David Botstein - US grants

Gene interactions among genes governing protein localization, DNA replication, and the biogenesis of the cell surface will be studied using the entire range of molecular genetic methods now available to attack difficult biological problems. The organisms under study are the bacterium Salmonella typhimurium, its temperate bacteriophage P22, and the lower eukaryotic yeast Saccharomyces cerevisiae. The ideas and methods which will be used include classical genetic analysis of mutants, pseudo-reversion methods for detecting gene interactions, recombinant DNA methods, (particularly highly localized in vitro mutagenesis) and the construction of gene fusions. The value of these studies in health science is through direct understanding of these basic processes and through the development of technology which can be directly applied to specific problems. Secretion of beta-lactamase in Salmonella and invertase in yeast are being analyzed by making mutations specifically in the signal sequences that have been shown to be essential in the protein translocation process. Revertants of such mutants will be analyzed; such revertants may occur intragenically or in genes which are involved in the translocation and processing machinery. The complexes of proteins which must function together to replicate DNA will be studied in both Salmonella and yeast. Pseudoreversion will be used to find suppressor genes whose products interact with enzymes known to be involved in DNA replication. The biosynthesis of the cell wall will be studied by finding the pathway to d-alanine and the basis for O-antigen variation. These studies directly address the health problems of pathogenicity and rational antibiotic design.

Genes which specify cytoskeletal proteins as well as genes involved specifically in the progress of the cell division cycle will be studied in the budding yeast Saccharomyces cerevisiae. Mutations in cytoskeletal proteins (actin and tubulins) will be made in vitro starting with the cloned structural genes. The genes will be mutagenized by highly specific segment-directed mutagenesis methods and the mutagenized gene used to replace the normal gene in a yeast chromosome. Pseudoreversion methods will be used to find genes whose products interact with known cytoskeletal or cell-division-cycle genes. The mutations obtained will be used to characterize the mutant phenotypes by a variety of physiological, morphological (including immunofluorescence and electron microscopy), and biochemical methods. The results of these characterizations can be used to deduce the function of these genes in vivo. Monoclonal antibodies will be produced against two purified proteins: bacterial beta-lactamase from E. coli and yeast hexokinase B. Differences in affinities to the several monoclonals exhibited by mutations in the structural genes for the proteins will be correlated with map position of the mutations, in an attempt to probe domain structure without direct crystallography. Results of these investigations will be applied, if warranted, to studies of the yeast cytoskeletal proteins.

The major goal of this research program is to contribute to the establishment of a set of restriction fragment length polymorphisms, to be used in the construction of a human genetic linkage map. Recent work has shown that standard human DNA libraries in bacteriophage lambda vectors appear systematically to be missing a large fraction of human DNA sequences which can be propagated only on hosts with mutations in the recB, recC, and sbcB genes. The results further suggest a possible relationship between lethality and/or instability of sequences in E. coli and polymorphism in the human genome. The major experimental projects are: 1. To test a variety of new hosts, particularly E. coli strains having mutations in various DNA functions, as well as S. cerevisiae (wild type and mutant), in order to identify cloning hosts with improved tolerance of human DNA sequences. 2. To examine sequences acquired through the use of these new cloning hosts and to determine whether as a class, these otherwise lethal and/or unstable sequences are also sites of polymorphism. 3. To elucidate the molecular organization of the highly polymorphic locus D14S1, and to explore the genome for other loci homologous to D14S1 in order to determine whether these are also sites of polymorphism. 4. As new highly polymorphic loci are uncovered, to determine their chromosomal locations, using in situ hybridization and somatic-cell hybrid. To follow the inheritance of the markers in families and, where possible, to determine linkage among RFLPs or between RFLPs and other inherited traits.

transposon /insertion element; fungal genetics; genome; Saccharomyces cerevisiae; polymerase chain reaction;

The overall goal of the proposed research is to acquire a set of mutations covering the entire surface of conserved proteins already known to have important cellular functions in yeast. After characterization, these mutations will be used to find interacting genes and proteins using a variety of genetic techniques, including isolation and analysis of pseudo-revertants, unlinked non-complementing mutants, and mutations displaying synthetic lethality. We expect to find different genetic interactions with mutations linking different regions of a protein's surface, and it is for this reason that we plan to start with a "synoptic" set of mutations made with a recent modification of a new technology, the charged-to-alanine scan. Genetic studies of mutations in the interacting genes will be supplemented with limited biochemical investigations aimed at characterizing the interactions of normal and mutant gene products. We hope in the end to make the finding of interacting genes and proteins by genetic methods as reliable and routine as mutagenesis has become. These technologies will be developed and applied to several specific projects: First, we will construct and characterize a synoptic set of new tubulin mutants of yeast. Second, we will identify genetically genes encoding ligands to the actin and tubulin cytoskeletons starting with a synoptic set of alanine-scanning mutants covering actin (already in hand) and the tubulins. Third, we will carry out homolog-scanning hybrid mutagenesis between pairs of genes encoding small GTP-binding proteins in yeast (e.g. the products of the YPTI and SEC4 genes) in order to try to identify, in complementation tests using these hybrids, the sub-domains of these highly similar proteins that determine their specificity for a particular function in protein secretion. Fourth, we will construct mutations with scorable non-lethal or conditional-lethal phenotypes within specificity-determining subdomains of the aforementioned GTP-binding proteins and use these to identify genetically genes encoding protein ligands to particular GTP-binding proteins.

