Creation of genome-wide protein expression libraries using random activation of gene expression.

Here we report the use of random activation of gene expression (RAGE) to create genome-wide protein expression libraries. RAGE libraries containing only 5 x 10(6) individual clones were found to express every gene tested, including genes that are normally silent in the parent cell line. Furthermore, endogenous genes were activated at similar frequencies and expressed at similar levels within RAGE libraries created from multiple human cell lines, demonstrating that RAGE libraries are inherently normalized. Pools of RAGE clones were used to isolate 19,547 human gene clusters, approximately 53% of which were novel when tested against public databases of expressed sequence tag (EST) and complementary DNA (cDNA). Isolation of individual clones confirmed that the activated endogenous genes can be expressed at high levels to produce biologically active proteins. The properties of RAGE libraries and RAGE expression clones are well suited for a number of biotechnological applications including gene discovery, protein characterization, drug development, and protein manufacturing.

Yeast HOS3 forms a novel trichostatin A-insensitive homodimer with intrinsic histone deacetylase activity.

Histone deacetylases such as human HDAC1 and yeast RPD3 are trichostatin A (TSA)-sensitive enzymes that are members of large, multiprotein complexes. These contain specialized subunits that help target the catalytic protein to histones at the appropriate DNA regulatory element, where the enzyme represses transcription. To date, no deacetylase catalytic subunits have been shown to have intrinsic activity, suggesting that noncatalytic subunits of the deacetylase complex are required for their enzymatic function. In this paper we describe a novel yeast histone deacetylase HOS3 that is relatively insensitive to the histone deacetylase inhibitor TSA, forms a homodimer when expressed ectopically both in yeast and Escherichia coli, and has intrinsic activity when produced in the bacterium. Most HOS3 protein can be found associated with a larger complex in partially purified yeast nuclear extracts, arguing that the HOS3 homodimer may be dissociated from a very large nuclear structure during purification. We also demonstrate, using a combination of mass spectrometry, tandem mass spectrometry, and proteolytic digestion, that recombinant HOS3 has a distinct specificity in vitro for histone H4 sites K5 and K8, H3 sites K14 and K23, H2A site K7, and H2B site K11. We propose that while factors that interact with HOS3 may sequester the catalytic subunit at specific cellular sites, they are not required for HOS3 histone deacetylase activity.

Transcriptional repression by UME6 involves deacetylation of lysine 5 of histone H4 by RPD3.

The histone deacetylase RPD3 can be targeted to certain genes through its interaction with DNA-binding regulatory proteins. RPD3 can then repress gene transcription. In the yeast Saccharomyces cerevisiae, association of RPD3 with the transcriptional repressors SIN3 and UME6 results in repression of reporter genes containing the UME6-binding site. RPD3 can deacetylate all histone H4 acetylation sites in cell extracts. However, it is unknown how H4 proteins located at genes near UME6-binding sites are affected, nor whether the effect of RPD3 is localized to the promoter regions. Here we study the mechanism by which RPD3 represses gene activity by examining the acetylation state of histone proteins at UME6-regulated genes. We used antibodies specific for individual acetylation sites in H4 to immunoprecipitate chromatin fragments. A deletion of RPD3 or SIN3, but not of the related histone-deacetylase gene HDA1, results in increased acetylation of the lysine 5 residue of H4 in the promoters of the UME6-regulated INO1, IME2 and SPO13 genes. As increased acetylation of this residue is not merely a consequence of gene transcription, acetylation of this site may be essential for regulating gene activity.

HDA1 and RPD3 are members of distinct yeast histone deacetylase complexes that regulate silencing and transcription.

