Our Success
Through a series of design-engineer cycles, BoviTect has managed to successfully produce a positive fluorescent result from multiple bTB biomarkers. We have used synthetic biology to successfully create a diagnostic tool for bTB.The figure below demonstrates this; there is a clear and rapid fluorescent signal that is significantly higher than the fluorescence from the negative control in response to four of our selected bTB biomarker targets.
[Figure 1] This shows the targets (PLAUR, CXCL8, RGS16, NR4A1 and FOSB) in the Cas13a system compared to the negative control where Cas13a has been removed, with a threshold at 10 nM fluorescein concentration equivalents (nM) (dotted black line), units given in fluorescein concentration equivalents (nM) and R squared values for the Michaelis-Menten model the data is around. PLAUR and NR4A1 only shows the top error bars and CXCL8 shows only the bottom to make the data more clear to read. Only 2 repeats were able to be done and this is due to time and budget limitations.
Figure 1 shows PLAUR and CXCL8 have surpassed the threshold value (>10-fold above negative control) and therefore are confirmed to be a positive result meaning our Cas13a system has multiple targets that can detect Mycobacterium tuberculosis biomarkers and report a significant positive fluorescent result. The threshold was picked as 10x the value of our highest negative control as our stakeholders have stated that reducing the chance of false-negatives is one of the largest priorities. As well as this target NR4A1 and RGS16 also had a significant increase from the negative control, however did not surpass the threshold.
Design, Build, Test, Learn Cycles
Our final working diagnotic tool is made up of many components which in turn were modified and tested in the process much like the Design, Build, Test, Learn process lays out. There were three significant design cycles that altered our engineering success the most, which are outlined below. We found that these cycles ended up being more like spirals, as the diagrams show. These have two or three design cycles feeding onto the next.
Sample Preparation Design Cycle
Design 1
Knowing that the eventual field-diagnostic would be working with biological fluids such as blood, we needed to assess whether the fluorescent signal would be blockedby the sample needed for the diagnostics test. We therefore designed a series of experiments to assess various methods of sample preparation in order to reduce sample opacity and decrease background noise.
Build 1
In order to abide by the 3Rs[1] as a key part of our scientific values in Human Practices, we began by testing on synthetic blood as the first R is Replacement. In this sense we were trying to replace the need for using real animal blood by using synthetic blood.
We trialed different methods to reduce the opacity by the blood including filtering and the use of a "hand centrifuge"[4].
We trialed a hand centrifuge instead of an electric one because we wanted to try and make our test more accessible in the field where we believed electricity may not be available.
Test 1
To mimic the signal we expected from our diagnostic, we spiked synthetic blood samples with fluorescein and measured fluorescencewith a fluorometer (Tecan Infinite m200 Pro) at excitation 488 nm and emission 519 nm which were then converted into fluorescein concentration equivalents.
[Figure] Our blood samples when conducting opacity tests.
Learn 1
From this we learnt that synthetic blood was not a suitable substitute to replace biological material as it lacked many of the key similarities to real blood, such as coagulants and serum.
However, we did learn that the background noise was very high and this interfered with the fluorescent reading, therefore we would need to find a solution to reduce opacity to allow our test the highest sensitivity possible as this was an important factor from our stakeholders and key to integrating human practices into our research process.
We also learnt from our Human Practices that according to veterinary practitioner Piers Pepperell, an electric centrifuge, or other methods that involve electricity, were feasible as there is often electricity available in the locations that bTB testing occurs.
Design 2
We redesigned our sample tests but this time using real blood. We ensured to keep this in line with our scientific values and the 3Rs[1] by following the principle of reduction. To do this we bought the minimum amount of pigs blood which was a waste product from an abattoir to trial. Buying from an abattoir also meant that this was a non-hazardous material as it was already deemed food-grade. Cattle blood was not used as it is not availble for sale in the UK due to biohazard restrictions.
Build 2
We didn’t try heating and coagulation with synthetic blood because we knew there weren't any coagulants in the synthetic blood, but now we had real blood we could trial methods such as coagulating the blood through waiting and heating as well as repeating filtering and centrifuging on real blood.
Also, as informed by Human Practices we now were trying to an electric powered centrifuge.
However we had to make some adaptations, for example using a 5 μm filter size for the first filter and then decreasing the filter size for each subsequent filter to 1.2 μm, 0.8 μm and 0.2 μm filter due to the red blood cells size blocking the smaller filters.
