/or for other coactivators not tested here. Reciprocally, it is possible that the repression properties of TR dimers and trimers may differ, despite our preliminary observations that both homodimeric and homotrimeric TRß0 complexes bind SMRT and N-CoR to comparable extents in vitro.

A series of recent studies have employed microarray analysis and genetic knockout mice to study isoform-specific gene regulation by TRs (4, 5, 6, 9, 10, 12, 48, 49, 50, 51, 52, 53). These studies demonstrate that different isoforms of the TR play distinct roles in different tissues at different times in development. When characterizing the target genes through which these isoform-specific receptor actions are manifested, it will be important to consider not just the classical dimeric TRE elements but also to consider the contribution that multimeric binding by the TRß isoforms may also play a role in the finely tuned regulation of gene expression.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
REFERENCES
 
Cell Culture and Transfections
CV-1 cells were grown in DMEM (Invitrogen Life Technologies, Carlsbad CA) supplemented with 10% fetal bovine serum (Hyclone, Logan, UT) at 37 C and buffered with a bicarbonate/5% CO2 system. Transfections were performed in the same medium in 24-well cell culture plates using Effectene (QIAGEN, Valencia, CA) per the manufacturer’s instructions. Each transfection employed approximately 3 x 104 cells, 100 ng of luciferase reporter plasmid carrying a single copy of the specified TRE, a pSG5 or pSG5-TR expression plasmid, 40 ng of pCH110 expressing ß-galactosidase as an internal transfection control, plus sufficient pUC18 to bring the total DNA to 250 ng. All pSG5 expression constructs were designed to produce native, full-length proteins, either avian (Gallus) TR{alpha}1, Gallus TRß0, Gallus TRß2, or human TRß1. Twenty-four hours post transfection, the medium was aspirated and replaced with fresh medium containing either T3 thyronine (Sigma, St. Louis, MO) or a corresponding volume of ethanol carrier. After a further 24 h, the cells were harvested and lysed. Luciferase activity was measured using the Luciferase Assay System (Promega, Madison, WI) and a Turner Design 20/20 luminometer; ß-galactosidase activity was measured using a CPRG (chlorophenol-red-ß-D-galactopyranoside monosodium salt) substrate (Roche, Indianapolis, IN) assay and a Molecular Devices SpectraMax 250 microplate reader.

EMSAs
Sf9 cells were infected with recombinant baculoviruses designed to express the relevant TR isoforms as native, full-length proteins: avian (Gallus) TR{alpha}1, Gallus TRß0, Gallus TRß2, or human TRß1; nuclear extracts were prepared as previously described (54). TR preparations were resolved by SDS-PAGE and were quantified by SYPRO Ruby staining (Bio-Rad, Hercules, CA) following the manufacturer’s protocol. Oligonucleotide probes representing the various TREs were prepared commercially (MWG Biotech, High Point, NC) and were radiolabeled by Klenow polymerase fill-in using 32P-{alpha}-deoxy-GTP (3000 Ci/mmol) and the remaining three unlabeled deoxynucleotide triphosphates. For EMSA, the TR of interest was incubated together with the radiolabeled probe in binding buffer [7.5 mM Tris-HCl (pH 7.5), 150 mM KCl, 2.3% glycerol, 10 mg/ml BSA, 1.5 mM MgCl2, 2 µg poly(deoxyinosine-deoxycytosine) and 2 mM dithiothreitol] for 15 min at room temperature. The reaction products were then separated by electrophoresis in a 5% polyacrylamide gel using a 0.5x TBE buffer system (45 mM Tris-borate, 1 mM EDTA) at 200 V for 75 min. Free and bound probe were visualized and quantified by Storm phosphorimager analysis (Amersham Biosciences, Piscataway, NJ). For antibody supershift experiments 2 µl of antibody was added to each binding reaction. Anti-TRß monoclonal antibody MA1–215 (Affinity Bioreagents, Inc., Golden, CO) and anti-RXR rabbit polyclonal 4RX-1D12.1.3 (provided by P. Chambon, Institut de Genetique et de Moleculaire et Cellulaire, Strasbourg, France) were used at a 1/4 dilution.

