Team:UofUppsala/Results
Wet Lab
On our wet lab page you will find information about the experiments that we performed in the lab. Furthermore, you will find our protocols and links to our notebooks. You can read more about the design behind it here.
Development of iGEM Type IIS Standard
Type IIS assembly was used for this project as it allows for easy and fast construction of many different transcriptional units. The flexibility of this cloning method is needed in order to tune the sensitivity of the biosensor since it allows us to comfortably modify the expression level of a certain coding sequence (CDS) by changing the promoter and RBS. This would allow the optimization of the expression level of all components which will help avoid false-positive results. Type IIS can also assemble the multiple transcriptional units that comprise the NANOFLEX system together in the same plasmid. However, iGEM Type IIS standard is fairly new, and is thus in need of improvement. In this section, we describe the experiments and new parts that were constructed for the sake of developing this standard. For more information about this assembly standard you can check the info in the iGEM registry and for more extensive details read our Type IIS guidebook.
Mutagenesis of pSB3C0# pEven Loop Vectors
The sequence of pEven Loop vectors (for Level 2 assemblies) contains an additional, illegal BsaI site. BsaI is the enzyme which is used to cleave the assembly out of a pEven vector to be cloned back to a pOdd vector. Therefore, the presence of this illegal site would impact the construction of Level 3 assemblies.
On one hand, this additional site should not be a huge problem as it merely destroys the plasmid you are no longer interested in. On the other hand, it means that there is an additional site where the BsaI enzyme cuts, which could decrease the efficiency of the cloning reaction.
Therefore, Site-directed Mutagenesis by PCR was conducted on the pSB3C0# plasmids in order to remove the illegal sites. The new plasmids were named pSB3C1# and they can be found in our Type IIS parts collection. Thus, pSB3C01 became pSB3C11 (BBa_K3425001), pSB3C02 became pSB3C12 (BBa_K3425002), and so on.
After PCR mutagenesis, screening was done by digestion of the plasmid DNA with BsaI (Figure 1). All backbones (pSB3C0# and pSB3C1#) contain a band that corresponds to the mRFP1 cassette (BBa_J04455) with a size of 1093bp. Then, pSB3C1# consist of one more band of 2725pb. Since pSB3C0# have one more BsaI site, they should present two more bands of 2434bp and 291bp. While the 291bp band is not visible, probably due to too low DNA amount, the shift in size before and after mutagenesis is clearly visible.
This difference is not visible, however, for pSB3C01 and pSB3C11. In addition, a band of longer DNA fragments can be seen for pSB3C01, but a comparison to well 10 shows that it corresponds to linearized (only single cut) plasmid, indicating that the digest was only partially complete. To investigate the lack of difference between pSB3C01 and pSB3C11, as well as to confirm the mutagenesis by another method, pSB3C01 and pSB3C1# were sequenced.
Sequencing results showed successful PCR mutagenesis for all of the pSB3C1# backbones, with the expected point mutation (Figure 2B). Furthermore, they also showed a point mutation in the original backbone, pSB3C01, (at a different point than the planned one) which eliminated the extra BsaI site (Figure 2A).
Afterwards, some assemblies into pSB3C1# were conducted. One example that was sequenced is the assembly of a dummy part into pSB3C11. You can read more about dummy parts here, and find the sequencing results in the Parts page. The results confirmed that these backbones would work as the new pEven set of vectors.
New pOdd Backbones
One of the problems of iGEM Type IIS standard is the lack of pOdd backbones for Level 3 assemblies. While assembling from pSB3C1# to pSB1K0# is not a problem in silico, it is highly likely that it would not work in vivo. The reason for this is the high copy number of the currently available pOdd vectors (pUC19-derived pMB1 ori, 100-300 copies), as Level 3 assemblies contain multiple transcriptional units, which would make the plasmid toxic to the cells and thus unstable (1).
Three sets of Level 3 plasmids have been designed in order to suit different situations. Each set comprises four plasmids which have the mRFP1 device for pOdd plasmids (BBa_J04454) flanked by the corresponding Type IIS fusion sites substituting the BioBrick sites of the base plasmid. The first one, pSB3K0#, is derived from the pSB3K3 plasmid. These pOdd backbones have the same copy number as the existing pEven backbones (p15A pMR101-derived ori, 10-12 copies).
