Recent advances in understanding chromatin remodeling by Swi/Snf complexes Joseph A Martens and Fred Winston Members of the Swi/Snf family of chromatin-remodeling complexes play critical roles in transcriptional control. Recent studies have made significant advances in our understanding of the fundamental aspects of Swi/Snf complexes, including the roles of specific subunits, the repression of transcription, and the mechanism of remodeling. In addition, new findings also indicate an important role for the Swi/Snf-related complex, RSC, in controlling gene expression. Addresses Department of Genetics, Harvard Medical School, 200 Longwood Avenue, Boston, Massachusetts, 02115, USA Correspondence: Fred Winston; e-mail: winston@rascal.med.harvard.edu Current Opinion in Genetics & Development 2003, 13:136–142 This review comes from a themed issue on Chromosomes and expression mechanisms Edited by John Tamkun and David Stillman 0959-437X/03/$ – see front matter ß 2003 Elsevier Science Ltd. All rights reserved. DOI 10.1016/S0959-437X(03)00022-4 Abbreviations BAF Brg1/Brm-associated factor BAP Brm-associated protein ChIP chromatin immunoprecipitation HDAC histone deacetylase HMG high mobility group MNase micrococcal nuclease RSC remodels the structure of chromatin SAGA Spt–Ada–Gcn5–acetyltransferase SWI/SNF mating type switching/sucrose non-fermenting TBP TATA-binding protein Introduction Chromatin-remodeling complexes carry out key enzy- matic activities, changing chromatin structure by altering DNA–histone contacts within a nucleosome in an ATP- dependent manner. These complexes can be divided into three classes on the basis of the similarities of their ATPase subunits to the Swi2/Snf2, Isw1, and Mi-2 pro- teins. To date, most studies of chromatin-remodeling complexes have focused on their roles in transcriptional control, although there is strong evidence that they also have roles in replication, repair, and recombination. Recent studies of all three classes of ATP-dependent chromatin-remodeling complexes have significantly advanced our understanding of the mechanism by which these complexes remodel nucleosomes and of the biolo- gical significance of this activity. This review focuses on the following fundamental aspects of the Swi/Snf com- plexes from Saccharomyces cerevisiae, Drosophila, and humans: the roles of specific subunits; the repression of transcription; functional genomic analysis of the S. cere- visiae RSC complex; and the mechanism of remodeling. (Additional information on all remodeling complexes can be found in other recent reviews [1–5].) Roles of Swi/Snf subunits in remodeling, recruitment, and chromatin association Swi/Snf complexes comprise nine or more proteins, including both conserved (core) and nonconserved com- ponents (Table 1). Previous studies had shown that Swi2/ Snf2 proteins possess both ATPase and remodeling activ- ities. Little was known about the roles of the other subunits, although genetic analyses in S. cerevisiae had shown that most Swi/Snf and RSC subunits are required in vivo [4,6,7  ]. Recent biochemical and genetic analyses have begun to identify distinct and sometimes multiple roles for different subunits. In addition to Swi2/Snf2 proteins, there is evidence that other core components are required for chromatin-remo- deling activity. Biochemical experiments have shown that two other Swi/Snf members, Swi3/BAF155/BAF170 and Snf5/Ini1, stimulate the remodeling activity of the human Swi2/Snf2 protein Brg1 in vitro [8]. As is consistent with this finding, a yeast snf5 mutant was recently shown to be proficient for Swi/Snf assembly and recruitment to the Swi/Snf-dependent SUC2 promoter, yet was unable to remodel chromatin, thus providing strong evidence that Snf5 is required for remodeling in vivo [9  ]. Another role for a subset of Swi/Snf subunits is in inter- actions with DNA-bound regulatory proteins. Previous studies have established the fact that Swi/Snf complexes are recruited to promoters by DNA-bound activators or repressors [10]. Direct interactions between activators and specific Swi/Snf subunits have been demonstrated ([11]; see also references cited in [12  ]). More recently, two additional human Swi/Snf members, BAF57 [13  ,14  ] and BAF60a [15], have been shown to interact directly with regulatory proteins. For S. cerevisiae Swi/Snf, three subunits, Swi2/Snf2, Snf5, and Swi1, have been shown to interact with a set of activators in vitro [12  ]. Interactions between specific Swi/Snf components and DNA-bound proteins may provide selectivity among related Swi/Snf complexes, as was recently suggested by the demonstration that human Swi/Snf-B, but not human Swi/Snf-A, can stimulate activation by nuclear hormone receptors in vitro [16]. The recruitment of 136 Current Opinion in Genetics & Development 2003, 13:136–142 www.current-opinion.com