Team:Thessaloniki/Demonstrate

<!DOCTYPE html>

Results

Kinetic Measurements

Currently, chemically synthesized DNA serves as the main material for DNA nanotechnology. However, the error-prone DNA synthesis process and the resulting imperfect oligonucleotides are believed to hinder the performance of dynamic DNA devices by causing side reactions. For example, “leak” reactions can result in the release of an output oligonucleotide even in the absence of a reaction trigger. Conversely, reactions often fail to reach the expected level of activation because some components do not trigger even in the presence of the intended input. To make the performance of DNA-based nanodevices comparable to their biological protein-based counterparts, such error modes need to be dramatically reduced.

In that accord, we’ve striven towards implementing our CRN into bacterial plasmids which could serve as a relatively cheap source of highly pure DNA for nanotechnological applications. Large amounts of DNA can be generated by replication in bacteria and the intrinsic proofreading abilities of living systems ensure the purity of the resulting DNA. In support of this rationale, there is proof in the literature that plasmid-derived gates outperform their synthetic counterparts even if the synthetic system is assembled from strands purified using polyacrylamide gel electrophoresis (PAGE)[1][2]. Likely, the improved performance of plasmid-derived gates is primarily due to the high purity of biological DNA. Synthetic DNA contains a variety of errors, in particular deletions that result in oligonucleotides of length n-1, and such side products are typically not completely removed in PAGE or high-performance liquid chromatography (HPLC) purification procedures. Additionally, it is a great way to openly distribute already created systems so that anyone can retrieve the already designed and tested gates effortlessly and cost-effectively.

Firstly, we performed a characterization of the reporter complex with varying degrees of output concentrations. Reporters were set at an excess concentration of 100nM.

Figure 1: Reporter characterization with varying concentrations of input

The fluorescence levels of signal C after correcting for bleaching at the measurement endpoint (200 min) shows a linear relationship with the initial concentration of signal C. From that data, we performed a calibration curve for our subsequent fluorescent data.

Figure 2: Bleaching of the fluorophore FAM after subsequent excitation during the course of kinetic measurements

The rate-limiting step of our CRN would, eventually, be the binding of the TF to the Input Gate, which would limit the working concentration of the gate in regards to the binding ability of the TF to its designated binding site. The best way to characterize the circuit's efficiency would be by titrating the Input Gate and analyzing its leakage and performance. Our model’s validation of the sequences through the Genetic Algorithm developed, indicates that no other binding site should be present in any other gates so that no other rate-limiting factors come into effect. Experiments were conducted with the concentration of Join and Fork Gates set at 20nM while auxiliary strands were set to 50nM and the Reporter complex at 100nM. Experiments were conducted at a set temperature of 37oC. The circuit leakage of plasmid derived and synthetic gates were also compared.

Figure 3: Circuit performance with varying concentrations of Input Gate for the CRN targeted at the p65 homodimer. The plasmid derived gates behave as intended after the proper digestion steps. Bacterial plasmid DNA can thus be used as a source for molecular programming applications. Additionally, the system’s leakage appears significantly reduced supporting plasmid derived gates as a better source for molecular computing logic gates.
Figure 4: Performance of the ELK1-designated circuit with varying concentrations of the Input Gate. Circuit leaks were observed to be similar while the circuit’s performance was observed to be better for similar time scales.

Afterward, we proceeded with the insertion of the Transcription Factor as input to the circuit. We predicted that the transcription factor would bind to the Input Gate, thus inhibiting the initiation of the strand displacement reactions. Accordingly, the circuit’s behavior would reflect a reduced input concentration. That concentration would indicate the number of input gates not bound by the TF giving us quantitative data on the TFs affinity to the corresponding sequence. We performed all experiments that included the TF as input with a set concentration of Input Gates at 20nM. Thus, we predict that the TF would interact with its binding site, inhibiting the strand displacement capabilities of the Input Gate and forcing the circuit to perform as if there were smaller concentrations of the Input Gate.

Figure 5: Comparison of the circuit’s kinetics with the p65 homodimer in regards to the titrations of the Input Gate’s concentration without the TF present. Input Gate was set at 20nM concentration.

The predicted shift in behavior is apparent. Thus, comparative estimates about the p65s homodimer binding affinity to its consensus binding site can be made.

