Results
Results
Our project set out to develop a sensitive and reliable tool for detecting Prymnesium parvum blooms. We can confidently say that we have achieved many successes in this endeavor. However, like any scientific journey, we have also encountered challenges and areas that require further refinement.
On this page, we share the results we have achieved during our iGEM year, as interpreted and understood by our team. We have included findings from each phase of the project, along with a brief summary of successes and areas needing improvement, followed by our plans for future work to build on these outcomes.
We can confidently say that we have succeeded in engineering functional SHERLOCK reaction components that allow for the detection of Prymnesium parvum DNA.
To program the system to target a specific region of the Prymnesium parvum genome, we needed to design an effective combination of RPA primers and a crRNA molecule. We carefully considered which region of the genome to focus on, ultimately selecting the ITS2 (Internal Transcribed Spacer 2) sequence for specific detection. The ITS2 region, located between the 5.8S and nuclear large subunit rRNA genes [1], is commonly used for species identification due to its variability between species and conservation within a species. This makes it an ideal target for species-specific primers, including those used to detect Prymnesium parvum [2].
We designed crRNA molecules to target ITS2, which we named PrymCrRNA1 (BBa_K5087022) and PrymCrRNA2 (BBa_K5087023). Anticipating that some crRNAs might not be functional, we designed two initial crRNA sequences to start our experiments.
Figure 1. Binding sites of PrymCrRNA1 and PrymCrRNA2 to the Prymnesium parvum genome.
We finalized three primer pairs, some of which were adapted from literature, while others were designed from scratch:
With our three primer pairs and two crRNA designs, we were left with 18 possible combinations of SHERLOCK components.
Figure 2. Positioning of all designed primers and crRNAs on the ribosomal cistron of Prymnesium parvum genomic DNA.
We validated the functionality of all but one of our primer-crRNA combinations using both fluorescence and Lateral Flow Assay (LFA) readouts.
The one combination we found to be non-functional was GalF-GalR + PrymCrRNA1, as it consistently produced false positives (positive results in negative controls), such as what can be observed in Figure 3. In this test, the “negative control + prymcrRNA1” sample produced a fluorescence signal. This sample included water added to an RPA mix containing Gal primers. This outcome underscored the critical importance of including negative controls in our experiments, which we had been diligently maintaining.
We hypothesize that both Gal primers inadvertently serve as a target for PrymCrRNA1, likely due to the forward primer’s proximity to the PrymCrRNA1 binding site (Figure 2). However, we did not consistently observe the same effect when the GalF primer was paired with a different reverse primer, leading us to conclude that GalF can be used in combination with other reverse primers.
Figure 3. SHERLOCK fluorescence readout result – false positives given by the GalF-GalR + PrymCrRNA1 combination (red line).
The rest of primer-crRNA combinations were validated through a mixture of preliminary testing using both LFA and fluorescence readouts. We selected only representative results from the multiple tests conducted (Figures 4-6).
Figure 4. SHERLOCK fluorescence readout result – GalF-GalR + PrymCrRNA2, performed on decreasing concentrations of Prymnesium parvum DNA.
The chart in Figure 4 shows that PrymCrRNA2 in combination with the Gal primers was effective in detecting Prymnesium parvum DNA, even at a 200 fM concentration. The detection of 200 fM by the assay was a promising sign for the LOD value.
Figure 5. SHERLOCK fluorescence readout result – Mod, Alt and Gal primers, performed on a 530 nM sample of Prymnesium parvum DNA.
Figure 6. SHERLOCK fluorescence readout result – Mod, Alt and Gal primers, performed on a 200 fM sample of Prymnesium parvum DNA.
All of the combinations shown in Figure 5 were effective in detecting Prymnesium parvum DNA. The best primer-PrymcrRNA pair appeared to be ModF-ModR + PrymCrRNA1, as it achieved high fluorescence intensity for both 530 nM and 200 fM samples. No other combination was effective in detecting the 200 fM target concentration.
After validating our designs, we could have concluded the testing process. However, we aimed to optimize the system further by identifying the most efficient primer-crRNA combination among all the possible configurations. We performed fluorescence-based assays where we assessed all of our candidate combinations in one assay so that we could compare their performance to each other (Figures 7-9).
Specifically, in this test we aimed to determine if mixing forward and reverse primers as well as PrymCrRNAs between the sets could achieve a lower limit of detection than the preliminarily observed LOD of 200 fM (ModF-ModR-PrymCrRNA1).
For simplicity, negative controls are not included in the graphs on Figures 7-9, as they showed no significant signal and their fluorescence intensity values have been subtracted from the obtained results.
Figure 7. SHERLOCK fluorescence readout result – primer and crRNA mixing screening, performed on a 480 nM sample of Prymnesium parvum DNA.
Figure 8. SHERLOCK fluorescence readout result – primer and crRNA mixing screening, performed on a 200 fM sample of Prymnesium parvum DNA.
As shown in Figures 7-8, the best combination was ModF + GalR + PrymCrRNA1 because it yielded high signals for both 480 nM and 200 fM target DNA concentrations.
To ultimately select the most promising primer pair for optimization, an assay was conducted using the primers that produced the most satisfactory results in the previous test: ModF-GalR; ModF-ModR; AltF-GalR.
Figure 9. SHERLOCK fluorescence readout result – primer and crRNA mixing screening, additional experiment.
Based on the results shown in Figure 9 and the previous assay (Figure 8), the best primer combination was indeed ModF-GalR + PrymCrRNA1. This combination provides the highest fluorescence intensity and effectively detects a 200 fM target DNA concentration, as demonstrated in the previous test.
