A Novel Assay to Quantitatively Detect Bacterial Endotoxin by Harnessing PAMP-Triggered Immunity of FRK1-LUC Arabidopsis thaliana

Category: Plant Science
Table: PLANT1802
Experimentation location: Reseach Institution
Regulated Research (Form 1c): No
Project continuation (Form 7): Yes

Display board image not available


The current practice of harvesting Limulus polyphemus (Atlantic horseshoe crab) to use in the Limulus amoebocyte lysate assay for the detection of bacterial endotoxins in medications, medical devices, and contaminated water is ravaging coastal ecosystems. This project develops an Arabidopsis gene expression-based quantitative endotoxin assay in order to minimize the environmental impacts of endotoxin testing.

I harnessed the PAMP-Triggered Immunity (PTI) response of Arabidopsis thaliana to pathogen-associated molecular patterns (PAMPs), also known as microbe-associated molecular patterns (MAMPs), for quantitative determination of endotoxin presence based on induction of the FRK1 promoter. I utilized transgenic FRK1-LUC Arabidopsis thaliana to express luciferase (LUC) upon activation of the flagellin receptor kinase 1 (FRK1) gene by exposure to gram-negative bacteria. I first tested luciferase enzyme control via plate reader, and recorded luminescence produced by varying enzyme quantities. This provided expected luminosity based on the amount of luciferase produced. Next, I infiltrated E. coli ranging from 6*10^5 to 10^3 CFU (colony-forming units)/mL as well as positive control flg22 peptide, into the leaf apoplastic space of FRK1-LUC, WRKY46-LUC, and wild-type plants. I measured luminescence of infiltrated leaf discs from plant samples after adding luciferin substrate to treated tissues to reconstitute functional luciferase.

This showed a direct relationship between infiltrated bacteria concentration and luminescence, a product of the luciferin-luciferase reaction. Using regression analysis of the FRK1-LUC results, the model yields a formula of y = 1518e^0.0196x (R^2 = 0.937), y representing luminescence and x representing endotoxin concentration. Data suggests this assay achieves endotoxin detection specificity down to 18 endotoxin units (EU)/mL, with p < 0.001 and a standard error of 1.76% as compared with bacteria concentration values calculated via optical density measurements. To determine whether these results were caused by LPS or another bacterial PAMP, I used the SeeSAR molecular modeling software to calculate the binding affinities of LPS and flg22 with LORE and FLS2. The results indicated LORE most strongly binds with LPS, indicating the luminesce results were caused by LPS concentration and there was not significant interference from other PAMPs.

The results from this proof-of-concept provide a promising indication that this method can be utilized as well as further investigated to improve detection at lower endotoxin concentrations and to demonstrate greater specificity.


Bertók, L (1998). Bakteriális endotoxinok és ezek hatása [Bacterial endotoxins and their effects]. Orvosi Hetilap volume 139, 1947-53.

Detection of naturally occurring bacterial endotoxins in water samples. European Pharmaceutical Review Issue 6 2014.

Kutschera, Alexander, and Stephanie Ranf. “The Multifaceted Functions of Lipopolysaccharide in Plant-Bacteria Interactions.” Biochimie, Elsevier, 2 Aug. 2018.

Kutschera, Alexander, et al. “Bacterial Medium-Chain 3-Hydroxy Fatty Acid Metabolites Trigger Immunity in Arabidopsis Plants.” Science, American Association for the Advancement of Science, 12 Apr. 2019.

Lee, Alex K.y., et al. “The 3-Hydroxy Fatty Acids as Biomarkers for Quantification and Characterization of Endotoxins and Gram-Negative Bacteria in Atmospheric Aerosols in Hong Kong.” Atmospheric Environment, vol. 38, no. 37, 2004, pp. 6307–6317..

Li, Yanyan, et al. “Differentiation of Bacteria Using Fatty Acid Profiles from Gas Chromatography-Tandem Mass Spectrometry.” Journal of the Science of Food and Agriculture, vol. 90, no. 8, 2010, pp. 1380–1383., doi:10.1002/jsfa.3931.

“Methods of Gene Transfer in Plants.” Methods of Gene Transfer in Plants, Agricultural Biosecurity.

