Nuclear receptor corepressor recruitment by unliganded thyroid hormone receptor in gene repression during Xenopus laevis development

Abstract

Thyroid hormone receptors (TR) act as activators of transcription in the presence of the thyroid hormone (T(3)) and as repressors in its absence. While many in vitro approaches have been used to study the molecular mechanisms of TR action, their physiological relevance has not been addressed. Here we investigate how TR regulates gene expression during vertebrate postembryonic development by using T(3)-dependent amphibian metamorphosis as a model. Earlier studies suggest that TR acts as a repressor during premetamorphosis when T(3) is absent. We hypothesize that corepressor complexes containing the nuclear receptor corepressor (N-CoR) are key factors in this TR-dependent gene repression, which is important for premetamorphic tadpole growth. To test this hypothesis, we isolated Xenopus laevis N-CoR (xN-CoR) and showed that it was present in pre- and metamorphic tadpoles. Using a chromatin immunoprecipitation assay, we demonstrated that xN-CoR was recruited to the promoters of T(3) response genes during premetamorphosis and released upon T(3) treatment, accompanied by a local increase in histone acetylation. Furthermore, overexpression of a dominant-negative N-CoR in tadpole tail muscle led to increased transcription from a T(3)-dependent promoter. Our data indicate that N-CoR is recruited by unliganded TR to repress target gene expression during premetamorphic animal growth, an important process that prepares the tadpole for metamorphosis.

FIG. 1.

Sequence and domain structure of xN-CoR. (A) Amino acid sequence of xN-CoR as predicted from the full-length cDNA sequence. The three RDs, located in the N-terminal part of the protein, are underlined. The receptor IDs, represented in bold, are located in the C-terminal part. The CoRNRs within the ID are underlined. (B) Schematic diagram of xN-CoR showing the various conserved domains: RDs I to III, SNOR (for present in SMRTER, SMRT, and N-CoR), SANT 1 and 2 (for present in SWI3, ADA2, N-CoR, and TFIIIB-B′), ITS (motif of isoleucine-threonine-serine), IDs (N and C terminal), CoRNR (1 to 3), and LSD (motif of leucine-serine-aspartic acid).

FIG. 2.

Various functional motifs are conserved among N-CoR, SMRT, and SMRTER. Alignments of xN-CoR, human N-CoR (hN-CoR; GenBank accession no. AF044209), mouse N-CoR (mN-CoR; GenBank accession no. U35312), human SMRT (hSMRT; GenBank accession no. AF125672), and Drosophila SMRTER (GenBank accession no. AF175223) are given for the SNOR, SANT, ITS, LSD, and CoRNR motifs. Dashes indicate sequences identical to that of xN-CoR, while blank spaces are gaps introduced for better alignments. Amino acid positions of each motif in the respective protein are also indicated.

FIG. 3.

High levels of xN-CoR gene transcripts are present during intestinal remodeling (A), tail resorption (B), and hindlimb morphogenesis (C). Each lane had 10 μg of total RNA, except for stage NF64 tail and stage NF56 hindlimb, which contained 5 μg. The same blot was probed with either an 800-bp xN-CoR cDNA fragment corresponding to aa 1992 to 2265 or the loading control gene rpl8, whose expression remains constant during development (44, 45, 58). The signals from the blots were quantified with a PhosphorImager and plotted on the right after normalization against the control rpl8 signals. The data represent one of two independent experiments with similar conclusions.

FIG. 4.

T₃ treatment of premetamorphic tadpoles increases xN-CoR mRNA levels in the intestine (A) and tail (B) but not in the hindlimb (C). NF55 tadpoles were treated with 5 nM T₃ for the indicated number of days. Total RNA was isolated and subjected to Northern blot analysis (10 μg/lane for tail and intestinal RNA) or RT-PCR analysis (for hindlimb RNA, due to the limited amount of RNA isolated from the small organ) for xN-CoR RNA levels. The signals were quantified with a PhosphorImager and plotted on the right after normalization against the control rpl8 signals. All the experiments were done at least twice with similar results.

FIG. 5.

