GLUT4 promoter activity is regulated by hormonal, metabolic, and tissue-specific controls. This complicates the study of GLUT4 gene transcription, as no cell culture model adequately recapitulates these extracellular regulators. While investigating cultured primary adipocytes as a model system for GLUT4 transcription, we observed that GLUT4 mRNA was specifically and rapidly downregulated upon tissue dispersal. Downregulation of GLUT4 mRNA was mediated in part by loss of regulatory control by the trans-acting factors that control GLUT4 transcriptional activity [the myocyte enhancer factor 2 (MEF2) transcription factor family and the GLUT4 enhancer factor] and their cognate DNA binding sites in transgenic mice. The differences in GLUT4 transcription when whole adipose tissue and cell culture model systems are compared can be correlated to a posttranslational phosphorylation of the transcription factor MEF2A. The difference in the MEF2A phosphorylation state in whole tissue vs. isolated cells may provide a further basis for the development of an in vitro system that could recapitulate fully regulated GLUT4 promoter activity. Development of an in vitro system to reconstitute GLUT4 transcriptional regulation will further efforts to discern the molecular mechanisms that underlie GLUT4 expression.
- glucose transporter-4 transcription
- myocyte enhancer factor 2
the primary physiological effect of insulin in peripheral tissues is to increase glucose uptake and metabolism. In adipose, cardiac, and skeletal muscle tissues, this process begins with redistribution of the facilitative glucose transporter GLUT4 from an intracellular vesicular pool to the cell surface. The level of insulin-responsive glucose uptake is dependent not only on translocation of GLUT4 but also on the total cellular content of the transporter. Previous studies have demonstrated that glucose homeostasis is highly sensitive to the level of GLUT4 expression. Transgenic mice engineered to overexpress human (h)GLUT4 show increased glycemic control in both normal and diabetic models (9, 15), whereas heterozygous GLUT4 knockout mice develop a diabetic phenotype characterized by insulin resistance (14, 33). Moreover, depletion of GLUT4 in one tissue can affect metabolism in other tissues. For example, fat-specific GLUT4 deficiency increases insulin resistance in muscle and liver, suggesting more of a regulatory role for GLUT4 (12). Taken together, these data demonstrate that GLUT4 levels play an important role in nutrient homeostasis and insulin sensitivity.
GLUT4 protein concentration is regulated in a variety of physiological states, and alterations in GLUT4 levels correlate with changes in glucose homeostasis. Exercise training increases GLUT4 expression by a transcriptional mechanism (20), and this correlates well with the increased insulin sensitivity associated with exercise (4). In fact, GLUT4 is required for changes in glucose uptake and glycogen synthesis that occur postexercise (32). In contrast to the effects of exercise training, GLUT4 mRNA and protein levels are decreased in diabetes mellitus (DM), reducing insulin sensitivity and contributing to the increased blood glucose levels (for review see Ref. 22). The reduction of GLUT4 is due, at least in part, to a decrease in the rate of gene transcription (8, 26). Overexpression of hGLUT4 in diabetic rodent models has been shown to improve insulin sensitivity, suggesting that GLUT4 is a potential therapeutic target for the treatment of insulin resistance (9, 13, 37).
In addition to metabolic and hormonal regulation, GLUT4 transcription is subject to tissue-specific controls. In DM the loss in GLUT4 mRNA is seen more rapidly in adipose than in muscle tissue, which is known to be due primarily to a decrease in transcriptional rate (19, 30). The study of GLUT4 transcription is therefore complicated by the fact that no cellular model adequately recapitulates proper GLUT4 transcriptional control in an intact animal. For example, in contrast to adipose tissue, cultured 3T3-L1 adipocytes express high levels of the glucose transporter GLUT1, and GLUT4 transcription decreases under chronic insulin treatment (7, 36). Thus, a metabolically regulated gene such as GLUT4 may depend on different regulatory mechanisms in cultured cells compared with intact tissue. To begin studying GLUT4 promoter regulation, transgenic mice expressing a CAT reporter gene under the control of various regions of the GLUT4 promoter were generated and studied under various metabolic states (24, 26). Using the cis DNA elements identified in these screens, we identified the two trans-acting elements required for proper control of GLUT4 transcription: the myocyte enhancer factor 2 (MEF2) transcription factor family (specifically MEF2A and MEF2D) and a novel positive regulatory protein we have termed the GLUT4 enhancer factor (GEF) (28, 35). In vitro, GEF and MEF2A cooperatively transactivate the GLUT4 promoter through their cognate binding sites, whereas MEF2D apparently inhibits transcriptional activity (10). This recapitulates previous data from transgenic mice in which we found that neither the GEF binding domain nor the MEF2 binding domain alone is sufficient to achieve a level of promoter activation and regulation similar to the two binding domains together (10, 35). Downregulation of GLUT4 gene expression in DM is correlated with a decrease in total MEF2 binding activity on the GLUT4 MEF2 binding domain, but the basis for this decreased binding has not been well established (18, 35). Clearly, further study is required to understand the control of the GLUT4 gene by the transcription factors MEF2A, MEF2D, and GEF. To examine the transcriptional regulation of GLUT4 in a more physiological context, we chose to examine GLUT4 expression in primary cultured adipocytes.
