Fluorescence Detection
The main purpose of our project was to detect pathogen genes using the SHERLOCK technique. We also wanted to monitor the kinetics of the reaction so we needed a quantitative readout that could be measured in real time. We thus decided to use fluorescence detection as a way to tackle this challenge. We had several goals to achieve: validate our guide RNAs to detect their targets, optimize the detection technique to have the best results possible, measure the reaction kinetics and assess the selectivity of Cas13. Also because fluorescence detection can be calibrated to get access to concentrations of cleaved RNAs , we can fit our experiments with the model we have developed and see how accurate it is.
1- Optimizing probe concentration:
Since we had a limiting amount of fluorescent probe, we wanted to test if we could minimize its concentration in the SHERLOCK assay. For this, we used the synthetic control target and guide RNAs. For that we tested three different conditions, the probe at 2µM, 1 µM, 0,1 µM.
As we can see the optimal probe concentration is at 2 µM like it was stated in the protocol we based our technique on.
2- Sensitivity of Shell’Lock:
The next step was to determine the optimal concentration of target we can detect and so see how sensitive our detection really is. We tested different target concentrations and from the results we obtained we could see that the best detection was at 1 µM. And we see that below 10 nM we can hardly detect anything (Fig.2). This helps us to know the actual limits of our test. Even though in the original Sherlock technique they were able to detect to the femto level we were not able to reach that level (1). In our case a number of things could be causing this failure of detection at low concentration, one of them being that maybe the Cas13a enzyme we are using is not as active as the one used in the article and thus it is not able to detect targets at the same level. However, if we need to improve our sensitivity, we can add an amplification step before Cas13 activation (1). This is something we did not have time to accomplish but that would be extremely useful. Indeed, we suspect that pathogens in water samples or contaminated oysters might be in low amounts and contain low levels of target RNAs.
3- Validation of our targets:
After optimizing our technique what was left was to detect the actual target genes we designed. We ran a plate with 100 wells containing all our samples. Each well contained a mix of reaction buffer and fluorescent probes. This was combined with our various guide and target RNAs (link to the lab notebook). The Cas13 enzyme was added to the side of each well. All our experiments were performed in triplicates. The plate was then placed in a spectrofluorometer. At time t=0 the plate was shaken to mix the reaction mix with the Cas13 enzyme, and fluorescence was recorded for 8 hours. The plate also contained a range of solutions of FITC at several known concentrations. This allowed us to convert the fluorescence intensity into concentrations of cleaved probes.
We can see in Fig.3 that the majority of the target genes are detected. The genes that we were able to detect belong to different families:
- - ToxR, DnaJ and flA belong to the genome of V.astereanus
- - SD and il17 are present in the genome of oysters
-
finally we were able to detect environmental probes like trigopus and algae present in the Thau lagoon and EF1a, a gene belonging to mussels. This shows that our technique is very versatile and can detect multiple genes, just by changing the guide RNA. This is very important as it opens the possibility to multiplex our detection. This is a perspective that we did not have time to explore, but would be extremely interesting.
4- Selectivity of Cas13
Seeing that our technique was capable of detecting target genes, we wanted to see how specific this detection is. We wanted to study the specificity of the Cas13 once it recognizes a target. We also wanted to see through our model whether once it recognizes a target will the reaction be spontaneous. And so we designed target genes containing mutations ranging from 1 to 10 mutations at specific locations. We can clearly see that after two mutations we don’t have any detection. Another thing we could have done was try and induce mutations at different locations and monitor how that change can also affect the detection.
Importantly, fitting these data with our model can help us understand to what extent the mismatch between the guide RNA and its target impacts the Cas13 reaction. In particular, we explored whether it affects binding to the target, cleavage of the non specific probe or both (model).
Conclusion
Our experiments show that we can detect the presence of target genes in our samples. Furthermore through optimization we were able to obtain better results. In addition, fluorescence allows quantitative monitoring of Cas13 collateral activity. We could thus fit these data with our model to better understand the SHERLOCK kinetic constants across a variety of targets.
References
1. M. J. Kellner, J. Koob, J. S. Gootenberg, O. O. Abudayyeh, et F. Zhang, « SHERLOCK: Nucleic acid detection with CRISPR nucleases », Nat Protoc, vol. 14, no 10, p. 2986‑3012, oct. 2019, doi: 10.1038/s41596-019-0210-2.
Latera Flow Result
To transfer the possibility to detect oyster pathogens directly to oyster farmers, we simplified the readout step and used a lateral-flow detection system.