Figure 6: Comparison of the ELK1 designated circuit when the TF is used as the rate-limiting factor. Once again, one can observe the change in kinetic behavior after the TF insertion, proof that the designed circuit can be assessed through our model’s Genetic Algorithm in order to receive various TFs as input. Input Gate was set at 20nM concentration as well

The same as above stands for the circuit’s behavior with the TF ELK-1 as input.

Figure 7: SNP comparison. While bound by the TF the Input Gates that were mutated in the p65 binding site exhibit different behavior. It is apparent that the change of the 10th nucleotide from cytosine to thymine retards fully the TF’s ability to bind to the site as the circuit performs identically as the one without the TF present for 20nM Input Gate. The other mutations exhibit gradient reduction of the TFs affinity which is always lower than the binding site’s consensus sequence.
Figure 8: Comparative results between different concentrations of Transcription Factor p65 as input to set concentrations of Logic Gates. The circuit can be used to not only predict the binding affinity of the Transcription Factor to a given sequence but, additionally, to quantify the concentration of Transcription Factors in a sample

To conclude, our data show verification of our proof of concept and demonstrates the ability of molecular programming to act as the analysis media for the intricate intramolecular and intracellular interplay between various components. As far as we know this is the first confirmation of DNA-only computation used in order to analyze DNA-protein interactions in excessive detail, providing valuable info for future analysis of such interplay for the elucidation of intracellular procedures involved in transcription, cellular behavior, and disease. Additionally, high-throughput analysis of such interactions can be facilitated by the intrinsic abilities of DNA circuitry in parallel programming and computation. We envision a much more expansive and complex analysis of such interactions in the future expanding from multi-protein complexes to drug-protein or drug-DNA interactions focused on certain diseases.

Electrochemical Assay

As an additional readout method, we developed a novel Electrochemical assay that can detect strand displacement reactions accurately all the while overcoming some of the limitations of the fluorescence measurements. Its cost-effective design and the bypassing of the explicit instruments needed for fluorescent readout, constitute it as a universal and versatile output detection module. Moreover, the intrinsic limitations of the fluorescent tags, like fluorophore bleaching, can be overcome as well. Having the advantage of being accessible, as it can be synthesized from commercial off-the-shelf products, it presents the perfect alternative to the standard measurement methods for strand displacement reactions.

Specifically, using the gold sensing pad we designed two separate experiments. Materials for both of them are commercially available, and the systems can be fabricated at a low cost.

The first is designed to measure capacitance variance, observed when specific DNA strands bind on the probe strand ( that is sulfur gold bonded on a gold plated electrode). The results obtained show that the system is able to measure concentrations as low as ~10nM. Also, the system’s response is very high, as the existence of the specific strand is almost immediately detected. In the following graphs, the measurements obtained for the interaction of various specific and non-specific strands is shown

Figure 9
  • A -> + Water
  • B -> + 15nM binding DNA strand
  • C -> + 15nM binding DNA strand
  • D -> + 60nM binding DNA strand
  • E -> + 30nM binding DNA strand
  • F -> + 30nM binding DNA strand
Figure 10
  • A -> + Water
  • B -> + 15nM non binding DNA strand
  • C -> + 15nM non binding DNA strand
  • D -> + 60nM non binding DNA strand
  • E -> + 50nM binding DNA strand

The second experiment is designed to measure conductance variance, observed when specific DNA strands with redox tag bonded on the probe strand ( that is sulfur gold bonded on a gold plated electrode) is displaced through a DSD reaction. The results obtained show that the system is not as sensitive as the previous method, as it requires higher concentrations. The system’s response is also very high, as the existence of the specific strand is almost immediately detected. In the following graphs, the measurements obtained for various specific and non-specific strands are shown.

Figure 11
  • A -> Only probes (single strand)
  • B -> Probes with binded strand (no redox)
  • C -> Probes with binded redox strand
  • D -> With displace addition 60nM

In accordance with our results, we have shown that the reporting complex used in our circuit, can be dissociated from its fluorescent output to an electrochemical one, enabling circuit output quantification as voltage or current intensity respectively. To our knowledge this is the first exhibition of DNA strand displacement readout through an electrochemical assay. This paves the road for future implementations of strand displacement circuitry translated back into electric signaling making use of the robustness and repeatability of such measures.