Having identified our chosen pair, our next step was to quantify the SHERLOCK system to determine if the fluorescence signal corresponds to the amount of target Prymnesium parvum DNA and, consequently, the number of algal cells.
The Limit of Detection (LOD) was defined as the lowest concentration of PCR-amplified Prymnesium parvum DNA that yielded a positive test result. This was indicated by a significant increase in fluorescence intensity over time, with the measured values exceeding those of the corresponding negative control.
Initial experiments on assessing the LOD were carried out using the GalF-GalR + PrymCrRNA2 combo, as they were performed before the adaptation of Mod and Alt primers.
Figure 10. SHERLOCK fluorescence readout result – preliminary LOD determination, repeat 1.
Figure 11. SHERLOCK fluorescence readout result – preliminary LOD determination, repeat 2.
Based on the charts in Figures 10-11, we concluded that the initial template DNA concentration does not show a linear relation with fluorescence intensity. We hoped that optimizing RPA primer concentration may solve the issue. We thought the detection of 200 fM by the assay to be a promising sign for the LOD value. From the first SHERLOCK test (Figure 10), it was concluded that the LOD is between 200 pM and 200 fM. The second SHERLOCK test (Figure 11) was performed to assess the LOD more precisely, and it was unexpectedly able to detect a 200 fM concentration, which was not the case for the previous test.
We conducted experiments to optimize the RPA primer concentration and therefore make SHERLOCK quantifiable; the details are described on the "Measurement" page. Unfortunately, these efforts were not successful, as we were unable to define a range of target DNA concentrations that consistently correlated with the final fluorescence intensity. As a result, we were also unable to establish a universal standard curve for the assay.
After ModF-GalR + PrymCrRNA1 was established as the most effective combination, an additional assay was conducted to determine the final LOD for our test.
This experiment was repeated twice. In Figures 12-15, the data were split into two charts (for higher and lower concentrations) to facilitate easier visual analysis. However, it's important to note that all samples were measured in the same experiment, allowing for direct comparison between them (the two charts simply reflect an approach to data presentation).
Figure 12. SHERLOCK fluorescence readout results – final LOD determination, higher DNA concentrations, repeat 1.
Figure 13. SHERLOCK fluorescence readout results – final LOD determination, lower DNA concentrations, repeat 1.
Figure 14. SHERLOCK fluorescence readout results – final LOD determination, higher DNA concentrations, repeat 2.
Figure 15. SHERLOCK fluorescence readout results – final LOD determination, lower DNA concentrations, repeat 2.
Despite optimizing the RPA primer concentration, it was not possible to define a range of target DNA concentrations that are proportional to the final fluorescence intensity.
The SHERLOCK reaction proved to be unrepeatable. Different fluorescence intensities were observed for samples from the same RPA, and varying LODs were determined for test duplicates: 1 pM for the first test and 200 fM for the second. This inconsistency undermines the ability to compare fluorescence intensities between tests.
Given these observations, the test does not appear suitable for quantitative measurements. If a range of Prymnesium parvum DNA concentrations proportional to fluorescence intensity could be established, then creating a standard curve for each test would provide reference points for quantifying DNA concentration.
After conducting these experiments, we discovered that commercial RPA kits are optimized for rapid amplification, but diluting the kit (e.g., 2x) may improve the proportional relationship between the amplified DNA and the initial target concentration. Therefore, future efforts to quantify the reaction could involve testing with a diluted RPA kit. Additionally, it would be beneficial to measure the DNA yield after RPA using more direct methods, such as the use of EvaGreen DNA dye [3], rather than relying on SHERLOCK, which depends on the activity of two proteins: T7 polymerase and Cas13.
Nevertheless, based on the collected data, we can conclude that the limit of detection (LOD) of our test is 1 pM. This concentration corresponds to 50 billion Prymnesium parvum cells per liter of water (see calculations below). Additionally, we successfully detected Prymnesium parvum genomic DNA in the sample taken from our algal cultures. These findings suggest that SHERLOCK can be used for screening the presence of Prymnesium parvum in water with a detection limit of 1 pM.
Based on bioinformatic analysis we learned that Prymnesium parvum from the Oder River has ITS1-5.8S-ITS2 sequences exclusively on chromosome 20, present in two haplotypes: one with 8 repetitions of the ITS1-5.8S-ITS2 sequence and another with 4 repetitions. Statistically, this means that 25% of the Prymnesium parvum population should have 8 copies of the ITS1-5.8S-ITS2 sequence (haplotypes 4 + 4), 50% should have 12 copies (8 + 4), and 25% should have 16 copies (8 + 8).
The LFA procedure was used to support and validate the conclusions gained from tests that utilized the fluorescence-based readout. However, there were also experiments conducted which were notable or specific to the LFA procedure.
We were successful in directly detecting Prymnesium parvum DNA isolated from algal culture samples as well as water samples taken from bodies of water affected by Prymnesium parvum blooms using the LFA procedure. The following test was repeated twice.
Figure 16. LFA test result on Prymnesium parvum DNA.
The strip numbers in Figure 16 correspond to the following samples (“Water” – sample of Prymnesium genomic DNA isolated directly from a water body affected by Prymnesium parvum blooms; “Culture” – sample of Prymnesium gDNA isolated from our laboratory cultures):
Figure 17. LFA test result on Prymnesium parvum DNA – repeat.
The strip numbers in Figure 17 correspond to the following samples (“Water” – sample of Prymnesium genomic DNA isolated directly from a water body affected by Prymnesium parvum blooms; “Culture” – sample of Prymnesium gDNA isolated from our laboratory cultures):
The tests successfully detected Prymnesium genomic DNA in samples isolated from our algal culture as well as in samples directly obtained from a water body affected by Prymnesium parvum blooms. This confirmed the efficacy and utility of our test for screening water bodies for the presence of Prymnesium parvum.