Test 2
We conducted these different tests on the pigs blood and did a fluorescence intensity scan at excitation 488 nm and emission 519 nm as decided on by previous experiments. These results were converted into percentage change of the fluorescein concentration equivalent when compared to a benchmark of untreated blood with 10 nM fluorescein. This data is availble on the measurement page.
Our method for testing the heating of blood was drawn from Arizti-Sanz, 2020 [2]
Learn 2
Analysis of our results showed that heating was the most effective at reducing the absorbency of the blood so that we can get a clear fluorescent result to the bTB targets. Our eventual diagnostic device would be required to include a heating elements for blood sample processing.
Click for References
[1]The 3Rs | NC3Rs [Internet]. Nc3rs.org.uk. The 3Rs | NC3Rs; 2017 [cited 2024 Sep 4]. Available from: https://nc3rs.org.uk/who-we-are/3rs
[2]Arizti-Sanz J, Freije CA, Stanton AC, Petros BA, Boehm CK, Siddiqui S, et al. Streamlined inactivation, amplification, and Cas13-based detection of SARS-CoV-2. Nature Communications [Internet]. 2020 Nov 20 [cited 2024 Oct 1];11(1). Available from: https://www.nature.com/articles/s41467-020-19097-x#Sec8
RNA Transcription Design Cycle
Design 1
Potential targets from either the Mycobacterium tuberculosis genome (for direct bTB detection) or the Bos taurus genome (sequenced from a Hereford cow, for identification of immune resonse biomarkers) were identified from literature[5,6]. These genes were processed by our custom python scripts (for either Cas12a or Cas13a respectively). Using string manipulation the scripts found all possible spacer sequences within genes/mRNA, that satisfy the Cas12a PAM site and Cas13a PFS site.
These spacer sequences were appended to the crRNA for the respective Cas protein, to make the full sgRNA sequence. The sgRNA sequences which had the correct hairpin folding were selected, and DNA templates for the following were synthesised by IDT: Cas12a sgRNA, Cas13a sgRNA, Cas13a targets and Cas12a target DNA was also ordered.
Build 1
We planned to amplify the 250 ng of each DNA template for the target sequences using PCR, so there would be enough template DNA to run a T7 in vitro transcription reaction of Cas13a targets. We also planned to order enough linear sgRNA DNA template oligomers to transcribe the sgRNA.
Therefore PCR amplification and T7 transcription was carried out on relevent sequences, as can be seen on the sgRNA tab of our Results page.
[Figure 2] Abstract representation of DNA sequences synthesised by IDT.
A) Template DNA to be used in transcription of Cas13a and Cas12a sgRNA. The sequence contains a T7 promoter, however unlike the target sequences does not contain a forward M13 primer. Since transcription needs to end abruptly after the spacer sequence, to ensure specific folding of the RNA, requried to form a complex with the Cas13a and Cas12a proteins.
B) Template DNA for Cas13a target, including a T7 promoter and M13 primer binding sites for amplification, via the polymerase chain reaction.
C) DNA sequence to be used as the Cas12a target direclty, including M13 primers binding sites for amplification, via the polymerase chain reaction.
Test 1
A High Sensitivity DNA Qubit quantification was carried out on the PCR product, which showed all sampled had a yield below <0.01 µg mL.
Nevertheless a T7 in vitro transcription was attempted, however no RNA was detected on a HSRNA Screentape (Agilent Tapestation 4200) - likely because there was not enough DNA template.
Learn 1
It became apparent that our transcription protocols were optimised for larger amounts of template DNA than we had, and due to budgetary constraints we could not order in enough linear DNA to carry out the transcription reactions.
At this point there was also concerns about removal of DNA from our transcription product, as template DNA would trigger the Cas12a system, leading to false positives in our final testing.
Design 2
The template DNA for; Cas12a sgRNA, Cas13a sgRNA, Cas13a targets and DNA for Cas12a targets, were re-designed and re-synthesised by IDT to include Type IIS restriction sites, so they could be cloned into plasmids.
In order to ensure transcription ended abruptly a Type IIS endonuclease site was added just after the sgRNA sequence ended, this in theory would cause T7 RNA polymerase to run off our linearised DNA templates. Klenow was used to blunt the ends of the linearised plasmids and restore any cleaved spacer nucleotides, and to prevent the plasmids from reforming.
