As we conducted our project, we implemented numerous iterations of the engineering design cycle, comprising of design, build, test and learn. This is implemented in all aspects of the project, including the plant miRNA extraction, hybridisation, RCA and dry lab.
A few key stages in our DBTL cycle is linked to below:
Verifying G-quadruplex
Farmer extraction
DNA Binding to Paper
Verifying Circularisation
Extraction
Gel Optimisation
Design - We hoped to optimise our gels for running RNA ladders, to yield the clearest images, such that we could learn as much as possible from future experiments.
Build - In order to test which gel would be best for testing our extraction products, we tested the same ladders (a high and low range molecular weight pair) on a TBE and TAE gel at 2% and 100V. We post stained with gel red for 30 minutes, then destained for a further 15 minutes, before visualising the gels.
Test - The TBE gel had cleaner bands and better separation (Figure 1), however the bands could still be sharper.
Learn - Hence we swapped to a lower 90V in order to produce better gel images, and used TBE for RNA gels in future.
Isopropanol
Isopropanol: Running extractions from ladder
Design - We wanted to test the viability of the isopropanol RNA extraction method
Build - We tested the isopropanol protocol
Test - The resuspended pellet, isopropanol and ethanol supernatants were all run on a 2% TBE gel at 90V for 90 minutes.
Learn - The isopropanol bands had the highest intensity while visualising the gel, then the pellet. The ethanol had no bands. This indicated that further efforts needed to be made in the precipitation step in order to ensure more RNA ends up in the pellet during the initial centrifugation step.
Isopropanol: Testing spin time
Design - We aimed to determine the rate at which RNA precipitates out of the isopropanol suspension in the first spin step.
Build - We once again used an RNA ladder as a standard. We then inspected the relative intensities from the resuspended pellets formed at every 5 minutes of centrifuge time for 30 minutes total.
Test - We conducted the experiment as described previously.
Learn - Diminishing returns were seen past 15 minutes, indicating that spinning for longer than 15 minutes did little to benefit RNA extraction.
Isopropanol: Extract from plant material
Design - Working with plant material for the first time, we were expressly aware of the risk to miRNA posed by RNAses.
Build - Hence we used a temperature controlled centrifuge at 4°C, worked with the samples and reagents on ice/ dry ice, and used RNAse-OUT inhibitor. We also spun for a full 30 minutes to maximise RNA extraction. Two homogenisation protocols were tested, one using a lysing matrix, and the other using a screw cap tube filled with larger glass beads and placed on a vortex. The latter method was cheaper, hence making it possibly more accessible to farmers.
Test - The RNA extraction protocols were conducted, and run on both a nanodrop and RNA high sensitivity tapestation to compare the two protocols. No RNA showed up on the ScreenTape, however this may have been an error in loading the sample, as the marker did not show up either. The nanodrop results ([RNA] = 6ng/μL, A260/280 = 1.69) were very low, so it would be unsurprising that the ScreenTape yielded poor results.
Learn - The lysing matrix worked significantly better, as such we dropped the vortex method, however still looked for alternative, cheaper methods, such as using a pestle and mortar. A lot more progress still needed to be made with pelleting the RNA.
Isopropanol: Extract from plant material
Design - RNases degrade miRNA, and they will inevitably be present in plant tissues.
Build - Hence, we used a temperature controlled centrifuge at 4°C, worked with the samples and reagents on ice/ dry ice, and used RNAse-OUT inhibitor. We also spun for a full 30 minutes to maximise RNA extraction. Two homogenisation protocols were tested, one using a lysing matrix, and the other using a screw cap tube filled with larger glass beads and placed on a vortex. The latter method was cheaper, hence making it possibly more accessible to farmers.
Test - The RNA extraction protocols were conducted, and run on both a nanodrop and RNA high-sensitivity TapeStation to compare the two protocols. No RNA showed up on the ScreenTape. However, this may have been an error in loading the sample, as the marker did not show up either. The nanodrop results ([RNA] = 6ng/μL, A260/280 = 1.69) were very low, so it would be unsurprising that the ScreenTape yielded poor results.
Learn - The lysing matrix worked significantly better and as such, we dropped the vortex method. However, we still looked for alternative, cheaper methods, such as using a pestle and mortar. A lot more progress still needed to be made with pelleting the RNA.
Isopropanol: Optimising lysate volume
Design - We have noticed that several different lysate volumes yielded different results on the nanodrop, so sought to test the effect of this.
Build - The extraction protocol was tested with several different lysate volumes in the total 200 μL, making up the difference with distilled water conducting extraction over a 30-minute spin step cooled to 4°C.
Test - Using a nanodrop we found that the yield improved with lysate volume - to be expected, but that the biggest difference came between 60 and 80μL.
| [RNA] ng/μL | 260/280 | |
|---|---|---|
| 20μL lysate | 2.8 | 1 |
| 40μL lysate | 2.9 | 1.4 |
| 60μL lysate | 17.4 | 1.43 |
| 80μL lysate | 29.1 | 1.68 |
| 100μL lysate | 22.6 | 2.16 |
Learn -Based on the results, we decided to use 80μL lysate for the subsequent isopropanol extractions.
Isopropanol: Optimising Isopropanol volume
Design - We sought to optimise the isopropanol volume in the extraction protocol
Build - We repeating the prior protocol, but this time with 60μL of lysate, 20μL of Na AcO and several (80/100/120μL) volumes of isopropanol
Test - We conducted the experiment as described.
| [RNA] ng/μL | 260/280 | |
|---|---|---|
| 80μL isopropanol | 45.5 | 1.62 |
| 100μL isopropanol | 26.8 | 1.86 |
| 120μL isopropanol | 13.5 | 1.58 |
Learn - It seemed the optimal isopropanol volume was 80μL, so this will be used in future experiments.
Whatman
Whatman: Running extractions from a ladder
Design - We aimed to conduct a preliminary assessment of the feasibility of the Whatman protocol and its efficacy.
Build - We extracted from a standard RNA ladder, followed by several elutions with the wash buffer.
Test - We ran a 2% TBE gel on the extraction and found that the RNA adsorption efficiency was not ideaLearn - most of the RNA stayed in the sample and it was easy to wash them off.
Learn - Whatman’s RNA binding capacity was not too high as many of the RNA was still in the remnant, and its elution step could not be easily controlled. We could consider incubating the paper for longer to get potentially better adsorption.
Whatman: Testing an elution step
Design - Similar protocols had an elution step, so we wanted to test the effect of this addition to the Whatman protocol.
Build - We added an elution step at 95°C to wash off more of the RNA and extended incubation time to 60 s.
Test - The incubation did not significantly improve the RNA binding capacity, but the high temperature elution did grant minor improvements to the elution efficiency.
Learn - Next things to try if we have time - Using lower RNA concentration; running a nanodrop/ScreenTape of the sample to determine actual extraction efficiency; and trying on Arabidopsis seedlings to see if it is capable of isolating the nucleic acids in a real world context.
Although its efficiency was not as promising, if it worked, it would be the simplest method to carry out. So, if there is time at the end, Whatman extraction might still be worth a try.
Salt extraction
Salt extraction: Running extractions from a ladder
Design - The protocol used PEG 8000, but it was quite an expensive reagent and we did not have access to it, so we used PEG 3350. This would have unfortunately compromised the pelleting efficiency.
Build - We followed a pre-existing protocol on salt precipitation of nucleic acids.
Test - We added salt and PEG to the crude lysate, followed by 30 minutes of cooling over ice and 30 minutes of centrifuge, collected supernatant and resuspended the pellet. There wasn’t a pronounced pellet, and the gel showed no significant pelleting and selection of nucleic acids.
Learn - We can increase the RNA concentration to try to get a better pellet.