Quantification technique
Our test is performed on a paper-strip using capillarity to migrate and capture antibodies linked to gold
nanoparticles for the detection. The detection is a qualitative process where the position of the band indicates
if the test is positive or negative. To quantify the test we use the software ImageJ(1) to extract absorbance
intensities and a python pipeline to perform the analysis. Further information on the functioning of the test
can be found here.
Paper-test is a rapid and easy to use detection method hence we used it as a benchmark to choose potential
targets and used it as a proof of concept of detection. Furthermore we optimized the experimental conditional in
hope to make it usable in the field and user friendly.
Proof Of Concept
In order to calibrate our test we first worked on “synthetic sequences” published by Kellner et al. (2). We followed the experimental procedure described in the article. To benchmark and optimize the process for all the experiments we used the synthetic sequence. Our goal was to be able to reproduce already published results to set the experimental conditions for the test.A result of such an experiment is shown below.
This result approved the experimental conditions as we see a positive read for the synthetic sequence. We continued our project by optimizing further the following conditions:
- Incubation time
- Temperature
- Salinity
- Target sequence
Incubation time
The first parameter we wanted to optimize is the incubation time. For a field-deployable test we deemed it important to fix this value. Our experiment consisted of different tests performed with different incubation times. The incubation time starts once the target has been added to the SHERLOCK reaction. Then the reaction is loaded onto the paper strip and imaged rapidly.
We see that we can detect starting from thirty minutes of incubation. We see partial detection after fifteen
minutes of incubation. As the level (relative) of detection is low compared to thirty minutes and one hour of
incubation we decided to consider it as negative. To further investigate this we would have needed to perform
more experiments with fifteen minutes of incubation.
For the following experiments we decided to set the incubation time at one hour. As these experiments were performed in-vitro we prefered to have a longer incubation time to maximize our detection signal. With the scope of developing a field-deployable test an incubation time of thirty minutes would be more suited for the user.
Temperature
The next parameter that we tested and that is crucial for field tests is the reaction temperature. As our test is composed of an enzyme (Cas13a) we first performed the reaction at physiological temperature thirty-seven degrees celsius. We wanted to see whether this value was crucial for the performance of the test.
We see that at 4°C the reaction doesn’t detect the target which is expected when working with enzymes. We see that the reaction works both at 25 and 37 °C. This information was crucial as 25 °C is a temperature that doesn’t require particular equipment to achieve. After this experiment we thus set the temperature of the test to 25°C.
Salinity
First attempt
Following the idea of developing a field-ready test we aimed at adapting our experimental condition to the field. We are working on the detection of a marine pathogen. This implies that our possible “test samples” will be in sea-water (more precisely lagoon water). The major parameter we need to control for the water in the Thau lagoon is the salinity. As the lagoon is connected to the sea the water can achieve concentration of salt equivalent to the sea, 40g/l. As our test is based on the activity of the Cas13a protein such concentration of salt could be an issue as it could introduce structural changes that might impact the enzymatic activity. Moreover, the Thau water has a particular pH of 6.4 that might also impact the activity of the Cas. Regarding the pH we are performing the test with a buffer that works in the pH of Thau’s water range (HEPES). To test the effect of salinity we decided to dilute Thau’s water in our test in increasing percentage (replacing nuclease free water). The results are shown below.
The percentage represents the fraction of Thau water in the total water used to perform the test. The sequence
tested is our positive control (sequence published by Kellner et al. nature protocol 2019)(2)
NC: negative control without Cas13a
The results show that the test did not work for the salinity percentages tested. Following this results we decided to test multiple hypotheses :
- The test did not work because of the pH of Thau’s water
- The test did not work because of the salt concentration
- The test did not work because of some other factor present in Thau’s water
Chemical salt solution
To verify the first hypothesis we verified using a pH-meter the pH of Thau’s water and verified it was in the range of our buffer (HEPES). Upon this verification we excluded this hypothesis as the pH of the water was close to the pH range of our buffer. To test the second hypothesis i.e the salt, we decided to prepare an “artificial” salt solution using pure NaCl and perform the same experiment. The results of this experiment if positive would also exclude the last hypothesis as we are working with “pure” salted water. The results of this experiment are described in figure 5.
We see that upon gradual addition of the salt solution there is a decrease in the detection. This is probably due to structural modification of the enzyme Cas13a by the salt. Salt plays an important role in protein folding and this equilibrium is changed in this experiment. Further experiments would be required to verify this claim. Our main conclusion however is that below 10 % of salt water the test is still detecting the target. This result confirmed our second hypothesis that the salt concentration affected the test.