Ferguson Analysis
The molecular weights of the various TR-TRE complexes were determined by a Ferguson analysis as previously described for RXRs (41, 55). Briefly, protein/DNA complexes were resolved by EMSA using 4, 5, 6, 7, and 8% polyacrylamide gels in 0.5x TBE buffer, employing as molecular mass standards soybean trypsin inhibitor (21.5 kDa), ovalbumin (45 kDa), BSA (monomer, 67 kDa, and dimer, 134 kDa), ferritin (440 kDa) and thyroglobulin (669 kDa). Protein standards were visualized with Coomassie Blue R-250. After Rf values for each standard on each percentage acrylamide gel were calculated, primary plots were made of 100 x log(Rf x 100) vs. the acrylamide concentration and the slopes of the resultant lines were calculated using the least-squares analysis function of Sigma-Plot. A secondary plot was then made of the negative slope vs. the molecular mass (in kilodaltons) of the standards, and the equation for this line was calculated with Sigma-Plot. Finally the Rf values of the EMSA complexes were calculated, primary plots were made as for the standards and the resultant slopes were inserted into the linear equation from the secondary plots of the standards to solve for the molecular weights of the EMSA complexes.


   ACKNOWLEDGMENTS
 
The authors thank Liming Liu for excellent technical assistance. We also thank Pierre Chambon for the generous gift of the anti-RXR antibody 4RX-1D12.1.3.


   FOOTNOTES
 
This work was supported by Public Health Service/National Institutes of Health Award R37-53394.

First Published Online September 30, 2004

Abbreviations: ACTR, Activator of thyroid and retinoic acid receptors; DR, direct repeat; ER, everted repeat; INV, inverted repeat; N-CoR, nuclear receptor corepressor; rGH, rat GH; RAR, retinoic acid receptor; RXR, retinoid X receptor; SMRT, silencing mediator of retinoic acid and thyroid hormone action; TR, thyroid hormone receptor; TRE, thyroid hormone response element.

Received for publication July 22, 2003. Accepted for publication September 20, 2004.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 

