MEF2 Is Restricted to the Male Gonad and Regulates Expression of the Orphan Nuclear Receptor NR4A1

Abstract

Leydig cell steroidogenesis is controlled by the pituitary gonadotropin LH that activates several signaling pathways, including the Ca²⁺/calmodulin kinase I (CAMKI) pathway. In other tissues, CAMKI regulates the activity of the myocyte enhancer factor 2 (MEF2) transcription factors. MEF2 factors are essential regulators of cell differentiation and organogenesis in numerous tissues but their expression and role in the mammalian gonad had not been explored. Here we show that MEF2 factors are expressed in a sexually dimorphic pattern in the mouse gonad. MEF2 factors are present in the testis throughout development and into adulthood but absent from the ovary. In the testis, MEF2 was localized mainly in the nucleus of both somatic lineages, the supporting Sertoli cells and the steroidogenic Leydig cells. In Leydig cells, MEF2 was found to activate the expression of Nr4a1, a nuclear receptor important for hormone-induced steroidogenesis. In these cells MEF2 also cooperates with forskolin and CAMKI to enhance Nr4a1 promoter activity via two MEF2 elements (−318 and −284 bp). EMSA confirmed direct binding of MEF2 to these elements whereas chromatin immunoprecipitation revealed that MEF2 recruitment to the proximal Nr4a1 promoter was increased following hormonal stimulation. Modulation of endogenous MEF2 protein level (small interfering RNA-mediated knockdown) or MEF2 activity (MEF2-Engrailed active dominant negative) led to a significant decrease in Nr4a1 mRNA levels in Leydig cells. All together, our results identify MEF2 as a novel testis-specific transcription factor, supporting a role for this factor in male sex differentiation and function. MEF2 was also positioned upstream of NR4A1 in a regulatory cascade controlling Leydig cell gene expression.

Members of the nuclear receptor (NR)4A subgroup of the nuclear hormone receptor superfamily were first identified as immediate early-response genes. Their expression is rapidly induced by a variety of physiological stimuli, including fatty acids, prostaglandins, growth factors, calcium, cytokines, peptide hormones, phorbol esters, and neurotransmitters (1–8). Members of this subgroup include NR4A1 (NUR77), NR4A2 (NURR1), and NR4A3 (NOR1). These transcription factors are known to bind DNA as monomers to a consensus nerve growth factor I-B response element sequence (AAAGGTCA) (9), or as homo- or heterodimers to a palindromic Nur response element sequence (9–11). NR4A1 and NR4A2, but not NR4A3, can heterodimerize with the retinoid X receptor (12, 13). In addition to the different DNA-binding activities (monomer, homodimer, heterodimer), the specificity of action of NR4A1 can also be regulated by phosphorylation. In fact, the phosphorylation status of NR4A1 is known to modulate its DNA-binding capacity, protein-protein interaction, and nuclear localization (reviewed in References 14 and 15).

We and others have shown that NR4A1 is strongly and rapidly induced in Leydig cells in response to LH/cAMP (4, 16–19). In these cells, NR4A1 was shown to be an important regulator of several steroidogenic genes including mouse Star (20–22), mouse Hsd3b1 (23), human HSD3B2 (4), and mouse Cyp17a1 (24). Furthermore, maximal expression of the Star gene in response to hormone stimulation was found to require NR4A1 (22). Despite its well-characterized role in Leydig cell function, the mechanisms that regulate Nr4a1 expression in these cells remain poorly understood. Previously, we have shown that Nr4a1 activation in response to cAMP in Leydig cells does not require de novo protein synthesis and rather relies on activation (eg, by posttranslational modifications) of transcription factors already present in the cell (25). Moreover, we have identified 3 distinct regions of the Nr4a1 promoter that we named NIR-A, B, and C (Nur77 important region) and found that NIR-B drives basal activity whereas NIR-A and NIR-C are essential for hormone responsiveness (25). Finally, we found that the LH/cAMP-mediated increase in Nr4a1 expression involves Ca²⁺/calmodulin kinase I (CAMKI) activity (20, 25), although the transcription factor(s) targeted by CAMKI remain unknown.

Myocyte enhancer factor 2 (MEF2) proteins are a family of 4 transcription factors, (MEF2A, 2B, 2C, 2D) encoded by distinct genes. MEF2 members display distinct, but overlapping, temporal and spatial expression patterns in embryonic and adult tissues and have partially redundant functions (reviewed in Reference 26). They are known to play critical roles in cell differentiation, proliferation, survival, and apoptosis in a wide range of cell types (reviewed in Reference 26). MEF2 factors have been identified as key downstream effectors of the Ca²⁺-signaling pathway and are involved in gene expression in muscle, neuronal, and immune cells (27). MEF2 factors bind as dimers to the palindromic AT-rich sequence (C/T)TA(A/T)₄TA(G/A) in the promoter region of target genes (reviewed in Reference 26). In T cells, MEF2 factors have been shown to bind to two consensus MEF2 elements important for the Ca²⁺-dependent activation of the Nr4a1 promoter (28). Despite their well-established roles in other tissues, no data are currently available regarding their expression in the gonads.

In the present study, we report that MEF2 is expressed in a sexually dimorphic pattern in the mouse gonad; it is found exclusively in somatic cells of the testis throughout development and into adulthood but not in the ovary. In Leydig cells, using overexpression, small interfering RNA (siRNA)-mediated knockdown as well as a MEF2-Engrailed active dominant negative, we found that endogenous MEF2 factors are functionally active and regulate Nr4a1 expression in cooperation with the adenylate cyclase activator forskolin (Fsk) or CAMKI. Our data identify MEF2 as a novel testis-specific factor involved in the regulation of hormone-dependent gene expression in steroidogenic Leydig cells.

