| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
, ß, and 
) Localize to the Endoplasmic Reticulum and Are Not Activated by Progesterone
Endokrinologikum Hamburg (T.K., M.H., B.G.), 20251 Hamburg, Germany; Institute of Reproductive and Developmental Biology (M.S.F., J.K., I.H., J.J.B.), and Cancer Research UK Labs, Department of Oncology (E.W.-F.L.), Imperial College London, Hammersmith Hospital, London W12 ONN, United Kingdom; and Department of Clinical Pharmacology (R.L.), School of Medicine Mannheim, University of Heidelberg, 68167 Mannheim, Germany
Address all correspondence and requests for reprints to: Birgit Gellersen, Ph.D., Endokrinologikum Hamburg, Falkenried 88, 20251 Hamburg, Germany. E-mail: gellersen{at}endokrinologikum.com; or Jan J. Brosens, M.D., Ph.D., Institute of Reproductive and Developmental Biology, Imperial College London, Hammersmith Hospital, London W12 ONN, United Kingdom. E-mail: j.brosens{at}ic.ac.uk.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
, ß, and 
) with structural similarity to G protein-coupled receptors have been identified. These mPR isoforms are expressed in a tissue-specific manner and belong to the larger, highly conserved family of progestin and adiponectin receptors found in plants, eubacteria, and eukaryotes. The fish mPR has been reported to mediate progesterone-dependent MAPK activation and inhibition of cAMP production through coupling to an inhibitory G protein. To functionally characterize the human homologs, we established human embryonic kidney 293 and MDA-MB-231 cell lines that stably express human mPR, ß, or . For comparison, we also established cell lines expressing the mPR cloned from the spotted seatrout (Cynoscion nebulosus) and Japanese pufferfish (Takifugu rubripes). Surprisingly, we found no evidence that human or fish mPRs regulate cAMP production or MAPK (ERK1/2 or p38) activation upon progesterone stimulation. Furthermore, the mPRs did not couple to a highly promiscuous G protein subunit, Gq5i, in transfection studies or provoke Ca2+ mobilization in response to progesterone. Finally, we demonstrate that transfected mPRs, as well as endogenous human mPR, localize to the endoplasmic reticulum, and that their expression does not lead to increased progestin binding either in membrane preparations or in intact cells. Our results therefore do not support the concept that mPRs are plasma membrane receptors involved in transducing nongenomic progesterone actions. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
The cloning and characterization in 2003 of an entire new family of membrane progestin receptors (mPR) with no structural homology to the nuclear PR or PGRMC1 was widely considered a major breakthrough in the field (19, 20). The prototypical mPR was first isolated from ovaries of the spotted seatrout (Cynoscion nebulosus) and predicted to be a seven transmembrane spanning G protein-coupled receptor (GPCR). Upon expression of the seatrout (st) mPR in human MDA-MB-231 breast cancer cells, short-term P4 treatment resulted in MAPK (ERK1/2) activation and pertussis toxin-sensitive inhibition of cAMP formation, indicating inhibitory G (Gi) protein coupling. In addition to mPR, two other isoforms, mPRß and mPR, have been identified in a wide variety of species, ranging from Xenopus to humans. Binding of P4 to bacterially expressed human (h) mPR and hmPR was competed for by 17-hydroxyprogesterone (17-OHP) and 20ß-hydroxyprogesterone (20ß-OHP), but not by other steroids including 17ß-estradiol, testosterone, cortisol, or by the antiprogestin RU486 (19, 20).
Based on amino acid sequence homology, the mPRs belong to a larger family of 11 highly conserved mammalian paralogs termed the PAQR (progestin and adipoQ receptors) family, which also includes the two adiponectin receptor isoforms AdipoR1 and AdipoR2. The PAQRs are uniquely identified by the PFAM (database of protein families) motif UPF0073 and appear to be an ancient family that evolved preceding the split of eukaryotes and eubacteria (21, 22). Within this nomenclature, mPR, ß, and are designated PAQR7, PAQR8, and PAQR5, respectively.
Northern blot analysis of human tissues revealed that mPR is primarily expressed in reproductive tissues such as testis, ovary, and placenta, whereas mPRß predominates in the brain and kidney and mPR in the kidney and intestinal tract (19). The same tissue distribution was reported for mPRß and mPR mRNAs in the rat, whereas mPR appears to be more widely expressed (23). We have performed extensive expression profiling of mPRs in human gestational tissues and found mPR to be the most highly expressed isoform in placenta (22). Interestingly, mPR and mPRß transcripts were found to be down-regulated in the myometrium before parturition, suggesting an important role for these novel receptors in P4 target tissues (22). These observations prompted us to functionally characterize the human mPR isoforms to elucidate their physiological roles and their potential as pharmacological targets. Surprisingly, our findings do not support the concept that the mPRs are plasma membrane receptors or that they modulate cAMP production, MAPK activation or Ca2+ mobilization in response to P4.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
, mPRß, or mPR, with or without an N-terminal hemagglutinin (HA) or a C-terminal V5 tag. For comparative analysis, we also generated a vector for expression of the previously characterized seatrout (st) mPR (20) with an N-terminal HA tag. Although mPR has been cloned from a number of other teleost species including the zebrafish (Danio rerio), goldfish (Carassius auratus), and channel catfish (Ictalurus punctatus) (19, 24, 25), the homologous transcript has not yet been isolated from the Japanese pufferfish (Takifugu rubripes, Fugu rubripes). Using the stmPR sequence in a homology search, we identified the corresponding genomic sequence in the Fugu Genome Project database and cloned the mPR cDNA from fugu (f) kidney total RNA. The fmPR cDNA sequence has been submitted to GenBank database under accession no. DQ400857. The coding region of fmPR cDNA exhibits 83.2% similarity to that of stmPR (see supplemental figure, Fig. S1, which is published as supplemental data on The Endocrine Society’s Journals Online web site at http://mend.endojournals.org). Comparative analysis of the fish homologs at the amino acid sequence level revealed that the fmPR protein is most closely related to stmPR (90.9%) (Fig. 1
). Notably, the fugu and seatrout proteins contain 352 amino acids as opposed to 354 amino acids in the other fish species, due to a deletion of the same two residues in the core of the mPR protein. Search for putative transmembrane domains (TM) with the TopPred II program (26) predicted seven TMs in the seatrout protein, as expected, but only six for the fugu homolog as it lacked the fourth TM of the seatrout mPR (Fig. 1).