Mürer, Erik H., et al. “Isolation and Studies of the Granules of the Amebocytes Of Limulus Polyphemus, the Horseshoe Crab.” Journal of Cellular Physiology, vol. 86, no. 3, 1975, pp. 533–542..

Owings, Meghan, et al. “Effects of the Biomedical Bleeding Process on the Behavior of the American Horseshoe Crab, Limulus Polyphemus, in Its Natural Habitat.” The Biological Bulletin, vol. 236, no. 3, 15 Apr. 2019, pp. 207–223..

Ranf, S., Gisch, N., Schäffer, M. et al. A lectin S-domain receptor kinase mediates lipopolysaccharide sensing in Arabidopsis thaliana. Nat Immunol 16, 426–433 (2015).

Stromberg, Loreen R., et al. “Detection Methods for Lipopolysaccharides: Past and Present.” IntechOpen, IntechOpen, 12 July 2017.

Tian, Wang, et al. “A Calmodulin-Gated Calcium Channel Links Pathogen Patterns to Plant Immunity.” Nature, vol. 572, no. 7767, 2019, pp. 131–135..

Tricoire, Ludovic, et al. “Calcium Dependence of Aequorin Bioluminescence Dissected by Random Mutagenesis.” PNAS, National Academy of Sciences, 20 June 2006.

Nascimento, Francisco X., et al. “Ethylene and 1-Aminocyclopropane-1-Carboxylate (ACC) in Plant–Bacterial Interactions.” Frontiers in Plant Science, vol. 9, Feb. 2018, doi:10.3389/fpls.2018.00114.

Ding, Zhong Jie, et al. “WRKY46 Functions as a Transcriptional Repressor of ALMT1, Regulating Aluminum-Induced Malate Secretion in Arabidopsis.” The Plant Journal, vol. 76, no. 5, 30 Sept. 2013, pp. 825–835., doi:10.1111/tpj.12337.

Mang, Hyunggon, et al. “Differential Regulation of Two-Tiered Plant Immunity and Sexual Reproduction by ANXUR Receptor-Like Kinases.” The Plant Cell, vol. 29, no. 12, Dec. 2017, pp. 3140–3156., doi:10.1105/tpc.17.00464.

“Atlantic States Marine Fisheries Commission.” Species - Atlantic States Marine Fisheries Commission.

Chesler, Caren. “This Crab's Blood Is the Reason You're Alive.” Popular Mechanics, 28 Aug. 2019.

Ding, Jeak L., and Bow Ho. “A New Era in Pyrogen Testing.” Trends in Biotechnology, vol. 19, no. 8, 2001, pp. 277–281.

“Endotoxin Timeline.” The Horseshoe Crab: Natural History, Anatomy, Conservation and Current Research.

Hardin, Garrett. “The Tragedy of the Commons.” Science, vol. 162, no. 3859, 13 Dec. 1968, pp. 1243–1248.

Krisfalusi-Gannon, Jordan, et al. “The Role of Horseshoe Crabs in the Biomedical Industry and Recent Trends Impacting Species Sustainability.” Frontiers in Marine Science, vol. 5, 5 June 2018.

Niles, Lawrence J., et al. “Effects of Horseshoe Crab Harvest in Delaware Bay on Red Knots: Are Harvest Restrictions Working?” BioScience, vol. 59, no. 2, Feb. 2009, pp. 153–164.

Park, Chul-Yong, et al. “Comparison of the Rabbit Pyrogen Test and Limulus Amoebocyte Lysate (LAL) Assay for Endotoxin in Hepatitis B Vaccines and the Effect of Aluminum Hydroxide.” Biologicals, vol. 33, no. 3, 2005, pp. 145–151.

Additional Project Information

Additional Resources: -- No resources provided --
Project files:
Project files

Research Plan:

