Wetlab

Asparagine Module: Whole-cell Bacterial Biosensor



Introduction


A central avenue of synthetic biology research is the development of whole-cell bacterial biosensors; genetically engineered bacteria that produce a predictable and measurable response upon introduction of a molecule of interest5. Unlike some traditional sensing methods, bacterial biosensors are low-cost, self-manufacturing, and biodegradable, making them ideal for a novel sensing technology4. However, whole-cell bacterial biosensing is limited by a major challenge: the difficulty of genetically engineering a new solution for each molecule of interest that is being detected4. In response to this challenge, Team Saptasense has developed a modular, customizable whole-cell bacterial biosensor that requires no further genetic engineering for the detection of a new compound. In this sense, we have engineered a universal whole-cell bacterial biosensor.

Our sensor contains two simple components: a genetically engineered strain of bacteria, and an antibody of choice. The read-out of the sensor is based on clumping, or “autoaggregation” of the engineered bacteria, which can be measured in any basic biology lab that possesses a spectrophotometer. A schematic of how the sensor was designed to work can be found in Figure 1.

Figure 1. Schematic of Universal Biosensor Design

First, E. coli must be transformed with our novel BioBrick BBa_K4130000. This BioBrick contains the coding region for the protein EibD (Escherichia coli Ig-binding protein D). EibD is an outer-membrane protein known to bind to immunoglobulin A/G (also known as antibodies). Previous studies have shown that expression of the protein results in the ability for bacteria to bind non-immunologically to the constant region of antibodies. This means that EibD can bind antibodies without disrupting the antigen-binding function. Unrelatedly, EibD also introduces a unique phenotype to the bacteria: autoaggregation. Due to self-self (homophilic) interactions between the EibD proteins, the bacteria begin to “clump” together. This bacterial clumping has been reported to be observed by the eye, and can also be measured quantitatively by optical density6.
Our designed biosensor leverages both the autoaggregation and Ig-binding properties of EibD. Upon induction and expression of EibD, the bacteria will visually clump together. This autoaggregation can be mitigated by incubation with an antibody of choice (Figure 1C). EibD-antibody interactions may outcompete the EibD-EibD interactions, reducing autoaggregation. As EibD binding to antibody leaves the variable region untouched, incubation with this antibody also introduces specificity to the biosensor towards a target molecule of interest.
Once antibodies have been bound to the bacterial surface, an antigenic sample may then be introduced. Multivalent interactions between antigen and antibody, and across multiple bacteria, may result in clumping or re-aggregation (Figure 1D)7. The degree of clumping can be measured as a function of concentration of antigen present.

Design Modification for Detection of Small Molecules


Our biosensor relies on the ability of a target molecule to bind multiple antibodies, themselves bound to multiple bacteria. For larger molecules such as proteins, a less specific or "polyclonal" antibody may be employed to this end. Small molecules such as amino acids however, contain limited binding sites and therefore present a unique challenge.
Therefore, Team Saptasense has employed a strategy modeled off of a previous publication by Riangrungroj et. al8. In this strategy, a non-antigenic protein is crosslinked to multiple small molecules of interest, forming an antigenic "bead". While the bead base may be composed of any protein which will not interact with the EibD bound antibody, we chose to use the highly characterized Bovine Serum Albumin (BSA). We chose glutaraldehyde as our crosslinker, enabling a reaction between the amine group of our small molecule of interest, asparagine, and the primary amine of lysine residues within BSA. The type of crosslinker will necessarily change depending on the small molecules of interest.
After successful synthesis, a bead can be employed for the detection of the small molecule in the designed whole-cell biosensor. The schematic showing incorporation of this bead can be seen in Figure 2.
Figure 2. Schematic of Universal Biosensor Design, Adapted for Small Molecule Detection

Just as in the detection of larger/complex molecules, EibD-expressing bacteria must bind an antibody of choice causing a disaggregation event. Subsequently, the “beads” must be added, allowing for multivalent interactions between antigens on the beads and antibody-bound bacteria. Binding events involving large numbers of bacteria will result in aggregation, and settling out of suspension. Finally, a test sample containing the small molecule of interest can be added to the system. The read-out of the sensor should be a disaggregation due to competition between bead-antibody interactions and free small molecule-antibody interactions.

Defective cabbage-tasting “buddy” maple syrup is produced at the end of each maple sap collection season, resulting in lost profits. The “buddiness” , associated with an increase in asparagine concentrations9, can only be detected after the sap has been processed into syrup, a time- and energy-intensive process for local sugarmakers. To our knowledge, a biosensor has yet to be produced for asparagine , especially in the context of maple sap. Therefore, our universal whole-cell biosensor employed for the detection of asparagine, will fill this gap.

Cultures of E. coli BL21 (+/- biobrick BBa_K4130000) were grown to mid-exponential phase (O.D.600 ~ 0.4 - 0.6), where growth and protein translation is maximal. Subsequently, to determine the differential effect of inducer concentration, bacteria were induced with 0.0% rhamnose, 0.001% rhamnose, or 0.01% rhamnose and grown for 2 hours at 30 C, 200rpm. After induction, the EibD-transformed bacterial cultures appeared markedly different from a BL21 control strain harboring no plasmid. A white precipitate accumulated on the bottom of the EibD cultures induced with 0.001% L-rhamnose, while the uninduced EibD culture exhibited no formation of white precipitate (Figure 3). This white precipitate is likely a result of autoaggregation due to previously reported homophilic interactions between EibD proteins on separate bacteria6. To investigate this, we used light microscopy to visualize the bacteria. Bacterial cultures were pelleted and washed in PBS, followed by incubation with acidic proteoglycan-staining safranin (Figure 4). Control E. coli BL21 without the EibD-containing plasmid appear evenly dispersed throughout the field of view. In contrast, EibD-expressing E. coli BL21 appear to aggregate in clumps within the field of view. This evidence is further indicative of autoaggregation. We quantified the degree of autoaggregation by measuring the settling rate of bacteria in solution over time. After brief mixing, the optical density at 600 nm (O.D.600) was recorded over a period of 30 minutes. As the bacteria settled and precipitated over time, a decrease in optical density was observed. The O.D.600 was observed to decrease more rapidly with increasing concentrations of L-rhamnose inducer, indicating that autoaggregation correlates positively with induction concentration. (Figure 5)
 Visual Observation of Autoaggregation: Mid-exponential phase E. coli BL21 were induced with 0.001% or 0% L-rhamnose and incubated for 2 hours at 30 C. Bacteria were harvested and resuspended in 1 mL PBS. Pictures were taken 0 minutes, 40 minutes, and 70 minutes post resuspension. At zero minutes, all cultures appear white in color. By 40 minutes, the EibD strain induced with 0.001% rhamnose appears noticeably more clear than the control cultures.
Figure 3. Visual Observation of Autoaggregation: Mid-exponential phase E. coli BL21 were induced with 0.001% or 0% L-rhamnose and incubated for 2 hours at 30 C. Bacteria were harvested and resuspended in 1 mL PBS. Pictures were taken 0 minutes, 40 minutes, and 70 minutes post resuspension. At zero minutes, all cultures appear white in color. By 40 minutes, the EibD strain induced with 0.001% rhamnose appears noticeably more clear than the control cultures.
Light Microscopy Observation of Autoaggregation: Induced bacteria were stained with safranin and visualized using a light microscope at 1000x magnification. Pictures were taken through the eyepiece using a phone camera. As can be seen on the left, control E. coli BL21 bacteria are evenly distributed throughout the field of view. On the right, EibD-induced bacteria appear to clump together, indicative of autoaggregation.
Figure 4. Light Microscopy Observation of Autoaggregation: Induced bacteria were stained with safranin and visualized using a light microscope at 1000x magnification. Pictures were taken through the eyepiece using a phone camera. As can be seen on the left, control E. coli BL21 bacteria are evenly distributed throughout the field of view. On the right, EibD-induced bacteria appear to clump together, indicative of autoaggregation.
Effect of Induction Concentration on Agglutination: Mid-exponential phase BL21 EibD-transformed cultures were induced with 0%, 0.001%, or 0.01% L-rhamnose and grown for 2 hours. 2 mL of pelleted bacteria was resuspended in PBS, briefly mixed, and the optical density at 600 nm was monitored for 30 minutes.
Figure 5. Effect of Induction Concentration on Agglutination: Mid-exponential phase BL21 EibD-transformed cultures were induced with 0%, 0.001%, or 0.01% L-rhamnose and grown for 2 hours. 2 mL of pelleted bacteria was resuspended in PBS, briefly mixed, and the optical density at 600 nm was monitored for 30 minutes.
EibD has previously been shown to be capable of non-immune binding to IgG10. The premise of our biosensor depends heavily upon this capability. Therefore, it was necessary to demonstrate that expression of BioBrick BBa_K4130000 in E. coli BL21 produces functional EibD capable of binding IgG. The EibD-harboring strain was induced with 0%, 0.001%, or 0.01% L-rhamnose and cultured for 2 hours. Bacteria were harvested and washed in PBS two times. Subsequently, bacteria were incubated in 200ug/mL mouse IgG-FITC for 24 hours. Bacteria were pelleted and washed twice, followed by fixation and staining for DNA by NucBlue. (Figure 4).
Comparing Figure 6A-C, as inducer concentration is increased, large clumps of bacteria become apparent. This is suggestive of autoaggregation, and further supports the quantitative data in Figure 5, and qualitative data in Figures 3 and 4.
Comparing Figure 6A and D, B and E, and C and F indicates localization of the IgG-FITC protein and the DNA. Throughout all induction levels, IgG-FITC signal is weak, except for exceedingly bright spots that do not colocalize with the DNA. These bright spots likely represent aggregates of the IgG-FITC protein in solution. The weaker “blots” of FITC fluorescence do, however, co-localize with the DNA stain in Figure 6, panels A-C. This suggests that IgG-FITC may be interacting with the EibD-expressing bacteria as observed in previous research.
IgG-FITC Co-Localizes With EibD-Induced Bacteria: Harvested bacterial cultures induced with 0%, 0.001%, and 0.01% rhamnose were incubated with mouse IgG-FITC for 24 hours and then washed. Fixed bacteria were stained for DNA using NucBlue Fixed Cell ReadyProbes Reagent and imaged. Panels A and D, B and E, and C and F correspond to the same region.
Figure 6. IgG-FITC Co-Localizes With EibD-Induced Bacteria: Harvested bacterial cultures induced with 0%, 0.001%, and 0.01% rhamnose were incubated with mouse IgG-FITC for 24 hours and then washed. Fixed bacteria were stained for DNA using NucBlue Fixed Cell ReadyProbes Reagent and imaged. Panels A and D, B and E, and C and F correspond to the same region.

In order to impart specificity towards a particular compound into the whole-cell bacterial biosensor, the EibD-expressing bacteria must be incubated with an antibody for the compound of interest. This will allow binding of the EibD protein to the antibody, thus creating a bacterium coated in antibodies. The intended application of our biosensor in the maple syrup industry is to detect buddy maple sap, which is associated with increases in free asparagine. Therefore, we incubated our EibD-expressing bacteria with an anti-asparagine antibody (Immusmol). To test the universality of our biosensor, we also incubated our EibD-expressing bacteria with an anti-GFP antibody.
Figure 7 displays the effect of incubation with the anti-asparagine antibody on autoaggregation. Bacteria were incubated with or without 2 mg/mL antibody in PBS for 20 hours. The bacteria were resuspended briefly and the O.D.600 was measured over 30 minutes. Experiments were performed in triplicate, and the average fraction O.D.600 was calculated for each time point. The error ribbon on the graph represents 1 standard deviation from the mean (n=3). The addition of the anti-asparagine antibody significantly reduced the amount of autoaggregation (Figure 7). This is consistent with the hypothesis that the antibodies may compete with the homophilic interactions between EibD proteins causing autoaggregation. The data suggests that the anti-asparagine antibody has successfully been bound by the EibD protein on the surface of the bacteria.
Figure 8 displays the effect of incubation with anti-GFP antibody on autoaggregation. Bacteria were treated identically as the anti-asparagine antibody treated samples (2 mg/mL antibody in PBS for 20 hours). Incubation with the anti-GFP antibody resulted in a significant reduction in autoagglutination. This is consistent with the results of incubation with the anti-asparagine antibody, and suggests that the anti-GFP antibody has successfully been bound by the EibD protein on the surface of the bacteria.
Bacteria were incubated with or without 2 ug/mL anti-asparagine antibody in PBS for 20 hours. The bacteria were resuspended briefly and the O.D.600 was measured over 30 minutes. Experiments were performed in triplicate, and the average fraction O.D.600 was calculated for each time point. The error ribbon on the graph represents 1 standard deviation from the mean.
Figure 7. The Effect of Incubation with Anti-Asparagine Antibody on Autoaggregation: Bacteria were incubated with or without 2 ug/mL anti-asparagine antibody in PBS for 20 hours. The bacteria were resuspended briefly and the O.D.600 was measured over 30 minutes. Experiments were performed in triplicate, and the average fraction O.D.600 was calculated for each time point. The error ribbon on the graph represents 1 standard deviation from the mean.
Bacteria were incubated with or without 2 ug/mL anti-GFP antibody in PBS for 20 hours. The bacteria were resuspended briefly and the O.D.600 was measured over 30 minutes. Experiments were performed in triplicate, and the average fraction O.D.600 was calculated for each time point. The error ribbon on the graph represents 1 standard deviation from the mean.
Figure 8. The Effect of Incubation with Anti-GFP Antibody on Autoaggregation: Bacteria were incubated with or without 2 ug/mL anti-GFP antibody in PBS for 20 hours. The bacteria were resuspended briefly and the O.D.600 was measured over 30 minutes. Experiments were performed in triplicate, and the average fraction O.D.600 was calculated for each time point. The error ribbon on the graph represents 1 standard deviation from the mean.

