 |
Understanding the Words of Chromatin Remodeling
Summary: Gerald Crabtree's laboratory is studying the interaction between the signaling pathways and genetic circuits regulating embryonic development. To modulate and explore these circuits, his lab is also designing small molecules that rapidly and reversibly activate or inhibit the products of specific genes critical to these circuits, thereby allowing one to discern the precise time during development when a gene is used for a specific purpose.
The DNA of mammalian cells is compacted about 5,000-fold by incorporation into chromatin, yet DNA must remain accessible to mechanisms that replicate and transcribe the genome. Recent studies have revealed that the vertebrate genome contains an extraordinary number of genes dedicated to chromatin regulation. These studies suggest that chromatin remodeling is a far more specific process than had been previously thought. The ATP-dependent chromatin-remodeling enzymes are one large family of chromatin regulators. These enzymes use the energy of ATP hydrolysis to control the formation and resolution of chromatin structures.
To understand the diversity and specificity of action of these chromatin-remodeling mechanisms, we have taken a proteomic approach to ATP-dependent chromatin-remodeling complexes in mammalian cells. Specifically, we have studied the prototypic mSWI/SNF or BAF complexes. Our work has demonstrated that these complexes are combinatorially assembled and has defined the rules of combinatorial assembly: (1) 21 genes encode the 11 subunits of the complex; (2) each position in the complex can be occupied by the product of only one member of the family of genes encoding the subunit; (3) each subunit is essentially nonexchangeable in vitro and the entire complex can only be dissociated by denaturing conditions. These studies suggest that there might be several hundred combinatorially assembled complexes and that they produce biologic specificity, just as letters in an 11-letter word produce meaning, by combinatorics. However, a rigorous genetic and biochemical test of this combinatoric model was needed.
An adequate test might be to genetically "misspell" the words (i.e., by switching subunits) in the complexes and determine if this modifies the biologic meaning of the complexes. Our studies of these complexes in the development of the nervous system provided the setting needed for a test of this combinatorial hypothesis. Proteomic analysis of the complexes in the developing nervous system led to a surprising discovery: Nature had deliberately changed the spelling. The complexes in neural progenitors or stem cells (npBAF complexes) are made of a specific composition, while complexes in neurons after mitotic exit have a second composition (nBAF complexes). The transition between these two complexes is accomplished by the exit of two subunits, BAF53a and BAF45a, and the insertion of two new homologous subunits, BAF53b and BAF45b. This occurs in the context of a specific subunit composition characteristic of neurons at all stages of their life (Figure 1).
Further studies by Julie Lessard and Jiang Wu in our lab led to the realization that both the npBAF and nBAF complexes are essential for the development of the nervous system. The combinatoric model predicts that they are essential for entirely different reasons, and indeed this proved to be the case. The npBAF complex is required for self-renewal of neural progenitors and represses genes such as Patched and others that play more hands-on roles in neural development. In contrast, the nBAF complex is required for the formation of dendritic processes, one of the most distinctive features of the nervous system. Thus the switch in subunit composition (respelling of the word) produced a dramatic change in function.
But what are the consequences of deliberately misspelling the word? Jiang Wu in our lab prepared mice that lacked BAF53b (one of two neural-specific components of nBAF complexes) but contained an extra copy of the BAF53a or BAF53b gene. She found that BAF53b rescues the lethal defect in the bAF53b mutant mice, but the highly related (81 percent identical) BAF53a was not able to rescue the lethal defects (Figure 2). This was true even though BAF53a was incorporated at the correct position in the npBAF complex. Thus misspelling the chromatin-remodeling word has lethal consequences.
Although these studies have provided the first evidence that combinatorial assembly of chromatin-remodeling complexes produces biologic specificity, they leave many questions unanswered. How general is this mechanism? Do all stem cells have specific complexes dedicated to stem cell self-renewal? If so, what is the chromatin-remodeling word for pluripotency and is it different in different types of stem cells? Finally, what are the biochemical mechanisms used by the specific complexes to enact specific biologic pathways of development?
