Search

We have identified Serine 73 as a novel GSK-3β site on SREBP-1c that alters its affinity for SCAP, and proteasomal degradation. Phosphorylation of Serine 73 by GSK-3β during starvation (insulin-depleted stat) may lead to lower levels of SREBP-1c; conversely, de-phosphorylation of this site may be involved in stabilizing SREBP-1c by insulin (by blocking GSK-3β action). A functional role of this site needs to be corroborated in vivo., Sterol regulatory element binding protein-1c (SREBP-1c) is a key transcription factor that regulates genes involved in the de novo lipid synthesis and glycolysis pathways. The structure, turnover and transactivation potential of SREBP-1c are regulated by macronutrients and hormones via a cascade of signalling kinases. Using MS, we have identified serine 73 as a novel glycogen synthase kinase-3 (GSK-3) phosphorylation site in the rat SREBP-1c purified from McA-RH7777 hepatoma cells. Our site-specific mutagenesis strategy revealed that the turnover of SREBP-1c, containing wild type, phospho-null (serine to alanine) or phospho-mimetic (serine to aspartic acid) substitutions, was differentially regulated. We show that the S73D mutant of pSREBP-1c, that mimicked a state of constitutive phosphorylation, dissociated from the SREBP-1c–SCAP complex more readily and underwent GSK-3-dependent proteasomal degradation via SCFFbw7 ubiquitin ligase pathway. Pharmacologic inhibition of GSK-3 or knockdown of GSK-3 by siRNA prevented accelerated degradation of SREBP-1c. As demonstrated by MS, SREBP-1c was phosphorylated in vitro by GSK-3β at serine 73. Phosphorylation of serine 73 also occurs in the intact liver. We propose that GSK-3-mediated phosphorylation of serine 73 in the rat SREBP-1c and its concomitant destabilization represents a novel mechanism involved in the inhibition of de novo lipid synthesis in the liver.

Msx1 promoter is known to be repressed by Msx1 protein [Shetty, Takahashi, Matsui, Iyengar and Raghow (1999) Biochem. J. 339, 751–758]. We show that in the transiently transfected C2C12 myoblasts, co-expression of Msx3 also causes potent repression of Msx1 promoter that can be relieved by exogenous expression of cAMP-response-element-binding protein-binding protein (CBP) and p300 in a dose-dependent manner. Co-immunoprecipitation and Western blot analyses revealed that Msx3 interacts with CBP and p300 and this interaction significantly decreases the histone acetyltransferase (HAT) activity of both proteins. We also discovered that Msx3-mediated repression of Msx1 promoter is synergized by the exogenous co-expression of histone deacetylase 1 (HDAC1). Furthermore, the repression of Msx1 promoter by Msx3 could be relieved by treating transfected cells with trichostatin A, an inhibitor of HDAC(s). Finally, we show that Msx3 and HDAC1 can be co-immunoprecipitated in a complex that does not contain CBP and that Msx3 and HDAC1 proteins are co-localized in the nucleus. Taken together, our results strongly suggest that two distinct multiprotein complexes are present within the nuclei of C2C12 cells: one containing Msx3 and HDAC(s) and another containing Msx3 and CBP and/or p300. On the basis of these results, we propose a dual mechanism of repression by Msx3 protein that involves the squelching of the HAT activity of co-activators, CBP and p300, and recruitment of HDAC(s).

The TATA-less murine Msx1 promoter contains two Msx1-binding motifs, located at -568 to -573 and +25 to +30, and is subject to potent autorepression [Takahashi, Guron, Shetty, Matsui and Raghow (1997) J. Biol. Chem. 272, 22667-22678]. To investigate the molecular mechanism by which Msx1 represses the activity of its own promoter, we transfected C2C12 myoblasts with Msx1-promoter-luciferase constructs and assessed reporter gene activity, with and without the exogenous expression of Msx1. We demonstrate that Msx1-mediated autorepression remained unaffected, regardless of the presence or absence of the Msx1 recognition motifs on the promoter. Furthermore, graded exogenous expression of TATA-binding protein (TBP), Sp1 or cAMP-response-element-binding protein-binding protein (CBP/p300) could counteract the autoinhibitory activity of Msx1. Finally, we demonstrate that Msx1 protein can be immunoprecipitated in a multiprotein complex containing TBP, Sp1 and CBP/p300. We hypothesize that the interaction of Msx1 protein with one or more ubiquitous or tissue-restricted transcription factors mediates transcriptional autorepression of the Msx1 gene.

