| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Obstetrics and Gynaecology,
2 Institute Growth and Development (GROW), University Hospital of Maastricht, Maastricht, The Netherlands
(Requests for offprints should be addressed to A Romano; Email: a.romano{at}og.unimaas.nl)
(D-C Fischer is now at Department of Pediatrics, Experimental Nephrology, Rostock University Medical Centre, Rostock, Germany)
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
All major actions of progesterone are mediated by the two progesterone receptor (PR) isoforms (PRA and PRB), which are ligand-dependent transcription factors (Beato & Klug 2000, Li & O’Malley 2003). PRB differs from PRA by the presence of 164 additional amino acids at the N-terminus. The remaining part is identical for both proteins and consists of the DNA-binding domain (DBD) and the C-terminal ligand-binding domain (LBD). A hinge region is located between DBD and LBD. This region is involved in nuclear translocation, dimerisation and cofactor binding, but its function in large part remains unclear (Beato & Klug 2000, Li & O’Malley 2003). After ligand binding and formation of hetero- or homodimers, PRA and PRB bind to progesterone-response elements (PRE) located in the promoter region of target genes and modulate their expression. Studies on cell lines (Richer et al. 2002) and animal models (Lydon et al. 1995, Mulac-Jericevic et al. 2000, 2003, Conneely et al. 2003) have shown that PRA and PRB have distinct functional activities.
The PR gene, located on chromosome 11q22–23, comprises eight exons and seven introns (A–G). One PR polymorphic variant, PROGINS (Rowe et al. 1995, Agoulnik et al. 2004) consists of a 320 bp PV/HS-1 Alu insertion in intron G and two single nucleotide polymorphisms (SNPs). SNP-G3432T (mRNA nucleotide counting, NCBI: X51730) affects exon 4 and causes an amino acid substitution (V660L), whereas SNP-C3764T affects exon 5 and is a silent mutation (H770H). These three aberrations are in complete linkage disequilibrium (De Vivo et al. 2002, Agoulnik et al. 2004, Schweikert et al. 2004, Pijnenborg et al. 2005) and a plethora of epidemiological studies indicate that PROGINS might modify the risk for several benign and malign gynaecological disorders (Dunning et al. 1999, Lattuada et al. 2004, Modugno 2004, Schweikert et al. 2004, Pijnenborg et al. 2005, Romano et al. 2006). However, the results are not conclusive.
Considering the possible consequences of each of the three PROGINS aberrations, only the Alu insertion and the V660L amino acid substitution are expected to affect the function of the receptor. In fact, RNA processing is not influenced by the two SNPs (G3432T in exon 4 or the C3764T in exon 5). Preliminary algorithm analyses to predict the secondary structures of RNA transcripts containing the two PROGINS SNPs showed that neither RNA folding, nor the binding sites for ribonucleoprotein complexes involved in mRNA maturation are affected by these two SNPs. Furthermore, the silent mutation (H770H) is not expected to alter the properties of the PR protein. Several authors have indicated that Alu insertions can be of functional relevance with respect to CpG islands DNA methylation, gene transcription and binding of transcription factors (Deininger & Batzer 1999, Gibbons & Dugaiczyk 2005) and the generation of splice variants due to Alu exonisation (Deininger & Batzer 1999, Lev-Maor et al. 2003, Sorek et al. 2004). Thus, the intron G Alu insertion could affect transcription and pre-mRNA maturation, including splicing, whereas the V660L amino acid substitution could only affect the activity of the protein.
Given that PROGINS occurs at frequencies up to 0.15 in Caucasians of different origins, the functional characterisation of this polymorphism is important to clarify its role in the pathogenesis. To this end, we have investigated the individual consequences of the two most likely relevant gene and protein aberrations associated with the PROGINS, the intronic Alu insertion and the V660L amino acid substitution, with respect to their effects on RNA maturation, protein activity and proliferation using different in vitro approaches.
In the following part of this study, the most common polymorphic allele or protein variant (i.e. no Alu insertion and V660) will be indicated as CP allele or protein.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
Human cervical cancer cell line HeLa, human ovarian cancer cell lines MDAH:2774 and Skov-3, human breast cancer cell lines T47D and MCF7 and Chinese hamster ovarian cell line CHO-K1 were purchased from the American Type Culture Collection (ATCC; Rockville, MD, USA). With the exception of T47D cells, none of the cell lines mentioned above express the PR. MCF7 cells are homozygous PROGINS, whereas T47D cells are heterozygous. The remaining cell lines are homozygous CP. Cells were maintained in Roswell Park memorial Institute medium (RPMI; Invitrogen Life Technologies, Inc.) supplemented with 10% foetal bovine serum (Hyclone, Gent, East Flanders, Belgium), 2 mM L-glutamine, 2 mM Na pyruvate, 100 U/l penicillin and 100 mg/l streptomycin).
Steroid hormones
Progesterone, org2058 and the antagonists (RU486 or org31710) were a gift from Organon NV (Oss, Brabant, The Netherlands). 17ß-Oestradiol and tamoxifen were purchased from Sigma-Aldrich Chemie BV. All hormones were dissolved in ethanol at a stock concentration of 10–3 M.