|
, hmPRß, and hmPR either as untagged proteins, or as fusions with HA or V5 tags. Furthermore, we stably transfected HA-tagged human, seatrout, or fugu mPR into MDA-MB-231 cells, the human breast cancer cell line originally used to demonstrate that stmPR mediates P4-dependent effects on cAMP and MAPK signaling (20).
Western blot analysis was used to select clones that highly express tagged mPR isoforms for further analysis. Notably, all mPRs migrated on SDS-PAGE faster than expected on basis of their theoretical mass, regardless the redox state of the samples (Fig. 2, and data not shown). Anomalous migration in SDS-PAGE could be due to the highly hydrophobic nature of the proteins, a known phenomenon (29). Alternatively, in the case of mPRs tagged with a C-terminal V5 epitope, lower apparent molecular mass might indicate the cleavage of an N-terminal signal peptide. This is, however, not consistent with the detection of mPRs with N-terminal HA tag. Cleavage of a signal peptide would lead to loss of the epitope and failure to immunodetect the mature protein. Furthermore, the theoretical size difference between the larger V5-tagged forms (which also include a C-terminal 6xhistidine epitope) and the corresponding smaller HA-tagged forms amounts to about 2 kDa, which is well reflected by the relative faster migration of the latter isoforms (Fig. 2A).
|
, when expressed in mammalian cells, mediates P4-dependent and pertussis toxin-sensitive inhibition of adenylyl cyclase, suggesting coupling to a Gi protein (20). We used HEK293-HA/hmPR, HEK293-HA/hmPRß and HEK293-HA/hmPR cells to test whether the hmPR isoforms also inhibit cAMP formation in response to P4. Like Zhu et al. (20) in 1993, we treated the cells with 10 nM to 1 µM P4 for 5 min. No decline in cAMP levels was observed in wt HEK293 cells upon P4 treatment (Fig. 3A). Surprisingly, such response was also lacking in HEK293 cells stably expressing HA-tagged hmPR, hmPRß, or hmPR (Fig. 3A). To exclude that the HA tag of the mPRs interfered with signal transduction, we repeated these experiments in HEK293 cells stably expressing hmPR, hmPRß, or hmPR with C-terminal V5 tag or without tag. Again, P4 treatment did not inhibit cAMP formation in any of the cell lines tested (Fig. S2A). In contrast, cAMP levels increased significantly in all HEK293 lines, wt or stably transfected clones, upon stimulation with 1 µM forskolin (data not shown).
|
and/or was due to the host cell used, we generated MDA-MB-231 breast cancer cells stably expressing stmPR to reproduce the conditions used by Zhu et al. (20). For comparison, we performed parallel experiments in MDA-MB-231 cells stably expressing human or fugu mPR. The fmPR is highly related to the seatrout protein (91% similarity), whereas the hmPR shares only 54% sequence similarity with the stmPR protein. As shown in Fig. 3B
, we found no evidence for P4-dependent inhibition of cAMP formation in either wt MDA-MB-231 cells or in cells overexpressing human, seatrout, or fugu mPR.
Stimulation with nicotinic acid, a ligand of the endogenous Gi-coupled receptor GPR109a (also designated PUMA-G) (30, 31), was performed as a positive control and resulted in a marked decrease of cAMP production within 5 min in HEK293 or MDA-MB-231 wt cells (Fig. S2B), and in cells stably transfected with HA/mPR of human or seatrout origin, whereas parallel incubations with P4 again failed to elicit such a response (Fig. 3C). Activity of the P4 preparation and the ability of MDA-MB-231 cells to respond to the hormone were confirmed by cotransfection of a reporter gene construct driven by a progesterone response element and expression vectors for either nuclear PR-A or PR-B. Stimulation with P4 at the same doses as used for the preceding experiment resulted in reporter gene activation in MDA-MB-231 cells transfected with PR-B, but not with PR-A or empty vector (Fig. S2C).