Increased histone acetylation has been correlated with increased transcription, and regions of heterochromatin are generally hypoacetylated. In investigating the cause-and-effect relationship between histone acetylation and gene activity, we have characterized two yeast histone deacetylase complexes. Histone deacetylase-A (HDA) is an approximately 350-kDa complex that is highly sensitive to the deacetylase inhibitor trichostatin A. Histone deacetylase-B (HDB) is an approximately 600-kDa complex that is much less sensitive to trichostatin A. The HDA1 protein (a subunit of the HDA activity) shares sequence similarity to RPD3, a factor required for optimal transcription of certain yeast genes. RPD3 is associated with the HDB activity. HDA1 also shares similarity to three new open reading frames in yeast, designated HOS1, HOS2, and HOS3. We find that both hda1 and rpd3 deletions increase acetylation levels in vivo at all sites examined in both core histones H3 and H4, with rpd3 deletions having a greater impact on histone H4 lysine positions 5 and 12. Surprisingly, both hda1 and rpd3 deletions increase repression at telomeric loci, which resemble heterochromatin with rpd3 having a greater effect. In addition, rpd3 deletions retard full induction of the PHO5 promoter fused to the reporter lacZ. These data demonstrate that histone acetylation state has a role in regulating both heterochromatic silencing and regulated gene expression.

HDA1 and HDA3 are components of a yeast histone deacetylase (HDA) complex.

Histone acetylation is maintained through the action of histone acetyltransferases and deacetylases and has been correlated with increased gene activity. To investigate the functional role of these enzymes in the regulation of transcription, we have purified from Saccharomyces cerevisiae two histone deacetylase activities, HDA and HDB, with molecular masses of approximately 350 and 600 kDa, respectively. In vitro, the HDA activity deacetylates all four core histones, has a preference for histone H3, and is strongly inhibited by trichostatin A (a specific inhibitor of histone deacetylases). HDB is considerably less sensitive to trichostatin A. We report the extensive purification of the HDA activity and the identification of peptides (p75, p73, p72, and p71) whose presence correlates with deacetylase activity on native polyacrylamide gels. An antibody to p75 immunoprecipitates peptides with molecular masses similar to those in the 350-kDa complex. Additionally, antibodies to p75 and p71 specifically precipitate histone deacetylase activity and co-immunoprecipitate each other. Gene disruptions of p75 (HDA1) or p71 (HDA3) cause the loss of the 350-kDa (but not the 600-kDa) activity from our chromatography profiles. These data argue strongly that HDA1 and HDA3 are subunits of the HDA complex, which is structurally distinct from the second, HDB complex.

Quantitative differences in androgen and glucocorticoid receptor DNA binding properties contribute to receptor-selective transcriptional regulation.

Androgen receptor (AR) and glucocorticoid receptor (GR) belong to the same subfamily of steroid/nuclear receptors and have been shown to bind qualitatively to the same hormone response element (HRE) DNA sequences. Despite this similarity in target gene recognition, AR and GR have differential affects on the transcriptional regulation of genes containing both simple and complex HRE control regions. Using HREs from the mouse mammary tumor virus (MMTV), tyrosine aminotransferase (TAT), prostatein (C3) or sex-limited protein (SLP) genes, linked to the thymidine kinase promoter, we found receptor-selective differences in the ability of rat AR and rat GR to induce transcription of these various reporter genes. Since AR and GR have a 20% amino acid sequence difference in their DNA binding domains (DBDs), which could result in altered DNA binding affinities, we measured the ability of purified AR and GR DBDs to bind selectively and with high affinity to these HRE sequences in vitro. Gel shift mobility assays showed that the GR DBD had a higher affinity for a consensus HRE than did the AR DBD, and quantitative DNase I footprinting revealed that AR and GR DBDs bound to the MMTV, TAT, C3 and SLP HREs with different affinities. It was found that AR had a dissociation constant (Kd) that was 2-3 times higher than GR on the TAT, C3 and SLP HREs and that the Kd of AR for the C3 and SLP HREs differed by an order of magnitude (43 nM and 460 nM, respectively). Taken together, these data suggest that amino acid differences in the AR and GR DBDs contribute to altered receptor-DNA interactions, however it is likely that non-receptor factors are involved in further modulating receptor-selective DNA binding and transactivation functions.

