| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Steroid Hormones Section, National Institute of Diabetes and Digestive and Kidney Diseases/Laboratory of Molecular and Cellular Biology, National Institutes of Health, Bethesda, Maryland 20892
Address all correspondence and requests for reprints to: Dr. S. Stoney Simons, Jr., Building 8, Room B2A-07, National Institute of Diabetes and Digestive and Kidney Diseases/Laboratory of Molecular and Cellular Biology, National Institutes of Health, Bethesda, Maryland 20892. E-mail: steroids{at}helix.nih.gov.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Yet additional targets of corepressors were suggested by the recent observations that corepressors decrease the total amounts of gene expression by receptor-agonist complexes (7, 15, 16, 17, 18, 19, 20). Also, the dose-response curve of steroid receptor-agonist complexes is influenced by corepressors (12, 13, 16, 17, 18, 19). This ability to change the position of the dose-response curve of a given gene in different cells means that the same concentration of circulating hormone can afford unequal levels of gene expression, which is invaluable for allowing differential control of gene expression during development, differentiation, homeostasis, and endocrine therapy (21, 22, 23).
These multiple actions of corepressors raise several mechanistic questions. p160 Coactivators are thought to augment the amount of receptor-mediated gene activation by binding to a highly conserved hydrophobic cleft in the receptor ligand binding domain (LBD), which is often called the activation function (AF)-2 domain and formed by four of the 12 common 
-helices (numbered 3, 4, 5, and 12) (24). A strong corepressor binding site exists in the LBD of nuclear receptors, which overlaps with but is not identical with the site used by p160 coactivators (25, 26, 27). It has been assumed that a similar binding mode exists for steroid receptors. Importantly, the actions of coactivators and corepressors are reversible both at the level of biological activities (i.e. dose-response curve, partial agonist activity, and fold increase of gene activation) (Ref. 17 and Szapary, D., and S. Simons, unpublished results) and for protein-protein interactions in whole-cell and cell-free environments (12). However, different mechanisms may dictate the changes in fold transactivation with agonists vs. modulation of the partial agonist activity of antisteroids and the dose-response curve of agonists. This is because the ability of coactivators to modify the total amount of transactivation is separable from their capacity to modulate the dose-response curve and partial agonist activity (12, 13, 17, 28, 29, 30, 31, 32). These data suggest that corepressors (and coactivators) interact with receptors at another site in addition to the one in the LBD. This hypothesis is supported by the report that the inhibitory region of PR, which is the first 140 amino acids of PR-A (=165–304 of PR-B), is important for PR binding to the receptor interaction domain (RID) of corepressors (33), which is essential for corepressor binding to steroid/nuclear receptors (12, 34). Similarly, N-terminal sequences of ARs (7, 20) and GRs (11) are reported to interact with corepressors. More recently, we found that neither the N- nor the C-terminal sequences of GR are sufficient for the whole-cell or cell-free binding of corepressors (12).
Another important mechanistic question is whether the above putative additional site(s) for corepressor binding to steroid receptors is as conserved among different receptors as the currently identified LBD site. The fact that the dose-response curves and partial agonist activities for PR and GR induction of the same gene in the same cells are unequally affected by overexpression of the same corepressor (19) suggests that different features of each receptor are important. This conclusion was reinforced when PR/GR chimeras having heterologous N- and C-terminal sequences were found to afford responses that were intermediate between those of the wild-type GR and PR (19).
Finally, the biological responses of steroid receptors to corepressors usually require ligand binding, whereas the physical interactions of corepressors with steroid receptors appear to be ligand independent (7, 10, 11, 12, 15, 20, 35). Given these unequal requirements for bound steroid, the relevance of the cell-free physical interactions of corepressors with steroid receptors to the biological actions of corepressors remains an open question.
The purpose of this study, therefore, was first to see whether a second binding site for corepressors does exist in GRs and PRs. Second, we wanted to determine why ligand is required for whole-cell but not cell-free binding of corepressors to GRs and PRs. Finally, we reasoned that a comparison of the features required for corepressor binding ± steroid to GRs vs. PRs might be helpful in understanding the different biological responses of GRs and PRs, even with the same gene. We now report that a second binding site for corepressors is present in the nonhomologous amino-terminal half of GRs and PRs. For both receptors, most of the ligand-free binding of receptors to corepressors is blocked by sodium molybdate, a known inhibitor of receptor-steroid complex activation (36, 37). However, several specific differences offer mechanistic clues about how unequal responses are evoked by GRs and PRs under otherwise identical conditions.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
show a robust interaction between NCoR or SMRT and either PR-A or PR-B complexes of the antiprogestin RU486. A much weaker, but statistically very significant, interaction is seen with agonist (R5020)-bound PRs. This is similar to what was observed for agonist- and antagonist-bound GR interactions with GAL/NCoR-RID (12). Western blots indicate that equal amounts of VP16/PR-A and -B, and GAL/NCoR-RID and GAL/SMRT-RID, are expressed in Cos-7 cells (data not shown).
|
) argue that the CoRNR box motifs are also essential for the association of both antagonist- and agonist-bound complexes of PR-A and PR-B with NCoR.
