Team:Rochester/Results

Loading...
Logo
IGEM Home
Model Logo

Results

Table of Contents

Click on a title to visit

Top
Aptamer Experiments
CRP, Lactoferrin and IL-1β
Characterization
Aptamers to rGO
Testing
TNF-α and IL-6
Future Directions
Stop Animations

Introduction

The goal of our project is to develop a way to continuously monitor the concentration of biomarkers (small molecules that are associated with a disease or condition), specifically to help with the diagnoses of sepsis. We designed a biosensor that uses aptamers (single stranded pieces of DNA that can bind to proteins and similar molecules) to detect the presence of sepsis biomarkers. These aptamers are attached to reduced graphene oxide (rGO), and the conformational change that results from aptamer binding alters the electrical properties of the rGO. By measuring these property changes, we can track the concentration of biomarkers with our sensor. You can read more about the hardware component here Aptasensor hardware. To maximize our efficiency, our wetlab team split into two subteams: one to design, produce, and test our aptamers, and one to engineer Shewanella oneidensis MR-1 to synthesize the rGO for our project. Click the buttons below to learn more about each part of our project.

Aptamer Experiments

Background Information

Once we chose sepsis as the model disease for which we were going to design a device that continuously monitors sweat, we had to choose proper target biomarkers. In collaboration with our modeling team ( Diagnostic Pipeline), we came up with a set of six biomarkers that are present in sweat and together can provide sufficient information regarding the patient’s health condition. The biomarkers can also predict specifically whether or not a bacterial infection is present.

We are creating a device that can detect certain molecules in the sweat of patients. We found six biomarkers that can tell us if a patient has bacteria in their system.

Although biomarkers are not enough to diagnose sepsis alone, researchers mentioned that combination of several biomarkers together would be beneficial for sepsis diagnosis. Furthermore, a real-time analysis for these biomarkers that could reflect patients’ immune status would be useful in a clinical setting15.

Through literature research, we found a set of biomarkers that have shown to be useful in clinical diagnosis of sepsis: human C-reactive protein (CRP), Procalcitonin (PCT), Tumour Necrosis Factor alpha (TNF-α), Interleukin 1 beta (IL-1β) and Interleukin-6 (IL-6)16. TNF-α, IL-1β, and IL-6 are pro-inflammatory cytokines, secreted by immune cells such as macrophages and helper T cells. They play an important role in the innate immune response and thus show elevated levels in sepsis due to the increased inflammatory response. CRP is secreted by the liver in response to IL-6, and thus also increases in sepsis patients17. CRP is useful to monitor postoperative patients who are prone to develop sepsis because CRP levels would not fall down if there was a postoperative infection present17. Although it is found that TNF-α, IL-1β, and IL-6 are not specific for sepsis (as many inflammatory responses could cause these cytokines’ secretion to increase) they could still be crucial to notify the doctors about a possible septic infection17. According to a review paper about sepsis biomarkers19, we have verified that IL-1β, IL-6, IL-8, TNF-α, and CRP have been evaluated as prognostic factors of sepsis19, and we further looked into these biomarkers’ presence in sweat. Currently, blood culture detection of biomarkers has been used as a reliable means for the diagnosis of sepsis, but this assay requires over 24 hours to get a result which is not optimal for sepsis patients who could develop septic shock and encounter life threatening conditions at any minute16. This is why we chose to use sweat instead of other body fluids for a faster, noninvasive and continuous monitoring of these biomarkers. Collecting and transporting sweat in microfluidic channels means that our device can bring enough sweat to the biosensor and detect an electrical change in significantly less time. By doing so, sepsis-related biomarkers are detected in real time with a relatively short lag, and fortunately, all of these biomarkers can be detected in eccrine sweat20 and can be analyzed from the sweat collected in a sweat patch.21,22

Furthermore, it was indicated that combining CRP, PCT and neutrophil CD64 would be more beneficial for sepsis diagnostics than CRP alone.23 Despite the high specificity of PCT for sepsis diagnosis, we decided to not include PCT because PCT and CD64 are undetectable in sweat.18 Meanwhile, we did more literature research to find another biomarker that could provide more insights on possible septic infection. Lactoferrin caught our attention due to the fact that elevated lactoferrin indicates intestinal inflammation, which predicts possible multi-organ failure.24,25 Lactoferrin, a protein involved in innate defense, has been indicated to have a major anti-inflammatory role in maintaining homeostasis.26,27 Since organ failure is a common outcome of severe sepsis infection, lactoferrin has been shown to play a crucial role in indicating an overactive immune response. Lastly, while IL-8 was initially one of our targets due to its anti-inflammatory activity, it was eventually excluded because only an RNA aptamer was found in literature. It is harder to synthesize RNA aptamers in vitro because RNA is prone to hydrolysis, so we would need to use 2’-fluoro-modified dNTPs, which are expensive. This is why we didn’t proceed to detect IL-8. Overall, IL-1β, IL-6, TNF-α, CRP, and lactoferrin were the final biomarkers selected for our project for a more accurate and efficient diagnosis of possible septic infection.

In literature, clear cut-off values for these biomarkers in sweat were unavailable due to the lack of studies using sweat as the sample body fluid. However, the cut-off values of CRP, lactoferrin, and IL6 are available in blood samples.28,29 With the cut-off values, medical professionals can take action as soon as possible when they observe possible indication of sepsis, hence, a model Sweat-Plasma was developed along with our diagnostic device to convert the concentration of biomarkers in sweat to the corresponding value in blood to better characterize sepsis for a more rapid diagnosis.

In hospitals, it would be useful to have a way to constantly monitor biomarker levels in patients. This could help doctors diagnose sepsis, which is an overwhelming immune response to infection that is often fatal.

We found 5 biomarkers that can help doctors diagnose sepsis, and they are called CRP, PCT, TNF-a, IL-1B, and IL-6. They are all involved in the body's immune response to sepsis, so they increase in the body when a patient has sepsis. Right now, these biomarkers can be detected in blood, but that process takes over 24 hours, which is too long because sepsis can kill patients very quickly. Our device uses sweat, which allows us to get results in less time.

Out of the biomarkers we found, we decided not to use PCT because it can't be found in sweat. However, lactoferrin is another biomarker that shows up in high levels in the body when there is swelling in the intestines, which can be a sign of sepsis. We also decided not to look at IL-8 because we would have to use an RNA aptamer instead of DNA, which would be more expensive. Overall, IL-1β, IL-6, TNF-α, CRP, and lactoferrin were the final biomarkers selected for our project.

We weren't able to find studies that provided information on biomarker levels in sweat, but we found this information in blood for CRP, lactoferrin, and IL6. Because of this, we made a model that will convert the concentration of biomarkers in blood to that in sweat. This will help doctors figure out if their patients might have sepsis based on their sweat samples.

In our search for the type of receptor that would work best for point-of-care diagnostic devices, we came across aptamer based diagnostic devices. Nucleic acid aptamers are either DNA or RNA fragments that adopt a secondary structure able to specifically bind a target molecule. They are discovered via Systematic Evolution of Ligands by Exponential Enrichment, which is a recurrent process starting with a large DNA or RNA library, and at each step, only high binding affinity sequences are separated from the library. One example is Silver Enhanced Fluorescence-Activated Cell Sorting, shown in Figure 1. In this process, the target protein or the control/negative proteins (which are used to ensure specificity of the aptamer) are immobilized on polystyrene microspheres and added to the large DNA/RNA initial library, modified with a fluorophore. Silver decahedral nanoparticles conjugated to a fluorophore are also added, and then these NPs will interact with the labelled DNA and clump together. Through fluorescence-associated cell sorting, one can then choose the cells that had DNA clones that tightly bound to the protein.30

We can observe that only aptamers that bind to the cell or protein of interest are selected from the initial library. This is the first round, where enrichment occurs. All other rounds are negative selection, where the protein or cell of interest reacts with silver nanoparticles, and the non-binding DNA sequences are directed to the waste container via an electric current. The collection container sequence is then amplified via PCR, and ssDNA or RNA aptamers have been generated as a result. They can then be sequenced via standard sequencing methods. Aptamers have the advantage that any molecules can be targeted, and there are no immunizations of animals required. This results in a high affinity and high specificity receptor that can be selected for and then synthesized completely in vitro.31 Moreover, since aptamers are just a DNA sequence, they can be ordered from DNA synthesis companies as an insert inside a vector, and then stored as plasmids in E.coli, so that unlimited template DNA can be maintained. Luckily, aptamers that strongly bind to all of our protein biomarkers already existed in literature, so we were able to select sequences that were shown to have high affinity (nM ranges for KD), and request those to be synthesized by IDT.