Salt extraction: Reaction mixture optimisation
Design - To increase precipitate, we altered the reaction mixture.
Build - We raised the concentration of salt to increase the likelihood of getting a pellet.
Test - Still, no pellet was formed.
Learn - We already had isopropanol as a working extraction protocol that relied on precipitation and did not need a 30 minute cooling step on ice. It seemed reasonable not to go forth with salt and focus more on isopropanol instead.
FTA
FTA: Extraction from plant material
Design - We set out to develop a paper-based extraction method to extract miRNA from plants - based on FTA elute paper.
Build - Our protocol was adapted from a paper (1), with additional insight gained through communication with the paper’s author. RNaseOUT was added to the lysis buffer to mitigate against RNA degradation.
Test - Extraction was performed, and results were analysed using a NanoDrop spectrophotometer. Low RNA concentrations were obtained (~30 ng/µL), with a sharp peak at 230 nm.
Learn - Potential problem areas were identified:
- Crude plant extract may have been too dilute
- Soaking FTA disc in supernatant may be saturating the FTA paper, compromising adsorption of RNA
- Residual ethanol on the FTA disc was likely responsible for the sharp peak at 230 nm, and may have also interfered with elution of RNAs
- No insight into the size of RNA extracted
- Water was observed to condense on the underside of the PCR tubes.
FTA: optimising homogenisation, application of lysate and drying conditions
Design - We sought to optimise conditions for homogenising plant material, application of lysate to the FTA disc, drying of the FTA disc and elution.
Build - The amount of plant material used was increased to 200 mg, and the ratio of plant material to PBS used was decreased from 1:5 to 1:3. Instead of soaking the paper disc in the supernatant, 20 µL of supernatant was added to the FTA disc using a pipette. The FTA disc was left to dry for a longer period of time (30 mins), and different elution times (30 mins vs 1 h) were tested. Elutions were also performed in a thermocycler with a heated lid, with a larger volume of nuclease-free water used (30 µL).
Test - Extraction was performed, and results were analysed using a NanoDrop spectrophotometer.
Learn - Better RNA yields were obtained with the longer elution time. However, we needed more insight into the type of RNA being extracted.
Screentape extraction comparison
Design - We wanted a more accurate method to compare the several extraction techniques in parallel.
Build - We compared efficacy of all RNA extraction techniques using an RNA ScreenTape.
Test - We used the same lysate sample between all 4 tests, and TRIzol as a gold standard, performing the prior described extraction protocols. (The Lysate was also run on the ScreenTape, Wash and Elute are both from the Whatman extraction).
Learn - The isopropanol extraction seemed the most promising for future adaptation to being viable for farmers.
Farmer extraction
Farmer extraction: Extraction from plant material
Design - Designed a protocol that farmers could do with minimum technical experience or equipmenTest - based on isopropanol extraction.
Build - Pre-measured reagents into 4 PCR tubes:
Tube 1 50μL PBS + 5μL RNAse out,
Tube 2 80μL isopropanol + 20μL3M sodium acetate
Tube 3 100μL ethanol
Tube 4 40 μL NF water
Test - We tested our extract on the nanodrop ([RNA] = 45.1ng/μL, A260/280 = 1.26).
A similar RNA concentration to lab standard extraction, however the A260/280 was a fair bit lower.
Learn - The polyester string was not durable enough to withstand the 15 minute spin step (snapped after 7 minutes) and the PCR tubes did not remain fixed in the centrifuge. Additionally, the transfer of volumes between PCR tubes proved difficult due to surface tension.
Farmer extraction: Testing a protocol with improved hardware
Design - We redesigned the centrifuge to better hold the PCR tubes in place, used a kevlar string and a syringe with a pipette tip inserted, to resolve the prior issues with string snapping and transferring small volumes.
Build - The kevlar string proved too thin so we doubled it up three times. We also chose the best clamp design for the centrifuge from three designs we tested.
Test - The string held up easily for both the 15 minute and 5 minute spin steps, and the syringe made transferring volumes much easier, though the transferring of supernatant after homogenisation was still fairly technical. The PCR tubes stayed fixed in the centrifuge well. We tested our extract on the nanodrop ([RNA] = 96.2ng/μL, A260/280 = 1.41).
Learn - This protocol was viable for extracting RNA for farmers. Next steps would involve testing how accessible the protocol is, giving a written version to someone who has not conducted the extraction before.
Farmer extraction: Testing the accessibility of the protocol
Design - We aimed to test the feasibility of farmers conducting this extraction for themselves.
Build - An accessible protocol was written, and the 4 PCR tubes were premeasured and refrigerated overnight.
Test - The new extractor tested the protocol, finding most of it feasible - including the 15-minute spin step. Their advice was adding pictures to the protocol, and including a way of holding the PCR tubes up. Running this on the nanodrop yielded worse results than the prior extraction ([RNA] = 55.8 ng/μL and 260/280 = 1.53).
Learn - Leaving some reagents in the fridge overnight, especially the RNAse out (instead of using it from the freezer) may have caused the decrease in RNA extracted, however other factors, such as less plant material, may also have caused this. Next test will use no RNAse-OUT, as it was greatly increasing the cost of a single tesTest - to see if it is necessary)
Comparing extraction methods using RT-qPCR
Design - A method to quantify the amount of miRNAs obtained from each extraction method was required, which would also enable comparison of the different extraction methods.
Build - RT-qPCR was incorporated to determine if our miRNA of interest (ath-miR399f) was present in extracts obtained from the different extraction methods.
Test - RT-qPCR was performed on all four extraction methods, along with TRIzol as the gold standard for extraction.
Learn - We found that the farmer extraction protocol performed to a comparable standard to a normal lab isopropanol protocol, as well as other standard lab techniques.
Hybridisation
qPCR melt test to optimise buffers
qPCR melt test to optimise buffers: Cycle 1
Design - Research was done into the properties of RNA/DNA duplexes and how this affects SYBR green binding, along with the ability of SYBR green to bind short double-stranded oligonucleotides. As RNA/DNA duplexes have been shown to have intermediate physical properties between those of a DNA/DNA or RNA/RNA duplex, it was assumed that SYBR green could bind weakly. However, due to the lack of testing the ability of SYBR green binding to RNA/DNA duplexes, we could not be sure of the accuracy of the results. Additionally, 3-dimensional structures of the miR99f/mobile probe were modelled using a web server. This demonstrated that the structures had a minor groove which is part of the DNA double helix that SYBR green typically binds to. In addition to this, EvaGreen a dye that has a similar binding mechanism to DNA as SYBR green had been used in prior literature to obtain melting temperature values for short RNA/RNA duplexes, and DNA/DNA duplexes using a method similar to that which we were using. The concentrations of miRNA and DNA probe to use were decided. The composition of the buffers and order in which components are added to the plate were determined. The plate layout was designed to vary miRNA, DNA, and buffer conditions. Triplicate repeats on the plate were performed.
Build - Master mixes for the buffers were prepared that contained SYBR Green. DNA and miRNA dilutions were also prepared at 10x concentrations. The plate was pipetted and held at 25°C overnight.
Test - The plate was tested in the qPCR machine the next day and fluorescence readouts were obtained
Learn - The temperature range that the qPCR machine detected at did not include the predicted melting temperature of the miRNA/DNA hybrid.
qPCR melt test to optimise buffers: Cycle 2
Design - A single concentration of DNA and miRNA were decided on and the plate layout was designed. Buffer 1 was used for all wells. The purpose of the test was to determine the correct temperature range. 9 repeats for each condition were performed.
Build - The DNA and miRNA dilutions were prepared along with the buffer master mixes. The plate was pipetted again and left for 3hrs at 25°C.
