The following is archived (~2004) information detailing some of the research activities within the Schreiber laboratory. For information on current research, please visit the Science Overview.
Introduction. Genetics has been a primary contributor to our understanding of biology, especially cancer biology. Both forward and reverse genetics rely upon mutant alleles to gain insights into pathways or processes of interest. Small molecules have also been used to gain insights into biology in ways that are analogous to either forward or reverse genetics. Many of these advances (for example, Carlsson's use of chlorpromazine to explore the dopamine receptor and Borisy's use of colchicine to discover tubulin) have been brought to light on a case-by-case basis. Chemical genetics is a new area of research that aims to use small molecules in a systematic way to explore biology.
Relating Small Molecules and Mutations in the Study of Cellular Protein Function. Determining the cellular function of a protein generally requires a means to alter the function. The most common way of doing so is an indirect one involving the use of mutations in the genes encoding proteins of interest. A complementary and direct approach involves the use of small molecules that alter the function of proteins to which they bind. As many proteins function inside of cells, cell permeable small molecules have been particularly valuable. Such small molecules exist that are capable of either inactivating or activating the function of the proteins.
The majority of mutations used to study protein function inactivate the encoded protein. The most common of these are deletion mutations -- widely used from bacterial to mammalian genetics. Also useful are conditional mutations, where the inactivating effect is only observed under "non-permissive" conditions accessible to the experimenter. An example is a temperature-sensitive folding mutation. Shifting to the non-permissive temperature prevents newly synthesized polypeptide chains from folding and, therefore, from acquiring their cellular function.
There are many small molecules that inactivate the function of their liganded protein target, and they do so in a conditional way - inactivation requires that the molecule be included in the experiment. The most common of these are ones that bind to an enzyme active site, thereby inhibiting a critical enzymatic activity. But there are many examples of small molecules that inactivate proteins by other means. Colchicine, for example, was used not only to discover the protein it binds to, tubulin, but also to investigate its cellular function. Cytochalasin and latrunculin have been used to inactivate, and thereby explore, the function of actin, another cytoskeletal protein. In signal transduction, rapamycin and trapoxin were used to discover and investigate the proteins they inactivate upon binding, FRAP and histone deacetylase (HDAC), respectively. Since these molecules have been used analogously to temperature-sensitive folding mutations, the specificity of the two techniques is of interest. Recent results examining the effects of small molecules and temperature shifts on a complete set of transcripts suggest exquisite specificity in the former, but not latter case.
Mutations that activate the function of encoded proteins are also known. The most famous of these are oncogenic mutations that cause constitutive activity in signaling proteins normally under cell cycle regulation. Conditionally activating mutations are extremely rare. Small molecules that conditionally activate their protein targets, however, represent a significant fraction of known ligands. Taxol and discodermolide are small molecules that bind and, in contrast to colchicine, activate tubulin. These small molecules facilitate tubulin's ability to polymerize into microtubules, a component of the cytoskeletal. Cyclosporin, rapamycin, and FK506 bind to immunophilin proteins and thereby induce them to form complexes with the signaling proteins calcineurin and FRAP. The most famous "activating" small molecules are the steroid homones, which activate transcription following their binding to steroid hormone receptors. More recently, a general way of constructing activating small molecules from inactivating ones has been developed. This involves synthesizing dimeric structures capable of binding two proteins simultaneously ("dimerizers"). By simply causing two proteins to have a proximal relationship in the cell (provided they also have a suitable relative orientation), activation of one of the proteins can ensue. This is due to the increased rate of chemistry (e.g., phosphorylation, proteolysis) that can occur between appropriately selected pairs of proteins when their small molecule-dependent effective molarity is high.
The preceeding examples provide illustrations of the equivalency of small molecules and mutations in the study of cellular protein function. Thus far, with only few exceptions, the small molecules used have been either natural products or their synthetic variants (e.g., synthetic dimers of natural products). To extend the small molecule-based approach, powerful methods of small molecule discovery are being developed. The principles defined by geneticists to identify mutations that illuminate protein function are proving to be of value in the search for new small molecules with similar properties. The geneticist generates large numbers of mutations, chooses from the myriad of methods to prepare the library of mutations, and selects the desired mutations through the use of an effective screen. Likewise, the "chemical geneticist" synthesizes vast numbers of small molecules (using a strategy named "split-and- pool" synthesis), chooses from the myriad of methods to synthesize complex, asymmetric, natural product-like molecules, and selects the desired small molecules through the use of screens compatible with the split-and-pool method of small molecule generation.
