Progesterone Regulation of Progesterone Receptor Membrane Component 1 (PGRMC1) Sumoylation and Transcriptional Activity in Spontaneously Immortalized Granulosa Cells

Abstract

Progesterone (P4) receptor membrane component (PGRMC)1 is detected as a 22-kDa band as well as higher molecular mass bands (>50 kDa) in spontaneously immortalized granulosa cells. That these higher molecular mass bands represent PGRMC1 is supported by the findings that they are not detected when either the primary antibody is omitted or the PGRMC1 antibody is preabsorbed with recombinant PGRMC1. Some but not most of the higher molecular mass bands are due to oligomerization. At least one of the higher molecular mass bands is sumoylated, because PGRMC1 coimmunoprecipitates with small ubiquitin-like modifier protein-1. Moreover, in situ proximity ligation assays reveal a direct interaction between PGRMC1 and small ubiquitin-like modifier protein-1. This interaction is increased by P4. Finally, the higher molecular mass forms of PGRMC1 localize to the nucleus. An analysis of transcription factor activity demonstrates that P4 suppresses T-cell factor/lymphoid enhancer factor (Tcf/Lef) activity through a PGRMC1-dependent mechanism, because treatment with PGRMC1 small interfering RNA depletes PGRMC1 levels and attenuates P4's effects on Tcf/Lef activity. In addition, transfection of a PGRMC1-Flag fusion protein enhances basal Tcf/Lef activity, which is suppressed by P4 treatment. Conversely, transfection of a PGRMC1-Flag protein in which all the sumoylation sites are mutated increases basal Tcf/Lef activity but attenuates P4's ability to suppress Tcf/Lef activity. Therefore, the ability to suppress Tcf/Lef activity is likely an essential part of the mechanism through which P4 activation of PGRMC1 regulates the gene cascades that control granulosa cell function with this action being dependent in part on the sumoylation status of PGRMC1.

Progesterone (P4) plays an essential role in regulating female reproduction by acting at numerous sites, including but not limited to 1) the hypothalamic-pituitary axis to regulate GnRH secretion (1), 2) the ovary to control granulosa cell function (2, 3), and 3) the uterus to prepare it for implantation and the maintenance of pregnancy (4, 5). In the ovary, P4 acts on granulosa cells to inhibit apoptosis and mitosis (3). Similar effects of P4 are also observed in cells derived from granulosa cells [i.e. spontaneously immortalized granulosa cells (SIGC)] (6).

Although it is clear that the nuclear P4 receptors are important mediators of P4's actions (7), their activation cannot account for P4's actions in granulosa cells of small antral follicles or SIGC, because these cells do not express the nuclear P4 receptors (8, 9). Granulosa cells and SIGC (10) do express P4 receptor membrane component (PGRMC)1, with PGRMC1 being localized to the plasma membrane, cytoplasm, and nucleus (11). Moreover, small interfering RNA (siRNA) depletion of PGRMC1 in either SIGC (6) or cell lines derived from human granulosa cells (hGL5 cells) (12) attenuates P4's antiapoptotic activity. These in vitro studies, along with the clinical observations that women with premature ovarian failure (13) and polycystic ovary syndrome (14) have lower levels of PGRMC1, provide compelling evidence that PGRMC1 is an important regulator of ovarian function and mediates P4's antiapoptotic actions in granulosa cells. However, the mechanism through which ligand activation of PGRMC1 regulates granulosa cell function is unknown.

Insight into how PGRMC1 regulates granulosa function is provided by the somewhat perplexing observation that Western blot analysis of PGRMC1 not only detects a specific band in the predicted 22-kDa range but also higher molecular mass bands (11). Interestingly, these higher molecular mass bands localize to the nucleus and are required for P4-induced gene expression (11). One possible explanation for the presence of the higher molecular mass forms of PGRMC1 is that PGRMC1 may undergo some type of posttranslational modification.

One type of posttranslational modification that could both increase the molecular mass of PGRMC1 and influence its nuclear function is sumoylation. Sumoylation involves the rapid and reversible covalent binding of small ubiquitin-like modifier (SUMO) proteins. There are four SUMO proteins that are approximately 10 kDa, with SUMO1 being the most common. Because in silico analysis indicates that PGRMC1 has three sumoylation sites (http://sumosp.biocuckoo.org/online.php) (15) and sumoylation could alter PGRMC1's stability, cytoplasmic-nuclear transport and/or putative transcriptional function (16), the first series of experiments used various biochemical approaches to characterize the nature of the higher molecular mass forms of PGRMC1, particularly because it relates to sumoylation. The second series of studies focused on determining which transcription factors were regulated by P4 activation of PGRMC1 and whether the sumoylation status affects PGRMC1's nuclear action.

Materials and Methods

SIGC culture

All the chemicals used in this study were purchased from Sigma Chemical Co. (St. Louis, MO) except for those specifically mentioned. SIGC used in these studies were derived from rat granulosa cells isolated from preovulatory follicles (17).

SIGC were cultured in DMEM/F12 supplemented with 5% fetal bovine serum (HyClone, Logan, UT), 100 U/ml penicillin G, and 100 μg/ml streptomycin (10, 18).