Materials and Methods

Plasmids

The rat −1013-bp Nr4a1 (nerve growth factor I-B) promoter reporter plasmid, its deletions and constructs harboring different regions called NIR (Nur77 important region) of the rat Nr4a1 promoter were described previously (25). The NIR-ABC and NIR-A reporters containing a mutation of the MEF2 element(s) at −318 and/or at −284 were obtained by site-directed mutagenesis using the QuikChange XL mutagenesis kit (Stratagene) using primers listed in Supplemental Table 1. Expression vector for MEF2D was generated by PCR using cDNAs prepared from mouse testis and primers listed in Supplemental Table 1, followed by cloning into the pcDNA3 expression vector (Life Technologies). Expression vectors encoding wild-type and constitutively active forms of CAMKI (29) were obtained from Dr Thomas Soderling (Oregon Health Sciences University, Portland, OR). The MEF2-Eng expression vector was generated by transferring the cDNA from a plasmid originally obtained from Dr Ilona S. Skerjanc (University of Ottawa, Canada) into the pcDNA3 expression vector. Expression vector for Eng was generated by PCR using the MEF2-Eng plasmid as template using primers listed in Supplemental Table 1 and subcloned in the pcDNA3 expression vector containing a hemagglutinin tag. All plasmids were verified by sequencing (Centre de Génomique de Québec, Québec City, Canada).

Protein purification and Western blots

For cAMP stimulations, MA-10 Leydig cells were treated with 0.5 mM 8-bromo-cAMP (8Br-cAMP, Sigma-Aldrich Canada) for time intervals ranging from 0–6 hours. Cells were then rinsed twice with ice-cold PBS, and total and nuclear proteins were prepared as described previously (20). Total (40 μg) and nuclear (15 μg) proteins were fractionated by SDS-PAGE and transferred onto polyvinylidene difluoride membrane (EMD Millipore). Immunodetection was performed using an avidin-biotin approach according to the manufacturer's instructions (Vector Laboratories, Inc) or an enhanced chemiluminescence approach according to the manufacturer's protocol (GE Healthcare), as indicated. Detection of MEF2, α-TUBULIN, and LAMIN B was performed using an anti-MEF2 antiserum (C-21, 1:500 dilution; Santa Cruz Biotechnology), a monoclonal anti-α-TUBULIN antibody (1:5000 dilution; Sigma-Aldrich Canada), and an anti-LAMIN B antiserum (C-20, 1:500 dilution; Santa Cruz Biotechnology).

Immunohistochemistry

Detection of MEF2 by immunohistochemistry was performed as described elsewhere (20) on testis or ovary sections from CD-1 mice at various developmental stages. MEF2 localization was assessed using an anti-MEF2 polyclonal antiserum (C-21, 1:500 dilution; Santa Cruz Biotechnology). The double immunohistochemistry to identify Leydig cells was performed with an anti-CYP17A1 antiserum (1:200 dilution, C-17, Santa Cruz Biotechnology). The counterstain with hematoxylin following MEF2 detection was omitted and replaced by incubation with the anti-CYP17A1 overnight at 25°C. Samples were washed twice with 1× PBS for 10 minutes, and an alkaline phosphatase-conjugated secondary antiserum (1:1000, Sigma-Aldrich) was added and incubated for 1 hour at 25°C. Detection was done using alkaline phosphatase detection solution (0.1 M Tris-HCl, pH 9.2; 1 mM Levamisole; Naphtol, 0.04%; and Fast Blue BB Diazonium salt, 0.2%). Omission of the primary antiserum was used as negative control. All experiments were conducted according to the Canadian Council for Animal Care and have been approved by the Animal Care and Ethics Committee of Laval University (protocol 2003-068).

Immunofluorescence

MA-10 cells (250 000 cells) were plated on poly-l-lysine-coated (200 μg/mL) glass coverslips in 60-mm petri dishes. Forty-eight hours later, cells were fixed in cold methanol, permeabilized with 0.1% Triton X-100 in PBS, and blocked in PBS with 1% BSA + 10% goat serum for 30 minutes with agitation. Cells were then washed 3 times with 1× PBS for 5 minutes at room temperature. Detection was performed using a rabbit polyclonal anti-MEF2 antiserum (C-21, Santa Cruz Biotechnology) and a monoclonal anti-α-TUBULIN antiserum (1:500, T5168, Sigma-Aldrich Canada) in 1× PBS + 1% BSA incubated overnight at 4°C with light agitation. The next day, cells were washed with 1× PBS and incubated with a goat antirabbit IgG conjugated to Alexa fluor 546 (1:300, A-11010; Life Technologies) and a goat antimouse IgG conjugated to Alexa fluor 488 (1:300, A-11001; Life Technologies) in 1× PBS with 1% BSA for 1 hour at room temperature. Cells were washed and the glass coverslips were put on a microscope slide with Slow Fade Gold Antifade Reagent (S36936, Life Technologies). The microscope slides were kept at 4°C until images were obtained using an Axioskop2 Plus microscope (Carl Zeiss Canada Ltd.) and the Image-Pro Plus software (MediaCybernetics).