Test - The plate was tested in the qPCR machine with an overall temperature range of 25 to 70°Ç.
Learn - Small peaks indicating the melting temperature appeared at around 30 to 40°C for wells containing both DNA and miRNA.
The controls with just DNA or miRNA or water (Plus buffer 1) showed melting temperatures of around 50 to 60°C. The temperature range of 25 to 70°C was a good one but the peak was small.
qPCR melt test to optimise buffers: Cycle 3
Design - Higher concentrations of miRNA and DNA were used to hopefully increase the signal. Only Buffer 1 was used again as the goal was to increase the height of the peak. The plate layout was designed. Triplicate repeats were performed.
Build - miRNA and DNA dilutions were prepared along with the mastermix. The plate was pipetted and left at 25°C overnight.
Test - Fluorescence read out was detected on a qPCR machine the following day with an overall temperature range of 25 to 70°C.
Learn - The melting temperature peaks remained small despite increasing concentrations. Inconsistent melting temperatures were observed between triplicate repeats which was thought to be due to condensation. The presence of a holding step of 70°C prior to melt curve testing may render the overnight hybridisation test irrelevant as the nucleic acids may degrade. SYBR Green may also not bind strongly to the hybrids due to short base pair overlap and SYBR Green not binding strongly to RNA/DNA duplexes. Troubleshooting was performed through contacts of biotech companies.
qPCR melt test to optimise buffers: Cycle 4:
Design - A DNA version of miR399f was ordered along with a longer version of the DNA probe (the new probe completely overlapped with the miR399f with a 21bp overlap). Concentrations of nucleic acids were increased using 250 nM, 500 nM, and 1000 nM. The hold step at 70°C was removed along with the pre-hybridisation step. The plate should instead be added straight to the qPCR machine, which should be heated to 45°C then cooled to 25°C at a rate of 0.015°C per second then increased to 70°C at the same temperature increase rate as prior. The plate layout was designed with controls.
Build - SYBR Green buffer was made and the plate was pipetted. DNA was first pipetted, then RNA, then the SYBR green into the plate. The plate was then sealed with an optical plate cover.
Test - The plate was added to the qPCR machine with the relevant temperature settings and run.
Learn - No clear melting peak was shown for the shorter probes. Clearer melting peaks were shown with the longer probes indicating that the length may be the issue. Volumes may be the issue as only 10 uL were used per well. The background reference was also set to ROX instead of None which may have caused the lack of a clear peak.
qPCR melt test to optimise buffers: Cycle 5:
Design - The SYBR Green concentrations used were increased to 2x and 5x in some tests. Concentrations of target RNA and DNA probes were decided on and Volumes were increased to 20 uL. The longer probe and DNA version of the miRNA were tested in this experiment alongside the normal miRNA and shorter probe due to the results from the prior experiment indicating that the lack of a defined melt curve is due to the overlap region being too short.
Build - The relevant SYBR green and DNA/RNA dilutions were prepared. The plate was pipetted according to the plate plan.
Test - The plate was run in the qPCR machine. The passive reference setting was set to none instead of ROX.
Learn - Unfortunately, due to the qPCR machine malfunctioning before we could run the plate, the plate had to be stored in a 4°C freezer for a week before another qPCR machine was found. This likely led to RNA degradation by RNAses and hence no defined melt curves in results.
In addition to this SYBR green method, another method for melting temperature was also determined utilising our Vantastar plate reader. This method involved 384-well UV transparent plates and the temperature control settings on the plate reader to mimic a temperature controlled UV spectrophotometer but with vastly lower volumes of solution. Two tests were performed as the experimental journey was limited by timing constraints.
UV spectrophotometry melt curve analysi
UV spectrophotometry melt curve analysis: Cycle 1
Design - Concentrations of oligonucleotides of 0, 1, and 4 uM were decided on based on our resources. Additionally, 40 uL was decided as the final volume for each well in an attempt to maximise the changes in the absorbance values upon melting. A buffer was found in literature. UV transparent plates were found and the plate reader settings were decided on. Controls were done with only RNA or only DNA at the concentrations used to validate that any changes seen in the wells with both RNA and DNA in them had occurred due to interactions between the RNA and DNA and not interactions due to the RNA and RNA, or DNA and DNA. Additional wells with no RNA or DNA and only buffer were also used.
Build - The plate reader programme was set up with relevant temperature increases and decreases. Oligonucleotides were diluted and the plate was pipetted as specified in the experiments tab.
Test - The plate was run in the plate reader on the specified programme set up prior.
Learn - The concentrations of nucleic acid used were too low. More repeats of each condition needed to be done to discern a clear trend (graphs attached in raw data).
UV spectrophotometry melt curve analysis: Cycle 2
Design - Concentrations of oligonucleotides were increased and 9 blank replicates were done. Additionally, 40 uL was decided as the final volume for each well in an attempt to maximise the changes in the absorbance values upon melting. The concentrations of the nucleic acids were increased during the second experiment, using concentrations of 0, 1, 5, 10, and 15 uM of nucleic acids. Negative controls were done to allow subtraction of the background reading of UV absorbance due to the buffer solution in both experiments. The concentrations and number of repeats that we were able to do were limited by the concentrations of the oligonucleotides that we had left. Additional controls were done with only RNA or only DNA at the concentrations used similar to those in cycle 1. Unfortunately, the plate reader was unable to obtain readings for the 260 nm spectrum for the wells containing both 15 uM of mobile DNA probe and miR399f.
Build - Relevant dilutions were formed and the plate was pipetted as specified in the plate plan in the experiments document.
Test - The plate was run in the plate reader on the specified programme set up prior.
Learn - The concentrations of nucleic acid were likely still too low. Additionally, due to the full volume of each well being used, the process of covering the plate with the optical plate cover may have led to solution spilling out of some of the wells and disrupting the seals. This may have led to increased variability between replicates. The plate reader detector got saturated on some wells with higher concentrations and resulted in an overflow reading, indicating that the gain may need to be altered (graphs attached in raw data).
Binding the APTMS to paper and testing with FITC
Binding the APTMS to paper and testing with FITC: Cycle 1
Design - The protocol was gathered from various literature sources online. We tried to skip the step where the FITC reaction is quenched by adding ammonium chloride and incubating for 2 hours to see if the washing step alone was sufficient for quenching.
Build - Paper discs were created, 95% ethanol was made up, protocol was followed and discs were incubated for different time periods in APTMS/ethanol solutions. Paper discs were dried in a vacuum chamber. Then they were incubated in FITC in a shaking incubator overnight. The Ammonium chloride step was skipped and the paper discs were washed with 95% ethanol.
Test - The discs were added to a half area 96 well plate and the fluorescence was measured using a plate reader.
Learn - The 3 hr incubation discs had the highest fluorescence indicating that this was the most efficient way to bind APTMS to paper. The FITC solution froze overnight with the paper inside it as it was incubated overnight at 4°C, so we modified the protocol to incubating at 25°C. When washing the discs with ethanol after overnight incubation with FITC, the ethanol still ran yellow after 6 washes and we did not wash it any more. The ammonium chloride may affect the quenching, so we decided to purchase some. Discs were only dried in a vacuum chamber for 30 minutes and were not completely dry when they came out. They were also difficult to put in and take out due to the narrow opening of the flasks. The 96-well half area plates were too small for the whole discs to fit in, however, we could not afford to get full area plates given our budget.
Binding the APTMS to paper and testing with FITC: Cycle 2
Design - The previous protocol was tested again. Additionally another test was designed which varied APTMS concentration. This test was designed to incubate the discs for 3hrs due to these discs having the highest fluorescence previously. The ammonium chloride quenching step is included in this protocol. Additionally, only Whatman No.1 was tested due to both discs of both types of paper showing similar fluorescence readings and a lack of time and resources to test both discs. Acetone was used to wash with instead of ethanol, similar to other existing protocols in literature as ethanol still ran yellow after the washes in the previous test.
