A novel thyroid hormone receptor-beta mutation that fails to bind nuclear receptor corepressor in a patient as an apparent cause of severe, predominantly pituitary resistance to thyroid hormone

A Novel Thyroid Hormone Receptor-

␤ Mutation That

Fails to Bind Nuclear Receptor Corepressor in a Patient

as an Apparent Cause of Severe, Predominantly

Pituitary Resistance to Thyroid Hormone

Sharon Y. Wu,* Ronald N. Cohen,* Enver Simsek, Dursun A. Senses, Nese E. Yar, Helmut Grasberger,

Janet Noel, Samuel Refetoff, and Roy E. Weiss

Departments of Medicine (S.Y.W., R.N.C., H.G., J.N., S.R., R.E.W.) and Pediatrics (S.R.), University of Chicago, Chicago Illinois 60645; and Department of Pediatrics (E.S., D.A.S., N.E.Y.), Abant Izzet Baysal University, Duzce Medical School, 14450 Duzce, Turkey

Context: Resistance to thyroid hormone (RTH) is a dominantly in-herited syndrome of variable tissue hyporesponsiveness to thyroid hormone (TH).

Objective: We report a newborn who presented with severe RTH (Mkar) with serum TSH 1500 mU/liter and free T3greater than 50 pM (normal 3.1–9.4) and free T425.3 pM(normal 12–22). We hypothesized that the RTH was due to reduced ligand binding and/or abnormal interaction with nuclear cofactors.

Design: These were prospective in vivo and in vitro studies. Setting: The study was conducted at a tertiary care university hospital. Patients: Patients included a newborn child and two other subjects with RTH.

Intervention: The effect of various TH-lowering agents in the subject with RTH was studied. In vitro studies including EMSA and mam-malian two-hybrid assay as well as in vitro transfection studies were conducted.

Main Outcome Measures: Sequencing of the TH receptor (TR)␤ and

in vitro measurements of receptor-cofactor interaction were measured.

Results: Sequencing of the TR␤ demonstrated a de novo heterozygous mutation, 1590_1591insT, resulting in a frameshift producing a mu-tant TR␤ (mutTR)-␤ with a 28-amino acid (aa) nonsense sequence and 2-amino acid carboxyl-terminal extension. The Mkar mutation was evaluated in comparison to three other TR␤ frameshift mutations in the carboxyl terminus. EMSA demonstrated that the Mkar mutTR␤1 had impaired ability to recruit nuclear receptor corepressor but intact association with silencing mediator of retinoid and thyroid receptor (SMRT).

Conclusion: Our data suggest that alterations in codons 436 – 453 in helix 11 result in significantly diminished association with nuclear receptor corepressor but not SMRT. This novel mutTR␤ demonstrates nuclear corepressor specificity that results in severe predominantly pituitary RTH due to impaired release of SMRT. (J Clin Endocrinol Metab 91: 1887–1895, 2006)

R

ESISTANCE TO THYROID hormone (RTH), an

inher-ited syndrome of tissue hyporesponsiveness to thyroid

hormone of variable degree, is characterized by elevated

serum thyroid hormone levels (TH), normal or elevated

se-rum TSH levels, and goiter (1). The molecular basis of RTH

in approximately 85% of subjects is heterozygous mutations

of the TH receptor (TR)-␤ gene, resulting in impaired T

3

binding and/or transactivation function.

The mechanism of RTH in subjects heterozygous for a

mutant TR␤ (mutTR) is interference with the function of the

wild-type TR

␤, known as a dominant-negative effect (DNE).

One mechanism of DNE is defective interaction of the mutTR

with a cofactor (2). The severity of resistance depends on the

degree of impaired ligand binding and interaction with

tis-sue-specific nuclear cofactors (3).

Most of mutTR

␤s are able to bind corepressor in the

ab-sence of ligand but cannot release corepressor in response to

ligand (4 –11). The two main corepressors are nuclear

recep-tor corepressor (NCoR) and silencing mediarecep-tor of retinoid

and thyroid hormone receptors (SMRT). Lack of hormone

binding results in impairment of corepressor release and

decreased coactivator recruitment. NCoR and SMRT are

ubiquitously expressed in vertebrate tissues but interact

dif-ferently with nuclear receptors. TR␤ preferentially recruits

NCoR instead of SMRT via the N-terminal interaction

do-mains (N3 and N2) (12, 13). SMRT, on the other hand, prefers

to bind to the retinoic acid receptor isoforms mediated by S2,

the most proximal interaction domain.

We describe a newborn with severe RTH (Mkar) who had

a heterozygous mutation in the TR

␤ gene, 1590_1591insT,

First Published Online February 7, 2006

* S.Y.W. and R.N.C. contributed equally to the study and both should be considered as first authors.