RNA isolation, traditional PCR, and quantitative PCR

Total RNA extraction and cDNA synthesis were performed as described elsewhere (20). For the various MEF2 family members, PCRs were done on a PXE 0.2 thermal cycler (Thermo Electron Corp) using Vent DNA polymerase (New England Biolabs) and the following conditions: 3 minutes at 95°C followed by 32 cycles (for Mef2a and Mef2b) or 30 cycles (for Mef2c and Mef2d) of denaturation (1 minute at 95°C), annealing (1 minute at 62°C for Mef2a and Mef2b or 60°C for Mef2c and Mef2d), and extension (30 seconds at 72°C), with a final extension step of 5 minutes at 72°C. PCR products were analyzed by agarose gel electrophoresis and ethidium bromide staining. Primers used are described in Supplemental Table 1. Quantitative RT-PCR was performed using a LightCycler 1.5 instrument and the LightCycler FastStart DNA Master SYBR Green I (Roche Diagnostics) according to the manufacturer's protocol. PCRs were performed using the previously described conditions for Nr4a1 (20) and were as follow for Mef2c and Mef2d using the primers listed in Supplemental Table 1: 10 minutes at 95°C followed by 35 cycles of denaturation (5 seconds at 95°C), annealing (5 seconds at 62°C), and extension (20 seconds at 72°C) with single acquisition of fluorescence at the end of each extension steps. As internal control, PCRs were performed using previously described Rpl19-specific primers (30) and conditions (20). Specificity of the PCR products was confirmed by analysis of the melting curve and agarose gel electrophoresis. Quantification of gene expression was performed using Relative Quantification Software (Roche Diagnostics) and is expressed as a ratio of target to Rpl19 levels. Each amplification was performed in duplicate using 3 different preparations of first-strand cDNAs for each of at least 3 different RNA extractions.

Chromatin immunoprecipitation (ChIP) assay

ChIP assay was performed as previously described (20, 31). Genomic DNA was immunoprecipitated with an anti-MEF2 antiserum and directly analyzed by quantitative PCR using primers (listed in Supplemental Table 1) specific for the −378 to −259 bp region of the mouse Nr4a1 promoter. The quantitative PCR conditions were the following: 10 minutes at 95°C followed by 35 cycles of denaturation (5 seconds at 95°C), annealing (5 seconds at 62°C), and extension (20 seconds at 72°C) with single acquisition of fluorescence at the end of each extension step. Absolute quantification of Nr4a1 promoter fragments immunoprecipitated with the anti-MEF2 antiserum was performed using a standard curve generated with known concentrations of the −1013-bp Nr4a1 promoter construct and is expressed as a ratio of Nr4a1 promoter-immunoprecipitated DNA to input DNA levels. Input DNA represents 5% of total DNA used for a ChIP experiment.

EMSAs

EMSAs were performed using nuclear extracts from MA-10 Leydig cells either untreated or treated with 0.5 mM 8Br-cAMP as described previously (4). The ³²P-labeled double-stranded oligonucleotides used as probes are listed in Supplemental Table 1. For the competition experiments, double-stranded oligonucleotides listed in Supplemental Table 1 (2- and 5-fold molar excess) corresponding to the wild-type or a mutated version of the MEF2 elements were used. For supershift experiments, 1.6 μg of anti-MEF2 antiserum was also added to the binding reaction.

Cell culture and transfections

Mouse MA-10 Leydig cells (32), provided by Dr Mario Ascoli (University of Iowa, Iowa City, IA), were grown and transfected as described previously (20, 33). In experiments with stimulations, cells were treated with 0.5 mM 8Br-cAMP or 10 μM Fsk (Sigma-Aldrich Canada) for different time intervals before harvesting. In knockdown experiments, MA-10 Leydig cells were transfected using JetPrime (PolyPlus-Transfections) with 200 nM siRNA (Life Technologies) directed against MEF2A/D or scrambled siRNAs as control and treated with 10 μM Fsk for 4 hours in serum-free media. The sequences of the siRNA oligonucleotides are listed in Supplemental Table 1.

Statistical analyses

To identify significant differences between multiple groups, statistical analyses were done using either a one-way ANOVA followed by Bonferroni or Newman-Keuls test or a nonparametric Kruskal-Wallis one-way ANOVA followed by a Dunn's multiple-comparisons test when conditions of normality and/or equal variance between groups were not met. Single comparisons between 2 experimental groups were done using either an unpaired Student's t test or a Mann-Whitney U-test when conditions of normality/variance failed. For all statistical analyses, P < .05 was considered significant. All statistical analyses were done using the GraphPad Prism software (GraphPad Software, Inc).

Results

MEF2 members are expressed in testicular somatic cells

We previously reported that cAMP-mediated Nr4a1 expression involves a CAMK-dependent transcription factor that acts through the NIR-A (−331 to −233 bp) and/or NIR-C (−121 to −65 bp) regions of Nr4a1 promoter (25). Interesting candidates include members of the MEF2 family of transcription factors for 3 reasons: 1) they are known effectors of the CAMK pathway in other tissues (reviewed in Reference 27); 2) there are 3 consensus MEF2 elements in the Nr4a1 promoter, 2 of which are conserved across species and located in the NIR-A region (Supplemental Figure 1); and 3) they have been implicated in calcium-dependent activation of the Nr4a1 promoter in T cells (28).

The presence of MEF2 proteins at the cellular level in vivo was therefore assessed by immunohistochemistry on mouse testis and ovary sections at various developmental stages (E18.5, P5, P21, and P32). MEF2 was detected (brownish staining) in the nucleus of both somatic cell populations of the testis: interstitial Leydig cells (arrows) and tubular Sertoli cells (arrowheads) (Figure 1, A, C, E, and G). This labeling is specific because no signal was observed when the primary antibody was omitted (data not shown). Surprisingly, MEF2 was absent from the ovary at all stages tested (Figure 1, B, D, F, and H). Thus, within the mammalian gonad, MEF2 is present almost exclusively in Leydig and Sertoli cells of the testis.

Figure 1.

