| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Université de Toulouse, Centre de Biologie du Développement, UMR 5547, CNRS, IFR109, Toulouse, France; and
Laboratoire de Biologie Cellulaire et Moléculaire du Contrôle de la Prolifération, UMR 5088, Toulouse, France.
1 Correspondence: Centre de Biologie du Développement, UMR 5547 CNRS/Université Paul Sabatier, 118 route de Narbonne, F-31062 Toulouse, France. E-mail: vandel{at}cict.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: p160 proteins • MMPs • AP-1 • transcription • mRNA stability • post-transcriptional regulation
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The vertebrate Fos family is composed of four members: c-Fos, FosB, Fra-1, and Fra-2. Although c-fos is constitutively expressed in a large number of tissues, its expression can also be activated rapidly and transiently in most cell types by a variety of stimuli. c-Fos is a proto-oncoprotein, as deregulating its expression leads to transformation in various lineages, in cultured cells as in living organisms (2)
. Interestingly, premalignant papillomas in c-fos knockout mice are incapable of progressing to malignant skin carcinomas, and this phenomenon correlates with the specific down-regulation of the expression of the matrix metalloproteinase (MMP) -1b and -3 (MMP-1b and MMP-3) genes (3)
, which are known in vivo c-Fos target genes (4)
. The 25 currently known MMPs degrade extracellular matrix components and activate growth factors, thereby contributing to physiological events (tissue remodeling, wound healing, or angiogenesis) but also to pathological conditions (cancer and arthritis). Understanding their regulation at the transcriptional and post-transcriptional level has been the focus of many studies and has led to the characterization of several of their cis-regulating elements and corresponding transactivators (for reviews, see refs. 5
, 6
). More recently, much effort has been devoted to understanding the role of chromatin modifiers in regulating gene expression, but this area has been understudied with respect to MMP gene regulation. We demonstrate in the present study that c-Fos-mediated transcription activation is regulated by protein arginine methyltranferase 4/coactivator associated arginine methyltransferase 1 (PRMT4/CARM1; hereafter referred to as CARM1). CARM1 was initially described as a factor that interacts with and further stimulates the transcription-enhancing function of the p160 family of nuclear receptor (NR) coactivators that comprise SRC-1/NCoA1, GRIP1/TIF2/SRC-2/NCoA2, and RAC3/ACTR/AIB1/SRC-3/NCoA3 (7
8
9
10
11
12
13
14)
. Recruitment of CARM1 to NR-responsive genes was found to coincide with their activation and with histone H3 specific methylation at their promoter region (15
, 16)
. That CARM1 can also act as a coactivator of transcription factors other than NRs, namely p53, NF-
B, LEF1/TCF4, and E2F-1 (17
18
19
20)
suggests that this enzyme plays pleiotropic roles in cell physiology, including differentiation, growth, and survival. Interestingly, in addition to its transcriptional role, CARM1 can also regulate gene expression by affecting the stability of mRNAs containing AU-rich elements (AREs) within their 3'-untranslated region (UTR) sequences. Indeed, CARM1 associates with and methylates HuR and HuD proteins, two members of the Hu family that bind to AREs and regulate mRNA stability (21
, 22)
. Here, we show that CARM1 regulates c-Fos-mediated transcriptional activation synergistically with each member of the p160 coactivators. Notably, our data reveal that CARM1 regulates the expression of c-Fos target genes not only at the transcriptional level but also at the post-transcriptional level. Hence, CARM1 integrates (at least) two distinct mechanistic levels to fine-tune the c-Fos-induced expression level of AP-1 target genes.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cell culture and treatments
U2OS and 3T3 cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with antibiotics and 10% fetal calf serum (FCS). Clone 20 of 3T3 cells was derived from carm1 knockout mouse embryonic fibroblasts, whereas clone 20.1 was derived from clone 20 after transfection of the flp-recombinase to restore carm1 expression and clone 13 [3T3 wild type used in chromatin immunoprecipitation (ChIP) experiments] were a kind gift from Mark Bedford (University of Texas, Smithville, TX, USA) (23)
.
U2OS or 3T3 cells were treated with either 50 ng/ml platelet-derived growth factor (PDGF) or 20% FCS or 100 ng/ml TPA as indicated.