Referencing the schema in Figure 1, the next step in building a biosensor for GFP is to incubate with known concentrations of GFP and observe the resulting autoaggregation. According to our model, due to the size and relative complexity of the structure of GFP, polyclonal antibodies should bind to multiple GFP locations. Since the polyclonal antibodies are attached to the bacteria via EibD, these multivalent antibody-GFP interactions should cause an increase in the clumping/autoaggregation behavior of the bacteria. The amount of autoaggregation should also be a function of the amount of GFP in the system.
To test our model, anti-GFP antibody-coated bacterial cultures were incubated with 0uM, 0.96uM, and 3.85uM GFP in PBS overnight for 16 hours. The cultures were briefly resuspended and the O.D.600 was monitored for 30 minutes. There appears to be an overall decrease in autoaggregation with the addition of any concentration of GFP (Figure 9). However, due to differences in original autoaggregation profile between samples, a more accurate measure of the data is to subtract the fraction O.D.600 after addition of GFP from the fraction O.D.600 from before the addition of GFP (of the same sample). This will eliminate any sample-to-sample variability that may contribute to false interpretations of data. The re-analyzed graph can be seen in Figure 10, displaying that the autoaggregation profile of the samples varies as a function of GFP concentration. When taken together with Figure 9, the data indicates that increasing concentrations of GFP results in decreases in autoaggregation.
Interestingly, the data are not supportive of the hypothesis that addition of GFP will lead to bivalent antibody-GFP interactions, increasing autoaggregation. Therefore, a new molecular model was proposed to better fit the data (Figure 16). In this new model, addition of the antibody does compete away homophilic EibD-EibD interactions. However, as indicated by the data in Figures 5 and 6, this does not result in zero autoaggregation. This may be because of additional aggregation that is imparted by the antibodies themselves. While the EibD proteins may no longer be interacting, the antibodies may be interacting with each other. This is supported by previous research that has indicated that the immunoglobulin Greek-key beta sandwich folding has been shown to be susceptible to edge-edge association1. Additionally, the complementarity determining regions of antibodies responsible for binding to antigens are often composed of hydrophobic and electrostatic residues, which can also contribute to aggregation2,3. These antibody-antibody interactions may not be as strong as the EibD-EibD interactions, causing a reduction, but not full absence, of autoaggregation. In this new model, addition of the antigen, GFP, may therefore cause a decrease in autoaggregation by out-competing the antibody-antibody interactions and increasing steric hindrance. This new model may explain the decrease in autoaggregation observed in Figures 9 and 10.
Overall, the data from Figure 10 demonstrates that we have been successful in creating a sensitive biosensor for the detection of GFP.
Bacterial cultures were incubated with 2ug/mL anti-GFP antibody for 3.5 hours and then washed, removing any unbound antibody. Subsequently, bacteria were resuspended in 0uM, 0.96uM, or 3.85uM GFP and incubated for 12 hours. Bacteria were briefly resuspended and the O.D.600 was measured for 30 minutes.
Figure 9. Effect of Incubation with GFP on Autoaggregation: Bacterial cultures were incubated with 2ug/mL anti-GFP antibody for 3.5 hours and then washed, removing any unbound antibody. Subsequently, bacteria were resuspended in 0uM, 0.96uM, or 3.85uM GFP and incubated for 12 hours. Bacteria were briefly resuspended and the O.D.600 was measured for 30 minutes.
Effect of GFP Concentration on Autoaggregation: </b>Bacterial cultures were incubated with 2ug/mL anti-GFP antibody for 3.5 hours and then washed, removing any unbound antibody. The O.D.600 was measured for 30 minutes (Fraction O.D.600 pre-GFP) Subsequently, bacteria were resuspended in 0uM, 0.96uM, or 3.85uM GFP and incubated for 12 hours. Bacteria were briefly resuspended and the O.D.600 was measured for 30 minutes (Fraction O.D.600 post-GFP). The difference between Fraction O.D.600 pre-GFP and post-GFP was calculated and plotted as a function of time for each sample.
Figure 10. Effect of GFP Concentration on Autoaggregation: Bacterial cultures were incubated with 2ug/mL anti-GFP antibody for 3.5 hours and then washed, removing any unbound antibody. The O.D.600 was measured for 30 minutes (Fraction O.D.600 pre-GFP) Subsequently, bacteria were resuspended in 0uM, 0.96uM, or 3.85uM GFP and incubated for 12 hours. Bacteria were briefly resuspended and the O.D.600 was measured for 30 minutes (Fraction O.D.600 post-GFP). The difference between Fraction O.D.600 pre-GFP and post-GFP was calculated and plotted as a function of time for each sample.

According to the original model of the designed biosensor (Figures 1 and 2), synthesis of BSA beads would be necessary for the detection of a small molecule like asparagine. This is because there is a low likelihood that multiple antibodies would be able to attach to the same asparagine molecule and cause autoaggregation. Such a problem would be common to all small molecules, and limit the universality of our biosensor. To solve this problem, we created BSA-asparagine “beads” (BSA-Asn beads). The central component of the beads is the protein BSA, which is crosslinked via glutaraldehyde to asparagine molecules. The glutaraldehyde reagent is composed of two carbonyl groups, one of which will bind the amine groups found on lysine side chains in BSA, and the other to the free amine group on asparagine. This will result in approximately 35 asparagine molecules crosslinked to each BSA protein. According to the original model, addition of the BSA-Asn beads to antibody-coated bacteria would cause autoaggregation, which could then be competed away by free asparagine.
BSA-Asn beads were synthesized by reacting BSA with glutaraldehyde and asparagine at 30C and after incubation, quenched with Tris. The cross-linking reaction proceeds via double reductive amination mechanism with iminium ions as intermediate. Three different incubation times were tested to find the optimal period of incubation when enough iminium ions were produced in the reaction mixture; thus maximizing cross-linking rate. The gel in Figure 11 shows our primary attempt at synthesizing the BSA-Asn beads under different conditions. The extensive smearing of bands in our experimental sample lane 7 suggests that the highest rate of cross-linking of BSA with glutaraldehyde and possibly asparagine was achieved at 0.05% glutaraldehyde with an incubation time of 30 minutes. This condition was later used to produce a larger quantity of the beads.
SDS-PAGE of BSA-Asn beads synthesis under different conditions
Figure 11. SDS-PAGE of BSA-Asn beads synthesis under different conditions

For our first attempt, we avoided using BSA+Glutaraldehyde control because we believed that it would produce large clumps of BSA cross-linked with other BSA molecules. However, we decided to add that extra control along with our large-scale synthesis of the beads to eliminate the possibility that glutaraldehyde has a similar effect to the BSA without the asparagine. However, the gel in Figure 12 shows the SDS-PAGE result of that experiment. We noticed the length and smearing pattern of our negative control with BSA+Glutaraldehyde matches exactly with our experimental samples. This result does not nullify the possibility of successful cross-linking of asparagine with BSA, however, crosslinking between BSA molecules may be occurring as well.
Large-scale Synthesis of BSA-Asn beads
Figure 12. Large-scale Synthesis of BSA-Asn beads

Before the addition of BSA-Asn beads to our whole-cell bacterial biosensor, we desalted the beads using a size-exclusion column (micron columns) to get rid of unreacted asparagine and glutaraldehyde, since these molecules may interfere with our biosensor detection since our biosensor reacts to different concentrations of asparagine.

According to the original model (Figures 1 and 2) of our biosensor, BSA-Asparagine “beads” would be necessary to detect a small molecule such as asparagine. Addition of the beads was expected to increase bacterial aggregation due to multiple bacteria binding the same bead. Thus, we designed experiments to examine the aggregative effect of addition of BSA-Asparagine beads (Figures 13 and 14) to the system.
We first ran a control experiment to examine the effect of addition of the beads themselves (Figure 13). EibD-expressing bacteria were incubated with 0.015uM and 0.15uM BSA-Asn beads alone. There is a slight dis-aggregation of the bacterial cultures, in a near negligible amount.
Bacterial cultures were incubated in PBS, 0.015uM beads in PBS, or 0.15uM beads in PBS for 6 hours. Bacteria were briefly resuspended and the O.D.600 was measured for 30 minutes.
Figure 13. Effect of Addition of Beads without Antibody: Bacterial cultures were incubated in PBS, 0.015uM beads in PBS, or 0.15uM beads in PBS for 6 hours. Bacteria were briefly resuspended and the O.D.600 was measured for 30 minutes.

We next examined the effect of the addition of the beads while in the presence of the antibody-incubated bacteria (Figure 14). According to the original biosensor model, this should cause an aggregation event between the bacteria. However, there is no observable response to the addition of the beads (Figure 14). This could be due to several reasons: 1) the antibody is not recognizing the asparagine bound to the beads, 2) there are self-self interactions formed between the antibodies, reducing the aggregation response upon addition of the beads, 3) bead synthesis failed, 4) taking into account the slight disaggregation in the control experiment (Figure 13), there may have been an aggregation response that simply “canceled out” the disaggregation response, resulting in the appearance of no response. Ultimately, it is likely that all three possible explanations contribute to the lack of response to the addition of BSA-Asparagine beads.
Bacterial cultures were first incubated with 2ug/mL antibody for 16 hours. Subsequently, bacterial cultures were incubated in PBS, 0.015uM beads in PBS, or 0.15uM beads in PBS for 6 hours. Bacteria were briefly resuspended and the O.D.600 was measured for 30 minutes.
Figure 14. The Effect of Addition of Beads with Antibody-Coated Bacteria: Bacterial cultures were first incubated with 2ug/mL antibody for 16 hours. Subsequently, bacterial cultures were incubated in PBS, 0.015uM beads in PBS, or 0.15uM beads in PBS for 6 hours. Bacteria were briefly resuspended and the O.D.600 was measured for 30 minutes.

Of greatest importance to the overall project goal is the determination of whether our biosensor can detect the differences between various concentrations of free asparagine. Accordingly, bacteria were incubated with anti-asparagine antibody for 16 hours, washed in PBS, incubated with 1.5uM beads for 6 hours, and incubated with free asparagine (30mM, 100mM, or 300mM) for 12 hours (Figure 15). Upon addition of the free asparagine, there is a disaggregation event. This is consistent with the results observed for GFP (Figures 9 and 10). Again, this result is contrary to what was expected with our original model of the biosensor, but is consistent with the new model (Figure 16), explaining that there are interactions between the antibodies, being out-competed upon addition of the antigen.

Importantly, the disaggregation event appears to be asparagine concentration dependent. The 30mM asparagine sample exhibited significantly less disaggregation than the 100mM and 300mM asparagine samples. There was no significant difference between 100mM and 300mM, indicating that these concentrations were out of the range of the biosensor. These experiments suggest that we have successfully created a biosensor for asparagine.

Bacterial cultures were first incubated with 2ug/mL antibody for 16 hours. Subsequently, bacterial cultures were incubated in 1.5uM beads in PBS for 6 hours. Bead-incubated bacteria were then incubated with 0mM, 3mM, 30mM, or 300mM asparagine. Bacteria were briefly resuspended and the O.D.600 was measured for 30 minutes. The gray ribbons represent the 95% confidence interval for the fitted line.
Figure 16. Sensitivity of Biosensor to Varying Concentrations of Asparagine Bacterial cultures were first incubated with 2ug/mL antibody for 16 hours. Subsequently, bacterial cultures were incubated in 1.5uM beads in PBS for 6 hours. Bead-incubated bacteria were then incubated with 0mM, 30mM, 100mM, or 300mM asparagine. Bacteria were briefly resuspended and the O.D.600 was measured for 30 minutes. The gray ribbons represent the 95% confidence interval for the fitted line.

Some of the experiments did not support the original hypothesized biosensor model. Specifically, upon addition of antigen, a decrease in aggregation was observed instead of a predicted increase in aggregation. This observation can be observed by our updated model below.
a) uninduced bacteria have relatively few interactions with one another b) upon induction of EibD expression, the bacteria “autoaggregate” together c) addition of the antibody outcompetes most of the EibD homophilic interactions, but also contributes to aggregation due to antibody-antibody interactions d) addition of the antigen outcompetes remaining interactions, reducing the observed autoaggregation
Figure 16. Hypothesized Mechanism of Action of Universal Whole-Cell Biosensor: a) uninduced bacteria have relatively few interactions with one another b) upon induction of EibD expression, the bacteria “autoaggregate” together c) addition of the antibody outcompetes most of the EibD homophilic interactions, but also contributes to aggregation due to antibody-antibody interactions d) addition of the antigen outcompetes remaining interactions, reducing the observed autoaggregation

In this new model, addition of the antibody does compete away homophilic EibD-EibD interactions. However, as indicated by the data in Figures 7 and 8, this does not result in zero autoaggregation. This may be because of additional aggregation that is imparted by the antibodies themselves. While the EibD proteins may no longer be interacting, the antibodies may be interacting with each other. These antibody-antibody interactions may not be as strong as the EibD-EibD interactions, causing a reduction, but not full absence, of autoaggregation. In this new model, addition of the antigen, GFP, may therefore cause a decrease in autoaggregation by out-competing the antibody-antibody interactions. This new model may explain the decrease in autoaggregation observed in Figures 9, 10, and 14.