Activation of liver Ito cells is characterized by increased proliferation, fibrogenesis, loss of cellular retinoid and change of cell-shape. Here, we have described fundamental differences between freshly isolated Ito cells (FIC) and long-term cultured Ito cells (LTIC). This process of activation correlates with the absence of expression of Pro alpha 1(I) gene in FIC. LTIC expressed abundant transcripts of Pro alpha 1(I) gene. Nuclear run-off experiments showed the inability of FIC to support Pro alpha 1(I) RNA transcription while LTIC transcribed it greater than 5-fold as compared with FIC. Transforming growth factor beta (TGF beta)-treated LTIC had a preferential increase in the rate of Pro alpha 1(I) gene transcription as compared with control LTIC. A human collagen type I promoter-enhancer construct (pCOL-KT) [Thompson, Simkevich, Holness, Kang and Raghow (1991) J. Biol. Chem. 266, 2549-2556] was readily expressed in LTIC but failed to be expressed in FIC. Furthermore, TGF beta treatment of LTIC resulted in an increased expression of pCOL-KT. The deletion of an activator protein-1 (AP-1) binding site (+598 to +604) in the 360 bp enhancer region of pCOL-KT (S360) caused decreased expression of the CAT reporter gene, suggesting that this bonafide AP-1 site can, at least in part, mediate the transactivation effect of TGF beta. Using DNAase I protection, we demonstrate a single foot-print located at +590 to +625 in the S360 fragment; nuclear extracts prepared from TGF beta-treated LTIC exhibited greater activity of these AP-1 binding proteins. Gel mobility assays corroborated and extended the footprinting observation. No AP-1-binding activity was found in the nuclear extracts of FIC. Double-stranded oligonucleotides containing the consensus AP-1 motif were able to compete out the binding; consensus NF-1 motif oligonucleotides failed to do so. The preincubation of nuclear extracts from control and TGF beta-treated LTIC with antibodies against c-jun and c-fos rendered a reduced binding of AP-1 proteins to the target S360 fragment.

The transcriptional activity of plasmid pCOL-KT, in which human pro alpha 1 (I) collagen gene upstream sequences up to -804 and most of the first intron (+474 to +1440) drive expression of the chloramphenicol acetyltransferase (CAT) gene [Thompson, Simkevich, Holness, Kang & Raghow (1991) J. Biol. Chem. 266, 2549-2556], was tested in a number of mesenchymal and non-mesenchymal cells. We observed that pCOL-KT was readily expressed in fibroblasts of human (IMR-90 and HFL-1), murine (NIH 3T3) and avian (SL-29) origin and in a human rhabdomyosarcoma cell line (A204), but failed to be expressed in human erythroleukaemia (K562) and rat pheochromocytoma (PC12) cells, indicating that the regulatory elements required for appropriate tissue-specific expression of the human pro alpha 1 (I) collagen gene were present in pCOL-KT. To delineate the nature of cis-acting sequences which determine the tissue specificity of pro alpha 1 (I) collagen gene expression, functional consequences of deletions in the promoter and first intron of pCOL-KT were tested in various cell types by transient expression assays. Cis elements in the promoter-proximal and intronic sequences displayed either a positive or a negative influence depending on the cell type. Thus deletion of fragments using EcoRV (nt -625 to -442 deleted), XbaI (-804 to -331) or SstII (+670 to +1440) resulted in 2-10-fold decreased expression in A204 and HFL-1 cells. The negative influences of deletions in the promoter-proximal sequences was apparently considerably relieved by deleting sequences in the first intron, and the constructs containing the EcoRV/SstII or XbaI/SstII double deletions were expressed to a much greater extent than either of the single deletion constructs. In contrast, the XbaI* deletion (nt -804 to -609), either alone or in combination with the intronic deletion, resulted in very high expression in all cells regardless of their collagen phenotype; the XbaI*/(-SstII) construct, which contained the intronic SstII fragment (+670 to +1440) in the reverse orientation, was not expressed in either mesenchymal or nonmesenchymal cells. Based on these results, we conclude that orientation-dependent interactions between negatively acting 5′-upstream sequences and the first intron determine the mesenchymal cell specificity of human pro alpha 1 (I) collagen gene transcription.