Structural and functional characterization of the interdomain interaction in the mineralocorticoid receptor

Abstract

The mineralocorticoid receptor (MR) plays a central role in electrolyte homeostasis and in cardiovascular disease. We have previously reported a ligand-dependent N/C-interaction in the MR. In the present study we sought to fully characterize the MR N/C-interaction. By using a range of natural and synthetic MR ligands in a mammalian two-hybrid assay we demonstrate that in contrast to aldosterone, which strongly induces the interaction, the physiological ligands deoxycorticosterone and cortisol weakly promote the interaction but predominantly inhibit the aldosterone-mediated N/C-interaction. Similarly, progesterone and dexamethasone antagonize the interaction. In contrast, the synthetic agonist 9alpha-fludrocortisol robustly induces the interaction. The ability of the N/C interaction to discriminate between MR agonists suggests a subtle conformational difference in the ligand-binding domain induced by these agonists. We also demonstrate that the N/C interaction is not cell specific, consistent with the evidence from a glutathione-S-transferase pull-down assay, of a direct protein-protein interaction between the N- and C-terminal domains of the MR. Examination of a panel of deletions in the N terminus suggests that several regions may be critical to the N/C-interaction. These studies have identified functional differences between physiological MR ligands, which suggest that the ligand-specific dependence of the N/C-interaction may contribute to the differential activation of the MR that has been reported in vivo.

Figure 1

9α-Fludrocortisol dose response for the MR N/C-interaction. COS-1 cells were transiently transfected with expression vectors and the GAL4-responsive luciferase reporter vector g5-LUC. The cells were treated 14–16 h after transfection with various concentrations of 9α-fludrocortisol as indicated. Luciferase activity was measured after a 24-h incubation and is represented as fold activation relative to luciferase activity in the absence of ligand. A, 9α-Fludrocortisol dose response with wild-type MRC. The open bars correspond to GAL4-MRC + pVP16, and the solid bars correspond to GAL4-MRC + pVP16-MRNT. B, 9α-Fludrocortisol dose response with MRC(E962A). The open bars correspond to GAL4-MRC(E962A) + pVP16, and the solid bars correspond to GAL4-MRC(E962A) + pVP16-MRNT. Each data point represents the mean ± sem derived from three independent experiments; *, P < 0.001; **, P < 0.01 denotes significance when compared with the same concentration of 9α-fludrocortisol in the absence of MRNT.

Figure 2

Dexamethasone dose response for the MR N/C-interaction, alone and in the presence of aldosterone. COS-1 cells were transiently transfected for the M-2-H assay as described in Fig. 1. The cells were treated with various concentrations of dexamethasone alone (A and B) or in the presence of 1 nm aldosterone (C and D) to examine the effect on the MR N/C-interaction with wild-type MRC (A and C) and MRC(E962A) (B and D). Luciferase activity is represented as fold activation relative to luciferase activity in the absence of ligand. Each data point represents the mean ± sem derived from three independent experiments; *, P < 0.001 denotes significance when compared with 1 nm aldosterone in the absence of MRNT (A and B) or when compared with aldosterone alone in the presence of MRNT (C and D).

Figure 3

DOC dose response for the MR N/C-interaction, alone and in the presence of aldosterone. COS-1 cells were transiently transfected for the M-2-H assay as described in Fig. 1. The cells were treated with increasing concentrations of DOC in the absence (A and B) and presence (C and D) of aldosterone (1 nm) to examine the effect on the MR N/C-interaction with wild type MRC (A and C) and MRC(E962A) (B and D). Luciferase activity is represented as fold activation relative to luciferase activity in the absence of ligand. Each data point represents the mean ± sem derived from three independent experiments; *, P < 0.001 denotes significance when compared with 1 nm aldosterone in the absence of MRNT (A and B) or when compared with aldosterone alone in the presence of MRNT (C and D).

Figure 4

Effect of progesterone on the aldosterone-mediated MR N/C-interaction. COS-1 cells were transiently cotransfected with GAL4-MRC as described in Fig. 1. The cells were treated with aldosterone (1 nm) and progesterone (1 and 10 nm) alone (A) or cotreated with aldosterone (1 nm) and progesterone (1 and 10 nm) as indicated. B, Luciferase activity is represented as fold activation relative to luciferase activity in the absence of ligand. Each data point represents the mean ± sem derived from three independent experiments; *, P < 0.001 denotes significance when compared with 1 nm aldosterone in the absence of MRNT (A) or when compared with aldosterone alone in the presence of MRNT (B).

Figure 5

The MR N/C-interaction in LLC-PK1 (A), H9C2 (B), COV-434 (C), and HEK293 (D) cells. The cells were transiently cotransfected with GAL4-MRC with or without VP16-MRNT in the presence or absence of aldosterone (10 nm). Luciferase activity is represented as fold activation relative to luciferase activity in the absence of ligand. Each data point represents the mean ± sem derived from three independent experiments; *, P < 0.001; and ***, P < 0.05 denote significance when compared with aldosterone in the absence of MRNT.

Figure 6

Analysis of the role of sumoylation in the MR N/C-interaction. COS-1 cells were transiently cotransfected with GAL4-MRC(E962A) with or without VP16-MRNT (wild type; 1–597) and VP16-MRNT in which the four sumoylation sites have been mutated. Where K1 represents the lysine mutation at position 89, K234 at 399, 428, and 494 and K1234 incorporates the lysine mutations at all four positions. The cells were treated with vehicle (open bars) or aldosterone (10 nm) (solid bars). Luciferase activity is represented as a percentage of the maximal response. Each data point represents the mean ± sem derived from three independent experiments; *, P < 0.001 denotes significance when compared with vehicle.

Figure 7

Structural determinants of the MR N/C-interaction in the NTD. COS-1 cells were transiently cotransfected with GAL4-MRC(E962A) with or without VP16-MRNT and with a series of deletions of VP16-MRNT as indicated in panels A and B. The cells were treated with vehicle (open bars) or aldosterone (10 nm) (solid bars). Each data point represents the mean ± sem derived from three independent experiments; *, P < 0.001 denotes significance when compared with wild type [VP16-MRNT (1–597) + MRC(E962A)] in the presence of aldosterone; ♦, P < 0.01; and †, P < 0.001 denote significance when compared with vehicle alone. C, Schematic representation of N/C-interaction between fragments of MRNT and MRC(E962A).

Figure 8

Deletion of amino acids 351–542 from the NTD. A, COS-1 cells were transiently cotransfected with GAL4-MRC(E962A) with or without VP16-MRNT and with deletions of VP16-MRNT. The cells were treated with vehicle (open bars) or aldosterone (10 nm) (solid bars). Each data point represents the mean ± sem derived from three independent experiments; *, P < 0.001 denotes significance when compared with wild type [VP16-MRNT (1–597) + MRC(E962A)] in the presence of aldosterone; Δ, P < 0.05; and †, P < 0.001 denote significance when compared with vehicle alone. B, Schematic representation of series of internal deletions in the MR NT that retain N/C-interaction with MRC(E962A).

Figure 9

GST pull-down assay with the N and C termini. The MR NTD expressed as GST fusion protein, which previously had been coupled to Sepharose glutathione beads, was incubated with [³⁵S]methionine-labeled MR-LBD in the presence (lanes 2) or absence (lane 3) of 1 μm aldosterone. As controls, incubation with SRC-1 (570–780) (positive; lanes 4 and 5), GST alone (lanes 6 and 7) and one tenth of the receptor input (lane 1) are shown. kd, Kilodaltons.