For studies involving Western blot analysis, 4 × 10⁶ SIGC were plated in 10 ml of medium in 100-mm cell culture dishes and cultured for 24, 72, or 96 h depending on the experimental design. For immunofluorescence analysis and in situ proximity ligation assays (PLA), 4 × 10⁵ SIGC were plated in 2 ml of medium on cover glasses, which were placed in 35-mm culture dishes and then fixed in 4% paraformaldehyde as previously described (19).

Western blot analysis

PGRMC1 was detected in lysates obtained from SIGC by Western blot analysis (11, 18). Unless otherwise stated, cells were lysed in radioimmunoprecipitation assay (RIPA) buffer [50 mm Tris-HCl (pH 7.4), 150 mm NaCl, 1 mm EDTA, 1% Nonidet P-40 (NP-40), and 0.25% sodium deoxycolate], supplemented with protease inhibitors and phosphatase inhibitors. All procedures were conducted on ice. Total amount of protein was determined using the bicinchoninic acid assay kit from Pierce Biotechnology (Rockford, IL), and unless stated otherwise, 10 μg of total protein/lane were used for Western blottings. PGRMC1 was detected by Western blotting using a rabbit polyclonal antibody raised against amino acids 48–130 of human PGRMC1 (catalog no. HPA002877, Prestige Antibodies; Sigma Chemical Co.) at a dilution of 1:750 (11). Primary PGRMC1 antibody was detected using an antirabbit horseradish peroxidase (HRP)-labeled antibody (1:2000) and enhanced chemiluminescence detection (GEHealth, Buckinghamshire, UK). To confirm that equal amounts of total protein were loaded during the procedure, the nitrocellulose membranes were placed in stripping buffer [100 mm 2-mercaptoethanol, 2% sodium dodecyl sulfate (SDS), and 62.5 mm Tris-HCl (pH 6.7)] at 50 C for 30 min and reprobed with a glyceraldehyde-3-phosphate dehydrogenase polyclonal antibody (catalog no. AM4300; Ambion, Inc., Austin, TX) at 1:4000 dilution. Negative controls were conducted by omitting the primary antibody.

To determine whether all of the higher molecular bands specifically bound the PGRMC1 antibody, cells were lysed in RIPA buffer and analyzed by Western blotting using an aliquot of the Sigma PGRMC1 antibody that was preincubated for 1 h at 22 C in 5% powered milk in Tris-buffered saline with or without 5 μg of full-length PGRMC1 recombinant protein (H00010857-P01; Novus Biologicals, Littleton, CO). This protein was used to absorb the PGRMC1 antibody and thereby reduce its ability to detect PGRMC1 when used for Western blottings.

To determine whether the higher molecular mass forms represent oligomers of the approximately 22-kDa band, Western blot analysis was conducted on whole-cell lysates prepared using RIPA buffer with or without 100 mm dithiothreitol (DTT). DTT is a strong reducing agent that cleaves the disulfide bonds that presumably account for the presence of PGRMC1 dimers (20). To assess which form of PGRMC1 was within cellular membranes, whole-cell lysates were prepared with 50 mm Tris-HCl (pH 7.4), 150 mm NaCl, and 1 mm EDTA buffer with or without ionic (sodium deoxycholate) and nonionic detergents (NP-40). Finally, the cellular localization of the different molecular mass forms of PGRMC1 was monitored by conducting Western blottings of nuclear and cytoplasmic proteins that were isolated using the Qproteome nuclear protein kit (QIAGEN, Valencia, CA) as previously described (11). Note that in this protocol the cytoplasmic fraction contains the plasma membranes as well as the membranes of the cytoplasmic organelles.

Sumoylation of PGRMC1

The sumoylation status of PGRMC1 was assessed by three methods. These included coimmunoprecipitation, colocalization, and in situ PLA.

Coimmunoprecipitation

To immunoprecipitate proteins to which SUMO1 was covalently linked, SIGC were cultured for 24 h and lysed in 1 ml of lysis buffer [150 mm NaCl, 1% Triton X-100, and 50 mm Tris-HCl (pH 8)] supplemented with proteases inhibitors and phosphatase inhibitors. After centrifugation, the supernatant of each sample was collected and incubated overnight at 4 C with 2 μg of rabbit monoclonal anti-SUMO1 or rabbit monoclonal IgG XP Isotype Control (catalog nos. 4931 and 3900, respectively; Cell Signaling Technology, Inc., Danvers, MA). The immune complex was then magnetically labeled with 50 μl of protein A (Miltenyi Biotec, Inc., Auburn, CA) and applied to a μColumn (Miltenyi Biotec, Inc.). Unbound material was removed by stringent washing, and the immune complex was eluted in SDS gel loading buffer [50 mm Tris-HCl (pH 6.8), 50 mm DTT, 1% SDS, 1 mm EDTA, 0.005% bromphenol blue, and 10% glycerol]. The eluted immunoprecipitate was analyzed by Western blotting using the rabbit monoclonal anti-SUMO1 (1:1000 dilution) and the rabbit polyclonal anti-PGRMC1 (1:750 dilution; Sigma Chemical Co.).