We previously documented differing behavior of PRs and GRs to corepressors (19). The much lower luciferase activity for NCoR-RID interacting with PR-B (Fig. 1A) than GR (12) suggests that NCoR-RID binding to PRs may be less avid than to GRs. This possibility was assessed by examining the relative ability of increasing amounts of NCoR or SMRT to reduce the interaction of GAL/NCoR-RID with PR-B and GR complexes bound by the antisteroid RU486 (Fig. 1C). The approximately equal levels of competition for NCoR-RID interactions with PR-B and GR by NCoR and SMRT suggest that no major differences exist in the affinity of PR- and GR-antagonist complexes for these corepressors.
N Terminus of PR Is Required for Corepressor Binding in Two-Hybrid Assays
Most studies have identified the LBD in the C-terminal half of steroid/nuclear receptors as being the corepressor binding site (14, 25, 26, 27). More recent studies indicate that GR binding to corepressors involves both N- and C-terminal domains (11, 12) and suggest that a combination of these two domains is required for the expression of biological activity of corepressors with both GR and PR (19). We therefore conducted a detailed examination of the PR and GR N-terminal sequences. Note that the numbering of the full-length PR-B (1–933) is used in all truncated PRs.
The chimera lacking the first 467 amino acids, VP16/PR468C (Fig. 2A), affords close to wild-type activity in the mammalian cell two-hybrid assay (Fig. 2B). However, deletion of the next 41 residues (468–508 to give VP16/PR509C) dramatically reduces the interaction with NCoR (Fig. 2B). Interestingly, this 41-amino-acid region is unable to mediate a productive response with GAL/NCoR-RID when present in either the larger sequence of PR395–634 (Fig. 2B) or the entire N-terminal domain of PR-B (1–535) or PR-A (165–535) (data not shown). In all cases, Western blots show that approximately equal amounts of VP16/PR chimera protein are being expressed (data not shown). The possibility that different amounts of the various expressed chimeric proteins are functionally active was eliminated by showing that the assays with 10-fold higher amounts of chimeric receptor plasmid afforded the same pattern as seen in Fig. 2B (data not shown). As a further check for functional expressed proteins, all chimeras containing an intact LBD were found to cause steroid-inducible transactivation of a progesterone-regulated luciferase reporter (GREtkLUC) (data not shown). Therefore, we propose that PR-B residues 468–508 are necessary, but not sufficient, for PR association with NCoR under our conditions.
|
|
). Again, no major differences are obtained when 10-fold more PR plasmid is used, indicating that sufficient functionally active chimeric protein is present (data not shown). Interestingly, sequences upstream of residue 468 appear to play a larger role in PR binding to SMRT than to NCoR (c.f. Fig. 2, C vs. B).
To confirm the importance of PR residues 468–508, we examined two internal deletion mutants lacking amino acids 468–508 or 479–508. As seen in Fig. 2D, these two deletions eliminate most of the interaction of PR with both NCoR and SMRT. These results support the conclusions from Fig. 2, B and C, that the 41 amino acids from 468–508 of PR-B are necessary for PR binding to corepressors.
Identification of the GR N Terminus Region Required for Corepressor Binding in Two-Hybrid Assays
The above approach with PR was next employed to determine the location of the putative corepressor binding domain in the GR N terminus (11, 12). The data of Fig. 3A imply that residues 206–236 are critical for GR binding to NCoR. However, GR constructs containing an internal deletion argue for a slightly larger domain (i.e. amino acids 154–236) (Fig. 3B). The equal levels of expressed protein from VP16/GR, /GR
206–236, and /GR154–236 (data not shown) indicate that the low activities of GR deletion constructs in Fig. 3B are not due to inadequate levels of receptor protein.
|
, could substitute for the comparable region of GR, i.e. amino acids 154–236. To do this, we created a hybrid chimera of VP16 and GR with PR-B residues 458–514 replacing GR residues 154–236. The fact that this hybrid receptor affords almost as much transactivation activity (and expressed protein; data not shown) as the wild-type VP16/GR (Fig. 3B) strongly argues that PR-B amino acids 458–514 comprise an NCoR-RID binding domain and that this domain can replace the comparable domain of GR.