Abbreviations: aa, Amino acid; AM, artificial mutant; BCS, bovine calf serum; DNE, dominant-negative effect; FT4, free T4; GST, glutathione-S-transferase; IVT, in vitro translated; Luc, luciferase; mutTR, mutant TR; NCoR, nuclear receptor corepressor; RTH, resistance to thyroid hor-mone; RXR, retinoid X receptor; SMRT, silencing mediator of retinoid and thyroid hormone receptor; SRC, steroid receptor coactivator; TH, thyroid hormone; TR, TH receptor; TRE, TH response element; TRIAC, 3,5,3⬘-triiodothyroacetic acid, also called tiratricol; WT, wild type. JCEM is published monthly by The Endocrine Society (http://www. endo-society.org), the foremost professional society serving the en-docrine community.

doi: 10.1210/jc.2005-2428

resulting in a frameshift that produces a nonsense amino acid

(aa) sequence 436 – 463 and a carboxyl terminal extension of

TR␤1 by 2 aa. In this study, we evaluated three natural and

one artificial frameshift mutations in helix 11 and 12 of TR

␤

to delineate the role of this C-terminal region in cofactor

interactions. We used EMSAs and mammalian two-hybrid

assays to evaluate the ability of TR␤s with mutations in helix

11 and 12 to bind to corepressors NCoR and SMRT and

coactivator, steroid receptor coactivator (SRC)-1. The

func-tional properties of transactivation, repression, and DNE

were evaluated in transient transfection studies. These

stud-ies demonstrated that mutations resulting in alterations in

the 436- to 453-aa sequence in the C terminus of TR␤ result

in significantly impaired association with NCoR but not

SMRT. This results in severe but predominantly pituitary

RTH.

Patients and Methods Case report of Mkar

Mkar is a Turkish male born at 38 wk to healthy nonconsanguinous parents. At 6 d of age, he had respiratory distress, tachycardia (190 beats/min), diaphoresis, and a gland four times the normal size. Serum TSH was 220 mU/liter (normal range⬍ 20) with free T3of 50 pmol/liter

(3.7– 8.6), and free T4(FT4) of 25.3 pmol/liter (12–22), consistent with severe RTH. TRH stimulation test demonstrated an increase in TSH from 405 to 1716 mU/liter at 30 min. Magnetic resonance imaging showed a normal pituitary gland. Echocardiogram was normal and knee x-ray showed an appropriate bone age. Neither parent has a history of thyroid disease and both have normal thyroid function.

Treatment with propylthiouracil decreased the FT4to 2.45 pmol/liter, resulting in a marked rise in TSH to 1172 mU/liter. (Fig. 1). l-T3was then started in an attempt to lower the TSH but had little effect and was stopped after 8 wk. Two weeks treatment with octreotide did not sup-press the serum TSH. 3,5,3⬘-Triiodothyroacetic acid (TRIAC) at escalat-ing doses up to 1.4 mg/d resulted in a decrease in heart rate and diaphoresis and marked decrease in TSH to 2.36 mU/liter (Fig. 1).

Case report of Mdbs

Mdbs was a newborn male initially misdiagnosed as having hyper-thyroidism who after a subtotal thyroidectomy had increased TSH of 77 mU/liter, normal total T4, and elevated total T3of 300 ng/dl. TRH test demonstrated robust response of TSH (60 min maximum 1500 mU/ liter). T3failed to suppress the TRH-stimulated TSH until 200␮g T3per day was administered for 3 sequential days.

Other human subjects

PV is a previously described male with RTH who presented at age 7 yr with a TSH of 4.2 mU/liter (0.5– 4), total T3of 369 ng/dl (88 –162) and FT4of 4.6 ng/dl (1–1.9) (14). The PV mutation, 1627_1628incC, produces

FIG. 1. Thyroid function tests of Mkar over 9 months and different treatments: propylthiouracil (PTU), T3, octreotide, and TRIAC. See text for details. Note that TRIAC cross reacts with T4in the FT4assay.

a frame shift resulting in a 16-aa nonsense sequence after the insertion point and extension of the length of the TR␤1 protein by 2 aa.

TR␤ sequencing

DNA was extracted from circulating white blood cells of Mkar and his parents and from skin fibroblasts of Mdbs. Studies carried out in humans were approved by the institutional review boards, and all individuals or parents of minors gave informed consent. PCR was per-formed on 100 ng of DNA from Mkar to amplify the coding exons 3–10 of the TR␤1 gene and exon 1 of TR␤2. The products of amplification were sequenced directly using automated fluorescence-based sequencing. Ex-ons 8, 9, and 10 of TR␤1 of Mkar’s parents and Mdbs were also amplified and sequenced. The set of primers used to amplify and sequence all coding exons of TR␤1 and -␤2 have been previously described (15, 16).

Fibroblast studies

Human skin was obtained by punch biopsy from Mkar after informed consent. This was approved by the institutional review boards. Fibro-blasts were grown in DMEM (Invitrogen, Carlsbad, CA) supplemented with 10% bovine calf serum (BCS) as previously described in detail (17). Isolation of RNA and preparation of cDNA have been previously de-scribed (18). PCR amplification of a 460-bp segment containing the T insertion at codon 436 as well as the wild-type (WT) sequence was carried out using Mkar cDNA and a control cDNA from fibroblasts of a normal subject. The primers used for this PCR were 5⬘-ATG GAG ATC ATG TCC CTT CGC-3⬘ (forward) and 5⬘-TCT AAT CCT CGA ACA CTT CCA AGA A-3⬘ (reverse). To generate cDNA, the PCR used 150 ng RNA as the template and 10 pmol each of the primers. The first denaturation was set at 94 C for 3 min, followed by 35 cycles of 1-min denaturation at 94 C, 1 min of annealing at 60 C, and 1 min of extension at 72 C. The final extension was for 5 min at 72 C.

