Identification of Phosphorylation Sites Unique to the B Form of Human Progesterone Receptor

The human progesterone receptor (PR), a member of the steroidlthyroid receptor superfamily of ligand-activated transcription factors, is expressed in most tissues as two forms that exhibit differential transcriptional activation potentials, full-length PR-B and NH,-terminally truncated PR-A. In human breast cancer cells (T47D) both forms of PR are constitutively phosphorylated but phosphorylation is increased in response to hormone treatment, suggesting that this modification has a role in regulating the activation state of the receptor. To more directly define the functional role of phosphorylation in the action of A and B receptors requires knowledge of the phosphorylated amino acid residues and the protein kinase(s) involved. Toward this end we have developed a strategy that combines isolation of PR phosphotryptic peptides by reverse phase high performance liquid chromatography, secondary analytical protease digestion, manual Edman degradation, and release of 32P that resulted in identification of two major phosphorylation sites, Ser” and Ser16’. Both sites are located in the amino-terminal region unique to PR-B, and one of these sites (Sersl) is encompassed in a casein kinase

hormone and undergo a n increase in phosphorylation upon hormonal stimulation. Treatment of T47D human breast cancer cells with a progestin induces a 2-foId increase in net phosphorylation of human PR (hPR), and a characteristic decrease (or upshift) of receptor mobility on SDS-gel electrophoresis (11). A time course study under steady-state labeling conditions elucidated a two-phase phosphorylation mechanism; a rapid phosphorylation that occurs between 5 and 10 min after addition of hormone, accounting for most of the net increase, followed by the PR upshift that begins at 20 min and requires 40-60 min for completion. Interestingly, the phosphorylation associated with PR upshifts occurs with little additional change in net 32P incorporation (11). Additional data have suggested that this late phase of phosphorylation may be both hormone-and DNAdependent (12,13).
It is becoming increasingly evident that phosphorylation of PR, as well as other steroid receptors, plays a role in regulating the activity of the receptors. Modulators of protein kinases and phosphatases such as 8-bromo-cAMP, okadaic acid, calyculin, vanadate, and epidermal growth factor have been used to assess the role of phosphorylation in regulating the activity of chicken PR (cPR) in transient transfection assays (14). These compounds are capable of activating receptor in the absence of hormone (14). This striking ligand-independent activation has also been reported for several other steroid receptors (15)(16)(17)(18). In contrast to cPR, human PR does not appear to be susceptible to ligand-independent activation by modulators of protein phosphorylation. However, modulators such as CAMP, okadaic acid, and phorbol esters do potentiate hormone-dependent activation of human PR (11). Interestingly, the progesterone receptor antagonist RU486 can be converted to a partial but potent agonist by treating cells with 8-bromo-CAMP (19). Whether this antagonist to agonist switch is the consequence of altered receptor phosphorylation or phosphorylation of another protein involved in PR-mediated transcriptional enhancement remains t o be determined.
There is increasing evidence that PR-A and PR-B have distinct functional properties that are dependent on the cell type and target genes with which they interact. PR-B when expressed alone in HeLa cells was reported to be able to mediate partial agonist activity of RU486 (20, 211, whereas PR-A alone was not capable of doing this. In addition, PR-A has been reported to be capable of exerting dual functional roles; to serve as a positive transcriptional activator or as a repressor (21,221. In cells and target genes where PR-A exhibits no transcriptional activation properties, it has been reported to act as a repressor of PR-B-mediated transcription. Surprisingly, PR-A but not PR-B, can also act as a trans-repressor of glucocorticoid, androgen, and mineralocorticoid receptor-mediated transcription in a cell type and promoter-dependent manner (22). The mechanism for these distinct functional properties of PR-A and PR-B is not known. Clearly, the different NH,-terminal se-quences present in PR-A and PR-B must play a role in directing the distinct activities of the two PR isoforms.
