| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Departments of Medicine (Division of Hematology, Oncology, and Transplantation) and Pharmacology, University of Minnesota Cancer Center, Minneapolis, Minnesota 55455
Address all correspondence and requests for reprints to: Carol A. Lange, Ph.D., University of Minnesota Cancer Center, 420 Delaware Street Southeast, MMC 806, Minneapolis, Minnesota 55455. E-mail: Lange047{at}umn.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
We and others have shown that growth factor activation of intracellular kinase pathways can induce and augment posttranslational modifications of PR, including phosphorylation (4) and ubiquitination (5); these events contribute to altered hormone responsiveness. For example, Ser294-phosphorylated PRs exhibit heightened transcriptional activity and increased turnover relative to phospho-mutant S294A PR, containing a Ser294 to Ala substitution (5, 6).
In addition to phosphorylation and ubiquitination, PRs are sumoylated in response to ligand (7). Sumoylation is a dynamic process involving conserved enzymatic cascades that rapidly catalyze the formation and cleavage of isopeptide bonds between the target protein lysine and the C-terminal glycine of small ubiquitin-like modifier (SUMO) proteins (8). SRs are among the transcription factors targeted for sumoylation, including PR, glucocorticoid receptors, androgen receptors (ARs), estrogen receptors (ERs) and mineralcorticoid receptors (7, 9, 10, 11, 12, 13). The majority of these receptors display diminished transcriptional activity when sumoylated. SUMO attachment to PR occurs primarily at Lys388 in the N-terminal region via the E3 ligase activity of protein inhibitor of activated STAT (PIAS)1 (14) or PIAS3 (15). In addition to Lys388, overexpression of PIAS3 induces PR-B sumoylation at Lys7 and Lys531 (15). PR sumoylation represses PR transcriptional activity on tandem PRE (progesterone response element)-driven promoters in addition to mediating the transrepressive actions of PR-A and PR-B receptors (7).
A distinctive characteristic of SUMO regulation of transcriptional repression is that it is controlled, in part, by phosphorylation events. For example, Elk-1 is sumoylated and transcriptionally repressed before MAPK phosphorylation, which leads to the loss of SUMO and transcriptional activation (16). C-Fos is similarly activated by phosphorylation-induced desumoylation (17). Phosphorylation-dependent reversal of SUMO repression has also been reported for the coactivator amplified in breast cancer 1 (steroid receptor coactivator 3), whereupon MAPK phosphorylation leads to SUMO removal and restores amplified in breast cancer 1 transactivation (18). In contrast, a phosphorylation-dependent variation of the 
KxE (19) consensus motif, known as phospho-dependent SUMO motif, controls the sumoylation of many transcription factors including MEF2 and GATA-1 (20, 21).
Herein, we sought to understand how PR sumoylation is regulated in the context of steroid hormone- and growth factor-initiated signaling events. We found that PR sumoylation is negatively regulated by phosphorylation events in response to mitogenic signaling pathways. We conclude that reversible PR sumoylation/desumoylation rapidly alters hormone responsiveness according to the hormonal milieu; growth factor stimulation of cells activates kinases, inducing phosphorylation of PR and favoring desumoylation, thereby sensitizing PR to subthreshold levels of progestins. Unimpeded PR transcriptional activity in response to up-regulated growth factor-induced signaling pathways has important implications for breast cancer biology. Heightened kinase activity is typical of advanced breast cancers, suggesting that phosphorylated and undersumoylated PRs may be highly active, but perhaps undetected due to their rapid turnover (22).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
).
|
, lane 5); this band was recognized by both PR and SUMO-1 antibodies. Notably, this band did not appear in lysates from cells transfected with wt PR-B alone (lane 1), EGFP-SUMO-1 alone (lane 2), or EGFP-SUMO-1 and K388R PR-B (lane 3), confirming that it is indeed sumoylated PR-B. In addition, sumoylated PR-B in PR immunoprecipitates was detected with both PR- and SUMO-1-specific antibodies (data not shown).
PR sumoylation has been reported to suppress PR transcriptional activity in transiently transfected HeLa cells and is detectable after 24 h of progestin treatment (7). We detected sumoylated PR, however, after just 1 h exposure to ligand (Fig. 1B). To further understand the kinetics of PR sumoylation, we performed a short time course of progestin treatment in HeLa cells transiently transfected with wt PR-B and EGFP-SUMO-1. PR expressed in HeLa cells closely mimics endogenous PR with regard to regulated changes in phosphorylation, location, and turnover (5, 6, 25, 26). Serum-starved cells were treated with R5020 (10–8 M) for 0–60 min (Fig. 1C). Interestingly, sumoylated PR-B was detected as early as 15 min after R5020 treatment, indicating that rapid sumoylation may modulate PR action almost immediately after ligand exposure. In a related experiment, the PR antagonist/partial agonist, RU486, also induced rapid PR sumoylation in HeLa cells (Fig. 1D). RU486-induced PR sumoylation as efficiently as R5020 and did not alter the effects of R5020 during cotreatment, suggesting that sumoylation of PR may not be relevant to antagonist-bound transcriptional repression. Together, these data demonstrate that liganded PR is rapidly sumoylated (15 min), suggesting a role for sumoylation-dependent modulation of PR in a fast-acting mechanism of hormone responsiveness.