Allele-specific digestion of the PCR products used the restriction enzyme BsrD1 (New England Biolabs, Ipswich, MA). The digested prod-ucts were run on a 1.5% agarose gel. Mkar BsrD1 digestion produces two fragments of 380 and 80 bp each in the presence of the Mkar mutation.

Construction of plasmids

TR␤ constructs. The wild-type (WT) TR␤1 expression vector and the

manner of its generation was previously described (3). This and other expression plasmids were cloned in pcDNAI/amp (Invitrogen).

Plasmids expressing the two mutant TR␤1s (Mkar and Mdbs) were constructed by exchange of PCR-amplified DNA fragments containing the mutations with the corresponding WT TR␤ insert in the pcDNAI/ amp plasmid. This was done using the PGEM-T-Easy vectors (Promega Corp., Madison, WI).

Plasmids expressing the other two mutant TR␤1s (AM and PV) were constructed by directed mutagenesis (Invitrogen GeneTailor site-directed mutagenesis system). The artificial mutant (AM) that contains only the upstream nonsense sequence of 11 aa of Mkar was constructed as follows. A “T” was inserted between nucleotide 1590 and 1591 to duplicate the mutation of Mkar, and the resulting frameshift was re-paired by deletion of the C at nucleotide 1626, corresponding to the PV mutation site. The PV mutant was created by inserting a “C” between nucleotide 1627 and 1628 in codon 448, creating a frameshift resulting in a 16-aa nonsense carboxyl terminal sequence.

All constructs were checked for correct DNA insertion and presence of mutation by sequencing.

The preparation of Gal4-TR, VP16-NCoR, and VP-16-SMRT has been previously described (19). Mutant Gal4-TR constructs, Gal4-Mkar and Gal4-PV, were constructed using site-directed mutagenesis (QuikChange site-directed mutagenesis kit, Stratagene, La Jolla, CA) on the Gal4-TR plasmid.

Reporter constructs

Four firefly luciferase (Luc) reporter constructs were used, two pos-itively (Pal x3Luc and F2) and two negatively regulated (TH␣SU-Luc and UAS Tk-Luc) prepared as previously described (19).

EMSAs

EMSAs were carried out as previously described with a32 P-radio-labeled DR⫹4 probe (20). Glutathione-S-transferase (GST)-fusion pro-teins GST-SMRT and GST-SRC-1 were prepared as previously described (13). Nuclear receptors [TR␤1 WT, Mkar, Mdbs, AM, and PV as well as retinoid X receptor RXR)-␣] and NCoR were in vitro translated (IVT) in reticulocyte lysate (Promega) using T7 polymerase. An equivalent GST NCoR construct containing all three interacting domains did not gen-erate usable amounts of protein. Therefore, IVT NCoR containing all three interacting domains was used in these experiments. EMSAs used 4␮l of the TR␤1 constructs, 4 ␮l of IVT NCoR, and 2 ␮l of IVT RXR␣. Incubations were carried out for 20 min, complexes were resolved on a 5% nondenaturing gel, and films were developed after 12–24 h exposure.

Mammalian two-hybrid assays

Mammalian two-hybrid transient transfections were performed in 293T cells maintained in DMEM supplemented with 10% BCS at 37 C 5% CO2with lipofectamine (Life Technologies, Inc., Gaithersburg, MD) such that each well contained equal amounts of plasmid DNA. One hundred nanograms each of reporter UAS-TK Luc, the appropriate construct (Gal4-TR, Gal4-Mkar, Gal4-PV, or empty Gal4), and the core-pressor (VP16-NCoR, VP16-SMRT, or empty VP16) were added to each well. Three hours after transfection, cells were washed with PBS and exposed to DMEM containing 10% steroid hormone-depleted BCS. Then 21–27 h after transfection, cells were lysed and assayed for luciferase activity. Experiments were performed in triplicate. Data are expressed as fold stimulation⫾ sem.

Transient transfection studies

CV-1 cells that lack endogenous TRs were used to evaluate ligand-independent activity and repression. For all other transfections, the human hepatoblastoma cell line (HepG2) was maintained in DMEM supplemented with 10% BCS at 37 C 5% CO2. Before transfection, cells were transferred to 12-well plates in the complete medium. Transfection was performed with TransFast transfection reagent (Promega) at a 1.3:1 charge ratio of TransFast reagent to DNA. The DNA and TransFast were mixed for 10 –15 min before addition to cells. After 24 h the medium was changed to one containing 5% of TH-depleted BCS without and with added T3, and cells were incubated for an additional 48 h before har-vesting. Firefly luciferase and Renilla-TK luciferase activities were de-termined sequentially (Promega dual-luciferase reporter assay system). TH-depleted BCS was prepared as previously described (3).