mPRs Do Not Activate ERK1/2 in Response to P4
Next, we examined whether the mPRs couple to MAPK activation. The wt HEK293 cells or clones stably expressing HA/hmPR, ß, or , were treated with P4 (10 nM to 1 µM) for 5 min before protein extracts were harvested for Western blot analysis. Control cells were stimulated with epidermal growth factor (EGF) (10 pM to 1 nM) for 10 min. Western blots were probed with antibodies specific for total ERK1/2 and for the phosphorylated active form of ERK1/2 (p-ERK1/2). No P4-dependent increase in the amount of p-ERK1/2 relative to total ERK1/2 was detectable in either wt HEK293 cells or any of the clones expressing the human mPR isoforms (Fig. 4A). Likewise, HEK293 cells stably transfected with C-terminally V5-tagged hmPR, ß, or , or with the untagged hmPR isoforms, did not mount a response to P4 in this assay (Fig. S3). In contrast, all cell lines were extremely sensitive to EGF as reflected by a massive increase in ERK1/2 activation, even at a concentration as low as 10 pM (Fig. 4A and Fig. S3).
|
to demonstrate a rapid and dose-dependent increase in p-ERK1/2 in response to P4 at a concentration range of 10–1000 nM, which did not occur in control cells transfected with empty vector or a vector containing a reversed insert. Recapitulating these experimental conditions, we found no evidence of P4-dependent ERK1/2 activation in MDA-MB-231 cells expressing mPR from seatrout, fugu or human origin, or in wt MDA-MB-231 cells (Fig. 4B). In contrast, EGF elicited an increase in ERK1/2 activation even at the lowest concentration (10 pM).
mPRs Do Not Activate p38 MAPK in Response to P4
A recent report demonstrated that treatment of human uterine myocytes with cell-impermeable P4-BSA or with free P4 elicits rapid phosphorylation of p38 MAPK, but not ERK1/2, and inferred that this response was mediated by endogenous hmPR and ß (32). To test whether such response was transduced by mPR, we incubated MDA-MB-231 cells stably expressing human, seatrout, or fugu mPR with P4 (0.1, 1 µM). No p38 activation was seen after 15 or 30 min of treatment in wt or mPR-expressing cells, whereas a robust increase in phosphorylated p38 was obtained with anisomycin, a bacterial antibiotic included as a positive control (Fig. 5). Likewise, no P4-induced p38 activation was elicited within 5 min of treatment (data not shown).
|
, of human or fish origin, is not involved in relaying P4 signaling to the p38 MAPK cascade.
mPRs Do Not Couple to a Promiscuous G Protein and/or Mediate Ca2+ Mobilization in Response to P4
To further assess whether mPRs are functional GPCRs and activated by P4, we made use of a Ca2+ reporter assay involving a promiscuous G protein -subunit, Gq5i (30, 33). This chimeric protein is based on a Gq backbone with a mutation of the glycine residue in linker I (G66K), and harbors a substitution of the five amino acids at the extreme C terminus by those of Gi (34). The G66 mutation confers upon the Gq the ability to be activated by Gi- and Gs-coupled receptors but does not affect the ability to reconstitute a functional Gq-phospholipase Cß-calcium signaling pathway (35). The C-terminal modification bestows vastly improved interaction, as compared with other promiscuous G proteins, with Gi-selective GPCRs (36). To monitor receptor activation, we used Chinese hamster ovary (CHO)-K1 cells that stably express a highly sensitive Ca2+ reporter, consisting of green fluorescent protein fused to aequorin (30, 37). These reporter cells were transiently transfected with expression vectors encoding Gq5i and either hmPR or stmPR. In control experiments, cells were cotransfected with Gq5i and GPR109a, a Gi-coupled receptor for nicotinic acid (30, 31). As shown in Fig. 6, nicotinic acid rapidly elevated intracellular Ca2+ levels in reporter cells transfected with promiscuous Gq5i and GPR109a. Notably, this response was approximately hundred-fold lower in cells expressing only GPR109a (data not shown) and undetectable in cells transfected with Gq5i alone (Fig. 6). Importantly, we found no evidence of Ca2+ mobilization upon treatment with P4 (10 nM to 1 µM) in reporter cells transfected with hmPR or stmPR in the presence or absence of Gq5i (Fig. 6, and data not shown). Furthermore, P4 failed to induce aequorin-mediated green fluorescent protein luminescence in reporter cells transfected with Gq5i and hmPRß, hmPR, or fmPR (Fig. S4). Although the above experiments were performed with HA-tagged mPR isoforms, transfection of untagged hmPR yielded identical results (data not shown).
|
Subcellular Localization of mPRs
The lack of P4 responses in our cell models prompted us to examine the subcellular localization of the mPR isoforms. To this end, we first transiently transfected COS-1 cells with either N-terminally HA-tagged or C-terminally V5-tagged hmPR (Fig. 7). Cells were either fixed in paraformaldehyde (PFA) (Fig. 7, c and d) or in permeabilizing acetone/methanol (Fig. 7, a and b). Staining with anti-V5 or anti-HA antibodies yielded strong immunoreactivity for tagged hmPR in acetone/methanol-fixed cells but not in PFA-fixed cells. However, permeabilization of PFA-fixed cells with either Triton X-100 (Fig. 7e) or digitonin (Fig. 7f) allowed the detection of tagged proteins, indicating that neither the N nor the C terminus of the protein is exposed on the cell surface. This could mean that the receptor is not inserted into the plasma membrane, or, alternatively, that it spans the membrane an even number of times such that both termini are located intracellularly. We reasoned that if transfected hmPR is expressed on the plasma membrane with classical GPCR topology, then the N-terminal tag should be detectable on living cells with an ELISA. To test this, MDA-MB-231 cells were transiently transfected with N-terminally HA-tagged hmPR or HA-tagged plexin-B2, a known cell surface protein that served as a positive control (38). As shown in Fig. 8, strong luminescence indicating surface expression was detected in cells transfected with HA-tagged plexin-B2. In contrast, the levels were 47 times lower in HA/hmPR expressing cells and comparable to those of mock transfected cells (Fig. 8).
