Our project is initiated by the screening and selection of appropriate variants of genes to build the pyruvate carboxylase (PYC), oxaloacetate acetylhydrolase (OAH), acetate-CoA ligase (ACS), and pyruvate synthase (PFOR) (POAP) cycle. In the report of the POAP cycle, researchers used variants of genes that yielded impressive carbon fixation rates at 50 degrees Celsius [1]. This operating condition is too hostile for the minimal cell (JCVI-Syn3B) and is hostile for many potential biological hosts. To overcome this obstacle, we aimed to find different variants of enzymes that constitute the POAP cycle, compare their characterized activities with regards to temperature and pH, and build a POAP cycle that is more suitable for biological conditions (< 45 degrees Celsius). Amongst the 4 enzymes that complete the POAP cycle, the biggest challenge is from PFOR which requires high operating temperatures to overcome the thermodynamic barriers of pyruvate synthesis. Using Braunschweig Enzyme Database (BRENDA), we searched through characterized enzymes from all facets of life and compared their reported activities in association with temperature and pH, yielding the following selection. It is worth noting that while some enzymes had multiple variants selected, our project eventually focused on the testing of one particular variant (bolded), and the other selections became backups in the case of failure.
| Gene | Organism | Organism Type | Working Temp (°C) | Gene Length (nt) |
|---|---|---|---|---|
| PFOR | Desulfocurvibacter africanus | Bacteria | 40 | 3699 |
| PYC | Bacillus subtilis | Bacteria | 37 | 3443 |
| OAH | Aspergillus niger | Fungus | ≤45 | 1026 |
| OAH | Moniliophthora perniciosa | Fungus | 30 | 993 |
| OAH | Dianthus caryophyllus | Plant | 20 | 1267 |
| ACS | Dunaliella tertiolecta | Algae | 37-45 | 2464 |
| ACS | Pseudomonas putida | Bacteria | 38 | 1962 |
| ACS | Bacillus subtilis | Bacteria | 40 | 6300 |
Figure 1.1: Table of selected genes to build a POAP cycle that is more compatible with the growth conditions of the minimal cell (JCVI-Syn3B).
After selection of desired variants of enzymes, we designed fragments of DNA and attached it to the selected genes. The first fragment is a complementary region to the pCC1BAC EcoRV fixed HindIII Exp Vector backbone, a copy control vector that we are using to integrate our genes of interest into the minimal cell. This complementary fragment allows our gene of interest to anneal onto the backbone and be assembled into the plasmid via Gibson assembly and E. Coli plasmid assembly. The second fragment is a mycoplasma promoter of intermediate strength. This allows our genes to be expressed once integrated into the minimal cell. Additionally, we also designed primers to PCR, and crossover PCR, our genes of interest which allows us to generate large quantities of desired genes and confirm its fidelity post plasmid assembly. The crossover PCR procedure is necessary because certain designed gene fragments such as ACS and PFOR were too large so we can only order them to be synthesized in 2 and 3 pieces respectively. After benchling designs, these parts were ordered for synthesis by Twist Biosciences. These components have also been registered as basic and composite parts. Please see the list of these parts in our Parts section of the wiki.
Amongst the 4 enzymes that build the POAP cycle, PYC from Bacillus subtilis was already integrated into the pCC1BAC EcoRV fixed HindIII Exp Vector and transformed into the minimal cell from a previous J. Craig Venter Institute (JCVI) project. This transformed minimal cell containing PYC is supplied to us by the Suzuki Lab at JCVI. Thus, we did not need to synthesize and transform this gene of interest into E. Coli and the minimal cell but had access to its designs nonetheless.
Upon arrival of our parts, we first performed PCR on composite parts containing genes of interest to both obtain larger quantities of it and to prepare for E. Coli plasmid assembly. It is important to point out that designed gene fragments such as ACS and PFOR were too large so Twist Bioscience divided ACS and PFOR fragments into 2 and 3 pieces respectively. Before moving onto E. Coli plasmid assembly, we first performed crossover PCR to combine the two fragments of ACS into one and the three fragments of PFOR into one.
Figure 2.1: Agarose gel image showing the PCR products of our composite parts.
