Daniel M. Weinreich, Ph.D. - US grants

Theory suggests that the variance-to-mean ratio in the number of nucleotide substitutions which accumulate on replicate lineages in a gene can give good insight into the action of natural selection on the locus. The Dispersion Index (R) is defined as this ratio, but published estimates of R come exclusively from metazoans and have been confounded by a number of factors, weakening the conclusions of these studies. Chief among these factors is the failure of diverging metazoan species to be truly controlled replicates of one another, as they naturally differ historically in many (unknowable) ecological and evolutionary respects. The goal of this project is to measure the Dispersion Index at all 13 protein-coding loci in laboratory populations of the bacteriophage phi6. The bacteriophage's short generation time and high mutation rate will permit the rapid accumulation of substitutions in truly replicate populations. Theory suggests that R may be sensitive to functional constraint and to population size; both of these predictions will be tested. From a clinical perspective, understanding the basis of rate variation is important because RNA viruses are pathogens. If the emergence of a new virus or the escape of one from a vaccine is facilitated by rapid evolution, it would be useful to know what factors, e.g. small or large populations, facilitate this process.

9981497
Rand

This research seeks to understand how genetic recombination and dominance can alter the effects of natural selection on replacement and silent mutations. In general, natural selection acts more strongly on replacement mutations. By sequencing many copies of a gene from each of two closely related species, one can document the amount of variation within, and between, species for silent and replacement mutations (known as the McDonald-Kreitman test). To test the hypothesis that levels of genetic recombination and dominance alter the effectiveness of natural selection, new DNA sequence data will be collected from nuclear genes that lie in regions experiencing little or no recombination in the fruit flies, Drosophila melanogaster and D. simulans. These data will be compared to existing data from regions of high recombination. The issue of dominance will be addressed by obtaining data from low recombination genes on X-chromosome, which is haploid in males, and low recombination genes on the fourth chromosome, which is diploid in both sexes. While studies of DNA variation have been conducted on the X- and 4th chromosomes, no studies have specifically focused on mutations in protein-coding genes in this context.

The genetic contexts provided by the chromosomes of Drosophila mean that the data are very likely to be general for all organisms, including humans. Since mutations in the DNA are the ultimate source of all genetic variation, understanding how genetic factors can alter natural selection's ability to eliminate deleterious mutations and preserve beneficial mutation is of general significance to the preservation of genetic biodiversity and the genetic health of the human population.

Intellectual Merit: Genetics is the study of how biological traits are transmitted from parents to offspring. For over 100 years it has been appreciated that owing to biological complexities, a mutations effect may vary with the genetic background in which it occurs. For example, imagine the following biochemical pathway. Here some compound X is converted by Enzyme 1 (coded by one gene) into compound Y and then Y is converted by Enzyme 2 (coded by another gene) into a third compound Z. In this case, a mutation that inactivates Enzyme 2 also inactivates the entire pathway, even in an organism in which Enzyme 1 is functional. On the other hand, the same mutation would have no effect in an organism in which Enzyme 1 was previously inactivated. Such interactions complicate the question, what does this mutation do since the effects of mutations in these cases are context-dependent? But at the same time, such interactions provide opportunities for experimentation to dissect underlying biological mechanisms. In our simple example, the observation that mutations inactivating Enzyme 2 have no effect when Enzyme 1 has been inactivated implies that Enzyme 2 mechanistically acts downstream of Enzyme 1 in some common pathway.

This project formalizes these intuitive notions using two theoretical approaches, one based on a quantitative model of how single enzymes operate and the other based on a quantitative model of whole-organism metabolism. This work will yield an analytic framework to sort pairs of mutations into those that act by a shared mechanism and those that act by distinct mechanisms. In addition it will provide an estimate of the number of distinct mechanisms influencing a biological trait. This research is extremely timely because recent high-throughput technical innovations in genomics are yielding vast datasets on mutational interactions in several microbial model systems (E. coli, S. cerevisiae and S. pombe), and prospects are good for similar datasets in multicellular model organisms such as D. melanogaster and C. elegans. These experimental innovations thus open the door to far more sophisticated mechanistic analyses. Critically, direct experimental attack on specific mechanistic interactions remains prohibitively expensive, further motivating the present theoretical approach. This work also promises to make contributions at several levels of biological organization, from enzymatics to whole organism reproductive success to ecological and biogeochemical resource fluxes. This follows because the theoretical model of single enzymes can also be applied to whole organisms, and because the model of metabolism can be applied to any network of chemical fluxes.

Broader Impacts. Beyond allowing inferences to be made regarding biological mechanisms, mutational interactions have theoretical implications for a diversity of biological processes, including constraints on adaptation, the evolution of sex and speciation. The PI has active research programs in several of these areas and so this research directly complements his ongoing work. Moreover this project directly supports the training of a graduate student at the interface of mathematics and biology, to develop expertise essential in this genomic and post-genomic era. Spin-off projects are planned to engage a number of undergraduate students in research working in the PIs laboratory each term and during the summer. The PI also has an ongoing commitment to the intellectual engagement of Providence public school students and teachers through an existing NSF-funded GK-12 program. This outreach work addresses current cultural barriers to understanding genetics and the implications of evolutionary thinking in the United States.