GR and PR N-Terminal Regions Display Functional Redundancy for Gene Induction
The properties of GR, GR154–236, and GR/PR/GR in conventional gene induction assays were next examined to determine whether the ability of PR sequences to substitute for GR residues is limited to corepressor binding or includes any of the more elaborate processes involved in transactivation. As shown in Table 1A, deletion of amino acids 154–236 to give GR has multiple consequences: the total amount of induced gene product is diminished, the fold induction by dexamethasone (Dex) is reduced, the partial agonist activity of the antiglucocorticoid DexOx (32) is decreased, and the EC50 of the dose-response curve is increased. Replacing the missing GR sequence in GR154–236 by amino acids 458–514 of PR-B to give GR/PR/GR restores each of these properties to nearly their wild-type levels (Table 1A).
|
, and GR/PR/GR was determined in 1470.2 cells. This time the deletion in GR, and the replacement by the PR sequence, has little effect on either the total amount of induced gene product or the fold induction by Dex, whereas the changes in partial agonist activity and EC50 are similar to those in CV-1 cells (Table 1B). However, the ability of NCoR to reduce the partial agonist activity of DexOx and to increase the EC50, due to a right shift in the dose-response curve, is lost upon deletion of amino acids 154–236 of GR and is reversed, but not completely restored, upon adding PR residues 458–514 (Table 1C). Therefore, amino acids 458–514 of PR-B and 154–236 of GR are functionally interchangeable for all of the properties of GRs that we have examined to date.
Competition of NCoR Binding to PRs and GRs by Coactivators
Coactivators and corepressors competitively inhibit the binding of each other to both agonist- and antagonist-bound GRs (12). Because coactivators and corepressors interact with PR-agonist and antagonist complexes (Figs. 1 and 2 and Refs. 18 and 19), we asked whether coactivators can also compete with corepressors for binding to PRs. We find that transcription intermediary factor (TIF) 2.0, an N-terminal fragment of the coactivator TIF2 (Fig. 4A), competes for GAL/NCoR-RID interactions both with PR-B and, as previously reported (12), with GR (Fig. 4B). Surprisingly, constructs containing the sequence of TIF2.0, such as TIF2 and a TIF2 mutant (TIF2m123) in which the RIDs have been altered to prevent binding to steroid receptors, still compete for corepressor binding to GRs (Fig. 4A and Ref. 12) but not PRs. Similarly, the interaction of RU486-bound PR-B with SMRT-RID is reduced by TIF2.0 but not by TIF2 or TIF2m123 (data not shown). Thus, sequences outside of the N terminus of TIF2 affect its ability to compete for corepressor binding to PR but not GR.
|
shows that the immobilization of [35S]methionine-labeled PR-B by glutathione-S-transferase (GST)/NCoR-RID is much greater than that by GST and thus depends upon the presence of NCoR-RID. This binding is relatively independent of ligand. Surprisingly, whereas the biological activities of full-length PR-B and PR468C are much greater than that of PR562C in mammalian cell two-hybrid assays (see Fig. 2B), densitometric scanning of paired experiments (Fig. 5, A and B) shows that 3–8% of the input PR, or PR562C, ± steroid was immobilized by NCoR-RID (data not shown). We therefore examined the binding to GST/NCoR-RID of N- and C-terminal fragments of both PR and GR. As expected (12), neither half of GR binds to NCoR-RID (Fig. 5C). In contrast, a PR-B C-terminal fragment (PR562C) binds to NCoR-RID, whereas the N-terminal half (PRN561) does not. Thus, there are significant differences in the ability of the C-terminal domains of GR and PR to bind to NCoR-RID in cell-free pull-down assays.
|
shows that PR-B is associated with GAL/NCoR-RID but only when NCoR-RID is present (lanes 8 vs. 7). As in the pull-down assays (Fig. 5A), the formation of a PR-B/NCoR-RID complex is ligand independent (lanes 8–10).
|
) could be recapitulated under whole-cell conditions. Here, VP16 chimeras of the PR fragments were used to facilitate immunochemical identification. VP16/PR-B is coimmunoprecipitated with GAL/NCoR-RID in the presence of the antiprogestin RU486 in a manner that requires NCoR-RID (compare lanes 7 vs. 6 of Fig. 6B), just like the wild-type PR-B (Fig. 6A). Thus, the VP16 moiety does not alter the ability of PR-B to associate with GAL/NCoR-RID in whole cells. Of the various PR fragments, the N-terminal species PR-BN is not coimmunoprecipitated (lane 10), whereas the C-terminal fragments PR468C and PR509C are, all in the presence of RU486. However, laser densitometry shows that the amount of receptor coimmunoprecipitated with NCoR-RID is about 6-fold less for PR509C than for full-length PR-B and PR468C (data not shown), suggesting that the efficiency of PR509C binding to NCoR is reduced and that more N-terminal sequences of PR are needed for full binding activity.
Finally, we asked whether the PR sequence 458–514 could restore corepressor binding in a coimmunoprecipitation assay to GRs that are deleted of their corepressor binding site. As expected from the data of Fig. 3B, the deletion of GR residues 154–236 dramatically reduces the binding of GR to NCoR-RID under whole-cell conditions (lanes 8 vs. 6 in Fig. 6C). However, replacement of the deleted GR sequence with PR amino acids 458–514 rescues the NCoR binding of GR (lanes 7 vs. 6). This reinforces our above conclusions about the importance of these regions for corepressor binding to PRs and GRs.
