Cycle 1

Design

At first, we looked for a protein marker that could be used to identify Entamoeba histolytica. We found a few promising candidates and, in the end, we chose cysteine proteinase 5 (CP5) and pyruvate phosphate dikinase (PPDK) as our analytes because they are unique markers of the E. histolytica. One of them has even been already considered as a marker for a test in previous papers [1,2]. We decided to put it in the pET28a(+) plasmid as in that way we could manipulate the insert in various ways and get the construct in both N-His and C-His forms.

Build

The cloning of PPDK was unsuccessful - the sequencing showed that there were a few mutations in our gene of interest that made it impossible to use the protein. We tried to clone it again. After four failed attempts we were running out of our ordered construct, so we decided to clone it into a pUC19 vector with simple blunt-end cloning. Since that worked at the first time, we decided that there was probably something wrong with our plasmid backbone. Finally, we successfully cloned PPDK into the pET28a(+) vector.

Test

We tested a few different E. coli strains and growing conditions for the induction. Some had no visible protein at all, some had a lot, but none of them had a substantial amount in the soluble fraction. After determining that the BL21 (DE3) strain had the most protein in the insoluble fraction, we tested out different lysis conditions varrying the saccharose, NP-40 and Triton X-100 concentrations in the lysis buffers.

Learn

After all the testing we found out that the best conditions for induction are in the E. coli strain BL21 (DE3) with a 1 mM IPTG concentration for 3 hours at 37°C in the TB medium [2]. The most protein was found in the soluble fraction when we used a lysis buffer with 50 mM NaH₂PO₄, 500 mM NaCl, 10 mM imidazole, and 0.5 % NP-40. The protein was not pure enough to start the SELEX and there was not much of it, therefore we needed to find an alternative.

Cycle 2

Design

Since we could not synthesize a significant amount of our target protein, we decided to perform our SELEX process on a denatured (and refolded) protein. We decided to denature the proteins in the sample blood on our test.

Build

At first, we needed to find out the specific concentration of a needed strong detergent. Then we required to find a buffer solution, which could keep the protein soluble and would not interfere with the SELEX process downstream.

Test

We tried to dissolve the PPDK containing biomass with a gradient of urea buffers to find out at what concentration the protein dissolves the best. After determining the denaturing lysis conditions, we dialyzed the lysate against a soluble lysis buffer containing 5 different additives - arginine (0.375 M), trehalose (0.75 M), proline (0.5 M), mannitol (0.5 M) and CuCl₂ (10 mM) [3, 4].

Learn

We found out that PPDK becomes soluble when a lysis buffer supplemented with 6 M urea is used. It is possible to transfer the solubilized protein in a phosphate buffer containing 500 mM of NaCl and 0.375 M arginine. The protein stays soluble in this solution and can be used in downstream applications.

Cycle 1

Design

At first, we looked for a protein marker that could be used to identify Entamoeba histolytica, we found a few promising candidates. In the end, we chose cysteine proteinase 5 (CP5) and pyruvate phosphate dikinase (PPDK) as our analytes because they are unique markers of E. histolytica and one of them has even been considered as a marker for a test already in previous papers [1,2]. We decided to put it in the pET28a(+) plasmid because we could manipulate the insert in various ways and get the construct in both N-His and C-His forms.

Build

The cloning of CP5 into the pET28a(+) vector was successful with both N-His and C-His tags, then it was transformed into the E. coli BL21 (DE3) strain.

Test

The protein synthesis was successful and we did not need to optimize the growing conditions besides the IPTG concentration – we compared the protein expression at 1 mM and 0.6 mM IPTG and we found that the lower one works better. We ran into a bunch of problems while trying to process the insoluble protein because it had to be denatured, renatured, and then activated [5, 6]. The protein was dissolved and purified successfully, but we were not able to refold or activate it. After a few consultations with relevant experts from the lab of our PI, we tried to increase our refolding buffer’s glycerol concentration by adding saccharose, changing the refolding conditions from 4 °C to 37 °C, but none of it worked. The only condition left that we could change was the ratio of buffer to protein. However, that was not feasible because we did not have the equipment to concentrate large volumes of the buffer after refolding. The method we did try was pulling out all the water from the buffer through a dialysis membrane with carboxymethyl cellulose. It took more than 4 days with constant cellulose clean-up and change for the buffer to go from 100 ml to about 20-30 mL.