Build - Paper discs were created, 95% ethanol was made up, protocol was followed and discs were incubated at different times and concentrations of APTMS/ethanol solutions. Each disc was added into solution such that they all finished incubation in the APTMS/ethanol solution at the same time.
Test - The discs were incubated at 25°C overnight instead of 4°C. Discs were washed with acetone instead of ethanol as suggested by other protocols in literature. Discs were kept in the vacuum chamber for one hour. To test the efficiency of drying paper discs in vacuum chambers, we soaked paper discs in 95% ethanol and placed them in either a weighing boat or vacuum chamber to dry.
Learn - The ammonium chloride quenching step made a difference in the results. Washing with acetone seemed to lead to a better result than with ethanol as by the end of the 6 washes it ran clear rather than yellow. The disks taken out of the vacuum were still wet and comparison between air drying and vacuum drying did not show significant difference. It was shown that the vacuum chamber is not greatly efficient in drying and rather time-consuming to place/take out, so we decided to air-dry instead.
Binding the APTMS to paper and testing with FITC: Cycle 3
Design - The previous two protocols were tested again. Additionally another test was designed which varied APTMS concentration. Additionally another test was performed to test a 2 step method using the addition of acid and alkali to the paper. This was tested at different alkali concentrations
Build - Paper discs were created, 95% ethanol was made up, protocol was followed and discs were incubated at different times and concentrations of APTMS/ethanol solutions. Each disc was added into solution such that they all finished incubation in the APTMS/ethanol solution at the same time.
Test - Discs were air dried instead of dried in a vacuum chamber. We put the discs in the half area plates. We attempted to cut out smaller circles out of the disks with a small holepuncher but the disks were too fragile to give good circular discs after all of the soaking in solution. Additionally, only the central 4 pixels were used to average the readings out from the plate reader data.
Learn - Holepunching small discs didn’t work and despite the larger discs not fitting into the half area plates, they still allowedsome fluorescence readout to be generated. Additionally, it was found that using the central 4 pixels of the matrix scan generated by the plate reader led to lower variability in results.
Binding the DNA probe to paper
Binding the DNA probe to paper: Cycle 1:
Design - We decided to use the one-step APTMS addition protocol due to this being faster than the two-step addition method and not leading to a significantly lower concentration of APTMS binding. A protocol for the ionic addition of DNA probes to cellulose paper was found in literature and followed. The concentration of the DNA probe added to the solution was increased 10-fold compared to literature to maximise DNA probe binding as we mathematically showed that if the prior APTMS addition step led to full binding then all of the DNA probe in the literature concentration could bind without saturating the APTMS. Different buffers were obtained from literature to test the best to dissolve DNA probes for the immobilisation reaction.
Build - Relevant buffers were made up and made nuclease free through filter sterilisation. Discs were treated with APTMS using the one step method. DNA probe solutions were added to the paper discs and incubated before washing.
Test - Discs were allowed to dry and placed in a half-area plate before detection of fluorescence levels in the plate reader.
Learn - The TE buffer worked best with the ionic method. This looked to be significantly better than with water and slightly better than with phosphate buffer. Not enough time was available and not enough data points were obtained to statistically test the buffers.
Binding the DNA probe to paper: Cycle 2:
Design - Our hypothesis that concentrations used in existing literature (on the ranges of 1-2μM) would not fully saturate the paper disc were tested by the addition of a range of concentrations of DNA, from 0-10µM, in order to test for any correlations between fluorescence and DNA probe concentration. In accordance to the previous results on the ionic addition of DNA probes, the DNA was added using a TE buffer.
Build - As before, TE buffer was made up nuclease free via filter sterilisation. The TE buffer was used to make a series of DNA dilutions, to produce DNA of 0, 1, 2, 3, 4, 5, 7.5, and 10µM concentration. Discs were treated with the one-step ethanol method to add APTMS, before being incubated with the different DNA concentrations before being washed.
Test - The discs were left to dry before being placed into half-area plates. The detection of the disc due to the Cy3 fluorophore on the immobilised DNA probe was then determined by a plate reader.
Learn - There was a clear initial trend, where increased concentrations of DNA led to increased concentrations of fluorescence. The three lowest fluorescences came from the 0, 1, and 2µM DNA concentrations, and the three highest fluorescences came from the 5, 7.5, and 10µM concentrations.
Binding the DNA probe to paper: Cycle 3
Design - Having optimised aspects of the ionic method of DNA probe addition to paper, we aimed to test whether the ionic method was able to bind similar amounts of DNA (and therefore produce similar fluorescence levels) as the more standard, and well documented, covalent method.
Build - Ionic DNA was made up in TE buffer, filter sterilised to make nuclease free, and in accordance with the previous results, we continued to make it up at a concentration of 10µM. DNA was then added to paper discs treated by the one-step ethanol method. Coupling was carried out via an EDC/sulfo-NHS method, where the APTMS bound paper was first pre-treated by a solution of terephthalic acid, EDC, and suflo-NHS, before being washed, and having the covalent DNA probe added. Covalent DNA was not made up in the TE buffer due to the free primary amine groups present in the TE buffer. These free amine groups would cause unintended coupling reactions, by themselves reacting with the coupling reagents used, instead of the amine group of the 5’ amine DNA. Instead, a PB buffer was used, once again filtered to make nuclease free. A negative control of APTMS paper without a DNA addition was also produced.
Test - The discs were allowed to dry before being added to a half-area plate to have the fluorescence read.
Learn - There was a very drastic difference between the fluorescence of the covalent and ionic discs, to the disc without DNA. This suggested that almost the entirety of the fluorescence of these two discs was due to the added DNA, and not due to innate fluorescence of the APTMS. There was not a large difference in fluorescence between the covalent and ionic methods, with the ionic actually having a slightly higher mean fluorescence. Considering this, along with the fact that ionic discs were both cheaper and easier to produce than the covalent discs, further experiments on the DNA sandwich hybridisation made a focus on whether we could continue to utilise the ionic method to potentially make it the primary method of DNA probe binding for the final test.
Sandwich hybridisation
Sandwich hybridisation: Cycle 1
Design - Concentrations for the DNA probe and target synthetic miRNA were decided on. The different tests run were decided and protocols were developed. This included finding a good blocking buffer and washing buffers for the paper and additionally determining a buffer for the pre-hybridisation test. We decided to use the one-step APTMS addition protocol due to this being faster than the two-step addition method and not leading to a significantly lower concentration of APTMS binding. Both Covalent and Ionic DNA probe addition methods were tested due to the possibility that one of them would interact with blocking to give a better final product than the other. TE was used to dissolve the immobilised DNA probe for the paper due to previous results in the binding of the DNA probe to paper test. PB was used to dissolve immobilised DNA probe for the paper addition in the covalent method due to the amine groups present in TE.
Sandwich hybridisation: Cycle 1
Design - Concentrations for the DNA probe and target synthetic miRNA were decided on. The different tests run were decided and protocols were developed. This included finding a good blocking buffer and washing buffers for the paper and additionally determining a buffer for the pre-hybridisation test. We decided to use the one-step APTMS addition protocol due to this being faster than the two-step addition method and not leading to a significantly lower concentration of APTMS binding. Both Covalent and Ionic DNA probe addition methods were tested due to the possibility that one of them would interact with blocking to give a better final product than the other. TE was used to dissolve the immobilised DNA probe for the paper due to previous results in the binding of the DNA probe to paper test. A phosphate buffer was used to dissolve immobilised DNA probe for the paper addition in the covalent method due to the amine groups present in TE.