However, there are situations where this low to medium copy number is still too high. One such example is NANOFLEX, where including all the components would bring the plasmid length over 15kb. This would leave two possible strategies: genomic integration or a very low copy number plasmid (~5 copies). Genomic integration was the chosen strategy when working with B. subtilis, and two sets of very low copy number plasmids were created for E. coli constructs.
One set, pSB4K0#, is derived from the pSB4K5 (pSC101 ori) plasmid. These plasmids keep the alternating Cm-Kan resistances on Type IIS backbones. However, there might be a situation in which a very low copy number plasmid has to be co-transformed with a high copy number plasmid. For this reason, pSB4A0# were derived from the pSB4A5 plasmid (pSC101 ori). This last set can be co-transformed with any of the already existing backbones, allowing even triple transformation as an option for special cases.
Gibson assembly was used to construct these backbones. Afterwards, it was needed to screen for plasmids which contained the mRFP1 device and had the correct antibiotic resistance and copy number. Antibiotic resistance was screened by plating the transformation, the presence of the mRFP1 device by PCR and the copy number by sequencing the ori and by controlled plasmid extraction.
The Gibson assembly did not yield too many colonies and there was no time to repeat it, so there were no samples of pSB3K03, pSB4K01 and pSB4K02. The mRFP1 device (1396bp amplification) was present in all available pSB3K0# and pSB4K02 (Figure 3A).
The full backbone amplification (3756bp for pSB3K0#, 3567bp for pSB4K0# and 4278bp for pSB4A0#) is positive for pSB4K04 and all pSB4A0# (Figure 3B). Although the band in pSB4A03 is barely visible amplification at the 4278bp length is observed, so it is considered as positive too. All these samples except pSB4A03 have multiple bands, which we believe are non-specific and are due to the long extension times required for full backbone amplification. Even with multiple bands, the strongest band corresponds to the expected size.
A smear can be seen in all samples (Figure 3), which we suspect is due to the use of Diamond Dye (Promega). The smear is present in samples which seem to have a lot of DNA and can also be seen in the MW marker. It is something we have consistently observed when using Diamond Dye.
It is concerning how in each gel half of the samples were negative, but it was different samples each time. Judging by the results (Figures 3-6 and 8), it seems to be an error in the handling of samples. It would be good to repeat it, but there was no time. From the available results, however, it seems that all backbones are correct.
To evaluate the copy number of these new backbones, two different experiments were conducted. First, sequencing of the origin of replication and then controlled plasmid DNA extraction.
A region crucial for copy number regulation, the origin of replication (ori), was sequenced (Figures 4-6).
In Figure 4, the sequencing results of pSB3K02 can be seen. It shows multiple single nucleotide changes in two different clones when comparing it to the expected sequence. It is unclear in which step the mutations were generated. Full length sequencing results are available in the Parts page.
In Figure 5, the sequencing results of pSB4A01-4 can be seen. They show no mutations in the ori of all four backbones when comparing it to the expected sequence. Full length sequencing results for all backbones are available in the Parts page.
In Figure 6, the sequencing results of pSB4K02 and pSB4K04 can be seen. They show no mutations in the ori when comparing it to the expected sequence. Full length sequencing results for both backbones are available in the Parts page.
According to QIAGEN, a simple way to distinguish high and low copy number plasmids is to perform a miniprep (2). Controlled plasmid extraction was conducted in order to see the difference between high (100-300), medium (15-20) and low (~5) copy number. This meant standardising plasmid DNA amounts to pellet weight that was used for its extraction, both the plasmids constructed by Gibson and controls for high (pSB1K01), medium (pSB3C11) and low (pSB4K5, pSMART (3)) copy number. The amount of DNA was measured in two different ways: via Nanodrop and via agarose gel electrophoresis.
Standardised plasmid amounts were loaded on an agarose gel (Figure 7). Strangely, the controls (except pSMART) have much higher amounts of DNA than the plasmids constructed by Gibson, in spite of both groups having medium and low copy number plasmids. This is also seen in standardised Nanodrop results (Figure 8), which also shows that the control for medium copy number has a higher amount of DNA than the control for high copy number. A box plot of the results (Figure 9) suggests no difference between medium and low copy number.
An ANOVA test was conducted, however, to see if there was any difference between groups (high, medium, low copy number). The results were non-significant (p=0.52). Even with the clear medium copy number outlier removed (pSB3C11), the test was non-significant (p=0.21). Even if the sample size was small, there is far too much overlap in the standardised values of different copy numbers. Moreover, there was a clear difference in RFP intensity between colonies with high and medium copy number plasmids in plates (Results not shown) when the control strains were used in other experiments. We can conclude that this method is not sensitive enough to distinguish between them, or that more replicates are needed.
