Synthetic biology is an emerging engineering discipline, focused on leveraging the intricate workings of biological systems in order to address contemporary challenges, spanning from climate change to critical medical phenomena. What sets synthetic biology apart from traditional biology is its emphasis on the engineering process, encompassing the designing, building, testing, and iterative learning from these created biological systems.

During our time in the lab, we successfully completed several Design-Build-Test-Learn (DBTL) cycles. These cycles involved addressing various challenges we faced, ranging from basic troubleshooting of standard procedures like clonings to more intricate aspects related to our overall design. On this page, we will showcase our engineering process and the impact it had on our final design.

To build our constructs, our team followed the GoldenBraid 2.01, an advanced and highly efficient cloning method widely used in synthetic biology and genetic engineering. This method involves several key steps, such as the domestication of the DNA parts to ensure compatibility with the GoldenBraid standard grammar and the assembly of parts into higher-level constructs, using universal domesticator plasmids (in our case pUPD2) and destination vectors (in our case pDGB3α1, pDGB3α2 and pDGB3ω1). One of its notable features is the ability to create multi-level constructs, facilitating the construction of intricate genetic circuits. The different levels of GoldenBraid correspond to different stages of construct assembly.

• Level 0: In this level, individual DNA parts are inserted into universal domesticator plasmids and prepared for higher level assembly. For that to happen a process known as domestication must occur, which involves modifying these DNA parts to remove unwanted internal restriction sites and add specific overhangs compatible with the GoldenBraid grammar. Level 0 parts are typically inserted into universal domesticator plasmids, creating what is known as "level 0 constructs." These constructs serve as the building blocks for higher-level assemblies.
• Level 1: Level 1 (alpha plasmids) involves the assembly of multiple level 0 constructs to create complete transcriptional units.
• Level 2: level 1 modules are combined to create more intricate genetic circuits or systems. Specifically, level alpha plasmids can be combined with an omega destination vector to create a construct with 2 transcriptional units.

The PhaJ gene encodes the enzyme trans-enoyl-CoA hydratase, an enzyme that catalyzes the linking step between β-oxidation and the PHA producing machinery2 (see Design page for more details). The overexpression of this enzyme in P. putida would facilitate the efficient conversion of the free fatty acids derived from OMW or E. coli, shifting them away from the β-oxidation pathway and redirecting them towards the PHA biosynthetic pathway3, making its efficient regulation essential for our system. For this reason, we decided to develop a free fatty acid (FFA)–inducible system for phaJ expression4, based on FadR, an Escherichia coli transcription factor regulated by fatty acyl-CoAs and the fadBA promoter (PfadBA), containing a FadR-recognition site5 ( Fig. 1.)

Fig. 1. Schematic representation of the PfadBA-FadR system. In absence of free fatty acids ( FFAs), the consistently produced FadR protein inhibits the function of the PfadBA promoter, suppressing the phaJ expression. When FFAs are present, they are activated to acyl-CoAs by acyl-CoA synthase (encoded by fadD). Acyl-CoAs bind to theFadR protein, allowing the PfadBA promoter to function properly, leading to the transcription of the phaJ gene.

Cycle 1 - Characterization of the fadBA promoter

Designing

Since we could not find enough data characterizing the fadBA promoter to an extend that could help us with our decision making, we decided to attempt a basic characterization, by firstly designing a construct with the PfadBA upstream of a reporter gene encoding the SYFP2 (BBa_K864100), a monomeric fluorescent protein with optimized folding, maturation and a narrow fluorescence emission spectrum6, contained in this year’s Distribution kit ( Fig. 2).

Fig. 2. Schematic representation of the level 1 (alpha) construct we designed for the characterization of the PfadBA.

Additionally, we designed a level 2 construct with two transcriptional units; the level 1 construct described before and an additional transcriptional unit of the fadR under the control of a constitutive promoter ( Fig. 3) to evaluate the level of inhibition the protein FadR exerts on the promoter.

Fig. 3. Schematic representation of the level 2 (omega) constructs we designed for the evaluation of the inhibition fadR imposes on the promoter PfadBA .

