Vimentin polymers are highly dynamic with an active subunit exchange between polymer and soluble subunits. Vimentin have been shown to undergo a dramatic and
specific reorganization during the differentiation of preadipocytes into adipocytes (Franke et al, 1987). During this reorganization, vimentin filaments surround the
nascent lipid droplets, forming a regularly spaced cage-like structure around them.
Similar structures of vimentin filaments also have been reported in cholesterol-loaded
macrophages (McGookey & Anderson, 1983) and the foam cells in the atherosclerotic lesions (Amanuma et al, 1986). It also has been reported that disruption of vimentin IFs during adipose differentiation of 3T3-L1 cells inhibits lipid droplet accumulation (Lieber & Evans, 1996).
14
Materials and Methods
Cell Culture
3T3-L1 fibroblasts were maintained in Dulbecco’s modified Eagle’s medium (DMEM; high glucose) plus 10% calf serum. Two days after confluence, differentiation
was induced by the addition of DMEM containing 10% fetal bovine serum (FBS), 172 nM insulin, 1μM dexamethasone (Dex), and 0.5mM methylisobutylxanthine(Mix) for 4 days. The medium was then replaced with DMEM containing 10% FBS for full
differentiation in 2–3 days. The effect of PPARγ ligand was assessed by inducing 3T3-L1 cells with either dexamethasone/insulin, with or without 0.5 μM BRL49653 or 10 μM 15-ketoPGE2.
Preparation of Whole Cell Extracts
Cell monolayers were washed with phosphate-buffered saline (PBS) and
harvested in a lysis buffer containing 20 mM Tris, pH 8.0, 137 mM NaCl, 1 mM EGTA, 5 mM EDTA, 1% (vol/ vol) Triton X-100, 10% (vol/ vol) glycerol, 1 mM
phenylmethylsulfonyl fluoride, 10 mM NaF, 1mM sodium pyrophos phate, 1 mM sodium orthovanadate, and proteinase inhibitor mixture (Roche, Basel, Switzerland).
Samples were extracted on ice for 30 min prior to centrifugation at 12,000 rpm for 30
min at 4 °C. The resulting supernatants were analyzed for protein content by BCA analysis (Pierce) and stored at -80 °C until used.
15
Preparation of Nuclear/Cytosolic Extracts
Cell monolayers were washed with PBS, scraped, swelled in hypotonic cell lysis buffer (5 mM HEPES, pH 8.0, 85 mM KCl ,0.5% Nonidet P-40 and proteinase inhibitor
mixture (Roche, Basel, Switzerland), and incubated on ice for 5 min followed by
centrifugation at 12,000 rpm for 15 min to pellet the nuclei, and the supernatant was saved as cytosolic extract. The nuclear pellet was washed twice with PBS, resuspended
in nuclear lysis buffer (50 mM Tris-HCl, pH 8.0, 10 mM EDTA, 1% SDS) , and
incubated on ice for 1 hr. After sonication, the sample was centrifuged at 12,000 rpm for 15 min at 4 °C. The resulting nuclear extract and the previously obtained cytosolic extract were analyzed for protein concentrations by BCA analysis (Pierce) and stored at -80 °C until used.
Preparation of Four Subcellular Fractions
Cell extracts were collected with Qproteome Cell Compartment kit (Qiagen). By sequential addition of different extraction buffers to a cell pellet, proteins in the
different cellular compartments can be selectively isolated. 3T3-L1 cells in 6 cm dish were washed, trypsinized, and the cell suspension containing 5x106 cells were
transferred into a conical tube followed by centrifugation. Extraction Buffer CE1 was added to cell pellet and selectively disrupted the plasma membrane without solubilizing
16
it, resulting in the isolation of cytosolic proteins. The pellet from the first step was resuspended in Extraction Buffer CE2, which solubilized the plasma membrane as well as all organelle membranes except the nuclear membrane. After solubilization, the
sample was centrifuged. The resulting supernatant contained membrane proteins and proteins from the lumen of organelles (e.g., the ER and mitochondria), while the pellet
consisted of nuclei.