Although phosphorylation is thought to have a regulatory function, the role of specific phosphorylation sites in PR function is still poorly understood. One of our laboratories has identified four phosphorylation sites in cPR (23, 24). All four sites are on serine residues that are common to both cPR, and cPR,, Two sites (SeP7 and Ser530) are hormone-dependent phosphorylation sites. The phosphorylation of human PR is not as well characterized. Initial phosphopeptide mapping analysis by Sheridan et al. (8) showed that hPR phosphorylation is more complex than cPR. They reported that there are at least five constitutive phosphopeptides common to PR-A and PR-B and a sixth one is unique to PR-B. One additional phosphopeptide, and possibly a second, were detected only after hormone treatment. In addition, all phosphorylations are on serine residues. The exact location of hPR phosphorylation sites, however, has not been reported.
In the present study, we have identified two major phosphopeptides which are unique to the B form of hPR and have developed a simple strategy that should be applicable t o other proteins and steroid receptors which requires a minimum amount of 32P-labeled protein to identify phosphorylation sites. Using this strategy, we have identified S e P and S e P as the major PR-B specific phosphorylation sites. In addition, we have performed i n vitro phosphorylation studies with casein kinase I1 (CKII) and purified PR-B showing that CKII preferentially phosphorylates S e P , which is part of a consensus sequence for CKII.

EXPERIMENTAL PROCEDURES
Materials"R5020 and carrier free [32PlH3P0, were obtained from DuPont-NEN. fy3'P1ATP was purchased from ICN (Irvine, VA). Protein A-Sepharose was obtained from Pharmacia Biotech Inc. Tosylphenylalanyl chloromethyl ketone-treated trypsin was obtained from Worthington. Sequencing grade endoproteinase Asp-N and Glu-C were purchased from Boehringer Mannheim. Phenylisothiocyanate and HPLC reagents were obtained from J. T. Baker Inc. Triethylamine, l-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDAC), and sequencing-grade trifluoroacetic acid were obtained from Sigma. Sequelon-AA membranes and Mylar sheets were obtained from Millipore Corp. (Milford, MA). AB-52 is a mouse monoclonal immunoglobulin G (IgG) produced against purified human PR that recognizes both A and B forms of receptor (25). Casein kinase I1 was purchased from Promega (Madison, WI). Minimum essential medium (MEM) was purchased from Irvine (Santa Ana, CAI. Phosphate-free MEM was obtained from Life Technologies, Inc. All other chemicals were reagent grade. Purification of Baculouirus-expressed Human PR-B-Human PR-B used in protein sequencing experiments and for in vitro phosphorylation was produced and purified as a full-length recombinant protein from the baculovirus expression system as described previously (26,27). Purification was by single-step monoclonal antibody affinity chromatography using the B-30 antibody specific for PR-B (25). As judged by single-dimension silver stained SDS-gels and Western blot, PR-B preparations are routinely purified to apparent homogeneity (27). Receptors were bound to hormone (R5020) in vivo during the last 4 h of infection of Spodaptera frugiperda (Sf91 cells with the PR-B expressing recombinant virus. Thus purified PR used in these experiments was bound to hormone. Baculovirus produced hPR as reported previously is functionally indistinguishable from hPR synthesized endogenously in mammalian cells (26). Cell Culture, Metabolic Labeling, a n d Receptor Preparations-T47D human breast cancer cells were cultured and grown for 2 weeks with frequent changes of media as described previously (11). Cells growing in 75-cmZ T flasks (Falcon, Oxnard, CA) were incubated for 24 h prior to harvest with MEM containing 5% fetal calf serum that had been stripped of steroid hormones by treatment with dextran-coated charcoal. For steady-state labeling with [32Plorthophosphate the serumcontaining medium was removed and cells were preincubated in phosphate-free serum-free medium for 1 h at 37 "C. Cells were then incubated for 6 h at 37 "C in phosphate-free MEM containing [32Plorthophosphate (0.83 mCi/ml). Cells were treated with 40 I~M R5020 for times indicated in the figures.
