During this project, many results were achieved, including making chemically competent cells, transforming a plasmid in the so-said cells, making de novo sequences through primer amplification, with sequences that are able to proceed a transcription step and even work with enzymes to make fluorescent detections.More details can be found below, including the steps, the results and the discussion of everything that was completed.
The full protocol can be seen in Experiments in the part Competent Rosetta BL21-DE3. According to the reference article for the plasmid, it was retrieved from the DH5 alpha and transformed into the Rosetta BL21-DE3 cells. To do so, it was mandatory to prepare some chemically competent cells. Here are the final results of this step:
As neither the cell nor the transformed plasmid (PUC 19) contains the resistance to the antibiotic Kanamycin, results can be seen in the figure above that no colony can be found on the plate.
On the contrary, the strain naturally got embedded in its system the resistance against the Chloramphenicol, leading to the positive control being overwhelmed by bacterias.
To check about the transformation, it is important to look on the plates for the Ampicillin resistance, as it’s the antibiotic that get rid of everything besides what contains our transformed test plasmid (PUC 19), and a few colonies are transformed, and can be seen on the photos, showing that the bacterias are competent, considering that they integrated the plasmid correctly.
Once it was deemed possible to actually transform the newly competent, it is time to go onto the next stage and perform the transformation using CasRx’s plasmid. Again, the protocol can be found here (inserer lien de experiment).
Those pictures were made on an agarose plate containing both chloramphenicol (for the natural resistance of the Rosetta strains) and kanamycin (for the natural resistance of the CasRx plasmid). Colonies can be seen as the white dots determine the fact that bacteria grew on the media.
We confirmed that the plasmid was properly transformed in the strains using antibiotic resistance and moved on to the next steps of the purification process, namely, investigating the proper expression of the protein.
The transformed cells were safely stored by making a glycerol stock and keeping it at -80C for later uses.
While part of the team was working on the expression and purification of the Cas, others were in charge of the production and purification of the guides and targets RNA. As they were ordered in the form of DNA, the first step was to amplify the sequences.
The expected bands for each sample is:
The amplification was made using a PCR Protocol and the Thermo Fisher superfi II kit. The amplification was verified by looking at the DNA band sizes on agarose gel
After proceeding with the DNA clean up, the next step was the transcription from DNA samples to RNA. It can be quite challenging to work with those as they are degraded easily thanks to all the RNAse present everywhere, which were taken care of using some RNAse away, found in Thermo Fisher’s website. After cleaning everything, from the benches to the pipettes, nuclease-free water was needed in order not to contaminate our samples, special tips were also used.
Once everything is done, the T7 transcription was ready to go using this protocol and it was purified using the clean up protocol found at the same place. We used the Nanodrop to verify the presence of DNA and determine its concentration.
Tab 1: Tables showing RNA concentration for both the first (old) RNA and the second (new) batch.
As the Nanodrop isn’t precise enough, every concentration below 50 ng/µL was considered negative.
Also, the conversion between the result provided (in ng/µL) is also present normalized (in µM).
Daniel J. Brogan et al.'s research [1] served as the foundation for CasRX production. Within the Cas13 enzyme family, RfxCas13d, a member of class-d, distinguishes itself with several notable features:
Those properties make it easier to produce in bacteria.
We ordered the high copy plasmid OA-1136J [2] from addgene as agar stab. It was growth in DH5alpha at 37°C.
First, we used the RfxCas13d Expression protocol mentionned in experiments, but while doing a SDS-page gel, we did not obtained significant results as a band for our protein was not visible. As a matter of fact, it seems that the CasRX protein is insoluble, so we did back the experience this time making an IPTG gradient for the induction as well as taking either the supernatant alone after centrifugation, or the pellet + supernatant.The purification step was not possible a second time, if you are interested by it, you will find the protocol in the experiments section. Below, is detailed the protocol we used for this expression.
Rosetta cells ON were transformed at 37°, using chloramphenicol and kanamycin to filter out other bacteria than our transformed Rosetta.A colony was taken from the petri dish to inoculate 10 ml of LB, and was grow at 37° with the same antibiotics.The preculture was then split in 4 tubes that we named : 1A,1B,1C,1D, by putting 10 ml in each.
