Synthetic biology is defined as “a growing discipline that has two subfields. One uses unnatural molecules to reproduce emergent behaviors from natural biology, with the goal of creating artificial life. The other seeks interchangeable parts from natural biology to assemble into systems that act unnaturally.” (1) To successfully develop engineering systems, synthetic biologists follow the Design Build Test Learn Cycle (see figure 1).
The problem we are trying to solve is the lack of rapide and simple detection of oyster pathogens in water samples. To do so, we adapted an existing detection technique called SHERLOCK to our purpose. This is how Shell’lock born.
In our project Shell’lock, we were able to design guide RNAs specific to target sequences of Vibrio Aestuarianus and RNA probes, successfully detect this pathogen through the recognition of the target sequence by the designed guide RNA and finally develop a fluorescent output and paper-based test.
To illustrate our way to reach these successful results, we decided to illustrate all the steps through the DBTL cycle.
Cycles
-
Cycle 1: design of all the guide RNA specific for the pathogen and test the positive control using paper strip
Design
The first step of this cycle consists in choosing V.Aestuarianus genes that will be targeted. Bibliography was needed to choose relevant genes, involved in pathogenicity or virulence of the pathogen and that are specific for V. aestuarianus. Because we wanted first to perform the test on synthetic sequence before testing it with true samples, the first step was to design synthetic target sequence as shown in figure 1. The second step consists in designing guide RNAs that are able to recognize a 28 nucleotides sequence present in each chosen gene. The design is done using bio informatic tools and literature. For this step, we developed a blast program that is able to find matches between designed guide RNAs and other species present in water.
In our idea, we wanted first to test the SHERLOCK reaction using a paper strip. To be able to perform these experiments, we ordered already purified Cas13a, a plasmid containing this enzyme and a FAM-biotin probe.
Finally, we started to design a model for the Cas13a kinetic and a data analysis codepipeline for the fluorescence experiment we planned to perform.
Figure 2: template of the design part of the first engineering cycle. Build
The build step of this first cycle consists in recombinant expression and purification of LwaCas13a.
Moreover, we did PCR amplify target and guide sequences, T7 transcription to transcribe them into RNA.
Both the purification and the PCR - transcription of target and guide sequences worked. We had all the material to perform our first paper-based test using Milenia Hybrideted strip and reagent.
Figure 3: build process of the first cycle Test
The first test we did is the positive control using the synthetic sequences described by Kelner et al. (2). This first test worked but we wanted to optimize the concentration. Meaning, that we tested several concentrations of Cas13a, guide RNA, target sequences and probes. The idea was to optimize the result without throwing a lot of reagents.
Learn
We were able to draw different conclusions from this first cycle. First, the paper-based test provides visual information about the presence of the pathogen in water but it is subjective and relies only on the analysis of the person performing the test. To better evaluate the response given by this test, we decided to use a gel imager to expose our readouts to white light and obtain a RGB image of our test. Moreover, we found out that the Cas13a already purified protein does not work. Finally, at the end of this first cycle all the concentrations were optimized.
-
Cycle 2: “quantify” paper-based test results and test it under several conditions with the positive control and with specific target / guide sequences
Design
The first step of this second cycle was to develop a way to analyze the paper-strip images obtained after white light exposition. We did that using Image J and python (see figure 3). Moreover, with this second cycle, we wanted to realize the paper-based test under several conditions. We design experiments to monitor the consequence of temperature, water salinity and incubation time under the results. Finally, during this second cycle we also tried the paper test with specific target and guide sequences.
We did positive and negative control tests again, used the gel imager and the python code to analyze the results obtained. This is our first proof of concept: the positive control works with our Cas13a protein.
Figure 4: SHERLOCK detection on synthetic sequence. PC: positive control (synthetic sequence), NC: negative control without the Cas13a. Then, we tested the influence of incubation time, water salinity and temperature of the reaction (results).
Concerning the V.aestuarianus target and guides, none of them produced positive results.Learn
This cycle approved the way we read the paper-based test results and gave us information about the test robustness under different conditions: 15 minutes is not enough to obtain a positive test, the reaction does not work at 4°C, the salinity has an influence on the result. However, by using a paper-strip to realize the test, we do not have access to the kinetic of the reaction. This is why in the third cycle, we decided to develop a fluorescence assay. Moreover, we found out that the design of the guide RNA was incorrect, it will be modified in the fourth cycle.
-
Cycle 3: Fluorescent assay to monitor the kinetic of the reaction
Design
For this new cycle we designed a fluorescent probe and we decided the protocol of the experiments. We also designed the fluorescence data analysis codepipeline. To perform the fluorescence experiment, fluorescent probes were ordered. As the latter provides the kinetic of the reaction, we also developed a kinetic model. Because we read the fluorescence output using a plate reader, we decided to test several experimental conditions at the same time.
Test
For the first fluorescent experiment, we tried several probes (see figure 4a and 4b) and target (see figure 4c and 4d) concentrations. Moreover, we fitted the data with the model and used it to have access to the unspecific catalytic constant.
Figure 4: A. Kinetic of the reaction with probe dilutions. B. End-point of the reaction with probe dilutions. C. Kinetic of the reaction with target dilutions. D. End-point of the reaction with target dilutions. What we learn from this cycle is that the efficiency of the detection technique strongly depends on the probe and target concentration. 2µM of probe and as low as 10nMof target are enough to obtain positive results. Moreover, the model allowed us to have a better understanding of the Cas13a kinetic.
-
Cycle 4: paper-test and fluorescence assay with new guide sequences
Design
As we found out an error in the design of the v.aestuarianus guide RNAs, we designed and re ordered them again.
Build
We did PCR to amplify the new guide sequences, T7 transcription to transcribe them into RNA.
Test
We tested both the paper-based test and the fluorescence assay on all the target sequences. Moreover, we modified the model to obtain the best one and fitted all the data with it. A remarkable result is the one of the detection of the toxR gene (see figure 5).
Figure 5: Shell’lock is able to detect V.aestuarianus gene using both readouts methods A. Lateral flow detection of ToxR sequence B. Fluorescent readout of ToxR sequences with decreasing concentration. Learn
We have a proof of concept. We are able to detect vibrio aestuarianus genes using both fluorescence assay and paper-based test.
Conclusion
Designing a project from scratch is complex and sometimes slow. Careful design is key, especially when time is limited. The learning part is extremely important to correct our mistakes, better understand our system and optimize it. And repeating the cycle is time consuming but necessary to bring the project to maturity. Here we detailed the process that we followed to achieve success.
References
1. A.Benner S, Sismour AM. Synthetic biology. 2005 Jul; Available from:
10.1038/nrg1637
2. J.Kellner M, Koob J, S.Gootenberg J, O.Abudayyeh O, Zhang F. SHERLOCK: Nucleic acid detection with CRISPR nucleases. 2020 Mar; Available from:
10.1038/s41596-019-0210-2