Results
Successfully constructed and validated toehold switches through a Golden Gate cloning approach.
Optimized mammalian cell culture protocols enabled reliable testing of our constructs in both AFP-positive and AFP-negative cell lines.
Our first flow cytometry and qPCR experiments provided promising indications that selected toehold switches may function as intended, laying the groundwork for further validation.
Chapter I: Cloning
Toehold Switch Assembly
Introduction
Toehold switches are advanced tools in synthetic biology that enable precise regulation of gene expression at the translational level. They are artificially engineered single-stranded RNAs that function as sensors within the 5′ untranslated region (5′ UTR) of a transcript. In their inactive state, a toehold switch adopts a complex secondary structure that occludes the ribosome binding site and the start codon, thereby preventing the initiation of translation. Only in the presence of a specific RNA sequence (the so-called trigger) does the toehold switch undergo a conformational change, allowing the ribosome to initiate protein synthesis.
Aim of this subproject
The objective of this experiment is to obtain 21 constructs, each consisting of Transcription Unit 1 (TU1) (BBa_25SH329F) and a genetic element comprising one of the twenty-one toehold switches placed upstream of the eGFP coding sequence. The prepared constructs will provide an effective platform for testing all of the designed toehold switches.
Key Steps and results
In the first stage, 21 genetic parts were ordered from Twist Biosciences, each consisting of a single toehold switch placed upstream of an eGFP coding sequence (Toehold switches collection, Parts). Each part was additionally flanked by sequences recognized by the restriction enzyme BsaI, enabling the generation of sticky ends suitable for cloning. Twist Biosciences also provided us TU1 (BBa_25SH329F) which contains sequences flanked by BsaI recognition sites, designed to allow the insertion of parts carrying toehold switches. All ordered elements were transformed into E. coli and subsequently plated on a selective medium containing the appropriate antibiotic. The resulting colonies were used for plasmid isolation, thus providing purified genetic parts.
Following unsuccessful cloning of TU2 (BBa_25VE60AA), we decided to simplify the process of constructing test systems for toehold switches. Using the restriction enzyme BsaI and T4 ligase, we inserted the gene encoding the α-peptide of β-galactosidase (LacZ, BBa_25DTJWTR) with elements enabling expression in a bacterial system into the TU1 region designated for toehold switches with eGFP (Fig. 1). To enable this, a PCR was performed on the L2TU2 construct (BBa_259T9MVI), which contains the LacZ cassette. We used modified primers containing additional nucleotides at their 5' ends, providing flanking sequences recognized by the BsaI restriction enzyme, thereby enabling the insertion of the element into TU1 (Fig. 2). This modification enabled us to perform blue-white screening.
Cloning of constructs containing toehold switches into TU1 was performed using the Golden Gate assembly system with the restriction enzyme BsaI (Fig. 3). The reaction mixture was transformed into E. coli and plated on a selective medium with the appropriate antibiotic. The white colonies were used for colony PCR to confirm construct attendace (Fig. 4). Subsequently confirmed colonies were used for plasmid isolation, thus providing purified genetic parts.
Discussion
We have obtained 20 genetic constructs that enable us to test the specificity of each designed toehold switch in a eukaryotic system based on eGFP expression.
Distribution kit
Introduction
The iGEM collection represents a remarkably rich repository of genetic parts, enabling research across a wide range of biological systems. When designing projects based on mammalian cellular systems, an especially valuable source of genetic components is the Asimov Mammalian Parts Collection, which provides elements such as promoters, 5' UTRs, 3' UTRs, poly(A) sequences, and coding sequences for fluorescent proteins.
Aim of this subproject
To obtain plasmid stocks of Asimov Mammalian Part Collection that we wanted to use for the cloning in our project.
Key Steps and results
To obtain plasmid stocks of Asimov Mammalian Part Collection we performed transformation of E. coli. Despite positive and negative controls being ok we barely saw any colonies on the plates. We minipreped them and sent them to Sanger sequencing to confirm the sequence fidelity. We obtained two clones of P3 (supposed to be Inert/synthetic1 3UTR, BBa_J433018) and one clone of H3 (supposed to be CMV_v1, BBa_J433000). Surprisingly, clone H3 and one of P3 clones turned out to be Inert/synthetic1 3UTR (Photo 1), but the second clone from P3 turned out to be bGH (BBa_J433021) (Photo 2). It indicates that it is a problem with a plate, because cross-contamination was impossible, since we never opened the well that was supposed to have bGH plasmid (well L5) (Photo 3)
Discussion
Our observations indicate an incorrect distribution of genetic components on the supplied plate. The lack of colonies may be in that case the result of using the wrong antibiotic or possible damage to the plasmids.