Activation Is Required for Receptor Binding to NCoR
The binding of PRs to NCoR in the pull-down and coimmunoprecipitation assays is steroid independent (Figs. 5A and 6A). In contrast, PR interactions with NCoR in whole-cell mammalian two-hybrid assays are steroid dependent (Figs. 1–4). An identical dichotomy was observed for GR interactions with NCoR-RID (12). It is well known that the DNA binding of steroid receptors requires prior activation of the receptor protein. Furthermore, it is thought that the DNA binding of activated receptors precedes the recruitment of cofactors (38). Therefore, we speculated that a ligand-independent activation of cell-free receptors might be permitting their binding to corepressors. To test this hypothesis, we examined the effect of sodium molybdate, which blocks receptor activation (36, 37), on GR and PR binding to NCoR in the presence and absence of added steroid in pull-down assays. Steroid (1 µM) was added to maximize the amount of steroid-bound receptors that are obtained during in vitro translation (39). Sodium molybdate (20 mM) was present during the in vitro translation reactions to prevent any possible activation of GRs or PRs during their synthesis. As a control for nonspecific ionic effects of molybdate (MoO4–2), we used the iso-electronic sulfate group (SO4–2). The presence of molybdate or sulfate does not alter the amount of in vitro-translated GR (Fig. 7A, lanes 1–6), nor do they increase the background binding of GRs to GST (Fig. 7A, lanes 10 and 11). However, 20 mM molybdate does selectively inhibit the cell-free binding of GRs ± agonist or antagonist ligands to NCoR-RID (Fig. 7A, lanes 7–9 vs. 12–14). Similarly, 20 mM molybdate, but not sulfate, dramatically reduces the cell-free binding of PRs ± agonist or antagonist to GST/NCoR-RID (Fig. 7B, lanes 7–9 vs. 12–14). Thus, it appears that the activation of receptors is required for binding to corepressors in pull-down assays and that the binding of ligand-free GRs and PRs to corepressors (Fig. 5A and Ref. 12) is due to the formation of ligand-free activated receptors.
|
) is also due to an activation of ligand-free receptors that can be blocked by molybdate. We did not try to prevent the activation of ligand-bound PRs within the cells because the ionic nature of sodium molybdate makes it difficult to get a sufficiently high intracellular concentration of molybdate. However, when added immediately after cell lysis, molybdate, but not sulfate, reduces the coimmunoprecipitation of ligand-free PRs and GRs (lanes 7 vs. 8 of Fig. 7C, top and bottom panels, respectively) by a factor of 4 (Fig. 7D). This argues that activation is responsible for most of the coimmunoprecipitation of ligand-free receptors with NCoR. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
and 3) in a manner that requires the corepressor RID sequence (Fig. 1B). However, this amino-terminal binding domain is not sufficient for wild-type whole-cell interactions of corepressor with either receptor (Fig. 2C) (12). Similar results are seen under cell-free pull-down (Fig. 5) and whole-cell coimmunoprecipitation (Fig. 6) conditions, although PR constructs lacking the amino-terminal domain show more association than in the whole-cell bioassays (Fig. 2). This may be due, in part, to the much higher concentrations of receptors and corepressors that are used in the pull-down and coimmunoprecipitation assays, which would permit the formation of more weakly bound complexes. It has been estimated that protein-protein interactions with affinities lower than 50–100 nM are not detected in two-hybrid assays (40).
We could discern no obvious sequence homology between the two corepressor binding regions of PR and GR. Nevertheless, the PR domain of 458–514 restores NCoR binding to the inert GR deletion mutant (GR154–236) under whole-cell coimmunoprecipitation and two-hybrid conditions (Figs. 3B and 6C) and largely replaces the deleted GR domain in a variety of transactivation responses (Table 1). This is strong evidence for the importance of these two, nonhomologous sequences in receptor binding and function of corepressors. These results further suggest that the tertiary structure of the protein, and not amino acid sequence, is critical for corepressor binding to, and the biological activities of, these amino-terminal domains (Fig. 8). In sum, these data nicely explain why both N- and C-terminal regions of PR and GR were found to be required for the full response to corepressors (19).
|
The corepressor binding sequences that we identified at amino acids 458–514 of human PR(hPR), and 154–236 of rat GR, overlap the AF-1 domains of GRs (residues 98–282) (42) and PRs (456–546) (43). Given the importance of the AF-1 domain for the amount of transactivation (30, 44, 45) and proper RNA splicing (46) in conjunction with the requirement of a corepressor binding sequence in the GR N-terminal domain for robust transactivation (Table 1), it is tempting to speculate that the competitive binding of corepressors and other factors to the AF-1 domain is sufficient to block many of the AF-1 activities.
Protein sumoylation has been suggested to facilitate corepressor binding (47). Rat GR is sumoylated at residues K297 and K313 (48), and human PR is sumoylated at K388 (49, 50). However, these residues are outside of the regions that contribute to corepressor binding: GR154–236 and PR458–514 (Figs. 2, B and C, and 3). Therefore, we conclude that the sumoylation sites of GR and PR are not required for corepressor binding.