Transactivation
U2OS cells (4x104) were plated in 12-well plates and transfected 18 h later by the calcium/phosphate coprecipitation procedure. Typically, 20 ng of c-Fos expression vector, 1 µg of SRC-1, TIF2 or RAC3 expression vectors, and 500 ng of CARM1 or CARM1 E/Q expression vector were used together with 1 µg of the pBS-Gal4 Firefly luciferase reporter plasmid (described in ref. 24
) and 20 ng of pGL4.74 Renilla luciferase plasmid (Promega) for normalization. The total amount of DNA was kept constant in all experiments. Forty-eight hours after transfection, cells were lysed and luciferase activities were measured using the Dual-Luciferase Reporter Assay System (Promega) according to the manufacturer’s instructions.
Pull downs
35S-Radiolabeled proteins were in vitro translated with the TNT coupled/reticulocyte lysate system (Promega) and glutathione S-transferase (GST) fusion proteins linked to glutathione Sepharose beads (Amersham Biosciences Corp., Piscataway, NJ, USA) were used in pull-down assays as described previously (25)
.
Coimmunoprecipitation
U2OS cells treated with 100 ng/ml TPA for 4 h were lysed in buffer L [50 mM Tris (pH 8), 300 mM NaCl, 0.4% Nonidet P-40, 10 mM MgCl2, and protease inhibitors], sonicated, and then centrifuged for 10 min at maximum speed. The supernatant was diluted twice with buffer D [50 mM Tris (pH 8), 0.4% Nonidet P-40, 2.5 mM CaCl2, and 5 µg/ml DNase). One milligram of protein extract was precleared with protein A- and G-conjugated agarose beads for 1 h at 4°C, and protein extract was then incubated overnight at 4°C with anti-c-Fos or an IgG isotype control antibody that had been cross-linked on beads. Beads were then washed in buffer W (buffer L/buffer D, 1:1) before being resuspended in loading buffer, and the supernatant was subjected to an SDS-PAGE. Coimmunoprecipitated proteins were then analyzed by Western blot.
Antibodies
Anti-CARM1 antibody was from U.S. Biological (Swampscott, MA, USA; Western blot and immunoprecipitation), anti-HDAC3 (recognizing HDAC-1, -2, and -3), and anti-RAC3 antibodies were from BD Transduction Laboratories (Franklin Lakes, NJ, USA), anti-c-Fos and anti-SRC-1 antibodies were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA), and anti-TIF2 antibody was from Bethyl Laboratories (Montgomery, TX, USA).
Small hairpin (shRNA) expressing U2OS cell lines
Oligonucleotides specific for CARM1 mRNA (shRNA CARM1) or that cannot recognize any known mRNA sequence (shRNA CTR) were synthesized by Sigma-Aldrich Corp. (St. Louis, MO, USA) and subcloned into the pSUPER-puro-EGFP vector (containing a puromycin-IRES-EGFP cassette) digested by BamHI/XhoI. pSUPER-puro-EGFP was modified by Jean-Marie Garnier and Cathie Erb (Institut de Génétique et de Biologie Moléculaire et Cellulaire, Illkirch, France) from the original pSUPER vector (a gift from René Bernards, The Netherlands Cancer Institute, Amsterdam, The Netherlands). shRNA sequences are listed in Supplemental Fig. 6.
The corresponding pSUPER-puro-EGFP constructs were transfected into U2OS cells, and individual clones were selected in the presence of 2 µg/ml puromycin. Clones showing a significant CARM1 expression knockdown by Western blot were amplified and used in transactivation experiments. Protein quantification was performed by densitometry using ImageJ software (U.S. National Institutes of Health, Bethesda, MD, USA) and calculated relative to the loading control.
Small interfering RNA (siRNA) electroporation
Cells (5x106) were electroporated with 3 µM siRNA using the Cell Line Nucleofector kit (Amaxa Biosystems, Cologne, Germany) according to the manufacturer’s instructions. siRNA-mediated knockdown efficiency was tested by quantitative polymerase chain reaction (qPCR) and by Western blot analysis. siRNA sequences used in the present study are listed in Supplemental Fig. 6.