For IL-1b, we decided to use the AptIL-1b aptamer39, as it was shown to bind IL-1b in a dose dependent manner in an ELISA-like assay.25 The binding affinity was not provided, but it detected as low as 10 ng IL-1b in the ELISA assay. For IL-6, we chose the aptamer used in Tertis et al., 2019, because they created an electrochemical aptasensor with detection limit 1.66 pg/mL26, which is sufficient enough for testing sweat samples, as IL-6 ranges are 1-20 pg/mL. For TNF-a, we found several aptamers characterized in Orava et al., 2013.41 We chose VR1127 because it had the smallest dissociation constant, thus the highest affinity. For CRP, we chose the aptamer used in Jarczewska et al., 201842 because it had a dissociation constant of 6 pM, which is the type of affinity we were looking for. Also, it has been used successfully in electrochemical biosensors before.42 For lactoferrin, we chose Ylac8 from Yu et al., 201930 because it had the lowest dissociation constant, thus highest affinity.

We researched receptors to find the best ones to use in our device. Receptors are proteins that can receive biological signals. Nucleic acid aptamers are pieces of DNA or RNA that can bind to receptors. Relevant aptamers are discovered by a process in which nucleic acids that bind well to receptors will be singled out. One example of this is Silver Enhanced Fluorescence-Activated Cell Sorting, shown in figure 1. The target proteins are modified and added to a library of DNA or RNA, and then the DNA that bind tightly to the protein can be chosen.

After the right aptamers are chosen, the DNA that doesn't bind well is disposed of, and the DNA aptamers we do want are amplified using a process called PCR that creates many copies of the DNA. We can order these aptamers from DNA synthesis companies and store them as plasmids (small circular DNA molecules) inside the E. coli. For our project we ordered aptamers that strongly bind to the biomarkers we chose. We chose these aptamers based on a variety of factors found in the literature we read, like how well they bind our biomarkers and how well they can differentiate between the biomarker and a similar but different molecule.

Figure 1: Process of Aptamer Screening by AgFACS - SELEX

As described earlier, all of the aptamers that recognize our target biomarkers are single-stranded DNA (ssDNA). RNA aptamers are harder to produce because they are highly unstable without modifications.43 Double-stranded (dsDNA) aptamers are not discovered as often44, possibly because there is less flexibility in its tertiary structure. Having only ssDNA aptamers means that we were able to use the same method to synthesize each one of them. For synthesizing ssDNA aptamers in vitro, we chose to do asymmetric PCR. Conventional asymmetric PCR procedures include using unequal concentration of the forward and reverse primers.45 The first phase of amplification is dsDNA production which occurs exponentially, which is then followed by the second phase of amplification that gives ssDNA. This ssDNA generation occurs until the increase in the number of DNA copies is restricted by the amount of enzyme present in the reaction mixture.32

A few factors were considered when designing asymmetric PCR. The most essential factor for the reaction to work was the ratio of primer amounts. Asymmetric PCR is unidirectional, depending on the presence of both primers in different amounts, using dsDNA as the template. The excess primer corresponds to the forward primer, which is the one that binds to the antisense strand, if we consider that the sense one is the one that has the aptamer sequence. The forward primer is in excess because that contains the sequence of the strand that we want. This is described in Figure 2. In the literature, excess primer concentrations were usually set to be 1000nM, and limited primer concentration was 50nM.33 Another factor that was considered was the melting temperature of primers and the annealing temperature of the reaction. In literature, annealing temperature was set to be 2°C below the melting temperature of the excess primer.33 We first ran gradient PCR reactions, as is discussed in results, in order to determine the optimal temperature. Table A shows the sequences of each primer and their melting temperatures, as well as the strength of heterodimer formation, if there is any (obtained using Vita Scientific32 calculator)46. Tabel B shows the determined specific annealing temperatures. We had to make sure that the primers are not too short to anneal/too low melting temperature, but also that they are not too long, which would cause them to anneal to each other given how short the aptamer sequences are. This was challenging, but we managed to find good enough primers for all aptamers.

There are a few advantages to using asymmetric PCR for synthesis of ssDNA. This type of PCR is the most cost effective method for ssDNA production and, given the team’s budget and the simplicity and cost of this reaction, we found it to be the most useful for purposes of our project. Unfortunately, asymmetric PCR amplification shows an overall efficiency of 60-70% in comparison to the efficiency obtained by the conventional PCR, which is 90% or more.33 Additionally, asymmetric PCR gives a mixture of ssDNA and dsDNA. This requires a good purification method for only ssDNA. Therefore, in our project, we aimed to optimize conditions of asymmetric PCR to produce ssDNA. This investigation can later be used by future iGEM teams or labs with limited resources, as we successfully purified ssDNA.

Figure 2. Representation of forward/reverse primers’ annealing in asymmetric PCR

Table A. Design and characteristics of the forward and reverse primers for each biomarker

Table B. Determined optimal specific annealing temperature

CRP, Lactoferrin and IL-1β

From the initial gradient PCR reaction, we observed that lactoferrin produced a higher-intensity band on agarose gel at higher temperature, so we tried a 58-68°C gradient PCR (Figure 3). Based on the agarose gel result, we decided that 60°C is the best annealing temperature for lactoferrin (Table 1 shows the temperature for each lane).

Figure 3. Lactoferrin gradient PCR

A B C D E F G H
Temperature (℃) 68 66.6 65.1 63.7 62.3 60.9 59.4 58

Table 1. Annealing temperature for gradient PCR

After purification of the lactoferrin product using the NEB PCR cleanup kit, we compared the concentration of the purified lactoferrin product from 15 cycles and lower Mg2+ versus 20 cycles and 2mM Mg2+. We found the best conditions for lactoferrin aptamer to be 60°C annealing temperature, 20 PCR cycles, and 2 mM MgCl2. After performing six different reactions with these conditions, we obtained an average concentration of 8.68 ng/μL and a ssDNA:dsDNA ratio of roughly 7:1. (Figure 6)

For IL-1β, the initial gradient (48-60°C) gave poor results, so we tried a lower gradient (40-50°C), as well as 15:1 and 30:1 Forward:Reverse primer ratios, instead of the initial 20:1. We observed that 46°C annealing temperature and 30:1 primer ratio gave the highest ssDNA concentration, as measured by the Quantifluor kit and visualized on gel (Figure 4). Finally, increasing the Mg2+ concentration and the number of cycles also improved the amount of ssDNA produced. Thus, the final PCR conditions for IL-1β aptamer were 46°C annealing temperature, 30:1 forward:reverse primer ratio, 20 PCR cycles, and 2 mM MgCl2. After performing four different reactions with these conditions, we obtained an average concentration of 9.24 ng/μL and a ssDNA:dsDNA ratio of roughly 5.5:1. (Figure 6)

Figure 4. Agarose gel of IL-1β at different primer ratios after gradient PCR

For CRP, a 55-68°C gradient of annealing temperature was tried first. We also tried 15:1 and 20:1 primer ratio, as we observed with previous aptamers that 20:1 might be too much forward primer in some cases. The agarose gel results are shown in Figure 5. Seeing the high intensity bands in the 15:1 primer ratio conditions (Figure 5), we decided to continue with this ratio and try annealing temperatures lower than 55°C. After measuring the concentration obtained from different Mg2+ concentrations and different number of cycles, we found that 20 cycles, 2 mM MgCl2 concentration and 53°C annealing temperature were the best conditions. We performed the reaction three times with these conditions, in 50 μL reaction volumes and purifying each reaction separately, and we obtained an average concentration of 27.39 ng/μL, and an average ssDNA:dsDNA ratio of 7.5:1 (CRP low, Figure 6). However, for our future aptamer applications, we needed a slightly higher concentration so we tried running 25 μL PCR reactions and purifying 100 μL at a time using the NEB PCR cleanup kit. After doing this three times, we obtained an average concentration of 75.64 ng/μL with an average ssDNA:dsDNA ratio of around 4:1 (CRP high, Figure 6). The final PCR conditions are shown in Table 2.