The whole-cell binding of corepressors to both PRs and GRs appears to be reversible (Fig. 1C and Ref. 12). We confirm that coactivators inhibit the interactions of NCoR with GRs (Fig. 4B). The amino-terminal portion of TIF2 also counteracts the association of NCoR with PR under conditions where the full-length TIF2 is inactive. Whether this is due to the higher level of expression of the fragment TIF2.0 vs. the full-length TIF2 (our unpublished observations), to an unidentified PR-associated protein that impedes TIF2 binding, or to some unique property of TIF2.0 remains to be determined. Nevertheless, the data with PR support our general hypothesis that coactivators and corepressors competitively inhibit the binding of each other to receptor-agonist and -antagonist complexes (12, 16, 17, 28).
Differences between PRs and GRs
Numerous examples exist for PRs and GRs eliciting different responses even within the same cells (19, 51). The molecular reasons for this are likely to be multifactorial. One component appears to be unequal interactions of corepressors with a combined receptor target of N- and C-terminal domains (19). The present study supports this hypothesis. The quantitative response of GRs in mammalian two-hybrid assays is much greater for NCoR than SMRT (12) but about the same with PRs (Fig. 1A). These results, in combination with the recent report that the magnitude of gene expression in two-hybrid assays is determined by the affinity of the two interacting proteins for each other (40), suggest that the affinity of corepressors for GRs and PRs are not equal. Thus, the same cellular concentration of corepressors would be predicted to afford different amounts of corepressor-bound receptors (Fig. 8). Cell-specific factors are also important, as shown here by the much greater reduction, upon deleting GR residues 154–236, in the total amount of induced luciferase, and the fold induction of luciferase, in CV-1 cells than in 1470.2 cells (Table 1). The identification of these cell-specific factors should be very informative.
We have not been able to identify the amino acids that participate in corepressor binding by the N-terminal domain of PRs and GRs because no interactions are observed for the isolated domains in two-hybrid (Fig. 2B and Ref. 12), pull-down (Fig. 5C and Ref. 12), or coimmunoprecipitation (Fig. 6B) assays. We propose that the C-terminal sequence of each receptor plus other cellular proteins contribute to the formation of the final complex. Nevertheless, the fact that the C-terminal half of PR has a higher avidity for NCoR-RID than does the GR C terminus in both pull-down (Fig. 5C) and coimmunoprecipitation (Fig. 6B) assays is yet another difference between the two receptors (Fig. 8).
Coactivators compete for corepressor binding to both PRs and GRs (Fig. 4 and Ref. 12), but the determinants of this competition are not identical. Sequences outside of TIF2.0, which is the N terminus of TIF2 (Fig. 4A), affect its ability to compete for corepressor binding to PR but not GR. Thus, one would expect that the same intracellular ratio of coactivators to corepressors would result in a different proportion of PRs and GRs being associated with coactivators (and corepressors), thereby resulting in unequal amounts of gene activation or repression of common target genes (Fig. 8). In summary, these differences in coactivator binding in combination with the unequal affinities of corepressors for binding to selected regions of PRs and GRs expand the mechanisms by which nonidentical responses can be generated by PRs and GRs.
Molybdate Inhibition of Receptor Binding to Corepressors
The present data indicate that the apparent contradiction of why the whole-cell biological responses of GRs and PRs with corepressors, but not the biochemical binding/association, require ligand-bound receptors is due, at least in part, to the activation of ligand-free receptors under cell-free conditions. Activation is defined as a still poorly understood process that converts steroid receptors from a non-DNA-binding to a DNA-binding form and is associated with the dissociation of chaperon proteins such as heat shock protein 90 (37). Steroid-free receptors can be activated (52) under several conditions (53) that do not occur in cells but can be blocked by sodium molybdate (36, 37). The observation that molybdate both prevents activation and reduces the cell-free binding/association of GRs and PRs with NCoR in pull-down and coimmunoprecipitation assays (Fig. 7) argues that only activated receptors can bind to corepressors. Whether receptor-bound chaperone proteins block corepressor binding is a logical, but untested, hypothesis even though heat shock protein 90 is thought to bind only to the GR LBD (54). Molybdate, but not sulfate, inhibits the formation of activated complexes in coimmunoprecipitation assays (Fig. 7C), suggesting that we are not witnessing a general salt effect and that most of the activation occurs after cell lysis. Molybdate does not have any effect on activated receptors (55, 56), so it is unlikely that molybdate disrupts receptor-corepressor complexes that were preformed in intact cells. Rather, we interpret the data of Fig. 7, C and D, as showing that the ligand-free PRs and GRs present in intact cells can be activated by various processes subsequent to cell rupture. Whether the small amount of ligand-free PR and GR that binds to corepressors in the coimmunoprecipitation assays in the presence of 20 mM molybdate (lanes 7 vs. 6 in Fig. 7C and lanes labeled "Mo/IP" in Fig. 7D) represents the existence of a low level of activated ligand-free receptors in intact cells remains to be established.