We propose to design, implement, and make available (through the National Library of Medicine) a genome database for budding yeast (Saccharomyces cerevisiae). The database will contain and correlate existing information about genes and other genomic objects: position on the genetic map, position on the physical map (especially which of the overlapping set of cloned DNA segments contain the gene), DNA sequence and amino acid sequence of gene products (where known), and supporting bibliographic citations. The database is intended for the working biologist. It will retrieve all relevant pieces of information about a gene or sequence in response to queries giving any of the following: the gene name, the neighborhood on the genetic or physical maps, or, most importantly, a fragment of DNA or amino acid sequence. Our distributed design mill facilitate supporting several types of computers and allow the database to evolve to handle the data for the entire yeast genome. We will use a commercial database system for data management tasks and develop an intuitive graphical interface for displaying genomic data. Standard homology searching tools will be employed. A limited amount of experimental work to improve the degree of correlation will also be necessary.

We propose to develop, using genes specifying essential proteins of the yeast Saccharomyces cerevisiae, a way to study protein structure/function relationships in the living cell. The method, which involves a new generation of mutagenesis strategies, will be applied to yeast proteins that are highly homologous (and functionally interchangeable) with a human counterpart and for which a highresolution X-ray structure is available. We plan to begin this study with the single yeast actin gene (ACT1) and later extend it to one or more members of the small GTP-binding or RAS oncogene superfamily. The method consists of three elements. The first (already accomplished in the case of actin) is a comprehensive "charged-toalanine" mutagenesis-scan of protein surface regions for mutations that allow synthesis and folding but which prevent function as judged by the ability to support the growth of a haploid yeast cell in vivo. The second is analysis by "random-replacement mutagenesis" of the regions indicated by the scan to be essential for function. The third is molecular modeling of a representative sample of the variants in each essential region that are consistent with function. Random replacement mutagenesis and molecular modeling of functional replacement sequences has been applied already to the secreted enzyme TEM-beta-lactamase, a major determinant of plasmid-borne penicillin resistance in bacteria. The information gained from this kind of comprehensive mutagenic study should allow the determination of which parts of each of these proteins contact their ligands. In each case, the new mutations generated should facilitate isolation of interacting genes and proteins by standard genetic and biochemical methods that connect specifically to different-essential parts of the protein surface.

We propose to develop, using genes specifying essential proteins of the yeast Saccharomyces cerevisiae, a way to study protein structure/function relationships in the living cell. The method, which involves a new generation of mutagenesis strategies, will be applied to yeast proteins that are highly homologous (and functionally interchangeable) with a human counterpart and for which a highresolution X-ray structure is available. We plan to begin this study with the single yeast actin gene (ACT1) and later extend it to one or more members of the small GTP-binding or RAS oncogene superfamily. The method consists of three elements. The first (already accomplished in the case of actin) is a comprehensive "charged-toalanine" mutagenesis-scan of protein surface regions for mutations that allow synthesis and folding but which prevent function as judged by the ability to support the growth of a haploid yeast cell in vivo. The second is analysis by "random-replacement mutagenesis" of the regions indicated by the scan to be essential for function. The third is molecular modeling of a representative sample of the variants in each essential region that are consistent with function. Random replacement mutagenesis and molecular modeling of functional replacement sequences has been applied already to the secreted enzyme TEM-beta-lactamase, a major determinant of plasmid-borne penicillin resistance in bacteria. The information gained from this kind of comprehensive mutagenic study should allow the determination of which parts of each of these proteins contact their ligands. In each case, the new mutations generated should facilitate isolation of interacting genes and proteins by standard genetic and biochemical methods that connect specifically to different-essential parts of the protein surface.