Figure 5. CRP on 3% agarose gel

Figure 6. Average DNA concentrations from Quantifluor measurement.

Aptamer Temperature Ratio Optimization
IL-1b 46°C 30:1 20 cycles, 2mM MgCl2
CRP 53°C 15:1 20 cycles, 2mM MgCl2
Lactoferrin 60°C 20:1 20 cycles, 2mM MgCl2

Table 2. Final PCR conditions for all aptamers

We analyzed the final PCR products on a 5% agarose gel, because it was reported that for small ssDNA the separation is more clear on higher-concentration gels.1 We also ran a 10 bp ladder, and the size of each band is shown in Figure 7.

Figure 7. 5% Agarose gel of all aptamer products.

Aptamer characterization:

Because it is important that we have mostly ssDNA since that is what the biomarkers are going to bind to, we used the Quantifluor assay to accurately determine the ssDNA to dsDNA ratio. This assay uses two different dyes, one with high affinity for ssDNA and one with high affinity for dsDNA, and the fluorescence is measured using a BioTek plate reader. The concentration is measured based on a standard curve that was run for each experiment. The average ratios obtained for the three aptamers that worked are shown in Figure 8. We got a ssDNA:dsDNA ratio between 5.5:1 to 8:1 for the optimized aptamers, which, to the best of our knowledge, is considerably higher than values reported in literature, since those were only around 3:1.2 Thus, we were able to optimize the PCR conditions for single-stranded DNA synthesis and the concentration could be further improved by modifying the primer sequence or by gel purifying with an optimized protocol.

Figure 8. Final ssDNA:dsDNA ratios for aptamers.

Attaching Aptamers to rGO

While trying to find the best method to attach our aptamers to reduced graphene oxide, we found several options: some are using azide-modified aptamers on PEI-modified rGO, because propargylactic acid covalently binds to NH2 groups on PEI via EDC-NHS linking and then azide-modified aptamers covalently attach via Cu(I) cross-link chemistry47; some are using thiol-modified aptamers, which covalently attach to gold nanoparticles.48 However, because of our limited budget and dedication to produce an affordable device, and also because of the clinical importance of using methods that allow for rapid replacement of aptamers on the surface, we chose to use a method that does not require any aptamer modifications. We found that aptamers can attach to rGO simply via π-π stacking interactions49, and if the interaction is T-shaped, this will allow the aptamer to adopt its proper tertiary structure and bind its target protein. We assumed that is the case because the paper describing this method was using the final device to detect a hormone, so the aptamer had to be able to bind properly. Alternatively, spacers can be used to ensure proper conformation. Also, our discussions with several professionals working with aptamer-based biomaterials confirmed that attaching aptamers via π-π stacking seems feasible and that the biomarker binding should result in a detectable electrical change on rGO. We ended up using a variation of the protocol described in Chergui et al., 202049 where the rGO was first coated with PEI and then aptamers were attached in solution, via end-over-end rotation, for 2 hours at room temperature.50 We hoped that this method will work without using the PEI coating, if we use a higher amount of rGO than used in the paper. Our confocal microscopy results do show that ssDNA was successfully attached to rGO.

Results of Aptamer attachment to rGO

As mentioned, after production of our biomarker-specific aptamers, we needed to attach them to reduced graphene oxide. We performed the attachment in solution by mixing 10 mL of 4 g/L rGO with 2 mL 10 nM aptamer solution. As a control, we also attached 2 mL of just the buffer solution. After 2 hours incubation at room temperature with end-over-end rotation, 1.5 mL of the sample was placed on a glass slide and dried at room temperature overnight.

We wanted to test whether the aptamers were actually attached to rGO. We added Quantifluor ssDNA dye to the slide, rinsed, and imaged under confocal microscopy. Images taken with the 20x objective are shown in Figure 9. For each image, we identified 30 regions with no aggregated dye (bright white spots were ignored), quantified the average fluorescence over each area for each region using ImageJ and then determined the mean value for each sample (Figure 10). rGO-CRP aptamer has a higher fluorescence over the entire area, as compared to rGO-buffer, which confirms that the aptamer is attached to the rGO.

Figure 9. Confocal microscope images of rGO - CRP aptamer (left) and rGO - buffer (right) slides (10 nM)

Figure 10. Average relative fluorescence of rGO+aptamer slide versus rGO+buffer slide (black asterix shows significance with respect to background; red asterix shows significance with respect to rGO + buffer slide; statistical test: Unpaired t test; Reported: two-tailed P value: 0.0006 for aptamer-buffer)

We also drop casted rGO-aptamer solution onto electrodes, and perform cyclic voltammetry measurements with our low cost potentiostat to show the aptamer successfully attached to rGO. We observed a significant decrease in peak current compared to the rGO-buffer electrode. Proof of Concept

Testing our Aptasensor

To determine if the lactoferrin aptamer successfully binds its target biomarker when attached to the rGO sheet, we tested the electrode modified with rGO + lactoferrin aptamer at different lactoferrin concentrations, to determine the change in resistance.

Our lactoferrin aptamer showed a monotonically increase in resistance in the 50 - 400 nM lactoferrin concentration range, while there was no detectable change in resistance in the electrode modified with rGO + buffer (Figure 11). This shows that lactoferrin aptamer and our device detects lactoferrin biomarkers up to 400 nM and can also distinguish between different concentrations, with detection limit of 50 nM.

The binding affinity of lactoferrin reported in literature is 1 ± 0.12 nM3 and concentrations between 1-25 nM indicate sepsis.4 Although our detection sensitivity might not be great in that range, it could be improved by optimizing the attachment of aptamer to rGO, for example by using spacers and a covalently linking method, such as azide linking to PEI-modified rGO.5 We could also increase aptamer density on the rGO sheet or change the degree of reduction of graphene oxide, and therefore changing the conductivity of the sheet (which could result in a more detectable change when biomarkers bind). We chose a simpler method of aptamer attachment in order to keep our device as accessible as possible, and even though our resistance dependence on concentration might not have an established relation yet, even being able to detect a change will have a significant impact if this data is analyzed together with the other biomarkers we targeted. Overall, our results show proof of concept of our aptasensor, because each module successfully integrated within the final device and we determined the detection limit of our device. Proof of Concept

Figure 11. The resistance of rGO sheet given the concentrations of lactoferrin biomarker

TNF-α and IL-6

For the remaining two aptamers, we optimized the asymmetric PCR by tuning the reaction parameters. First of all, we adjusted the temperature gradients accordingly to find the optimal annealing temperature for each aptamer. Specifically for Tumour Necrosis Factor alpha (TNF-α) and Interleukin-6 (IL-6), the gradient first used was 8 temperatures in the range of 45-60℃ (Table 3). As we saw darker bands in the high temperature range (Figure 12A and 12B), we narrowed the temperature range to 50-60℃. After gel purifying the bands and analyzing the concentration by Nanodrop, we found the best temperature for TNF-α to be 52.2℃ and the best for IL-6 to be 57.2℃.

We also performed several experiments to identify the best forward-reverse primer ratios, including 15:1, 20:1, and 30:1. To further optimize the condition, we also adjusted the number of cycles from denaturation to elongation. It is mentioned in previous research that asymmetric PCR would perform better to not exceed 20 cycles, since increasing the PCR cycles to 30 could produce spurious products.1 Therefore we have tried running 15 to 20 cycles of amplification for each aptamer. We used a PCR cleanup kit from NEB to purify the PCR product with high yield, because the kit has a protocol that works well for oligonucleotides. After cleanup, we determined the amount of ssDNA and dsDNA using sensitive fluorescent dyes that specifically bind to single stranded DNA or double stranded DNA, from the Quantifluor kit.

Figure 12. Initial Gradient PCR Result. Top: TNF-α agarose gel; Bottom: IL-6 agarose gel. (A-H is referring to the annealing temperature.)