|
|
and stained the cells with an affinity-purified antipeptide antibody that specifically recognizes the C terminus of the human isoform (22). As shown in Fig. 9, a–c, the pattern of wt hmPR immunofluorescence was indistinguishable from that of the tagged proteins, characterized by strong perinuclear staining that fizzled out into a cytoplasmic tubuloreticular network. Furthermore, overexpressed wt hmPR was only detectable upon permeabilization of the cells.
|
isoforms. In addition, we cotransfected combinations of the receptor isoforms because it has been reported for various GPCRs that heteromerization is essential for proper trafficking and membrane insertion (39). To this end, COS-1 cells were transiently transfected with expression vectors that encode HA-tagged hmPRß or either alone or in combination with wt hmPR. Like hmPR, the ß and isoforms were detected in the cytoplasm (Fig. 9, d and e). Furthermore, coexpression of hmPR with either hmPRß or did not alter the cellular localization pattern, and overlay of images revealed a high degree of colocalization between the family members (Fig. 9, f and g).
Previous studies have reported that transfection of stmPR into human cells results in plasma membrane expression (20), raising the possibility that the cellular localization of fish mPRs differs from that of their mammalian homologs. To test this hypothesis, we also transfected COS-1 cells with expression vectors that code for N-terminally HA-tagged seatrout or fugu mPR. Surprisingly, no immunoreactivity was detected in nonpermeabilized PFA-fixed cells. Furthermore, the cellular distribution of both fish homologs was identical with that of hmPRs (Fig. 9, h and i). It has been proposed that stmPR has a classical GPCR topology exposing the N terminus on the cell surface (20). To validate these results, we also performed ELISA on living cells transfected with N-terminally HA-tagged stmPR, but again, the level of luminescence was comparable to that of mock-transfected cells (Fig. 8). The cytoplasmic localization of exogenous mPRs was not an artifact of COS-1 cells as an identical pattern of expression was observed in transfected Ishikawa and MDA-MB-231 cells (data not shown, and Fig. S5). Furthermore, no difference in the cellular expression pattern was observed between transiently and stably transfected hmPR (Fig. S5). Together, the data unequivocally demonstrate that mPRs do not localize to the cell surface, at least not when overexpressed.
Heterologous expression of GPCRs can result in retention of the receptor in the endoplasmic reticulum (40, 41). In an RT-PCR screen for endogenous mPR expression in a range of cell lines, we had observed that transcripts for hmPR were readily detected in Ishikawa and MDA-MB-231 cells (data not shown). This prompted us to examine the subcellular localization of the endogenous hmPR in these cell types. Ishikawa cells were first double stained for hmPR and pan-cadherin, a plasma membrane marker. As shown in Fig. 10, a–c, endogenous hmPR immunofluorescence was found to be almost exclusively cytoplasmic with little or no overlap with pan-cadherin staining. In contrast, staining of cells with an endoplasmic reticulum marker (KDEL) yielded an almost complete overlap with endogenous hmPR immunofluorescence (Fig. 10, d–f). To visualize the mitochondria, Ishikawa cells were loaded with Mitotracker Red CMXRos, but the pattern of staining was distinct from that of the endogenous hmPR (Fig. 10, g–i). These colocalization studies were repeated with MDA-MB-231 cells and again a high degree of overlap was found between hmPR and KDEL staining but not with pan-cadherin or Mitotracker Red CMXRos (Fig. S6).
|
immunofluorescence can be attributed to its expression in the endoplasmic reticulum. Furthermore, the expression patterns of the transfected hmPR, ß and isoforms were very similar, and the distribution of transiently or stably transfected hmPR was indistinguishable from that of the endogenous protein.
Progesterone Binding Assays
To determine the binding characteristics of eukaryotically expressed mPRs, we incubated crude membrane preparations from stably transfected MDA-MB-231 or HEK293 cells with tritiated P4 as the ligand, and compared the amount of displaceable binding to that of the corresponding wt cells (Fig. 11). In addition to P4, we employed 17-OHP and 20ß-OHP for competition studies as these steroids have been reported to bind to recombinant stmPR and hmPR, whereas 17ß-estradiol was used as a nonspecific competitor (19, 20). If the mPRs were bona fide P4 receptors, then the amount of tritiated P4 that remains bound in the presence of excess cold P4 could be considered nonspecific binding, and the difference between total binding and nonspecific binding would represent specific binding. However, because the ligand specificity of fish and human mPRs has not been determined in a eukaryotic system, we decided to avoid these terms and substitute them with nondisplaceable binding and displaceable binding. In wt MDA-MB-231 cells, about 30–50% of total bound [3H]-P4 was displaced by P4, 17-OHP or 20ß-OHP, whereas less than 10% was displaced by 17ß-estradiol. However, neither total binding nor the amount of displaceable binding was elevated in MDA-MB-231 cells expressing HA-tagged mPR of human, seatrout, or fugu origin, irrespective of the competitor used (Fig. 11, A–D).
|
, ß, or did not show higher displaceable binding of P4 than the corresponding wt cells (Fig. 11E). In this binding experiment, a preparation of porcine liver microsomes, which are known to contain high-affinity P4 binding sites, was included as a positive control (42). Similar results were obtained with V5-tagged hmPRs, and when using microsomal fractions instead of crude membrane preparations (data not shown).