Creation of Dummy Parts
Type IIS iGEM standard allows for fast and efficient construction of transcriptional units (TU) and multi-transcriptional units (MTU) in groups of four components. These components are needed in Level 1 assemblies (promoter, RBS, CDS and terminator) but might not be needed in higher level assemblies if the amount of transcriptional units in your system is not a multiple of four.
For this reason, placeholders called dummy parts were created based on an existing Modular Assembly method (4). These parts mimic a basic Type IIS part or a transcriptional unit and can be included in the assembly whenever needed. Two different dummies (TU-DY (BBa_K3425017) and MTU-DY (BBa_K3425018)) were ordered as primers with the corresponding fusion sites to be cloned into all different pOdd and pEven plasmids, with four nucleotides (AACA) in between.
Therefore, pSB1K01-DY is the dummy cloned into pSB1K01 which mimics TU1. It can be cloned to a pEven plasmid like pSB3C11 together with other transcription units or dummies which represent TU2, TU3 and TU4. pSB3C11-DY mimics MTU1 or a promoter part and it can be cloned to a pOdd plasmid in a similar way.
In Figure 10, the sequencing of pSB3C11-DY can be seen as one example of these dummy parts. All the dummy clonings were successful, and the results can be found in the corresponding Parts.
In order to check whether these would work as dummy parts, all four pSB1K0#-DY mimicking four TUs were cloned together into pSB3C11. Similarly all four dummies mimicking MTUs pSB3C1#-DY were cloned into pSB1K02. Sequencing (Figures 11 and 12) revealed that only one out of six clones contains the expected sequence (the last clone of Figure 11). The remaining five clones contain deletions or insertions of various sizes, all of them in the area of the “multi-dummy” sequence.
This is probably due to the short and repetitive sequence generated by the dummies, which might have caused polymerase slippage. This phenomenon can occur when there are reiterated sequences, even if they are imperfect (5). Since the fusion sites at the ends, next to the BsaI and SapI recognition sequences, are present in all the clones it is likely that the cloning was successful and the polymerase slippage happened afterwards, during the growth of the transformed cells.
Dummy parts are only supposed to be fillers that allow the cloning reaction to ligate properly when you have less than four parts to assemble. In this sense, this dummy sequence works as intended. However, it is not ideal to have unpredictable mutations occurring systematically.
A simple solution would be not cloning the dummy sequences next to each other. There might be a situation where this is not possible, so in that case it might be good to use longer dummy parts that are all different from each other. These parts were designed in silico by retrieving random 8bp DNA sequences from the Random DNA Generator (6) and adding the corresponding fusion sites. They were added in the parts registry as pSB1K0#-DY (BBa_K3425023 to 26) and pSB3C11 (BBa_K3425027 to 30).
First Generation
DNA Binding Domains
In our minimal system, we use DNA binding domains (DBDs) for analyte detection as they are naturally occurring in Escherichia coli ( E. coli). The DBDs presented here are designed to get activated and induce transcription of downstream genes upon the presence of caffeine. To test the design, we performed the following experiments.
Assemblies of pSB1C3+DBD-Caff, pSB3K3+pCad and Double Transformations
For cloning of the two constructs pSB1C3+DBD-Caff, pSB3K3+pCadBA+mRFP we used the Gibson assembly method.
DBD-Caff was assembled with pSB1C3 (see Notebook), a high copy number plasmid. The construct was transformed into the E. coli strain DH5α as well as into the E. coli Top10 strain. The presence of the plasmid was confirmed by colony PCR for two colonies of each strain. However, the sequencing results after miniprep showed a frameshift mutation for the plasmid of DH5α cells, and the experiment was henceforth continued with the Top10 strain.
The operator pCadBA was assembled with mRFP and pSB3K3 (see Notebook), a low copy number plasmid. The expression of mRFP is therefore controlled by the binding of DBD-Caff to the pCadBA operator upon the presence of caffeine. After transformation of the plasmid into DH5α, ten colonies were inspected. Several of these colonies were sequenced and confirmed as having assembled the desired construct.