RNA extraction and reverse transcription
Total RNAs were extracted from cells using the RNeasy Mini Kit (Qiagen, Valencia, CA, USA) and were eluted in RNase-free water. Ten micrograms of each preparation of RNA was reversed-transcribed using Superscript II Reverse Transcriptase (Invitrogen, Carlsbad, CA, USA) according to the supplier’s instructions, and cDNAs were then analyzed by qPCR.
qPCR
qPCR analysis was performed on an i-Cycler device (Bio-Rad Laboratories, Hercules, CA, USA) using SYBR green JumpStart Taq Ready mix (Sigma-Aldrich Corp.), according to the manufacturer’s instructions. All experiments included a standard curve. Samples were analyzed in triplicate, and the mean and SD were calculated. Samples were normalized to the number of P0 mRNA copies. Sequences of primer pairs used in gene expression analysis are listed in Supplemental Fig. 7.
ChIP
Chromatin from 1 x 107 3T3 cells that were treated with 100 ng/ml TPA and 20% FCS for the indicated time points was prepared to obtain DNA fragments of a size ranging from 200 to 1000 bp according to the ChIP Assay Kit protocol from Upstate Biotechnology (Lake Placid, NY, USA). ChIPs were then performed with 3 x 106 cells per immunoprecipitation with either 3 µg of anti-c-Fos antibody (Santa Cruz Biotechnology) or 3 µg of anti-CARM1 antibody (a gift from Hinrich Gronemeyer, Institut de Génétique et de Biologie Moléculaire et Cellulaire, Illkirch, France) or with beads (as a control) according to the ChIP assay kit protocol from Upstate Biotechnology. Analysis of the immunoprecipitated products was performed by qPCR by using MMP-1b, -2, -3, and -13 promoter-specific primers. All primers used in ChIPs are listed in Supplemental Fig. 8. Values were determined by subtracting values obtained with beads-only immunoprecipitations and normalizing to the total amounts of each MMP promoter DNA (input).
Analysis of transcript stability
3T3 carm1-rescued or carm1–/– cells were treated with 5 µg/ml actinomycin D dissolved in ethanol (time 0). Cells were harvested at the indicated times (from 0 to 8 h) and mRNA level was assessed by qPCR and normalized to the number of P0 mRNA copies. The mean and SD were calculated from 3 independent experiments.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
suggests that c-Fos may regulate transcription by mechanisms similar to those of NRs. Indeed, the transactivation in U2OS cells of a cognate luciferase reporter gene by c-Fos chimeras harboring the DNA binding of Gal4 was significantly enhanced by coexpressing not only SRC-1 but also either of the three p160 proteins (Fig. 1
A, B). Pull-down experiments revealed that all three p160 proteins interacted, albeit with different efficiency, directly with c-Fos (Fig. 1C
). That SRC-1 and TIF2 bound c-Fos more efficiently than RAC3 suggests a direct correlation between binding and coactivation (compare Fig. 1A, B with C
). Coimmunoprecipitation experiments performed after induction of c-fos expression by TPA confirmed the interaction in vivo of the endogenous p160 proteins with c-Fos (Fig. 1D
). Taken together, these data support a model according to which c-Fos-dependent transcription is regulated in a way similar to that of NRs, i.e., p160 proteins interact with c-Fos and would themselves eventually recruit the arginine methyltransferase CARM1.
|
CARM1 coactivates c-Fos-dependent transcription in a methyltransferase activity-dependent manner and interacts with c-Fos
If c-Fos-mediated transcription activation is similar to NR-dependent transactivation, CARM1 should coactivate c-Fos transcription. Indeed, overexpression of CARM1 in U2OS cells increased by 2- to 3-fold the transactivation induced by a chimeric Gal4-c-Fos protein of a cognate luciferase reporter gene, whereas the methyltransferase-deficient mutant of CARM1 (CARM1 E/Q) did not show any effect (Fig. 2
A). These data support very strongly the facts that 1) CARM1 is a bona fide coactivator of c-Fos, and 2) this coactivation depends on its methyltransferase activity. To assess whether CARM1 is required for c-Fos-dependent transactivation, we established U2OS cell lines carrying a vector expressing either a specific shRNA against carm1 mRNA or a control shRNA. In one representative clone, CARM1 expression was down-regulated by approximately 2-fold when compared with that of cells expressing the control shRNA (Fig. 2B
, bottom panel). Importantly, the capacity of Gal4-c-Fos to activate transcription was strongly impaired, especially at high amounts of expression vector, when endogenous CARM1 expression was restricted by RNA interference (Fig. 2B
). Coimmunoprecipitation experiments in U2OS cells confirmed that endogenous c-Fos and CARM1 do indeed interact in vivo (Fig. 2C
). However, although we expected that, similarly to NRs, CARM1 interaction with c-Fos would be mediated by p160s, pull-down assays with in vitro-translated CARM1 or CARM1 E/Q showed that both proteins could bind GST-c-Fos in vitro, indicating that p160 proteins could be dispensable for this interaction (Fig. 2D
). Altogether, these results reveal a coactivator function of CARM1 for c-Fos-induced transcription that is apparently due to direct interaction with c-Fos.