After incubation with the primary antibodies, the blots were incubated for 1 h at room temperature with mouse antirabbit IgG conformation-specific monoclonal antibody (catalog no. 3678; Cell Signaling Technology, Inc.) at a 1:2000 dilution. This antibody only reacts with the native rabbit IgG (i.e. the primary antibodies used in the Western blotting), whereas it does not recognize the denatured and reduced rabbit IgG (heavy or light chains) that were used for the immunoprecipitation reaction. The mouse antirabbit IgG was revealed using an antimouse HRP-labeled antibody (1:2000) and enhanced chemiluminescence detection (GEHealth). Negative controls were done by omitting the primary antibody.

Immunofluorescence and colocalization studies

Immunofluorescence and double-immunofluorescence (colocalization) staining were conducted to localize PGRMC1 and SUMO1 and to colocalize PGRMC1 with SUMO1. After fixation in 4% paraformaldehyde, cells were permeabilized with 0.1% Triton X-100 in PBS for 7 min and blocked with 5% normal goat serum in PBS for 1 h at room temperature. The samples were incubated overnight at 4 C with the Sigma rabbit polyclonal anti-PGRMC1 antibody (1:250), or with a mixture of the anti-PGRMC1 antibody and a mouse anti-SUMO1 monoclonal antibody (1:50 dilution, mouse anti-GMP1, catalog no. 33-2400; Invitrogen, Carlsbad, CA). Primary antibodies were detected with either an Alexa Fluor 546-labeled antirabbit or an Alexa Fluor 488-labeled antimouse antibodies (1:800 dilution; Invitrogen). All primary and secondary antibodies were diluted in 0.1% BSA in PBS. Negative controls were performed by omitting one or both of the primary antibodies. After DNA counterstaining with Hoechst 33342, samples were mounted on slides with ProLong antifade reagent (Invitrogen) and observed using a Zeiss Axio Observer inverted microscope equipped with a Lumen 200 Fluorescence Illumination Systems (Prior Scientific, Inc., Rockland, MA). The images were captured with a QImaging Retiga EXi CCD digital camera (QImaging, Surrey, British Columbia, Canada).

In situ PLA

The PLA was used to reveal the cellular sites at which PGRMC1 and SUMO1 interact. The PLA was performed according to the manufacturer's instructions (OLINK Bioscience, Uppsala, Sweden), using the anti-Sumo1 and PGRMC1 antibodies as the primary antibodies. This technology detects protein-protein interaction in fixed cells using a complementary pair of oligonucleotide-labeled secondary antibodies (PLA probes), which generate a signal only when the two probes hybridize and therefore are in close proximity (OLINK Bioscience). The PLA signal was visualized as an individual fluorescent (red) spot. Negative controls were performed omitting one of the two primary antibodies.

PLA was also performed to assess the role of P4 in regulating PGRMC1-SUMO1 interaction. For this purpose, cells were plated in serum-supplemented medium, cultured overnight, then washed in serum-free medium and incubated for 5 h in serum-supplemented medium or in serum-free medium with ether vehicle or 1 μm P4. As a control, cells cultured in serum were also subjected to this PLA assay. The samples were then processed as described above. For each sample, images were captured in five different areas of the slides using identical exposure times and gain settings. The number of fluorescent spots per cell was determined with the iVision-Mac Image Acquisition and Analysis Software (BioVision Technologies, Exton, PA). The mean number of spots per cell was calculated for each treatment and used as an indicator of the degree of PGRMC1-SUMO1 interaction.

Assessment of transcription factor activity

The transcriptional activity that was regulated by P4 and PGRMC1 was assessed by three methods. These include transcription factor profiling, filter assays, and luciferase-based reporter assays.

Transcription factor profiling

To determine whether P4 regulated transcription factor activity, a TF Activation Profiling Plate Array assay was used according to the protocol provided by Signosis, Inc. (Sunnyvale, CA). Briefly, nuclear proteins were isolated from serum-free or P4-treated SIGC using the reagents and protocol provided by QIAGEN (Qproteome Nuclear Protein kit). Then, 12 μg of nuclear extract were added to a mixture of DNA sequences that encoded 48 different transcription factor-binding sites. This mixture was incubated at 16 C for 30 min to allow for the formation of transcription factor-DNA complexes and then passed through an isolation column to separate the transcription factor-DNA complexes from free DNA probes. One hundred microliters of elution buffer were then added to the column to elute the transcription factor-DNA complexes, and the transcription factor-DNA complexes were subsequently denatured by incubation at 98 C.

Ninety-five microliters of each sample were then added to each well of a 96-well plate that contained an immobilized complementary sequence to one of the 48 transcription factors. The 96-well plate was sealed and incubated overnight at 42 C to hybridize the complementary strands of DNA. After hybridization, the plate was washed three times, streptavidin-HRP conjugate and substrate added, and the resulting chemiluminescence detected using a multidetection microplate reader.