A B C D E F G H
Temperature (℃) 60 57.8 55.7 53.5 51.4 49.3 47.1 45

Table 3. Gradient annealing temperature for TNF-α and IL-6 PCR

In spite of our attempts to enhance the single-stranded aptamer production, the results from ssDNA and dsDNA quantification, and agarose gel were not ideal for TNF-α and IL-6 compared to other aptamers. None of the three primer ratios that we tried for TNF-α PCR reactions result in high concentration of product in quantification, or a band of the right size on the gel (Figure 13). Similarly for IL-6, only primer dimers appear on the gel but not the aptamer in the correct size, and concentrations of single-stranded products were less than 3 ng/μL for TNF-α and less than 6 ng/μL for IL-6 (Fig. 14).

Figure 13. Agarose gel of IL-1β, TNF-α, CRP, IL-6, and lactoferrin with 10 bp ladder, at determined annealing temperatures

As a last resort, we tried using a high-fidelity Q5 polymerase instead of Taq polymerase to see if the problem was polymerase processivity. Despite this effort, no product appeared on the agarose gel (Figure 15). Eventually, because of the low concentration of single-stranded products along with being unable to visualize the expected size of the band, we decided to depart from TNF-α and IL-6, and continue to work only with IL-1β, CRP, and lactoferrin.

Figure 14. Average ssDNA concentration (ng/μL)

Figure 15. Q5 Polymerase Agarose Gel Result for TNF-α and IL-6 with 100bp and 10bp ladders

Future Directions

RCA

A useful next step in order to increase the binding affinity of the aptamers would be to use rolling circle amplification to synthesize several aptamer molecules next to each other, since that would increase the number of binding sites. It has been shown that it has a great effect on the sensitivity of biosensors.6,7,36,37

rGO Experiments

Background Information

What is rGO?

rGO is a 2-dimensional sheet of six-membered carbon rings. The sheets are decorated with a few oxygen-containing groups, like hydroxyls, epoxides, and carboxylic acids. Graphene oxide has a similar structure but with more oxygen groups. When GO is reduced, the oxygen groups are replaced with carbon-carbon double bonds, eventually forming an extensive system of conjugated p-orbitals. This makes rGO much more conductive than GO, since electrons can flow through the conjugated pi system.

rGO is a 2-dimensional sheet of six-membered carbon rings. Some of the carbons are connected to each other with double bonds, and others only have single bonds. A lot of the single-bonded carbons have bonds to other kinds of atoms, like oxygen. Graphene oxide (GO) has a similar structure to rGO, but it has more bonds to oxygen. Reduction is the process of removing bonds to oxygen. In this case, the single bonds to oxygen are replaced with double bonds to carbon. Electrons have an easier time flowing through double bonds than single bonds, so rGO is more conductive than GO.

Why rGO?

When we were deciding on a material for a biosensor we needed one that was conductive enough for us to be able to monitor small changes, inexpensive enough to be able to be widely available in hospitals, and easy to produce. We eventually settled on rGO because it is much more conductive than GO, and also less expensive and easier to modify than metals like gold or silver.

We also chose to use rGO because of its versatility. Since we planned to use aptamers to detect biomarkers, we needed to find a way to attach the aptamers to our sensor. There are a few literature methods to attach aptamers to sensors: the first is to chemically modify the aptamers and put a functional group on one of the ends, which can then be used to attach it to the biosensor; the second method is to rely on pi-pi stacking between the aptamer and the sensor material to keep them connected. rGO is able to do both. The oxygen functional groups can be used to covalently bind aptamers to the rGO, while the conjugated pi-system can be used to attach the aptamers through pi-pi stacking.

One of the benefits of rGO is that we can attach aptamers to it in a couple different ways. First, we can modify the aptamer by attaching something called a functional group, which is a small collection of atoms that have a certain purpose. In this case, the purpose would be to form a direct (covalent) bond to the rGO. The second option to attach aptamers is something called pi-pi stacking (named for the pi orbitals that make double bonds). Pi-pi stacking is when two double bonds overlap and interact with each other. Our aptamers are made out of DNA, which has a few double bonds to allow for this kind of interaction. Both methods have their advantages: covalent bonding is much stronger than pi-stacking, but is also much more expensive. That's why we wanted a material like rGO, which can be used for both kinds of attachments.

Figure X: Chemical Structure of rGO and GO.

Thermal Reduction

When making reduced graphene oxide (rGO), there are three ways to reduce graphene oxide (GO) to rGO: thermally, chemically, or with bacteria. Thermal reduction of GO requires high temperatures (up to 450℃), which needs specialized equipment that we were unable to obtain. Additionally, we were concerned about the possibility of high heat causing the pi-bonds in rGO to interact with one another and cause defects in the structure of our rGO sheets.

Chemical Reduction

Chemical synthesis of rGO presented its own share of challenges. The traditional method of chemical reduction uses hydrazine as a reducing agent, which is a caustic, flammable, toxic, and possibly carcinogenic reagent. Recently, safer methods have been developed using reducing reagents like ascorbic acid. Unfortunately, ascorbic acid reduces almost all of the oxygen groups on GO, making it very difficult to functionalize the rGO with aptamers or other chemical groups. Since our ultimate goal is to attach aptamers to rGO, we needed to find a milder reducing agent than ascorbic acid.

Biological Reduction

Bacterial reduction of GO is a developing field in synthetic biology. Some bacteria, like Shewanella oneidensis MR-1, are capable of exporting electrons into a surrounding medium and reducing insoluble compounds, like GO. Shewanella naturally export compounds called flavins which act as electron shuttles to reduce various chemicals. Unlike thermal reduction, bacterial reduction can be done at room temperature, and unlike chemical reduction, no hazardous reagents are needed. Even better, the flavins are strong enough reducing agents to remove hydroxyl and epoxy groups from the GO sheets, but not strong enough to interact with carbonyl groups like ketones or carboxylic acid. That means that we can eliminate the nonessential oxygen groups to maximize the conductivity of our rGO while also leaving enough functional groups for us to manipulate the rGO by attaching our aptamers. We decided to use bacterial reduction because it is safer, more accessible, and more effective for our project than any of the other reduction methods.

Bacterial reduction of GO is a developing field in synthetic biology. Some bacteria, like Shewanella oneidensis MR-1, are capable of exporting electrons into a surrounding medium. They do this by releasing molecules that can accept electrons from the bacteria's respiration pathway. Those electrons can then be used to reduce nearby molecules. Bacterial reduction has a few advantages. First, it can be done at room temperature and requires much less specialized equipment than the thermal reduction method. Second, it doesn't need any potentially dangerous reagents like the chemical method. Finally, the electron shuttles can only remove single bonds to oxygen, so any double bonds to oxygen get left behind. While this may seem like a disadvantage, we need oxygen groups to covalently bind aptamers to our rGO. Bacterial reduction removes some oxygen groups to make double bonds, which makes our material much more conductive than the starting GO. Those double bonds can also be used for pi-pi stacking, or the remaining oxygen groups can be used for covalent bonding.

Our team decided to use a bacteria called Shewanella oneidensis MR-1 to synthesize our rGO. Other methods, like chemicals, are more thorough than bacteria, but we wanted to keep our material as versatile as possible. As there are two main ways to attach aptamers to rGO (covalent bonding and pi stacking), we wanted to develop a material that could use both methods, which meant that we needed a method that would leave some oxygen groups behind; a method like bacterial reduction.

Shewanella oneidensis MR-1 possesses the natural capability of reducing extracellular materials and has previously been shown to be capable of reducing graphene oxide.8 Therefore we decided to use this as our chassis organism to carry out microbial reduction to produce reduced graphene oxide. We engineered transformed strains of S. oneidensis MR-1 to inducibly express YdeH, oprF, cymA, ribF, or the riboflavin synthesis gene cluster (rib clus) to either speed up the reduction process or improve the quality of reduction within a 48-hour period as compared to S. oneidensis MR-1 wild-type.