PR Sumoylation Alters Hormone Sensitivity and Promoter Selectivity
PR sumoylation has only previously been studied using transiently expressed PR in HeLa or Cos-1 cell lines (7) (Fig. 1), as well as CV-1 (Green Monkey kidney cells) (27), 293T (human renal epithelial cell line) (15), and human endometrial stromal cells (14). To further examine SUMO-dependent regulation of PR-B transcriptional activity in a breast cancer cell model, we generated T47D human breast cancer cell lines stably expressing either wt PR-B or nonsumoylated K388R PR-B (Fig. 2A). Western blotting indicated that both wt PR-B and K388R PR-B underwent ligand binding and subsequent phosphorylation events, as demonstrated by slight gel-mobility up-shift in the presence of R5020. K388R PR-B also underwent rapid ligand-dependent down-regulation (6 h). Densitometric ratios of PR to β-actin loading control demonstrated comparable levels of PR expression in the cell lines; the values are 1.00, 1.27, and 0.91 for wt PR-B, KR-7, and KR-27, respectively. Ligand-induced transcriptional activity of K388R PR-B was then compared with wt PR-B using a PRE-luciferase reporter assay. T47D-Y (PR-null) cells (28) containing stably integrated wt PR-B (T47D-YB) or K388R PR-B (clones 7, 24, and 28) were transiently transfected with 2XPRE-luciferase reporter and Renilla control plasmids and treated for 24 h with R5020 (10–8 M) (Fig. 2B). As expected, in the presence of R5020, cells stably expressing wt PR-B exhibited a greater than 20-fold increase in PR transcriptional activity relative to vehicle-treated or vector-containing controls. Notably, K388R PR-B exhibited an average 80-fold increase in transcriptional activity relative to vehicle controls. In multiple experiments using separately derived clones, K388R PR-B was consistently at least 10-fold more active than wt PR-B in the presence of progestin, consistent with a similar study using HeLa cells (7). These data demonstrate that modification of PR by SUMO-1 attachment negatively regulates PR-B transcriptional activity in breast cancer cells.
|
). As expected, modest levels of HB-EGF transcripts were measured in R5020-treated cells expressing wt PR-B, whereas transcripts were undetectable in R5020-treated cells stably expressing S294A PR-B (22). In contrast, R5020 treatment of cells stably expressing K388R PR-B resulted in elevated HB-EGF transcript levels relative to wt PR-B control cells. MUC1, a PR target gene containing few functional PRE half-sites (32), was regulated by R5020 in cells expressing wt, S294A, and K338R PR-B (Fig. 3A). Real-time quantitative PCR was used to confirm RT-PCR data (Fig. 3B), which demonstrated a 5-fold increase in HB-EGF levels in K388R PR-B-expressing cells compared with wt-expressing cells. These data suggest that a subset of PRE-containing promoters is sensitive to hormonally regulated changes in PR phosphorylation and sumoylation status.
|
KXE) was first identified as a transcriptional "synergy control motif" in GR (30), before it was characterized as the site of SUMO attachment (19). We hypothesized that reversible SUMO attachment may regulate aspects of receptor cross talk with other signaling pathways and/or hormone sensitivity. We tested whether nonsumoylated PR-B receptors were sensitive to low hormone concentrations. As above, a PRE-luciferase reporter assay was performed with T47D cells stably expressing either wt PR-B or K388R PR-B, but cells were instead exposed to a concentration curve of progestin (Fig. 3C). At low hormone concentration (10–14–10–11 M), wt PR-B is transcriptionally inactive; weak activity is visible at 10–10 M R5020. In contrast, K388R PR-B exhibited significantly elevated transcriptional activity in response to 10–11 M R5020. Multiple K388R PR-B clones were examined at the 10–11 M R5020 concentration consistently producing a greater than 20-fold increase in transcriptional activity (Fig. 3C, inset). These data suggest that undersumoylated PRs are markedly hypersensitive to subphysiological progestin concentrations, a phenomenon similar to that observed with hyperphosphorylated PR in the presence of activated growth factor pathways (22).
Long-term exposure of breast cancer cells to progestins has been shown to promote changes in proliferation (33) and increased survival (34). The transcriptional activity of PR-B is required for progestin-induced, anchorage-independent growth of T47D cells (35). The above transcriptional data (Fig. 3C) suggest that relative to wt PR-B, K388R PR-B may display increased proliferation in response to subphysiological R5020 concentrations. To examine this possibility, T47D cells stably expressing wt PR-B, S294A PR-B, or K388R PR-B were plated in soft agar containing either vehicle control, 10–11 M R5020, or 10–8 M R5020. Colonies were counted 21 d later (Fig. 3D). Cells stably expressing wt PR-B displayed a significant increase in colony formation in the presence of 10–8 M R5020, but remained unresponsive to low hormone levels. K388R PR-B expressing cells exhibited higher basal ligand-independent colony formation compared with wt PR-B-expressing cells. Notably, colony formation increased in response to 10–11 M R5020, consistent with the hypersensitive transcriptional responsiveness of this mutant (Fig. 3D). Breast cancer cells stably expressing S294A PR-B failed to respond to R5020, reflective of their inability to activate transcription via select promoters (6). These data suggest that SUMO-1 modification of PR may serve to protect cells from illicit proliferation by dampening ligand-induced transcriptional activation, whereas undersumoylated PR may function to promote increased cell survival and proliferation of breast cancer cells, even in the presence of low levels of steroid hormone.
