Cooperation between C/EBPalpha TBP/TFIIB and SWI/SNF recruiting domains is required for adipocyte differentiation

Abstract

Chromatin remodeling is an important step in promoter activation during cellular lineage commitment and differentiation. We show that the ability of the C/EBPalpha transcription factor to direct adipocyte differentiation of uncommitted fibroblast precursors and to activate SWI/SNF-dependent myeloid-specific genes depends on a domain, C/EBPalpha transactivation element III (TE-III), that binds the SWI/SNF chromatin remodeling complex. TE-III collaborates with C/EBPalpha TBP/TFIIB interaction motifs during induction of adipogenesis and adipocyte-specific gene expression. These results indicate that C/EBPalpha acts as a lineage-instructive transcription factor through SWI/SNF-dependent modification of the chromatin structure of lineage-specific genes, followed by direct promoter activation via recruitment of the basal transcription-initiation complex, and provide a mechanism by which C/EBPalpha can mediate differentiation along multiple cellular lineages.

Figure 1

C/EBPα domains required for adipogenesis in vitro. (A) Domain structure of the C/EBPα protein. The transactivation elements (TE) I through III, the phosphorylation sites for glycogen synthase kinase 3 (GSK3), and protein kinase C (PKC), and the basic region-leucine zipper (BR-LZ) DNA binding domain are indicated. (B) NIH3T3 fibroblasts were transduced with pBabePuro (vector), pBabePuro-based retrovirus encoding wild-type C/EBPα, or C/EBPα derivatives lacking TE-I (D1–70), TE-II (D70–96), TE-III (D126–200), or the GSK3 and PKC phosphorylation sites (D200–256). After 2 wk of selection for expression of the retroviral construct total cellular RNA was analyzed for the presence of the following mRNAs by Northern blotting: lipoprotein lipase (LPL), peroxisome-proliferator activating receptor γ (PPARγ), aP2–422 (aP2), and glyceraldehyde phosphate dehydrogenase (GAPDH). (C) Parallel cultures were lysed in SDS sample buffer and equivalent amounts of total cellular protein analyzed for C/EBPα expression by Western blotting using the C103 anti-C/EBPα antibody. Bands corresponding to C/EBPα derivatives are indicated with asterisks.

Figure 2

Interaction between TE-III and the SWI/SNF complex. (A) Structure of the C/EBPα derivatives used for coimmunoprecipitation assays. The transactivation elements (TE−) I through III, and the basic region-leucine zipper DNA binding domain (BR-LZ) are indicated. CR1 is the conserved region 1 from C/EBPβ (Kowenz-Leutz et al. 1994). All C/EBPα derivatives were C-terminally FLAG tagged for coimmunoprecipitation analysis. (B,C) 5 × 10⁶ C33A cells were calcium phosphate-transfected with 5μg of the indicated CMV-driven C/EBPα-FLAG and HA-hBrm expression vectors. Lysates were prepared and immunoprecipitated (IP) with M2 anti-FLAG antibody (FLAG) or with nonimmune serum (non-immune). Precipitated proteins were detected by Western blotting with PRB-205C antiHA monoclonal antibody (HA–hBrm), M2 anti-FLAG monoclonal antibody (C/EBPα-FLAG) and anti-BAF155 polyclonal antibody (BAF155). Western blot analysis of the lysates used for immunoprecipitation is shown (lysate) to verify the expression of the various input proteins.

Figure 3

TE-III dependent activation of chromosomal myeloid-specific loci. (A) Poly(A) RNA from 8 × 10⁶ QT6 fibroblasts transfected with CMV–C/EBPα expression vectors and pCRNCM–Myb expression vector (+) or the corresponding empty expression vectors, was analyzed by Northern blotting for expression of the myeloid-specific mim-1 and #126 transcripts, and for expression of GAPDH mRNA as an internal control. (B) Poly(A) RNA from 2 × 10⁷ HD3 cells transfected with the indicated CMV–C/EBPα expression vectors was analyzed by Northern blotting for expression of the myeloid-specific mim-1, #325, and GAPDH transcripts.

Figure 4

Rescue of adipogenesis by the C/EBPβ CR1 domain. NIH3T3 cells were infected with pBabePuro control virus (vector) or virus encoding the indicated C/EBPα derivatives and allowed to differentiate for 2 wk. Adipocyte morphology was analyzed by Oil Red O staining (A), and expression of the LPL, PPARγ, and aP2 adipocyte mRNAs and C/EBPα protein determined (B) as in Figure 1B. (C) Level of adipocyte differentiation determined by quantitative Oil Red O staining (red columns; error bars indicate standard deviations; n = 3) and measurement of aP2 mRNA accumulation (after normalization to GAPDH) for the indicated C/EBPα alleles. The background value obtained from cells transduced with empty vector have been subtracted.

Figure 5

Mutation analysis of CR1. (A) Amino acid sequence of mouse CR1, indicating the positions of residues mutated to alanines in the mutLL, mutWD, and mutFR CR1 variants, respectively. (B) Coimmunoprecipitation analysis of C/EBPα D126–200+CR1 and its mutant derivatives was performed as in Figure 2B. The HA–hBrm coprecipitated with the C/EBPα–FLAG-derivatives is shown in panel a, the precipitated C/EBPα–FLAG in panel b. The levels of the same proteins in the input lysate for the coimmunoprecipitation are shown in panels c and d, respectively. (C) The ability of C/EBPα D126–200+CR1 and its mutant derivatives to induce mim-1 expression in HD3 erythroblasts was analyzed as in Figure 3B. (D) The induction of aP2 mRNA in retrovirally transduced NIH3T3 cells determined as in Figure 1B.

Figure 6

Requirement of C/EBPα TBP/TFIIB interaction motifs for adipogenesis. NIH3T3 cells were infected with pBabePuro control virus (vector), virus encoding wild-type (WT) C/EBPα, or the Y67A, FL77,78AA C/EBPα mutant (YFL) and allowed to differentiate for 2 wk. Parallel cultures were analyzed for morphological adipocyte differentiation (A) and expression of the LPL, PPARγ, and aP2 adipocyte mRNAs (B) as in Figures 4A and 1B, respectively. Quantification of accumulated triglyceride (by Oil Red O staining) and aP2 mRNA (C) was done as in Figure 4C, and C/EBPα protein (D) as in Figure 1C.