The expression of xSMRT is similar to that of xN-CoR during development. Total RNA was isolated from intestine, tail, and hindlimb from tadpoles at the indicated stages. RT-PCR analysis was carried out to determine the mRNA levels of xSMRT and the internal control gene rpl8. Note that similar to xN-CoR (Fig. 3), xSMRT was upregulated during intestinal remodeling (stage NF60) and tail resorption (stage NF64). The results are from one of two independent experiments with similar results.

FIG. 6.

xN-CoR interacts in vivo with TR only in the absence of T₃. Xenopus mature oocytes were microinjected (+) with in vitro-transcribed mRNA encoding Xenopus TRβ with or without Xenopus RXRα. The oocytes were incubated with or without T₃ (50 nM) as indicated. The oocyte extracts were precipitated with antibodies specific for either TRβ (A) or xN-CoR (B). The precipitation products were assayed by Western blotting for xN-CoR (A) and TRβ (B). A part of the oocyte extracts was also analyzed directly by Western blotting for the expression of xN-CoR (A) or TRβ (B) proteins (input control). Uninjected oocytes were used as controls because they lack any detectable level of TRβ (B) or RXRα (data not shown) protein.

FIG. 7.

ChIP assays show that T₃ treatment leads to the release of xN-CoR from and increase of histone acetylation at T₃ response gene promoters. (A) Schematic diagrams showing the regions that control T₃-dependent transcription of the two T₃-responsive genes TRβ and TH/bZIP. The positions of the transcription start site, T₃REs, and primers used for PCR are indicated. (B) xN-CoR is associated with T₃ response genes and released upon T₃ treatment. Stage NF55 tadpole tail and intestinal nuclei were isolated after treatment with 10 nM T₃ for 48 h. Chromatin was cross-linked with formaldehyde, fragmented, immunoprecipitated with antibody against xN-CoR, and analyzed by PCR for the presence of a T₃RE-containing fragment of TRβA and TH/bZIP promoters or for the presence of IFABP promoter, which is not regulated by T₃ and thus served as a negative control. (C) In the tail, TR binds to the T₃REs constitutively while local histone acetylation is upregulated by T₃ treatment. The experiments were done as for panel B except for the use of the antibody against TR or acetylated histone H4 for the ChIP assay. Aliquots of the chromatin before immunoprecipitation were used for PCR as the control for DNA quantity (input). The data represent one out of several independent experiments with identical results. Ct, control; AB, antibody.

FIG. 8.

Dominant-negative N-CoR (CoRNR) abolishes the repression of T₃RE-containing promoter by unliganded TR. (A) Overexpression of CoRNR, but not the control peptide (Ct), upregulates the expression of T₃RE-tk-fLuciferase. An aliquot (0.5 μg) of the CoRNR vector, expressing the dominant-negative N-CoR, or the control vector, expressing a nonspecific peptide, was coinjected with 0.5 μg of the T₃RE-tk-fLuciferase reporter and 0.1 μg of the control reporter phRL-SV40 into Xenopus dorsal tail muscle. Two days later, the luciferase activities were assayed from tail muscle homogenates. The ratio of the activity of the firefly luciferase (fLuciferase) to that of the control Renilla luciferase (rLuciferease) was plotted together with standard errors. (B) CoRNR had no effect on the firefly luciferase expression under the control of the CMV promoter, which is not regulated by T₃. The experiments were done as for panel A except that T₃RE-tk-fLuciferase was replaced by CMV-fLuciferase. (C) The dominant-negative N-CoR does not affect the expression of Renilla luciferase reporter under the control of a constitutive promoter. The experiments were done as for panel A except that T₃RE-tk-fLuciferase was omitted. (D) Overexpression of TRα represses T₃RE-tk-fLuciferase reporter while high levels of CoRNR overexpression alleviate this repression. The T₃RE-tk-fLuciferase reporter (0.5 μg) and the control reporter plasmid phRL-SV40 (0.1 μg) were coinjected with or without the indicated amounts of pCMV-TRα, which expressed TRα, and the N-CoRNR vector or the Ct vector into Xenopus dorsal tail muscle. The activities of the luciferases were measured and plotted as for panel A. The data are given as means ± standard errors of the means. Each points represents at least seven animals. **, significant changes (P < 0.01) between the two sets of data. All the experiments were done at least twice with similar results.