Section A: Research Rationale

Arabidopsis thaliana is a member of the Brassicaceae family, the immune systems of which have been characterized. Of particular interest to this project is the response to lipopolysaccharides (LPS) on the outer membrane of gram-negative bacteria. These plant immune defenses produce products such as reactive oxygen species (ROS) and Ca2+, which respond to cellular damage. A lectin S-domain receptor kinase mediates lipopolysaccharide sensing in Arabidopsis thaliana (Ranf S., et al., 2015) initially proposed that LORE (AT1G61380) detected LPS, as a microbe-associated molecular pattern (MAMP). However, in The multifaceted functions of lipopolysaccharide in plant-bacteria interactions (Kutschera, A. and Ranf, S., 2019), they refine their conclusion to state that the 3-OH fatty acid in LPS is being detected. They state that 3-OH FA/LPS presence induces pathogen-related (PR) genes in Arabidopsis, by LORE or FRK1, which responds to molecules on gram-negative bacteria. Distribution of 3-Hydroxy Fatty Acids in Tissues after Intraperitoneal Injection of Endotoxin (Szponar, B., et al., 2003) confirms that 3-OH FA is an integral component of the endotoxin LPS, and therefore the new studies published do not conflict with the project’s premises. 

This research intends to solve the multifaceted problem of endotoxin testing in medical devices and drugs. Currently, the limulus amoebocyte lysate (LAL) assay is the only widely-approved test. However, the main component of this is the blood from L. polyphemus, a keystone species in the coastal Atlantic ecosystem. Overfishing for medical use has severely impacted their population and threatened the ecosystem (Owings, M., et al., 2019). In addition, blood prices average $60,000 per gallon, making the assay expensive. Harnessing the response of A. thaliana to endotoxin is a viable alternative. Ascertaining modified gene activation with a luciferase assay is a similarly efficient method to observing coagulation of L. polyphemus blood when it contacts LPS. Moreover, costs are reduced, considering the FRK1-LUC transgenic variant can cheaply be made ubiquitous.

This project is vital for performing the lifesaving role of testing for endotoxins on medical devices and drugs. While the RfC assay, using the Recombinant factor C protein from L. polyphemus, has been developed, it is not widely approved. Therefore, this assay using A. thaliana promises to be the solution. It can provide specificity and sensitivity to LPS while detecting in a cost-effective and environmentally-friendly manner. Moreover, it is easy because endotoxin presence is quantifiable based on luminescence intensity, correlated with level of FRK1 gene expression. 

Section B: Research Question(s), Hypothesis(es), Engineering Goal(s)

One of the main questions in this project is what is the maximum sensitivity and specificity of Arabidopsis thaliana to endotoxin. While the literature referenced above, such as The multifaceted functions of lipopolysaccharide in plant-bacteria interactions (Kutschera, A. and Ranf, S., 2019) provide a mostly qualitative analysis of LPS detection by Arabidopsis with a primary focus on the genetic components involved, current studies have not ascertained the limits of detection at low concentrations. I hypothesize that detection levels of 0.01EU/mL, similar to the LAL assay, can be achieved in Arabidopsis, and that all gram-negative bacteria, possessing the LPS pathogen-associated molecular pattern (PAMP) will induce expression of the FRK1-LUC gene based on concentration. Higher luminescence hence should indicate higher bacteria/LPS concentration.

Another goal of this project is to determine how to achieve maximum sensitivity in Arabidopsis. Current literature has encountered obstacles with detection due to the presence of the relatively impermeable cell-wall, as well as the comparatively high threshold of activation for the defense system in the first level of pathogen-triggered immunity (PTI). I hypothesize that by inducing effector-triggered immunity (ETI) and selectively favoring this process, sensitivity will increase. Moreover, I predict that pressure infiltration of the sample into the apoplastic space should overcome hindrances to sensitivity caused by the cell wall.

Section C: Experimental Methods

1. Procedures

Part 1: Creating Transgenic Plant

Transgenic FRK1-LUC Arabidopsis thaliana will be obtained from Dr. Ping He at Texas A&M University, which were created to study differences in PTI and ETI responses. This was due to time constraints of inserting a gene via Agrobacterium vector and selectively reproducing multiple generations. (**Work done by others**)

Part 2: Control Experiment

Luminometer will be tested and calibrated using samples of 1mg luciferase in 200μL of buffer. 50 μL of this solution will be added to 0 progressing in intervals of 0.5 up to 5μL of 5mg/μL luciferin. Luminescence will be quantified by the luminometer, which will then be used to create a standard curve for calibration purposes and to compare data obtained from different intensities of Arabidopsis luminescence.