However, due to time constraints, the whole level 2 (omega) construct ( Fig. 3), was not assembled. So, we decided to test only the reporter transcription unit ( Fig. 2), in the E. coli BL21 (DE3) strain. This decision was also based on the fact that FadR is a transcription factor found natively in Escherichia coli 7, so its inhibitory effects could still be noticeable on the fluorescence output of the construct. The sequence of PfadBA was acquired and then domesticated, using the GoldenBraid Domesticator tool, which removes any internal restriction sites that did not comply with the GoldenBraid standard and adds the appropriate 4-nt 3’ and 5’ flanking overhangs in order for the inserts to be compatible with our level 0 pUPD2 cloning vector.

Building
After the domestication process, the part was ordered and synthesized by IDT. When received, the DNA was resuspended and the cloning procedures began. To get to the desired reporter construct, we began by cloning the PfadBA into a level 0 cloning vector, pUPD2. Through digestion-ligation reaction, DH5α transformation, plasmid isolation and restriction-digestion confirmation ( Fig. 4), we successfully constructed the level 0 construct.

Fig. 4. Diagnostic digestion of pUPD2_pfadBA with EcoRI and EcoRV, expected bands (bp): 1309 and 896. Lane 2: pUPD2 (no insert).

The level 0 construct was then combined, via a digestion-ligation reaction, with the Distribution kit’s level 0 constructs of the B0030 RBS (BBa_J428032), the syfp2 (BBa_K864100) and the B0015 double terminator (BBa_J428092), in order to build the complete transcriptional unit that we could then test. We successfully built the level 1 (alpha) construct, which was confirmed with a restriction digestion reaction ( Fig. 5).

Fig. 5. Diagnostic digestion of pDGB3α1_pfadBA-syfp-rrnB T1/T7TE with BsaHI, expected bands (bp): 2327, 2026, 1628 and 1291. Lane 2: pDGB3α1 (no insert).

Once we had the complete construct, we transformed E. coli BL21 (DE3) chemically competent cells with the isolated plasmid. To further confirm that our insert was correct and functional, a single colony was picked, inoculated in LB medium and cultured O/N at 37οC and 210 rpm. The next morning, 5 x 200 ul of the culture were loaded into a black plate with transparent bottom (with the appropriate blanks) and then a fluorescence intensity measurement was taken at wavelengths of 511 nm (excitation) and 529 nm (emission). The tested culture showed distinct fluorescence when compared to the negative control (non-transformed E. coli BL21 (DE3)), so we knew that the construct was indeed functional.

Testing

To adequately characterize PfadBA , we included two controls: 1. a positive control construct which we designed and built, carrying the same reporter gene (syfp2) under the control of the well-documented constitutive Anderson promoter J23118, and 2. non-transformed BL21 (DE3) cells as a negative control. The detailed protocol we followed can be found on our Experiments page. In short, single colonies were picked for both constructs, as well as a colony of non-transformed BL21 (DE3) cells and inoculated into 5 ml of LB with the appropriate antibiotic. The cultures were grown O/N at 37οC and 210 rpm. The next day, a dilution was performed to reach an OD600 of approximately 0.02 and a black plate with a transparent bottom was prepared, blanks included. The construct was tested in three different conditions: with the addition of non-dissolved oleic acid, addition of oleic acid dissolved in DMSO8,9 and without oleic acid. Four biological repeats were performed for each condition, the plate was placed into the plate reader and measurements of absorbance and fluorescence were taken after 6h incubation. The results we got are shown below ( Fig. 6).

Fig. 6. Normalized fluorescence intensity measurement for pfadBA-syfp-rrnB T1/T7TE construct on three different conditions ( No oleic acid addition, addition of non-dissolved oleic acid, addition of oleic acid dissolved in DMSO) after 6h incubation.

Learning
From the data we got, we noticed that the PfadBA resulted in low expression levels of the SYFP2 protein when compared to our positive control, the PJ23118 Anderson promoter, a result for which we were previously alerted to by Dr. Tsampika, suggesting that a promoter with higher expression rates would be better suited for the adequate transcription of the phaJ gene since the low expression levels could hamper the efficiency of our system. So, we set out to repeat the DBTL cycle in search of a better suited promoter for the phaJ.