The two assembled constructs pSB1C3+DBD-Caff and pSB3K3+pCadBA were then used for double transformation (see Notebook). The gel electrophoresis of colony PCR products and sequencing results showed that the double transformation was successful. However, the colonies of the double transformation appeared still red. This indicates the constitutive expression of mRFP without induction. The expected results were white colonies with very low to no expression of mRFP in absence of caffeine and red colonies with high mRFP expression in the presence of caffeine. Comparing our system to the system presented in the paper, we noticed that the expression of DBD-Caff in our construct is much higher (7). The high expression level of DBD-Caff derives from a strong promoter in a high copy number plasmid. This design can reasonably cause non-specific activation of the pCadBA promoter. Taking this explanation as the working hypothesis, the next objective was to reduce the expression of DBD-Caff to reduce the non-specific activation of the expression of mRFP, and increase the copy number of pCadBA to increase the range of expression of mRFP.
Subcloning of pSB3K3+DBD-CaFF and pSB1C3+pCadBA
Due to the result of the double transformation, the previously assembled plasmids underwent subcloning (see Notebook). The results were confirmed by a change in the phenotype (Figure 13B). The colonies containing both plasmids had soft-red colour, and adding caffeine increased the expression of mRFP in a visually appreciable manner. The increased expression of mRFP was observed in several colonies.
Caffeine Experiments
A caffeine assay was also performed (see Notebook). The molecule detection system that has been designed is a basic example in which caffeine is used as the analyte of interest. The presence of caffeine in the environment triggers the activation of the system.
The proof that the system is inducible under presence of caffeine is shown in Proof of concept.
Next Generation
Detection Module with an Integrated Amplification Step
Testing the System in Escherichia coli
As a proof-of-concept of a functioning detection system, the aim was to design and express an anti-caffeine nanobody (VHHac) fused to the transmembrane domain of the histidine kinase NarX in E. coli. When two VHHac bind to caffeine, it brings two VHHac-NarX molecules in close proximity, activating NarX through phosphorylation. This in turn phosphorylates the response regulator, NarL, which binds to and activates its respective promoter PyeaR. In our designed system, PyeaR acts as a promoter for the GFP coding sequence thus, GFP will be expressed when caffeine is present. After developing the proof-of-concept in E. coli, the system would be tested and implemented in B. subtilis.
In order to execute this, we designed a sequence consisting of PyeaR with GFP and two copies VHHac fused to NarX, each with the constitutive promoter BBa_J23100. The ribosomal binding site for all three proteins was BBa_B0034. The transcriptional terminator of GFP was BB1_B0010. The design of the system also consists of NarL, which is constitutively expressed and intrinsically present in E. coli. In order to make the construct compatible with Gibson Assembly, it was designed with complementary sequences to the pSB1C3 backbone plasmid. Thus, three Gibson Assembly fragments consisting of PyeaR-GFP and two copies of VHHac-NarX was designed (Figure 14).
A second goal was to optimize the system in E. coli by introducing two different point mutations in NarX to reduce noise. Since the NarX kinases dimerize upon proximity, Mazé and Benenson (8) introduced a system where two different point mutations are introduced in separate NarX kinase proteins in order to reduce phosphorylation between kinases. Thus, the two copies of VHHac-NarX were differentiated by codon optimization in order to perform site-directed mutagenesis.
In order to express and test the system in E. coli the first step was to purify the pSB1C3 plasmid followed by PCR linearization (see Purifying pSB1C3 in Notebook). To assemble the three fragments with the backbone of pSB1C3, Gibson Assembly was performed (see Gibson Assembly in Notebook) and the product was transformed into E. coli (see Transformation in Notebook).
To verify that the system was transformed, colony PCR was performed on a number of colonies (see colony PCR in Notebook) followed by gel electrophoresis (see Gel Electrophoresis in Notebook). The gel electrophoresis indicated that the product of two colonies, referred to as “O” and ”Q” was closer to the desired 3.3 kb than other products (Figure 15). Although the products were less than 3.3 kb, they were observed under UV light where colony “O” showed fluorescence indicating a successful transformation of GFP (Figure 16). However, GFP was expressed without the presence of caffeine, indicating leakage.
To further investigate these colonies, the transformed products were sequenced (see Sequencing in Notebook). The results showed two alignments of primers VF2 and VR, meaning that two of the three fragments were successfully integrated into pSB1C3 and were transformed into the cell.
