GnRH Increases c-Fos Half-Life Contributing to Higher FSHβ Induction

Abstract

GnRH is a potent hypothalamic regulator of gonadotropin hormones, LH and FSH, which are both expressed within the pituitary gonadotrope and are necessary for the stimulation of gametogenesis and steroidogenesis in the gonads. Differential regulation of LH and FSH, which is essential for reproductive fitness, is achieved, in part, through the varying of GnRH pulse frequency. However, the mechanism controlling the increase in FSH during the periods of low GnRH has not been elucidated. Here, we uncover another level of regulation by GnRH that contributes to differential expression of the gonadotropins and may play an important role for the generation of the secondary rise of FSH that stimulates folliculogenesis. GnRH stimulates LHβ and FSHβ subunit transcription via induction of the immediate early genes, Egr1 and c-Fos, respectively. Here, we determined that GnRH induces rapidly both Egr1 and c-Fos, but specifically decreases the rate of c-Fos degradation. In particular, GnRH modulates the rate of c-Fos protein turnover by inducing c-Fos phosphorylation through the ERK1/2 pathway. This extends the half-life of c-Fos, which is normally rapidly degraded. Confirming the role of phosphorylation in promoting increased protein activity, we show that a c-Fos mutant that cannot be phosphorylated by GnRH induces lower expression of the FHSβ promoter than wild-type c-Fos. Our studies expand upon the role of GnRH in the regulation of gonadotropin gene expression by highlighting the role of c-Fos posttranslational modification that may cause higher levels of FSH during the time of low GnRH pulse frequency to stimulate follicular growth.

FSH and LH are key hormones that regulate mammalian reproduction through stimulation of steroidogenesis, ovulation, and follicular growth. Their critical role is illustrated by the fact that either LH- or FSH-deficient humans and mice are infertile (1–5). LH and FSH are heterodimers consisting of a common α-subunit and a unique β-subunit. Regulation of the β-subunit transcription is the limiting component in the synthesis of mature hormones and, therefore, a key step in the control of their concentration in circulation (6, 7).

FSH and LH are produced by the same gonadotrope cell in the same hormonal milieu. However, LH and FSH are differentially regulated throughout the menstrual and estrous cycle, and this differential regulation is critical for reproductive fitness. Both LH and FSH are synthesized and secreted simultaneously during the ovulatory surge, under the influence of high frequency and high amplitude of GnRH (8–10). Then, FSH in particular, exhibits a secondary rise in serum concentrations, which is preceded by an increase of the β-subunit transcription (10–13). In rodents, due to a relatively short, 4-d estrous cycle, FSHβ transcription and FSH synthesis exhibit a secondary increase at estrus. FSHβ mRNA concentration increases 3-fold in late estrus through the afternoon of metestrus, in addition to a 4-fold increase in the afternoon of proestrus (10, 13). This secondary rise of FSH is required for follicle selection and maturation (10, 14, 15). In humans, the FSH bioactivity levels increase during the middle of the luteal phase until the midfollicular phase of the menstrual cycle. This increase in circulating FSH coincides with a critical time during folliculogenesis when the next cohort of follicles is being recruited (16).

GnRH from the hypothalamus regulates transcription and secretion of the gonadotropin hormones. Stimulation with GnRH induces LHβ and FSHβ transcription, through the activation of MAPK-signaling pathways (17–19). However, the induction of LHβ and FSHβ is not direct but occurs via the induction of intermediates, immediate-early genes (reviewed in Refs. 20 and 21). c-Fos is an immediate-early gene that heterodimerizes with c-Jun to form the activator protein 1 (AP1) transcription factor and it is induced by GnRH both in vivo (22) and in immortalized gonadotrope cells (23) to stimulate the transcription of the FSHβ gene (24–26). Another immediate-early gene, Egr1, is stimulated by GnRH to induce LHβ transcription (27–30). c-Fos and Egr1 are prototypical immediate-early genes in that they are activated rapidly and transiently with very short half-lives. Both genes have mRNA that is unstable with a half-life of as little as 9 min (31) and equally unstable proteins, exhibiting rapid turnover and degradation (32). The short half-life allows for strong temporal regulation of gene targets, which in gonadotropes are the gonadotropin β-subunits.

Differential regulation of LHβ and FSHβ is achieved in part through varying GnRH pulse frequencies. A higher GnRH pulse frequency, as occurs before the ovulatory surge, favors LHβ, whereas a lower GnRH frequency only regulates FSHβ (7, 33–35). Evidence suggests that LHβ is more sensitive to GnRH frequency, whereas FSHβ synthesis is less frequency dependent (36, 37). This differential synthesis of β-subunits may stem from divergent regulation of the intermediates. Given that GnRH induces LHβ and FSHβ through the immediate-early genes, Egr1 and c-Fos, respectively, turnover and half-lives of these transcription factors would contribute to rates of LHβ and FSHβ transcription. In particular, it is still unclear how the FSHβ secondary rise is achieved during a time of low GnRH frequency. Thus, elucidating the regulation of key players in FSHβ synthesis is necessary to understand the etiology of infertility.

Here, we uncover an additional level of regulation by GnRH, whereby GnRH-induced signaling pathways posttranslationally modify c-Fos to increase its half-life. This increase in half-life is specific for c-Fos, not occurring to other GnRH-induced immediate-early genes, and is due to c-Fos phosphorylation through the ERK1/2 signaling pathway. This allows for higher transcriptional activity and higher induction of FSHβ. It is intriguing to consider that this mechanism may function during the time of the cycle characterized by low GnRH pulse frequency to extend c-Fos half-life and increase specifically FSHβ transcription.

