Discussion surrounding iGEM Guelph's 2023 wet lab results
______________________________________________________________________
pBeloBac11loxP2272 extraction and confirmation
Due to the size of the vitamin B12 gene cluster, ~16kbp, a vector that can function efficiently with a large insert was needed (Godiska et al., 2013). A bacterial artificial chromosome (BAC) was chosen as the vector because it possesses these properties. Based on an Escherichia coli F-plasmid, it was discovered that these plasmids could replicate efficiently down generations and avoid recombination due to low copy number (1 or 2) while possessing upwards of one megabase worth of bacteria DNA (Shizuya et al., 1992). Originally used for the construction of libraries of complex genomes, such as humans, they are now also used for cloning novel genes into bacteria such as E. coli (Shizuya et al., 1992; Godiska et al., 2013). The BAC chosen for our vitamin B12 experiments was the pBeloBac11loxP2272 vector (Coren, 2017)
Figure 1: Plasmid map of pBeloBac11loxP2272 with vitamin B12 gene cluster inserted.
Note: The inserted vitamin B12 gene cluster interrupts the lacZα gene in the BAC.
One drawback of the BAC vector is its low copy number. Because of this, there were some issues with extracting a high enough concentration of plasmid using a miniprep kit, necessitating multiple attempts to obtain a good quality plasmid extraction. After two attempts to extract sufficient vector concentration using 4.5mL of overnight culture, a third attempt using 10mL was performed. This third attempt was not successfully confirmed through digestion with restriction enzyme BsaI HF-v2 as no bad appeared on the gel. However, as the tube was partly melted in the thermocycler after the incubation for digestion, it was decided that a PCR of the vector would still be done, which was successful.
As the experiments progressed and the extracted vector was used, it was anticipated that more vector would be required, so more of the E. coli containing the vector and plasmid was grown overnight, and a miniprep was performed. This miniprep was done using a mix of reagents from different kits (the original kit ran out of some reagents), and this resulted in a usable product, but one with some RNA contamination as evidenced by the A260/A280 ratio being above 2.00. Ultimately, none of these vector extracts were used as the Gibson Assembly reagent was depleted before more of the vector was required.
______________________________________________________________________
Generation of Gibson Assembly Fragments with PCR
The vitamin B12 gene cluster, genes CbiA to CobT in the Salmonella Typhimurium LT2 genome, is large for a PCR reaction, so the gene cluster was amplified in two parts. The length of the gene fragments and vector also necessitated the use of a high-fidelity polymerase, and Q5 polymerase from New England Bioscience (NEB) was chosen for this reason. PCR amplification of the vector was also done for two reasons. As previously stated, the vector is a low copy plasmid, and as the vector plus inserts were to be digested and ligated using Gibson Assembly, a high concentration was needed to lower the volume of fragments in the assembly reaction. Second, the Gibson Assembly requires overlapping regions to anneal for ligation, and the primers were designed to create the needed overlap with joining fragments.
The reverse primer for the vector and the forward primer for the first gene fragment (CbiA to CbiH) also introduced a promoter region that was shown to express at high levels in the murine gut environment (Armetta et al., 2021). This caused the primers to be long, 59bp for the CbiA-CbiH forward primer and 58bp for the vector reverse primer. The corresponding annealing temperatures were likewise high; therefore, a 2-step PCR reaction was done with the annealing and extension steps at the same time at 72℃. All other primers were shorter in length, 27-35bp, but due to both the properties of the Q5 polymerase and needing to anneal at similar temperatures, they were also designed with higher annealing temperatures in mind, primarily by increasing GC content.
The first attempt at generating all three amplicons failed, likely due to the PCR tubes partly melting in the thermocycler as it reached temperatures of 98℃ during the denaturation steps. After a discussion with the lab technician in the Emma Allen-Vercoe lab at the University of Guelph, it was thought that the low number of tubes in the thermocycler was likely causing the melting tubes. This proved to be the case as subsequent attempts at creating the desired amplicons were successful after adding 8 PCR tubes with water to the thermocycler when running the PCR protocols. This strategy likely succeeded because it allowed the heat to be distributed among more tubes.