Assembly of dual-reporter construct
(TU1‑eGFP + TU2‑RFP)
Introduction
Red fluorescent proteins (RFPs) are a class of coral-derived proteins that emit red light when illuminated with light of a suitable wavelength. RFPs are highly valuable in a range of biological applications, from fluorescence microscopy and protein localization studies to in vivo imaging and molecular biotechnology. The compatibility of RFPs with other fluorescent proteins enables multi-color imaging and facilitates a deeper understanding of complex cellular processes.
Aim of this subproject
To obtain the construct with both the transcription unit 1 having insert site for introducing toehold switches with eGFP and transcription unit 2 constitutively expressing RFP (BBa_25V2E0HW). We wanted to use it first as a control if the lack of GFP signal from AFP negative cells is from the lack of transfection or the lack of translation. Secondly, it would also allow us to normalize the expression of eGFP in the population of cells, since every cell would be transfected with a slightly different amount of pDNA.
Key Steps and results
We ordered from IDT the TU1 on a cloning vector pTwist Amp High Copy, ready for Golden Gate Assembly. We also wanted to order an already assembled TU2, but the sequence was too complex for the company to synthesize. We decided to split it into 3 parts with 20 nt overhangs and use HiFi cloning to assemble them and then use this product for Golden Gate Assembly. To do so we first tested a non-classical approach without placing an insert in a vector but it failed, probably due to the low amount of correctly assembled product. We used primers specific to ends to amplify it from the reaction mixture and we obtained a band with corresponding length that we extracted and sequenced (Fig. 5). We used it in the Golden Gate Assembly of TU1, TU2 and backbone, but it continuously failed, even after modifying protocol. We also tried preparing a HiFi reaction with backbone, but we obtained no colonies. At this point we decided to omit this cloning step due to time constraints and perform flow cytometry experiment without this additional control.
Discussion
The construct containing the RFP coding sequence could not be obtained due to significant issues with the cloning reaction. This might have resulted from the linear nature of our assembly, since it cannot self replicate as stably as a plasmid. In the future we plan to first fully sequence the transcript and order new primers to extend TU2 more, since currently the SapI recognition site is close to the end of dsDNA, which may decrease the restriction digestion efficiency. We also ordered a fully assembled TU2 from Ansa Biotechnologies cloned into a vector, so it may increase efficiency of our next attempts.
Chapter II: Transfection
Background information
Transfection is the process of introducing nucleic acids such as DNA or RNA into eukaryotic cells to study gene function or protein expression. One of the most commonly used chemical approaches is lipofection, which relies on positively charged lipids forming complexes with negatively charged nucleic acids. These lipid-nucleic acid complexes, called lipoplexes, facilitate cellular uptake through endocytosis or membrane fusion. A widely used reagent for this purpose is Lipofectamine, a lipid-based formulation that ensures efficient delivery while maintaining high cell viability. This method enables both transient and stable gene expression across a wide range of mammalian cell types, making it ideal for testing the functionality of our toehold constructs.
Introduction
Our project leverages toehold switch reporter constructs encoded on plasmids to selectively detect and respond to the presence of alpha-fetoprotein (AFP) inside mammalian cells. We engineered 20 distinct plasmid variants, each bearing a different toehold switch. To assess the specificity and efficacy of our toehold switch designs we will need to perform transfection and cytometry to quantify eGFP expression. Therefore we need to optimize our transfection method for best usage of our resources. Due to this, we performed two transfection methods: forward and reverse in four conditions: (-)control, (+)GFP expressing control, 0.15 μL of lipofectamine, 0.3 μL of lipofectamine on four human cell lines: HepG2, WI38, A549, HEK293T. By comparing fluorescent signals from post-transfection cells, we aim to decide and optimize the transfection method for performing toehold switch screening. Protocol used: Cell culture protocols, Transfection
Hypothesis
There is no significant difference between the two tested transfection methods for each cell line used in the experiment.