Transcriptional activity as assessed by filter assay

To confirm the transcription factor-profiling assay, which indicated that P4 influenced the activity of seven different transcription factors, filter assays were used. This assay was performed using reagents and protocols provided by Signosis, Inc. Briefly, 4 μg of nuclear extract were added to binding buffer and incubated at 16 C for 30 min with a DNA probe that encoded one of the following transcription factor-binding sites: neurofibromatosis type 1, pituitary-specific positive transcription factor 1, estrogen receptor, paired box protein-5, thyroid receptor, transcription factor IID, and T-cell factor/lymphoid enhancer factor (Tcf/Lef). The mixture was then transferred to filter plate to separate transcription factor bound DNA probes from free probe. The isolated DNA was hybridized to a complementary sequence and detected by chemiluminescence using a multidetection microplate reader.

Luciferase-based reporter assay

Because the filter assay only confirmed that P4 suppressed Tcf/Lef activity, studies were conducted to determine whether P4's ability to suppress Tcf/Lef activity was dependent on PGRMC1. To achieve this goal, a Tcf/Lef luciferase reporter assay was used (SA Biosciences, Fredrick, MD). In this assay, cells were transfected using a mixture of Lipofectamine 2000 and 100 ng of Tcf/Lef reporter DNA (or 100 ng of negative and positive reporter control) and either 2 pmol siRNA PGRMC1 (siRNA ID, 253165) or scramble control. Then, 2 × 10⁴ cells were plated in each well of a 96-well plate. Immediately after plating, the cells were incubated with serum-free Opti-MEM with or without 1 μm P4. After 24 h, 1 μl of RNasin Plus, a ribonuclease inhibitor, was added and dual Luciferase Reporter Assay (Promega, Madison, WI) performed. The RNA was then isolated from each well using the RNeasy Plus Mini kit (QIAGEN, Fredrick, MD) and assayed by PGRMC1 mRNA content by quantitative PCR (qPCR).

An additional study was conducted with the only modification being that the cells were transfected with either the scramble or PGRMC1 siRNA for 48 h before transfection with the Tcf/Lef luciferase report construct. As in the initial experiment, Tcf/Lef activity and PGRMC1 mRNA levels were assessed 24 h after transfection of the Tcf/Lef reporter construct.

To further elucidate PGRMC1's genomic action, an expression construct (p3XFLAG-CMV-14-hmnPGRMC1) that encodes a PGRMC1-Flag (3x) fusion protein was obtained from James Pru of Washington State University (Pullman, WA). This PGRMC1-Flag expression construct was generated using the mammalian expression system provided by Sigma Chemical Co. Using the QuikChange Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA) and this PGRMC1-Flag expression construct, a PGRMC1-Flag expression construct was generated, in which the lysine residues at amino acid 136, 187, and 193 (i.e. the sumoylation sites) were mutated to arginine (i.e. ΔSUMO-PGRMC1-Flag). That the mutagenesis correctly altered the sumoylation sites was confirmed by DNA sequencing using either the universal cytomegalovirus forward primer (5′-CGCAAATGGGCGGTAGGCGTG-3′) or the PGRMC1 forward primer (5′-GAGGATGTGGTGGCGACT-3′). The cellular localization and the ability of the PGRMC1-Flag and the ΔSUMO-PGRMC1-Flag fusion proteins to modulated P4-regulated Tcf/Lef activity was determined as described above.

qPCR analysis of PGRMC1 mRNA

qPCR was performed with the primers as published by Peluso et al. (11) using the Script one-site RT-PCR kit with SYBR green (Bio-Rad Laboratories, Hercules, CA). Relative level of Pgrmc1 mRNA was evaluated with Bio-Rad CFX96 software using the ΔCT method. Values were expressed as a percentage of the scramble serum-free control treatment.

Statistical analysis

All experiments were repeated two to three times with each experiment yielding essentially identical results. When appropriate, data from the replicate experiments were pooled and the data expressed as a mean ± se. A paired Student's t test was used to determine whether the means were different between two groups. When the means of three or more groups were compared, then either a one-way ANOVA followed by a Fisher's least significant difference test or a two-way ANOVA was used depending on the experimental design. A Fisher's exact test was used to analyze percentage data. All statistical analysis was done using GraphPad Prism 5.0 software (GraphPad, San Diego, CA). Regardless of the statistical test, P values of less than 0.05 were considered to be significant.

Results

Western blottings of whole-cell lysates prepared from SIGC detected a major band at approximately 22 kDa as well as two major bands between 50 and 75 kDa. Other bands that were greater than 75 kDa were also detected depending on the experimental conditions. The intensity of all of these bands was reduced by preabsorbing the PGRMC1 antibody with recombinant PGRMC1. No bands were detected in the absence of the primary antibody (Fig. 1A). Moreover, treatment with 100 mm DTT, which disrupts disulfide bonds that form dimers, did not greatly reduce the intensity of more than 50-kDa PGRMC1-bands but did increase the intensity of the approximately 22-kDa band (Fig. 1B). Although the more than 50-kDa bands were detected in the absence of detergents, the approximately 22-kDa band was only detected if the detergents, NP-40 and sodium deoxycholate, were added to the lysate buffer (Fig. 1C). Finally, the approximately 22-kDa band was detected in the cytoplasmic fraction, which contains plasma and organelle membranes, whereas the more than 50-kDa bands were detected in the nuclear fraction (Fig. 1D).

