During development in most animals, tubes form as precursors to complex organs such as the neural tube, digestive system, and vasculature. To create a tube, cells within a sheet, or epithelium, must coordinate specific shape changes and movements. This coordination requires each cell to establish and maintain directional identity, thereby distinguishing the ‘top’ of the sheet from the ‘bottom’. While extensive research on a group of proteins, called ‘polarity proteins’, has elucidated how cells establish directional identity, little is known about how they maintain that orientation during the shape changes and rearrangements that occur during tube formation. To address this gap in our understanding, I am studying how these polarity proteins contribute to proper tube morphogenesis during the formation of specialized structures on Drosophila melanogaster eggshells called dorsal appendages (DAs). These appendages, which provide the developing embryo with oxygen, are formed from an epithelium that wraps into a tube, elongates, and then fills with eggshell protein. The epithelium sloughs off when the egg is laid, leaving the appendages as a visualization of the earlier tube formation. I used RNA interference (RNAi) to assess the role of 24 candidate proteins in DA formation. My initial results led me to hypothesize that one protein, Crumbs (crb), regulates the tube’s directional elongation. To explore this role, I am studying crb protein localization during tube elongation, assaying DA defects after knocking down expression using RNAi in subsets of cells, and analyzing the distribution of adhesion, motor, and other polarity proteins when crb is completely absent in null clones. These analyses will add to our understanding of the role of polarity proteins in the conserved development of epithelial sheets into tubes.

Genes encoding the RNA portion of the ribosome (rDNA) are present in essentially all eukaryotic genomes as tandem repeated arrays. In humans, rDNA copy number is highly variable and an undervalued potential source of genetic disease. Changes in rDNA copy number can occur through DNA breakage and repair as well as through errors in DNA replication. High transcriptional activity at the rDNA locus poses challenges for replication; all tested eukaryotes have evolved replication fork barriers (RFBs), ensuring that replication machinery does not collide with transcribing RNA polymerases. In yeast, the RFB is a specific sequence to which the protein Fob1 binds, blocking replication forks that converge with transcription. Mutants lacking Fob1 have greatly reduced variation in rDNA copy number. There are currently two models to explain how Fob1 binding to the RFB produces rDNA copy number instability. One model suggests that binding of Fob1 actively recruits DNA break and repair machinery which induces recombination between rDNA repeats. Another model proposes that the stalled replication fork at the RFB is inherently fragile, increasing the likelihood of breakage. To distinguish between these two models, I am generating yeast strains where Fob1 binds to the RFB but does not arrest forks. Using CRISPR/Cas9 gene editing technology, I am reversing the direction of the RFB in each of the 150 rDNA repeats in yeast to prevent replication fork stalling. By confirming the absence of replication fork stalling and determining whether rDNA instability has also been reduced, I can distinguish whether Fob1 binding to the RFB in the absence of fork blocking contributes to rDNA copy number changes. Clarifying involvement of the RFB in rDNA copy number changes will additionally provide insights into the connections between transcriptional activity, replication fork stalling and genome instability.
 

Hybrid vigor or heterosis describes superiority of a hybrid compared to its parents; however, the genetic mechanisms underlying this phenotype remain largely unresolved. One potential mechanism is loss of heterozygosity (LOH), a process where there is loss of one copy of a gene or surrounding chromosomal region. In our previous work, we evolved hybrids between two species of yeast, Saccharomyces cerevisiae and Saccharomyces uvarum - which differ in their temperature preference - by growing them in a chemostat at different temperatures in phosphate limited media for several hundred generations. We repeatedly observed LOH events in these hybrids in response to changes in temperature. Each LOH event incorporated the gene region encoding the Pho84 membrane bound inorganic phosphate transporter protein on chromosome 13, which is important when phosphate is limited in the environment. These repeated LOH events all affect fitness based on environmental temperature; however the events span various lengths with some as short as a few kilobases and others as large as 200 kilobases. Because they are different in length we also know that they include different numbers of genes. To investigate whether these varying LOH lengths may include other genes that affect hybrid fitness, we have used CRISPR/Cas9 to create double strand breaks at specific sites along the S. cerevisiae chromosome 13, resulting in DNA repair using the S. uvarum chromosome as a template and formation of LOH events of different lengths. Our aim is to create a pool of hybrids with varying LOH tracts and let them compete in a phosphate limited environment to assess the relationship between the different LOH regions and fitness. This will allow us to narrow down genes that may be responsible for temperature sensitivity or that contribute to higher hybrid fitness in organisms that are heterozygous or homozygous.