Conclusions and Future Work


Team Saptasense has made strides toward the development of a novel, antibody-based universal whole-cell bacterial biosensor. When applied to the detection of the protein GFP and the small molecule asparagine, our biosensor was successful in differentiating between different concentrations of the molecules of interest. In further experiments, we hope to apply our new technology to detect other compounds such as biomarkers for disease or environmental hazards. We hope to offer our strain as a “kit” for other scientists to use with their desired antibody, enabling universal detection. Towards the central goal of our project, we intend for sugarmakers to be able to employ this technology for the accurate detection of asparagine in “buddy” sap.

References

  1. Richardson J.S., Richardson D.C. Natural beta-sheet proteins use negative design to avoid edge-to-edge aggregation. Proc. Natl. Acad. Sci. USA. 2002;99:2754–2759. doi: 10.1073/pnas.052706099.
  2. Wu S.J., Luo J., O’Neil K.T., Kang J., Lacy E.R., Canziani G., Baker A., Huang M., Tang Q.M., Raju T.S., et al. Structure-based engineering of a monoclonal antibody for improved solubility. Protein Eng. Des. Sel. 2010;23:643–651. doi: 10.1093/protein/gzq037.
  3. Wang X., Das T.K., Singh S.K., Kumar S. Potential aggregation prone regions in biotherapeutics: A survey of commercial monoclonal antibodies. mAbs. 2009;1:254–267. doi: 10.4161/mabs.1.3.8035.
  4. Chang, HJ., Zúñiga, A., Conejero, I. et al. Programmable receptors enable bacterial biosensors to detect pathological biomarkers in clinical samples. Nat Commun 12, 5216 (2021). https://doi.org/10.1038/s41467-021-25538-y
  5. Gui, Q., Lawson, T., Shan, S., Yan, L. & Liu, Y. The application of whole cell-based biosensors for use in environmental analysis and in medical diagnostics. Sensors 17, 1623 (2017).
  6. Leo, J. C., Lyskowski, A., Hattula, K., Hartmann, M. D., Schwarz, H., Butcher, S. J., Linke, D., Lupus, A. N., Goldman, A. 2011. The Structure of E. coli IgG-Binding Protein D Suggests a General Model for Bending and Binding in Trimeric Autotransporter Adhesins. Structure 19(7):1021-1030.
  7. Kylilis, N. et al. Whole-cell biosensor with tuneable limit of detection enables low-cost agglutination assays for medical diagnostic applications. ACS sensors, https://doi.org/10.1021/acssensors.8b01163 (2019).
  8. Riangrungroj, P., Bever, C.S., Hammock, B.D. et al. A label-free optical whole-cell Escherichia coli biosensor for the detection of pyrethroid insecticide exposure. Sci Rep 9, 12466 (2019). https://doi.org/10.1038/s41598-019-48907-6
  9. Garcia EJ, McDowell T, Ketola C, Jennings M, Miller JD, Renaud JB. Metabolomics reveals chemical changes in Acer saccharum sap over a maple syrup production season. PLoS One. 2020 Aug 20;15(8):e0235787. doi: 10.1371/journal.pone.0235787. PMID: 32817615; PMCID: PMC7444596.
  10. Sandt, C. H., and C. W. Hill. 2000. Four different genes responsible for nonimmune immunoglobulin-binding activities within a single strain of Escherichia coli. Infect. Immun. 68:2205-2214.

Dextran Module: Reducing Syrup Waste by Repurposing Ropy Syrup into Dextran Hydrogels For Agricultural Use



Background



Many types of textural defects may arise during the production of maple syrup that render the final product unfit for consumption. Sugarmakers that produce defective syrup must discard these syrups as food waste, leading to a significant financial loss 1. One such textural defect is known as “ropy syrup”, named for its ability to create strings of 10 cm or greater1 (Figure 1). This viscous, stringy texture is created by the presence of bacterial residue in the sap prior to evaporation. These bacteria are capable of using the sucrose and other sugars in the sap to produce larger polysaccharides, such as dextran, arabinogalactan and rhamnogalacturonan1. It is estimated that the production of ropy syrup alone has produced an economic loss of over $4 million USD since 20081.
Figure 1: Comparison of ropy and non-ropy maple syrups. (1A) The ropy syrup sample shows a 10 cm string characteristic of this type of defective syrup. (2A) Comparison of ropy syrup (left) and non-ropy syrup (right). Ropy syrup from Cornell Maple Program; non-ropy maple syrup from Wegmans Food Markets

Dextran is a high molecular mass homo-exopolysaccharide made from sucrose by lactic acid bacteria2. The molecule is a useful and versatile product in medical and biomaterial research. Dextran consists primarily of 𝛼-1→6 chain and 𝛼-1→3 branching linkages between the glucose monomers2. The long chain structure allows the molecule to effectively crosslink with itself and other small molecules. High efficiency crosslinking using dextran is broadly employed by many industries. Some common uses of dextran include treating hemorrhage and burns, drug delivery, and radiological imaging3. The molecule is also used as a stabilizing and texturizing agent in food processing4. In industry, dextran is synthesized by growing the bacterial species Leuconostoc mesenteroides in sucrose-rich media and isolating the polysaccharide using acid hydrolysis1. >br>The synthesis of dextran is catalyzed by dextransucrase, an extracellular Class II enzyme expressed in these bacterial strains. Dextransucrase is a glucosyltransferase that catalyzes the transfer of glucosyl residues from sucrose to the dextran polymer, as described in the following chemical reaction5:
n Sucrose → n Fructose + Dextran [(glucose)n]
We believe this ropy syrup can be harnessed for other industrial purposes for its sucrose-rich content. Earlier this year, it was utilized in the creation of bioethanol fuels1; there is great potential in the recycling of the texturally defective syrup. In our project, a dextransucrase from L. mesenteroides NRRL B-512F is expressed in Escherichia coli and incubated in diluted off-flavor ropy syrup. This syrup serves as the sucrose media necessary for the bacteria to synthesize dextran. In using off-flavor syrup that is otherwise discarded, we have repurposed maple industry food waste for the novel use of producing valuable dextran.
Inspired by the versatility of dextran across various industries, Team Saptasense has created a way of repurposing the dextran created from off-flavor syrup to benefit the agricultural industry. The synthesized dextran from the syrup incubation is isolated and purified in order to create hydrogels. Hydrogels are cross-linked, hydrophilic polymers that have a variety of uses across many industries. Once formed, these gels are able to hold water without dissolving, making them a great tool for water and material delivery. We are particularly interested in creating suitable dextran hydrogels for use in the agricultural industry. Hydrogels are being increasingly utilized during germination and initial seedling growth to provide a consistent supply of water to the growing plant. Hydrogel usage is especially vital in areas of the world that experience recurring dry seasons and droughts6. We propose using the purified dextran from off-flavor syrup to create such hydrogels. Through this new technology, syrup that was once classified as waste is now given a novel function that will benefit its industry.

Experimental Design and Results


Dextransucrases are natively expressed only in lactic acid bacteria, most notably L. mesenteroides. The numerous variations of dextransucrases are categorized based on the weight of dextran they produce, the types of branching ɑ-linkages, and the strain of lactic acid bacteria in which they are natively found [5]. For this project, we chose to work with a dextransucrase encoded by a smaller gene for ease of assembly. The chosen dextransucrase is known as DexYG, encoded by the dexYG gene found in L. mesenteroides NRRL B-512F [7]. The 4,584 bp gene encodes a 1,527 aa protein that has a molecular weight of 170 kDa [7]. This enzyme is secreted by the host cell so that it can synthesize extracellular dextran when in a sucrose-rich environment.
The DexYG biobrick (composite part BBa_K4130013) created by Team Saptasense was designed for constitutive, extracellular expression of the DexYG enzyme by transformed E. coli. The part contains a strong viral T7 promoter (part BBa_I712074) and ribosomal binding site (part BBa_B0034) for increased expression in vivo. The DexYG gene was optimized for translation in E. coli and contains a double transcription terminator sequence (part BBa_B0015).
Due to its relatively large size, the dexYG gene was purchased in two parts, dexYG1 and dexYG2, of equal length. To facilitate successful ligation, the fragments were PCR amplified using overlapping primers. The final fragments contained >20 bp 5’ and 3’ overhangs, respectively, that were complementary to one another. Additionally, overlapping primers were used to create overhangs on each fragment complementary to the linear pSB1C3 vector. To assemble the plasmid, we used the NEBuilder® HiFi DNA Assembly Cloning Kit from New England Biolabs. Similarly to a Gibson assembly, the NEBuilder® HiFi is able to assemble multiple fragments of DNA together with high efficiency given that the fragments contain overlaps that are complementary to one another. Once the assembly reaction was complete, we confirmed that our DexYG plasmid was completely assembled using gel electrophoresis. Transformation of E. coli was done using chemically competent DH5-ɑ from New England Biolabs.
The DexYG biobrick was successfully ligated to the linear vector to form a fully assembled plasmid, as confirmed by gel electrophoresis (Fig. 2). Transformation of the plasmid into E. coli could not be completed due to time constraints; however, ropy syrup samples were acquired and used as the media from which dextran was isolated. Through this ropy syrup experimentation, we are able to show how the dextran, once produced by the recombinant bacteria, can be utilized for the creation of hydrogels.
Figure 2:Gel image of the final plasmid assembly reaction product using the NEBuilder® HiFi DNA Assembly Cloning Kit.

Dextran is classified as a homo-exopolysaccharide; that is, it is a long saccharide chain of glucose monomers that is synthesized extracellularly by lactic acid bacteria. Therefore, it can easily be purified from a culture using common exopolysaccharide isolation procedures. These procedures involve numerous days of precipitation and centrifugation followed by purification by dialysis and size-exclusion chromatography. To isolate dextran from our cell cultures and ropy syrup alike, we first added 95% ethanol following brief centrifugation at 12,000 x g before allowing the samples to precipitate overnight at 4℃. The samples were centrifuged at 12,000 x g to produce our polysaccharide pellet, which was redissolved in water before undergoing a second ethanol precipitation step. The final pellet was redissolved in water and underwent a 24 hour dialysis. We used dialysis tubing with a 3,500 Da molecular weight cut off to remove smaller molecular weight saccharides from the dextran. The retentate was dried slowly in an oven for 12 hours to isolate the dextran product.
Figure 3: Precipitation of dextran out of solution following centrifugation and the addition of 95% ethanol. Non-ropy maple syrup is used as a negative control.

50 mL of ropy syrup was used during this isolation protocol. 3.2 mg of dried product was isolated from the dialyzed retentate (Figure 4). If we apply this isolation protocol to a full gallon (3.78 L) of ropy syrup, then we would expect to isolate 64 mg of dextran. Given that over 500 barrels of ropy maple syrup have been produced per year in Canada alone since 2014 [9], and each barrel can hold 30-45 gallons (CDL Sugaring Equipment, St. Albans, VT, USA), then, at a minimum, 960 g of dextran can be isolated from ropy maple syrup per year in Canada. If recombinant DexYG E. coli were to be incubated in the syrup to upregulate the production of dextran, then the amount of dextran that can be isolated from the ropy syrup would significantly increase. This shows that there is great potential in harnessing the defected ropy syrup as an alternative substrate for the isolation of large quantities of dextran.
Figure 4: Dextran isolated from ropy syrup

The synthesized dextran will be used to create hydrogels for use with seed germination in the agricultural industry. A hydrogel is a crosslinked hydrophilic polymer that is highly absorbent, maintains structure, and doesn’t dissolve in water. Unfortunately, the high production cost limits the advantages of hydrogels. Therefore, our aim was to produce hydrogels by crosslinking dextran isolated from the ropy syrup with the least number of crosslinkers (less reagents thus cheaper), to make affordable and efficient hydrogels. We have chosen N,N′-methylenebisacrylamide (MBAm) as our hydrogel crosslinker. MBAm has been previously used as a crosslinker in dextran-based hydrogels [6] and has been shown to be a safe and effective crosslinker in the creation of hydrogels specifically for agricultural purposes [7].
To create these hydrogels, solutions of 10% (w/v) and 20% (w/v) dextran were created by dissolving dextran in 2.8M NaOH solution. Varying percentages (w/w) of MBAm were added to each solution to create unique hydrogels of various dextran and MBAm compositions. These gels were allowed to solidify overnight at 25℃ before swelling in deionized water at room temperature for 24 hrs. The mass of water absorbed by the gels was measured after the swelling period. Various assays were performed on the swelled gels to characterize their ability to successfully retain and carry water and nutrients to growing plants. These assays included a swelling test in solutions of various pH values to simulate different chemical compositions in soil and a diffusion test to determine how gel pore size created through MBAm crosslinking affected how well nutrients could travel through the gel.
Our first five hydrogels were made with 20% (w/v) dextran and 40%, 50%, 60%, 70%, 80% (w/w) MBAm crosslinker, respectively, which is consistent with hydrogels made in previous literature [8]. While the gels did form and held the shape of the disc in which they were formed, structurally they were hard (not gel-like), brittle, and more translucent-opaque rather than transparent. This structure and formation of gels can be reasoned by the overabundance of crosslinker, producing a brittle hydrogel (Fig. 5).
Figure 5: 20% dextran gels made with 40%, 50%, and 60% MBAm after incubation at 25℃. The gels remained opaque, indicating an overabundance of the MBAm crosslinker.