Characterization of a panel of rat ventral prostate epithelial cell lines immortalized in the presence or absence of androgens.

We have transfected rat ventral prostate (RVP) epithelial cells with a plasmid containing the SV40 large T-antigen in an attempt to establish a panel of cell lines that will be useful in molecular genetic studies of prostate cell function. Since the distribution of cell types in the RVP is dramatically affected by androgen withdrawal and replacement, cells isolated from normal, castrated, or castrated rats that were given daily injections of testosterone were used in these experiments. Cell lines were established in media that were supplemented or depleted of androgens to accommodate the possible requirements of different prostate cell types. Numerous cell lines were isolated which retain characteristics of RVP epithelial cells and five of these cell lines were studied in detail. All five cell lines express the SV40 large T-antigen, supporting the role of this viral protein in immortalization. The RVP cell lines were shown to contain high levels of functional glucocorticoid receptors, but very low levels of androgen binding activity even though androgen receptor RNA could be detected. It was determined that the decreased androgen receptor activity in the RVP cells was apparently due to low receptor expression based on the results of transient transfection assays using androgen receptor cDNA. Taken together, the biochemical, cytological, and morphological characterizations of the RVP cell lines suggest that they may all have been derived from basal prostate epithelial cells despite the initial differences in androgen status of the animal and the level of androgens in the culture media.

Functional characterizations of the androgen receptor confirm that the molecular basis of androgen action is transcriptional regulation.

In an effort to understand the molecular basis of androgen action in the prostate, we isolated androgen receptor (AR) cDNA from rat ventral prostate cells and analyzed the transcriptional regulatory activity of the encoded protein in a cotransfection assay. We found that AR is capable of inducing chloramphenicol acetyltransferase activity more than 20-fold using the mouse mammary tumor virus LTR as a source of androgen response elements. This induction was observed in both monkey CV1 cells and human HeLa cells, neither of which contains endogenous functional AR, and was entirely dependent on added androgens. Deletion mapping studies showed that carboxy-terminal deletions of approximately 250 amino acids convert AR into a constitutive activator of transcription. In addition, a chimeric receptor protein containing the amino-terminus and DNA-binding domains of AR fused to the previously defined ligand domain of the glucocorticoid receptor was found to be fully functional based on dexamethasone-induced chloramphenicol acetyltransferase activity. Our results support the prediction that androgens modulate rates of transcriptional initiation, suggesting that posttranscriptional effects of androgens are secondary responses. Moreover, these data reveal that, like other steroid receptors, AR contains a number of distinct regulatory regions important for normal activity. The isolation and characterization of fully functional AR sequences will facilitate the use of molecular genetics to study complex androgen responses in target tissues such as the prostate.

Aphidicolin resistance in herpes simplex virus type I reveals features of the DNA polymerase dNTP binding site.

We describe the mapping and sequencing of mutations within the DNA polymerase gene of herpes simplex virus type 1 which confer resistance to aphidicolin, a DNA polymerase inhibitor. The mutations occur near two regions which are highly conserved among DNA polymerases related to the herpes simplex enzyme. They also occur near other herpes simplex mutations which affect the interactions between the polymerase and deoxyribonucleoside triphosphate substrates. Consequently, we argue in favor of the idea that the aphidicolin binding site overlaps the substrate binding site and that the near-by conserved regions are functionally required for substrate binding. Our mutants also exhibit abnormal sensitivity to another DNA polymerase inhibitor, phosphonoacetic acid. This drug is thought to bind as an analogue of pyrophosphate. A second-site mutation which suppresses the hypersensitivity of one mutant to phosphonoacetic acid (but not its aphidicolin resistance) is described. This second mutation may represent a new class of mutations, which specifically affects pyrophosphate, but not substrate, binding.