The possible presence of ligand-free, activated receptors in intact cells, coupled with the requirement of steroid for GR binding to coactivators in pull-down assays (28), permits an interesting mechanism for the deinduction of steroid-regulated gene transcription (Fig. 8). The data of Fig. 7 suggest that the dissociation of steroid from activated receptors bound to hormone response elements will cause a major decrease in coactivator affinity with little change in the affinity of corepressors. Given the ability of corepressors and coactivators to competitively inhibit the binding of each other to GRs in an equilibrium fashion (12, 17), this decrease in coactivator affinity would result in a markedly increased ratio of corepressors to coactivators that are associated with the hormone response element-bound receptors and a concomitant decrease in transactivation activity. Given the similar corepressor-binding properties of ligand-free PRs and GRs, we propose that the ligand-independent binding of the other classical steroid receptors to corepressors (7, 10, 11, 12, 15, 20, 35) also depends upon the activation state of the receptor. In this case, the above-proposed model for gene deinduction may prove to be general for all of the classical steroid receptors. How many other proteins bind to steroid receptors only after activation remains to be seen.
Some studies with ERs support the hypothesis that cofactor binding to receptors can modify receptor affinity for ligands (Ref. 57 but not Ref. 58). Comparable studies have not been conducted with GRs. However, steroids cannot bind to activated GRs (59). Therefore, the fact that activation is needed for corepressor binding to GRs strongly argues that corepressor binding to GRs cannot alter the affinity of steroid binding to ligand-free GRs. Similarly, the binding of coactivator TIF2 to DNA-bound GRs occurred only with GR-steroid complexes (60). At this point, steroid binding to GRs is again not reversible, and the affinity of steroid binding to GRs cannot be affected. Instead, some step downstream of steroid binding, and GR activation (32), is a more probable target for corepressors and coactivators in their modulation of the dose-response curve of agonists and the partial agonist activity of antagonists (12, 16, 60).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Chemicals
Dex was obtained from Sigma (St. Louis, MO), and promegestone (R5020) was from PerkinElmer Life Sciences (Boston, MA). RU486 was a gift from Etienne Baulieu (Paris, France). Restriction enzymes and DNA polymerase were from New England Biolabs (Beverly, MA), Amersham Biosciences (Piscataway, NJ), or Promega (Madison, WI). [35S]Methionine was from Amersham Biosciences, and sodium molybdate and sodium sulfate were from Baker Chemical (Phillipsburg, NJ).
Plasmids
The Renilla null luciferase reporter was purchased from Promega (Madison, WI), pM vector and GAL/VP16 from CLONTECH (Palo Alto, CA), and pFR-LUC reporter from Stratagene (La Jolla, CA). Human serum albumin/pS65, VP16/GR, VP16/GR361C, VP16/GR407C, pRBAL-GR, pRBAL-GRN523, pRBAL-GR486C have been described previously (12). Chimeras of the VP16 activation domain and PR-B, PR-A, PR-BN, PR-AN, and PR-LBD were gifts from Dean P. Edwards (University of Colorado Health Sciences Center, Denver, CO). Other donated plasmids were received from Hinrich Gronemeyer (Institut de Génétique et de Biologie Moléculaire et Cellulaire, Strasbourg, France: hPR-B, TIF2, TIF2.0, and TIF2m123), Michael Rosenfeld (University of California-San Diego, San Diego, CA: NCoR), Ron Evans (Salk Institute, La Jolla, CA: s-SMRT), Keith Yamamoto (University of California-San Francisco, San Francisco, CA: GR), and Mitch Lazar (University of Pennsylvania School of Medicine: GAL/NCoR-RID [amino acids 1944–2453], GAL/NCoR-RIDm12, GAL/SMRT-RID [amino acids 982–1495], and GST/NCoR-RID).
Several chimeras were synthesized as follows. For VP16/PR468C, the AscI-(filled in) and PstI digest of VP16/PR-B was inserted into EcoRI-(filled in) and PstI-cleaved VP16. For VP16/PR545C, the Bsu36I-(filled in) and SapI fragment of VP16/PR-B was inserted into HincII- and SapI-treated VP16. For VP16/PR562C, the PleI-(filled in) and PstI-restricted VP16/PR-B was inserted into the SalI (filled in) and PstI product from VP16. For VP16/PR395C, the StuI- and XbaI-generated fragment of VP16/PR-B was inserted into the MluI-(filled in) and XbaI of VP16. For VP16/PR395–634, StyI-(fill-in) and EcoRI of VP16/PR395C was inserted into HindIII-(filled in) and EcoRI-treated VP16. VP16/PR479C and VP16/PR509C were constructed using PCR amplification of VP16/PR-B with the forward primers 5'-AGAATTCCCCTGCAAGGCGCCGGGC-3' and 5'-AGAATTCCCCGCGCTCTACCCTGCACTC-3', respectively, and the same reverse primer 5'-GCTCTAGAGCTTTTTATGAAAGAGAAG-3'. The resulting PCR product was digested with EcoRI/XbaI, and inserted into VP16 using the same restriction sites. VP16/PR468–508 was prepared using PCR amplification of VP16/PR-B with the forward primer 5'-GCCCCCGCGCTCTACCCTGCACTC-3' and the same reverse primer as for VP16/PR479C. The resulting PCR products were digested with XbaI and inserted into VP16/PR-B using AscI (filled in) and XbaI. VP16/PR479–508 was synthesized using PCR amplification of VP16/PR-B with two separate sets of primers: one set primers was 5'-GGAATTCATGACTGAGCTGAAGGCAAAG-3' and 5'-TGGAGGTGGCGCGAACGGGCCCTG-3', the other set was the same as for VP16/PR468–508. The resulting PCR product was digested with EcoRI and XbaI and inserted into VP16 using the same restriction sites.