Materials and Methods

Whole-cell extract and Western blotting

Cells and treatments

An immortalized LβT2 cell line, kindly provided by Pamela Mellon (University of California San Diego, La Jolla, CA) was cultured in 10-cm plates in DMEM (Mediatech, Inc., Herndon, VA) with 10% fetal bovine serum (Gemini Bio-Products, West Sacramento, CA) and penicillin/streptomycin antibiotics (GIBCO/Invitrogen, Carlsbad, CA) at 37 C. Cells were passaged using 1× Trypsin-EDTA (Sigma-Aldrich, St. Louis, MO). The LβT2 cells were starved overnight in serum-free DMEM with 0.1% BSA before treatment. The cells were treated with 10 nm GnRH throughout the experiment for times indicated (tonic treatment) or for 5 min (pulse) after which the GnRH-containing medium was removed and replaced by serum-free DMEM with 0.1% BSA for times indicated. To study the effect of GnRH on c-Fos degradation and half-life without its influence on c-Fos induction, endogenous c-Fos was induced by GnRH for 2 h or exogenous c-Fos was overexpressed by transfection of expression vector containing the mouse c-Fos cDNA under the cytomegalovirus promoter for 24 h. GnRH containing media (for endogenous c-Fos induction) or serum containing media (for overexpression) was then removed and replaced with serum-free media containing actinomycin D (ActD, 5 μg/ml) to inhibit transcription and Cycloheximide (200 μg/ml) to inhibit new protein synthesis. GnRH or vehicle was added to assess GnRH effect of degradation.

Western blot

Cells were rinsed with 1× PBS and lysed with a buffer containing: 20 mm Tris (pH 7.4), 140 mm NaCl, protease inhibitors (Sigma), 1 mm PMSF, 10 mm NaF, 1% Nonidet P-40, 0.5 mm EDTA, and 1 mm EGTA. Bradford reagent was used to determine protein concentrations, and the concentrations were calculated using a standard curve. Equal amounts of protein from whole-cell extracts were loaded with 4× sample buffer, resolved by gel electrophoresis, and transferred to a polyvinylidene fluoride membrane. The membranes were blocked with 10% milk in wash buffer (20 mm Tris, pH 7.4; 1% Tween, 150 mm NaCl, and 0.5% BSA) and then probed with antibodies to: c-Fos (sc-52), Egr1 (sc-110), c-Jun (sc-1694) all from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA); TATA-binding protein (antibody 63766), or β-tubulin (antibody 6046) both from Abcam (Cambridge, UK). Proteins were detected with a secondary antibody to rabbit or mouse IgG linked to horseradish peroxidase (Amersham, GE Healthcare, Piscataway, NJ) and Enhanced Chemiluminescence Western Blotting Detection Reagent (GE Healthcare). To assure equal loading, membranes were stripped at 60 C for 1 h with strip buffer (50 mm Tris, pH 6.8; 5% sodium dodecyl sulfate, and 100 mm β-mercaptoethanol and reexposed to enhanced chemiluminescence and autoradiography to ensure complete removal of the antibody and then blocked again with milk and reprobed for TATA-binding protein or β-tubulin.

Analysis

Bands were quantified using the GeneGnome Bio Imaging Chemiluminescence (Syngene, Frederick, MD) reader and analyzed by ImageJ software available from NIH. Each experiment was repeated at least three times, the quantifications obtained for c-Fos, Egr1, or c-Jun were normalized to β-tubulin quantifications, and the average of all experiments was presented in the graphs. For the analysis of the rates of degradation (K), Ln[A]_t = −Kt + Ln[A]₀ formula was used with normalized values of Western blot quantifications as indicators of protein concentration. Slope of the fit curve was calculated using GraphPad Prism software (GraphPad Software, Inc., San Diego, CA). Statistical significance, P < 0.05, was determined by using ANOVA and Tukey's post hoc test and indicated with an asterisk in the figure.

mRNA extraction and quantitative PCR

RNA was obtained with Trizol reagent (Invitrogen/GIBCO) according to the manufacturer's instructions after GnRH treatment for the times indicated as described above. Contaminating DNA was removed with DNA-free reagent (Ambion, Austin, TX) and 2 μg RNA were reverse transcribed using Superscript III First-strand Synthesis System (Invitrogen). Quantitative Real-Time PCR was performed in an iCycler with SybrGreen Supermix from Bio-Rad Laboratories, Inc. (Hercules, CA) and the following primers: c-Fos forward, GGCAAAGTAGAGCAGCTATCTCCT; reverse, TCAGCTCCCTCCTCCGATTC; glyceraldehyde 3-phosphate dehydrogenase (GAPDH) forward, TGCACCACCAACTGCTTAG; reverse, GGATGCAGGGATGATGTTC under the following conditions: 95 C for 15 min, followed by 40 cycles at 95 C for 15 sec, 56 C for 30 sec, and 72 C for 30 sec. Each sample was assayed in triplicate and the experiment was repeated three times. A standard curve with dilutions of 10 pg/well, 1 pg/well, 100 fg/well, and 10 fg/well of a plasmid containing c-Fos cDNA was generated in each run with the samples. In each experiment, the amount of c-Jun or GAPDH was calculated by comparing a threshold cycle obtained for each sample with the standard curve generated in the same run. Replicates were averaged and divided by the mean value of GAPDH in the same sample. After each run, a melting curve analysis was performed to confirm that a single amplicon was generated in each reaction.

