INTRODUCTION

Members of the steroid hormone receptor superfamily are ligand-regulated transcription factors that regulate multiple biological processes ranging from reproduction to behavior (1, 2, 3, 4, 5, 6, 7). The superfamily includes the classical steroid receptors (such as receptors for progesterone, estrogen, glucocorticoid, mineralocorticoid, and androgen), nonsteroid receptors (such as receptors for thyroid, retinoic acid, and vitamin D), and orphan receptors whose ligands have not been identified (e.g. Nur77) or whose ligand binding property has been lost (e.g. thyroid receptor 2) (8). Most of the steroid receptors are phosphoproteins, and the importance of phosphorylation in the transcriptional activation of the receptors has been suggested by studies using kinase activators, phosphatase inhibitors, growth factors, and neurotransmitters to modulate receptor activity (9, 10, 11). Studies directly addressing the significance of the individual phosphorylation sites are limited (12, 13, 14, 15, 16, 17).

Chicken progesterone receptor (cPR)¹ is one of the earliest classical steroid receptors to be identified and studied. Like other steroid receptors, cPR contains a highly conserved DNA binding domain composed of two zinc fingers, a carboxyl-termi-nal hormone binding domain, and an amino-terminal domain (or the A/B region), which is the most variable region among the members of the steroid receptor superfamily (Fig. 1). Both the amino-terminal domain and the hormone binding domain contain autonomous activation functions (AF-1 and AF-2, respectively), which regulate the transcription of target genes in a cell- and promoter-specific manner (6, 18). The hinge region between the DNA and the hormone binding domains contains a typical nuclear localization signal and has been shown to be involved in interaction with heat shock proteins (19, 20).

cPR is naturally expressed as two forms, cPR_B and cPR_A which lacks the amino-terminal 128 amino acids of cPR_B (Fig. 1). Both forms are phosphorylated (21, 22). The level of phosphorylation is significantly enhanced in response to progesterone treatment, resulting in reduced mobility of the receptor on SDS-PAGE (21). This mobility change can be reversed by phosphatase treatment of the receptor before gel electrophoresis, confirming that it is caused by phosphorylation of the receptor (21).

Four phosphorylation sites have been identified in both cPR_A and cPR_B (Fig. 1) (21, 23). These sites were identified by purifying the receptor from oviduct tissue minces labeled with [³²P]orthophosphate, isolating individual phosphopeptides after specific protease digestion and directly sequencing the isolated phosphopeptides (21, 23). All four sites are within Ser-Pro motifs, and they account for all the Ser-Pro motifs in cPR. Among the four sites, Ser²¹¹ and Ser²⁶⁰ in the amino-terminal domain are basally phosphorylated, and the level of phosphorylation is enhanced in response to progesterone treatment (21). Ser³⁶⁷ and Ser⁵³⁰ are phosphorylated primarily in response to progesterone stimulation (21, 23).

A number of roles have been suggested for phosphorylation in the regulation of steroid receptor function. These include regulation of hormone binding (24, 25, 26), DNA binding (27, 28, 29, 30, 31), transcriptional activity (12, 13), and receptor stability (32). We have shown previously that mutation of Ser⁵³⁰ to alanine, a nonphosphorylatable amino acid, resulted in reduced transcriptional activity of cPR_A at low hormone concentrations but did not affect the maximal activity of the receptor at saturating levels of hormone, suggesting that the phosphorylation at Ser⁵³⁰ influences the response of the receptor to its ligand (12). The decreased sensitivity of the mutant receptor toward the ligand is not due to a decrease in hormone binding affinity, leading to the hypothesis that Ser⁵³⁰ phosphorylation stabilizes the receptor in its active state, perhaps by preventing its reassociation with heat shock proteins.

In the current study, Ser²¹¹ was mutated to alanine, and the mutant receptor was compared with the wild type both for transcriptional activity and for changes in mobility in SDS-PAGE as a result of progesterone treatment. Our data demonstrate that mutation of Ser²¹¹ to alanine prevents the hormone-induced change in mobility. In addition, the mutation reduces the transcriptional activity of cPR_A in a cell- and reporter-specific manner.