One method of comparing strain fitness is to compete strains head to head; over time, the competitor with a fitness advantage will increase in frequency in the population. To track the frequency of each strain, they must be differentially marked such that their frequencies can be measured by plating the culture and counting colonies of each type - a time and labor-intensive process. One common marker used by our lab and others is using pigment production to produce colonies of different colors. My project is to create a continuous culture monitoring device named a chromostat that uses a color sensor to measure the relative abundance of different colored strains in solution, removing the need for plating and increasing the automation of competition experiments. By comparing the color of the individual yeast strains to the color of them mixed together, the chromostat can calculate, in real time, the relative abundance of each strain in solution and determine which strain is more fit and by how much. I built the chromostat on a raspberry pi minicomputer using an open source Java library, pi4j, to control the attached color sensor. This sensor converts light waves to red, green, and blue color values, which are then converted to frequency values for each strain and displayed to the user. The chromostat is controlled through a text-based interface that operates on the command line and has a variety of functions to modify data acquisition to increase overall accuracy and allow for data analysis. It will be used as part of a high school teaching laboratory in which students conduct evolution experiments and later compete different yeast strains against each other to generate fitness data.

Large and complex regions of the human genome (a complete set of genetic material) are difficult to sequence and characterize due to the inherent biases in shotgun sequencing technologies (by far the most prevalent sequencing technology today) or short read length limitations. These regions have also been shown to associate with many human diseases such as neurodevelopmental disorders. One category of such regions is variable number tandem repeat (consecutive repetitive sequences of variable size, aka VNTRs). In previous studies, Chaisson and colleagues discovered a group of complex regions in the human genome that had variable lengths in the human population using single molecule, real-time (SMRT) sequencing. Unlike older DNA sequencing technologies, SMRT can sequence with high accuracy long stretches of complex genomic regions. We compared these regions between human and non-human primates (chimpanzee, gorilla, and orangutan) in order to detect human-specific DNA sequences. We generated six haplotypes (we each have two copies of the genome, one from mom, one from dad, where each copy is a haplotype) resolved non-human primate genomes. We also developed a customizable, automated, Python program which used a variety of sequence analysis methods such as dot plots (a method to visualize repetitive sequences), average length statistics, repeat motif (pattern) analysis, and gene annotation to characterize VNTR size differences and possible human-specific expansions across different samples. We discovered a group of human ab initio VNTR expansions that were enriched in intronic regions of genes with established roles in human early fetal neurodevelopment. This study elucidated for the first time the haplotype-resolved sequence of the rapidly expanded VNTRs of the human genome, which was a vastly under-ascertained form of genetic variation in the human genome. The results of this study could potentially further our understanding of diseases associated with VNTR expansions.

Recently, microfluidic technologies have been developed to allow higher throughput collection of yeast replicative lifespan data. Adoption of these devices has been limited, in part, due to the high cost of the motorized microscopy instrumentation from mainline manufacturers. Inspired by recent development of open source microscopy hardware and software, we developed minimal-cost hardware attachments to provide long-term focus stabilization for lower-cost microscopes and open-source software to manage concurrent time-lapse image acquisition from multiple microscopes. This functional 3D printed scaffold along with attached electronics has shown to capture accurate yeast lifespans without costing as much as other third party autofocusing system. We hope that this can help spur the wider adoption of microfluidic technologies for the study of yeast replicative aging.

