Design

Selection of Chassis and Plasmid

Our goal for this project was to genetically modify an organism to produce chocolate compounds in order to combat the climate and labor issues related with chocolate production. When we originally researched the pathways needed to create these compounds, we were debating between choosing E. coli or Saccharomyces cerevisiae, as many of the papers we read had successfully created these compounds in those species.

However, we ended up choosing Lactococcus lactis cremoris (CCUG 21954, ATCC 9625, NCTC 7019, LMG 7932, IFO 3427) as the host organism when we found that one of our compounds could be produced in L. lactis, specifically 4-Hydroxy-2,5-dimethyl-3(2H)-furanone (HDMF) (Hayashida et al. 2001). Furthermore, we chose L. lactis because it has a long-standing reputation for being food-safe, producing many of the dairy products we enjoy such as cheese, butter, and buttermilk (Madera et al. 2003). Additionally, the pathways for the production of other chocolate compounds documented in E. coli had common precursors presumed to also be present in L. lactis (Xu et al. 2020).

We chose our plasmid, pNZ8148, because nisin, which induces our promoter, is also commonly used in the food industry (Zhang et al. 2024). Nisin is a broad-spectrum bacteriocin produced by L. lactis used extensively as a food preservative (O’Reilly et al. 2023). The combination of pNZ8148 and the use of L. lactis, on top of the purification of our compound, would ensure that our product would pose a lower risk to human consumption.

Diagram of pNZ8148 Vector
Fig. 1: a diagram of our chosen pNZ8148 plasmid, containing the PnisA promoter, chloramphenicol resistance, a repC and a repA region.

Design of Pathways

We created four pathways for the production of four chocolate compounds: 2-Phenylethylamine, theobromine, 4-Hydroxy-2,5-dimethyl-3 (2H)-furanone (HDMF) (also known as the trademark name Furaneol), and 3-ethyl-2,5-dimethylpyrazine. Each coding region is preceded by a ribosome binding site (either BBa_B0032 or BBa_J34801), which is further preceded by the inducible nisin promoter (BBa_K4307009). Downstream of the coding region is the terminator found inside of the original pNZ8148 plasmid (Accession CP068658.2, Region 399302 - 399354).

Synthesis pathway for 2-phenylethylamine
Fig. 2: Pathway for the production of 2-phenylethylamine. Phenylalanine decarboxylase (PDC) is an enzyme which catalyzes the production of 2-phenylethylamine and carbon dioxide from the amino acid phenylalanine. The PDC coding region is placed in between the nisin promoter, which is inducible with the biochemical nisin, and a terminator (Xu et al. 2020).

Even though there’s only one PDC pathway, we had three different amino acid sequences which all coded for PDC from three different organisms: PDC from Psychroflexus gondwanensis, Psychroflexus torquis, and Solanum lycopersicum.