After initial results of the hydrogels, we decided to remake three 20% dextran gels using the following decreased w/w crosslinker amounts: 10%, 20%, and 30% MBAm, respectively. The second set of gels appeared to be more gel-like as they were transparent, held their shape, and were not as brittle as the initial gels with higher concentration of crosslinkers, therefore the decreased amounts of MBAm were qualitatively better forming hydrogels (Figure. 6A).
Figure 6: Photographs of the hydrogels created by Team Saptasense. (6A) 20% (w/v) dextran hydrogels; (6B) 10% (w/v) dextran hydrogels; (6C) “Upside down” test performed on the 10% (w/v) dextran hydrogels.

Following the crosslinking, the hydrogels were immersed in water for overnight bath to get rid of the excess MBAm that didn’t crosslink with dextran. After the overnight water bath, the three gels plus a 0% MBAm control gel were massed to find their water absorbance capacity. All three gels showed very similar water absorption capabilities as they retained water ranging from 2.66 g to 2.84 g (Fig. 7). The water absorption capacity is one of the most essential properties of a hydrogel as it characterizes the holding capacity of the gel and thus predicts its uses. Therefore, based on the water holding capacity we chose to move forward with our hydrogel assays using 20% dextran and 10% MBAm. The water absorption difference between the highest crosslinker concentration (30%) and the lowest MBAm concentration (10%) was calculated as 7.41%. This difference compared to the overall cost of production between the two hydrogels convinced us that the 10% crosslinking agent was much better for the dextran hydrogels for two reasons: one that it would reduce the amount of MBAm used per hydrogel, therefore reducing cost, and two the lesser mass of chemical crosslinker, the more organic is the hydrogel. Therefore, the 20% dextran and 10% MBAm was identified as the better composition that was keeping the gels simple while still delivering a relatively large amount of water to seeds and seedlings.
After observing the effects of various concentrations of the cross-linking agent, we were curious to observe the effect of varying concentrations of dextran in producing hydrogels. Thus, our next steps included making hydrogels using 10% dextran. We primarily chose 10%, as our overall goal was to create the best characteristic hydrogel (water holding capacity, pore size, pH stability, and diffusion) at the cheapest price and thus with the least amount of the reactants. To test this, three gels were created using 5%, 10%, and 20% MBAm crosslinker. Similarly to the successful 20% gels, the 10% dextran gels formed with a gel-like structure, transparent-translucent in composition, and held their structure when turned upside down (Fig. 6B, 6C). All of these properties, therefore, initially suggested that a 10% dextran hydrogel is able to form a gel very similar to a 20% dextran hydrogel. The 10% hydrogels were then quantitatively tested for their properties by immersing them in an overnight incubation of water to get rid of excess MBAm and to measure the water holding capacity post the incubation period. Following the overnight incubation, the three gels plus a 0% MBAm control were massed to calculate water absorbance. The 10% dextran with 20% MBAm showed the greatest absorbance of water (1.95 g) compared to the other 10% dextran gels (Fig. 7).
Figure 7: The water absorbance capabilities of both 10% and 20% dextran hydrogels created using varying percent MBAm.

This mass of absorbed water is comparable to 2.66 g absorbed by the previously chosen 20% dextran with 10% MBAm gel. For this reason, we chose to move forward with our assays using 20% MBAm crosslinker in our 10% dextran gels.

This assay was performed to determine the swelling capabilities of the hydrogels in different environments. Solutions of various pH values were used to simulate the diverse chemical compositions of soil. The four pH values were chosen to represent a wide variety of soil conditions without being too acidic or basic, which is not viable for plant growth. The swelling ratios of the 10% and 20% MBAm gels remained similar across all four solution pH values tested (Fig. 8). For pH 2.5, pH 5.5, and pH 7, the 20% MBAm gel swelled more than the 10% MBAm gel; the opposite observation was made for pH 8.5. For all four solution pH values, the 30% MBAm gel had the highest swelling ratio and therefore increased retention of the aqueous solution. The pH 7 swelling parallels the previous absorption data collected for the 20% dextran + 30% MBAm gel (Fig. 7), except in the previous assay, the water absorption for all three 20% dextran gels was similar. A likely cause of the discrepancy is that the pH swelling assay allowed only 3 hours of soaking compared to the 12 hours in the absorption assay. Further replicates of both assays would need to be performed to determine the swelling ratio of a 20% dextran + 30% MBAm gel in pH 7 solution. Ultimately, the data shows that the 10% and 20% dextran hydrogels are capable of absorbing a consistent amount of water across a wide range of pH values, and suggests that the gels are able to be implemented into agricultural work in diverse environments.
Figure 8: Calculated swelling ratios of three 20% dextran hydrogels using 10%, 20%, and 30% MBAm in aqueous solutions of four different pH values.

For structurally analyzing the hydrogels, we observed the pore sizes and pore density using microscopy. The pore sizes and density provide information on the rate of diffusion. The larger the pore sizes with closer proximity, the faster the rate of diffusion. The rate of diffusion determines the rate of which water and nutrients would travel through the hydrogel, which is an essential characteristic of the hydrogel and its use in enhancing maple seed germination. Small samples of the 20% dextran + 10% MBAm and 10% dextran + 20% MBAm gels were imaged at 10X magnification to characterize their respective pore sizes and calculate the average number of pores in a specified area. The 10% dextran gel (Fig. 9A) showed an increased pore density compared to the 20% dextran gel (Fig. 9C).
Figure 9: Images of dextran hydrogel samples at 10X magnification (9A) 10% dextran + 20% MBAm; (9B) 15% dextran + 15% MBAm; (9C) 20% dextran + 10% MBAm

In addition, the pores in the 10% dextran gel have an average pore width larger than that in the 20% dextran gel (Fig. 10).
Figure 10: Average pore diameter of the three tested hydrogels. The 15% and 20% dextran gels had an average pore diameter (0.13 mm) less than that of the 10% gel (0.20 mm). Error bars represent one standard deviation from the mean diameter.

Because the 20% dextran gel maintained its structural integrity post-swelling compared to the 10% dextran gel (Fig. 11), we decided to create an intermediary 15% dextran gel that would resemble a 20% gel in its structure while also maintaining the increased porosity of a 10% gel.
Figure 11: Comparison of 20% and 10% dextran gels after overnight water bath. The 20% gel is better able to maintain its structure after water absorption. This observation motivated us to create an intermediate 15% dextran gel.

The cross linker also plays a significant role in determining the structural integrity of the hydrogel, therefore the final 15% dextran gel used an intermediate 15% amount of MBAm crosslinker to provide it with a structure stronger than that of a 10% MBAm crosslinked hydrogel. As predicted by the patterns of a 10% and 20% gel, the 15% dextran hydrogel does show an increase in the pore density compared to the 20% dextran gel, but the average pore diameter remained the same as the 20% gel (Fig. 9B, 10). From these results we can conclude that the concentration of dextran may have an effect on the overall pore density within the gel. Further testing is needed to determine if dextran concentration has a significant impact on the average gel pore diameter. Since the 10% and 20% dextran gels had similar water retention capabilities, it is likely that the differences in pore size and density will have a greater effect on the ability of the gels to deliver this water and other nutrients to the plants.

A diffusion assay was performed on the swelled gels to observe how the gel pore sizes could affect how nutrients travel through the gel to the seeds. For beneficial agricultural application, hydrogels must be able to release what they absorb to the seeds in a controlled manner. Surrounding the plant roots with too many or too few nutrients and water can be detrimental to growth [7]. Application of colored food dye was used to simulate soil nutrients traveling through the swelled gels (Fig. 12). The 10% dextran gel showed the fastest diffusion rate (0.016 mm/sec) whereas the 20% dextran hydrogel has a slower diffusion rate (0.011 mm/sec) (Fig. 13). This is supported by the qualitative pore analysis described above. The 10% dextran gel contains an increased number of large pores compared to the 20% dextran gel; therefore, the dye should be able to travel through it faster. The 15% dextran + 15% MBAm gel was also tested in this assay and showed a diffusion rate of 0.014 mm/sec that is also supported by the pore images (Fig. 13). As FD&C Blue #1 dye is 749.9 g/mol (National Center for Biotechnology Information, PubChem Compound Summary), we can conclude that our hydrogels are capable of rapid diffusion of nutrients of similar molecular weight.
Figure 12: Photos of the 10% dextran + 20% MBAm gel and 20% + 10% MBAm gel before and after the competition of the diffusion assay.

Figure 13: Diffusion rates of the three hydrogels tested in the diffusion assay. A slight decrease in the diffusion rate is observed as the percent dextran used in the hydrogel increases.

We performed a germination assay to determine whether seed germination is possible on our gels. Ten chia seeds were placed on 9 g of swollen gel. Two controls were created: one where seeds were grown with just water, and one where the seeds grew on a damp paper towel. 5 mL of additional water was added to all five conditions on each day of growth (Fig. 14). On the third day after addition of the seeds to the conditions, we only observed the germination of one seed on the 10% dextran gel (Fig. 15); germination in the other four conditions was not observed by this day . These results show that germination of seeds is possible using our dextran hydrogels.
Figure 14: Image of the five germination assay conditions. The top two petri dishes represent the control conditions.

Figure 15: Germinating chia seed on 10% dextran gel

Conclusion



Dextran is a widely utilized exopolysaccharide with application across various industries. It is the molecule responsible for the stringy texture found in ropy syrup, an increasingly common off-flavor maple syrup that is responsible for millions of dollars in losses over the past decade. Team Saptasense has shown that there is great potential in harnessing this texture-defective ropy maple syrup which would otherwise be discarded as food waste. We have shown that the numerous barrels of ropy syrup produced each year in the U.S. and Canada are capable of producing over 1 kg of dextran per year through isolation techniques alone; this number can increase exponentially with the use of recombinant bacteria to convert the remaining sucrose in the syrup to dextran. We also demonstrate that this isolated dextran can be an incredibly useful material for the production of hydrogels. The hydrogels we have created are capable of absorbing large quantities of water relative to their sizes across a wide range of pH values. We also show that seeds are able to germinate on the gels, supporting the practicality of using these gels in the agricultural industry.

References



  1. de Medieros Dantas, J.M. et al., Bioethanol Production as an Alternative End for Maple Syrup with Flavor Defects. Fermentation 2022, 8: 58. https:// doi.org/10.3390/fermentation8020058
  2. Díaz-Montes, E. Dextran: Sources, Structures, and Properties. Polysaccharides 2021, 2, 554–565. https://doi.org/10.3390/ polysaccharides2030033
  3. Miao, K.H.; Guthmiller, K.B. Dextran. National Library of Medicine via StatPearls, 2022. PMID: 32491563
  4. Zhang, H., Hu, Y., Zhu, C. et al. Cloning, sequencing and expression of a dextransucrase gene (dexYG) from Leuconostoc mesenteroides. Biotechnol Lett 2008, 30, 1441–1446. https://doi.org/10.1007/s10529-008-9711-8
  5. Dols, M.; Remaud-Simeon, M. et al. Characterization of the Different Dextransucrase Activities Excreted in Glucose, Fructose, or Sucrose Medium by Leuconostoc mesenteroides. Appl Environ Microbiol 1998, 64(4), 1298-1302. https://doi.org/10.1128/AEM.64.4.1298-1302.1998
  6. Dhanapal, V. et al. Design, synthesis and evaluation of N,N1-methylenebisacrylamide crosslinked smart polymer hydrogel for the controlled release of water and plant nutrients in the agriculture field. Materials Today: Proceedings 2021, 45(2): 2491-2497. https://doi.org/10.1016/j.matpr.2020.11.101
  7. Zhang, H. et al. Cloning, sequencing and expression of a dextransucrase gene (dexYG) from Leuconostoc mesenteroides. Biotechnol Lett 2008, 30, 1441–1446. https://doi.org/10.1007/s10529-008-9711-8
  8. Imren, D. et al. Synthesis and characterization of dextran hydrogels prepared with chlor- and nitrogen-containing crosslinkers.J. Appl. Polym. Sci., 102: 4213-4221. https://doi.org/10.1002/app.24670
  9. Pelletier, M. et al., Ropy Maple syrup. Centre ACER, 2018. https://mapleresearch.org/wp-content/uploads/ropy-maple-syrup-presented-at-the-namsc-tech-sessions-2018.pdf

Choline Module



Background


The maple syrup season in Upstate New York generally starts in early February and can last for three to eight weeks depending on different weather conditions or the species of tree being tapped [1]. As the seasons change, the weather warms up, causing the maple trees to begin the budding process. This budding process leads to changes in the chemical composition of sap within the tree, which when boiled down to make syrup forms compounds that cause an off-taste to appear. This taste, which is usually described as cabbage-like, renders the syrup unsellable and results in a profit loss for the sugarmakers [2]. However, the presence of these buddy chemicals is not currently detectable in the sap, meaning that the sugarmakers have no way to know if they will be producing buddy syrup until time and energy has already been invested in making the syrup.