Figure 2.1.2: Two Agarose gel showing successful assembly of Pseudomonas putida (Pp) ACS fragments and Desulfocurvibacter africanus (Da) PFOR from their fragments using crossover PCR.
After obtaining large stocks of our composite parts containing genes of interest, we began the integration of our genes into the plasmid vector backbone through E. Coli plasmid assembly. This assembly works similarly to the Gibson Assembly (fun fact, Gibson Assembly is developed at J. Craig Venter Institute (JCVI) by Daniel Gibson - a JCVI Professor!) but occurs directly inside of E. Coli (please see our Notebook and Experiments for the protocol). Although there was a process of trial and error, we eventually obtained 3 plasmids containing ACS, OAH, and PFOR respectively. E. Coli colonies with successful assemblies of the complete plasmid gains chloramphenicol resistance and appears on a chloramphenicol plate. These colonies were then picked, upscaled, and mini-prepped to confirm plasmid fidelity. It is worth noting that we are assembling individual genes into distinct constructs to ensure that each component works and as a proof of concept. If we assemble all 4 genes into one large construct, it will be immensely more difficult to troubleshoot when encountering an issue.
Figure 2.2: Transformed E. Coli colonies growing on a chloramphenicol plate. These E. Coli colonies successfully assembled our gene of interest into the vector backbone and are hence transformed. On the left is E. Coli with ACS in the plasmid. In the middle is E. Coli with OAH in the plasmid. On the right is E. Coli with PFOR in the plasmid.
Figure 2.3: Plasmids obtained from mini-prepping transformed E. Coli culture. On the left gel, the labels 9.1, 9.2, 9.3, 9.4 are 4 separate colonies picked from the plate containing E. Coli cultures with the ACS gene inside of the plasmid. On the right gel are plasmids obtained from E. Coli colonies containing the PFOR gene inside of the plasmid, 2 separate colonies picked from the plate containing Aspergillus niger OAH, 2 separate colonies picked from the plate containing Moniliophthora perniciosa OAH, and the pCC1BAC EcoRV fixed HindIII Exp Vector backbone.
In Figure 2.3, we observe that the plasmids are heavy and are the size we expect them to be. On the gel containing PFOR and OAH variants, we see that the vector backbone has two bands with one heavier than the other. This is likely due to supercoiling of the vector which, based on whether the plasmid is supercoiled or not, separates the two forms of plasmid on an agarose gel. While this is unpleasant to look at, we simply wanted confirmation that our assembled plasmids have sizes within expectation which is indeed confirmed form these results.
After having some confirmation that our plasmids are present and correctly assembled, we sent these plasmids for sequencing by Plasmidsaurus. This is also the point at which we prioritized testing for just one variant of a gene for future steps, and plasmids containing other variants of the same gene were stored as a backup.
Figure 2.4: Sequenced plasmid containing OAH (SD2c1) aligned with benchling plasmid design (pCC1BAC_EcoRV_fixed_HindIII_Exp_Vector_file2_AnOAH). The OAH gene is present in the assembled plasmid from nucleotide 2285 to nucleotide 3325 which aligns perfectly with the benchling design, showing that our plasmid is correctly assembled according to our initial design.
Figure 2.5: Sequenced plasmid containing PFOR (SD1c1) aligned with benchling plasmid design (pCC1BAC_EcoRV_fixed_HindIII_Exp_Vector_file2_PFOR). The PFOR gene is present in the assembled plasmid from nucleotide 2285 to nucleotide 5983 which aligns perfectly with the benchling design, showing that our plasmid is correctly assembled according to our initial design.
Figure 2.6: Sequenced plasmid containing ACS (SD9c1) aligned with benchling plasmid design (pCC1BAC_EcoRV_fixed_HindIII_Exp_Vector_file2_PpACS). The ACS gene is present in the assembled plasmid from nucleotide 2285 to nucleotide 4246 which aligns perfectly with the benchling design, showing that our plasmid is correctly assembled according to our initial design.