Additionally, during the Gibson Assembly, the first copy of VHH-NarX was almost entirely deleted. Thus, the resulting construct comprised the PyeaR-GFP and only the second copy of VHH-NarX (Figure 17). After further analyzing the sequence, we found out that the second fragment of the Gibson Assembly reaction had homology to itself. Therefore, since this assembly method is dependent on sequence homology for joining the fragments together, these regions may have annealed with each other during the reaction, causing a "self-excision" by homologous recombination. Luckily, this did not affect the sequence of VHH-NarX: it solely implies that there is one copy of this CDS instead of the intended two copies. This also explains why the gel showed a product of 2.4 kb, corresponding to only one copy construct.
Three caffeine assays were performed on colonies “O” and “Q” (see Caffeine Assay in Notebook) in order to test the system. In the first assay, all cultures including the one without caffeine showed fluorescence and the intensity and absorbance were independent of the caffeine concentration. The hypothesis was that activation of the natural NarX/NarL two-component system had occured due to the presence of nitrates and nitrites in the LB-medium. Another caffeine assay using M9-medium was performed though with no conclusive results, partially explained by the slow cell growth and inability to reach OD 0.3. The use of a high copy number plasmid might be the explanation of the constant expression of GFP, leading to leakage. Thus, future experiments to improve the system may be to change to a low copy number plasmid and only design one copy of the desired genes in the construct to avoid self-excision.
Gram-positive Host
Transformation of the Module in Bacillus subtilis
Before transformation, the B. subtilis plasmids pBS1C and pBS2E were submitted to PCR amplification, to linearize them and remove the mRFP that these plasmids originally contained. Afterwards, the Detection Module constructs were inserted in the respective vectors through Gibson Assembly.
A B. subtilis culture was then transformed* with both plasmids at the same time and plated on LB-agar supplemented with the appropriate concentrations of Erythromycin, Lincomycin and Chloramphenicol. On the following day, no growth was observed in these plates. So, a two-step approach was adopted: B. subtilis was first transformed with pBS1C; upon overnight incubation in Chloramphenicol plates, 6 colonies were picked, submitted to colony PCR (to confirm the presence of PydfJ115_GFP) and then transformed with pBS2E and respective insert.
After plating in double resistance LB-plates, each plate yielded 1-3 colonies, including a negative control where a strain transformed only with pBS1C was inoculated in double resistance plates, pointing towards a failed transformation. This was further confirmed by a colony PCR, where the results obtained were inconclusive.
We suspect that the negative outcome of pBS2E transformation may have been caused by the toxicity of this plasmid’s insert (Rewired NarL + VHH-NarX). A possible explanation for growth inhibition of transformed cells could stem from the size of the VHH-NarX protein alongside the strength of the respective constitutive promoter.
*The B. subtilis transformation protocol used by our team – ISN 2.0 – is a more consistent and practical version of another protocol. If you would like to have access to the original protocol and know more about the optimization process that led to this current version, please check the B. subtilis Handbook, done as a collaboration with iGEM team Generation Mendel, from Brno University, Czech Republic.
Signal Amplification and Noise Reduction
As described above, the presence of target ligand specifically activates the detection module and triggers the expression of the reporter protein, β-galactosidase (β-gal). The related experiments showed that essential components of the "minimal biosensor" work as expected. However, they also confirmed our initial hypothesis that at the absence of the ligand there is a background signal - noise (see Proof of concept). To increase the signal to noise ratio, the system needs to be optimised. This can be achieved by implementing several signal amplification and noise cancelling mechanisms. As we have foreseen this issue, we have already started building components for optimization.
Reporter System: β-galactosidase
Our reporter system is based on the enzymatic cleavage of X-gal or ONPG by β-galactosidase.
Before testing the β-galactosidase activity in vivo we tested the reaction time of the enzyme in vitro. For this, different concentrations of β-galactosidase ranging from 0.5 to 3 U/ml were added to 0.2 mg/ml X-gal or ONPG. The absorbance was measured for 1 h at OD610 for X-gal or at OD420 for ONPG. From the values of three biological replicates a standard curve was created. By mixing β-galactosidase with X-gal in vitro , a visible blue color change of the media could be observed in approximately 30-40 minutes (Figure 19), while cleavage of ONPG resulted in a yellow colored media in 5-10 minutes (Figure 20). The absorbance for the in vitro samples in ONPG or X-gal raised with increasing amounts of β-galactosidase as expected. This is also visible by naked eye as shown in Figure 18. Despite the faster colour change using ONPG as a substrate we decided to use X-gal for the final system because yellow is harder to see by naked eye as the color of the culture media is slightly yellow.