Oligonucleotides
Oligonucleotides used for PCR amplification were purchased from MWG-Biotech AG (Ebersberg, Bayern, Germany) and are listed in Table 1
.
|
Plasmid for the expression of the reporter gene ß-galactosidase (pSV-ß-galactosidase) was purchased from Promega GmbH. PR-responsive reporter plasmid MMTV-luciferase (MMTV-LUC; Muller et al. 2000) was a gift from Prof. Schuele. Expression plasmids for PRA-V660 (CP), PRA-L660 (PROGINS protein variant), PRB-V660 and PRB-L660 (Agoulnik et al. 2004) were based on plasmid pcDNA3.1 (Invitrogen Life Technologies, Inc.), and were kindly provided by Dr Agoulnik. Minigenes for the expression of PRA-IntG-CP, PRB-IntG-CP and PRA-IntG-Alu, PRB-IntG-Alu were constructed by re-inserting the complete intron G derived either from the CP allele (2672 nucleotides) or from the PROGINS (2992 nucleotides) allele at its original genomic position in the cDNAs of PRA and PRB. The complete intron G was cloned by PCR in four steps from genomic DNA isolated from cell lines Skov-3 (CP homozygous) and MCF7 (PROGINS homozygous) as templates. The Expand High Fidelity Kit (Roche) was used for PCR as recommended by the manufacturer. A detailed description of all cloning steps is given in Fig. 1. Figure 2A shows the structure of the intron G and relevant restriction sites used for cloning.
|
|
Stable and transient transfection
Stable transfection of CHO-K1 cells with either one of the plasmids encoding PRA-V660, PRB-V660, PRA-L660 or PRB-L660 were performed using jetPEI (Q-Biogene, Heidelberg, Baden-Württemberg, Germany) as recommended by the manufacturer. For the selection of stably transfected clones, transfected cells, as well as their untransfected counterpart, were plated in RPMI supplemented with 5 mg/l blasticidin-S-HCl (Invitrogen, Life Technologies, Inc.) 3 days after transfection. After 2 weeks of incubation with blasticidin-containing medium, only transfected cells were viable. Monoclonal cells resistant to blasticidin were obtained by plating serial dilutions (0.01–10 cells/well) on 96-well plates, isolating those colonies derived from a single cell (evaluated by microscopic observation) and subsequent sub-cloning of those clones that showed PR expression (western blotting). Cell lines derived from transfection with plasmid PRA-V660 were labelled CHO-A series, PRB-V660 lines CHO-B, PRA-L660 lines CHO-C and PRB-L660 lines CHO-D. Transient transfection of HeLa, MDAH:2774, Skov-3 and MCF7 cells as well as of maternal and stably transfected CHO-K1 was performed using jetPEI (Q-Biogene) as recommended by the manufacturer.
Genomic DNA (gDNA) extraction
One million cells were lysed in 300 µl DNA cell lysis buffer (200 mM Tris–HCl (pH 8.0), 250 mM NaCl, 25 mM EDTA, 0.5% SDS). DNA was extracted with one volume of TE-equilibrated phenol (Amresco, Solon, Ohio, USA) and precipitated with 0.7 volumes of isopropanol. Genotype screening of the mutation V660L substitution was performed by PCR–RFLP using primers exon-4 S and exon-4 AS (Pijnenborg et al. 2005). The presence of the 320 bp PV/HS Alu insertion in intron G was diagnosed by PCR using primers PROGINS-I and PROGINS-II (Wang-Gohrke et al. 2000).
RNA extraction and cDNA synthesis
RNA was isolated using the Trizol reagent (Invitrogen, Life Technologies, Inc.) as recommended by the manufacturer. cDNA was synthesised using the M-MLV reverse transcriptase (Invitrogen, Life Technologies, Inc.) and poly-T oligonucleotide (MBI Fermentas GmbH, St Leon-Rot, Baden-Württemberg, Germany) for priming. Five micrograms of total RNA were used for cDNA synthesis. Prior to cDNA synthesis, RNA was treated with 3 U DNaseI (MBI Fermentas GmbH). To check the expression of the PR gene, semi-quantitative RT-PCR was performed using primers PR-GA-FW and PR-GD-RV. Primer pairs exon-4 S/exon-4 AS as well as exon-4 S/exon-5 AS were used to estimate the relative expression of CP PR and PROGINS alleles by utilising PCR–RFLP analysis on cDNA. For Alu exonisation, forward primer PR-GA-FW or exon-4 S was used in combination with the reverse primer Alu-Ex-RV. Expression of the oestrogen-responsive PS2 gene was investigated with primer pair pS2-2/pS2-3. Expression of the RASSF1 gene was evaluated using primers RSF-3 and RSF-4. For PCR normalisation, expression of the cyclophilin A gene (housekeeping) was analysed using CycloA S/CycloA AS for amplification of the corresponding cDNA.