DESCRIPTION (provided by applicant): Antibiotic resistance evolution is commonly mediated by missense mutations in a single protein (e.g. antifolate resistance mutations are often in dihydrofolate reductase, quinoline resistance mutations are often in DNA gyrase, rifamycin resistance mutations are often in RNA polymerase). This fact draws attention to the mechanistic details of protein evolution: how do individual mutations influence protein structure and function so as to allow pathogenic microbes to survive in the presence of the drug? b-lactam antibiotics (e.g. penicillin and the cephalosporins) represent 65% of the world antibiotic market [65], and the TEM serine b-lactamases are the chief source of clinical, plasmid mediated b-lactam resistance [64]. The biochemical and biophysical determinants of TEM b-lactamase activity are experimentally accessible [e.g. 2, 7], making this an ideal model system in which to dissect the mechanistic process of protein evolution. The present proposal is to use a novel, geometric model of protein evolution to do just this. This work simultaneously advances the field of evolutionary genetics, which is now beginning to move from statistical tests for evidence of natural selection to questions of molecular mechanism [4]. Recent work has highlighted the fact that to be successful a protein must perform its function (e.g. enzymes must catalyze their chemical reactions) but simultaneously it is under stabilizing selection (i.e. must fall within narrow tolerances) for structural traits such as folding stability, aggregation and degradation potentials and likely others. Moreover, most missense mutations exhibit pleiotropy, meaning that they perturb more than one such trait. These biophysical and biochemical considerations motivate a picture of protein evolution in which adaptation is a succession of mutations, each with net beneficial effect that act by substantially improving some traits while modestly degrading others [5]. This model decomposes protein evolution into its mechanistically most proximal components, and is formally analogous to a geometric model of adaptation first proposed by RA Fisher [6]. In the present proposal, this theory is extended to allow inference into mutational mechanisms of action based on their effect on organismal fitness. This theoretical innovation is important because measuring mutational effects on fitness is often far more practical than is the identification of their biochemical and biophysical mode of action. This theory will then be applied to a novel panel of mutations in TEM b-lactamase, whose effect on drug resistance will be characterized as a proxy for fitness. Finally, capitalizing on the experimental tractability of TEM b- lactamases, predictions made about these b-lactamase mutations using this theory will be validated via the bottom-up, biochemical and biophysical characterization of the mechanism of these same mutations.

This project will provide understanding of the fundamental principles that influence adaptation in changing environments. Evolutionary theory predicts that natural selection in changing environments will favor traits so long as they confer a net benefit over time. This context of adaptation in changing environments is central to understanding many traits, such as the ability to sense and respond to the environment, and drug resistance of pathogenic microorganisms. This research will therefore help to elucidate the evolutionary processes underlying medically important traits. The researchers will also develop and distribute a lesson plan aimed at teaching the basic tenets of adaptation in changing environments to high school students, thus helping to raise public awareness about a biologically relevant problem.

Biological traits such as phenotypic plasticity and bet-hedging are presumed to evolve by increasing the long-term geometric mean fitness of a lineage. This framework serves as a useful description of the average selective advantage of traits in a variable environment. However, there are known theoretical limitations with this approach since it fails to explicitly capture evolutionary dynamics of evolving populations and thus does not provide a mechanistic description of the evolutionary origin and maintenance of traits such as bet-hedging. Using laboratory yeast populations adapted to rapid periods of environmental stress, the researchers will experimentally measure evolutionary dynamics of competing bet-hedging strategies in a variable environment. The researchers will use a combination of genome sequencing and reverse genetics to construct a panel of fluorescently labeled yeast strains that express differing bet-hedging strategies. By carefully tracking the abundance of competing strains during the course of competition experiments in variable environments, the researchers will directly compare evolutionary changes in the laboratory to the predictions of computational models of adaptation in varying environments. In doing so, the research will generate data necessary to directly test of the geometric mean fitness concept and will experimentally determine the extent of its limitations for describing adaptation to varying environments.

Genetic variation, the raw material for evolution, is ultimately generated by mutation. The rate of mutation is influenced by particular genes, e.g. for DNA replication and repair enzymes, which can themselves be altered by mutations. Some variants of those genes, known as mutators, increase the genomic rate of mutation. Even when mutators have no direct effect on survival and reproduction of the individual, they can experience indirect selection via genetic linkage to beneficial and deleterious mutations that occur elsewhere in the genome. This proposal explores a novel hypothesis that population size may have an effect on indirect selection experienced by mutators. Understanding the role of population size and spatial structure in shaping the evolution of mutation rate may resolve a longstanding empirical puzzle: why mutators frequently emerge in well-mixed microbial laboratory populations, yet are only sporadically found in nature. This project has important implications for understanding both natural and directed genetic changes in basic, health, conservation, and agriculture research. For example, this research may illuminate the role of mutators during disease progression of cancer tumors. The investigators will also work directly with K-12 teachers and students to improve their understanding of genetics and scientific inquiry.

Building on several of their previous publications, the applicants present preliminary calculations and experimental results suggesting that large populations enrich for mutators and small populations inhibit them. In specific aim 1, the applicants propose to extend their calculations by developing computer simulations and mathematical analysis to model the evolution of mutators in complex and realistic scenarios, including spatially structured environments. In specific aim 2 the applicants propose to experimentally test their predictions in the laboratory by competing mutator and non-mutator strains of Saccharomyces cerevisiae yeast. Competitions will be monitored by fluorescence and conducted across a range of population sizes and spatial structures.