Synthesis pathway for theobromine
Fig. 3: Pathway for the production of Theobromine. The enzyme 7-methylxanthosine synthase (XMT) catalyzes the reaction of xanthosine into 7-methylxanthosine. The coenzyme S-adenosyl-L-methionine (SAM) is also turned into S-adenosyl-L-homocysteine (SAH) in the function of the XMT enzyme. Once creating 7-methylxanthosine, XMT can once turn 7-methylxanthosine into 7-methylxantine using hydrolysis to split 7-methylxanthosine and water into 7-methylxantine and a ribose sugar. Upstream of the XMT coding region are the nisin promoter and a RBS. Before the next coding region, a 15 base pair spacer and another RBS are attached. The next coding region is the enzyme theobromine synthase (MXMT) which can turn 7-methylxantine into theobromine using the same SAM coenzyme. Downstream of the MXMT coding region is the terminator (Leibrock et al. 2022). *note: while only XMT and MXMT enzymes are required to synthesize theobromine, we accidentally included an extra DXMT enzyme which would have turned our theobromine into caffeine. We are working on fixing this issue in the future.
Synthesis pathway for HDMF
Fig. 4: Pathway for the production of 4-Hydroxy-2,5-dimethyl-3 (2H)-furanone (HDMF). Fructose-bisphosphate aldolase (FBA) splits fructose-1,5-bisphosphate into the triose phosphates dihydroxyacetone phosphate and glyceraldehyde 3-phosphate. Dihydroxyacetone is the preferred product, as it can be turned into furaneol by a not-fully understood metabolic pathway (Pérez et al. 1999). The triose-phosphate isomerase (TPI) enzyme further increases the yield of furaneol by creating more dihydroxyacetone phosphate from glyceraldehyde 3-phosphate. The nucleotide sequence that codes for these enzymes is structured very similarly to previous pathways. The nisin promoter precedes, with an RBS, the first coding region (FBA), a 15-base pair spacer, a second RBS, the second coding region (TPI), and a terminator.
Synthesis pathway for pyrazine
Fig. 5: L-threonine 3-dehydrogenase (TDH) converts the amino acid threonine to 2-amino-3-ketobutyrate (AKB) and AKB is metabolized to glycine by the acetyltransferase activity of 2-amino-3-ketobutyrate coA ligase (KBL) in usual conditions. However, in poor nutrition, at low CoA concentrations, and when the enzyme is overloaded, unstable AKB is gradually decarboxylated to aminoacetone (red) and the lyase activity of KBL decomposes threonine to acetaldehyde and Gly. 3-ethyl-2,5-dimethylpyrazine is synthesized chemoenzymatically via the condensation reaction of two molecules of aminoacetone and one molecule of acetaldehyde. Despite 3-ethyl-2,5-dimethylpyrazine not being the major product of the reaction, yields have been reported as high as 20.2% under normal conditions (not explicitly creating conditions with poor nutrition or low CoA concentrations) (Motoyama et al. 2021). The structure of the nucleotide is as follows: the nisin promoter precedes, with an RBS, the first coding region (TDH), a 15 base pair spacer, a second RBS, the second coding region (KBL), and a terminator.

In addition to these pathways, the nisin signal transduction pathway found in our nisin promoter is as follows:

NICE system
Fig. 6: The nisin-controlled gene expression (NICE) system in L. lactis. Nisin first binds to NisK, a membrane-bound receptor protein. NisK autophosphorylates and transfers a phosphate group to NisR, which activates the protein. Once activated, NisR can interact with the pNisA promoter, turning on expression of the gene. The gene is transcribed into mRNA, where it can then be translated into our chocolate-producing enzymes via free-floating ribosomes (Bakari et al. 2014).

Build

We successfully assembled our vectors inside of pNZ8148. We did this through HiFi assembly. We first designed primers specific to each pathway which would amplify pNZ8148 with a 5’ overhang homologous to the 5’ end of the coding region. For example, for the TDH-KBL pathway (Fig. 5) a reverse primer would have homology with the 5’ end of the TDH coding region and a forward primer would have homology with the 5’ end of the KBL coding region (see Fig. 7).

Primer diagram
Fig. 7: A diagram showing the positioning of primers. A reverse primer amplifies pNZ8148 and includes a 20-base pair 5’ homology with the start of the first coding region. A forward primer would do the same at the end of the KBL gene in the TDH-KBL pathway.

This 20 base pair overhang is crucial for HiFi assembly. 3 enzymes constitute HiFi assembly: 5’ exonuclease, DNA polymerase, and DNA ligase. 5’ exonuclease chews back more of the 5’ end strand to create a longer overhang so that the two fragments anneal together, DNA polymerase fills in the gaps, and DNA ligase seals any nicks and covalently links the DNA fragments together.

In order to prepare the PCR products for HiFi assembly for higher efficiencies, two additional steps were taken: DpnI to digest methylated DNA, and column purification to further purify our DNA.

Once the plasmid was constructed, we first cloned the plasmid inside of E. coli. Then, we could extract the plasmid and transform it into our final host organism, L. lactis. Only then could we induce the plasmid by adding nisin to the media and start producing our chocolate compounds.

Test

Unfortunately, we had many issues with cloning, both in E. coli and L. lactis.

We first tried cloning inside of DH5a E. coli cells. While each gene construct ended up growing up on an LB + chloramphenicol plate (Fig. 8), when we performed colony PCR, hardly any colonies were positive for our gene insert (Fig. 9).