Part 3: Bacteria Cultures

Escherichia coli DH-5α and Pseudomonas syringae will be cultured to later be used for testing of gene activation and resulting luminescence. 5μL of bacteria will be incubated in 5mL of LB overnight at 37 degrees C. 1mL of the sample will then be drawn out and serial diluted up to 10-5 of initial concentration. Intermediate dilutions may be made based on OD values at λ = 600nm. Each dilution will then be placed into a cuvette and tested for absorbance. After this, 20-40μL of each dilution (based on concentration) will be spread and grown on a Petri dish under aseptic conditions. Colonies will be counted after incubating overnight. From this, an OD vs CFU curve can be generated, which will be used for later quantification of bacterial concentration based on plant luminescence. 

Part 4: Assay Testing

Three genetic backgrounds of plants will be tested: Wild-type Col-0 (control), FRK1-LUC (gene induction during PTI), and WRKY46-LUC (gene induction during ETI). Different concentrations of gram-negative bacteria will be injected into the leaf apoplastic space using pressure infiltration, on flowering-stage A. thaliana plants. Samples will range from 108 CFU/mL (175 EU/mL) through 103 CFU/mL (~0 EU/mL). Luciferin + DMSO solution will be sprayed onto leaves to serve as the substrate to produce luminescence based on gene induction, and ATP is supplied by the plant cells. A luminometer will be used to test replicates of each concentration on plants with each genetic background to create a reliable indicator of luminescence based on sample bacterial concentration. This will be repeated with isolated endotoxin/LPS to determine if there are any differences in results. Decreasing sample concentrations up to a luminescence of 0 will be tested to demonstrate sensitivity. A standard curve for the assay will be ascertained based on the averages of luminescence at each sample concentration in each genetic background. Controls will be run with pure LB and dH2O. 

Part 5: Refining Assay

Next, the PTI response with FRK1-LUC, and ETI response with WRKY46-LUC as a negative control will be tested. Seeds will be obtained from the more sensitive plants and selected for more effective use. 

Part 6: Computational Analysis

Finally, in silico protein modeling and molecular binding interactions will be studied and quantified in order to confirm the efficacy of the aforementioned methods.The SeeSAR software will be used to quantify the strength of binding (affinity), torsion, and intramolecular and intermolecular clashes. The interactions between LPS and the LORE protein, as well as flg22 and the FLS2 protein will be studied in order to determine whether bacterial flagellin binding alters the relative results of LPS concentration-based luminescence in FRK1-LUC Arabidopsis.

2. Risk and Safety

One risk with this project is that I will be working with LPS-producing gram-negative bacteria, which could be pathogenic in some instances. However, the strain being used is non-pathogenic to human for precautionary reasons. This is not expected to hinder LPS detection since this molecule is present on all gram-negative bacteria. These are necessary for my project because I need them to test the assay I have developed. In order to minimize risk, Class 1 hood will be used during all procedures with these bacteria. Bacteria will be stored in a -80 degree Celsius environment in which they will be inactive. Gloves and lab coat will be worn when working with bacteria and proper eyeware will be worn.

In addition, precautions will be taken when working with the reagents. Gloves will be worn when handling luciferase and luciferin substrate.

All lab work will be done under the supervision of the mentor or a lab member to ensure safe lab technique and safe operation of equipment. 

3. Data Analysis 

From control experiment, measurements of luciferin substrate quantities and luciferase volumes will be noted down. Repeated controls will be run for consistency.

OD600 measurements will be taken from the results of absorbance testing for each concentration of bacteria samples. The original suspension will be diluted in factors of 10 to generate different concentrations before taking absorbance measurements. Samples will then be used to determine their Colony Forming Units (CFU) by plating on standard LB (Luria Broth) agar plates. Equal spreading will be verified, and then will be counted for colonies. The number of CFU per mL in the original bacteria suspension will then be deduced. A graph of OD vs CFU/mL graph will then be generated with averaged measurements from replicates and standard deviations.

In assay testing, repeated trials will be done at each sample concentration in each genetic background. From luminometer readings, a graph of luminescence vs. sample concentration based on averages over repetitions will be generated. Statistical significance will be quantified using t-test and regression analysis to test null hypothesis of no control vs. transgenic difference.

Finally, in the computational analysis portion, I will use the visualizations generated by SeeSAR to measure the interactions of each component of LPS with amino acids in the active site of LORE and flg22’s interactions with the amino acids in the active site of FLS2. I will compare the standard deviations and averages of the in silico binding affinity ranges generated by the program to determine the relative strength of binding of each of these bacterial PAMPs.