In the next step, nuclei were solubilized using Extraction Buffer CE3 in which all soluble and most membrane-bound nuclear proteins were extracted. Addition of
Benzonase allowed release of the proteins which were tightly bound to nucleic acids (e.g., histones). After another centrifugation, Extraction Buffer CE4 was used to
solubilize all residual, mainly cytoskeletal, proteins in the pellet. To further concentrate proteins, the extracts were then subjected to acetone precipitation to remove salts and lipid soluble contaminants. In brief, pre-cool the required volume of acetone to -20°C.
Then, place protein sample in acetone-compatible tube and add four times the sample volume of cold (-20°C) acetone to the tube. Vortex tube, incubate for 30 minutes on ice
and centrifuge 10 minutes at 12,000 rpm. Remove the supernatant and allow the acetone to evaporate from the uncapped tube at room temperature for 30 minutes to dry the pellet. Add Tris buffer (pH 8.0) and vortex thoroughly to dissolve protein pellet for western blot analysis or immunoprecipitation assay.
17
Immunoprecipitation
After 3T3-L1 cells were fully differentiated, whole cell lysates and
nuclear/cytosolic/membrane/cytoskeletal extracts were collected as described above.
For co-precipitation with PPARγ, 1 mg of protein was incubated with 2 μg of anti-PPARγ-agarose conjugate (Santa Cruz Biotechnology) and the samples were rotated overnight at 4 °C on a rotating platform. For immunoprecipitation with anti-vimentin and nonspecific mouse IgG, 2 μg of each antibody was first incubated with 40 μl protein G-agarose beads for 1 hr. Then, 1 mg of cellular extracts was added to the antibody-beads conjugates. After incubation for overnight at 4 °C on a rotating platform, the samples were centrifuged at 12,000 rpm for 1 minute at 4°C and
supernatants were removed. The beads were washed four times in cold PBS, boiled in 2X Laemmli Sample Buffer (125 mM Tris, pH 6.8, 4% SDS, 0.02% bromophenol blue and 20% glycerol) before being resolved by SDS-PAGE, as outlined below.
Western blot
Cell extracts were resolved on 7.5, 10 or 15% SDS polyacrylamide gels and transferred onto nitrocellulose membranes. Membranes were incubated at room
temperature for 1 hr in a Tris-buffered saline with 0.1% Tween 20 (TBS-T) containing 5
% skimmed milk. After being washed three times for 5 min each with TBS-T, the
18
membrane was incubated with primary antibodies in the TBS-T containing 3% bovine serum albumin (BSA) overnight at 4°C. Bound antibody was detected with horseradish
peroxidase-conjugated secondary antibodies and enhanced chemiluminescence using the ECL Detection Kit (Pierce).
Immunofluorescence microscopy
3T3-L1 cells were grown and differentiated on glass coverslips. Cells were fixed on coverslips in 4% (wt/vol) paraformaldehyde in PBS, followed by a 1-min
permeabilization in 0.1% (vol/ vol) Triton X-100 in PBS at room temperature. After
blocking with phosphate-buffered saline containing 10% normal goat serum, cells were
incubated with antibodies directed against vimentin (Abcam Ltd.), pPPARγ (Chemicon Ltd.), and PPARγ (Santa Cruz Biotechnology, Inc.) for 1 hour at 37°C. Preparations
were then incubated with a combination of tetramethyl rhodamine isothiocyanate (TRITC)-conjugated goat anti-mouse IgG and fluorescein isothiocyanate
(FITC)-conjugated goat anti-rabbit IgG (Sigma). Nuclei were visualized upon a 15-min incubation with 0.1 mg/ml 4,6-diamidino-2-phenylindole (DAPI). ER and mitochondria
were stained with ER-Tracker™ Blue-White DPX (Invitrogen) and Mito-Tracker® Red (Invitrogen).Cells were observed, and images were acquired with an LSM510 confocal
laser-scanning microscope using a Zeiss 63X oil immersion lens. At least two
19
independent experiments were performed, and 5 or more fields per sample were analyzed in each experiment. For studying the effect of leptomycin B on intracellular
localization of PPARγ, pPPARγ and vimentin, preadipocyte or differentiated adipocytes were pretreated with or without 20 nM leptomycin B (Sigma) for 8 h prior to the
addition of BRL49653 for 1 h.