Cells harvested with I mM EDTA in Earle's balanced salt solution were homogenized a t 4 "C in a Teflon-glass Potter-Elvehjem homogenizer in KPFM buffer (50 mM potassium phosphate (pH 7.4), 50 mM sodium fluoride, l mM EDTA, l mM EGTA, and 12 mM monothioglycerol) containing 0.5 M NaCl and a mixture ofproteinase inhibitors as described previously (11). The homogenates were centrifuged at 100,000 x g for 30 min, and the supernatant (whole cell extract) was dialyzed against KPFM to reduce salt concentration before immunoprecipitation.
Immunoprecipitation and Gel Purification of PR-Protein A-Sepharose was prebound with the receptor-specific monoclonal antibody, AB-52 (25) as described previously (11). Receptors in dialyzed whole cell extracts were then incubated with AB-52-coated protein A-Sepharose on an end over end rotator for 4 h at 4 "C. Protein A-Sepharose was washed repeatedly by centrifugation in buffer containing 0.3 M NaCl. Bound receptors were eluted with 2% SDS sample buffer and electrophoresed on 7.0% discontinuous SDS-polyacrylamide gels as described previously (11). 32P-Labeled receptors were detected by autoradiography of wet gels and the gel pieces corresponding to the PR-A and PR-B isoforms were excised and incorporated radioactivity was measured by Cerenkov counting of the gel pieces.
HPLCAnalysis of Zkypsin-digested PR-SDS gel slices containing PR were cut and placed in 1.5-ml microcentrifuge tubes, washed with 50% methanol for 1 h, with H,O for 30 min, and with 50 mM ammonium bicarbonate for 5 min. Twenty pg of trypsin were added to the tubes containing gel slices and 500 pl of 50 mM ammonium bicarbonate and incubated at 37 "C; four additional aliquots of trypsin were added at 1.5-h intervals. Tryptic phosphopeptides were dried and redissolved in 50% formic acid and applied to a Vydac C18 reverse phase column in 0.1% trifluoroacetic acid in water, run at a flow rate of 1 mumin, and eluted with a linear gradient from 0 to 45% acetonitrile over 90 min. 32P-Labeled peptides were identified on-line with a model IC Flo-One radioactive flow detector (Radiomatic Instruments, Inc., Tampa, FL) (23).
Characterization of Dyptic Phosphopeptides by Proteinase Digestion a n d Manual ,7zP Release-Fractions containing tryptic phosphopeptides resolved by reverse phase HPLC were further separated by electrophoresis on a 40% alkaline polyacrylamide gel (28). The gel was dried and autoradiographed. Bands containing tryptic phosphopeptides were excised and eluted with H,O overnight, and dried in a Speedvac. Eluted peptides were subsequently subjected to digestion with the endoproteinases Glu-C and Asp-N which cut, respectively, on the COOH-terminal side of Glu and NHz-terminal side of Asp residues followed by peptide gel electrophoresis or manual Edman degradation. Asp-N digestion was performed using 0.2 pg of Asp-N in 200 pl of 50 mM sodium phosphate buffer, pH 8.0, for 4 h at 37 "C. Glu-C digestion was performed using 1 pg of Glu-C in 200 pl of 25 mM ammonium bicarbonate, pH 7.8, for 8 h at 37 "C.
Manual Edman degradation studies were performed to localize the position of phosphoamino acids in the peptides using the method described by Sullivan and Wong (29). Briefly, the peptide to be analyzed was dissolved in 30 pl of 50% acetonitrile, and spotted on an arylamine-Sequelon disc, which was placed on a Mylar sheet on top of a heating block set at 50 "C. After 5 min, the aqueous solvent was evaporated, and the disc was removed from the heating block. 5 pl of EDAC solution (1 mg in 0.1 M Mes, pH 5.0) was added to the disc to allow the peptide to covalently link to the disc, and the disc was placed at RT for 30 min. The disc was then washed five times with water and five times with trifluoroacetic acid to remove unbound peptide. The disc was then washed three times with methanol, and subjected to Edman degradation. The disc was treated at 50 "C for 10 min with 0.5 ml of coupling reagent (methano1:water:triethylamine: phenylisothiocyanate; 7:1:1:1, v/v). After five washes with 1 ml of methanol, the disc was treated at 50 "C for 6 min with 0.5 ml trifluoroacetic acid to cleave the amino-terminal amino acid. The trifluoroacetic acid solution was placed in a scintillation vial and the disc was washed with 1 ml of trifluoroacetic acid and 42.5% phosphoric acid (9:1, v/v). The wash was combined with the trifluoroacetic acid solution, and the released 32P was determined by Cerenkov counting. At this stage, the disc was either stored in methanol at -20 "C or washed five times with 1 ml of methanol before the next cycle was started.