Questions and Answers


1. What was the major objective of your project and what was your plan to achieve it? 

The objective of my project was to create an alternative to the environmentally damaging, expensive, and/or inaccessible current methods of bacterial endotoxin contamination testing in pharmaceutical products and drinking water. In order to do so, I sought to ascertain and optimize a novel system that could quantitatively detect bacterial contamination in samples at similar sensitivities to the current Limulus amoebocyte lysate (LAL) system, but at a fraction of the cost and without damaging the environment by requiring the harvesting of key organisms in the ecosystem. To do so, I focused on harnessing the response of Arabidopsis thaliana to endotoxin, and determining bacterial endotoxin concentration by measuring gene expression via a luciferin-luciferase luminescence assay.

    a. Was that goal the result of any specific situation, experience, or problem you encountered?  

I set this as my goal after seeing the problem of excessive horseshoe crab harvesting in my local ecosystem on the Jersey shore. I wanted to determine a way to ensure patient and consumer safety through preventing exposure to endotoxin while also preventing the environmental impact of the LAL method. Thus, a combination of factors led me to set out on the endeavor of creating a more environmentally friendly and cost-effective endotoxin assay.

    b. Were you trying to solve a problem, answer a question, or test a hypothesis?

I was trying to solve the problem of environmental damage and high costs of the current endotoxin testing method (Limulus amoebocyte lysate). I created the hypothesis that I could harness the Arabidopsis thaliana immune response to bacterial endotoxin as a means to measure its concentration in a sample, and pursued that line of scientific questioning throughout my research project.

2. What were the major tasks you had to perform in order to complete your project?

I initiated this research project after observing the problem of declining horseshoe crab populations in the coastal ecosystem. I developed all the initial concepts, did the preliminary literature review, and ideation without guidance from a mentor. I began by examining existing endotoxin assay methods and the regulations surrounding endotoxin testing of medicines and medical devices. After determining that the horseshoe crab blood-based assay (Limulus amoebocyte lysate) is the only one in widespread use, I read papers describing the pathogen response systems of organisms from different kingdoms. I centralized on Arabidopsis thaliana due to its ease of growth and well-understood genetics combined with its strong response to pathogen-associated molecular patterns (PAMPs) such as endotoxin. Based on my initial findings, I developed a method to measure calcium ion levels in A. thaliana and determine endotoxin concentration by correlating the concentration of calcium ions with known infiltrated endotoxin concentration. After this method proved insensitive, I pivoted to a gene expression assay approach and devised a method using FRK1-LUC A. thaliana with a luciferase-based gene expression assay. I did not have any external input in developing or initiating the purpose of the research and was free to take the project in any direction.

I designed all of the procedures myself, based on my existing knowledge of plant pathology and Arabidopsis thaliana physiology. My initial calcium ion concentration method emerged from my observation that this ion’s levels are strongly correlated with endotoxin concentration. While the procedures I developed were feasible, the method itself was not specific enough to endotoxin since many factors could have initiated a change in calcium ion concentration. I then designed a new procedure that would be more specific and sensitive to endotoxin by measuring FRK1 gene expression, induced after endotoxin specifically binds to the LORE membrane protein. While I knew this gene could also be induced by other PAMPs such as the flg22 peptide from bacterial flagellin (potential false positive), I knew that non-endotoxin inducers would still indicate bacterial presence and hence endotoxin presence, so I determined this was still a viable method. While I designed all of the experimental procedures, my research mentor reviewed them primarily for safety purposes and feasibility. He held weekly meetings to review my plans, consult existing literature, and provide suggestions and critiques on new procedural ideas I developed.