Peptide Identification by Mass Spectrometry and Bioinformatics Analysis The gel pieces containing polypeptides of interest were first reduced and
pyridylethylated as previously described (Tsay et al, 2000). Up to 0.2 μg of trypsin (Promega) was added to the dried gel to incubate overnight. The supernatant was
removed and the gel was extracted with the adequate amount of 0.1% formic acid. After
formic acid extraction, supernatant and extracts were combined together and dried in Speed-Vac. Electrospray mass spectrometry was performed using a Finnigan Met LCQ ion trap mass spectrometer interfaced with an ABI 140D HPLC (Perkin-Elemer). A 150 x 0.5 mm PE Brownlee C18 column (Perkin-Elemer) (5 mm particle diameter, 300 pore size) with mobile phases of A (0.1% formic acid in water) and B (0.085% formic acid in
aceteonitrile) were used. The peptides were then eluted using the aceteonitrile gradient and analyzed by “triple-play” experiment as described (Tsay et al, 2000). Data
interpretation and correlation between the spectra and amino acid sequences within a
20
human EST database was done by Finnigan Corporation software package, the SEQUEST Browser.
21
Results
Identification of PPARγ interacting proteins
Our previous study has demonstrated that the prostaglandin (PG) metabolite 15-keto-PGE2 is a newly identified ligand of PPARγ. Binding of 15-keto-PGE2 to PPARγ can increase coactivator recruitment, thus activating PPARγ-mediated
transcription and enhancing adipogenesis of 3T3-L1 cells (Chou et al, 2007). To further investigate whether these two distinct ligand–receptor complexes exhibit differential interactions with any co-factors, histones, or other transcription factors, we treated 3T3-L1 cells with 15-keto-PGE2 and the canonical PPARγ ligand BRL. It has been
reported that treating 3T3-L1 preadipocytes with insulin (I) and dexamethasone (Dex, D) is unable to induce adipocyte differentiation unless methylisobutylxanthine (Mix, X) is
added together to stimulate the generation of endogenous PPARγ ligands via cAMP
signaling pathways. This DI system allows us to evaluate the effect of supplementation of potential PPARγ ligands on promoting adipogenesis in the absence of Mix.
Cell lysates were then collected from the 3T3-L1 cells treated with DI only, DI
and 15-keto PGE2 or DI and BRL to evaluate different ligand-induced effect. To identify the interacting proteins with PPARγ, we performed immunoprecipitation using
anti-PPARγ antibody, followed by SDS-PAGE and the gel was subsequently stained with Coomassie blue. As shown in Figure 1, there was a 57-kd band which showed
22
differential activity in interaction with PPARγ under BRL treatment as compared to 15-keto-PGE2 treatment.
To characterize the nature and function of the interacting protein of PPARγ, we then isolated two protein bands, as indicated as number 1and 2, for LC/MS/MS analysis.
The results revealed that these protein bands contain several cytoskeleton proteins (Table 1). Among them, the intermediate filament vimentin is the major cytoskeleton
protein. We then focus to investigate its probable underlying role in regulating the metabolism of PPARγ in adipocyte.
Expression and interaction of vimentin with PPARγ during 3T3-L1 adipocyte differentiation
The expression level of vimentin was increased during the differentiation of 3T3-L1 cells at protein level (Figure 2), showing similar expression pattern to PPARγ and a typical PPARγ target gene aP2 (Figure 2). By use of immunoprecipitations with
reciprocal antibodies, we confirmed the interaction between PPARγ and vimentin upon ligand treatment (Figure 3A, B). Notably, vimentin seemed to preferentially interact with PPARγ2 isoform (Figure 3B).