Build - Paper discs were holepunched out and kept nuclease free. Relevant buffers were made up and filtered. Discs were added into vials and the DNA probe was added using relevant methods. Blocking was applied to relevant discs. The pre-hybridisation step was performed and a DNA probe/ target mIRNA solution was added to the discs.
Test - The discs were placed in a full-area plate and run in a plate reader.
Learn - Running the whole experiment from start to finish took more than 9 hours. Discs were still wet upon addition of the pre-hybridisation solution and therefore the fluorescence values obtained were unreliable.
Sandwich hybridisation: Cycle 2:
Design - We requested to be allowed in the lab for a longer period of time by our lab supervisor to allow sufficient time for the discs to dry before addition of the pre-hybridisation solution. In addition to this, fewer discs were tested on the same day to shorten overall washing times of the discs and therefore allow them all more time to dry and make incubation time between disc tests more consistent. The rest of the protocol was followed as usual.
Build - Paper discs were holepunched out and kept nuclease free. Relevant buffers were made up and filtered. Discs were added into vials and the DNA probe was added using relevant methods. Blocking was applied to relevant discs The pre-hybridisation step was performed and a DNA probe/ target mIRNA solution was added to the discs.
Test - The discs were placed in a full area plate and run in a plate reader.
Learn - The discs had time to dry before the pre-hybridisation solution was added. The discs prepared using the covalent DNA probe addition on the same day as testing with the pre-hybridisation solution took 12 hours to test in total. We were not allowed in the lab for this long except on the day that this method was tested therefore no experimental repeats for this condition were done. Instead, the ionic DNA probe addition method was used to ensure that the experiment had time to run in a day.
RCA
Protein purification
Protein purification: Cycle 1:
Design - The protocol was developed according to manufacturer’s recommendations (1) and scientific literature (2, 3), but subsequently modified by adding 5% Glycerol in 1X PBS during protein concentration and supplementing all the buffers with 1 mM final concentration of DTT to prevent any disulphide formation between proteins upon cell lysis.
- Binding/Lysis buffer: 20 mM sodium phosphate, 1 mM DTT, 300 mM NaCl, pH 7.4
- Wash Buffer: 20 mM sodium phosphate, 300 mM NaCl, 1 mM DTT, 5 mM Imidazole, pH 7.4
- Elution Buffer: 20 mM sodium phosphate,300 mM NaCl, 1 mM DTT, 500 mM Imidazole, pH 7.4)
- We decided to employ a homogenisation technique to lyse the bacterial cells to obtain the lysate. Due to the immense amount of products, we increased the number of washes from 3 to 6 times.
Build - We lysed the bacterial cells via homogenisation, centrifuged to clarify the lysate and performed the purification on the supernatant with the Ni Spin columns the recommended buffer compositions. After purification, the eluted products were washed to remove remaining imidazole with 5% Glycerol in 1X PBS and were subsequently concentrated using a centrifugal filter with an appropriate molecular cut-off of 10 kDa.
Test - 6 µL of samples were taken from the products at different stages, namely supernatant, flow-through, wash, elute 1 and 2, and concentrated protein. The samples were mixed with 4X NuPAGE loading dye and then analysed with SDS-PAGE at 200 V for 40 mins to determine the presence of the protein and their purity. Additionally, the concentrated protein was measured of their concentration and A260/A280 using a Nanodrop spectrophotometer.
Learn - The results from the Nanodrop spectrophotometer were as follow:
| Proteins | Final Concentration 1 | Final Concentration 2 | Final Concentration 3 | Average Final Concentration | A260/A280 |
|---|---|---|---|---|---|
| NLuc | 1.315 g/L | 1.328 g/L | 1.302 g/L | 1.315 g/L | 1.02 |
| mNeonGreen | 1.244 g/L | 1.127 g/L | 1.142 g/L | 1.171 g/L | 1.07 |
- The A260/A280 ratios indicated the contamination of DNA in the concentrated protein products as the zinc-finger proteins could bind to their target sequence of DNA. Thus, we should add DNase to resolve the contamination issue and increase the salt concentration which facilitates the precipitation of DNA out of the cell lysate. The SDS-PAGE result is shown in Figure 20, and it highlighted several problems: overloading of the sample and no discrete bands identified as the protein of interest. Hence, to avoid overloading of the sample which leads to a smear, we should load less sample relative to the column volume.
Protein purification: Cycle 2:
Design - Due to the large quantity of the product, the Ni spin columns used in Cycle 1 were overwhelmed and led to sample overloading. We thus changed the protocol and adapted one used by the lab from which we borrowed equipment. The protocol employed buffers of different compositions:
- Buffer A: 500 mM NaCl, 20 mM Imidazole, 50 mM Tris-HCl, 5% Glycerol, pH 7.5
- Buffer B: 500 mM NaCl, 250 mM Imidazole, 50 mM Tris-HCl, 5% Glycerol, pH 7.5
- Lysis buffer: 250 mM Buffer A, 3 mM DTT, 1-2 tablets crushed with P1000 tips PIC (cOmplete™ Protease Inhibitor Cocktail), 0.1 mg/mL Lysozyme, 0.1 mg/mL DNase
- DNase was added to the lysis buffer to remove any DNA contamination, and the Ni-NTA column volume was increased to 0.5 mL to accommodate the amount of product.
Build - We lysed the bacterial cells via homogenisation, centrifuged to clarify the lysate and performed the purification on the supernatant. The purification also included another step to incubate the supernatant with the resin for at least 1 hr, which allowed the coordination between the 6X His Tag with the Ni ion. Similarly, the eluted products were washed to remove remaining imidazole with 5% Glycerol in 1X PBS and were subsequently concentrated using a centrifugal filter with an appropriate molecular cut-off of 10 kDa.
Test - 6 µL of samples were taken from the products at different stages, namely supernatant, flow-through, wash, elute, and concentrated protein. The samples were mixed with 4X NuPAGE loading dye and then analysed with SDS-PAGE at 200 V for 40 mins to determine the presence of the protein and their purity. Additionally, the concentrated protein was measured of their concentration and A260/A280 using a Nanodrop spectrophotometer.
Learn - The results from the Nanodrop spectrophotometer were as follow:
| Proteins | Final Concentration 1 | Final Concentration 2 | Final Concentration 3 | Average Final Concentration | A260/A280 |
|---|---|---|---|---|---|
| NLuc | 0.799 mg/mL | 0.774 mg/mL | 0.785 mg/mL | 0.786 mg/mL | 0.97 |
| mNeonGreen | 0.655 mg/mL | 0.687 mg/mL | 0.650 mg/mL | 0.664 mg/mL | 1.36 |
- These results provided evidence that the concentrated NLuc product was less contaminated with DNA than the previous attempt, but the concentrated mNeonGreen protein saw a more severe DNA contamination. Thus, a higher salt concentration e.g. LiCl ought to be added to precipitate DNA out. For SDS-PAGE, Figure 21 shows the improvement of the overloading problem from the previous cycle, and the tentative mNeonGreen fusion protein was labelled with an arrow. However, like the previous attempt, both fusion proteins were not evident on the gel. This may be due to a missing Alanine residue at the 188th position of the mNeonGreen expressed. Hence, in the future, a plasmid with the correct amino acid sequence should be used for transformation and protein expression. Also, the expression conditions e.g. IPTG concentrations and expression temperature could be altered to increase protein expression level. For the higher purity of fusion proteins, additional tags could be introduced, and an extra purification process could be executed for that tag. Western Blot could be performed to detect the presence of the tagged fusion protein with better clarity.