Although this approach to small molecule discovery has only recently been tested in the laboratory, it has been used widely in nature to produce the natural product small molecules described above. For example, bacterial geneticists have uncovered the global outline of polyketide synthesis, which leads to polyketide natural products such as rapamycin and FK506. These molecules are synthesized by an iterative sequence involving a Claisen condensation, ketone reduction, dehydration, and enone reduction. The polyketide synthases containing the enzyme modules that perform these functions are encoded in single bacterial operons. These modules appear to have been shuffled throughout evolution by genetic recombination, which split-and-pool synthesis emulates. Other types of gene modifications, such as mutations in the ketoreductase modules, further enhance the structural complexity of the natural polyketide library. Finally, the process of natural selection leads to the existing members of this family of polyketide small molecules. Their frequent use in present day cell biological studies stems from their selection, over millions of years, as protein ligands.
Chemical Genetics in the Schreiber Laboratory. Research in this laboratory results from the melding of synthetic organic chemistry with cell biology. In these studies, cell permeable molecules have been synthesized and used to understand and control signal transduction pathways involved in cell cycle regulation. Part of the goal of the laboratory is to extend chemical genetics as an approach to the study of all proteins, and to discover new cell permeable small molecules. Diversity-oriented organic syntheses of "natural product"-like substances and miniaturization techniques for assaying their intracellular protein-binding properties are being developed in the Howard Hughes Medical Institute laboratory in the Harvard University Department of Chemistry and Chemical Biology and in the Institute of Chemistry and Cell Biology in the Harvard Medical School.
The current interest of the lab in discovering new non-natural "natural product"-like molecules, originated from work that defined the molecular mechanisms of the immunosuppressive agents cyclosporin A, FK506, and rapamycin. The synthetic chemists in the laboratory not only completed total syntheses of the three immunosuppressants, but prepared the reagents used to co-discover (with scientists at Merck) the FKBP family of immunophilin proteins and to characterize them in functional and structural terms. A designed, synthetic small molecule named 506BD was used to show that the immunosuppressants cause a gain in the function of an associated immunophilin following receptor binding. The molecular basis for this gain in function was clarified with the discovery in 1991 by a postdoctoral fellow, Jun Liu, that both FKBP12-FK506 and cyclophilin- CsA bind to and inhibit the protein phosphatase calcineurin. This finding led to the discovery that calcineurin is a key molecule in the T cell receptor signaling pathway that activates resting T cells for the cell cycle.
A Schreiber graduate student, Peter Belshaw, demonstrated a new strategy that permits structural variants of CsA and FK506 to inhibit calcineurin only in targeted tissues or organs in transgenic animals, and therefore to understand this phosphatase's function in these locations. This strategy involves creating new receptor-small molecule pairs using site-directed mutagenesis and synthetic chemistry ("bumps and holes"), and it has been used effectively more recently in the laboratory of Professor Kevan Shokat to explore the functions of kinases.
The FKBP12-rapamycin complex was shown by two graduate students, Eric Brown and Mark Albers, to bind to a previously unrecognized regulator of the G1 phase of the cell cycle now named FRAP. The structural basis for rapamycin's ability to bond two proteins simultaneously was illuminated by x-ray crystallographic studies performed in Professor Jon Clardy's lab at Cornell. A short movie has been prepared to visualize this unusual molecular recognition by a small molecule. Using rapamycin as an equivalent of a loss-of-function temperature-sensitive allele of FRAP, the protein's kinase activity was shown to be necessary for the activation of the p70 S6 kinase. A human checkpoint homolog originally named FRP1 (FRAP-Related Protein) was and recently renamed ATR (ATM-related protein), was identified by Karlene Cimprich, Tae Bum Shin, and Curtis Keith. The lab has begun to shed light on the function of other members of this fascinating family of "PIK-related kinases".
