The CCAAT/enhancer-binding protein (C/EBP) family of transcriptional regulators is critically important for the activation of adipogenic genes during differentiation. The C/EBPβ and δ isoforms are rapidly induced upon adipocyte differentiation and are responsible for activating the adipogenic regulators C/EBPα and peroxisome proliferator activated receptor (PPAR)γ2, which together activate the majority of genes expressed in differentiating adipocytes. However, mitosis is required following the induction of adipogenesis, and the activation of C/EBPα and PPARγ2 gene expression is delayed until cell division is underway. Previous studies have used electromobility shift assays to suggest that this delay is due, at least in part, to a delay between the induction of C/EBPβ protein levels and the acquisition of DNA binding capacity by C/EBPβ. Here we used in vivo chromatin immunoprecipitation analysis of the C/EBPα, PPARγ2, resistin, adiponectin, and leptin promoters to examine the kinetics of C/EBP protein binding to adipogenic genes in differentiating cells. In contrast to prior studies, we determined that C/EBPβ and δ were bound to endogenous regulatory sequences controlling the expression of these genes within 1–4 h of adipogenic induction. These results indicated that C/EBPβ and δ bind not only to genes that are induced early in the adipogenic process but also to genes that are induced much later during differentiation, without a delay between induction of C/EBP protein levels and DNA binding by these proteins. We also showed that each of the genes examined undergoes a transition in vivo from early occupancy by C/EBPβ and δ to occupancy by C/EBPα at times that correlate with the induction of C/EBPα protein levels, demonstrating the generality of the transition during adipogenesis and indicating that the binding of specific C/EBP isoforms does not correlate with timing of expression from each gene. We have concluded that C/EBP family members bind to adipogenic genes in vivo in a manner that follows the induction of C/EBP protein synthesis.
The major transcription factor families involved as key regulators of adipocyte differentiation include the nuclear hormone receptor peroxisome proliferator activated receptor (PPAR) γ and the CCAAT/enhancer-binding proteins (C/EBPs) (Darlington et al. 1998, Fajas et al. 1998, Lane et al. 1999, Morrison & Farmer 2000, Rangwala & Lazar 2000, Rosen et al. 2000, Camp et al. 2002, MacDougald & Mandrup 2002). The C/EBP family members belong to the basic leucine zipper (bZIP) class of transcription factors, and bind to specific DNA sequences as dimers with other C/EBPs (reviewed by Lekstrom-Himes & Xanthopoulos 1998, Ramji & Foka 2002). When adipocyte differentiation is induced in preadipocyte cell lines, C/EBPβ and δ are rapidly and transiently induced (Cao et al. 1991, Yeh et al. 1995). These regulators act synergistically to modulate the expression of C/EBPα and PPARγ2 via interaction with the C/EBP regulatory elements present in the proximal promoters of these genes (Christy et al. 1991, Zhu et al. 1995, Clarke et al. 1997, Tang et al. 1999). Subsequently, C/EBPα and PPARγ2 play a prominent role regulating the expression of adipocyte genes necessary for the development of functional, mature adipocytes (Lin & Lane 1994, Tontonoz et al. 1994, Fajas et al. 1998, Wu et al. 1999, Rosen et al. 2000).
The essential role of the C/EBP proteins in adipocyte differentiation has been established. Ectopic expression of C/EBPβ or C/EBPα is able to force non-adipogenic cell lines to differentiate into adipocytes (Freytag et al. 1994, Wu et al. 1995, Yeh et al. 1995). In contrast, expression of antisense C/EBPα RNA in preadipocyte cell lines prevents the differentiation program (Lin & Lane 1992). Additionally, analysis of promoter regions of adipogenic genes as well as studies of knockout mice have demonstrated the involvement of this family of transcription factors in regulating adipogenesis and other important physiological processes (reviewed by Cornelius et al. 1994, MacDougald & Lane 1995, Tanaka et al. 1997, Darlington et al. 1998, Gregoire et al. 1998, Lane et al. 1999, Rangwala & Lazar 2000, Ramji & Foka 2002).