Furthermore, we performed binding assays on intact whole cells, using COS-1 cells transiently transfected with untagged hmPR, and with PR-B as a positive control. As a competitor, we used 17-OHP, which has been reported to bind to mPR and with high specificity (19, 20). In addition, 17-OHP is a known agonist on PR-B and is clinically used for prevention of recurrent preterm delivery (43, 44). Whereas cells expressing PR-B readily bound tritiated P4, and binding was largely displaced by excess 17-OHP, the amount of total and displaceable binding was extremely low in cells expressing hmPR and not different from that seen in untransfected cells or cells transfected with empty vector (Fig. 11F).
In summary, we found no evidence of increased P4 binding in mammalian cells that overexpress mPRs.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
and mPRß. According to this model, activation of mPRs leads to a decrease in the level of SRC2 and a modulation of PR-B transcriptional activity. Moreover, the authors reported increased expression of mPR at term and, through coupling to Gi protein, reduced cAMP formation, resulting in increased phosphorylation of myosin light chain, which would lead to enhanced contractile activity of myometrial smooth muscle cells (32). This proposed model elegantly integrates P4 actions mediated by membrane and nuclear PR isoforms and explains how high P4 concentrations could serve to promote the onset of labor.
However, our findings do not lend support to the concept that mPRs are plasma membrane receptors that mediate P4 actions on cAMP production and MAPK activation. First, we could not confirm that mPRs localize to the plasma membrane. By immunofluorescence, all isoforms were found to reside primarily in the endoplasmic reticulum, and there was little or no overlap in immunoreactivity upon double staining for endogenous hmPR and a plasma membrane marker. Arguably, residence of mPRs in the endoplasmic reticulum would not preclude activation by P4, as steroid hormones freely enter cells due to their lipophilic nature. However, it makes it difficult to interpret the findings of Karteris et al. (32), who used a membrane-impermeable P4 conjugate (P4-BSA) to elicit mPR-mediated events. These included an increase in the phosphorylation of p38, a decrease in cAMP formation, an increase in myosin light chain phosphorylation and a reduction in SRC2 expression in cultured human myocytes. The latter two effects were abrogated by silencing of hmPR expression by small interfering RNA and therefore attributed to the action of this receptor isoform (32). Although it is conceivable that the P4-BSA preparation was contaminated with sufficient free P4 (48) to elicit these responses upon binding to mPRs in the endoplasmic reticulum, Karteris et al. (32) further showed surface binding of fluorescently labeled P4-BSA conjugate (P4-BSA-fluorescein isothiocyanate) to cultured myocytes. They also raised antibodies against mPR and mPRß. In Western blot analyses, these antibodies detected bands of 40 and 80 kDa in MDA-MB-231 cells transfected with human mPR and mPRß, which were interpreted as mono- and dimeric forms. Surprisingly, the immunoreactive bands were detected in plasma membrane extracts of transfected, but not of nontransfected MDA-MB-231 cells, although we found that this cell line expresses substantial amounts of endogenous hmPR. These antibodies were also used for indirect immunofluorescent detection of hmPR and ß in myometrial tissue sections and cultured myocytes. The pattern of immunoreactivity was interpreted to show plasma membrane localization of the mPRs, but no dual staining with organelle markers was performed to support this notion (32).
Second, Zhu et al. (20) reported that P4 inhibits adenylyl cyclase activity in MDA-MB-231 cells transfected with stmPR, but we found no such effect with any of the human isoforms. We reasoned that this apparent discrepancy could be due to species-specific differences in mPR signal transduction. To test this, we recapitulated the experimental design of Zhu et al. (20) by stably transfecting MDA-MB-231 cells with expression vectors for the seatrout or the closely related fugu mPR. However, we failed to observe an effect of P4 on cAMP production. As mentioned, hmPR is expressed endogenously in MDA-MB-231 cells raising the question why cAMP levels did not change in the mock-transfected cells in the study by Zhu et al. (20) or in our parental wt MDA-MB-231 cells upon P4 treatment.
Third, previous studies described P4-dependent activation of ERK1/2 through transfected stmPR and of p38 MAPK through endogenous human mPR and/or mPRß (20, 32). Again, in our hands P4 did not activate ERK1/2 or p38 MAPK in MDA-MB-231 cells expressing mPR of seatrout, fugu, or human origin, or in HEK293 cells expressing the human mPR, ß, or isoforms. Notably, the evidence implicating hmPR and/or hmPRß in P4-BSA-dependent p38 phosphorylation in myocytes was circumstantial (32), raising the possibility that other P4-binding plasma membrane moieties could account for this effect. For instance, the components of a membrane progestin binding complex, PGRMC1 and PAIRBP1, have been found to be rather ubiquitously expressed in rat tissues, at least at the transcript level (23), and we found PAIRBP1 and PGRMC1 to be expressed in cultured human myometrial cells (our unpublished observations).