Application to Detection of Buddy Maple Sap


One of the molecules associated with buddy syrup is choline. Choline (aka bilineurine) is an organic compound containing a N, N, N-trimethylethanolammonium cation and has a molecular weight of 104.17 g/mol [8]. It is an important metabolite in humans (as a precursor to the neurotransmitter acetylcholine), but it is found in maple sap as well [8]. Right before the maple trees start to bud, the choline concentration increases in the sap from 0.1 uM to around 20 uM [7]. This makes choline an appropriate choice of molecule to detect when sap is buddy. Additionally, choline can be broken down into glycine betaine and hydrogen peroxide [9]. Team Saptasense has developed an electrochemical biosensor that utilizes the enzyme choline oxidase (similar to using glucose oxidase in a glucometer) to detect choline concentrations in sap, thereby determining if the sap is fit for further processing.

Description


Biological Parts

We obtained choline oxidase by expressing the choline oxidase gene from Arthrobactor globiformis in Escherichia coli. Our A. globiformis choline oxidase biobrick contains the coding sequence for the enzyme plus a 6x Histidine tag on the C-terminal end for easy protein purification (BBa_K13004). Because the goal of choline oxidase expression was to produce large amounts of enzyme, we designed it to include the choline oxidase part with a strong T7 promoter (BBa_I712074) and a strong ribosome binding site (BBa_B0034) for high protein yield. Our part also contains a TAA double terminator (BBa_B0015). We designed the part with restriction sites for EcoRI, XbaI, SpeI, and PstI corresponding with the pSB1C3 plasmid so we could easily insert the part into the plasmid using 3A assembly.

Experiments/Results


The choline biobrick (BBa_K130005) was inserted into a CIP-treated pSB1C3 plasmid by 3A assembly, and cloned into E. coli highly competent NEB 5-alpha. We selected this strain for cloning because of its high transformation efficiency [3]. We confirmed successful transformation of our genes of interest by performing colony PCR using our biobrick prefix and suffix primers, and by extracting and sequencing our plasmids.
Figure 1. Results of transformation of the choline oxidase biobrick (BBa_K413005) into NEB 5-alpha. Left: selection plate containing colonies of NEB 5-alpha E. coli into which the choline oxidase biobrick has been transformed. Right: Results of colony PCR for the choline oxidase biobrick displayed on agarose gels. In all three colonies, there is a band at the expected size of 1906 bp.
Following successful cloning of the choline biobrick, we transformed the choline oxidase plasmid into E. coli BL21(DE3) for protein purification. We used this strain to maximize protein production using the IPTG-inducible T7 promoter on our biobrick (BBa_I712074). By adding differing amounts of IPTG to growing cultures of choline oxidase-transformed E. coli BL21(DE3), we differentially increased the production of choline oxidase by the cells. Samples of each culture were run on a SDS-PAGE gel to see which concentration of IPTG produced the maximum amount of choline oxidase production (Fig. 2). After a small-scale induction test demonstrated that the optimal concentration of IPTG for choline oxidase induction is 0.5 mM (Fig. 2), we performed a large-scale induction to generate our supply of enzymes.
Figure 2. SDS-PAGE gel of the cellular lysates of choline oxidase-transformed E. coli BL21 (DE3) induced at various concentrations of IPTG. Successful induction of choline oxidase was confirmed by the band at around 63.5 kDa that is thick in all the induced samples and thinner in the uninduced samples.
As recipients of the Promega Grant, we obtained the Promega MagneHis Purification Kit to use in our His tag purification process. Upon successful induction of choline oxidase expression (an example of which can be seen in lanes 3-7 of Fig. 2) we purified our enzymes using the MagneHis Protein Purification System, which uses nickel-coated magnetic beads to capture histidine-tagged proteins [4]. Following the purification, we buffer-exchanged the proteins to remove the imidazole from our sample using Microcon columns.
Figure 3. SDS-PAGE gel of samples from all steps of the purification of choline oxidase. Lane 9 contains the purified choline oxidase that we used in our activity assays (see below).

After purification, we began testing our choline oxidase in an activity assay to test whether our choline oxidase biobrick was producing a functional version of the enzyme. We did this using the Cayman Chemical Hydrogen Peroxide kit, which produces a colorimetric readout corresponding to the amount of hydrogen peroxide in a sample. Because hydrogen peroxide is a byproduct of the oxidation of choline, we can use the kit to measure the activity of our enzyme. The results of this assay could be observed both qualitatively in the darkness of the color change (Figure 4) and quantitatively by plotting the absorbance over the course of the reaction (Figures 5, 8, and 9).
Figure 4. Example qualitative end results of an activity assay. The top row contains the hydrogen peroxide standard with no choline. From left to right, the uM concentration of hydrogen peroxide in the well is increasing, leading to the observably darker pink color. The bottom row shows the assay containing the enzyme choline oxidase and varying concentrations of choline. The leftmost well contains samples with the highest concentration of choline and the rightmost well contains no choline. The darker pink wells contain more hydrogen peroxide and thus correspond to the wells loaded with a higher concentration of choline.
To investigate activity of the purified choline oxidase, we measured hydrogen peroxide production in the presence and absence of the choline substrate (Figure 5). 100mM choline conditions resulted in high relative absorbance, indicating the production of the hydrogen peroxide byproduct of choline oxidation. Comparatively, conditions lacking choline substrate display constant low absorbance, conveying a lack of hydrogen peroxide production. These results demonstrate the activity of purified choline oxidase, and the functionality of BioBrick BBa_K413005.
Figure 5. Initial activity testing of choline oxidase. 1 uM samples of choline oxidase were incubated with either no choline (red line) or in saturating levels of choline (black line). Both samples also contained the colorimetric hydrogen peroxide indicator. While there was no observable change in absorbance in the sample with no choline, the sample with saturating levels of choline produced a demonstrable increase in absorbance. In this sample, the absorbance peaked at around 1.5.

Critical to the function of the proposed biosensor, is the sensitivity of choline oxidase to differing choline concentrations. To examine sensitivity to choline, we tested the activity of the purified enzyme at various substrate concentrations. We calculated the initial reaction rates for each choline concentration by fitting the first four minutes of each reaction to a linear fit model. The slope of each linear fit equation was graphed alongside the choline concentration (Figure 6). As substrate (choline) concentrations were increased, a parallel increase in catalytic activity was observed. These data were fitted according to Michaelis Menten kinetics, producing a standard curve that can be used to predict samples of unknown choline concentrations.
To test the functionality of our bioassay, we designed an experiment that would replicate the conditions in which an end-user would interact with our product. “Test samples” containing choline in various concentrations were applied to our bioassay and the choline concentration was predicted. Briefly, various choline concentrations were prepared, and 0.1 uM choline oxidase was added. The rate of the proceeding reaction was measured via the colorimetric assay. Measured reaction rates were applied to the experimentally determined Michaelis-Menten curve (Figure 6), and a predicted choline concentration was produced (Figure 7). Within the range of 0-100uM choline, our assay produced accurate predictions with minimal percent error. At higher concentrations, error in the predictions increased. This indicates that in its current iteration, the bioassay is best suited for lower concentrations of choline. This range of detection corresponds similarly to the range of concentrations of choline found in maple sap. While normal maple sap has negligible concentrations of choline, buddy sap exhibits concentrations closer to 20uM choline. Thus, these data demonstrate that the range of our choline biosensor is appropriate for the detection of buddy maple sap.
Figure 6. Michaelis-Menten curve generated from the incubation of .1 uM choline oxidase in varying concentrations. The plot generated a Vmax value of 9236 uM/sec and a Km 102.4 sec-1. The curve fits strongly with the data, having an R-squared value of .9603. This plot was used to determine the choline concentrations of the test solutions (Figure 7).
Table 1. The predicted values of the mystery choline solutions by the Michaelis-Menten curve to the actual concentrations of the mystery choline solutions.

All of the previous assays were performed in a phosphate buffered saline (PBS) buffer that does not provide sap-like conditions. Our next assay therefore investigated whether choline oxidase would still be functional in a solution more similar to sap. We tested our enzyme in mock sap solutions containing different weight percentages of sucrose dissolved in water, since sap is made up of about 90% water and anywhere between 1-5% sucrose [6]. Figure 8 shows that the choline oxidase enzyme is still functional in 1-5% sucrose concentrations and performs similarly to choline oxidase enzyme when it is in the non-sap-like buffer.
Figure 7. Assay testing the functionality of choline oxidase in media containing different concentrations of sucrose. All samples contained .1 uM choline oxidase and 200 uM choline. The positive control was a sample run in PBS buffer instead of sucrose media. The negative control contained no choline, and was run in PBS buffer. The experimental concentrations of 1%, 1.5%, and 5% were chosen because maple sap contains a sucrose content of 1-5% [6]. There is no discernible difference in the size or shape of any of the experimental or positive control curves, indicating that sucrose does not affect the activity of choline oxidase.
Since the choline oxidase enzyme shows similar activity in sucrose media and in PBS buffer, we next investigated whether choline oxidase could still distinguish between varying concentrations of choline in the sucrose buffer (Figure 9). We incubated 1 uM choline oxidase with varying concentrations of choline in a 1% sucrose solution. We observed that solutions containing smaller concentrations of choline produced very little signal while the solutions containing higher concentrations of choline produced higher levels of signal at a faster rate than solutions containing lower concentrations (Fig. 9). These results demonstrate that sucrose does not affect the impact of varying choline concentrations on choline oxidase activity.
Figure 8. Choline oxidase activity at different concentrations of choline in a 1% sucrose solution. In all samples, the concentration of choline oxidase was .1 uM. The magnitude of absorbance was directly related to the concentration of choline, with the highest concentration producing the largest increase in absorbance and the remaining absorbance values decreasing in descending order of choline concentration. These results suggest that our choline oxidase is able to distinguish between varying choline concentrations in sap-like conditions.

Conclusions


Team Saptasense has made strides toward the development of a novel, enzymatic biosensor for the small molecule choline. Our sensor is capable of reliably predicting unknown concentrations of choline in the range of 0-100uM. Importantly, this range of reliability encompasses the concentrations of choline found in normal and “buddy” sap. Further, we have demonstrated the choline-dependent activity of the enzyme in sap-like conditions. When taken together, our data indicate the suitability of our biosensor for choline detection in maple sap and prediction of “buddiness”. To our knowledge, we have developed the first known biosensor for the detection of “buddy” maple sap. In future experiments, we will improve the accessibility of our technology to sugarmakers, implementing hardware to produce an easily-interpretable electrochemical readout.

References


  1. County, Washington County NY Tourism. “New York Maple Season Is Here.” Washington County NY, 18 Mar. 2022, https://washingtoncounty.fun/new-york-maple-season-2021/#:~:text=Tap%20Into%20Spring%3A%20New%20York%20Maple%20Season%20is%20Here&text=Every%20February%20and%20March%2C%20producers,3%2D8%20weeks%20on%20average.
  2. Miller, David. “What Causes Buddy Syrup and What Can Be Done to Prevent It?” Maple Research, Maple Syrup Digest, 1 Mar. 2021, https://mapleresearch.org/pub/buddy0321/.
  3. “NEB® 5-alpha Competent E. coli (High Efficiency).” New England Biolabs, New England Biolabs, Inc. https://www.neb.com/products/c2987-neb-5-alpha-competent-e-coli-high-efficiency#Product%20Information
  4. Godat, B, et al. “MagneHis™ Protein Purification System: Purification of His-Tagged Proteins in Multiple Formats.” Attractive Protein Purification, 2003.
  5. Fan, Fan, and Giovanni Gadda. “On the Catalytic Mechanism of Choline Oxidase.” Journal of the American Chemical Society, vol. 127, no. 7, 2005, pp. 2067–2074., https://doi.org/10.1021/ja044541q.
  6. Taylor, Fred H. Variation in Sugar Content of Maple Sap. University of Vermont and State Agricultural College, Mar. 1956, https://www.uvm.edu/~uvmaple/sapsugarcontentvariation.pdf.
  7. Garcia, E. Jose, et al. “Metabolomics Reveals Chemical Changes in Acer Saccharum SAP over a Maple Syrup Production Season.” PLOS ONE, vol. 15, no. 8, 2020, https://doi.org/10.1371/journal.pone.0235787.
  8. “Showing Compound Choline (FDB000710).” FooDB, The Metabolomics Innovation Centre, 17 Sept. 2020, https://foodb.ca/compounds/FDB000710#:~:text=Choline%2C%20also%20known%20as%20bilineurine,many%20plants%20and%20animal%20organs.
  9. Gadda, Giovanni. “Choline Oxidases.” PubMed.gov, U.S. National Library of Medicine, 18 July 2020, https://pubmed.ncbi.nlm.nih.gov/32951822/#:~:text=Choline%20oxidase%20catalyzes%20the%20four,cells%20to%20counteract%20osmotic%20pressure.