EXPERIMENTAL PROCEDURES

All cell culture reagents were purchased from Life Technologies, Inc. except for Nutridoma SR, which is from Boehringer Mannheim. [³H]Chloramphenicol is from DuPont NEN, and N-butyryl coenzyme A is from Pharmacia Biotech Inc. The T7-Gen in vitro mutagenesis kit and Sequenase version 2.0 sequencing kit are from U. S. Biochemical Corp. The oligonucleotides used in the mutagenesis and sequencing were synthesized by GenoSys (The Woodlands, TX). Polybrene and 2,6,10,14-tetra-methyl-pentadecane were purchased from Sigma, and Xylene was purchased from Fisher. Monoclonal antibody to the chicken progesterone receptor (PR22) was kindly provided by Dr. David Toft. All other chemicals were of reagent grade.

Ser²¹¹ was mutated to alanine using a conventional, non-polymerase chain reaction mutagenesis strategy as described in previous studies (12), and the mutation was confirmed by direct sequencing using a Sequenase version 2.0 sequencing kit. The subcloning of the mutant and the wild type cPR_A into the expression vector, p91023B, has been fully described (12). GRE₂E1bCAT (provided by Dr. John Cidlowski) is a simple promoter-based reporter composed of two progesterone/glucocorticoid response elements (PRE/GREs), the TATA box from the adenovirus Elb gene, and the cDNA sequence for chloramphenicol acetyltransferase (CAT) (33). PREtkCAT is a complex promoter-driven reporter that is composed of two PRE/GREs followed by the promoter from the thymidine kinase gene and the cDNA for chloramphenicol acetyltransferase (34).

The conditions for cell culture and transfection have been described previously (11, 12). Cell lines were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with fetal bovine serum and antibiotics (penicillin and streptomycin). 24 h before transfection, cells were plated in 10-cm dishes at a density of 1 × 10⁶ cells per dish in the same medium. 4 h after plating, cells were rinsed with Hanks' balanced salt solution (HBSS), and 10 ml of DMEM supplemented with 1% Nutridoma was added to each dish. The next morning, cells were washed again with HBSS, and 10 ml of serum-free DMEM was added to each dish. The indicated amount of DNA in 0.8 ml of HBS (1 mM MgCl₂, 0.9 mM CaCl₂, 137 mM NaCl, 5 mM KCl, 1 mM Na₂HPO₄·7H₂O, 23 mM HEPES, pH 7.05) was mixed with 5 µl of polybrene (10 mg/ml) and added dropwise to each dish. After incubation with DNA for 4 h, the cells were treated with 25% glycerol in HBSS for 1 min, washed twice with HBSS, and grown in serum-free DMEM containing 1% Nutridoma SR and progesterone at the indicated concentration for an additional 48 h before being harvested.

Chloramphenicol Acetyltransferase and -Galactosidase Assays

Cells were harvested by scraping and lysed by three cycles of freeze-thawing. The protein concentrations were determined using the Bio-Rad protein assay reagent according to the manufacturer's protocols. Equal amounts of protein (usually 100 µg) were then used for both CAT and -galactosidase assays.

The liquid CAT assay has been described (35) and used in our previous studies (11, 12). The samples were heated at 60 °C for 8 min before the addition of the substrates. The -galactosidase activity was determined as follows. Cell lysates containing the same amount of protein used for the CAT assay were mixed with 125 µl of o-nitrophenyl -D-galactopyranoside (4 mg/ml in H₂O), and the reaction mixtures were brought to a final volume of 250 µl with buffer (pH 7.0) containing 60 mM Na₂HPO₄, 40 mM Na₂H₂PO₄, 10 mM KCl, 1 mM MgCl₂, 50 mM -mercaptoethanol. The reactions were incubated at 30 °C until the yellow color became obvious. The reactions were then stopped by adding 125 µl of 1 M Na₂CO₃, and 200 µl of the stopped reaction mixture was removed and transferred to a 96-well microplate; the absorption at 410 nM was determined using a Dynatech microplate reader. Before plotting, the CAT activities were normalized using the A₄₁₀ readings from the corresponding -galactosidase assays.