Chromatin immunoprecipitation (ChIP)
ChIP was performed as described elsewhere (Shang et al. 2000). Briefly, T47D cells were cultivated to 80% confluence (75 cm2 culture flasks) in RPMI without phenol red (Invitrogen, Life Technologies, Inc.) supplemented with 10% hormone-stripped serum (c.c.pro GmbH, Neustadt, Saxony, Germany) for 3 days. Cells were treated with vehicle (ethanol) or with 17ß-oestradiol (10–8 M) for 50 min. Subsequently, cells were fixed (RPMI with 1% formaldehyde, 10 min at 37 °C) washed twice with cold PBS and scraped in 1 ml cold PBS supplemented with Complete protease inhibitor (Roche). Cells were lysed in 10 mM HEPES (pH 6.5), 10 mM EDTA, 0.25% Triton X-100 supplemented with complete protease inhibitor. Nuclei were pelleted and subsequently lysed in 300 µl lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris–HCl (pH 8.1) plus Complete protease inhibitor) and sonicated. Debris was pelleted, aliquots of the nuclear lysates (50 µl) were diluted ten times in immunoprecipitation (IP) buffer (1% Triton, 2 mM EDTA, 150 mM NaCl, 20 mM Tris–HCl (pH 8.1)) and pre-cleared (2 h, 4 °C) with 100 µl protein-G beads (Active motif, Rixensart, Brabant, Belgium). For IP of chromatin–protein complexes, specific antibodies (2 µg) were added and the tubes were continuously rotated during incubation (overnight, 4 °C). For IP of ER-
, RNA Pol-II and Sp1-containing chromatin complexes, antibodies F10 (Santa Cruz Biotechnology, Santa Cruz, CA, USA), 4H8 and Sp1 (Active motif) were used respectively. Antibodies bound to chromatin complexes were precipitated by the addition of protein-G beads (100 µl, 2 h, 4 °C). Beads were washed six times (100 mM Tris–HCl (pH 8.0) containing 500 mM LiCl, 1% NP-40, 1% deoxycholic acid) and chromatin–protein complexes were eluted from the beads with 100 µl 25 mM NaHCO3 supplemented with 1% SDS. Crosslinking was reversed by the addition of 4 µl 5 M NaCl supplemented with 10 U RNaseA and incubation overnight at 65 °C. Subsequently, the samples were treated with proteinase-K (400 mg/l) in 20 mM Tris–HCl (pH 6.5) and 10 mM EDTA for 2 h at 42 °C and DNA was purified using the Qiaquick reaction cleanup kit (Qiagen GmbH). Binding of the RNA Pol-II to the GAPDH promoter was used as positive control of the ChIP procedure, and it was assessed using primers ChIP-pos-FW and ChIP-pos-RV. ER- binding to the PS2 promoter was used as a positive control for ChIP with ER- and it was assessed using primers PS2-0 and PS2-1. For ER- and Sp1 binding to the Alu-ERE/Sp1 site, primers Alu-ERE-Sp1-FW and PROGINS-II were used (Table 1
and Fig. 2B). ChIP PCR signals were normalised with an unspecific negative control, using primers ChIP-neg-FW and ChIP-neg-RV that flank cytogenetic location 12p13.3 where no transcription factors bind to.
5-aza-2'-deoxycytidine (ADC) treatment
Cells were plated in six-well plates at 50% confluence and treated for 0, 48, 96 and 120 h with 100 µM ADC (Sigma-Aldrich Chemie BV). ADC causes DNA demethylation. At the end of the treatment, RNA was isolated as described earlier. Medium containing ADC was refreshed every 24 h.
RNA stability assay
Native T47D cells or MDAH:2774 cells transiently transfected with the PR minigenes were plated in six-well plates at 70% confluence. One day later, actinomycin-D (Sigma-Aldrich Chemie BV) as an inhibitor of RNA synthesis was added to the medium (6 µM final concentration) and cells were cultured for the indicated period of time prior to isolation of RNA.
Immunocytofluorescence
Cells were cultured on glass coverslips (six-well plates, overnight) and subsequently fixed in buffered formaldehyde (4% paraformaldehyde in PBS). Cells were washed with PBS and permeabilised with 0.1% Triton X-100 diluted in PBS. To reduce unspecific binding, cells were incubated with PBS containing 1% BSA and 1% goat serum (1 h, room temperature). To detect PR expression, cells were incubated with the PR-specific antibody AB-52 (Santa Cruz Biotechnology). Bound antibody was visualised using Cy-3-labelled goat anti-mouse antibodies (Jackson Immunoresearch, Cambridge, UK).
Transactivation assay
Cells were cultivated for at least 3 days in RPMI without phenol red (Invitrogen, Life Technologies, Inc.) supplemented with 10% hormone-stripped serum (c.c.pro GmbH). One day before transfection, cells were plated in 12-well plates at 40–50% confluence. In each well, 1 µg MMTV-LUC reporter plasmid, 800 ng of either PR-L660, PR-V660 or pcDNA3.1 (empty vector; Invitrogen, Life Technologies, Inc.) together with 200 ng pSV-ß-galactosidase were co-transfected. In all experiments, ß-galactosidase was used for normalisation of luciferase measurements with transfection efficiency. Twenty-four hours after transfection, the culture medium was replaced and supplemented with the ligand as indicated in Fig. 3. After an additional 24-h incubation, cells were lysed in 60 µl 100 mM Tris–HCl (pH 7.8), 0.5% Triton X-100 and 40 µl extract were assayed for luciferase activity by adding one volume of Steady-GLO substrate (Promega GmbH). The remaining extract was used for ß-galactosidase assay and western blot analysis. For ß-galactosidase assay, 5–10 µl extract were incubated in 96-well plates with 180 µl PBS supplemented with 0.7 mg/ml of o-nitrophenyl-ß-galactopyranoside (Sigma-Aldrich Chemie BV) and 0.4% ß-mercaptoethanol (2–4 h, 37 °C). ß-Galactosidase activity was measured spectrophotometrically at 450 nm. For western blot, replicates from one single experiment were pooled and western blot was performed as described below. Each data point was tested in triplicate using multiple plasmid preparations of PR-expressing constructs and significance of results was confirmed in at least two independent experiments.
|
Variations in the three-dimensional structure of a protein can be visualised by partial trypsin digests. Changes in the tertiary structure potentially mask or expose cutting sites for trypsin, which is known to cut after the amino acid arginine. Cells stably expressing PRA-V660 (CHO-A25_2), PRB-V660 (CHO-B34), PRA-L660 (CHO-C4_3) and PRB-L660 (CHO-D9_14) were induced in vitro with vehicle (ethanol) or ligand (progesterone 10–8 M or RU486 10–8 M) for 1 h. Cell extracts were prepared in PBS and hormone induction was continued for 20 min at room temperature in silico with the same ligand used in vitro. Equal amounts of extracts (20 µl) were incubated for 10 min at room temperature with increasing amounts of trypsin (from 0 to 25 µg/ml and 0 to 5 µg/ml for PRA and PRB respectively). Reactions were stopped by adding the SDS-PAGE loading dye and boiling immediately. The total volume was used for western blot analysis for PRA and PRB using antibody AB-52. Trypsin (Sigma-Aldrich Chemie BV) was dissolved in water.