Reporter plasmids regulated positively by TH, PALx3-Luc, and F2x3-Luc and negatively by TH and␣SU-Luc were transfected into the cells at concentrations of 1␮g/well. Empty pcDNA vectors, WT TR␤1, and mutTR␤1 constructs were transfected at 25–50 ng/well. To test for a DNE, equal amounts of WT TR␤1 and mutTR␤1s at 25 ng each per well were cotransfected. An additional transfection using the PALx3-Luc reporter and varying ratios of mutTR␤1 to WT TR␤1 constructs, from 1:1 (25:25 ng) to 0.0625:1 (1.56:25 ng), was performed.

Experiments were performed in triplicate and expressed as fold in-duction⫾ sem.

Results

Two novel frameshift mutations were found in the TR␤ of

Mkar and Mdbs

Sequencing all coding exons of the TR

␤1 and TR␤2 genes

of Mkar revealed the subject to be heterozygous for a novel

single nucleotide insertion (1590_1591insT), resulting in the

extension of the length of the protein by 2 aa with 28 nonsense

aa after the insertion point (Fig. 2). This is a de novo mutation

because both parents have a normal thyroid phenotype and

absence of the TR

␤ gene mutation identified in the

propos-itus. The mutation results in a new restriction enzyme site for

BsrD1. Digestion of the amplified PCR product of exon 10 of

absence of the mutation in his parents (data not shown). Both

WT and mutant alleles were detected in skin fibroblasts of

MKar.

Sequencing of the TR

␤1 gene of Mdbs revealed a single

nucleotide insertion, (1643_1644insC) (Fig. 2).

Homodimerization and heterodimerization were intact in mutTRs

To determine whether the mutant TRs retained the ability

to bind DNA, an EMSA analysis was performed with a

32

_{P-radiolabeled DR}

_{⫹4 probe (Fig. 3A). WT TR␤1 and all}

mutTR

␤1s demonstrated intact homodimerization. Addition

of 100 nm of T

3

dissociated WT TR

␤1 homodimer but had no

significant effect on mutTR

␤1 homodimers due to lack of T

3

binding. All mutTR

␤1s also retained the ability to bind RXR

(Fig. 3A).

Codons 436 – 453 are necessary for NCoR binding

When the receptors were studied with IVT NCoR in

EMSAs, the WT TR

␤1 bound to NCoR in the absence of T

3

and released NCoR when 100 nm of T

3

were added. The

Mdbs mutant interacted with NCoR in the absence of T

3

as

did the WT TR

␤1 but continued to interact with NCoR when

100 nm of T

3

were added. Surprisingly, however, in the

presence of IVT NCoR and the absence of T

3

, the Mkar

mutant did not bind NCoR (Fig. 3B).

To determine the region of the carboxyl-terminal domain

that is involved in corepressor binding, we studied the

in-teractions of IVT NCoR in the two other mutTR

␤1s AM and

PV also in gel shift assays. AM contains the proximal

non-sense aa sequence of Mkar, codons 436 – 447, whereas PV

shares the distal nonsense aa sequence from codon 448 to 463

with Mkar (Fig. 2). AM and PV also did not bind NCoR in

the absence or the presence of 100 nm of T

3

(Fig. 3B). The

absence of N-CoR binding by Mkar, AM, and PV point to an

abnormality in the C-terminal domain of TR

␤ localized to the

region between codons 436 and 453.

To confirm the absence of NCoR binding in the Mkar

mutant, a mammalian two-hybrid assay was used to

eval-uate interactions between VP-16-bound NCoR and the

mutTR␤1s, Mkar and PV, both bound to a Gal4 DNA binding

domain. Three different doses of NCoR were used: 0.01, 0.05,

and 0.1

␮g. Luc activity of the mutTR␤1s was corrected for

basal empty Gal4 activity and expressed as fold interaction.

Gal4 Mkar and Gal4 PV were both found to have significantly

less interaction with all three doses of the VP16-NCoR

con-struct, compared with Gal 4 WT (P

ⱕ 0.01), confirming the

EMSA findings (Fig. 3C). NCoR failed to bind to the

Mkar-RXR

␣ heterodimer (Fig. 3D).

Increasing amounts of WT TR

␤1, when added to a constant

amount of Mkar mutant (simulating a possible in vivo

situ-ation), increased binding to NCoR as demonstrated (Fig. 4A).

When the opposite experiment was performed and

increas-ing amounts of Mkar mutant were added to a constant dose

of WT TR␤1, binding to NCoR diminished (Fig. 4B). This is

consistent with diminished binding of NCoR to the Mkar

mutTR.

Ligand-independent repression does not reflect absence of

NCoR association with the mutTR␤1s

Ligand-independent repression and activation of

posi-tive and negaposi-tive TH response elements (TREs) by the

mutant receptors was evaluated in transient transfection

studies using CV-1 cells. Mean Luc activity in the presence

of the PALx3Luc reporter for WT TR

␤1, Mkar, Mdbs, AM,

and PV mutTR

␤1s were 0.576 ⫾ 0.03, 0.5 ⫾ 0.06, 0.45 ⫾

0.01, 0.41

⫾ 0.06, 0.52 ⫾ 0.07, respectively (Fig. 5A). Basal

repression by Mkar was not significantly different from

WT. Mdbs and AM had significantly greater repression

than WT. These findings demonstrate that basal repression

on a positively regulated TRE is not impaired despite the

EMSA findings of absent NCoR interaction with the

mu-tants Mkar, AM, and PV.