We selected the previously mentioned genes based on their roles in different stages of the reduction pathway. Some of the genes were selected to increase the electrons that are shuttled across the membrane, and these include oprf and cyma.9,10 The others were selected to increase the production of the electron shuttle, flavin, for export out of the cell. These genes include ribf, oprf and the rib clus.11,12 Lastly, ydeh was selected for its ability to increase biofilm production to increase direct contact between the outer membrane electron transport proteins and the graphene oxide.13

Figure 1: Electron transport pathway including the inner membrane cytochrome c protein (cymA) and the flavin electron shuttles flavin adenine dinucleotide (FAD) exported into the periplasm to eventually be converted to riboflavin (RF).14,15

Golden Gate Cloning

Our selected genes were synthesized as gBlocks by IDT and Twist Bioscience. To insert our genes into the appropriate vector, pcD8, for transformation of S. oneidensis MR-1, we utilized the Golden Gate cloning method. Here, BsaI sites were added to our vector via touchdown PCR, and cloning was carried out to insert our gBlocks which already contained the Golden Gate BsaI sites. To confirm that our genes were inserted into the vector, we transformed the cloned products into S. oneidensis MR-1, performed screening digests with KpnI, and then carried out Sanger sequencing.

Once it was confirmed that our genes were correctly inserted into the vectors, we began to carry out microbial reduction of graphene oxide with our newly transformed strains. This reduction was executed over a 48-hour period where reduction was measured by optical density (O.D.600), Raman spectroscopy, and X-Ray Photoelectron Spectroscopy (XPS).

Optical Density

Previous studies have shown that the magnitude of reduction of graphene oxide can be measured using optical density measurements at a wavelength of 600 nm. It has also been shown that the presence of graphene oxide has no significant effect on the growth rate of S. oneidensis MR-1.16 So, we utilized this technique to measure and compare the magnitude of reduction with our different transformed strains and wild-type.

Expression of our genes in our IPTG-inducible vector, pcD8, was initially induced with 1.5mM IPTG. Immediately following induction, reduction of graphene oxide was carried out under constant shaking at 200 rpm, and the O.D.600 was measured every hour for 48 hours for TSB only, graphene oxide only, graphene oxide and TSB only, and graphene oxide and TSB with each of our transformed strains and the wildtype. Here, the TSB only , graphene oxide only, and graphene oxide and TSB only are our negative controls, and wild type MR-1 is the positive control to which all our transformed strains are compared. The O.D.600 measurements obtained from these experiments reflect both bacterial absorbance as well as reduced graphene oxide absorbance. So, the absolute O.D.600 due to reduced graphene oxide is obtained by correcting for O.D.600 of graphene oxide and TSB only and bacteria and TSB only.

1 / 3

Figure 2: O.D.600 of microbial reduction (bacterial strains+TSB+GO) with wild-type MR-1 (deep purple) compared to the transformed strains cyma, oprf, pcD8 (empty vector), ydeh, ribf and riboflavin cluster for the reduction period. Here, time zero reflects the start of induction with 1.5mM IPTG.

2 / 3

Figure 3: O.D.600 of bacteria and TSB only with wild-type MR-1 (dark gray) compared to the transformed strains cyma, oprf, pcD8 (empty vector), ydeh, ribf and riboflavin cluster over a 48-hour period. Here, time zero reflects the start of induction with 1.5mM IPTG.

Since the O.D.600 measurement obtained reflects both the reduced graphene oxide absorbance and the bacterial absorbance, and Figure 3 shows that the bacterial O.D.600 differs between strains and changes over time, the direct measured values may not be an accurate representation of reduction. To correct for this we subtract out the absorbance values for bacteria and TSB only which should then result in only the absorbance due to reduced graphene oxide that can then be compared between strains (See figure 4).

3 / 3

Figure 4: O.D.600 of microbial reduction (bacterial strains+TSB+GO) with wild-type MR-1 (deep purple) compared to the transformed strains cyma, oprf, pcD8 (empty vector), ydeh, ribf and riboflavin cluster adjusting for the O.D.600 values due to bacterial growth. Here, time zero reflects the start of induction with 1.5mM IPTG.


Since ydeh, oprf and ribf showed slower growth than wild-type MR-1 (Figure 3), we hypothesized that the induced genes may have been slightly toxic to the cells. We then observed the growth curves when the concentration of IPTG was decreased to 1.0mM to reduce the amount of expression for all genes.

1 / 3

Figure 5: O.D.600 of microbial reduction (bacterial strains+TSB+GO) with wild-type MR-1 (deep purple) compared to the transformed strains cyma, oprf, pcD8 (empty vector), ydeh, ribf and riboflavin cluster for the reduction period. Here, time zero reflects the start of induction with 1.0mM IPTG.

2 / 3

Figure 6: O.D.600 of bacteria and TSB only with wild-type MR-1 (dark gray) compared to the transformed strains cyma, oprf, pcD8 (empty vector), ydeh, ribf and riboflavin cluster for the 48-hour period. Here, time zero reflects the start of induction with 1.0mM IPTG.

Since the O.D.600 measurement obtained reflects both the reduced graphene oxide absorbance and the bacterial absorbance, and Figure 6 shows that the bacterial O.D.600 differs between strains and changes over time, the direct measured values may not be an accurate representation of reduction. To correct for this we subtract out the absorbance values for bacteria and TSB only which should then result in only the absorbance due to reduced graphene oxide that can then be compared between strains (See figure 7).

Figure 6 shows that there was no significant alteration in the growth for ydeh and oprf for induction with 1.0mM IPTG as compared with 1.5mM IPTG (Figure 3). There appears to be a longer delay in ribf growth for induction with 1.0mM IPTG but a greater increase in growth following this delay as compared with 1.5mM IPTG (Figure 3).

3 / 3

Figure 7: O.D.600 of microbial reduction (bacterial strains+TSB+GO) with wild-type MR-1 (deep purple) compared to the transformed strains cyma, oprf, pcD8 (empty vector), ydeh, ribf and riboflavin cluster adjusting for the O.D.600 values due to bacterial growth. Here, time zero reflects the start of induction with 1.0mM IPTG.

The results showed that ydeh, cyma and riboflavin cluster had comparable reduction rates to wild-type MR-1 and the empty vector, pcD8 for induction with 1.5mM IPTG (Figure 4). ydeh and riboflavin cluster also had comparable reduction rates to the wildtype and pcD8 with 1.0mM IPTG induction (Figure 7).


We saw that the transformed strains oprf, ribf and riboflavin cluster showed a high maximal rate of reduction that occurred later than the maximal rate of reduction for wild-type MR-1. We therefore hypothesized that an induction time of 5 hours prior to reduction would allow the transformed strains to already have their enhanced electron transport pathways fully active before we began the reduction reactions. This was carried out with either 1.5mM or 0.75mM of IPTG under the same conditions of shaking at 200 rpm (Figure 8,11).

1 / 3

Figure 8: O.D.600 of microbial reduction (bacterial strains+TSB+GO) with wild-type MR-1 (deep purple) compared to the transformed strains cyma, oprf, pcD8 (empty vector), ydeh, ribf and riboflavin cluster for the reduction period. Here, reduction was initiated following a 5 hour induction with 1.5mM IPTG (5hr induction)

2 / 3

Figure 9: O.D.600 of bacteria and TSB only with wild-type MR-1 (red) compared to the transformed strains cyma, oprf, pcD8 (empty vector), ydeh, ribf and riboflavin cluster for the 48-hour period. Here, reduction was initiated following a 5 hour induction with 1.5mM IPTG (5hr induction)

Since the O.D.600 measurement obtained reflects both the reduced graphene oxide absorbance and the bacterial absorbance, and Figure 9 shows that the bacterial O.D.600 differs between strains and changes over time, the direct measured values may not be an accurate representation of reduction. To correct for this we subtract out the absorbance values for bacteria and TSB only which should then result in only the absorbance due to reduced graphene oxide that can then be compared between strains (See figure 10).

The bacterial growth measured by O.D.600 (Figure 9) shows that bacteria expressing the ydeh gene (green) still had less overall growth over the 48-hour period and more variability in its growth curve as compared to the wildtype and the other expressed genes. There is still a delay, albeit a decreased one, in the growth of ribf (deep purple) and oprf (light purple).

