Design

First, we decided on which oscillator would be most robust for generating the oscillations we needed to design our bacterial system. The first oscillator we considered was designed by Hasty et. al in 2010 [1]. After feedback from our mentors, we decided against using this oscillator for two reasons:

1. The oscillator promotes synchrony between individual cells— which involves diffusion of transcription factors across cells. This may have been too many moving parts and variables, which could have contributed to failure of the oscillations.


2. The 2008 oscillator [2] is easier to tune— with different levels of arabinose and IPTG. We thought this would let us control the oscillations and gene expression with a higher degree of control than with the 2017 oscillator.

Initially, based on literature review [3], we assumed that since the only coding regions of nitrogenase were nifH, nifD, and nifK, that these were the only genes necessary for full expression of the enzyme. Thus, our original plan was to simply put those three genes as a fused complex under the control of one hybrid promoter. However, based on further literature search, we found that there were multiple genes that comprise the full expression of nitrogenase [4], and that the quantity and type of genes that each diazotroph possessed differed. From here, we decided on the best functioning and minimal cluster our engineered bacteria could possess to have a fully functional enzyme, drawing inspiration mainly from past iGEM teams and papers who considered doing the same.

After more literature searching, we settled on the 9 gene minimal MoFe nitrogenase cluster from the 2018 iGEM team from Nanjing-China. We found that their work had several advantages: (a) it was previously successfully assembled, (b) it was characterized to have some function and (c) the sequence was relatively shorter than most minimal clusters.

From then: we developed a plan to incorporate the cluster to have it expressed under the dual plasmid system. We developed three separate plans to do so:

1. Placing nifH, nifD, and nifK under the control of the hybrid promoter and the rest of the operon in a separate plasmid.


2. Putting the entire operon on a separate plasmid with the hybrid promoter cloned alongside it


3. Integrating our operon and hybrid promoter into the genome

We strongly pursued the third plan, which would prevent the population of E. coli with successfully cloned plasmids from being lost due to fitness costs.

Build

We started by coming up with a plan to clone the operon. Since IDT and TWIST both had a limitation on how many base pairs could be chemically synthesized at once, we had to think of ways to clone our genes altogether.

We considered this problem from the three plans that we planned to pursue. Due to the number of and large base pair sizes of the genes and complete operon, we had to eliminate strategies such as using restriction enzymes to stitch together the genes. We strongly considered other strategies that have been popular amongst other iGEM teams, such as Gibson Assembly and HiFi Assembly. Ultimately, based on feedback from our mentors, in particular AJ Sillato, we considered other ways to clone our operon. AJ gave us the ideas of pursuing cloning with in-vivo assembly and sequence and ligation independent cloning [5][6], which were both cost-effective, less time-consuming, and less hands-on work, which could improve our results and introduce less confounding variables to the experiments. Then, primers to pursue these methods were designed. For future teams, we characterized these primers as well as described how they were designed.

Another issue with our plans involving plasmids (plans 1 and 2), was the conflict between origins of replications. It is known that there are three major categories of origins of replication, and excluding some exceptions, origins of replications from the same category cannot be co-transfected [7]. Due to our activator plasmid and repressor plasmids, (pJS167 and pJS169), 2 out of 3 origins of replications were already accounted for. Thus, a third origin of replication was needed to clone the genes into. This severely limited our choices for plasmid characteristics we could choose from to control important factors such as copy number.

Test

After designing the necessary primers for both in-vivo assembly (IVA) and sequence and ligation independent cloning (SLIC), we tested our designs by running PCR and gel electrophoresis with our obtained constructions. We found that our primers were very successful in amplifying the genes with overhangs, although with slightly less specificity than we would have preferred with some genes. For those, with feedback from our mentors, we decided to run gel extractions or readjust the PCR conditions we had used, especially in regard to annealing temperature and time.

We then set out to clone our operon in its entirety. Although we had all the components necessary to clone the operon in one experiment cycle, we decided against that plan based on feedback we received and literature-based evidence. Both cloning methods, IVA and SLIC, have significantly reduced efficiency when more fragments are attempted to be assembled all at once. Out of the two methods, SLIC had the higher efficiency when 10 or more fragments are assembled [8].