Primary cultured (or dispersed) adipocytes have classically been used to study insulin signaling and GLUT4 translocation (6, 11, 29, 34). As these cells have not been immortalized through tissue culture, we hypothesized that insights into GLUT4 transcriptional regulation might be easily elucidated in this more physiological system. In the present study, we find that GLUT4 transcription is specifically and rapidly downregulated in primary culture in a manner dependent on the promoter's functional regulatory elements. A low level of unregulated GLUT4 promoter activity is maintained if the cis elements are deleted in transgenic mice. We also find that MEF2A undergoes an apparent hyperphosphorylation when adipose tissue is dispersed into primary culture that correlates with downregulation of MEF2-dependent transcriptional activity of the GLUT4 promoter. These findings are consistent with a model whereby the differences in transcriptional control of GLUT4 in vivo vs. tissue culture models hinge on regulated posttranscriptional modifications of the known trans-acting elements of the GLUT4 promoter.
MATERIALS AND METHODS
Animal care and housing.
All procedures were approved by the Institutional Animal Care and Use Committee of the University of Oklahoma Health Sciences Center.
Isolation of primary mouse adipocytes.
Mouse perigonadal fat pads were pooled and diluted to 1 g/ml in prewarmed Krebs-Ringer-HEPES (KRHB) buffer (117 mM NaCl, 5 mM NaHCO3, 4 mM KH2PO4, 1 mM CaCl2, 1 mM MgSO4, 30 mM HEPES, pH 7.4, 1 mM Na pyruvate, 1% BSA), minced, and digested with type I collagenase (2 mg/ml; Worthington) at 37°C for 40 min with gentle agitation. After digestion, adipocytes were washed four times in fresh warmed KRHB buffer and then incubated for the times indicated in DMEM containing 25 mM glucose and 10% fetal bovine serum at 37°C and 5% CO2. The stromal tissue fraction was collected after the first KRHB wash by removing the cells pelleted under the floating adipocyte fraction to a new tube.
Nuclear run-on transcription assay.
Nuclear run-on transcription assays of isolated primary mouse adipocytes were performed as previously described (8).
RNA isolation, RNAse protection assay, and Northern blot analysis.
Total cellular RNA was isolated from frozen tissue by guanidinium isothiocyanate extraction followed by purification over a CsCl gradient (3) as described previously (25). RNA concentration was determined by absorbance at 260 nm and stored as an ethanol precipitate at −80°C. RNase protection and Northern blot analysis were carried out as previously described (35).
Preparation of nuclear extracts.
Nuclear extracts from primary mouse adipocytes, stromal tissue, and whole adipose tissue were prepared using the NE-PER nuclear protein extraction kit (Pierce). Whole adipose tissue was initially homogenized by 10 passes in a Tenbrook tissue homogenizer. Total protein concentration was determined by Commassie Plus Protein Assay Reagent (Pierce), and aliquots were stored at −80°C.
Western blot analysis.