Once the OD was around 0.5-0.6, we induced tubes :
Afterwards, the culture were grow at 18°C overnight to induce production.Once the induction was one, we run a SDS Page using Bis-Tris 10 % precast gel, and MOPS buffer.
This is the SDS-page gel, we obtain for the RfxCas13d protein
It look like our protein (MW = 155 kDA) was much less present than expected. We did some ImageJ analysis for the intensity of the bands, and the higher one seems to be at 0.25 mM induction which is the concentration recommanded in the CasRX paper. On the other hand the 1mM IPTG induction seems to be lethal for the bacteria as no band are observed. You will find below a table that resume the result of the intensity value for each band.
As the expression did not work the first time, we ask ourselves if the plasmid was mutated, so we did a nanopore sequencing.
As you noticed there is only 1 red spot on the sequencing which would indicate a mutation in the ORI region, but looking back at addgene sequence of CasRX plasmid there was a nucleotide in this region that was not mentionned. Nevertheless a mutation in the ORI region would not be able to modify CasRX expression pattern according to one of our advisors : Maxime Bello.
sub parts of each.
This first graph shows plate reader fluorescence detection for every sample and our positive control (sequence 1.1, green). Interestingly, a difference above 25k units can be seen between sequences 1.4 and 1.1, which are highly similar. This observation indicates that small nucleotide differences in the guides can lead to big detection differences.
While the numbers of each type of nucleotides (A,T,C,G) remains relatively constant for each sequence, a higher fluorescence can be seen for both sequences 2.1 and 2.2, indicating that it might be a better sequence for fluorescence detection.
According to Konnerman et al. (2019), Cas13d should have a preference for cutting uracil bases.The results found here were that a higher A and/or G concentration provides more fluorescence (both sequence 2.1 and 2.2 contains more than 10 A and 10 G in their sequences), which means at the same time a higher cutting proficiency. It would be interesting to investigate these results in the future, to check how the nucleotides A, U and G lead to the highest cutting efficiency.
After looking at the overall sequence efficiency and responses, we started looking into the kinetics of the reactions and the overall impact of concentration on the system.
batch in Graph 2 or second batch in Graph 3) and the method used for the concentration calcul (ng / µL or µM / µL respectively in Graph 2
and Graph 3)
In those graphics, different concentrations and batches of RNA preparation can be seen. The objective was to determine if there are differences between the batches at first, but in a second time, the main goal became to check which is the best method to proceed with our analysis (concentration in ng/µL or normalization in µM).
Unexpected behaviors were observed in both graphs 2 and 3, where the fluorescence signal decreased. (once the probe is cut, it will emit fluorescence and would not be able to bind back to reduce the units emitted). Nonetheless, the sequences 2.1 and 2.2 can be seen as really homogeneous no matter the conditions and the batch, with a fluorescence around 27k RFU (Relative Fluorescence Unit).
Despite observing overall more noise and fewer positive signals using the new RNA batch compared to the old one, we can see that once everything is normalized, either the sequence is working or not, and the results are constant no matter the sequence used.
The normalization process from ng/µL to µM was made by using the concentration obtained after the Nanodrop in ng/µL, and then converting it into g/L (*10^-9 to go from ng to g, then /10^-6 to go from µL to L). It was then divided with the molar mass of the sequence, leading to mol / L as a unit, or molar M. Last step is to scale this result down to µM by multiplying it with 10^6. By doing so, we reduced by half the quantity of target used (1 µM of target), but the quantity of guides got upped, in order to get the 2:1 ratio between guides and targets present in the CasRx reference article.
Considering that the normalization method is using up to 4 times less materials, and provides better reproductivity and better fluorescence, it will be adopted for all the future experiments, and is also advised to do so for whomever will think about doing a similar project.
In the end, the sequence 1.2 from the new RNA will be used as a reference sequence in further experiments according to the results last shown (RFU around 37k, compared to the 15k in the old RNA), but mostly considering that, paired with those fluorescent data, it also uses less materials compared to the others.