Once we optimized our reaction components, we aimed to simplify the SHERLOCK procedure as much as possible. In this experiment, we compared the effectiveness of incubating the RPA and SHERLOCK reactions using body heat versus thermocycler incubation. This was done to assess whether the reaction can be effectively carried out by the user holding it in their hand, therefore making the LFA assay portable and possible to use for in the field screening.
Figure 18. LFA test results – body heat vs thermocycler incubation experiment. Samples incubated using the thermocycler are labeled in red, while those incubated by hand (palm of the hand) are denoted in black.
The strip numbers in Figure 18 correspond to the following samples:
The test results for both thermocycler and hand incubation methods do not show significant differences in band intenstity, suggesting that hand incubation is as effective as thermocycler incubation for this protocol. However, only test one from the tested conditions yielded a positive result, indicating that there may be issues with either the RPA reaction reagents or the DNA templates used in the other tests, since “water” and “culture” samples have been positive in previous tests (Figures 16 and 17). This inconsistency in results points to the need for further optimization or troubleshooting of the reaction components to ensure reliable outcomes across all tests.
The Limit of Detection (LOD) was determined as the lowest concentration of PCR-amplified Prymnesium parvum DNA for which the test result remained positive, indicated by the presence of a visible T-line. For this experiment, samples were incubated in a PCR thermocycler.
Figure 19. LFA test results – LOD determination experiment.
The strip numbers in Figure 19 correspond to the following samples:
By comparing strips 1-13 with test 16, a negative control, it can be estimated that the limit of detection (LOD) for our Sherlock assay is approximately 10 pM, as indicated by test 7. However, this conclusion is somewhat subjective, given that band intensity was not quantified.
To obtain a more precise and reliable LOD determination, this test should be repeated. The positive control (stripe 17) yielded a positive result. However, the test performed on older RPA samples previously provided a more definitive positive result (lower C-line was not visible). This suggests there may be inefficiencies in the RPA procedure or potential contamination of its reagents, which could be affecting the consistency of the results.
The tests for both the “Water” and “Culture” samples returned negative results. This is likely due to unidentified issues within the procedure, possibly linked to the RPA step or sample handling.
We successfully developed a Cas13 purification protocol that efficiently yields pure Cas13 while also being cost-effective and time-efficient. We purified both huLwCas13a and CcaCas13b proteins on the first attempt. This page presents the results of the purification process and the functionality of Cas13a as this is the protein we used in all our SHERLOCK assays (detailed results for CcaCas13b can be found in our Notebook: ProteinLab section).
After each step of the protein purification process, results were visualized using SDS-PAGE (Figures 20-22).
Figure 20. SDS-PAGE gel for huLwCas13a – samples after IMAC*.
Wells key: unind. – uninduced bacterial culture, ind. – culture after 18h IPTG
induction, sonic. – supernatant after sonication and centrifugation, prec. – sediment after sonication and
centrifugation, FT – flow-through from sample application on His-resin column, wash – flow-through from column
washing before elution; E1, E2, E3… – fractions eluted with 150 mM imidazole buffer, *E1, *E2, *E3 – fractions
eluted with 300 mM imidazole buffer.
*Note: Some electropherograms show bands shifted downward from their expected positions, likely due to well overloading. To confirm the proteins' sequences, mass spectrometry would be ideal, but budget constraints prevented us from performing this analysis.
The expression step was effective as after induction the band of a mass corresponding to the produced protein was observed (for Cas13a – 155 kDa).
IMAC was effective in partially purifying Cas13a – percentage content of contaminants in the solution decreased (compared to wells ind. and sonic), although they are still present in significant amounts. Moreover, fractions were enriched with the produced protein (thicker bands corresponding to the produced protein – marked by violet arrows).
The SUMO digestion step was conducted, as planned, on half of the collected IMAC fractions. SUMO digested and undigested pooled fractions were used in the second purification protocol step – IEC.
Figure 21. SDS-PAGE gel for huLwCas13a – samples after IEC.
Cas13a with the tag is indicated by the violet arrow, Cas13a without the tag – blue arrow. The purple rectangle
marks two bands that correspond to the cleaved 6xHis-TwinStrep-SUMO tag.
Wells key: entry – pooled fractions after IMAC, sample appl. – flow-through from
sample application on the column, wash – flow-through from column washing, E1, E2, E3… – eluted fractions. Samples
with * correspond to samples after SUMO protease digestion.
For Cas13a, for both SUMO protease digested and undigested samples, a reduction in protein contaminants, approximately 2-fold, can be observed, although they still remain in significant amounts. Concentration of Cas13a in the collected fractions is visible.
After IEC and concentration of pooled fractions to less than 0.5 ml the final purification step – SEC – could be conducted.
Figure 22. SDS-PAGE gel for huLwCas13a – samples after SEC.
The protein with the tag is marked by the violet arrow, while the protein without the tag is marked by the blue
arrow.
Wells are labeled as follows: E1, E2, E3, etc., representing eluted fractions.
Samples marked with * are those that underwent SUMO protease digestion.
After the 3-step purification process, we obtained pure huLwCas13a, as confirmed by SDS-PAGE (Figure 22). As expected, a single band was observed for the undigested protein. In SUMO-digested fractions, two bands were observed, indicating incomplete tag cleavage.
As planned, we conducted a comparative analysis of tagged and untagged proteins to test their functionality.
Figure 23. Dependence of mean residue molar ellipticity for huLwCas13a with and without the tag in the far-UV range.