Transient transfections

FSHβ-luciferase reporter and c-Fos expression vector were described previously (38). Transfection was performed 1 d after the plating into 12-well plates in DMEM with 10% FBS with FuGENE 6 (Roche Molecular Biochemicals, Indianapolis, IN) transfection reagent following the manufacturer's instructions. Each well was transfected with 0.5 μg of a luciferase-reporter plasmid, 0.2 μg of expression vector, and 0.1 μg of Herpes virus thymidine kinase promoter-driven β-galactosidase, as a control for transfection efficiency. The cells were starved overnight in serum-free DMEM with 0.1% BSA and penicillin/streptomycin antibiotics. Cells were washed 48 h after transfection with 1X PBS and lysed with 100 nm K-PO₄ buffer containing 0.2% Triton X-100. Black plates (96-well) were loaded with 20 μl of each lysate, and luciferase activity was measured on a luminometer (Veritas Microplate luminometer from Turner Biosystems/Promega Corp., Madison, WI) by injecting 100 μl of buffer containing 25 mm Tris (pH 7.8), 15 mm MgSO₄, 10 mm ATP, and 65 μm luciferin per well. The Tropix Galacto-light β-galactosidase assay (Applied Biosystems, Foster City, CA) was used to measure β-galactosidase activity on the luminometer following the manufacturer's instructions. Transfections were performed in triplicate and repeated three times. Transfection efficiency was controlled by dividing the luciferase values by β-galactosidase, and individual experiments were normalized by dividing the luciferase/β-galactosidase ratio by control vector pGL3-luciferase/β-galactosidase ratio. Normalized luciferase/β-galactosidase values from three experiments were averaged and ANOVA followed by the Tukey's post hoc test was performed using the JMP program with significance set at P ≤ 0.05.

Mutagenesis

Mutagenesis of the c-Fos expression vector was performed using the QuikChange Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA) following the manufacturer's instructions, with PCR (95 C for 30 sec, 95 C for 30 sec, 55 C for 1 min, 68 C for 14 min) for 18 cycles followed by 37 C for 1 h of Dpn treatment, and products were transformed into XL1 Supercompetent Cells (Stratagene, La Jolla, CA). Mutations were confirmed by dideoxyribonucleotide sequencing.

Results

GnRH maintains c-Fos protein, but not Egr1 or c-Jun

GnRH induces FSHβ and LHβ gonadotropin genes through intermediates, immediate-early genes c-Fos and Egr1, respectively. Given that c-Fos and Egr1 are rapidly induced upon treatment and quickly degraded in other cell types, we analyzed the time course of protein synthesis and degradation in LβT2 gonadotrope cells. We started our analysis by comparing tonic and pulse GnRH treatments. The same amount of protein from LβT2 whole-cell lysates treated with tonic GnRH or a single 5-min GnRH pulse was analyzed by Western blotting for c-Fos, Egr1, and c-Jun expression. Interestingly, peak amplitude of response was observed 2 h after beginning of either tonic or pulse treatment (Fig. 1, A–C). In contrast to the same timing of the response, tonic treatment caused higher induction of all three proteins at the peak amplitude than a single pulse treatment. In quiescent cells c-Fos and Egr1 are not present without stimulus, and we were unable to analyze fold induction from the control. Further, it is not possible to compare different proteins because intensity depends on the antibody affinity. We are, however, able to quantify relative differences between tonic and pulse treatment in the same protein probed with the same antibody, by normalizing each immediate-early gene western to β-tubulin control. We determined that induction of c-Fos by tonic treatment is 3.4-fold higher than induction by a single pulse, whereas induction of Egr1 by tonic treatment is 2.6-fold higher from the induction by a single pulse. c-Jun, which is a c-Fos binding partner and forms an AP1 transcription factor that induces FSHβ, has detectable basal expression and is induced 7.3-fold by tonic treatment and 3.6-fold by a GnRH pulse.

Fig. 1.

c-Fos (A), Egr1 (B), and c-Jun (C) proteins are induced rapidly by GnRH but have different degradation rates. LβT2 cells were treated with 10 nm GnRH, either tonic or a single 5-min pulse. Whole-cell protein extracts were obtained after indicated time, and the same amount of protein was loaded in each lane and run on SDS-PAGE. Blots were stripped and reprobed with antibody to β-tubulin (β-tub) to assure equal loading. D–F, Cells were treated with either tonic GnRH (solid line) or a single 5-min GnRH pulse (dashed line). Western blots were quantified and c-Fos (D), Egr1 (E), or c-Jun (F) protein amount was normalized to β-tubulin in each experiment. The quantified protein at each time point was presented as percent from the maximal amount at 2 h of treatment. The experiment was repeated three times. *, Significant difference between tonic and pulse GnRH treatment calculated by ANOVA and Tukey's post hoc test using JMP program.

Because we were interested in the half-life and degradation of these short lived proteins, we loaded the same amounts of protein on the gel and quantified and normalized the amount of c-Fos, Egr1, and c-Jun to the amount of β-tubulin in each sample. To observe the rate of turnover, the maximal level at 2 h was set to 100% for each experiment (Fig. 1, D–F). We determined that c-Fos had a lower rate of degradation with tonic treatment than with pulse treatment (Fig. 1D). Such a finding may not have been surprising initially, because with tonic treatment, the GnRH stimulus was present throughout the experiment. However, this effect was not observed in the turnover of Egr1 and c-Jun proteins (Fig. 1, E and F). At 4 h, with tonic treatment, 90% of the c-Fos protein remains, but only 60% with pulse treatment, whereas at 8 h 70% was preserved with tonic treatment, and 43% with a pulse (Fig. 1D). Egr1, induced by GnRH to stimulate transcription of LHβ, is a labile protein that was rapidly degraded after either tonic or pulse treatment (Fig. 1E). At 4 h, 48% of the Egr1 protein remained after tonic treatment and 44% after a pulse, which was not a statistically significant difference between treatments. After 8 h, 10% of the protein remained regardless of the treatment. The degradation of c-Jun is slower than the degradation of Egr1, because c-Jun is a more stable protein. Despite the longer half-life of c-Jun compared with Egr1, there is no difference in the turnover rates for c-Jun either, between the two treatments (Fig. 1F). At 4 h 78% and 75% of c-Jun remained after tonic and pulse treatment, respectively, and 59% and 56% at 8 h, neither of which yielded a significant difference. Thus, three important immediate-early genes have different time courses of protein turnover in the gonadotrope cells. However, only the c-Fos protein degradation changes depending on the manner of GnRH stimulus.

