Determinants of the heightened activity of glucocorticoid receptor translational isoforms

Abstract

Translational isoforms of the glucocorticoid receptor α (GR-A, -B, -C1, -C2, -C3, -D1, -D2, and -D3) have distinct tissue distribution patterns and unique gene targets. The GR-C3 isoform-expressing cells are more sensitive to glucocorticoid killing than cells expressing other GRα isoforms and the GR-D isoform-expressing cells are resistant to glucocorticoid killing. Whereas a lack of activation function 1 (AF1) may underlie the reduced activity of the GR-D isoforms, it is not clear how the GR-C3 isoform has heightened activity. Mutation analyses and N-terminal tagging demonstrated that steric hindrance is probably the mechanism for the GR-A, -B, -C1, and -C2 isoforms to have lower activity than the GR-C3 isoform. In addition, truncation scanning analyses revealed that residues 98 to 115 are critical in the hyperactivity of the human GR-C3 isoform. Chimera constructs linking this critical fragment with the GAL4 DNA-binding domain showed that GR residues 98 to 115 do not contain any independent transactivation activity. Mutations at residues Asp101 or Gln106 and Gln107 all reduced the activity of the GR-C3 isoform. In addition, functional studies indicated that Asp101 is crucial for the GR-C3 isoform to recruit coregulators and to mediate glucocorticoid-induced apoptosis. Thus, charged and polar residues are essential components of an N-terminal motif that enhances the activity of AF1 and the GR-C3 isoform. These studies, together with the observations that GR isoforms have cell-specific expression patterns, provide a molecular basis for the tissue-specific functions of GR translational isoforms.

Figure 1.

The human GR-C3 (hGR-C3) isoform has heightened transactivation activity. A, Diagram of the human GRα isoforms. Numbers on the left indicate the locations of the start codons responsible for the expression of the corresponding GR isoforms. For example, GRα-A contains amino acids 1 through 777. τ-1c, transactivation function 1 core (187–244). B, Transactivation activity of GR isoforms. Increasing amounts of GR isoform-expressing pTRE constructs (5–50 ng) were cotransfected with 200 ng of pMMTV-Luc reporter genes and 100 ng of pGL3-hRL into COS-1 cells on 12-well plates in glucocorticoid-free media (100,000 cells/well in triplicate). At 24 hours after transfection, cells were treated with vehicle or DEX (10 nM). After another 24 hours, cells were harvested and lysed. Renilla (transfection control) and firefly luciferase activities were determined on a BioTek luminometer. Relative light units (RLU) indicate the activity of firefly luciferase normalized by Renilla luciferase activity. Shown is a representative of at least 4 experiments. Expression levels of GR isoforms in each well are similar as indicated by a representative Western blot analysis. *, Significantly greater than GRα; +, significantly less than GRα; P < .05, ANOVA followed by Tukey post-hoc tests.

Figure 2.

Steric hindrance rather than an inhibitory motif in the N terminus underlies the differential transactivation activities of GR isoforms. A, Comparison of transactivation activity of the GR-C3 isoform and the GR-C1 isoform harboring various mutations. The GR-C1 isoform was used as a representative of GR-A, -B, and -C2 isoforms, and they all have less than half of the activity of the GR-C3 isoform. Individual or double mutations of residues upstream of Met98 were as indicated. Twenty nanograms of plasmid DNA was transfected into COS-1 cells. B, Addition of unrelated domains upstream of the GR-C3 start codon suppresses transactivation activity. FLAG, Asp-Tyr-Lys-Asp-His-Asp; HA, Tyr-Pro-Tyr-Asp-Val-Pro-Asp-Tyr; YFP, yellow fluorescent protein. Representative Western blots of the protein levels are shown. Luciferase activities (percentage, normalized to protein levels) of the constructs represented by the solid bars are significantly less than those of the GR-C3 isoform; P < .05, ANOVA followed by Tukey post-hoc tests. RLU, relative light unit.