Figure 2: Thermocycler setup to decrease the chance of melting tubes.
Note: Thermocyclers can cause tubes to melt with only one of two tubes, but adding more tubes decreases tube melting. Created with BioRender.com
Two attempts were made at amplifying the entire vitamin B12 gene cluster to reduce the fragment number in the Gibson Assembly reaction. Both of these attempts failed and resulted in a large number of fragments of varying sizes. This was likely because the primers used were designed to amplify the fragment in two sections. The forward CbiA primer and the reverse CobT primer were not designed to function together in a reaction and may have bound to each other, creating increasingly long amplicons with each PCR cycle. This would offer an explanation for why the gel to confirm the amplicons had bands from 100bp to 12,000bp.
Figure 3: Vitamin B12 gene fragments to be amplified with PCR.
Note: Created with BioRender.com
______________________________________________________________________
Assembling PCR Fragments with Gibson Assembly
The Gibson Assembly of the two vitamin B12 gene fragment amplicons and vector amplicon was unsuccessful after seven attempts under varying conditions. The primers for vector amplification were designed to disrupt the lacZα gene and, along with the competent cells chosen, would allow for white/blue colony screening based on the metabolism of X-gal, which was added to the surface of the media.
The first attempt followed the default protocol for three fragments provided with the Gibson Assembly kit from NEB. This involved incubating the fragments for 15 minutes. However, the standard protocol did not result in the formation of colonies on the agar media after attempting to transform the assembly product into NEB α-competent cells. The transformation protocol and competent cells were also tested using the unaltered vector, and many blue colonies formed on all control plates. As all three fragments are quite large (7.55, 7.86, 8.73kbp), it was determined that a longer incubation time would likely improve the chance of the reaction success (Li et al., 2018). The second and third attempts incubated the assembly mixture for 60 and 75 minutes, respectively, and both produced white colonies after incubation overnight.
Initially, to test for successful insertion of gene fragments, the white colonies were grown overnight in broth media, and a miniprep kit was used to extract the plasmid before digestion with restriction enzymes and gel electrophoresis to check for expected banding. This method failed to confirm the second and third assembly reactions. As an alternative, PCR confirmation was attempted with the miniprep product and colony PCR using the same primers that originally generated the fragments. These PCR confirmations were not successful, as the length of the PCR product was thought to be a potential issue. New PCR primers (three sets) were ordered and designed based on the in silico vector model with inserts. Each pair of primers flanked a different region of the plasmid where the fragments joined together, creating amplicons 350-400bp in length. Unfortunately, after multiple attempts, these primers also failed to confirm the second or third attempts at Gibson Assembly.
A member of the iGEM Guelph team mentioned that the lab they worked in used much longer incubation times for their Gibson Assemblies, so our next attempt was done with an incubation time of 3.5 hours. After transformation, the assembly appeared successful, with three white colonies forming on the full concentration plate. We then carried out our confirmation: digestion and gel, PCR with primers from fragment generation, and finally, PCR with primers flanking regions where fragments would join. As with attempts two and three, neither of these confirmation attempts were successful.
In a meeting with our team's principal investigators, it was brought up that the PCR reagents may be interfering with the Gibson Assembly, so the gene fragments were cleaned up using a PCR cleanup kit. After cleanup, the concentration of the fragments was reduced but still high enough to move forward. Gibson assembly attempts five, six and seven all used the cleaned-up fragments. For attempt five, the incubation temperature was reduced back to one hour based on the success of the Li et al. (2018) protocol, which saw success with similarly large assembly fragments. Again, some white colonies grew on the media after incubation overnight, but confirmation using the previously stated techniques returned negative results.
The next two attempts, six and seven, were done in a sequential fashion with two gene fragments added to the reaction mixture and incubated before the third fragment was added later on. This was based on a conversation with a graduate student in the Molecular and Cellular Biology department who has had success with this technique in the past. The first attempt at sequential assembly (attempt six overall) was done with 30-minute incubation times for each stage of incubation but produced no colonies on the agar media. The second sequential assembly attempt (seventh overall) was done with 40-minute incubation times and initially appeared successful, with a few colonies forming on multiple plates after transformation and overnight growth. However, following 24 hours of growth, all colonies turned blue. This final result of blue colonies forming on all the plates indicates that the PCR cleanup likely did not work as intended. The colonies could not turn blue without the unbroken lacZα gene and could not grow on the antibiotic-infused media without the vector's antibiotic resistance gene, suggesting that the unaltered vector was transformed and not removed during PCR cleanup.