Learn

In the end, we could not find out, what was the problem with CP5 processing due to the following reasons. It is not possible to determine if the protein folded correctly or did not fold at all, and whether the problem is in the protein activation process - all we could see was the 32 kDa band on our SDS gels. Since we could not produce the soluble protein, but could purify insoluble protein relatively easily, we decided to change our approach to the test itself. We considered doing our SELEX process on the denatured CP5 protein and then made a test in which we would denature the proteins in the blood sample too.

Cycle 2

Design

The design of the test stayed the same, we just needed a different approach to get the soluble and active CP5 protein.

Build

Since we did not manage to produce an active and soluble protein, we consulted with a relevant expert in our PI’s lab and got advised that we maybe could find some fraction of the soluble protein inside of the cells. Every time we grew our E. coli cells with the CP5 construct we saw a noticeable decrease in the weight of the biomass compared to other proteins. Therefore we made a conclusion that there must have been some fraction of the proteinase that lysed the cells from the inside.

Test

We tried seven different E. coli strains: BL21(DE3), Rosetta-gami (De3), Rosetta (De3), C41, Arctic express, HMS(17) and KRX. None of them produced any soluble CP5 protein in any considerable amount, although the biomass was smaller in every instance.

Learn

The protein was obviously activated and folded correctly in some cells because every could only grow about half as much biomass compared to other proteins with identical growth conditions. Since a foreign proteinase was activated inside E. coli cell, the cells were lysed and the active protein was probably denatured and lost in the growth medium.

Cycle 3

Design

As we could not produce any active protein for the SELEX process, we tried to produce some soluble denatured protein in the pro-form and perform the SELEX process with it and then process the blood sample in such way that the proteins in it would be similar to our denatured analytes in the lab.

Build

At first, we lysed the cells in denaturing guanidine-hydrochloride and urea buffers to solubilized inclusion bodies. Then we decided to just dialyze the whole cell lysate against the soluble lysis buffer with various different additives that might help to stabilize the protein in solution. Then we centrifuged the insoluble cell debris and considered to use this solution for downstream SELEX application as the lysate.

Test

The additives we tried out were arginine (0.375 M), trehalose (0.75 M), proline (0.5 M), mannitol (0.5 M) and CuCl₂ (10 mM) [5,6]. Some of them were clear as water after the centrifugation - this result showed that nothing from our lysate was soluble in them, but a few were cloudy to some degree which was promising.

Learn

We found out that the protein was stable in a soluble lysis buffer containing 10mM CuCl₂, but the amount of the stable protein was very small and not enough for further applications. In the end, we decided not to use CP5 as a biomarker for Entamoeba histolytica.

Cycle 1

Design

Since one of the parts of our project was to create a fusion protein system for the bottleneck reaction of naringenin synthesis, we decided to model it in silico. The model’s primary purpose is to check whether the shorter distance between active sites leads to the higher production of naringenin. In the process, we found out that even the modeling part required to cover the main steps of engineering: design, build, test, and learn. Therefore, we began with the already existing modeling workflow described for a fusion construct with a malaria pre-erythrocytic vaccine candidate [7]. The described modeling approach was initially designed as follows:

1. Choosing the protein candidates
2. Selecting linkers
3. Primary structural analysis of the fusion system
4. Protein modeling
5. Scoring modeled complexes to choose the model for visualization
6. Plotting Ramachandran graphs before structure refinement
7. Structure refinement
8. Plotting Ramachandran graphs after structure refinement

Build

The building stage of the modeling flow consisted of performing the initial design flow. In this cycle we have raised a hypothesis that homology-based protein modeling approaches should not be the most suitable method for our case, since there are few fusion models existing.