Cycle 2 - A more suitable promoter for the phaJ gene

Designing
Going back to the designing stage, we searched for alternative promoters, regulated by the FadR protein. Utilizing the iGEM registry and after thorough literature review, we discovered two synthetic promoters,PLR andPAR5. These promoters had previously been characterized as having higher relative strength than PfadBA. Our objective was to evaluate and compare their intensity using the same reporter gene and compare them to the same positive control. To achieve this, we designed two additional reporter constructs, one for PLR and the other for PAR, both regulating the expression of syfp2 (Fig. 7). Again, the promoter sequences were first domesticated for the removal of internal restriction sites not compatible with the GoldenBraid cloning method, as well as the addition of the 3’ and 5’ overhangs.

Fig. 7. Schematic representation of the standard construct for the characterization of PLR and PAR .

Building

The same cloning strategy was followed, as described above. The -1 domesticated part sequences were successfully inserted into a pUPD2 vector ( Fig. 8).

Fig. 8. Diagnostic digestion of (1) pUPD2_pLR and (2) pUPD2_pAR with EcoRI and EcoRV, expected bands (bp): (1) 1300 and 896, and (2) 1305 and 896. Lane 3: pUPD2 (no insert).

Then, the level 0 parts were used to build the 2 new reporter constructs. Finally, both constructs were built and the inserts were confirmed by restriction-digestion reactions (Fig. 9).

Fig. 9. Diagnostic digestion of (1) pDGB3α1_pLR-syfp-rrnB T1/T7TE and (2) pDGB3α1_pAR-syfp-rrnB T1/T7TE with BsaHI, expected bands (bp): (1) 2327, 2017, 1628, 1291 and 58, and (2) 2327, 2022, 1628, 1291 and 58. Lane 3: pDGB3α1 (no insert).

The isolated plasmids were used to transform E. coli BL21 (DE3) chemically competent cells. To further ensure that the correct functional insert was built, we performed a preliminary fluorescence intensity measurement, as in the case of PfadBA. Both constructs exhibited significant fluorescence levels, so we concluded that the clonings were indeed successful.

Testing

Having all the devices transformed into the same E. coli strain, we repeated the experiment with the same conditions. The results we got are shown below ( Fig. 10).

Fig. 10. Normalized fluorescence intensity measurement for pLR-syfp-rrnB T1/T7TE and pAR-syfp-rrnB T1/T7TE constructs on three different conditions ( No oleic acid addition, addition of non-dissolved oleic acid, addition of oleic acid dissolved in DMSO) after 6h incubation.

Learning

Comparing the two new promoters to the PfadBA and the PJ23118 we can clearly see the elevated levels of fluorescence intensity ( Fig. 11). However, as in the first cycle, no significant difference can be seen with the addition of oleic acid. We assume that if we were to test the complete level 2 construct ( Fig. 3), the presence of the additional FadR regulator would cause greater downregulation of the system in the absence of the inducer. However, we must mention that additional experiments are required to clarify which of the two promoters is the most suitable for our system. Considering those results, we deemed the second engineering cycle as a success, since our initial goal to find a higher intensity fatty acyl-CoA-responsive promoter, was achieved.

Fig. 11. The results of both engineering cycles for all fatty acyl-CoA-responsive promoters in a common chart.

Future

Having acquired these data, to decide on the appropriate promoter for our design, we would have to build two level 2 (omega) constructs, each containing one of the promoters regulating the reporter gene and another transcription unit containing the fadR under the control of the AraC/PBAD system. These two constructs would eventually be transformed into P. putida and then be tested with the addition of oleic acid and a gradient of concentrations of L-arabinose. The ideal promoter would be the one with lower levels of leakiness combined with high induced gene expression at a concentration of oleic acid that is similar to the levels found in OMW. Due to time constraints, we were not able to perform this experiment, but we hope that a future team will be able to.