Most regulatory sequences controlling the expression of adipocyte-specific genes contain at least one functional C/EBP binding site, from which transactivation is mediated by members of the C/EBP family (reviewed by Hwang et al. 1997, Gregoire et al. 1998, Cowherd et al. 1999, Morrison & Farmer 2000, Rangwala & Lazar 2000). Work on the PPARγ2 and C/EBPα promoters has focused on the role of C/EBPβ and δ as primary inducers of the expression of key regulators. PPARγ2 and C/EBPα are expressed by day 2 of the differentiation process, following one or two rounds of mitotic clonal expansion. The induction of C/EBPβ and δ protein levels, however, occurs almost immediately after addition of differentiation inducers at the onset of differentiation (Cao et al. 1991, Yeh et al. 1995, Darlington et al. 1998, Tang & Lane 1999). Thus, even though both C/EBPβ and δ are expressed at high levels at the beginning of the differentiation program, the target genes C/EBPα and PPARγ2 are not expressed until nearly 2 days later (Lane et al. 1999, Rosen et al. 2000). Previous work using electrophoretic mobility shift assay (EMSA) analysis (Tang & Lane 1999) determined that the lag in C/EBPα expression is due to a delay in acquisition of C/EBPβ and δ DNA binding activity, therefore pausing the transcriptional activation of the gene. The need for such a delay fits well with the numerous observations that C/EBPα is anti-mitotic in both preadipocytes as well other cell types (Umek et al. 1991, Lin et al. 1993, Timchenko et al. 1996, Wang et al. 2001).
The transcriptional activity of the C/EBPβ protein is regulated at several levels, including transcription, translation, association with other proteins, and post-translational modification, which includes the regulation of the phosphorylated state by multiple effectors. Multiple phosphorylation sites have been characterized on C/EBPβ, some of which result in attenuation or enhancement of DNA binding and transactivation activity (Mahoney et al. 1992, Trautwein et al. 1993, 1994, Park et al. 2004a, Tang et al. 2005 and references therein). At least some of these phosphorylation events occur almost simultaneously with induction of C/EBPβ levels, and it has been suggested that phosphorylation causes a conformational change in C/EBPβ that transforms it from a repressor to an activator (Kowenz-Leutz et al. 1994).
We previously examined the temporal interactions of C/EBP family members, modified histones, and subunits of the SWI/SNF family of ATP-dependent chromatin remodeling enzymes at the PPARγ2 promoter as a function of adipocyte differentiation (Salma et al. 2004). In that study, we used chromatin immunoprecipitation (ChIP) assays to determine that C/EBPβ was bound at the PPARγ2 promoter at 24 and 48 h following the initiation of the differentiation program in both differentiating 3T3-L1 cells and in fibroblasts forced to differentiation into adipocytes by ectopic expression of C/EBPα (Salma et al. 2004). Continuation of these ChIP studies at earlier time-points revealed that the C/EBPβ and δ isoforms were found on the regulatory regions of numerous adipocyte-specific genes in differentiating 3T3-L1 cells within a few hours of the onset of differentiation. These genes included PPARγ2, C/EBPα, and several genes expressed later in the differentiation process. Thus the binding of C/EBPβ and δ to adipogenic genes in cells correlates with the kinetics of C/EBPβ and δ expression. In addition, binding of C/EBPβ and δ was replaced at each of the loci by the binding of C/EBPα. C/EBPα binding was noted on each promoter between 20 and 48 h post-differentiation, indicating that it was associating with promoters as soon as it was expressed and that binding did not strictly correlate with the time at which the locus became transcriptionally active. We have concluded that the binding of C/EBP transcription factors to regulatory sequences controlling the expression of adipogenic genes in vivo occurs rapidly and without significant delay following the induction of each isoform during adipogenic differentiation.
Materials and methods
Cell lines and differentiation methods
3T3-L1 preadipocytes were purchased from the American Type Culture Collection (Manassas, VA, USA), maintained in growth medium consisting of Dulbecco’s minimum essential medium containing 10% calf serum, and induced to differentiate as described previously (Wu et al. 1995). Cells were collected at 0, 1, 2, 4, 8, 12, 16, 20, and 24 h and then at 24-h intervals for 7 days after addition of differentiation cocktail (Salma et al. 2004) for western blot, RT-PCR, and ChIP analysis. In experiments where 3T3-L1 cells over-expressed C/EBPβ, cells were infected with pBABE retrovirus containing C/EBPβ or empty vector as described previously (Tontonoz et al. 1994, Salma et al. 2004). The generation of the retrovirus has been described previously (Pear et al. 1993, Salma et al. 2004). Samples were collected at 0, 4, 24, 48, and 168 h in the presence or absence of differentiation cocktail for western blots, RT-PCR, and ChIP analysis.