1. The protein candidates were chosen to be 4-coumarate-CoA ligase 2 from Glycine max and chalcone synthase from Arabidopsis thaliana
2. Seven linkers were chosen for our fusion system. Four of them were flexible glycine and serine linkers (GSG, (GGGGS)n, where n is equal to 1, 2, and 3) and the other three were rigid ((EAAAK)n, where n is equal to 1, 2, and 3)
3. For a primary structural analysis we studied PDB: 3TSY structure - an experimentally determined structure of the fusion system that has 4CL protein in it. We took this structure as a starting point of what we can expect from our models
4. Protein modeling was initially attempted to perform using homology-based modeling program SWISS-MODEL [8]
5. The modeled structures in the first cycle were evaluated using VoroMQA [9], QMEAN [10], and QMEANDisCo [11] scores
6. The plot was not drawn in this cycle
7. The structure refinement was not performed in this cycle
8. The plot was not drawn in this cycle

Test

The output of SWISS-MODEL showed that sequence identity between the templates found and the sequence that is modeled is 66.42 - 69.36%. According to the specialists in protein modeling, homology-based methods are suitable, when the sequence identity between the template and the modeled sequence is more than 25%, belonging to the “daylight" zone [13], thus, in theory, the method could be applied.

The modeling flow was tested with PyMOL [12] visualization program using commands `cealign`, `set seq_view`. We were looking for an accurate representation of the linked proteins in the fusion system by comparing them to their distinct versions. The main focus on the system was the linker region, which in the homology-based modeling case was composed of the uncoiled domain of 4CL that made up a highly disordered massive linker region. We stopped the flow after the visualization because we decided that structure refinement using molecular dynamics will not introduce any significant changes to the structure.

Learn

Nonetheless in theory the homology-based modeling method could be applied, these modeling results proved that this approach could not be applied to our fusion protein system. Therefore, we decided to try running ab initio protein modeling programs.

Cycle 2

Design

The design (main steps) of the modeling flow stayed the same as it was in the first cycle, yet also we included one more step - energy minimization with Yasara [16] - that was inserted after the protein modeling step. The main changes in the second cycle were the choice of protein modeling programs. In this engineering cycle we took ab initio modeling programs trRosetta [14] and RoseTTAFold [15] to model our proteins.

Build

1. Protein candidates stayed the same as in the first cycle
2. Linkers stayed the same as in the first cycle
3. Primary structure analysis was not required to be redone in the second cycle
4. Protein modeling was performed using trRosetta and RoseTTAFold simultaneously
5. The modeled structures were evaluated using VoroMQA score
6. Energy minimization with Yasara
7. The Ramachandran plots were drawn before molecular dynamics (MD) simulations
8. We performed short MD simulations for the trivial linker (GSG) case
9. The Ramachandran plots were drawn before molecular dynamics (MD) simulations

Test

The modeled structures visually were not satisfactory - the disordered domain of 4CL protein was still present in the structure. The energy minimization step with Yasara and MD simulations did not introduce any significant changes to the structure.

Learn

After this step we decided to consult the bioinformatician who works with protein modeling. He suggested we include multiple sequence alignment files into our modeling flow. Additionally, we got advice to use more scoring functions in the evaluation step. Yet also we consulted the protein molecular dynamics specialist to get a professional insight into how we should run the MD for our system to get a more significant output. After the consultation we adjusted the box size of the system and the duration of the simulations.

Cycle 3

Design

In the third engineering cycle, we included our own generated multiple sequence alignment (MSA) files as input to the protein modeling programs of our choice. In addition, we took into account the advice from the specialists and included QMEAN and QMEANDisCo scores into the evaluation of the structures. The structure refinement was not performed in this cycle due to the lack of computational resources at the time.

Build

1. Protein candidates stayed the same as in the previous cycle
2. Linkers stayed the same as in the previous cycle
3. Primary structure analysis was not required to be redone
4. Protein modeling was performed using RoseTTAFold
5. The modeled structures were evaluated using QMEAN, QMEANDisCo, and VoroMQA scores
6. The plot was not drawn in this cycle
7. The structure refinement was not performed in this cycle
8. The plot was not drawn in this cycle
9. The distances between active sites in fusion protein systems with rigid linkers were calculated

Test

The modeled structures were visually satisfactory - the disordered domain of 4CL protein was less disordered in the modeled structure.

Learn

The multiple sequence alignment (MSA) files have a significant impact on the modeling output.