|
CARM1 and p160 proteins synergize to activate c-Fos-dependent transcription, but p160 proteins are not required for the interaction between CARM1 and c-Fos in vivo
The additional expression of CARM1 boosted the transactivation induced by Gal4-c-Fos in the presence of either of the p160 proteins (Fig. 3
A–C), indicating a cooperative coactivation of c-Fos-induced transcription between p160s and CARM1. This synergy between CARM1 and the p160s is reminiscent of that observed for NR- or E2F-1-dependent transcription (14
, 19
, 27)
. However, whereas both CARM1 and p160 proteins appear to interact directly and independently of each other with c-Fos, in the case of NRs or E2F-1 CARM1 is recruited indirectly via the p160 proteins. This fact enticed us to test whether p160 proteins might influence the binding of CARM1 to c-Fos. Accordingly, GST pull downs were performed with recombinant full-length c-Fos protein and radiolabeled in vitro-translated CARM1 in the absence or in the presence of an excess of each nonradiolabeled in vitro-translated p160. Contrary to what has been observed for the binding of CARM1 to E2F-1 in similar experiments (19)
, none of the p160s appeared to modify CARM1 binding to c-Fos (Fig. 3D
). Finally, we asked whether CARM1 binding to c-Fos depended on the p160 proteins in vivo. Thus, siRNA duplexes directed against the transcripts of all three p160 proteins or a control siRNA were nucleofected into U2OS cells, and the interaction of CARM1 with c-Fos was analyzed by coimmunoprecipitation. siRNA directed against all three p160 mRNAs strongly reduced the levels of endogenous p160 proteins (Fig. 3E
, left panel). Nevertheless, the amount of CARM1 that coimmunoprecipitated with c-Fos was similar in p160 knocked-down cells and in cells nucleofected with the control siRNA (Fig. 3E
, right panel). Hence, it seems that CARM1 associates with c-Fos in a p160-independent manner both in vitro and in vivo.
|
MMP-1b, -3, and -13 are c-Fos target genes and CARM1 regulates their expression in vivo
Several lines of evidence indicate that MMP-1b, -3, and -13 (but not MMP-2) are c-Fos target genes in human and mouse cells (for a review, see ref. 5
). Also, the analysis of MMP-1b, -2, -3, and -13 promoters in mouse and in other mammalian species showed that all of them, except MMP-2, contain consensus sites for AP-1 binding (Supplemental Fig. 1
). Hence, the expression of these genes might provide a useful readout of c-Fos transcriptional activity in vivo. Indeed, qPCR experiments showed that there was a significant reduction of MMP-1b, -3, and -13 but not of MMP2 expression when mouse 3T3 cells were nucleofected with a mix of two siRNAs targeting c-fos compared with cells treated with a control siRNA (Fig. 4
A). Hence MMP-1b, -3, and -13 are bona fide target genes of c-Fos in both mouse 3T3 cells and human cells, whereas MMP-2 is not regulated by c-Fos. To assess the impact of CARM1 on the regulation of endogenous c-Fos target genes, we then compared the expression of these MMPs between carm1-deficient (carm1–/–) 3T3 cells and the corresponding cells re-expressing carm1 (carm1-rescued) (23)
. We found a striking decrease in the expression of the three MMPs in the absence of carm1: the MMP-1b level dropped 
110-fold, the MMP-3 level dropped 450-fold, and the MMP-13 level dropped 30-fold. On the contrary, the expression of endogenous MMP-2 was not down-regulated but was slightly up-regulated (2-fold) in the absence of CARM1. Similarly, we observed a 2-fold increase in the expression of c-fos in carm1–/– cells, indicating that the drops in MMP-1b, -3, and -13 expression were not due merely to a reduction of c-fos expression (Fig. 4B
). These data indicate that CARM1 plays a major role in the regulation of MMP-1b, -3, and -13 expression.
