Repression of cardiac phospholamban gene expression is mediated by thyroid hormone receptor-{alpha}1 and involves targeted covalent histone modifications

Abstract

Phospholamban (PLB) is a critical regulator of Ca(2+) cycling in heart muscle cells, and its gene expression is markedly down-regulated by T(3). Nonetheless, little is known about the molecular mechanisms of T(3)-dependent gene silencing in cardiac muscle, and it remains unclear whether thyroid hormone receptors (TRs) directly bind at the PLB gene in vivo and facilitate transcriptional repression. To investigate the regulatory role of TRs in PLB transcription, we used a physiological murine heart muscle cell line (HL-1) that retains cardiac electrophysiological properties, expresses both TRalpha1 and TRbeta1 subtypes, and exhibits T(3)-dependent silencing of PLB expression. By performing RNA interference assays with HL-1 cells, we found that TRalpha1, but not TRbeta1, is essential for T(3)-dependent PLB gene repression. Interestingly, a PLB reporter gene containing only the core promoter sequences -156 to +64 displayed robust T(3)-dependent silencing in HL-1 cells, thus suggesting that transcriptional repression is facilitated by TRalpha1 via the PLB core promoter, a regulatory region highly conserved in mammals. Consistent with this notion, chromatin immunoprecipitation and in vitro binding assays show that TRalpha1 directly binds at the PLB core promoter region. Furthermore, addition of T(3) triggered alterations in covalent histone modifications at the PLB promoter that are associated with gene silencing, namely a pronounced decrease in both histone H3 acetylation and histone H3 lysine 4 methylation. Taken together, our data reveal that T(3)-dependent repression of PLB in cardiac myocytes is directly facilitated by TRalpha1 and involves the hormone-dependent recruitment of histone-modifying enzymes associated with transcriptional silencing.

Figure 1

HL-1 cells express TRs and exhibit T₃-dependent silencing of PLB mRNA and protein expression. A, Total RNA was extracted from HL-1 cells and then assayed by semiquantitative RT-PCR using primers specific for TRα1, TRβ1, or β-actin. As a negative control, reverse transcriptase (RT) was omitted as indicated. B, T₃ inhibits PLB mRNA expression. HL-1 cells cultured with or without T₃ (1 μm) for 24 h were analyzed by RT-PCR using primers specific for PLB or β-actin. The PCR signals were quantitated using by gel documentation, normalized against β-actin expression, and then displayed graphically. Error bars represent the ±sd calculated from three separate experiments. C, Whole-cell lysate was prepared from HL-1 cells cultured with or without T₃ (1 μm) for 72 h and then probed by immunoblot with antibodies against PLB, SERCA2a, or α-tubulin.

Figure 2

T₃ inhibits PLB gene expression in a dose- and time-dependent fashion. HL-1 cells were cultured with different concentrations of T₃ for 24 h and then analyzed by RT-PCR semiquantitatively (A) or in real time (B). The semiquantitative RT-PCR results (A) were quantitated and displayed graphically as described in the Fig. 1 legend. C, HL-1 cells cultured with or without T₃ (1 μm) for different lengths of time were analyzed by real-time RT-PCR. For both real-time RT-PCR experiments (A and B), values were normalized to β-actin expression. Error bars represent the ±sd calculated from three separate experiments. T₃-untreated controls were set at 1.

Figure 3

Acceleration of decay kinetics in spontaneous Ca²⁺ waves indicates increased SERCA2a function in T₃-treated HL-1 cells. Spontaneous Ca²⁺ waves in HL-1 cells that were either treated with T₃ (1 μm) or vehicle control for 72 h were analyzed for their kinetic properties. A, Representative images of the fluo-4 fluorescent signal from control Hl-1 cells (top panels) with a trace (bottom panels, black line) of the mean fluorescent signal from the white boxed area. The trace data presented has been smoothed by a fast Fourier transform filter (4 point). A gray dashed line shows the function used to fit the fluorescence signal trace for analysis of kinetic properties. B, Same as in A for T₃-treated HL-1 cells. C, T₃ treatment (open bar) resulted in a decrease in the decay time constant (DTC) of Ca²⁺ waves when compared with a vehicle control (black bars). This suggests that Ca²⁺ clearance from the cytosol was accelerated due to T₃ treatment, which is consistent with an increase of SERCA2a activity, resulting in increased uptake of Ca²⁺ into the sarcoplasmic reticulum. D, T₃-treated cells also displayed a decreased duration of individual Ca²⁺ waves, which is also consistent with accelerated Ca²⁺ clearance for the cytosol due to elevated SERCA2a activity. E, The amplitude of individual Ca²⁺ waves in T₃-treated cells was not significantly altered from those of vehicle-treated cells. Therefore, it is not likely that the alterations to the kinetics of the Ca²⁺ waves is due to alterations in the Ca²⁺ release machinery of, Ca²⁺ storage within the sarcoplasmic reticulum. Data are presented as mean ± sem. Statistical significance is determined by t test (**, P < 0.01).