Twenty micrograms of total nuclear extract protein were solubilized in an equal volume of 2× Laemmli sample buffer (120 mM Tris, pH 6.8, 4% SDS, 20% glycerol, and 140 mM β-mercaptoethanol) and boiled for 5 min at 100°C. Proteins were separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) using 10% acrylamide gels. Separated proteins were transferred to Immobilon-P polyvinylidene difluoride (PVDF) membranes (Millipore) overnight in buffer (25 mM Tris and 190 mM glycine, pH 8.5) at 0.2 A and 4°C. PVDF membranes were blocked for 1 h at room temperature with Block Buffer (Li-Cor Biosciences). Blocked membranes were probed with rabbit polyclonal α-MEF2A (Santa Cruz Biotechnology) or mouse monoclonal α-MEF2D antibodies (Transduction Laboratories). Immunoreactive peptides were visualized by the Odyssey Infrared Imaging System (Li-Cor Biosciences) following incubation of the blot with the respective Alexa fluor 680 goat anti-rabbit IgG or goat anti-mouse IgG antibodies (Molecular Probes).
Electrophoretic mobility shift assay.
EMSA analysis was performed using total nuclear protein extracts as previously described (28). DNA supershifts were performed with the corresponding rabbit polyclonal α-MEF2A or mouse monoclonal α-MEF2D antibodies.
Biotinylated oligonucleotide precipitations.
After initial EMSA analysis using Oct1 binding as a measure of general DNA binding activity, equivalent amounts of adipose and dispersed adipose nuclear extracts were incubated with biotinylated oligonucleotides correseponding to two repeats of the GLUT4 promoter MEF2 binding site (Invitrogen) in EMSA binding buffer [15 mM HEPES, pH 7.4, 40 mM KCl, 5 mM MgCl2, 1 mM EDTA, 1 mM DTT, 0.4 mg/ml poly(dI-dC), 5% glycerol] for 1 h at 4°C. Streptavidin-agarose (Fluka) was then added, and mixing occurred for 1 h at 4°C. The beads were then pelleted (500 g, 5 min, 4°C) and washed three times in ice-cold 1× PBS. Protein was eluted in 2× Laemmli buffer with boiling for 5 min and analyzed by 10% SDS-PAGE and immunoblotting. For analysis of stromal nuclear extract DNA binding activity, equal total protein levels of adipose or stromal nuclear extracts were incubated with the biotinylated oligonucleotide as above.
Calf intestinal alkaline phosphatase treatment.
Twenty micrograms of adipose tissue or dispersed adipose nuclear extracts were incubated with 2 U of calf intestinal alkaline phosphatase (CIAP; Fisher) for 1 h at 37°C. The reaction was stopped by addition of Laemmli buffer and boiling for 5 min, and proteins were analyzed by 10% SDS-PAGE and immunoblot. Heat inactivation of CIAP was performed according to manufacturer's specifications.
Whole cell extraction and immunoprecipitation.
Whole cell extracts of transfected COS-7 cells were prepared as in Lu et al. (17). In short, cells were washed in 1× PBS and resuspended in ice-chilled WCE buffer [1× PBS, 0.5% Triton X-100, 1 mM EDTA, 1 mM PMSF, 1× Complete-Mini EDTA-free protease inhibitors (Roche)]. The extracts were sonicated for 10 s on ice, and the remaining insoluble debris was pelleted by centrifugation for 10 min at 4°C. After removal of an input fraction, 40 μl of EZview Red ANTI-FLAG M2 Affinity Gel (Sigma) prewashed in WCE buffer were added to each extract, and the immunoprecipitations were incubated overnight at 4°C with end-over-end rocking. Samples were washed five times in fresh WCE buffer, and samples were eluted with boiling for 5 min in 2× Laemmli sample buffer.
Linear regression analysis was performed with Analyze-it for Microsoft Excel (Analyze-it Software). Statistical significance was determined using Student's t-test for paired data.
GLUT4 gene transcription is lower in dispersed adipocytes compared with adipose tissue.
As previously shown, GLUT4 mRNA levels declined rapidly following the dispersion of adipocytes by collagenase treatment of adipose tissue in rodents (8). The decline of GLUT4 mRNA was accompanied by a rapid increase in GLUT1 mRNA. To begin to understand the molecular basis for this rapid change in GLUT4 mRNA expression, we have further characterized the time-dependent changes in mRNA expression in the primary cultured adipocytes. The reciprocal changes in GLUT4 and GLUT1 mRNA levels were specific since actin mRNA levels were unchanged during the first 12 h of primary adipocyte culture (Fig. 1A). The loss of GLUT4 mRNA over time was significant in both the presence and absence of cycloheximide (P < 0.0001 and P = 0.0040, respectively).