Observations - Fluorescent microscope
HepG2
We can observe green signals from (+) GFP expressing control, 0.15 μL of lipofectamine, 0.3 μL of lipofectamine, and there is no signal on (-) control. There is more signal spotted on samples transfected with 0.3 μL of lipofectamine than in samples transfected with 0.15 μL of lipofectamine.
There was no significant difference observed in signal between the two tested transfection methods.
HEK293T
Similarly like with the HepG2 cell line, green signals were spotted from (+) GFP expressing control, 0.15 μL of lipofectamine, 0.3 μL of lipofectamine, and there is no signal on (-) control. There is more signal spotted on samples transfected with 0.3 μL of lipofectamine than in samples transfected with 0.15 μL of lipofectamine.
There was no significant difference spotted in signal between the two tested transfection methods.
However, there was more signal observed for the HEK293T cell line than for the HepG2 cell line.
WI38
Similarly like with the other cell lines, green signals were spotted from (+) GFP expressing control, 0.15 μL of lipofectamine, 0.3 μL of lipofectamine, and there is no signal on (-) control. There is more signal spotted on samples transfected with 0.3 μL of lipofectamine than in samples transfected with 0.15 μL of lipofectamine.
There was no significant difference spotted in signal between the two tested transfection methods.
The signal observed for the WI38 cell line was higher than the HepG2 cell line, but lower than the HEK293T cell line.
A549
Similarly like with the previously observed cell lines, green signals were spotted from (+) GFP expressing control, 0.15 μL of lipofectamine, 0.3 μL of lipofectamine, and there is no signal on (-) control. There is more signal spotted on samples transfected with 0.3 μL of lipofectamine than in samples transfected with 0.15 μL of lipofectamine.
There was no significant difference spotted in signal between the two tested transfection methods.
The signal observed for the A549 cell line was similar to the WI38 cell line, higher than the HepG2 cell line, but lower than the HEK293T cell line.
Results
There was no visible difference between the two tested transfection methods. Both had sufficient efficiency for our experiments.
The transfection with 0.3 μL of lipofectamine had higher efficiency than the transfection with 0.15 μL of lipofectamine.
The transfection efficiency was the highest for the HEK293T cell line and the lowest, but relatively good, for the HepG2 cell line, which is considered to be hard to transfect.
Conclusions
As expected, there was no clear difference in transfection efficiency between the two tested methods - forward and reverse transfection. Considering its shorter duration and simpler workflow, the reverse transfection protocol was chosen as the preferred method for future large-scale experiments. A noticeable difference was observed between the two tested reagent volumes: transfections performed with 0.3 µL of Lipofectamine per well resulted in visibly higher efficiency than those using 0.15 µL, and this condition was therefore selected for further work. It should be noted that this experiment was performed only once, without technical or biological replicates. The aim of this test was exploratory - to obtain a general impression of which transfection approach and reagent amount appeared more effective - rather than to generate quantitative or validated data. Future experiments with additional replicates and quantitative analysis are planned to confirm these initial observations and establish a more reliable protocol.
Chapter III: Cytometry
Introduction
Toehold switches are sophisticated single-stranded RNA sensors that precisely regulate gene expression at the translational level. In their default, inactive state, they form a secondary structure that blocks the ribosome binding site, preventing protein synthesis. The switch is activated only when it binds to a specific trigger RNA sequence, causing a conformational change that exposes the binding site and initiates translation. Motivated by the urgent need for effective therapeutic strategies against advanced Hepatocellular Carcinoma, we vowed to develop highly selective and adaptable therapies, using tools of synthetic biology. Understanding that HCC is often associated with the high expression of Alpha-fetoprotein (AFP) mRNA-a fetal protein re-expressed in HCC tumours, which became the marker for the disease, as it is not found in cells outside of fetal development. We decided to target this characteristic to design toehold switches to activate, should the sensor find the trigger sequence uniquely associated with the mRNA, expressing the protein of interest, which will help eradicate the cancer cells. The primary objective of this project was to test unique toehold switches, their expression levels, sensitivity and specificity, after delivering them to cells of choice. This was achieved by developing 21 distinct genetic constructs, each integrating one toehold switch positioned upstream of the eGFP coding sequence. This library allows for the systematic evaluation of each switch's function, using eGFP fluorescence as a quantifiable output.