To make VP16/GR154–236 and VP16/GR206–236 were prepared by amplifying VP16/GR with the same forward primer (5'-GGAATTCATGGACTCCAAAGAATCCTTAGC-3') coupled with the reverse primers 5'-GAAGATCTGTGGGATACAATTTCACACTGCC-3' for VP16/GR206–236 or 5'-GAAGATCTGTCGACCTATTGAGGTTTG-3' for VP16/GR154–236. Each resulting PCR product was digested with EcoRI and BglII and ligated with two other fragments: the 1.7-kb fragment obtained by BglII and XbaI digestion of VP16/GR and the 3.3 kb fragment from EcoRI/XbaI digestion of VP16. pSVLGR154–236 was prepared by cutting pSVLGR with AccI and BglII and then self-ligating after filling-in the 3'-recessed termini with Klenow fragment. The VP16/GR/PR/GR chimera was prepared by ligating three fragments: 1) the PCR amplified product from VP16/PR-B with forward primer 5'-CCATCCAGACCCGGGGAAGCG-3' and reverse primer 5'-GAAGATCTGCAGGGTAGAGCGCGGG-3' that was then digested with SalI and BglII, 2) the 1.7-kb fragment obtained by BglII and XbaI digestion of VP16/GR, and 3) the 3.7-kb fragment from SalI/XbaI digestion of VP16/GR. VP16/GR/PR/GR was digested with SalI and BglII and the 170-bp fragment was inserted into pSVLGR using the same restriction sites to yield pSVLGR/PR/GR.
For pSG5/PR-B, the EcoRI and XbaI (filled in) digestion product of VP16/PR-B was inserted into EcoRI- and BamHI-(filled in) treated pSG5. PSG5/PR562C and pSG5/PR561N were both made by PCR, using VP16/PR-B as the template. For pSG5/PR562C, the primers were 5'-GGAATTCTTACCTCAGAAGATTTGTTTAATC-3' and 5'GAAGATCTCACTTTTTATGAAAGAGAAG-3'. For pSG5/PR561N, the primers were 5'-GGAATTCATGACTGAGCTGAAGGCAAAGG-3' and 5'-GAAGATCTGACTCGAAGCTGTATTGTGG-3'. In both cases, the amplified DNA was then cut with EcoRI and BglII and inserted into the same sites of pSG5.
Cell Culture and Transient Transfection
Monolayer cultures of CV-1, Cos-7, and 1470.2 cells were grown as previously described (18, 30). The total transfected DNA was adjusted to 300 ng/well of a 24-well plate (or 3 µg/60-mm dish) with pBluescriptII SK+ (Stratagene). Renilla TS (Promega) (5–10 ng/well of a 24-well plate) was included as an internal control. Cells were incubated with plasmid DNA, Opti-MEM I, and LipofectAMINE or FuGENE 6 reagent (Roche Molecular Biochemicals, Indianapolis, IN) for 24 h, after which this mixture was replaced by the normal media (10% fetal calf serum, DMEM). The cells were induced with steroids for 20–24 h and then harvested. The cells were lysed and assayed for reporter gene activity using the Dual Luciferase Assay reagents according to manufacturer’s instruction (Promega). Luciferase activity was measured by an EG&G Berthhold (Oak Ridge, TN) luminometer (Microlumat LB 96P). The data were normalized either for total protein or Renilla null luciferase activity.
Mammalian Two-Hybrid Assays
The recommended procedure for the Mammalian Matchmaker two-hybrid assay kit (CLONTECH) was modified slightly by changing from a chloramphenicol acetyltransferase reporter to the luciferase reporter pFRLuc (Stratagene), which is under the control of five repeats of the upstream activating sequence for the binding of GAL4.
Bacterial Expression of Proteins
The pGEX series of plasmids were transformed into Escherichia coli (BL21[DE3]; Stratagene) according to the manufacturer’s procedure. A single colony was picked and inoculated into 3 ml Luria Bertani broth with 100 mM ampicillin. After overnight culture, 1 ml of bacterial culture was diluted into 50 ml of Luria Bertani broth containing 100 mM ampicillin, shaken at 37 C for 2 h, adjusted to 0.5 mM isopropyl-ß-D-thiogalacto-pyranoside, and shaken at 25 C for another 3 h. The cells were harvested by centrifugation, washed once with PBS, resuspended in 10 ml of PBS, and sonicated for 30 cycles at 30% of maximum power (Fisher Scientific sonic dismembrator, model 500). The supernatant was collected for use after centrifugation (5000 x g for 20 min).
