Nuclear receptor DAX-1 recruits nuclear receptor corepressor N-CoR to steroidogenic factor 1

Abstract

The orphan nuclear receptor steroidogenic factor 1 (SF-1) is a critical developmental regulator in the urogenital ridge, because mice targeted for disruption of the SF-1 gene lack adrenal glands and gonads. SF-1 was recently shown to interact with DAX-1, another orphan receptor whose tissue distribution overlaps that of SF-1. Naturally occurring loss-of-function mutations of the DAX-1 gene cause the human disorder X-linked adrenal hypoplasia congenita (AHC), which resembles the phenotype of SF-1-deficient mice. Paradoxically, however, DAX-1 represses the transcriptional activity of SF-1, and AHC mutants of DAX-1 lose repression function. To further investigate these findings, we characterized the interaction between SF-1 and DAX-1 and found that their interaction indeed occurs through a repressive domain within the carboxy terminus of SF-1. Furthermore, we demonstrate that DAX-1 recruits the nuclear receptor corepressor N-CoR to SF-1, whereas naturally occurring AHC mutations of DAX-1 permit the SF-1-DAX-1 interaction, but markedly diminish corepressor recruitment. Finally, the interaction between DAX-1 and N-CoR shares similarities with that of the nuclear receptor RevErb and N-CoR, because the related corepressor SMRT was not efficiently recruited by DAX-1. Therefore, DAX-1 can serve as an adapter molecule that recruits nuclear receptor corepressors to DNA-bound nuclear receptors like SF-1, thereby extending the range of corepressor action.

FIG. 1

Repression of SF-1 transactivation by DAX-1 requires the most carboxy-terminal repressive domain within SF-1. (A) JEG-3 cells were cotransfected with full-length SF-1 (50 ng) and increasing amounts of the wild type, mutant R267P, or mutant del 369 DAX-1 (0, 5, 15, or 100 ng), along with 250 ng of the SF-1/Luc reporter, which harbors two SF-1 response elements upstream of the prolactin minimal promoter TATA box. (B) JEG-3 cells were cotransfected with empty expression vector (control), full-length SF-1 (amino acids 1 to 462), SF-1 (amino acids 1 to 451), or SF-1 (amino acids 1 to 277) (30 ng) in the presence or absence of exogenous wild-type DAX-1 (10 ng) and 250 ng of SF-1/Luc reporter. Luciferase activities were measured and standardized as described in Materials and Methods.

FIG. 2

Interaction between SF-1 and DAX-1 requires two domains within SF-1. CV-1 cells were cotransfected with fusions of GAL4 DBD and SF-1 carboxy terminus (20 ng), fusions of VP16 and DAX-1 (wild type [amino acids 1 to 472]), DAX-1 (R267P), or DAX-1 (del 369) (50 or 200 ng), and 250 ng of GAL4 reporter. (A) Fusions of GAL4 DBD with SF-1 amino acids 120 to 462, 120 to 451, 120 to 447, 120 to 442, 120 to 437, or 120 to 380 were tested with a two-hybrid assay against all three forms of VP16–DAX-1. Only activity from VP16–DAX-1 (wild type) is shown for GAL4–SF-1 (amino acids 120 to 462). (B) Fusions of GAL4 DBD with SF-1 amino acids 120 to 462, 220 to 462, 226 to 462, 230 to 462 or 245 to 462 were tested in a two-hybrid assay against VP16–DAX-1 (wild type). Luciferase activities were measured and standardized as described in Materials and Methods. Fold activities were calculated by determining the degree of enhancement by VP16–DAX-1 of the activity of each GAL4–SF-1 construct compared to the degree of enhancement observed with GAL4 DBD (see Materials and Methods). (C) Schematic diagram of SF-1, demonstrating the DBD, T/A box, PID, R domain, and AF-2-AH. See text for details.

FIG. 3

DAX-1 acts as a repressor of transcription when tethered to DNA. Expressors driving GAL4 DBD alone or fusions of GAL4 DBD and DAX-1 (amino acids 256 to 472, 256 to 472 [R267P], and 256 to 369) (20 ng) were transfected with 250 ng of GAL4 reporter as described in Materials and Methods. Fold repression relative to the activity that observed with GAL4 DBD is given in parentheses. Luciferase activities were determined and standardized as described above.