FIG. 2. C-terminal aa sequence of TR␤1 constructs, WT, and mutants Mkar, Mdbs, PV, and AM, from codons 435 to 463. A thymine (T 1591) insertion at codon 436 in mutant Mkar produces a frameshift and nonsense of 28 subsequent aa with extension of the TR␤1 and TR␤2 proteins by 2 aa. Mdbs has a cytosine (C 1644) insertion at codon 453, producing a frameshift and extension of the TR␤ protein. Mdbs shares with Mkar the distal abnormal aa sequence from codons 453 to 463. PV contains a more proximal insertion of a C 1627 at codon 448 and shares with Mkar the aa sequence from codons 448 to 463. The artificial mutant, AM, contains, like Mkar, a T 1591 insertion at codon 436 but also has a C 1626 deletion to revert the distal AA sequence back to that of WT TR␤s. Mutated aa sequences resulting from frameshifts are in gray. Clinical severity of the RTH is rated from mild (_{⫹) to very severe (⫹⫹⫹). Underlined gray aa correlate with the portion of helix 11 and 12 thought to be involved} in NCoR binding, as described in the text.

mutTR␤1s exhibit impairment of ligand-dependent transactivation and repression on positive and negative TREs, respectively, and demonstrate strong DNE

T

3

-dependent transactivation by the mutTR

␤1s was tested

at a T

3

concentration of 10 nmol/liter. This dose was chosen

because it was shown to result in a significant increase or

decrease from baseline with positive and negative reporters,

respectively, yet still allowed differences to be seen among

mutations (3). All the mutTR

␤1s exhibited severe

impair-ment of T

3

-induced transactivation with the two positively

regulated reporters, PALx3-Luc and F2x3. Relative Luc

ac-tivity of the WT TR

␤1 using PALx3Luc in the presence of 10

nm of T

3

was 762

⫾ 111, compared with only 9.5 ⫾ 1.5 with

the mutTR

␤1s. No significant differences were seen among

the mutTR

␤1s. Similarly, the WT TR␤1 exhibited 14- to

23-fold greater relative Luc activity using the F2x3 Luc reporter,

compared with the mutTR

␤1s (Fig. 5B).

T

3

-dependent repression of the

␣SU-Luc reporter, a

neg-atively regulated TRE, was evaluated. Transfection of WT

and mutTR

␤1s in the presence of 10 nm of T

3

demonstrated

appropriate repression of activity in the WT TR

␤1 and

sig-nificantly impaired repression with the mutTR

␤1s (Fig. 5C).

Data were normalized to 100% to compare the degree of

repression. Ligand-independent activation was not

signifi-cantly different between the mutTR

␤1s and WT (data not

shown).

FIG. 3. Evaluation of mutTR␤1 homo- and heterodimerization and their interaction with NCoR, assessed by EMSA. A, All C-terminal frameshift

mutants retain the ability to form homodimers and heterodimers. In the absence of T3, intact homodimer and heterodimer formation is seen (lanes 1–5 and 11–15). Addition of 100 nMof T3dissociates the WT TR␤1 homodimer appropriately (absence of band in lane 6), whereas T3does not dissociate the mutTR␤1 homodimers (persistent band in lanes 7–10) due to reduced ability to bind ligand. The corresponding mutant aa sequences are shown in gray. B, Mutants Mkar, PV, and AM are unable to associate with NCoR, localizing the area involved in NCoR binding to codons 436 – 453. WT and Mdbs TR␤1s bind to IVT NCoR (lanes 6 and 8), whereas Mkar, AM, and PV demonstrate no association with NCoR (lanes 7, 9, and 10). When 100 nMof T3are added, WT TR␤1 releases NCoR (lane 11), whereas Mdbs retains binding (lane 13). The aa sequence shared in common by Mkar, AM, and PV is localized between codons 436 and 453. C, Absence of NCoR binding is confirmed in a mammalian two-hybrid assay using Gal4-TR, Gal4-Mkar, and Gal4-PV. VP-16-NCoR is added in increasing concentrations from 0.01 to 0.1_{␮g. Gal4-TR␤1} has significantly greater interaction with all doses of VP-16-NCoR than mutants Gal4-Mkar and Gal4-PV. D, NCoR does not bind to the Mkar-RXR_{␣ heterodimer. IVT WT TR␤1 and Mkar are combined with RXR␣ IVT NCoR to determine whether NCoR has preferential binding} to the mutant heterodimer, compared with the mutant homodimer. WT TR␤1 binds to NCoR as a homo- or heterodimer with RXR␣ (lane 3), but no binding is seen to Mkar-RXR_{␣ (lane 4).}

DNE of the mutTR

␤1s was determined by cotransfecting WT

TR

␤1 with mutTR␤1s with the reporter PALx3Luc (Fig. 5D). All

mutants exhibited a significant DNE over the WT TR

␤1. As the

ratio of mutant to WT TR

␤1 decreased, diminishing

dominant-negative inhibition of activity was seen (Fig. 5D).

Thus, all the mutant receptors demonstrate a similar in

vitro function despite differences in NCoR association. All

mutants have severely impaired ligand-dependent

transac-tivation and repression on positively and negatively

regu-lated TREs, respectively, as well as strong DNE.