As the difference in fluorescence between the reference and the changed concentrations is noticeable, we concluded that the quantity of target and guides in the mix could be optimized further and provide clearer results. In order to understand our system better, we decided to investigate the impact of concentration variations on total fluorescence. Fluorescence was not observed for all of the tested sequences. It is possible that some of the RNA concentrations were in reality lower than what we measured and not sufficient for Cas detection. Nanodrop is indeed known for not being the most precise way to measure RNA. It could be interesting to measure concentrations again by using the Qbit method to validate this hypothesis.
Also, normalizing the data seems to give overall better results compared to the concentration analysis, considering that the sequences either work really well, or do not give any detection at all. Henceforth, the molarity acts as a huge factor in the fluorescence detection.
For this experiment, the goal was to vary either the targets or the guides concentration (twice or halved) to gather more intel on the optimization possibilities of those parameters.
While we expect less fluorescence when scaling down the concentration, we also expect better fluorescence values when increasing the quantity.
On the contrary of the expectations, the results came out as higher fluorescence for all the conditions compared to the reference, no matter if the parameters were doubled or halved.
Those results can be explained because the concentrations used (1 µM of target, 2 µM of guide) might provide different results depending on the sequence used.
In the future, it would be interesing to push the optimisation further by systematically testing different concentration ratios of guides and targets. this would help understanding if the concentration impact is sequence specific or general for Cas13 functions.
As we decided to get more insight on the quantitative aspect of this analysis, we developed a short python script that is able to fit the curves and sort out parameters such as the angle of the slope (alpha), the value of the plateau, and the time it took to reach the plateau.
Thanks to that, it can be seen that amongst the different conditions tested, higher concentrations generally means higher fluorescence (above 16k units of fluorescence in the plateau for both conditions, target and guide) and the slope also as an higher alpha angle (above 15 and above 14.5 relative fluorescent unit per second, RFU / s). When lowering the concentration of guides and targets we can see that, despite getting lower fluorescence values than when the concentrations are doubled, they are still above the reference concentrations, with an alpha unit above by more than 2 points (with the lowest value, corresponding to the half target condition) , but the plateau also gets lower, achieving a whole 3k units of fluorescence difference between the low target and the reference.
Those results mean that more optimisation can be made concerning the optimal concentration of targets and guides, and also that a lower amount of materials is needed to proceed with the detection.
The purpose for this experiment was to test the fluorescence on the reference sequence (1.2, orange curve) and check for the first time whether or not it is possible to detect viruses in in more realistic conditions, with the digested mosquito instead of the water and also the presence of deactivated viral RNA (green curve). After that, tests were made simultaneously with purified viral RNA from the mosquitoes with either the reference guide only, or a mix containing all the guides. The reason behind it was to check if multiplexing was possible, to have a fluorescence signal when all the guides are present in the mix.
The expected results are that both the reference only and the one containing the mosquito juice and viral RNA are working just fine. For the guide’s reference with the viral RNA from mosquitoes, a few fluorescence units are expected, and even less for the condition including the mix of all the guides.
While what was expected got obtained when using the reference guides and targets in both conditions, nothing was detected when no synthetic targets were added to the mix.
As the detection is proved to be possible, the lack of fluorescence can be the results of a couple of parameters. The first one being the fact that the viral charge of our target is too low to be detected, while the second one being that the virus got a mutation on this specific part of his sequence. The third possibility is that our target sequence could have been degraded over time. Last but not least, our sample might have no traces of the target as the experiments did not took place the day the mosquitoes were received and, even though anti-RNAse was added, the viral mRNA in the mosquitoes could have been degraded by the time we used them.
In order to understand if the viral charge was too low, in the future we want to perform RT PCR on the mosquito juice and with that verify if the viruses were present, and at the same time, validate that our guide can bind to the viruses properly. Using this PCR, a sequencing can also be done to verify the sequences of the viruses and if they match our guide properly. Also, checking the state of the RNA throughout our experiments in a bioanalyzer could also give insight on the quality of the RNA. Lastly, if the detection works, it can be interesting to test if the same results can be obtained using isothermal PCR (RPA), as this method would amplify the viral material without requiring a PCR machine and could be done out of a lab, to investigate if whether or not, it could be implemented in an automatic device.