Cycle 4

Design

According to the output of the third cycle, the fourth engineering cycle was not required. However, after the latter cycle was finished, AlphaFold2 [17] became available for public use. Therefore, we decided to apply this highly evaluated tool in our modeling workflow. The structure refinement was not performed in this cycle due to the insignificant impact of MD for structures recorded in the studies [18].

Build

1. Protein candidates stayed the same as in the previous cycle
2. Linkers stayed the same as in the previous cycle
3. Primary structure analysis was not required to be redone
4. Protein modeling was performed using AlphaFold2
5. The modeled structures were evaluated using QMEAN, QMEANDisCo, ProQ2D, ProQRosCenD, ProQRosFAD, ProQ3D scores
6. The plots were drawn in this cycle
7. The structure refinement was not performed in this cycle
8. The distances between active sites in fusion protein systems with rigid linkers were calculated

Test

The output of the fourth cycle had the less disordered domain just as the output of the third cycle.

Learn

We were visually satisfied with the modeled complex. Nonetheless, AlphaFold2 models do not require the usage of molecular dynamics for structure refinement, we considered applying them to get a better insight of the system with flexible linkers.

Cycle 1

Design

Before beginning SELEX with purified soluble proteins we chose a few SELEX protocols available from the internet. One in the industry called XELEX [19] and another from an article specifically intended for fool-proof aptamer discovery[20].

Build

We tried the initial SELEX runs with the default parameters but for some reason, it kept failing, so we chose as many parameters as we could in order to optimize the process. We settled on the bead and protein amount in the SELEX mixture, the heating of aptamers before the incubation length, the storage buffer of the aptamers post-cycle, and the polymerase itself.

Test

We lowered the amount of protein and beads because we suspected that it might interfere with both the SELEX process and the downstream PCR reaction. We also lowered the heating of the aptamers before the selection round from 10 to 5 minutes based on another source [21]. We also substituted the TE exchange buffer with distilled water because the EDTA in the TE buffer interferes with the PCR reaction polymerase. Finally, we tested the efficiency of the PCR reactions with both DreamTaq and Phusion polymerases.

Learn

In the end, we did manage to optimize the SELEX process enough for it to produce aptamers. We determined that more protein and beads result in obstruction of SELEX. Half of the initial quantity should be applicable for second and further rounds. we also found out that we should heat the aptamer pool for 5 minutes instead of 10 minutes after the first round of SELEX. Heating for too long results in degradation or other forms of loss of our dsDNA from the previous round. The aptamer pool post-selection round should be stored in water, not in TE buffer as the EDTA interferes with the functions of polymerase, Phusion polymerase seems to not be affected as much as DreamTaq, and works better in both water and TE buffer.

Cycle 2

Design

For the second cycle, we researched that less bead and protein complex should have better effect in regenerating aptamers. The first SELEX round should have higher quantity of beads for making sure that all possible aptamers bind. However, second and further rounds should have half of the initial volume. It also increases stringency parallel to other parameters.

Build

In order to check which quantity of complex suits us best, we developed a plan to try changes on separate initial and move on to further rounds.

Test

We tested unchanged bead and protein complex quantities for a few rounds and tried various complex volumes for second and further rounds side by side. Comparing volumes reduced by ⅓, ½ and ⅔, we found that change to half of the initial volume suited our SELEX protocol the best.

Learn

We found that first round should have larger volume of bead complex and lower in later rounds. More protein and beads result in obstruction of SELEX. Half of the initial quantity should be applicable for second and further rounds.