Given that tonic and pulse GnRH treatment elicited different levels of c-Fos protein, we analyzed whether this difference stems from the mRNA induction. Cells were treated as described above, and mRNA was extracted, reverse transcribed, and quantified using real-time PCR. In each sample, the amount of c-Fos mRNA was normalized to the amount of the housekeeping gene GAPDH. Although the fold induction is variable between experiments, in each experiment the tonic treatment resulted in a higher level of induction than pulse treatment, 82.6-fold vs. 37.6-fold, respectively (Fig. 2A). This induction is rapidly diminished, such that after 2 h of tonic treatment c-Fos induction measured 16-fold compared with 2-fold in response to pulse treatment. As with the protein amount, there is no difference in the timing of the response between treatments. The peak amplitude was achieved at 45 min, whereas the levels returned to basal by 4 h of either treatment.

Fig. 2.

Tonic GnRH causes higher c-Fos mRNA induction. A, LβT2 cells were treated with tonic (solid line) or a single 5-min pulse (dashed line) GnRH treatment (10 nm), and mRNA was obtained at the indicated time. c-Fos mRNA levels were determined by quantitative PCR, calculated based on the standard curve from a serial dilution of the c-Fos cDNA-containing plasmid, and normalized to the GAPDH mRNA level in each sample. *, Significant difference between tonic and pulse GnRH treatment calculated by ANOVA and Tukey's post hoc test. B, c-Fos/GAPDH ratio at each time point was presented as percent from the peak amplitude observed at 0.75 h (45 min) of treatment. Experiment was repeated three times and half-life (t_1/2) was calculated from the curve, using time at peak amplitude as the starting point for degradation.

Due to the variable fold induction, we set the maximal level of c-Fos/GAPDH ratio at 100% to observe the resolution of the signal. We determined that the half-life (t_1/2) of c-Fos mRNA was 12.4 min with tonic treatment and 8.5 min with pulse treatment, a difference that did not achieve significance. With both treatments mRNA levels decreased rapidly, to 20% of maximum with tonic treatment and 9.7% with a pulse at 2 h, and 16% and 5.3%, respectively, at 3 h (Fig. 2B). Although with both treatments mRNA levels diminished rapidly, the initial higher peak amplitude in transcriptional response after tonic treatment, may underlie the higher level of c-Fos protein at 4 h.

GnRH increases the half-life of the c-Fos protein

GnRH causes an increase in transcription and translation of its target genes in LβT2 cells, which compounds the analysis of its effect on degradation. To observe the effect of hormone treatment on degradation separately, the cells were first stimulated with GnRH for 2 h, which is the time point at which the maximal c-Fos protein level was observed. Then, the GnRH-containing media were removed and replaced with media supplemented with ActD and cyclohexamide (Chx) to prevent new transcription and translation. After 15 min vehicle or GnRH was added (time zero for degradation studies, Fig. 3). We conducted preliminary experiments to determine optimal doses of ActD and Chx that completely abrogate transcription and translation (data not shown). Expression of endogenous c-Fos in LβT2 cells was highest at time zero and progressively decreased over the 8-h experiment. Western blot analysis illustrates that the addition of GnRH decreased c-Fos degradation (Fig. 3A). The rate of degradation was determined by normalizing the amount of c-Fos to β-tubulin. To eliminate the effect of changes in absolute levels of protein that varies from experiment to experiment, the amount at time zero was set to 100%. After GnRH removal and the addition of ActD and Chx to stop transcription and translation, c-Fos protein in vehicle-treated samples degraded so that 65% remained after 2 h and 44% after 8 h. Addition of GnRH lowered the rate of degradation compared with vehicle-treated cells, so that 90% remained at 2 h and 65% after 8 h (Fig. 3B). The difference in degradation was significant at 2 h and remained significantly slower in the presence of GnRH for the duration of the 8-h experiment. Without the compounding effect of new transcription and translation, we calculated that the rate of degradation (K) significantly decreased (P < 0.0001) from 0.1059 to 0.05729 with the addition of GnRH (Fig. 3C).

Fig. 3.

GnRH decreases c-Fos degradation. A, c-Fos was induced by GnRH for 2 h, after which the media was changed and ActD and Chx were added to prevent new transcription and translation. Vehicle (two top panels) or GnRH (two bottom panels) was added at time zero to assess the role of GnRH in c-Fos turnover. After LβT2 whole-cell protein extracts were obtained, equal amounts of protein were run on gel and Western blots for c-Fos and β-tubulin (β-tub) were performed. B, After quantification, the c-Fos amount was normalized to β-tubulin and presented graphically. The amount of c-Fos was set to 100% at time zero for degradation, which corresponds to the time of addition of vehicle (dashed line) or GnRH (solid line). The experiment was repeated three times, and statistical analysis was performed and significance determined by ANOVA and Tukey's post hoc test. *, Significant difference in the amount of c-Fos between vehicle- and GnRH-treated samples at each time point. C, Rates of degradation were calculated by the formula Ln[A]_t = −Kt + Ln[A]_0. Slope of the fit curve was determined using the GraphPad Prism software. D, c-Fos was overexpressed by transfection of a plasmid containing c-Fos cDNA. After 24 h, media were changed to media containing ActD and Chx, and the experiment was performed as presented in panel A. E, The amount of c-Fos was calculated as in panel B. F, The rates of degradation of overexpressed c-Fos in the presence and absence of GnRH were calculated as in panel C.

