Design

Intro

---

As explained in the project Description page, our team's mission is to biomanufacture decursin, for the purpose of decreasing its price, pave the way for it to become more widely used, and reach its full potential as a substance with an abundance of therapeutic properties, specifically to be used to treat CIA.

Umbelliferone, the fifth metabolite in the eight-step synthesis pathway of decursin in the plant Anglecia gigas (AG), is our starting point. From umbelliferone to decursin we need three enzymes; The first is Umbelliferone-6 prenyltransferase (U6PT), which mediates the reaction from umbelliferone to 7-demethylsuberosin (DMS). The second and third enzymes are XimD and XimE, which mediate the reaction from DMS to decursinol. And finally, through a simple esterification reaction on decursinol, we get decursin, our end product.

In this page we discuss in detail our circuit designs, and the thought process behind each one of them.

Umbelliferone-6 prenyltransferase (U6PT)

---

Prenyltransferases (PTs) are a class of enzymes that mediate prenylation reactions, in which an isoprenoid moiety, most commonly a prenyl group, is transferred from a donor molecule to an acceptor^[1]. In our project, we needed a PT with umbelliferone as the prenyl acceptor, and dimethylallyl diphosphate (DMAPP) as the prenyl donor, to yield our desired product, DMS, the sixth metabolite in the decursin anabolic pathway.

A study by Reddy et al.^[2] appraised for the first time the anabolic pathway of the major active pyranocoumarins in AG, and specifically decursin. However, they did not state the specific enzyme which catalyzes the prenylation reaction converting umbelliferone to DMS, as they only stated that the reaction was mediated by U6PT present natively in AG. Upon extensive review of the literature, we found four U6PT homologues, all from plant sources:

PcPT

Petroselium crispum prenyltransferase (PcPT) was identified in parsley leaves and was found to have a strong substrate specificity to DMAPP as a prenyl donor. It showed dominant prenylation activity on the C6 position of umbelliferone, yielding DMS, and a minor activity on the C8 position, yielding the unwanted product, osthenol^[3]. PcPT had been successfully expressed in a transient expression system in Nicotiana benthamiana by Munakata et al., who also claimed that this enzyme cannot be functionally expressed in yeast, nor bacteria, due to the fact that all plant PTs, including PcPT, are membrane-bound proteins, localized at the plastid membrane. Bu et al. refuted the former and managed to express PcPT in E. coli, showing successful prenylation with it^[4].

PsPT1

P. sativa prenyltransferase 1 (PsPT1) was identified in parsnip and was shown to favor the production of DMS over osthenol, when umbelliferone and DMAPP were introduced into the system^[5]. PsPT1 had been previously expressed in N. benthamiana, in a transient expression system, by Karamat F et al. However, upon further research in the literature, we could not find a successful attempt to express this enzyme in a bacterial system.

FcPT1

F. carica prenyl transgerase 1 (FcPT1) was identified in fig and was shown to have a high prenylation specificity to umbelliferone as the acceptor, and DMAPP as the donor^[6]. FcPT1 had been previously transiently expressed in an N. benthamiana system. We were unable to find in the literature, a successful attempt in cloning this enzyme's gene into a bacterial system.

UDT

Umbelliferone prenyltransferase (UDT) is a member of the PT family of enzymes. Its gene was mined from the genome of Glehnia littoralis, based on the local Blast alignment with homologous sequences as described by Song et al. ^[7].

As expressing PTs in bacterial or yeast systems was scarce in the literature, we turned to another approach. We designed two U6PTs using Blast sequence alignment and docking tools. Both U6PTs were designed using preexisting PTs that catalyze the reaction with either the right substrate and the wrong donor, or the right donor and the wrong substrate.