  1. Aranda A, Pascual A 2001 Nuclear hormone receptors and gene expression. Physiol Rev 81:1269–1304[Abstract/Free Full Text]
  2. Yen PM 2001 Physiological and molecular basis of thyroid hormone action. Physiol Rev 81:1097–1142[Abstract/Free Full Text]
  3. Zhang J, Lazar MA 2000 The mechanism of action of thyroid hormones. Annu Rev Physiol 62:439–466[CrossRef][Medline]
  4. Gullberg H, Rudling M, Forrest D, Angelin B, Vennstrom B 2000 Thyroid hormone receptor ß-deficient mice show complete loss of the normal cholesterol 7{alpha}-hydroxylase (CYP7A) response to thyroid hormone but display enhanced resistance to dietary cholesterol. Mol Endocrinol 14:1739–1749[Abstract/Free Full Text]
  5. Gullberg H, Rudling M, Salto C, Forrest D, Angelin B, Vennstrom B 2002 Requirement for thyroid hormone receptor ß in T3 regulation of cholesterol metabolism in mice. Mol Endocrinol 16:1767–1777[Abstract/Free Full Text]
  6. Mansen A, Yu F, Forrest D, Larsson L, Vennstrom B 2001 TRs have common and isoform-specific functions in regulation of the cardiac myosin heavy chain genes. Mol Endocrinol 15:2106–2114[Abstract/Free Full Text]
  7. Gloss B, Trost S, Bluhm W, Swanson E, Clark R, Winkfein R, Janzen K, Giles W, Chassande O, Samarut J, Dillmann W 2001 Cardiac ion channel expression and contractile function in mice with deletion of thyroid hormone receptor {alpha} or ß. Endocrinology 142:544–550[Abstract/Free Full Text]
  8. Gothe S, Wang Z, Ng L, Kindblom JM, Barros AC, Ohlsson C, Vennstrom B, Forrest D 1999 Mice devoid of all known thyroid hormone receptors are viable but exhibit disorders of the pituitary-thyroid axis, growth, and bone maturation. Genes Dev 13:1329–1341[Abstract/Free Full Text]
  9. Ng L, Hurley JB, Dierks B, Srinivas M, Salto C, Vennstrom B, Reh TA, Forrest D 2001 A thyroid hormone receptor that is required for the development of green cone photoreceptors. Nat Genet 27:94–98[Medline]
  10. Wikstrom L, Johansson C, Salto C, Barlow C, Campos Barros A, Baas F, Forrest D, Thoren P, Vennstrom B 1998 Abnormal heart rate and body temperature in mice lacking thyroid hormone receptor {alpha} 1. EMBO J 17:455–461[CrossRef][Medline]
  11. Forrest D, Erway LC, Ng L, Altschuler R, Curran T 1996 Thyroid hormone receptor ß is essential for development of auditory function. Nat Genet 13:354–357[CrossRef][Medline]
  12. Gauthier K, Chassande O, Plateroti M, Roux JP, Legrand C, Pain B, Rousset B, Weiss R, Trouillas J, Samarut J 1999 Different functions for the thyroid hormone receptors TR{alpha} and TRß in the control of thyroid hormone production and post-natal development. EMBO J 18:623–631[CrossRef][Medline]
  13. Fraichard A, Chassande O, Plateroti M, Roux JP, Trouillas J, Dehay C, Legrand C, Gauthier K, Kedinger M, Malaval L, Rousset B, Samarut J 1997 The T3R {alpha} gene encoding a thyroid hormone receptor is essential for post-natal development and thyroid hormone production. EMBO J 16:4412–4420[CrossRef][Medline]
  14. Rastinejad F, Perlmann T, Evans RM, Sigler PB 1995 Structural determinants of nuclear receptor assembly on DNA direct repeats. Nature 375:203–211[CrossRef][Medline]
  15. Perlmann T, Rangarajan PN, Umesono K, Evans RM 1993 Determinants for selective RAR and TR recognition of direct repeat HREs. Genes Dev 7:1411–1422[Abstract/Free Full Text]
  16. Williams GR, Harney JW, Moore DD, Larsen PR, Brent GA 1992 Differential capacity of wild type promoter elements for binding and trans-activation by retinoic acid and thyroid hormone receptors. Mol Endocrinol 6:1527–1537[Abstract/Free Full Text]
  17. Naar AM, Boutin JM, Lipkin SM, Yu VC, Holloway JM, Glass CK, Rosenfeld MG 1991 The orientation and spacing of core DNA-binding motifs dictate selective transcriptional responses to three nuclear receptors. Cell 65:1267–1279[CrossRef][Medline]
  18. Umesono K, Murakami KK, Thompson CC, Evans RM 1991 Direct repeats as selective response elements for the thyroid hormone, retinoic acid, and vitamin D3 receptors. Cell 65:1255–1266[CrossRef][Medline]
  19. Kato S, Sasaki H, Suzawa M, Masushige S, Tora L, Chambon P, Gronemeyer H 1995 Widely spaced, directly repeated PuGGTCA elements act as promiscuous enhancers for different classes of nuclear receptors. Mol Cell Biol 15:5858–5867[Abstract]
  20. Forman BM, Casanova J, Raaka BM, Ghysdael J, Samuels HH 1992 Half-site spacing and orientation determines whether thyroid hormone and retinoic acid receptors and related factors bind to DNA response elements as monomers, homodimers, or heterodimers. Mol Endocrinol 6:429–442[Abstract/Free Full Text]