To determine whether the change in gene expression required protein synthesis, we incubated dispersed adipocytes in the absence or presence of 0.01 mg/ml cycloheximide to inhibit protein synthesis. Cycloheximide treatment increased GLUT4 mRNA and slightly inhibited the increase in GLUT1 mRNA as a function of time (Fig. 1). The cycloheximide effect did not represent a global effect, since actin mRNA was not changed (Fig. 1). To determine whether the changes in GLUT4 mRNA were due to changes in transcriptional activation of the gene, we performed a nuclear run-on transcription assay from nuclei isolated from freshly dispersed adipose tissue (time 0) or dispersed adipocytes incubated in complete medium for 1 h without or with 0.01 mg/ml cycloheximide (Fig. 2A). The transcription rate of the GLUT4 gene was directly compared with total RNA transcription (genomic) and found to decrease threefold compared with total RNA transcription rate following 1 h in culture medium (Fig. 2A, lanes 1 and 2). The addition of cycloheximide to medium increased the transcription rate of the GLUT4 gene sixfold after 1 h of incubation (Fig. 2A, lanes 2 and 3). In contrast, the transcription rates of c-Fos decreased twofold following 1 h in culture medium (Fig. 2A, lanes 1 and 2), and cycloheximide treatment did not significantly increase the transcription rate of c-Fos genes (Fig. 2A, lanes 2 and 3). Actin gene transcription rate relative to total RNA transcription rate did not change under any of the experimental conditions (Fig. 2A).
Because the downregulation of GLUT4 appears, in part, to be mediated at the level of transcription, we made use of transgenic animals engineered to express a CAT reporter gene under the control of mutations of the human GLUT4 promoter (23, 26). Adipose tissues from several lines of these transgenic mice were removed, dispersed by collagenase treatment, and incubated for either 0 or 9 h in the absence or presence of 0.01 mg/ml cycloheximde. Initially, three lines of transgenic mice were studied. The first line (2.4-hG4-CAT) contains 2.4 kb of DNA found immediately upstream of the major transcription start site of the human GLUT4 gene (1). This transgenic construct has been shown to be regulated in a manner identical to the endogenous mouse GLUT4 gene with respect to both tissue-specific pattern of expression and regulated expression observed during insulin deficiency (16). As expected, isolation of adipocytes from mice carrying the 2.4-hG4-CAT construct showed a time-dependent decrease in transgene expression, and treatment of the dispersed adipocytes with cycloheximide increased transgene expression (Fig. 2B). As a loading control, actin mRNA was measured and was not downregulated in dispersed adipocytes.
Next, we examined the expression of two transgenic constructs that were truncated to position −730 or −412 relative to the transcription initiation site. These truncations remove the important regulatory elements required for tissue-specific expression and metabolic regulation (10, 26). Transgenic mRNA expression of these truncated promoter fragments was not downregulated when the adipocytes were dispersed and held in culture medium for 9 h (Fig. 2, C and D), suggesting that the downregulation of GLUT4 gene expression under cellular dispersal conditions is mediated through the important regulatory elements of the gene promoter. Cycloheximide treatment markedly increased transgenic mRNA expression. Thus, the cycloheximide effect is not mediated by the same regulatory elements that support the downregulation of GLUT4. Nuclear run-on transcription assays monitoring CAT transcription via the −730-bp hGLUT4 promoter showed similar CAT and endogenous GLUT4 mRNA levels 7 h postdispersal in both the absence and presence of cycloheximide (data not shown), indicating that the cycloheximide effect is most likely due to changes in mRNA stability, not transcription.
To confirm that there is a sequence-specific requirement for downregulation of GLUT4 gene transcription, we examined the decay curves for endogenous mouse GLUT4 mRNA and transgenic mRNA driven by various fragments of the hGLUT4 regulatory region (Fig. 3). In this experiment, adipose tissue from four transgenic lines was dispersed and incubated in medium for 0, 5, 10, or 15 h. Steady-state levels of mouse GLUT4 mRNA or transgenic mRNA were measured (Fig. 3A), and the proportional decay was plotted as a function of time (Fig. 3B). The slopes of the decay curves fell into two discrete groups. Endogenous mouse GLUT4 and transgenic mRNA expressed under control of either 2.4 or 1.1 kb of hGLUT4 promoter sequences showed negative slopes (−0.03, r2 = 0.83), whereas the decline of the transgenic mRNA under the control of truncation mutants deleting the important regulatory elements had slopes equal to zero, indicating that there was no change in steady-state transgenic mRNA as a function of time in culture following dispersion.