Verifying circularisation
Verifying circularisation: Finding appropriate agarose concentration
Design - Lucigen’s protocol on CircLigase™ II ssDNA Ligase (1) for the circularisation of linear padlock probes involved verifying the reaction had taken place with PAGE and adding exonucleases to the reaction mixture to degrade remaining linear probes. However, in an effort to save time and reagents, we decided to test out both verification and purification in the same step with Agarose gel electrophoresis. After separating the circularised and linear probes on a gel such that they are resolved as separate bands, we could use a DNA Gel extraction kit to obtain the circularised probe. (DNA Gel extraction kits can only be used to retrieve DNA product from agarose and not PAGE gels). To achieve satisfactory separation of the bands, as the two forms of the probe differ only in topology and not in the number of base pairs, we recognised the need to use gels of high concentrations. Hence, we tested out agarose gels of different, high concentrations in resolving bands for a low range DNA ladder.
Build - ThermoFisher’s GeneRuler Low Range DNA Ladder was run in gels of 2.5%, 2.75% and 3% for 60, 90 and 120 minutes respectively, in 1X TBE buffer at 100 V.
Test - The gels were stained by 1X GelRed in 1X TBE buffer, then visualised with a transilluminator.
Learn - Figure 22, 23 and 24 represented the Agarose gel electrophoresis results at 2.5%, 2.75% and 3%, respectively. The bands on the 2.75% and 3% Agarose gels appeared warped, presumably because the gels did not set properly. We realised that getting the agarose to dissolve and set properly was generally a problem with gels of such high concentrations. Additionally, the particular fuzziness of bands in the 3% Agarose gel may be because the gel was run for an excessive amount of time.
On the contrary, the bands on the 2.5% Agarose gel seemed clean and regular although there was a concern over the lowest bands having poor resolution. Still, we decided that it would be an appropriate concentration to load our circularisation products.
Verifying circularisation: Attempts on 2.5% Agarose
Design - After confirming that 2.5% Agarose gel was potentially a suitable concentration for obtaining circularised probes, we tried running our circularisation reaction products on 2.5% Agarose gels.
Build - After circularising our probes with CircLigase™ II ssDNA Ligase, we ran the reaction mixture alongside the GeneRuler Low Range DNA Ladder and its linear DNA probe form on a 2.5% Agarose gel. This was conducted at 100 V for 60 min in 1X TBE buffer.
Test - The gels were stained with 1X GelRed in 1X TBE buffer, then visualised with a transilluminator.
Learn - For all of the following gels, their lanes contained (from left to right): ladder, linear DNA, circularisation products after 60 min and after 90 min. Figure 25, 26 and 27 represented our 2.5% Agarose gel electrophoresis results of Attempt 1, 2 and 3, respectively. Our three attempts to resolve the circularised probe from the linear probe on a 2.5% Agarose gel did not work. The bands on all three gels were warped, indicating the gels did not set properly. More importantly, even though bands were present in the lanes where the circularisation product was loaded, they were fuzzy and inconclusive as to whether or not circularisation took place. In our first two attempts, only one band was essentially visible. In our third attempt, there appeared to be two close bands, but as there was a line running through the entire gel where the bands seemed to separate, it was inconclusive as to whether it represented 5’-adenylated intermediates of the circularisation reaction or actually linear padlock probes.
Hence, we decided to pivot to using a smaller portion of the reaction mixture for verifying circularisation with PAGE, and proceed to use an RNA cleanup kit from New England Biolabs (NEB) to obtain circularised probes from the rest of the reaction mixture.
Verifying circularisation: Initial PAGE attempt
Design - As per Lucigen’s protocol on CircLigase™ II ssDNA Ligase [@lucigen_ma298e-circligase-ii-ssdna-ligase_jul2019_2indd_2019], we decided to try verifying circularisation with a 20% Acrylamide/8 M Urea denaturing PAGE gel. The protocol was also adapted from [@thermofisher_man0011970_denaturing_polyacrylamideurea_gel_electrophoresis_ug_2012] and [@summer_denaturing_2009].
Verifying circularisation: Initial PAGE attempt
Design - As per Lucigen’s protocol on CircLigase™ II ssDNA Ligase, we decided to try verifying circularisation with a 20% Acrylamide/8 M Urea denaturing PAGE gel, with the protocol adapted from literature (2, 3).
Build - The gel was pre-run for 15 minutes to heat the gel up and to remove remaining urea and unpolymerised acrylamide from the gel at the constant voltage of 35 V. We then ran the circularisation product alongside the GeneRuler Low Range DNA Ladder and linear DNA probe on the gel at 60 V for 1 hr.
Test - The gels were stained by 1x GelRed in 1X TBE buffer, then visualised with a transilluminator.
Learn - As seen in Figure 30, the smearing of the bands was quite serious across all lanes, and this may be a problem of an overly high voltage. Therefore, we decided to rerun the gel at a lower voltage but for a longer time. Particularly serious smearing in the lane containing linear DNA might be due to degradation of the probe.
Verifying circularisation: PAGE with Lower Voltage
Design - Keeping with a 20% Acrylamide/8 M Urea denaturing PAGE gel, we lowered the voltage that the gel was run on from 60 V to 42 V and increased the running time from 1 hr to 200 mins.
Build - The gel was pre-run for 30 minutes to heat the gel up and to remove remaining urea and unpolymerised acrylamide from the gel at the constant voltage of 35 V. We ran the circularisation product alongside the GeneRuler Low Range DNA Ladder and linear DNA probe on the gel at 42 V for 200 mins.
Test - The gels were stained by 1x GelRed in 1X TBE buffer, then visualised with a transilluminator.
Learn - From Figure 31, although the ladder took on a strange shape (possibly due to overloading), two bands were clearly visible and separate in the lane containing the circularisation reaction mixture, representing the circularised (above) and linear (below) probe respectively. Hence, probe circularisation was successfully verified. We decided to keep to these reaction conditions for verifying circularisation from this point onwards.
Circularisation
Circularisation: Cycle 1
Design - The protocol was derived from the supporting [@lucigen_ma298e-circligase-ii-ssdna-ligase_jul2019_2indd_2019] document from Lucigen on CircLigase™ II ssDNA Ligase. We followed the protocol for the volumes of all components and decided to incubate the reactions for 60 mins as recommended and 90 mins to investigate whether or not 60 and 90 mins would yield a similar amount of circular products. The padlock probe used in this experiment was designed for miR399f and BRET strategy.
Circularisation: Cycle 1
Design - The protocol was derived from the supporting document from Lucigen on CircLigase™ II ssDNA Ligase (1). We followed the protocol for the volumes of all components and decided to incubate the reactions for 60 mins as recommended and 90 mins to investigate whether or not 60 and 90 mins would yield a similar amount of circular products. The padlock probe used in this experiment was designed for miR399f and BRET strategy.
Build - We performed the reactions as mentioned above and inactivated the enzymes by incubating the mixture at 80℃.
Test - We analysed the results with both 2.5% Agarose gel electrophoresis and 20% Acrylamide/8 M Urea PAGE, including the linear probe for comparison, and both gels were stained by 1x GelRed in 1X TBE buffer. They were then visualised with a transilluminator.
Learn - Both agarose gel and acrylamide gel as shown in Figure 32 and 33 showed that some probes may have formed a 5’-adenylated intermediate but not validated, so extending the incubation time would be appropriate for the next cycle.
Circularisation: Cycle 2
Design - The protocol was modified from Lucigen by extending the incubation time to 2 hr. The probe used for the experiment was also changed to the one designed for miR399f and G-quadruplex strategy.
Build - We performed the experiment, following the protocol with the incubation time amendment.
Test - We analysed the result with 20% Acrylamide/8 M Urea PAGE, including the linear probe for comparison, and the gel was stained by 1x GelRed in 1X TBE buffer. It was then visualised with a transilluminator.