Mutation of PR Lys388 Rescues Phospho-Mutant S294A PR-B
We have shown previously that phosphorylation of PR-B Ser294 mediates transcriptional synergy in the presence of progestins and agents that activate MAPKs (6), although the mechanism was unknown. We speculated that phosphorylation of PR-B on Ser294 may sensitize receptors to ligand via negative regulation of SUMO attachment at PR Lys388. S294A PR-B phospho-mutant receptors are transcriptionally impaired and resistant to ligand-induced ubiquitination and degradation (5, 6), phenotypes consistent with increased sumoylation of substrate proteins. To measure the degree of S294A PR-B sumoylation relative to wt PR-B, we performed SUMO assays. Cos-1 cells were transiently transfected with wt PR-B or either of the two mutants, K388R or S294A PR-B, in combination with EGFP-SUMO-1 (Fig. 4A). Notably, a much larger percentage of S294A PR-B mutant receptors were sumoylated in response to R5020 relative to wt PR-B. However, due to the greatly increased stability of S294A PR-B, more total PR was also available for sumoylation.
|
). As expected, in the presence of R5020, S294A PR-B was transcriptionally impaired relative to wt PR-B, whereas K388R PR-B exhibited greatly elevated transcriptional activity. Remarkably, blockade of PR Lys388 sumoylation in the context of the S294A phospho-mutant receptor (KRSA PR-B) completely rescued the transcriptional defect of the S294A PR-B mutant. We next generated T47D-Y (PR-null) breast cancer cells stably expressing KRSA PR-B mutant receptors and examined their ligand-dependent down-regulation and transcriptional activity relative to wt PR-B and single-mutant receptors. Western blot analysis demonstrated that KRSA PR-B bound to ligand and underwent global phosphorylation, as indicated by gel-mobility up-shift (Fig. 4C). Notably, KRSA PR-B also underwent ligand-dependent down-regulation with a time course similar to that of wt PR-B (18 h). These studies suggest that liganded mutant S294A is resistant to degradation (Ref. 5 and see Fig. 7B) because it is heavily sumoylated. Inhibition of SUMO-1 attachment restores rapid ligand-dependent PR-B turnover. We also tested the transcriptional activity of KRSA PR-B relative to wt PR-B stably expressed in breast cancer cells transiently transfected with 2XPRE-luciferase and treated with R5020 for 24 h (Fig. 4D). The ligand-activated KRSA PR-B mutant exhibited heightened transcriptional activity (75-fold) compared with wt PR-B (5-fold). These data indicate that sumoylation of S294A PR-B receptors is transcriptionally repressive. Inhibition of SUMO-1 attachment restored both S294A PR-B turnover and transcriptional activity; these data are consistent with the notion that PR Ser294 phosphorylation negatively regulates PR sumoylation.
Sentrin (SUMO)-Specific Protease 1 (SENP1) Removes SUMO-1 from wt PR-B and Rescues S294A PR-B Transcriptional Activity and Turnover
Sumoylation of target proteins is a dynamic process wherein conjugation and removal of SUMO occurs rapidly and continuously (24). SENPs are desumoylating enzymes that cleave the isopeptide bond linkage between SUMO and target proteins (36). Liganded PRs are rapidly sumoylated (Fig. 1C), undergo nucleo-cytoplasmic shuttling, and accumulate in the nucleus. SENP1 functions in the nucleus, whereas SENP2 molecules shuttle between the cytoplasm and the nucleus (37).
To examine whether PR-B Ser294 phosphorylation results in PR-B desumoylation via a SENP-mediated mechanism, we assessed the action of SENPs on PR desumoylation and, subsequently, PR transcriptional activity and turnover. PR SUMO assays were performed in Cos-1 cells overexpressing EGFP-SUMO-1, as well as SENP1, SENP2, catalytically inactive mutant SENP1 (SENP1m), or vector control (Fig. 5A). Sumoylated wt PR-B was again clearly visible in R5020-treated vector controls. Notably, PR-B sumoylation was undetectable in cells overexpressing SENP1, whereas SENP2-dependent PR-B desumoylation was somewhat less complete. Catalytically inactive SENP1 (SENP1m) did not block the accumulation of sumoylated PR-B. Rather, sumoylated PR species in SENP1m-expressing cells exceeded control levels, suggesting that SENP1m may act as a dominant-negative SENP for endogenous activities.
|
). In a similar reporter assay, expression of SENP2 (50 ng) only modestly affected PR transcriptional activity relative to SENP1 (Fig. 5C), perhaps because of its location-dependent (i.e. subnuclear) activity (37). SENP1m inhibited PR-B transcriptional activity, again suggesting that this molecule acts as a dominant negative inhibitor of endogenous PR-specific SENPs expressed in breast cancer cells. In addition to PR, SENPs may act upon other members of the transcriptional machinery, such as histone deacetylases (38), to augment transcriptional responses. As a control for the direct effects of SENP1 on PR, the transcriptional activity of the nonsumoylated K388R PR-B was examined in the presence of SENP1 or catalytically inactive SENP1m. In stable PR-expressing cells, SENP1 expression increased wt PR-B transcriptional activity but caused no significant change in the transcriptional activity of K388R PR-B (Fig. 5D).
We then predicted that desumoylation of PR-B by SENP1 would increase the expression of select endogenous PR target genes, such as HB-EGF. To examine HB-EGF transcript levels, we transiently transfected vector control, SENP1, or SENP1m into T47D cells stably expressing wt PR-B. Cells were then treated with R5020 for 6 h before RNA isolation for real-time quantitative PCR (Fig. 5E). Despite low transfection efficiency, SENP1 expression increased HB-EGF transcript levels relative to the vector control and the catalytically inactive SENP1m.
To test the requirement of Ser294 phosphorylation in SENP1-mediated PR desumoylation, we performed SUMO assays in HeLa cells coexpressing wt or S294A PR-B with SENP1 (Fig. 5F). Again, S294A PR-B appeared more sumoylated relative to wt PR-B. Both wt and S294A PR-B were sensitive to SUMO removal in SENP1-expressing cells, suggesting that increased sumoylation of S294A receptors is not likely to be due to an inability to interact with desumoylating machinery.