  21. Forman BM, Evans RM 1995 Nuclear hormone receptors activate direct, inverted, and everted repeats. Ann NY Acad Sci 761:29–37[Medline]
  22. Williams GR, Zavacki AM, Harney JW, Brent GA 1994 Thyroid hormone receptor binds with unique properties to response elements that contain hexamer domains in an inverted palindrome arrangement. Endocrinology 134:1888–1896[Abstract/Free Full Text]
  23. Muscat GE, Griggs R, Downes M, Emery J 1993 Characterization of the thyroid hormone response element in the skeletal {alpha}-actin gene: negative regulation of T3 receptor binding by the retinoid X receptor. Cell Growth Differ 4:269–79[Abstract]
  24. Muscat GE, Mynett-Johnson L, Dowhan D, Downes M, Griggs R 1994 Activation of myoD gene transcription by 3,5,3'-triiodo-L-thyronine: a direct role for the thyroid hormone and retinoid X receptors. Nucleic Acids Res 22:583–591[Abstract/Free Full Text]
  25. Williams GR, Harney JW, Forman BM, Samuels HH, Brent GA 1991 Oligomeric binding of T3 receptor is required for maximal T3 response. J Biol Chem 266:19636–19644[Abstract/Free Full Text]
  26. Brent GA, Larsen PR, Harney JW, Koenig RJ, Moore DD 1989 Functional characterization of the rat growth hormone promoter elements required for induction by thyroid hormone with and without a co-transfected ß type thyroid hormone receptor. J Biol Chem 264:178–182[Abstract/Free Full Text]
  27. Brent GA, Harney JW, Chen Y, Warne RL, Moore DD, Larsen PR 1989 Mutations of the rat growth hormone promoter which increase and decrease response to thyroid hormone define a consensus thyroid hormone response element. Mol Endocrinol 3:1996–2004[Abstract/Free Full Text]
  28. Ye ZS, Forman BM, Aranda A, Pascual A, Park HY, Casanova J, Samuels HH 1988 Rat growth hormone gene expression. Both cell-specific and thyroid hormone response elements are required for thyroid hormone regulation. J Biol Chem 263:7821–7829[Abstract/Free Full Text]
  29. Norman MF, Lavin TN, Baxter JD, West BL 1989 The rat growth hormone gene contains multiple thyroid response elements. J Biol Chem 264:12063–12073[Abstract/Free Full Text]
  30. Murray MB, Zilz ND, McCreary NL, MacDonald MJ, Towle HC 1988 Isolation and characterization of rat cDNA clones for two distinct thyroid hormone receptors. J Biol Chem 263:12770–12777[Abstract/Free Full Text]
  31. Koenig RJ, Warne RL, Brent GA, Harney JW, Larsen PR, Moore DD 1988 Isolation of a cDNA clone encoding a biologically active thyroid hormone receptor. Proc Natl Acad Sci USA 85:5031–5035[Abstract/Free Full Text]
  32. Wood WM, Dowding JM, Haugen BR, Bright TM, Gordon DF, Ridgway EC 1994 Structural and functional characterization of the genomic locus encoding the murine ß 2 thyroid hormone receptor. Mol Endocrinol 8:1605–1617[Abstract/Free Full Text]
  33. Shi YB, Yaoita Y, Brown DD 1992 Genomic organization and alternative promoter usage of the two thyroid hormone receptor ß genes in Xenopus laevis. J Biol Chem 267:733–738[Abstract/Free Full Text]
  34. Showers MO, Darling DS, Kieffer GD, Chin WW 1991 Isolation and characterization of a cDNA encoding a chicken ß thyroid hormone receptor. DNA Cell Biol 10:211–221[Medline]
  35. Lin BC, Wong CW, Chen HW, Privalsky ML 1997 Plasticity of tetramer formation by retinoid X receptors. An alternative paradigm for DNA recognition. J Biol Chem 272:9860–9867[Abstract/Free Full Text]
  36. Chen H, Privalsky ML 1995 Cooperative formation of high-order oligomers by retinoid X receptors: an unexpected mode of DNA recognition. Proc Natl Acad Sci USA 92:422–426[Abstract/Free Full Text]
  37. Lazar MA, Berrodin TJ, Harding HP 1991 Differential DNA binding by monomeric, homodimeric, and potentially heteromeric forms of the thyroid hormone receptor. Mol Cell Biol 11:5005–5015[Abstract/Free Full Text]
  38. Force WR, Tillman JB, Sprung CN, Spindler SR 1994 Homodimer and heterodimer DNA binding and transcriptional responsiveness to triiodothyronine (T3) and 9-cis-retinoic acid are determined by the number and order of high affinity half-sites in a T3 response element. J Biol Chem 269:8863–8871[Abstract/Free Full Text]
  39. Yen PM, Brubaker JH, Apriletti JW, Baxter JD, Chin WW 1994 Roles of 3,5,3'-triiodothyronine and deoxyribonucleic acid binding on thyroid hormone receptor complex formation. Endocrinology 134:1075–1081[Abstract/Free Full Text]
  40. Miyamoto T, Suzuki S, DeGroot LJ 1993 High affinity and specificity of dimeric binding of thyroid hormone receptors to DNA and their ligand-dependent dissociation. Mol Endocrinol 7:224–231[Abstract/Free Full Text]