Nuclear extracts isolated from 3T3-L1 cells differentiated in the presence or absence of cocktail were prepared as described (Hasegawa et al. 1997). The binding reaction contained 6 μg nuclear extract and 5 fmol 32P-labeled double-stranded oligonucleotide probe corresponding to the region from −343 to −306 bp from the mRNA start site in the mouse PPARγ2 promoter. This region contains a functional C/EBP binding site (Zhu et al. 1995, Clarke et al. 1997). Binding reactions contained 10 mM Tris–HCl (pH 7.5), 5% glycerol, 50 mM NaCl, 0.5 mM dithiothreitol, 1 mM MgCl2, and 0.5 mg/ml Poly (dI-dC) in a volume of 10 μl. Reactions were incubated for 30 min at room temperature and separated electrophoretically on 4% non-denaturing polyacrylamide gels made with 0.5 × Tris–borate-EDTA buffer. Some reactions were preincubated for 10 min with 1 μl IgG or anti-C/EBPβ antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA; sc-7962) prior to addition of the labeled oligonucleotides. The sequence of the oligonucleotide probe was: 5′-TAAAAAGCAATCAATATTGAACAATCTCTGCTCTGGTAA-3′.
RNA isolation and analysis by northern blotting have been described previously (de la Serna et al. 2001b). Probes were derived from plasmids containing PPARγ (provided by B Spiegelman, Dana Farber Cancer Institute, Boston, MA, USA) and the ribosomal phosphoprotein 36B4 (obtained by RT-PCR) and labeled by random priming. Washed blots were exposed to a PhosphorImager (GE Healthcarem Chalfont St. Giles, UK). For RT-PCR, total RNA (3 μg) was reverse transcribed with Moloney murine leukemia virus RT (Invitrogen). cDNA was amplified by PCR with QIAGEN HotStart Taq master mix in the presence of 2 μCi [α32P]dATP. The sequences of the primers were as follows: 5′-CCG GCC GCC TTC AAC GAC-3′ and 5′-CTC CTC GCG GGG CTC TTG TTT-3′ for C/EBPα (288 bp product); 5′-GAA CTG AGT TGT GTC CTG CT-3′ and 5′-TGC ACA CTG GCA GTG ACA-3′ for resistin (340 bp product); 5′-GAT CAA TGA CAT TTC ACA CA-3′ and 5′-GGA CGC CAT CCA GGC TCT CT-3′ for leptin (281 bp product); 5′-CAG TGG ATC TGA CGA CAC CA-3′ and 5′-CGA ATG GGT ACA TTG GGA AC-3′ for adiponectin (433 bp product); and 5′-CTC CAA GCA GAT GCA GCA GA-3′ and 5′-TCA ATG GTG CCT CTG GAG AT-3′ for the ribosomal phosphoprotein 36B4 (351 bp). The PCR conditions for leptin and 36B4 were: 95 °C for 15 min, followed by 24 cycles of: 95 °C for 30 s; 62 °C for 40 s; 72 °C for 30 s, and a final round of extension for 5 min. The PCR conditions for adiponectin were the same, except that the number of cycles was 20. For resistin, the conditions were the same as for leptin except that the annealing temperature was 58 °C. For C/EBPα, the PCR conditions were: 95 °C for 15 min, followed by 25 cycles of: 95 °C for 50 s; 66 °C for 55 s; 72 °C for 50 s, and a final round of extension for 5 min.
Protein extracts and Western analysis
Isolation of protein and western blotting have been described (de la Serna et al. 2001a). Antibodies utilized included the following from Santa Cruz Biotechnology: C/EBPα (sc-61), C/EBPβ (sc-7962), C/EBPδ (sc-151), and cyclin A (sc-596). Phosphatidylinositol 3-kinase (PI-3K) antibody (06–496) was obtained from Upstate (Charlottesville, VA, USA).
The ChIP procedure was adopted from the Upstate protocol and was performed as described by Salma et al.(2004). One-tenth of the immunoprecipitated DNA and 1% of the input DNA were analyzed by PCR. Antibodies used included Santa Cruz antibodies: C/EBPα (sc-61), C/EBPβ (sc-7962), and C/EBPδ (sc-151). PCRs were performed with QIAGEN Hot Start Taq master mix in the presence of 2 μCi [α-32P]dATP under the following conditions: a preheating at 94 °C for 15 min, followed by 24–30 cycles of 94 °C for 30 s, 62 °C for 40 s (except for PPARγ2 which was 49.5 °C), 72 °C for 30 s, followed by a 72 °C extension for 5 min. PCR products were resolved in 6–8% polyacrylamide–1 × Tris–borate–EDTA gels, dried, and exposed to a PhosphorImager. Primers used were the following: β-actin, 5′(+31) GCTTCTTTGCA GCTCCTTCGTTG-3′ and 5′ (+135) TTTGCACAT GCCGGAGCCGTTGT-3′ (Rayman et al. 2002); PPARγ2 promoter, 5′ (−413) TACGTTTATCGGT GTTTCAT-3′ and 5′ (−247) TCTCGCCAGTGA CCC-3′; upstream region of the PPARγ2 promoter, 5′ (−1871) GGGCGTTAAAACACAATCCT-3′ and 5′ (−1707) TCTCTTCCTCCTTCCCTTCC-3′; C/EBPα promoter, 5′ (−315) TGACTTAGAGGCTTAAA GGA-3′ and 5′ (−32) CGGGGACCGCTTTTATAG AG-3′; resistin promoter, 5′ (−177) CACCATGGTC CCTGGTGTTA-3′ and 5′ (+26) CTCAGTTCTGGG TATTAGCTC-3′; adiponectin promoter, 5′ (−272) ATTGTCCTTACCCTTGCCCC-3′ and 5′ (−15); and leptin promoter, 5′ (−323) GCCTTCTGTAGCC TCTTGCT-3′ and 5′ (−22) GCTCCATGCCCTG CCTGC-3′. Representative experiments from at least three independent experiments are shown.