The whole cell hormone binding assay has been described (36) and used in our previous studies (12). In brief, 1 × 10⁶ COS M6 cells were transfected with 10 µg/dish receptor DNA. 1 day after transfection, cells were incubated with [³H]progesterone at concentrations of 0-4.5 nM for about 16 h. The cells were then washed five times with ice-cold phosphate-buffered saline and extracted with ethanol. The bound [³H]progesterone, which is extracted into the ethanol, was counted in a liquid scintillation counter. Specific binding at each progesterone concentration was calculated by subtracting counts in mock-transfected cells from the counts obtained from cells transfected with receptor DNA.

COS M6 cells transfected with equal amounts (10 µg) of either wild type or Ala²¹¹ mutant receptor DNA were cultured in the Nutridoma SR medium with or without 10 nM progesterone for the indicated times. High salt cell extracts in TESH (10 mM Tris, 1 mM EDTA, 0.1% monothioglycerol, pH 7.7) containing 0.3 M NaCl were then prepared, and the protein concentrations were determined. 15-50 µg of protein from each sample was separated on a SDS-PAGE and transferred to a nitrocellulose membrane. The receptor was detected using the monoclonal antibody, PR22, as described (37).

RESULTS

As shown in Fig. 1, Ser⁸³ of cPR_A corresponds to Ser²¹¹ of cPR_B. To be consistent with the literature, the phosphorylation sites of cPR are numbered according to the sequence of cPR_B, although the mutational analysis was performed on cPR_A in this study. After Ser²¹¹ was mutated to alanine, the mutant receptor was first characterized by immunoblotting analysis. As shown in Fig. 2, the expression level of the mutant and the wild type receptors is comparable in COS cells either in the presence or in the absence of hormone, indicating that the mutation does not affect the stability of the receptor. However, the mutation of Ser²¹¹ to alanine prevents the hormone-induced mobility shift of cPR_A. Neither the small amount of reduced mobility form observed in the absence of hormone (Fig. 2, lane 1) nor the hormone-induced reduction in mobility (Fig. 2, lane 2) observed for the wild type receptor is detected in the cells transfected with the Ala²¹¹ mutant receptor (Fig. 2, lanes 3 and 4). The mobility of the Ala⁵³⁰ mutant is also shown, demonstrating that the mobility change does not require phosphorylation of Ser⁵³⁰ (Fig. 2, lanes 5 and 6).

A

Transcriptional activation is an end point assay. If lack of phosphorylation affects any aspect of receptor activation, the change will be reflected in altered transcriptional activity. To determine whether substituting alanine for Ser²¹¹ affects the transcriptional activity of cPR, various amounts of either the wild type or mutant receptor DNA were cotransfected with 10 µg of GRE₂E1bCAT reporter DNA into HeLa cells. The transcriptional activities of both the mutant and the wild type receptors were measured by assaying the CAT activity in the transfected cells. As shown in Fig. 3, the transcriptional activity of the mutant receptor is significantly less than that of the wild type when the amount of the receptor DNA used in the transfection is less than 1 µg. The degree of reduction varies in different experiments, and the average of the result from five independent experiments showed that the activity of Ala²¹¹ in the responsive range is about 25% of the wild type with GRE₂E1bCAT in HeLa cells.

Alterations in receptor activation up to and including binding to consensus response elements are likely to be cell and promoter independent, whereas changes in interaction with specific transcription factors may be cell and/or promoter dependent. To determine whether the difference in activity between the mutant and the wild type receptors varies with different reporters, the transcriptional activities of both receptors were compared in HeLa cells using both GRE₂E1bCAT and PREtkCAT. As shown in Fig. 4, the activity of the mutant receptor is lower than that of the wild type with both reporters. However, the difference in activity is more substantial with GRE₂E1bCAT (Fig. 4, panel A) than with PREtkCAT (Fig. 4, panel B). In this experiment, the activity of the mutant receptor is 20% of the wild type with GRE₂E1bCAT and 45% with PREtkCAT.

The results described in Figs. 3 and 4 were obtained from experiments using HeLa cells. To further determine whether the decreased activity of the mutant receptor is cell specific, the activities of the mutant and wild type receptors were compared in CV1 cells as shown in Fig. 5. Compared with the activity of the wild type, the mutant receptor showed a significant reduction in CAT activity with GRE₂E1bCAT (Fig. 5, panel A). The activity of the mutant receptor is 41% of the wild type, which is comparable to that observed with PREtkCAT in HeLa cells. However, little difference in activity (80% of the wild type for the Ala²¹¹ mutant) was detected in CV1 cells when PREtkCAT was used in the study (Fig. 5, panel B).