MEF2 factors are specifically expressed in male gonad. MEF2 proteins are expressed in somatic cells of the male but not the female gonad. Immunohistochemistry was performed on embryonic (E18.5), neonatal (P5), prepubertal (P21) and P32 mouse testis (A, C, E, and G) and ovary (B, D, F, and H) sections using an anti-MEF2 antiserum. MEF2 was detected in interstitial (arrows) and Sertoli (arrowheads) cells of the testis at all stages tested (brownish staining). Magnification, ×400.

To further characterize MEF2 expression in the testis, a detailed developmental profile was performed using immunohistochemistry on mouse testis sections at embryonic day 14.5 (E14.5), E18.5, E19.5, postnatal day 0 (P0), P2, P5, P10, P21, P30, P32, P36, and P120. MEF2 was detected in the nucleus of Sertoli (solid arrowheads) and interstitial/Leydig (arrows) cells at all ages tested, although the signal appears stronger in Sertoli cells (Figure 2). Peritubular cells and gonocytes/germ cells were negative at all times (Figure 2). Interstitial cells were identified as Leydig cells by double labeling with the steroidogenic enzyme CYP17A1 at E19.5, P2, P36, and P120 (Figure 3). Although most interstitial cells were positive for both MEF2 and CYP17A1 (arrows in Figure 3), some were not, which indicates that MEF2 is also expressed in interstitial cells that do not correspond to steroidogenically active fetal Leydig cells (open arrowheads in Figure 3). Thus, MEF2 is a novel transcription factor present in testicular somatic cells throughout development and into adult life.

Figure 2.

MEF2 factors are present in testicular somatic cells throughout life. A–L, Immunohistochemistry was performed on embryonic (E14.5, E18.5, E19.5), neonatal (P0, P2, P5), prepubertal (P10, P21), pubertal (P30, P32), adult (P36), and aging (P120) mouse testis sections using an anti-MEF2 antiserum. MEF2 was detected in interstitial (arrows) and Sertoli (solid arrowheads) cells at all stages tested (brownish staining). Omission of the primary antibody served as negative control (data not shown). Magnification, ×400.

Figure 3.

MEF2 factors colocalize with CYP17A1 in Leydig cells. A–D, Double-immunohistochemistry was performed on embryonic (E19.5), neonatal (P2), adult (P36), and aging (P120) mouse testis sections using an anti-MEF2 and an anti-CYP17A1 antiserum. MEF2 (brownish staining) and CYP17A1 (blue staining) colocalization was detected in Leydig cells (arrows). MEF2 was also detected in non-Leydig interstitial cells (open arrowheads) and in Sertoli (solid arrowheads) cells. Omission of the primary antibody served as negative control (data not shown). Magnification, ×400.

To address the role of MEF2 factors in Nr4a1 expression in Leydig cells, we determined whether MEF2 factors are expressed in these cells. As shown in Figure 4A, a band was detected by Western blot using nuclear extracts from MA-10 Leydig cells, and its intensity was not affected by cAMP treatment. Because the antiserum recognizes MEF2A, -2C, and -2D (with a better affinity for MEF2A), we performed RT-PCR for the different MEF2 family members using cDNAs from mouse testis, MA-10 Leydig cells, and mouse brain. As shown in Figure 4B, the mRNA for all four MEF2 members (Mef2a, 2b, 2c, and 2d) was detected in MA-10 Leydig cells as well as in whole testis and brain. The presence of MEF2 in the nucleus of MA-10 Leydig cells was also confirmed by immunofluorescence (Figure 4C). Thus, all four MEF2 family members are expressed in the testis and in the nucleus of MA-10 Leydig cells.

Figure 4.

Members of the MEF2 family of transcription factors are expressed in Leydig cells. A, MA-10 Leydig cells were treated with 0.5 mM 8Br-cAMP for the indicated times. Proteins were then isolated, separated by SDS-PAGE, transferred to polyvinylidene difluoride membrane. Immunodetection were performed using antisera specific for MEF2 family members or α-TUBULIN (as a loading control). B, Semiquantitative RT-PCR using primers specific for different MEF2 family members was performed using cDNAs from mouse brain, testis, and MA-10 Leydig cells. Results are representative of 3 individual experiments. C, Immunofluorescence was used to confirm the nuclear localization of MEF2 (red) in MA-10 Leydig cells. The cytoplasm (green) was labeled using α-TUBULIN. Magnification, ×400.

MEF2 binds to the NIR-A region of Nr4a1 promoter

Next we tested whether MEF2 could bind to the Nr4a1 promoter by EMSA. In addition to the 2 previously characterized MEF2 elements at −318 and −284 bp (28), we located another MEF2 consensus element at −958 bp (Supplemental Figure 1). As shown in Figure 5A, a single complex from MA-10 Leydig cell nuclear extracts was observed with all three MEF2 probes. This complex was not affected by a 2 hours 8Br-cAMP treatment (Figure 5A, lanes 1–2, 7–8, and 13–14). This complex was competed by unlabeled oligonucleotides corresponding to the probe (Figure 5A, lanes 3–4, 9–10, and 15–16) but not by oligonucleotides harboring a mutation that destroys the MEF2 site (Figure 5A, lanes 5–6, 11–12, and 17–18). To confirm that this complex contained a MEF2 protein, an anti-MEF2 antibody was added to the binding reaction. As shown in Figure 5B, the binding was supershifted for all 3 probes (Figure 5B, lanes 5, 10, and 15) indicating that endogenous MEF2 present in MA-10 Leydig cells can bind to all three MEF2 consensus elements in the Nr4a1 promoter.

Figure 5.