Preparation of protein extracts and western blot
Cells were washed with cold PBS and lysed in RIPA buffer (10 mM sodium phosphate (pH 7.0), 150 mM NaCl, 2 mM EDTA, 1% Na deoxycholate, 1% NP-40 and 0.1% SDS) supplemented with Complete protease inhibitor (Roche). Proteins were separated by SDS-PAGE (7–10% polyacrylamide) and blotted on a nitrocellulose membrane (Protran, Schleicher and Schuell, Germany). PR was visualised using the AB-52 monoclonal antibody (Santa Cruz Biotechnology) or PR-A0098 polyclonal (Dako, Glostrup, Denmark). ER- and Sp1 were visualised with antibody F10 (Santa Cruz Biotechnology) and Sp1 (Active motif) respectively. Antibody AC-15 (Sigma-Aldrich Chemie BV) was used to detect ß-actin. For visualisation of bound antibodies, HRP-conjugated rabbit anti-mouse antibodies (DAKO, Netherlands BV, Eindhonen, Brabant, The Netherlands) and goat anti-rabbit antibodies (Pierce, Aalst, Flanders, Belgium) in conjunction with the super signal-R West-Femto kit (Pierce) were used. The proteasome inhibitor, MG132, was purchased from Biomol (Exeter, UK).
For determination of the protein concentration, the Bradford dye reagent (BCA assay; Uptima, Interchim, Montluçon, Allier, France) was used. Alkaline phosphatase was purchased from Fermentas GmbH.
Cell viability
Cell viability, which is a measure for the total number of cells, was determined with the MTT assay (3-[4,5-dimethylthiazole-2-yl] 2,5-diphenyl tetrazolium bromide (Sigma-Aldrich Chemie BV), as described previously (You et al. 2001).
Cell proliferation
Proliferative activity was assessed by determining 5-bromo-2-deoxyuridine (BrdU) incorporation. Cells were cultured for 3 days in RPMI without phenol red (Invitrogen, Life Technologies, Inc.) supplemented with 10% hormone-stripped serum (c.c.pro GmbH). Subsequently, cells were plated on six-well plates at 50% confluence and cells were synchronised by incubation for 20 h in RPMI (without serum) supplemented with nocadozole (20 µg/ml; Sigma-Aldrich Chemie BV). Cells were washed twice with PBS and cultured in RPMI supplemented with 10% hormone-stripped serum and containing 10 µM BrdU (Sigma-Aldrich Chemie BV). Forty minutes after the addition of BrdU-containing medium, cells were fixed (in methanol) to measure the labelling index. Afterwards, cells were treated with org2058 or vehicle and fixed after 8, 24 and 36 h of culture to measure BrdU incorporation. Fixation of cells and measure of the BrdU incorporation by flow cytometric analysis (FACSort Beckton and Dickinson, Sunnyvale, CA, USA) were performed as previously described (Schutte et al. 1997).
Image quantification analyses
For all agarose gel and western blot band quantification, the ImageJ programme was used (ImageJ 1.32j, Wayne Rasband, National Institutes of Health, USA, http://rsb.info.nih.gov/ij/).
GenBank accession numbers
PR mRNA accession number: NCBI X51730. PR mRNA G3432T and C3764T: NCBI AF016381 [GenBank] . The accession number of the PR partial gene and the Alu insertion into intron G is NCBI Z49816. [GenBank] SNP accession numbers: SNP-G3432T: Reference SNP Cluster Report (ref SNP ID): rs1042838; SNP-C3764T: ref SNP ID: rs1042839. Chromosome 11 complete sequence accession number is NC000011.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
Mature mRNA does not contain the polymorphic intron G (with or without the Alu element). Nevertheless, the mRNA transcribed from the PROGINS allele can be distinguished from the CP transcript because it still contains the two SNPs in exon 4 (G3432T) and in exon 5 (C3764T) that are in linkage disequilibrium with the intronic Alu insertion. Using cell lines heterozygous for PROGINS, the relative abundance of the PROGINS and CP mRNAs (cDNA) can be assessed based on the PCR–RFLP method used to diagnose the SNP in exon 4 in epidemiological studies (Pijnenborg et al. 2005; Fig. 4). For further analyses, the human breast cancer cell line T47D, which is heterozygous for PROGINS (Fig. 4A, lanes 3 and 4) and expresses the PR protein (Fig. 4B, lane 1), was selected.
|
Alu elements have high CpG content that can be methylated and thus reduce gene transcription (Deininger & Batzer 1999). The PROGINS allele in cell line T47D is transcribed with reduced efficiency compared with the CP allele (Figs 4A, lane 5 and 4C). To investigate whether this reduced transcription is due to CpG-Alu methylation, T47D cells were treated with ADC, which demethylates DNA. The relative expression of the PROGINS versus the CP allele was assessed by PCR-RFLP on cDNA and did not vary after ADC treatment for different periods of time (0, 48, 96 and 120 h; Fig. 4C and D).