Learn - The gel image in Figure 34 demonstrated the successful circularisation of the padlock probe with a band representing a circular probe above the linear probe’s band. For the next cycle, we should increase the incubation time further to examine whether or not increasing the incubation time would improve the yield.
Circularisation: Cycle 3
Design - The protocol was modified again by extending the incubation period to 3 and 3.5 hr to increase the yield of the reaction. The same probe as Cycle 2 was used to investigate the effect of incubation time.
Build - We performed the experiment again, following the protocol with the incubation time amendment.
Test - We analysed the results with 20% Acrylamide/8 M Urea PAGE, including the linear probe for comparison, and the gel was stained by 1X GelRed in 1X TBE buffer. It was then visualised with a transilluminator.
Learn - Both conditions gave similar results in terms of yield as observed in Figure 35. Hence, we concluded that by 3 hr, the reaction might have already reached saturation.
Circularisation: Cycle 4
Design - The protocol was modified by reducing the amount of CircLigase™ II ssDNA Ligase by half to save the cost of the final product and due to the limited amount of enzyme available. We also doubled the incubation time from 3 to 6 hr and decided to use the probes both designed for miR399f, one for BRET and the other for G-quadruplex.
Build - We implemented the new protocol with both the reduced CircLigase™ II ssDNA Ligase and doubled incubation time.
T: We analysed the results with 20% Acrylamide/8 M Urea PAGE, including both linear probes for comparison, and the gel was stained by 1X GelRed in 1X TBE buffer. It was then visualised with a transilluminator.
Learn - The outcomes were as seen in Figure 36: there were faint bands that should correspond to circular probes although the yield seemed to be lower than when the recommended volume of the enzyme was used. We proceeded with this protocol for the rest of the project.![][image35]
G-quadruplex
G-quadruplex: Cycle 1
Design - From the protocol of [@li_portable_2019], we incorporated a step for adding Thioflavin T (ThT) into the RCA mixture using the concentration used in [@renaud_de_la_faverie_thioflavin_2014]. We then designed an experiment in solution phase investigating how fluorescence signal changes with varying miR399f concentrations - a 10-fold serial dilution from 75 nM to 750 fM.
Build - We performed the aformentioned serial dilution on miR399f standard and conducted RCA reactions with ThT. Nuclease-free water was added as a negative control.
G-quadruplex: Cycle 1
Design - From the protocol of Li et al. (1), we incorporated a step for adding Thioflavin T (ThT) into the RCA mixture using the concentration used in Renaud de la Faverie et al. (2). We then designed an experiment in solution phase investigating how fluorescence signal changes with varying miR399f concentrations - a 10-fold serial dilution from 75 nM to 750 fM.
Build - We performed the aforementioned serial dilution on miR399f standard and conducted RCA reactions with ThT. Nuclease-free water was added as a negative control.
Test - After incubation for 2 hr and 15 mins (from the previous experiment adding ThT simultaneously and after RCA), the products were loaded into a black, half-area 96-well plate. Fluorescence was measured by a microplate reader with the excitation wavelength of 425 nm (Bandwidth 16 nm) and the emission scan function from 470 to 550 nm. The gain set was 10% as the enhancement dynamic range was not compatible with spectral scans.
Learn - As seen in Figure 37, the negative control resulted in a relatively high signal possibly due to miR399f contamination. Thus, the negative control condition should be mixed first prior to any work with miR399f, and miR399f dilution must be added into the RCA mixture in increasing dilution. Gloves must also be changed after the serial dilution of miR399f before adding miR399f to any RCA mixture.
G-quadruplex: Cycle 2
Design - We followed the previous protocol in cycle 1 and adjusted the order that miRNA and water was added to initiate RCA reactions. We prioritised adding water as the negative control first and adding progressively higher concentrations of miR399f to limit any cross-contamination from miR399f standards of higher concentrations. Additionally, we included a control by adding a G-quadruplex Assay buffer without Thioflavin T (ThT) to examine the inherent fluorescence activity of the RCA product. To test the selectivity of the test for miR399f, we also performed an RCA reaction with 75 nM miR16 standard. miR16 is present in humans and is 56% identical to miR399f when aligned together. The miR399f concentrations used ranged from 75 nM to 7.5 pM, and the 750 fM was excluded due to the extra test conditions implemented in the experiment.
Build - We performed the aformentioned serial dilution on miR399f standard and conducted RCA reactions with ThT whilst implementing measures to mitigate any contamination. Nuclease-free water was added as a negative control.
Test - After incubation for 2 hr and 15 mins, the products were loaded into a black, half-area 96-well plate as in cycle 1. Fluorescence was measured by a microplate reader with the excitation wavelength of 425 nm (Bandwidth 16 nm) and the emission scan function from 470 to 550 nm. The gain was set to 20%.
Learn - Figure 38 showed the miR399f-concentration dependence of the fluorescence readouTest - the higher the miR399f concentration, the more intense the fluorescence signal. The graphs confirmed that without ThT, there was a substantially lower fluorescence readout. With regards to selectivity, at the same concentration, miR399f produced a signal at 486 nm that was around 11 times that of miR16, which could be concluded that there was some selectivity for miR399f to which the probe is designed to bind. Compared to the negative control, the signal emitted from the condition with 75 nM (effective concentration 2.5 nM) miR399f was around 38-fold that of the negative control.
Dry Lab
Kinetic modelling
Kinetic modelling: Designing the ODEs
Design - We used basic equilibrium, and michaelis menten kinetics and incorporated quantum yield and BRET efficiency transfer mathematics to model the overall BRET RCA system.
Build - We wrote the code to solve the ODEs for the system.
Test - The shapes of the curves were in line with what is expected for enzyme kinetics.
Learn - The next steps are to use this model to optimise our actual assay.
Kinetic modelling: Adding sampling optimisation
Design - We decided to start by optimising our sampling conditions.
Build - We added an algorithm that screens through all emitted light for a range of wavelengths throughout the reaction time, to find the maximum signal to noise ratio.
Test - We wrote and ran the code.
Learn - The optimal wavelength to sample was very similar to the corresponding peak emission of MNeon Green - indicating the optimisation was successful.
Kinetic modelling: Adding saturation kinetics
Design - We added saturation kinetics following monomolecular growth to the model.
Build - We did the calculations and derived a new equation for the rate of RCA product increase, determining the saturation constant.
Test - We wrote and ran the code.
Learn - Adding the saturation kinetics resulted in apparently long term stability of the concentrations of several of the reagents, indicating the improved accuracy of the model.
Kinetic modelling: Determining RET parameters
Design - We aimed to determine distance in AlphaFold between luminescent substrate site and fluorescent residues for foster distance calculations.
Build - We used AlphaFold, the fluorescent residue of mNeon and 3 residues in NLuc active site to determine distance for BRET (an improvement on using the distances between the Zinc Finger binding sites)
Issues were the linker was low confidence, and NLuc barrel did not fully form in simulations.
Test - Our new distance was lower than the Forster DIstance, unlike the original, making it likely a more realistic estimate of the distances seen in the source paper.
Learn - This greatly improved the signal to noise ratio, demonstrating how effective BRET is for signal amplification.
Hand Centrifuge
Hand Centrifuge: Cooling capacity test 1
Design - Starting from the established hand centrifuge, we noted that loaded samples got fairly warm while spinning, so sought a way to keep the sample cool.
Build - To reduce this effect we designed a new centrifuge with added water channels, next to the sample, which could then be frozen.
Test - In order to test the cooling effect we scaled up models to fit a thermometer in, and recorded temperature changes over time.
Learn - When spinning our first iteration for any length of time, we realised the ice melted very quickly, and then fell out. In order to reduce the rate of ice melting we would need to cover it, reducing the drag it experiences, as well as keeping it from falling out.
