PFAS Detection

Research plan summary

PFAS (per and poly fluorinated alkyl substances) are an increasingly prevalent threat to humans and animals. PFAS are industrial products that are extremely persistent, resilient to degradation, and widespread in the environment, especially in the United States and our home state of Kentucky. PFAS has been responsible for a variety of human diseases, such as cancers, birth defects, autoimmune responses, and many more characterized and potentially uncharacterized impacts. PFAS testing is also very inaccessible, requiring mass spectroscopy machines that not many places have. Our project combats this problem by creating a gene circuit that causes transformed bacteria to fluoresce in the presence of PFAS, providing an accessible in situ test for PFAS. Additionally, our team also sought to gain a deeper understanding of the methods available to detect PFAS through literature searches of the molecular mechanisms of PFAS toxicity, reverse screening of proteins affected by PFAS, kinetic modeling of our gene circuit, and molecular dynamics modeling of PFAS interactions with potential signal proteins. Our team also had to overcome challenges related to the size and complexity of our gene circuit, problems associated with being a first year team, and delays in lab proceedings. Although we were unable to finish lab testing, we successfully completed a kinetic simulation of our gene circuit which can be viewed or modified by anyone else freely. We also gained a deeper understanding of the molecular mechanisms of PFAS through computational simulations and literature searches, the results of which are published on this wiki.

Dubbed per and polyfluorinated alkyl substances, PFAS are an issue in our environment. These chemicals were first discovered during experimentation with refrigerator gases and soon became commonplace in every citizen’s life. DuPont chemical company became a behemoth with the help of PFOA purchases from 3M to make their revolutionary nonstick spray, Teflon. Our group began by talking to prospective PIs at our local university (University of Louisville). We saw that one of them had researched PFAS detection in wastewater in association with graduate students. This research inspired us to pursue a PFAS-related topic, and we wanted to address the root of the issue in the United States: the ubiquity of the chemical across all spheres of life. From firefighters using aerosol cans to the typical, household, nonstick cooking spray, PFAS is everywhere. The issue with this chemical lies in its fundamental structure. PFAS chemicals contain carbon-fluorine chains, which are strong bonds. These bonds hold the chemical together despite environmental intervention, causing the chemical to be resistant to degradation. When ingested through the residue from the nonstick cooking spray on food, for example, the chemical bioaccumulates, which is increasingly problematic. The bioaccumulation of this chemical can lead to the effects listed above in the abstract and is prevalent at a community level in West Virginia. In this state, PFAS waste was dumped into local waterways, leading to congenital disabilities in children. With the DuPont factory being one of the largest employers in the area, many employees were also affected by chronic diseases like diabetes. Our group was shocked to see disparate access barriers in PFAS testing kits. After reading the story the movie “Dark Waters” was based on, we were appalled by the loss of livestock for farmers within seconds due to polluted waterways. Although a first-year team, we set out on an ambitious goal to model a mass detection system for PFAS using genetic engineering, which would create equitable access to clean water. Even though we live in Louisville and are fortunate enough to have a company as great as the Louisville Water Company filter our water, we wanted to make an impact across all communities that might not be as fortunate as ours and ensure that clean water was available to those in future generations to come.

The USAFA 2019 team used a PFAS-sensitive promoter to upregulate the transcription of mRFP mRNA. Our PFAS-detecting gene construct uses the AHL quorum sensing molecule as a way to amplify the signal produced by the PFAS-sensitive prmA-promoter. AHL first needs to bind with a protein called LuxR before it can influence transcription of the pLux promoter. This amplification scheme was attempted by the Stockholm 2020 team and while they tested individual components of the system, they never tested it all together. Our gene circuit will contain LuxR under constitutive transcription, LuxI (an AHL synthase) under the prmA promoter, and GFP under pLux promoter. This way, PFAS should increase the level of AHL in the cell which then binds to the abundant LuxR which finally induces large scale transcription of GFP. Each LuxI mRNA molecule transcribed as a result of PFAS detection will eventually have a larger impact on cell fluorescence than if 1 molecule of GFP mRNA was produced instead of LuxI mRNA.

We also created an alternative circuit with the pLac promoter instead of prmA promoter to act as a way to check if the prmA promoter was the failure point in the circuit, however we did not have enough time to test it.

