A mechanism of AP-1 suppression through interaction of c-Fos with lamin A/C

Abstract

AP-1 (Activating Protein 1) transcription factor activity is tightly regulated at multiple levels, including dimer formation (i.e., Fos/Jun). Here we show that the intermediate filament protein lamin A/C suppresses AP-1 function through direct interaction with c-Fos, and that both proteins can interact and colocalize at the nuclear envelope (NE) in mammalian cells. Perinuclear localization of c-Fos is absent in Lmna-null cells but can be restored by lamin A overexpression. In vitro, preincubation of c-Fos with lamin A prior to the addition of c-Jun inhibits AP-1 DNA-binding activity. In vivo, overexpression of lamin A reduces the formation of c-Fos/c-Jun heterodimers, and suppresses AP-1 DNA-binding and transcriptional activity. Notably, c-Fos colocalizes with lamin A/C at the NE in starvation-synchronized quiescent cells lacking detectable AP-1 DNA binding. In contrast, serum-induced AP-1 DNA-binding activity coincides with abundant nucleoplasmic c-Fos expression without changes in lamin A/C localization. We also found that Lmna-null cells display enhanced proliferation. In contrast, lamin A overexpression causes growth arrest, and ectopic c-Fos partially overcomes lamin A/C-induced cell cycle alterations. We propose lamin A/C-mediated c-Fos sequestration at the NE as a novel mechanism of transcriptional and cell cycle control.

Figure 1.

The interaction between lamin A/C and c-Fos is mediated by leucine heptad repeats. (A) GST fusion proteins spanning different regions of rat lamin A/C (LMN A/C) were tested for their interaction with hexahistidine-tagged c-Fos basic-leucine-zipper (His₆-bZIPc-Fos, rat amino acids 116–211). Protein complexes were precipitated using glutathione-sepharose 4B and washed extensively. The presence of His₆-bZIPc-Fos and GST proteins in the precipitated material was monitored using anti-His antibody and Coomassie blue staining, respectively. (B) Summary of pull-down experiments showing the domains of rat lamin A/C. (NLS) Nuclear localization signal; (+) positive interaction; (-) negative interaction. (C) Pull-down assay using His₆-bZIPc-Fos and GST fused to the murine lamin A/C region containing a portion of coil 1 (amino acids 117–227), either wild-type (WT) or the mutant sequence L158G-E159T-L172S-E173T (Mutant) in which one “leucine heptad repeat” has been disrupted (see Supplementary Fig. S1C). (Left) Representative example showing the amount of GST-lamin A/C 117–227 (wild type [WT] and Mutant) and His₆-bZIPc-Fos in the precipitate. (Right) Quantification of the interaction as determined by densitometric analysis of four independent experiments (wild type [WT] set as 100%).

Figure 2.

(A) Localization of endogenous lamin A/C (LMN A/C) (red) and c-Fos (green) in HeLa cells examined by confocal indirect double immunofluorescence microscopy. Colocalization (yellow) was noted at the nuclear periphery. Arrows in three-dimensional reconstructions show examples of coexpression at specific perinuclear locations. (B) Fluorescence line profile analysis of c-Fos (red) and LMN A/C (green) expression across the nuclear section depicted by the discontinuous white line. Arrows in the graph indicate points of prominent colocalization at the perinuclear rim. (C,D, top) Indirect immunofluorescence analysis of c-Fos expression in asynchronously growing wild-type (Lmna^+/+) and Lmna-null (Lmna^-/-) MEFs. (Middle) Nuclear Hoechst counterstain of the same microscopic field. (Bottom) Merged image. Asterisks and arrowheads mark cells displaying low and high levels of nucleoplasmic c-Fos expression, respectively. (E) Quantification of the percentage of cells with c-Fos perinuclear localization in starvation-synchronized (0.5% FBS) and in asynchronously growing (10% FBS) MEFs (n = 300 of each Lmna^+/+ and Lmna^-/- cells were scored). (F) Representative photomicrographs of Lmna^-/- MEFs transfected with YFP-Fos (control) or YFP-Fos + CFP-LMN A and grown in 10% FBS showing the distribution of both proteins. Bar, 8 μm. The graph shows the percentage of cells that displayed c-Fos perinuclear localization.

Figure 3.