DESCRIPTION: (Applicant's Description) The goal of this research is to use a first-generation DNA microarray system developed at Stanford to study gene expression patterns in normal and neoplastic human tissues. Rapidly growing knowledge of human expressed gene sequences will be applied to the problem of cancer by following the expression (at the mRNA level) of 10 to 20 thousand genes at once. Microarray hybridization will be used to study expression in a series of 60 human cancer cell lines already under intensive study by the NCI in connection with drug-susceptibility screening. Experience with these cell lines will aid in the design of experiments dealing with expression in samples of tumors and normal tissues provided by clinical collaborators. Arrangements have been made to obtain a large number of normal tissue samples from the genome anatomy project at NCI. Arrangements have also been made with clinical collaborators for preliminary surveys (ca. 100 samples) of lymphoma, breast cancer, colon cancer and leukemia. Suitably analyzed aggregate gene expression data should provide examples of genes whose expression, individually or as part of a complex pattern, is characteristic of particular types or sub-types of cancer. Information systems for acquisition, analysis and database storage of expression data will he put in place so that correlations of gene expression pattern with growth, physiology, disease state, drug susceptibility and eventually treatment outcome can be made. Mathematical algorithms for manipulating large datasets and testing similarity of expression patterns for many thousands of genes will be adapted or developed as needed.

DESCRIPTION (provided by applicant): The overarching goal of the proposed Center for Quantitative Biology is to instantiate at Princeton a research and teaching environment that fully meets the challenge and opportunity to practice a usefully quantitative biological science. The Center will increase the bandwidth of communication among researchers from different disciplines and departments (including Molecular Biology, Ecology and Evolutionary Biology, Computer Science, Chemical Engineering and Physics), some of whom are already collaborating on quantitative projects and others who are planning to do so. For both undergraduate and graduate students, the Center will provide a focus for a new way of multidisciplinary teaching and learning, where the quantitative and biological are integrated from the beginning. The goal is to provide an education that prepares students with equal facility in biological and quantitative concepts. The specific aims are: (1) to develop realistic and quantitative models of biological processes; (2) to collect large-scale data sets comprehensively describing biological processes.; (3) to devise new and improved methods for computational analysis and display of complex models, structures and data; (4) to institute at Princeton new curricula and course of quantitative biology education for undergraduates, graduate students, and the larger scientific community; and (5) to reduce these ideas to practice in several collaborative and multi-disciplinary projects, each aimed at a specific biological question in the subject areas of (i) spatial patterning during development; (ii) intracellular signaling and transcriptional networks, and (iii) virus-host interactions. All these projects have common quantitative goals, require a common infrastructure (i.e. substantial computation, microarray and imaging core facilities) and will benefit from the intellectual synergy and multi-disciplinary cooperation that the proposed Center will bring. An important aspect of this aim is to make all underlying data publicly available upon publication.

Abstract Not Provided.

Abstract Not Provided.

[unreadable] DESCRIPTION (provided by applicant): Support is requested for a new multi-disciplinary training program in genomics at Princeton University, aimed at educating genome scientists with the quantitative and computational tools likely to be required for the biology of the future. The genome training program will support students in Princeton's new Graduate Program in Quantitative and computational Biology (QCB), a joint undertaking between the new Lewis-Sigler Institute for Integrative Genomics and four Princeton departments (Computer Science, Ecology and Evolutionary Biology, Molecular Biology, and Physics) and administered by the Institute. Genomics trainees will be able to do their thesis research with any of 54 QCB faculty in 10 departments united by common interests in quantitative and computational biology. Five genomics training slots are requested for the first year, 10 for the second, and 15 for each year thereafter, so that a steady-state population of 15 genomics trainees can be supported. In addition to fulfilling the requirements of one of the participating departments, genomics trainees will do research in any of a wide array of functional genomics projects in bacteria, eukaryotic models and mammalian systems; computational projects ranging from basic bio-informatics and molecular evolution to high-throughput data visualization; systems biology projects ranging from microbial metabolism to systems neuroscience; biophysical projects from basic structural biology to the construction of novel proteins and regulatory switches; and theoretical projects ranging from basic dynamical modeling to trying to model signal transduction in epithelia or neurons. Trainees will have an individualized training plan administered by the Executive Committee of the QCB and a Genomics Committee; formal training will include new courses in genomics and genomic analysis, a seminar series, a student-run journal club, and other multi-disciplinary activities centered in the Institute. Trainees will have the opportunity to teach in an innovative new multidisciplinary introductory program for undergraduates at Princeton. Finally, trainees and eligible faculty will participate in a number of activities designed to recruit and teach individuals who are members of under-represented minorities, and to extend the genomics message to high school teachers. [unreadable] [unreadable]