These data, coupled with the previous observation that a single point mutation, abolishing MEF2 binding activity, completely blocks GLUT4 promoter activity (35), led us to examine changes in MEF2 activity in dispersed adipocytes.
GLUT4 expression correlates with hyperphosphorylation of MEF2A.
To begin to understand the molecular mechanism that underlies downregulation of GLUT4 gene expression in cultured cells, we performed EMSA using nuclear extracts from adipose tissue and dispersed primary adipocytes (Fig. 4). MEF2 binding activity was higher in adipose tissue compared with dispersed adipocytes (Fig. 4A, lanes 2 and 3). Supershift assays using isoform-specific MEF2 antibodies revealed subtle differences when mouse adipose tissue was compared with primary adipocytes. Pretreatment of adipose tissue nuclear extracts with MEF2A-specific antibody left behind two discrete bands that were not evident when dispersed adipocyte extracts were pretreated with the MEF2A antibody (Fig. 4A, lanes 4 and 5). On the other hand, pretreatment with MEF2D antibody left behind a single band that was decreased in dispersed adipocytes to the same extent as the decrease in total MEF2 binding activity (Fig. 4A, lanes 6 and 7). Sequence specificity of the protein-DNA interaction was determined using competition with 20-fold molar excess unlabeled oligonucleotide (Fig. 4A, lanes 8 and 9). Downregulation of MEF2 binding activity was specific, as there was no difference in serum response factor (SRF) binding activity using the identical extracts (Fig. 4B).
To determine the isoform composition of MEF2-DNA binding complexes in adipose and dispersed adipose tissues, we performed DNA coprecipitations utilizing biotinylated oligonucleotides containing two repeats of the GLUT4 promoter MEF2 binding site (Fig. 5). Nuclear extracts from whole and dispersed adipose tissue were incubated with the biotinylated oligonucleotide and streptavidin-agarose. DNA-precipitated protein complexes were then analyzed by immunoblotting. In extracts with similar total serum response element (SRE) binding activity (as determined by EMSA), the quantities of DNA-bound MEF2A and MEF2D were not significantly different. However, MEF2A isolated from dispersed adipocytes by binding the biotinylated oligonucleotide resolved as two bands (Fig. 5A, lane 2), whereas MEF2A isolated from whole adipose tissue resolved as only one band (Fig. 5A, lane 1), indicating a possible posttranslational modification of MEF2A initiated after tissue dispersal. To determine whether this band corresponded with a phosphorylation of MEF2A, dispersed adipose nuclear extracts were incubated with CIAP (Fig. 5B). Upon CIAP treatment, this higher-molecular-weight MEF2A species in dispersed adipocytes decreased in intensity (Fig. 5B, lane 6), and there was a slight downward shift in mobility in both adipose and dispersed adipose MEF2A bands, indicating that the phosphorylation state had decreased in both tissues. We confirmed the efficacy of CIAP treatment in adipose tissue nuclear extracts by performing a mixing experiment in which phosphorylated MEF2 was added to the nuclear extract (data not shown). Heat-inactivated CIAP did not cause a change in mobility of MEF2A (data not shown). A faint higher-molecular-weight band was apparent after CIAP treatment of adipose tissue. This most likely corresponds to a nonspecific protein provided by stromal contaminants from the adipose tissue extraction, as the protein did not bind the biotinylated MEF2 oligonucleotide (Fig. 5C). Therefore, the posttranslational hyperphosphorylation of DNA-bound MEF2A was most apparent in dispersed adipocytes.
Phosphorylation increases the affinity of MEF2A for MEF2D.