Figure 3.

The heightened activity of the GR-C3 isoform is dependent on residues 98 to 115. Serial truncation of the N-terminus was introduced in the GR-C3 isoform. The fragment deleted was immediately upstream of the start codons (M) indicated. A representative Western blot of the protein levels is shown. Luciferase activities (percentage, normalized to protein levels) of the constructs represented by the solid bars are significantly less than those of the GR-C3 isoform; P < .05, ANOVA followed by Tukey post-hoc tests. RLU, relative light unit.

Figure 4.

Charged and polar residues are critical for the heightened activity of the GR-C3 isoform. Asp101, Gln106, and Gln107 but not other mutations inhibited the activity of the GR-C3 isoform. Consistent with the observation that fragment 98 to 115 is critical for the enhanced activity of the GR-C3 isoform (Figure 3), mutations of residues downstream of residue 115 did not reduce the activity of the GR-C3 isoform. A representative Western blot of the protein levels is shown and luciferase activities (percentage) were normalized to protein levels. *, Significantly less than the activity of the GR-C3 isoform; P < .05, ANOVA followed by Tukey post-hoc tests. RLU, relative light unit.

Figure 5.

Motif 98 to 186 enhances the activity of the τ-1 core. COS-1 cells were transfected with pBind and pG5luc. pBind constructs express GR or ER and GAL4 DBD chimeras as well as Renilla luciferase. pG5luc expresses firefly luciferase driven by 5 copies of GAL4 response elements. Transfection procedure and luciferase assays were performed as described in Materials and Methods. Fragment 98 to 115 (construct 2) had no independent transactivation activity as demonstrated in a GR GAL4 DBD chimera construct. However, the addition of residues 98 to 186 (construct 7) enhanced the activity of τ-1 core (construct 4). The fragment 98 to 186 harboring the mutation Asp101Lys (construct 8) did not enhance the activity of the τ-1 core, further demonstrating the importance of this charged residue. Addition of residues upstream of the GR-C3 isoform (constructs 5 and 6) did not increase the activity of the τ-1 core, confirming the role of steric hindrance from the region upstream of Met98. The addition of residues 98 to 115 to ER AF1 (construct 10) did not enhance transactivation. A representative Western blot of the protein levels is shown. Luciferase activities (percentage) were normalized to protein levels. *, Significantly greater than construct 4 (187 to 244-GAL4); ANOVA, P < .05, Tukey post-hoc tests. CON, control; RLU, relative light unit.

Figure 6.

Residue Asp101 is critical for the function of the GR-C3 isoform. A, A representative Western blot shows that comparable levels of receptor proteins were expressed in 2 clones of U-2 OS cells expressing the GR-C3 isoform and in 2 additional clones expressing the GR-C3 isoform harboring the Asp101Lys mutation. B, Propidium iodide (PI) (or 4′,6-diamidino-2-phenylindole [DAPI]) staining of dead cells and annexin V staining of apoptotic cells indicated that the Asp101Lys mutations impaired the glucocorticoid-induced cell death. Histograms and dot plots are from flow cytometric analyses. Bar graphs are averages of at least four experiments. C, Real-time RT PCR experiments show that granzyme A (GZMA) expression was impaired by Asp101Lys mutation (n = 2). D, Immunoprecipitation (IP) with anti-FLAG and immunoblotting (IB) with anti-GR showed that Asp101Lys did not impair GR interaction with CBP (n = 3). E, ChIP assay demonstrated that GR binding to DNA was not impaired by the mutation. In contrast, CBP and p300, but not BRG1, recruitment in the context of DNA was significantly inhibited by Asp101Lys mutation. Shown are representatives of 3 experiments. Bar graphs are averages of real-time PCR quantification. *, Significantly less than the GR-C3 isoform (DEX); P < .05, ANOVA followed by Tukey post-hoc tests (B) or Student's t tests (C, E). CON, control.