DESCRIPTION (provided by applicant): The explosion of genome-wide research over the last decade has made clear the need for a new paradigm for the way in which biological information is integrated and disseminated to the research community. Thus, the overall goal of this proposal is to devise and implement systems that make possible continuing integration of the information contained within genomic data sets, including, at a minimum, gene expression, 2-hybrid, synthetic lethality, and systematic deletion studies, and to make the integrated information accessible to the research community, initially through the Saccharomyces Genome Database (SGD), and later through other model organism databases as we create a generally applicable system and submit it to the Generic Model Organism Database (GMOD) project. We envision that information from the data sets will become accessible through 2 general routes: first, the data integration system will be a guide to curators in assessing what is known about the function of genes, especially those genes for which little is known; second, we will organize the integrated data, on a continuing basis, into a schema that can be queried directly. By providing multiple sources of information and the tools to integrate and query them in new and better ways, we aim to help biologists progress toward seeing individual genes or pathways of interest in a more global context. Equally important, we hope to add to our knowledge of the biological roles of individual genes from knowledge of their behavior in the increasing variety of different conditions that are coming under study using the tools of functional genomics.

PROJECT 1 - INTRACELLULAR NETWORKS. Project Leader: J. Broach (Molbio) Six research groups in the Center work on distinct examples of intracellular networks. In all cases the approach to modeling a signaling network consists of a joint effort of experimental biologists and mathematically oriented theorists. The dynamic interplay between the experimentalists and the theorists occurs on a daily basis and is made possible by the close physical proximity and the collaborative mind set (see Figure 1) of everyone associated with the Center. In addition, all the groups draw heavily on the core facilities of the Center; microarray, mass spectrometry and imaging facilities provide the experimentalists with the means of acquiring requisite quantitative data and the computational core provides a means of storing and analyzing those data. Two fundamental problems in cellular biology are being addressed: transcriptional networks (i.e. defining and analyzing how cells coordinate their complex transcriptional changes that occur under changing environments, and coordination of cell growth (i.e. how cells manage to maintain balanced growth over a wide range of growth rates and environmental perturbations).

DESCRIPTION (provided by applicant): Support is requested for a renewal of a multi-disciplinary training program in genomics at Princeton, aimed at educating genome scientists with the quantitative and computational tools likely to be required for the biology of the future. The genome training program will support students in Princeton's Graduate Program in Quantitative and Computational Biology (QCB), a joint undertaking between the Lewis-Sigler Institute for Integrative Genomics and five Princeton departments (Computer Science, Ecology and Evolutionary Biology, Molecular Biology, Chemistry and Physics) and administered by the Institute. Genomics trainees will be able to do their thesis research with any of 49 QCB faculty in 10 departments united by common interests in quantitative and computational biology. Trainees will do research in functional genomics in bacteria, eukaryotic models and mammalian systems; computational projects ranging from bio-informatics and molecular evolution to high-throughput data visualization and systems biology projects ranging from microbial metabolism to systems neuroscience; and theoretical projects ranging from basic dynamical modeling to modeling signal transduction in epithelia or neurons. Because the training program, completely new when the training grant was awarded, is still growing, and because it is still one of the few of its kind, we request a modest increase from 9 positions (the current level), ramping up one position per year to a new steady state of 12 slots. Trainees will have an individualized training plan administered by the Executive Committee of the QCB and a Genomics Committee; formal training will include new courses in genomics and genomic analysis, a seminar series, a student-run journal club, and other multi-disciplinary activities centered in the Institute. Trainees will have the opportunity to teach in a new innovative multidisciplinary introductory program for undergraduates at Princeton. Finally, trainees and eligible faculty will participate in a number of activities designed to recruit and teach individuals who are members of under-represented minority groups. PUBLIC HEALTH RELEVANCE: The genome and the computer have revolutionized biomedical science. In order that the medical science of the future takes full advantage of this revolution, a new kind of education for Ph.D. biomedical scientists is required, one in which biology and the more quantitative sciences, especially computation, are given equal weight. The NHGRI training program at Princeton is one of the pioneer programs fulfilling this role.

BIOINFORMATICS: DEVELOPMENT OF COMPUTATIONAL METHODS FOR ROBUST FUNCTIONAL ANALYSIS Project Leader: O. Troyanskaya, (CompSci/LSI) Our Center provides a highly collaborative environment and necessary computing and experimental Infrastructure for the development and global dissemination of bioinformatics methods that help us to answer diverse systems-level questions ranging from genome-scale evolutionary dynamics to functional annotation to pathway modeling. These methods range from addressing nucleotide variation in the genome, to its functional characterization, to inferring regulatory networks. Subproject 1: Global identification of genome sequence variation (Kruglyak and Botstein) Progress Report: Three years ago, the Botsteln and Kruglyak laboratories developed SNPScanner, a method for comprehensively assessing nudeotide-variation on a global scale. The basis of this approach was to compare hybridization data from two yeast strains with diverged genomes: the