To test the ability of SENP1 to restore S294A PR transcriptional activity, we next performed PRE-luciferase transcription assays in T47D breast cancer cells transiently expressing either wt or S294A PR-B and cotransfected with SENP1 (Fig. 5G). In the presence of progestin, SENP1-dependent desumoylation of S294A PR-B rescued the transcriptional activity of this mutant receptor, similar to our results with KRSA PR-B (Fig. 4, B and D). Furthermore, desumoylation by SENP1 increased the ligand-dependent down-regulation of the highly stable S294A PR-B mutant (see Fig. 7B) relative to SENP1m (Fig. 5H). These results demonstrate that PR sumoylation is favored by prevention of PR-B phosphorylation at Ser294, which in turn results in stabilized PR with repressed transcriptional activity. Prevention of SUMO attachment (as in KRSA PR-B) or SENP1-mediated removal of SUMO reverses this phospho-mutant PR phenotype.
Phosphorylation of PR-B Negatively Regulates Sumoylation
We and others have shown that Erk 1/2 and cyclin-dependent protein kinase-2 (CDK2) strengthen PR signaling, in part via direct phosphorylation (6, 39, 40). To further explore how phosphorylation events negatively regulate PR sumoylation (Fig. 4), we evaluated the effect of two kinases upstream of PR Ser294 phosphorylation. Cos-1 cells were transfected with the R4F constitutively active mutant (41) of MAPK kinase 1 (Mek1), an activating kinase directly upstream of Erk 1/2, or CDK2-TY, a mutant form of CDK2 that cannot be inactivated by Wee1-induced phosphorylation (42). Western blotting demonstrated that expression of either kinase resulted in robust PR Ser294 phosphorylation (Fig. 6A). Cos-1 cells were next transfected with wt PR-B and EGFP-SUMO-1, and Mek1-R4F or vector control (Fig. 6B). Activation of Erk 1/2 in response to Mek1-R4F expression clearly blocked PR-B sumoylation in the absence or presence of the proteasome inhibitor, MG132. In a similar set of experiments, Cos-1 cells were transfected with wt or S294A PR-B and EGFP-SUMO-1, as well as Cdk2-TY or a vector control, followed by treatment with progestin R5020 (10–8 M) for 1 h (Fig. 6C, upper panel). Similar to our results with Mek1-R4F, Cdk2-TY expression prevented wt PR-B sumoylation in response to R5020 relative to vector control. In contrast, Cdk2-TY expression failed to prevent S294A PR-B sumoylation in response to R5020. To examine the specificity of PR Ser294 phosphorylation in the negative regulation of PR sumoylation, HeLa cells were transfected with S400A PR-B, a receptor mutated at a known Cdk2 phospho-site (40); S400A PR-B behaved like wt PR-B in that it displayed greatly decreased ligand-induced sumoylation upon overexpression of Cdk2-TY (Fig. 6C, lower panel).
|
). Again, progestin treatment induced PR-B sumoylation relative to vehicle control, whereas EGF alone had no such effect. In contrast, PR-B sumoylation was completely blocked in cells treated with EGF and R5020 compared with R5020 alone. Treatment with both EGF and progestin initiated activation of p42/p44 MAPKs and phosphorylation of PR Ser294, as indicated by Western blots for phospho-ERK1/2 and phospho-Ser294 (Fig. 6D).
After establishing that EGF functions as an upstream regulator of PR-B sumoylation, we evaluated its effect on gene expression. HB-EGF transcript levels were evaluated in T47D breast cancer cells stably expressing wt, K388R (nonsumoylated), or S294A PR-B treated with EGF and R5020 (Fig. 6E). In cells expressing wt PR-B, HB-EGF transcript levels were increased after exposure to R5020 alone and were further increased after exposure to both EGF and R5020, as measured by RT-PCR. Cells expressing K388R PR-B exhibited increased levels of transcript with R5020 but no further increase with EGF. S294A PR-B expressing cells failed to induce an increase in HB-EGF transcript levels in response to R5020. To assess the ability of EGF to increase PR hormone sensitivity, via phosphorylation and desumoylation, we performed luciferase assays. HeLa cells transiently expressing wt PR-B were pretreated with EGF and increasing concentrations of R5020 (Fig. 6F). EGF increased PR-B transcriptional activity in response to both 10–10 and 10–8 M R5020. Based on these data, we conclude that growth factor activation of kinase pathways rapidly alters the sumoylation state of liganded PR-B (Fig. 6D). Phosphorylated and thus undersumoylated receptors exhibit heightened transcriptional activity (Figs. 3 and 6, E and F).
PR-A and PR-B Are Differentially Phosphorylated, Down-Regulated, and Sumoylated
Posttranslational modifications may contribute to functional differences in PR isoforms (43). To measure Ser294 phosphorylation of PR-B and PR-A isoforms, T47D cells stably expressing either PR-B or PR-A were treated with EGF for a short time course (Fig. 7A). As reported previously (26, 44), PR-B was rapidly phosphorylated on Ser294 in response to ligand binding as well as upon treatment of cells with EGF (Fig. 7A). MAPK activation in response to EGF was measured using phospho-Erk 1/2 antibodies, and total Erk 1/2 levels are shown as a protein loading control (Fig. 7A). Notably, phospho-Ser294 PR-A was weakly visible relative to phospho-Ser294 PR-B when treated with EGF or R5020. PR-A was rapidly phosphorylated on Ser345 in response to progestin compared with PR-B, suggesting that EGF selectively regulates PR-B Ser294 phosphorylation relative to PR-A. Consistent with previous studies (43), we also found that PR-A was markedly underphosphorylated on Ser294 relative to PR-B in T47D cells coexpressing both endogenous PR isoforms (data not shown).
|
). As expected, coexpressed endogenous PR isoforms simultaneously undergo ligand-induced down-regulation (35); in contrast, PR-A expressed alone failed to turn over rapidly relative to PR-B expressed alone. Notably, S294A PR-B mimicked wt PR-A in this regard and was relatively resistant to ligand-induced down-regulation (Fig. 7B and Ref. 5). Based on the increased stability of PR-A relative to PR-B in the presence of ligand and the lack of detectable phosphorylation on PR-A Ser294, we hypothesized that PR-A is more sumoylated than PR-B. PR-A and PR-B expressed individually in HeLa cells are clearly sumoylated upon ligand binding (7). However, PR isoforms are coexpressed developmentally, and simultaneous measurement of receptor isoform sumoylation has not been conducted previously. Notably, PR-B is a more active transcription factor relative to PR-A on PRE-driven and selected endogenous promoters (45).