To eliminate the need of GnRH to induce c-Fos, we expressed c-Fos protein by transfecting a vector that contains mouse c-Fos cDNA, because c-Fos is undetectable in cells in the absence of a stimulus. By using this approach we circumvented the effect of GnRH to induce mRNA as observed in Fig. 2, and concentrated solely on the GnRH effect on degradation. Cells were treated 24 h after transfection with vehicle or GnRH in the presence of ActD and Chx to prevent new transcription and translation. In the vehicle-treated cells, c-Fos degraded more rapidly, and difference in the c-Fos turnover between GnRH and vehicle treatment was more pronounced (Fig. 3D). At 2 h, the c-Fos protein in vehicle-treated samples was 44% of maximum vs. 71% in the presence of GnRH; at 4 h it was 19% vs. 59%, and at 8 h it was 5% vs. 51% in vehicle vs. GnRH-treated cell extracts, respectively (Fig. 3E). The rate of degradation (K) decreased significantly from 0.3747 to 0.0805 in the presence of GnRH (P < 0.0001; Fig. 3F), indicating that GnRH increases c-Fos protein half-life.

We examined whether the effect of GnRH on protein degradation is specific for c-Fos by determining degradation of another immediate-early gene, Egr1, that is also rapidly induced by GnRH. c-Fos or Egr1 were induced with GnRH in LβT2 cells for 2 h to achieve the maximal protein level, after which the media were changed to a media containing ActD and Chx to prevent new transcription and translation. Vehicle or GnRH was added at time zero for degradation studies, and the cells were lysed at the times indicated. c-Fos and Egr1 levels were assessed by Western blots, quantified, and normalized to β-tubulin (Fig. 4A). Similar to the results in Fig. 3, after 4 h the presence of GnRH lowered c-Fos degradation preserving 84% of the protein compared with 58% in vehicle-treated samples (Fig. 4B). In contrast, 17% of Egr1 protein remained regardless of treatment, suggesting that GnRH did not have an effect on stabilizing Egr1 and preventing its degradation (Fig. 4C). Therefore, GnRH-induced increase in c-Fos protein stability is specific for c-Fos and does not have a global effect on other GnRH-responsive genes in the gonadotrope.

Fig. 4.

GnRH-induced decrease in protein turnover is specific for c-Fos. c-Fos and Egr1 proteins were induced by GnRH for 2 h after which the media were changed to a media containing ActD and Chx to prevent new transcription and translation. Vehicle or 10 nm GnRH was added at time zero to assess the role of GnRH in protein turnover. After whole-cell extracts were obtained, the same amount of protein was run on gel and Western blots for c-Fos, Egr-1, and β-tubulin (β-tub) were performed. The experiments were performed three times, the amount of c-Fos (B) or Egr1 (C) was quantified on a chemiluminescence reader, normalized to β-tubulin in each blot, and the maximal value at time zero was set to 100%. *, Statistical difference between vehicle- and GnRH-treated samples at each time point.

GnRH phosphorylates c-Fos via ERK1/2 pathway

To determine the mechanism of c-Fos degradation and identify potential players in the stabilization by GnRH, we used the proteasome inhibitor MG-132, because short-lived proteins like c-Fos are usually degraded by the proteasome in other cells. We determined that overexpressed c-Fos protein degraded rapidly without new transcription and translation after the addition of ActD and Chx. However, c-Fos levels remained the same in the cells treated with MG-132 throughout the experiment, even after the treatment with ActD and Chx (Fig. 5A). c-Fos protein was not degraded in the presence of the proteasome inhibitor, whereas it rapidly diminished in the control cells. Therefore, in gonadotrope-derived cells c-Fos is degraded by proteasomes rather than other protein degradation systems, such as lysosomes, calpain, or caspases.

Fig. 5.

c-Fos protein is phosphorylated and degraded by the proteasome. A, After c-Fos protein was expressed, and new transcription and translation prevented with ActD and Chx, to observe c-Fos degradation, cells were treated with dimethylsulfoxide vehicle (ctrl) or 50 μm MG132 proteasome inhibitor (MG132). Whole-cell extracts were obtained after addition of vehicle or the inhibitor at times indicated, and c-Fos and β-tubulin Western blots were performed. B, An increase in the molecular weight of the c-Fos protein that changed its migration through the gel was observed after a longer run on the high-resolution gel, indicating posttranslational modification. C, c-Fos was induced with 10 nm GnRH and whole-cell extracts were obtained after times indicated. Treatment of protein extracts with alkaline phosphatase (ALP) results in the reduction of molecular weight, indicating that the molecular weight shift is due to phosphorylation. Each experiment was repeated three times and representative blots are shown.

Usually, proteins are posttranslationally modified for targeting to the proteasome. Before Western blotting, we ran the proteins on high-resolution gels to better observe putative posttranslational modification. We determined that in addition to induction, c-Fos protein exhibited a shift in molecular weight during GnRH treatment, indicating a potential posttranslational modification (Fig. 5B). We analyzed whether c-Fos was ubiquitinated, sumoylated, or acetylated, as these potential modifications are associated with protein degradation, by immunoprecipitating c-Fos and blotting with ubiquitin, sumo, or acetylation antibodies. However, we did not detect a change in these modifications on the c-Fos protein with GnRH treatment (data not shown). We then treated the extracts with alkaline phosphatase to assess whether the increase in molecular weight was due to c-Fos phosphorylation. Indeed, treatment of cell lysates with alkaline phosphatase abrogated the increase in molecular weight of c-Fos (Fig. 5C). Thus, in addition to the rapid induction of c-Fos transcription by GnRH, c-Fos is also posttranslationally modified by phosphorylation.