Studies of cell cycle signaling pathways sensitive to natural products led to the discovery of other new signaling proteins (e.g., originally histone deacetylase HDAC1, more recently HDACs 1-6, the target of trapoxin, trichostatin, and depudecin; protein palmitoyl transferase, a target of didemnin), and to the identification of valuable probes of known signaling proteins (microtubles, discodermolide; proteasome, lactacystin), and have revealed that the approach has broad generality. The case of trapoxin provides an illustration of the importance of synthetic chemistry in these studies. A total synthesis of trapoxin by a graduate student, Jack Taunton, was adapted to a synthesis of a tritium-labeled analog and, even more importantly, to an immobilized variant that was used as an affinity reagent. This reagent led to the discovery of human histone deacetylase-1 (HDAC1). This previously unkown protein provides a critical link between two active areas of research - transcriptional activation and chromatin remodeling. The most recent work by a graduate student, Christian Hassig, has demonstrated that gene regulation occurs in cells by the targeting of HDAC1 to specific genes through a DNA-protein complex, and Chris and Jeff Tong have shown that HDAC 1/2 is part of a nucleosome remodeling and deacetylating (NRD) complex. Histone deacetylase resisted molecular characterization for over 30 years after Allfrey and co-workers first demonstrated its existance in crude nuclear extracts. The laboratory's success illustrates how synthetic organic chemistry can be applied to a problem in cell biology.
Research in the laboratory has also demonstrated that chemical approaches to signal transduction can also be used to control signaling pathways. A key insight came with the recognition that small molecule-induced protein dimerization and oligomerization constitute a common means of initiating information transfer, rivaling the role of small molecule-induced allosteric change. In collaboration with Dr. Gerald R. Crabtree and members of his laboratory at the Howard Hughes Medical Institute in Stanford, a method has been devised that permits controlled intracellular dimerization or oligomerization of proteins with cell-permeable, dumbbell-shaped, synthetic small molecules. Like the immunosuppressive natural products that inspired their design, these molecules have two protein-binding surfaces. This approach has been used in the Harvard-Stanford collaboration to activate proliferative and death pathways involving the T cell, PDGF, insulin, and Fas receptors, and to regulate transcription, protein translocation, and protein degradation. Using this approach, it was demonstrated in 1996 that synthetic dimerizers are able to ablate CD4/CD8 double positive thymocytes in a transgenic mouse expressing a rationally designed, conditional allele of the Fas receptor. This work illustrates for the first time the use of small molecules to achieve spatial and temporal control over a specific signaling pathway in an animal. Ligand-regulated activation and termination of cellular pathways has illustrated the importance of proximity and orientation of proteins in biology and that offers new opportunities in research in biology and medicine. The technique also provides an illustration of how the chemical genetic technique can be extended to proteins for which small molecule partners are not known, including (through the use of small molecule dimerizers and conditional alleles) the T cell, Fas, insulin, PDGF, EPO, and TGFB receptors, the SOS, Src, Lck, Raf, and ZAP70 intracellular signaling proteins, and the ATR, p53, HNF-1 nuclear proteins.
Techniques are now being developed to discover small molecules to any target protein, thereby allowing the preparation of direct inhibitors and dimeric activators. In an early study, millions of potential small molecules were synthesized, and the small subset that bind to an SH3 domain of interest were selected. Multidimensional NMR was used by two graduate students, Sibo Feng and Hongtao Yu, to determine the structure of a variety of complexes, even ones with peptide small molecules that bind in opposite orientations. By comparing the structures of different complexes, the origin of SH3-binding preferences became evident, and the structural information obtained in this three-step process has been used to design highly biased libraries of nonpeptide, cell permeable small molecules. These have yielded the first non-peptide small molecules to SH3 proteins. In more recent studies, the laboratory is patterning small molecule discovery efforts after those used in genetics to discover mutations. Split-pool syntheses of complex, natural product-like molecules, using modern asymmetric synthesis, are being developed. Techniques for assaying the binding and cellular properties of these small molecules on a vast scale, including the simultaneous screening of large collections of synthetic molecules against huge collections of proteins (for example, all proteins encoded in human cDNA libraries) are being developed. This latest work is providing a blueprint for synthetic chemists to discover their own unique molecules to explore biology.
Summary. The activities described above are being undertaken to achieve a single goal - to develop a chemical approach to understand and control the cellular function of proteins. In the future, the ability of small molecules to control the function of proteins in cells, tissues, organs, and animals may allow the links between chemistry, biology, and medicine to be more direct.
For information regarding current research activities within the Schreiber lab, see the Science Overview and the Chemical Biology community pages.