(A,B) Representative photomicrographs of COS-7 cells cotransfected with either YFP-c-Fos (YFP-Fos) + pcDNA-HA-lamin A (HA-LMN A) or YFP-Fos + pcDNA. YFP-Fos is shown in green and HA-LMN A in red (visualized by indirect immunofluorescence using anti-HA monoclonal antibody). (C) Western blot analysis of the soluble and extraction-resistant nuclear fractions of COS-7 cells transfected with HA-LMN A (+) or with empty pcDNA (-) using the indicated antibodies. (Nup50) Nucleoporin 50-kDa/nuclear pore-associated protein 60L. The short exposure of the lamin A/C blot precludes the visualization of the endogenous proteins, which requires longer exposures (cf. Fig. 6A). (D) Immunoprecipitation (IP) in HeLa cell lysates was performed using mouse monoclonal anti-lamin A/C or control (anti-HA) antibody. Both the input (5% of starting material) and the whole immunoprecipitated material were subjected to Western blot (WB) analysis using anti-c-Fos and anti-lamin A/C antibodies. (E) Immunoprecipitation (IP) in HeLa cells was performed with rabbit polyclonal anti-c-Fos antibodies or normal rabbit IgG (control), and the presence of lamin A/C in the precipitate was examined by Western blot (WB).

Figure 4.

FRET analysis demonstrates the interaction between A-type lamins and c-Fos and inhibition of c-Fos/c-Jun heterodimerization by ectopic lamin A. FRET was determined using the acceptor photobleaching method in COS-7 cells cotransfected with various combinations of expression vectors (see details in Materials and Methods). The graphs show the quantification of protein–protein interactions calculated as the percentage of pixels in the donor channel that increased their intensity value after photobleaching in 20–30 cells from two to five independent experiments. Images show representative examples. Bar, 8 μm. (A) Cells were cotransfected with CFP-LMN A and either YFP (negative control) or YFP-Fos. The low-magnification images show CFP-LMN A and YFP-Fos in the same cell before and after photobleaching. CFP-LMN A fluorescence in the perinuclear rim regions in the three squares is shown at higher magnification before and after photobleaching. (B) To examine the effect of lamin A on c-Fos/c-Jun heterodimerization, cells were cotransfected with YFP-Fos + CFP-LMN A + CFP-Jun (red and dark-blue bars). The photomicrographs show representative examples of cells negative only for CFP-LMN A (left), positive for all three fluorescent proteins (middle), and negative only for CFP-Jun (right). The arrowhead and arrow indicate the localization of CFP-LMN A (perinuclear rim) and CFP-Jun (nucleoplasm), respectively (cf. Supplementary Fig. S5). The red bars represent the nucleoplasmic CFP-Jun/YFP-Fos interaction, and the dark blue bar represents the CFP-LMN A/YFP-Fos interaction at the NE. The purple and light-blue bars represent the amount of nonspecific FRET in cells cotransfected with YFP + CFP-Jun or YFP + CFP-LMN A, respectively; (purple) measurements in nucleoplasm; (light blue) measurements in NE. Results were analyzed using one-way ANOVA and Bonferroni's post hoc test. For simplicity, only comparisons with appropriate controls are shown.

Figure 5.

Dynamics of serum-dependent regulation of c-Fos expression and AP-1 DNA-binding activity in primary cultures of rat VSMCs. Cells were serum starved and restimulated with 10% FBS for the indicated time. (A–F) Subcellular localization of endogenous A-type lamins (A–C) and c-Fos (D–F) assessed by indirect immunofluorescence. Both lamin A/C and c-Fos disclosed perinuclear staining at all time points examined. Note strong c-Fos staining throughout the nucleus at 2 h of serum restimulation. (G) Confocal immunofluorescence microscopy confirmed the accumulation of c-Fos at the perinuclear rim in serum-starved cultures. (H,I) Expression levels of c-Fos mRNA and protein as assessed by qRT–PCR (H) and Western blot (I), respectively. Results of qRT–PCR are represented as the c-Fos/18S rRNA ratio relative to serum-deprived cells (set as 1; n = 3 independent measurements). (J) Western blot analysis of the extraction-resistant and soluble nuclear fractions. (K) The same soluble nuclear fractions were analyzed by EMSA using a consensus AP-1 radiolabeled probe. (Lane 1) A binding reaction without nuclear extract. AP-1 DNA binding was undetectable under conditions of colocalization of lamin A/C and c-Fos at the NE (cf. D,F) and lack of nucleoplasmic c-Fos expression. Maximum AP-1 DNA-binding activity correlated with highest amount of soluble c-Fos protein (cf. E,I,J). (Lane 9) Same as lane 3, but a 50-fold molar excess of unlabeled AP-1 oligonucleotide was added to the reaction mixture as specific competitor. (L) DNA synthesis in serum-deprived and serum-restimulated VSMCs was investigated by [³H]-thymidine incorporation assay (n = 3 independent assays).