In vitro binding of C/EBPβ at a C/EBP site in the PPARγ2 promoter occurs as early as 3 h post-differentiation
While investigating the role of C/EBP isoforms in the activation of adipogenic genes, we examined in vitro binding of C/EBPβ at a C/EBP regulatory element present in the PPARγ2 promoter by EMSA during a short time-course of 3T3-L1 preadipocytes induced to differentiate into adipocytes. We found that C/EBPβ present in nuclear extracts prepared from cells differentiated for 3, 18, and 24 h was able to bind to a C/EBP site in the PPARγ2 promoter (Fig. 1, lanes 3–5). Confirmation that the shifted band in the EMSA was C/EBPβ was demonstrated by the appearance of a supershifted band upon addition of C/EBPβ antibody to the reaction while addition of purified IgG had no effect (Fig. 1, lanes 6 and 7). The results revealed that the C/EBPβ regulator has the capacity to bind to DNA as early as 3 h following the induction of differentiation. In addition, we observed that the apparent level of C/EBPβ binding under the reaction conditions used did not increase between 3 and 24 h post-differentiation (Fig. 1, lanes 3–5).
These results diverged somewhat from data published previously where the in vitro binding activity of C/EBPβ was delayed until 12–16 h following the stimulation of adipogenic differentiation (Tang & Lane 1999). We note that the previous report demonstrated that C/EBPβ binding could be observed at 4 h post-differentiation but that binding was significantly induced at the 12–16 h time-points. This dissimilarity in results is possibly due to differences in reaction conditions, or perhaps to the use of different oligonucleotide probes. In the previous study (Tang & Lane 1999), EMSA was performed using a C/EBP site from the C/EBPα promoter, while our experiment utilized a probe that contained a C/EBP binding site from the PPARγ2 promoter. Nevertheless, the results presented in Fig. 1 raised a question about the timing of C/EBPβ binding to adipogenic regulatory sequences. To better address this issue, we decided to perform ChIP experiments at different times following the induction of adipocyte differentiation in 3T3-L1 cells in order to examine the binding of C/EBP isoforms in vivo.
Kinetics of C/EBP expression in differentiating 3T3-L1
Since the EMSA data presented in Fig. 1 indicated a potential difference from previously published studies, we first performed a series of control experiments to analyze the expression levels of C/EBPβ, δ, and α during adipocyte differentiation of 3T3-L1 cells by western blot in order to eliminate the possibility that our results might be due to differences in the experimental handling of the differentiating 3T3-L1 cells. It has been well established that initiating differentiation of 3T3-L1 preadipocytes activates a cascade of gene expression events. Among the initial events are the rapid induction of C/EBPβ and δ, which are stimulated by components of the differentiation cocktail, followed by the induction of C/EBPα on the second day of differentiation (Yeh et al. 1995, Wu et al. 1996). Western blot analyses corroborated that induced protein levels of C/EBPβ and δ were detectable at 1 h, reached a maximum at 4 h, and were declining at 48 and 120 h respectively (Fig. 2). Expression of C/EBPα occurred later; significant induction began at about 48 h post-differentiation and expression was maintained throughout the time-course (Fig. 2).
To corroborate that the cells began the mitotic expansion phase, we performed a western blot with antibodies directed to cyclin A (Fig. 2). As was expected, cyclin A levels were increased between 16–24 h post-differentiation, indicating that the cells had entered the cell cycle prior to 16 h and had exited by approximately 48 h. All of the data presented in Fig. 2 indicate that the differentiation of the 3T3-L1 cells occurred as expected and confirm previously published results (Morrison & Farmer 1999, Tang et al. 2003a,b). Thus we have demonstrated the integrity of the 3T3-L1 cells and the differentiation protocol used for this and subsequent experiments.