Flow Cytometry Results
To compare the performance of these constructs, a total of 144 samples of HepG2 (AFP+) and HEK293T (AFP-) cells were analyzed using flow cytometry. GFP fluorescence, which measures the expression driven by the toehold switches, was detected in the FITC channel after standard quality control and single-cell gating. The experiment's core data involved quantifying the mean fluorescence intensity (MFI) of GFP to compare the expression levels of the toehold riboswitch constructs between the AFP+ and AFP- cell lines. Two lines, each of 3 repetitions were transfected with a plasmid containing. The results were then scaled to correctly convey the effectiveness of each toehold, regardless of transfection efficiency. The results were then compared with the maximum possible expression in the cells, which was the positive control, transfected with TU1 plasmids containing the eGFP sequence without a toehold. The result of the analysis can be seen in the figure below. The toeholds 15 and 16 were excluded due to mistakes during experiment process.
As the mammalian cells are a challenging environment for riboswitches, due to different translation initiation mechanisms and else, we can see that a significant portion of the toeholds, has low specificity, despite targeting the sequence specific only for the AFP. However among many failed toeholds, we may see that there are few with expression levels higher for cells of interest. However, before any conclusion is drawn, we performed statistical analysis to ensure the significance of the results. In the table below the statistically significant toeholds are listed, which returned the p-value of less than 0.1 for the t-student’s test between the expression levels between the cell lines for each toehold.
| Toehold | P-value | Fold Change | Relative expression in HepG2 |
|---|---|---|---|
| 21 | 0.009 | 31.9 | 0.4 |
| 2 | 0.049 | 14.4 | 0.195 |
| 10 | 0.073 | 3.1 | 1.19 |
| 14 | 0.074 | 3.3 | 1.13 |
To analyse the reason for the success of these toeholds, we tried to correlate their results with parameters available to us.
Discussion
The results confirmed GFP expression within the transfected cells, indicating that the construct is functional. The differences in expression between the toehold variants also suggest that they have a significant impact on expression levels in the human cell environment.
Despite the lower efficiency of riboswitches in mammalian cells documented in the literature, we were able to find a statistically significant difference in expression among the toehold variants between cells containing alpha-fetoprotein mRNA, indicating the constructs' sensitivity in recognizing the target sequence. However, in many cases, this specificity is absent, and variances in expression make it impossible to draw conclusions about the constructs' effectiveness in target sequence recognition. This was further complicated by significant differences in transfection efficiency. While both positive controls fluoresced, the difference in the amount of plasmid taken up by HEK cells compared to HepG2 cells was many times greater (approx. 100-fold), causing the results to be scaled relative to expression.
Correlating parameters related to the toeholds, although we do not have strong correlations between the results, we can note the strongest ones. The average fluorescence in HepG2 cells is positively correlated with the co-fold energy of the toeholds and moderately negatively correlated with the ON/OFF ratio. The second characteristic can be interpreted simply: as the riboswitch leaks and loses specificity, the average expression increases.
The co-fold energy in toeholds signifies the stability of the hybridized state of the toehold molecule and its trigger sequence. The RNA-RNA structure they form is considered more stable the lower the co-fold energy is, which can be interpreted as another way of calculating free energy. Interestingly, because the energy of this bond is negative, a positive correlation between protein expression and this energy means that the stronger the binding and the toehold-trigger duplex, the lower the expression. This result is not necessarily intuitive, as opening the toehold and keeping it open does not necessarily mean higher expression. Toeholds with the lowest co-fold energy had very low expression in both HEK and HepG2 cells, suggesting an increase in specificity at the cost of the construct's efficiency. The final negative correlation between the proportion of GC pairs in the trigger sequence and the co-fold energy points to a relatively simple conclusion that the stability of the hybridization between the toehold and the recognition sequence is directly dependent on the number of hydrogen bonds between the bases.
Analyzing the individual toeholds, the results immediately draw attention to toehold 21, which has very good statistics in every aspect. Its high specificity and accuracy give it excellent parameters and a high ON/OFF ratio; however, this result should be treated with some caution due to the low OFF value. Nevertheless, this structure and result indicate a very good candidate for further experiments and for various modifications for optimization.
Toehold 2 has twice the expression of toehold 21, which is also reflected in its ON/OFF ratio, yet it is also a satisfactory result of our experiment. Worthy of note are toeholds 10 and 14, with sufficiently good statistical significance; they are an example of high sensitivity but lower specificity. Understanding the mechanisms causing their parameters would be important for the further design of riboswitches used in mammalian cells.