Table 1. Proportions of various secondary structure types determined from Circular Dichroism spectra for huLwCas13a.
|
huLwCas13a |
Mean proportions of secondary structure types determined using algorithms from the CDPro package ± standard deviation |
|||
|
α-helix |
β-sheet |
β-turn |
unordered |
|
|
undigested |
64.5 ± 2.1 |
3.8 ± 0.9 |
10.2 ± 0.7 |
20.9 ± 2.5 |
|
digested |
58.1 ± 10.2 |
4.7 ± 6.3 |
18.2 ± 9.9 |
19.5 ± 7.2 |
The spectra in Figure 23 overlap significantly, and the contribution of individual secondary structures to the protein structure (Table 1) does not differ substantially. This indicates (contrary to our expectations) that the tag does not have a significant impact on the structure of Cas13a.
Figure 24. Dependence of the first derivative of the fluorescence
intensity change from temperature for huLwCas13a.
Key: D – sample digested using SUMO protease, UD – sample undigested
Contrary to our expectations, NanoDSF results indicated no statistically significant difference between tagged and untagged proteins in thermal stability.
Figure 25. A comparison of the activity of huLwCas13a with and without
6xHis-TwinStrep-SUMO tag.
Figure key: 1 – SUMO protease digested huLwCas13a added to SHERLOCK mix; 2 – SUMO
protease undigested huLwCas13a added to SHERLOCK mix; 3 – water added instead of Cas13a (negative control); 4 –
RNase A added instead of Cas13a (positive control); 5 –target RNA and crRNA replaced with water in SHERLOCK mix,
for SUMO digested huLwCas13a; 6 – target RNA and crRNA replaced with water in SHERLOCK mix, for SUMO undigested
huLwCas13a.
Contrary to our expectations the Cas13a protein, both digested and undigested by SUMO protease, showed similar levels of activity.
Our optimized protocol successfully produced pure, active Cas13 for future experiments. We obtained 3.6 mg of purified protein, sufficient for nearly 29,000 SHERLOCK reactions.
During the purification process, we observed that a single-step purification using IMAC yielded only around 15% Cas13 in the collected fractions, making it insufficient for our needs. We recommend a three-step purification process—IMAC, IEC, and SEC—to achieve pure Cas13 detailed protocol can be found here. This improvement builds upon existing protocols and offers a valuable contribution for future iGEM teams.
Notably, we concluded that the SUMO protease cleavage step is unnecessary. This conclusion was supported by several tests:
Omitting the SUMO digestion step makes the purification protocol more cost-effective and time-efficient.
We produced and purified Cas13b as a backup in case Cas13a faced challenges such as inefficient purification or loss of activity. (The data presented here pertains to Cas13a, with a full description of the purification process for both proteins available in the Protein Lab documentation in "Notebooks"). This strategy ensured the project could continue without delays. However, due to budget constraints, we did not order synthetic DNA for Cas13b testing, as Cas13a's activity had already been confirmed.
Looking ahead, future work could involve testing our purified CcaCas13b and using it in SHERLOCK assays. This could allow for multiplex detection of the ITS2 region of Prymnesium parvum alongside other genomic fragments.
In conclusion, our Cas13a protein met our purity standards and was used for all SHERLOCK tests throughout the project.
Prymnesium parvum is a non-model organism, and there are no established protocols for culturing this species. Through a comprehensive system of trial and error, combined with guidance from our supervisors and external collaborators, we developed a method for maintaining Prymnesium parvum under laboratory conditions. Additionally, we tested various techniques for isolating DNA from Prymnesium parvum samples.
We successfully set up cultures from various water samples collected from bodies of water affected by Prymnesium parvum blooms.
Figure 26. Microscoping image of water sample containing Prymnesium parvum cells.
Figure 27. Fluorescent microscopic image of water sample containing Prymnesium parvum cells.
Characteristic morphological features of Prymnesium parvum, such as the haptonema and the two chloroplasts, are visible in Figure 26.
Despite observing these features, differentiating Prymnesium parvum cells from other phytoplankton in the water samples proved challenging. Therefore, we confirmed its presence via PCR using primers designed by Galluzzi et al. [1], which target the ITS2 sequence specific to Prymnesium parvum. This PCR product is 132 bp in length.
We isolated DNA and performed PCR, both following the protocol.
Figure 28. Gel electrophoresis of the PCR product – result.
Legend to Figure 28:
“Gz” and “Kz” are names of DNA samples derived from different Prymnesium parvum cultures. Samples Gz and Kz contain the ITS2 sequence of 132 bp, as indicated by the clear bands visible between the 200 bp and 100 bp markers. No specific product was observed in row 2. Based on these results, we can conclude that the algal DNA isolated contains Prymnesium parvum genomic DNA, confirming the presence of Prymnesium parvum cells in the respective water samples.
Based on our optimization of culture conditions, we developed a method for maintaining Prymnesium parvum under laboratory conditions. A detailed description of the recommended culture conditions can be found on the “Contribution” page, as we hope it will be useful for future iGEM teams working with Prymnesium parvum.
Throughout our project, we used three different DNA isolation methods:
Isolation using the NEB kit yielded a slightly higher DNA concentration compared to the QIAGEN kit. Additionally, the NEB isolation kit proved to be more convenient, as it did not require a lengthy initial incubation step of 5-7 hours like the QIAGEN kit. Tables 2 and 3 display the DNA yields obtained for selected samples with the NEB and QIAGEN kits, respectively.