As none of the assembly attempts could be confirmed using the three techniques mentioned previously, it was decided that the most successful attempt based on the number of white colonies would be sent to Plasmidsaurus for sequencing. The chosen experiment was attempt 5, unfortunately the sequencing did not work, and the company made two attempts with no results. A higher copy plasmid linearized with restriction enzymes may have been a better approach as this would not have required as many PCR fragments to be assembled for a working product to be constructed.
Figure 4: Visual Representation of Gibson Assembly of Gene Fragments and Vector Into a Circular Plasmid
Note: Three DNA fragments generated via PCR with corresponding overlapping regions (color-coded) for Gibson Assembly. After gene fragments are mixed in a tube and placed in an incubator, enzymes and reagents in Gibson Assembly Mix ligate fragments together based on overlapping regions within each fragment. Created with BioRender.com
______________________________________________________________________
Use of Golden Gate Assembly to Insert Human Gastric Intrinsic Factor into pGGAselect
Golden Gate assembly was used for inserting the human gastric intrinsic factor gene (hGIF) into a pGGAselect plasmid. The hGIF gene fragment was synthesized by Twist Bioscience after having the codons optimized for expression in E. coli using IDTs codon optimization tool. On top of the optimized gene, a signal peptide and translation initiation region were needed to ensure the protein was efficiently expressed and exported to the periplasm after peptide synthesis. Periplasmic expression was important to increase the chances that the protein was exported out of the cell and to increase the frequency of proper protein folding. The cytoplasm of E. coli is a reducing environment that, unlike the periplasm, does not provide a suitable environment for disulphide bond formation, which hGIF contains (Gąciarz et al., 2017). The PelB signal peptide chosen was demonstrated to efficiently result in the export to the periplasm of a recombinant human protein, growth hormone, which also contains disulphide bonds (Mirzadeh et al., 2020).
As with Gibson Assembly, the first attempt at inserting the hGIF into pGGAselect was based on the default protocol provided by NEB with the Golden Gate BsaI HF-v2 kit and involved a very short 5-minute incubation time. Following transformation, multiple colonies grew on overnight agar plates. Colonies were selected from the plates and grown in broth overnight, and then a miniprep kit was used to extract the plasmids for digestion with EcoRI and confirmation via gel electrophoresis. The miniprep results were poor, with mostly low concentrations and all having bad A260/A280 ratios (1.27-1.44), which was potentially a result of using reagents from multiple different kits and manufacturers. After imaging the gel, it was clear that the unaltered plasmid had been transformed into all tested colonies. This result was likely because the extremely short incubation time was lower than the digestion protocol for the enzyme in the assembly mixture (15 minutes versus the 5 minutes used).
A second attempt followed with a 15-minute incubation time, and this once again resulted in multiple colonies forming on the plate after overnight growth. The results appeared promising after growth in broth overnight, miniprep, and digestion. A sharp band at approximately 2.1kbp and a very smudged lower band at 1.1-1.4kbp were observed, though an expected band at 33bp was missing. This result was promising enough that the miniprep with the sharpest lower band was sent to Plasmidsaurus for sequencing and further confirmation. Unfortunately, when the sequence was returned, it was discovered that the unaltered plasmid was transformed and extracted.
