Signalling by steroid hormones is mediated by receptor proteins that bind hormonal ligands and regulate the transcription of specific genes. The heat-shock protein hsp90 seems to associate selectively with unliganded receptors (aporeceptors), but it has not been determined whether this interaction affects receptor function in vivo. To address the role of hsp90, we have taken advantage of the capacity of mammalian steroid receptors to function in yeast. We constructed a strain of Saccharomyces cerevisiae in which hsp90 expression was regulatable and could be reduced more than 20-fold relative to wild type. At low levels of hsp90, aporeceptors seem to be mostly hsp90-free, yet fail to enhance transcription; on hormone addition, the receptors are activated but with markedly reduced efficiency. Thus hsp90 does not inhibit receptor function solely by steric interference; rather, hsp90 seems to facilitate the subsequent response of the aporeceptor to the hormonal signal. This is the first biological evidence that hsp90 acts in the signal transduction pathway for steroid receptors.

Site-specific mutagenesis of the highly conserved milk box (-140 to -110) region suggested that I-casein expression is regulated by a hormone-mediated relief of repression (M. Schmitt-Ney, W. Doppler, R. K. Ball, and B. Groner, Mol. Cell. Biol. 11:3745-3755, 1991). However, when this sequence was placed upstream of a heterologous thymidine kinase promoter, it activated reporter gene expression. This apparent paradox was resolved when the trans-acting factor YY1, capable of acting as both a positive and negative regulator, was shown to interact with the milk box region, using bacterially expressed WY1 and specific oligonucleotide and antibody competition experiments. Second, it was demonstrated that extracts prepared from several cell types contained a protein(s) interacting with the mammary gland-specific factor (MGF) binding site, previously shown to be required for j-casein promoter activity (Schmitt-Ney et al., Mol. Cell. Biol. 11:3745-3755, 1991 construct. The second of these interesting regions, originally termed the milk box consensus (17), at -140 to -110, was shown to repress reporter gene expression. Deletion or sitespecific mutagenesis of the milk box region led to increased basal activity of a 3-casein-chloramphenicol acetyltransferase (CAT) construct. To date, however, no single binding region which is able to confer proper hormonal regulation of a reporter gene in cells has been characterized, suggesting that cooperative factor interactions may be required for appropriate hormonal regulation of r-casein transcription.To begin to decipher this complex factor interplay, the trans-acting factor-DNA interactions at these two important sites in the 13-casein promoter were further characterized. Initially, the ubiquitous trans-acting factor YY1 (yin and yang 1) (41) was identified in mammary epithelial cell extracts and shown to bind in the ,B-casein repressor milk box region. Second, an MGF DNA binding activity was observed in whole cell extracts (WCE) and nuclear extracts of several different cell types, suggesting that a protein(s) interacting with this site is ubiquitous and not mammary gland specific. Sequence analysis of the MGF binding site revealed a striking similarity to the recently described gamma interferon (IFN--y)-activated sequence (GAS), present in a number of genes (27

No abstract

Plant genes usually have multiple potential poly(A) addition signals, and different sites are used for 3' processing of transcripts from a single gene. In contrast, we show here that four barley genes that also have multiple poly(A) addition signals conforming to the plant consensus use only one signal. In each of these genes, the region of cleavage for poly(A) addition is centered on a conserved pentanucleotide. This AGGCA is followed by a conserved sequence homologous to sequences involved in self-cleavage of plant viroid RNA precursors; immediately following, in turn, are four or five nucleotides complementary to the nucleotides immediately preceding AGGCA in each gene. The presence of these conserved sequences and their association with a single region for poly(A) addition in three different gene types (high-pI and low-pI a-amylase genes and a thiol protease) that otherwise are not homologous in their 3' untranslated/flanking sequences suggest that they might participate in some common regulatory mechanism shared by these genes.