Table 2. Concentration of the DNA isolated with the NEB kit.
|
Sample name |
DNA concentration [ng/µl] |
A260/A280 |
A260/A230 |
|
NOW 5 |
16 |
1,68 |
1,34 |
|
BOW 2 |
24,1 |
1,79 |
1,56 |
Table 3. Concentration of the DNA isolated with the QIAGEN kit.
|
Sample name |
Concentration [ng/µl] |
A260/A280 |
|
NOW 5 |
5,8 |
1,66 |
|
NOW 5.1 |
4,9 |
1,65 |
|
P |
12,7 |
1,67 |
|
BOW 1 |
5,6 |
1,53 |
|
BOW 2 |
6,4 |
1,55 |
|
2W |
11,2 |
1,28 |
|
Gdańsk 2 |
5,2 |
1,8 |
We also tested a nucleic acid purification method designed for use outside the laboratory setting [4]. Our objective was to adapt this method for field-based nucleic acid isolation within the PrymChip. The technique involves preparing cellulose dipsticks coated with wax for nucleic acid purification. The dipsticks we created are presented in Figure 29.
Figure 29. The dipsticks made out of Wattman paper.
Figure 30. Equipment for the in-field isolation with instructions. In Eppendorf tubes lysis buffer (number 1) and wash buffer (number 2) are presented.
Despite our efforts, we were unable to successfully isolate DNA using this method within the PrymChip system. However, we highlight that this cost-effective field-based method for nucleic acid isolation exists and could be further optimized by future iGEM teams for environmental testing purposes.
During the design of our crRNA spacers for detecting Prymnesium parvum, we recognized that the process was quite inefficient. While we were only changing a small portion of the crRNA sequence with each design, we still had to order the full sequence each time. To simplify this process, we aimed to create a system where users would only need to focus on designing their specific spacer, while the T7 and direct repeat (DR) fragments—common to all crRNAs used with Cas13a—would remain unchanged. This led to the creation of SynLOCK, a user-friendly crRNA synthesis system for SHERLOCK. SynLOCK – the crRNA Synthesis System for SHERLOCK – is envisioned as a modular set of components that allows users to easily insert a custom crRNA spacer of their choice.
To do this, we initially designed the SynLOCK Cassette (BBa_K5087017) as shown in the schematic in Figure 31.
Figure 31. Initial SynLOCK Cassette Design.
Initially, our team used the PSB1C5C plasmid backbone for assembly due to its size, high copy number, and availability in the 2024 iGEM Distribution kit, which was accessible to us and other teams. We ordered the Cassette DNA, assembled it by hybridizing two DNA strands to create sticky ends, and successfully cloned it into the pSB1C5C plasmid backbone using cloning with BsaI.
Figure 32. The schematic representation of the initially assembled device.
We verified the sequence of our design using Sanger sequencing to ensure that the clones we selected contained the Cassette. The results confirmed that the Cassette was present and capable of accepting the crRNA spacers.
However, we realized that the sequencing is required to verify if the Cassette is integrated into the plasmid backbone, which is not time-efficient. Including a reporter within the Cassette could solve that problem. Moreover, the plasmid backbone we used is not ideal. It is designed for storing CDS parts and is not a LVL1 plasmid suitable for Type IIS assembly. Additionally, it lacks the biobrick prefix and suffix, making it impossible to transfer the Cassette between vectors after cloning.
Therefore, we refined our design and included a reporter unit. The reporter was strategically placed between the SapI sites to ensure it would be excised when a custom crRNA spacer was incorporated, allowing for easy visual screening of colonies.
Figure 33. Improved SynLOCK Cassette Design.
Additionally, we chose the pSB1C3 plasmid as the new backbone. This plasmid is a standard for storing BioBrick parts, meeting our criteria for being high-copy and accessible, and it supports cassette transfer using BioBrick prefix and suffix cutting sites.
We introduced the reporter unit into the Cassette by amplifying it via PCR to obtain compatible sticky ends. We then amplified the complete Cassette with the reporter and transferred it into the pSB1C3 plasmid backbone.
Figure 34. The schematic representation of the final assembled device.
We then assembled the system by cloning in the custom crRNA spacers used in our PrymDetect test. We employed the Type IIS assembly standard to insert the crRNA spacers, ensuring that there were no scar sites in our final IVT template construct that could impact the length or sequence of the crRNA.
Figure 35. Schematic depiction of the improved system in action. On the right, a fragment of a Petri dish showing visible white and purple bacteria, indicating which colonies are ready for picking.
We conducted IVT on the linearized plasmids and performed RNA electrophoresis to visualize the results.
Figure 36. RNA electrophoresis – in vitro transcription (IVT) results using SVR and USV plasmids containing the Cassette with crRNA spacers. "SVR" indicates that the crRNA from that well was stored in the pSB1C5C backbone, while "USV" signifies that it was stored in the pSB1C3 backbone.
We also tested the Cassette’s transferability between different backbones by excising it from the pSB1C3 backbone using the enzymes present in the BioBrick prefix and suffix. This step added to the system’s standardization.
Figure 37. Electrophoresis results showing the digestion of USV plasmid with the enzymes marked in the image.
Table 5. Expected band size the digestion of the USV plasmid.
|
Enzyme/s used |
Expected Band Size (bp) |
|
EcoRI & SpeI |
2047 & 1031 |
The correct bands were obtained, demonstrating that the Cassette can now be transferred between other systems compatible with the BioBrick assembly standard. This allows our design to be used with a variety of plasmid backbones, giving users the flexibility to choose the one that best suits their specific needs.