The Alu element in intron G increases in-cis the transcription of the PR-PROGINS allele
The Alu element in intron G of the PROGINS allele (Fig. 2A and B) contains the site TGA CCn nnn nnn CCG CCC (Alu-ERE/Sp1), located 1132 nucleotides downstream the 5' splice site of intron G. This site is similar to the half-ERE/Sp1 transcription site (Fig. 2B, bottom) located in the promoter region of the PR gene and responsible for the oestrogen/Sp1-stimulated transcription of the PR gene (Petz & Nardulli 2000). We hypothesised that the Alu-ERE/Sp1 site may affect the transcription rate of the PROGINS allele. Since T47D cells express high levels of ER- and Sp1 transcription factor (Fig. 4B, lanes 2 and 3), these cells were treated with 17ß-oestradiol for different periods of time. Subsequently, the relative abundance of mRNA derived from the CP and the PROGINS alleles were assessed by cDNA-PCR–RFLP (Fig. 2C). After incubation with 17ß-oestradiol (10–8 M) for 1 and 3 h, the relative amount of PROGINS derived mRNA gradually increases from 21 to 53% compared with the mRNA derived from the C P. PR gene. Furthermore, ChIP indicated that ER- and Sp1 bind to the Alu-ERE/Sp1 site (Fig. 2D). Whereas Sp1 transcription factor appears to bind independently of the presence of 17ß-oestradiol, the binding of ER- to this sequence is ligand dependent.
The Alu element reduces the stability of the transcript
Mobile elements like Alus transpose via a RNA intermediate that is transcribed by the RNA polymerase III (RNA Pol-III) and large differences have been reported on the stability of Alu RNAs belonging to different classes of Alu elements (Aleman et al. 2000, Li & Schmid 2004). However, nothing is known on the influence that intronic Alu elements might have on pre-mRNAs, in which they are co-transcribed by the RNA Pol-II. The folding of the PR-pre-mRNA with or without the Alu insertion was predicted using the ‘Vienna RNA Secondary Structure Prediction’ algorithm (http://rna.tbi.univie.ac.at). Differences were seen in the free energies and the predicted structures of the two RNA molecules (not shown). To check whether these predicted differences affect the half-life of the corresponding RNAs, we compared the stability of the two PR transcripts in T47D cells (Fig. 5A). Cells were treated with actinomycin-D, an inhibitor of RNA synthesis. Three and eighteen hours of treatment with actinomycin-D (6 µg/ml) results in 50 and 80% decrease of the transcript generated from the PROGINS allele compared with the one generated from the CP allele. (Fig. 5B; The amounts of CP and PROGINS cDNAs shown in the chart, do not represent the relative amount of the two molecules like in Figs 4 and 2, but are both fixed to 100 to clearly show the degradation rates.) To confirm that the observed trend was not a consequence of other known (such as SNPs in exons 4 and 5) or unknown genetic variations in linkage disequilibrium with the PROGINS-Alu insertion, human ovarian cancer MDAH:2774 cells (PR negative) were transfected with minigenes encoding the PR CP and PROGINS variant (PRA-IntG-CP, PRA-IntG-Alu, PRB-IntG-CP or PRB-IntG-Alu). These minigenes, differing exclusively with respect to the presence of the Alu element in intron G, were correctly transcribed and translated (not shown). After the treatment of MDAH:2774 cells with actinomycin-D for 18 h, we observed a reduction of approximately 20% of the amount of RNA generated from minigenes bearing the PROGINS intron G compared with the amount of RNA generated from minigenes expressing PR with the CP intron G (Fig. 5B). Although this reduction was less important compared with what we observed in T47D cells (see discussion for one possible explanation), it was seen for PRA and PRB.
|
Alu elements contain 3' and 5' splice consensus sites (squared in Fig. 2B). When Alus are integrated into introns of actively transcribed genes, they can generate alternative splice variants by ‘exonisation’ of the approximately 100 nucleotides comprised between the 3' and 5' splice sites (Deininger & Batzer 1999, Deininger et al. 2003, Lev-Maor et al. 2003, Sorek et al. 2004). To test whether the PROGINS Alu is ‘exonised’, cDNA from native cell line T47D and cell line Skov-3 or MDAH:2774 transiently transfected with the PR minigenes (PRA-IntG-CP, PRA-IntG-Alu, PRB-IntG-CP or PRB-IntG-Alu) was analysed by PCR. When using primer pair overlapping the exon 7/exon 8 boundary and annealing inside the PR coding sequence, no product larger than the correctly spliced PR cDNA was amplified. To further exclude the possibility that the cDNA form containing the ‘exonised’ Alu would not compete with the normally spliced PR cDNA because of too low expression, primers specific for the ‘exonised’ Alu were used (one primer, Alu-Ex-RV, anneals at the putative 5' site for alternative splicing in the Alu element, underlined in Fig. 2B). However, no product could be amplified although the non-spliced PCR product derived from the minigene containing the complete intron G+Alu element could be efficiently amplified (not shown).
Taken together, on the one hand, the Alu insertion in intron G moderately increases the transcription rate of the PR-PROGINS allele in response to 17ß-oestradiol, while, on the other hand, the presence of the Alu element is accompanied by a decreased stability of the corresponding mRNA. No other effects related to the presence of the Alu insertion were detectable.
Consequences of the V660L amino acid substitution in exon 4
The following part of the study focuses on the consequences of the amino acid V660L substitution. Thus, when we refer to the PROGINS variant in this part of the study (both in the results and in the discussion sections), we consider exclusively the PROGINS protein variant characterised only by the leucine substitution at codon 660, but we do not refer to the intronic Alu insertion.
Generation of stably transfected cell lines
Stably transfected CHO-K1 cells (Chinese hamster ovary) expressing PRA and PRB, CP (V660) and PROGINS protein variant (L660) were generated and four cell lines that had similar expression (western blot analysis and immunocytofluorescence; Fig. 3) of the V660 and L660 variants of PRA and PRB isoforms were selected: CHO-A25_2 (expressing PR-A V660), CHO-C4_3 (PR-A L660), CHO-B34 (PR-B V660) and CHO-D9_14 (PR-B L660). In cell line CHO-A21, the integration of the plasmid into the genomic DNA was confirmed by PCR but no PR expression was observed. Line CHO-A21, together with the maternal cell line CHO-K1, served as negative control.
