Measurement

Overview

Our plan is to use samples from patients that are whole blood where the miRNA’s are circulating. In order to get data from those samples, we run into the problem that blood has an opaque red color, meaning that it’s hard to extract information from it directly without specialized equipment found in labs. Our goal here is to develop a method(s) which could be used anywhere (in particular in homes) to accurately and quantitatively measure the presence of reporters produced by toehold switches within blood.

Project 1: Developing Colorimetry as a detection method for chromoproteins

Research Question: One of the initial goals of this project was to evaluate colorimetry via RGB colors as a method of quantifying color changes. Specifically, we wanted to know if a colorimeter app could quantify changes caused by expression of a chromoprotein as a reporter.

Motivation: We initially wanted to use a chromoprotein as our detection method because this could be seen by people without any other equipment. A colorimeter app that people could access on their phone would allow us to provide a quantitative output from our detection system rather than relying on people being able to see a color change by eye.

Materials: We tested 3 different chromoproteins: aeBlue, G62M amilCP (a mutant created in our lab that is reddish in color), and sfGFP (superfolder GFP). Traditionally, GFP is used only for fluorescence, but it is actually a yellow Chromoprotein in ambient light. We used 5 replicates for each dilution of each Chromoprotein.

Controls and Calibration: We first studied the properties of the CarolinaRGB colorimeter app. This app allows users to hold their phone camera over anything and get Red, Green, Blue, (or RGB) values for it. We chose this app specifically to ensure accessibility, as it’s available across both android devices on google play store, and IOS through the App Store. We discovered a few major considerations: firstly, the values are constantly changing in real time, and so we had to use screenshots to capture specific time points. In our case, we used Apple Siri, and directed it to take the screenshot, meaning the movement of our phones after pressing the buttons wouldn’t affect the reading. Secondly, we saw that when we turned on a light, we saw the values change for about three seconds, but then cameras on all phones adjusted and went back to the old value. However, this wasn’t true for specifically sfGFP, which retained approximately 15 points more on all values when the light was illuminating it. We found that in more opaque samples, background didn’t change the results much, but for more dilute samples background has a dramatic effect, so we used a white piece of printer paper as a background for all measurements. We found that we had to sometimes slightly tilt the tubes onto another surface so that the liquid condensed to the bottom before reading. On the left is what a typical result looks like on the app itself and on the right is a graph of the 3 color dimensions for different chromoproteins as graphed in Google Sheets.

Results: We can see that the chromoproteins show color patterns that are very different from the patterns in blood.

We then graphed the color intensity in each of the channels for R, G, and B and saw significant differences.

We could see this even better by looking at a ratio of the intensity between two different channels. There was a noticeable increase in the green/blue ratio in sfGFP when compared to water, with an average of 2.07 compared to 1.10 for water. This was statistically significant with a p-value of 2.1x10-7in a two-sample t-test.

Protocol

We then wanted to see whether we could see a difference between sfGFP and water in a more dilute sample, so we created dilutions of sfGFP-expressing bacteria (OD600=1.0, 0.5, and 0.25).

Results: We can see that the difference in green/blue ratio decreases as the sample is diluted. But, the difference between all of the samples and water is still significant with p-values of 2.10x107, 0.017, and 3.97x10-6.

We plan to now test this app on samples where sfGFP is experimentally induced over time.

Usefulness to Other Teams: So many iGEM teams use a fluorescent reporter of some kind for their experiments and theoretical products. GFP is one of the most popular fluorescent reporters in biology as a whole. However, generally the only way to get quantitative data from GFP is by using a plate reader, and there are exceptionally few high schools (in particular public schools) which have access to that machinery or something similar. This creates a huge barrier to entry for high school iGEM teams, especially in the United States. Being able to get equivalent information with a simple app is super helpful, and democratizes synthetic biology to a greater number of people. Our data is very consistent between samples, meaning this is a very reliable measure over a range of OD’s. Additionally, this achieves our original goal of creating a miRNA testing method that can be used outside of lab settings. This can work for both our final product to test for TNBC at home, but can also be used by any other team in very similar ways.