23
Subcellular compartment of interactions for vimentin and PPARγ during 3T3-L1 adipocyte differentiation
Vimentin is exclusively localized in the cytoplasm; however, it is well-studied that PPARγ is mainly localized in the nucleus (Berger et al, 2000; Gurnell et al, 2000).
Intriguingly, how do these two proteins interact with each other? Previously, it has been shown that, under certain stimulations, downregulation of the PPARγ’s genomic activity occurs via MEK/ ERK signaling cascade, which attenuates PPARγ’s transactivation function either by an inhibitory phosphorylation of PPARγ (pPPARγ) or by modulating PPARγ’s nuclear-cytoplasmic compartmentalization (Burgermeister et al, 2007).
Therefore, we fractionated total cellular extracts into cytosolic and nuclear extracts to further analyze whether the interactions occurs in specific cellular compartments. As shown in Figure 4, the expression level of PPARγ and pPPARγ in different cellular compartments is very different during adipocyte differentiation. The cytosolic level of pPPARγ is quite stable at different time point of adipocyte differentiation, but not that of
PPARγ (Figure 4A), and is slightly increased upon BRL treatment (Figure 4A, lane 2,4,6). On the contrary, the nuclear level of PPARγ could only be detected upon ligand treatment (Figure 4B, lane 2,4,6) , whereas the nuclear level of pPPARγ was hardly
detected in the early stage of differentiation (Day 2 & 4; Figure 4B, lane 2 and 4) but was highly expressed in the terminally differentiated cells (Day 6; Figure 4B, lane 6).
24
There were low level of PPARγ detected in the nuclear extracts in early stage of
differentiation (Day 2 and 4; Figure 4B, lane 1 and 3) when the cells was not treated with BRL49653.
To further study the regulation of phosphorylation and localization of PPARγ upon PPARγ agonist, we studied the effect of BRL49653 treatment for 1 hr in the adipocytes induced with regular DIM treatment. As shown in Figure 5, the expression level of both PPARγ1 and PPARγ2 isoforms were induced during differentiation (Figure 5A), and the localization of the PPARγ and pPPARγ was mostly in the nucleus, although pPPARγ protein could be readily detected in the cytosol (Figure 5A). The distribution of
vimentin is more complex as it was more preferentially purified in the nuclear extracts at day 1 after induction of differentiation, but return to basal levels after day 3 (Fig. 5A).
The presence of vimentin in the nuclear extracts are possibly due to incomplete
separation of the cytoplasm and nucleus because the presence of vimentin in the perinuclear zone evidenced by immunocytochemistry (Fig. 5E). To illustrate
compartments for the interactions of vimentin and PPARγ, we further performed
immunoprecipition assays in nuclear and cytosolic extracts. As shown in Figure 5B, the interaction between vimentin and pPPARγ was clearly detected in the cytosol (Figure 5B, upper panel). However, these interactions were also observed in the nuclear extracts (Figure 5C, D, lower panel), probably due to incomplete separation of the cytoplasm
25
and nucleus (Fig. 5E).
To confirm the interaction between vimentin and PPARγ in selective subcellular compartment, we performed immunocytochemistry studies in pre-adipocytes and differentiated 3T3-L1 adipocytes to demonstrate if vimentin and PPARγ/pPPARγ
colocalized in specific cellular compartments. In the pre-adipocytes, vimentin was localized in the cytoplasm (Figure 6A, E), whereas PPARγ (Figure 6B) and pPPARγ (Figure 6F) were localized exclusively in the nucleus. There is no colocalization of
vimentin and PPARγ/pPPARγ in the 3T3-L1 pre-adipocytes, indicating there is no interaction between these molecules (Figure 6C and 6G). Following differentiation, we
found that both vimentin and PPARγ/pPPARγ were gradually increased in the
well-differentiated 3T3-L1 adipocytes. Notably, PPARγ was almost entirely localized in the nucleus (Figure 7B, F, J) whereas pPPARγ was found in both nucleus and cytosol.