Figure 4

Loss of TRα1 expression abolishes T₃-dependent repression of PLB gene expression. A, HL-1 cells were electroporated with siRNA specific for either TRα1 or TRβ1 or with a nonspecific control siRNA. Total RNA was then extracted and assayed by semiquantitative RT-PCR using primers specific for TRα1, TRβ1, or β-actin. B and C, HL-1 cells were electroporated with siRNA specific for either TRα1 or TRβ1 or with a nonspecific control siRNA. The cells were then cultured with or without T₃ (1 μm) for an additional 24 h. Total RNA was then extracted and analyzed by RT-PCR for PLB mRNA expression semiquantitatively (B) or in real time (C). Error bars represent the ±sd calculated from three separate experiments. In C, T₃-untreated controls were set at 1. Quantitation and graphical representation were as outlined in the Fig. 2 legend.

Figure 5

T₃-dependent silencing of PLB gene expression is localized to the conserved core promoter region. A, Alignment of mammalian PLB core promoter sequences. Nucleotide sequences from mouse (36), rat (37), and human (38) are shown. Numbers indicate relative nucleotide positions upstream of the transcription start site (+1). B, rat PLB core promoter sequences showing potential TR-binding half-sites (underlined arrows), potential TATA boxes (italics), GATA element (bold), CCAAT element (boxed), and an E-box (bold and italics). Numbers indicate relative nucleotide positions up- and downstream of the transcription start site (+1). C, HL-1 cells were transfected with either the rat pGL3-PLB (−156 to +64) promoter luciferase vector or an empty pGL3 vector along with pSV-β-gal as an internal control. The cells were then cultured with or without T₃ (1 μm) for additional 24 h and then harvested for determination of luciferase activity via a luminometer. Luciferase activity was normalized for both protein concentration as well as β-galactosidase activity. Error bars represent the ±sd calculated from three separate experiments. *, Statistical significant difference, P < 0.01.

Figure 6

TRα1 directly binds at the PLB core promoter in HL-1 cells, and addition of T₃ triggers changes in covalent histone modifications associated with transcriptional repression. A, Schematic representation of the mouse PLB promoter including the relative positions of the three PCR primer sets. B and C, TRα1 binds at the PLB core promoter region and its occupancy increases in the presence of T₃. Formaldehyde cross-linked chromatin was prepared from HL-1 cells cultured in normal serum (B) or additionally cultured in the presence of T₃ (1 μm) for 24 h (C). ChIP was then carried out using an antibody specific for TRα1 or a nonspecific control (Con.) rabbit IgG and then analyzed by semiquantitative PCR using all three PCR primer sets (B) or primer set B alone (C). D and E, T₃-dependent histone H3 deacetylation and demethylation at the PLB core promoter. ChIP was carried out on formaldehyde cross-linked chromatin prepared from HL-1 cells cultured with or without T₃ using antibodies specific for H3-Ac or H3K4–2me and then analyzed by PCR using primer set B. The ethidium bromide-stained PCR bands in C–E were quantitated by gel documentation and are presented as the fold enrichment over the PCR signal generated by the control IgG. Error bars represent the ±sd calculated from three separate experiments. Statistical significant difference: *, P < 0.001; **, P < 0.005; ***, P < 0.004.

Figure 7

TRα1 and RXRα directly bind to the PLB core promoter region in vitro. A, Schematic representation of the PLB promoter and intragenic region including the relative positions of the two biotinylated templates. B, TRα1 binds at the PLB core promoter, and its occupancy increases in the presence of RXRα and T₃. The biotinylated PLB core promoter and intragenic templates were conjugated to streptavidin beads and then incubated with recombinant baculovirus-expressed FLAG-tagged TRα1 or RXRα in the presence or absence of T₃ (10⁻⁷ m). The DNA-protein complexes were isolated, washed, fractionated by SDS-PAGE, and then probed by immunoblot using anti-FLAG antibodies.