  41. Bollag M, Edelstein SJ 1991 Gel electrophoresis under nondenaturing conditions. In: Protein methods. New York: John Wiley and Sons, Inc.; 143–160
  42. Glass CK 1996 Some new twists in the regulation of gene expression by thyroid hormone and retinoic acid receptors. J Endocrinol 150:349–357[Abstract/Free Full Text]
  43. Yang Z, Hong SH, Privalsky ML 1999 Transcriptional anti-repression. Thyroid hormone receptor ß-2 recruits SMRT corepressor but interferes with subsequent assembly of a functional corepressor complex. J Biol Chem 274:37131–3718[Abstract/Free Full Text]
  44. Yoh SM, Privalsky ML 2001 Transcriptional repression by thyroid hormone receptors. A role for receptor homodimers in the recruitment of SMRT corepressor. J Biol Chem 276:16857–16867[Abstract/Free Full Text]
  45. Kurokawa R, Yu VC, Naar A, Kyakumoto S, Han Z, Silverman S, Rosenfeld MG, Glass CK 1993 Differential orientations of the DNA-binding domain and carboxy-terminal dimerization interface regulate binding site selection by nuclear receptor heterodimers. Genes Dev 7:1423–1435[Abstract/Free Full Text]
  46. Williams GR 2000 Cloning and characterization of two novel thyroid hormone receptor ß isoforms. Mol Cell Biol 20:8329–8342[Abstract/Free Full Text]
  47. Furlow JD, Brown DD 1999 In vitro and in vivo analysis of the regulation of a transcription factor gene by thyroid hormone during Xenopus laevis metamorphosis. Mol Endocrinol 13:2076–2089[Abstract/Free Full Text]
  48. Feng X, Jiang Y, Meltzer P, Yen PM 2000 Thyroid hormone regulation of hepatic genes in vivo detected by complementary DNA microarray. Mol Endocrinol 14:947–955[Abstract/Free Full Text]
  49. Flores-Morales A, Gullberg H, Fernandez L, Stahlberg N, Lee NH, Vennstrom B, Norstedt G 2002 Patterns of liver gene expression governed by TRß. Mol Endocrinol 16:1257–1268[Abstract/Free Full Text]
  50. Morte B, Manzano J, Scanlan T, Vennstrom B, Bernal J 2002 Deletion of the thyroid hormone receptor {alpha} 1 prevents the structural alterations of the cerebellum induced by hypothyroidism. Proc Natl Acad Sci USA 99:3985–3989[Abstract/Free Full Text]
  51. Amma LL, Campos-Barros A, Wang Z, Vennstrom B, Forrest D 2001 Distinct tissue-specific roles for thyroid hormone receptors ß and {alpha}1 in regulation of type 1 deiodinase expression. Mol Endocrinol 15:467–475[Abstract/Free Full Text]
  52. Dellovade TL, Chan J, Vennstrom B, Forrest D, Pfaff DW 2000 The two thyroid hormone receptor genes have opposite effects on estrogen-stimulated sex behaviors. Nat Neurosci 3:472–475[CrossRef][Medline]
  53. Plateroti M, Chassande O, Fraichard A, Gauthier K, Freund JN, Samarut J, Kedinger M 1999 Involvement of T3R{alpha}- and ß-receptor subtypes in mediation of T3 functions during postnatal murine intestinal development. Gastroenterology 116:1367–1378[CrossRef][Medline]
  54. Chen H, Smit-McBride Z, Lewis S, Sharif M, Privalsky ML 1993 Nuclear hormone receptors involved in neoplasia: erb A exhibits a novel DNA sequence specificity determined by amino acids outside of the zinc-finger domain. Mol Cell Biol 13:2366–2376[Abstract/Free Full Text]
  55. Kersten S, Gronemeyer H, Noy N 1997 The DNA binding pattern of the retinoid X receptor is regulated by ligand-dependent modulation of its oligomeric state. J Biol Chem 272:12771–12777[Abstract/Free Full Text]

NURSA Molecule Pages Link:

Nuclear Receptors:   TRα  |  TRβ  |  RXRα
Coregulators:   AIB1  |  NCOR  |  SMRT
Ligands:   Thyroid hormone



This article has been cited by other articles:


Home page
 EndocrinologyHome page
E. R. Nelson and H. R. Habibi
Functional Significance of a Truncated Thyroid Receptor Subtype Lacking a Hormone-Binding Domain in Goldfish
Endocrinology, September 1, 2008; 149(9): 4702 - 4709.
[Abstract] [Full Text] [PDF]

Home page
 EndocrinologyHome page
P. de Lange, A. Feola, M. Ragni, R. Senese, M. Moreno, A. Lombardi, E. Silvestri, R. Amat, F. Villarroya, F. Goglia, et al.
Differential 3,5,3'-Triiodothyronine-Mediated Regulation of Uncoupling Protein 3 Transcription: Role of Fatty Acids
Endocrinology, August 1, 2007; 148(8): 4064 - 4072.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow NURSA Molecule Pages Link
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Mengeling, B. J.
Right arrow Articles by Privalsky, M. L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Mengeling, B. J.
Right arrow Articles by Privalsky, M. L.

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