MEF2 family members bind to the proximal region of Nr4a1 promoter. A, EMSA was used to assess protein binding from MA-10 nuclear extracts to double-stranded ³²P-labeled oligonucleotides corresponding to the 3 MEF2 elements at −958 bp, −318 bp, or −284 bp of Nr4a1 promoter. Binding was then challenged by increasing doses (black triangles; molar excesses of 2-fold and 5-fold) of unlabeled oligonucleotides corresponding to the wild-type (WT) or mutated (Mut) MEF2 element as indicated. B, MEF2 binding was confirmed by supershift (lanes 5, 10, and 15) using a MEF2 antiserum specific for MEF2 family members (αMEF2). P.I., preimmune serum. Nuclear extracts used in EMSAs were from MA-10 Leydig cells treated or not for 2 hours with 0.5 mM 8Br-cAMP. C, MA-10 Leydig cells were treated with 0.5 mM 8Br-cAMP for the indicated times and ChIP assays were performed. Chromatin was immunoprecipitated with either a MEF2 (black bars) antiserum or IgG (open bars), which was used as control. A 120-bp fragment of the proximal Nr4a1 promoter encompassing the two MEF2 elements at −318 and −284 bp was amplified by quantitative PCR immediately following the ChIP assay. An aliquot (5%) of chromatin preparation before immunoprecipitation (input) was used as positive control. Results are presented as a ratio of Nr4a1 promoter-immunoprecipitated DNA to input DNA levels from 3 independent experiments (±SEM). An asterisk (*) indicates a statistically significant difference (P ≤ .001).

To investigate whether MEF2 was recruited to the NIR-A region of the Nr4a1 promoter in vivo, we performed ChIP assays followed by quantitative PCR. As shown in Figure 5C, an increase in MEF2 association with the proximal Nr4a1 promoter was detected 1 hour after cAMP stimulation and was more evident at 2 hours. Thus, in Leydig cells, MEF2 associates with the NIR-A region of the Nr4a1 promoter, which contains the −318 and −284 MEF2 consensus elements.

MEF2 factors and CAMKI cooperate to activate Nr4a1 transcription

Knowing that MEF2 is recruited to the Nr4a1 promoter, we next tested whether MEF2D could activate a −1013-bp rat Nr4a1 promoter by transient transfections in MA-10 Leydig cells. As shown in Figure 6, MEF2D on its own cannot activate the −1013-bp Nr4a1 promoter. This can be explained by the already high levels of MEF2 protein in MA-10 cells or by the requirement of posttranslational modifications. Because MEF2 is known to be phosphorylated and activated by CAMK in other systems (34) and because CAMKI is expressed in Leydig cells (20), we therefore assessed the role of CAMKI and MEF2 in Nr4a1 promoter activation. MA-10 cells were cotransfected with a −1013-bp Nr4a1 reporter construct along with expression vectors for MEF2D and wild-type (WT) or constitutively active (CA) CAMKI. As previously reported (25), CAMKI on its own activated the Nr4a1 reporter: 2-fold for wild-type CAMKI and 6-fold for the CA form (Figure 6). Combination CAMKI with MEF2 led to a more robust activation of the Nr4a1 promoter, up to 4-fold for CAMKI WT/MEF2D and up to 9-fold for CAMKI CA/MEF2D (Figure 6). Similar results were obtained using MEF2C (data not shown). In order to locate which region of the Nr4a1 promoter is involved in MEF2D/CAMKI CA transcriptional cooperation, various Nr4a1 promoter constructs were transfected in combination with MEF2D and/or CAMKI CA. As shown in Figure 7A, MEF2D/CAMKI cooperation was only observed in constructs containing the NIR-A region. Furthermore, a construct containing only the NIR-A region was activated (∼4-fold) by MEF2D alone (Figure 7B). These data support a role for the −318 and −284 MEF2 elements. To confirm this, mutations known to abrogate MEF2 binding (Figure 5) were introduced in the MEF2 elements either individually or in combination. As shown in the left panel of Figure 7B, mutation of the −318 or the −284 MEF2 element individually caused a 2-fold decreased in CAMKI-dependent activation of the NIR-A reporter, suggesting that CAMKI acts, at least in part, by targeting endogenous MEF2 factors. The CAMKI/MEF2D cooperation was decreased but not abolished when each MEF2 element was individually mutated (left panel of Figure 7B). However, when both MEF2 elements were mutated, the MEF2D/CAMKI cooperation was abrogated indicating that both MEF2 elements are functional and necessary for the cooperation with CAMKI. We have previously reported that stimulation of Leydig cells with Fsk/cAMP leads to increased Ca²⁺ release from internal stores (33) and CAMKI activation (20). Furthermore, Fsk-dependent activation of Nr4a1 was inhibited by blocking CAMKI activity (20). These data established that CAMKI is activated following Fsk/cAMP treatment. Consistent with this, Fsk could also cooperate with MEF2 to increase the activity of the NIR-A region of the Nr4a1 promoter (Figure 7B, right panel).

Figure 6.

MEF2D transcriptionally cooperates with CAMKI on the Nr4a1 promoter. MA-10 Leydig cells were cotransfected with a −1013 to +52-bp rat Nr4a1 promoter construct along with an empty expression vector (−) or an expression vector for MEF2D in the absence or presence of expression vectors encoding either wild-type (Wt) or a constitutively active (CA) form of CAMKI. The number of experiments, each performed in duplicate, is indicated. Results are shown as fold activation over control (± SEM). An asterisk (*) indicates a statistically significant difference (P < .05).

Figure 7.

