TRANSCRIPTIONAL REPRESSION, A NOVEL FUNCTION FOR 3'-UNTRANSLATED REGIONS

The transcription rates of the rat serine protease inhibitor 2.3 and 2 .1 genes (spi 2.3 and spi 2.1), which are normally very low and high, respectively, are inversely modulated during inflammation. Two growth- hormone-response elements (GHRE-I and GHRE-II) maintain the spi 2.1 ge ne under the stringent control of growth hormone [Le Cam, A., Pantescu , V., Paquereau, L., Legraverend, C., Fauconnier, G. & Asins, G. (1994 ) J. Biol. Chem. 269, 21532-21539], whereas spi 2.3 appears to escape control by this hormone, despite the presence in its promoter of a fun ctional GHRE-I. A major difference between these two otherwise very si milar genes is the presence in spi 2.3 of a specific 348-bp extension of the 3' untranslated region (3' UTR). Inserting this 3' UTR element downstream of the polyadenylation signal or upstream of the spi 2.3 pr omoter in constructs containing the chloramphenicol acetyltransferase gene strongly decreases basal transcription and inhibits growth-hormon e-stimulated transcription, but poorly affects transcriptional stimula tion by dexamethasone or interleukin-6. The spi 2.3 3' UTR extension a lso inhibits, basal and growth-hormone-induced transcription from the spi 2.1 promoter. Repressor activity appears to be distributed through out the specific extension of the 3' UTR and seems to involve interact ions with two types of 5' cis-acting promoter elements. The first is t he GAGA box, a key control spi promoter element, whose mutation faithf ully reproduces the effects of the 3' UTR silencer on spi 2.1 and spi 2.3 promoters. The second is represented by CCAAT enhancer-binding-pro tein-(C/EBP)-binding sites, whose functions are severely impaired by t he spi 2.3-specific 3' UTR extension. The presence of this silencer in the spi 2.3 gene very likely accounts for the lack of basal of transc ription in vivo and for induction of the gene during acute inflammatio n.