Hand Centrifuge: Cooling capacity test 2
Design - We needed a new centrifuge where the ice does not fall out after only short periods of spinning.
Build - We designed a new centrifuge model (right), where the water channels were covered.
Test - we found that this prevented the ice from melting as fast, prevented it from falling out, and also kept the sample cool. We then recorded the rate at which the two versions lost weight (melted), over 14 minutes.
L- Securing the tubes themselves did prove difficult however. Normally they would be taped down, but the cooled centrifuge resulted in some condensation on the surface, hence the tape failed to stay fixed, and the tubes fell out.
Hand Centrifuge: Securing PCR tubes.
Design - Three new designs for fixing the tubes in place were planned.
Build - Partial centrifuge models with the 3 mechanisms were printed out.
Test - The 3 designs and original model were compared and rankeDesign - the original was lowest, then the twisting one (middle), then the round clamp (top), then the straight line clamp (bottom) was the best fit.
Learn - The bottom design gave the best fit to the PCR tube, so was integrated into the final centrifuge prototype.
Hand Centrifuge: Improving the string integrity
Design - We needed a new knot and string to withstand high speeds for a longer duration.
Build - We used a square reef knot with kevlar string, instead of polyester.
Test - We used this string and knot for several full extractions without it breaking.
Learn - This improved kevlar string with the square reef knot is sturdy enough for successful RNA extraction.
Dark box
Dark Box: Proof of concept
Design - We needed a standardised but cheap detection method, so decided on developing a darkbox.
Build - We 3D-printed a basic dark box model. We also designed an app to determine the magnitude of the differences in brightness between two photos, via a subtraction method.
Test - To test it we used a paper disc with water and with 5 μM Thioflavin T.
Learn - We found that increasing the angle to overhead lighting improved the difference score of the subtraction image. We also made the box taller, to make it easier to focus the image. We added a stand to hold the rest of the phone stable, to try to make pictures as reproducible as possible.
Dark Box: Improvements to the dark box
Design - We sought to integrate our prior findings into an improved and functional box.
Build - We designed and printed the new dark box. In order to test whether darker walls made the subtraction of images more effective, we tested the box as printed (white), with black tape on the walls, and with the remainder coloured in with black marker.
Test - We tested the difference detected under the 3 circumstances with a range of Thioflavin T concentrations.
| Tht concentration (μM) | <- Rank | white | <- Rank | black | <- Rank | black on black | <- Rank | |
|---|---|---|---|---|---|---|---|---|
| 1 | -ve | - | - | - | ||||
| 0 | 7 | 78 | 2 | 112 | 5 | 86 | 6 | |
| 0.1 | 6 | 25 | 5 | 98 | 6 | 32 | 7 | |
| 2 | 0.5 | 5 | 5 | 7 | 135 | 3 | 99 | 3 |
| 1 | 4 | 13 | 6 | 47 | 7 | 99 | 3 | |
| 3 | 5 | 3 | 35 | 4 | 161 | 1 | 106 | 2 |
| 10 | 2 | 52 | 3 | 122 | 4 | 119 | 1 | |
| 100 | 1 | 90 | 1 | 153 | 2 | 95 | 5 | |
| SRCC values: | 0.3571428571 | 0.5357142857 | 0.5628780358 |
For n=8, p=0.05, one-tailed Spearman’s Rank test, the critical value is 0.6429 - the data had too much noise.
Learn - The software still had quite a lot of random noise from slight changes to lighting, the white PLC seemed best for distinguishing colour. We will modify the sample tray to allow for image cropping around the sample only, then repeat this test.
Dark Box: Improvements to the smartphone app
Design - We now wanted to improve the app of the detection system to reduce the detected noise.
Build - We tweaked the app and added dots to mark a square on the control part of the tray, to crop the image before subtraction, to reduce noise, as well as improving the camera focus.
Test - we repeated the prior experiment (with an additional 50μM Thioflavin T measurement). The black on black box was too dark to distinguish the dots marking the square for image cropping, so we did not use it.
Learn - Both of these graphs showed a marked improvement overalLearn - due to the software changes. The phone (Redmi Note 9 - 5G) was able to detect an increase in fluorescence as concentration increased. However, the 100μΜ showed lower brightness than the 50μΜ, showing that both the app and the box are far from perfect.
We suspected that the plateau at 50,000 pixels is representative of the size of the sample disc, in the images taken, and here we see the lightbox have a more consistent plateau across several concentrations, indicating that it was better for discerning the sample from the rest of the dark box.
References
FTA extraction
- Leong, S. M., Tan, K. M.-L., Chua, H. W., Huang, M.-C., Cheong, W. C., Li, M.-H., Tucker, S., & Koay, E. S.-C. (2017). Paper-Based MicroRNA Expression Profiling from Plasma and Circulating Tumor Cells. Clinical Chemistry, 63(3), 731–741. https://doi.org/10.1373/clinchem.2016.264432
Protein purification
- NEBExpress® Ni Spin Column Reaction Protocol (NEB #S1427) | NEB. (n.d.). https://www.neb.com/en-gb/protocols/2019/08/28/nebexpress-ni-spin-column-reaction-protocol-neb-s1427
- Li, Y., Zhou, L., Ni, W., Luo, Q., Zhu, C., & Wu, Y. (2019). Portable and Field-Ready Detection of Circulating MicroRNAs with Paper-Based Bioluminescent Sensing and Isothermal Amplification. Analytical Chemistry, 91(23), 14838–14841. https://doi.org/10.1021/acs.analchem.9b04422
- Li, Y., Yang, P., Lei, N., Ma, Y., Ji, Y., Zhu, C., & Wu, Y. (2018). Assembly of DNA-Templated Bioluminescent Modules for Amplified Detection of Protein Biomarkers. Analytical Chemistry, 90(19), 11495–11502. https://doi.org/10.1021/acs.analchem.8b02734
Verifying Probe circularisation
- Lucigen (Cambio). (2019, July). CircLigase II ssDNA ligase: Protocol manual (MA298E‑CircLigase II ssDNA Ligase). https://www.cambio.co.uk/library/images/html_images/2022%20protocols/manual_CircLigase-II-ssDNA-Ligase.pdf
- Thermo Fisher Scientific. (2012). Denaturing polyacrylamide/urea gel electrophoresis (Manual No. MAN0011970). https://assets.fishersci.com/TFS-Assets/LSG/manuals/MAN0011970_Denaturing_PolyacrylamideUrea_Gel_Electrophoresis_UG.pdf
- Summer H, Grämer R, Dröge P. Denaturing urea polyacrylamide gel electrophoresis (Urea PAGE). J Vis Exp. 2009 Oct 29;(32):1485. doi: 10.3791/1485. PMID: 19865070; PMCID: PMC3329804.
Circularisation
- Lucigen (Cambio). (2019, July). CircLigase II ssDNA ligase: Protocol manual (MA298E‑CircLigase II ssDNA Ligase). https://www.cambio.co.uk/library/images/html_images/2022%20protocols/manual_CircLigase-II-ssDNA-Ligase.pdf
G-quadruplex
- Li, Y., Zhou, L., Ni, W., Luo, Q., Zhu, C., & Wu, Y. (2019). Portable and Field-Ready Detection of Circulating MicroRNAs with Paper-Based Bioluminescent Sensing and Isothermal Amplification. Analytical Chemistry, 91(23), 14838–14841. https://doi.org/10.1021/acs.analchem.9b04422
- Renaud de la Faverie, A., Guédin, A., Bedrat, A., Yatsunyk, L. A., & Mergny, J. L. (2014). Thioflavin T as a fluorescence light-up probe for G4 formation. Nucleic acids research, 42(8), e65. https://doi.org/10.1093/nar/gku111
Follow us!