Project 2: Quantifying How Blood Affects Reporter Strength

Research Question: How does blood affect reporter signal strength? How well can we detect reporter (i.e. chromoproteins or fluorescent protein) signals within diluted blood?

Motivation: Our final product for TNBC patients utilizes fluorescence for measurement of miRNA’s 21 and 155 circulating in blood. Due to blood’s thickness and opacity, we predict that it would dampen the amount of fluorescent signal when compared with just water. For our system, we will probably need to dilute the blood sample before testing to reduce the color impact. However, this will also dilute the miRNAs that we want to detect. So, we wanted to test the lowest dilution level of blood that when added to a fluorescent sample, the fluorescence was still visible.

Materials: To produce fluorescence, we used the MiniPCR “Biobits” kit. This kit is used in classrooms to illustrate the central Dogma of Molecular Biology. Biobits uses a cell free system to produce fluorescence after addition of 5uL of “DNA A” and 2 of water. The results are analyzed by eye in a portable blue light viewer. We used this kit, but replaced the water traditionally used with different dilutions of blood. For our initial experiments, we decided not to use real blood but instead to use VOETIR-brand fake blood, available on Amazon.

Results: For whole blood and dilutions of 1:1 and 1:3, the addition of blood renders the fluorescence undetectable or very faint. The small quantity of what appears like fluorescence is caused by a reflection of the light coming from below. You can see this by viewing the blood only sample in which no fluorescent material was present. There is slight fluorescence detectable in the 1:3 sample, but it’s too slight to be distinguishable from basal levels in blood. Based on these results, we must use a minimum 1:7 dilution. Importantly, this suggests that we may need to lower the detectable threshold of our measurement target (miRNAs 21 and 155) to a value lower than our target in standard whole blood.

Protocol

More quantitative ways to see the effect of blood on GFP:

Although Biobits is designed to be visualized by eye, we wanted to get more standardized and numerical data so we used 2 different methods.

Method 1: Plate reader for samples with diluted blood

First, we measured fluorescence in a plate reader. We used the Molecular Devices Filtermax plate reader, and we had our excitation wavelength at 488nm and our emission wavelength at 535nm. We see a nice upward trend of fluorescence as we dilute the blood further, which is what we expected. However, there is some variability between samples. We performed a two sample t test, and got a p-value of 0.08 for our most diluted sample. While that’s not statistically significant, it is close to the 0.05 cutoff and it shows that there is a difference between samples with and without blood.

Method 2: Colorimeter app for samples with diluted sfGFP

Motivation: Instead of diluting the blood, we wanted to perform this next test to see if we could still detect GFP when the GFP cells are diluted in blood.

Results: We can dilute the cells significantly (1:1000) and the change seems to be detectable and different from samples with no GFP as evident by a subtle but distinct downward trend as we lower GFP concentration. However, we need to repeat this experiment more times to see if the result is statistically significant.

Protocol

Usefulness to Other Teams: We’ve quantitatively tested how signal is dampened by fake blood, and something major we’ve identified is that Chromoproteins are very sensitive to both dilution and blood. sfGFP, whether used as a Chromoprotein or as a fluorescent protein, is probably one of if not the best option when choosing a reporter although it seems to be the best of the options. Usage as a Chromoprotein may be preferable to a lot of teams in regions or educational levels that don’t have access to a plate reader. Alternatively, they could use fluorescence even though blood also has a major effect on fluorescence - granted to a lesser extent than Chromoprotein.

Project 3: Determining if miRNAs can cross the cell membrane

Research question: How much miRNA makes it into bacterial cells? Does that influence reporter output from the toehold switches?

Motivation: During our literature review, Zhao et al. reported that miRNA-155 was able to cross the cell membrane of E.coli [Zhao L, Zhou T, Chen J, Cai W, Shi R, Duan Y, Yuan L, Xing C. Colon specific delivery of miR-155 inhibitor alleviates estrogen deficiency related phenotype via microbiota remodeling. Drug Deliv. 2022 Dec;29(1):2610-2620. doi: 10.1080/10717544.2022.2108163.], there is no equivalent for miRNA-21. If the miRNA are unable to cross the membrane, we would be unable to use a possible cell-based test. The purpose of the following experiment is to see whether miRNA-21 can cross the cell membrane by using qPCR to see miRNA levels inside cells. But first, we wanted to generate a Ct value curve with known amounts of miRNAs 21 and 155.