In Vitro Transcription and Translation Assays
For each reaction, 1 µg of plasmid DNA was mixed with 2 µl of [35S]-methionine and 40 µl of TNT T7 (or SP6) master mix (Promega) and brought up to a total volume of 50 µl with H2O. The reaction was conducted at 30 C for 90 min. Various steroid hormones (Dex, R5020, or RU486), with or without 20 mM sodium molybdate or sodium sulfate were added into the transcription-translation reaction during the 30 C incubation.
Pull-Down Assays
Sonicated bacterial lysates (0.5 ml) containing overexpressed GST or GST/NCoR-RID were incubated with 20 µl of glutathione-Sepharose 4B beads for 1 h at 0 C. The mixture was centrifuged (12,000 x g), the supernatant discarded, and the matrix was washed with PBS (4 x 1 ml). Each 20 µl sample of immobilized GST or GST-chimera was then incubated overnight at 4 C with 10 µl of hormone prebound, activated, [35S]labeled in vitro-translated GR. The matrix was washed with Buffer H (4 x 1 ml). The immobilized proteins were removed from the beads by heating at 90 C for 5 min in 20 µl of 2x sodium dodecyl sulfate (SDS) loading buffer. The proteins were then separated on 8 or 10% SDS-PAGE gels, and the bound receptor was located by autoradiography.
Coimmunoprecipitation Assays
The day before transfection, Cos-7 cells were seeded into 150-mm dishes at 200,000 cells per dish containing 20 ml of media. On the next day, about 30 µg of DNA/dish was transfected with 60 µl of FuGene reagent. After 2 d, the cells were treated with EtOH ± 1 µM RU486. Cells were lysed 2 h later at room temperature with CytoBuster Protein Extraction Buffer (Novagen, Madison, WI) and clarified by centrifugation at 16,000 x g for 5 min at 4 C. For complexes with GAL-DBD tagged proteins, immunoprecipitation was achieved by incubating with anti-GAL4DBD (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) while rocking (100 cycles/min) at 4 C overnight and then immobilizing the antibody complexes on Protein G Plus/Protein A-Agarose (Oncogen Research Products, San Diego, CA) with rocking (100 cycles/min) at 4 C for another 2 h. In both cases, the agarose beads were then washed three times with 1 ml of ice-cold wash buffer [50 mM Tris (pH 7.5), 150 mM NaCl, 0.1% Nonidet P-40, complete protease inhibitor cocktail (Roche)]. The proteins were extracted with 2x SDS-loading buffer (95 C for 3 min) and separated by gel electrophoresis (8–10% SDS-PAGE), and visualized by Western blotting with anti-PR (PR-22) or anti-GR (BUGR-2) antibody (Affinity Bioreagents, Golden, CO).
Western Blotting
SDS-PAGE gels were equilibrated in transfer buffer for 15 min at room temperature before electrophoretic transfer of the proteins to nitrocellulose membranes (Schleicher & Schuell BioScience, Keene, NH) in a Bio-Rad (Hercules, CA) small (150–200 mA overnight) or large (350 mA overnight) Transblot apparatus. The nitrocellulose membranes were stained with Ponceau S (0.02% Ponceau S and 0.04% glacial acetic acid in water) to localize the molecular weight markers. The VP16 fusion proteins were probed with rabbit VP16 polyclonal antibody (CLONTECH) and the GAL4 fusion proteins were probed with mouse GAL4 DBD monoclonal antibody (CLONTECH). Antibody complexes were visualized by ECL detection reagents as described by the manufacturer (Amersham Biosciences). Western blot films were scanned using a Bio-Rad Model GS-800 Calibrated Imaging Densitometer and the densitometry quantitated using Bio-Rad Quantity One software.
Statistical Analysis
Unless otherwise noted, all experiments were performed in triplicate several times. The values of n independent experiments were then analyzed for statistical significance by the two-tailed Student’s t test using the program InStat 2.03 for Macintosh (GraphPad Software, San Diego, CA). When the difference between the SDs of two populations is significantly different, then the Mann-Whitney test or the Alternate Welch t test is used.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Abbreviations: AD, Activation domain; AF, activation function; AR, androgen receptor; GAL, GAL4 DNA binding domain; DBD, DNA binding domain; Dex, dexamethasone; ER, estrogen receptor; GST, glutathione-S-transferase; GR, glucocorticoid receptor; hPR, human PR; LBD, ligand binding domain; NCoR, nuclear receptor corepressor; PR, progesterone receptor; RID, receptor interaction domain; SMRT, silencing mediator of retinoid and thyroid receptor; TIF, transcription intermediary factor; VP16, VP16 activation domain.
Received for publication January 7, 2005. Accepted for publication March 7, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
0.0052; ***, 
, 