Cycle 1

Design

One of our project’s parts is genetically modified probiotics that are capable of performing de novo synthesis of natural flavonoid naringenin. This flavonoid has reached our attention as it remarkably reduces Entamoeba histolytica viability [22] and does not cause any harm to humans. Conversely, it has been found that naringenin has antioxidant, antitumor, antiviral, antibacterial, anti-inflammatory, anti-adipogenic, and cardioprotective features [23]. To create an efficiently working naringenin production system, we have been searching for various tools - the synthetic protein quality control (ProQC) system, the most efficient metabolic pathway enzymes, promoters of particular strength, etc. However, it is not only important to construct a fully functional system but also to ensure its stability in the chosen chassis. For this reason, we chose to insert naringenin synthesis metabolic pathway into our probiotic strains: E. coli Nissle 1917 and Lactobacillus casei BL23. Insertion into genomic DNA has several advantages. Firstly, it provides stability to the system. If the new metabolic pathway does not cause too much burden to the chassis, it should be maintained in genome active (otherwise there might be some mutations suppressing its functionality, reduced cell growth, enhanced cell sensitivity to nutrient deficiency [24]). Surprisingly the integration of the genetic circuit of the metabolic pathway into the chassis genome contributes to the burden reduction as it serves as DNA copy number tuning. Secondly, by inserting the desired sequences in the genome we eliminate the need for antibiotics or any other environmental pressure usage to maintain needed genetic circuits in the chosen chassis. This is very important when the designed naringenin-producing bacteria will live in the human intestine.

In order to reach this goal, we did a huge research in the genome editing techniques field. With the help of our advisor Inga Songailienė, who is focusing on CRISPR-Cas systems in her Ph.D. thesis, we have chosen to use the pCas-pTarget plasmids system which enables the usage of CRISPR-Cas9 and Lambda Red recombination combinment into the one system to acquire the efficient E. coli Nissle 1917 genome editing [25]. For L. casei BL23 we have chosen a single plasmid pLCNICK based system relying on CRISPR-Cas9 D10A Nickase-Assisted genome editing [26]. For the genomic insertion site in L. casei BL23 we decided to follow the pLCNICK already provided sequences required for genome editing. We followed this approach because we have run out of resources for new DNA sequences, and for this chassis, there is a need for longer homology arms. Meanwhile, we turned our all focus on E. coli Nissle 1917 to design the functional and robust metabolic pathway. In the future, we would adapt it to the L. casei BL23, too.

During the design stage, we first considered genomic sequences that would be used for metabolic pathway insertion. From literature, we learned that 16S rRNA genes might be used for this purpose [27]. However, it would not let us implement a ProQC system due to 16S rRNA transcript being created as polycistronic RNA. For this reason, we targeted non-essential gene - colicin. We have also considered possible metabolic flux enhancement toward naringenin synthesis. It covers acetate kinase (ackA), phosphate acetyltransferase (pta) knockouts creation to channel carbon flows toward acetyl-CoA [28]. Subsequently, it leads to the enhanced malonyl-CoA amounts needed for bottle-neck reaction in naringenin production [29]. Also, tyrosine-specific transporter (tyrP) knockout mutants are potentially able to produce 10% higher amounts of L-tyrosine (naringenin synthesis precursor) than that of the original strains [30]. In addition, aldehyde-alcohol dehydrogenase (adhE) knockout disrupts the conversion of acetyl-CoA to ethanol in the cell, therefore more acetyl-CoA can be converted to malonyl-CoA [31] further enhancing naringenin synthesis.

Build

To fulfill our experimental plan, we have designed sgRNA for genomic insertion and earlier mentioned knockouts generation. For this reason, we used CRISPOR tool [32] to assess the probability of sgRNA efficiency and off-targets generation. Furthermore, after selecting sgRNA for each gene we have checked how many other similar DNA sequences exist in the E. coli Nissle 1917 genome. For this reason, we have used NCBI Blast tool [33] to analyse whether there are any other identical sequences to the seed region of the chosen sgRNA. It is important to be aware of this because it might contribute to possible off-targets [34]. Also, for genome insertion we decided to use approximately 70-90 bp long homology sequences flanking the inserts as it should work on this system [35]. By following these considerations we have constructed sgRNAs specific to ackA (BBa_K3904401), pta (BBa_K3904402), adhE (BBa_K3904403), tyrP (BBa_K3904404), colicin (BBa_K3904405) genes.

Test

We have performed genome editing as it is described in the literature [25]. However, genome editing has not been very efficient as very few cells have survived and not all of them contained desired modifications.

Learn

As we have received these results, we have hypothesized that some of the cells might lose pCas plasmid as it has a temperature-sensitive replicon.

Cycle 2

Design

In the second genome editing cycle we have used the same parts but changed some experimental conditions. We have contacted our advisor Inga Songailienė, who gave us some insights about this system.