To examine the relative levels of PR-A and PR-B sumoylation, we performed SUMO assays in HeLa cells expressing PR-A and PR-B, either together or alone. In cells coexpressing both PR isoforms, PR-A and PR-B were clearly sumoylated in response to R5020 (Fig. 7C). However, PR-A consistently appeared to be the predominant sumoylated isoform, with a darker band representing the sumoylated species relative to PR-B in multiple experiments conducted in HeLa (Fig. 7C) and Cos-1 cells (data not shown). Similar results were obtained when either isoform was expressed alone, with PR-A consistently more sumoylated relative to PR-B. Quantification of the relative sumoylation of PR isoforms using the LI-COR infrared system (see Materials and Methods) indicated that PR-A, when expressed alone or with PR-B, was consistently significantly more sumoylated relative to PR-B (Fig. 7C, table).
Resistance of PR-A to Ser294 phosphorylation in response to hormones (Fig. 7A) suggests that PR-A may also be insensitive to negative regulation of SUMO-1 attachment in response to EGF. We therefore performed a SUMO assay in HeLa cells expressing PR-A in the presence and absence of R5020 and EGF (Fig. 7D). PR-A was robustly sumoylated after only 15 min of R5020 treatment. However, in contrast to our results with PR-B (Fig. 6D), PR-A was relatively resistant to EGF-induced blockade of SUMO attachment (Fig. 7D). In addition, EGF treatment did not alter PR-A transcriptional activity in the presence of ligand, as measured by PRE-driven reporters (data not shown). Our data suggest that PR isoform differences in Ser294 phosphorylation confer differential isoform sumoylation and responsiveness to external stimuli upstream of protein kinase effectors.
Resistance of PR-A Ser294 to phosphorylation appears to insulate this isoform from EGF-regulated desumoylation (Fig. 7D). We therefore examined the susceptibility of PR-A to desumoylation by SENP1. SENP1 desumoylated PR-A in Cos-1 cells transiently transfected with PR-A and EGFP-SUMO-1 (Fig. 7E). As described above (Fig. 5), T47D breast cancer cells transiently expressing wt PR-B or PR-A were cotransfected with 2XPRE-luciferase and β-gal reporters and either control vector or SENP1 (Fig. 7F). In the presence of progestin, SENP1 markedly increased PR-A transcriptional activity, which approached the magnitude of PR-B activity in the absence of SENP1. However, during SENP1 overexpression, although its activity was increased, PR-A remained a weaker transcription factor relative to PR-B under similar conditions. Finally, we examined the ability of the nonsumoylated K388R PR-A to respond to low concentrations of ligand in a PRE-luciferase assay. HeLa cells transiently expressing wt PR-A or K388R PR-A were treated with increasing concentrations of progestin R5020 (Fig. 7G). The nonsumoylated mutant PR-A displayed increased transcriptional activity relative to wt PR-A at both 10–10 and 10–8 M R5020, indicating that desumoylated PR-A is hypersensitive to ligand despite a lack of regulation by Ser294 phosphorylation. Collectively, these data suggest that sumoylation of PR-A may account, in part, for its diminished transcriptional activity on selected promoters relative to PR-B. However, other mechanisms also clearly contribute to differential PR-A/PR-B function (46).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
).
|
Kinase activation leading to PR Ser294 phosphorylation acts to inhibit SUMO attachment (Fig. 6) and increase receptor ubiquitination (5), thereby altering both transcriptional activity on select promoters and subsequent turnover of liganded receptors. Nonsumoylated mutant (K388R) and double-mutant (K388R/S294A) receptors are down-regulated in response to ligand binding, and their degradation is blocked by the proteasome inhibitor, MG132 (data not shown), confirming proteasome-dependent loss. The ability of the double mutant to undergo proteasomal degradation indicates that although Ser294 phosphorylation targets PR for efficient ubiquitination (5), it is not absolutely required. PR sumoylation may prevent or retard ubiquitination by sequestration of PR or by interference with ubiquitination or proteasomal machinery (48). Recently, CUEDC2 expression has been shown to regulate ubiquitination of PR-B at Lys388 (49), the primary site of SUMO attachment. PR phosphorylation may enhance CUEDC2/PR interaction, thereby favoring Lys388 ubiquitination over sumoylation. However, K388R PRs clearly turn over (Figs. 2A and 4C), indicating alternate sites of ubiquitin attachment. Interplay between the opposing actions of PR ubiquitination and sumoylation in response to phosphorylation events provides a mechanism for rapid positive and negative regulation of gene expression in response to a continuously varying hormonal milieu.
Importantly, our data suggest that PR Ser294 phosphorylation does not alter the ability of SENPs to recognize and desumoylate PR (Fig. 5). Alternatively, PR phosphorylation may weaken the receptor’s ability to interact with conserved sumoylating enzymes, thereby favoring the undersumoylated state. In this case, phosphorylated PRs may resist modification by UBC9 or PIAS family members, which are predicted to rapidly sumoylate PRs. Down-regulation of PIAS1 has been shown to modulate PR-A sumoylation in the uterus upon cAMP activation over a period of many days (14), whereas PIAS3 may regulate PR-B (15) in breast cancer cells. Although PKA activation by cAMP or forskolin had no effect on PR sumoylation at rapid time points (data not shown), it is possible that PIAS enzyme levels constitute a rate-limiting step in PR sumoylation. During kinase activation the availability of PIAS proteins to interact with PR may be decreased due to changes in PR localization, altered PR conformation (i.e. via phosphorylation), and/or sequestering of PIAS to other target proteins; we are currently investigating these possibilities.