 Antibiotic resistance is a growing threat to public health. This resistance, coupled with a dearth of new antibiotics, makes the study and development of antibiotics of critical importance. Many antibiotics derive from microbial pathways that synthesize complex natural products, and engineering these pathways to produce new antibiotics is an exciting prospect. I aim to use such a strategy to produce variants of the klebsidin lasso peptide, known for its distinct lariat knot structure and its antibiotic activity. Klebsidin is produced by three biosynthetic enzymes that modify and export a ribosomally-synthesized precursor peptide. We have developed an assay that relies on cellular growth to (1) investigate how mutation influences the antibiotic activity and biosynthesis of klebsidin, and to (2) identify variants of klebsidin that overcome a known resistance mutation. When expressed in E. coli, klebsidin inhibits growth of its host cell. I expressed a library of approximately 10,000 klebsidin variants, each within a single cell, and used next-generation DNA sequencing to count the frequency of each variant before and after a growth selection. Functional klebsidin variants should decrease in frequency after selection, whereas null variants should increase in frequency. Using these sequencing data, I generated functional scores for a majority of single amino acid mutations within klebsidin, identifying positions that tolerate mutation and positions for which mutation abolishes bioactivity. To better understand these trends, I am performing mass spectrometry experiments to study the biosynthesis of key variants. In addition, we are interested in how mutagenesis of klebsidin can combat known resistance to lasso peptides. By performing the same growth selection in an E. coli strain resistant to wild type klebsidin, we hope to identify variants that overcome this resistance. Together, these studies will provide a framework for understanding lasso peptide engineering and identifying novel peptide-based antibiotics.

PTEN is a tumor suppressing protein that carries out important cell functions such as inhibiting cell growth and promoting genomic stability. Somatic variants of PTEN can lead to cancer, and PTEN mutational status has shown to be an indicator of patient survival and prognosis. However, it is not clear whether cancer-associated PTEN variants affect cell growth, genome stability, or both. Here, we demonstrate that simple competition assays can quantitatively assess PTEN variants for their effect on these two important cellular functions. Cells expressing cancer-associated PTEN variants tagged to blue fluorescent protein are mixed with cells expressing wild-type (WT) PTEN tagged with a red fluorescent protein. The proportion of blue and red cells are analyzed over several days using flow cytometry. If the variant does not repress cell growth, variant (blue) cells will outcompete their WT (red) counterparts. To modify the competition assay for genome stability assessment, cells are treated with a PI3K inhibitor and the genotoxic chemotherapeutic temozolomide. These drugs isolate the genomic stability function of PTEN by removing its role in cell growth and causing genome instability, respectively. Here, cells harboring variants that cannot repair temozolomide-induced DNA lesions will be outcompeted by their WT counterparts. The assay generates a score that is based on the rate of change of variant populations relative to the WT population to quantitatively define the phenotype. Results can be interpreted to establish a relationship between a PTEN variant and its quantitative effect on the cell growth or genomic stability functionality of PTEN. Furthermore, the growth-based nature of these assays means that in future work they can be adapted to a pooled library format, allowing the simultaneous, quantitative assessment of thousands of PTEN variants. Data from both low-throughput and high-throughput experiments bring clarity to the relationship between specific PTEN functions and patient prognosis.

Meier-Gorlin Syndrome (MGS), a form of human proportionate dwarfism, arises from mutations in proteins needed for chromosome replication, including the origin recognition complex protein Orc4. Yeast cells (S. cerevisiae) with the mutant allele (orc4MGS) have altered origin activity across the genome, but most dramatically, origin activity in the rDNA is abolished. The orc4MGS cells also display secondary phenotypes such as slow growth, temperature sensitivity, and sensitivity to the drugs hydroxyurea and cycloheximide. To deal with the lack of rDNA origin activity yeast with fewer rDNA repeats (about 10) overtake the culture. My research focuses on understanding whether the secondary yeast phenotypes are due to fewer rDNA repeats, or other consequences of mutant Orc4. To explore the distinction, I am using CRISPR/-Cas9, a system for precise gene editing, to replace the origins of replication in the rDNA region with more efficient origins (ARS1 and ARS1^max). CHEF gel electrophoresis provides a reliable way to quantify the copy number of rDNA repeats in my new strains. The copy number of rDNA increased in my mutant strains to about equal, or even above that of the parent strain. With the rDNA copy number of these new strains restored, I am retesting the previously secondary phenotypes of the orc4MGS strain. Testing is ongoing, but the data suggest that the strains I created now display intermediate phenotypes of growth rate, temperature sensitivity, and drug sensitivity. I conclude from the data that fewer rDNA repeats, as well as the mutant Orc4 protein contribute to the phenotypes observed in the original Meier-Gorlin yeast cells. I am currently determining the efficiency of the new origins in the rDNA, and asking how genome wide origin use has changed. With these experiments I hope to gain insights into some of the cellular mechanisms of Meier-Gorlin Syndrome.