Preparation and Isolation ofpeptides for Sequencing-Purified baculovirus expressed PR-B (10 pg) was digested with trypsin (0.5 pg). Tryptic peptides were separated by reverse phase HPLC. Fractions with retention times corresponding to 32P-labeled tryptic phosphopeptides from T47D cells were collected, dried in a Speedvac, and sequenced using an automated sequencer (23).
In Vitro Phosphorylation of PR-B with Casein Kinase II-Purified baculovirus expressed PR-B was incubated at 37 "C for 30 min in a buffer containing 50 mM Tris-HC1, pH 7.5, 100 mM NaCl, 12 mM MgCl,, 10 p~ [y-32PlATP (specific activity, 100,000 cpdpmol), and 5 units of casein kinase 11. 32P-Labeled PR-B was separated by SDS-gel electrophoresis and detected by autoradiography. The band containing PR-B was excised and subjected to subsequent analyses consisting of tryptic digestion, HPLC and peptide gel separations, protease digestion, and manual 32P release.

RESULTS
llyptic Phosphopeptide Maps of Human PR-PR-A and PR-B were metabolically labeled to steady-state in T47D cells as described previously (11). Radiolabeled receptors were immunoprecipitated from cell lysates, and the A and B receptor isoforms were separated by SDS-gel electrophoresis and eluted from the gel pieces by digestion with trypsin. The eluted phosphopeptides were then separated by HPLC on a C18 reverse phase column. Shown in Fig. 1 are tryptic phosphopeptide maps of PR-A (lower panel) and PR-B (upper panel) isolated from cells that were treated for 2 h with the synthetic progestin R5020. The phosphopeptide maps therefore represent the fully phosphorylated state of hPR. Peaks with assigned numbers are the abundant phosphopeptides that have been observed most consistently. PR-B contains at least 12 tryptic phosphopeptides, and a minimum of nine phosphopeptides are common to PR-A and PR-B. The three peptides designated 3,4, and 6 are unique

Alkaline Polyacrylamide
Gel Electrophoresis of llyptic Phosphopeptides-Tryptic phosphopeptides of PR-A and PR-B were separated by HPLC and collected in fractions. Those fractions corresponding to each of the major peaks were pooled, dried, and electrophoresed on alkaline 40% acrylamide gels (Fig. 2). While most of the peaks contain only one peptide, we found that several of the peaks (4B, 5A, 7B, 8,9A, and 8,9B) actually contain multiple peptides that were not resolved by reverse phase HPLC. These data suggest that there may be more phosphorylation sites than indicated by the HPLC analysis alone. While most peptides are common to PR-A and PR-B, several are unique to PR-B (3B, 4B, and 6B) and one ( 5 A ) may be unique to PR-A.
Identification of Phosphorylation Sites: Se$', Ser'62-As a first step to identify the major PR-B specific sites, tryptic pep-to PR-B. tide 3 and peptide 6, after isolation from HPLC, were redigested with two other endoproteinase, Asp-N and Glu-C. This was followed by alkaline peptide gel electrophoresis. Fig. 3 shows that the mobility of peptide 3 increased after either Asp-N or Glu-C digestion, indicating that peptide 3 contains both Asp and Glu. The mobility of peptide 6 remained the same after enzyme digestion, indicating the absence of Asp and Glu in peptide 6. These secondary digestions of tryptic phosphopeptides are useful for narrowing the possible sequence locations for peptide 3 and 6.
In addition, to determine the position of the phosphoamino acid in tryptic peptides 3 and 6, manual Edman degradation was performed. Since HPLC peaks 3 and 6 each contain only one phosphopeptide as judged by alkaline gel electrophoresis (Fig. 21, samples from the HPLC analysis were used for this experiment. Fig. 4 shows the cycle at which 32P was released. The majority of the 32P in peptide 3 and peptide 6 was released at cycle 8 and cycle 3, respectively.