PR Sumoylation Alters Gene Regulation on PRE-Rich Promoters
Sumoylation of transcription factors can modulate target gene promoter selectivity, causing increased expression from some promoters and silencing of others (48). Holmstrom et al. (31) have shown that promoters with multiple identical hormone response elements are regulated in response to sumoylation of cognate transcription factors; sumoylated GR is transcriptionally repressed on a minimal promoter containing two or more tandem GREs, whereas promoters containing a single element are insensitive to GR sumoylation (31). SUMO-mediated transcriptional inhibition has been mapped to a highly conserved region in SUMO1 and 2 (9) that binds to the SUMO-interacting motif (SIM) of coregulatory proteins (50, 51). For example, sumoylated glucocorticoid receptors interact with the SIM of the corepressor Daxx, mediating transcriptional repression on multiple glucocorticoid response elements (50). In the present study, we observed that the HB-EGF promoter is rich in PRE half-sites, and expression of HB-EGF mRNA is highly sensitive to alterations in PR sumoylation, whereas MUC1 expression is insensitive to PR sumoylation/desumoylation (Fig. 3, A and B). These findings suggest that the Kd for receptor-ligand interaction is unaltered by sumoylation (3). The transcriptional synergy control observed here in the presence of multiple sumoylated PR molecules at the HB-EGF promoter region is likely due to recruitment of corepressor (31, 50) and histone deacetylase (52, 53) complexes via their SIMs. The increased availability of corepressor binding surfaces provided by SUMO molecules at promoters containing multiple PREs may shift the corepressor/coactivator balance, thereby altering the EC50 of gene activation (3); we are currently further exploring these mechanisms.
Interestingly, we observed increased soft-agar colony formation in T47D breast cancer cells stably expressing K388R PR-B (Fig. 3D), thus revealing a ligand-independent activity of nonsumoylated PR. Consistent with this finding, PR expression induces changes in gene expression independently of exogenously added ligands (54), and ligand-independent PR transcriptional activity in response to phosphorylation events is well documented (39, 40). Thus, in the absence of ligand, PR sumoylation may serve to repress the inappropriate expression of cell growth and survival genes in mammary epithelial cells. Identification of promoter contexts that confer responsiveness to PR sumoylation/desumoylation will provide further insight into how SRs and growth factors synergize in the regulation of developmental processes as well as during cancer progression in hormonally responsive tissues.
Rapid Signaling Events Favor PR-B Desumoylation
SR family members, including PR, localize to cell membranes, where they are capable of rapidly activating membrane-associated protein kinase pathways. Liganded PR, ER, and AR activate c-Src within 3–5 min, leading to rapid and transient (5–15min) Erk1/2 MAPK activation. Liganded PRs also activate PI3K phosphatidylinositol-3-kinase (55), p38 MAPKs (56), and CDK2 (40). Rapid activation of kinase pathways in response to PR ligand binding provides a mechanism for direct phosphorylation of PRs, perhaps ensuring that this fraction of receptors are undersumoylated and thus acutely and robustly active on selected promoters. The initiation of rapid PR signaling may account for the difficulty in detecting sumoylated PR-B relative to PR-A at early time points; PR-A does not participate in rapid signaling in response to progestins (57, 58) and is highly resistant to Ser294 phosphorylation (Ref. 43 and Fig. 7). These differences in the ability of PR isoforms to activate selected kinase pathways that, in turn, regulate their differential phosphorylation likely translate, in part, to their considerable functional differences (discussed below).
Notably, PR-B Ser294 phosphorylation is differentially regulated in response to progestins relative to EGF (Fig. 7A and Ref. 22), suggesting there are multiple kinase pathways upstream of this site. Phosphorylation of PR Ser294 by different kinases in response to growth factors or steroid hormone inputs may allow for temporal control of gene regulation in response to changes in PR sumoylation status. In this case, receptor sumoylation serves to modulate the transcriptional response of PRs to hormone in magnitude (Fig. 3C), through selective initiation on PRE-rich promoters (Fig. 3, A and B), and ultimately, according to the timing of kinase/phosphatase inputs that also affect PR Ser294.
PR Isoforms Are Differentially Sumoylated
Surprisingly, we identified disparate regulation of PR-A and PR-B sumoylation that provides insight into their distinct transcriptional activities and protein stabilities in the presence of ligand (Ref. 35 and Fig. 7B). PR-A, unlike PR-B, is weakly phosphorylated on Ser294 and is the more stable PR isoform (Refs. 35 and 43 and Fig. 7). Notably, coexpressed PR isoforms turn over with similar kinetics (Fig. 7B), suggesting that hypophosphorylated PR-A is targeted for increased turnover via association with PR-B in heterodimeric complexes. These data suggest that PR in heterodimeric complexes can turn over regardless of the sumoylation state of individual isoforms. Additionally, PRs are stabilized by interaction with chaperone molecules, including heat shock protein 90, independently of Ser294 phosphorylation (5).
PR-A isoforms resistant to Ser294 phosphorylation may confer increased PR-A basal sumoylation and sustain PR-A sumoylation in response to progestins during periods when intracellular kinase pathways are also activated (Fig. 7D). Because PR sumoylation is required for its transrepressive properties (7), increased PR-A sumoylation is likely responsible for the greater PR-A-induced transrepression of ER
relative to PR-B (59). Persistent PR-A sumoylation may also serve to dampen PR-A transcriptional activity relative to PR-B on selected promoters (Fig. 7F) and may insulate this receptor from rapid degradation. Interestingly, PR-A has been shown to repress a larger proportion of progestin-regulated genes relative to PR-B (54); this property of PR-A may be, in part, sumoylation dependent.
