My project will create a biochemical batch oscillator and accompanying design framework for tuning oscillator dynamics in complex environments (e.g., period, amplitude, and damping at different temperatures, pH ranges, and/or heterogeneous systems). Our target oscillator combines autocatalytic trypsin generation (positive feedback) with enzyme-mediated inhibitor formation (delayed negative feedback) to achieve oscillators for ~100 hours or more. We will study oscillator dynamics and formalize design rules with kinetic models supported by parameters obtained from in vitro kinetic assays and central databases (e.g., BRENDA). Our experimental analyses will use fluorometric assays to validate our model and assess environmental effects (pH, temperature, and buffer composition), and microfluidic systems to encapsulate oscillators into independently addressable droplets with cell-like functions. It will yield an experimental and theoretical framework for building tunable biochemical oscillators with applications in synthetic biology and materials science.
Abstract of first year rotation work with Drs. Maggie Stanislawski and Luke Evans at CU Anschutuz and CU Boulder: Body mass index (BMI) is a clinical measure routinely used to capture weight status. It can be influenced by environmental and genetic factors. BMI dysregulation, as described by large fluctuations or high values, has been associated with weight gain, obesity, and negative health conditions such as metabolic syndrome, heart disease, and high blood pressure that can follow. Utilizing longitudinal, multi -omic data collected by the Multi-Ethnic Study of Atherosclerosis (MESA), we first our data of over 900 individuals into training (70% of individuals from the main set), and testing (the remaining 30% of individuals) groups. With the testing data, we first employed LASSO regression to select the most significant features of each -omic dataset contributing to change in BMI. We then created a risk score of these significant features to include in linear regression models predicting the change in BMI over 5 years. The testing set was used to evaluate the predictive power of individual -omic layers as well as a combined multi-omic model. Additionally, we explored the contribution of individual proteins and metabolites to BMI regulation, many of which have direct health implications. Ultimately, our models demonstrated the contribution of proteomic, metabolic, and genetic metrics to changes in BMI.
Abstract of first year rotation work with Dr. Xuedong Liu's Lab at CU Boulder: PRDX6 is a prolific protein in cells that has a multitude of roles in the cell. One of its most prevalent roles is its peroxidase activity. In cancerous cells, activity from PRDX6 can hide them from the body’s immune system. Following a live cell imaging assay involving photochemical activation of lipid peroxidation, this study showed that PRDX6 does have an effect on peroxidase which can be quantified. In most species, the only cysteine residue in the PRDX6 sequence is located at the peroxidase site. However, in human PRDX6, there is a cysteine residue at site 91. This raises the question of whether this extra sulfur atom serves a purpose. To determine if C91 is involved in PRDX6’s peroxidase activity, an intact mass spectroscopy experiment was performed via co-incubation of different purified PRDX6 mutants with PX12. PX12 is a thioredoxin inhibitor that modulates the activity of proteins that protect the cell from reactive oxygen species (ROS).
Abstract of first year rotation work with Dr. Karolin Luger's Lab at CU Boulder: Exploring specific inhibitors of the protein complex between PARP1 (Poly(ADP-ribose) polymerase 1) and HPF1 (histone PARylation factor 1) is a promising strategy in cancer therapy and other diseases associated with DNA repair dysregulation. This study utilized an automated liquid handling system to efficiently prepare 384-well plates with diverse compound libraries targeting PARP1/HPF1 interactions. The pipetting robot enabled high-throughput compound dispensing, ensuring precision and reproducibility in the preparation of test solutions. These plates were subsequently used to perform a smear gel electrophoresis to assess the inhibition or modulation of PARP1/HPF1 activity in DNA repair processes compared to activity of PARP1 by itself. This method was used to identify potential lead compounds with selective action against PARP1/HPF1, providing valuable insights into their therapeutic potential. The integration of automation in compound preparation and analysis streamlines the process, allowing for the rapid screening of large compound libraries in drug discovery efforts focused on targeted DNA repair pathways.
Abstract of first year rotation work with Dr. Michael Shirts' Lab at CU Boulder: Poly (ADP-ribose) polymerase 1 (PARP1) is a crucial enzyme involved in DNA repair processes and has become an important therapeutic target in cancer treatment. This study investigates the interactions of PARP1 with two different inhibitors (Olaparib and AZD5303) through molecular dynamics (MD) simulations. I prepared the protein-ligand system for the GROMACS molecular dynamics simulation by preparing the protein and inhibitor structures and adding solvent to the model system. After energy minimization and equilibration, I ran a production run to simulate the system’s dynamics and analyze protein-inhibitor interactions. With this model, I explored the structural dynamics, binding affinity, and stability of the inhibitor-protein complexes over time.
The peptidoglycan layer in bacterial cells is a popular target for antibiotic development. The membrane protein MraY and peripheral membrane protein MurG are part of critical steps in the synthesis of peptidoglycan. Lipid I, a lipid precursor formed by MraY, is recognized by MurG through its soluble domain. Currently, there is no structure of MurG with bound Lipid I, and the residues required for this interaction have not been conclusively defined. Crystallographic methods and Cryo-Electron Microscopy were applied to study the interactions between MurG and the soluble domain of Lipid I by binding Park’s Nucleotide, Lipid II, or a Lipid I analog were used to study the interactions of MurG and MraY with the aforementioned substrates. By adding Park’s Nucleotide, Murgocil, Lipid II, the Lipid I analog, or a combination of the listed additives to concentrated MurG, crystals formed under optimized conditions. We aim to obtain electron-density maps from these techniques to model the structure of MurG.
https://thesis.library.caltech.edu/16214/