GnRH activates three branches of MAPK signaling pathways: ERK1/2, p38, and c-Jun N-terminal kinase (JNK) (17–19). All three of these pathways have been implicated in c-Fos phosphorylation in other cell types (39–41). We used inhibitors of these pathways to determine which one leads to c-Fos phosphorylation after GnRH treatment in LβT2 cells, by observing the shift in molecular weight of the c-Fos protein between vehicle-treated and GnRH-treated cells. We transfected the cells with a c-Fos expression vector to eliminate the need for GnRH to induce c-Fos protein. Cells were treated with inhibitors 15 min before GnRH treatment and lysed after 2 h at the time of maximal phosphorylation. In control cells, GnRH treatment completely shifted the c-Fos protein to the slower mobility, phosphorylated form. Inhibition of the ERK1/2 pathway diminished the shift of c-Fos protein, leaving a substantial proportion of the protein as the faster mobility, unmodified form (Fig. 6A). On the other hand, inhibition of p38 and JNK in GnRH-treated cells did not prevent the shift to the slower mobility, phosphorylated form. Surprisingly, p38 and JNK inhibition in control, vehicle-treated cells caused the shift to the phosphorylated form, which may be due to the inhibition of the dephosphorylation pathway. Nevertheless, because our interest is in the GnRH-induced changes in c-Fos phosphorylation, we determined that the ERK1/2 pathway plays a role GnRH phosphorylation of c-Fos.

Fig. 6.

GnRH phosphorylates c-Fos through the ERK1/2 pathway. A, LβT2 cells were transfected with a c-Fos expression vector, 24 h before addition of ActD and Chx to prevent new transcription and translation. Cells were treated with dimethylsulfoxide vehicle or inhibitors of the following MAPK pathways to identify the pathways involved in c-Fos phosphorylation: p38 (20 μm SB202190 to inhibit p38); ERK1/2 (5 μm UO126 to inhibit MEK1 and activation of ERK1/2), or JNK (10 μm SP600125 to inhibit JNK). GnRH or vehicle control was added for 2 h, after which the extracts were obtained and run on high-resolution gels to observe changes in molecular weight. B, Three separate experiments were performed and the amount of lower (unphosphorylated) and higher (phosphorylated) bands were quantified. The graph represents the ratio of lower mobility, phosphorylated c-Fos to the total c-Fos amount. Vehicle-treated samples are represented with white bars, whereas GnRH-treated samples are presented by black bars. *, Statistically significant decrease of phosphorylated c-Fos in GnRH and ERK1/2 inhibitor-treated cells compared with GnRH and dimethylsulfoxide-treated samples. #, Significant change caused by p38 and JNK inhibitors in vehicle samples (−GnRH) compared with dimethylsulfoxide vehicle samples (−GnRH). Ctrl, Control; Unmod., unmodified protein.

Three separate experiments were quantified and the slower migration, phosphorylated form was presented as a portion of the total c-Fos protein (Fig. 6B). In the vehicle-treated control lane, 20% of the c-Fos protein was phosphorylated, and GnRH treatment led to phosphorylation of approximately 100% of the c-Fos protein. Inhibition of the ERK1/2 pathway did not significantly change the proportion of the phosphorylated form in vehicle-treated lane, but it lowered the proportion of the phosphorylated form in the GnRH-treated lane by 47%. Inhibition of p38 and JNK pathways increased phosphorylation of c-Fos in vehicle-treated lanes. These pathways may activate the phosphatase that dephosphorylates c-Fos and serve in a negative feedback for ERK1/2. On the other hand, more relevant to our studies, p38 and JNK inhibition did not affect GnRH-induced phosphorylation. Thus, GnRH phosphorylates c-Fos in an ERK1/2 pathway-dependent manner.

C-terminal region of c-Fos is phosphorylated by GnRH and regulates c-Fos half-life

c-Fos protein has several phosphorylation sites in the C-terminal transactivation domain (Fig. 7A, TAD), located in three separate functional regions. We created three mutations in the c-Fos protein, mutating separately these three regions to assess which of them is phosphorylated by GnRH. Mutant 1 (m1) contains a mutation of the threonine at the position 232 (T232) to alanine, which is close to the leucine zipper domain important for dimerization. Mutant 2 (m2) has a double mutation in the closely positioned threonines T325 and T331 to alanines, and mutant 3 (m3) contains mutations of serines S362 and S374 on the C terminus to alanines. We transfected expression vectors containing the wild-type c-Fos sequence or mutations listed above in LβT2 cells. Cells were treated 24 h later with vehicle or GnRH to observe phosphorylation, in the presence of ActD and Chx to prevent compounding effects of GnRH on the induction of endogenous c-Fos. Whole-cell extracts were run on high resolution gels. Although m1 and m2 did not have an effect on the phosphorylation by GnRH, mutant m3 prevented a complete shift to the phosphorylated form of c-Fos that we observed for the wild-type protein after GnRH treatment (Fig. 7B).

Fig. 7.