Does PR Sumoylation/Desumoylation Contribute to Breast Cancer Biology?
Modulation of reversible PR sumoylation in response to hormonal and growth factor signals provides a mechanism for rapid alteration of hormone responsiveness in breast cancer cells (Fig. 8). PR sumoylation may protect cells from inappropriate proliferation (Fig. 3D) by dampening transcriptional responses. In advanced breast cancers where heightened kinase activity is common (60), phosphorylated and undersumoylated PR-B is predicted to be transcriptionally hyperactive and rapidly turned over (Figs. 2–4). An increased PR-A/PR-B ratio reported in a majority of breast cancers (61) may be, in fact, indicative of hyperactive, yet unstable/undetectable, PR-B. Indeed, PR loss is a marker of activated growth factor signaling in human breast tumors (62). Additionally, low levels of progestins previously considered to be subphysiological may actually provide cells with significant survival (34, 63) and growth (Ref. 22 and Fig. 3D) advantages via activation of unsumoylated PR. Notably, as SUMO attachment to PR is required for transrepression of ER, undersumoylated PR may also contribute to heightened estrogen sensitivity (7). Interestingly, prostate cancers overexpress SENP1 (64), a condition predicted to mediate increased AR activity (38). SENP1 expression is also induced by progestin treatment of T47D breast cancer cells (data not shown).
Progestins increase breast cancer risk when taken as part of hormone replacement therapy (65, 66). Recently, Poole et al. (67) have shown that RU486 prevents mammary tumor development in BRCA1/p53-deficient mice. Our data showing that desumoylated PRs mediate increased breast cancer cell growth in soft agar (Fig. 3D) and RU486-induced PR sumoylation (Fig. 1D) provide further rationale for targeting PR as part of breast cancer treatment. The development of ligands that increase PR sumoylation while antagonizing its transcriptional activity could additionally block ER activity by augmenting SUMO-dependent transrepression (7). Further study of PR sumoylation/desumoylation and PR/ER cross talk in the context of breast cancer progression is clearly warranted (59, 68).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Immunoblotting
For SUMO assays, Cos-1 or HeLa cells were seeded at 7.5 x 105 cells in 100-mm dishes on d 1. Cells were starved in IMEM 5% dextran-coated charcoal-stripped serum (HyClone Laboratories, Inc., Logan, UT) before transfection using FuGENE 6 (Roche) on d 2. Cells were either treated with 10 nM R5020 (Sigma Chemical Co., St. Louis, MO), 100 nM RU486 (Sigma), EGF (Sigma), and/or 10 µM MG132 (Sigma) on d 3 for 24 h or treated on d 4 for 0–2 h. Cells were washed with cold PBS supplemented with 5 mM N-ethylmaleimide (Sigma) and lysed in radioimmune precipitation assay buffer (35) (without β-mercaptoethanol) containing 15 mM N-ethylmaleimide. Lysates containing equal protein levels were separated by SDS-PAGE and transferred to Immobilon-P membrane for chemiluminescence or to Immobilon-FL membrane (Millipore Corp., Bedford, MA) for infrared Western blot analysis.
Membranes were probed with primary antibodies recognizing total PR (MS-298, Lab Vision Corp., Fremont, CA), phospho-Ser294 PR (MS-1332, Lab Vision Corp.), phospho-Ser345 PR (commissioned from Invitrogen, Carlsbad, CA), SUMO-1 (sc-9060; Santa Cruz Biotechnology Inc., Santa Cruz, CA), GFP (sc-8334, Santa Cruz Biotechnology, Inc.), ERK1/2 (9102, Cell Signaling Technology, Danvers, MA), phospho-ERK1/2 (9101, Cells Signaling Technology) and β-actin (A 4700, Sigma). For chemiluminescence blotting, secondary antibodies conjugated to horseradish peroxidase (Bio-Rad Laboratories, Inc., Hercules, CA) and SuperSignal West Pico (Pierce Chemical Co., Rockford, IL) were used. For infrared Western blotting, secondary antibodies labeled with infrared fluorophores were detected at 700 nm (red) or 800 nm (green). To detect transfected SUMO-1, a rabbit polyclonal anti-SUMO-1 antibody was used in conjunction with a goat antirabbit infrared Dye 800CW secondary antibody (green). PR-B was detected using a mouse monoclonal anti-PR antibody followed by a goat antimouse Alexa Fluor 680 secondary antibody (red). Western blots were scanned on two channels simultaneously with the Odyssey infrared imaging system (LI-COR Biosciences). Quantification was determined with Odyssey software; protein bands recognized by the antibodies were selected and corrected for background; relative intensity values were then assigned. Values for the unsumoylated band and sumoylated band were added to obtain total PR-A or PR-B in each lane; the value for the unsumoylated PR isoform band was then divided by the value for total PR isoform to obtain percent sumoylation value. The average of triplicate lanes was determined (±SD), and statistical significance was assessed by using unpaired Student’s t tests.
Transcription Assays
Luciferase assays were performed as described previously (35) in triplicate using the Dual-Luciferase Reporter Assay (Promega). Relative luciferase units were normalized to Renilla or β-galactosidase activity (±SD).