Finally, in contrast to previous reports (20, 32), we found no evidence that P4 induces G protein-dependent signaling of mPRs, irrespective of isoform or species of origin. Employing the highly promiscuous chimeric Gq5i protein in a Ca2+ reporter assay, we not only observed a lack of G protein activation upon P4 treatment but did also not detect Ca2+ mobilization mediated by any of the mPRs. This is remarkable because P4 is known to stimulate Ca2+ influx in spermatozoa in the process of capacitation and acrosome reaction (4, 49), and a study on sperm from the Atlantic croaker (Micropogonias undulatus) had suggested an involvement of mPR (50). Recently, ovine mPR, when overexpressed in CHO cells, was found to mediate Ca2+ mobilization in response to 1 nM P4 (51). It remains to be determined whether this discrepancy to our data reflects a species-specific difference, or is due to differences in methodology or the doses of steroid used.
The mPRs form a subgroup of the newly established PAQR family of multispan transmembrane proteins (21). It is still controversial if they are true classical GPCRs, not only based on functional analyses as discussed above, but also on structural criteria. For instance, although some of the family members are predicted to have seven putative TMs, they bear little sequence homology to well-characterized GPCRs and carry no obvious N-terminal signal peptide sequence (48). By Western blot analysis of N- and C-terminally tagged human mPR, ß, or , we also found no evidence of signal peptide cleavage.
To date, experimental confirmation of structural predictions is lacking for any of the vertebrate PAQR proteins (21). In contrast to the proposed model of the seatrout mPR (20), the N terminus of the human homolog was predicted to be intracellular by TopPred II. Such reverse orientation has also been reported for two other members of the PAQR family, the adiponectin receptors AdipoR1 and AdipoR2 (PAQR1, PAQR2). In nonpermeabilized fixed cells, only the C-terminal tag of transfected receptors was detectable by immunofluorescence, whereas upon permeabilization, the N-terminal tag became accessible (52). These results, however, may need to be reevaluated in light of the fact that in this study a truncated form of AdipoR2 was used that lacked 87 residues at the N terminus (21). Of note, a yeast homolog of the PAQR family, the product of the YOL101c locus in Saccharomyces cerevisiae, was included in another study that combined prediction and experimental confirmation to determine the topology of 37 membrane proteins. The result suggested 7 TM regions and an extracellular C terminus for the YOL101c product, i.e. inside-out orientation as described above for the adiponectin receptors (53).
Our detailed immunofluorescent studies revealed that both the N and C terminus of hmPR are intracellular. By dual staining with organelle markers, we demonstrated predominant localization of hmPR to the endoplasmic reticulum. Likewise, ovine mPR has been localized to the endoplasmic reticulum, at least when overexpressed in CHO cells (51). Of note, hmPRß has originally been cloned as the brain-specific lysosomal membrane protein C6orf33 (54), again pointing to an intracellular membranous localization of endogenous mPR.
Interestingly, a similar controversy on the subcellular localization and signal transduction coupling surrounds another recently identified putative membrane steroid receptor, GPR30 (55, 56, 57). Whereas one group characterized this molecule as a plasma membrane-bound estrogen receptor coupled to a stimulatory G protein, another study localized GPR30 to the endoplasmic reticulum and showed that it mediates estrogen-induced calcium mobilization and nuclear phosphatidyl inositol-3,4,5-triphosphate accumulation (58, 59). With regard to our own study, we are aware that overexpressed plasma membrane-spanning molecules may be artifactually retained in the endoplasmic reticulum if chaperones or heteromerization partners required for proper folding and/or trafficking are limiting or lacking in a particular cell type (39). However, like GPR30 in the latter study, hmPR displayed predominant localization to the endoplasmic reticulum not only when transfected but also as an endogenous protein.
Another issue which has not been exhaustively resolved is the ligand-specificity of the mPRs. Zhu et al. (19, 20) had used bacterially expressed recombinant mPRs to determine the steroid binding characteristics. It is however unusual that the highly hydrophobic mPRs retained their structural features and binding abilities when extracted as recombinant proteins from Escherichia coli (48). Crude membrane preparations from CHO cells transfected with ovine mPR have been reported to bind P4 and 17-OHP, but not other steroids (51). The only, albeit indirect, evidence for P4 binding to human mPRs in a homologous cellular background was based on a single point radioreceptor assay using membrane preparations from human myocytes. Binding of tritiated P4 was reduced in the presence of excess cold P4 and upon transfection of short interfering RNAs to hmPR and ß (32). We assessed the binding characteristics of mPRs more directly by performing radioligand assays and competition experiments on cells expressing specific receptor isoforms but failed to detect displaceable binding of P4 exceeding that seen in the corresponding wt cell lines. Notably, lack of specific mPR-mediated P4 binding was observed both in membrane preparations and in intact cells and was independent of the mPR isoform, the species of origin, or the presence of a protein tag.