|
CARM1 and p160 proteins synergistically regulate MMP-1b, -3, and -13 in vivo
We next determined whether p160 proteins control the expression of MMP-1b, -3, and -13 in vivo and whether they synergize with CARM1 to regulate these genes. Thus, we first compared the expression levels of MMP-1b, -2, -3, and -13 in 3T3 carm1-rescued cells nucleofected either with a mix of siRNAs directed against all three p160 proteins or with a control siRNA (Fig. 4C
). As with c-fos (Fig. 4A
), the simultaneous knockdown of the three p160s induced a significant decrease in the expression of MMP-1b, -3, and -13, whereas MMP-2 expression was not affected, suggesting that p160 proteins, as CARM1, act as coactivators of c-Fos target genes in vivo (Fig. 4C
). To address whether CARM1 and p160s cooperate in this process, we then performed a similar experiment in carm1–/– 3T3 cells (Fig. 4D
). Knockdown of the three p160 proteins led to a dramatic reduction of MMP-1b, -3, and -13 expression (but not MMP-2 expression). Interestingly and consistent with a synergistic effect, the consequence of the p160 knockdown on MMP-1b, -3, and -13 expression was much more pronounced in the carm1-deficient cell line (from 20- to >100 fold reductions) than in the rescued cell line (3- to 4-fold reductions). Thus, these results suggest that p160 proteins synergize with CARM1 to regulate the expression of c-Fos target genes in vivo.
CARM1 is required for c-Fos-induced MMP-1b and MMP-3 expression
To determine whether CARM1 is required for the expression of MMP-1b, -3, and -13 after c-Fos induction, c-fos and MMP expression levels were analyzed in the presence or absence of CARM1 after induction of c-Fos by either FCS or PDGF, two well-described potent inducers of c-Fos activity (4)
(Fig. 5
A–C). We found that c-fos expression was strongly increased (>30-fold) with FCS or PDGF treatment both in 3T3 carm1-rescued and 3T3 carm1–/– cells, indicating that CARM1 does not regulate c-fos induction in these conditions (Fig. 5A
). Furthermore, consistent with our data indicating that c-Fos regulates MMP-1b, -3, and -13 but not MMP-2 expression, treatment of carm1-rescued cells with FCS or PDGF led to a significant increase of MMP-1b, -3, and -13 expression (from 2- to 3.5-fold), whereas MMP-2 was not affected (Fig. 5B, C
). However, when the same experiments were performed in carm1–/– cells, MMP-1b and -3 levels did not increase but slightly decreased, indicating that CARM1 is required for the activation of MMP-1b and -3 in response to c-Fos induction in vivo. Unexpectedly, MMP-13 expression was more strongly induced in 3T3 carm1–/– cells (8- to 10-fold) than in 3T3 carm1-rescued cells (2.5- to 3.5-fold) after FCS or PDGF treatment. Thus, it seems that CARM1 does not act predominantly as a transcriptional coactivator for MMP-13 expression on PDGF or FCS addition and that other mechanisms are probably involved. Finally, it is worth noting that, as in carm1-rescued cells, MMP2 expression was not affected on PDGF or FCS treatment of carm1–/– cells. In conclusion, these data indicate that CARM1 is required for the induction of MMP-1b and MMP-3 by c-Fos.