RT-PCR and Real-Time Quantitative PCR
T47D-YB, T47D-YB-S294A, or T47D-YB-K388R cells were plated at 6 x 105 cells per 60-mm dish, starved for 2 d, and treated with R5020 for 6 or 8 h. Total RNA was isolated with Trizol (Invitrogen); cDNA was synthesized from 2 µg of total RNA using Moloney murine leukemia virus-reverse transcriptase and random primers (BD Biosciences, Mountain View, CA). Equal amounts of cDNA were used for each PCR with HB-EGF, Muc-1, β-actin, or SENP1-specific primers. Real-time quantitative PCR was performed as previously described (35) using a Light Cycler (Roche Diagnostics, Indianapolis, IN) and FastStart Master DNA Master SYBR Green 1 kit according to the manufactures protocol. Results in triplicate were normalized to β-actin (±SD). Identification of PRE half-sites (69) in the HB-EGF promoter region was performed using Vector NTI software (Invitrogen) to search the 4000 bp upstream sequence of Exon 1 (ENSG00000113070, www.ensembl.org).
Anchorage-Independent Growth Assay
Soft agar assays were performed as previously described (70). T47D variant cells lines were suspended in 0.45% SeaPlaque agarose in IMEM 5% dextran-coated charcoal containing EtOH, 10–11 M R5020, or 10–8 M R5020 and plated at 1 x 104 cells per well over the bottom agar layer. Each condition was plated in triplicate wells and incubated for 21 d and colonies were counted. The average number of colonies per nine fields is shown (±SD); statistical significance was determined using an unpaired Student’s t test.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Disclosure Statement: The authors have nothing to disclose.
First Published Online August 23, 2007
Abbreviations: AR, Androgen receptor; CDK2, cyclin-dependent protein kinase-2; EGF, epidermal growth factor; EGFP, enhanced green fluorescent protein; ER, estrogen receptor; HB-EGF, heparin-binding EGF-related growth factor; MEK, MAPK kinase; PIAS, protein inhibitor of activated STAT; PRE, progesterone response element; SENP1, sentrin(SUMO)-specific protease 1; SIM, SUMO-interacting motif; SR, steroid hormone receptor; SUMO, small ubiquitin-like modifier; wt, wild type.
Received for publication May 10, 2007. Accepted for publication August 16, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
hinge region regulates its transcriptional activity. Mol Endocrinol 19:2671–2684
B cascade in human breast cancer cells. Endocrinology 146:4917–4925NURSA Molecule Pages Link:
This article has been cited by other articles:
 |
|
 |
|
  G. E Dressing, C. R Hagan, T. P Knutson, A. R Daniel, and C. A Lange Progesterone receptors act as sensors for mitogenic protein kinases in breast cancer models Endocr. Relat. Cancer, June 1, 2009; 16(2): 351 - 361. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Guo and A. D. Sharrocks Extracellular Signal-Regulated Kinase Mitogen-Activated Protein Kinase Signaling Initiates a Dynamic Interplay between Sumoylation and Ubiquitination To Regulate the Activity of the Transcriptional Activator PEA3 Mol. Cell. Biol., June 1, 2009; 29(11): 3204 - 3218. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Abdel-Hafiz, M. L. Dudevoir, and K. B. Horwitz Mechanisms Underlying the Control of Progesterone Receptor Transcriptional Activity by SUMOylation J. Biol. Chem., April 3, 2009; 284(14): 9099 - 9108. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Kaikkonen, T. Jaaskelainen, U. Karvonen, M. M. Rytinki, H. Makkonen, D. Gioeli, B. M. Paschal, and J. J. Palvimo SUMO-Specific Protease 1 (SENP1) Reverses the Hormone-Augmented SUMOylation of Androgen Receptor and Modulates Gene Responses in Prostate Cancer Cells Mol. Endocrinol., March 1, 2009; 23(3): 292 - 307. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W.-H. Yang, J. H. Heaton, H. Brevig, S. Mukherjee, J. A. Iniguez-Lluhi, and G. D. Hammer SUMOylation Inhibits SF-1 Activity by Reducing CDK7-Mediated Serine 203 Phosphorylation Mol. Cell. Biol., February 1, 2009; 29(3): 613 - 625. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Gellersen, M.S. Fernandes, and J.J. Brosens Non-genomic progesterone actions in female reproduction Hum. Reprod. Update, January 1, 2009; 15(1): 119 - 138. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Meednu, H. Hoops, S. D'Silva, L. Pogorzala, S. Wood, D. Farkas, M. Sorrentino, E. Sia, P. Meluh, and R. K. Miller The Spindle Positioning Protein Kar9p Interacts With the Sumoylation Machinery in Saccharomyces cerevisiae Genetics, December 1, 2008; 180(4): 2033 - 2055. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. R. Levin Rapid signaling by steroid receptors Am J Physiol Regulatory Integrative Comp Physiol, November 1, 2008; 295(5): R1425 - R1430. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. R. Holmstrom, S. Chupreta, A. Y.-L. So, and J. A. Iniguez-Lluhi SUMO-Mediated Inhibition of Glucocorticoid Receptor Synergistic Activity Depends on Stable Assembly at the Promoter But Not on DAXX Mol. Endocrinol., September 1, 2008; 22(9): 2061 - 2075. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H.-S. Byun, G.-S. Lee, B.-M. Lee, S.-H. Hyun, K.-C. Choi, and E.-B. Jeung Implantation-Related Expression of Epidermal Growth Factor Family Molecules and Their Regulation by Progesterone in the Pregnant Rat Reproductive Sciences, September 1, 2008; 15(7): 678 - 689. [Abstract] [PDF]
|
|
|
|
|
|
E. J. Faivre, A. R. Daniel, C. J. Hillard, and C. A. Lange Progesterone Receptor Rapid Signaling Mediates Serine 345 Phosphorylation and Tethering to Specificity Protein 1 Transcription Factors Mol. Endocrinol., April 1, 2008; 22(4): 823 - 837. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