In vivo recruitment of C/EBPβ, δ, and α to the PPARγ2 and C/EBPα promoters during adipogenesis
The results presented in Fig. 1 show that C/EBPβ was able to bind to C/EBP binding sites on the PPARγ2 promoter at times that correlated with the induction of C/EBPβ levels. To determine whether binding occurs on endogenous adipocyte promoters at early times after the induction of adipocyte differentiation, we performed ChIP experiments and temporally analyzed binding, not only of C/EBPβ, but also of C/EBPδ and C/EBPα, to specific adipocyte promoters. First, we chose to evaluate binding at the C/EBPα and the PPARγ2 promoters, since these two essential adipogenic regulators are expressed early during differentiation, on day 2 (Fig. 3A and B).
Regulation of the C/EBPα and the PPARγ2 promoters by C/EBP family members has been characterized previously (Christy et al. 1991, Zhu et al. 1995, Clarke et al. 1997, Tang & Lane 1999, Tang et al. 1999, 2004, Elberg et al. 2000, Yang & Chow 2003). The C/EBPα proximal promoter contains a C/EBP regulatory element at −187 relative to the transcriptional start site that mediates transactivation by C/EBPs. The PPARγ2 promoter contains two previously characterized C/EBP recognition elements at −340 bp and −327 bp relative to the transcriptional start site in addition to other potential sites that diverge from the C/EBP consensus. As shown in Fig. 3A, recruitment of C/EBPβ as well as C/EBPδ at the C/EBP regulatory element on the PPARγ2 promoter was induced in a manner consistent with the protein expression patterns of both proteins (Fig. 2), with binding of both proteins apparent 2 h after the onset of differentiation. Subsequently, C/EBPβ and C/EBPδ were replaced by C/EBPα, which initiated binding to the promoter at 48 h and was maintained throughout the time-course. These results confirmed the transition of binding from C/EBPβ/δ to C/EBPα at this promoter in vivo (Salma et al. 2004, Tang et al. 2004).
Analysis of the C/EBPα promoter revealed the same pattern found at the PPARγ2 promoter, except that binding occurred even earlier following differentiation (Fig. 3B). C/EBPβ and C/EBPδ were bound as early as 1 h post-differentiation and remained present until 48 h and 120 h respectively. Induction of C/EBPα binding to the C/EBPα promoter began at 48 h as reported previously (Tang et al. 2004). Therefore, the transition in binding of the C/EBP isoforms that was observed at the PPARγ2 promoter also occurred at the C/EBPα promoter.
The appearance and disappearance of binding of the different C/EBP isoforms at the PPARγ2 and C/EBPα promoters at different times post-differentiation indicate specificity of binding. No antibody controls provide further evidence for specific binding (Fig. 3A and B). As an additional control, we analyzed C/EBP factor interactions at the β-actin locus (Fig. 3C) and at sequences 1.8 kb upstream of the PPARγ2 start site (data not shown). Neither β-actin sequences nor sequences upstream of the PPARγ2 promoter were immunoprecipitated by any of the antibodies used in the ChIP procedure.
In vivo recruitment of C/EBPβ, δ, and α to additional adipogenic gene promoters during adipogenesis
We next examined if C/EBPβ, δ, and α are recruited to the resistin, adiponectin, and leptin promoters. These adipocyte-secreted peptides, collectively referred to as adipocytokines, have generated considerable interest since they are important regulators of body mass and their misregulation may play a role in obesity (Miner 2004). The proximal promoters of these genes contain C/EBP binding sites that are necessary for expression. The resistin promoter contains a C/EBP site at −56 relative to the transcriptional start site, adiponectin contains two identified C/EBP sites at −775 and −264, and two potential binding sites at −117 and −73, and leptin has three consensus C/EBP binding sites at nucleotides −55, −211, and −292. The activity of these C/EBP sites has been confirmed by reporter assays for each of these genes (de la Brousse et al. 1996, Hwang et al. 1996, Hartman et al. 2002, Park et al. 2004b).
Binding of C/EBPβ and δ to the C/EBP site at the resistin proximal promoter was evident at 2–4 h post-differentiation and did not decline until after 48 h (Fig. 4A). A modest increase in binding of C/EBPα was observed from 20 to 48 h, and was further increased at 120 h (Fig. 4A). These results indicate that C/EBPβ and δ bind early after differentiation to the resistin promoter and suggest that all three C/EBP isoforms play a role regulating this adipocyte gene. It was previously demonstrated that C/EBPα binds specifically to the C/EBP element on the resistin promoter that is essential for expression (Hartman et al. 2002); however, a transition in binding of these regulators has not been previously demonstrated.