In order to speed up the process of attempting different conditions, a high-throughput method was conceived where five reaction tubes were assembled and incubated simultaneously but the reactions were taken out of the incubator at different times (20, 30, 40, 50, and 60 minutes). Based upon colony forming units (cfu), these incubation times were the most successful, with all plates having 100+ colonies after overnight growth. Because of our previous attempts, this was an unexpected cfu outcome. It was thought that one reason for the large increase in colonies was the degradation of the chloramphenicol in the media, so multiple colonies from each plate were restreaked onto newly made agar containing chloramphenicol. Growth was seen on all plates after the transfer, ruling out chloramphenicol degradation as a reason for the large increase in colonies. Two colonies from each incubation time were selected for overnight growth and miniprep. After digestion with EcoRI and gel electrophoresis, the results were inconclusive. The gel image showed bands at approximately 3.4kbp and 2.1kbp for all ten minipreps. These sizes were unexpected as the 3.4kbp band corresponds to the size of the vector construct, but the construct has three EcoRI recognition sites, so it should not show up as a single band on the gel. Also, the 2.1kbp band corresponds to the size of the unaltered plasmid, which also has three EcoRI recognition sites.
Based on the corresponding sizes of these bands, it is possible that the E. coli got transformed with both the unaltered vector and the vector with the gene inserted (Tomoiaga et al., 2022). This result would also require that the digestion was incomplete for only one band to appear for each plasmid. This could result from insufficient incubation time, leftover Golden Gate reagents contaminating the reaction, or incorrect substrate volume (ThermoFisher, n.d.). Substrate volumes were determined with a Nanodrop machine. However, as it cannot differentiate between circular plasmid and DNA fragments, it is likely to have overestimated how much circular plasmid DNA was present in the sample. A cleanup step before plasmid quantification and digestion with a miniprep kit after Golden Gate assembly may alleviate this issue by removing any leftover proteins and DNA fragments (Pryor et al., 2020). Though the NEB Golden Gate Assembly Tool showed 100% ligation fidelity in silico for this reaction, it is also possible that the gene fragments ligated together in the reaction before ligating with the vector. Based on the observed banding, this would require three fragments to be joined and for the digestion to be incomplete, as there would still be three EcoRI recognition sites. Care must be taken when designing restriction sites to minimize the chances of DNA fragments ligating together when not desired.
As the fall semester was beginning, it was not possible to follow up on these last ambiguous results with the lab space needed for classes. The “Future Directions” section has information about plans to tackle these ambiguous results and resolve the issues faced in inserting the hGIF gene into pGGAselect and testing the expression and function of the protein.
______________________________________________________________________
References
Armetta, J., Schantz-Klausen, M., Shepelin, D., Vazquez-Uribe, R., Bahl, M. I., Laursen, M. F., Licht, T. R., & Alexander, O. (2021). Escherichia coli Promoters with Consistent Expression throughout the Murine Gut. ACS Synthetic Biology, 10(12), 3359–3368. https://doi.org/10.1021/acssynbio.1c00325
Coren, J. S. (2017). Retrofitting the BAC cloning vector pBeloBAC11 by the insertion of a mutant loxP site. BMC Research Notes, 10(1). https://doi.org/10.1186/s13104-017-2631-8
Gąciarz, A., Khatri, N. K., Velez-Suberbie, M. L., Saaranen, M. J., Uchida, Y., Keshavarz-Moore, E., & Ruddock, L. W. (2017). Efficient soluble expression of disulfide bonded proteins in the cytoplasm of Escherichia coli in fed-batch fermentations on chemically defined minimal media. Microbial Cell Factories, 16(1). https://doi.org/10.1186/s12934-017-0721-x
Godiska, R., Wu, C.-C. ., & Mead, D. A. (2013). Genomic Libraries. Brenner’s Encyclopedia of Genetics, 306–309. https://doi.org/10.1016/b978-0-12-374984-0.00641-0
Li, L., Jiang, W., & Lu, Y. (2018). A Modified Gibson Assembly Method for Cloning Large DNA Fragments with High GC Contents. Methods in Molecular Biology (Clifton, N.J.), 1671, 203–209. https://doi.org/10.1007/978-1-4939-7295-1_13
Shizuya, H., Birren, B., Kim, U. J., Mancino, V., Slepak, T., Tachiiri, Y., & Simon, M. (1992). Cloning and stable maintenance of 300-kilobase-pair fragments of human DNA in Escherichia coli using an F-factor-based vector. Proceedings of the National Academy of Sciences, 89(18), 8794–8797. https://doi.org/10.1073/pnas.89.18.8794