Transactivation of PR PROGINS protein variant (L660) on an MMTV-luciferase reporter varies in a cell-dependent manner compared with the CP PR
To determine whether the V660L substitution modifies the transactivation of PR, cells stably expressing PRA and PRB, V660 or L660 (see above) were transiently transfected with the MMTV-luciferase reporter plasmid. In addition, PR-negative cell lines derived from female gynaecological cancers (HeLa: human cervical cancer; Skov-3 and MDAH-2774: human ovarian cancer; MCF7: human breast cancer) were transiently transfected with equal amounts of PR expression plasmids along with the MMTV-luciferase reporter plasmid.
Progesterone-induced PRA leads to poor transactivation of the luciferase reporter gene (Fig. 6A). In HeLa and MCF7 cells (Fig. 6A), progesterone-induced PRA-V660 and PRA-L660 equally reduce (approximately 20%) luciferase activity compared with unligated receptor. On the contrary, in Skov-3 and in cells stably transfected with PRA (CHO-A25_2 and CHO-C4_3), progesterone induction of PRA-V660 (CP PR) leads to a 1.2-fold increase of luciferase activity compared with unligated PRA-V660, whereas progesterone induction of PRA-L660 (PROGINS protein variant) does not change luciferase activity compared with the unligated receptor. Given that the expression of the receptor protein was confirmed by western blot analysis (not shown) we concluded that the reduced activity of PRA-L660 reflects a real functional consequence of the amino acid substitution.
|
) compared with the unligated receptor. PRB activated with the more potent agonist, org2058, leads to 5- to 20-fold induction of the luciferase activity (Fig. 6C). Large differences in the relative induction of luciferase activities are seen among cell lines. Transactivation activities of PRB-V660 and PRB-L660 are similar in HeLa, MCF7 and MDAH:2774 cell lines. However, similarly to what we observed for the PRA isoform, in Skov-3 and in cells stably expressing PRB (CHO-B34 and CHO-D9_14) ligand-bound PRB-L660 (progesterone, Fig. 6B and org2058, Fig. 6C) is significantly less active than ligand-bound PRB-V660 (CP receptor) on the MMTV promoter.
Titration of the antagonist RU486 in the presence of a fixed concentration of progesterone reveals no difference in transactivation activity of PRB-V660 and PRB-L660 (Fig. 6D). Similar results were obtained for all cell lines regardless of whether another antagonist (org31710) and agonist (org2058) were examined and for PRA as well (data not shown).
Transrepression is widely described among steroid hormone receptors and consists in the ability of one nuclear receptor to inhibit the activity of a second nuclear receptor. Given the crucial role of PRA, PRB and ER- and their corresponding crosstalks, we investigated: (a) how the transactivation of either PRB isoform (V660 and L660), on the MMTV-luciferase is transrepressed by PRA (V660 or L660) and (b) how the transactivation of ER- on an ERE-TK-luciferase reporter construct is influenced by PR (A or B isoform, V660/L660 variant). Transactivation of PRB on the MMTV-luciferase reporter gene is reduced by PRA in a dose-dependent manner (Fig. 6E). Transactivation of 17ß-oestradiol-induced ER- is reduced about 40–50% by PRA alone, PRB alone or by the presence of both receptors (data not shown). In the transrepression experiments (on PRB and on ER-), no differences between the CP and the PROGINS protein variant were observed.
PRA and PRB PROGINS (L660) protein variants show different post-translational modifications compared with the CP PR-V660
The V660L substitution does not affect DBD, LBD or activating functions (AF) of the PR. However, fine-tuning of PR activity is modulated by phosphorylation and protein degradation (Lange et al. 2000, Qiu & Lange 2003, Qiu et al. 2003). We hypothesised that post-translational modifications of PR may explain why the activities of the PROGINS protein variant compared with the CP PR differ in different cellular contexts. Cells stably transfected with PRA-V660 (CHO-A25_2), PRA-L660 (CHO-C4_3), PRB-V660 (CHO-B34) and PRB-L660 (CHO-D_14) were stimulated with different ligands and for different periods of time. Subsequently, cell extracts were analysed by western blotting to discriminate between non-phosphorylated and phosphorylated PR. Upon binding of an agonist (progesterone or org2058) or an antagonist (RU486, not shown), PRA and PRB are phosphorylated in a dose-dependent manner (Fig. 7A and B), and this occurs regardless of V660 or L660. Phosphorylation occurs within 20 min after ligand induction and phosphorylated forms are still visible 24 h later. However, at this time point, PRB-V660 (CP) is more extensively phosphorylated after either agonist or agonist plus antagonist induction compared with PRB-L660 (PROGINS protein variant; Fig. 7C). The same trend, although less pronounced, is observed for PRA upon agonist induction but not after agonist plus antagonist induction (Fig. 7D).