The adiponectin promoter contains two C/EBP sites at −775 and −264 and two potential binding sites at −117 and −73; however, the C/EBP element at −264 and the potential C/EBP sites at −117 and −73 confer promoter activity as defined in transient promoter studies, EMSA, DNase I footprinting, and ChIP assays (Park et al. 2004b, Seo et al. 2004). Consequently, we performed ChIPs in the region of the proximal promoter that contains these C/EBP sites. The analysis of binding of C/EBPβ, C/EBPδ, and C/EBPα to the adiponectin promoter showed a pattern nearly identical to that observed for the resistin promoter, despite the fact that adiponectin expression initiated later than resistin expression (Fig. 4B). C/EBPβ and δ were bound to the promoter at 2 h and maintained until 48 h. Definitive binding of C/EBPα was present at 48 h, although a modest increase was noticed at 20–24 h.
Finally, we assessed the recruitment of the C/EBP members on the leptin promoter. The proximal promoter contains three consensus C/EBP binding sites. DNase I footprint analysis, reporter gene assays, and EMSA studies have demonstrated that one of these C/EBP sites, located at −53 relative to the transcriptional start site, is functional (de la Brousse et al. 1996, Hwang et al. 1996, Mason et al. 1998). However, because the C/EBP sites are near each other, we designed PCR primers to amplify a region of the proximal promoter containing all three sites. The recruitment of C/EBPβ occurred at 4 h and remained relatively constant until 48 h (Fig. 4C). In contrast, C/EBPδ was detectable from 12 h to 48 h. Binding of C/EBPα was observed on the promoter at 20 h, but was not robust until 120 h post-differentiation, which coincides with the start of leptin mRNA accumulation. These results indicate that C/EBPβ and δ bind quickly after the induction of differentiation to adipogenic promoters that are not expressed until much later in the differentiation process. Furthermore, the transition from C/EBPβ and δ to C/EBPα binding on these promoters also occurred prior to gene expression, suggesting that C/EBP factor interations with adipogenic promoters is independent of the time at which gene expression is initiated.
Overexpression of C/EBPβ is not sufficient to promote C/EBP protein binding to the PPARγ2 promoter in differentiating 3T3-L1 cells
The detection of C/EBPβ on adipogenic promoters in vivo within a few hours of the onset of adipogenic stimulation caused us to evaluate whether over-expression of C/EBPβ in 3T3-L1 preadipocytes would be sufficient to induce binding of C/EBPβ to the regulatory sequences examined. 3T3-L1 preadipocytes were infected with a retroviral vector expressing C/EBPβ and were allowed to reach confluence, but no differentiation cocktail was added. Instead, cells were maintained in 10% calf serum as a confluent plate, and C/EBP binding was assessed at 0, 4, 24, 48, and 168 h. Despite the ectopic expression of C/EBPβ (Fig. 5A), no binding of C/EBPβ, δ, or α was observed at the PPARγ2 promoter, whereas control plates treated with differentiation cocktail showed C/EBPβ and δ binding to the PPARγ2 promoter within 4 h post-differentiation and C/EBPα binding by 48 h post-differentiation (Fig. 5B).
The data suggested that overexpression of C/EBPβ in cells is not sufficient to promote C/EBPβ binding to adipogenic promoters in the absence of differentiation cocktail. A potential caveat to this conclusion is that ectopic expression did not provide a high enough level of C/EBPβ protein to surpass a threshold level of C/EBPβ protein necessary to achieve binding. We note that the inoculum of pBABE-C/EBPβ retrovirus used in these experiments is the same as we have previously used to trans-differentiate fibroblast lines into adipocyte-like cells; thus the levels of C/EBPβ provided are sufficient to reprogram cells of a different lineage (Salma et al. 2004). However, to more directly address this concern, we compared the levels of C/EBPβ protein present in uninfected cells and in cells infected with pBABE-C/EBPβ or with the pBABE empty retrovirus that were differentiated in the presence or absence of differentiation cocktail for 4 h. A western blot (Fig. 5C) demonstrates that cells infected with the C/EBPβ virus contained greater levels of C/EBPβ protein in both the presence and absence of differentiation cocktail (compare lane 2 with lane 3 and lane 5 with lane 6), as expected. We also observed that C/EBPβ levels in vector-infected cells that were differentiated in the presence of cocktail were lower than C/EBPβ levels in the pBABE-C/EBPβ-infected cells that were differentiated in the absence of cocktail (compare lanes 4 and 3). The levels of C/EBPβ protein in pBABE-infected cells treated with differentiation cocktail were sufficient to permit C/EBPβ binding to the PPARγ2 promoter, whereas higher levels of C/EBPβ in the C/EBPβ-infected cells differentiated in the absence of cocktail were not (Fig. 5D, compare lanes 4 and 3). The results exclude the possibility that insufficient levels of C/EBPβ were present in the cells not treated with differentiation cocktail. The C/EBPβ protein undergoes a number of post-translational modifications that are associated with the induction of adipogenesis; it is likely that such modifications are induced by addition of the differentiation cocktail and are necessary to promote rapid binding to adipogenic gene regulatory sequences in vivo. Thus, simple overexpression of C/EBPβ is not sufficient to induce C/EBPβ binding in 3T3-L1 preadipocytes.