As MEF2A phosphorylation correlated with a decrease in GLUT4 promoter activity in dispersed adipocytes, and previous data from our laboratory indicated that the adipose isoform of MEF2D played a negative regulatory role in human GLUT4 promoter activity as well, we sought to determine whether phosphorylation could change the affinity of MEF2A for MEF2D. We had previously observed that MEF2A, when transiently transfected in COS-7 cells, expressed as at least two species by apparent molecular weight on SDS-PAGE analysis. To determine whether the higher-molecular-weight form of MEF2A indicated a phosphorylation event, we incubated nuclear extracts at 37°C for 1 h in the presence or absence of CIAP (Fig. 6A). Because the upper-molecular-weight form of MEF2A decreased in intensity when incubated with CIAP, this species most likely is indicative of a phosphorylation of MEF2A in basal COS-7 cells. We then prepared whole cell extracts (WCE) of COS-7 cells transfected with plasmids encoding HA-tagged MEF2A or Flag-tagged MEF2D or cotransfected with both vectors. Upon immunoprecipitation with α-Flag-agarose, we observed an increase in the intensity of the higher-molecular-weight band of MEF2A in the coimmunoprecipitation lane compared with the input WCE of the cotransfected sample (Fig. 6B). Flag-MEF2A immunoprecipitations of MEF2D did not appreciably change the ratio of the two MEF2A bands (data not shown). We calculated the ratios of the higher-molecular-weight MEF2A to the lower-molecular-weight form from the Flag-MEF2D coprecipitations and determined that there was a statistically significant increase in the phosphorylated form of MEF2A in the MEF2D-precipitated fraction (Fig. 6C).
In the present report, we have presented data that demonstrate a specific, cis sequence-dependent transcriptional downregulation of the hGLUT4 promoter in adipose tissue that is dispersed and held in primary culture. This finding provides a molecular basis for the rapid decline in GLUT4 mRNA observed in primary cultured adipocytes. It also provides the framework for beginning to determine why GLUT4 gene expression in cultured cells is often regulated in a manner quite different from that seen in intact tissues. The latter observation has been a major impediment to studying GLUT4 gene expression in cultured cell models (for review see Ref. 23).
Several pieces of evidence indicate that the downregulation of GLUT4 mRNA is a specific, transcriptional event that appears to be mediated through the same regulatory elements that are responsible for regulated expression of GLUT4 in both a tissue-specific and a hormonal/metabolic manner (10, 24, 26, 28, 35). The downregulation of GLUT4 is specific for GLUT4, as other mRNA levels are either unchanged or increased following collagenase dispersion (Fig. 1). Using transgenic mice engineered to express mutations and deletions in the hGLUT4 promoter, we show that downregulation of transgenic mRNA occurs only when the promoter fragment driving expression of the construct is fully functional with regard to tissue-specific expression and regulation in experimental diabetes (Figs. 2 and 3 and Ref. 26). Our previous work (26, 28, 35) has shown that regulation of the hGLUT4 promoter appears to be mediated by a cooperative interaction between the domain 1 binding site and MEF2 binding sites within the first 895 bp immediately upstream of the human GLUT4 transcription initiation site. Truncation of these regulatory domains causes the promoter to assume a lower level of unregulated and ubiquitous expression. Similarly, in dispersed primary adipocytes, the truncated promoters did not appear to be downregulated (Figs. 2 and 3). These data are entirely consistent with the conclusion that downregulation of GLUT4 mRNA expression is dependent on the identified functional regulatory elements. The data also support the model that regulation of GLUT4 transcription in the primary cultured adipocytes at steady state is not dependent on the cooperative functional relationship between domain 1 and the MEF2 domain (Fig. 2 and 3). In primary adipocytes, we observed that truncation of the hGLUT4 promoter to delete domain 1 (−730-hG4-CAT) results in a level of transcriptional activity that was not downregulated over time in culture and that the new steady expression of transgenic mRNA was much higher under control of this truncated promoter compared with a fully functional promoter previously defined in transgenic mice (Fig. 3 and Refs. 26, 28, and 35). These data point to a molecular mechanism that underlies the differential regulation of GLUT4 mRNA expression in intact animals compared with cultured cells (23); that is, the basic regulated transcriptional machinery is different in tissues vs. cultured cells.
Given that the same cis-acting elements required for regulated expression of the GLUT4 gene promoter in vivo were required for downregulation of the promoter in isolated adipocytes (10), we reasoned that the transcriptionally active ligands binding to the cis-acting DNA sequences might be different in cultured cells compared with intact tissue. This was supported by our EMSA analysis of nuclear extracts from adipose tissue and dispersed adipocytes (Fig. 4). The primary difference that we observed between adipose and dispersed adipocytes detectable by EMSA was the specific decrease in MEF2 binding activity. To further determine proteins bound to DNA in these tissue systems, we examined MEF2 proteins bound to DNA by immunoblot. Although there was no obvious loss in total MEF2D bound to DNA in dispersed adipose tissue by the affinity purification technique, there was an appearance of a MEF2A band with increased molecular weight upon adipose dispersal (Fig. 5). This increase in molecular weight most likely represents a hyperphosphorylation of MEF2A, as treatment with CIAP caused a collapse of the MEF2A bands to a smaller apparent molecular weight. The specific sites of hyperphosphorylation that result in the mobility shift of MEF2A in SDS-PAGE have not yet been identified. This work is underway, and once we have a candidate phosphorylation site, we can carry out functional studies to determine the consequence of this phosphorylation.