The cytosolic pPPARγ was only noted after ligand treatment in a time-dependent manner (Figure 7N, R, V). Interestingly, there was no obvious colocalization between vimentin and PPARγ (Figure 7G, K), however, pPPARγ showed clear colocalization with vimentin in the cytoplasm, esp. after PPARγ ligand treatment (Figure 7R, V), indicating the interaction of vimentin and pPPARγ occurs in cytoplasmic compartment following activation with PPARγ agonist.
26
Identification of nuclear-cytoplasmic translocation of pPPARγ upon ligand treatment It has been reported that the export of some nuclear proteins is sensitive to the treatment of leptomycin B, an inhibitor that blocks the NES-receptor protein
exportin-1/CRM-1 at the nuclear pore. Therefore, we treated cells with leptomycin B for indicated time to study whether the translocation of PPARγ/pPPARγ is dependent on this shuttle or not. As shown in figure 8, PPARγ was stably localized in the nucleus after PPARγ ligand treatment irrespective of presence of leptomycin B or vehicle (Figure 8B,
F). By contrast, the cytoplasmic distribution of pPPARγ upon ligand treatment was remarkably inhibited by leptomycin B treatment (Figure 8J, 8N). The colocalization
between vimentin and pPPARγ decreased dramatically when the nuclear export was blocked with leptomycin B (Figure 8O) compared to vehicle control (Figure 8K).
Sqaured areas are enlarged as shown in Figure 8I’, K’, M’, O’, respectively.
Interaction of vimentin and PPARγ/pPPARγ in cytoplasmic compartments
To further demonstrate the compartments for the interactions of vimentin and pPPARγ, we fractionated cellluar extracts into four different fractions, i.e. cytosol, membranes, nucleus, and cytoskeleton portions under canonical DIM treatment.
Immunoblots were performed using antibody against pPPARγ, vimetin, and proteins specific to each fraction as shown in Figure 9. Vimentin was preferentially present in
27
the cytoskeleton fraction (Figure 9A, lane 10, 11, 12), esp. at the day 0 and 7. The
protein level of vimentin was decreased at day3 (Figure 9A, lane 11), but it could then be detected in the cytosol and nuclear fractions (Figure 9A, lane 2 and 8), suggesting that the intermediate filament vimentin may undergo marked reorganization during adipocyte differentiation. It is notable that both pPPARγ and vimentin could be also detected in the membrane fractions under long exposure (Figure 9A, lane 5 and 6). To establish the compartments where vimentin interacts with pPPARγ, we then performed immunoprecipitation assays in the four subcellular fractions of cellular extracts. The interaction of these two proteins was clearly detected in the cytoskeletal fraction as well as in the membrane fraction (Figure 9B).
The membrane fraction contains membrane and lumen proteins of the organelles, including endoplasmic reticulum (ER), mitochondria, and Golgi apparatus. To further distinguish where vimentin-pPPARγ interacting complex associates with,
immunocytochemisrty studies were performed using makers of ER, mitochondria, and, Golgi, as shown in Figure10-12. PPARγ did not colocalize with vimentin (Figure 10C,
D) as shown before, and also not colocalize with markers for ER (Figure 10F), mitochondrial (Figure 11G, H), or Golgi (Figure 12 G, H). However, vimentin and pPPARγ showed clear colocalization with markers for ER (Figure 10M, N, O, P) and mitochondria (Figure 11C, D, K, L), but not Golgi (Figure 12 C, D, K, L).
28
Discussion
Our present study shows for the first time that an interaction between vimentin and phospho-PPARγ occurs during differentiation of 3T3-L1 adipocytes. Our data also suggest that phosphorylation of PPARγ appears after ligand treatment which leads to subsequent export of pPPARγ to cytoplasm. Based on our immunocytochemistry data, the vimentin-pPPARγ interacting complexes were colocalized to certain organelles, such as mitochondria and ER, in addition to the expected site of cytoskeleton where vimentin locates.