ChimeraI and ChimeraII

The desired reaction is the prenylation of umbelliferone on the C6 position using DMAPP as a prenyl donor. We discovered PT1a-Citrus limon ^[8] that was successfully expressed in yeast, which prompted us to attempt expressing it in bacteria. This enzyme uses the desired acceptor but not donor. We decided to identify the active prenylation site and change the donor binding site of PT1a for the purpose of increasing its affinity to DMAPP, the desired prenyl donor.

The UbiA domain is shared by a family of prenyltransferases, each catalyzing prenylation reaction ^[9]. For that reason, we deduced that this is the domain we should target. We used the Pfam database ^[10] to find the amino acids positions that constitute the UbiA domain and then turned to identify the specific site responsible for donor docking. We found two other enzymes, G4DT and PT1L, originating from soybean (Glycine max) ^[11] and hop (Humulus lupulus) ^[12], respectively, that were expressed in yeast and use the correct donor, DMAPP, with the wrong substrate acceptor. We again located the UbiA domains and used a Blast algorithm to compare the two sequences expecting to find homology due to a shared donor. We found a site consisting of 16 amino acids having high homology. We also noticed that in the middle of the motif are two aspartate residues, a commonplace feature in UbiA domains of other PTs. Reviewing the literature revealed that aspartates are crucial for the action of the prenyltransferase as they interact with Mg⁺² that is a cofactor for the enzyme ^[13]. All of the above led us to believe we identified the motif responsible for binding the desired donor (DMAPP). In order to identify the site in PT1a enzyme (Citrus limon) responsible for binding the donor, geranyl diphosphate, we used Blast to compare the UbiA of PT1a enzyme and PcPT. We saw several homologies and narrowed our search space for sequences containing aspartates. We found a smaller homologous sequence, which was expected due to the difference in the two enzymes’ donor molecule. We proceeded to assume that the motif size should be 16 amino acids long and identified the site to be the homologous part (10 amino acids) with an additional 6 amino acids upstream. 

We switched between the identified sites to create two chimera enzymes: ChimeraI - PT1a enzyme - Citrus limon (right substrate, wrong donor) with the donor binding site from PT1L - hop enzyme (wrong substrate, right donor)

ChimeraII - PT1a enzyme - Citrus limon (right substrate, wrong donor) with the donor binding site from G4DT - soybean enzyme (wrong substrate, right donor).

	Acceptor	Donor
PT1a - Citrus lemon	V	X
PT1L - hop	X	V
G4DT - soybean	X	V

Table.2-Donor, acceptor switch for ChimeraII:

	Acceptor site	Donor site
CimeraI	PT1a- Citrus lemon	PT1L- hop
ChimeraII	PT1a- Citrus lemon	G4DT- soybean

We predicted a proper folding of our enzymes using Phyre2^[14] and tested the chimeras' capability to bind the right donor, DMAPP, using Dockthor^[15]. The results showed that our chimeras bind the prenyl group even better than the original enzyme bind the geranyl group.
For Blast results and docking parameters ( see appendix).

U6PT genetic circuit

With four native enzymes and two synthetic chimeras as candidates, we proceeded to design the genetic circuits for these enzymes, to be expressed in E. coli. First and foremost, we optimized the codons of all PT candidates to E. coli using Integrated DNA Technologies -IDT codon optimization tool. All PTs were ordered as DNA fragments except for PcPT which was amplified from a codon-optimized plasmid that was used in E. coli^[4]. We decided to express the enzyme under the rhlR-regulated promoter. Downstream to the U6PT gene we added a P2A sequence, a self-cleaving peptide element which induces ribosomal skipping during translation of a protein in the cell, upstream to an mCherry reporter gene. In such a manner, the mCherry expression is tied directly to the start codon of the U6PT candidates while not being covalently bound to the candidates themselves. As P2A has been shown to have the highest cleavage efficiency compared to other 2A sequences^[16], an mCherry signal corresponds to a 1:1 ratio, at the very least, which allows for a lower estimate of the candidates' expression.
The A133_rhlR_tdPP7_mCherry plasmid we used as a backbone was a gift by Prof. Roee Amit's lab^[17].