Our previous studies have shown that mutation of Ser⁵³⁰ to alanine reduces the transcriptional activity of the receptor preferentially under conditions when the hormone is limited. To determine whether Ser²¹¹ phosphorylation preferentially modulates the activity of the receptor at low levels of hormone, the transcriptional activities of the Ala²¹¹ receptor and the wild type receptor at 0.1 and 10 nM hormone were compared. As shown in Fig. 5, the degree of reduction in CAT activity for the Ala²¹¹ mutant as compared with that of the wild type at 0.1 nM (low level) progesterone is the same as that at 1 nM (saturating level) with both reporters, suggesting that Ser²¹¹ phosphorylation regulates the activity of the receptor by a mechanism distinct from Ser⁵³⁰ phosphorylation.

To further characterize the Ala²¹¹ mutant receptor, a whole cell hormone binding experiment was performed to determine whether mutation of Ser²¹¹ to alanine altered the affinity of cPR_A for progesterone. As shown in Fig. 6, panel A, this assay detected saturable and specific progesterone binding for the Ala²¹¹ mutant receptor. Scatchard analyses of these data (Fig. 6, panel B) yielded a value of 1 nM for the dissociation constant, which is identical to that of the wild type receptor as was determined and reported in our previous studies using the same whole cell hormone binding assay (36). Thus, the hormone binding affinity is not altered by the mutation of Ser²¹¹ to alanine.

To determine whether phosphorylation of the site responsible for the hormone-induced mobility change of the chicken progesterone is a fast or slow process, cells were transfected with the wild type receptor DNA, harvested after being treated with progesterone for different lengths of time, and analyzed by immunoblotting. As shown in Fig. 7, the faster migrating form is the dominant one in untreated cells. The ratio between the amounts of the faster and slower migrating forms remains unchanged in cells treated with progesterone for 30 min. After being treated with progesterone for 1 h, the amounts of the two forms become comparable. At the 2-h point, the ratio between the amounts of the two forms is reversed, and the slower migrating form becomes the dominant one. Thus, the phosphorylation resulting in the mobility change in SDS-PAGE occurs rather slowly.

A

DISCUSSION

These studies demonstrate that phosphorylation of Ser²¹¹ is important for the overall transcriptional activity of cPR_A and for the change in mobility in SDS-PAGE. The original phosphorylation analyses performed in oviduct tissue minces by immunopurification of receptor in the presence of phosphatase inhibitors showed that this site was 19% phosphorylated in the absence of hormone and that treatment of the minces with hormone for 1 h increased the phosphorylation to 36% (21). It was assumed initially that phosphorylation of one of the highly hormone-dependent sites, either Ser³⁶⁷ or Ser⁵³⁰, caused the change in mobility on SDS gels because conventional purification of receptor from oviducts that had not been treated with progesterone yielded a single band on SDS gels (38). However, the purification procedure was performed in the absence of phosphatase inhibitors (38), and subsequent studies have shown that omission of the inhibitors allows dephosphorylation to occur during purification. Using the change in mobility as an assessment of Ser²¹¹ phosphorylation, the amount of basal phosphorylation in COS cells appears to be similar to that in oviduct.

The slow time course of the phosphorylation as assessed by changes in mobility in SDS gels (Fig. 7) suggests that there is a limiting factor for this phosphorylation step. Although it is formally possible that this is due to overexpression of receptor in COS cells, this is unlikely since a similarly slow time course has been observed in chickens (data not shown). Phosphorylation studies of PR_A expressed in yeast (Saccharomyces cerevisiae) showed that a cPR mutant that does not bind to DNA was not phosphorylated on either Ser²¹¹ or Ser³⁶⁷, although the phosphorylation of Ser²⁶⁰ and Ser⁵³⁰ still occurred (39). This implies that phosphorylation of Ser²¹¹ requires either DNA binding or a type of nuclear localization that is lost when the receptor cannot bind to DNA. Interestingly, our previous studies using the same transient transfection procedure have demonstrated that the transcriptional activity of cPR_A can be detected 6 h after progesterone treatment but not at 2 or 4 h (9). This suggests that Ser²¹¹ phosphorylation occurs prior to receptor activation, consistent with the idea that Ser²¹¹ phosphorylation is important for the transcriptional activity of the receptor.