|
CARM1 is recruited on MMP-1b, -3, and -13 promoters in vivo
To definitely confirm the transcriptional role of CARM1 for endogenous c-Fos target gene transcriptional activation, ChIP experiments were performed after induction of c-fos by addition of serum and TPA with c-Fos and CARM1 antibodies on MMP promoter regions that contain the AP-1 binding site (Fig. 5D-G
). After this treatment, c-fos expression was strongly increased in 3T3 cells at 1 h after induction (20-fold) and then decreased as expected, whereas CARM1 expression was not affected (Supplemental Fig. 4). In the same set of experiments, MMP expression analyses showed strong activation of MMP-1b, -3, and -13 but no induction of MMP-2 (Fig. 5D
--G, left panels) consistent with our data indicating that c-Fos regulates MMP-1b, -3, and -13 but not MMP-2 expression in vivo. ChIP experiments using either c-Fos or CARM1 antibodies showed that c-Fos and CARM1 were recruited on the MMP-1b, -3, and -13 gene promoters on their AP-1 binding site region when activated but not on the MMP-2 promoter, which does not contain any AP-1 binding site (Fig. 5D-G
, middle and right panels) or an irrelevant β-actin promoter (data not shown). The three c-Fos target genes showed a recruitment of c-Fos that was optimal between 2 and 6 h after induction, whereas CARM1 recruitment was increased at 2 h but was maximal at 6 h postinduction. Thus, these data indicate that CARM1 plays a transcriptional role on MMP-1b, -3, and -13 activation as it is indeed recruited onto the AP-1 binding site within their promoters concomitantly with c-Fos in vivo.
CARM1 regulates transcript stability
That MMP-13 expression was constitutively decreased by 30-fold in carm1–/– cells compared with the carm1-rescued cell line (Fig. 4B
) and that CARM1 is recruited on MMP-13 promoter on c-fos induction (Fig. 5G
) suggested that CARM1 plays an important role for MMP-13 activation. However, our data also showed that CARM1 does not act predominantly as a transcriptional coactivator for MMP-13 expression on PDGF or FCS addition (Fig. 5C
), suggesting that other mechanisms are probably involved. Because CARM1 has been shown to regulate transcript stability by methylating AU-rich binding proteins such as the Hu family protein HuD (21)
, we hypothesized that CARM1 might control MMP-13 mRNA stability. Thus, MMP-13 transcript stability was assessed in the presence of actinomycin D in carm1–/– vs. carm1-rescued 3T3 cells (Fig. 6
A). Interestingly, MMP-13 transcript was strongly stabilized in 3T3 carm1–/– cells compared with 3T3 carm1- rescued cells, with its half-life shifting from 3 h 15 min to >8 h in the absence of CARM1 (Fig. 6A
). These data indicate that CARM1 destabilizes MMP-13 transcript and might explain the increase in MMP13 level observed in carm1–/– cells after c-fos induction (Fig. 5C
; see Discussion). On the contrary, the half-life stability of MMP-3 was greater than 8 h in the 3T3 carm1-rescued cell line, whereas it was less than 4 h 30 min in carm1–/– cells, suggesting that CARM1 is significantly stabilizing the MMP-3 transcript (Fig. 6B
). Therefore, CARM1 plays a double-positive role on MMP-3 expression, both at the transcriptional and post-transcriptional levels, and this might explain the very low level of MMP-3 expression in carm1–/– cells (Fig. 4B
). When MMP-2 transcript stability was assessed, it was slightly increased in the absence of CARM1 (half-life of the transcript of 5 h 30 min compared with 3 h 20 min) (Fig. 6C
). These data show that although CARM1 does not regulate MMP-2 transcriptionally, it may regulate to some extent its stability (Figs. 4B
, 5B
, and 6C)
. c-fos transcript stability was also slightly increased in 3T3 carm1–/– cells compared with the carm1-rescued cells with half-lives of 35 and 29 min, respectively (Fig. 6D
). Thus, CARM1 slightly regulates MMP-2 and c-fos transcript degradation, possibly explaining the increase of their expression observed in carm1–/– cells (Fig. 4B
). However, when we tried to assess MMP-1 stability in the same set of experiments, its low expression level in carm1–/– cells dropped very quickly below the detection limit of qPCR on actinomycin D treatment, thus preventing the analysis of its transcript stability (data not shown). Finally, to determine whether the effect of CARM1 was specific to the MMP family or whether it could be observed on other c-Fos target gene types, we also tested the transcript stability of unrelated genes (S16, S18, and RPL32 ribosomal protein transcripts) and of other known c-Fos target genes that do not belong to the MMP family such as Dnmt1, Fra-1, Rab11a, VEGFD, and CD44 (28