C-terminal c-Fos residues S362/S374 are phosphorylated by GnRH and important for c-Fos transcriptional activity. A, Schematic representation of the c-Fos protein structure. DBD, DNA-binding domain; LZ, leucine zipper; TAD, transactivation domain. The residues in TAD that can be phosphorylated are identified above the white bar and were mutated to alanine to prevent phosphorylation. Three mutants were created representing mutations in different regions of TAD in the c-Fos protein. B, LβT2 cells were transfected with wild-type (WT) and mutant c-Fos expression vectors. After 24 h, ActD and Chx were added to prevent new transcription and translation, and vehicle or GnRH was added for 2 h, before cell lysis. Extracts were run on high-resolution gels to determine whether GnRH phosphorylates c-Fos in the C-terminal domain. The mutation m3 prevented a shift in molecular weight (of unmodifed protein, unmod.) caused by phosphorylation (phospho). C, Wild-type (WT) and mutant c-Fos proteins were overexpressed without (white bars) and with c-Jun (black bars) to induce the AP1 target reporter, FSHβ-luciferase. *, Significant difference in FSHβ promoter induction with c-Fos mutants compared with the wild-type c-Fos.

To assess the functional significance of these phosphorylation sites, we analyzed the transcriptional activity of c-Fos mutants compared with the wild type. Using transient transfection assays, we evaluated their activity alone or with c-Fos binding partner, c-Jun, with which it creates AP1 transcription factor. We cotransfected the c-Fos expression vectors with the FSHβ proximal promoter, containing an AP1 response element, linked to a luciferase reporter. As we reported previously, wild-type c-Fos induces FSHβ by 2-fold, and by 5.1-fold in the presence of c-Jun (Fig. 7C). T232 mutation (m1) significantly reduced induction by 80% when c-Fos was transfected alone and by 66% when it was cotransfected with c-Jun, compared with the wild-type c-Fos. T232 site has been associated with transcriptional activity of the c-Fos protein; however because it was not phosphorylated by GnRH, it is unlikely that it plays a role in GnRH-regulated c-Fos degradation. Mutant m2 did not have an effect on the transcriptional activity of c-Fos. Mutation m3 decreased induction of the FSHβ promoter by AP1 by 36%, suggesting that the S362/S374 site contributes to the function of the c-Fos transcription factor. Thus, the m3 mutation not only prevents c-Fos phosphorylation by GnRH, but also reduces its transcriptional activity.

Because we determined that the T232 mutation in m1 and the mutation of the S362/S374 site in m3 lower the functional activity of c-Fos, we examined whether these effects stem from the faster turnover of mutated proteins compared with wild-type c-Fos. Although we postulate that the transcriptional activity of m1 decreases due to a change in affinity for its binding partner c-Jun, because T232 is in the proximity of the leucine zipper, we analyzed the degradation of m1. We analyzed the degradation of m3, because its mutation had an effect on the transcriptional activity of c-Fos and is the mutation that affects c-Fos phosphorylation by GnRH. These proteins were overexpressed with the corresponding expression vectors for 24 h, ActD and Chx were added at time zero to prevent new transcription and translation, and the amount of c-Fos proteins was monitored for 8 h to assess degradation (Fig. 8A). Mutant m3 exhibited more rapid degradation than wild-type or m1. After quantification and normalization to β-tubulin, the amount at time zero was set to 100%. There was no difference between the degradation of wild-type c-Fos and m1, whereas m3 degraded significantly faster (Fig. 8B). After 1 h, 77.1% of wild-type c-Fos remained and 69.9% of m1, compared with 52.4% of m3. After 4 h, 14% of wild-type c-Fos was present, 12.6% of m1, compared with 2.2% of m3. After 8 h, 4.9% of wild-type and 6% of m1 are detected, compared with 0.3% of m3. We calculated the rates of degradation and determined that the rate of degradation of m1 is K = 0.4090, which is similar to the wild-type rate of K = 0.4147, but significantly lower that the rate of m3, K = 0.7852 (Fig. 8C). The half-life of wild-type c-Fos protein is 100.3 min, whereas the half-life of m3 is 52.9 min in the gonadotrope-derived cells. Although we determined that m3 is not phosphorylated by GnRH, we analyzed whether GnRH treatment can affect its half-life as it affected the half-life of the wild-type c-Fos protein. The rate of degradation of m3 is the same in the presence or absence of GnRH (Fig. 8D). Collectively, our results demonstrate that mutation of the S362/S374 sites in m3 prevents c-Fos phosphorylation by GnRH, exhibits an increased rate of turnover and shorter half-life, and lowers transcriptional activity of c-Fos and induction of its target gene, the FSHβ promoter.

Fig. 8.

C-terminal mutation of S362/S374 lowers c-Fos half-life. A, Wild-type (WT) and mutant c-Fos expression vectors were transfected in LβT2 cells to express these proteins for 24 h. ActD and Chx were added to prevent new transcription and translation at time zero for degradation studies. Cell extracts were obtained after the times indicated, and c-Fos and β-tubulin Western blots were performed. B, c-Fos Western blots were quantified on the chemiluminescence reader, normalized to β-tubulin in each sample, and the amount at time zero was set to 100% for each experiment to observe protein turnover. *, Significant difference between wild-type and mutant m3. C, Rates of degradation (K) for each c-Fos protein were calculated by the formula Ln[A]_t = −Kt + Ln[A]_0. Slope of the fit curve was determined using GraphPad Prism software. D, Mutant m3 was overexpressed in LβT2 cells for 24 h. The cells were treated with ActD and Chx, and vehicle or GnRH for times indicated above the lanes. Cells were lysed after times indicated and c-Fos Western blots were performed. Experiments were performed three times and representative blots are shown.

Discussion

FSH is required for folliculogenesis in mammals, and tight regulation of FSH levels is crucial for reproductive fitness (4). GnRH is a key regulator of FSHβ gene expression and therefore FSH synthesis. However, the mechanisms controlling the secondary rise of FSHβ at the time of low GnRH is not completely understood and cannot be solely explained by regulation of GnRH frequency. GnRH regulation of FSHβ is not direct; rather it occurs through an induction of the immediate-early genes that make up AP1 (25), of which the amount of c-Fos is strictly regulated. Herein, we identify another stratum of GnRH regulation, by decreasing the turnover rate of the c-Fos protein. This mechanism may function to selectively increase the amount of c-Fos protein and consequently transcription of its target gene, FSHβ, during the period of slow GnRH pulse frequency.