All the tryptic peptides in PR-B that have a serine in cycle 8 or cycle 3 and thus have the potential to release 32P in the manner observed are listed in Table I. There are three possible peptides that have serines in position 8 and that also have sites for cleavage by Asp-N and Glu-C. These are the peptides that start with residue 63, 74, and 547. Since this peptide is found only in PR-B, it is likely that it is either the peptide beginning with residue 63 or with residue 74. Peptide 6 contained a phosphoserine in cycle 3 and was not cut by Asp-N or by Glu-C. The peptide starting with residue 160 is the only peptide located in the unique NH, terminus of PR-B which contains a serine in the third position, that also lacks Asp-N and Glu-C digestion sites. Thus, we concluded that Ser'@ is the only possibility for phosphorylation of peptide 6.
To distinguish between the two likely possibilities for peptide 3 and to confirm the deduced identity of peptide 6, we have performed amino acid sequencing of peptide 3 and 6 using purified baculovirus PR-B as the source of receptor. After trypsin digestion, peptides were separated by HPLC and fractions which had the same retention time as either [32Plphosphopeptide 3 or peptide 6 of PR-B from T47D cells were collected, and the amino acid sequence was determined by an automated microsequencer. The sequencing results shown in Table I1 reveal that peptide 3 contains the sequence beginning with residue 74 identifying the phosphorylation site as Sel.8'; peptide 6 contains the sequence beginning with residue 160. A few minor amino acid sequences, either from trypsin or receptor, were found in the peptide 3 and peptide 6 preparations. However, these sequences did not match with the 32P release and endoproteinase digestion results. Thus, the secondary protease digestion, 32P release, and direct amino acid sequencing results collectively provide unambiguous identification that Sera' and Ser16' are phosphorylated in vivo.
In Vitro Phosphorylation of PR-B-We noticed that Sera' is within a casein kinase I1 (CKII) phosphorylation motif; therefore, we decided to test CKII for its ability to phosphorylate PR-B in vitro. As a substrate for in vitro kinase assays we utilized a highly purified preparation of human PR-B expressed as a full-length recombinant protein in a baculovirus system (27). The receptor was phosphorylated as described under "Materials and Methods" and separated by SDS-gel electrophoresis. Phosphorylated PR-B was detected by autoradiography and quantified by counting the radioactivity in the band containing labeled PR-B. The stoichiometry of receptor phosphorylation (>25%) was calculated based on the total amount of PR-B used and moles of phosphate incorporated. Fig. 5 shows that PR-B was readily phosphorylated only in the presence of CKII, indicating that there is no endogenous kinase activity  Peptides 3 and 6 were covalently coupled to arylamine membrane discs using carbodiimide and subjected to manual Edman degradation. The radioactivity in the released amino acid was determined after each cycle using a scintillation counter. The background count (24 -c 4, n > 10) was not subtracted from all the counts. The cycle containing the released "Pis the cycle containing the phosphoamino acid. associated with purified PR-B, and that it is an excellent substrate for CKII. The tryptic phosphopeptide map of the in vitro phosphorylated PR-B shown in Fig. 6 reveals one major "P peak which corresponds to peptide 3 isolated from in vivo phosphorylated PR. The fraction was collected and dried, digested with Asp-N and Glu-C and run on a peptide gel. The mobility of this peptide treated or untreated with endoproteinase was the same as that of peptide 3 (data not shown). To further confirm the identity of this single major phosphopeptide, we performed  Fig. 7, 32P was released in cycle 8 confirming that Ser*' was preferentially phosphorylated by CKII in vitro. In contrast, purified PR-A was not phosphorylated in vitro by CKII (data not shown), further confirming the preference of this enzyme for the PR-B specific Ser". DISCUSSION We have identified two PR-B-specific phosphorylation sites, Ser" and SerI6* located within the 164-amino acid amino terminus of PR-B. We have also found that casein kinase I1 preferentially phosphorylates SerR1 in vitro. Analysis of phosphopeptide maps of PR from hormone-treated T47D cells was previously reported by Sheridan (8) who indicated that there might be at least five common phosphopeptides between PR-A and PR-B with a single site unique to PR-B in the absence of hormone and one or two more additional ones after hormone treatment. Our study has shown that hPR phosphorylation may be more complex than initially reported. We find a t least 12 phosphopeptides including the B-specific peptides. The difference in the total number of peptides may be due to our use of an on-line radioactive flow detector, which gives higher resolution than counting individual fractions after HPLC.