, 29)
. We found that CARM1 either destabilizes them (Dnmt1 and VEGFD) or does not change their stability (Fra-1, Rab11a, CD44, S16, S18, and RPL32) (Supplemental Fig. 5 and data not shown). Altogether, our data show that in addition to transcription regulation, CARM1 regulates c-Fos target gene expression at the post-transcriptional level also, either positively (MMP-3) or negatively (MMP-13, Dnmt1, and VEGFD). Hence, the expression of a given gene depends on the result of this dual activity.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Overexpression of each p160 protein stimulated c-Fos-dependent transcription in a reporter system assay, whereas the simultaneous depletion of the three p160 proteins by siRNA specifically compromised the activation of direct c-Fos target genes (MMP-1b, -3, and -13). Interestingly, SRC-1 and TIF2 cooperated and interacted with c-Fos more efficiently than RAC3 did. Although the endogenous p160s are expressed at different levels in the cell line used in the present study (unpublished results), these differences are not sufficient to account for the distinct coactivating properties we observed. Rather, it is likely that these differences reflect an intrinsic selectivity of c-Fos for each of the three p160 proteins. For instance, it was shown that depending on the target genes, a given transcription factor, such as the estrogen receptor, was not recruiting the three p160 proteins equally (30)
. In line with these results, we found that individual knockdown of each p160 by specific siRNA led to compensatory effects in the expression of the nontargeted p160 members that precluded the unambiguous analysis of the effects of each p160 protein on MMP expression (unpublished results), as reported in single knockout mice (31
32
33
34)
.
We presented several lines of evidence demonstrating that CARM1 is a coactivator of c-Fos and that it synergizes with p160s to activate c-Fos target genes. However, unlike what was described for NRs or E2F-1 (19
, 27)
, we found that CARM1 interacts directly with c-Fos and its binding was independent of the presence of p160 proteins in vitro as well as in vivo. This situation is reminiscent of the one described for p53 or for the p65 subunit of NF-B, which are both able to bind CARM1 directly, although the involvement of the p160 proteins had not been tested in those cases (17
, 18)
. Hence, one can hypothesize that there are two groups of transcription factors: the first one requires p160s to recruit CARM1, whereas the second binds CARM1 directly, but in both cases p160 proteins and CARM1 activate transcription synergistically.
Our data indicated that 1) the coactivation of c-Fos by CARM1 depends on its methyltransferase activity, 2) CARM1 is recruited on c-Fos target genes concomitantly with c-Fos in vivo, and 3) this recruitment correlates with the transcriptional activation of these genes. It is well established now that when recruited to the promoter by a specific transcription factor, CARM1 methylates the histone H3 tail and this methylation triggers the activation of target genes (15
, 16
, 18
, 19)
. However, CARM1 also regulates transcription by methylating a variety of nonhistone proteins, including the acetyltransferases cAMP response element-binding protein (CREB) binding protein (CBP)/p300 (35
36
37)
and the p160 proteins (38
, 39)
. It is tempting to speculate that the coactivating activity of CARM1 reported in this study may be due in part to its capacity to methylate these two coactivators of c-Fos (present study and ref. 40
). Methylation by CARM1 of CBP/p300 within distinct regions has been shown to be involved in CREB signaling (37)
, NR-dependent transcription (35
, 36)
, and CIITA-dependent gene activation (25)
whereas methylation of RAC3/AIB1 by CARM1 has been reported to decrease NR-dependent gene activation (38
, 39)
. Thus, it would be of interest to test whether CBP or p160 proteins mutated on the various sites methylated by CARM1 might interfere with c-Fos target gene activation.
