Polyamines are organic polycations essential for a wide variety of cellular functions, including nuclear integrity and chromosome condensation. Here we present genetic evidence that depletion of cellular polyamines partially alleviates the defects in HO and SUC2 expression caused by inactivation of the GCN5 histone acetyltransferase. In addition, the combination of polyamine depletion and a sin(-) allele of the histone H4 gene leads to almost complete bypass of the transcriptional requirement for GCN5. In contrast, polyamine depletion does not alter the transcriptional requirements for the SWI/SNF chromatin remodeling complex nor does depletion lead to global defects in transcriptional regulation. In addition to these genetic studies, we show that polyamines facilitate oligomerization of nucleosomal arrays in vitro, and that polyamine-mediated condensation requires intact core histone N-terminal domains and is inhibited by histone hyperacetylation. Our studies suggest that polyamines are repressors of transcription in vivo, and that one role of histone hyperacetylation is to antagonize the ability of polyamines to stabilize highly condensed states of chromosomal fibers.

The compaction of the eukaryotic genome into a highly folded chromatin structure necessitates cellular mechanisms for allowing access of regulatory proteins to the DNA template. Recent advances in the fields of gene silencing, transcription, recombination, and DNA repair have led to the identification of two distinct families of chromatin remodeling enzymes--nuclear histone acetyltransferases and multisubunit complexes that harbor a SWI2/SNF2 ATPase family member. This paper reviews the current notion of how these enzymes function in remodeling chromatin; we then discuss some tantalizing lines of evidence that lead to the hypothesis that members of both families may actually function in concert to facilitate cellular processes in the context of chromatin.

The compaction of the eukaryotic genome into a complex, highly folded chromatin structure necessitates cellular mechanisms for allowing access of regulatory proteins to the DNA template. Recent advances have led to the identification of two distinct families of chromatin remodeling enzymes--multi-subunit complexes that harbor a SWI2/SNF2 ATPase family member, and the nuclear acetyltransferases. The Saccharomyces cerevisiae SWI/SNF complex, the prototype for the ATP-dependent chromatin remodeling machines, is required for expression of a subset of genes in yeast. This 2MDa multimeric assembly is believed to facilitate transcriptional enhancement by antagonizing chromatin-mediated transcriptional repression through disruption of histone-DNA contacts. In an attempt to identify components or regulators of the SWI/SNF complex, we have cloned three previously identified genes, ADA2, ADA3, and GCN5, that encode subunits of a complex distinct from SWI/SNF. During the course of this thesis work, one of these gene products, GCN5, was identified as the first catalytic nuclear histone acetyltransferase. The goal of this thesis work was to determine the role of the ADA/GCN5 complex in transcriptional activation in Saccharomyces cerevisiae. Using in vivo functional and genetic analysis, we have found that mutations in ADA2, ADA3, and GCN5 cause phenotypes strikingly similar to those of swi/snf mutants. ADA2, ADA3, and GCN5 are required for full expression of all SWI/SNF-dependent genes tested, including HO, SUC2, INO1, and Ty elements. Furthermore, mutations in the SIN1 gene, which encodes a non-histone chromatin component, or mutations in histones H3 or H4, alleviate the transcriptional defects caused by ada/gcn5 or swi/snf mutations. We have also found that ada2 swi1, ada3 swi1, and gcn5 swi1 double mutants are inviable and that mutations in SIN1 allow viability of these double mutants. To determine the biochemical activities of the native GCN5-containing complex in yeast, we have partially purified three chromatographically distinct GCN5-dependent acetyltransferase activities. We have found that these three acetyltransferase complexes demonstrate unique substrate specificities for free histones and histones assembled into nucleosomal arrays. Additionally, we found that these enzymes not only acetylate histones, but also purified yeast Sin1 protein, a non-histone chromatin component that resembles HMG1. We have also established a functional relationship between GCN5-dependent histone acetylation and polyamine-dependent chromatin condensation. We have found that depletion of cellular polyamines alleviates transcriptional defects caused by inactivation of the GCN5 histone acetyltransferase. In contrast, polyamine depletion does not alter the transcriptional requirements for the SWI/SNF chromatin remodeling complex. We have also found that polyamines facilitate oligomerization of nucleosomal arrays in vitro. Furthermore, this polyamine-mediated condensation reaction requires intact N-terminal domains of the core histones, and is inhibited by hyperacetylation of these domains. The results presented throughout this thesis support roles for the ADA/GCN5 products in antagonizing chromatin. In vivo analysis suggests a functional relationship between the ADA/GCN5 acetyltransferase complex (or complexes) and the SWI/SNF complex. These comp1exes may operate in concert at nucleosomes within specific promoters to facilitate activated transcription. Furthermore, our studies suggest that polyamines are repressors of transcription in vivo, and that an additional role of histone hyperacetylation is to antagonize the ability of polyamines to stabilize highly condensed states of chromosomal fibers.

The Saccharomyces cerevisiae SWI/SNF complex is a 2-MDa multimeric assembly that facilitates transcriptional enhancement by antagonizing chromatin-mediated transcriptional repression. We show here that mutations in ADA2, ADA3, and GCN5, which are believed to encode subunits of a nuclear histone acetyltransferase complex, cause phenotypes strikingly similar to that of swi/snf mutants. ADA2, ADA3, and GCN5 are required for full expression of all SWI/SNF-dependent genes tested, including HO, SUC2, INO1, and Ty elements. Furthermore, mutations in the SIN1 gene, which encodes a nonhistone chromatin component, or mutations in histone H3 or H4 partially alleviate the transcriptional defects caused by ada/gcn5 or swi/snf mutations. We also find that ada2 swi1, ada3 swi1, and gcn5 swi1 double mutants are inviable and that mutations in SIN1 allow viability of these double mutants. We have partially purified three chromatographically distinct GCN5-dependent acetyltransferase activities, and we show that these enzymes can acetylate both histones and Sin1p. We propose a model in which the ADA/GCN5 and SWI/SNF complexes facilitate activator function by acting in concert to disrupt or modify chromatin structure.