For the expression of β-galactosidase in vivo its corresponding gene lacZ was amplified from the E. coli MG1655 chromosome via PCR. The agarose gel electrophoresis of the PCR product is shown in Figure 21. The band indicates a DNA fragment length of > 3000bp which resembles the size of lacZ (3075bp).
For cloning, two different methods were used: Type IIS assembly and Gibson assembly. With Type IIS assembly lacZ was cloned into the pSB1K01 plasmid combined with a constitutive promoter. With Gibson assembly lacZ was cloned into pBAD24 which includes the inducible pBAD promoter. Both plasmids were transformed into chemically competent E. coli DH5α cells. We chose this strain as our model organism as it lacks the natural chromosomal expression of β-galactosidase to avoid false positive reactions.
LacZ was only inserted correctly into pBAD24. Type IIS assembly was implemented several times but remained unsuccessful. Figure 22 shows the gel electrophoresis of the colony PCR products from the target genes after the plasmids were cloned and transformed. The second of the two clones transformed with pBAD24 shows a band at the size of 3000 bp which indicates the insertion of lacZ.
The DNA sizes in lane 3 and 4 indicate an unspecific insertion of DNA fragments and ligation of pSB1K01. Vaguely speculating, the reason for the difficult integration of lacZ into pSB1K01 via Type IIS could reside in the length of the gene.
The successfully transformed cells were subjected to 1 mg/ml X-gal or ONPG. First, cells grew in LB media mixed with glucose overnight, and were diluted the next morning. Glucose silences the expression of lacZ. To start the expression of lacZ 0.2% arabinose was added. As a negative control CaFF-I pSB1C3 DH5α cells were used. To test if X-gal/ ONPG was cleaved intra- or extracellular, the reactant was added to the cells directly or to the supernatant after centrifugation of the cells. Absorbance was measured with a spectrophotometer every 30 minutes for a total of 180 minute. Figure 23 shows that the change in absorbance is higher for ONPG than for X-gal. A visible color change was observed only for ONPG samples after 60 minutes.
The absorbance measurements of the in vitro and in vivo experiments could not be compared. The reason for this is that for the in vivo samples no colour change could be observed by naked eye when adding 0.2 mg/ml substrate as we did in the in vitro measurements. Therefore, 1 mg/ml of X-gal/ ONPG was added to the cells.
In order to achieve an even stronger signal, the final step would have been to work on the compatibility of the β-galactosidase system with the rest of the amplification loop. The β-galactosidase system was not combined with the detection system due to lack of time. The reporters used for testing the system were RFP and GFP.
Qβ Replicase Amplification System
Qβ
To increase the sensitivity of the system, Qβ replicase was introduced (see Design). Initially, the β-subunit of Qβ replicase was cloned under a constitutive promoter to simply test the protein expression. Since this resulted in large insertions of genomic DNA into the Qβ sequence (Figure 25A), Qβ is likely toxic to the cell in the amount we tried to express it in. Previously, Yao et al. (9) expressed Qβ under an inducible promoter. Therefore in the next step, we cloned Qβ replicase under the pCAD inducible promoter (Figure 25B), which prevented Qβ from mutating. pCAD is the promoter activated by the detection module and linking it directly to QB replicase expression is the simplest variant of incorporating the Qβ amplification system in the biosensor. This construct is needed for testing the Qβ amplification. However, in the final design, Qβ is controlled by T7 RNA polymerase, which allows for implementing the T7 lysozyme as the noise reducing mechanism.
Qβ Recognition Sequences
Qβ requires specific recognition sequences on either side of the sequence you want to amplify. 5'Qβ and 3'Qβ were shown to be the most efficient pair of recognition sequences (9) and therefore we chose to use these.
Type IIS iGEM assembly standard allows easy cloning of transcriptional units which consist of four typical parts - promoter, RBS, CDS and terminator (You can read more about Type IIS here.) However, the template of Qβ replicase is atypical in a sense that it is composed of more than four parts: promoter, 5'Qβ recognition site, RBS, CDS, 3'Qβ recognition site and terminator. To simplify the cloning procedure for initial experiments, the 5' recognition site was ordered together with medium strength RBS (B0034) and 3' recognition sequence was fused with the terminator obtained from the Distribution Kit 2019.