I implemented almost all of the experimental procedures in my project. The only step I didn’t do was creating the FRK1-LUC transgenic variant due to age restrictions. I connected with a Texas A&M University professor who provided seeds of this genotype and WRKY46-LUC (for comparison). However, before implementation, my mentor made sure to teach me safety steps that I needed to take to use equipment and initial standard laboratory procedures such as aseptic technique and properly cultivating Arabidopsis thaliana in the growth chamber. When I was measuring bacteria concentrations in diluted culture as verification to my Arabidopsis-based assay, my mentor taught me how to use the spectrophotometer to measure bacteria concentration using light absorbance. In the final stages of my assay development, he also taught me how to use the plate reader to obtain luminescence measurements of my Arabidopsis leaf disc samples infiltrated with E. coli. I also grew eight generations of FRK1-LUC A. thaliana throughout my project, selected for individuals most sensitive to endotoxin, prepared bacteria samples, infiltrated samples into leaves, and implemented my designed protocol to carry out the enzymatic reactions necessary to determine the level of gene expression.

I did all of the data gathering independently. After running experiments, I recorded my data and qualitative observations, both electronically and on paper. Since I conducted all of the experiments myself, I also gathered all of the data from these experiments. For experiments that took several hours or needed to be run overnight, I began the experiment in the evening, and my mentor took pictures of the results in the morning, so I could then record them. Occasionally, my mentor reviewed my data to make sure that I was moving on the right track and that there were no machine-generated anomalies that I didn’t catch immediately after experimentation.

To conduct all of the data analysis independently, I self-studied statistical techniques relevant to my project and biological research in general before beginning experimentation. After gathering the data from each of my experiments, I used computational methods to analyze them. I primarily used Google Sheets and Microsoft Excel to generate graphs of my data, such as the relationship between observed luminescence and infiltrated endotoxin concentration. For statistical analysis such as standard error and p-value calculations, I used a calculator and obtained results manually, double-checking with spreadsheet equations. After I did the statistical analysis for each part of my project, my mentor looked at an overview of them to make sure I was analyzing the data correctly.

I formulated all of the conclusions of this project after examining the data and my statistical analysis. After completing experimentation on each stage of my research plan, I evaluated what the data from that stage signified and then determined whether to proceed as planned or reevaluate my initial framework. In doing so, I revised my plan several times. By the end of my project, I concluded that FRK1-LUC A. thaliana is, in fact, a viable potential replacement for the horseshoe crab blood-based LAL assay. My mentor verified my conclusions to make sure they correctly connected back to the data and that my statistical analysis supported my conclusions.

a. For teams, describe what each member worked on.

3. What is new or novel about your project?

My project develops a novel bacterial endotoxin contamination test that is substantially more beneficial than the current method due to preventing environmental damage, being significantly more cost-effective, and having similar sensitivity levels to existing assays. The method of harnessing the endotoxin immune response of Arabidopsis thaliana has not been explored before, and my project makes headway both in understanding this system as well as harnessing and optimizing it for quantitative endotoxin testing in drinking water and pharmaceutical products such as vaccines.

    a. Is there some aspect of your project's objective, or how you achieved it that you haven't done before?

My project introduced me to the field of genetic engineering and viewing biological systems as a platform for molecular-scale modification and optimization. I had previously studied biology for years, but my project’s objective of engineering Arabidopsis thaliana to quantify endotoxin concentration in a sample based on luminescence brought me into the new frontier of bioengineering.

    b. Is your project's objective, or the way you implemented it, different from anything you have seen?

I have never seen scientists using the plant response to pathogens as a means of detecting bacterial contamination in samples. As one of the leading causes of sickness and death, bacterial contamination tests are extremely important, yet all have their own drawbacks, from environmental damage to expense to being inaccessible for people to run tests in their own homes. The way I implemented my bacterial endotoxin contamination test essentially allows people to grow their own testing platforms, these genetically engineered Arabidopsis thaliana plants.  

    c. If you believe your work to be unique in some way, what research have you done to confirm that it is?

First, I did extensive reading of literature such as scientific papers in the field of bacterial contamination testing. In all the research I read, none found an effective solution that met the three conditions of environmental-friendliness, low cost, and accessibility. I additionally investigated into the clinical practices of hospitals and pharmaceutical companies, and the availability of amateur testing kits, and found that none of these satisfy all these three conditions. Moreover, recent water-quality related disasters such as the Legionnaires disease outbreak during the Flint, Michigan water crisis confirmed the need for a novel bacterial contamination test that meets these conditions. Thus, I confirmed the uniqueness and novel nature of my research through reading past literature and assessing currently available testing solutions.