After confirming the fidelity of both our gene of interest and our plasmid through sequencing and alignment with our benchling designs, we proceed to transformation of the minimal cell. What is particularly interesting about this step is that once the plasmid enters the minimal cell, it becomes a suicide plasmid that can no longer replicate. However, our plasmid vector contains a cre-loxp site. The same cre-loxp site is also engineered into the genome of the minimal cell which is known as the landing pad. Our gene of interest, along with an antibiotic resistance gene, will be integrated into the genome of the minimal cell before the plasmid is degraded. Thus, successful transformations will yield minimal cell colonies on an antibiotic SP4 plate in which the plasmid is no longer present, but the gene of interest (and an antibiotic resistance gene) it carries is inserted into the genome of the minimal cell and can be expressed.
Figure 2.7: SP4 antibiotic agar plates that selects for minimal cells that have integrated a gene of interest and antibiotic resistance gene into their genome from the suicide plasmid. On the left, a positive and negative control plate is used to ensure that our protocol works correctly. We expect no colonies to form in the No-DNA side because no antibiotic resistance gene is supplied. We expect colonies to form in the pSG023 side as that is a control vector that has been shown to deliver an antibiotic resistance gene into the minimal cell genome. In the middle, the 9.2 side are minimal cell colonies that now have the ACS gene. The 4.3 side are minimal cell colonies that now have the OAH gene. On the right, the 2.1 and 1.1 side are both minimal cell colonies that now have the PFOR gene. The difference in 2.1 and 1.1 is just that the plasmids were mini-prepped from two separate E. Coli colonies that both have assembled the PFOR plasmid.
As mentioned above, because our genes of interest are integrated into the minimal cell genome and the plasmids are soon eliminated/degraded, we lose the ability to mini-prep our minimal cells and send the results off for sequencing. Instead, we use PCR with primers we have designed to both see if a product is produced and if that product is of the expected size. Successful PCR products of the expected size are sufficient for us to be confident in our integration of the genes of interest into the minimal cell and proceed to activity screening.
Figure 2.8: PCR products obtained from performing PCR on transformed minimal cells. 2.1 are PCR products obtained from minimal cells containing the PFOR gene. 4.3 are PCR products obtained from minimal cells containing the OAH gene. 9.2 are PCR products obtained from minimal cells containing the ACS gene. VC is vector control which affirms that PCR of our transformed colonies yield a product that is heavier, showing that our gene of interest is present.
After confirming that our minimal cells now have genes of interest contained in their genome, we move on to screening of enzyme activity. Our enzymes do not contain tags which allow for protein purification and SDS-PAGE analysis. Instead, we wanted to rely on gas chromatography-mass spectrometry (GC-MS) to screen for metabolic perturbations. All of these enzymes perturb the central metabolism of the minimal cell, and because the minimal cell has the simplest metabolism that is very well characterized, we plan on using the GC-MS to compare the metabolic profiles of a normal minimal cell and metabolic profiles of transformed minimal cells. Knowing exactly what each enzyme does, we also have a general expectation of what kind of perturbation will occur and to which metabolites it will affect. To better find our target metabolites in the profile, we also created standards of pyruvate, acetate, acetyl-CoA, and oxalate, though not all of them worked as expected (discussed later).
Figure 3.1: Pyruvate standard (derivatized and dissolved in solution of pyridine, methoxylamine, and MSTFA) profile by GC-MS.
Figure 3.2: Derivatized pyruvate elution time (top image), mass-spectrometry (top image), and predicted structure (bottom image).
Figure 3.3: Oxalate standard (derivatized and dissolved in solution of pyridine, methoxylamine, and MSTFA) profile by GC-MS.
Figure 3.4: Derivatized oxalate elution time (top image), mass-spectrometry (middle image), and predicted structure (bottom image).
The development of standards makes it much easier to screen for specific metabolites in the minimal cell profile as we will know exactly when our target compound elutes and what the derivatized structure will look like. The key metabolites that will be directly perturbed by the enzymes involved in the POAP cycle include acetyl-CoA, pyruvate, oxaloacetate, oxalate, and acetate. We discovered that derivatization using pyridine and methoxylamine does not work well with acetyl-CoA and acetate since the GC profile of these standards show no product (all visible peaks are from pyridine, methoxylamine, and MSTFA). We are working actively to find another derivatizing agent that can work well with these metabolites. Concurrently, we did not have access to oxaloacetate which is a rather uncommon and unstable (typically stored at -20 Celsius) compound. We have placed an order for oxaloacetate and will make a standard for it after arrival.