|
). PRA-PROGINS protein variant (L660) appears to be degraded with reduced efficiency compared with the CP receptor (PRA-V660) after induction with an agonist plus an antagonist (Fig. 7E). This effect was not observed with PRB (not shown). Taken together, these observations suggest that the different responses observed for the PR-L660 variant (both A and B isoforms) compared with the CP receptor, arise secondary to different post-translational modifications. The three-dimensional structures of PROGINS variant and CP PR are different
Different affinities of PR-V660 and PR-L660 proteins for kinases/phosphatases and components of the proteasomal degradation pathway may be related to different three-dimensional structures, which in turn may alter the accessibility of epitopes required for protein–protein interactions. To check this hypothesis, we performed partial trypsin digestion assays (Fig. 8). The differences in the extent of trypsin digestions suggest that binding of progesterone (Fig. 8B and E) or RU486 (Fig. 8C and F) induces different three-dimensional conformations in PRA and PRB. Furthermore, the same stimulation conditions resulted in different degradation patterns of PR-V660 or PR-L660 variants for both A and B isoforms, suggesting that ligand-induced changes in the three-dimensional structures of PR-V660 and PR-L660 occur. In fact, in the absence of any induction, trypsin degradation of the PRA-L660 (PROGINS protein variant) is increased compared with PRA-V660 (Fig. 8A), whereas it is slightly reduced after incubation with progesterone (Fig. 8B) and unaffected after incubation with RU486 (Fig. 8C). PRB-L660 (PROGINS protein variant) appears to be less stable compared with the CP PR in the absence of any ligand or upon progesterone stimulation (Fig. 8D and E). By contrast, stimulation with RU486 only slightly reduces the stability of PRB-L660 compared with PRB-V660 (Fig. 8F).
|
The inhibitory effects of the progestin org2058 on the proliferation rate of cell lines stably expressing PRA-V660 (CHO-A25_2), PRA-L660 (CHO-C4_3), PRB-V660 (CHO-B34) and PRB-L660 (CHO-D_14) were measured. Maternal cell line CHO-K1 and cell line CHO-A21 (PR negative) were used as negative controls. In the absence of ligand, the cell line expressing PRA-V660 (CP) shows lower cell numbers compared with cell lines not expressing PR or cells expressing PRA-L660 (PROGINS protein variant; Fig. 9A). Cells expressing PRA (either CP or PROGINS protein variant) also had reduced cell numbers in the presence of a PR agonist (org2058; Fig. 9B–C). However, in the presence of low concentration of the agonist org2058 (10–10 M; Fig. 9B), cells expressing PRA-V660 show a 12% reduction in total cell number compared with the response obtained in the absence of ligands. On the contrary, the same agonist concentration has no effect on the total cell number in cells expressing PRA-L660 (PROGINS protein variant); a 100-fold higher agonist concentration (org2058 10–8 M; Fig. 9C) was required to obtain a similar (13%) inhibition. To verify that the effects on total cell number were the result of diminished proliferative activity, the S-phase fractions were determined using BrdU incorporation and flow cytometry. BrdU incorporation in cells expressing PRA-V660 (CP) was reduced by treatment with org2058 10–10 M from 69.6% (vehicle) to 55.7%, whereas in cells expressing PRA-L660 BrdU incorporation is reduced from 64.5% (vehicle) to 61.1% (org2058) only. No change in BrdU incorporation is seen in the PR-negative control cell line after org2058 treatment. After 24 and 36 h (not shown) of culture, BrdU incorporation reaches saturation.
|
). |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
Consequences of the PV/HS-1 Alu element in intron G of the PROGINS genetic variant
Alu elements are human-specific transposable units present in approximately 1.4 million copies in the human genome (Deininger & Batzer 2002, Deininger et al. 2003). They are retrotransposons that have amplified over the past 56 million years via a RNA intermediate transcribed by RNA Pol-III (Deininger et al. 2003) and it is estimated that in humans, Alu elements contribute to 0.1% of genetic diseases (Deininger & Batzer 1999). Several functional consequences can be ascribed to the presence of Alu elements (Deininger & Batzer 2002). Among these, Alus can influence gene transcription by modifying proximal promoter or distal enhancer. Owing to the high CpG content, Alus can lead to DNA methylation ((Deininger & Batzer 1999) and references therein) and, in several cases, intronic Alus can be exonised (Lev-Maor et al. 2003, Sorek et al. 2004) leading to the generation of splice variants.
We have investigated whether the Alu insertion in intron G (PROGINS polymorphic variant) affects the transcription rate and RNA maturation due to (a) CpG methylation, (b) modification of transcription factor-binding sites, (c) affecting RNA stability and d) the generation of splice variants.
Although the PROGINS allele in T47D cells is transcribed with reduced efficiency compared with the CP PR (Figs 4A, lane 5 and 4C), this is not caused by methylation of the PROGINS-Alu because treatment with a DNA demethylating agent (ADC) does not increase its transcription. Furthermore, reduced transcription of the PROGINS allele appears not to be a general phenomenon, because cDNA-PCR–RFLP based screening on mRNA isolated from human endometrium of fertile healthy women showed that the PROGINS and the CP alleles are expressed at comparable levels (A Romano unpublished). Even if transcription of the PROGINS allele is not modified by DNA methylation, it is increased by 17ß-oestradiol upon ER- activation and subsequent binding, together with the Sp1 transcription factor, to the Alu-ERE/Sp1 site. Similar cases have been reported in literature. Norris et al.(1995) have shown that the BRCA-1 gene responds to oestrogen because of an Alu element containing an imperfect ERE, which is responsible for ER binding. The half-ERE/Sp1 site in the promoter of the PR gene induces its transcription in response to oestrogens (Petz & Nardulli 2000). Furthermore, Sp1 and ER- can bind to DNA sequences located in proximal promoter and in distal enhancers (Norris et al. 1995, Bouwman & Philipsen 2002, Carroll et al. 2005, Goffin et al. 2005) and the binding of ER- to ‘non-canonical’ half-ERE is more common than until recently expected (Carroll et al. 2005).
