Designing flexible cloning cassettes for CYP enzymes

Modularity in construct design is important to allow the simple exchange of single parts for flexible assemblies of different related constructs. Particularly, for the design of our fusion protein it was essential to enable the exchange of cytochrome P450 and reductase enzymes, e.g. to analyse the effect of the linker sequence.

Level 1: Initially we designed our parts to be assembled by golden gate assembly for each construct (e.g. fusion protein expression cassette) individually. After assembling the first constructs we realised that to test our system we needed to exchange parts of the sequence. This was not really possible with our construct design and we had to assembly the whole construct again. Which made it difficult to exchange parts of the construct like promoters. Therefore we aimed to improve the modularity of our construct design.
Level 2: In a first step to allow more variability and easier combination of cytochrome P450 enzymes and reductases, restriction enzymes sites for MfeI (CAATG) and BamHI (GGATCC) were introduced at the N-terminus and C-Terminus of the cytochrome P450 enzymes and reductases respectively. Linkers were then created by annealing DNA oligos with AATG and GATC overhangs. Flexible linkers mainly containing glycine and rigid linkers mainly containing proline of different length were created. Introducing the MfeI and BamHI restriction sites to the construct leads to restriction in the flexibility of the linker sequence, since not all amino acids can be chosen freely. Two amino acids at the N-Terminus (Glu + Leu)) and two amino acids at the C-terminus (Gly + Ser) were fixed. This new design improved the ability of our system in creating fusion constructs but all our other parts were still assembled individually. There was no common assembly syntax specified so for each assembly new primers needed to be obtained.
Fig. 1: Linker design V 1.0. Expression cassette, MfeI and BamHI restriction sites
Level 3: To even improve the usefulness of our parts, we then aimed to make our parts compatible with the MoClo standard of golden gate based IIS restriction enzyme assembly. Thereby we expanded the Common Genetic Syntax for fusion sites to allow the creation of a) fusion proteins connected by linker sequences and b) multiple CDS expressed in an operon. All parts created according to this syntax are compatible with other MoClo collections like the Marburg collection. This advancement of the

Fig. 2: Common fusion site assembly syntax for our part design.

Designing a CYP1A1-Activity Assay

One of our projects’ goals was building a flexible platform for the development of CYP/CPR-fusion proteins. This requires a cheap, fast and reliable assay to assess the functionality of different fusion constructs. Since the CYP enzymes we used in our project require expensive substrates and produce difficult to detect products they are not suitable for this purpose and we needed to select a model enzyme. As a model enzyme we chose CYP1A1 which catalyzes the dealkylation of 7-ethoxycoumarin, producing 7-hydroxycoumarin (fig. 3). This substrate is an ideal model, as it is readily available and can be tracked by fluorescence measurements. Fig. 3: Convertion of 7-Ethoxycoumarine to 7-Hydroxycoumarine
Before testing our CYP1A1 constructs we first verified how well we could qualitatively and quantitatively measure both 7-ethoxycoumarin and 7-hydroxycoumarin. We started by measuring a dilution series of 7-ethoxycoumarin and 7-hydroxycoumarin in water, using wavelengths indicated in the literature [1] (exc: 390 nm; emi: 465nm) but observed no fluorescence. For our next test cycle we instead used the excitation wavelength (exc: 323 nm) indicated by the supplier [2] and used the plate-readers’ scanning mode to determine the emission spectra of 7-ethoxycoumarin and 7-hydroxycoumarin. The emission maxima from the emission spectra were then used for further assay development.
With our completed CYP1A1 control construct (BBa-K3846369) we looked to iteratively optimize our assay through multiple design-build-test cycles. Thanks to our opentrons pipetting robot we were able to combine these design-build-test cycles and perform them in parallel in 96-well plates. We investigated different starting concentrations of 7-ethoxycoumarin, different dilutions of the cell-suspension and different cell-lines. While using wild type E. coli as a negative control and cell-suspensions diluted with water as a blank. Our experiments showed rapid conversion of the substrates in both E. coli transformed with our constructs and wild-type E. coli. No differences in the conversion rate could be detected.
These results show that even after our optimization cycles the assay still does not achieve the required reliability. Clearly the bacteria can already metabolize 7-ethoxycoumarin on their own, though one could expect that additional enzymes with catalytic activity towards 7-ethoxycoumarin would result in a measurable increase in the conversion rates. There remain two possible explanations for the results. Either our constructs are non-functional, possibly due to misfolding during expression or low expression concentrations. Alternatively our constructs are functional but the assays’ signal-to-noise ratio is too low to resolve the difference in activity.
To eliminate issues with the assays’ signal-to-noise ratio, an in vitro assay using purified protein is the logical next step. For this purpose we added 6xHis-tags to our proteins, but so far were unable to obtain sufficient quantities of the purified proteins for the next design-build-test cycle.

Quantifying Terpenoid production

After optimizing our expression platform using CYP1A1 as a model substrate we will still need to verify if our actual terpene oxidizing CYP P450 enzymes work as expected and then quantify their output. As a measurement technique we settled on gas chromatography coupled mass spectrometry. This technique enables seperation of the terpenoid from complex metabolite mixtures and also enables accurate identification and quantification. For their identification, terpenoid spectra can be compared to spectra of authentic samples, which can either be obtained from the literature or measured on our own device. For the quantification gas chromatography peaks can be identified and integrated. These integrals can then be compared to the integrals of genuine analytical standards with known quantities.

Parallel approach to testing fusion constructs

There are two key aspects to consider when engineering novel metabolic pathways in microorganisms. First, the DNA sequences of the enzymes need to be considered. Second, the expression conditions need to be optimized.
When optimizing the DNA sequences of enzymatic fusion proteins the assembly, cloning and subsequent expression and purification can pose significant bottlenecks when working with sequential design-build-test cycles. Therefore we decided to take a partial parallel approach to planning design-build-test cycles. We designed a library of each of our CYP enzymes with their native CPR (if available), the CPR from A. thaliana and the BM3R domain, both as fusion constructs and as individually expressed constructs on one plasmid. Sadly we did not successfully implement the production of terpenoid backbones and therefore could not test the designs we created. When taking this project further towards a real world application there are additional parameters that could be included in the library design. Different linker designs and different truncation patterns of the CYP and CPR enzymes could be used. Due to the exponential growth of library size when considering multiple parameters these were not considered while pursuing a proof of concept in this project.
For each CYP-CPR fusion construct individual expression conditions need to be determined, while the linkers and truncation patterns should have little impact on the optimal expression conditions. The expression conditions are dependent on several factors which can be optimized in sequential screening cycles. The following conditions should be investigated: inducer concentration, optical density at induction, induction time before cell harvesting and temperature during protein expression. For each of these parameters multiple different conditions should be investigated and both protein concentration and product titer should be determined. After selecting the optimal condition it should be used while screening for the next parameter.

[1] C. Manjon, B. J. Troczka, M. Zaworra, K. Beadle, E. Randall, G. Hertlein, K. S. Singh, C. T. Zimmer, R. A. Homem, B. Lueke, R. Reid, L. Kor, M. Kohler, J. Benting, M. S. Williamson, T. G. E. Davies, L. M. Field, C. Bass, R. Nauen, Curr. Biol. 2018, 28, 1137-1143.
[2] Sigma-Aldrich, can be found under

© iGEM Hamburg 2021