Materials and Methods: To quantify miRNA, we tested a new, cost effective, method developed by New England Biolabs called SPLICER (SplintR ligase-mediated ligation of complementary-pairing probes enhanced by RNase H-qPCR)1 which replaces reverse transcription with DNA probes. For each miRNA, two DNA sequences are designed that each have complementarity to the miRNA. When miRNA binds, it brings the DNA sequences together at which point they are ligated with SplintR enzyme. The ligated DNA probe serves as a template for qPCR. The amount of ligated DNA probe available as a template should reflect the amount of miRNA in solution. We generated probes for miRNAs 21 and 155, and used the probes previously designed in the original SPLICER paper for miRNA 159 as a positive control. Using in vitro transcription, we produced each of the corresponding miRNAs and measured the resulting concentrations via nanodrop. This miRNA was used to test how well we could get the SPLICER system to work in our hands. Each sample was run in triplicate. We used a set of 3 dilutions for each miRNA, specifically 1 nM, 1pM, and 1 fM concentrations, reflecting a range of values found in the SPLICER paper and literature on concentrations of miRNA in circulating blood, which range from 0.8 to 107 copies/ng of serum2, 3. We also included a control where no miRNA was added.

Protocol

Results: We were able to observe fluorescence curves in many, but not all samples. Importantly there was no fluorescence detected in samples where no template was added. The program associated with the qPCR machine calculated Ct values for each sample. We observed the Ct values within each miRNA subgroup at their different concentrations including samples where probes were present but no miRNA was added. Trendlines showed a slight trend towards lower CT values at the higher the concentration as expected. This was especially pronounced for miRNA 159.

Exp 2: Given the high CT values from Exp 1, we decided to try higher concentrations to see if we can get any signal differences between the concentrations. We found a paper that uses 10 nM concentrations of miRNA when adding to E. coli cultures4. While this may be higher than concentrations found in blood, it might be a good starting point to validate our system. We re-ran the experiment using concentrations of 1 uM, 100 nM, 10 nM, and 1 nM miRNA to see if CT values would get any lower and the differences between concentrations might be more separable. CT values were lower for some samples than for exp 1 likely reflecting the higher starting concentrations. We also once again noticed a slight trend towards lower CT values as the concentration of miRNA added increased. This was especially true for miR21 whose CT values remained high in exp 1.

Controls and Calibrations: When testing the SPLICER system, we included the probe generated in the original paper to use a system that was previously demonstrated to work. We included two negative controls in our qPCR tests. The first was adding all qPCR reagents, but no probes. Therefore, any fluorescence observed would be due to contamination in the qPCR premix or water used. Secondly, we included controls where we did not add any miRNA to each probe mix. Theoretically, in that case, no probes should have ligated and thus there should be no template. In the future, we would like to add positive controls since we artificially ligated the probes together by ordering the full probe sequence that would occur upon ligation from IDT and use that as a template.

Usefulness for Others: SPLICER is a fairly new technique which is much more cost-effective than other methods like miRCURY and Taqman5. However, it has not been used in many studies so far, perhaps because of how recently it has been described. Therefore, testing out this method can help others decide whether or not it is a good option for their studies.

Next Steps: Now that we have demonstrated the ability of the SPLICER system for our miRNAs and can build standard curves, we can next use this design to ask how much miRNA can enter bacterial cells and how quickly they can do so. Our plan was to add known amounts of our in vitro transcribed miRNA to bacterial cells, allow incubation for differing amounts of time, and then measure the quantity of miRNA found within the cells. We planned to pellet the cells, rinse multiple times with PBS, and then extract RNA using a trizol kit and quantify with SPLICER. Furthermore, we could correlate fluorescent outputs of our toehold switches with the amount of miRNA measured within those cells. We hope to perform these experiments in the future.

Bibliography