Taken together, our data imply that the mPRs are intracellular orphan receptors at best, or taking it farther, that they may not be receptors at all. Yet we believe that they are important molecules, considering the striking conservation of the PAQR family from eubacteria, yeast and plants to Homo sapiens (21, 60). Their physiological role in fishes and mammals clearly needs further evaluation. In yeast, a PAQR homologous subgroup of four proteins has been identified that were designated IZH1–4 (implicated in zinc homeostasis) (61). IZH2 and IZH4 are encoded by the YOL002c and YOL101c loci, respectively (53). The IZH proteins are involved in zinc homeostasis and sterol metabolism, and IZH2 has also been implicated in lipid and phosphate metabolism (61, 62). Importantly, the IZH proteins share highly conserved metal-binding motifs with human AdipoR1 and mPR, ß, and (61). Furthermore, these conserved residues are also found in hemolysin III from Bacillus cereus, which is involved in formation of transmembrane oligomeric pores in erythrocytes and constitutes the founding member of the hemolysin III-related subgroup of PAQR proteins (22, 61). Meanwhile, the functions of the mPRs remain unknown, and elucidation of their true physiological roles may have to await the generation of mouse models carrying single or combined ablations of the encoding genes.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Generation of Expression Vectors
Human mPR
The human mPR cDNA (pos. –97/+1079 relative to the start ATG, GenBank accession no. AF313620) was amplified by RT-PCR from T47D human breast cancer cells and cloned into pCR-BluntII-TOPO (Invitrogen, Karlsruhe, Germany) to yield pCR-Blunt/hmPR. The insert was excised with BamHI and XhoI and inserted into the respective sites of pcDNA3.1(+) (Invitrogen) to yield pcDNA/hmPR. The hmPR cDNA without start ATG was amplified from pCR-Blunt/hmPR with a sense primer mutating the ATG codon and introducing an ApaI site, and an antisense primer anchored in the 3'-untranslated region (UTR). The PCR product was digested with ApaI and inserted into the ApaI/SmaI sites of a modified pDISPLAY vector (Invitrogen), pDISPLAY-TMD. This had been generated by removal of the transmembrane domain by BsmI/NotI digest, polishing and religation. Insertion of the PCR product into pDISPLAY-TMD yielded pSP-HA/hmPR, carrying the Ig
signal peptide, followed by an HA tag and the hmPR cDNA without start codon. By EcoRV/ApaI restriction, the Ig signal peptide and the HA tag were removed and replaced by a double-stranded oligonucleotide encoding the HA tag only with an added start codon. This double-stranded oligonucleotide had a 5' blunt EcoRV half-site and a 3' ApaI sticky end. The resulting expression vector, pHA/hmPR, encodes hmPR with a 5' HA tag. This full-length insert was excised with EcoRI and BamHI and inserted into the respective sites of pEF-IRES-puro6 to yield pEFIRESp6/HA/hmPR. The bicistronic vector pEF-IRES-puro6 (kindly provided by J. Wallace, University of Adelaide, Australia) is driven by the human EF1 promoter. The multiple cloning site for insertion of the cDNA of choice is followed by an internal ribosomal entry site (IRES) and the puromycin resistance gene for selection of stably transfected cells (63, 64). To generate pEFIRESp6/hmPR for stable transfection of untagged hmPR, the full-length insert was excised from pcDNA/hmPR with BamHI/EcoRV and inserted into the respective sites of pEF-IRES-puro6. To generate an expression vector for hmPR with a 3' V5 and 6xHis tag, PCR was performed on template pCR-Blunt/hmPR with a sense SP6 primer, and an antisense primer mutating the stop codon and introducing a 3' XhoI site. The PCR product was digested with BamHI/XhoI and inserted into the respective sites in pcDNA3.1/V5-HisB (Invitrogen) to yield pcDNA/hmPR/V5. The full-length insert including 3' V5 and 6xHis tags was excised with EcoRI/PmeI and inserted into the EcoRI/EcoRV sites of pEF-IRES-puro6, yielding pEFIRESp6/hmPR/V5 for stable transfection of V5-tagged hmPR.
Human mPRß
The cDNA for C6orf33, which is identical with hmPRß, in pBSSKII(–) was kindly provided by K. Yamakawa and T. Suzuki (RIKEN Brain Science Institute, Saitama, Japan) (54). The cDNA (pos. –164/+1772 relative to the start codon, GenBank accession no. NM_133367) was excised with XbaI/HindIII and inserted into the respective sites of pcDNA3.1(–) (Invitrogen) to yield pcDNA/hmPRß. The EST clone BC030664 in pBluescriptR (MRC Geneservice, Cambridge, UK) was used as template for PCR with a sense T7 primer and an antisense primer mutating the stop codon of hmPRß and introducing a 3' XhoI site. The PCR product was digested with EcoRI/XhoI and inserted into the respective sites in pcDNA3.1/V5-HisB. The resulting vector pcDNA/hmPRß/V5 contains 149 bp of 5'-UTR, full-length hmPRß coding region without stop codon, and 3' V5 and 6xHis tags. C6orf33 in pBSSKII(–) was used as the template for PCR with a sense primer mutating the start codon of hmPRß and introducing a 5' ApaI site, and an antisense primer anchored in the 3'-UTR. The PCR product was digested with ApaI and inserted into the ApaI/SmaI sites of pDISPLAY-TMD to yield pSP-HA/hmPRß. The Ig signal peptide and HA tag were removed from this construct and replaced by an HA tag as described above for hmPR, to yield pHA/hmPRß. The HA-tagged hmPRß cDNA was excised from this vector with EcoRI/BamHI and inserted into the respective sites of pEF-IRES-puro6 to yield pEFIRESp6/HA/hmPRß. The untagged insert from pcDNA/hmPRß was excised with BamHI/PmeI and inserted into the BamHI/EcoRV sites of pEF-IRES-puro6 to generate pEFIRESp6/hmPRß. The hmPRß cDNA with 3' V5 and 6xHis tag was excised from pcDNA/hmPRß/V5 with EcoRI/PmeI and inserted into the EcoRI/EcoRV sites of pEF-IRES-puro6 to yield pEFIRESp6/hmPRß/V5.