It is well established that the c-Fos protooncogene encodes a transcription factor that is a very unstable protein, exhibiting rapid turnover and a short half-life (32). Short-lived proteins are usually degraded by a proteasome, and not by lysosomes, calpain, or caspase systems (42, 43). Several posttranslational modifications of the protein can lead to degradation, of which ubiquitination targets the proteins to the 26S proteasome. c-Fos, however, can be degraded with or without ubiquitination, depending on the system and stage of the cell cycle (32, 44). Sumoylation is considered an inactivation step that leads to degradation. Of interest, c-Fos can be modified by addition of Sumo-1, -2, and -3 at lysine 257, which results in a lower AP1 transactivation activity (45). Although we did not detect a change in ubiquitination, sumoylation, or acetylation of the c-Fos protein after GnRH treatment, we did identify that c-Fos is phosphorylated by GnRH signaling pathways. There is evidence that phosphorylation of c-Fos protects the protein from being degraded by the proteasome, by changing its intracellular localization (41).

Previously, we demonstrated that p38 and ERK1/2 are involved in c-Fos induction by GnRH (38), whereas herein, we reveal that ERK1/2 is involved in c-Fos phosphorylation, also. Preceding literature reports that MAPK kinases are not only involved in c-Fos induction, but in c-Fos protein phosphorylation as well. All three branches of MAPK (i.e. ERK1/2, p38, and JNK MAPKs) can associate with c-Fos and phosphorylate its transactivation domain on multiple residues both in vitro and in vivo (39–41). This phosphorylation changes the transcriptional ability of c-Fos, and MAPK-activated c-Fos enhances AP1-driven gene expression. Although GnRH can activate these three branches of MAPK in the gonadotrope cell line, GnRH utilizes ERK1/2, and not p38 or JNK, to phosphorylate c-Fos at the C-terminal residues.

We determine that both T232 and S362/S374 in the c-Fos protein contribute to its activity, of which S362/S374 is phosphorylated by GnRH, whereas T232 is not. S362/S374 also contributes to the stability of the protein, whereas T232 does not. The phosphorylation status of c-Fos's partner, c-Jun, has been more extensively studied and reveals that c-Jun is phosphorylated by JNK at two sites, S63 and S73, which increases its transcriptional activity. A structurally similar site on c-Fos contains the T232 site and likewise allows for increased transcriptional activity after phosphorylation by MAPK (46). Both the S63/S73 site in the c-Jun protein and T232 site in the c-Fos protein are localized in proximity to their respective leucine zipper domains that are necessary for dimerization. Thus, phosphorylation of these sites may contribute to transcriptional activity by increasing dimerization, and not by affecting protein degradation. We demonstrate that T232 contributes to the functional activity of c-Fos because the mutation to alanine lowers the induction of the c-Fos target gene, FSHβ. However, we find that this site is not phosphorylated by GnRH based on the Western blots, and it does not affect c-Fos turnover. Additional MAPK sites identified by Monje et al. (40), T325 and T331, contribute to transcriptional activity of c-Fos in NIH3T3 cells. Yet, we did not observe GnRH phosphorylation of these sites, and their mutation does not affect expression of the FSHβ reporter. In contrast, we determine that GnRH phosphorylates c-Fos at S362/S374, the residues found at the C terminus. This area has been deemed important, as 20 amino acids (361–380) on the C terminus that make up the PEST domain (proline, aspartate, serine, and threonine-rich) have been implicated in the destabilization of the c-Fos protein (47). However, in contrast to our findings, the stability of c-Fos was not dependent on PEST phosphorylation. Furthermore, the critical residues for destabilization were mapped to the last six (375–380) amino acids. Similar to our results, Chen et al. (48) reported that phosphorylation of S362/S374 could lead to stabilization during the G-S transition of the cell cycle. We observe that GnRH phosphorylates S362/S374 in the c-Fos protein, which leads to increased half-life, because preventing phosphorylation by mutation increases turnover.

Our finding that GnRH protects c-Fos, without altering Egr1 degradation, is interesting in light of the role of GnRH to differentially regulate two gonadotropin β-subunit genes. We postulate that this differential regulation may stem from diverse accumulation of their respective intermediates. Specifically, Egr1 is an immediate early gene, induced rapidly by GnRH to activate LHβ transcription. Contrary to c-Fos, Egr1 degradation is not affected by GnRH. This is surprising, because Egr1 protein can be phosphorylated by AKT and MAPK pathways in other cell types, which increases its accumulation and transcriptional activity (49, 50). This also underscores the specificity of the GnRH response by demonstrating the selective posttranslational modification of c-Fos. Understanding the mechanism of specificity of GnRH signaling to precisely phosphorylate c-Fos and to increase its half-life, but not the half-life of Egr1, remains a topic of future investigation.

We can postulate from the results presented here that every pulse of GnRH causes a new wave of immediate-early gene transcription, which leads to the induction of gonadotropin β-subunits. High GnRH pulse frequency leads to elevated accumulation of Egr1 and c-Fos, and transcription of both LHβ and FSHβ, as observed during the preovulatory period. In contrast, during a period of low GnRH pulse frequency, the half-life of the c-Fos protein is specifically increased due to its phosphorylation by the ERK1/2 pathway, allowing sufficient accumulation of the c-Fos protein to increase transcription of its target. This leads to selective induction of the FSHβ subunit and its secondary rise, which is necessary for follicular growth.