3 / 3

Figure 10: O.D.600 of microbial reduction (bacterial strains+TSB+GO) with wild-type MR-1 (deep purple) compared to the transformed strains cyma, oprf, pcD8 (empty vector), ydeh, ribf and riboflavin cluster adjusting for the O.D.600 values due to bacterial growth. Here, reduction was initiated following a 5 hour induction with 1.5mM IPTG (5hr induction)


1 / 3

Figure 11: O.D.600 of microbial reduction (bacterial strains+TSB+GO) with wild-type MR-1 (deep purple) compared to the transformed strains cyma, oprf, pcD8 (empty vector), ydeh, ribf and riboflavin cluster for the reduction period. Here, reduction was initiated following a 5 hour induction with 0.75mM IPTG (5hr induction).

2 / 3

Figure 12: O.D.600 of bacteria and TSB only with wild-type MR-1 (red) compared to the transformed strains cyma, oprf, pcD8 (empty vector), ydeh, ribf and riboflavin cluster for the 48-hour period. Here, reduction was initiated following a 5 hour induction with 0.75mM IPTG (5hr induction).

Since the O.D.600 measurement obtained reflects both the reduced graphene oxide absorbance and the bacterial absorbance, and Figure 12 shows that the bacterial O.D.600 differs between strains and changes over time, the direct measured values may not be an accurate representation of reduction. To correct for this we subtract out the absorbance values for bacteria and TSB only which should then result in only the absorbance due to reduced graphene oxide that can then be compared between strains (See figure 13).

3 / 3

Figure 13: O.D.600 of microbial reduction (bacterial strains+TSB+GO) with wild-type MR-1 (deep purple) compared to the transformed strains cyma, oprf, pcD8 (empty vector), ydeh, ribf and riboflavin cluster adjusting for the O.D.600 values due to bacterial growth. Here, reduction was initiated following a 5 hour induction with 0.75mM IPTG (5hr induction).


The slopes at the steepest point on the curves were obtained for each induction condition to give the maximum reduction rate measured during the reduction period, and this was compared to that of wild-type MR-1.

Figure 14: Maximum rates of the bacterial reduction curves adjusted for bacterial growth at the IPTG induction conditions of 1.5mM and 1.0 mM IPTG with 0hr induction and 1.5mM and 0.75mM IPTG with 5 hour induction.

We can see that after a 5hr induction with 1.5mM IPTG as well as 0.75mM IPTG, ydeh had the fastest rate of reduction as compared to the wild-type MR-1 and the other transformed strains (Figure 10,13). Figure 14 shows that the maximum rate obtained for ydeh at these conditions were 8.836 hr-1 and 8.468 hr-1 respectively whereas the wildtype had a rate of 7.732 hr-1. There are comparable reduction rates for ydeh with 1.5mM and 1.0mM IPTG for reduction immediately following induction (Figure 4,7). The reduction with ydeh appears to be more similar to wild-type MR-1 at the lower IPTG concentration, 1.0mM (Figure 7).

The growth curves of ydeh show stochastic patterns for each of the IPTG conditions (Figure 3,6,9,12). It is thought that such patterns are observed due to the production of biofilm where increased biofilm production may cause bacteria to adhere to the sides of the plate wells, reducing the amount of ydeh cells in the TSB for O.D.600 measurements.

All the IPTG conditions show that the riboflavin cluster and cyma have reduction rates comparable to each other, the wildtype, and pcD8, especially at 5hr induction with 1.5mM (Figure 10). oprf and ribf are seen to lag behind in reduction at all IPTG conditions, potentially due to the initial lag in their growth phase. The maximum reduction rates for riboflavin cluster, oprf, and ribf are greater than that of the wildtype for 1.0mM IPTG 0 hr induction. At 1.5mM IPTG 0hr induction the just riboflavin cluster and ribF also have maximum reduction rates greater than the wildtype (Figure 14).

The negative controls (TSB only, GO only, and GO and TSB only) show an insignificant change in O.D.600 over time indicating that the bacteria are responsible for the increase in O.D.600, which is the measurement of the change in reduction.

Overall, transforming S. oneidensis MR-1 with our inserts allowed for reduction of graphene oxide, where ydeH was shown to have the greatest rate of reduction, exceeding that of wild-type MR-1.

With these results, Raman spectroscopy was carried out to investigate the amount of carbon-carbon single bonds of the reduced graphene oxide produced by each of the transformed strains over the 48 hour period.

Raman Spectra

We used Raman spectroscopy to characterize our rGO. We primarily relied on the D/G ratio, which is the intensity of the D peak divided by the intensity of the G peak. The D peak is associated with double-bonded carbon, and the G peak is associated with single-bonded carbon. Since our goal is to make reduced graphene oxide, we wanted to have a high D/G ratio, since that would mean that we have a higher proportion of double bonds than single bonds. Because we were only able to take a limited number of spectra, we supplemented our characterization with other techniques, including the optical density described above and XPS described later on this page.

Raman spectroscopy is a characterization technique. Chemical bonds are constantly vibrating, and these vibrations are very specific to the type of bond. When light hits a molecule, the vibrations of the bonds scatter the light. Raman spectroscopy measures this scattering to determine the presence and amount of different kinds of bonds.

First, we set up a few controls. Chemically reduced rGO (crGO) was reduced using ascorbic acid. This was predicted to be a much more thorough reduction than the bacterial rGO, because of the high concentration of our reducing agent and the more intense conditions that it was produced under. This sample was our positive control. We also had two negative controls: untreated GO and GO that incubated in sterile TSB media while our other samples were reduced. The incubations in media are referred to as “blanks” and were used as a control to make sure the growth media was not interfering with our results.

When we made our microbially reduced rGO, we tested the reductive capabilities of wild type S. oneidensis MR-1, S. oneidensis MR-1 with an empty pcD8 vector, strains with CymA, Ydeh, Oprf, the Rib Cluster, and RibF plasmids. After reduction, we compared the D/G ratios to determine how well each had done in reducing the number of Sp2 carbons.17

Figure 15: Comparisons between each type of rGO after 48 hours.

As figure 15 shows, the crGO was by far the most reduced, with an average D/G ratio of around 1.2. The microbial reductions all ended up with very similar D/G ratios of around 1, which demonstrates that our parts worked as intended. None of our genetic modifications were intended to increase the degree of reduction, just the speed. Our 48 hour incubations confirmed that all of our S. oneidensis MR-1 strains reduced the GO to around the same degree.

Figure 16: Comparisons between wild-type, pcD8, CymA, and Rib Cluster strains after 12 hours»

As figure 16 shows, not all of the bacteria reduced the GO at the same rate. After 12 hours, cymA and the Rib Cluster were by far the most reduced compared to the pcD8 and wild type strains. This shows that these two modified bacteria strains are faster at reducing the GO than either the wild type strain or the strain with the empty expression vectors. CymA is likely faster than wild type because it has increased expression of an electron-exporting protein, which should increase the amount of electrons in solution to reduce the GO. The Rib Cluster strain is likely faster than wild type or pcD8 because it increases the amount of flavins synthesized, which means that there are more electron shuttles available to reduce GO. With these two strains of S. oneidensis MR-1, we have accomplished our goal of quickly synthesizing rGO in a safe and sustainable manner. We hope that these findings will help make our aptasensor less expensive and accessible to anyone who needs it.

Figure 17: Raman Spectrum I of crGO

Figure 18: Raman Spectrum II of crGO

Figure 19: Raman Spectrum of GO

Figure 20: Raman Spectrum I of Blank(GO+TSB)

Figure 21: Raman Spectrum II of Blank(GO+TSB)

Figure 22: Raman Spectrum I of Ydeh

Figure 23: Raman Spectrum II of Ydeh

Figure 24: Raman Spectrum I of pcD8

Figure 25: Raman Spectrum II of pcD8

Figure 26: Raman Spectrum I of Wild Type

Figure 27: Raman Spectrum II of Wild Type

Figure 28: Raman Spectrum I of CymA

Figure 29: Raman Spectrum II of cymA

Figure 30: Raman Spectrum I of ribF

Figure 31: Raman Spectrum II of ribF

Figure 32: Raman Spectrum I of Oprf

Figure 33: Raman Spectrum II of Oprf

Figure 34: Raman Spectrum I of rib cluster

Figure 35: Raman Spectrum II of rib cluster


Figure 36: Raman Spectrum of wild type

Figure 37: Raman Spectrum of pcD8

Figure 38: Raman Spectrum of cymA

Figure 39: Raman Spectrum of rib cluster


Future Directions

In the future, it would be interesting to see if an S. oneidensis strain could be engineered to overexpress both the cymA and rib cluster genes. This should theoretically increase the speed of reduction even more, because there would be both an increase in electrons and electron shuttles. In case the two plasmids are too toxic for the cells to tolerate, we would like to explore co-culturing our cymA strain with an E. coli strain that has been engineered to produce riboflavin. This would hopefully also increase the concentration of electrons and electron shuttles, while also reducing the risk of harming the cells.