All mutTR␤1s bind SMRT in the absence of T3and

demonstrate impaired ligand-dependent release of SMRT

A second corepressor, SMRT, was then studied as a

pos-sible mechanism for the ligand-independent repression and

strong DNE seen in the transient transfections. All mutTR

␤1s

were found to interact with SMRT in the absence of T

3

and

to have impaired release of SMRT when 100 nm of T

3

was

added (Fig. 6A). The interaction between SMRT and Mkar

was confirmed in a mammalian two-hybrid assay using

VP-16 SMRT (Fig. 6B).

MutTR␤1s have impaired ligand-dependent recruitment

of SRC-1

Given the presence of SMRT binding and impaired release

in the presence of T

3

, EMSA was then used to evaluate

whether or not coactivator recruitment was defective. At a

dose of 1.5

␮g of GST-SRC-1, the WT TR␤1 had normal

recruitment and binding of SRC-1, whereas the mutTR

␤1s,

Mkar and Mdbs, were unable to recruit SRC-1 in the presence

FIG. 4. Mkar TR␤1 interferes with WT TR␤1 binding to NCoR. A, Increasing doses of WT TR␤1 (0–8 ␮l) are added to a constant dose of Mkar (2␮l) and NCoR. As the dose of WT TR␤1 increases, interaction with NCoR increases. B, Increasing doses of mutant Mkar (0–8 ␮l) are added to a constant dose of WT TR␤1 (2 ␮l) and NCoR. Interaction of the WT TR␤1-Mkar heterodimer with NCoR diminishes as the dose of Mkar increases, consistent with interference of corepressor interaction by Mkar.

of 1000 nm of T

3

and even 10,000 nm of T

3

(Fig. 6C). These

results confirm that the C-terminal domain is an area

in-volved in exchange of corepressors and coactivators,

de-pending on the absence or presence of T

3

.

Discussion

The interaction of corepressors and coactivators with

sev-eral mutTR

␤1s has previously been studied. Almost all

mutTR

␤s causing RTH tested in these studies were able to

bind to corepressor in the absence of hormone but failed to

release corepressor with the addition of T

3

(4, 6 –9, 11, 21, 22).

This impaired corepressor release prevents conformational

change of the TR

␤ to allow coactivator recruitment (8, 23).

Thus, the mechanism of RTH and DNE may be due to

im-paired ligand-dependent corepressor release, coactivator

re-cruitment, or both. Many groups (4 –9) note that impairment

of corepressor release is the more important mechanism of

impaired TH action. Tagami et al. (7) showed that

dominant-negative potency of mutTR

␤1s strongly correlated with TR

interactions with NCoR (r

⫽ 0.987), whereas in contrast,

correlation of dominant-negative potency with SRC-1

bind-ing was very low (r

⫽ 0.051). Studies have also shown a

strong correlation between DNE and the inability of

mutTR

␤1s to release the corepressor, SMRT, in the presence

of ligand (4, 24).

Our studies demonstrate a novel finding of absent NCoR

binding in a naturally occurring mutTR␤. The impairment of

NCoR interaction with the Mkar mutant was evaluated by

both EMSA and mammalian two-hybrid assays. By studying

other mutant TRs with altered aa sequences in helix 11 and

12, we were able to localize the area responsible for this

abnormality in NCoR binding to codons 436 – 453 of TR

␤1.

FIG. 5. All mutTR␤1s have similar in vitro activity with severe impairment of transactivation and strong DNE despite differences in ability to interact with NCoR. Experiments performed are transient transfection assays in CV-1 or HepG2 cells using reporter vectors regulated positively and negatively by T3. Results are expressed as relative Luc activity. See text for details. A, Ligand-independent repression is intact in Mkar, AM, and PV despite inability to bind NCoR. Ligand-independent repression using PALx3 Luc was measured in a transient transfection assay in CV-1 cells and found to be similar in WT TR␤1, Mkar, and PV receptors. Mdbs and AM actually demonstrated significantly greater repression than WT TR_{␤1. B, Ligand-dependent transactivation in the presence of 10 n}Mof T3using the two reporters, PALx3 Luc and F2x3 Luc, positively regulated by T3, was found to be severely impaired in the mutTR␤1s, compared with the WT TR␤1. C, Ligand-dependent repression on a reporter negatively regulated by T3, TSH␣, was significantly impaired in Mkar, Mdbs, and AM in the presence of 10 nMof T3. Data were normalized to 100% to compare the degree of repression. D, All mutTR␤1s exert strong DNE. Mkar, Mdbs, and AM were transfected in varying amounts with a constant amount of WT TR_{␤1 and 10 n}Mof T3. As the ratios of mutTR␤1 to WT TR␤1 decreases, the relative luciferase activity increases.

The mutated aa sequence shared by the three mutTR

␤1s is

unable to bind to NCoR. This may be the mechanism by

which these extensive frameshift mutations diminish or

abol-ish corepressor binding. These findings have not been

de-scribed to date in any other case of RTH.