Build 2
Our supervisor then cloned our DNA fragments into a high copy number plasmid, pX1800, then transformed the plasmids into DH5α E.coli, which in turn were grown and a plasmid extraction was carried out.
The plasmids were cleaved and blunted in a combined reaction, then transcribed using a T7 in vitro transcription reaction. We also compared different ways of cleaning up the DNA templates; using a DNase digestion or a DNase digestion followed by a AMPure bead extraction (to remove the DNase). This data can be viewed on the sgRNA section of our Results page.
[Figure 3] First successful transcription results, of Cas13a targets and Cas12a sgRNA, measured on a HSRNA tape (with Agilent Tapestation 4200)
A) Shows tape columns: B1- #8 CXCL8 target, C1 - #9 RGS16 target, D1 – #11 EthA_b sgRNA, E1 – #15 RD4_c. Blue box surrounds bands where desired peaks are observed.
B) Shows the normalised sample intensity graph of #9 RGS16 target. Due to a slightly degraded lower marker, a peak at 45 bp is observable. A 219 bp peak, close to our desired 153 bp (and due to the unreliability of the degraded sample buffer, this is likely our desired product).
C) Shows the normalised sample intensity graph of #15 RD4_c. Due to the degraded lower marker, and the small size of the transcribed fragment, the desired product peak at 38 bp (44 bp expected) is adjoined to the lower marker. The 124 bp peak is likely from the transcription of a plasmid templates that has not been cleaved properly (127 bp expected).
Test 2
Transcribed samples were ran on a RNA Screentape using an Agilent Tapestation 4200, showing RNA of our desired size was present. Both DNase digestion and the DNase digestion followed by a AMPure bead extraction, had similar results.
Our final Cas13a testing resulted in a functioning detection system, which reinforces the information of the Tapestation as the sgRNA and targets must be functioning as intended.
[Figure 4] Shows the Agilent Tapestation 4200 RNA ScreenTape results from final T7 in vitro transcription. Where; Cas13a sgRNA sequences (1,2,4,5), Cas13a target sequences (6,7,8,9,10) and Cas12a sgRNA sequences (11,12,13,14,15) were tested. Against a contorl, where the T7 polymerase was not included in the reaction mixture. This figure includes an image of the RNA ScreenTape and a representative normalised sample intensity sample graph.
A)Shows the tape gel – showing uniformity among varients of each RNA component.
Learn 2
Dnase digestion followed by a denaturation cycle was used, due to budgetary restraints concerning AMPure bead use.
We found we could produce high quality sgRNA and target RNA for our Cas13a test, using our transcription protocol.
Cas13a Expression and Purification Design Cycle
Design 1
The coding sequence for LwCas13a was taken from the plasmid pC029-Lw2Cas13a from Leptotrichia wadei F0279 (Addgene plasmid #91919)[1] and codon optimised for expression in E. coli using the IDT tool. Any forbidden restriction enzyme sites were removed. A 6xHis tag and TEV protease cleavage site was added at the N-terminal, Type IIS cloning prefix and suffixes were added and the complete sequence was synthesised as a g-Block by IDT. This plasmid was made with the intention to transform it into BL21(DE3) and Rosetta.
Build 1
This sequence was cloned into a medium copy plasmid (origin of replication from pBR322[2]) carrying an ampicillin selection marker with a composite part (IPTG inducible T7 promoter system) comprising the T7 promoter, lac operator and the RBS from bacteriophage T7 gene 10 and the transcription terminator from bacteriophage T7 RNA polymerase.
Plasmids were transformed and the protein expressed. As a positive control for protein expression we also transformed a plasmid coding for His-Tagged mCherry. As a negative control for protein expression we transformed an empty plasmid pX1900.
E.coli strain: BL21(DE3) and Rosetta.
Induction method: Addition of 100mM IPTG (37°C overnight) to 250mL cultures.
Purification: Gravity Nickel Affinity column only, washed with 20mM Imidazole buffer, eluted with 250mM Imidazole buffer.
Test 1
Western Blot Result:
[Figure 5]: A western blot of purified proteins from gravity nickel affinity column. First 4 proteins produced in BL21(DE3) and others produced in Rosetta.