In our results, we successfully obtained crRNA transcripts of the correct length, as indicated by the single band on the gel for most wells at around 64 bp. Some contamination is visible in certain wells, likely due to the plasmid IVT template remaining in the sample. Additionally, some bands appeared faint, suggesting that the vectors may not have been fully digested with BbsI or that the DNase I treatment was only partially effective. For future experiments, it would be ideal to include an RNA ladder with bands below 200 bp in the electrophoresis samples; unfortunately, the RiboRuler High Range RNA Ladder (Thermo Fisher Scientific) was the only option available to us.
The correct single band at an expected transcript length confirms that the system is functioning as expected, establishing a proof-of-concept for SynLOCK. Overall, while further technical optimization is necessary to obtain purer molecules, our results support the system's functionality and potential for future development.
We aimed to address the need for accessible, field-based environmental sample testing by creating a device that allows individuals, regardless of research experience, to test their water samples for Prymnesium parvum near water sources. Leveraging the versatility and cost-effectiveness of 3D printing, we focused on developing a low-cost prototype as part of our iGEM project. Our efforts culminated in the development of a device called the PrymChip.
The PrymChip is a 3D-printed device specifically designed to excite, detect, and quantify green fluorescence signals, with data processed using a Python script. The primary objective was to create an affordable, field-deployable tool capable of detecting Prymnesium parvum in environmental water samples using the SHERLOCK detection system.
We successfully designed all the 3D-printed parts in TinkerCAD and printed using filament fused deposition modeling (FDM) technology with a Prusa i3 MK3 Printer. We went through multiple iterations of the DBTL (Design-Build-Test-Learn) cycle before landing on a final design (Figure 38). The design was carefully planned to incorporate all the components necessary for detecting Prymnesium parvum. Consequently, the PrymChip consists of three main parts (Figure 38):
Figure 38. Schematic of the PrymChip.
A comprehensive description of the PrymChip has been included in the ‘Hardware’ page.
Due to time constraints and limited reagent availability, we were unable to demonstrate algae detection using SHERLOCK within the PrymChip. Instead, we successfully performed proof-of-concept experiments, demonstrating the device’s functionality with fluorescein solutions. Testing the software on fluorescein solutions was the most convenient and reliable option as the excitement (~490 nm) and emission maximums (~520 nm) suggest that the fluorescein or its derivative is used in formation of the RNase Alert utilized normally in SHERLOCK tests conducted by our team.
Fluorescein from the Fluorescence Measurement Kit was used to prepare different concentrations of the fluorescein. A series of photos were then taken, corresponding to fluorescein solutions prepared in PBS, with concentrations ranging from 0.78 µM to 100 µM, including a blank sample containing only PBS solution (Figures 39-40).
Figure 39. Series of fluorescein solutions and the PBS solution. Photos taken by a Google Pixel smartphone.
Figure 40. Series of fluorescein solutions and the PBS solution. Photos taken by a Samsung smartphone.
As illustrated in Figures 39-40, the fluorescent signal from fluorescein can be detected using both smartphone devices, although minor differences in fluorescence intensity are observed between the models.
The calibration curve images served as training data for a linear regression model in the Python software. After training, the model was tested to evaluate its ability to predict the concentration of green fluorescence based on fluorescence intensity values obtained from images with unknown fluorescein concentration.
Table 6. Concentrations of the tested fluorescein solutions and the results obtained from the calibration curves using different phones.
|
Real fluorescein concentration (µM) |
Concentration detected by Samsung phone (µM) |
Concentration detected by Google Pixel phone (µM) |
|
8 |
7.22 |
0 |
|
2 |
2.84 |
0 |
|
50 |
53.60 |
42.88 |
The results indicated that the software accurately analyzed the photos, yielding fluorescein concentration values that closely matched the actual concentrations. This demonstrates that our PrymChip software is effective and reliable. However, the sensitivity of measurement highly depends on the smartphone’s camera. It would be advisable to test the software amongst a wider range of devices.
Encouraged by the promising results from the fluorescein tests, we decided to test the RNaseAlert™ substrate (from the RNaseAlert™ Lab Test Kit v2, Invitrogen), which is the RNA probe used in our fluorescence-based SHERLOCK assays. We introduced RNase A to cleave the probe and emit a fluorescent signal.
Unfortunately, this experiment did not yield successful results, as we did not detect a fluorescent signal (Figure 41). Several factors may have contributed to this outcome, including the possibility that the LEDs used in the PrymChip design may not effectively excite the fluorophore in RNaseAlert. This experiment needs to be repeated to further investigate the cause of this issue; however, we were unable to do so due to limited availability of RNaseAlert™.
Figure 41. Photos from the PrymChip – RNase alert testing.
The current PrymChip serves as a proof-of-concept but requires several improvements to detect Prymnesium parvum in environmental samples. One key enhancement is incorporating nucleic acid isolation via cellulose dipsticks, which could be seamlessly integrated into the design. Looking ahead, key future experiments will include further testing with RNaseAlert, as well as final experiments utilizing all optimized SHERLOCK reaction components to detect Prymnesium parvum DNA.
In terms of device design, several modifications can be made. Adding a drawer to the black detection box would simplify solution insertion and reduce spillage risk. Redesigning the lid to house an external camera, connecting to a phone via USB, would standardize image capture and eliminate the need for calibration curves from individual phone cameras. However, the current lid option should remain for those preferring phone-based photography. A mobile app for image analysis is essential to avoid the inefficiency of manual photo transfer. Finally, the connection between the boxes could be improved by using a magnetic mechanism, which would be easier to incorporate into the thick PrymChip walls and would enhance overall usability.