Sarcosine Module



Background


An increasingly prevalent problem sugarmakers face is the production of off-flavor buddy syrup, a cabbage-tasting syrup created from sap tapped late in the season. According to the U.S. Department of Agriculture, buddy syrup “fails to meet requirements of Grade A syrup” [29] and is therefore deemed unfit for human consumption. This occurs when maple trees near the end of the season and start developing buds. During this period, known as bud break, particular amino acids and amino acid derivatives including sarcosine, methionine, asparagine, choline, lysine, and others increase in concentration [1]. Research suggests that these compounds act as methyl (-CH3) donors and therefore, at rising concentrations, alter the tree's metabolic profile to the point of affecting the flavors of the syrup boiled from sap [2]. Currently, no preventative detection methods for buddy syrup exist, meaning that sugarmakers boil down thousands of gallons of sap, causing up to 10% loss of annual income [3]. To combat this problem, our team has targeted 3 of these small molecular compounds - sarcosine, choline, and asparagine - with unique detection assays for determining each molecule’s presence in sap. By combining detection methodologies for each of these small molecules into one unit, Team Saptasense provides a comprehensive biosensor kit that can test whether or not sap is buddy, and therefore prevent waste caused by the unnecessary production of buddy, inedible syrup.
To target sarcosine, a known buddy flavor-causing compound, our team developed an electrochemical aptasensor. The aptasensor utilizes previously established aptamers, single-stranded DNA (ssDNA) sequences that can fold to form pin-loop structures, that selectively and strongly bind to sarcosine. To precisely detect sarcosine levels in maple sap samples, aptamers are placed on electrodes to exploit the folding “aptasensor” mechanism, which causes a change in electrode resistance upon the aptamer binding its target, sarcosine. This binding event induces a difference in the electrical current that can be detected as a signal by an electrical circuit mechanism, like an Arduino Uno circuit board. Collectively, our aptasensor has 3 different components that contribute to sarcosine detection: the aptamer component, an electrode component, and the electrochemical modification components.

Application to Detection of Buddy Maple Sap


Sarcosine (aka N-methylglycine, Fig 1.) is an amino acid derivative with a molecular weight of 89.09 g/mol and with an amino acid-like structure maintaining the typical 𝜶 carbon, carboxyl group, and hydrogen in the R-group position[4]. However, the molecule has a secondary amine replacing the position of the typical 𝜶 amino group, similar to glycine. It is a byproduct of creatine hydrolysis [8] and an intermediate and byproduct of both glycine and choline metabolism [4]. Because sarcosine is additionally found to increase in concentration from 0.01uM to 0.12 uM over the course of the 4-6 week maple season, we decided to aim our sensor at detecting this molecule to sense levels of sap “buddiness”[1, 9].
Figure 1. Depicts the composition of the sarcosine molecule [4].


Aptasensor Design Components


Our sarcosine aptasensor has 3 primary components in order to detect sarcosine concentrations in sap. In order to develop an effective and fully operating aptasensor we utilized the following design components:

  • Basic Biological Parts: We applied the strong, specific binding properties of our novel BioBricks, BBa_K4130014 and BBa_K4130015, to detect free sarcosine molecules. To learn more about our modified ssDNA aptamers, Sar09-3 and Sar11-5, see the Biological Parts section below.
  • Screen Printed Electrodes (SPEs): SPEs are “disposable, low-cost and portable [electrochemical] devices” that are made by screen printing conductive ink materials onto plastic or ceramic [21]. To learn more about Screen Printed Electrodes, the types we used, and how we used them, see our Hardware page
  • Electrochemical Modifications: Modifications are critical to refining the detection threshold of our sensor to detect micromolar concentrations of sarcosine. Modifications by materials such as nanoparticles can increase electrode surface area and conductivity, and therefore improve detection sensitivity [21]. See the Method Design Process section below to learn more.
    • Electrochemical methods: In order to detect our target molecule electrochemically, our team utilized cyclic voltammetry (CV) and chronoamperometry, and differential pulse voltammetry (DPV) techniques by means of a potentiostatic instrument that circulates and measures electrical current. See the Method Design Process section below to learn more about how we used our potentiostat, and to learn more about these electrochemistry was used to characterize our aptasensor, see our Hardware Page

Based on previous studies, when aptamers are deposited onto an electrode through dropcasting or electrodeposition* and introduced to a target molecule, the aptamer will fold and create steric hindrance that triggers a detectable change in electrical resistance on the electrode [6, 7, 10, 12, 13, 14,18, 25]. These electrodes are often modified with nanocomposite materials to immobilize aptamers to them and to improve their sensitivity. Change in resistance can then be modeled to a detection curve that produces a target-concentration readout. For the purposes of our aptasensor, this range would be between the thresholds of sarcosine in normal sap (low concentration) up to the amount of sarcosine in unusable buddy sap (high sarcosine concentration).
Figure 2. Graphic depicting the aptasensor design process

Biological Parts

Two basic biobricks are the primary biological components of this module. They are ssDNA aptamers for free sarcosine with a 3’ and 5’ amine modification, respectively, named Sar09-3 (see part registry [19]: BBa_K4130014) and Sar 11-5 (see part registry [20]: BBa_K4130015). Information about these basic parts, why we chose them, and how they were used for our project can be found below.

Aptamers are single strands of RNA (ssRNA) or, in our project’s case, DNA (ssDNA) that have selective binding capabilities to target proteins or small molecules [10]. Aptamers take on 3-dimensional structures in order to tightly bind to target molecules. These oligonucleotides can be synthesized in vitro with PCR and selected through a process called Systematic Evolution of Ligands by Exponential Enrichment (SELEX) [11]. Taking into consideration these methods and the efforts of previous iGEM teams [12], we prioritized the design, build, and testing of our composite aptasensor over the time-intensive process of synthesizing and selecting our own aptamers via SELEX. Therefore our team performed intensive research to select existing ssDNA sarcosine aptamer sequences that were then ordered to be synthesized by IDT, which ultimately streamline our experimental design.
A primary goal during the brainstorming and implementation stages of our project was to develop an expansive selection of buddy sap detection methods. We chose to target multiple molecules (sarcosine, choline, and asparagine) via three distinct detection methods. After designing modules for applying enzymatic and whole cell agglutination biosensors, we were inspired by the rapidly developing applications of aptamers in conjunction with electrochemical biosensors [22]. One particular study utilizing aptasensors for detecting tryptophan also drew us toward the usage of screen printed electrodes (SPEs), which you can read more about in the SPEs and Electrode Modification section below [7].

To date there have only been 3 total studies that have developed sarcosine aptamers. One of these studies, however, developed aptamer sequences based on glycine and did not determine a singular optimal aptamer sequence, while also lacking aptamer binding and specificity data [13]. Therefore, we selected two different aptamer sequences from the remaining sarcosine-binding aptamer studies to identify which aptamer would bind better to sarcosine in sap samples. The first study was conducted in 2018 and selected an aptamer with a dissociation constant of 134.8 ± 1.12nM in standard buffer (approx pH 7.5) conditions via a modified affinity chromatography capture-SELEX method, indicating strong binding properties [14]. The secondary structure of this aptamer, which we named Sar11-5 (Fig. 3) has the following sequence, 5’-3’: CTCAGTTCGGGACGACCACGCAAATACGAATAGTGTGAACGCGGGAGTCCCGAA*
Figure 3. Predicted secondary structure of Sar11-5 as modeled by the Mfold web server [14]. Highlighted in blue is the consensus sequence identified by Clustalx 1.8.3, which is shown to be responsible for binding to the target molecule[16].

Figure 4. Predicted secondary structure of Sar09-3 as modeled by the Mfold web server[6]. Highlighted in blue is the consensus sequence identified by the MEME suite, which is shown to be responsible for binding to the target molecule[16].

The second study, conducted in 2019, discovered a high-specificity aptamer via the GO-SELEX selection method. The selective aptamer indicated even stronger binding than that of the previous study, with a dissociation constant of 0.33 ± 0.05nM, and was developed in urine-like buffer conditions [6]. The secondary structure of this aptamer, which we named Sar09-3 (Fig. 4) has the following sequence, 5’-3’: TAGGGAAGAGAAGGACATATGATGTGCCGCGCTTCCCTTGCCGCTCAAAACAGACCACCCACTTTGACTAGTACATGACCACTTGA*
*Underlined portions indicate the consensus sequences of all developed sarcosine aptamers in the respective aptamer selection studies
Since freshly tapped sap can range between pH 3.9 to pH 7.9 [15], we also wanted to test two different aptamers in order to consider whether the more neutral [14] or acidic [6] selection conditions impacted the binding capabilities of aptamers in sap-like conditions.

In addition to selecting Sar09-3 and Sar11-5 based on binding criteria, we wanted to ensure that our aptamers would be compatible with the desired biosensor design and therefore chose to add an amine modification to their respective sequences. Modifications are desirable for applications like fluorescent dye labeling, increasing detection of aptamer folding, and for attachment to solid surfaces [17].To develop a reliable aptasensor mechanism, the aptamer must be able to properly stick to or immobilize on the electrode [18]. The amine modification ensures that the aptamers covalently bond to functional groups present on screen printed electrodes (SPEs) and the nanomaterials selected to modify and improve SPE performance. Therefore, we engineered Sar09-3 to be synthesized with a 3’ “Amino Modifier” (Fig. 5) and Sar11-5 with a 5’ “Amino Modifier C6” (Fig. 6) [17]. To learn more about how we chose the specific modification locations for our aptamers, see the Sar09-3 and Sar11-5 Part Pages [19, 20].
Figure 5. The 3’ amination applied to S09-3. The 3’ end was chosen for modification because of literature truncation testing that, based on ΔGo values, indicated 3’ truncation did not hinder aptamer folding [6].

Figure 6. The 5’ amination applied to Sar11-5. The 5’ end was chosen for modification instead of the 3’ end due to the literature usage of the aptamer with 5’ FAM labeling [14], suggesting the 5’ end is more ideal for modification.


SPEs and Electrode Modifications


To detect when our aptamer is bound to sarcosine, an electrode such as a screen printed electrode (SPE) is necessary to produce a quantifiable electrochemical signal. We were inspired to develop our sarcosine aptasensor to use SPEs in comparison to more expensive alternatives, similar to a recent tryptophan-aptasensor study [7]. In addition, we applied nano-material electrode modifications to our design as to improve SPE performance. This was similarly inspired by material usage of chitosan, carbon-based and gold-based nano-particles (MCWNTs and AuNPs respectively) that were used in the aforementioned study.


Essential Concepts

We executed electrochemical techniques such as cyclic voltammetry (CV), chronoamperometry, and differential pulse voltammetry to quantitatively characterize aptamers binding to ourSPEs. Brief descriptions of these methods, including how and why they are used, are discussed below.

Cyclic Voltammetry (CV) is an electrochemical technique that measures current under excess voltage conditions, and can be used to evaluate electrochemical activity and characterize materials. This is done by cycling the potential of a working electrode between two potential ranges. The y-coordinates of the graph will decrease in magnitude and approach the x-axis in the case of increased resistance, and the y-coordinates will increase in magnitude in the case of decreased resistance. For our project, we primarily applied this technique to characterize electrode modifications such as electrochemical reduction and deposition methods.
While this current is cycled through an electrode, the working electrode is immersed in a redox solution. Redox reagents are ionic compounds that, under application of an electrical current, undergo a reduction reaction that unfavorably introduces 1 electron to the compound, which reverts back to an oxidized state. One of these redox reagents, potassium ferricyanide, is commonly used for its ability to reduce to potassium ferrocyanide [31] and is the redox reagent we chose to utilize for our CV purposes. Methylene blue is also a commonly used redox reagent that we chose to employ.
Like CV, chronoamperometry is another electrochemical technique that depends on current and applies a square-wave potential to an electrode under excess voltage conditions. The current of the electrode fluctuates according to the diffusion of an analyte from the bulk solution toward the sensor surface. Similarly, it is run in the presence of a redox solution and can be used to examine the electrochemical activity and stability of electrocatalysts. This method was applied to aptamer binding event measurement rather than electrode characterization.
In addition to CV and chronoamperometry, differential pulse voltammetry, a potentiostatic method was used to offer some additional advantages in which the waveform is a series of pulses increasing over a linear baseline instead of producing square-wave potentials under voltage conditions. DPV is a pulse technique that is designed to minimize background charging currents. The base potential value is chosen and applied to the electrode and is increased with equal increments over a time period, during which the current is immediately measured before the pulse application and the end of the pulse forming a pulse wave by recording the difference between the start and the end of the pulse [32]. DPV is therefore a first derivative of a linear voltammogram in which the formation of a peak is observed for a given redox process. In general, DPV is a more sensitive method than the linear sweep methods (CV) because there is minimization of capacitive current. Therefore, the combination of CV, chronoamperometry, and DPV provides essential information to characterize our SPEs and aptamer binding.
Figure 7. the PalmSens4 potentiostat, the blue rectangular device on the black table, that was donated to our team’s usage being set up to in preparation of testing carbon ZP SPEs

In order to run analytical electrochemical techniques like those outlined above, a current must be consistently applied. These instruments are called potentiostats (Fig. 6), and more specifically contain multiple internal circuits that can generate, hold out, and measure such potentials and currents. Our team was donated a PalmSens4 Potentiostat in order to run these protocols, seen in the image above [33].

Dropcasting is the method of depositing a desired material onto a working electrode by means of pipetting it from solution. Though the technique is synonymous with pipetting, dropcasting avoids the direct touching of the electrode to preserve its structural integrity.
Electrodeposition is the method of depositing desired solutions onto a working electrode by means of submerging the working electrode in a solution of the desired material and applying either a static or CV/chronoamperometric current. This usually entails adding a redox reagent into solution to ensure proper electron transfer and material-to-electrode assembly.
Comparatively, dropcasting usually takes longer than electrodeposition to apply materials to the electrode due to the associated drying (and sometimes washing) steps. It is also less consistent due to the increased margin of human error, to include non-homogenous solution dispersion or over or under-covering the electrode surface area.
For the purposes of our project, aptamers were only dropcasted onto already modified electrodes. However, we tested both dropcasting and electrodepositing of our nano-composite materials to determine the most optimal electrode modifications, as seen in our Method Design Process below.