The nuclear receptor corepressor specificity demonstrated

by this novel mutation has been observed in another

mutTR

␤1, namely R429Q (13). This naturally occurring

mutTR

␤1 was found to be defective in its ability to

ho-modimerize. That mutation was unable to recruit NCoR as

either a homodimer or heterodimer with RXR but had the

ability to recruit SMRT as a heterodimer only. Clinically the

subject had primarily pituitary manifestations of RTH, as did

Mkar with elevated serum TSH, which failed to suppress

with high TH levels, causing thyrotoxicosis at the level of

peripheral tissues. Furthermore, R429Q is associated with

impaired ligand-dependent repression on TRH and TSH

␤-negative TREs but normal to enhanced function on positive

TREs (25). R429Q differs from our mutation in that

ho-modimerization was not defective in Mkar and Mkar had

impaired ligand-dependent function on positive and

nega-tive TREs.

Our data add to the observation that SMRT

␣, one of the

multiple SMRT isoforms recently described, may be an

ex-ample of how corepressor function is adapted to different

cells (26). This case of severe RTH demonstrates nuclear

corepressor specificity in TH action and the importance of the

C-terminal domain, in particular codons 436 – 453, in NCoR

interaction.

Acknowledgments

We thank Xiao-Hui Liao for her expert assistance in the construction of vectors.

Received November 7, 2005. Accepted January 31, 2006.

Address all correspondence and requests for reprints to: Roy E. Weiss, M.D., Ph.D., Thyroid Study Unit, Department of Medicine, University of Chicago, 5841 South Maryland Avenue, Mail Code 3090, Chicago, Illinois 60645. E-mail: rweiss@medicine.bsd.uchicago.edu.

This work was supported by National Institutes of Health Grants RR18372, RR00055, DK07011, and DK15070.

FIG. 6. MutTR␤1s demonstrate impaired release of SMRT and impaired recruitment of SRC-1. A, All mutTR␤1s bind to SMRT. EMSA was performed using 200 ng and 1␮g of SMRT and 4 ␮l of IVT WT TR␤1 and mutTR␤1s. A supershifted band representing TR bound to SMRT is seen in lanes 1–10 in the absence of T3. When T3is added, WT TR␤1 homodimer dissociates and releases SMRT as expected (lane 11), whereas the mutTR␤1s demonstrate intact homodimers and continue to bind strongly to SMRT (lanes 12–15). B, Mkar interaction with SMRT is confirmed in a mammalian two-hybrid assay. Gal4-TR␤1 and Gal4-Mkar were transfected with VP16-SMRT at a dose of 0.1 ␮g. Both Gal4-TR␤1 and Gal4-Mkar demonstrate strong interaction with VP16-SMRT. Interaction between VP16-SMRT and Gal4-Mkar is significantly greater than with Gal4-WT␤1. C, MutTR␤1s were unable to recruit SRC-1 in the presence of largely supraphysiological doses of 1,000 nMof T3and 10,000 nMof T3. EMSA was performed using IVT WT TR␤1, Mkar, Mdbs, and 1.5 ␮g of GST-SRC-1. WT TR␤1 recruits SRC-1 at both T3doses (lanes 4 and 10), whereas Mkar and Mdbs are unable to associate with SRC-1 at either T3dose (lanes 5, 6, 11, and 12).

Disclosure of Potential Conflicts of Interest: S.Y.U., R.N.C., E.S., D.A.S., N.E.Y., H.G., J.N., and R.E.W. have nothing to declare. S.R. is a consultant for Quest.

References

1. Refetoff S, Weiss RE, Usala SJ 1993 The syndromes of resistance to thyroid hormone. Endocr Rev 14:348 –399

2. Chatterjee VK, Nagaya T, Madison LD, Datta S, Rentoumis A, Jameson JL 1991 Thyroid hormone resistance syndrome. Inhibition of normal receptor function by mutant thyroid hormone receptors. J Clin Invest 87:1977–1984 3. Hayashi Y, Weiss RE, Sarne DH, Yen PM, Sunthornthepvarakul T, Marcocci

C, Chin WW, Refetoff S1995 Do clinical manifestations of resistance to thyroid hormone correlate with the functional alteration of the corresponding mutant thyroid hormone-␤ receptors? J Clin Endocrinol Metab 80:3246–3256 4. Yoh SM, Chatterjee VK, Privalsky ML 1997 Thyroid hormone resistance

syndrome manifests as an aberrant interaction between mutant T3 receptors and transcriptional corepressors. Mol Endocrinol 11:470 – 480

5. Yoh SM, Privalsky ML 2000 Resistance to thyroid hormone (RTH) syndrome reveals novel determinants regulating interaction of T3receptor with core-pressor. Mol Cell Endocrinol 159:109 –124

6. Tagami T, Jameson JL 1998 Nuclear corepressors enhance the dominant neg-ative activity of mutant receptors that cause resistance to thyroid hormone. Endocrinology 139:640 – 650

7. Tagami T, Gu WX, Peairs PT, West BL, Jameson JL 1998 A novel natural mutation in the thyroid hormone receptor defines a dual functional domain that exchanges nuclear receptor corepressors and coactivators. Mol Endocrinol 12:1888 –1902