Figure 6.

Lamin A/C suppresses AP-1 DNA-binding and transcriptional activity. (A) Soluble nuclear extracts were prepared from control and HA-lamin A-transfected 293 and COS-7 cells and analyzed by Western blot (WB) with the indicated antibodies and EMSA using an AP-1 consensus probe (only retarded bands are shown). Addition of cold AP-1 competitor demonstrated binding specificity (first lane in each EMSA). The graph shows the relative DNA-binding activity in COS-7 cells averaged from three independent EMSAs. (B) COS-7 cells were analyzed as in A, but a consensus Sp1 probe was used. (C) EMSA with recombinant Fos(bZIP) (His₆-c-Fos, rat amino acids 116–211), Jun(bZIP) (His₆-c-Jun, rat amino acids 257–318), and AP-1 probe. (Lane 1) Fos(bZIP) alone. (Lane 2) Jun(bZIP) alone. (Lane 3) Fos(bZIP) + Jun(bZIP). (Lanes 4,5) Fos(bZIP) was preincubated with GST or GST-lamin A (amino acids 4–361, rod domain), respectively, before the addition of Jun(bZIP) and AP-1 probe. The amount of each recombinant protein was 200 ng. (D) Luciferase activity in extracts from HeLa cells cotransfected with the AP-1-responsive coll73-luciferase or the E2F-responsive 4xE2F1bTATA-luciferase reporters, and either empty pMEX (control, white bars) or pMEX-lamin A (LMN A, black bars). Cells were serum-starved and restimulated with 10% FBS for the indicated time. Luciferase is expressed relative to serum-starved cells (=1). Bars represent the mean ± SE of six independent transfections.

Figure 7.

Ectopic lamin B1 does not impair c-Fos/c-Jun heterodimerization in vivo. (A) FRET was determined using the acceptor photobleaching method in transiently transfected COS-7 cells (see details in Materials and Methods). The graph shows the quantification of protein–protein interactions calculated as the percentage of pixels in the donor channel that increase their intensity value after photobleaching in 20–30 cells from two to four independent experiments. Cells were cotransfected with CFP-Jun + YFP-Fos (open bars) or CFP-Jun + YFP-Fos and either HA-LMN A or HA-LMN B1 (black bars). For negative controls, cells were cotransfected with YFP + CFP-Jun (striped bar). Results were analyzed using one-way ANOVA and Bonferroni's post hoc test. For simplicity, only comparisons with appropriate controls are shown. (NS) No significant difference (p > 0.05). (B) Lysates of COS-7 cells were subjected to immunoprecipitation with either anti-c-Fos or anti-lamin A/C antibodies as indicated, and the precipitated material was subjected to Western blot analysis (WB) to examine lamin A/C and lamin B1 expression.

Figure 8.

Lamin A overexpression affects cell cycle dynamics. (A) Asynchronously growing COS-7 cells transfected with HA-lamin A expression vector were labeled for 4 h with 50 μM BrdU and then analyzed by indirect immunofluorescence microscopy to detect HA-lamin A expression (green) and BrdU incorporation (red). The graph shows the percentage of BrdU-positive cells in the indicated number of HA-lamin A-negative and HA-lamin A-positive cells from four independent transfections. (B–F) Flow cytometric analysis of asynchronously growing COS-7 cells cotransfected with HA-lamin A and either pYFP (control) (C–E) or pYFP-Fos (F). HA-lamin A was detected using monoclonal anti-HA antibody. The gray overlay in B represents HA-negative cells, and the empty overlay shows two subpopulations with low and high HA-lamin A expression that were gated separately for cell cycle distribution analysis. Graphs and percentage of cells in each phase of the cell cycle are shown as representative results of two experiments.