Human mPR
The hmPR cDNA (pos. –79/+1061 relative to the start ATG, GenBank accession no. NM_017705) was amplified by RT-PCR from T47D breast cancer cells and cloned into pCR-BluntII-TOPO to yield pCR-Blunt/hmPR. The insert was excised with BamHI/XhoI and inserted into the respective sites of pcDNA3.1(+) to generate pcDNA/hmPR. Construct pCR-Blunt/hmPR was used as the template for PCR with a sense primer mutating the hmPR start codon and introducing a 5' ApaI site, and an antisense primer anchored in the 3'-UTR. The PCR product was digested with ApaI and ligated into the ApaI/SmaI sites of pDISPLAY-TMD to yield pSP-HA/hmPR. From this vector, the Ig signal peptide and HA tag were removed and replaced by an HA tag as described above for hmPR, to generate pHA/hmPR. The insert including the 5' HA tag was excised with EcoRI/BamHI and inserted into the respective sites in pEF-IRES-puro6 to yield pEFIRESp6/HA/hmPR. The insert from pcDNA/hmPR was excised with BamHI/EcoRV and inserted into the respective sites of pEF-IRES-puro6 to yield pEFIRESp6/hmPR. The insert including 3' V5 and 6xHis tags was recovered from pcDNA/hmPR/V5 with EcoRI/PmeI and ligated into the EcoRI/EcoRV sites of pEF-IRES-puro6 to generate pEFIRESp6/hmPR/V5.
Seatrout mPR
The mPR cDNA from spotted seatrout (C. nebulosus) in pBK-CMV was kindly provided by P. Thomas (University of Texas, Austin, TX). The insert (stmPR) was amplified by PCR with a sense primer mutating the start codon and introducing a 5' ApaI site, and an antisense primer to the 3'-end of the coding region including the stop codon, which introduced a 3' XhoI site. The PCR product was restricted with ApaI/XhoI and used to replace the hmPR insert (ApaI/XhoI fragment) in pHA/hmPR. This resulted in the construct pHA/stmPR, which carries a 5' HA tag and the stmPR coding region from the second amino acid to the stop codon (GenBank accession no. AF262028). The entire insert including HA tag was excised with EcoRI/XhoI, the XhoI site was polished, and the fragment inserted into the EcoRI/EcoRV sites of pEF-IRES-puro6 to yield pEFIRESp6/HA/stmPR.
Fugu mPR
Total RNA from kidney tissue of the Japanese pufferfish (F. rubripes) was obtained from MRC GeneService (Cambridge, UK). Fugu mPR cDNA was amplified by RT-PCR with the following primers:
sense, 5'-GATCAACTGCCCGTCTCATTT-3'; antisense, 5'-TGTGTCTTTACTCCTTCTTCTC-3'. Primer sequences were based on genomic sequence retrieved from the Fugu Genome Project Web page (http://www.fugu-sg.org/), which had been searched for sequences with the closest similarity to stmPR. The cDNA, including 54 bp of 5'-UTR, the full coding region, and 7 bp of 3'-UTR, was ligated into pCR-BluntII-TOPO to yield pCR/fmPR. On this template, PCR was performed with a sense primer mutating the start codon and introducing a 5' ApaI site, and an antisense SP6 primer. The product was restricted with ApaI/NsiI and used to replace the hmPR insert (excised with ApaI/PstI) in pHA/hmPR. This resulted in the construct pHA/fmPR that carries a 5' HA tag and the fmPR coding region from the second amino acid on. This insert, including the HA tag, was retrieved by EcoRI digestion and inserted into the EcoRI site of pEF-IRES-puro6 to yield pEFIRESp6/HA/fmPR. A partial sequence of the fugu mPR mRNA can now also be found on the Fugu Genome Project Web page (Ensembl Transcript ID SINFRUT00000149401).
All inserts were verified by sequencing. Sequences of primers used for PCR cloning are available upon request
Generation of Clonal Cell Lines Stably Transfected with mPR Expression Constructs
For stable transfection, HEK293 cells were transfected with hmPR expression plasmids in pEF-IRES-puro6 (encoding hmPR, hmPRß or hmPR without tag, with 5' HA tag or with 3' V5 tag) using PolyFect reagent (QIAGEN, Hilden, Germany) according to the supplier’s instructions. Two days after transfection medium was changed and transfected cells were selected with 3 µg/ml puromycin. The resultant surviving clones were picked with cloning discs (Sigma, Deisenhofen, Germany) and expanded. MDA-MB-231 cells were transfected with mPR expression plasmids in pEF-IRES-puro6 (encoding hmPR, stmPR, or fmPR with 5' HA tag) using PolyFect reagent (QIAGEN), and isolation of clones was done in the presence of puromycin at 1 µg/ml. Expression levels of tagged mPRs in selected HEK293 and MDA-MB-231 clones were monitored by Western blot analysis, and clonality of cell lines was verified by immunofluorescence with HA or V5 antibody, respectively (see Supplemental Materials and Methods, published as supplemental data on The Endocrine Society’s Journals Online web site at http://mend.endojournals.org). HEK293 cells stably expressing untagged hmPR (HEK293-hmPR) were examined by immunofluorescent analysis using a specific antipeptide antibody validated in our previous study (22). However, this antibody does not allow detection of tagged or untagged hmPR