2 3 4 6 B A B A B A B A B A B A B B B B -T --m n -n ' A B "
Analysis of phosphorylation of steroid receptors has been hampered by both the complexity of the phosphorylation (30, 31) and the low abundance of receptor both of which make the use of conventional protein chemistry techniques to identify the sites both very expensive and difficult. Phosphorylation sites in chicken progesterone receptors, isolated from "'P-labeled oviduct tissue minces (23) and in mouse glucocorticoid receptors (32) overexpressed in Chinese hamster ovary cells, have been identified by HPLC isolation of phosphopeptides followed by amino acid sequencing. Some of the phosphorylation sites in Peptide 6 ( I Amino acid position within the PR-B. -PR-B 1 2 FIG. 5. In vitro phosphorylation of PR-B by casein kinase 11. Phosphorylation of PR-B was done under the conditions described in the text of Methods. After incubating at 37 "C for 30 min, the reaction was terminated by addition of sample buffer, heated at 90 "C for 5 min, separated by SDS-gel electrophoresis, and autoradiographed. Lane 1, reaction without casein kinase 11; lune 2, reaction with casein kinase 11. Position of PR-B is indicated. the estrogen receptor have been identified by site directed mutagenesis. Regions of estrogen receptor were deleted and phosphorylation measured to locate regions containing phosphorylation sites. Potential candidate serine residues within those regions were mutated and confirmation that the substitutions actually result in a loss of expected phosphopeptide was done by phosphopeptide mapping. The limitation of this approach is that deletion of a region and/or mutation of an amino acid may alter the structure of the protein in a way that results in reduced phosphorylation a t a distal site. As reported for the vitamin D receptor (331, mutation of an authentic site can in fact result in phosphorylation of an alternate site. The approach described in this paper permits direct, unambiguous identification of phosphorylation sites in wild type, endogenously ex- The elution time of the major peak corresponds to that of peptide 3. The phoresis, digested with trypsin, and separated by reverse phase HPLC. elution time differs somewhat from that in Fig. 1 but was confirmed by comparison with known phosphopeptides. Peptide CKII was subjected to nine cycles of manual n2P release using the procedure described under "Materials and Methods." pressed receptors. Use of secondary protease digestion of tryptic phosphopeptides isolated from HPLC as well as release of 32P by manual Edman degradation, enabled us to identify one of the major phosphorylation sites located in the unique NH, terminus of PR-B (Serlfi2) and narrowed the possibilities for the second (Ser") to two sites. Amino acid sequencing of corresponding tryptic peptides isolated from baculovirus expressed PR-B provided confirmation of SerI6, and identified Ser"' as the second site. However, if the carrier protein were unavailable we could have performed manual Edman degradation on the Asp-N digested peptide 3 to distinguish between the two remaining possibilities. Phosphopeptide mapping has shown that baculovirus expressed hPR is correctly phosphorylated on all sites but one (peptide 9)2 and thus is a suitable carrier protein for identification of all other sites by amino acid sequencing of phosphopeptides, including peptides 3 and 6 in this study. Hilliard et al. (33) successfully used vitamin D receptor isolated from a yeast expression system as a carrier protein for amino acid sequencing to identify a phosphorylation site in the human vitamin D receptor, and chicken PR expressed in yeast was found to be correctly phosphorylated on all of the sites that have been identified in the endogenously expressed receptor (34). Thus, correct phosphorylation of steroid receptors expressed as a recombinant protein in heterologous eukaryotic systems, such as baculovirus or yeast, may be more the rule than the exception. Therefore, the approach developed in our C. A. Beck, Y. Zhang, N. L. Weigel, and D. P. Edwards, manuscript in preparation. study of using endogenously expressed 32P-labeled protein as a tracer combined with unlabeled carrier protein purified from an overexpression system for amino acid sequencing may be a generally effective means to identify phosphorylation sites on steroid receptors and other proteins.