Modulation of PPARγ activity via phosphorylation
As a central transcriptional regulator of metabolism, the activity of PPARγ is highly regulated by various mechanisms (Rochette-Egly, 2003). Mitogenic hormones, growth factors, stress and pro-inflammatory signals are all known to reduce the ability
of PPARγ to respond to ligand stimulation. The mechanisms underlying this downregulation are complex and comprise a set of post-translational modifications
including phosphorylation (Diradourian et al, 2005; Rochette-Egly, 2003), ubiquitina-tion (Genini & Catapano, 2006), sumoylation (Yamashita et al, 2004), and cytoplasmic shuttling (Burgermeister et al, 2007). These regulatory modifications are in concordance
29
with the anti-proliferative and anti-inflammatory role of PPARγ.
Various MAPKs participate in a key downregulating machinery via
serine/threonine phosphorylation of the substrates. It was shown that PPARγ can be phosphorylated by ERKs, JNKs and p38, resulting in inhibition of transactivating
activity of PPARγ (Diradourian et al, 2005). Phosphorylation is directly against Ser82/Ser112(Ser82/112) within a MAPK consensus motif (PXSPP) located in the AF1
domain of PPARγ1/PPARγ2 (Adams et al, 1997; Camp & Tafuri, 1997), and decreases basal and ligand-dependent transactivation through PPARγ.
Several factors inhibit PPARγ-mediated regulations by the induced
phosphorylation. Thus the phosphorylation of PPARγ upon epidermal growth factor (EGF) and platelet-derived growth factor (PDGF) treatment further inhibits adipogenic differentiation of murine fibroblasts (NIH-3T3) and preadipocytes (3T3-L1) (Adams et al, 1997; Aouadi et al, 2006; Camp & Tafuri, 1997; Hu et al, 1996). Moreover, It was proposed that upon Ser82/112 phosphorylation evoked by IFNγ, PPARγ is subjected to the
subsequent poly-ubiquitination and proteasomal degradation in adipocytes (Floyd &
Stephens, 2002).
On the other hand, it was also demonstrated that the transcriptional activity of PPARγ and adipogenesis were increased upon Ser82/112 phosphorylation mediated by
30
cyclin-dependent kinase 9 (CDK9) (Iankova et al, 2006)and CDK7 (Compe et al, 2005).
Thus, phosphorylation of PPARγ at the same residue by different kinases may result in either activation or repression of its activity. The exact explanation of this discrepancy is still lacking.
The nuclear-cytoplasmic shuttle of PPARγ
According to a previously mentioned model (Burgermeister & Seger, 2007), upon mitogenic or PPARγ ligand stimulation, cytoplasmic MEKs and ERKs are released from scaffold proteins of the MAPK-module to rapidly translocate to the nucleus while ERKs
phosphorylate PPARγ at Ser82/112 leading to subsequent MEK1-dependent nuclear export of PPARγ which was demonstrated in an exportin-1/CRM1-dependent manner in our study (Burgermeister & Seger, 2007). This massive nuclear export then further
reduces the genomic activity of PPARγ through its removal from target gene promoters involved in cell cycle control, differentiation or apoptosis. The cytoplasmic
redistribution may facilitate the cytoplasmic effects of PPARγ including its association with caveolae and other cytoplasmic proteins. It was shown that PPARγ can also associate with caveolin-1 and this association upregulates the expression of caveolin-1
in human MCF-7 breast and HT-29 colon adenocarcinoma and leukemia cells
31
(Burgermeister et al, 2003; Chintharlapalli et al, 2004; Llaverias et al, 2004). However, no other interacting protein in the cytosol has been identified so far.
Our current study showed that upon ligand stimulation, PPARγ was subjected to
serine phosphorylation and was subsequently transported to the cytoplasm. Upon exporting out of the nucleus, pPPARγ subsequently interacted with a cytoskeleton
protein, vimentin. This is the first evidence linking the nuclear-cytoplasmic shuttling of
PPARγ to a certain interacting partner in the cytoplasm. This mechanism was confirmed
by our immunostaining data which was demonstrated that the level of cytoplasmic
pPPARγ and also the colocalization with vimentin in the cytoplasm were decreased
pPPARγ and also the colocalization with vimentin in the cytoplasm were decreased