Plant aromatic PTs all belong to the UbiA superfamily of intramembrane PTs. Specifically, U6PTs are members of this family, which catalyze prenylation reactions with phenolic compounds as their substrates. Those enzymes possess similar characteristics, such as the two conserved aspartate-rich motifs which bind Mg⁺² resulting in the stabilization of the diphosphate moiety of the prenyl donor. Additionally, they all contain between 6 to 9 transmembrane helices, and possess a transit peptide sequence at their N-terminus region. Which leads us to the topic of transit peptides.

Transit peptides

A transit peptide (TP) is a short peptide sequence, ranging between 16 to 30 amino acids long, that is usually present at the N-terminus of most newly synthesized proteins. This peptide sets the fate of the protein towards a certain organelle or the secretion pathways, or aids in the post-translational insertion of the protein into cellular as well as intracellular membranes.

As previously mentioned, U6PTs all contain a TP located at their N-terminus region. Multiple literature sources showed that expression of plant PTs in yeast systems without the TP sequence (truncated form) increases the efficiency of their expression and thus yields better metabolic results^[8][18], however the reason for this is not fully understood.

Assuming adequate similarities between yeast and bacterial expression systems in comparison to plant expression systems, we will express the six PT candidates, without their TPs (delta TP-U6PT), and compare the results between the original and the truncated form of each enzyme.

XimD & XimE

---

The XimD enzyme together with XimE enzyme are responsible for turning 7-demethylsuberosin (DMS) into decursinol. XimD is a flavin-dependent monooxygenase that mediates the oxygenation of DMS to give an epoxide intermediate. With the presence of XimE enzyme, which is a SnoaL-like cyclase, this intermediate undergoes cyclization to give decursinol^[19].The genes of both these enzymes are cloned into one plasmid.

Xim example animation

XimD & XimE genetic circuit

When designing the genetic circuit for XimD and XimE, we wanted to incorporate them in a regulatory system. The aim was to have each one of the two enzymes under a different regulation, XimD regulated by the Lac system, and XimE regulated by the Tet-On system. That way we can control the expression ratio between the two xim enzymes.

According to this approach, we designed the genetic circuit to have XimD and XimE , each one under T7 promoter, with LacO or TetO placed between the promoter and XimD or XimE, respectively.

To complete the regulatory system, the repressor genes were added to the design as independent fragments, each one with its own promoter. pJ23119 promoter for TetR, and pLac promoter for LacI. And with that our design ended.

Relating on our modeling results, ximE enzyme should be expressed constitutively to approach a higher bio-manufacturing yield of the desired metabolite - decursinol (To read more about the impact of the model on the designs click here). So, to implement these results in our design, the endmost cloning step of the TetR repressor was denied (see fig.5-ximD & ximE final genetic circuit).

Final system- Decursinol Bio-Production Machine

---

Both circuits mentioned above (fig.2 and fig.4) will be transformed into E. coli BL21 DE3 strain (figure.6). This specific strain was chosen due to its ability to express T7 polymerase that employs the T7 promoter for transcription. Each plasmid has a different origin of replication and antibiotic resistance, ensuring the maintenance of both plasmids after selection.

Cloning Strategy

---

In the beginning stages of our project, when we were reviewing the literature on the topic of decursin, we saw that previous efforts were made to introduce the synthesis pathway of plant pyranocoumarins into microbial systems, and amongst those pyranocoumarin was decursin ^[4]. However, due to the fact that we were interested specifically in decursin, this is what we focused on. As the purpose of our project is biomanufacturing decursin on a large scale, we decided to use the information acquired from previous literature as our starting point and build upon it by focusing our efforts on optimizing the system using modelling tools and software. Thus, pETDuet-PcPT-XimD-XimE plasmid was utilized for this motive. It was the backbone for one of our genetic circuits; the Xim design.

U6PT

XimD & XimE