Fig. 1.

Characterization of the higher molecular mass bands observed in Western blottings of PGRMC1. All Western blottings in this and other figures used the antibody provided by Sigma Chemical Co. The effect of preabsorbing the PGRMC1 antibody with recombinant PGRMC1 (rPGRMC1) is shown in A, In this study, Western blottings were conducted using either 5 or 10 μg of whole-cell lysate protein. The effect of adding 100 mm DTT to the RIPA buffer is shown in B, whereas the effect of the detergents, NP-40 and sodium deoxylcholate, is shown in C. The presence of the higher molecular mass bands in the cytoplasm (cyto) and nuclear fractions is shown in D. The Western blottings shown in A and B were replicated twice, whereas the Western blottings shown in C and D were replicated three and four times, respectively.

Interestingly, the cellular distribution of PGRMC1 changed depending on cell density. After 24 h of culture, the cells were about 50% confluent, were actively undergoing mitosis (4.6% cells in metaphase), and had detectable levels of PGRMC1 in both the cytoplasm and nucleus of virtually all of the cells (Fig. 2A). After 72 and 96 h of culture, cell density increased; the percentage of cells in metaphase decreased to 1.3 and 0.5%, respectively, and PGRMC1 was only detected in cytoplasm of many cells (Fig. 2A). This change in the distribution of PGRMC1 corresponded to decrease in the intensity of the more than 50-kDa bands and an increase in the intensity of the approximately 22-kDa band (Fig. 2B).

Fig. 2.

The effect of SIGC density on the localization (A) and presence of the higher mass bands associated with PGRMC1 Western blottings (B). The experiments shown in A and B were replicated twice. GAPDH, Glyceraldehyde-3-phosphate dehydrogenase.

PGRMC1 has three putative sumoylation sites, which are well conserved being present in both rat (Fig. 3A) and human PGRMC1. If functional, their presence could account for the PGRMC1-specific bands that are more than 50 kDa. To test this concept, Western blottings were used to confirm that SUMO1 is expressed in SIGC (Fig. 3B). Moreover, the SUMO1 antibody coimmunoprecipitated two specific bands (compare lanes 1 and 2 of Fig. 3C). The approximately 100-kDa band was also specifically detected by PGRMC1 antibody (compare lane 3 with 4 of Fig. 3C). Note that this approximately 100-kDa band was not present in any of the lanes used for the various controls required for this immunoprecipitation protocol (lanes 2 and 4–6 of Fig. 3C).

Fig. 3.

Sumoylation of PGRMC1. A, A diagram of rat PGRMC1 is shown, in which putative sumoylation and phosphorylation sites are shown. The phosphorylation of serine (p-ser), threonine (p-Thr) and tyrosine (p-Tyr) residues are marked and preceded by their amino acid number. The location of the transmembrane (TM) and cytochrome b5 binding domain (Cyt b5 bind) is also shown. The expression of SUMO1 in SIGC is shown in the Western blotting in B. The negative control (CTRL) was obtained by omitting the SUMO1 antibody. C, Presence of higher molecular mass forms of PGRMC1 that coimmunoprecipitate with SUMO1. The immunoprecipitates (IP), obtained with either SUMO1 antibody or IgG control were probed with the SUMO1 (lanes 1 and 2) or the PGRMC1 antibody (lanes 3 and 4). Lanes 5 and 6 represent the controls of the Western blotting, in which only the secondary antibodies were used. Note that two secondary antibodies were used in these Western blottings. Lane 5 shows that, in the absence of a primary antibody, the mouse antirabbit IgG conformation-specific monoclonal antibody detects an approximately 150-kDa band in both immunoprecipitates and an additional approximately 60-kDa band in the SUMO1 immunoprecipitates, indicating that these bands are not specific. The experiments shown in B and C were replicated twice.

Immunofluorescence studies revealed that both SUMO1 and PGRMC1 were present in the cytoplasm and nuclei of SIGC (Fig. 4A and, for control images, see Supplemental Fig. 1, published on The Endocrine Society's Journals Online web site at http://endo.endojournals.org). The PLA detected a direct interaction between SUMO1 and PGRMC1 with this interaction being detected in both the cytoplasm and nucleus (Fig. 4B and, for control images, see Supplemental Fig. 1). Finally, the PGRMC1-SUMO1 interaction was significantly decreased in the absence of serum but could be restored to levels comparable with that of serum by P4 treatment (Fig. 4C).

Fig. 4.

Colocalization and interaction of SUMO1 and PGRMC1 as revealed by immunofluorescence (A) and in situ PLA (B). The site of interaction is revealed by a red dot. All images are presented at the same magnification with the magnification bar shown in B. The experiments shown in A and B were replicated three times. The effect of serum, serum-free, and serum-free plus 1 μm P4 on PGRMC1-SUMO1 interaction is shown in C. This experiment was replicated three times, with a total of 495, 535, and 479 cells being analyzed for the serum, serum-free, and serum-free + P4 treatments. *, Value is different from both the serum and serum-free plus P4 groups (P < 0.05).