For the time being, our GC-MS screens of transformed minimal cells are limited to PFOR and PYC genes. Both OAH and ACS lack critical standards for screening - oxaloacetate, acetate, and acetyl-CoA. Still, we collected GC-MS data for all of our transformed minimal cells (including OAH and ACS) but manually searching for derivatized oxaloacetate, acetate, and acetyl-CoA proved difficult.Therefore, we mainly conducted GC-MS screens and analysis on PFOR and PYC genes, and will revisit the collected profiles of OAH and ACS after making appropriate standards (we have attached their profiles in Supplemental Materials). We adapted an untargeted metabolite derivatization and screening protocol from plant natural products to use on the minimal cell, and obtained the following profiles for normal minimal cells and transformed minimal cells.
Figure 3.5: Metabolic profile of untransformed minimal cell by GC-MS.
Figure 3.6: Derivatized pyruvate and lactic acid profile of untransformed minimal cell by GC-MS.
metabolite extraction and derivatization protocol we used provided a comprehensive coverage of biological compounds. Similarly, the GC-MS method we developed, with the help of our PI Dr. Immo Burkhardt, also reinforces this comprehensive coverage of biological compounds. In Figure 3.5, we were able to find metabolites such as pyruvate and lactic acid in the earlier elution phase. Further down the profile, we find a collection of amino acids. Near the middle and end of the profile, we find lipids and large organic molecules. Finally, near the end, we find cholesterol. This coverage of polar and primary metabolites to amino acids then to non-polar lipids and cholesterol gives us confidence that our derivatization and GC-MS is holistically capturing the metabolic landscape of our cell. Further, our profile contains little to no noise and contaminants, allowing us to cleanly distinguish between metabolites and even capture compounds with small concentrations. For a more detailed illustration of the metabolites captured in Figure 3.5, including which peak corresponds to what compound, their mass spectrometry, and the predicted structure using mass spectrometry, please visit our supplementary materials.
Figure 3.7: Metabolic profile of PFOR transformed minimal cell by GC-MS.
Figure 3.8: Derivatized pyruvate and lactic acid profile of PFOR transformed minimal cell by GC-MS.
Figure 3.9: Metabolic profile of PYC transformed minimal cell by GC-MS.
Figure 3.10: Derivatized pyruvate and lactic acid profile of PYC transformed minimal cell by GC-MS.
For PYC and PFOR, we are focusing on observing the perturbation to pyruvate and lactic acid because these are the metabolites that the two enzymes will perturb directly and indirectly, respectively. Lactic acid relates to pyruvate in the minimal cell because the minimal cell converts pyruvate into lactic acid which is a dead-end metabolite. Thus, we expect the levels of lactic acid to change in response to changes in levels of pyruvate. PYC converts pyruvate into oxaloacetate, so we expect a decrease to pyruvate and lactic acid levels. PFOR should facilitate pyruvate synthesis, so we expect an increase to pyruvate and lactic acid levels. The area under the curve, or integration of an elution peak, tells us the abundance of a metabolite profiled by GC-MS. The integration of all of our profiles have been compiled in tables below.
| Sample | Cell Pellet Mass (mg) | Pyruvate Area | Lactic Acid Area | Pyruvate Area per mg of cell pellet mass | Lactic Acid Area per mg of cell pellet mass |
|---|---|---|---|---|---|
| Base Syn3B | 0.8 | 44270.99 | 800901.88 | 55338.7375 | 1001127.35 |
| Syn3B-PFOR | 0.7 | 31699.12 | 1033966.48 | 45284.45714 | 1477094.971 |
| Syn3B-PYC | 1.1 | 13599.86 | 283049.79 | 12363.50909 | 257317.9909 |
| Sample | Percent change of pyruvate area per cell pellet mass compared to Base Syn3B | Percent change of lactic acid area per cell pellet mass compared to Base Syn3B |
|---|---|---|
| Syn3B-PFOR | -18.2% | +47.5% |
| Syn3B-PYC | -77.7% | -74.3% |
From Table 3.2 we see that PFOR transformed minimal cell (Syn3B) did not increase pyruvate levels, but it did greatly increase levels of lactic acid by nearly 50%. While this seems conflicting, we believe this still makes sense because we can imagine the minimal cell attempting to stabilize an increase to pyruvate levels by directing more pyruvate to be converted into other metabolites, including the one of the main waste products produced from pyruvate - lactic acid. Thus, the impressive increase to lactic acid levels may allude to the pyruvate synthesis reaction catalyzed by PFOR. Because we are not attempting to characterize the activity rates of PFOR and are merely looking for some sign of perturbation that gives us confidence to assemble the pathway, we feel this is sufficient for us to incorporate this PFOR into the POAP pathway assembly. While the PFOR data may not be definitive, we feel that the PYC transformed minimal cells give us high confidence in proceeding to the next step. We expect decreases to both pyruvate and lactic acid, and this indeed what we observe - a quite significant decrease to both metabolic concentrations.