The use of E. coli for the secretion of a functional laccase enzyme would take us a step further towards the holistic detoxification of the OMW10,11 ( Fig. 12). Although we performed experiments to evaluate the activity of the enzyme (see Experiments), we mainly focused on tackling the secretion aspect in an engineering cycle, by reviewing the literature and actually building the constructs to test the secretion of the laccase, using the native peptide from Trametes versicolor as well as 3 different well-documented signal peptides, NSP4 (BBa_K3606042), PelB (BBa_K208004) and MalE (BBa_K1012004) in search of a more efficient secretion system11,12.

Fig. 12. The laccase secretion system in E. coli. The laccase is under the regulation of the araBAD promoter and is secreted from the cell (pac-man), thanks to the fusion with a signal peptide, suitable for protein secretion in bacterial expression systems.

Cycle 1 -Testing the original secretion peptide.

Designing

Our first objective was to test the native signal peptide (Na.SP), originating from Trametes versicolor, since, belonging to different biological kingdoms, we could not be sure about its functionality in E. coli. To test its functionality, we designed a construct where the coding sequence of the laccase was fused with the sequence of the fluorescent protein SYFP2 at the 3’-terminal ( Fig. 13). This sequence was then domesticated to the GoldenBraid standards and ordered to be synthesized by IDT .

Fig. 13. Schematic representation of the level 1 (alpha) construct of the laccase fused with syfp and carrying its native peptide.

Building

Once the level -1 DNA fragment arrived, we began the process of cloning the insert into the pUPD2 vector, turning it into a level 0 construct. Once this procedure was successful, with confirmation from restriction-digestion reaction and gel electrophoresis ( Fig. 14), the isolated plasmid was combined by digestion-ligation reaction with the level 0 constructs syfp2 and B0015 double terminator provided by this year’s iGEM distribution kit, and the AraC/PBAD regulatory system we had previously constructed. The complete construct was successfully built, verified by restriction-digestion reaction and gel electrophoresis. ( Fig. 15).

Fig. 14. Diagnostic digestion of pUPD2_Na.SP-laccase with EcoRV and NotI, expected bands (bp): 1562, 1143 and 903. Lane 2: pUPD2 (no insert).

Fig. 15. Diagnostic digestion of pDGB3α1_pBAD-Na.SP-laccase-syfp-rrnB T1/T7TE with BsaHI, expected bands (bp): 2327, 2049, 1628, 1361, 1291, 610, 423, 185, 163, 81 and 58. Lane 2: pDGB3α1 (no insert).

The isolated plasmid was transformed into E. coli BL21 (DE3) chemically competent cells and to further ensure that the insert is functional we performed the diagnostic fluorescence measurement, as described previously. The tested samples exhibited significant fluorescence levels which further confirmed the insertion of the functional construct.

Testing

The detailed protocol we followed can be found on our Experiments page. In short, a single colony carrying the level 2 (omega) construct was used to inoculate 5 ml of LB medium, with the appropriate antibiotic, and the culture was grown O/N at 30οC and 160 rpm. E. coli BL21 (DE3) cells with the construct containing the syfp2 under the control of the J23118 Anderson promoter and non-transformed E. coli BL21 (DE3) cells were used as negative controls. The next day, O/N cultures were diluted in order to reach the same OD600 and the addition of L-arabinose to a final concentration of 10 mM followed. Finally, after a 6h incubation the cultures were centrifuged in order to measure absorbance and fluorescence of the supernatant and the pellet fraction, after cell pellet resuspension14. Four biological repeats were performed for each condition, the plate was placed into the plate reader and measurements were taken. The results of the normalized fluorescence intensity measurements for each fraction are depicted in Fig. 16.

Fig. 16. Normalized fluorescence intensity of the supernatant and pellet fraction for pBAD-Na.SP-laccase-syfp-rrnB T1/T7TE construct after 6h incubation with 1,5% arabinose, (-) control 1: pJ23118-syfp-rrnB T1/T7TE, (-) control 2: non-transformed cells.

Learning

From those results we can conclude that the signal peptide from Trametes versicolor is not suitable for laccase secretion from E. coli, since supernatant fluorescence intensity exhibited low values compared with the negative controls. Thus, we decided to go through another cycle, in order to find and test additional, well characterized signal peptides for protein secretion in bacterial expression systems.