The MEF2 elements within the NIR-A region of the Nr4a1 promoter mediate the MEF2/CAMKI cooperation. A, The NIR-A region of Nr4a1 promoter is required for the MEF2D/ CAMKI cooperation. MA-10 Leydig cells were cotransfected with either an empty expression vector (open bars) or expression vectors for MEF2D (hatched bars), CAMKI CA (gray bars), and MEF2D+CAMKI CA (black bars) along with a −1013-bp Nr4a1 reporter, a minimal −65-bp Nr4a1 reporter, and various reporter constructs containing different combinations of the NIR-A, -B, and -C regions. B, MEF2 activates the Nr4a1 promoter via the −318 and −284 MEF2 elements. MA-10 Leydig cells were cotransfected with a reporter construct containing the wild-type NIR-A region or a mutated version harboring a 2-nucleotide mutation in the MEF2 element at −318 bp (CTATATTTAG to CgATATTTcG), and/or a 2-nucleotide mutation in the MEF2 element at −284 bp (CTATTTATAG to CgATTTATcG) along with either an empty expression vector (open bars) or expression vectors for MEF2D (hatched bars), CAMKI CA, or Fsk (gray bars) and MEF2D+CAMKI CA or MEF2D+Fsk (black bars). The mutated elements are represented by a large X. The number of experiments, each performed in duplicate, is indicated. Results are shown as fold activation over control (± SEM). An asterisk (*) indicates a statistically significant difference (P < .05). CTL, control.

MEF2 factors in Leydig cells are required for proper Nr4a1 gene transcription

Our data suggest that CAMKI mediates part of its effects via endogenous MEF2 factors (Figure 7B). Although PCR data suggest the presence of all 4 MEF2 factors in MA-10 Leydig cells (Figure 4B), immunoprecipitations followed by LC-MS/MS revealed that MA-10 Leydig cells express mainly MEF2A and MEF2D (data not shown). To further assess the contribution of endogenous MEF2 factors to Nr4a1 transcription, we first used an active dominant-negative MEF2 construct, MEF2-Eng, in which the C-terminal MEF2 activation domain was replaced with the Engrailed repressor domain (35). This fusion protein has been shown to bind to MEF2 element (Reference 36 and our unpublished data) and to actively repress promoter activity in cardiomyocytes (36). We found that MEF2-Eng could efficiently compete the cooperation between wild-type MEF2D and CAMKI (Figure 8A) or Fsk (Figure 8B) on the NIR-A reporter, whereas overexpression of the Eng repressor domain alone had no impact on the cooperation (Figure 8, A and B).

Figure 8.

A MEF2-Eng dominant negative represses the CAMKI-dependent activation of the Nr4a1 promoter. A, A MEF2-Eng fusion protein behaves as a dominant negative. MA-10 Leydig cells were cotransfected with a reporter construct containing the NIR-A region of the Nr4a1 promoter or a minimal −65-bp Nr4a1 reporter along with either an empty expression vector (open bars) or expression vectors for MEF2D + CAMKI CA (hatched bars), MEF2D + CAMKI CA + MEF2-Eng (gray bars), and MEF2D + CAMKI CA + Eng (black bars). B, Same as described in panel A except 10 μM Fsk was used instead of the CAMKI CA expression vector. C, CAMKI-dependent activation of the Nr4a1 promoter requires endogenous MEF2 factors. MA-10 Leydig cells were cotransfected with either an empty expression vector (open bars) or expression vectors for CAMKI CA (hatched bars), MEF2-Eng (gray bars), and CAMKI CA + MEF2-Eng (black bars) along with a −1013-bp Nr4a1 reporter, the NIR-A reporter, and a minimal −65-bp Nr4a1 reporter. D, Fsk activates Nr4a1 promoter activity via endogenous MEF2 factors. The same experiments described in panel C were performed except that 10 μM Fsk was used instead of the CAMKI CA expression vector. The number of experiments, each performed in duplicate, is indicated. Results are shown as fold activation over control (± SEM). An asterisk (*) indicates a statistically significant difference (P < .05). CTL, control.

Knowing that MEF2-Eng can specifically repress the MEF2D/CAMKI and Fsk/CAMKI cooperation on the NIR-A region of the Nr4a1 promoter, MEF2-Eng was next used to investigate the contribution of endogenous MEF2 factors to Nr4a1 transcription. MA-10 Leydig cells were transfected with MEF2-Eng with or without CAMKI CA or Fsk along with various Nr4a1 reporter constructs. As expected, overexpression of CAMKI CA alone (Figure 8C) or stimulation with Fsk (Figure 8D) resulted in an activation of the −1013-bp Nr4a1 reporter as well as of a reporter containing only the NIR-A region (which harbors two MEF2 elements) but not of a minimal −65-bp reporter construct devoid of MEF2 elements. The CAMKI/Fsk-induced activation was decreased by approximately 50% when MEF2-Eng was cotransfected (Figure 8, C and D). Expression of MEF2-Eng alone had no effect on promoter activity (Figure 8, C and D). These data indicate that endogenous MEF2 factors do not contribute to basal Nr4a1 promoter activity but are activated/targeted by CAMKI and Fsk in these cells.

Having established that MEF2 factors are important for Nr4a1 promoter activity, we next studied the involvement of MEF2 factors on Nr4a1 expression by analyzing endogenous Nr4a1 mRNA levels. Overexpression of MEF2D or of the MEF2-Eng dominant negative had no significant effect on basal Nr4a1 mRNA levels (Figure 9A). As expected, treatment with Fsk caused an increase in Nr4a1 mRNA levels that was further enhanced by MEF2D (Figure 9A). However, in the presence of MEF2-Eng, the Fsk-mediated increase in Nr4a1 mRNA levels was completely abolished (Figure 9A). To complement these data, we used an siRNA knockdown approach. MA-10 Leydig cells were transfected with scrambled siRNAs or siRNAs directed against MEF2A/2D and treated or not with Fsk. As shown in the upper panel of Figure 9B, MEF2 protein levels were dramatically reduced in cells exposed to siRNAs directed against MEF2A/2D compared with controls. In MEF2-depleted MA-10 Leydig cells, basal Nr4a1 mRNA levels remained unchanged whereas the Fsk-mediated increase was eliminated (Figure 9B, lower panel). Altogether, these results confirm a role for endogenous MEF2 factors exclusively in hormone/Fsk/cAMP-induced Nr4a1 gene expression in MA-10 Leydig cells.