Because nuclear PGRMC1 mediates P4-regulated gene expression in SIGC (11), PGRMC1's role in P4-regulated transcription factor activity was assessed. A screen of transcription factor activity suggested that P4 enhanced the activity of neurofibromatosis type 1 and pituitary-specific positive transcription factor 1 and decreased the activity of estrogen receptor, paired box protein-5, thyroid receptor, transcription factor IID, and Tcf/Lef (for details, see Supplemental Fig. 2). Subsequent studies revealed that only P4-induced suppression of Tcf/Lef activity could be confirmed by both a filter-based (Fig. 5A) and a luciferase reporter (Fig. 5B) assay. Moreover, treatment with PGRMC1 siRNA for 24 h reduced PGRMC1 mRNA levels to 25–50% of scramble control but only slightly attenuated P4's ability to suppress Tcf/Lef activity. However, PGRMC1 siRNA treatment for 72 h suppressed PGRMC1 levels to 7% of scramble control and completely blocked the capacity of P4 to suppress Tcf/Lef activity (Fig. 6, A and B).

Fig. 5.

The effect of P4 on Tcf/Lef transcription factor activity as assessed by a filter assay (A) or a luciferase reporter assay (B). Experimental details are presented Materials and Methods. Briefly, the filter assay used a biotinylated Tcf/Lef-binding site (5′-biotin-ACGTTACTTTGATCTGATCAGGCC-3′) that was incubated with nuclear proteins isolated from cells cultured under serum-free conditions. The biotinylated Tcf/Lef probe that bound nuclear protein (i.e. an indication of activated Tcf/Lef transcription) was isolated by centrifugation through a filter. The amount of bound-Tcf/Lef probe was detected with streptavidin-HRP and the luciferase activity monitored. The data in A represent the mean ± se of nine samples obtained from serum-free-treated cells and nine samples from serum-free plus P4-treated cells. In each experiment, a serum-free and serum-free plus P4 sample were assessed for Tcf/Lef activity. The Tcf/Lef luciferase reporter assay used a TCF/LEF-responsive luciferase construct that encodes both the renilla luciferase reporter, which serves as an internal normalizing control and the firefly luciferase reporter gene under the control of a minimal cytomegalovirus promoter and tandem repeats of the TCF/LEF transcriptional response element. The data in B were obtained from three separate experiments with each experiment run in triplicate or quadruplicate. Values were expressed as the percentage of serum-free control for each experiment. Means and se were calculated from each value (n = 11). In both A and B, a paired Student's t test was used to assess differences in mean values. *, Value that is different from the serum-free control group (P < 0.05).

Fig. 6.

The effect of PGRMC1 siRNA and P4 treatment on PGRMC1 mRNA levels and Tcf/Lef transcriptional activity as assessed by a luciferase-based reporter assay. In these experiments, SIGC were either transfected with siRNA and Tcf/Lef luciferase reporter construct, plated, and assayed for both PGRMC1 mRNA and luciferase activity or plated, transfected with siRNA for 48 h, and then transfected with the Tcf/Lef luciferase reporter construct. After an additional 24 h, PGRMC1 mRNA levels and luciferase were determined. The data in A and B were obtained from four separate experiments with each experiment run in triplicate. Means and se were calculated from each value (n = 12). In both A and B, an ANOVA was used to assess differences in mean values. *, Value that is different from the scramble control group at each time period (P < 0.05).

To establish a causal relationship between the sumoylation of PGRMC1 and PGRMC1's ability to mediate P4's suppressive effect on Tcf/Lef activity, a Flag-tagged PGRMC1 fusion protein was transfected into SIGC. Western blot analysis of lysates prepared 24 h after transfection with PGRMC1-Flag detected Flag-labeled proteins at approximately 25 kDa as well as protein bands more than 50 kDa. Stripping and reprobing this blot with the PGRMC1 antibody detected bands that corresponded to the bands detected with the Flag antibody. In addition, an approximately 22-kDa band was observed using the PGRMC1 antibody. This band likely represents endogenous PGRMC1 (Fig. 7A). Note that the higher molecular mass forms of PGRMC1 cannot be differentiated from the PGRMC1-Flag fusion proteins, because the addition of the three Flag peptides only adds 3 kDa to PGRMC1, and this difference cannot be detected in the higher molecular mass range of the blot.

Fig. 7.

The expression of PGRMC1-Flag (A) and ΔSUMO-PGRMC1-Flag fusion (B) and their effect on P4's ability to regulate Tcf/Lef activity (C). In A, SIGC were transfected with the PGRMC1-Flag expression construct, and 24 h later, expression was assessed by Western blotting, which used the Flag antibody. The blot was stripped and reprobed with the PGRMC1 antibody. Note that PGRMC1-Flag fusion proteins were detected as both the lower and the higher molecular mass forms. The arrow points to the endogenous PGRMC1 protein that was detected by the PGRMC1 antibody but not the Flag antibody. In B, SIGC were transfected with either the PGRMC1-Flag construct or the Δ SUMO-PGRMC1-Flag construct. After 24 h, cytoplasmic (C) and nuclear (N) fractions were prepared and processed for Western blot analysis using the Flag antibody. The Western blottings shown in A and B were replicated twice. In C, SIGC were transfected with the Tcf/Lef luciferase reporter construct and either vehicle, PGRMC1-Flag, or ΔSUMO-PGRMC1-Flag constructs, plated, and monitored for luciferase activity 24 h after plating. Each treatment was run in quadruplicate and repeated twice. Means ± se from each value (n = 8/group). *, Values that are different from mock controls (P < 0.05); **, value that is less than its no treatment control (φ) (P < 0.05); ***, value that is less than its no treatment control (φ) but greater that P4 treatment of both the mock control and the P4-PGRMC1-Flag treatment (P < 0.05).