In addition to its role as a transcriptional coactivator of MMP-1b, -3, and -13, CARM1 also acts at the post-transcriptional level. Indeed, we found that MMP-2 and c-fos transcript stability was significantly increased in the absence of CARM1, and this might explain the difference of expression of these two transcripts detected between carm1-rescued and carm1–/– cycling cells. Furthermore, the MMP-13 transcript was also stabilized in carm1–/– cells, but its steady-state level was reduced by more than 30-fold compared with that of carm1-rescued cells, indicating that MMP-13 regulation by CARM1 is prominently transcriptional in cycling cells. It was shown that CARM1 methylates the RNA-binding proteins HuR and HuD that control transcript stability by binding AREs present in 3'-UTR sequences (21
, 22)
. HuD methylation by CARM1 decreased p21 mRNA stability by inhibiting the binding of HuD to p21 AREs (21)
. Interestingly, it has been shown that c-fos, MMP-1b, MMP-2, and MMP-13 transcripts contain AREs in their 3'-UTR (41
42
43)
, whereas no such motif has been described for MMP-3 to date. It is possible that CARM1 regulates MMP-2, MMP-13, and c-fos transcript stability in a manner similar to that for p21, i.e., binding of HuR to the AREs present in their 3'-UTR would stabilize them, but the methylation of HuR by CARM1 would lead to their degradation by inhibiting HuR binding. In contrast, we observed that CARM1 stabilizes MMP-3 transcripts, indicating that CARM1 can act either in stabilization or in destabilization of mRNAs, possibly by methylating other ARE-binding proteins (BPs) (for reviews, see refs. 44
, 45
). Although beyond the scope of the present study, it would be of great interest to analyze the regulation of other ARE-BP activity by CARM1.
Contrary to that of MMP-1b and MMP-3, the activation of MMP-13 expression on PDGF or FCS treatment was not blocked but enhanced in carm1–/– cells. Although the lack of activation of MMP-1b and MMP-3 is consistent with the role of CARM1 as an essential coactivator of c-Fos, it is probable that MMP-13 induction (and the slight reduction of MMP-1b and MMP-3 levels) reflects further post-transcriptional regulations. Indeed, it has been shown that PDGF or FCS induced the relocation of the ARE-BPs HuR, K homology-type splicing regulatory protein (KSRP), and tristetraprolin from the nucleus to the cytoplasm (46
, 47)
and that MMP-13 transcript was destabilized by KSRP in osteoblasts (42)
.
Finally, it has been shown that NF-B target genes IL-6 or COX-2 were up-regulated in the absence of CARM1 despite CARM1 being a transcriptional coactivator of NF-B for these two genes (18)
. In view of our results, one can suggest that COX-2 and IL-6 transcripts, two targets of ARE-BPs (41
, 43
, 48)
, are regulated similarly to MMP-13, i.e., CARM1 acts at the transcriptional level as a coactivator but also actively contributes to the degradation of the corresponding transcripts. The integration of CARM1 transcriptional and post-transcriptional activities then dictates whether these genes are apparently up-regulated (MMP-13) or down-regulated (IL-6 and COX-2) by CARM1 on certain conditions. Hence, our data highlight the crucial role of CARM1 in the post-transcriptional regulation of c-Fos target genes, a role that turns out to be nearly as important as the well-described transcriptional one.
c-fos plays a crucial role in bone homeostasis as c-fos knockout mice displayed osteopetrosis with an associated loss of osteoclasts (49)
(for a review see ref. 50
). Furthermore, ectopic c-fos expression resulted in osteosarcoma formation (51)
. Interestingly, MMP-3 and MMP-13 have also been shown to be involved in bone remodeling and chondrocyte resorption (for reviews, see refs. 6
, 52
). It would be of interest to determine whether CARM1 participates in this process during development by analyzing bone remodeling and chondrogenesis in carm1–/– mice (23)
.
In summary, the findings presented in this study show the key role of the methyltransferase CARM1 and its cooperation with the p160 coactivators in the regulation of c-Fos target genes in vivo. Strikingly, our data indicate that CARM1 regulates gene expression by a variety of mechanisms from the transcriptional to the post-transcriptional level. Therefore, the fine tuning of CARM1 activity might allow integration of the transcriptional and post-transcriptional regulations of its target genes.
|
ACKNOWLEDGMENTS |
|---|
Received for publication December 14, 2007. Accepted for publication April 24, 2008.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
B-dependent gene expression. EMBO J. 24,85-96[CrossRef][Medline]
-inducible MHC-II gene expression. Proc. Natl. Acad. Sci. U. S. A. 102,16321-16326This article has been cited by other articles:
|
|
 |
|
  D. Kim, J. Lee, D. Cheng, J. Li, C. Carter, E. Richie, and M. T. Bedford Enzymatic Activity Is Required for the in Vivo Functions of CARM1 J. Biol. Chem., January 8, 2010; 285(2): 1147 - 1152. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
-c-Fos and 