Plated transformed colonies
Fig. 8: Colonies of our transformed genes in dh5a cells. We see visible colonies in genes 1, 2, 3, 4, 6, and 7 in diagrams a), b), c), d), e), and f), respectively. Gene 5 is omitted due to a labeling error early on. All of our genes have colonies.
Result of colony PCR on transformed cultures
Fig. 9: Colony PCR results. 7 colonies were picked from each of our 6 gene constructs. Gel a) has gene #1 on the left of the middle ladder, with gene #2 on the right of the middle ladder. Gel b) has gene #3 on the left and gene #4 on the right. The left gel in diagram c) is gene #6, and the right gel is gene #7. The genes correspond to: 1) SlPDC, 2) XMT-MXMT-DXMT, 3) TDH-KBL, 4) PtPDC, 6) FBA-TPI, 7) PgPDC.

We had successful insertion and cloning of gene #2 (five out of seven with a positive result), gene #3 (one out of seven), and #4 (one out of seven). However, we also learned that our plasmid required cloning inside of recA+ strains of E. coli, and DH5a competent cells lack the recA protein. The lack of recA results in plasmid instability and the potential deletion of the chloramphenicol resistance gene (NICE Expression Handbook and Volff et al. 1997). Therefore, we moved all future cloning attempts to be strictly inside of MC1061 cells.

Plated clones of transformed cells
Fig. 10: Cloning results in MC1061 cells. Only three genes grew: XMT-MXMT-DXMT (a), TDH-KBL (b), and FBA-TPI (c).
Result of colony PCR on clones
Fig. 11: Colony PCR results for our MC1061 and DH5a cloning. The top 2 gels in (a) as well as the first 6 non-ladder lanes on the bottom left gel were taken from 20 DH5a PtPDC colonies. No colonies show positive results. The last lane in the bottom left gel in (a), the bottom right gel in (a), and the first 7 lanes in the top gel in (b) show positive results for 15/15 colonies MC1061 colonies with the FBA-TPI gene. The rest of the lanes in (b) indicate negative colonies for DH5a genes FBA-TPI and PgPDC. The left gel in (c) shows the gene XMT-MXMT-DXMT in MC1061. We hypothesize that the gel either punctured or the DNA stain wasn’t mixed properly which resulted in the empty middle. The right gel in (c) is the TDH-KBL gene in MC1061.

The colony PCRs show significant improvement of MC1061 compared to DH5a. While DH5a resulted in 0/45 colonies, MC1061 had positive results in every single lane (excluding the mishap on the left gel in Fig. 11 (c)). All of our positive colony PCR results in this latest round came from MC1061, indicating plasmid stability inside of MC1061.

However, cloning inside of MC1061 was far from perfect. Despite attempting transformation six different times, only one attempt resulted in colonies that grew. This highlights our underlying issues with the replication of our plasmid inside of E. coli. After growing liquid cultures from the one successful attempt and purifying that plasmid, we attempted transformation into L. lactis. Unfortunately, nothing grew.

Learn

In addition to realizing the lack of plasmid stability in the DH5a recA- strain, there was an underlying issue with the overall replication and viability of our plasmid in both E. coli and L. lactis. After contacting NovoPro labs, the supplier of our plasmid, we determined that there was a mutation within the repC region. NovoPro labs conducted three generations of full-length sequencing on the pNZ8148 to find the mutation. This could be one of the major reasons why we’ve been having a lot of trouble cloning inside of both E. coli and L. lactis. They have since sent us a plasmid with the fixed repC region, and we plan to repeat many of our experiments with the fixed plasmid.

Design Round 2

Going back to square one, we started on using re-building our construct with the fixed plasmid. We took this time to address many of our previous issues as well. Firstly, we designed new primers that would add BioBrick assembly prefixes and suffixes in order to make them BioBrick compatible. We also designed primers to address removing the DXMT enzyme (see the note in Fig. 3). Finally, we added a pathway for the expression of EGFP in order to characterize both the strength of our promoter (we would see EGFP expression at varying concentrations of nisin) and to check for any unwanted promoter expression when it is not being induced, as it is possible that our promoter is not fully repressed in E. coli (NICE Expression Handbook).