Like endogenous PGRMC1, the lower molecular mass form of PGRMC1-Flag fusion protein localized to the cytoplasm, whereas the higher molecular mass (>50 kDa) forms localized to the nucleus (Fig. 7B). Interestingly, the ΔSUMO-PGRMC1-Flag fusion protein was also detected as low and higher molecular mass forms (>50 kDa), which were localized to the cytoplasm and nucleus, respectively (Fig. 7B).

The presence of either PGRMC1-Flag or ΔSUMO-PGRMC1-Flag increased basal Tcf/Lef activity by nearly 2-fold (Fig. 7C). P4 suppressed the PGRMC1-Flag-induced increase in basal Tcf/Lef activity, but P4's ability to suppress Tcf/Lef activity was attenuated in the presence of the ΔSUMO-PGRMC1-Flag fusion protein (Fig. 7C).

Discussion

PGRMC1 is predicted to be a 22-kDa protein, and a corresponding band of about 22 kDa is detected on Western blottings. Interesting, many PGRMC1 Western blottings also detect higher molecular mass bands (i.e. >50 kDa) (6, 20, 21). The present study supports the concept that both the 22 kDa and higher molecular mass bands are PGRMC1, because they are not detected when either the primary antibody is omitted or the PGRMC1 antibody is preabsorbed with recombinant PGRMC1. This conclusion is also consistent with the finding that PGRMC1 siRNA treatment depletes both the lower and higher molecular mass forms of PGRMC1 (11). It is assumed that some of the higher molecular mass forms of PGRMC1 are oligomers of the 22-kDa form. This assumption is based on the finding that 100 mm DTT, which is known to disrupt the disulfide bonds that join dimers, increases the proportion of the 22-kDa form and decreases the proportion of higher molecular mass forms in preparations of the porcine liver membranes (20). DTT treatment also increases the proportion of the 22-kDa form in whole-cell lysates prepared from SIGC. This agrees with the concept that PGRMC1 forms oligomers. However, most of the higher molecular mass forms of PGRMC1 in whole-cell lysates are still present even after treatment with 100 mm DTT. This indicates that in addition to dimerization, other mechanisms account for the presence of the higher molecular mass forms of PGRMC1.

Insight into the nature of higher molecular mass forms of PGRMC1 is provided by comparing lysates prepared with and without the detergents, NP-40 and sodium deoxycholate. This comparison demonstrates that the 22-kDa form is only detected if detergents are used to prepare the lysate, whereas the higher molecular mass forms are observed even in the absence of detergents. Because detergents solubilize membrane-bound proteins (22), this finding suggests that only the 22-kDa form of PGRMC1 is membrane bound. Moreover, cell fractionation studies indicate that the 22-kDa form of PGRMC1 localizes to the cytoplasmic fraction that includes plasma and organelle membranes, whereas the higher molecular mass forms reside in the nucleus (present study and Ref. 11). Furthermore, the cellular distribution of PGRMC1 appears to play an important physiological role, because nuclear PGRMC1 is present in mitotically active SIGC but not in cells that are in mitotic arrest due to being contact inhibited.

That PGRMC1's cellular localization changes under different physiological conditions raises the question as to whether the transmembrane domain of PGRMC1 must be cleaved in order for PGRMC1 to translocate to the nucleus. Interestingly, a potential enzymatic cleavage site exists between amino acids 48 and 49. The amino acid sequence that encodes the predicted cleavage site is subsequent to the transmembrane domain (see http://www.cbs.dtu.dk/services/SignalP/). This site is predicted to be cleaved by an Asp-N endopeptidase (EC 3.4.24.33; http://web.expasy.org/peptide_cutter/). Because granulosa cell extract has endopeptidase activity (23, 24) and the ovary is known to express at least one EC 3.4.24.33 endopeptidase, matrix metalloproteinase-7 (25), it is possible that enzymatic cleavage of the extracellular and transmembrane domains of PGRMC1 is part of the mechanism that facilitates nuclear localization of PGRMC1.