The C/EBP family of transcription factors is widely expressed and is a key regulator of a variety of target genes important in physiological events, including energy metabolism, inflammation, hematopoiesis, cellular proliferation, and differentiation (Darlington et al. 1998, Lekstrom-Himes & Xanthopoulos 1998, Rosen et al. 2000, Ramji & Foka 2002, Kovacs et al. 2003). Of note is the essential role C/EBP family members play during adipogenesis (reviewed by Darlington et al. 1998, Lane et al. 1999). Almost immediately upon induction of adipogenesis, the C/EBP family members β and δ are induced in a manner dependent on several signal transduction cascades that result in phosphorylation of these proteins (Park et al. 2004a, Bezy et al. 2005, Tang et al. 2005 and references therein). However, the activation of the early adipogenic regulators, PPARγ2 and C/EBPα, which are dependent on C/EBPβ and δ, does not occur until day 2 of the differentiation process. During the first 2 days, cells undergo one or two rounds of mitosis, a process termed mitotic clonal expansion (Bernlohr et al. 1985, Cornelius et al. 1994, MacDougald & Lane 1995). Thus, the transcriptional activation potential of the C/EBPβ and δ proteins are masked or repressed until clonal expansion commences.
Over the past several years data have accumulated that suggest that the binding capacity of C/EBPβ for its cognate binding site is delayed 12–20 h post-induction of adipocyte differentiation (Lane et al. 1999, Tang & Lane 1999, Tang et al. 2003b, 2005). This observation fits well with the need to delay expression of C/EBPα, which has anti-mitotic properties (Umek et al. 1991, Lin et al. 1993, Timchenko et al. 1996, Wang et al. 2001), until clonal expansion occurs. Moreover, the kinetics of DNA binding activation fit well with the kinetics of other events that occur at this time, including the appearance of phosphorylated Rb and the localization of C/EBPβ to pericentric heterochromatin, which contains numerous C/EBP binding sites in the satellite DNA sequences (Tang & Lane 1999). How this change in sub-nuclear distribution relates to gene expression has not been well established, but the observation raises the possibility that localization of proteins to specific nuclear compartments contributes to the complexity of adipocyte gene regulation.
In the course of examining the activation of the PPARγ2 promoter during adipogenesis, we noted that in vitro binding of C/EBPβ to a C/EBP binding site in the PPARγ2 promoter did not appear to change significantly between 3 and 24 h post-differentiation (Fig. 1). Given the wide range of variable conditions that can affect protein binding in a gel shift assay, we initially did not view this as contradictory to the existing models explaining the delay in activation of C/EBPα and PPARγ2 gene expression. However, ChIP assays, which specifically detect in vivo protein:DNA interactions at endogenous loci, clearly demonstrated that C/EBPβ and C/EBPδ were capable of binding to both the C/EBPα and the PPARγ2 promoter at very early times post-differentiation. Moreover, C/EBPβ and δ could also bind at early times to adipocyte specific promoters that do not begin to transcribe until much later in the differentiation process. Our results do not alter the original conclusion that mechanisms exist during the time of mitotic clonal expansion to delay activation of C/EBPα and PPARγ2 gene expression and the target genes that they subsequently activate. Instead, they indicate that the rate-limiting step is not the interaction of the C/EBPβ protein with binding sites at the endogenous target gene promoters.
Numerous possibilities to restrict the transcriptional activating properties of C/EBPβ exist even if the protein is DNA bound. The exact isoform of C/EBPβ that is bound could influence the transcriptional potential. Interaction between C/EBPβ and repressor proteins (Ron & Habener 1992, Tang et al. 1999) would not necessarily be restricted to solution interactions; repression could occur via interactions at promoter sequences as shown previously (Mo et al. 2004). DNA-bound C/EBPβ could be and likely is still subject to post-translational modifications, including phosphorylation, acetylation, and sumoylation (Kim et al. 2002, Eaton & Sealy 2003, Xu et al. 2003, Park et al. 2004a, Tang et al. 2005), that may modulate transcriptional capacity. C/EBPβ-bound loci could remain transcriptionally silent because other activators, RNA polymerase II (pol II)- or pol II-associated general transcription factors have not been synthesized, are spatially restricted, have not yet undergone the appropriate post-translational modification, or cannot bind in the absence of specific chromatin modifications or alterations. In support of this last possibility, Wiper-Bergeron et al.(2003) showed that histone deacetylase1 (HDAC1) could affect the acetylation status of the C/EBPα promoter in a manner regulated by glucocorticoids. Finally, C/EBPβ-bound loci may change position within the nucleus, becoming associated or disassociated with specific sub-nuclear structures such as pericentric heterochromatin or splicing domains. None of these possibilities are mutually exclusive, and it is likely that multiple mechanisms are acting cooperatively to control the timing of expression for each individual target gene.