Since the mechanism of prmA-promoter activation in response to PFAS has not been explicitly described, we conducted a literature search to propose possible mechanisms. We also conducted reverse screening of proteins affected by PFAS since PFAS toxicity has not been fully characterized yet and to find potential receptors that could be sensitive to PFAS for future works.

We created a model in Virtual Cell (VCell) to simulate the kinetics of our gene circuit. With this model we could predict the detection threshold for PFAS as well as provide a tool to future teams that wish to work with PFAS detection or the AHL-LuxR transcription control network.

We conducted literature searches for the rate constants used in each step of the model and estimated uncharacterized reaction constants. We also performed sensitivity analysis for certain estimated reaction rates, though sensitivity analysis for all rate constants could be conducted with VCell. Our VCell model is freely available to anyone.

Our initial objective with molecular dynamics simulations was to investigate whether PFAS (Per- and Polyfluoroalkyl Substances) could induce similar conformational changes in receptor proteins as their natural ligands. Molecular dynamics simulations involve the detailed modeling of molecular interactions and motions at the atomic level. In our case, these simulations allowed us to observe how PFAS molecules interact with and influence the structural dynamics of receptor proteins. We chose to simulate a docked complex produced by a server instead of just using the docked complex straight out of the server because of many reasons. OpenMM provides us with a high degree of customization and control over the simulation parameters. This level of flexibility allows us to fine-tune the simulation conditions to closely mimic real-world scenarios and experimental conditions. By adjusting factors such as temperature, pressure, and force field parameters, We really want the simulation to run under physiological conditions, STP. Simulating the docked complex with OpenMM enables us to gain dynamic insights into the behavior of the complex over time. This dynamic perspective is crucial for understanding how the PFAS and receptor proteins interact, evolve, and adapt within the binding site. It provides a more comprehensive view of the binding process compared to a static pre-generated complex. By running our own simulations, we can validate and verify the accuracy and reliability of the binding configuration. This process helps ensure that the docked complex is energetically stable and that the binding interactions are consistent with our expectations and scientific hypotheses. OpenMM allows us to explore a range of simulation scenarios. We can conduct multiple simulations with variations in parameters or starting conditions to assess the robustness and reliability of the predicted binding, and to consider different binding pathways or binding site conformations. Running our simulations provides a means of quality assurance. It allows us to independently verify the results, ensuring that the complex generated by the server is a valid starting point for further analysis. This exploration aimed to shed light on whether PFAS would bind as expected or introduce potential disruptions, potentially unveiling proteins capable of effectively binding to PFAS, thus serving as a fundamental mechanism for detection. To address this inquiry, we chose to simulate the binding of PFOA to the LuxR protein, primarily because of its resemblance to the nonpolar nature of LuxR's natural ligand, AHL. If PFOA could successfully bind to LuxR, it would establish a direct means of detecting PFAS within cells, significantly streamlining our research process and minimizing the need for extensive modifications. However, it's essential to acknowledge that our current findings remain speculative, and ongoing efforts are dedicated to refining our molecular dynamics simulations, ensuring increased accuracy, and generating more comprehensive insights and data.

In addition to the network modeling and molecular dynamic simulations that we did, we also performed some protocols in the lab to ensure that our project functioned from a practical approach. Throughout our lab experience, we ligated the parts of our gene insert together** for expression and created storage vectors for later use**. In addition to testing aspects of our project, we learned basic lab skills, from personal safety to using a NanoDrop and running a chip on a BioAnalyzer. From our lab experience, we were able to identify weaknesses in our project design and improve on them through the use of the engineering cycle and iterations in our project. We were also able to learn crucial techniques that would allow us to better perform more in-depth experiments in the future. From our work in the lab, we were able to create storage vectors ( 3 plates of transformed bacteria each with a different part of our gene insert) that can be Miniprepped in the future for easy access to our insert. We were also able to attempt our adapted ligation procedure for 3 insert parts, and although it failed, we are now able to improve our design and iterate the engineering cycle to attempt putting our insert together.

We were using a method called reverse searching. We reverse-searched many databases using a smile string for PFAS and PFOA. We found targets that bound or were related to PFOA or PFAS which in the long run were very beneficial in working on this project as we could understand what types of proteins are related and we could use them to help build our gene circuit. These databases provided lists of proteins that are related and some of them like androgen receptors were very useful and we related them to our project. We used many databases like superPRED and pharmapper. We were also able to tie these targets to diseases or viruses and what part of the body they originate. The database searching also revealed the functions of these targets and how they could maybe be related to PFAS and PFOA.