References

  1. Agarose gel electrophoresis. Addgene. (n.d.). Retrieved October 21, 2021, from https://www.addgene.org/protocols/gel-electrophoresis/.
  2. Heiat, M., Ranjbar, R., Latifi, A.M., Rasaee, M.J. and Farnoosh, G. (2017), Essential strategies to optimize asymmetric PCR conditions as a reliable method to generate large amount of ssDNA aptamers. Biotechnology and Applied Biochemistry, 64: 541-548. https://doi.org/10.1002/bab.1507
  3. Fang Yu, Hui Li, Wei Sun, Yaju Zhao, Danke Xu, Fuchu He, Selection of aptamers against Lactoferrin based on silver enhanced and fluorescence-activated cell sorting, Talanta, Volume 193, 2019, Pages 110-117, ISSN 0039-9140, https://doi.org/10.1016/j.talanta.2018.09.063.
  4. Barkhatova NA. [The use of plasma lactoferrin in the diagnosis of pyonecrotic infections of soft tissues and sepsis]. Klin Med (Mosk). 2008;86(10):36-8. Russian. PMID: 19069457.
  5. Urmann, Katharina, Modrejewski, Julia, Scheper, Thomas and Walter, Johanna-G.. "Aptamer-modified nanomaterials: principles and applications" BioNanoMaterials, vol. 18, no. 1-2, 2017, pp. 20160012. https://doi.org/10.1515/bnm-2016-0012
  6. Lizardi, P., Huang, X., Zhu, Z. et al. Mutation detection and single-molecule counting using isothermal rolling-circle amplification. Nat Genet 19, 225–232 (1998). https://doi.org/10.1038/898/li>
  7. Zhao W, Cui CH, Bose S, Guo D, Shen C, Wong WP, Halvorsen K, Farokhzad OC, Teo GS, Phillips JA, Dorfman DM, Karnik R, Karp JM. Bioinspired multivalent DNA network for capture and release of cells. Proc Natl Acad Sci U S A. 2012 Nov 27;109(48):19626-31. doi: 10.1073/pnas.1211234109. Epub 2012 Nov 12. PMID: 23150586; PMCID: PMC3511714.
  8. Lin, T.; Ding, W.; Sun, L.; Wang, L.; Liu, C.-G.; Song, H. Engineered Shewanella Oneidensis-Reduced Graphene Oxide Biohybrid with Enhanced Biosynthesis and Transport of Flavins Enabled a Highest Bioelectricity Output in Microbial Fuel Cells. Nano Energy 2018, 50, 639–648.
  9. Marritt, S. J.; Lowe, T. G.; Bye, J.; McMillan, D. G. G.; Shi, L.; Fredrickson, J.; Zachara, J.; Richardson, D. J.; Cheesman, M. R.; Jeuken, L. J. C.; et al. A Functional Description of Cyma, an Electron-Transfer Hub Supporting Anaerobic Respiratory Flexibility in Shewanella. Biochemical Journal 2012, 444, 465–474.
  10. Brutinel, E. D.; Dean, A. M.; Gralnick, J. A. Description of a Riboflavin Biosynthetic Gene Variant Prevalent in the Phylum Proteobacteria. Journal of Bacteriology 2013, 195, 5479–5486.
  11. Liu, T.; Yu, Y.-Y.; Deng, X.-P.; Ng, C. K.; Cao, B.; Wang, J.-Y.; Rice, S. A.; Kjelleberg, S.; Song, H. Enhanced Shewanella Biofilm Promotes Bioelectricity Generation. Biotechnology and Bioengineering 2015, 112, 2051–2059.
  12. Dundas, C. M.; Walker, D. J. F.; Keitz, B. K. Tuning Extracellular Electron Transfer by Shewanella Oneidensis Using Transcriptional Logic Gates. ACS Synthetic Biology 2020, 9, 2301–2315.
  13. Covington, E. D.; Gelbmann, C. B.; Kotloski, N. J.; Gralnick, J. A. An Essential Role for Usha in Processing of Extracellular Flavin Electron Shuttles by Shewanella Oneidensis. Molecular Microbiology 2010, 78, 519–532.
  14. Lehner, B. A.; Janssen, V. A.; Spiesz, E. M.; Benz, D.; Brouns, S. J.; Meyer, A. S.; van der Zant, H. S. Creation of Conductive Graphene Materials by Bacterial Reduction Using Shewanella Oneidensis. ChemistryOpen 2019, 8, 888–895.
  15. Toward Resolving the Challenges of Sepsis Diagnosis | Clinical Chemistry | Oxford Academic. (n.d.). Retrieved August 31, 2021, from https://academic.oup.com/clinchem/article/50/8/1301/5640218
  16. Faix, J. D. Biomarkers of Sepsis. Critical Reviews in Clinical Laboratory Sciences 2013, 50 (1), 23–36.
  17. Hussain, J. N.; Mantri, N.; Cohen, M. M. Working up a good sweat - the challenges of standardising sweat collection for metabolomics analysis. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5548369/ (accessed Sep 17, 2021).
  18. Pierrakos, C., Vincent, JL. Sepsis biomarkers: a review. Crit Care 14, R15 (2010). https://doi.org/10.1186/cc8872
  19. Dai, X., Okazaki, H., Hanakawa, Y., Murakami, M., Tohyama, M., Shirakata, Y., & Sayama, K. (2013). Eccrine sweat contains IL-1α, IL-1β and IL-31 and activates epidermal keratinocytes as a danger signal. PloS one, 8(7), e67666. https://doi.org/10.1371/journal.pone.0067666
  20. Zee, H. H. van der, Ruiter, L. de, Broecke, D. G. van den, Dik, W. A., Laman, J. D., & Prens, E. P. (2011). Elevated levels of tumour necrosis factor (TNF)-α, interleukin (IL)-1β and IL-10 in hidradenitis suppurativa skin: A rationale for targeting TNF-α and IL-1β. British Journal of Dermatology, 164(6), 1292–1298. https://doi.org/10.1111/j.1365-2133.2011.10254.x
  21. Jagannath, B., Lin, K.-C., Pali, M., Sankhala, D., Muthukumar, S., & Prasad, S. (2020). A Sweat-based Wearable Enabling Technology for Real-time Monitoring of IL-1β and CRP as Potential Markers for Inflammatory Bowel Disease. Inflammatory Bowel Diseases, 26(10), 1533–1542. https://doi.org/10.1093/ibd/izaa191
  22. Bauer, P. R., Kashyap, R., League, S. C., Park, J. G., Block, D. R., Baumann, N. A., Algeciras-Schimnich, A., Jenkins, S. M., Smith, C. Y., Gajic, O., & Abraham, R. S. (2016). Diagnostic accuracy and clinical relevance of an inflammatory biomarker panel for sepsis in adult critically ill patients. Diagnostic Microbiology and Infectious Disease, 84(2), 175–180. https://doi.org/10.1016/j.diagmicrobio.2015.10.003
  23. Boone, J. H., DiPersio, J. R., Tan, M. J., Salstrom, S.-J., Wickham, K. N., Carman, R. J., Totty, H. R., Albert, R. E., & Lyerly, D. M. (2013). Elevated lactoferrin is associated with moderate to severe Clostridium Difficile disease, STOOL TOXIN, AND 027 INFECTION. European Journal of Clinical Microbiology & Infectious Diseases, 32(12), 1517–1523. https://doi.org/10.1007/s10096-013-1905-x
  24. Bossink, A. W. J., Groeneveld, A. B. J., & Thijs, L. G. (1999). Prediction of Microbial Infection and Mortality in Medical Patients with Fever: Plasma Procalcitonin, Neutrophilic Elastase-α1-Antitrypsin, and Lactoferrin Compared with Clinical Variables. Clinical Infectious Diseases, 29(2), 398–407. https://doi.org/10.1086/520222
  25. Sánchez, L., Calvo, M., & Brock, J. H. (1992). Biological role of lactoferrin. Archives of disease in childhood, 67(5), 657–661. https://doi.org/10.1136/adc.67.5.657