Learn 1
Cas13a did express in Rosetta but with a low yield. Rosetta transformation efficiency was low (Cas12a did transform and express in BL21(DE3)), and competent cell preparation proved difficult.
Design 2
Decided to continue with BL21(DE3) rather than Rosetta due to the better transformation results. Changed to autoinduction media [3]to see if that gave better expression results.
Build 2
Transformed and expressed again.
E.coli strain: BL21(DE3) only
Induction method: Autoinduction media (20°C for 24 hours) to 1L cultures.
Purification: Gravity Nickel Affinity column only but poured fresh columns but still washed with 20mM Imidazole buffer, eluted with 250mM Imidazole buffer.
Test 2
Western to show BL21(DE3) in autoinduction:
Figure 6: Western blots of proteins from gravity nickel affinity column. All proteins produced in BL21(DE3) and lysate is before purification and elutions are after purification. Concentration of Cas13a elution (measured via Qubit 2.0 Fluorometer): 612 µg/ml
Figure 7: SDS-PAGE of samples collected at every step during purification. Multiple bands indicate impurities at every step, showing that protein purification was not successful.
Concentration of Cas13a elution sample (measured via Qubit 2.0 Fluorometer): 612 µg/ml
Concentration of Cas13a: 612 x 0.0626 = 38.3112 = 38.3 µg/ml (3 s.f.)
Concentration of Cas 13a was calculated by importing each image using a custom Matlab script, convert to a numerical matrix of RGB intensities, and counting blue pixels given a certain threshold (Red <=100, Green <=100, Blue >= 100). Final values were plotted, the full process is available on the Protein Purification section of our Results page.
Learn 2
Yield increased of Cas13a but columns got clogged up in larger flask sizes - back to 250mL just multiple.
End results still have poor purity, so now we have better yield, need to improve the purification process. For this we talked to experts in the area from our university.
Design 3
To increase purity we decided to use a pressurised 1 mL His-Trap column before further purifcation off all pooled protein containg fractions using size exclusion chromatography (AKTA Pure).
Build 3
Transformed and expressed again.
E.coli strain: BL21(DE3) only
Induction method: Autoinduction media (20°C for 24 hours) to 250mL cultures.
Purification: Supernatant was loaded onto a His-Trap FF crude 1mL column (Cytiva) Eluted across a gradient of 20mM to 250mM Imidazole in 15 column volumes. Protein containing fractions were identified and concentrated to 2mL.
This was loaded onto a Superdex 200 10/300 120mL (Cytiva) (shortened to SEC) Eluted with 1 column volume of 20mM Tris-HCl, 500mM NaCl, 5% glycerol, pH 7.5.
Protein containing fractions were identified and pooled to give our final usable samples of Cas13a.
Test 3
The figures below show the progression of protein purification after sequential purification steps, with an eventual purity of 37.45% being achieved. Although this was a lower purity than we had hoped for, subseqeunt experiments determined it was sufficient for our diagnostic system to be sucessful.
Figure 8: SDS-PAGE gels of the fractions collected from the His-Trap columns, with the most concentrated protein containing fractions collected and pooled.
Figure 9: Western Blots of the fractions collected from the Superdex Size Exclusion Gel column, with the most concentrated protein containing fractions collected and pooled.
Figure 10: SDS-PAGE gel and Western Blot of the different stages of the purification process.
Concentration of Cas13a elution sample (measured via Qubit 2.0 Fluorometer): 183 µg/ml
Concentration of Cas13a: 183 x 0.3745 = 68.5335 = 68.5 µg/ml (3 s.f.)
Concentration of Cas 13a was calculated by importing each image using a custom Matlab script, convert to a numerical matrix of RGB intensities, and count blue pixels given a certain threshold (Red <=100, Green <=100, Blue >= 100). Final values were plotted, refer to figue 8.
Figure 11: A - SDS-PAGE showing lanes after the His-Trap and SEC. B - graph showing percentage of Cas13a protein after the His-Trap and after the SEC. There is a significant increase in the percentage of Cas13a in the overall protein content.
Learn 3
We used this Cas13a sample in our test of the overall system, and got a positive result for some of the targets, and a negative result for the negative controls, meaning the test worked! Purity could still be improved for a real life test. Cas13a band on the SDS-PAGE look clearer and show lower levels of impurities than the first iteration. Lower concentration of protein samples indicates more unwanted proteins has been filtered out.
