We are interested in two different yet related questions:
(1) How is chromatin structure modulated to achieve the precise regulation of nuclear processes such as transcription, DNA replication, recombination and repair?
(2) What are the physical and chemical principles underlying the action of large macromolecular machines?
We have begun addressing these questions by investigating the mechanisms of a fascinating class of molecular machines: ATP-dependent nucleosome remodeling complexes.

Compaction of eukaryotic DNA in the form of chromatin represents a major organizational feat allowing the packaging of approximately 2 meters of DNA into a space of the order of 5 microns. This compaction, however also makes the DNA less accessible for the central nuclear processes of replication, transcription, recombination and repair. DNA accessibility is greatly reduced even in the smallest unit of chromatin, the nucleosome, which consists of roughly 150 bp of DNA wrapped around an octamer of histone proteins. Modulation of chromatin structure is thus a crucial component of the regulation of any nuclear process that deals with DNA.

In the last decade, several categories of multi-subunit complexes have been identified that alter or "remodel" chromatin structure and contribute to making it dynamic in vivo. A major class of remodeling enzymes consists of complexes that expose nucleosomal DNA by using ATP hydrolysis to alter nucleosome conformation. The SWI/SNF family and the ISWIÐfamily are two families of ATP-dependent remodelers that play central roles in regulation of gene expression during normal development. Another major class consists of complexes that site-specifically modify histone residues by adding or removing acetyl, methyl, ubiquityl or phosphoryl groups. It is thought that specific combinations of modifications in one or more histones are recognized by particular regulatory proteins, leading to additional downstream events. This has been described as the "histone code" hypothesis. Several of the subunits of both categories of complexes have been identified as have many of their in vivo targets. The molecular mechanisms by which these complexes catalyze nucleosome remodeling however, are far from understood.

We are using biochemical and biophysical approaches to quantitatively dissect the mechanisms of ATP-dependent chromatin remodeling complexes. Unraveling how these enzymes work is crucial for understanding their biological impact and the levels at which they are regulated. At the same time, these enzymes provide a new class of molecular machines for studying how ATP hydrolysis is coupled to catalyzing a complex structural change. Some specific areas of investigation are outlined below.

Reaction Framework for Remodeling

Most of the biochemical activities of the ~11-subunit human SWI/SNF complex can be performed by the ATPase subunit, BRG1. As a starting point for defining the remodeling process, we have established a minimal reaction framework for BRG1 action on single nucleosomes. We find that BRG1 as well as SWI/SNF generate multiple distinct remodeled states and continuously interconvert them. SWI/SNF action can thus create multiple opportunities for binding of distinct regulatory factors. The reaction framework leads to several questions, which we plan to address:
(1) How many distinct products are generated and how is the distribution of products altered in the presence of other nuclear factors that interact with chromatin?
(2) What is the role of the conserved SWI/SNF subunits? Do any of them directly participate in the remodeling reaction and for example, alter the product distribution, or final structure(s) of the product(s)?
(3) Can we identify additional steps in the remodeling reaction that can shed light on how ATP-hydrolysis is coupled to remodeling?

Nature of the Remodeled Products

The ISWI-family and SWI/SNF family of remodelers have distinct biological roles. To fully comprehend the in vivo consequences these remodelers, we need to know what the remodeled nucleosomes look like. The ISWI-family appears to expose DNA sites by catalyzing sliding of the histone octamer away from the sites. In this case the products have a canonical nucleosome structure but altered nucleosome position. This strategy provides an efficient way to alter and specify nucleosome position in vivo but cannot explain how DNA is exposed in regions of closely spaced nucleosomes that have little room to slide. Our recent work with the SWI/SNF family suggests that these remodelers open up DNA on the surface of the histone octamer, without a requirement for sliding away the octamer. These results provide a biochemical basis to explain the distinct biological roles of the two families. An important unanswered question that we are currently addressing is whether SWI/SNF products have an altered histone octamer structure in addition to the alteration in DNA conformation as this can result in exposure of specific histone regions for covalent modification.

Remodeling of Higher Order Chromatin Structures

In the nucleus, chromatin is organized in arrays of nucleosomes, which fold into several levels of higher order structure. In vivo data suggests that in addition to remodeling individual nucleosomes, SWI/SNF plays an important role in modulating higher order chromatin structures. We plan to use the reaction framework with mononucleosomes as a foundation to tackle the more complex problem of understanding SWI/SNF action on higher order chromatin structures.

Interplay Between ATP-dependent Remodeling and Histone Acetylation

Genetic and biochemical studies have suggested a strong functional link between SWI/SNF and histone acetyl transferase (HAT) complexes such as the SAGA complex in yeast and humans. Remodeling of the nucleosomes by SWI/SNF may increase the accessibility of the histone N-termini for acetylation or deacetylation. This may then facilitate recruitment of activator and repressor proteins that specifically bind acetylated and deacetylated nucleosomes, respectively. Alternatively, SWI/SNF might simply bind more strongly to, or fall off more slowly from nucleosomes having N-termini acetylated at specific positions as suggested by recent biochemical experiments. This may keep the chromatin in an accessible state longer for activator proteins. It is also possible that direct physical interactions between ATP-dependent remodeling proteins and HAT complexes modulate each other's activities. We plan to address these questions by investigating how the in vitro reaction parameters for SWI/SNF mediated remodeling are altered in the presence of HAT complexes and vice-versa.

Selected Publications:
1. Phelan, M.L., Sif, S, Narlikar, G.J., Kingston, R.E., (1999) Mol Cell 3, 247-53. "Reconstitution of a core chromatin remodeling complex from SWI/SNF subunits."
2. Guyon, J.R., Narlikar, G.J., Sullivan, E.K., Kingston R.E. (2001) Mol Cell Biol. 21, 1132-1144. "Stability of a Human SWI-SNF Remodeled Nucleosomal Array."
3. Narlikar, G.J., Phelan, M.L., Kingston, R.E. (2001) Mol Cell 8, 1219-30. "Generation and Interconversion of Multiple Distinct Nucleosomal States as a Mechanism for Catalyzing Chromatin Fluidity."
4. Narlikar, G.J., Fan, H-Y., Kingston R.E. (2002) Cell. 108, 475-87. "Cooperation between complexes that regulate chromatin structure and transcription."