Since synthesis of terminator sequences is non-trivial, the 3´ recognition site was synthesised separately and then fused to the terminator obtained from the Distribution Kit 2019 (see Notebook) .
When the fusion construct of 3´site and terminator was sequenced, deletions of 29bp were found in the middle of 3' recognition site (Figure 26A). Since the 3´ recognition site was synthesised, we first suspected that the mutations emerged during PCR (Notebook). However, subsequent sequencing of the 3' recognition site cloned alone into Level 0 backbone revealed the set of non-identical deletions of 24 or 31 bp localised exactly in the same region (Figure 26B). Since the 3' recognition site is an RNA functional element, any change in its sequence might compromise its function. The sequence itself seems to be prone to mutate. Since in Level 0 the 3' recognition sequence is not transcribed, the mutation pressure probably acts on the DNA level. Without a working set of two recognition sequences, no further experiments measuring the activity of Qβ could be done.
Noise Reduction
TEV Protease
In order to reduce the basal level of the Qβ replicase, we designed the Qβ sequence with the LAA degradation tag at the 3’ end. In between the Qβ sequence and the degradation tag, a TEV cleavage site was inserted. This would allow to cut off the degradation tag by TEV protease when the system is activated resulting in a stable level of Qβ replicase.
The TEV protease sequence was ordered with the Type IIS CDS prefix and suffix and it was cloned into the Level 0 backbone. However, we did not proceed further with the implementation of the TEV protease into our system.
T7 RNA Polymerase and T7 Lysozyme
Apart from adding TEV tag to avoid leaking problem, more approaches were added. Since the Qβ replicase is highly sensitive, even a small amount of leakiness will be amplified and can cause a false-positive result. To obstruct this condition, T7 RNA polymerase was introduced to express Qβ. T7 RNA polymerase is extremely promoter-specific and more importantly, it can be reliably controlled by T7 lysozyme. This makes the T7 lysozyme a key component of the noise-reducing mechanisms.
In the final design, we place a constitutive promoter before the T7 lysozyme, so it is produced at a stable level and blocks the activity of the basal level of T7 RNA polymerase. The basal level of T7 RNA polymerase was determined by the level of noise measured in the detection module (see Proof of concept). Subsequently, the level of T7 lysozyme would be set so as to be higher than that, in order to block all T7 RNA polymerase molecules.
T7 RNA polymerase and T7 lysozyme were amplified from the E. coli(BL21)pLysS (Notebook) and the size was verified by agarose gel electrophoresis. For the purpose of experiments verifying the expression of T7 lysozyme, we first cloned it under LacUV5 and T7M5 inducible promoters. Since single amino acid of change was identified in both clones [Gly(119)->Val(119)] work was not continued due to the time constraints.
Checking Nanobodies: Agglutination Approach
With the agglutination approach we wanted to develop a biosensor based on an agglutination reaction using caffeine-binding nanobodies. These nanobodies are recombinant antibodies (VHH) and bind specific to caffeine. Agglutination can be triggered in the presence of large antigens such as BSA-caffeine haptens. Caffeine can compete for the binding of nanobodies, stopping agglutination from happening and triggers pelleting of the cells instead. The agglutination test as described by Riangrungroj et al. (10) aims to figure out the amount of caffeine that occupies enough nanobodies to trigger clumping. This reaction serves the function of a biosensor.
For detecting the minimal concentration, anti-caffeine VHH expressing cells where opposed to BSA-caffeine as well as to different concentrations of caffeine. When BSA-caffeine binds to the nanobodies, cells will agglutinate as BSA serves as a space holder between the E. coli cells. When a higher concentration of caffeine than BSA-caffeine is present in the sample, caffeine will interact with the nanobodies instead. Caffeine as a much smaller molecule reduces the space between the cells consequently resulting in a cell pellet.
For performing the experiments, the nanobodies were designed to be placed in the outer membrane of E. coli DH5α cells. Therefore, anti-caffeine VHH was attached to intimin which is a cell membrane protein and important for the transportation of the nanobody to the outer membrane. To express this fusion protein called CaFF-I, the coding sequence was cloned into the high copy number plasmid pSB1C3. Subsequent, the construct was transformed into chemically competent E. coli DH5α cells for amplification and storage. However, the sequence of CaFF-I showed frame-shifting mutations repeatedly after sequencing. The difficulty of the assembly and the mutations indicate that this construct might be toxic to the cell when overexpressed constitutively.