Figure 9.

MEF2 is required for Fsk-induced Nr4a1 endogenous gene transcription. A, MA-10 Leydig cells were transfected with either an empty expression vector (open bars) or expression vectors for MEF2D (hatched bars) or MEF2-Eng (black bars) and treated or not with 10 μM Fsk for 2 hours. Total RNAs were isolated and Nr4a1 mRNA levels were determined by quantitative PCR and results were normalized to Rpl19 cDNA (± SEM). Results are representative of 4 individual experiments. B, MA-10 Leydig cells were transfected with siRNA directed against MEF2A/2D (hatched and black bars) or with a control nontargeting siRNA (white and gray bars). Prior to harvesting, cells were treated (gray and black bars) or not (white and hatched bars) with 10 μM Fsk for 4 hours. Western blot (top panel) was used to confirm MEF2 protein level; α-TUBULIN was used as a loading control. Quantitative PCR (lower panel) was used to determine Nr4a1 mRNA levels, and results were normalized to Rpl19 cDNA (± SEM). An asterisk (*) indicates a statistically significant difference (P < .05). DMSO, dimethylsulfoxide.

Discussion

Signaling pathways regulating Nr4a1 expression

Accumulating evidence indicates that the orphan nuclear receptor NR4A1 is an important regulator of basal and hormone-induced gene expression in testicular Leydig cells (20, 21, 25). Although Nr4a1 expression is rapidly induced in response to LH/cAMP in these cells (4, 16–19), the mechanisms that regulate its expression in Leydig cells remain poorly understood compared with other tissues and cell types. For instance, different signaling pathways are involved in Nr4a1 expression, including cAMP and CAMK in adrenal cells (37, 38), and LH and PGF2α via ERK1/2 and CAMK in the ovary (39). In Leydig cells, we found that induction of Nr4a1 in response to LH/cAMP was preferentially mediated through a CAMKI-dependent pathway (20, 25). However, the exact target(s) of CAMKI in the regulation of Nr4a1 expression in Leydig cells remained to be fully elucidated.

MEF2 factors are present exclusively in the male gonad

Our previous analyses of the rat Nr4a1 promoter have revealed that it contains 3 distinct regions (NIR-A, -B, and -C) regulating basal activity and hormone responsiveness (25). With respect to cAMP responsiveness, the rapid and strong induction of Nr4a1 promoter activity is consistent with the fact that cAMP-mediated up-regulation of Nr4a1 transcription does not require de novo protein synthesis and rather relies on posttranslational modifications of transcription factors already present in the cell (25). Analysis of NIR-A, one of the cAMP-responsive regions in the Nr4a1 promoter, revealed the presence of binding sites for members of the MEF2 family of transcription factors. These factors have been reported to activate the Nr4a1 promoter in a CAMK-dependent manner in T-cell apoptosis (40). Up until now, no data were available regarding the expression of MEF2 in the gonads or testicular Leydig cells and whether they contribute to Nr4a1 transcription in these cells. We found that MEF2 factors are present in testis but not the ovary. This is an exciting finding because it suggests that MEF2 factors might be involved in testis differentiation. So far, only a handful of transcription factors have been shown to regulate this essential developmental process and include SRY, SOX9, NR5A1/SF1, and GATA4 (reviewed in Reference 41). Such a role for MEF2 factors would be consistent with their well-established function as critical developmental regulators, particularly in cardiac and skeletal muscles (reviewed in Reference 26).

Within the testis, MEF2 was found almost exclusively in the nucleus of both somatic cell lineages, Sertoli and Leydig/interstitial cells, throughout fetal development and into adulthood. Colabeling with CYP17A1, a marker of Leydig cells, revealed that most MEF2-positive interstitial cells are indeed Leydig cells. However, some interstitial cells were positive for MEF2, but negative for CYP17A1, indicating that this factor is also present in cells other than steroidogenically active Leydig cells. Some of these cells are likely blood vessels because MEF2 factors are known to be expressed in endothelial and smooth muscle cells (26). Some of these cells in the fetal testis could also represent precursors of the adult Leydig cell population as suggested by Barsoum et al (42). At E14.5, MEF2 signal was already clearly visible in Sertoli cells and much stronger than in interstitial cells, which would suggest that MEF2 might be activated in Sertoli cells prior to interstitial cells. The fact that Sertoli cell specification occurs prior to, and is essential for, Leydig cell formation (reviewed in Reference 43) further supports a role for this family of transcription factors in early testis formation. Of the four MEF2 factors, MEF2A and MEF2D appear to be the predominant members expressed in Leydig and Sertoli cell lines. Targeted deletion of both genes in the mouse has been reported; Mef2a-null mice die soon after birth due to cardiac defects (44) whereas Mef2d-deficient mice have no overt developmental phenotype likely due to redundancy by other MEF2 factors (45). The use of a MEF2-Eng dominant-negative protein that prevents the action of all MEF2 factors proved useful in our cell line experiments and represents a promising approach to decipher the role of MEF2 factors in testis formation and function in transgenic mice. Overexpression of MEF2-Eng in cardiomyoblasts was successfully used in vivo to show that MEF2 factors are essential for cardiomyogenesis (35).