4. What was the most challenging part of completing your project?

The most challenging part of completing my project was determining an accurate and reliable way of assessing FRK1 gene expression of Arabidopsis thaliana as a means of measuring endotoxin concentration in a sample. While there are several precise methods of doing so within a laboratory setting, very few are both sufficiently accurate and effective while being feasible to implement in a testing protocol that anyone from ordinary people to pharmaceutical companies can use. After deciding to pursue the luciferin-luciferase reaction method, the experimental testing and design procedures were also difficult, making sure to maintain consistency between tests of different endotoxin concentrations and improving sensitivity in each subsequent generation of Arabidopsis.

   a. What problems did you encounter, and how did you overcome them?

Another problem I encountered was determining a way to measure the Arabidopsis immune response to bacterial endotoxins at all. My initial approach was to measure the metabolites generated by this response, such as the influx of Ca2+ ions that occurs after bacterial exposure to the plant cells. However, these methods were unreliable due to the production of these metabolites from other processes as well. I overcame these by assessing more precise ways to measure the immune response, and ultimately settled upon the very accurate genetic approach since there is a clearly delineated pathway of protein cascade and gene induction that occurs after the endotoxin molecule binds to a cell membrane receptor. I also encountered the issue of improving sensitivity in my system since Arabidopsis inherently has a lower sensitivity to bacterial endotoxin than the Limulus proteins used in LAL. To do so, I had to use artificial selection in order to improve sensitivity over several generations of Arabidopsis in the lab, which I was fortunately able to do.

   b. What did you learn from overcoming these problems?

Overcoming these problems was both stimulating and invigorating, since pursuing this line of innovative research allowed me to integrate my interests in biology, chemistry, and physics to create a novel innovation that has not been developed before. I learned the power of thinking through problems in interdisciplinary ways and stretching the boundaries to engineer my effective system. For instance, if I had restricted myself to only focusing on the metabolite production approach, I would never have developed this sensitive and effective system. This project and its implications truly stretch through a vast range of fields, from the minutiae of biophysics at the molecular level with the endotoxin-receptor interactions all the way to the environmental benefits and ecosystem improvements that the use of my system over the current LAL test brings. This project thus taught me that I can combine my interdisciplinary interests in the life and physical sciences to create novel innovations that solve seemingly unsolvable problems.

5. If you were going to do this project again, are there any things you would you do differently the next time?

One thing I would do differently were I to repeat this project would be to be less hesitant of switching approaches. As aforementioned, after my initial approach of detecting metabolites produced from the immune response did not function, I initially thought of not pursuing the project any further. Only by switching to a genetic approach was I able to create a reliable and sensitive system.

6. Did working on this project give you any ideas for other projects? 

Working on this project gave me several ideas for improving it with further research. In order to improve the sensitivity of my genetically engineered Arabidopsis plants, I plan to upregulate the expression of the LORE receptor proteins that bind with LPS. This way, the same level of gene expression will be induced by a lower endotoxin concentration. To improve specificity and prevent interference from any other pathogen-associated molecular patterns like flagellin, I plan to also engineer the plants to downregulate expression of other cell membrane receptors like FLS2. Finally, I will also determine the effectiveness of making the plant constitutively express luciferin as well, so that the user does not need to add luciferin themselves, and sample application to the plant leaf will generate a quantitative luminescence response in a one-step, “automated” process.

7. How did COVID-19 affect the completion of your project?

COVID-19 affected the completion of my project mainly because I was not able to explore and further develop my system to be even more effective with the methods outlined above. The laboratory I was working in became restricted only to university faculty or postdoctoral researchers, meaning I could not conduct any laboratory work after March 2020. Additionally, while I tested my system using hundreds of samples over eight generations of Arabidopsis thaliana, having the extra months to work on my project would have allowed me to decrease the standard error even further down from 1.76% through testing with more samples and employing artificial selection for a few more generations of plants. In lieu of these, I turned my focus to using computational methods to verify the efficacy of my system. Specifically, I used the SeeSAR software to assess the binding affinities of different pathogen-associated molecular patterns (PAMPs) from bacteria, including LPS, with the different Arabidopsis cell membrane receptor proteins. My goal with this was to assess if there was any significant interference from these other molecules and to confirm whether LPS concentration itself was the cause of the relationship between luminescence and bacteria concentration I measured in my laboratory experiments.