FIG. 4

DAX-1 LBD interacts with N-CoR. (A) CV-1 cells were cotransfected with fusions of GAL4 DBD and RARα (amino acids 162 to 462) or DAX-1 LBD (amino acids 256 to 472, 256 to 472 [R267P], or 256 to 369 [del 369]) (20 ng), fusions of VP16 and N-CoR (1550 to 2453) or SMRT (565 to 1495) (20 or 800 ng), and 200 ng of GAL4 reporter. Luciferase activities were measured and standardized as described in Materials and Methods. Fold activities were calculated by determining the degree of enhancement by VP16–N-CoR or VP16-SMRT of the activity of each GAL4–DAX-1 (or GAL4-RARα) construct, compared to the degree of enhancement observed with GAL4 DBD (see Materials and Methods). (B) In vitro interaction between N-CoR and nuclear receptors. The experiment was performed as described in Materials and Methods. Precipitation of N-CoR (amino acids 1510 to 2453) by RARα (as well as 1/10 loaded input N-CoR) was exposed to film for 15 min. Precipitation of N-CoR by the DAX-1 wild type (amino acids 256 to 472) or mutant (amino acids 256 to 369) was exposed to film for 2 h. Where indicated, 10⁻⁷ M atRA was added to transfection media or to the protein incubation. M, molecular mass.

FIG. 5

DAX-1 potentiates two-hybrid interaction between SF-1 and N-CoR. CV-1 cells were cotransfected with GAL4 DBD fusions of the SF-1 carboxy terminus (amino acids 120 to 462, 120 to 447, 120 to 437, or 230 to 462) (20 ng), VP16–N-CoR (20 ng [Fig. 5]); increasing amounts of wild-type, mutant R267P, or del 369 DAX-1 (0, 5, 10, 15, or 100 ng); and GAL4 reporter (250 ng). Luciferase activities were measured and standardized as described in Materials and Methods. Fold activities were calculated by determining the degree of enhancement by VP16–N-CoR of the activity of each GAL4–SF-1 construct in the presence or absence of DAX-1 constructs compared to the degree of enhancement observed with GAL4 DBD (see Materials and Methods).

FIG. 6

Conservation of N-CoR-interactive domains within DAX-1 and RevErb. (A) Two of the three previously delimited domains within RevErb required for interaction with N-CoR (66, 67) share similarities with regions of DAX-1 which are mutated in AHC and which are also required for interaction with N-CoR. Arrows demarcate residues analyzed in this study: R267 of DAX-1 is mutated in some AHC kindreds (R267P), and F449 is one of three shared residues in the carboxy-terminal-interactive domain of RevErb (B, C, and D). (B) Mutation F449D prevents the two-hybrid interaction between DAX-1 and N-CoR. A two-hybrid experiment was performed and results were analyzed as described in the legend to Fig. 5A (500 ng of VP16–N-CoR used in this experiment). (C) Mutation F449D in the context of full-length DAX-1 abrogates the potentiation of the two-hybrid SF-1–N-CoR interaction. Ten nanograms of DAX-1 (wild type [WT]) or F449D was transfected with 20 ng of VP16–N-CoR, 20 ng of GAL4–SF-1 (120 to 462), and 200 ng of GAL4 reporter. (D) Mutant F449D of DAX-1 is not a potent repressor of SF-1 transactivation. Fifty nanograms of SF-1 expressor was cotransfected with 20 ng of DAX-1 (wild-type) or DAX-1 (F449D) mutant, along with 250 ng of SF-1/Luc reporter. The experiment was analyzed as described in the legend to Fig. 2A. (E) Western blotting (performed as described in Materials and Methods with anti-GAL4 DBD antibody) demonstrates expression levels of GAL4 DBD–DAX-1 mutant fusions. Lanes: 1, GAL4–DAX-1 (amino acids 256 to 472); 2, GAL4–DAX-1 (amino acids 256 to 472 [R267P]); 3, GAL4–DAX-1 (amino acids 256 to 369); 4, GAL4–DAX-1 (amino acids 256 to 472 [F449D]). MW, molecular mass.

FIG. 7

Schematic model of mechanism of DAX-1 repression of SF-1. SF-1, bound to its response element on DNA, recruits DAX-1, requiring the PID and the R domain within the carboxy terminus of SF-1. To interact with SF-1, DAX-1 employs its structurally degenerate DBD (30). DAX-1 in turn recruits N-CoR to DNA, requiring at least two conserved regions within DAX-1 LBD, either of which is mutated or eliminated in different AHC kindreds. Other molecules may stabilize or modify the interaction among the three proteins, depending on the cell or promoter context.