In summary, at the current development stage, the PrymChip excites green fluorescence signals, captures images via a smartphone, and analyzes the data using a Python script to calculate fluorescein concentration from the captured images. We have estimated the final cost of the functional device at €20.38. We envision this device as a promising, reusable, and cost-effective tool that enables individuals, regardless of their research experience, to test water samples for Prymnesium parvum near water sources. Additionally, it represents a valuable contribution to future iGEM teams, providing an excellent foundation for innovation. We anticipate seeing numerous exciting applications and modifications in the future.
Our modeling approach is focused on qualitative intuition enabling for grasping the key concepts of the enzyme dynamics in the PrymDetect system. The model and calculations were divided into 4 sections:
We successfully developed a model based on the law of mass action that allowed us to relate the chemical reaction rate to concentration for all reactants in the SHERLOCK reaction, aiding the understanding of the overall dynamics of the system. This system of 6 ordinary differential equations was subsequently simplified to a single Michaelis-Menten equation using a series of assumptions based on literature findings such as the quasi-steady-state assumption. The time scale which can be interpreted as an estimate of time to detection was calculated for literature values for the same CRISPR enzyme and fluorescent reporter as used in the PrymDetect system giving an estimate of 18 minutes. However, this estimate is likely to be longer due to an anticipated higher value of the Michaelis constant in our assay compared to the assay in the publication. Finally, the qualitative progression of the RPA and IVT reactions was modeled to illustrate the dynamics of these reactions specifically in the PrymDetect system.
17 out of the 18 primer-crRNA combinations we engineered are viable and functional, which has exceeded our expectations. Contrary to what we anticipated, both crRNA designs proved to be effective, as confirmed by both fluorescence and LFA readouts. Additionally, after encountering issues with false positive results from one of the combinations, we feared this might be a widespread issue across many others; fortunately, that was not the case. Ultimately, we selected the most effective combination – ModF-GalR + PrymCrRNA1 – to focus our efforts on. However, we believe that the wide array of combinations offers numerous opportunities for future optimization and fine-tuning of the system.
Our protein purification was successful on the first attempt, allowing us to move forward with the project. Additionally, we made adaptations to the protocol that we hope will help future iGEM teams purify this protein more effectively and at a lower cost.
Working with an understudied, non-model organism, relying on environmental samples, proved to be extremely challenging—something we did not fully anticipate when brainstorming our project idea. Nevertheless, we successfully maintained our cultures, isolated and sequenced the Prymnesium parvum DNA as well as optimized culture conditions and DNA isolation methods.
We successfully engineered a system that will simplify the work of future iGEM teams and other researchers using SHERLOCK. We established a proof-of-concept for this user-friendly and reliable system for crRNA synthesis.
We developed an inexpensive, 3D-printed device that has potential to be used with the SHERLOCK method to test for the presence of Prymnesium parvum in water samples. We established a proof-of-concept for the device and successfully engineered Python software that allows a smartphone camera to detect fluorescence signals using the device.
We developed a mathematical model that accurately represents SHERLOCK reaction kinetics and conducted back-of-the-envelope checks on published literature data to enhance our understanding of the SHERLOCK assay and inform its practical application and refinement.
As an iGEM team and as young researchers, we place great importance on honesty and scientific integrity. In evaluating our results and the project as a whole, we have identified several areas for improvement, along with elements that could be considered failures.
Firstly, all of our fluorescence-based SHERLOCK data present fluorescence intensity on the y-axis. This poses a problem because absolute fluorescence values can vary between plates and instruments, making it impossible to directly compare results across different setups. Unfortunately, we did not perform a calibration using a fluorescence standard curve, which we acknowledge as a mistake on our part. By the time we recognized this issue, near the end of our project, we had already used up all the fluorescein from the Fluorescence Measurement Calibrants Kit to calibrate our PrymChip. Although we ensured that comparisons between experimental samples were made within the same read from the start, we recognize that the lack of calibration may complicate the reproducibility of our results for future teams.
From the outset, we aimed to make our method field-applicable. This goal was a constant consideration as we worked to optimize the test for maximum efficiency. We explored practical solutions, including incubation using body heat, and simplified the overall procedure as much as possible. However, one key challenge remained: the DNA isolation step. While we identified a protocol that could be used for in-field DNA isolation and began testing it, we were unable to fully optimize it due to time constraints, and the method did not function as intended. Nevertheless, we believe there is potential for further optimization, and our proposed approach can still be adapted for use with samples containing Prymnesium parvum.
Another challenge affecting the full in-field application of our test is the relatively high Limit of Detection (LOD) that we were able to establish. In our fluorescence-based test, the LOD for detecting Prymnesium parvum DNA was determined to be 1 pM, which corresponds to approximately 50 billion Prymnesium parvum cells per liter of water. This is significant, as fish kills typically occur at cell counts exceeding 50-100 million cells per liter [5]. To address this issue, we propose centrifuging the water sample before applying the SHERLOCK detection procedure. However, this complicates field application since a centrifuge would require access to electricity. Another possible solution is using a syringe with a filter. On a positive note, Cas13-SHERLOCK fluorescence detection has been shown to demonstrate even attomolar sensitivity[6]. Therefore, we believe that with further optimization of the procedure and more thorough LOD determination, it would be possible to eliminate the need for condensing the samples and enable direct application of the method.
In the original paper by Kellner et al., the authors describe a one-pot SHERLOCK procedure that combines the RPA and Cas13 detection reactions into a single mixture [6]. We aimed to test this method, as its success would further simplify the LFA detection procedure. However, the authors also note that one-pot SHERLOCK is less sensitive and more challenging to optimize [6]. Our findings confirmed this – after testing the procedure twice, both with the original reagent proportions and a modified version, we received completely negative results for all samples, including the positive controls.