8. Safer JD, Cohen RN, Hollenberg AN, Wondisford FE 1998 Defective release of corepressor by hinge mutants of the thyroid hormone receptor found in patients with resistance to thyroid hormone. J Biol Chem 273:30175–30182 9. Nagaya T, Fujieda M, Seo H 1998 Requirement of corepressor binding of

thyroid hormone receptor mutants for dominant negative inhibition. Biochem Biophys Res Commun 247:620 – 623

10. Collingwood TN, Wagner R, Matthews CH, Clifton-Bligh RJ, Gurnell M,

Rajanayagam O, Agostini M, Fletterick RJ, Beck-Peccoz P, Reinhardt W, Binder G, Ranke MB, Hermus A, Hesch RD, Lazarus J, Newrick P, Parfitt V, Raggatt P, de Zegher F, Chatterjee VK1998 A role for helix 3 of the TR␤

ligand-binding domain in coactivator recruitment identified by characteriza-tion of a third cluster of mutacharacteriza-tions in resistance to thyroid hormone. EMBO J 17:4760 – 4770

11. Clifton-Bligh RJ, de Zegher F, Wagner RL, Collingwood TN, Francois I, Van

Helvoirt M, Fletterick RJ, Chatterjee VK1998 A novel TR␤ mutation (R383H)

in resistance to thyroid hormone syndrome predominantly impairs corepres-sor release and negative transcriptional regulation. Mol Endocrinol 12:609 – 621 12. Makowski A, Brzostek S, Cohen RN, Hollenberg AN 2003 Determination of nuclear receptor corepressor interactions with the thyroid hormone receptor. Mol Endocrinol 17:273–286

13. Cohen RN, Putney A, Wondisford FE, Hollenberg AN 2000 The nuclear corepressors recognize distinct nuclear receptor complexes. Mol Endocrinol 14:900 –914

14. Parrilla R, Mixson AJ, McPherson JA, McClaskey JH, Weintraub BD 1991 Characterization of seven novel mutations of the c-erbA␤ gene in unrelated kindreds with generalized thyroid hormone resistance. Evidence for two “hot spot” regions of the ligand binding domain. J Clin Invest 88:2123–2130 15. Pohlenz J, Weiss RE, Macchia PE, Pannain S, Lau IT, Ho H, Refetoff S 1999

Five new families with resistance to thyroid hormone not caused by mutations in the thyroid hormone receptor␤ gene. J Clin Endocrinol Metab 84:3919–3928 16. Adams M, Matthews C, Collingwood TN, Tone Y, Beck-Peccoz P, Chatterjee

KK1994 Genetic analysis of 29 kindreds with generalized and pituitary re-sistance to thyroid hormone. Identification of thirteen novel mutations in the thyroid hormone receptor␤ gene. J Clin Invest 94:506–515

17. Murata Y, Ceccarelli P, Refetoff S, Horwitz AL, Matsui N 1987 Thyroid hormone inhibits fibronectin synthesis by cultured human skin fibroblasts. J Clin Endocrinol Metab 64:334 –339

18. Moeller LC, Dumitrescu AM, Walker RL, Meltzer PS, Refetoff S 2005 Thy-roid hormone responsive genes in cultured human fibroblasts. J Clin Endo-crinol Metab 90:936 –943

19. Cohen RN, Wondisford FE, Hollenberg AN 1998 Two separate NCoR (nu-clear receptor corepressor) interaction domains mediate corepressor action on thyroid hormone response elements. Mol Endocrinol 12:1567–1581 20. Hollenberg AN, Monden T, Flynn TR, Boers ME, Cohen O, Wondisford FE

1995 The human thyrotropin-releasing hormone gene is regulated by thyroid hormone through two distinct classes of negative thyroid hormone response elements. Mol Endocrinol 9:540 –550

21. Collingwood TN, Adams M, Tone Y, Chatterjee VK 1994 Spectrum of tran-scriptional, dimerization, and dominant negative properties of twenty differ-ent mutant thyroid hormone␤-receptors in thyroid hormone resistance syn-drome. Mol Endocrinol 8:1262–1277

22. Lin BC, Hong SH, Krig S, Yoh SM, Privalsky ML 1997 A conformational switch in nuclear hormone receptors is involved in coupling hormone binding to corepressor release. Mol Cell Biol 17:6131– 6138

23. Liu Y, Takeshita A, Misiti S, Chin WW, Yen PM 1998 Lack of coactivator interaction can be a mechanism for dominant negative activity by mutant thyroid hormone receptors. Endocrinology 139:4197– 4204

24. Matsushita A, Misawa H, Andoh S, Natsume H, Nishiyama K, Sasaki S,

Nakamura H2000 Very strong correlation between dominant negative activ-ities of mutant thyroid hormone receptors and their binding avidity for core-pressor SMRT. J Endocrinol 167:493–503

25. Flynn TR, Hollenberg AN, Cohen O, Menke JB, Usala SJ, Tollin S, Hegarty

MK, Wondisford FE1994 A novel C-terminal domain in the thyroid hormone receptor selectively mediates thyroid hormone inhibition. J Biol Chem 269: 32713–32716

26. Goodson ML, Jonas BA, Privalsky ML 2005 Alternative mRNA splicing of SMRT creates functional diversity by generating corepressor isoforms with different affinities for different nuclear receptors. J Biol Chem 280:7493–7503 JCEM is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the