Cycle 2 -Testing signal peptides NSP4, PelB and MalE

Designing

After a thorough evaluation of the related literature, the 3 signal peptides NSP415, PelB16 and MalE17 were selected to test with the laccase. To evaluate the secretion efficiency of these peptides, we substituted the sequence of the native signal peptide of the laccase with each of the three peptides. These 3 coding sequences were then fused with the sequence of the syfp as illustrated in Fig. 17.

Fig. 17. Schematic representation of the standard level 1 (alpha) construct of the laccase fused with the syfp and carrying one of the selected signal peptides (NSP4, PelB and MalE).

Building

Once we designed the constructs, the sequences of the 3 domesticated signal peptides with the appropriate overhangs were ordered from IDT. The cloning procedures began with the insertion of the -1 parts into the pUPD2 vector, to create the level 0 constructs. Once successfully built, with confirmation by restriction-digestion reaction ( Fig. 18 and 19), the isolated level 0 parts of the signal peptides were combined with the level 0 parts for the laccase, the syfp2, the B0015 double terminator and the AraC/PBAD , to build the complete constructs ( Fig. 17).

Fig. 18. Diagnostic digestion of (1) pUPD2_NSP4 and (2) pUPD2_PelB with EcoRV and NotI, expected bands (bp): (1) 1143, 903 and 122 and (2) 1143, 903 and 128. Lane 3: pUPD2 (no insert).

Fig. 19. Diagnostic digestion of pUPD2_MalE with EcoRI and EcoRV, expected bands (bp): 1290 and 896. Lane 2: pUPD2 (no insert).

Following the standard protocol for digestion-ligation reaction, we were able to build the complete transcriptional units, again confirmed by restriction-digestion reaction and electrophoresis ( Fig. 20).

Fig. 20. Diagnostic digestion of (1) pDGB3α1_pBAD-NSP4-laccase-syfp-rrnB T1/T7TE, (2) pDGB3α1_pBAD-PelB-laccase-syfp-rrnB T1/T7TE and (3) pDGB3α1_pBAD-MalE-laccase-syfp-rrnB T1/T7TE with BsaHI, expected bands (bp): (1) 2327, 2049, 1628, 1361, 1291, 625, 423, 185, 163, 81 and 58, (2) 2327, 2049, 1628, 1361, 1291, 613, 423, 185, 163, 81 and 58 and (3) 2327, 2049, 1628, 1361, 1291, 625, 423, 185, 163, 81 and 58. Lane 4: pDGB3α1 (no insert).

Testing

We once again began by transforming all 3 test devices into bacteria of the same E. coli strain and we repeated the same steps in order to measure absorbance and fluorescence of both the supernatant and the pellet fraction. The protocol we followed for this experiment is available in detail in Experiments page. The results are depicted in Fig. 21. It can be seen clearly that NSP4 consistently gave the best results followed by the signal peptide MalE, while PelB and the native signal peptide constructs gave the lower secretion output.

Fig. 21. Normalized fluorescence intensity of the supernatant and pellet fraction for pBAD-NSP4-laccase-syfp-rrnB T1/T7TE, pBAD-PelB-laccase-syfp-rrnB T1/T7TE and pBAD-MalE-laccase-syfp-rrnB T1/T7TE constructs after 6h incubation with 1,5% arabinose, (-) control 1: pJ23118-syfp-rrnB T1/T7TE, (-) control 2: non-transformed cells.

Learning

After processing the results, we concluded that NSP4 signal peptide might be the most suitable for our purposes, since both supernatant fluorescence intensity value and supernatant fluorescence intensity per pellet fluorescence intensity value were the highest among the tested signal peptides (Fig. 22).

Fig. 22 . The results of both engineering cycles, for the normalized fluorescence intensity of the supernatant and the pellet fraction, in a common chart.

After completing two DBTL cycles, each comprising two iterations, we have gained a significantly deeper understanding of the oPHAelia system. This progress brings us closer to envisioning the promising future it may hold.