Lower dilution
Moving on we decided to retry using Thau’s water but using lower concentration of salted water. We thus performed the experiment using respectively 0, 5, 10 and 15 percent of Thau’s water in the total amount of water in the test. The results are shown in figure 6.
We see that at 5 and 10% we can detect the target. This result showed that we can use our test using Thau water and thus in the field. Moving forward we performed every test using 10% of Thau’s water. The next thing we wanted to try was the pathogens sequences.
Proof of concept
We used the “new targets” and performed the SHERLOCK reaction. The results can be seen in figure 7.
We see that two of the pathogenic sequences namely ToxR and DNAJ, V.aestuarianus infection genes showed positive output. This was our proof of concept that we can detect pathogenic sequences using lateral flow SHERLOCK.
Conclusion
Our results show that through cycles of optimization we characterized the experimental conditions to perform the test. Moreover, we showed that we could use Thau’s water to perform the test. Furthermore, we showed a proof of concept that we can detect pathogenic sequences with our test.
References
1. C. A. Schneider, W. S. Rasband, et K. W. Eliceiri, « NIH Image to ImageJ: 25 years of image analysis », Nat Methods, vol. 9, no 7, Art. no 7, juill. 2012, doi: 10.1038/nmeth.2089.
2. M. J. Kellner, J. Koob, J. S. Gootenberg, O. O. Abudayyeh, et F. Zhang, « SHERLOCK: Nucleic acid detection with CRISPR nucleases », Nat Protoc, vol. 14, no 10, p. 2986‑3012, oct. 2019, doi: 10.1038/s41596-019-0210-2.
Guide RNA and Target RNA Amplification and Transcription.
The second step of our work plan was to produce and purify our guide and target RNAs. As a reminder, we had ordered the sequences of interest as DNA templates. So we first had to amplify these DNAs before transcribing them in vitro. The guides have a size of around 80 bps while the targets are around 300 bps. We followed this protocol ( PCR Protocol) using Thermo Fisher superfi kit to perform our PCR and ran our products on an agarose gel. All guide and target probes were properly amplified (Figures 1 and 2)
| Name Of The Guide | Name Of The PCR Tube | Full Name Of The Guide |
|---|---|---|
| p38 | 1 | EF1_alpha_new |
| p39 | 2 | 16s04 |
| p40 | 3 | DnaJ-g10 |
| p41 | 4 | 40s |
| p42 | 5 | 60s |
| p43 | 6 | Hsp70 |
| p44 | 7 | FlaA |
| p45 | 8 | SD |
| p46 | 9 | DnaJ-G07 |
| p47 | 10 | guide trigriopus |
| p48 | 11 | VAM |
| p49 | 12 | IL17 |
| p50 | 13 | DnaJ-G09 |
| p51 | 14 | 16s03 |
| p52 | 15 | ToxR |
| p53 | 16 | VspR |
| Name of the Target | Full Name Of The Target |
|---|---|
| A2 | Hsp70 |
| A3 | SD |
| A4 | Il17 |
| A5 | 16s_03 |
| A6 | 16s_04 |
| A7 | DnaJ07 |
| A8 | DnaJ09 |
| A9 | DnaJ10 |
| A10 | ToxR |
| B1 | ToxRx1 |
| B2 | ToxRx2 |
| B3 | ToxRx4 |
| B4 | ToxRx6 |
| B5 | ToxRx8 |
| B6 | ToxRx10 |
| B7 | Synthetic DNA |
| B8 | Synthetic DNA_full |
| B9 | 40s_oyster |
| B10 | EF1a_mussel |
| C1 | trigopus |
| C2 | 60s_oyster |
| C3 | flaA |
| C4 | vspR |
Next we cleaned our PCR samples essentially to remove the polymerase and the primers that were used for the amplification. For that we used this protocol (DNA clean up protocol).
Next we transcribed our DNA to RNA. Working with RNA is quite difficult because of the challenge of RNAse that is present ominously and degrades the sample quickly. We thus need to carefully clean the bench and pipets to remove all RNAse traces, and use nuclease-free water, tips and tubes. For the transcription we used this protocol ( Transcription protocol).
The next step in our work was to now purify the RNAs that we obtained to make sure we had no contaminations and remove transcription components. To do that we used this protocol ( RNA purification)
Next we ran a gel to make sure that our samples were actually present in solution.
Conclusion
After transcribing and purifying our RNA targets successfully, the next step is to begin the shell’lock detection, but first we need to program and test the specificity of our Cas13 enzyme.
Protocols
-
PCR Protocol
-
DNA Clean Up Protocol
-
Transcription protocol
-
RNA purification
-
Cas13a Purification Protocol