Studies of human progesterone receptor reveal that there is an analogous slow phosphorylation that results in altered mobility on SDS gels (40). This change in mobility has been correlated with phosphorylation of Ser³⁴⁵ (41) and does not occur when the DNA binding domain is mutated or when the cells are treated with the antiprogestin ZK98299 (42). Despite these similarities, these two phosphorylation sites do not appear to be conserved either with respect to adjacent amino acid sequence or with respect to the location of other phosphorylation sites in the receptor. The role of the phosphorylation at this site in human progesterone receptor function has not yet been determined.

Phosphorylation of Ser²¹¹ affects neither the hormone binding affinity nor the expression level of cPR_A. Moreover, the reduction in activity relative to wild type is independent of hormone concentration. In previous studies, mutation of Ser⁵³⁰ to alanine was shown to reduce the ability of the receptor to respond to low levels of hormone despite a lack of change in hormone binding affinity (12). The activity of the Ala²¹¹ mutant is reduced compared to wild type regardless of the cell type or promoter tested. However, the magnitude of the decrease in activity is cell and promoter dependent, suggesting that the region surrounding Ser²¹¹ plays a role in interaction with other proteins and that the complement of proteins in the two cell types differs. It should be noted that the degree of reduction in transcriptional activity of the mutant receptor varies somewhat from experiment to experiment, presumably depending on the condition of the cultured cells and the levels of the interacting factors expressed in these cells. However, the cell- and promoter-specific effect of the Ser²¹¹ phosphorylation on the activity of the receptor was consistently observed in multiple experiments, including parallel experiments performed with two reporters analyzed in the same cells or with both CV1 and HeLa cells plated and cultured under identical conditions.

Our results are consistent with the previous findings that the two activation regions in progesterone receptor and in other steroid receptors contribute differentially to the overall activation depending upon the cell type and promoters examined (18, 43). Interestingly, mutation of Ser¹¹⁸, a phosphorylation site in the amino terminus of the estrogen receptor, displays a similar phenotype in that the reduction in activity is cell and promoter specific (13). A more recent study has identified Ser¹¹⁸ as the phosphorylation site responsible for the hormone-dependent change in mobility of estrogen receptor in SDS-PAGE (44). It is possible that these changes in mobility reflect a significant change in conformation that alters receptor function in both progesterone receptors and estrogen receptors, although this remains to be determined.

To date, little is known about the specific kinases that phosphorylate steroid receptors. Human estrogen receptor was recently reported to be phosphorylated by mitogen-activated protein kinase at Ser¹¹⁸ (45), and a few receptors have been shown to be phosphorylated by casein kinase II (16, 46, 47, 48). The finding that many of the sites contain the consensus sequence Ser-Pro indicates that cyclin-dependent kinases (49), mitogen-activated protein kinase (50, 51), and/or stress-activated protein kinases (52) phosphorylate and regulate steroid receptors. The location of the phosphorylation sites in the steroid receptors (predominantly in the amino-terminal region (53, 54, 55)) and the finding that some of these sites affect receptor activity in a cell- and promoter-specific manner suggest that phosphorylation serves to aid in activation of specific subsets of genes. Thus, in stages of the cell cycle in which specific cyclin-dependent kinases are active, receptors may efficiently activate genes dependent on phosphorylation of a specific subset of sites. Similarly, activation of a kinase that phosphorylates another subset of sites may result in preferential activation of another set of genes. Although studies to directly test this hypothesis have not yet been done, several pieces of evidence are consistent with this possibility. First, only a subset of the Ser-Pro phosphorylation sites in human progesterone receptor are phosphorylated by Cdk2 (56), suggesting that multiple kinases are involved in receptor phosphorylation. Second, Hsu et al. (57) have shown, using synchronized cells, that the glucocorticoid receptor is transcriptionally inactive in the G₂ phase of the cell cycle and that the phosphorylation pattern is altered relative to that in asynchronous cells. That the loss of activity as a result of mutation of Ser²¹¹ to alanine is profound in some cases but minimal in others suggests that Ser²¹¹ phosphorylation can contribute to differential expression of progesterone-inducible genes under different physiological conditions.