Another area of improvement concerns testing our PrymChip hardware with actual SHERLOCK reagents andPrymnesium parvum DNA. Unfortunately, we were unable to carry out these tests due to time constraints and limited availability of reagents, which meant we had to conclude our work at the proof-of-concept stage. The test we did perform, using the RNaseAlert™ probe typically used in SHERLOCK (which was digested by an RNase), did not yield the expected results, as no fluorescence was detected. Despite this, we believe there is still room for optimization, and we encourage future iGEM teams to pursue this line of experimentation.
Lastly, we were disappointed that our efforts to optimize RPA primer concentrations for quantifiable SHERLOCK fluorescence detection did not yield the desired results. From the outset, we recognized the potential quantifiability of the SHERLOCK method as a significant advantage in applying it to detect Prymnesium parvum. Despite our attempts to optimize the RPA primer concentration, we found it challenging to establish a range of target DNA concentrations that would correlate proportionally with the final fluorescence intensity or any other measurement point. On the “Measurement” page we provided our insights what changes in conducting experiments can be done to make SHERLOCK quantifiable.
Firstly, throughout our experiments, we observed that our SHERLOCK method exhibited poor repeatability, particularly in the fluorescence-based readout. Although we cannot directly compare intensity values between different reads for the reasons mentioned earlier, we noted that the trends exhibited by the samples sometimes varied across reads. Additionally, the limit of detection (LOD) differed between test duplicates, with one test showing an LOD of 1 pM and another yielding 200 fM. Consequently, establishing a universal standard curve for the assay was not feasible.
Our results have underscored the importance of including relevant controls, particularly negative controls, when validating SHERLOCK reaction components. We recommend using a sample in which water is added to the RPA reaction instead of DNA, with all other components—especially primers—retained, followed by SHERLOCK detection. We believe this control is preferable to simply adding water to an 'empty' RPA reaction mix for the SHERLOCK reaction. While the latter control can also be useful, we emphasize that the former is more critical due to the risk of primers inadvertently serving as templates for the Cas13:crRNA complex, which can lead to false positive results, as we observed in our project. As a positive control, we recommend including the synthetic oligonucleotides described in the Kellner et al. paper [6]. In our experience, these oligonucleotides have proven to be a reliable positive control for both fluorescence-based and LFA detection methods.
Secondly, we advise any future iGEM teams intending to use SHERLOCK in their projects to be mindful of RNase contamination and to seek supervisors and lab space with experience handling RNA. As a team of students, most of whom lacked prior experience with RNA samples, we were often concerned that some of our seemingly ‘strange’ results might be attributed to RNase contamination in our samples or reagents. Therefore, we recommend regularly checking for RNA contamination using kits like the RNaseAlert™ Lab Test Kit (Invitrogen), which we also utilized during our project. Additionally, we strongly suggest implementing preventative measures, such as using RNase-free reagents and materials, and thoroughly wiping down surfaces, pipettes, and other equipment with RNase decontamination solutions like RNaseZap™ (Thermo Fisher Scientific).
We believe our project establishes a solid foundation for future efforts to rapidly and reliably detect Prymnesium parvum at scale. However, we recognize that several aspects need further investigation before the test can be implemented on a larger scale.
To ensure project advancement beyond iGEM, we have spoken with Prof. Dariusz Dziga from the Laboratory of Metabolomics at Jagiellonian University. He has shown enthusiasm for potentially continuing the project.
Key areas for future research and improvement could include:
Overall, we consider the PrymDetect project a significant success. What began as an initial idea to apply SHERLOCK for detecting Prymnesium parvum expanded to include many unexpected aspects, such as the development of the PrymChip or the mathematical modeling of SHERLOCK reaction kinetics. It’s important to highlight that all idea development and planning were carried out by our student team members.
29 000 SHERLOCK reactions possible using Cas13 we purified
650 individual SHERLOCK reactions performed
160 hours of 3D-printing
112 LFA tests performed
30 parts added to the Registry
25 passages of algae cultures
All of this was carried out during… 304 days
By 10 undergraduate and graduate students fully planning and executing the project!
[1]L. Galluzzi et al., “Detection and quantification of Prymnesium parvum (Haptophyceae) by real‐time PCR,” Letters in Applied Microbiology, vol. 46, no. 2, pp. 261–266, Feb. 2008, doi: 10.1111/j.1472-765X.2007.02294.x.
[2]N. Luo, H. Huang, and H. Jiang, “Establishment of methods for rapid detection of Prymnesium parvum by recombinase polymerase amplification combined with a lateral flow dipstick,” Front. Mar. Sci., vol. 9, p. 1032847, Oct. 2022, doi: 10.3389/fmars.2022.1032847.
[3]P. Valloly and R. Roy, “Nucleic Acid Quantification with Amplicon Yield in Recombinase Polymerase Amplification,” Anal. Chem., vol. 94, no. 40, pp. 13897–13905, Oct. 2022, doi: 10.1021/acs.analchem.2c02810.
[4]Y. Zou et al., “Nucleic acid purification from plants, animals and microbes in under 30 seconds,” PLoS Biol, vol. 15, no. 11, p. e2003916, Nov. 2017, doi: 10.1371/journal.pbio.2003916.
[5]“Aquatic Invasive Species (AIS) Control Plan: Golden Alga”, [Online]. Available: https://www.fishandboat.com/Conservation/Plans/Documents/ais-control-plan-golden-alga.pdf
[6]M. J. Kellner, J. G. Koob, J. S. Gootenberg, O. O. Abudayyeh, and F. Zhang, “SHERLOCK: nucleic acid detection with CRISPR nucleases,” Nat Protoc, vol. 14, no. 10, pp. 2986–3012, Oct. 2019, doi: 10.1038/s41596-019-0210-2.