Thus, we attempted to clone our genes in chunks of three. Nine nif genes were divided into three separate segments; nifB-nifH-nifD, nifK-nifE-nifN, and nifX-hesA-nifV. We were able to successfully clone these three segments, and attempted to construct the entire operon by ligating these three segments together.

Learn

We learned and improved many parts of our system along the way. Most significantly, based on the results we obtained from cloning our assembly together, we realized a few major points about the techniques we used.

1. IVA -
1. Limitations in Fragment Assembly. Although assembly of the three segments (nifB-nifH-nifD, nifK-nifE-nifN, and nifX-hesA-nifV), individually did not require much troubleshooting— i.e. most attempts worked either by the first round of assembly or second, the assembly of the total operon required much troubleshooting to see a possible product. We think that this is possibly due to two issues:


1. One, the larger products is not as efficiently taken up by E. coli as the smaller constructs. We tried electroporating the cells for delivery of the segments, and found that this generally worked better than chemical transformation.


2. Two, larger products are harder to assemble by E. coli. Using a strain designed for large fragment assembly, such as DH10B, increased efficiency. CFUs were extremely low for transformed strains with DH5-alpha, where majority products were either annealed vector backbones or if fragments were assembled, only one or two of the fragments were ligated together into the vector.


2. Possible Construction of Plasmid Library. We did not notice this issue for smaller scale assemblies, but we noticed an interesting phenomenon for the large operon assembly. Instead of one consistent plasmid product like we usually notice/expect, an unequal mixture of correct product and incorrect assemblies were found within one isolated colony. Our PCR screening step first brought this result to our attention, and our sequencing results cemented them (see our results page for further details). It is a drawback that a plasmid library could be constructed within one cell instead of one homogenous product— and perhaps should be combated by deploying in-vitro assembly methods instead.


2. SLIC -
1. Need for troubleshooting “chew-back” digestion times. While generally 30 minutes of incubation with T4 exonuclease is recommended for a 20 basepair homology overhang, we found that 30 minutes was not sufficient for our gene products. Although multiple attempts were done to ligate our genes into a vector, we found that most were either empty vectors that were results of the 2BcT vector re-ligating with itself. (See our results page for how our attempts with SLIC progressed.) A time-course should be run with different periods of incubation are compared for efficiency.

References

1. Danino, T., Mondragón-Palomino, O., Tsimring, L., & Hasty, J. (2010). A synchronized quorum of genetic clocks. Nature, 463(7279), 326–330. https://doi.org/10.1038/nature08753


2. Stricker, J., Cookson, S., Bennett, M. R., Mather, W. H., Tsimring, L. S., & Hasty, J. (2008). A fast, robust and tunable synthetic gene oscillator. Nature, 456(7221), 516–519. https://doi.org/10.1038/nature07389


3. Scolnik, P. A., & Haselkorn, R. (1984). Activation of extra copies of genes coding for nitrogenase in Rhodopseudomonas capsulata. Nature, 307(5948), Article 5948. https://doi.org/10.1038/307289a0


4. Scolnik, P. A., & Haselkorn, R. (1984). Activation of extra copies of genes coding for nitrogenase in Rhodopseudomonas capsulata. Nature, 307(5948), Article 5948. https://doi.org/10.1038/307289a0


5. García-Nafría, J., Watson, J. F., & Greger, I. H. (2016). IVA cloning: A single-tube universal cloning system exploiting bacterial in vivo assembly. Scientific Reports, 6(1). https://doi.org/10.1038/srep27459


6. Li, M. Z., & Elledge, S. J. (2012). SLIC: A method for sequence- and ligation-independent cloning. Methods in Molecular Biology (Clifton, N.J.), 852, 51–59. https://doi.org/10.1007/978-1-61779-564-0_5


7. Plasmids 101: Origin of Replication. (n.d.). Retrieved October 11, 2023, from https://blog.addgene.org/plasmid-101-origin-of-replication


8. Li, M. Z., & Elledge, S. J. (2007). Harnessing homologous recombination in vitro to generate recombinant DNA via SLIC. Nature Methods, 4(3), 251–256. https://doi.org/10.1038/nmeth1010