While the data generated from transformed PFOR may not be as convincing as transformed PYC, we are reasoned to proceed with the construction of the big plasmid containing all four enzymes. What would be a true deal breaker would be insignificant perturbation to the metabolic profile compared to the untransformed minimal cell or the inability for transformed minimal cells to grow and proliferate, none of which are observed. Thus, we feel that there is a proof of concept and have proceeded with the assembly of the POAP pathway into one plasmid.
The cloning of an enzyme gene downstream of another enzyme gene turned out to be difficult, possibly due to the end sequences not suitable for homologous recombination in E. coli. We devised a cloning method to combine enzyme genes one at a time using crossover PCR and the self-closure of the generated PCR products within E. coli. So far, we have completed the construction of a 2-gene construct (OAH and ACS), confirmed through agarose gel and soon to be sequenced by Plasmidsaurus. We are conducting the final assembly soon to have a 4-gene construct (PYC, PFOR, OAH, and ACS). Subsequent to this, we will introduce the construct into the JCVI-Syn3B cell, confirm the introduction using PCR, and analyze the metabolites produced using GC-MS. Despite our ongoing work, we have demonstrated a proof of concept that the POAP pathway can be implemented in vivo within the minimal cell.
An ongoing project parallel to the work on carbon fixation is an adaptive laboratory evolution (ALE) project aimed at evolving thermal tolerance in the minimal cell (JCVI-Syn3B). The goal of this ALE experiment is to complement the POAP cycle because of reports that show the POAP cycle, especially the PFOR enzyme, exhibiting higher activity at higher temperatures ( > 45 Celsius). The ancestral/normal minimal cell grows optimally at 39 degrees Celsius and can not grow at this temperature, but we planned to evolve thermal tolerance through continual adaptive laboratory evolution. After a little under 4 months of continual evolution at bottleneck temperatures, we have obtained strains of the minimal cell that grows overnight at 44.2 degrees Celsius. On one hand, we successfully obtained strains of minimal cells more compatible with the POAP cycle. On the other hand, we were intrigued by the genetic changes that gave the evolved minimal cells the ability to grow overnight at a temperature that would kill the ancestral strain. Thus, we have collected cell pellets of our thermal tolerant minimal cell and sent them for gDNA extraction and sequencing by Plasmidsaurus. We report our sequencing results and preliminary mutation callouts below.
Our success with ALE of the minimal cell has yielded us a more compatible host for the POAP cycle which we will soon transform into, once the POAP pathway has been assembled. At the same time, this project opens up new research directions in the discovery of genes and mutations that are responsible for thermal tolerance. Further, the evolution of a synthetic and minimal cell is always exciting as we can learn and compare the evolution of primitive life forms. In summary, the genetic analysis of this ALE experiment is ongoing and will be continually spearheaded by the JCVI-UCSD iGEM team concurrent to carbon fixation. We, and our PIs, envision this to eventually fruition into a publication.
In our literature search, there was little literature that consolidates information on carbon fixation and synthetic biology to aid researchers in the development of in vivo carbon fixation and cellular carbon capture. To help future researchers tackle the climate crisis using synthetic biology approaches, we drafted a mini-review article that will be submitted for review in Frontiers: Bioengineering and Biotechnology. We have attached a snippet of our manuscript.