Method Design Process


Once we selected our project and target molecules at the end of the brainstorming stage, we searched to find the most effective manner of binding aptamers while increasing electrode conductivity and sensitivity. In order to optimize our procedure for developing a final, functional aptasensor, we evaluated and designed multiple protocols.

Table 1.
AuNPs on carbon SPEs
Why this protocol? The article this protocol was based on inspired our team’s initial development of SPE modification processes, which remained constant throughout the project until our final prototype development [7].
SPE Type Modification Materials Pros Cons
Carbon Gold Nanoparticles (AuNPs) Increases sensitivity, thiol binding properties ideal for aptamer immobilization Requires HAuCL4, a hazardous and expensive reagent
Multi Walled Carbon Nanotubes (MWCNTs) Increases electrode conductivity as well as surface area on the WE Expensive reagent making consumer accessibility non-feasible
Chitosan Maintains film-formation properties ideal for aptamer adhesion and NP layer assembly Difficult to disperse in solution for WE deposition

Results


After eliminating MWCNTs due to cost and replacing them with BSA, we chose not to use this protocol due to the price and hazardous nature of the AuNP synthesizing reagent, HAuCl4. Therefore this method did not reach the experimental stage. Because of this method, however, we wanted to continue using chitosan and rGO.

Chitosan, as seen above, is an amino-polysaccharide/biopolymer with high adhesion and film-formation properties shown to increase surface area and provide a hybridization matrix ideal for ionic and covalent binding interactions [23]. In addition, we decided to approach an alternative to MCWNTs, called reduced graphene oxide, for further method iterations. Reduced graphene oxide (rGO), is a 2D sheet of carbon atoms similar to that of graphene, but is derived from graphene oxide (GO) reduction, resulting in some functional group presence. It can be made via chemical, electrochemical, or biochemical reduction, and it is a cheaper alternative to graphene that maintains mechanical and thermal properties, in addition to electrical conductivity [24].
Table 2.
Chemically reduced GO (crGO)/chitosan/GA/aptamer on carbon SPEs
Why this protocol? After deciding not to use MCWNTs or AuNPs, we looked into Rochester 2021 BioSpire team’s SPE use and additionally found a study that also used chitosan in compound with reduced graphene oxide (rGO) and glutaraldehyde (GA) [25]. Based on that study, we decided to devise a novel protocol using those nano-materials to modify our aptasensor. Ferricyanide was the utilized redox reagent/mediator for this protocol.
SPE Type Modification Materials Pros Cons
Carbon SPEs: Both Zimmer and Peacock [ZP] and DropSens brands.ZP Hyper Value SPEs had significantly less surface area for drop-casting (pipetting) modification reagents/materials, but were significantly cheaper than DropSens at 99 cents per SPE [26]. However, DropSens electrodes (Fig. 9) maintain more surface area as well as cover other sensitive electrode components [27]. Reduced Graphene Oxide Increases electrode conductivity as well as surface area on the WE, cost-friendly modification More fragile and lower quality than alternative carbon based NP materials
Chitosan See Method 1 See Method 1
Glutaraldehyde Effectively crosslinks amine groups to one another, contributing to effective aptamer immobilization Hazardous reagent that non-selectively crosslinks; aggressively condenses amines with well known medical properties in tissue fixation [28].

Results


Figure 8. CV of an unmodified carbon SPE (blue curve) to a crGO/chitosan modified SPE (light grey curve). Aptamer immobilization or target solution testing not depicted by this curve. This was done with a DropSens carbon SPE

This method indicated that electrode modification with crGO increased conductivity of our aptasensor mechanism, which in turn increases electrode sensitivity. After CV was performed (10 cycles at a scan rate of 0.2 V/s) for each of the two conditions on the same SPE, an increase of 0.4 uA (Fig. 8) occurred as indicated by the increase in y-axes between the unmodified (bluish curve) and the crGO modified (greyish curve) electrode. When the aptamers were deposited (1uM solution) a minor increase in resistance of less than 0.1 uA (scan not depicted) occurred as expected, however, after incubating the electrode in multiple target molecule solutions (1uM, 100uM, and 1mM sarcosine to dH2O), there was no change in resistance, indicating either human error in dropcasting, chemical reduction of GO, or deficit in the detection threshold of the aptasensor itself.

Right, a ZP hypervalue carbon SPE (left) next to a DropSens carbon SPE (right). Left, The oven drying method of crGO. Middle, oven-dried crGO in 1% chitosan that would not disperse under excessive sonication or vortexing
Figure 9, 10, 11. Right, a ZP hypervalue carbon SPE (left) next to a DropSens carbon SPE (right). Left, The oven drying method of crGO. Middle, oven-dried crGO in 1% chitosan that would not disperse under excessive sonication or vortexing

Despite dropcasted crGO/chitosan solution increasing carbon SPEs conductivity, the film it formed was very fragile due the grainy and insoluble nature of the crGO. Initially, after reducing 4mg/mL of GO to rGO (Fig. 10), we approached an oven-based drying method that resulted in expected yield losses between 40-80% that created flakes incapable of homogeneous solution dispersion (Fig. 11). We confirmed that our GO was reduced to crGO by analyzing the loss of functional groups via IR spectra on the sample before and after reduction. After multiple trials of chemical reduction and trying different oven-drying conditions, we attempted lyophilization to reduce flake formation, which ultimately produced a powder-like substance that still had difficulties dispersing in solution with chitosan. Homogeneous dispersion is critical to producing reproducible results with equal layers of nanocomposite. However, we continued to utilize a 1% chitosan solution that was made by dissolving chitosan in 2% acetic acid, vortexing and incubating at 37 degrees C until dissolved, and then bringing it up to a pH of 5 using NaOH.
Ultimately, because of the time consuming nature of the synthesis and recovery of rGO, we employed three more protocols that eliminate this time consuming step.
*Tip: When working with rGO or GO you should always use filter tips so as not to stain or clog your pipettes with graphene, which increases the likelihood of contamination.

Because we wanted to continue working with rGO to improve electrode sensitivity, we developed an electrochemical approach in which to reduce GO without the excess equipment and materials that chemical reduction entails.
Table 3.
On-SPE electrochemical reduction of GO, followed by dropcasting chitosan + GA/aptamer (2 steps)
Why this protocol? By dropcasting water-dispersed GO on the WE, then running the SPE through a CV cycle to reduce the film to rGO, the reduction step is reduced by multiple hours and creates a more refined rGO film. The first method assembles the NP film in 2 steps where electrochemical reduction occurs before dropcasting chitosan. This protocol uses methylene blue as the redox reagent.
SPE Type Modification Materials Pros Cons
Carbon SPEs: Both ZP Hyper Value and DropSens Reduced Graphene Oxide See Method 2 See Method 2
Chitosan See Method 1 See Method 1
Glutaraldehyde See Method 2 See Method 2

Results


Figure 12. The CV plot indicating that rGO electrochemical reduction decreases in WE resistance. The red inner curves show the potential to current of unmodified carbon SPEs, and the yellow outer curves are rGO modified carbon SPEs. Aptamer immobilization or target solution testing not depicted by this curve.

This method indicated that the electrochemical reduction of GO to rGO on the electrode significantly increases electrode conductivity, in which the yellow curve indicates that the electrodeposited/reduced rGO SPE produces 0.3 mA more conductivity than the unmodified SPE in the red curve (Fig. 12). Run with the same CV conditions as Method 2, it is clear that this electrodeposition is both more efficient and effective (Fig. 13), resulting in nearly 750x magnitude of conductivity than the modifications done in Method 2 (0.3mA = 300uA, compared to 0.4uA).

To achieve this result, modification entailed the following before preparing aptamer immobilization.

  • Cleaning: Each carbon electrode was washed with DI water
  • Graphene oxide (GO):
    • approximately 5uL of 2mg/mL GO was dropcasted onto the electrode
    • 24 DPV cycles were ran in MB from 0 to -1.5 at 0.15V/s to electrochemically reduce the GO to rGO. Electrode was rinsed and dried with PBS
  • Chitosan solution for DPV:
    • a 1% chitosan solution was created by dissolving chitosan in 2% acetic acid, vortexing and incubating at 37C until dissolved. 5uL was dropcasted onto the electrode and let dry.
Figure 13. Carbon DropSense SPE with dried dropcasted GO before applied current (left) and after the current was applied, reducing the GO to rGO (right). The rGO appearance resembles the color of pencil lead, indicating possession graphene-like characteristics.

Conclusively, this method was applied primarily for characterization purposes, and demonstrates that the electrochemical reduction of GO onto an electrode produces an effective alternative to chemical reduction of the functional groups from GO to convert it into rGO. Despite the lack of aptamer-target, this method was the basis for the innovation of Method 4 seen below, which not only saved multiple hours of protocol time, but also saved on cost and amount of reagents used.

To combine these separate modification steps of rGO electrochemical reduction and chitosan electrodeposition, we developed a novel method for applying these modification steps simultaneously.
Table 4.
Novel On-SPE electrochemical reduction of GO in chitosan solution, followed by GA/aptamer (1 step)
Why this protocol? By dropcasting water-dispersed GO on the WE, then running the SPE through a CV cycle to reduce the film to rGO, the reduction step is reduced by multiple hours and creates a more refined rGO film. In comparison, this method combines the electrochemical reduction with electrodeposition by combining GO/chitosan in solution and running a CV electrodeposition to make the NP film in one step, in comparison to the previous step. We therefore developed a novel modification protocol based on our previous knowledge and application of modification methods. It also utilizes methylene blue.
SPE Type Modification Materials Pros Cons
Carbon SPEs: Both ZP Hyper Value and DropSens Reduced Graphene Oxide See Method 2 See Method 2
Chitosan See Method 1 See Method 1
Glutaraldehyde See Method 2 See Method 2
Methylene Blue (MB) It is a larger organic molecule with amino groups allowing gold and aminated aptamer binding. Studies show that incorporating MB with aptasensors increases the magnitude of folding effects [30] Could potentially oversaturate the sensor readout, bind to other MB molecules in the presence of crosslinker, or oversaturate with aptamers to the point of omitting any signaling of aptamer binding.

Results


Figure 14. DPV curves for aptamer binding testing. The red and dark blue curves show control chitosan/MB carbon SPEs ran in PBS, the green and purple curves show control chitosan/MB/aptamer carbon SPEs ran in PBS, and the orange and light blue curves show chitosan/MB/aptamer gold SPEs ran in 1M sarcosine/PBS solution. The orange and light blue peaks indicate successful binding events between Sar09-3 and sarcosine on the two separate electrodes.

Three conditions were run to test aptamer-sarcosine binding (Fig. 14). For context, the y-axis represents the current in µA and the x-axis represents the potential (V) of the DPV trial runs. For all trials, DPV was run for two cycles between 0.0V and 0.8V and used 1X PBS as the redox solution. Two control conditions were run in PBS where 2 trials were chitosan/MB modified gold SPEs (red and dark blue) and 2 trials were control chitosan/MB and aptamer modified gold SPEs (purple and light blue). The other 2 trials were chitosan/MB and aptamer modified gold SPEs (orange and green curve) that were incubated with 100uL of 1M sarcosine solution at room temperature for 15 mins.
After the two tests were run, results show that a significant resistance change of almost 3 uA (orange curve) and 1uA (green curve) for the respective trials occured after contact with the sarcosine solution. This demonstrates that Sar09-3 successfully bound to the target molecule, sarcosine. When compared to the other methods in this module, it has maintained the most positive and significant results, and is therefore established Method as the most successful aptasensor development protocol in this module.
This novel and effective carbon SPE modification was developed by performing 2 new steps. The values for DPV were the same for this modification method as Method 3.

  1. GO/Chitosan solution for DPV: 50% of 1% chitosan solution was combined with a 2G/L GO solution (make a final composition of a 0.5% chitosan, 1% acetic acid, 1G/L GO solution)
  2. Methylene blue intercalation: 10uL of 20mM Methylene blue was dropped on the dry modified carbon electrode and allowed to rest for 5 minutes before thorough washing with pH 8 Tris HCL buffer.
Further comparative analysis: Raman Spectroscopy

Another way we compared carbon SPE modifications to determine optimal deposition methods was by performing Raman Spectroscopy. Raman Spectroscopy (RS) is an important tool for investigating the unique “chemical fingerprints” of solid and liquid samples. It investigates the vibrational, rotational, and other low-frequency interactions in a molecule, providing characteristic information about the structure and formation of the sample. The principle of Raman, described as the “Raman Shift”, is the change in wavelength of the photon scattered based on the characteristic of the sample’s molecules [34]. Specifically, when our samples were exposed to a light source during RS, the light interacted with the sample molecules and either got absorbed or scattered (Fig. 15). Additionally, RS provides information based on the energy changes experienced by interacting photons, which is a result of the rotational or vibrational energies of a molecule. These energy changes are specific to the Carbon-Carbon chemical bond, and therefore provide information about the unique chemical properties of a molecule. Altogether, the recorded Raman shift and energy changes of our sample provided significant evidence for comparing the methods for reducing graphene oxide and SPE modification.