The identification of the major phosphorylation sites located in the unique NH, terminus of hPR-B is a potentially important finding that may help to explain the distinct functional properties of the two hPR isoforms that appear to be dependent upon cell type-specific factors and on target promoter context (20-22). Of particular interest is the recent discovery that PR-A can under certain circumstances act as a repressor of transcription mediated by PR-B as well as other steroid receptors in the closely related glucocorticoid receptor subfamily (22). The mechanism of the repressor activity of PR-A is unknown. By performing site-directed mutagenesis of Sel-8' and S e F , it will now be possible to determine t o what extent phosphorylation of these NH,-terminal sequences in PR-B contribute to the distinctly different biological activities of the two PR isoforms. Is it possible that PR-B lacking phosphorylation of Ser'l and will behave in a manner more closely resembling that of PR-A than wild type PR-B? It is also worth noting how the phosphorylation states of cPR and hPR differ. Although cPR is expressed as an A and B isoform, no PR-B-specific phosphorylation sites have been detected. As shown here, hPR phosphorylation is much more complex exhibiting as many as 9-12 sites compared t o 4 sites in cPR. Additionally, all hPR phosphorylation sites appear to be located in the ME% region NH,-terminal to the DNA binding domain (8); whereas a major hormonedependent site that is involved in modulation of cPR function is located in the COOH terminus between the DNA and steroid binding domain^.^ This illustrates the importance of phosphorylation site identification for individual steroid receptors and that one may not be able to extrapolate phosphorylation site data between even closely related receptors such as cPR and hPR.
A computer search of hPR revealed that Ser" is in one of 11 potential CKII sites (X-SerPThr-X-X-GldAsp) (35). CKII is a ubiquitous multifunctional enzyme which phosphorylates and regulates a variety of proteins including those involved in the regulation of transcription and translation (36). CKII may also play a role in regulating steroid receptors. It was reported that the thyroid hormone receptor encoded by the chicken c-erbAa gene is phosphorylated at a single site by CKII (37). In addition, a CKII site in human vitamin D receptor has been identified as a phosphorylation site both in vivo and in vitro (33,38). We therefore tested the ability of CKII to phosphorylate PR in vitro, finding that CKII preferentially phosphorylated Sera' without phosphorylating any of the other potential CKII sites in hPR. SerI6' is one of 15 potential phosphorylation sites in hPR that fit the consensus motif (X-SerPThr-Pro-X) of the prolinedirected kinases. The SA"P motifs are largely located in the amino-terminal region of PR, the region considered important for interaction with other transcription factors. Interestingly, five Ser-Pro sites are located in the B specific amino terminus of PR. It is possible that these S/l"P motifs play a role in PR interaction with other transcriptional factors and B specific functions. It is intriguing to note that many of the identified phosphorylation sites in steroid receptors such as chicken PR, human estrogen receptor, and mouse glucocorticoid receptor also contain Ser-Pro motifs which are part of the consensus sequences for mitogen-activated protein (  cyclin-dependent kinases. M A P kinases, present in both cytoplasm and nucleus, are important intermediates in signal transduction pathways that are initiated by many types of cell surface receptors. There is evidence that MAP kinases play a key role in the transduction of signals through both protein kinases and protein phosphatases. Whether MAP kinases regulate the activity of PR remains to be established. Ser-Pro is also a minimum consensus for cyclin-dependent kinases; cyclins and cyclin-dependent kinases are key regulators of cell cycle progression in eukaryotic cells. There is also evidence that progesterone receptor regulates expression of cyclin genes in T47D cells (39,40). The data suggest that the activity of human PR may be regulated by different kinases that are actively involved in either signal transduction or cell cycle regulation and that, in turn, PR may regulate proteins involved in cell cycle control.