The next steps to attain a successful agglutination biosensor are:
- Use inverse mutagenic PCR to fix the frameshift mutations
- Redo assembly, accounting for transformation efficiency of a control. This will give us more knowledge on the possible toxicity of CaFF-I
- Reassemble CaFF-I on an inducible vector (e.g. pET)
- Develop an assay to measure the binding affinity of the nanobody to caffeine
- Once the previous steps are successful, determine the Limit Of Detection (LOD), the lowest concentration that our cellular biosensor is capable of detecting
Detailed protocols for experimental implementation can be accessed below.
Protocols
| Download | Protocol |
|---|---|
 |
Agarose Gel Electrophoresis |
|
ISN 2.0 Transformation Protocol (B. subtilis Handbook, pages 16 and 17) |
|
Caffeine assay |
|
Cryostock |
|
Determination in vitro of B-gal activity |
|
Development of Competence in B. subtilis and Transformation |
|
Distribution kit DNA resuspension |
|
Electroporation - Transformation of B. subtilis |
|
Eurofins Mix2seq kit |
|
GeneJET PCR Purification Kit, K0701 |
|
GeneJET Plasmid Miniprep Kit, K0502 |
|
Genomic DNA purification |
|
Gibson Assembly |
|
Lyophilization |
|
Luria-Bertani (LB) media liquid and agar |
|
NEBuilder HiFi DNA Assembly Reaction |
|
Overnight liquid culture |
|
PCR - Phusion™ High–Fidelity DNA Polymerase |
|
PCR reaction - Phusion and Colony PCR |
|
Plasmid Isolation |
|
Promegakit - Assembly of Restriction Enzymes Digestion |
|
Sample Preparation for B. subtilis - Colony PCR |
|
Sequential Double Digest |
|
Spheroplasts formation |
|
Subcloning: digestion and ligation reaction |
|
Transformation of chemically competent E. coli DH5α cells |
|
Type IIS cloning (MoClo) |
References
- Jones, K. L., Kim, S.-W., and Keasling, J. D. (2000) Low-Copy Plasmids can Perform as Well as or Better Than High-Copy Plasmids for Metabolic Engineering of Bacteria. Metabolic Engineering. 2, 328–338
- How do I know if my plasmid is a high- or low copy number type? - QIAGEN [online] https://www.qiagen.com/no/resources/faq?id=1f42840e-fbd7-4734-b0cd-e17372a9e5a4&lang=en (Accessed October 24, 2020)
- CopyRight v2.0 BAC Cloning Kits (pSMART BAC and pEZ BAC) [online] https://www.lucigen.com/CopyRight-v2.0-BAC-Cloning-Kits-pSMART-BAC-and-pEZ-BAC/#subcat-tabs3 (Accessed October 21, 2020)
- Binder, A., Lambert, J., Morbitzer, R., Popp, C., Ott, T., Lahaye, T., and Parniske, M. (2014) A Modular Plasmid Assembly Kit for Multigene Expression, Gene Silencing and Silencing Rescue in Plants. PLOS ONE. 9, e88218
- Levinson, G., and Gutman, G. A. (1987) Slipped-strand mispairing: a major mechanism for DNA sequence evolution. Mol Biol Evol. 4, 203–221
- Random DNA Generator [online] https://www.faculty.ucr.edu/~mmaduro/random.htm (Accessed October 16, 2020)
- Chang, H. J., Mayonove, P., Zavala, A., De Visch, A., Minard, P., Cohen-Gonsaud, M., and Bonnet, J. (2018). A modular receptor platform to expand the sensing repertoire of bacteria. ACS synthetic biology, 7(1), 166-175.
- Mazé, A., and Benenson, Y. (2019) Artificial signaling in mammalian cells enabled by prokaryotic two-component system. Nat Chem Biol. 16, 179–187
- Yao, Y., Zhang, W., Zhang, M., Jin, S., Guo, Y., Zu, Y., Ren, K., Wang, K., Chen, G., Lou, C., and Wu, Q. (2019) A Direct RNA-to-RNA Replication System for Enhanced Gene Expression in Bacteria. ACS Synth. Biol. 8, 1067–1078
- Riangrungroj, P., Bever, C. S., Hammock, B. D., and Polizzi, K. M. (2019). A label-free optical whole-cell Escherichia coli biosensor for the detection of pyrethroid insecticide exposure. Scientific reports. 9(1), 1-9