Differential binding of cAMP-responsive-element(CRE)-binding protein-1 and activating transcription factor-2 to a CRE-like element in the human tissue-type plasminogen activator (t-PA) gene promoter correlates with opposite regulation of t-PA by phorbol ester in HT-1080 and HeLa cells

The human tissue-type plasminogen activator gene (t-PA) is induced by the phorbol ester, phorbol 12-myristate 13-acetate (PMA), in HeLa cells. Previous studies in transfected HeLa cells identified two cis-acting regulatory elements within the t-PA gene promoter responsible for both constitutive and PMA-inducible expression. One element differs from the consensus cAMP response element (CRE) by a single nucleotide substitution (referred to in this report as t-PACRE) and another which bears similarity to the AP-2 recognition sequence. In HT-1080 fibrosarcoma cells, t-PA mRNA levels are expressed at higher constitutive levels and are suppressed by PMA. Nuclear run-on transcription experiments indicate that PMA-mediated suppression of t-PA in these cells is associated with a decrease in t-PA gene template activity. We designed experiments to determine whether nuclear t-PACRE or AP-2-like binding proteins were differentially expressed in HeLa and HT-1080 cells and, accordingly, if these could be correlated with the opposite effect of PMA on t-PA expression. Band shift analyses indicated that the migration profiles of HeLa and HT-1080 nuclear proteins interacting with the AP-2-like site were indistinguishable; however, those produced with the t-PACRE binding site were qualitatively and quantitatively distinct. The distribution of t-PACRE binding proteins in these cells was investigated in a supershift assay using specific antibodies against members of the fos/jun and CRE-binding protein (CREB)/activating transcription factor (ATF) families. In HT-1080 cells, CREB-1 was the most prominent t-PACRE-binding activity detected and was greatly increased in cells treated with PMA. In contrast, CREB-1 activity was absent in HeLa cells, but antibodies specific for ATF-2 produced a marked supershifted complex which was unaffected by PMA treatment. Since CREB-1 can repress transcription of other target genes (including c-jun) via association with identical cis-acting CRE-like sequences, we suggest that the mechanism for the transcriptional down-regulation of t-PA by PMA in HT-1080 cells requires CREB-1 binding to the t-PACRE while ATF-2, by associating with the same site, plays a role in PMA-mediated induction of t-PA in HeLa cells.