CACCC and GATA-1 sequences make the constitutively expressed alpha-globin gene erythroid-responsive in mouse erythroleukemia cells.

Although the human alpha-globin and beta-globin genes are co-regulated in adult life, they achieve the same end by very different mechanisms. For example, a transfected beta-globin gene is expressed in an inducible manner in mouse erythroleukemia (MEL) cells while a transfected alpha-globin gene is constitutively expressed at a high level in induced and uninduced MEL cells. Interestingly, when the alpha-globin gene is transferred into MEL cells as part of human chromosome 16, it is appropriately expressed in an inducible manner. We explored the basis for the lack of erythroid-responsiveness of the proximal regulatory elements of the human alpha-globin gene. Since the alpha-globin gene is the only functional human globin gene that lacks CACCC and GATA-1 motifs, we asked whether their addition to the alpha-globin promoter would make the gene erythroid-responsive in MEL cells. The addition of each of these binding sites to the alpha-globin promoter separately did not result in inducibility in MEL cells. However, when both sites were added together, the alpha-globin gene became inducible in MEL cells. This suggests that erythroid non-responsiveness of the alpha-globin gene results from the lack of erythroid binding sites and is not necessarily a function of the constitutively active, GC rich promoter.     

Structure and transcriptional regulation of the mouse ferrochelatase gene

Ferrochelatase (EC.4.99.1.1), the final step in the biosynthesis of heme, is widely expressed in various tissues and is induced in erythroid cells. We determined the structure of the mouse ferrochelatase gene after isolation and characterization of lambda phage clones mapping discrete regions of the cDNA. The gene spans about 25 kb and consists of 11 exons. The exon/intron boundary sequences conform to consensus acceptor (GTn)/donor (nAG) sequences, and exons in the gene encode functional protein domains. The promoter region contains multiple Sp1 sites, a CACCC box and GATA-1 binding sites. Function analysis of the promoter by transient transfection assay demonstrated that one Sp1 binding site located at -37/-32 is essential for basic expression of the ferrochelatase gene in both mouse erythroleukemia (MEL) and non-erythroid EL4 cells. In addition, the region (-66/-51) containing a CACCC box and the neighboring GC box partly contributes to the inducible activity of the reporter in MEL cells upon induction with dimethylsulfoxide. It appears that at least two promoter regions of the mouse ferrochelatase gene function in basic and inducible expression.     

Activation of cAMP-dependent protein kinase alters the chromatin structure of the urokinase-type plasminogen activator gene promoter.

In LLC-PK1 cells, the urokinase-type plasminogen activator (uPA) gene is induced by two of the major signal transduction pathways, the protein kinase C (PKC) and the cAMP-dependent protein kinase (PKA) pathways. We have analyzed the chromatin structure of 26 kb of the uPA gene locus and have shown that PKA activation but not PKC activation induce major chromatin structural alterations in the uPA gene promoter. In uninduced cells, several DNase I hypersensitive (HS) sites were detected in the 5' and 3' flanking regions but not in the transcribed region. Two of the sites correspond to previously characterized regulatory sites: a cAMP responsive site at nucleotide position -3500 with respect to the initiation site, and the PEA3/AP1 site at -2100 that mediates PKC activation. After the activation of PKA but not PKC, a strong HS site was induced at -2600. Functional analysis of this region revealed cAMP responsive activity. Chromatin structural alterations again brought about specifically by PKA but not by PKC were were also detected in the upstream of the promoter by topoisomerase I cleavage site analysis, with two prominent sites appearing at -2800 and -3300. These results suggest that the strong cAMP induction of the uPA gene requires structural alterations that permit cooperative interactions between the multiple cAMP responsive sites.     

Differential effect of hemin-controlled eIF-2 alpha kinases from mouse erythroleukemia cells on protein synthesis

Cultured mouse erythroleukemia (MEL) cells can be induced to erythroid differentiation by a variety of chemical agents. This differentiation process is marked by the onset of globin mRNA and hemoglobin synthesis. In rabbit reticulocytes, globin synthesis is regulated by a hemin-controlled translational inhibitor (HCI) which acts via phosphorylation of the alpha subunit of eukaryotic initiation factor 2 (eIF-2). From both uninduced and induced MEL cells, hemin-controlled eIF-2 alpha kinases have been partially purified. They resemble HCI with respect to their chromatographic behaviour and their sensitivity towards physiological concentrations of hemin (5-10 microM). Further purification on phosphocellulose, however, reveals that the eIF-2 alpha kinase from uninduced MEL cells is chromatographically distinct from HCI, whilst the eIF-2 alpha kinase activity from induced MEL cells represents a mixture of the former and the HCI-type eIF-2 alpha kinase. The latter inhibits protein synthesis in a fractionated system from rabbit reticulocytes which is free of, but sensitive to, HCI, whereas the eIF-2 alpha kinase from uninduced MEL cells does not show any inhibitory activity. This observation is supported by the finding that induced MEL cells respond in vivo to iron depletion with a shut-off of protein synthesis (as do rabbit reticulocytes), whilst uninduced MEL cells do not.     

Characterization of trans-acting factor(s) regulating beta-globin gene expression by in vivo competition

A beta-globin/TK fusion gene was microinjected into non-erythroid cells (Ltk- cells) and erythroid cells (murine erythroleukemia (MEL) cells), and the interactions of the regulatory cellular factors with the beta-globin sequences were investigated by the in vivo competition experiment. The fusion gene was expressed efficiently in Ltk- cells. This expression was inhibited by a co-injection with a three-fold molar excess of the 5'-flanking sequence of the beta-globin gene or with a nine-fold molar excess of the mammary tumor virus LTR, but not with the alpha-globin gene. The fusion gene was expressed very poorly in the uninduced MEL cells and highly in the induced MEL cells. The co-injection of the beta-globin gene did not affect expression in the MEL cells in either uninduced or induced conditions.     

Inhibition of expression of a mouse alpha-globin gene by plasmids that include antisense oligonucleotides.

Plasmid-borne DNAs, corresponding to 68-base oligodeoxynucleotides, synthesized in the antisense or sense configuration and based on the nucleotide sequences of various regions of the mouse alpha-globin mRNA, were introduced with the gene for xanthine-guanine phosphoribosyl transferase from E. coli (Ecogpt) into mouse erythroleukemia (MEL) cells by protoplast fusion. Specific inhibition of the synthesis of alpha-globin was observed only in the cells transformed with the plasmids with antisense 68-mers that corresponded to the cap site as well as the site of initiation of translation of alpha-globin mRNA (Oligo-A); Other plasmids with antisense 68-mers that corresponded to the regions of the exon/intron junctions, the individual exons, or the 3' untranslated region were ineffective. This antisense RNA efficiently reduced the production of alpha-globin to 9-18% of the endogenous level after induction with hexylmethylene-bis-acetoamide (HMBA). Moreover, most of the antisense transformants did not show any decrease in the expression of the c-myc gene at the early phases of differentiation of MEL cells. Thus, we propose a hypothesis that the early decline in levels of c-myc mRNA may be independent of and uncoupled from the program of globin synthesis during the differentiation of MEL cells.