Converting Analog Signals to Digital Outcomes
Different cell types often emerge in a morphogenic gradient. This observation has raised a general question in intracellular signaling: How are the graded (analog) signals converted to distinct (digital) outcomes? Genetics approaches to this question have localized the sensing and distribution points, but it has been difficult to understand the biochemical mechanisms because the decisions are often made in only a few progenitor cells, where purification is not possible. One exception to this is in T lymphocyte development, where the decision of a cell to be positively or negatively selected occurs in a population of several million cells. In the thymus, weak signals generated by an interaction of an effectively rearranged T cell receptor (TCR) on the surface of a T cell with the matching major histocompatibility complex (MHC) protein on the antigen-presenting cell lead to differentiation, while strong signals generated by an effectively rearranged TCR with MHC bound to self-antigens generate an intense signal and cell death. In this way an immune system is made that avoids self-reactivity but can bind and inactivate bacteria, viruses, and foreign tissues.
Previous genetic studies had shown that calcineurin/NFAT as well as Ras/GRP, Erk1/2, SAP1, and Egr1 are all dedicated to positive selection while Bim, which executes cell death, is dedicated to negative selection. Studies by Elena Gallo, Joel Neilson, and Monte Winslow demonstrated that calcineurin and NFATc2 and NFATc3 are essential for Erk activation and that mice lacking calcineurin show about a 20-fold reduction in maximinal Erk activity. This was found to be due to the calcineurin/NFAT-dependent induction of a Raf modulator that is selectively present in developing T cells. Additional studies showed that the calcineurin-deficient phenotype could be rescued with either a constitutively active Raf transgene or a hypersensitive Erk allele, called sevenless. Thus, early calcineurin signaling is essential to set the dynamic range of Erk signaling, allowing the cell to perceive weak positively selecting signals generated by the T cell receptor binding matching MHC molecules. To test this model, Gallo, Neilson, and Winslow produced double-mutant mice lacking both Bim (essential for cell death in response to negatively selecting signals) and calcineurin (essential to sensitize cells to Erk signaling). In the double-mutant mice, self-antigens resulted in positive selection. These studies indicate that converting an analog signal to a discrete outcome involves first resetting the dynamic range of the Erk pathway, such that minute signals can be perceived within a specific developmental window (Figure 3).
Development of Conditional Alleles of Murine Genes by Small-Molecule-Regulated Proximity
To accurately and quantitatively understand signaling pathways and genetic circuits, it is essential to develop means of perturbing them. If a quantitative model is correct, it should predict the results of the perturbation. If a component or a modifier is missing, it should be possible with a specific perturbation and careful modeling to define the nature of a missing modulator(s). To develop such perturbation mechanisms, we have worked with Stuart Schreiber's group (HHMI, Harvard University) to make small molecules that rapidly and reversibly regulate protein-protein interactions. In the past we developed cell-membrane permeable molecules that induce proximity of genetically tagged proteins, allowing them to be rapidly and reversibly activated or inactivated.
Recently, Kryn Stankunas and Henri Bayle in our lab have developed a method to use rapamycin and nontoxic derivatives to reversibly stabilize proteins. The approach relies on a 97–amino acid unstable tag, which is stabilized by a small molecule. Using this approach, we made mice with this tag knocked into the GSK3β and Pax6 loci. Addition of drug stabilizes the protein and rescues the mutant phenotype only for the brief period of drug administration. We synthesized a partially stable molecule (C20MaRAP) that allows analysis of very brief periods during development. Using this approach, Bayle demonstrated that Pax6 is required for distinct intervals for different aspects of central nervous system development. Also, Karen Liu and Joseph Arron in our lab defined precise developmental execution points for specific phenotypes produced by GSK3β function. For example, they found that GSK3β-mutant mice have cleft palates and that GSK3β functions for only a brief period to result in proper palate development. Additional refinements of this approach in our lab and that of Thomas Wandless (Stanford University) are improving the drugs and tags to allow more rapid and efficient control in yeast, flies, worms, and mice. In addition, we are studying gene therapeutic approaches that make use of small-molecule-induced proximity to control transcription.
Last updated June 24, 2008
|
|
|
|