To begin to examine the significance of the hyperphosphorylation of MEF2A, we made use of the Cos-7 cell in vitro transcription system that we had previously used to recapitulate GLUT4 promoter activity using a luciferase reporter (10). We observed that MEF2A-MEF2D heterodimers had lower transactivational activity in an in vitro transcription assay using the hGLUT4 promoter (10). When overexpressed in Cos-7 cells, MEF2A is hyperphosphorylated and demonstrates a mobility shift in SDS-PAGE that is similar to MEF2A from dispersed adipocytes (Fig. 6A). This suggests that Cos-7 cells are a useful model for understanding the role of MEF2A hyperphosphorylation in dispersed adipocytes. When MEF2A and -2D were overexpressed in Cos-7 cells, MEF2D coimmunoprecipitated a greater proportion of the slower migrating band of MEF2A (Fig. 6, B and C). We therefore conclude that the slower mobility of MEF2A results from posttranslational phosphorylation and that the hyperphosphorylated form of MEF2A is more likely to form a heterodimer with MEF2D. On the basis of this experiment, we speculated that the formation of a heterodimer with phospho-MEF2A and -MEF2D in vivo might have slightly altered DNA binding characteristics, which lead to the decreased MEF2 binding activity in dispersed adipocytes observed using the EMSA technique (Fig. 4A). It has been shown in multiple cell lines that MEF2A binding activity of exogenously expressed MEF2A is enhanced when the nuclear extracts are CIAP treated, suggesting that the dephosphorylated form of the protein has increased affinity for DNA (27). In our system, we propose that the decreased MEF2 binding activity leads to decreased MEF2-dependent transcriptional activity in dispersed adipocytes compared with adipose tissue. This mechanism of regulation is consistent with the proposed mechanism of regulation of GLUT4 gene expression in diabetes that is accompanied by decreased MEF2 binding activity (35). Further work is required to determine whether MEF2A is similarly phosphorylated in diabetic tissues.
In addition to understanding the role of phosphorylation of MEF2A in gene expression, it is also important to understand the mechanisms that underlie the hyperphosphorylation of MEF2A. In Cos-7 cells, exogenously expressed MEF2A is phosphorylated by p38, and expression of a constitutively active p38 causes a quantitative mobility shift of MEF2A to the slower migrating form (5). Culturing adipocytes activates inflammatory mediators, consistent with a stress response (such as TNF-α and IL-6) that is likely to lead to p38 activation (21, 31). Enhanced p38 activity in 3T3-L1 adipocytes results in a downregulation of GLUT4 expression in those cells (2). Our current findings may play a significant role in understanding the mechanism by which inflammatory processes contribute to the development of insulin resistance in vivo through downregulation of GLUT4 expression in adipose tissue.
In summary, the data presented herein identify a promoter sequence-dependent transcriptional mechanism that results in a switch in regulation of GLUT4 gene expression in intact tissue compared with isolated cells. Using dispersed adipocytes, we have identified a potential mechanism by which transcriptional activity of the MEF2 DNA-protein complexes can be downregulated through posttranslational modification. It is possible that this mechanism may play a more universal role in regulation of MEF2-dependent genes such as GLUT4 in more physiologically relevant states such as fasting or insulin deficiency. Further work is required to test this model for GLUT4 gene transcription. Understanding the molecular mechanisms that underlie the changes in transcriptional regulation will likely provide a basis for developing a model system in vitro that recapitulates GLUT4 gene expression in vivo, thereby providing a method for confirming the transcriptional mechanisms that we have learned from transgenic mice.
The work herein was supported by National Institute of Diabetes and Digestive and Kidney Diseases research grants DK-62341 and DK-68438 and a predoctoral fellowship from the American Diabetes Association awarded to D. P. Sparling.
We thank Noah Abbas for contributions to this work.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 by American Physiological Society