If removal of the extracellular and transmembrane domains of PGRMC1 is required for nuclear localization, then there must be other posttranslational modifications that account for nuclear PGRMC1 being detected as higher molecular mass bands. In silico analysis of PGRMC1 predicts the presence of three sumoylation sites located at amino acids 138, 187, and 193 (http://sumosp.biocuckoo.org/online.php) (15). The present data indicate that sumoylation accounts in part for at least one of the higher molecular mass forms of PGRMC1. This conclusion is based on the findings that PGRMC1 coimmunoprecipitates with SUMO1 as revealed by a specific band at approximately 100 kDa. It is important to appreciate that in addition to SUMO1, there are three other SUMO family members (16) that could be covalently bonded to PGRMC1. PGRMC1 that is sumoylated with SUMO2, SUMO3, or SUMO4 could explain why not all of the higher molecular mass forms of PGRMC1 coimmunoprecipitate with SUMO1.

The data from the PLA also support the concept that PGRMC1 is sumoylated. PLA detects a direct interaction between PGRMC1 and SUMO1. Some of the PGRMC1-SUMO1 complexes are associated with a nuclear localization. This fits with the presence of the higher molecular mass forms of PGRMC1 within the nucleus. However, some PGRMC1-SUMO1 complexes are observed within the cytoplasm. This implies that PGRMC1 is sumoylated within the cytoplasm and then is transported to the nucleus. This mechanism would be similar to the requirement for Ran GTPase activating protein 1 to be sumoylated to localize to the nucleus (16). However, sumoylation is not required for PGRMC1 to localize to the nucleus, because ΔSUMO-PGRMC1-Flag is among the nuclear proteins as assessed by Western blotting.

The present studies also show that P4 can stimulate PGRMC1 sumoylation. How P4 mediates this effect is not clear. It is known that the activation of the Erk pathway in COS7 cells suppresses the sumoylation of the E-twenty six-like transcription factor 1 (26). Interestingly, P4 has been shown to suppress Erk activity in SIGC (27). If a similar relationship exists between Erk activity and PGRMC1 sumoylation, then P4's ability to suppress Erk activity could account for an increase in PGRMC1 sumoylation.

Regardless of the mechanisms that account for its sumoylation and localization, PGRMC1 plays an essential role in P4's ability to alter gene expression (11). To begin to define PGRMC1's capacity to regulate gene transcription, a series of studies were conducted to determine which transcription factors are regulated by P4. The first approach determined the ability of P4 to influence the activity of 48 different transcription factors. This approach revealed that P4 influenced the activity of several transcription factors, but only Tcf/Lef activity was confirmed to be suppressed by P4 using both filter and luciferase reporter assays. Moreover, P4 regulation of Tcf/Lef activity is mediated through a PGRMC1-dependent mechanism, because treatment with PGRMC1 siRNA depletes PGRMC1 levels and attenuates P4's ability to suppress Tcf/Lef activity.

The ability of P4 to regulate Tcf/Lef activity is likely to be an essential component of P4's capacity to regulate both mitosis and apoptosis. An increase in Tcf/Lef activity stimulates the expression of early immediate genes, such as activator protein 1 and c-myc, thereby promoting entry into the G₁ stage of the cell cycle (28, 29). Once entry into the cell cycle is initiated, the expression of these early immediate genes in granulosa cells as well as other cell types is suppressed (30). In granulosa cells, mitogen-induced expression of these early immediate genes is suppressed by P4 (31). Taken together, these findings suggest that P4's ability to suppress these early immediate genes is likely due to suppression of Tcf/Lef activity, which is mediated through a PGRMC1-dependent mechanism and ultimately accounts in part for P4 being able to suppress mitosis (10, 19, 31).

PGRMC1's ability to modulate Tcf/Lef activity is complex. The present studies suggest that PGRMC1 acts as a P4-dependent switch. In the absence of P4, PGRMC1 stimulates Tcf/Lef activity as revealed by the observation that Tcf/Lef activity is increased by overexpression of PGRMC1-Flag fusion protein and its nuclear presence. This stimulatory effect of PGRMC1 is not dependent on sumoylation, because ΔSUMO-PGRMC1-Flag fusion protein also enters the nucleus and enhances basal Tcf/Lef activity. In contrast, P4 activation of PGRMC1 changes its effect, such that Tcf/Lef activity is suppressed, and this suppressive action is dependent at least in part on PGRMC1 being sumoylated. It remains to be determined which sites are required to be sumoylated and which SUMO family members are involved in PGRMC1's capacity to inhibit Tcf/Lef activity in the presence of P4.

It is important to appreciate that mitosis and apoptosis have similar initiating events (32), and many granulosa cells enter mitosis before undergoing apoptosis (33). Because Tcf/Lef activation stimulates genes that promote entry into the cell cycle (29), P4-PGRMC1 activation likely prevents either an inappropriate entry into the cell cycle and/or a prolongation of early immediate gene expression, thereby preventing a “mitotic catastrophe” and subsequent apoptosis (34).

In summary, the present studies indicate that the higher molecular mass forms of PGRMC1 are due in part to sumoylation, which is stimulated by P4. Although the precise posttranslational modifications that account for the majority of the higher molecular mass forms remain unknown, these higher molecular forms of PGRMC1 are important, because they mainly localize to the nucleus, where they are involved in regulating Tcf/Lef activity through a P4-independent and P4-dependent mechanisms. Therefore, the ability to suppress Tcf/Lef activity is likely an essential part of the mechanism through which P4-PGRMC1 interactions regulate both mitosis and apoptosis.