Build

Here is experimental conditions which we have decided to change:

1. Lower growth temperature from 30°C to 27.5°C because a water filled incubator might show a bit lower temperature than it is in it.
2. Cells have been induced for recombination at 0.5-0.8 OD instead of 0.4-0.6 OD during preparation for co-electroporation with pTarget and recombination template.
3. Enhanced the L-arabinose concentration till 0.2 % (previously it was 0.15 %).

Test

We get pta and ackA gene knockouts. However, even after several repeats and recombination template amount increase (600 ng instead of 400 ng), new recombination template preparation, adhE and tyrP genes left unmodified.

Learn

By these changes, we learned that the timing of recombination induction is very important. However, we still had some unsolved problems with two left genes.

Cycle 3

Design

As we saw there were still some undetected problems. After further literature research, we found some alterations to this system [26]. Also, by this time we had constructed our kill-switch system, so we decided to insert it and superfolder green fluorescent protein (sfGFP) into the colicin gene. We used sfGFP to check how well the designed system is working and kill-switch VapXD system integration into genomic DNA is important for evaluation of how this system would perform in the final product.

Build

In this engineering cycle, we have changed the induction timing - L-arabinose to final 0.2 % concentration has been added at 0.2-0.3 OD600 and cells harvested by centrifugation then the OD600 has reached 0.6-0.8. Recombination template amount for kill-switch constructs, GFP - 400 ng, while for tyrP and adhE - 800 ng as previously there were many colonies resistant to both antibiotics but none of them insertion-positive.

Test

Both kill-switch systems and sfGFP constructs have been inserted into genomic DNA with almost 100 % accuracy leading to 7-12 insertion-positive colonies and just two survived colonies in the control plate. Meanwhile, adhE and tyrP genes knockout mutants still have not been generated. There were from hundreds to thousands of colonies in control plates, of those two mutant creation experiments. A bit smaller numbers occurred in plates where mutants were expected but not found.

Learn

The results of kill-switch and sfGFP insertion into the genome proved the system and chosen sgRNA for the colicin gene is working. However, just a bit of GFP fluorescence could be seen by the eye after blue light use. This might be because very little GFP is produced from only one DNA copy of this construct. Thus,urther investigation needed to be done.

The amount of colonies raised after several tries to generate tyrP and adhE genes knockouts lead to two possible explanations. Firstly, there was a huge possibility that chosen sgRNA is incapable to properly lead Cas9 endonuclease to a particular genomic site and causing double-strand break. This lead to almost all bacteria surviving even if they have Cas9 and gene-specific sgRNA inside. Second possible explanation would be that those two genes are too important for bacteria surveillance, so either it escapes double strand break occurrence, or it dies because of nonfunctional gene creation after successful genomic recombineering.

Cycle 4

Design

After taking earlier mentioned considerations together, we decided to design new sgRNA for adhE and tyrP genes to test whether our previously designed sgRNA are truly inefficient or can not be created mutants of tyrP and adhE genes knockouts. Moreover, as GFP levels from colonies with genomic insertion of its construct did not generate any clear fluorescence, we looked deeper into the pattern of genome wide E. coli transcriptional map [27]. From the literature and consultation with our advisor Inga Songailienė, we have chosen another genomic target for metabolic pathway insertion. nupG gene is responsible for broad-specific nucleoside permease synthesis.

Build

nupG specific sgRNA have been selected by CRISPOR [22] tool. Recombination template created by three steps PCR with long overhangs.

Test

GFP, linked 4CL-CHS, TAL and CHI genes have been inserted into nupG and colicin genes. tyrP, adhE knockouts have not been obtained.

Learn

From GFP fluorescence measurements we have learned that nupG it is transcriptionally more active. Also, we have found out that our designed primers pair for the third PCR round of recombination template creation for insertion into nupG gene are not suitable. These primers do not generate any PCR product as they probably do not anneal to the genomic DNA. This might be because of some nucleotides variation in the nupG gene in the Nissle genome. In order to evade this problem firstly for genomic editing chosen gene should be sequenced and only after that all over needed DNA constructs should be designed.