Figure 15.Raman Spectroscopy curve comparing material similarities between an electrode with electrodeposited chitosan/rGO, a crGO sample, a GO sample, and a rGO+chitosan dropcasted electrode (method 2)

Gold SPEs have very high conductivity compared to carbon SPEs; its thiol groups have functional binding properties ideal for immobilizing aptamers and other organic materials. These SPEs are, however, more expensive and are intended to assist product development to approach a more cost effective and therefore commercially desirable sensor. Regardless, we wanted to test whether the quality of the electrode would affect target signal readout and approached methods of applying gold SPEs.
Table 5.
Crosslinking aptamers intercalated with Methylene Blue to chitosan covered gold SPE
Why this protocol? Protocol Link: Gold SPEs are much more sensitive than carbon SPEs, but are slightly more expensive due to quality. When the aptamer binding event occurs and the aptamer folds, it will be bonded with MB, which allows for the folding event to maintain greater steric hindrance and therefore increase resistance (i.e. greater signal readout) of electric current passing through the electrode. We utilized gold rather than carbon SPEs due to their binding properties to amine groups as well.
SPE Type Modification Materials Pros Cons
Gold SPEs by DropSens. Methylene Blue (MB) See Method 4 See Method 4
Chitosan See Method 1 See Method 1
Glutaraldehyde See Method 2 See Method 2

Results


Figure 16.DPV curves for aptamer binding testing with MB. The red and dark blue curves show control chitosan/MB gold SPEs ran in PBS, the light blue and purple curves show control chitosan/MB/aptamer gold SPEs ran in PBS, and the orange and green curves show chitosan/MB/aptamer gold SPEs ran in 1M sarcosine/PBS solution. The orange peak indicates successful binding event between Sar09-3 and sarcosine.

Three conditions were run to test aptamer-sarcosine binding (Fig. 16). In all data, the y-axis represents the current in µA and the x-axis represents the potential (V) of the DPV trial runs. For all trials, DPV was run for two cycles between 0.0V and 0.8V and used 1X PBS as the redox solution. Depicted above are one experimental and two control conditions where two control trials were chitosan/MB modified gold SPEs (red and dark blue), and two control trials were chitosan/MB and aptamer modified gold SPEs (purple and light blue), and the two experimental trials were chitosan/MB and aptamer modified gold SPEs (orange and green) that were incubated with 100uL of 1M sarcosine solution at room temperature for 15 mins.
After the two tests were run, results showed a significant resistance change of more than 1.5 uA for the orange curve after contact with the sarcosine solution, indicating that the aptamer successfully bound to the target molecule. Generally, the curves are less favorable than that of the results in Method 4, reinforcing Method 4 as the more optimal method for aptamer immobilization in this module. Reasoning for the second test electrode not binding can be attributed to human error with the crosslinking or rinsing steps regarding aptamer immobilization.
For preparing the gold electrodes for this protocol, the below steps were followed.

  • Each gold electrode was washed with DI water and cleaned by electrochemical cycling in 0.1M sulfuric acid for 0.6 to 1.6V at scan rate 0.5V/s for 100 scans.
  • Chitosan was applied to the electrode by submerging the electrode in the 1% chitosan solution and applying a constant current of 25μA (200 μA/cm2) for 120s
  • Electrode was washed with DI water and then soaked in 1M NaOH for 5 minutes.
  • MB intercalation (as seen in Method 4)
Preparing Aptamers for Electrode Immobilization

For all aptasensor protocols, the aptamers needed to be in a solution compatible with electrode immobilization. Once the ordered aptamers were received, they were reconstituted in ddH2O and vortexed vigorously for 30 seconds (or until completely mixed) to make a 1M stock solution that was stored in -20oC for future usage. The aptamers were always thawed on ice and were stored in aliquots to prevent excessive freeze-thaw cycles. Regardless of electrode modifications, in order to execute the final step of aptamer immobilization, the WE of the SPE was always submerged in glutaraldehyde, quenched with 1X PBS, then left to completely dry. Once dry the aptamer solution was always dropcasted in volumes no greater than uL, then rinsed again in PBS once dried.


Conclusion and Future Directions


After applying four full cycles of collective protocol research, design, building, and execution, this module ultimately demonstrated that both of our aptamers were able to immobilize on screen printed electrodes(SPEs) modified with innovative nanocomposite materials, as well as showing that our basic Sar09-3 part can definitively detect sarcosine. While developing our sarcosine-specific aptasensor, we also devised a novel protocol that applies groundbreaking SPE modification research to accommodate the goal of achieving a cost-effective prototype. By using carbon SPEs, we offer a competitive alternative to more expensive electrode options on the market, like glassy carbon electrodes and the gold SPEs we tested for our aptasensor use. In addition, comparing the data between our new carbon-based SPE aptasensor to that of our gold aptasensor, the exposure of the same sarcosine concentration to both Method 4 electrodes and Method 5 electrodes revealed that Method 4 yielded a current change (avg. 1.7uA) over two times that in Method 5 (avg. 0.7uA). This suggests that carbon SPEs can achieve the high sensitivity and conductive qualities of more expensive, inaccessible electrode options like gold SPEs. Therefore Method 4 achieved the most significant data suggesting sarcosine binding, making it our team’s most preferred aptasensor development method. To learn more about the success of our aptasensor and how it was developed, please check out our Hardware Page.

Future goals of our project would entail application of more variable and intensive aptamer modification trials to achieve a more expansive characterization of the aptasensor. In addition, a comparison test to evaluate the differences in folding events between the experimentally tested Sar09-3 with that of the theoretical applied basic part, Sar11-5, would further optimize the fidelity of a final aptasensor prototype.

From these results, Team Saptasense introduces an opportunity to establish more inclusive and widely available engineering solutions for marginalized groups. This attributes not only to the local sugarmakers in our region, but to all small local businesses, disabled and indigenous communities, and to many more that we worked with throughout the course of our project.

References


  1. N’guyen, Guillaume Quang, et al. “A Systems Biology Approach to Explore the Impact of Maple Tree Dormancy Release on Sap Variation and Maple Syrup Quality.” Nature News, Nature Publishing Group, 2 Oct. 2018, https://www.nature.com/articles/s41598-018-32940-y.
  2. Garcia, E Jose, et al. “Metabolomics Reveals Chemical Changes in Acer Saccharum SAP over a Maple Syrup Production Season.” PloS One, Public Library of Science, 20 Aug. 2020, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7444596/.
  3. Miller, David. “What Causes Buddy Syrup and What Can Be Done to Prevent It?” Maple Research, Maple Syrup Digest, 1 Mar. 2021, https://mapleresearch.org/pub/buddy0321/.
  4. “Sarcosine.” National Center for Biotechnology Information. PubChem Compound Database, U.S. National Library of Medicine, https://pubchem.ncbi.nlm.nih.gov/compound/Sarcosine.
  5. Zhu, Guizhi, and Xiaoyuan Chen. “Aptamer-Based Targeted Therapy.” Advanced Drug Delivery Reviews, U.S. National Library of Medicine, Sept. 2018, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6239901/.
  6. Özyurt, Canan, et al. “A Highly Sensitive DNA Aptamer-Based Fluorescence Assay for Sarcosine Detection down to Picomolar Levels.” International Journal of Biological Macromolecules, Elsevier, 6 Feb. 2019, https://www.sciencedirect.com/science/article/pii/S0141813018364171?via%3Dihub#bb0175..
  7. Hashkavayi, Ayemeh Bagheri, and Jahan Bakhsh Raoof. “Ultrasensitive and Reusable Electrochemical Aptasensor for Detection of Tryptophan Using of [Fe(Bpy)3](p-CH3C6H4SO2)2 as an Electroactive Indicator.” Journal of Pharmaceutical and Biomedical Analysis, Elsevier, 3 Oct. 2018, https://www.sciencedirect.com/science/article/pii/S0731708518318065?via%3Dihub
  8. “Sarcosine.” Sarcosine - an Overview | ScienceDirect Topics, https://www.sciencedirect.com/topics/chemistry/sarcosine.
  9. “How Maple Syrup Is Made.” How Maple Syrup Is Made - Vermont Maple Sugar Makers, https://vermontmaple.org/how-maple-syrup-is-made.
  10. Liu, Ling Sum, et al. “Recent Developments in Aptasensors for Diagnostic Applications.” ACS Applied Materials & Interfaces, vol. 13, no. 8, 2020, pp. 9329–9358., https://doi.org/10.1021/acsami.0c14788.
  11. Drabek, Rafal. “Aptamers vs. Antibodies - Antibody Alternatives.” Base Pair Biotechnologies, 19 Jan. 2020, https://www.basepairbio.com/aptamers-vs-antibodies/.
  12. “BioSpire: A Diagnostic Microfluidic Device That Detects Sepsis in Sweat.” Team:Rochester/Results, University of Rochester 2021 IGEM, https://2021.igem.org/Team:Rochester/Results#Top.
  13. Guedez, Andrea. “Directed Evolution of Synthetic Riboswitches and a Leucyl tRNA Synthetase,” Texas Christian University, Ann Arbor, 2022. ProQuest, https://www.proquest.com/dissertations-theses/directed-evolution-synthetic-riboswitches-leucyl/docview/2663480714/se-2
  14. Luo, Yu, et al. “In Vitro Selection of DNA Aptamers for the Development of Fluorescent Aptasensor for Sarcosine Detection.” Sensors and Actuators B: Chemical, Elsevier, 23 Aug. 2018, https://www.sciencedirect.com/science/article/pii/S0925400518315430.
  15. Bates, Dewonica. “Monitoring Ph in Reverse Osmosis in the Maple Industry.” Hanna Instruments Blog, Hanna Instruments, 2 Aug. 2022, https://blog.hannainst.com/ph_maple_reverse_osmosis.
  16. Platt, Mark, et al. “Analysis of Aptamer Sequence Activity Relationships.” OUP Academic, Oxford University Press, 12 Nov. 2008, https://doi.org/10.1039/b814892a.
  17. “Attachment Chemistry / Linkers Modifications.” IDT, https://www.idtdna.com/site/Catalog/Modifications/Category/2.
  18. Oberhaus, Franziska V, et al. “Immobilization Techniques for Aptamers on Gold Electrodes for the Electrochemical Detection of Proteins: A Review.” Biosensors, MDPI, 28 Apr. 2020, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7277302/.
  19. https://parts.igem.org/Part:BBa_K4130014 (SAR09-3)
  20. https://parts.igem.org/Part:BBa_K4130015 (SAR11-5)
  21. “Screen-Printed Electrodes.” Encyclopedia, https://encyclopedia.pub/entry/8814.
  22. Huang, Jie, et al. “Advances in Aptamer-Based Biomarker Discovery.” Frontiers in Cell and Developmental Biology, Frontiers Media S.A., 16 Mar. 2021, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8007916/.
  23. Congur, Gulsah, et al. “Chitosan Modified Graphite Electrodes Developed for Electrochemical Monitoring of Interaction between Daunorubicin and DNA.” Sensing and Bio-Sensing Research, Elsevier, 19 Dec. 2018, https://www.sciencedirect.com/science/article/pii/S221418041830134X.
  24. “Reduced Graphene Oxide: An Introduction.” Graphene, https://www.graphene-info.com/reduced-graphene-oxide-introduction.
  25. Rezaei, Behzad, et al. “An Ultrasensitive and Selective Electrochemical Aptasensor Based on RGO-Mwcnts/Chitosan/Carbon Quantum Dot for the Detection of Lysozyme.” Biosensors and Bioelectronics, Elsevier, 9 May 2018, https://www.sciencedirect.com/science/article/pii/S0956566318303518.
  26. “Hyper Value Screen Printed Electrodes.” Zimmerandpeacock, https://www.zimmerpeacocktech.com/products/electrochemical-sensors/hyper-value-screen-printed-electrodes/.
  27. “DropSens Carbon SPEs.” ..:: Metrohm Dropsens ::.. Screen-Printed Electrodes, https://www.dropsens.com/en/screen_printed_electrodes_pag.html.
  28. “Crosslinking Applications: Thermo Fisher Scientific - US.” Crosslinking Applications | Thermo Fisher Scientific - US, https://www.thermofisher.com/us/en/home/life-science/protein-biology/protein-biology-learning-center/protein-biology-resource-library/pierce-protein-methods/crosslinking-applications.html.
  29. “United States Standards for Grades of Maple Syrup.” USDA, 2 Mar. 2015, https://www.ams.usda.gov/sites/default/files/media/MapleSyrupStandards.pdf.
  30. Tang, Chuyue, et al. “Methylene Blue Intercalated Aptamer to Amplify Signals toward Sensitively Electrochemical Detection of Dopamine Released from Living Parkinson's Disease Model Cells.” Sensors and Actuators Reports, Elsevier, 3 Feb. 2022, https://www.sciencedirect.com/science/article/pii/S2666053922000078.
  31. “Cyclic Voltammetry At Solid Electrodes .” Https://Www.asdlib.org/, Experiments in Analytical Electrochemistry, https://www.asdlib.org/onlineArticles/elabware/kuwanaEC_lab/PDF-19-Experiment1.pdf.
  32. “Differential Pulse Voltammetry.” Differential Pulse Voltammetry - an Overview | ScienceDirect Topics, https://www.sciencedirect.com/topics/chemistry/differential-pulse-voltammetry.
  33. “What Is a Potentiostat and How Does It Work?” Pine Research Instrumentation Store, 24 May 2021, https://pineresearch.com/shop/kb/theory/instrumentation/what-potentiostat-does/.
  34. https://www.sciencedirect.com/topics/neuroscience/raman-spectroscopy