We also note that, in other cell types, C/EBPβ-mediated activation of endogenous target genes can occur without induction of C/EBPβ levels or induction of post-translational modifications. Lipopolysaccharide-mediated induction of C/EBPβ target genes in B cell-derived lines lacking C/EBP proteins could be accomplished by constitutive expression of the bZIP domain of C/EBPβ (Hu et al. 2000). More recently, C/EBPβ binding to and subsequent activation of target genes in lipopolysaccharide-stimulated macrophages were shown to occur prior to induction of C/EBPβ protein levels and in the absence of induced C/EBPβ phosphorylation or nuclear translocation (Bradley et al. 2003). Thus, different mechanisms likely control C/EBPβ function in different cell types.
The in vivo binding capacity of C/EBPβ and δ to adipogenic genes as early as 1 h post-differentiation seems likely to be modulated by components of the differentiation cocktail. ChIP assays performed with 3T3-L1 overexpressing C/EBPβ in the absence of differentiation inducers showed that this regulator is not recruited to the PPARγ2 (Fig. 5). This result shows that the physical presence of C/EBPβ alone is insufficient to be recruited to specific promoters. Further studies will be required to determine the nature of modifications required for both C/EBPβ and δ to rapidly bind to adipogenic regulatory sequences following differentiation signaling as well as the relative importance of C/EBPβ homodimers and C/EBPβ and δ heterodimers.
Limited data on the binding of C/EBPβ and δ to gene promoters other than C/EBPα and PPARγ2 exist. In differentiating 3T3-L1 cells, C/EBPβ binds to the aP2 promoter between 24–72 h post-differentiation (Tang et al. 2004), and a recent report shows both C/EBPβ and δ on the adiponectin promoter in mouse adipose tissue (Park et al. 2004b). These data raise the question of why C/EBPβ and δ are bound to such promoters when evidence strongly suggests that the genes are not activated until later times when C/EBPα has replaced C/EBPβ and δ at these loci. We speculate that the presence of C/EBPβ may be part of the process by which C/EBPα is recruited to adipogenic promoters. Alternatively, or perhaps, additionally, all of the adipogenic loci undergo structural changes at the onset of differentiation, and the binding of C/EBPβ to these loci serves as a mark for subsequent gene activation.
The data have also clearly demonstrated that the C/EBP binding sites on each of the genes undergo a transition in binding from C/EBPβ and δ to C/EBPα in vivo. We and others have previously demonstrated that this occurred on the PPARγ2 promoter during adipogenesis (Salma et al. 2004, Tang et al. 2004). From the data presented here, we predict that this transition occurs on all adipocyte genes containing C/EBP binding sites with similar kinetics, indicating the importance of the transition and a role for C/EBPβ and δ in both early and late events of the differentiation program. We note that the binding of C/EBPβ and δ correlated with the induction of overall cellular levels of C/EBPβ and δ. Similarly, C/EBPβ and δ were replaced on each promoter between 24 and 48 h post-differentiation, the time at which C/EBPα protein levels begin to be induced. Thus the binding of the specific C/EBP proteins to adipocyte specific genes in vivo does not correlate with the time of target gene expression but instead occurs rapidly after the induction of C/EBP protein levels.
We are grateful to Y Ohkawa and the members of our laboratory for advice and discussion. This work was supported by a Scholar Award from the Leukemia and Lymphoma Society and by grants from the NIH and the University of Massachusetts Medical School Diabetes Endocrine Research Center to A N I. N S was supported in part by a Zelda Haidak Memorial Scholar Fellowship. The authors declare that there is no conflict of interest that would prejudice the impartiality of this scientific work.
ParkBH2004a Phosphorylation of C/EBPbeta at a consensus extracellular signal-regulated kinase/glycogen synthase kinase 3 site is required for the induction of adiponectin gene expression during the differentiation of mouse fibroblasts into adipocytes. Molecular and Cellular Biology248671–8680.