MEF2 factors as downstream effectors of Ca²⁺/CAMKI in Leydig cells

The 2 consensus MEF2 elements at −318 and −284 bp in the NIR-A region of the Nr4a1 promoter have been implicated in the calcium-dependent activation of Nr4a1 in T cells (28). Here we have shown that these elements are involved in MEF2/CAMKI cooperation to regulate Nr4a1 transcription in Leydig cells. We have also identified a previously uncharacterized potential MEF2 element at −958 bp further upstream in the Nr4a1 promoter. Although this distal element can be recognized by MEF2 factors present in MA-10 Leydig cells, it appears to be dispensable for Nr4a1 promoter activation in these cells because its deletion has no major impact on basal activity and cAMP responsiveness or on the MEF2D/CAMKI cooperation. This is consistent with the fact that of the three MEF2 elements, the distal MEF2 element is the least conserved across species (Supplemental Figure 1). Although the MEF2D/CAMKI cooperation requires the NIR-A region of Nr4a1 promoter, some of the CAMKI-dependent activation was independent of the NIR-A region, suggesting that CAMKI action on the Nr4a1 promoter in Leydig cells likely involves other transcriptional partners that, in addition to MEF2, remain to be identified.

Although we have detected members of the MEF2 family in Leydig cells and showed that they are essential for Fsk/cAMP-induced Nr4a1 expression, MEF2 protein levels are not increased by cAMP stimulation. No increase in MEF2 binding to its elements in the Nr4a1 promoter was detected in vitro by EMSA using nuclear extracts from cAMP-stimulated MA-10 Leydig cells. However, our ChIP data revealed a cAMP-dependent increase in MEF2 recruitment to the proximal Nr4a1 promoter. Several possibilities can explain these results. For instance, whereas MEF2 factors are constitutively bound to their elements in target promoters, they need to be activated, or unmasked, to increase gene transcription (26, 46). Under basal conditions, DNA-bound MEF2 factors are associated with corepressors such as CABIN1 (46, 47), histone deacetylase (HDAC)4, -5, -7, and -9 (48, 49), and MITR (50) silencing expression of target genes, including Nr4a1. Following an increase in intracellular Ca²⁺ levels, a Ca²⁺/calmodulin-dependent dissociation of MEF2 corepressors ensues, leading to activation of MEF2 target genes in various tissues. Activated calmodulin competes with MEF2 for interaction with corepressors leading to the recruitment of the p300 coactivator (40, 46). In addition, phosphorylation of HDAC4 and -5 by CAMKs leads to their nuclear export, chromatin unpacking, and promoter accessibility (51, 52). In Leydig cells, we and others have reported that treatment with human chorionic gonadotropin/Fsk/cAMP leads to a significant increase in intracellular Ca²⁺ levels, which is essential for optimal steroidogenesis (33, 53). Furthermore, we have identified several HDACs, including HDAC4 and HDAC5 in Leydig cells, which supports the concept that upon stimulation MEF2 factors are unmasked and activated (data not shown).

Alternatively, the increased recruitment observed by ChIP may be attributed to indirect recruitment of additional MEF2 protein to the Nr4a1 promoter through protein-protein interactions with other DNA-bound transcription factors following a hormonal stimulation. This would have no impact on MEF2 direct DNA binding activity observed by EMSA but would nonetheless translate into an increase in MEF2 recruitment on a given promoter as detected by ChIP. Such DNA binding-independent recruitment has been reported in monocytes in which MEF2D and SP1 physically interact and synergistically activate the CD14 promoter, which lacks MEF2 binding elements (54). In addition, transcriptional enhancer factor 1 and MEF2 interact to regulate various cardiomyocyte-specific promoters (55). Interestingly, the Nr4a1 promoter contains a highly conserved transcriptional enhancer factor 1-binding element near the MEF2 elements. Furthermore, MEF2 is also known to interact with nuclear factor of activated T-cells factors to up-regulate Nr4a1 expression following an increase in Ca²⁺ levels in T cells (56). Nuclear factor of activated T-cells 2 is expressed in Leydig cells (57), and Ca²⁺ levels are increased after Fsk/cAMP stimulation of Leydig cells (33). Another potential partner is the transcription factor GATA4. This factor is strongly expressed in Leydig cells (58) and is known to cooperate with MEF2 in cardiomyocytes (59). Interestingly, the Nr4a1 promoter contains a putative GATA element. It is likely therefore that similar interactions between MEF2 and these various transcription factors might also contribute to Nr4a1 transcriptional regulation in Leydig cells.

In conclusion, our analysis of the Fsk/cAMP and Ca²⁺ responsiveness of the Nr4a1 promoter in Leydig cells led to the identification of the MEF2 family of transcription factors in somatic cells of the male gonad. In addition, our findings show that the mechanisms regulating Nr4a1 expression in Leydig cells are diverse and consist of both transcriptional and posttranslational events that include MEF2, cAMP, and the CAMKI-signaling pathway. Finally, our ChIP data, in which a cAMP-dependent increase in MEF2 recruitment to the proximal Nr4a1 promoter was observed, support the concept that, upon stimulation, MEF2 factors are unmasked and activated. It is therefore likely that several mechanisms, which are not mutually exclusive, might be involved in MEF2 activation following stimulation of Leydig cells.

Funding Statement

This work was supported by grants from the Canadian Institutes of Health Research (funding reference number MOP-81387) and the Natural Sciences and Engineering Research Council of Canada (funding reference number 262224) (to J.J.T.). L.J.M. held a doctoral studentship from the Fonds de Recherche du Québec-Santé (FRQS). J.J.T. holds a Chercheur-boursier sénior Scholarship from FRQS.