Once transcribed by the RNA Pol-III, individual Alu-RNAs belonging to different families generate molecules with different stabilities (Aleman et al. 2000, Li & Schmid 2004). However, the effect that Alus have on the stability of (pre)-mRNA of those genes in which they are integrated (transcribed by the RNA Pol-II), has not been studied. We observed that the transcript derived from the PROGINS allele is less stable compared with the transcript derived from the CP allele, both in native T47D cells and in cells transfected with minigenes (Fig. 5). RNA decay pathways are controlled by RNA–protein complexes that assemble at specific sites on the target transcripts (Lemaire et al. 2002, Meisner et al. 2004, Huang & Steitz 2005, Moore 2005). For example, Lemaire et al.(2002) have shown that SR proteins (a family of splicing factors) can reduce the stability of the mRNA molecules by binding to purine-rich regions in the 3'UTR. A purine-rich region is present in Alu elements, and, although the PROGINS Alu is not located in the 3'UTR, it is still very proximal to it. Alternatively, Alus transcribed by RNA Pol-III (in their sense orientation) could pair with the PROGINS Alu, which is transcribed by RNA Pol-II in the pre-mRNA-Alu as its anti-sense sequence (Fig. 2A), and trigger dsRNA-mediated degradation. In our experiments, the reduced stability of the PROGINS transcript compared with the CP pre-mRNA is more pronounced in native T47D cells than in cells transfected with the minigenes (Fig. 5). This can be explained by the presence of specific enzymatic mechanisms leading to pre-mRNA-Alu degradation. Probably, transient transfection triggers higher RNA transcription in a few cells, thereby saturating the RNA degradation machinery in these cells.
Following pre-mRNA transcription, intronic Alus can generate alternative splice variants (Alu ‘exonisation’ (Deininger & Batzer 1999, Deininger et al. 2003, Lev-Maor et al. 2003, Sorek et al. 2004)). The 3' and 5' sites (shown in Fig. 2B) for alternate and constitutive ‘exonisation’ of Alus have been exhaustively studied by one group (Lev-Maor et al. 2003, Sorek et al. 2004). According to these studies, the PROGINS Alu is not expected to generate ‘exonised’ splice variants. Nevertheless, flanking sequences can influence alternate splicing (e.g. the branching site for the Alu ‘exonisation’ is located in the region of gDNA flanking the putative 3' splice site of the Alu). According to our results, the generation of any splice variants by the PROGINS Alu insertion can be excluded.
Consequences of the SNP-G3432T in exon 4 causing the amino acid substitution (V660L)
It is plausible that the translation efficiency of the PROGINS transcript is reduced by sub-optimal codon usage for leucine-660 in the PROGINS variant. The PROGINS leucine-660 is coded by TTG with a preference of approximately 11%, which is much lower than the 42% preference of the optimal codon for leucine, CTG (Zhao et al. 2003).
Consequences of the protein V660L substitution
The V660L substitution is located in the hinge region of the receptor, i.e. the region between the DBD and the LBD. So far, no specific function has been ascribed to the hinge region and it may be involved in nuclear translocation, receptor dimerisation in solution, protein stability and cofactor interaction (Jackson et al. 1997, Tetel et al. 1997, Li & O’Malley 2003). Our preliminary investigations indicate that the V660L substitution does not affect translocation of PRA and PRB to the nucleus (results not shown).
Agoulnik et al.(2004) have described an increased transactivation activity of PR-L660 (PROGINS protein variant) on a GRE2-E1b-CAT reporter in COS-1 cells. We have extended these studies and present evidence that the net activity of PR-V660 (CP) and PR-L660 (PROGINS) variant varies in a cell- and ligand-dependent manner. The transactivation activity on the MMTV-luciferase reporter induced by PR-L660 (PRO-GINS protein variant) compared with the CP PR-V660 (both A and B isoforms) is reduced only in Skov-3 cells and in cells stably transfected with PRA (CHO-A25_2 and CHO-C4_3) and PRB (CHO-B34 and CHO-D9_14). Interestingly, these cells are of ovarian origin (human and hamster), and epidemiological studies indicate that the PROGINS allele might increase the risk for ovarian cancer (see below). Transrepression of PRA on PRB and of PR on ER- is not influenced by the V660L substitution.
Different responses of the CP and PROGINS PR are a consequence of different post-translational modifications of the proteins
PR is subjected to several post-translational modifications. It contains 18 serines as potential phosphorylation sites and at least seven residues are phosphorylated in vivo (Knotts et al. 2001). Bagchi et al.(1992) have classified two types of phosphorylation sites. Fast phosphorylation sites, that are basally phosphorylated (without PR induction) and their phosphorylation increases upon ligand binding within minutes (ser-81, ser-162 present only in PRB, ser-190 and ser-400), and slow phosphorylation sites, that become phosphorylated in a ligand-dependent manner within a few hours after induction (ser-102 present only in PRB, ser-294 and ser-345). Although the activity of PR is controlled by the activating functions (AFs) in the LBD and in the N-terminal domain (unaffected by the V660L substitution), its fine regulation is influenced by phosphorylation (Bagchi et al. 1992, Takimoto et al. 1996, Weigel 1996, Knotts et al. 2001, Qiu & Lange 2003, Qiu et al. 2003, Pierson-Mullany & Lange 2004), protein degradation (Lange et al. 2000, Shen et al. 2001) and other post-translational modifications.
We have shown that the activity of the PROGINS protein variant compared with the cognate CP PR depends on the cellular context. Considering that the expression of kinases and proteasomal components is closely related to the cellular context, we hypothesised that this could determine the different actions of the two PR variants. Here, we show that PR CP and PROGINS protein variant become phosphorylated with similar efficiency within a few minutes after induction with a ligand (fast phosphorylation sites). By contrast, phosphorylation of the late phospho residues, examined 24 h after induction, occurs with reduced efficiency in the PROGINS protein variants (both PRA and PRB isoforms; Fig. 7) compared with the CP PR. PRA-L660 (PROGINS protein variant) also shows mino