  26. Kruzel, M. L., Zimecki, M., & Actor, J. K. (2017). Lactoferrin in a Context of Inflammation-Induced Pathology. Frontiers in Immunology, 8, 1438. https://doi.org/10.3389/fimmu.2017.01438
  27. Huang, K., Du, G., Wei, C., Gu, S., & Tang, J. (2013). Elevated serum lactoferrin and neopterin are associated with postoperative infectious complications in patients with acute traumatic spinal cord injury. Archives of medical science : AMS, 9(5), 865–871. https://doi.org/10.5114/aoms.2013.38680
  28. Póvoa, P., Almeida, E., Moreira, P., Fernandes, A., Mealha, R., Aragão, A., & Sabino, H. (1998). C-reactive protein as an indicator of sepsis. Intensive care medicine, 24(10), 1052–1056. https://doi.org/10.1007/s001340050715
  29. Fang Yu, Hui Li, Wei Sun, Yaju Zhao, Danke Xu, Fuchu He, Selection of aptamers against Lactoferrin based on silver enhanced and fluorescence-activated cell sorting, Talanta, Volume 193, 2019, Pages 110-117, ISSN 0039-9140, https://doi.org/10.1016/j.talanta.2018.09.063.
  30. Shatunova, Elizaveta A et al. “Aptamers for Proteins Associated with Rheumatic Diseases: Progress, Challenges, and Prospects of Diagnostic and Therapeutic Applications.” Biomedicines vol. 8,11 527. 22 Nov. 2020, doi:10.3390/biomedicines8110527
  31. Simple and effective method for generating single-stranded DNA targets and probes Xing Tang, Sheldon L. Morris, John J. Langone, and Larry E. Bockstahler BioTechniques 2006 40:6, 759-763. https://doi.org/10.2144/000112154
  32. Linear-After-The-Exponential (LATE)–PCR: An advanced method of asymmetric PCR and its uses in quantitative real-time analysis J. Aquiles Sanchez, Kenneth E. Pierce, John E. Rice, Lawrence J. Wangh Proceedings of the National Academy of Sciences Feb 2004, 101 (7) 1933-1938; https://doi.org/10.1073/pnas.0305476101
  33. Tominaga, J. et al. “Asymmetric PCR for good quality ssDNA generation towards DNA aptamer production.” Songklanakarin Journal of Science and Technology 34 (2012): 125-131. https://doaj.org/article/4f163019a88843cda1df87782a62e5af
  34. Lorenzo-Gómez, R.; Miranda-Castro, R.; de-los-Santos-Álvarez, N.; Lobo-Castañón, M.J. Improving the Analytical Performance of Weak Aptamers: DNA Isothermal Amplification Approaches. Proceedings 2020, 60, 26. https://doi.org/10.3390/IECB2020-07026
  35. Ramón Lorenzo-Gómez, Noelia Fernández-Alonso, Rebeca Miranda-Castro, Noemí de-los-Santos-Álvarez, María Jesús Lobo-Castañón, Unravelling the lipocalin 2 interaction with aptamers: May rolling circle amplification improve their functional affinity?,Talanta, Volume 197, 2019, Pages 406-412,ISSN 0039-9140, https://doi.org/10.1016/j.talanta.2019.01.057.
  36. Ali, M. Monsur, Feng Li, Zhiqing Zhang, Kaixiang Zhang, Dong-Ku Kang, James A. Ankrum, X. Chris Le, and Weian Zhao. "Rolling Circle Amplification: A Versatile Tool for Chemical Biology, Materials Science and Medicine." 10.1039/C3CS60439J. Chemical Society Reviews 43, no. 10 (2014): 3324-41. https://doi.org/10.1039/C3CS60439J. http://dx.doi.org/10.1039/C3CS60439J
  37. Korlach, J., Bibillo, A., Wegener, J., Peluso, P., Pham, T. T., Park, I., Clark, S., Otto, G. A., & Turner, S. W. (2008). Long, processive enzymatic DNA synthesis using 100% dye-labeled terminal phosphate-linked nucleotides. Nucleosides, nucleotides & nucleic acids, 27(9), 1072–1083. https://doi.org/10.1080/15257770802260741
  38. Ho, D.-R., Chang, P.-J., Lin, W.-Y., Huang, Y.-C., Lin, J.-H., Huang, K.-T., . . . Chen, C.-S. (2020). Beneficial Effects of Inflammatory Cytokine-Targeting Aptamers in an Animal Model of Chronic Prostatitis. International journal of molecular sciences, 21(11), 3953. https://doi.org/10.3390/ijms21113953
  39. Tertis, M., Leva, P. I., Bogdan, D., Suciu, M., Graur, F., & Cristea, C. (2019). Impedimetric aptasensor for the label-free and selective detection of Interleukin-6 for colorectal cancer screening. Biosensors and Bioelectronics, 137, 123-132. https://doi.org/https://doi.org/10.1016/j.bios.2019.05.012
  40. Orava, E. W., Jarvik, N., Shek, Y. L., Sidhu, S. S., & Gariépy, J. (2013). A Short DNA Aptamer That Recognizes TNFα and Blocks Its Activity in Vitro. ACS Chemical Biology, 8(1), 170-178. https://doi.org/10.1021/cb3003557
  41. Jarczewska, M., Rębiś, J., Górski, Ł., & Malinowska, E. (2018). Development of DNA aptamer-based sensor for electrochemical detection of C-reactive protein. Talanta, 189, 45-54. https://doi.org/https://doi.org/10.1016/j.talanta.2018.06.035
  42. R.R. White, B.A. Sullenger, C.P. Rusconi, Developing aptamers into therapeutics, J. Clin. Invest., 106 (2000) 929–934
  43. Blind, M., & Blank, M. (2015). Aptamer Selection Technology and Recent Advances. Molecular Therapy - Nucleic Acids, 4, e223. https://doi.org/https://doi.org/10.1038/mtna.2014.74
  44. Gyllensten, U. B., & Erlich, H. A. (1988). Generation of single-stranded DNA by the polymerase chain reaction and its application to direct sequencing of the HLA-DQA locus. Proceedings of the National Academy of Sciences, 85(20), 7652. https://doi.org/10.1073/pnas.85.20.7652
  45. https://www.vitascientific.com/information/calculator
  46. Oberhaus, F. V., Frense, D., & Beckmann, D. (2020). Immobilization Techniques for Aptamers on Gold Electrodes for the Electrochemical Detection of Proteins: A Review. Biosensors, 10(5). https://doi.org/10.3390/bios10050045
  47. Huang, Y.-F., Lin, Y.-W., Lin, Z.-H., & Chang, H.-T. (2009). Aptamer-modified gold nanoparticles for targeting breast cancer cells through light scattering. Journal of Nanoparticle Research, 11(4), 775-783. https://doi.org/10.1007/s11051-008-9424-x
  48. Chergui, S., Rhili, K., Poorahong, S., & Siaj, M. (2020). Graphene Oxide Membrane Immobilized Aptamer as a Highly Selective Hormone Removal. Membranes, 10(9). https://doi.org/10.3390/membranes10090229
  49. Subramanian, P., Lesniewski, A., Kaminska, I., Vlandas, A., Vasilescu, A., Niedziolka-Jonsson, J., . . . Szunerits, S. (2013). Lysozyme detection on aptamer functionalized graphene-coated SPR interfaces. Biosensors and Bioelectronics, 50, 239-243. https://doi.org/https://doi.org/10.1016/j.bios.2013.06.026
  50. del Toro, M., Gargallo R Fau - Eritja, R., Eritja R Fau - Jaumot, J., & Jaumot, J. Study of the interaction between the G-quadruplex-forming thrombin-binding aptamer and the porphyrin 5,10,15,20-tetrakis-(N-methyl-4-pyridyl)-21,23H-porphyrin tetratosylate. (1096-0309 (Electronic)).
  51. Jing, M., & Bowser, M. T. Methods for measuring aptamer-protein equilibria: a review. (1873-4324 (Electronic)).